1 ACPI ACPI support is enabled. 2 AGP AGP (Accelerated Graphics Port) is enabled. 3 ALSA ALSA sound support is enabled. 4 APIC APIC support is enabled. 5 APM Advanced Power Management support is enabled. 6 APPARMOR AppArmor support is enabled. 7 ARM ARM architecture is enabled. 8 ARM64 ARM64 architecture is enabled. 9 CLK Common clock infrastructure is enabled. 10 CMA Contiguous Memory Area support is enabled. 11 CPU_FREQ CPU frequency scaling is enabled. 12 CPU_IDLE CPU idle time management is enabled. 13 DRM Direct Rendering Management support is enabled. 14 DYNAMIC_DEBUG Build in debug messages and enable them at runtime 15 EARLY Parameter processed too early to be embedded in initrd. 16 EDD BIOS Enhanced Disk Drive Services (EDD) is enabled 17 EFI EFI Partitioning (GPT) is enabled 18 EVM Extended Verification Module 19 FB The frame buffer device is enabled. 20 FTRACE Function tracing enabled. 21 GCOV GCOV profiling is enabled. 22 HIBERNATION HIBERNATION is enabled. 23 HW Appropriate hardware is enabled. 24 HYPER_V HYPERV support is enabled. 25 IMA Integrity measurement architecture is enabled. 26 IP_PNP IP DHCP, BOOTP, or RARP is enabled. 27 IPV6 IPv6 support is enabled. 28 ISAPNP ISA PnP code is enabled. 29 ISOL CPU Isolation is enabled. 30 JOY Appropriate joystick support is enabled. 31 KGDB Kernel debugger support is enabled. 32 KVM Kernel Virtual Machine support is enabled. 33 LIBATA Libata driver is enabled 34 LOONGARCH LoongArch architecture is enabled. 35 LOOP Loopback device support is enabled. 36 LP Printer support is enabled. 37 M68k M68k architecture is enabled. 38 These options have more detailed description inside of 39 Documentation/arch/m68k/kernel-options.rst. 40 MIPS MIPS architecture is enabled. 41 MOUSE Appropriate mouse support is enabled. 42 MSI Message Signaled Interrupts (PCI). 43 MTD MTD (Memory Technology Device) support is enabled. 44 NET Appropriate network support is enabled. 45 NFS Appropriate NFS support is enabled. 46 NUMA NUMA support is enabled. 47 OF Devicetree is enabled. 48 PARISC The PA-RISC architecture is enabled. 49 PCI PCI bus support is enabled. 50 PCIE PCI Express support is enabled. 51 PCMCIA The PCMCIA subsystem is enabled. 52 PM Power Management support is enabled. 53 PNP Plug & Play support is enabled. 54 PPC PowerPC architecture is enabled. 55 PPT Parallel port support is enabled. 56 PS2 Appropriate PS/2 support is enabled. 57 PV_OPS A paravirtualized kernel is enabled. 58 RAM RAM disk support is enabled. 59 RDT Intel Resource Director Technology. 60 RISCV RISCV architecture is enabled. 61 S390 S390 architecture is enabled. 62 SCSI Appropriate SCSI support is enabled. 63 A lot of drivers have their options described inside 64 the Documentation/scsi/ sub-directory. 65 SDW SoundWire support is enabled. 66 SECURITY Different security models are enabled. 67 SELINUX SELinux support is enabled. 68 SERIAL Serial support is enabled. 69 SH SuperH architecture is enabled. 70 SMP The kernel is an SMP kernel. 71 SPARC Sparc architecture is enabled. 72 SUSPEND System suspend states are enabled. 73 SWSUSP Software suspend (hibernation) is enabled. 74 TPM TPM drivers are enabled. 75 UMS USB Mass Storage support is enabled. 76 USB USB support is enabled. 77 NVME NVMe support is enabled 78 USBHID USB Human Interface Device support is enabled. 79 V4L Video For Linux support is enabled. 80 VGA The VGA console has been enabled. 81 VMMIO Driver for memory mapped virtio devices is enabled. 82 VT Virtual terminal support is enabled. 83 WDT Watchdog support is enabled. 84 X86-32 X86-32, aka i386 architecture is enabled. 85 X86-64 X86-64 architecture is enabled. 86 X86 Either 32-bit or 64-bit x86 (same as X86-32+X86-64) 87 X86_UV SGI UV support is enabled. 88 XEN Xen support is enabled 89 XTENSA xtensa architecture is enabled. 90 91In addition, the following text indicates that the option 92 93 BOOT Is a boot loader parameter. 94 BUGS= Relates to possible processor bugs on the said processor. 95 KNL Is a kernel start-up parameter. 96 97 98Kernel parameters 99 100 accept_memory= [MM] 101 Format: { eager | lazy } 102 default: lazy 103 By default, unaccepted memory is accepted lazily to 104 avoid prolonged boot times. The lazy option will add 105 some runtime overhead until all memory is eventually 106 accepted. In most cases the overhead is negligible. 107 For some workloads or for debugging purposes 108 accept_memory=eager can be used to accept all memory 109 at once during boot. 110 111 acpi= [HW,ACPI,X86,ARM64,RISCV64,EARLY] 112 Advanced Configuration and Power Interface 113 Format: { force | on | off | strict | noirq | rsdt | 114 copy_dsdt | nospcr } 115 force -- enable ACPI if default was off 116 on -- enable ACPI but allow fallback to DT [arm64,riscv64] 117 off -- disable ACPI if default was on 118 noirq -- do not use ACPI for IRQ routing 119 strict -- Be less tolerant of platforms that are not 120 strictly ACPI specification compliant. 121 rsdt -- prefer RSDT over (default) XSDT 122 copy_dsdt -- copy DSDT to memory 123 nocmcff -- Disable firmware first mode for corrected 124 errors. This disables parsing the HEST CMC error 125 source to check if firmware has set the FF flag. This 126 may result in duplicate corrected error reports. 127 nospcr -- disable console in ACPI SPCR table as 128 default _serial_ console on ARM64 129 spcr -- enable console in ACPI SPCR table as 130 default _serial_ console on x86 131 For ARM64, ONLY "acpi=off", "acpi=on", "acpi=force" or 132 "acpi=nospcr" are available 133 For RISCV64, ONLY "acpi=off", "acpi=on" or "acpi=force" 134 are available 135 136 See also Documentation/power/runtime_pm.rst, pci=noacpi 137 138 acpi_apic_instance= [ACPI,IOAPIC,EARLY] 139 Format: <int> 140 2: use 2nd APIC table, if available 141 1,0: use 1st APIC table 142 default: 0 143 144 acpi_backlight= [HW,ACPI] 145 { vendor | video | native | none } 146 If set to vendor, prefer vendor-specific driver 147 (e.g. thinkpad_acpi, sony_acpi, etc.) instead 148 of the ACPI video.ko driver. 149 If set to video, use the ACPI video.ko driver. 150 If set to native, use the device's native backlight mode. 151 If set to none, disable the ACPI backlight interface. 152 153 acpi_force_32bit_fadt_addr [ACPI,EARLY] 154 force FADT to use 32 bit addresses rather than the 155 64 bit X_* addresses. Some firmware have broken 64 156 bit addresses for force ACPI ignore these and use 157 the older legacy 32 bit addresses. 158 159 acpica_no_return_repair [HW, ACPI] 160 Disable AML predefined validation mechanism 161 This mechanism can repair the evaluation result to make 162 the return objects more ACPI specification compliant. 163 This option is useful for developers to identify the 164 root cause of an AML interpreter issue when the issue 165 has something to do with the repair mechanism. 166 167 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG] 168 acpi.debug_level= [HW,ACPI,ACPI_DEBUG] 169 Format: <int> 170 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI 171 debug output. Bits in debug_layer correspond to a 172 _COMPONENT in an ACPI source file, e.g., 173 #define _COMPONENT ACPI_EVENTS 174 Bits in debug_level correspond to a level in 175 ACPI_DEBUG_PRINT statements, e.g., 176 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ... 177 The debug_level mask defaults to "info". See 178 Documentation/firmware-guide/acpi/debug.rst for more information about 179 debug layers and levels. 180 181 Enable processor driver info messages: 182 acpi.debug_layer=0x20000000 183 Enable AML "Debug" output, i.e., stores to the Debug 184 object while interpreting AML: 185 acpi.debug_layer=0xffffffff acpi.debug_level=0x2 186 Enable all messages related to ACPI hardware: 187 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff 188 189 Some values produce so much output that the system is 190 unusable. The "log_buf_len" parameter may be useful 191 if you need to capture more output. 192 193 acpi.poweroff_on_fatal= [ACPI] 194 {0 | 1} 195 Causes the system to poweroff when the ACPI bytecode signals 196 a fatal error. The default value of this setting is 1. 197 Overriding this value should only be done for diagnosing 198 ACPI firmware problems, as the system might behave erratically 199 after having encountered a fatal ACPI error. 200 201 acpi_enforce_resources= [ACPI] 202 { strict | lax | no } 203 Check for resource conflicts between native drivers 204 and ACPI OperationRegions (SystemIO and SystemMemory 205 only). IO ports and memory declared in ACPI might be 206 used by the ACPI subsystem in arbitrary AML code and 207 can interfere with legacy drivers. 208 strict (default): access to resources claimed by ACPI 209 is denied; legacy drivers trying to access reserved 210 resources will fail to bind to device using them. 211 lax: access to resources claimed by ACPI is allowed; 212 legacy drivers trying to access reserved resources 213 will bind successfully but a warning message is logged. 214 no: ACPI OperationRegions are not marked as reserved, 215 no further checks are performed. 216 217 acpi_force_table_verification [HW,ACPI,EARLY] 218 Enable table checksum verification during early stage. 219 By default, this is disabled due to x86 early mapping 220 size limitation. 221 222 acpi_irq_balance [HW,ACPI] 223 ACPI will balance active IRQs 224 default in APIC mode 225 226 acpi_irq_nobalance [HW,ACPI] 227 ACPI will not move active IRQs (default) 228 default in PIC mode 229 230 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA 231 Format: <irq>,<irq>... 232 233 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for 234 use by PCI 235 Format: <irq>,<irq>... 236 237 acpi_mask_gpe= [HW,ACPI] 238 Due to the existence of _Lxx/_Exx, some GPEs triggered 239 by unsupported hardware/firmware features can result in 240 GPE floodings that cannot be automatically disabled by 241 the GPE dispatcher. 242 This facility can be used to prevent such uncontrolled 243 GPE floodings. 244 Format: <byte> or <bitmap-list> 245 246 acpi_no_auto_serialize [HW,ACPI] 247 Disable auto-serialization of AML methods 248 AML control methods that contain the opcodes to create 249 named objects will be marked as "Serialized" by the 250 auto-serialization feature. 251 This feature is enabled by default. 252 This option allows to turn off the feature. 253 254 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump 255 kernels. 256 257 acpi_no_static_ssdt [HW,ACPI,EARLY] 258 Disable installation of static SSDTs at early boot time 259 By default, SSDTs contained in the RSDT/XSDT will be 260 installed automatically and they will appear under 261 /sys/firmware/acpi/tables. 262 This option turns off this feature. 263 Note that specifying this option does not affect 264 dynamic table installation which will install SSDT 265 tables to /sys/firmware/acpi/tables/dynamic. 266 267 acpi_no_watchdog [HW,ACPI,WDT] 268 Ignore the ACPI-based watchdog interface (WDAT) and let 269 a native driver control the watchdog device instead. 270 271 acpi_rsdp= [ACPI,EFI,KEXEC,EARLY] 272 Pass the RSDP address to the kernel, mostly used 273 on machines running EFI runtime service to boot the 274 second kernel for kdump. 275 276 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS 277 Format: To spoof as Windows 98: ="Microsoft Windows" 278 279 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead 280 of 2 which is mandated by ACPI 6) as the supported ACPI 281 specification revision (when using this switch, it may 282 be necessary to carry out a cold reboot _twice_ in a 283 row to make it take effect on the platform firmware). 284 285 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings 286 acpi_osi="string1" # add string1 287 acpi_osi="!string2" # remove string2 288 acpi_osi=!* # remove all strings 289 acpi_osi=! # disable all built-in OS vendor 290 strings 291 acpi_osi=!! # enable all built-in OS vendor 292 strings 293 acpi_osi= # disable all strings 294 295 'acpi_osi=!' can be used in combination with single or 296 multiple 'acpi_osi="string1"' to support specific OS 297 vendor string(s). Note that such command can only 298 affect the default state of the OS vendor strings, thus 299 it cannot affect the default state of the feature group 300 strings and the current state of the OS vendor strings, 301 specifying it multiple times through kernel command line 302 is meaningless. This command is useful when one do not 303 care about the state of the feature group strings which 304 should be controlled by the OSPM. 305 Examples: 306 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent 307 to 'acpi_osi="Windows 2000" acpi_osi=!', they all 308 can make '_OSI("Windows 2000")' TRUE. 309 310 'acpi_osi=' cannot be used in combination with other 311 'acpi_osi=' command lines, the _OSI method will not 312 exist in the ACPI namespace. NOTE that such command can 313 only affect the _OSI support state, thus specifying it 314 multiple times through kernel command line is also 315 meaningless. 316 Examples: 317 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)' 318 FALSE. 319 320 'acpi_osi=!*' can be used in combination with single or 321 multiple 'acpi_osi="string1"' to support specific 322 string(s). Note that such command can affect the 323 current state of both the OS vendor strings and the 324 feature group strings, thus specifying it multiple times 325 through kernel command line is meaningful. But it may 326 still not able to affect the final state of a string if 327 there are quirks related to this string. This command 328 is useful when one want to control the state of the 329 feature group strings to debug BIOS issues related to 330 the OSPM features. 331 Examples: 332 1. 'acpi_osi="Module Device" acpi_osi=!*' can make 333 '_OSI("Module Device")' FALSE. 334 2. 'acpi_osi=!* acpi_osi="Module Device"' can make 335 '_OSI("Module Device")' TRUE. 336 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is 337 equivalent to 338 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"' 339 and 340 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!', 341 they all will make '_OSI("Windows 2000")' TRUE. 342 343 acpi_pm_good [X86] 344 Override the pmtimer bug detection: force the kernel 345 to assume that this machine's pmtimer latches its value 346 and always returns good values. 347 348 acpi_sci= [HW,ACPI,EARLY] ACPI System Control Interrupt trigger mode 349 Format: { level | edge | high | low } 350 351 acpi_skip_timer_override [HW,ACPI,EARLY] 352 Recognize and ignore IRQ0/pin2 Interrupt Override. 353 For broken nForce2 BIOS resulting in XT-PIC timer. 354 355 acpi_sleep= [HW,ACPI] Sleep options 356 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig, 357 s4_nohwsig, old_ordering, nonvs, 358 sci_force_enable, nobl } 359 See Documentation/power/video.rst for information on 360 s3_bios and s3_mode. 361 s3_beep is for debugging; it makes the PC's speaker beep 362 as soon as the kernel's real-mode entry point is called. 363 s4_hwsig causes the kernel to check the ACPI hardware 364 signature during resume from hibernation, and gracefully 365 refuse to resume if it has changed. This complies with 366 the ACPI specification but not with reality, since 367 Windows does not do this and many laptops do change it 368 on docking. So the default behaviour is to allow resume 369 and simply warn when the signature changes, unless the 370 s4_hwsig option is enabled. 371 s4_nohwsig prevents ACPI hardware signature from being 372 used (or even warned about) during resume. 373 old_ordering causes the ACPI 1.0 ordering of the _PTS 374 control method, with respect to putting devices into 375 low power states, to be enforced (the ACPI 2.0 ordering 376 of _PTS is used by default). 377 nonvs prevents the kernel from saving/restoring the 378 ACPI NVS memory during suspend/hibernation and resume. 379 sci_force_enable causes the kernel to set SCI_EN directly 380 on resume from S1/S3 (which is against the ACPI spec, 381 but some broken systems don't work without it). 382 nobl causes the internal blacklist of systems known to 383 behave incorrectly in some ways with respect to system 384 suspend and resume to be ignored (use wisely). 385 386 acpi_use_timer_override [HW,ACPI,EARLY] 387 Use timer override. For some broken Nvidia NF5 boards 388 that require a timer override, but don't have HPET 389 390 add_efi_memmap [EFI,X86,EARLY] Include EFI memory map in 391 kernel's map of available physical RAM. 392 393 agp= [AGP] 394 { off | try_unsupported } 395 off: disable AGP support 396 try_unsupported: try to drive unsupported chipsets 397 (may crash computer or cause data corruption) 398 399 ALSA [HW,ALSA] 400 See Documentation/sound/alsa-configuration.rst 401 402 alignment= [KNL,ARM] 403 Allow the default userspace alignment fault handler 404 behaviour to be specified. Bit 0 enables warnings, 405 bit 1 enables fixups, and bit 2 sends a segfault. 406 407 align_va_addr= [X86-64] 408 Align virtual addresses by clearing slice [14:12] when 409 allocating a VMA at process creation time. This option 410 gives you up to 3% performance improvement on AMD F15h 411 machines (where it is enabled by default) for a 412 CPU-intensive style benchmark, and it can vary highly in 413 a microbenchmark depending on workload and compiler. 414 415 32: only for 32-bit processes 416 64: only for 64-bit processes 417 on: enable for both 32- and 64-bit processes 418 off: disable for both 32- and 64-bit processes 419 420 alloc_snapshot [FTRACE] 421 Allocate the ftrace snapshot buffer on boot up when the 422 main buffer is allocated. This is handy if debugging 423 and you need to use tracing_snapshot() on boot up, and 424 do not want to use tracing_snapshot_alloc() as it needs 425 to be done where GFP_KERNEL allocations are allowed. 426 427 allow_mismatched_32bit_el0 [ARM64,EARLY] 428 Allow execve() of 32-bit applications and setting of the 429 PER_LINUX32 personality on systems where only a strict 430 subset of the CPUs support 32-bit EL0. When this 431 parameter is present, the set of CPUs supporting 32-bit 432 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0 433 and hot-unplug operations may be restricted. 434 435 See Documentation/arch/arm64/asymmetric-32bit.rst for more 436 information. 437 438 amd_iommu= [HW,X86-64] 439 Pass parameters to the AMD IOMMU driver in the system. 440 Possible values are: 441 fullflush - Deprecated, equivalent to iommu.strict=1 442 off - do not initialize any AMD IOMMU found in 443 the system 444 force_isolation - Force device isolation for all 445 devices. The IOMMU driver is not 446 allowed anymore to lift isolation 447 requirements as needed. This option 448 does not override iommu=pt 449 force_enable - Force enable the IOMMU on platforms known 450 to be buggy with IOMMU enabled. Use this 451 option with care. 452 pgtbl_v1 - Use v1 page table for DMA-API (Default). 453 pgtbl_v2 - Use v2 page table for DMA-API. 454 irtcachedis - Disable Interrupt Remapping Table (IRT) caching. 455 nohugepages - Limit page-sizes used for v1 page-tables 456 to 4 KiB. 457 v2_pgsizes_only - Limit page-sizes used for v1 page-tables 458 to 4KiB/2Mib/1GiB. 459 460 461 amd_iommu_dump= [HW,X86-64] 462 Enable AMD IOMMU driver option to dump the ACPI table 463 for AMD IOMMU. With this option enabled, AMD IOMMU 464 driver will print ACPI tables for AMD IOMMU during 465 IOMMU initialization. 466 467 amd_iommu_intr= [HW,X86-64] 468 Specifies one of the following AMD IOMMU interrupt 469 remapping modes: 470 legacy - Use legacy interrupt remapping mode. 471 vapic - Use virtual APIC mode, which allows IOMMU 472 to inject interrupts directly into guest. 473 This mode requires kvm-amd.avic=1. 474 (Default when IOMMU HW support is present.) 475 476 amd_pstate= [X86,EARLY] 477 disable 478 Do not enable amd_pstate as the default 479 scaling driver for the supported processors 480 passive 481 Use amd_pstate with passive mode as a scaling driver. 482 In this mode autonomous selection is disabled. 483 Driver requests a desired performance level and platform 484 tries to match the same performance level if it is 485 satisfied by guaranteed performance level. 486 active 487 Use amd_pstate_epp driver instance as the scaling driver, 488 driver provides a hint to the hardware if software wants 489 to bias toward performance (0x0) or energy efficiency (0xff) 490 to the CPPC firmware. then CPPC power algorithm will 491 calculate the runtime workload and adjust the realtime cores 492 frequency. 493 guided 494 Activate guided autonomous mode. Driver requests minimum and 495 maximum performance level and the platform autonomously 496 selects a performance level in this range and appropriate 497 to the current workload. 498 499 amd_prefcore= 500 [X86] 501 disable 502 Disable amd-pstate preferred core. 503 504 amd_dynamic_epp= 505 [X86] 506 disable 507 Disable amd-pstate dynamic EPP. 508 enable 509 Enable amd-pstate dynamic EPP. 510 511 amijoy.map= [HW,JOY] Amiga joystick support 512 Map of devices attached to JOY0DAT and JOY1DAT 513 Format: <a>,<b> 514 See also Documentation/input/joydev/joystick.rst 515 516 analog.map= [HW,JOY] Analog joystick and gamepad support 517 Specifies type or capabilities of an analog joystick 518 connected to one of 16 gameports 519 Format: <type1>,<type2>,..<type16> 520 521 apc= [HW,SPARC] 522 Power management functions (SPARCstation-4/5 + deriv.) 523 Format: noidle 524 Disable APC CPU standby support. SPARCstation-Fox does 525 not play well with APC CPU idle - disable it if you have 526 APC and your system crashes randomly. 527 528 apic [APIC,X86-64] Use IO-APIC. Default. 529 530 apic= [APIC,X86,EARLY] Advanced Programmable Interrupt Controller 531 Change the output verbosity while booting 532 Format: { quiet (default) | verbose | debug } 533 Change the amount of debugging information output 534 when initialising the APIC and IO-APIC components. 535 536 apic_extnmi= [APIC,X86,EARLY] External NMI delivery setting 537 Format: { bsp (default) | all | none } 538 bsp: External NMI is delivered only to CPU 0 539 all: External NMIs are broadcast to all CPUs as a 540 backup of CPU 0 541 none: External NMI is masked for all CPUs. This is 542 useful so that a dump capture kernel won't be 543 shot down by NMI 544 545 apicpmtimer Do APIC timer calibration using the pmtimer. Implies 546 apicmaintimer. Useful when your PIT timer is totally 547 broken. 548 549 autoconf= [IPV6] 550 See Documentation/networking/ipv6.rst. 551 552 apm= [APM] Advanced Power Management 553 See header of arch/x86/kernel/apm_32.c. 554 555 apparmor= [APPARMOR] Disable or enable AppArmor at boot time 556 Format: { "0" | "1" } 557 See security/apparmor/Kconfig help text 558 0 -- disable. 559 1 -- enable. 560 Default value is set via kernel config option. 561 562 arm64.no32bit_el0 [ARM64] Unconditionally disable the execution of 563 32 bit applications. 564 565 arm64.nobti [ARM64] Unconditionally disable Branch Target 566 Identification support 567 568 arm64.nogcs [ARM64] Unconditionally disable Guarded Control Stack 569 support 570 571 arm64.nohaft [ARM64] Unconditionally disable Hardware managed Access 572 Flag for Table descriptors support 573 574 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory 575 Set instructions support 576 577 arm64.nompam [ARM64] Unconditionally disable Memory Partitioning And 578 Monitoring support 579 580 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension 581 support 582 583 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication 584 support 585 586 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix 587 Extension support 588 589 arm64.nosve [ARM64] Unconditionally disable Scalable Vector 590 Extension support 591 592 ataflop= [HW,M68k] 593 594 atarimouse= [HW,MOUSE] Atari Mouse 595 596 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess, 597 EzKey and similar keyboards 598 599 atkbd.reset= [HW] Reset keyboard during initialization 600 601 atkbd.set= [HW] Select keyboard code set 602 Format: <int> (2 = AT (default), 3 = PS/2) 603 604 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar 605 keyboards 606 607 atkbd.softraw= [HW] Choose between synthetic and real raw mode 608 Format: <bool> (0 = real, 1 = synthetic (default)) 609 610 atkbd.softrepeat= [HW] 611 Use software keyboard repeat 612 613 audit= [KNL] Enable the audit sub-system 614 Format: { "0" | "1" | "off" | "on" } 615 0 | off - kernel audit is disabled and can not be 616 enabled until the next reboot 617 unset - kernel audit is initialized but disabled and 618 will be fully enabled by the userspace auditd. 619 1 | on - kernel audit is initialized and partially 620 enabled, storing at most audit_backlog_limit 621 messages in RAM until it is fully enabled by the 622 userspace auditd. 623 Default: unset 624 625 audit_backlog_limit= [KNL] Set the audit queue size limit. 626 Format: <int> (must be >=0) 627 Default: 64 628 629 bau= [X86_UV] Enable the BAU on SGI UV. The default 630 behavior is to disable the BAU (i.e. bau=0). 631 Format: { "0" | "1" } 632 0 - Disable the BAU. 633 1 - Enable the BAU. 634 unset - Disable the BAU. 635 636 bdev_allow_write_mounted= 637 Format: <bool> 638 Control the ability to open a mounted block device 639 for writing, i.e., allow / disallow writes that bypass 640 the FS. This was implemented as a means to prevent 641 fuzzers from crashing the kernel by overwriting the 642 metadata underneath a mounted FS without its awareness. 643 This also prevents destructive formatting of mounted 644 filesystems by naive storage tooling that don't use 645 O_EXCL. Default is Y and can be changed through the 646 Kconfig option CONFIG_BLK_DEV_WRITE_MOUNTED. 647 648 bert_disable [ACPI] 649 Disable BERT OS support on buggy BIOSes. 650 651 bgrt_disable [ACPI,X86,EARLY] 652 Disable BGRT to avoid flickering OEM logo. 653 654 blkdevparts= Manual partition parsing of block device(s) for 655 embedded devices based on command line input. 656 See Documentation/block/cmdline-partition.rst 657 658 boot_delay= [KNL,EARLY] 659 Milliseconds to delay each printk during boot. 660 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled, 661 and you may also have to specify "lpj=". Boot_delay 662 values larger than 10 seconds (10000) are assumed 663 erroneous and ignored. 664 Format: integer 665 666 bootconfig [KNL,EARLY] 667 Extended command line options can be added to an initrd 668 and this will cause the kernel to look for it. 669 670 See Documentation/admin-guide/bootconfig.rst 671 672 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards) 673 bttv.radio= Most important insmod options are available as 674 kernel args too. 675 bttv.pll= See Documentation/admin-guide/media/bttv.rst 676 bttv.tuner= 677 678 bulk_remove=off [PPC] This parameter disables the use of the pSeries 679 firmware feature for flushing multiple hpte entries 680 at a time. 681 682 c101= [NET] Moxa C101 synchronous serial card 683 684 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection. 685 Sometimes CPU hardware bugs make them report the cache 686 size incorrectly. The kernel will attempt work arounds 687 to fix known problems, but for some CPUs it is not 688 possible to determine what the correct size should be. 689 This option provides an override for these situations. 690 691 carrier_timeout= 692 [NET] Specifies amount of time (in seconds) that 693 the kernel should wait for a network carrier. By default 694 it waits 120 seconds. 695 696 ca_keys= [KEYS] This parameter identifies a specific key(s) on 697 the system trusted keyring to be used for certificate 698 trust validation. 699 format: { id:<keyid> | builtin } 700 701 cca= [MIPS,EARLY] Override the kernel pages' cache coherency 702 algorithm. Accepted values range from 0 to 7 703 inclusive. See arch/mips/include/asm/pgtable-bits.h 704 for platform specific values (SB1, Loongson3 and 705 others). 706 707 ccw_timeout_log [S390] 708 See Documentation/arch/s390/common_io.rst for details. 709 710 cfi= [X86-64] Set Control Flow Integrity checking features 711 when CONFIG_FINEIBT is enabled. 712 Format: feature[,feature...] 713 Default: auto 714 715 auto: Use FineIBT if IBT available, otherwise kCFI. 716 Under FineIBT, enable "paranoid" mode when 717 FRED is not available. 718 off: Turn off CFI checking. 719 kcfi: Use kCFI (disable FineIBT). 720 fineibt: Use FineIBT (even if IBT not available). 721 norand: Do not re-randomize CFI hashes. 722 paranoid: Add caller hash checking under FineIBT. 723 bhi: Enable register poisoning to stop speculation 724 across FineIBT. (Disabled by default.) 725 warn: Do not enforce CFI checking: warn only. 726 debug: Report CFI initialization details. 727 728 cgroup_disable= [KNL] Disable a particular controller or optional feature 729 Format: {name of the controller(s) or feature(s) to disable} 730 The effects of cgroup_disable=foo are: 731 - foo isn't auto-mounted if you mount all cgroups in 732 a single hierarchy 733 - foo isn't visible as an individually mountable 734 subsystem 735 - if foo is an optional feature then the feature is 736 disabled and corresponding cgroup files are not 737 created 738 {Currently only "memory" controller deal with this and 739 cut the overhead, others just disable the usage. So 740 only cgroup_disable=memory is actually worthy} 741 Specifying "pressure" disables per-cgroup pressure 742 stall information accounting feature 743 744 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1 745 Format: { { controller | "all" | "named" } 746 [,{ controller | "all" | "named" }...] } 747 Like cgroup_disable, but only applies to cgroup v1; 748 the blacklisted controllers remain available in cgroup2. 749 "all" blacklists all controllers and "named" disables 750 named mounts. Specifying both "all" and "named" disables 751 all v1 hierarchies. 752 753 cgroup_v1_proc= [KNL] Show also missing controllers in /proc/cgroups 754 Format: { "true" | "false" } 755 /proc/cgroups lists only v1 controllers by default. 756 This compatibility option enables listing also v2 757 controllers (whose v1 code is not compiled!), so that 758 semi-legacy software can check this file to decide 759 about usage of v2 (sic) controllers. 760 761 cgroup_favordynmods= [KNL] Enable or Disable favordynmods. 762 Format: { "true" | "false" } 763 Defaults to the value of CONFIG_CGROUP_FAVOR_DYNMODS. 764 765 cgroup.memory= [KNL] Pass options to the cgroup memory controller. 766 Format: <string> 767 nosocket -- Disable socket memory accounting. 768 nokmem -- Disable kernel memory accounting. 769 nobpf -- Disable BPF memory accounting. 770 771 check_pages= [MM,EARLY] Enable sanity checking of pages after 772 allocations / before freeing. This adds checks to catch 773 double-frees, use-after-frees, and other sources of 774 page corruption by inspecting page internals (flags, 775 mapcount/refcount, memcg_data, etc.). 776 Format: { "0" | "1" } 777 Default: 0 (1 if CONFIG_DEBUG_VM is set) 778 779 checkreqprot= [SELINUX] Set initial checkreqprot flag value. 780 Format: { "0" | "1" } 781 See security/selinux/Kconfig help text. 782 0 -- check protection applied by kernel (includes 783 any implied execute protection). 784 1 -- check protection requested by application. 785 Default value is set via a kernel config option. 786 Value can be changed at runtime via 787 /sys/fs/selinux/checkreqprot. 788 Setting checkreqprot to 1 is deprecated. 789 790 cio_ignore= [S390] 791 See Documentation/arch/s390/common_io.rst for details. 792 793 clk_ignore_unused 794 [CLK] 795 Prevents the clock framework from automatically gating 796 clocks that have not been explicitly enabled by a Linux 797 device driver but are enabled in hardware at reset or 798 by the bootloader/firmware. Note that this does not 799 force such clocks to be always-on nor does it reserve 800 those clocks in any way. This parameter is useful for 801 debug and development, but should not be needed on a 802 platform with proper driver support. For more 803 information, see Documentation/driver-api/clk.rst. 804 805 clock= [BUGS=X86-32, HW] gettimeofday clocksource override. 806 [Deprecated] 807 Forces specified clocksource (if available) to be used 808 when calculating gettimeofday(). If specified 809 clocksource is not available, it defaults to PIT. 810 Format: { pit | tsc | cyclone | pmtmr } 811 812 clocksource= Override the default clocksource 813 Format: <string> 814 Override the default clocksource and use the clocksource 815 with the name specified. 816 Some clocksource names to choose from, depending on 817 the platform: 818 [all] jiffies (this is the base, fallback clocksource) 819 [ACPI] acpi_pm 820 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2, 821 pxa_timer,timer3,32k_counter,timer0_1 822 [X86-32] pit,hpet,tsc; 823 scx200_hrt on Geode; cyclone on IBM x440 824 [MIPS] MIPS 825 [PARISC] cr16 826 [S390] tod 827 [SH] SuperH 828 [SPARC64] tick 829 [X86-64] hpet,tsc 830 831 clocksource.arm_arch_timer.evtstrm= 832 [ARM,ARM64,EARLY] 833 Format: <bool> 834 Enable/disable the eventstream feature of the ARM 835 architected timer so that code using WFE-based polling 836 loops can be debugged more effectively on production 837 systems. 838 839 clocksource.verify_n_cpus= [KNL] 840 Limit the number of CPUs checked for clocksources 841 marked with CLOCK_SOURCE_VERIFY_PERCPU that 842 are marked unstable due to excessive skew. 843 A negative value says to check all CPUs, while 844 zero says not to check any. Values larger than 845 nr_cpu_ids are silently truncated to nr_cpu_ids. 846 The actual CPUs are chosen randomly, with 847 no replacement if the same CPU is chosen twice. 848 849 clocksource-wdtest.holdoff= [KNL] 850 Set the time in seconds that the clocksource 851 watchdog test waits before commencing its tests. 852 Defaults to zero when built as a module and to 853 10 seconds when built into the kernel. 854 855 cma=nn[MG]@[start[MG][-end[MG]]] 856 [KNL,CMA,EARLY] 857 Sets the size of kernel global memory area for 858 contiguous memory allocations and optionally the 859 placement constraint by the physical address range of 860 memory allocations. A value of 0 disables CMA 861 altogether. For more information, see 862 kernel/dma/contiguous.c 863 864 cma_pernuma=nn[MG] 865 [KNL,CMA,EARLY] 866 Sets the size of kernel per-numa memory area for 867 contiguous memory allocations. A value of 0 disables 868 per-numa CMA altogether. And If this option is not 869 specified, the default value is 0. 870 With per-numa CMA enabled, DMA users on node nid will 871 first try to allocate buffer from the pernuma area 872 which is located in node nid, if the allocation fails, 873 they will fallback to the global default memory area. 874 875 numa_cma=<node>:nn[MG][,<node>:nn[MG]] 876 [KNL,CMA,EARLY] 877 Sets the size of kernel numa memory area for 878 contiguous memory allocations. It will reserve CMA 879 area for the specified node. 880 881 If it is setup together with upper 'cmd_pernuma=' 882 (unlikely), its size setting takes priority for the 883 specified numa nodes. 884 885 With numa CMA enabled, DMA users on node nid will 886 first try to allocate buffer from the numa area 887 which is located in node nid, if the allocation fails, 888 they will fallback to the global default memory area. 889 890 cmo_free_hint= [PPC] Format: { yes | no } 891 Specify whether pages are marked as being inactive 892 when they are freed. This is used in CMO environments 893 to determine OS memory pressure for page stealing by 894 a hypervisor. 895 Default: yes 896 897 coherent_pool=nn[KMG] [ARM,KNL,EARLY] 898 Sets the size of memory pool for coherent, atomic dma 899 allocations, by default set to 256K. 900 901 condev= [HW,S390] console device 902 conmode= 903 904 con3215_drop= [S390,EARLY] 3215 console drop mode. 905 Format: y|n|Y|N|1|0 906 When set to true, drop data on the 3215 console when 907 the console buffer is full. In this case the 908 operator using a 3270 terminal emulator (for example 909 x3270) does not have to enter the clear key for the 910 console output to advance and the kernel to continue. 911 This leads to a much faster boot time when a 3270 912 terminal emulator is active. If no 3270 terminal 913 emulator is used, this parameter has no effect. 914 915 console= [KNL] Output console device and options. 916 917 tty<n> Use the virtual console device <n>. 918 919 ttyS<n>[,options] 920 ttyUSB0[,options] 921 Use the specified serial port. The options are of 922 the form "bbbbpnf", where "bbbb" is the baud rate, 923 "p" is parity ("n", "o", or "e"), "n" is number of 924 bits, and "f" is flow control ("r" for RTS or 925 omit it). Default is "9600n8". 926 927 See Documentation/admin-guide/serial-console.rst for more 928 information. See 929 Documentation/networking/netconsole.rst for an 930 alternative. 931 932 <DEVNAME>:<n>.<n>[,options] 933 Use the specified serial port on the serial core bus. 934 The addressing uses DEVNAME of the physical serial port 935 device, followed by the serial core controller instance, 936 and the serial port instance. The options are the same 937 as documented for the ttyS addressing above. 938 939 The mapping of the serial ports to the tty instances 940 can be viewed with: 941 942 $ ls -d /sys/bus/serial-base/devices/*:*.*/tty/* 943 /sys/bus/serial-base/devices/00:04:0.0/tty/ttyS0 944 945 In the above example, the console can be addressed with 946 console=00:04:0.0. Note that a console addressed this 947 way will only get added when the related device driver 948 is ready. The use of an earlycon parameter in addition to 949 the console may be desired for console output early on. 950 951 uart[8250],io,<addr>[,options] 952 uart[8250],mmio,<addr>[,options] 953 uart[8250],mmio16,<addr>[,options] 954 uart[8250],mmio32,<addr>[,options] 955 uart[8250],0x<addr>[,options] 956 Start an early, polled-mode console on the 8250/16550 957 UART at the specified I/O port or MMIO address, 958 switching to the matching ttyS device later. 959 MMIO inter-register address stride is either 8-bit 960 (mmio), 16-bit (mmio16), or 32-bit (mmio32). 961 If none of [io|mmio|mmio16|mmio32], <addr> is assumed 962 to be equivalent to 'mmio'. 'options' are specified in 963 the same format described for ttyS above; if unspecified, 964 the h/w is not re-initialized. 965 966 hvc<n> Use the hypervisor console device <n>. This is for 967 both Xen and PowerPC hypervisors. 968 969 { null | "" } 970 Use to disable console output, i.e., to have kernel 971 console messages discarded. 972 This must be the only console= parameter used on the 973 kernel command line. 974 975 If the device connected to the port is not a TTY but a braille 976 device, prepend "brl," before the device type, for instance 977 console=brl,ttyS0 978 For now, only VisioBraille is supported. 979 980 console_msg_format= 981 [KNL] Change console messages format 982 default 983 By default we print messages on consoles in 984 "[time stamp] text\n" format (time stamp may not be 985 printed, depending on CONFIG_PRINTK_TIME or 986 `printk_time' param). 987 syslog 988 Switch to syslog format: "<%u>[time stamp] text\n" 989 IOW, each message will have a facility and loglevel 990 prefix. The format is similar to one used by syslog() 991 syscall, or to executing "dmesg -S --raw" or to reading 992 from /proc/kmsg. 993 994 consoleblank= [KNL] The console blank (screen saver) timeout in 995 seconds. A value of 0 disables the blank timer. 996 Defaults to 0. 997 998 coredump_filter= 999 [KNL] Change the default value for 1000 /proc/<pid>/coredump_filter. 1001 See also Documentation/filesystems/proc.rst. 1002 1003 coresight_cpu_debug.enable 1004 [ARM,ARM64] 1005 Format: <bool> 1006 Enable/disable the CPU sampling based debugging. 1007 0: default value, disable debugging 1008 1: enable debugging at boot time 1009 1010 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver 1011 Format: 1012 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>] 1013 1014 cpuidle.off=1 [CPU_IDLE] 1015 disable the cpuidle sub-system 1016 1017 cpuidle.governor= 1018 [CPU_IDLE] Name of the cpuidle governor to use. 1019 1020 cpufreq.off=1 [CPU_FREQ] 1021 disable the cpufreq sub-system 1022 1023 cpufreq.default_governor= 1024 [CPU_FREQ] Name of the default cpufreq governor or 1025 policy to use. This governor must be registered in the 1026 kernel before the cpufreq driver probes. 1027 1028 cpu_init_udelay=N 1029 [X86,EARLY] Delay for N microsec between assert and de-assert 1030 of APIC INIT to start processors. This delay occurs 1031 on every CPU online, such as boot, and resume from suspend. 1032 Default: 10000 1033 1034 cpuhp.parallel= 1035 [SMP] Enable/disable parallel bringup of secondary CPUs 1036 Format: <bool> 1037 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise 1038 the parameter has no effect. 1039 1040 crash_kexec_post_notifiers 1041 Only jump to kdump kernel after running the panic 1042 notifiers and dumping kmsg. This option increases 1043 the risks of a kdump failure, since some panic 1044 notifiers can make the crashed kernel more unstable. 1045 In configurations where kdump may not be reliable, 1046 running the panic notifiers could allow collecting 1047 more data on dmesg, like stack traces from other CPUS 1048 or extra data dumped by panic_print. Note that some 1049 configurations enable this option unconditionally, 1050 like Hyper-V, PowerPC (fadump) and AMD SEV-SNP. 1051 1052 crashkernel=size[KMG][@offset[KMG]] 1053 [KNL,EARLY] Using kexec, Linux can switch to a 'crash kernel' 1054 upon panic. This parameter reserves the physical 1055 memory region [offset, offset + size] for that kernel 1056 image. If '@offset' is omitted, then a suitable offset 1057 is selected automatically. 1058 [KNL, X86-64, ARM64, RISCV, LoongArch] Select a region 1059 under 4G first, and fall back to reserve region above 1060 4G when '@offset' hasn't been specified. 1061 See Documentation/admin-guide/kdump/kdump.rst for further details. 1062 1063 crashkernel=range1:size1[,range2:size2,...][@offset] 1064 [KNL] Same as above, but depends on the memory 1065 in the running system. The syntax of range is 1066 start-[end] where start and end are both 1067 a memory unit (amount[KMG]). See also 1068 Documentation/admin-guide/kdump/kdump.rst for an example. 1069 1070 crashkernel=size[KMG],high 1071 [KNL, X86-64, ARM64, RISCV, LoongArch] range could be 1072 above 4G. 1073 Allow kernel to allocate physical memory region from top, 1074 so could be above 4G if system have more than 4G ram 1075 installed. Otherwise memory region will be allocated 1076 below 4G, if available. 1077 It will be ignored if crashkernel=X is specified. 1078 crashkernel=size[KMG],low 1079 [KNL, X86-64, ARM64, RISCV, LoongArch] range under 4G. 1080 When crashkernel=X,high is passed, kernel could allocate 1081 physical memory region above 4G, that cause second kernel 1082 crash on system that require some amount of low memory, 1083 e.g. swiotlb requires at least 64M+32K low memory, also 1084 enough extra low memory is needed to make sure DMA buffers 1085 for 32-bit devices won't run out. Kernel would try to allocate 1086 default size of memory below 4G automatically. The default 1087 size is platform dependent. 1088 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB) 1089 --> arm64: 128MiB 1090 --> riscv: 128MiB 1091 --> loongarch: 128MiB 1092 This one lets the user specify own low range under 4G 1093 for second kernel instead. 1094 0: to disable low allocation. 1095 It will be ignored when crashkernel=X,high is not used 1096 or memory reserved is below 4G. 1097 crashkernel=size[KMG],cma 1098 [KNL, X86, ARM64, RISCV, PPC] Reserve additional crash 1099 kernel memory from CMA. This reservation is usable by 1100 the first system's userspace memory and kernel movable 1101 allocations (memory balloon, zswap). Pages allocated 1102 from this memory range will not be included in the vmcore 1103 so this should not be used if dumping of userspace memory 1104 is intended and it has to be expected that some movable 1105 kernel pages may be missing from the dump. 1106 1107 A standard crashkernel reservation, as described above, 1108 is still needed to hold the crash kernel and initrd. 1109 1110 This option increases the risk of a kdump failure: DMA 1111 transfers configured by the first kernel may end up 1112 corrupting the second kernel's memory. 1113 1114 This reservation method is intended for systems that 1115 can't afford to sacrifice enough memory for standard 1116 crashkernel reservation and where less reliable and 1117 possibly incomplete kdump is preferable to no kdump at 1118 all. 1119 1120 cryptomgr.notests 1121 [KNL] Disable crypto self-tests 1122 1123 cs89x0_dma= [HW,NET] 1124 Format: <dma> 1125 1126 cs89x0_media= [HW,NET] 1127 Format: { rj45 | aui | bnc } 1128 1129 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU 1130 function call handling. When switched on, 1131 additional debug data is printed to the console 1132 in case a hanging CPU is detected, and that 1133 CPU is pinged again in order to try to resolve 1134 the hang situation. The default value of this 1135 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT 1136 Kconfig option. 1137 1138 dasd= [HW,NET] 1139 See header of drivers/s390/block/dasd_devmap.c. 1140 1141 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port 1142 (one device per port) 1143 Format: <port#>,<type> 1144 See also Documentation/input/devices/joystick-parport.rst 1145 1146 debug [KNL,EARLY] Enable kernel debugging (events log level). 1147 1148 debug_boot_weak_hash 1149 [KNL,EARLY] Enable printing [hashed] pointers early in the 1150 boot sequence. If enabled, we use a weak hash instead 1151 of siphash to hash pointers. Use this option if you are 1152 seeing instances of '(___ptrval___)') and need to see a 1153 value (hashed pointer) instead. Cryptographically 1154 insecure, please do not use on production kernels. 1155 1156 debug_locks_verbose= 1157 [KNL] verbose locking self-tests 1158 Format: <int> 1159 Print debugging info while doing the locking API 1160 self-tests. 1161 Bitmask for the various LOCKTYPE_ tests. Defaults to 0 1162 (no extra messages), setting it to -1 (all bits set) 1163 will print _a_lot_ more information - normally only 1164 useful to lockdep developers. 1165 1166 debug_objects [KNL,EARLY] Enable object debugging 1167 1168 debug_guardpage_minorder= 1169 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this 1170 parameter allows control of the order of pages that will 1171 be intentionally kept free (and hence protected) by the 1172 buddy allocator. Bigger value increase the probability 1173 of catching random memory corruption, but reduce the 1174 amount of memory for normal system use. The maximum 1175 possible value is MAX_PAGE_ORDER/2. Setting this 1176 parameter to 1 or 2 should be enough to identify most 1177 random memory corruption problems caused by bugs in 1178 kernel or driver code when a CPU writes to (or reads 1179 from) a random memory location. Note that there exists 1180 a class of memory corruptions problems caused by buggy 1181 H/W or F/W or by drivers badly programming DMA 1182 (basically when memory is written at bus level and the 1183 CPU MMU is bypassed) which are not detectable by 1184 CONFIG_DEBUG_PAGEALLOC, hence this option will not 1185 help tracking down these problems. 1186 1187 debug_pagealloc= 1188 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this parameter 1189 enables the feature at boot time. By default, it is 1190 disabled and the system will work mostly the same as a 1191 kernel built without CONFIG_DEBUG_PAGEALLOC. 1192 Note: to get most of debug_pagealloc error reports, it's 1193 useful to also enable the page_owner functionality. 1194 on: enable the feature 1195 1196 debugfs= [KNL,EARLY] This parameter enables what is exposed to 1197 userspace and debugfs internal clients. 1198 Format: { on, off } 1199 on: All functions are enabled. 1200 off: Filesystem is not registered and clients 1201 get a -EPERM as result when trying to register files 1202 or directories within debugfs. 1203 This is equivalent of the runtime functionality if 1204 debugfs was not enabled in the kernel at all. 1205 Default value is set in build-time with a kernel configuration. 1206 1207 debugpat [X86] Enable PAT debugging 1208 1209 default_hugepagesz= 1210 [HW] The size of the default HugeTLB page. This is 1211 the size represented by the legacy /proc/ hugepages 1212 APIs. In addition, this is the default hugetlb size 1213 used for shmget(), mmap() and mounting hugetlbfs 1214 filesystems. If not specified, defaults to the 1215 architecture's default huge page size. Huge page 1216 sizes are architecture dependent. See also 1217 Documentation/admin-guide/mm/hugetlbpage.rst. 1218 Format: size[KMG] 1219 1220 deferred_probe_timeout= 1221 [KNL] Debugging option to set a timeout in seconds for 1222 deferred probe to give up waiting on dependencies to 1223 probe. Only specific dependencies (subsystems or 1224 drivers) that have opted in will be ignored. A timeout 1225 of 0 will timeout at the end of initcalls; a negative 1226 value is treated as an infinite timeout value. If the 1227 timeout hasn't expired, it'll be restarted by each 1228 successful driver registration. This option will also 1229 dump out devices still on the deferred probe list after 1230 retrying. 1231 1232 delayacct [KNL] Enable per-task delay accounting 1233 1234 dell_smm_hwmon.ignore_dmi= 1235 [HW] Continue probing hardware even if DMI data 1236 indicates that the driver is running on unsupported 1237 hardware. 1238 1239 dell_smm_hwmon.force= 1240 [HW] Activate driver even if SMM BIOS signature does 1241 not match list of supported models and enable otherwise 1242 blacklisted features. 1243 1244 dell_smm_hwmon.power_status= 1245 [HW] Report power status in /proc/i8k 1246 (disabled by default). 1247 1248 dell_smm_hwmon.restricted= 1249 [HW] Allow controlling fans only if SYS_ADMIN 1250 capability is set. 1251 1252 dell_smm_hwmon.fan_mult= 1253 [HW] Factor to multiply fan speed with. 1254 1255 dell_smm_hwmon.fan_max= 1256 [HW] Maximum configurable fan speed. 1257 1258 dfltcc= [HW,S390] 1259 Format: { on | off | def_only | inf_only | always } 1260 on: s390 zlib hardware support for compression on 1261 level 1 and decompression (default) 1262 off: No s390 zlib hardware support 1263 def_only: s390 zlib hardware support for deflate 1264 only (compression on level 1) 1265 inf_only: s390 zlib hardware support for inflate 1266 only (decompression) 1267 always: Same as 'on' but ignores the selected compression 1268 level always using hardware support (used for debugging) 1269 1270 dhash_entries= [KNL] 1271 Set number of hash buckets for dentry cache. 1272 1273 disable_1tb_segments [PPC,EARLY] 1274 Disables the use of 1TB hash page table segments. This 1275 causes the kernel to fall back to 256MB segments which 1276 can be useful when debugging issues that require an SLB 1277 miss to occur. 1278 1279 disable= [IPV6] 1280 See Documentation/networking/ipv6.rst. 1281 1282 disable_radix [PPC,EARLY] 1283 Disable RADIX MMU mode on POWER9 1284 1285 disable_tlbie [PPC] 1286 Disable TLBIE instruction. Currently does not work 1287 with KVM, with HASH MMU, or with coherent accelerators. 1288 1289 disable_ddw [PPC/PSERIES,EARLY] 1290 Disable Dynamic DMA Window support. Use this 1291 to workaround buggy firmware. 1292 1293 disable_ipv6= [IPV6] 1294 See Documentation/networking/ipv6.rst. 1295 1296 disable_mtrr_cleanup [X86,EARLY] 1297 The kernel tries to adjust MTRR layout from continuous 1298 to discrete, to make X server driver able to add WB 1299 entry later. This parameter disables that. 1300 1301 disable_mtrr_trim [X86, Intel and AMD only,EARLY] 1302 By default the kernel will trim any uncacheable 1303 memory out of your available memory pool based on 1304 MTRR settings. This parameter disables that behavior, 1305 possibly causing your machine to run very slowly. 1306 1307 disable_timer_pin_1 [X86,EARLY] 1308 Disable PIN 1 of APIC timer 1309 Can be useful to work around chipset bugs. 1310 1311 dis_ucode_ldr [X86] Disable the microcode loader. 1312 1313 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support, 1314 this option disables the debugging code at boot. 1315 1316 dma_debug_entries=<number> 1317 This option allows to tune the number of preallocated 1318 entries for DMA-API debugging code. One entry is 1319 required per DMA-API allocation. Use this if the 1320 DMA-API debugging code disables itself because the 1321 architectural default is too low. 1322 1323 dma_debug_driver=<driver_name> 1324 With this option the DMA-API debugging driver 1325 filter feature can be enabled at boot time. Just 1326 pass the driver to filter for as the parameter. 1327 The filter can be disabled or changed to another 1328 driver later using sysfs. 1329 1330 reg_file_data_sampling= 1331 [X86] Controls mitigation for Register File Data 1332 Sampling (RFDS) vulnerability. RFDS is a CPU 1333 vulnerability which may allow userspace to infer 1334 kernel data values previously stored in floating point 1335 registers, vector registers, or integer registers. 1336 RFDS only affects Intel Atom processors. 1337 1338 on: Turns ON the mitigation. 1339 off: Turns OFF the mitigation. 1340 1341 This parameter overrides the compile time default set 1342 by CONFIG_MITIGATION_RFDS. Mitigation cannot be 1343 disabled when other VERW based mitigations (like MDS) 1344 are enabled. In order to disable RFDS mitigation all 1345 VERW based mitigations need to be disabled. 1346 1347 For details see: 1348 Documentation/admin-guide/hw-vuln/reg-file-data-sampling.rst 1349 1350 dm_verity.keyring_unsealed= 1351 [KNL] When set to 1, leave the dm-verity keyring 1352 unsealed after initialization so userspace can 1353 provision keys. Once the keyring is restricted 1354 it becomes active and is searched during signature 1355 verification. 1356 1357 driver_async_probe= [KNL] 1358 List of driver names to be probed asynchronously. * 1359 matches with all driver names. If * is specified, the 1360 rest of the listed driver names are those that will NOT 1361 match the *. 1362 Format: <driver_name1>,<driver_name2>... 1363 1364 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>] 1365 Broken monitors, graphic adapters, KVMs and EDIDless 1366 panels may send no or incorrect EDID data sets. 1367 This parameter allows to specify an EDID data sets 1368 in the /lib/firmware directory that are used instead. 1369 An EDID data set will only be used for a particular 1370 connector, if its name and a colon are prepended to 1371 the EDID name. Each connector may use a unique EDID 1372 data set by separating the files with a comma. An EDID 1373 data set with no connector name will be used for 1374 any connectors not explicitly specified. 1375 1376 dscc4.setup= [NET] 1377 1378 dt_cpu_ftrs= [PPC,EARLY] 1379 Format: {"off" | "known"} 1380 Control how the dt_cpu_ftrs device-tree binding is 1381 used for CPU feature discovery and setup (if it 1382 exists). 1383 off: Do not use it, fall back to legacy cpu table. 1384 known: Do not pass through unknown features to guests 1385 or userspace, only those that the kernel is aware of. 1386 1387 dump_apple_properties [X86] 1388 Dump name and content of EFI device properties on 1389 x86 Macs. Useful for driver authors to determine 1390 what data is available or for reverse-engineering. 1391 1392 dyndbg[="val"] [KNL,DYNAMIC_DEBUG] 1393 <module>.dyndbg[="val"] 1394 Enable debug messages at boot time. See 1395 Documentation/admin-guide/dynamic-debug-howto.rst 1396 for details. 1397 1398 early_ioremap_debug [KNL,EARLY] 1399 Enable debug messages in early_ioremap support. This 1400 is useful for tracking down temporary early mappings 1401 which are not unmapped. 1402 1403 earlycon= [KNL,EARLY] Output early console device and options. 1404 1405 When used with no options, the early console is 1406 determined by stdout-path property in device tree's 1407 chosen node or the ACPI SPCR table if supported by 1408 the platform. 1409 1410 cdns,<addr>[,options] 1411 Start an early, polled-mode console on a Cadence 1412 (xuartps) serial port at the specified address. Only 1413 supported option is baud rate. If baud rate is not 1414 specified, the serial port must already be setup and 1415 configured. 1416 1417 uart[8250],io,<addr>[,options[,uartclk]] 1418 uart[8250],mmio,<addr>[,options[,uartclk]] 1419 uart[8250],mmio32,<addr>[,options[,uartclk]] 1420 uart[8250],mmio32be,<addr>[,options[,uartclk]] 1421 uart[8250],0x<addr>[,options] 1422 Start an early, polled-mode console on the 8250/16550 1423 UART at the specified I/O port or MMIO address. 1424 MMIO inter-register address stride is either 8-bit 1425 (mmio) or 32-bit (mmio32 or mmio32be). 1426 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed 1427 to be equivalent to 'mmio'. 'options' are specified 1428 in the same format described for "console=ttyS<n>"; if 1429 unspecified, the h/w is not initialized. 'uartclk' is 1430 the uart clock frequency; if unspecified, it is set 1431 to 'BASE_BAUD' * 16. 1432 1433 pl011,<addr> 1434 pl011,mmio32,<addr> 1435 Start an early, polled-mode console on a pl011 serial 1436 port at the specified address. The pl011 serial port 1437 must already be setup and configured. Options are not 1438 yet supported. If 'mmio32' is specified, then only 1439 the driver will use only 32-bit accessors to read/write 1440 the device registers. 1441 1442 liteuart,<addr> 1443 Start an early console on a litex serial port at the 1444 specified address. The serial port must already be 1445 setup and configured. Options are not yet supported. 1446 1447 meson,<addr> 1448 Start an early, polled-mode console on a meson serial 1449 port at the specified address. The serial port must 1450 already be setup and configured. Options are not yet 1451 supported. 1452 1453 msm_serial,<addr> 1454 Start an early, polled-mode console on an msm serial 1455 port at the specified address. The serial port 1456 must already be setup and configured. Options are not 1457 yet supported. 1458 1459 msm_serial_dm,<addr> 1460 Start an early, polled-mode console on an msm serial 1461 dm port at the specified address. The serial port 1462 must already be setup and configured. Options are not 1463 yet supported. 1464 1465 owl,<addr> 1466 Start an early, polled-mode console on a serial port 1467 of an Actions Semi SoC, such as S500 or S900, at the 1468 specified address. The serial port must already be 1469 setup and configured. Options are not yet supported. 1470 1471 rda,<addr> 1472 Start an early, polled-mode console on a serial port 1473 of an RDA Micro SoC, such as RDA8810PL, at the 1474 specified address. The serial port must already be 1475 setup and configured. Options are not yet supported. 1476 1477 sbi 1478 Use RISC-V SBI (Supervisor Binary Interface) for early 1479 console. 1480 1481 smh Use ARM semihosting calls for early console. 1482 1483 s3c2410,<addr> 1484 s3c2412,<addr> 1485 s3c2440,<addr> 1486 s3c6400,<addr> 1487 s5pv210,<addr> 1488 exynos4210,<addr> 1489 Use early console provided by serial driver available 1490 on Samsung SoCs, requires selecting proper type and 1491 a correct base address of the selected UART port. The 1492 serial port must already be setup and configured. 1493 Options are not yet supported. 1494 1495 lantiq,<addr> 1496 Start an early, polled-mode console on a lantiq serial 1497 (lqasc) port at the specified address. The serial port 1498 must already be setup and configured. Options are not 1499 yet supported. 1500 1501 lpuart,<addr> 1502 lpuart32,<addr> 1503 Use early console provided by Freescale LP UART driver 1504 found on Freescale Vybrid and QorIQ LS1021A processors. 1505 A valid base address must be provided, and the serial 1506 port must already be setup and configured. 1507 1508 ec_imx21,<addr> 1509 ec_imx6q,<addr> 1510 Start an early, polled-mode, output-only console on the 1511 Freescale i.MX UART at the specified address. The UART 1512 must already be setup and configured. 1513 1514 ar3700_uart,<addr> 1515 Start an early, polled-mode console on the 1516 Armada 3700 serial port at the specified 1517 address. The serial port must already be setup 1518 and configured. Options are not yet supported. 1519 1520 qcom_geni,<addr> 1521 Start an early, polled-mode console on a Qualcomm 1522 Generic Interface (GENI) based serial port at the 1523 specified address. The serial port must already be 1524 setup and configured. Options are not yet supported. 1525 1526 efifb,[options] 1527 Start an early, unaccelerated console on the EFI 1528 memory mapped framebuffer (if available). On cache 1529 coherent non-x86 systems that use system memory for 1530 the framebuffer, pass the 'ram' option so that it is 1531 mapped with the correct attributes. 1532 1533 linflex,<addr> 1534 Use early console provided by Freescale LINFlexD UART 1535 serial driver for NXP S32V234 SoCs. A valid base 1536 address must be provided, and the serial port must 1537 already be setup and configured. 1538 1539 earlyprintk= [X86,SH,ARM,M68k,S390,UM,EARLY] 1540 earlyprintk=vga 1541 earlyprintk=sclp 1542 earlyprintk=xen 1543 earlyprintk=serial[,ttySn[,baudrate]] 1544 earlyprintk=serial[,0x...[,baudrate]] 1545 earlyprintk=ttySn[,baudrate] 1546 earlyprintk=dbgp[debugController#] 1547 earlyprintk=mmio32,membase[,{nocfg|baudrate}] 1548 earlyprintk=pciserial[,force],bus:device.function[,{nocfg|baudrate}] 1549 earlyprintk=xdbc[xhciController#] 1550 earlyprintk=bios 1551 1552 earlyprintk is useful when the kernel crashes before 1553 the normal console is initialized. It is not enabled by 1554 default because it has some cosmetic problems. 1555 1556 Use "nocfg" to skip UART configuration, assume 1557 BIOS/firmware has configured UART correctly. 1558 1559 Append ",keep" to not disable it when the real console 1560 takes over. 1561 1562 Only one of vga, serial, or usb debug port can 1563 be used at a time. 1564 1565 Currently only ttyS0 and ttyS1 may be specified by 1566 name. Other I/O ports may be explicitly specified 1567 on some architectures (x86 and arm at least) by 1568 replacing ttySn with an I/O port address, like this: 1569 earlyprintk=serial,0x1008,115200 1570 You can find the port for a given device in 1571 /proc/tty/driver/serial: 1572 2: uart:ST16650V2 port:00001008 irq:18 ... 1573 1574 Interaction with the standard serial driver is not 1575 very good. 1576 1577 The VGA output is eventually overwritten by 1578 the real console. 1579 1580 The xen option can only be used in Xen domains. 1581 1582 The sclp output can only be used on s390. 1583 1584 The bios output can only be used on SuperH. 1585 1586 The optional "force" to "pciserial" enables use of a 1587 PCI device even when its classcode is not of the 1588 UART class. 1589 1590 edac_report= [HW,EDAC] Control how to report EDAC event 1591 Format: {"on" | "off" | "force"} 1592 on: enable EDAC to report H/W event. May be overridden 1593 by other higher priority error reporting module. 1594 off: disable H/W event reporting through EDAC. 1595 force: enforce the use of EDAC to report H/W event. 1596 default: on. 1597 1598 edd= [EDD] 1599 Format: {"off" | "on" | "skip[mbr]"} 1600 1601 efi= [EFI,EARLY] 1602 Format: { "debug", "disable_early_pci_dma", 1603 "nochunk", "noruntime", "nosoftreserve", 1604 "novamap", "no_disable_early_pci_dma" } 1605 debug: enable misc debug output. 1606 disable_early_pci_dma: disable the busmaster bit on all 1607 PCI bridges while in the EFI boot stub. 1608 nochunk: disable reading files in "chunks" in the EFI 1609 boot stub, as chunking can cause problems with some 1610 firmware implementations. 1611 noruntime : disable EFI runtime services support 1612 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose) 1613 attribute may cause the kernel to reserve the 1614 memory range for a memory mapping driver to 1615 claim. Specify efi=nosoftreserve to disable this 1616 reservation and treat the memory by its base type 1617 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM"). 1618 novamap: do not call SetVirtualAddressMap(). 1619 no_disable_early_pci_dma: Leave the busmaster bit set 1620 on all PCI bridges while in the EFI boot stub 1621 1622 efi_no_storage_paranoia [EFI,X86,EARLY] 1623 The kernel reserves 5KB of EFI variable storage for 1624 safety, because some UEFI implementation may fail to 1625 boot if there's insufficient space in the EFI variable 1626 storage. 1627 1628 Using this parameter, you can use the 5KB reservation 1629 in the EFI variable storage. 1630 1631 However, Use this parameter only if you are really 1632 sure that your UEFI does sane gc and fulfills the spec 1633 otherwise your board may brick. 1634 1635 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT 1636 that is to be dynamically loaded by Linux. If there are 1637 multiple variables with the same name but with different 1638 vendor GUIDs, all of them will be loaded. See 1639 Documentation/admin-guide/acpi/ssdt-overlays.rst for details. 1640 1641 1642 eisa_irq_edge= [PARISC,HW] 1643 See header of drivers/parisc/eisa.c. 1644 1645 ekgdboc= [X86,KGDB,EARLY] Allow early kernel console debugging 1646 Format: ekgdboc=kbd 1647 1648 This is designed to be used in conjunction with 1649 the boot argument: earlyprintk=vga 1650 1651 This parameter works in place of the kgdboc parameter 1652 but can only be used if the backing tty is available 1653 very early in the boot process. For early debugging 1654 via a serial port see kgdboc_earlycon instead. 1655 1656 elfcorehdr=[size[KMG]@]offset[KMG] [PPC,SH,X86,S390,EARLY] 1657 Specifies physical address of start of kernel core 1658 image elf header and optionally the size. Generally 1659 kexec loader will pass this option to capture kernel. 1660 See Documentation/admin-guide/kdump/kdump.rst for details. 1661 1662 enable_mtrr_cleanup [X86,EARLY] 1663 The kernel tries to adjust MTRR layout from continuous 1664 to discrete, to make X server driver able to add WB 1665 entry later. This parameter enables that. 1666 1667 enable_timer_pin_1 [X86] 1668 Enable PIN 1 of APIC timer 1669 Can be useful to work around chipset bugs 1670 (in particular on some ATI chipsets). 1671 The kernel tries to set a reasonable default. 1672 1673 enforcing= [SELINUX] Set initial enforcing status. 1674 Format: {"0" | "1"} 1675 See security/selinux/Kconfig help text. 1676 0 -- permissive (log only, no denials). 1677 1 -- enforcing (deny and log). 1678 Default value is 0. 1679 Value can be changed at runtime via 1680 /sys/fs/selinux/enforce. 1681 1682 erst_disable [ACPI] 1683 Disable Error Record Serialization Table (ERST) 1684 support. 1685 1686 ether= [HW,NET] Ethernet cards parameters 1687 This option is obsoleted by the "netdev=" option, which 1688 has equivalent usage. See its documentation for details. 1689 1690 evm= [EVM] 1691 Format: { "fix" } 1692 Permit 'security.evm' to be updated regardless of 1693 current integrity status. 1694 1695 early_page_ext [KNL,EARLY] Enforces page_ext initialization to earlier 1696 stages so cover more early boot allocations. 1697 Please note that as side effect some optimizations 1698 might be disabled to achieve that (e.g. parallelized 1699 memory initialization is disabled) so the boot process 1700 might take longer, especially on systems with a lot of 1701 memory. Available with CONFIG_PAGE_EXTENSION=y. 1702 1703 failslab= 1704 fail_usercopy= 1705 fail_page_alloc= 1706 fail_skb_realloc= 1707 fail_make_request=[KNL] 1708 General fault injection mechanism. 1709 Format: <interval>,<probability>,<space>,<times> 1710 See also Documentation/fault-injection/. 1711 1712 fb_tunnels= [NET] 1713 Format: { initns | none } 1714 See Documentation/admin-guide/sysctl/net.rst for 1715 fb_tunnels_only_for_init_ns 1716 1717 floppy= [HW] 1718 See Documentation/admin-guide/blockdev/floppy.rst. 1719 1720 forcepae [X86-32] 1721 Forcefully enable Physical Address Extension (PAE). 1722 Many Pentium M systems disable PAE but may have a 1723 functionally usable PAE implementation. 1724 Warning: use of this parameter will taint the kernel 1725 and may cause unknown problems. 1726 1727 fred= [X86-64] 1728 Enable/disable Flexible Return and Event Delivery. 1729 Format: { on | off } 1730 on: enable FRED when it's present, the default setting. 1731 off: disable FRED. 1732 1733 ftrace=[tracer] 1734 [FTRACE] will set and start the specified tracer 1735 as early as possible in order to facilitate early 1736 boot debugging. 1737 1738 ftrace_boot_snapshot 1739 [FTRACE] On boot up, a snapshot will be taken of the 1740 ftrace ring buffer that can be read at: 1741 /sys/kernel/tracing/snapshot. 1742 This is useful if you need tracing information from kernel 1743 boot up that is likely to be overridden by user space 1744 start up functionality. 1745 1746 Optionally, the snapshot can also be defined for a tracing 1747 instance that was created by the trace_instance= command 1748 line parameter. 1749 1750 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo 1751 1752 The above will cause the "foo" tracing instance to trigger 1753 a snapshot at the end of boot up. 1754 1755 ftrace_dump_on_oops[=2(orig_cpu) | =<instance>][,<instance> | 1756 ,<instance>=2(orig_cpu)] 1757 [FTRACE] will dump the trace buffers on oops. 1758 If no parameter is passed, ftrace will dump global 1759 buffers of all CPUs, if you pass 2 or orig_cpu, it 1760 will dump only the buffer of the CPU that triggered 1761 the oops, or the specific instance will be dumped if 1762 its name is passed. Multiple instance dump is also 1763 supported, and instances are separated by commas. Each 1764 instance supports only dump on CPU that triggered the 1765 oops by passing 2 or orig_cpu to it. 1766 1767 ftrace_dump_on_oops=foo=orig_cpu 1768 1769 The above will dump only the buffer of "foo" instance 1770 on CPU that triggered the oops. 1771 1772 ftrace_dump_on_oops,foo,bar=orig_cpu 1773 1774 The above will dump global buffer on all CPUs, the 1775 buffer of "foo" instance on all CPUs and the buffer 1776 of "bar" instance on CPU that triggered the oops. 1777 1778 ftrace_filter=[function-list] 1779 [FTRACE] Limit the functions traced by the function 1780 tracer at boot up. function-list is a comma-separated 1781 list of functions. This list can be changed at run 1782 time by the set_ftrace_filter file in the debugfs 1783 tracing directory. 1784 1785 ftrace_notrace=[function-list] 1786 [FTRACE] Do not trace the functions specified in 1787 function-list. This list can be changed at run time 1788 by the set_ftrace_notrace file in the debugfs 1789 tracing directory. 1790 1791 ftrace_graph_filter=[function-list] 1792 [FTRACE] Limit the top level callers functions traced 1793 by the function graph tracer at boot up. 1794 function-list is a comma-separated list of functions 1795 that can be changed at run time by the 1796 set_graph_function file in the debugfs tracing directory. 1797 1798 ftrace_graph_notrace=[function-list] 1799 [FTRACE] Do not trace from the functions specified in 1800 function-list. This list is a comma-separated list of 1801 functions that can be changed at run time by the 1802 set_graph_notrace file in the debugfs tracing directory. 1803 1804 ftrace_graph_max_depth=<uint> 1805 [FTRACE] Used with the function graph tracer. This is 1806 the max depth it will trace into a function. This value 1807 can be changed at run time by the max_graph_depth file 1808 in the tracefs tracing directory. default: 0 (no limit) 1809 1810 fw_devlink= [KNL,EARLY] Create device links between consumer and supplier 1811 devices by scanning the firmware to infer the 1812 consumer/supplier relationships. This feature is 1813 especially useful when drivers are loaded as modules as 1814 it ensures proper ordering of tasks like device probing 1815 (suppliers first, then consumers), supplier boot state 1816 clean up (only after all consumers have probed), 1817 suspend/resume & runtime PM (consumers first, then 1818 suppliers). 1819 Format: { off | permissive | on | rpm } 1820 off -- Don't create device links from firmware info. 1821 permissive -- Create device links from firmware info 1822 but use it only for ordering boot state clean 1823 up (sync_state() calls). 1824 on -- Create device links from firmware info and use it 1825 to enforce probe and suspend/resume ordering. 1826 rpm -- Like "on", but also use to order runtime PM. 1827 1828 fw_devlink.strict=<bool> 1829 [KNL,EARLY] Treat all inferred dependencies as mandatory 1830 dependencies. This only applies for fw_devlink=on|rpm. 1831 Format: <bool> 1832 1833 fw_devlink.sync_state = 1834 [KNL,EARLY] When all devices that could probe have finished 1835 probing, this parameter controls what to do with 1836 devices that haven't yet received their sync_state() 1837 calls. 1838 Format: { strict | timeout } 1839 strict -- Default. Continue waiting on consumers to 1840 probe successfully. 1841 timeout -- Give up waiting on consumers and call 1842 sync_state() on any devices that haven't yet 1843 received their sync_state() calls after 1844 deferred_probe_timeout has expired or by 1845 late_initcall() if !CONFIG_MODULES. 1846 1847 gamecon.map[2|3]= 1848 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad 1849 support via parallel port (up to 5 devices per port) 1850 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5> 1851 See also Documentation/input/devices/joystick-parport.rst 1852 1853 gamma= [HW,DRM] 1854 1855 gart_fix_e820= [X86-64,EARLY] disable the fix e820 for K8 GART 1856 Format: off | on 1857 default: on 1858 1859 gather_data_sampling= 1860 [X86,INTEL,EARLY] Control the Gather Data Sampling (GDS) 1861 mitigation. 1862 1863 Gather Data Sampling is a hardware vulnerability which 1864 allows unprivileged speculative access to data which was 1865 previously stored in vector registers. 1866 1867 This issue is mitigated by default in updated microcode. 1868 The mitigation may have a performance impact but can be 1869 disabled. On systems without the microcode mitigation 1870 disabling AVX serves as a mitigation. 1871 1872 force: Disable AVX to mitigate systems without 1873 microcode mitigation. No effect if the microcode 1874 mitigation is present. Known to cause crashes in 1875 userspace with buggy AVX enumeration. 1876 1877 off: Disable GDS mitigation. 1878 1879 gbpages [X86] Use GB pages for kernel direct mappings. 1880 1881 gcov_persist= [GCOV] When non-zero (default), profiling data for 1882 kernel modules is saved and remains accessible via 1883 debugfs, even when the module is unloaded/reloaded. 1884 When zero, profiling data is discarded and associated 1885 debugfs files are removed at module unload time. 1886 1887 goldfish [X86] Enable the goldfish android emulator platform. 1888 Don't use this when you are not running on the 1889 android emulator 1890 1891 gpio-mockup.gpio_mockup_ranges 1892 [HW] Sets the ranges of gpiochip of for this device. 1893 Format: <start1>,<end1>,<start2>,<end2>... 1894 gpio-mockup.gpio_mockup_named_lines 1895 [HW] Let the driver know GPIO lines should be named. 1896 1897 gpt [EFI] Forces disk with valid GPT signature but 1898 invalid Protective MBR to be treated as GPT. If the 1899 primary GPT is corrupted, it enables the backup/alternate 1900 GPT to be used instead. 1901 1902 grcan.enable0= [HW] Configuration of physical interface 0. Determines 1903 the "Enable 0" bit of the configuration register. 1904 Format: 0 | 1 1905 Default: 0 1906 grcan.enable1= [HW] Configuration of physical interface 1. Determines 1907 the "Enable 0" bit of the configuration register. 1908 Format: 0 | 1 1909 Default: 0 1910 grcan.select= [HW] Select which physical interface to use. 1911 Format: 0 | 1 1912 Default: 0 1913 grcan.txsize= [HW] Sets the size of the tx buffer. 1914 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0. 1915 Default: 1024 1916 grcan.rxsize= [HW] Sets the size of the rx buffer. 1917 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0. 1918 Default: 1024 1919 1920 hardened_usercopy= 1921 [KNL] Under CONFIG_HARDENED_USERCOPY, whether 1922 hardening is enabled for this boot. Hardened 1923 usercopy checking is used to protect the kernel 1924 from reading or writing beyond known memory 1925 allocation boundaries as a proactive defense 1926 against bounds-checking flaws in the kernel's 1927 copy_to_user()/copy_from_user() interface. 1928 The default is determined by 1929 CONFIG_HARDENED_USERCOPY_DEFAULT_ON. 1930 on Perform hardened usercopy checks. 1931 off Disable hardened usercopy checks. 1932 1933 hardlockup_all_cpu_backtrace= 1934 [KNL] Should the hard-lockup detector generate 1935 backtraces on all cpus. 1936 Format: 0 | 1 1937 1938 hash_pointers= 1939 [KNL,EARLY] 1940 By default, when pointers are printed to the console 1941 or buffers via the %p format string, that pointer is 1942 "hashed", i.e. obscured by hashing the pointer value. 1943 This is a security feature that hides actual kernel 1944 addresses from unprivileged users, but it also makes 1945 debugging the kernel more difficult since unequal 1946 pointers can no longer be compared. The choices are: 1947 Format: { auto | always | never } 1948 Default: auto 1949 1950 auto - Hash pointers unless slab_debug is enabled. 1951 always - Always hash pointers (even if slab_debug is 1952 enabled). 1953 never - Never hash pointers. This option should only 1954 be specified when debugging the kernel. Do 1955 not use on production kernels. The boot 1956 param "no_hash_pointers" is an alias for 1957 this mode. 1958 1959 For controlling hashing dynamically at runtime, 1960 use the "kernel.kptr_restrict" sysctl instead. 1961 1962 hashdist= [KNL,NUMA] Large hashes allocated during boot 1963 are distributed across NUMA nodes. Defaults on 1964 for 64-bit NUMA, off otherwise. 1965 Format: 0 | 1 (for off | on) 1966 1967 hd= [EIDE] (E)IDE hard drive subsystem geometry 1968 Format: <cyl>,<head>,<sect> 1969 1970 hest_disable [ACPI] 1971 Disable Hardware Error Source Table (HEST) support; 1972 corresponding firmware-first mode error processing 1973 logic will be disabled. 1974 1975 hibernate= [HIBERNATION] 1976 noresume Don't check if there's a hibernation image 1977 present during boot. 1978 nocompress Don't compress/decompress hibernation images. 1979 no Disable hibernation and resume. 1980 protect_image Turn on image protection during restoration 1981 (that will set all pages holding image data 1982 during restoration read-only). 1983 1984 hibernate.compressor= [HIBERNATION] Compression algorithm to be 1985 used with hibernation. 1986 Format: { lzo | lz4 } 1987 Default: lzo 1988 1989 lzo: Select LZO compression algorithm to 1990 compress/decompress hibernation image. 1991 1992 lz4: Select LZ4 compression algorithm to 1993 compress/decompress hibernation image. 1994 1995 hibernate.pm_test_delay= 1996 [HIBERNATION] 1997 Sets the number of seconds to remain in a hibernation test 1998 mode before resuming the system (see 1999 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 2000 is set. Default value is 5. 2001 2002 hibernate_compression_threads= 2003 [HIBERNATION] 2004 Set the number of threads used for compressing or decompressing 2005 hibernation images. 2006 2007 Format: <integer> 2008 Default: 3 2009 Minimum: 1 2010 Example: hibernate_compression_threads=4 2011 2012 highmem=nn[KMG] [KNL,BOOT,EARLY] forces the highmem zone to have an exact 2013 size of <nn>. This works even on boxes that have no 2014 highmem otherwise. This also works to reduce highmem 2015 size on bigger boxes. 2016 2017 highres= [KNL] Enable/disable high resolution timer mode. 2018 Valid parameters: "on", "off" 2019 Default: "on" 2020 2021 hlt [BUGS=ARM,SH] 2022 2023 hostname= [KNL,EARLY] Set the hostname (aka UTS nodename). 2024 Format: <string> 2025 This allows setting the system's hostname during early 2026 startup. This sets the name returned by gethostname. 2027 Using this parameter to set the hostname makes it 2028 possible to ensure the hostname is correctly set before 2029 any userspace processes run, avoiding the possibility 2030 that a process may call gethostname before the hostname 2031 has been explicitly set, resulting in the calling 2032 process getting an incorrect result. The string must 2033 not exceed the maximum allowed hostname length (usually 2034 64 characters) and will be truncated otherwise. 2035 2036 hpet= [X86-32,HPET] option to control HPET usage 2037 Format: { enable (default) | disable | force | 2038 verbose } 2039 disable: disable HPET and use PIT instead 2040 force: allow force enabled of undocumented chips (ICH4, 2041 VIA, nVidia) 2042 verbose: show contents of HPET registers during setup 2043 2044 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET 2045 registers. Default set by CONFIG_HPET_MMAP_DEFAULT. 2046 2047 hugepages= [HW,EARLY] Number of HugeTLB pages to allocate at boot. 2048 If this follows hugepagesz (below), it specifies 2049 the number of pages of hugepagesz to be allocated. 2050 If this is the first HugeTLB parameter on the command 2051 line, it specifies the number of pages to allocate for 2052 the default huge page size. If using node format, the 2053 number of pages to allocate per-node can be specified. 2054 See also Documentation/admin-guide/mm/hugetlbpage.rst. 2055 Format: <integer> or (node format) 2056 <node>:<integer>[,<node>:<integer>] 2057 2058 hugepagesz= 2059 [HW,EARLY] The size of the HugeTLB pages. This is 2060 used in conjunction with hugepages (above) to 2061 allocate huge pages of a specific size at boot. The 2062 pair hugepagesz=X hugepages=Y can be specified once 2063 for each supported huge page size. Huge page sizes 2064 are architecture dependent. See also 2065 Documentation/admin-guide/mm/hugetlbpage.rst. 2066 Format: size[KMG] 2067 2068 hugepage_alloc_threads= 2069 [HW] The number of threads that should be used to 2070 allocate hugepages during boot. This option can be 2071 used to improve system bootup time when allocating 2072 a large amount of huge pages. 2073 The default value is 25% of the available hardware threads. 2074 2075 Note that this parameter only applies to non-gigantic huge pages. 2076 2077 hugetlb_cma= [HW,CMA,EARLY] The size of a CMA area used for allocation 2078 of gigantic hugepages. Or using node format, the size 2079 of a CMA area per node can be specified. 2080 Format: nn[KMGTPE] or (node format) 2081 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]] 2082 2083 The size must be a multiple of the gigantic page size. 2084 When using node format, this applies to each per-node size. 2085 Missaligned values are dropped with a warning. 2086 2087 Reserve a CMA area of given size and allocate gigantic 2088 hugepages using the CMA allocator. If enabled, the 2089 boot-time allocation of gigantic hugepages is skipped. 2090 2091 hugetlb_cma_only= 2092 [HW,CMA,EARLY] When allocating new HugeTLB pages, only 2093 try to allocate from the CMA areas. 2094 2095 This option does nothing if hugetlb_cma= is not also 2096 specified. 2097 2098 hugetlb_free_vmemmap= 2099 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP 2100 enabled. 2101 Control if HugeTLB Vmemmap Optimization (HVO) is enabled. 2102 Allows heavy hugetlb users to free up some more 2103 memory (7 * PAGE_SIZE for each 2MB hugetlb page). 2104 Format: { on | off (default) } 2105 2106 on: enable HVO 2107 off: disable HVO 2108 2109 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y, 2110 the default is on. 2111 2112 Note that the vmemmap pages may be allocated from the added 2113 memory block itself when memory_hotplug.memmap_on_memory is 2114 enabled, those vmemmap pages cannot be optimized even if this 2115 feature is enabled. Other vmemmap pages not allocated from 2116 the added memory block itself do not be affected. 2117 2118 hung_task_panic= 2119 [KNL] Number of hung tasks to trigger kernel panic. 2120 Format: <int> 2121 2122 When set to a non-zero value, a kernel panic will be triggered if 2123 the number of detected hung tasks reaches this value. 2124 2125 0: don't panic 2126 1: panic immediately on first hung task 2127 N: panic after N hung tasks are detected in a single scan 2128 2129 The default value is controlled by the 2130 CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time option. The value 2131 selected by this boot parameter can be changed later by the 2132 kernel.hung_task_panic sysctl. 2133 2134 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC) 2135 terminal devices. Valid values: 0..8 2136 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs. 2137 If specified, z/VM IUCV HVC accepts connections 2138 from listed z/VM user IDs only. 2139 2140 hv_nopvspin [X86,HYPER_V,EARLY] 2141 Disables the paravirt spinlock optimizations 2142 which allow the hypervisor to 'idle' the guest 2143 on lock contention. 2144 2145 hw_protection= [HW] 2146 Format: reboot | shutdown 2147 2148 Hardware protection action taken on critical events like 2149 overtemperature or imminent voltage loss. 2150 2151 i2c_bus= [HW] Override the default board specific I2C bus speed 2152 or register an additional I2C bus that is not 2153 registered from board initialization code. 2154 Format: 2155 <bus_id>,<clkrate> 2156 2157 i2c_touchscreen_props= [HW,ACPI,X86] 2158 Set device-properties for ACPI-enumerated I2C-attached 2159 touchscreen, to e.g. fix coordinates of upside-down 2160 mounted touchscreens. If you need this option please 2161 submit a drivers/platform/x86/touchscreen_dmi.c patch 2162 adding a DMI quirk for this. 2163 2164 Format: 2165 <ACPI_HW_ID>:<prop_name>=<val>[:prop_name=val][:...] 2166 Where <val> is one of: 2167 Omit "=<val>" entirely Set a boolean device-property 2168 Unsigned number Set a u32 device-property 2169 Anything else Set a string device-property 2170 2171 Examples (split over multiple lines): 2172 i2c_touchscreen_props=GDIX1001:touchscreen-inverted-x: 2173 touchscreen-inverted-y 2174 2175 i2c_touchscreen_props=MSSL1680:touchscreen-size-x=1920: 2176 touchscreen-size-y=1080:touchscreen-inverted-y: 2177 firmware-name=gsl1680-vendor-model.fw:silead,home-button 2178 2179 i8042.debug [HW] Toggle i8042 debug mode 2180 i8042.unmask_kbd_data 2181 [HW] Enable printing of interrupt data from the KBD port 2182 (disabled by default, and as a pre-condition 2183 requires that i8042.debug=1 be enabled) 2184 i8042.direct [HW] Put keyboard port into non-translated mode 2185 i8042.dumbkbd [HW] Pretend that controller can only read data from 2186 keyboard and cannot control its state 2187 (Don't attempt to blink the leds) 2188 i8042.noaux [HW] Don't check for auxiliary (== mouse) port 2189 i8042.nokbd [HW] Don't check/create keyboard port 2190 i8042.noloop [HW] Disable the AUX Loopback command while probing 2191 for the AUX port 2192 i8042.nomux [HW] Don't check presence of an active multiplexing 2193 controller 2194 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX 2195 controllers 2196 i8042.notimeout [HW] Ignore timeout condition signalled by controller 2197 i8042.reset [HW] Reset the controller during init, cleanup and 2198 suspend-to-ram transitions, only during s2r 2199 transitions, or never reset 2200 Format: { 1 | Y | y | 0 | N | n } 2201 1, Y, y: always reset controller 2202 0, N, n: don't ever reset controller 2203 Default: only on s2r transitions on x86; most other 2204 architectures force reset to be always executed 2205 i8042.unlock [HW] Unlock (ignore) the keylock 2206 i8042.kbdreset [HW] Reset device connected to KBD port 2207 i8042.probe_defer 2208 [HW] Allow deferred probing upon i8042 probe errors 2209 2210 i810= [HW,DRM] 2211 2212 i915.invert_brightness= 2213 [DRM] Invert the sense of the variable that is used to 2214 set the brightness of the panel backlight. Normally a 2215 brightness value of 0 indicates backlight switched off, 2216 and the maximum of the brightness value sets the backlight 2217 to maximum brightness. If this parameter is set to 0 2218 (default) and the machine requires it, or this parameter 2219 is set to 1, a brightness value of 0 sets the backlight 2220 to maximum brightness, and the maximum of the brightness 2221 value switches the backlight off. 2222 -1 -- never invert brightness 2223 0 -- machine default 2224 1 -- force brightness inversion 2225 2226 ia32_emulation= [X86-64] 2227 Format: <bool> 2228 When true, allows loading 32-bit programs and executing 32-bit 2229 syscalls, essentially overriding IA32_EMULATION_DEFAULT_DISABLED at 2230 boot time. When false, unconditionally disables IA32 emulation. 2231 2232 2233 idle= [X86,EARLY] 2234 Format: idle=poll, idle=halt, idle=nomwait 2235 2236 idle=poll: Don't do power saving in the idle loop 2237 using HLT, but poll for rescheduling event. This will 2238 make the CPUs eat a lot more power, but may be useful 2239 to get slightly better performance in multiprocessor 2240 benchmarks. It also makes some profiling using 2241 performance counters more accurate. Please note that 2242 on systems with MONITOR/MWAIT support (like Intel 2243 EM64T CPUs) this option has no performance advantage 2244 over the normal idle loop. It may also interact badly 2245 with hyperthreading. 2246 2247 idle=halt: Halt is forced to be used for CPU idle. 2248 In such case C2/C3 won't be used again. 2249 2250 idle=nomwait: Disable mwait for CPU C-states 2251 2252 [ARM64,EARLY] 2253 Format: idle=wfi, idle=yield, idle=nop 2254 2255 idle=wfi: Use the WFI (Wait For Interrupt) hint 2256 instruction in the idle loop. This is the default and 2257 allows the CPU to enter a low-power state until an 2258 interrupt arrives. 2259 2260 idle=yield: Use the YIELD hint instruction instead of 2261 WFI. CPUs supporting simultaneous multi-threading (SMT), 2262 can continue executing another thread when the current 2263 thread reaches the idle loop. This will make the CPUs 2264 eat more power, but may be useful to get slightly better 2265 performance in some applications, since the CPUs will 2266 not enter a low-power state. 2267 2268 idle=nop: Do not execute any idle instruction in the 2269 idle loop. This is useful on platforms where WFI 2270 misbehaves, leading to system instability or loss of CPU 2271 state. This will make the CPUs eat more power, but may 2272 give slightly better performance in some applications, 2273 since the CPUs will not enter a low-power state. 2274 2275 idxd.sva= [HW] 2276 Format: <bool> 2277 Allow force disabling of Shared Virtual Memory (SVA) 2278 support for the idxd driver. By default it is set to 2279 true (1). 2280 2281 idxd.tc_override= [HW] 2282 Format: <bool> 2283 Allow override of default traffic class configuration 2284 for the device. By default it is set to false (0). 2285 2286 ieee754= [MIPS] Select IEEE Std 754 conformance mode 2287 Format: { strict | legacy | 2008 | relaxed | emulated } 2288 Default: strict 2289 2290 Choose which programs will be accepted for execution 2291 based on the IEEE 754 NaN encoding(s) supported by 2292 the FPU and the NaN encoding requested with the value 2293 of an ELF file header flag individually set by each 2294 binary. Hardware implementations are permitted to 2295 support either or both of the legacy and the 2008 NaN 2296 encoding mode. 2297 2298 Available settings are as follows: 2299 strict accept binaries that request a NaN encoding 2300 supported by the FPU 2301 legacy only accept legacy-NaN binaries, if supported 2302 by the FPU 2303 2008 only accept 2008-NaN binaries, if supported 2304 by the FPU 2305 relaxed accept any binaries regardless of whether 2306 supported by the FPU 2307 emulated accept any binaries but enable FPU emulator 2308 if binary mode is unsupported by the FPU. 2309 2310 The FPU emulator is always able to support both NaN 2311 encodings, so if no FPU hardware is present or it has 2312 been disabled with 'nofpu', then the settings of 2313 'legacy' and '2008' strap the emulator accordingly, 2314 'relaxed' straps the emulator for both legacy-NaN and 2315 2008-NaN, whereas 'strict' enables legacy-NaN only on 2316 legacy processors and both NaN encodings on MIPS32 or 2317 MIPS64 CPUs. 2318 2319 The setting for ABS.fmt/NEG.fmt instruction execution 2320 mode generally follows that for the NaN encoding, 2321 except where unsupported by hardware. 2322 2323 ignore_loglevel [KNL,EARLY] 2324 Ignore loglevel setting - this will print /all/ 2325 kernel messages to the console. Useful for debugging. 2326 We also add it as printk module parameter, so users 2327 could change it dynamically, usually by 2328 /sys/module/printk/parameters/ignore_loglevel. 2329 2330 ignore_rlimit_data 2331 Ignore RLIMIT_DATA setting for data mappings, 2332 print warning at first misuse. Can be changed via 2333 /sys/module/kernel/parameters/ignore_rlimit_data. 2334 2335 ihash_entries= [KNL] 2336 Set number of hash buckets for inode cache. 2337 2338 ima_appraise= [IMA] appraise integrity measurements 2339 Format: { "off" | "enforce" | "fix" | "log" } 2340 default: "enforce" 2341 2342 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead. 2343 The builtin appraise policy appraises all files 2344 owned by uid=0. 2345 2346 ima_canonical_fmt [IMA] 2347 Use the canonical format for the binary runtime 2348 measurements, instead of host native format. 2349 2350 ima_flush_htable [IMA] 2351 Flush the IMA hash table when deleting all the 2352 staged measurement records, to achieve maximum 2353 memory saving at the cost of having duplicate 2354 records across the staged measurement lists. 2355 2356 ima_hash= [IMA] 2357 Format: { md5 | sha1 | rmd160 | sha256 | sha384 2358 | sha512 | ... } 2359 default: "sha1" 2360 2361 The list of supported hash algorithms is defined 2362 in crypto/hash_info.h. 2363 2364 ima_policy= [IMA] 2365 The builtin policies to load during IMA setup. 2366 Format: "tcb | appraise_tcb | secure_boot | 2367 fail_securely | critical_data" 2368 2369 The "tcb" policy measures all programs exec'd, files 2370 mmap'd for exec, and all files opened with the read 2371 mode bit set by either the effective uid (euid=0) or 2372 uid=0. 2373 2374 The "appraise_tcb" policy appraises the integrity of 2375 all files owned by root. 2376 2377 The "secure_boot" policy appraises the integrity 2378 of files (eg. kexec kernel image, kernel modules, 2379 firmware, policy, etc) based on file signatures. 2380 2381 The "fail_securely" policy forces file signature 2382 verification failure also on privileged mounted 2383 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE 2384 flag. 2385 2386 The "critical_data" policy measures kernel integrity 2387 critical data. 2388 2389 ima_tcb [IMA] Deprecated. Use ima_policy= instead. 2390 Load a policy which meets the needs of the Trusted 2391 Computing Base. This means IMA will measure all 2392 programs exec'd, files mmap'd for exec, and all files 2393 opened for read by uid=0. 2394 2395 ima_template= [IMA] 2396 Select one of defined IMA measurements template formats. 2397 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" | 2398 "ima-sigv2" } 2399 Default: "ima-ng" 2400 2401 ima_template_fmt= 2402 [IMA] Define a custom template format. 2403 Format: { "field1|...|fieldN" } 2404 2405 ima= [IMA] Enable or disable IMA 2406 Format: { "off" | "on" } 2407 Default: "on" 2408 Note that disabling IMA is limited to kdump kernel. 2409 2410 indirect_target_selection= [X86,Intel] Mitigation control for Indirect 2411 Target Selection(ITS) bug in Intel CPUs. Updated 2412 microcode is also required for a fix in IBPB. 2413 2414 on: Enable mitigation (default). 2415 off: Disable mitigation. 2416 force: Force the ITS bug and deploy default 2417 mitigation. 2418 vmexit: Only deploy mitigation if CPU is affected by 2419 guest/host isolation part of ITS. 2420 stuff: Deploy RSB-fill mitigation when retpoline is 2421 also deployed. Otherwise, deploy the default 2422 mitigation. 2423 2424 For details see: 2425 Documentation/admin-guide/hw-vuln/indirect-target-selection.rst 2426 2427 init= [KNL] 2428 Format: <full_path> 2429 Run specified binary instead of /sbin/init as init 2430 process. 2431 2432 initcall_debug [KNL] Trace initcalls as they are executed. Useful 2433 for working out where the kernel is dying during 2434 startup. 2435 2436 initcall_blacklist= [KNL] Do not execute a comma-separated list of 2437 initcall functions. Useful for debugging built-in 2438 modules and initcalls. 2439 2440 initramfs_async= [KNL] 2441 Format: <bool> 2442 Default: 1 2443 This parameter controls whether the initramfs 2444 image is unpacked asynchronously, concurrently 2445 with devices being probed and 2446 initialized. This should normally just work, 2447 but as a debugging aid, one can get the 2448 historical behaviour of the initramfs 2449 unpacking being completed before device_ and 2450 late_ initcalls. 2451 2452 initrd= [BOOT,EARLY] Specify the location of the initial ramdisk 2453 2454 initrdmem= [KNL,EARLY] Specify a physical address and size from which to 2455 load the initrd. If an initrd is compiled in or 2456 specified in the bootparams, it takes priority over this 2457 setting. 2458 Format: ss[KMG],nn[KMG] 2459 Default is 0, 0 2460 2461 init_on_alloc= [MM,EARLY] Fill newly allocated pages and heap objects with 2462 zeroes. 2463 Format: 0 | 1 2464 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON. 2465 2466 init_on_free= [MM,EARLY] Fill freed pages and heap objects with zeroes. 2467 Format: 0 | 1 2468 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON. 2469 2470 init_pkru= [X86] Specify the default memory protection keys rights 2471 register contents for all processes. 0x55555554 by 2472 default (disallow access to all but pkey 0). Can 2473 override in debugfs after boot. 2474 2475 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver 2476 Format: <irq> 2477 2478 int_pln_enable [X86] Enable power limit notification interrupt 2479 2480 integrity_audit=[IMA] 2481 Format: { "0" | "1" } 2482 0 -- basic integrity auditing messages. (Default) 2483 1 -- additional integrity auditing messages. 2484 2485 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option 2486 on 2487 Enable intel iommu driver. 2488 off 2489 Disable intel iommu driver. 2490 igfx_off [Default Off] 2491 By default, gfx is mapped as normal device. If a gfx 2492 device has a dedicated DMAR unit, the DMAR unit is 2493 bypassed by not enabling DMAR with this option. In 2494 this case, gfx device will use physical address for 2495 DMA. 2496 strict [Default Off] 2497 Deprecated, equivalent to iommu.strict=1. 2498 sp_off [Default Off] 2499 By default, super page will be supported if Intel IOMMU 2500 has the capability. With this option, super page will 2501 not be supported. 2502 sm_on 2503 Enable the Intel IOMMU scalable mode if the hardware 2504 advertises that it has support for the scalable mode 2505 translation. 2506 sm_off 2507 Disallow use of the Intel IOMMU scalable mode. 2508 tboot_noforce [Default Off] 2509 Do not force the Intel IOMMU enabled under tboot. 2510 By default, tboot will force Intel IOMMU on, which 2511 could harm performance of some high-throughput 2512 devices like 40GBit network cards, even if identity 2513 mapping is enabled. 2514 Note that using this option lowers the security 2515 provided by tboot because it makes the system 2516 vulnerable to DMA attacks. 2517 2518 intel_idle.max_cstate= [KNL,HW,ACPI,X86] 2519 0 disables intel_idle and fall back on acpi_idle. 2520 1 to 9 specify maximum depth of C-state. 2521 2522 intel_pstate= [X86,EARLY] 2523 disable 2524 Do not enable intel_pstate as the default 2525 scaling driver for the supported processors 2526 active 2527 Use intel_pstate driver to bypass the scaling 2528 governors layer of cpufreq and provides it own 2529 algorithms for p-state selection. There are two 2530 P-state selection algorithms provided by 2531 intel_pstate in the active mode: powersave and 2532 performance. The way they both operate depends 2533 on whether or not the hardware managed P-states 2534 (HWP) feature has been enabled in the processor 2535 and possibly on the processor model. 2536 passive 2537 Use intel_pstate as a scaling driver, but configure it 2538 to work with generic cpufreq governors (instead of 2539 enabling its internal governor). This mode cannot be 2540 used along with the hardware-managed P-states (HWP) 2541 feature. 2542 force 2543 Enable intel_pstate on systems that prohibit it by default 2544 in favor of acpi-cpufreq. Forcing the intel_pstate driver 2545 instead of acpi-cpufreq may disable platform features, such 2546 as thermal controls and power capping, that rely on ACPI 2547 P-States information being indicated to OSPM and therefore 2548 should be used with caution. This option does not work with 2549 processors that aren't supported by the intel_pstate driver 2550 or on platforms that use pcc-cpufreq instead of acpi-cpufreq. 2551 no_hwp 2552 Do not enable hardware P state control (HWP) 2553 if available. 2554 hwp_only 2555 Only load intel_pstate on systems which support 2556 hardware P state control (HWP) if available. 2557 support_acpi_ppc 2558 Enforce ACPI _PPC performance limits. If the Fixed ACPI 2559 Description Table, specifies preferred power management 2560 profile as "Enterprise Server" or "Performance Server", 2561 then this feature is turned on by default. 2562 per_cpu_perf_limits 2563 Allow per-logical-CPU P-State performance control limits using 2564 cpufreq sysfs interface 2565 no_cas 2566 Do not enable capacity-aware scheduling (CAS) on 2567 hybrid systems 2568 2569 intremap= [X86-64,Intel-IOMMU,EARLY] 2570 on enable Interrupt Remapping (default) 2571 off disable Interrupt Remapping 2572 nosid disable Source ID checking 2573 no_x2apic_optout 2574 BIOS x2APIC opt-out request will be ignored 2575 nopost disable Interrupt Posting 2576 posted_msi 2577 enable MSIs delivered as posted interrupts 2578 2579 iomem= Disable strict checking of access to MMIO memory 2580 strict regions from userspace. 2581 relaxed 2582 2583 iommu= [X86,EARLY] 2584 2585 off 2586 Don't initialize and use any kind of IOMMU. 2587 2588 force 2589 Force the use of the hardware IOMMU even when 2590 it is not actually needed (e.g. because < 3 GB 2591 memory). 2592 2593 noforce 2594 Don't force hardware IOMMU usage when it is not 2595 needed. (default). 2596 2597 merge 2598 Do scatter-gather (SG) merging. Implies "force" 2599 (experimental). 2600 2601 nomerge 2602 Don't do scatter-gather (SG) merging. 2603 2604 biomerge 2605 Do scatter-gather (SG) merging. Implies "force" 2606 (experimental). [same as "merge"] 2607 2608 panic 2609 Always panic when IOMMU overflows. 2610 2611 nopanic 2612 Don't panic on IOMMU overflows. 2613 2614 pt 2615 Use passththrough mode by default 2616 (Equivalent to iommu.passthrough=1) 2617 2618 nopt 2619 Use translated mode for DMA by default 2620 (Equivalent to iommu.passthrough=0) 2621 2622 soft 2623 Use software bounce buffering (SWIOTLB) (default for 2624 Intel machines). This can be used to prevent the usage 2625 of an available hardware IOMMU. 2626 2627 usedac 2628 Use the DAC on VIA PCI bridge 2629 (default: disable the VIA PCI bridge DAC) 2630 2631 AMD Gart HW IOMMU-specific options: (CONFIG_GART_IOMMU) 2632 2633 <size> 2634 Set the size of the remapping area in bytes. 2635 2636 allowed 2637 Overwrite iommu off workarounds for specific chipsets 2638 2639 force 2640 Overwrite iommu off workarounds for specific chipsets 2641 2642 fullflush 2643 Flush IOMMU on each allocation (default). 2644 2645 nofullflush 2646 Don't use IOMMU fullflush. 2647 2648 memaper[=<order>] 2649 Allocate an own aperture over RAM with size 2650 32MB<<order. (default: order=1, i.e. 64MB) 2651 2652 noaperture 2653 Ask the IOMMU not to touch the aperture for AGP. 2654 2655 noagp 2656 Don't initialize the AGP driver and use full aperture. 2657 2658 iommu= [PPC/POWERNV] 2659 2660 nobypass 2661 Disable IOMMU bypass, using IOMMU for PCI devices. 2662 2663 iommu.forcedac= [ARM64,X86,EARLY] Control IOVA allocation for PCI devices. 2664 Format: { "0" | "1" } 2665 0 - Try to allocate a 32-bit DMA address first, before 2666 falling back to the full range if needed. 2667 1 - Allocate directly from the full usable range, 2668 forcing Dual Address Cycle for PCI cards supporting 2669 greater than 32-bit addressing. 2670 2671 iommu.strict= [ARM64,X86,S390,EARLY] Configure TLB invalidation behaviour 2672 Format: { "0" | "1" } 2673 0 - Lazy mode. 2674 Request that DMA unmap operations use deferred 2675 invalidation of hardware TLBs, for increased 2676 throughput at the cost of reduced device isolation. 2677 Will fall back to strict mode if not supported by 2678 the relevant IOMMU driver. 2679 1 - Strict mode. 2680 DMA unmap operations invalidate IOMMU hardware TLBs 2681 synchronously. 2682 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}. 2683 Note: on x86, strict mode specified via one of the 2684 legacy driver-specific options takes precedence. 2685 2686 iommu.passthrough= 2687 [ARM64,X86,EARLY] Configure DMA to bypass the IOMMU by default. 2688 Format: { "0" | "1" } 2689 0 - Use IOMMU translation for DMA. 2690 1 - Bypass the IOMMU for DMA. 2691 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH. 2692 2693 iommu.debug_pagealloc= 2694 [KNL,EARLY] When CONFIG_IOMMU_DEBUG_PAGEALLOC is set, this 2695 parameter enables the feature at boot time. By default, it 2696 is disabled and the system behaves the same way as a kernel 2697 built without CONFIG_IOMMU_DEBUG_PAGEALLOC. 2698 Format: { "0" | "1" } 2699 0 - Sanitizer disabled. 2700 1 - Sanitizer enabled, expect runtime overhead. 2701 2702 io7= [HW] IO7 for Marvel-based Alpha systems 2703 See comment before marvel_specify_io7 in 2704 arch/alpha/kernel/core_marvel.c. 2705 2706 io_delay= [X86,EARLY] I/O delay method 2707 0x80 2708 Standard port 0x80 based delay 2709 0xed 2710 Alternate port 0xed based delay (needed on some systems) 2711 udelay 2712 Simple two microseconds delay 2713 none 2714 No delay 2715 2716 ip= [IP_PNP] 2717 See Documentation/admin-guide/nfs/nfsroot.rst. 2718 2719 ipcmni_extend [KNL,EARLY] Extend the maximum number of unique System V 2720 IPC identifiers from 32,768 to 16,777,216. 2721 2722 ipe.enforce= [IPE] 2723 Format: <bool> 2724 Determine whether IPE starts in permissive (0) or 2725 enforce (1) mode. The default is enforce. 2726 2727 ipe.success_audit= 2728 [IPE] 2729 Format: <bool> 2730 Start IPE with success auditing enabled, emitting 2731 an audit event when a binary is allowed. The default 2732 is 0. 2733 2734 irqaffinity= [SMP] Set the default irq affinity mask 2735 The argument is a cpu list, as described above. 2736 2737 irqchip.gicv2_force_probe= 2738 [ARM,ARM64,EARLY] 2739 Format: <bool> 2740 Force the kernel to look for the second 4kB page 2741 of a GICv2 controller even if the memory range 2742 exposed by the device tree is too small. 2743 2744 irqchip.gicv3_nolpi= 2745 [ARM,ARM64,EARLY] 2746 Force the kernel to ignore the availability of 2747 LPIs (and by consequence ITSs). Intended for system 2748 that use the kernel as a bootloader, and thus want 2749 to let secondary kernels in charge of setting up 2750 LPIs. 2751 2752 irqchip.gicv3_pseudo_nmi= [ARM64,EARLY] 2753 Enables support for pseudo-NMIs in the kernel. This 2754 requires the kernel to be built with 2755 CONFIG_ARM64_PSEUDO_NMI. 2756 2757 irqchip.riscv_imsic_noipi 2758 [RISC-V,EARLY] 2759 Force the kernel to not use IMSIC software injected MSIs 2760 as IPIs. Intended for system where IMSIC is trap-n-emulated, 2761 and thus want to reduce MMIO traps when triggering IPIs 2762 to multiple harts. 2763 2764 irqfixup [HW] 2765 When an interrupt is not handled search all handlers 2766 for it. Intended to get systems with badly broken 2767 firmware running. 2768 2769 irqhandler.duration_warn_us= [KNL] 2770 Warn if an IRQ handler exceeds the specified duration 2771 threshold in microseconds. Useful for identifying 2772 long-running IRQs in the system. 2773 2774 irqpoll [HW] 2775 When an interrupt is not handled search all handlers 2776 for it. Also check all handlers each timer 2777 interrupt. Intended to get systems with badly broken 2778 firmware running. 2779 2780 isapnp= [ISAPNP] 2781 Format: <RDP>,<reset>,<pci_scan>,<verbosity> 2782 2783 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance. 2784 Format: [flag-list,]<cpu-list> 2785 2786 Specify one or more CPUs to isolate from disturbances 2787 specified in the flag list (default: domain): 2788 2789 nohz 2790 Disable the tick when a single task runs as well as 2791 disabling other kernel noises like having RCU callbacks 2792 offloaded. This is equivalent to the nohz_full parameter. 2793 2794 A residual 1Hz tick is offloaded to workqueues, which you 2795 need to affine to housekeeping through the global 2796 workqueue's affinity configured via the 2797 /sys/devices/virtual/workqueue/cpumask sysfs file, or 2798 by using the 'domain' flag described below. 2799 2800 NOTE: by default the global workqueue runs on all CPUs, 2801 so to protect individual CPUs the 'cpumask' file has to 2802 be configured manually after bootup. 2803 2804 domain 2805 Isolate from the general SMP balancing and scheduling 2806 algorithms. Note that performing domain isolation this way 2807 is irreversible: it's not possible to bring back a CPU to 2808 the domains once isolated through this boot time 2809 configuration. Use cpusets for a dynamic configuration 2810 which can be altered at runtime. For details see 2811 Documentation/admin-guide/cpu-isolation.rst. 2812 2813 You can move a process onto or off an "isolated" CPU via 2814 the CPU affinity syscalls or cpuset. 2815 <cpu number> begins at 0 and the maximum value is 2816 "number of CPUs in system - 1". 2817 2818 managed_irq 2819 2820 Isolate from being targeted by managed interrupts 2821 which have an interrupt mask containing isolated 2822 CPUs. The affinity of managed interrupts is 2823 handled by the kernel and cannot be changed via 2824 the /proc/irq/* interfaces. 2825 2826 This isolation is best effort and only effective 2827 if the automatically assigned interrupt mask of a 2828 device queue contains isolated and housekeeping 2829 CPUs. If housekeeping CPUs are online then such 2830 interrupts are directed to the housekeeping CPU 2831 so that IO submitted on the housekeeping CPU 2832 cannot disturb the isolated CPU. 2833 2834 If a queue's affinity mask contains only isolated 2835 CPUs then this parameter has no effect on the 2836 interrupt routing decision, though interrupts are 2837 only delivered when tasks running on those 2838 isolated CPUs submit IO. IO submitted on 2839 housekeeping CPUs has no influence on those 2840 queues. 2841 2842 The format of <cpu-list> is described above. 2843 2844 iucv= [HW,NET] 2845 2846 ivrs_ioapic [HW,X86-64] 2847 Provide an override to the IOAPIC-ID<->DEVICE-ID 2848 mapping provided in the IVRS ACPI table. 2849 By default, PCI segment is 0, and can be omitted. 2850 2851 For example, to map IOAPIC-ID decimal 10 to 2852 PCI segment 0x1 and PCI device 00:14.0, 2853 write the parameter as: 2854 ivrs_ioapic=10@0001:00:14.0 2855 2856 Deprecated formats: 2857 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0 2858 write the parameter as: 2859 ivrs_ioapic[10]=00:14.0 2860 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and 2861 PCI device 00:14.0 write the parameter as: 2862 ivrs_ioapic[10]=0001:00:14.0 2863 2864 ivrs_hpet [HW,X86-64] 2865 Provide an override to the HPET-ID<->DEVICE-ID 2866 mapping provided in the IVRS ACPI table. 2867 By default, PCI segment is 0, and can be omitted. 2868 2869 For example, to map HPET-ID decimal 10 to 2870 PCI segment 0x1 and PCI device 00:14.0, 2871 write the parameter as: 2872 ivrs_hpet=10@0001:00:14.0 2873 2874 Deprecated formats: 2875 * To map HPET-ID decimal 0 to PCI device 00:14.0 2876 write the parameter as: 2877 ivrs_hpet[0]=00:14.0 2878 * To map HPET-ID decimal 10 to PCI segment 0x1 and 2879 PCI device 00:14.0 write the parameter as: 2880 ivrs_ioapic[10]=0001:00:14.0 2881 2882 ivrs_acpihid [HW,X86-64] 2883 Provide an override to the ACPI-HID:UID<->DEVICE-ID 2884 mapping provided in the IVRS ACPI table. 2885 By default, PCI segment is 0, and can be omitted. 2886 2887 For example, to map UART-HID:UID AMD0020:0 to 2888 PCI segment 0x1 and PCI device ID 00:14.5, 2889 write the parameter as: 2890 ivrs_acpihid=AMD0020:0@0001:00:14.5 2891 2892 Deprecated formats: 2893 * To map UART-HID:UID AMD0020:0 to PCI segment is 0, 2894 PCI device ID 00:14.5, write the parameter as: 2895 ivrs_acpihid[00:14.5]=AMD0020:0 2896 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and 2897 PCI device ID 00:14.5, write the parameter as: 2898 ivrs_acpihid[0001:00:14.5]=AMD0020:0 2899 2900 js= [HW,JOY] Analog joystick 2901 See Documentation/input/joydev/joystick.rst. 2902 2903 kasan_multi_shot 2904 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print 2905 report on every invalid memory access. Without this 2906 parameter KASAN will print report only for the first 2907 invalid access. 2908 2909 keep_bootcon [KNL,EARLY] 2910 Do not unregister boot console at start. This is only 2911 useful for debugging when something happens in the window 2912 between unregistering the boot console and initializing 2913 the real console. 2914 2915 keepinitrd [HW,ARM] See retain_initrd. 2916 2917 kernelcore= [KNL,X86,PPC,EARLY] 2918 Format: nn[KMGTPE] | nn% | "mirror" 2919 This parameter specifies the amount of memory usable by 2920 the kernel for non-movable allocations. The requested 2921 amount is spread evenly throughout all nodes in the 2922 system as ZONE_NORMAL. The remaining memory is used for 2923 movable memory in its own zone, ZONE_MOVABLE. In the 2924 event, a node is too small to have both ZONE_NORMAL and 2925 ZONE_MOVABLE, kernelcore memory will take priority and 2926 other nodes will have a larger ZONE_MOVABLE. 2927 2928 ZONE_MOVABLE is used for the allocation of pages that 2929 may be reclaimed or moved by the page migration 2930 subsystem. Note that allocations like PTEs-from-HighMem 2931 still use the HighMem zone if it exists, and the Normal 2932 zone if it does not. 2933 2934 It is possible to specify the exact amount of memory in 2935 the form of "nn[KMGTPE]", a percentage of total system 2936 memory in the form of "nn%", or "mirror". If "mirror" 2937 option is specified, mirrored (reliable) memory is used 2938 for non-movable allocations and remaining memory is used 2939 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror" 2940 are exclusive, so you cannot specify multiple forms. 2941 2942 kfence.burst= [MM,KFENCE] The number of additional successive 2943 allocations to be attempted through KFENCE for each 2944 sample interval. 2945 Format: <unsigned integer> 2946 Default: 0 2947 2948 kfence.check_on_panic= 2949 [MM,KFENCE] Whether to check all KFENCE-managed objects' 2950 canaries on panic. 2951 Format: <bool> 2952 Default: false 2953 2954 kfence.deferrable= 2955 [MM,KFENCE] Whether to use a deferrable timer to trigger 2956 allocations. This avoids forcing CPU wake-ups if the 2957 system is idle, at the risk of a less predictable 2958 sample interval. 2959 Format: <bool> 2960 Default: CONFIG_KFENCE_DEFERRABLE 2961 2962 kfence.fault= [MM,KFENCE] Controls the behavior when a KFENCE 2963 error is detected. 2964 report - print the error report and continue (default). 2965 oops - print the error report and oops. 2966 panic - print the error report and panic. 2967 2968 kfence.sample_interval= 2969 [MM,KFENCE] KFENCE's sample interval in milliseconds. 2970 Format: <unsigned integer> 2971 0 - Disable KFENCE. 2972 >0 - Enabled KFENCE with given sample interval. 2973 Default: CONFIG_KFENCE_SAMPLE_INTERVAL 2974 2975 kfence.skip_covered_thresh= 2976 [MM,KFENCE] If pool utilization reaches this threshold 2977 (pool usage%), KFENCE limits currently covered 2978 allocations of the same source from further filling 2979 up the pool. 2980 Format: <unsigned integer> 2981 Default: 75 2982 2983 kgdbdbgp= [KGDB,HW,EARLY] kgdb over EHCI usb debug port. 2984 Format: <Controller#>[,poll interval] 2985 The controller # is the number of the ehci usb debug 2986 port as it is probed via PCI. The poll interval is 2987 optional and is the number seconds in between 2988 each poll cycle to the debug port in case you need 2989 the functionality for interrupting the kernel with 2990 gdb or control-c on the dbgp connection. When 2991 not using this parameter you use sysrq-g to break into 2992 the kernel debugger. 2993 2994 kgdboc= [KGDB,HW] kgdb over consoles. 2995 Requires a tty driver that supports console polling, 2996 or a supported polling keyboard driver (non-usb). 2997 Serial only format: <serial_device>[,baud] 2998 keyboard only format: kbd 2999 keyboard and serial format: kbd,<serial_device>[,baud] 3000 Optional Kernel mode setting: 3001 kms, kbd format: kms,kbd 3002 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud] 3003 3004 kgdboc_earlycon= [KGDB,HW,EARLY] 3005 If the boot console provides the ability to read 3006 characters and can work in polling mode, you can use 3007 this parameter to tell kgdb to use it as a backend 3008 until the normal console is registered. Intended to 3009 be used together with the kgdboc parameter which 3010 specifies the normal console to transition to. 3011 3012 The name of the early console should be specified 3013 as the value of this parameter. Note that the name of 3014 the early console might be different than the tty 3015 name passed to kgdboc. It's OK to leave the value 3016 blank and the first boot console that implements 3017 read() will be picked. 3018 3019 kgdbwait [KGDB,EARLY] Stop kernel execution and enter the 3020 kernel debugger at the earliest opportunity. 3021 3022 kho= [KEXEC,EARLY] 3023 Format: { "0" | "1" | "off" | "on" | "y" | "n" } 3024 Enables or disables Kexec HandOver. 3025 "0" | "off" | "n" - kexec handover is disabled 3026 "1" | "on" | "y" - kexec handover is enabled 3027 3028 kho_scratch= [KEXEC,EARLY] 3029 Format: ll[KMG],mm[KMG],nn[KMG] | nn% 3030 Defines the size of the KHO scratch region. The KHO 3031 scratch regions are physically contiguous memory 3032 ranges that can only be used for non-kernel 3033 allocations. That way, even when memory is heavily 3034 fragmented with handed over memory, the kexeced 3035 kernel will always have enough contiguous ranges to 3036 bootstrap itself. 3037 3038 It is possible to specify the exact amount of 3039 memory in the form of "ll[KMG],mm[KMG],nn[KMG]" 3040 where the first parameter defines the size of a low 3041 memory scratch area, the second parameter defines 3042 the size of a global scratch area and the third 3043 parameter defines the size of additional per-node 3044 scratch areas. The form "nn%" defines scale factor 3045 (in percents) of memory that was used during boot. 3046 3047 kmac= [MIPS] Korina ethernet MAC address. 3048 Configure the RouterBoard 532 series on-chip 3049 Ethernet adapter MAC address. 3050 3051 kmemleak= [KNL,EARLY] Boot-time kmemleak enable/disable 3052 Valid arguments: on, off 3053 Default: on 3054 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y, 3055 the default is off. 3056 3057 kprobe_event=[probe-list] 3058 [FTRACE] Add kprobe events and enable at boot time. 3059 The probe-list is a semicolon delimited list of probe 3060 definitions. Each definition is same as kprobe_events 3061 interface, but the parameters are comma delimited. 3062 For example, to add a kprobe event on vfs_read with 3063 arg1 and arg2, add to the command line; 3064 3065 kprobe_event=p,vfs_read,$arg1,$arg2 3066 3067 See also Documentation/trace/kprobetrace.rst "Kernel 3068 Boot Parameter" section. 3069 3070 kpti= [ARM64,EARLY] Control page table isolation of 3071 user and kernel address spaces. 3072 Default: enabled on cores which need mitigation. 3073 0: force disabled 3074 1: force enabled 3075 3076 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires 3077 CONFIG_KUNIT to be set to be fully enabled. The 3078 default value can be overridden via 3079 KUNIT_DEFAULT_ENABLED. 3080 Default is 1 (enabled) 3081 3082 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs. 3083 Default is 0 (don't ignore, but inject #GP) 3084 3085 kvm.eager_page_split= 3086 [KVM,X86,RISCV] Controls whether or not KVM will try to 3087 proactively split all huge pages during dirty logging. 3088 Eager page splitting reduces interruptions to vCPU 3089 execution by eliminating the write-protection faults 3090 and MMU lock contention that would otherwise be 3091 required to split huge pages lazily. 3092 3093 VM workloads that rarely perform writes or that write 3094 only to a small region of VM memory may benefit from 3095 disabling eager page splitting to allow huge pages to 3096 still be used for reads. 3097 3098 The behavior of eager page splitting depends on whether 3099 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If 3100 disabled, all huge pages in a memslot will be eagerly 3101 split when dirty logging is enabled on that memslot. If 3102 enabled, eager page splitting will be performed during 3103 the KVM_CLEAR_DIRTY ioctl, and only for the pages being 3104 cleared. 3105 3106 On x86, eager page splitting is only supported when 3107 kvm.tdp_mmu=Y. 3108 3109 On RISCV, eager page splitting is supported by default. 3110 3111 Default is Y (on). 3112 3113 kvm.enable_pmu=[KVM,X86] 3114 If enabled, KVM will virtualize PMU functionality based 3115 on the virtual CPU model defined by userspace. This 3116 can be overridden on a per-VM basis via 3117 KVM_CAP_PMU_CAPABILITY. 3118 3119 If disabled, KVM will not virtualize PMU functionality, 3120 e.g. MSRs, PMCs, PMIs, etc., even if userspace defines 3121 a virtual CPU model that contains PMU assets. 3122 3123 Note, KVM's vPMU support implicitly requires running 3124 with an in-kernel local APIC, e.g. to deliver PMIs to 3125 the guest. Running without an in-kernel local APIC is 3126 not supported, though KVM will allow such a combination 3127 (with severely degraded functionality). 3128 3129 See also enable_mediated_pmu. 3130 3131 Default is Y (on). 3132 3133 kvm.enable_virt_at_load=[KVM,ARM64,LOONGARCH,MIPS,RISCV,X86] 3134 If enabled, KVM will enable virtualization in hardware 3135 when KVM is loaded, and disable virtualization when KVM 3136 is unloaded (if KVM is built as a module). 3137 3138 If disabled, KVM will dynamically enable and disable 3139 virtualization on-demand when creating and destroying 3140 VMs, i.e. on the 0=>1 and 1=>0 transitions of the 3141 number of VMs. 3142 3143 Enabling virtualization at module load avoids potential 3144 latency for creation of the 0=>1 VM, as KVM serializes 3145 virtualization enabling across all online CPUs. The 3146 "cost" of enabling virtualization when KVM is loaded, 3147 is that doing so may interfere with using out-of-tree 3148 hypervisors that want to "own" virtualization hardware. 3149 3150 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface. 3151 Default is false (don't support). 3152 3153 kvm.nx_huge_pages= 3154 [KVM] Controls the software workaround for the 3155 X86_BUG_ITLB_MULTIHIT bug. 3156 force : Always deploy workaround. 3157 off : Never deploy workaround. 3158 auto : Deploy workaround based on the presence of 3159 X86_BUG_ITLB_MULTIHIT. 3160 3161 Default is 'auto'. 3162 3163 If the software workaround is enabled for the host, 3164 guests do need not to enable it for nested guests. 3165 3166 kvm.nx_huge_pages_recovery_ratio= 3167 [KVM] Controls how many 4KiB pages are periodically zapped 3168 back to huge pages. 0 disables the recovery, otherwise if 3169 the value is N KVM will zap 1/Nth of the 4KiB pages every 3170 period (see below). The default is 60. 3171 3172 kvm.nx_huge_pages_recovery_period_ms= 3173 [KVM] Controls the time period at which KVM zaps 4KiB pages 3174 back to huge pages. If the value is a non-zero N, KVM will 3175 zap a portion (see ratio above) of the pages every N msecs. 3176 If the value is 0 (the default), KVM will pick a period based 3177 on the ratio, such that a page is zapped after 1 hour on average. 3178 3179 kvm-{amd,intel}.enable_mediated_pmu=[KVM,AMD,INTEL] 3180 If enabled, KVM will provide a mediated virtual PMU, 3181 instead of the default perf-based virtual PMU (if 3182 kvm.enable_pmu is true and PMU is enumerated via the 3183 virtual CPU model). 3184 3185 With a perf-based vPMU, KVM operates as a user of perf, 3186 i.e. emulates guest PMU counters using perf events. 3187 KVM-created perf events are managed by perf as regular 3188 (guest-only) events, e.g. are scheduled in/out, contend 3189 for hardware resources, etc. Using a perf-based vPMU 3190 allows guest and host usage of the PMU to co-exist, but 3191 incurs non-trivial overhead and can result in silently 3192 dropped guest events (due to resource contention). 3193 3194 With a mediated vPMU, hardware PMU state is context 3195 switched around the world switch to/from the guest. 3196 KVM mediates which events the guest can utilize, but 3197 gives the guest direct access to all other PMU assets 3198 when possible (KVM may intercept some accesses if the 3199 virtual CPU model provides a subset of hardware PMU 3200 functionality). Using a mediated vPMU significantly 3201 reduces PMU virtualization overhead and eliminates lost 3202 guest events, but is mutually exclusive with using perf 3203 to profile KVM guests and adds latency to most VM-Exits 3204 (to context switch PMU state). 3205 3206 Default is N (off). 3207 3208 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in 3209 KVM/SVM. Default is 1 (enabled). 3210 3211 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables, 3212 a.k.a. Two-Dimensional Page Tables. Default is 1 3213 (enabled). Disable by KVM if hardware lacks support 3214 for NPT. 3215 3216 kvm-amd.ciphertext_hiding_asids= 3217 [KVM,AMD] Ciphertext hiding prevents disallowed accesses 3218 to SNP private memory from reading ciphertext. Instead, 3219 reads will see constant default values (0xff). 3220 3221 If ciphertext hiding is enabled, the joint SEV-ES and 3222 SEV-SNP ASID space is partitioned into separate SEV-ES 3223 and SEV-SNP ASID ranges, with the SEV-SNP range being 3224 [1..max_snp_asid] and the SEV-ES range being 3225 (max_snp_asid..min_sev_asid), where min_sev_asid is 3226 enumerated by CPUID.0x.8000_001F[EDX]. 3227 3228 A non-zero value enables SEV-SNP ciphertext hiding and 3229 adjusts the ASID ranges for SEV-ES and SEV-SNP guests. 3230 KVM caps the number of SEV-SNP ASIDs at the maximum 3231 possible value, e.g. specifying -1u will assign all 3232 joint SEV-ES and SEV-SNP ASIDs to SEV-SNP. Note, 3233 assigning all joint ASIDs to SEV-SNP, i.e. configuring 3234 max_snp_asid == min_sev_asid-1, will effectively make 3235 SEV-ES unusable. 3236 3237 kvm-arm.mode= 3238 [KVM,ARM,EARLY] Select one of KVM/arm64's modes of 3239 operation. 3240 3241 none: Forcefully disable KVM. 3242 3243 nvhe: Standard nVHE-based mode, without support for 3244 protected guests. 3245 3246 protected: Mode with support for guests whose state is 3247 kept private from the host, using VHE or 3248 nVHE depending on HW support. 3249 3250 nested: VHE-based mode with support for nested 3251 virtualization. Requires at least ARMv8.4 3252 hardware (with FEAT_NV2). 3253 3254 Defaults to VHE/nVHE based on hardware support. Setting 3255 mode to "protected" will disable kexec and hibernation 3256 for the host. To force nVHE on VHE hardware, add 3257 "arm64_sw.hvhe=0 id_aa64mmfr1.vh=0" to the 3258 command-line. 3259 "nested" and "protected" are experimental and should be 3260 used with extreme caution. 3261 3262 kvm-arm.vgic_v3_group0_trap= 3263 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-0 3264 system registers 3265 3266 kvm-arm.vgic_v3_group1_trap= 3267 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-1 3268 system registers 3269 3270 kvm-arm.vgic_v3_common_trap= 3271 [KVM,ARM,EARLY] Trap guest accesses to GICv3 common 3272 system registers 3273 3274 kvm-arm.vgic_v4_enable= 3275 [KVM,ARM,EARLY] Allow use of GICv4 for direct 3276 injection of LPIs. 3277 3278 kvm-arm.wfe_trap_policy= 3279 [KVM,ARM] Control when to set WFE instruction trap for 3280 KVM VMs. Traps are allowed but not guaranteed by the 3281 CPU architecture. 3282 3283 trap: set WFE instruction trap 3284 3285 notrap: clear WFE instruction trap 3286 3287 kvm-arm.wfi_trap_policy= 3288 [KVM,ARM] Control when to set WFI instruction trap for 3289 KVM VMs. Traps are allowed but not guaranteed by the 3290 CPU architecture. 3291 3292 trap: set WFI instruction trap 3293 3294 notrap: clear WFI instruction trap 3295 3296 kvm_cma_resv_ratio=n [PPC,EARLY] 3297 Reserves given percentage from system memory area for 3298 contiguous memory allocation for KVM hash pagetable 3299 allocation. 3300 By default it reserves 5% of total system memory. 3301 Format: <integer> 3302 Default: 5 3303 3304 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables, 3305 a.k.a. Two-Dimensional Page Tables. Default is 1 3306 (enabled). Disable by KVM if hardware lacks support 3307 for EPT. 3308 3309 kvm-intel.emulate_invalid_guest_state= 3310 [KVM,Intel] Control whether to emulate invalid guest 3311 state. Ignored if kvm-intel.enable_unrestricted_guest=1, 3312 as guest state is never invalid for unrestricted 3313 guests. This param doesn't apply to nested guests (L2), 3314 as KVM never emulates invalid L2 guest state. 3315 Default is 1 (enabled). 3316 3317 kvm-intel.flexpriority= 3318 [KVM,Intel] Control KVM's use of FlexPriority feature 3319 (TPR shadow). Default is 1 (enabled). Disable by KVM if 3320 hardware lacks support for it. 3321 3322 kvm-intel.nested= 3323 [KVM,Intel] Control nested virtualization feature in 3324 KVM/VMX. Default is 1 (enabled). 3325 3326 kvm-intel.unrestricted_guest= 3327 [KVM,Intel] Control KVM's use of unrestricted guest 3328 feature (virtualized real and unpaged mode). Default 3329 is 1 (enabled). Disable by KVM if EPT is disabled or 3330 hardware lacks support for it. 3331 3332 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault 3333 CVE-2018-3620. 3334 3335 Valid arguments: never, cond, always 3336 3337 always: L1D cache flush on every VMENTER. 3338 cond: Flush L1D on VMENTER only when the code between 3339 VMEXIT and VMENTER can leak host memory. 3340 never: Disables the mitigation 3341 3342 Default is cond (do L1 cache flush in specific instances) 3343 3344 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor 3345 Identification feature (tagged TLBs). Default is 1 3346 (enabled). Disable by KVM if hardware lacks support 3347 for it. 3348 3349 l1d_flush= [X86,INTEL,EARLY] 3350 Control mitigation for L1D based snooping vulnerability. 3351 3352 Certain CPUs are vulnerable to an exploit against CPU 3353 internal buffers which can forward information to a 3354 disclosure gadget under certain conditions. 3355 3356 In vulnerable processors, the speculatively 3357 forwarded data can be used in a cache side channel 3358 attack, to access data to which the attacker does 3359 not have direct access. 3360 3361 This parameter controls the mitigation. The 3362 options are: 3363 3364 on - enable the interface for the mitigation 3365 3366 l1tf= [X86,EARLY] Control mitigation of the L1TF vulnerability on 3367 affected CPUs 3368 3369 The kernel PTE inversion protection is unconditionally 3370 enabled and cannot be disabled. 3371 3372 full 3373 Provides all available mitigations for the 3374 L1TF vulnerability. Disables SMT and 3375 enables all mitigations in the 3376 hypervisors, i.e. unconditional L1D flush. 3377 3378 SMT control and L1D flush control via the 3379 sysfs interface is still possible after 3380 boot. Hypervisors will issue a warning 3381 when the first VM is started in a 3382 potentially insecure configuration, 3383 i.e. SMT enabled or L1D flush disabled. 3384 3385 full,force 3386 Same as 'full', but disables SMT and L1D 3387 flush runtime control. Implies the 3388 'nosmt=force' command line option. 3389 (i.e. sysfs control of SMT is disabled.) 3390 3391 flush 3392 Leaves SMT enabled and enables the default 3393 hypervisor mitigation, i.e. conditional 3394 L1D flush. 3395 3396 SMT control and L1D flush control via the 3397 sysfs interface is still possible after 3398 boot. Hypervisors will issue a warning 3399 when the first VM is started in a 3400 potentially insecure configuration, 3401 i.e. SMT enabled or L1D flush disabled. 3402 3403 flush,nosmt 3404 3405 Disables SMT and enables the default 3406 hypervisor mitigation. 3407 3408 SMT control and L1D flush control via the 3409 sysfs interface is still possible after 3410 boot. Hypervisors will issue a warning 3411 when the first VM is started in a 3412 potentially insecure configuration, 3413 i.e. SMT enabled or L1D flush disabled. 3414 3415 flush,nowarn 3416 Same as 'flush', but hypervisors will not 3417 warn when a VM is started in a potentially 3418 insecure configuration. 3419 3420 off 3421 Disables hypervisor mitigations and doesn't 3422 emit any warnings. 3423 It also drops the swap size and available 3424 RAM limit restriction on both hypervisor and 3425 bare metal. 3426 3427 Default is 'flush'. 3428 3429 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst 3430 3431 l2cr= [PPC] 3432 3433 l3cr= [PPC] 3434 3435 lapic [X86-32,APIC,EARLY] Enable the local APIC even if BIOS 3436 disabled it. 3437 3438 lapic= [X86,APIC] Do not use TSC deadline 3439 value for LAPIC timer one-shot implementation. Default 3440 back to the programmable timer unit in the LAPIC. 3441 Format: notscdeadline 3442 3443 lapic_timer_c2_ok [X86,APIC,EARLY] trust the local apic timer 3444 in C2 power state. 3445 3446 libata.dma= [LIBATA] DMA control 3447 libata.dma=0 Disable all PATA and SATA DMA 3448 libata.dma=1 PATA and SATA Disk DMA only 3449 libata.dma=2 ATAPI (CDROM) DMA only 3450 libata.dma=4 Compact Flash DMA only 3451 Combinations also work, so libata.dma=3 enables DMA 3452 for disks and CDROMs, but not CFs. 3453 3454 libata.ignore_hpa= [LIBATA] Ignore HPA limit 3455 libata.ignore_hpa=0 keep BIOS limits (default) 3456 libata.ignore_hpa=1 ignore limits, using full disk 3457 3458 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume 3459 when set. 3460 Format: <int> 3461 3462 libata.force= [LIBATA] Force configurations. The format is a comma- 3463 separated list of "[ID:]VAL" where ID is PORT[.DEVICE]. 3464 PORT and DEVICE are decimal numbers matching port, link 3465 or device. Basically, it matches the ATA ID string 3466 printed on console by libata. If the whole ID part is 3467 omitted, the last PORT and DEVICE values are used. If 3468 ID hasn't been specified yet, the configuration applies 3469 to all ports, links and devices. 3470 3471 If only DEVICE is omitted, the parameter applies to 3472 the port and all links and devices behind it. DEVICE 3473 number of 0 either selects the first device or the 3474 first fan-out link behind PMP device. It does not 3475 select the host link. DEVICE number of 15 selects the 3476 host link and device attached to it. 3477 3478 The VAL specifies the configuration to force. As long 3479 as there is no ambiguity, shortcut notation is allowed. 3480 For example, both 1.5 and 1.5G would work for 1.5Gbps. 3481 The following configurations can be forced. 3482 3483 * Cable type: 40c, 80c, short40c, unk, ign or sata. 3484 Any ID with matching PORT is used. 3485 3486 * SATA link speed limit: 1.5Gbps or 3.0Gbps. 3487 3488 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7]. 3489 udma[/][16,25,33,44,66,100,133] notation is also 3490 allowed. 3491 3492 * nohrst, nosrst, norst: suppress hard, soft and both 3493 resets. 3494 3495 * rstonce: only attempt one reset during hot-unplug 3496 link recovery. 3497 3498 * [no]dbdelay: Enable or disable the extra 200ms delay 3499 before debouncing a link PHY and device presence 3500 detection. 3501 3502 * [no]ncq: Turn on or off NCQ. 3503 3504 * [no]ncqtrim: Enable or disable queued DSM TRIM. 3505 3506 * [no]ncqati: Enable or disable NCQ trim on ATI chipset. 3507 3508 * [no]trim: Enable or disable (unqueued) TRIM. 3509 3510 * trim_zero: Indicate that TRIM command zeroes data. 3511 3512 * max_trim_128m: Set 128M maximum trim size limit. 3513 3514 * [no]dma: Turn on or off DMA transfers. 3515 3516 * atapi_dmadir: Enable ATAPI DMADIR bridge support. 3517 3518 * atapi_mod16_dma: Enable the use of ATAPI DMA for 3519 commands that are not a multiple of 16 bytes. 3520 3521 * [no]dmalog: Enable or disable the use of the 3522 READ LOG DMA EXT command to access logs. 3523 3524 * [no]iddevlog: Enable or disable access to the 3525 identify device data log. 3526 3527 * [no]logdir: Enable or disable access to the general 3528 purpose log directory. 3529 3530 * max_sec=<sectors>: Set the transfer size limit, in 3531 number of 512-byte sectors, to the value specified in 3532 <sectors>. The value specified in <sectors> has to be 3533 a non-zero positive integer. 3534 3535 * max_sec_128: Set transfer size limit to 128 sectors. 3536 3537 * max_sec_1024: Set or clear transfer size limit to 3538 1024 sectors. 3539 3540 * max_sec_lba48: Set or clear transfer size limit to 3541 65535 sectors. 3542 3543 * external: Mark port as external (hotplug-capable). 3544 3545 * [no]lpm: Enable or disable link power management. 3546 3547 * [no]setxfer: Indicate if transfer speed mode setting 3548 should be skipped. 3549 3550 * [no]fua: Disable or enable FUA (Force Unit Access) 3551 support for devices supporting this feature. 3552 3553 * dump_id: Dump IDENTIFY data. 3554 3555 * disable: Disable this device. 3556 3557 If there are multiple matching configurations changing 3558 the same attribute, the last one is used. 3559 3560 liveupdate= [KNL,EARLY] 3561 Format: <bool> 3562 Enable Live Update Orchestrator (LUO). 3563 Default: off. 3564 3565 lockd.nlm_grace_period=P [NFS] Assign grace period. 3566 Format: <integer> 3567 3568 lockd.nlm_tcpport=N [NFS] Assign TCP port. 3569 Format: <integer> 3570 3571 lockd.nlm_timeout=T [NFS] Assign timeout value. 3572 Format: <integer> 3573 3574 lockd.nlm_udpport=M [NFS] Assign UDP port. 3575 Format: <integer> 3576 3577 lockdown= [SECURITY,EARLY] 3578 { integrity | confidentiality } 3579 Enable the kernel lockdown feature. If set to 3580 integrity, kernel features that allow userland to 3581 modify the running kernel are disabled. If set to 3582 confidentiality, kernel features that allow userland 3583 to extract confidential information from the kernel 3584 are also disabled. 3585 3586 locktorture.acq_writer_lim= [KNL] 3587 Set the time limit in jiffies for a lock 3588 acquisition. Acquisitions exceeding this limit 3589 will result in a splat once they do complete. 3590 3591 locktorture.bind_readers= [KNL] 3592 Specify the list of CPUs to which the readers are 3593 to be bound. 3594 3595 locktorture.bind_writers= [KNL] 3596 Specify the list of CPUs to which the writers are 3597 to be bound. 3598 3599 locktorture.call_rcu_chains= [KNL] 3600 Specify the number of self-propagating call_rcu() 3601 chains to set up. These are used to ensure that 3602 there is a high probability of an RCU grace period 3603 in progress at any given time. Defaults to 0, 3604 which disables these call_rcu() chains. 3605 3606 locktorture.long_hold= [KNL] 3607 Specify the duration in milliseconds for the 3608 occasional long-duration lock hold time. Defaults 3609 to 100 milliseconds. Select 0 to disable. 3610 3611 locktorture.nested_locks= [KNL] 3612 Specify the maximum lock nesting depth that 3613 locktorture is to exercise, up to a limit of 8 3614 (MAX_NESTED_LOCKS). Specify zero to disable. 3615 Note that this parameter is ineffective on types 3616 of locks that do not support nested acquisition. 3617 3618 locktorture.nreaders_stress= [KNL] 3619 Set the number of locking read-acquisition kthreads. 3620 Defaults to being automatically set based on the 3621 number of online CPUs. 3622 3623 locktorture.nwriters_stress= [KNL] 3624 Set the number of locking write-acquisition kthreads. 3625 3626 locktorture.onoff_holdoff= [KNL] 3627 Set time (s) after boot for CPU-hotplug testing. 3628 3629 locktorture.onoff_interval= [KNL] 3630 Set time (s) between CPU-hotplug operations, or 3631 zero to disable CPU-hotplug testing. 3632 3633 locktorture.rt_boost= [KNL] 3634 Do periodic testing of real-time lock priority 3635 boosting. Select 0 to disable, 1 to boost 3636 only rt_mutex, and 2 to boost unconditionally. 3637 Defaults to 2, which might seem to be an 3638 odd choice, but which should be harmless for 3639 non-real-time spinlocks, due to their disabling 3640 of preemption. Note that non-realtime mutexes 3641 disable boosting. 3642 3643 locktorture.rt_boost_factor= [KNL] 3644 Number that determines how often and for how 3645 long priority boosting is exercised. This is 3646 scaled down by the number of writers, so that the 3647 number of boosts per unit time remains roughly 3648 constant as the number of writers increases. 3649 On the other hand, the duration of each boost 3650 increases with the number of writers. 3651 3652 locktorture.shuffle_interval= [KNL] 3653 Set task-shuffle interval (jiffies). Shuffling 3654 tasks allows some CPUs to go into dyntick-idle 3655 mode during the locktorture test. 3656 3657 locktorture.shutdown_secs= [KNL] 3658 Set time (s) after boot system shutdown. This 3659 is useful for hands-off automated testing. 3660 3661 locktorture.stat_interval= [KNL] 3662 Time (s) between statistics printk()s. 3663 3664 locktorture.stutter= [KNL] 3665 Time (s) to stutter testing, for example, 3666 specifying five seconds causes the test to run for 3667 five seconds, wait for five seconds, and so on. 3668 This tests the locking primitive's ability to 3669 transition abruptly to and from idle. 3670 3671 locktorture.torture_type= [KNL] 3672 Specify the locking implementation to test. 3673 3674 locktorture.verbose= [KNL] 3675 Enable additional printk() statements. 3676 3677 locktorture.writer_fifo= [KNL] 3678 Run the write-side locktorture kthreads at 3679 sched_set_fifo() real-time priority. 3680 3681 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver 3682 Format: <irq> 3683 3684 loglevel= [KNL,EARLY] 3685 All Kernel Messages with a loglevel smaller than the 3686 console loglevel will be printed to the console. It can 3687 also be changed with klogd or other programs. The 3688 loglevels are defined as follows: 3689 3690 0 (KERN_EMERG) system is unusable 3691 1 (KERN_ALERT) action must be taken immediately 3692 2 (KERN_CRIT) critical conditions 3693 3 (KERN_ERR) error conditions 3694 4 (KERN_WARNING) warning conditions 3695 5 (KERN_NOTICE) normal but significant condition 3696 6 (KERN_INFO) informational 3697 7 (KERN_DEBUG) debug-level messages 3698 3699 log_buf_len=n[KMG] [KNL,EARLY] 3700 Sets the size of the printk ring buffer, in bytes. 3701 n must be a power of two and greater than the 3702 minimal size. The minimal size is defined by 3703 LOG_BUF_SHIFT kernel config parameter. There 3704 is also CONFIG_LOG_CPU_MAX_BUF_SHIFT config 3705 parameter that allows to increase the default size 3706 depending on the number of CPUs. See init/Kconfig 3707 for more details. 3708 3709 logo.nologo [FB] Disables display of the built-in Linux logo. 3710 This may be used to provide more screen space for 3711 kernel log messages and is useful when debugging 3712 kernel boot problems. 3713 3714 lp=0 [LP] Specify parallel ports to use, e.g, 3715 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses 3716 lp=reset first parallel port). 'lp=0' disables the 3717 lp=auto printer driver. 'lp=reset' (which can be 3718 specified in addition to the ports) causes 3719 attached printers to be reset. Using 3720 lp=port1,port2,... specifies the parallel ports 3721 to associate lp devices with, starting with 3722 lp0. A port specification may be 'none' to skip 3723 that lp device, or a parport name such as 3724 'parport0'. Specifying 'lp=auto' instead of a 3725 port specification list means that device IDs 3726 from each port should be examined, to see if 3727 an IEEE 1284-compliant printer is attached; if 3728 so, the driver will manage that printer. 3729 See also header of drivers/char/lp.c. 3730 3731 lpj=n [KNL] 3732 Sets loops_per_jiffy to given constant, thus avoiding 3733 time-consuming boot-time autodetection (up to 250 ms per 3734 CPU). 0 enables autodetection (default). To determine 3735 the correct value for your kernel, boot with normal 3736 autodetection and see what value is printed. Note that 3737 on SMP systems the preset will be applied to all CPUs, 3738 which is likely to cause problems if your CPUs need 3739 significantly divergent settings. An incorrect value 3740 will cause delays in the kernel to be wrong, leading to 3741 unpredictable I/O errors and other breakage. Although 3742 unlikely, in the extreme case this might damage your 3743 hardware. 3744 3745 lsm.debug [SECURITY] Enable LSM initialization debugging output. 3746 3747 lsm=lsm1,...,lsmN 3748 [SECURITY] Choose order of LSM initialization. This 3749 overrides CONFIG_LSM, and the "security=" parameter. 3750 3751 machtype= [Loongson] Share the same kernel image file between 3752 different yeeloong laptops. 3753 Example: machtype=lemote-yeeloong-2f-7inch 3754 3755 maxcpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 3756 will bring up during bootup. maxcpus=n : n >= 0 limits 3757 the kernel to bring up 'n' processors. Surely after 3758 bootup you can bring up the other plugged cpu by executing 3759 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus 3760 only takes effect during system bootup. 3761 While n=0 is a special case, it is equivalent to "nosmp", 3762 which also disables the IO APIC. 3763 3764 max_loop= [LOOP] The number of loop block devices that get 3765 (loop.max_loop) unconditionally pre-created at init time. The default 3766 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead 3767 of statically allocating a predefined number, loop 3768 devices can be requested on-demand with the 3769 /dev/loop-control interface. 3770 3771 mce= [X86-{32,64}] 3772 3773 Please see Documentation/arch/x86/x86_64/machinecheck.rst for sysfs runtime tunables. 3774 3775 off 3776 disable machine check 3777 3778 no_cmci 3779 disable CMCI(Corrected Machine Check Interrupt) that 3780 Intel processor supports. Usually this disablement is 3781 not recommended, but it might be handy if your 3782 hardware is misbehaving. 3783 3784 Note that you'll get more problems without CMCI than 3785 with due to the shared banks, i.e. you might get 3786 duplicated error logs. 3787 3788 dont_log_ce 3789 don't make logs for corrected errors. All events 3790 reported as corrected are silently cleared by OS. This 3791 option will be useful if you have no interest in any 3792 of corrected errors. 3793 3794 ignore_ce 3795 disable features for corrected errors, e.g. 3796 polling timer and CMCI. All events reported as 3797 corrected are not cleared by OS and remained in its 3798 error banks. 3799 3800 Usually this disablement is not recommended, however 3801 if there is an agent checking/clearing corrected 3802 errors (e.g. BIOS or hardware monitoring 3803 applications), conflicting with OS's error handling, 3804 and you cannot deactivate the agent, then this option 3805 will be a help. 3806 3807 no_lmce 3808 do not opt-in to Local MCE delivery. Use legacy method 3809 to broadcast MCEs. 3810 3811 bootlog 3812 enable logging of machine checks left over from 3813 booting. Disabled by default on AMD Fam10h and older 3814 because some BIOS leave bogus ones. 3815 3816 If your BIOS doesn't do that it's a good idea to 3817 enable though to make sure you log even machine check 3818 events that result in a reboot. On Intel systems it is 3819 enabled by default. 3820 3821 nobootlog 3822 disable boot machine check logging. 3823 3824 print_all 3825 print all machine check logs to the console. 3826 3827 monarchtimeout (number) 3828 sets the time in us to wait for other CPUs on machine 3829 checks. 0 to disable. 3830 3831 bios_cmci_threshold 3832 don't overwrite the bios-set CMCI threshold. This boot 3833 option prevents Linux from overwriting the CMCI 3834 threshold set by the bios. Without this option, Linux 3835 always sets the CMCI threshold to 1. Enabling this may 3836 make memory predictive failure analysis less effective 3837 if the bios sets thresholds for memory errors since we 3838 will not see details for all errors. 3839 3840 recovery 3841 force-enable recoverable machine check code paths 3842 3843 Everything else is in sysfs now. 3844 3845 3846 md= [HW] RAID subsystems devices and level 3847 See Documentation/admin-guide/md.rst. 3848 3849 mds= [X86,INTEL,EARLY] 3850 Control mitigation for the Micro-architectural Data 3851 Sampling (MDS) vulnerability. 3852 3853 Certain CPUs are vulnerable to an exploit against CPU 3854 internal buffers which can forward information to a 3855 disclosure gadget under certain conditions. 3856 3857 In vulnerable processors, the speculatively 3858 forwarded data can be used in a cache side channel 3859 attack, to access data to which the attacker does 3860 not have direct access. 3861 3862 This parameter controls the MDS mitigation. The 3863 options are: 3864 3865 full - Enable MDS mitigation on vulnerable CPUs 3866 full,nosmt - Enable MDS mitigation and disable 3867 SMT on vulnerable CPUs 3868 off - Unconditionally disable MDS mitigation 3869 3870 On TAA-affected machines, mds=off can be prevented by 3871 an active TAA mitigation as both vulnerabilities are 3872 mitigated with the same mechanism so in order to disable 3873 this mitigation, you need to specify tsx_async_abort=off 3874 too. 3875 3876 Not specifying this option is equivalent to 3877 mds=full. 3878 3879 For details see: Documentation/admin-guide/hw-vuln/mds.rst 3880 3881 mem=nn[KMG] [HEXAGON,EARLY] Set the memory size. 3882 Must be specified, otherwise memory size will be 0. 3883 3884 mem=nn[KMG] [KNL,BOOT,EARLY] Force usage of a specific amount 3885 of memory Amount of memory to be used in cases 3886 as follows: 3887 3888 1 for test; 3889 2 when the kernel is not able to see the whole system memory; 3890 3 memory that lies after 'mem=' boundary is excluded from 3891 the hypervisor, then assigned to KVM guests. 3892 4 to limit the memory available for kdump kernel. 3893 3894 [ARC,MICROBLAZE] - the limit applies only to low memory, 3895 high memory is not affected. 3896 3897 [ARM64] - only limits memory covered by the linear 3898 mapping. The NOMAP regions are not affected. 3899 3900 [X86] Work as limiting max address. Use together 3901 with memmap= to avoid physical address space collisions. 3902 Without memmap= PCI devices could be placed at addresses 3903 belonging to unused RAM. 3904 3905 Note that this only takes effects during boot time since 3906 in above case 3, memory may need be hot added after boot 3907 if system memory of hypervisor is not sufficient. 3908 3909 mem=nn[KMG]@ss[KMG] 3910 [ARM,MIPS,EARLY] - override the memory layout 3911 reported by firmware. 3912 Define a memory region of size nn[KMG] starting at 3913 ss[KMG]. 3914 Multiple different regions can be specified with 3915 multiple mem= parameters on the command line. 3916 3917 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel 3918 memory. 3919 3920 memblock=debug [KNL,EARLY] Enable memblock debug messages. 3921 3922 memchunk=nn[KMG] 3923 [KNL,SH] Allow user to override the default size for 3924 per-device physically contiguous DMA buffers. 3925 3926 memhp_default_state=online/offline/online_kernel/online_movable 3927 [KNL] Set the initial state for the memory hotplug 3928 onlining policy. If not specified, the default value is 3929 set according to the 3930 CONFIG_MHP_DEFAULT_ONLINE_TYPE kernel config 3931 options. 3932 See Documentation/admin-guide/mm/memory-hotplug.rst. 3933 3934 memmap=exactmap [KNL,X86,EARLY] Enable setting of an exact 3935 E820 memory map, as specified by the user. 3936 Such memmap=exactmap lines can be constructed based on 3937 BIOS output or other requirements. See the memmap=nn@ss 3938 option description. 3939 3940 memmap=nn[KMG]@ss[KMG] 3941 [KNL, X86,MIPS,XTENSA,EARLY] Force usage of a specific region of memory. 3942 Region of memory to be used is from ss to ss+nn. 3943 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG], 3944 which limits max address to nn[KMG]. 3945 Multiple different regions can be specified, 3946 comma delimited. 3947 Example: 3948 memmap=100M@2G,100M#3G,1G!1024G 3949 3950 memmap=nn[KMG]#ss[KMG] 3951 [KNL,ACPI,EARLY] Mark specific memory as ACPI data. 3952 Region of memory to be marked is from ss to ss+nn. 3953 3954 memmap=nn[KMG]$ss[KMG] 3955 [KNL,ACPI,EARLY] Mark specific memory as reserved. 3956 Region of memory to be reserved is from ss to ss+nn. 3957 Example: Exclude memory from 0x18690000-0x1869ffff 3958 memmap=64K$0x18690000 3959 or 3960 memmap=0x10000$0x18690000 3961 Some bootloaders may need an escape character before '$', 3962 like Grub2, otherwise '$' and the following number 3963 will be eaten. 3964 3965 memmap=nn[KMG]!ss[KMG,EARLY] 3966 [KNL,X86] Mark specific memory as protected. 3967 Region of memory to be used, from ss to ss+nn. 3968 The memory region may be marked as e820 type 12 (0xc) 3969 and is NVDIMM or ADR memory. 3970 3971 memmap=<size>%<offset>-<oldtype>+<newtype> 3972 [KNL,ACPI,EARLY] Convert memory within the specified region 3973 from <oldtype> to <newtype>. If "-<oldtype>" is left 3974 out, the whole region will be marked as <newtype>, 3975 even if previously unavailable. If "+<newtype>" is left 3976 out, matching memory will be removed. Types are 3977 specified as e820 types, e.g., 1 = RAM, 2 = reserved, 3978 3 = ACPI, 12 = PRAM. 3979 3980 memory_corruption_check=0/1 [X86,EARLY] 3981 Some BIOSes seem to corrupt the first 64k of 3982 memory when doing things like suspend/resume. 3983 Setting this option will scan the memory 3984 looking for corruption. Enabling this will 3985 both detect corruption and prevent the kernel 3986 from using the memory being corrupted. 3987 However, it's intended as a diagnostic tool; if 3988 repeatable BIOS-originated corruption always 3989 affects the same memory, you can use memmap= 3990 to prevent the kernel from using that memory. 3991 3992 memory_corruption_check_size=size [X86,EARLY] 3993 By default it checks for corruption in the low 3994 64k, making this memory unavailable for normal 3995 use. Use this parameter to scan for 3996 corruption in more or less memory. 3997 3998 memory_corruption_check_period=seconds [X86,EARLY] 3999 By default it checks for corruption every 60 4000 seconds. Use this parameter to check at some 4001 other rate. 0 disables periodic checking. 4002 4003 memory_hotplug.memmap_on_memory 4004 [KNL,X86,ARM] Boolean flag to enable this feature. 4005 Format: {on | off (default)} 4006 When enabled, runtime hotplugged memory will 4007 allocate its internal metadata (struct pages, 4008 those vmemmap pages cannot be optimized even 4009 if hugetlb_free_vmemmap is enabled) from the 4010 hotadded memory which will allow to hotadd a 4011 lot of memory without requiring additional 4012 memory to do so. 4013 This feature is disabled by default because it 4014 has some implication on large (e.g. GB) 4015 allocations in some configurations (e.g. small 4016 memory blocks). 4017 The state of the flag can be read in 4018 /sys/module/memory_hotplug/parameters/memmap_on_memory. 4019 Note that even when enabled, there are a few cases where 4020 the feature is not effective. 4021 4022 mempool_debug [MM] 4023 Enable mempool debugging. This enables element 4024 poison checking when freeing elements back to the 4025 pool. Useful for debugging mempool corruption. 4026 4027 memtest= [KNL,X86,ARM,M68K,PPC,RISCV,EARLY] Enable memtest 4028 Format: <integer> 4029 default : 0 <disable> 4030 Specifies the number of memtest passes to be 4031 performed. Each pass selects another test 4032 pattern from a given set of patterns. Memtest 4033 fills the memory with this pattern, validates 4034 memory contents and reserves bad memory 4035 regions that are detected. 4036 4037 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control 4038 Valid arguments: on, off 4039 Default: off 4040 mem_encrypt=on: Activate SME 4041 mem_encrypt=off: Do not activate SME 4042 4043 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst 4044 for details on when memory encryption can be activated. 4045 4046 mem_sleep_default= [SUSPEND] Default system suspend mode: 4047 s2idle - Suspend-To-Idle 4048 shallow - Power-On Suspend or equivalent (if supported) 4049 deep - Suspend-To-RAM or equivalent (if supported) 4050 See Documentation/admin-guide/pm/sleep-states.rst. 4051 4052 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when 4053 the BIOS has incorrectly applied a workaround. TinyBIOS 4054 version 0.98 is known to be affected, 0.99 fixes the 4055 problem by letting the user disable the workaround. 4056 4057 mga= [HW,DRM] 4058 4059 microcode= [X86] Control the behavior of the microcode loader. 4060 Available options, comma separated: 4061 4062 base_rev=X - with <X> with format: <u32> 4063 Set the base microcode revision of each thread when in 4064 debug mode. 4065 4066 dis_ucode_ldr: disable the microcode loader 4067 4068 force_minrev: 4069 Enable or disable the microcode minimal revision 4070 enforcement for the runtime microcode loader. 4071 4072 mini2440= [ARM,HW,KNL] 4073 Format:[0..2][b][c][t] 4074 Default: "0tb" 4075 MINI2440 configuration specification: 4076 0 - The attached screen is the 3.5" TFT 4077 1 - The attached screen is the 7" TFT 4078 2 - The VGA Shield is attached (1024x768) 4079 Leaving out the screen size parameter will not load 4080 the TFT driver, and the framebuffer will be left 4081 unconfigured. 4082 b - Enable backlight. The TFT backlight pin will be 4083 linked to the kernel VESA blanking code and a GPIO 4084 LED. This parameter is not necessary when using the 4085 VGA shield. 4086 c - Enable the s3c camera interface. 4087 t - Reserved for enabling touchscreen support. The 4088 touchscreen support is not enabled in the mainstream 4089 kernel as of 2.6.30, a preliminary port can be found 4090 in the "bleeding edge" mini2440 support kernel at 4091 https://repo.or.cz/w/linux-2.6/mini2440.git 4092 4093 mitigations= 4094 [X86,PPC,S390,ARM64,EARLY] Control optional mitigations for 4095 CPU vulnerabilities. This is a set of curated, 4096 arch-independent options, each of which is an 4097 aggregation of existing arch-specific options. 4098 4099 Note, "mitigations" is supported if and only if the 4100 kernel was built with CPU_MITIGATIONS=y. 4101 4102 off 4103 Disable all optional CPU mitigations. This 4104 improves system performance, but it may also 4105 expose users to several CPU vulnerabilities. 4106 Equivalent to: if nokaslr then kpti=0 [ARM64] 4107 gather_data_sampling=off [X86] 4108 indirect_target_selection=off [X86] 4109 kvm.nx_huge_pages=off [X86] 4110 l1tf=off [X86] 4111 mds=off [X86] 4112 mmio_stale_data=off [X86] 4113 no_entry_flush [PPC] 4114 no_uaccess_flush [PPC] 4115 nobp=0 [S390] 4116 nopti [X86,PPC] 4117 nospectre_bhb [ARM64] 4118 nospectre_v1 [X86,PPC] 4119 nospectre_v2 [X86,PPC,S390,ARM64] 4120 reg_file_data_sampling=off [X86] 4121 retbleed=off [X86] 4122 spec_rstack_overflow=off [X86] 4123 spec_store_bypass_disable=off [X86,PPC] 4124 spectre_bhi=off [X86] 4125 spectre_v2_user=off [X86] 4126 srbds=off [X86,INTEL] 4127 ssbd=force-off [ARM64] 4128 tsa=off [X86,AMD] 4129 tsx_async_abort=off [X86] 4130 vmscape=off [X86] 4131 4132 Exceptions: 4133 This does not have any effect on 4134 kvm.nx_huge_pages when 4135 kvm.nx_huge_pages=force. 4136 4137 auto (default) 4138 Mitigate all CPU vulnerabilities, but leave SMT 4139 enabled, even if it's vulnerable. This is for 4140 users who don't want to be surprised by SMT 4141 getting disabled across kernel upgrades, or who 4142 have other ways of avoiding SMT-based attacks. 4143 Equivalent to: (default behavior) 4144 4145 auto,nosmt 4146 Mitigate all CPU vulnerabilities, disabling SMT 4147 if needed. This is for users who always want to 4148 be fully mitigated, even if it means losing SMT. 4149 Equivalent to: l1tf=flush,nosmt [X86] 4150 mds=full,nosmt [X86] 4151 tsx_async_abort=full,nosmt [X86] 4152 mmio_stale_data=full,nosmt [X86] 4153 retbleed=auto,nosmt [X86] 4154 4155 [X86] After one of the above options, additionally 4156 supports attack-vector based controls as documented in 4157 Documentation/admin-guide/hw-vuln/attack_vector_controls.rst 4158 4159 mminit_loglevel= 4160 [KNL,EARLY] When CONFIG_DEBUG_MEMORY_INIT is set, this 4161 parameter allows control of the logging verbosity for 4162 the additional memory initialisation checks. A value 4163 of 0 disables mminit logging and a level of 4 will 4164 log everything. Information is printed at KERN_DEBUG 4165 so loglevel=8 may also need to be specified. 4166 4167 mmio_stale_data= 4168 [X86,INTEL,EARLY] Control mitigation for the Processor 4169 MMIO Stale Data vulnerabilities. 4170 4171 Processor MMIO Stale Data is a class of 4172 vulnerabilities that may expose data after an MMIO 4173 operation. Exposed data could originate or end in 4174 the same CPU buffers as affected by MDS and TAA. 4175 Therefore, similar to MDS and TAA, the mitigation 4176 is to clear the affected CPU buffers. 4177 4178 This parameter controls the mitigation. The 4179 options are: 4180 4181 full - Enable mitigation on vulnerable CPUs 4182 4183 full,nosmt - Enable mitigation and disable SMT on 4184 vulnerable CPUs. 4185 4186 off - Unconditionally disable mitigation 4187 4188 On MDS or TAA affected machines, 4189 mmio_stale_data=off can be prevented by an active 4190 MDS or TAA mitigation as these vulnerabilities are 4191 mitigated with the same mechanism so in order to 4192 disable this mitigation, you need to specify 4193 mds=off and tsx_async_abort=off too. 4194 4195 Not specifying this option is equivalent to 4196 mmio_stale_data=full. 4197 4198 For details see: 4199 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst 4200 4201 <module>.async_probe[=<bool>] [KNL] 4202 If no <bool> value is specified or if the value 4203 specified is not a valid <bool>, enable asynchronous 4204 probe on this module. Otherwise, enable/disable 4205 asynchronous probe on this module as indicated by the 4206 <bool> value. See also: module.async_probe 4207 4208 module.async_probe=<bool> 4209 [KNL] When set to true, modules will use async probing 4210 by default. To enable/disable async probing for a 4211 specific module, use the module specific control that 4212 is documented under <module>.async_probe. When both 4213 module.async_probe and <module>.async_probe are 4214 specified, <module>.async_probe takes precedence for 4215 the specific module. 4216 4217 module.enable_dups_trace 4218 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set, 4219 this means that duplicate request_module() calls will 4220 trigger a WARN_ON() instead of a pr_warn(). Note that 4221 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s 4222 will always be issued and this option does nothing. 4223 module.sig_enforce 4224 [KNL] When CONFIG_MODULE_SIG is set, this means that 4225 modules without (valid) signatures will fail to load. 4226 Note that if CONFIG_MODULE_SIG_FORCE is set, that 4227 is always true, so this option does nothing. 4228 4229 module_blacklist= [KNL] Do not load a comma-separated list of 4230 modules. Useful for debugging problem modules. 4231 4232 mousedev.tap_time= 4233 [MOUSE] Maximum time between finger touching and 4234 leaving touchpad surface for touch to be considered 4235 a tap and be reported as a left button click (for 4236 touchpads working in absolute mode only). 4237 Format: <msecs> 4238 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices 4239 reporting absolute coordinates, such as tablets 4240 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices 4241 reporting absolute coordinates, such as tablets 4242 4243 movablecore= [KNL,X86,PPC,EARLY] 4244 Format: nn[KMGTPE] | nn% 4245 This parameter is the complement to kernelcore=, it 4246 specifies the amount of memory used for migratable 4247 allocations. If both kernelcore and movablecore is 4248 specified, then kernelcore will be at *least* the 4249 specified value but may be more. If movablecore on its 4250 own is specified, the administrator must be careful 4251 that the amount of memory usable for all allocations 4252 is not too small. 4253 4254 movable_node [KNL,EARLY] Boot-time switch to make hotplugable memory 4255 NUMA nodes to be movable. This means that the memory 4256 of such nodes will be usable only for movable 4257 allocations which rules out almost all kernel 4258 allocations. Use with caution! 4259 4260 MTD_Partition= [MTD] 4261 Format: <name>,<region-number>,<size>,<offset> 4262 4263 MTD_Region= [MTD] Format: 4264 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>] 4265 4266 mtdparts= [MTD] 4267 See drivers/mtd/parsers/cmdlinepart.c 4268 4269 mtouchusb.raw_coordinates= 4270 [HW] Make the MicroTouch USB driver use raw coordinates 4271 ('y', default) or cooked coordinates ('n') 4272 4273 mtrr=debug [X86,EARLY] 4274 Enable printing debug information related to MTRR 4275 registers at boot time. 4276 4277 mtrr_chunk_size=nn[KMG,X86,EARLY] 4278 used for mtrr cleanup. It is largest continuous chunk 4279 that could hold holes aka. UC entries. 4280 4281 mtrr_gran_size=nn[KMG,X86,EARLY] 4282 Used for mtrr cleanup. It is granularity of mtrr block. 4283 Default is 1. 4284 Large value could prevent small alignment from 4285 using up MTRRs. 4286 4287 mtrr_spare_reg_nr=n [X86,EARLY] 4288 Format: <integer> 4289 Range: 0,7 : spare reg number 4290 Default : 1 4291 Used for mtrr cleanup. It is spare mtrr entries number. 4292 Set to 2 or more if your graphical card needs more. 4293 4294 multitce=off [PPC] This parameter disables the use of the pSeries 4295 firmware feature for updating multiple TCE entries 4296 at a time. 4297 4298 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card 4299 4300 netdev= [NET] NE2000 ISA network devices parameters 4301 Format: <irq>,<io>,<mem_start>,<mem_end>,<name> 4302 4303 netpoll.carrier_timeout= 4304 [NET] Specifies amount of time (in seconds) that 4305 netpoll should wait for a carrier. By default netpoll 4306 waits 4 seconds. 4307 4308 nf_conntrack.acct= 4309 [NETFILTER] Enable connection tracking flow accounting 4310 0 to disable accounting 4311 1 to enable accounting 4312 Default value is 0. 4313 4314 nfs.cache_getent= 4315 [NFS] sets the pathname to the program which is used 4316 to update the NFS client cache entries. 4317 4318 nfs.cache_getent_timeout= 4319 [NFS] sets the timeout after which an attempt to 4320 update a cache entry is deemed to have failed. 4321 4322 nfs.callback_nr_threads= 4323 [NFSv4] set the total number of threads that the 4324 NFS client will assign to service NFSv4 callback 4325 requests. 4326 4327 nfs.callback_tcpport= 4328 [NFS] set the TCP port on which the NFSv4 callback 4329 channel should listen. 4330 4331 nfs.delay_retrans= 4332 [NFS] specifies the number of times the NFSv4 client 4333 retries the request before returning an EAGAIN error, 4334 after a reply of NFS4ERR_DELAY from the server. 4335 Only applies if the softerr mount option is enabled, 4336 and the specified value is >= 0. 4337 4338 nfs.idmap_cache_timeout= 4339 [NFS] set the maximum lifetime for idmapper cache 4340 entries. 4341 4342 nfs.max_session_cb_slots= 4343 [NFSv4.1] Sets the maximum number of session 4344 slots the client will assign to the callback 4345 channel. This determines the maximum number of 4346 callbacks the client will process in parallel for 4347 a particular server. 4348 4349 nfs.max_session_slots= 4350 [NFSv4.1] Sets the maximum number of session slots 4351 the client will attempt to negotiate with the server. 4352 This limits the number of simultaneous RPC requests 4353 that the client can send to the NFSv4.1 server. 4354 Note that there is little point in setting this 4355 value higher than the max_tcp_slot_table_limit. 4356 4357 nfs.nfs4_disable_idmapping= 4358 [NFSv4] When set to the default of '1', this option 4359 ensures that both the RPC level authentication 4360 scheme and the NFS level operations agree to use 4361 numeric uids/gids if the mount is using the 4362 'sec=sys' security flavour. In effect it is 4363 disabling idmapping, which can make migration from 4364 legacy NFSv2/v3 systems to NFSv4 easier. 4365 Servers that do not support this mode of operation 4366 will be autodetected by the client, and it will fall 4367 back to using the idmapper. 4368 To turn off this behaviour, set the value to '0'. 4369 4370 nfs.nfs4_unique_id= 4371 [NFS4] Specify an additional fixed unique ident- 4372 ification string that NFSv4 clients can insert into 4373 their nfs_client_id4 string. This is typically a 4374 UUID that is generated at system install time. 4375 4376 nfs.recover_lost_locks= 4377 [NFSv4] Attempt to recover locks that were lost due 4378 to a lease timeout on the server. Please note that 4379 doing this risks data corruption, since there are 4380 no guarantees that the file will remain unchanged 4381 after the locks are lost. 4382 If you want to enable the kernel legacy behaviour of 4383 attempting to recover these locks, then set this 4384 parameter to '1'. 4385 The default parameter value of '0' causes the kernel 4386 not to attempt recovery of lost locks. 4387 4388 nfs.send_implementation_id= 4389 [NFSv4.1] Send client implementation identification 4390 information in exchange_id requests. 4391 If zero, no implementation identification information 4392 will be sent. 4393 The default is to send the implementation identification 4394 information. 4395 4396 nfs4.layoutstats_timer= 4397 [NFSv4.2] Change the rate at which the kernel sends 4398 layoutstats to the pNFS metadata server. 4399 4400 Setting this to value to 0 causes the kernel to use 4401 whatever value is the default set by the layout 4402 driver. A non-zero value sets the minimum interval 4403 in seconds between layoutstats transmissions. 4404 4405 nfsd.inter_copy_offload_enable= 4406 [NFSv4.2] When set to 1, the server will support 4407 server-to-server copies for which this server is 4408 the destination of the copy. 4409 4410 nfsd.nfs4_disable_idmapping= 4411 [NFSv4] When set to the default of '1', the NFSv4 4412 server will return only numeric uids and gids to 4413 clients using auth_sys, and will accept numeric uids 4414 and gids from such clients. This is intended to ease 4415 migration from NFSv2/v3. 4416 4417 nfsd.nfsd4_ssc_umount_timeout= 4418 [NFSv4.2] When used as the destination of a 4419 server-to-server copy, knfsd temporarily mounts 4420 the source server. It caches the mount in case 4421 it will be needed again, and discards it if not 4422 used for the number of milliseconds specified by 4423 this parameter. 4424 4425 nfsaddrs= [NFS] Deprecated. Use ip= instead. 4426 See Documentation/admin-guide/nfs/nfsroot.rst. 4427 4428 nfsroot= [NFS] nfs root filesystem for disk-less boxes. 4429 See Documentation/admin-guide/nfs/nfsroot.rst. 4430 4431 nfsrootdebug [NFS] enable nfsroot debugging messages. 4432 See Documentation/admin-guide/nfs/nfsroot.rst. 4433 4434 nmi_backtrace.backtrace_idle [KNL] 4435 Dump stacks even of idle CPUs in response to an 4436 NMI stack-backtrace request. 4437 4438 nmi_debug= [KNL,SH] Specify one or more actions to take 4439 when a NMI is triggered. 4440 Format: [state][,regs][,debounce][,die] 4441 4442 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels 4443 Format: [panic,][nopanic,][rNNN,][num] 4444 Valid num: 0 or 1 4445 0 - turn hardlockup detector in nmi_watchdog off 4446 1 - turn hardlockup detector in nmi_watchdog on 4447 rNNN - configure the watchdog with raw perf event 0xNNN 4448 4449 When panic is specified, panic when an NMI watchdog 4450 timeout occurs (or 'nopanic' to not panic on an NMI 4451 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set) 4452 To disable both hard and soft lockup detectors, 4453 please see 'nowatchdog'. 4454 This is useful when you use a panic=... timeout and 4455 need the box quickly up again. 4456 4457 These settings can be accessed at runtime via 4458 the nmi_watchdog and hardlockup_panic sysctls. 4459 4460 no4lvl [RISCV,EARLY] Disable 4-level and 5-level paging modes. 4461 Forces kernel to use 3-level paging instead. 4462 4463 no5lvl [X86-64,RISCV,EARLY] Disable 5-level paging mode. Forces 4464 kernel to use 4-level paging instead. 4465 4466 noalign [KNL,ARM] 4467 4468 noapic [SMP,APIC,EARLY] Tells the kernel to not make use of any 4469 IOAPICs that may be present in the system. 4470 4471 noapictimer [APIC,X86] Don't set up the APIC timer 4472 4473 noautogroup Disable scheduler automatic task group creation. 4474 4475 nocache [ARM,EARLY] 4476 4477 no_console_suspend 4478 [HW] Never suspend the console 4479 Disable suspending of consoles during suspend and 4480 hibernate operations. Once disabled, debugging 4481 messages can reach various consoles while the rest 4482 of the system is being put to sleep (ie, while 4483 debugging driver suspend/resume hooks). This may 4484 not work reliably with all consoles, but is known 4485 to work with serial and VGA consoles. 4486 To facilitate more flexible debugging, we also add 4487 console_suspend, a printk module parameter to control 4488 it. Users could use console_suspend (usually 4489 /sys/module/printk/parameters/console_suspend) to 4490 turn on/off it dynamically. 4491 4492 no_debug_objects 4493 [KNL,EARLY] Disable object debugging 4494 4495 nodsp [SH] Disable hardware DSP at boot time. 4496 4497 noefi [EFI,EARLY] Disable EFI runtime services support. 4498 4499 no_entry_flush [PPC,EARLY] Don't flush the L1-D cache when entering the kernel. 4500 4501 noexec32 [X86-64] 4502 This affects only 32-bit executables. 4503 noexec32=on: enable non-executable mappings (default) 4504 read doesn't imply executable mappings 4505 noexec32=off: disable non-executable mappings 4506 read implies executable mappings 4507 4508 no_file_caps Tells the kernel not to honor file capabilities. The 4509 only way then for a file to be executed with privilege 4510 is to be setuid root or executed by root. 4511 4512 nofpu [MIPS,SH] Disable hardware FPU at boot time. 4513 4514 nofsgsbase [X86] Disables FSGSBASE instructions. 4515 4516 nofxsr [BUGS=X86-32] Disables x86 floating point extended 4517 register save and restore. The kernel will only save 4518 legacy floating-point registers on task switch. 4519 4520 nogbpages [X86] Do not use GB pages for kernel direct mappings. 4521 4522 no_hash_pointers 4523 [KNL,EARLY] 4524 Alias for "hash_pointers=never". 4525 4526 nohibernate [HIBERNATION] Disable hibernation and resume. 4527 4528 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,RISCV,SH] Forces the kernel to 4529 busy wait in do_idle() and not use the arch_cpu_idle() 4530 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP 4531 to be effective. This is useful on platforms where the 4532 sleep(SH) or wfi(ARM,ARM64) instructions do not work 4533 correctly or when doing power measurements to evaluate 4534 the impact of the sleep instructions. This is also 4535 useful when using JTAG debugger. 4536 4537 nohpet [X86] Don't use the HPET timer. 4538 4539 nohugeiomap [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge I/O mappings. 4540 4541 nohugevmalloc [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge vmalloc mappings. 4542 4543 nohz= [KNL] Boottime enable/disable dynamic ticks 4544 Valid arguments: on, off 4545 Default: on 4546 4547 nohz_full= [KNL,BOOT,SMP,ISOL] 4548 The argument is a cpu list, as described above. 4549 In kernels built with CONFIG_NO_HZ_FULL=y, set 4550 the specified list of CPUs whose tick will be stopped 4551 whenever possible. The boot CPU will be forced outside 4552 the range to maintain the timekeeping. Any CPUs 4553 in this list will have their RCU callbacks offloaded, 4554 just as if they had also been called out in the 4555 rcu_nocbs= boot parameter. 4556 4557 Note that this argument takes precedence over 4558 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 4559 4560 noinitrd [Deprecated,RAM] Tells the kernel not to load any configured 4561 initial RAM disk. Currently this parameter applies to 4562 initrd only, not to initramfs. But it applies to both 4563 in EFI mode. 4564 4565 nointremap [X86-64,Intel-IOMMU,EARLY] Do not enable interrupt 4566 remapping. 4567 [Deprecated - use intremap=off] 4568 4569 noinvpcid [X86,EARLY] Disable the INVPCID cpu feature. 4570 4571 noiotrap [SH] Disables trapped I/O port accesses. 4572 4573 noirqdebug [X86-32] Disables the code which attempts to detect and 4574 disable unhandled interrupt sources. 4575 4576 noisapnp [ISAPNP] Disables ISA PnP code. 4577 4578 nokaslr [KNL,EARLY] 4579 When CONFIG_RANDOMIZE_BASE is set, this disables 4580 kernel and module base offset ASLR (Address Space 4581 Layout Randomization). 4582 4583 no-kvmapf [X86,KVM,EARLY] Disable paravirtualized asynchronous page 4584 fault handling. 4585 4586 no-kvmclock [X86,KVM,EARLY] Disable paravirtualized KVM clock driver 4587 4588 nolapic [X86-32,APIC,EARLY] Do not enable or use the local APIC. 4589 4590 nolapic_timer [X86-32,APIC,EARLY] Do not use the local APIC timer. 4591 4592 nomce [X86-32] Disable Machine Check Exception 4593 4594 nomfgpt [X86-32] Disable Multi-Function General Purpose 4595 Timer usage (for AMD Geode machines). 4596 4597 nomodeset Disable kernel modesetting. Most systems' firmware 4598 sets up a display mode and provides framebuffer memory 4599 for output. With nomodeset, DRM and fbdev drivers will 4600 not load if they could possibly displace the pre- 4601 initialized output. Only the system framebuffer will 4602 be available for use. The respective drivers will not 4603 perform display-mode changes or accelerated rendering. 4604 4605 Useful as error fallback, or for testing and debugging. 4606 4607 nomodule Disable module load 4608 4609 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to 4610 shutdown the other cpus. Instead use the REBOOT_VECTOR 4611 irq. 4612 4613 nopat [X86,EARLY] Disable PAT (page attribute table extension of 4614 pagetables) support. 4615 4616 nopcid [X86-64,EARLY] Disable the PCID cpu feature. 4617 4618 nopku [X86] Disable Memory Protection Keys CPU feature found 4619 in some Intel CPUs. 4620 4621 nopti [X86-64,EARLY] 4622 Equivalent to pti=off 4623 4624 nopv= [X86,XEN,KVM,HYPER_V,VMWARE,EARLY] 4625 Disables the PV optimizations forcing the guest to run 4626 as generic guest with no PV drivers. Currently support 4627 XEN HVM, KVM, HYPER_V and VMWARE guest. 4628 4629 nopvspin [X86,XEN,KVM,EARLY] 4630 Disables the qspinlock slow path using PV optimizations 4631 which allow the hypervisor to 'idle' the guest on lock 4632 contention. 4633 4634 norandmaps Don't use address space randomization. Equivalent to 4635 echo 0 > /proc/sys/kernel/randomize_va_space 4636 4637 noresume [SWSUSP] Disables resume and restores original swap 4638 space. 4639 4640 no-scroll [VGA] Disables scrollback. 4641 This is required for the Braillex ib80-piezo Braille 4642 reader made by F.H. Papenmeier (Germany). 4643 4644 nosgx [X86-64,SGX,EARLY] Disables Intel SGX kernel support. 4645 4646 nosmap [PPC,EARLY] 4647 Disable SMAP (Supervisor Mode Access Prevention) 4648 even if it is supported by processor. 4649 4650 nosmep [PPC64s,EARLY] 4651 Disable SMEP (Supervisor Mode Execution Prevention) 4652 even if it is supported by processor. 4653 4654 nosmp [SMP,EARLY] Tells an SMP kernel to act as a UP kernel, 4655 and disable the IO APIC. legacy for "maxcpus=0". 4656 4657 nosmt [KNL,MIPS,PPC,EARLY] Disable symmetric multithreading (SMT). 4658 Equivalent to smt=1. 4659 4660 [KNL,LOONGARCH,X86,ARM64,PPC,S390] Disable symmetric multithreading (SMT). 4661 nosmt=force: Force disable SMT, cannot be undone 4662 via the sysfs control file. 4663 4664 nosoftlockup [KNL] Disable the soft-lockup detector. 4665 4666 nospec_store_bypass_disable 4667 [HW,EARLY] Disable all mitigations for the Speculative 4668 Store Bypass vulnerability 4669 4670 nospectre_bhb [ARM64,EARLY] Disable all mitigations for Spectre-BHB (branch 4671 history injection) vulnerability. System may allow data leaks 4672 with this option. 4673 4674 nospectre_v1 [X86,PPC,EARLY] Disable mitigations for Spectre Variant 1 4675 (bounds check bypass). With this option data leaks are 4676 possible in the system. 4677 4678 nospectre_v2 [X86,PPC_E500,ARM64,EARLY] Disable all mitigations 4679 for the Spectre variant 2 (indirect branch 4680 prediction) vulnerability. System may allow data 4681 leaks with this option. 4682 4683 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES,RISCV,LOONGARCH,EARLY] 4684 Disable paravirtualized steal time accounting. steal time 4685 is computed, but won't influence scheduler behaviour 4686 4687 nosync [HW,M68K] Disables sync negotiation for all devices. 4688 4689 no_timer_check [X86,APIC] Disables the code which tests for broken 4690 timer IRQ sources, i.e., the IO-APIC timer. This can 4691 work around problems with incorrect timer 4692 initialization on some boards. 4693 4694 no_uaccess_flush 4695 [PPC,EARLY] Don't flush the L1-D cache after accessing user data. 4696 4697 novmcoredd [KNL,KDUMP] 4698 Disable device dump. Device dump allows drivers to 4699 append dump data to vmcore so you can collect driver 4700 specified debug info. Drivers can append the data 4701 without any limit and this data is stored in memory, 4702 so this may cause significant memory stress. Disabling 4703 device dump can help save memory but the driver debug 4704 data will be no longer available. This parameter 4705 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP 4706 is set. 4707 4708 no-vmw-sched-clock 4709 [X86,PV_OPS,EARLY] Disable paravirtualized VMware 4710 scheduler clock and use the default one. 4711 4712 nowatchdog [KNL] Disable both lockup detectors, i.e. 4713 soft-lockup and NMI watchdog (hard-lockup). 4714 4715 nowb [ARM,EARLY] 4716 4717 nox2apic [X86-64,APIC,EARLY] Do not enable x2APIC mode. 4718 4719 NOTE: this parameter will be ignored on systems with the 4720 LEGACY_XAPIC_DISABLED bit set in the 4721 IA32_XAPIC_DISABLE_STATUS MSR. 4722 4723 noxsave [BUGS=X86] Disables x86 extended register state save 4724 and restore using xsave. The kernel will fallback to 4725 enabling legacy floating-point and sse state. 4726 4727 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended 4728 register states. The kernel will fall back to use 4729 xsave to save the states. By using this parameter, 4730 performance of saving the states is degraded because 4731 xsave doesn't support modified optimization while 4732 xsaveopt supports it on xsaveopt enabled systems. 4733 4734 noxsaves [X86] Disables xsaves and xrstors used in saving and 4735 restoring x86 extended register state in compacted 4736 form of xsave area. The kernel will fall back to use 4737 xsaveopt and xrstor to save and restore the states 4738 in standard form of xsave area. By using this 4739 parameter, xsave area per process might occupy more 4740 memory on xsaves enabled systems. 4741 4742 nr_cpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 4743 could support. nr_cpus=n : n >= 1 limits the kernel to 4744 support 'n' processors. It could be larger than the 4745 number of already plugged CPU during bootup, later in 4746 runtime you can physically add extra cpu until it reaches 4747 n. So during boot up some boot time memory for per-cpu 4748 variables need be pre-allocated for later physical cpu 4749 hot plugging. 4750 4751 nr_uarts= [SERIAL] maximum number of UARTs to be registered. 4752 4753 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86, EARLY] 4754 Disable NUMA, Only set up a single NUMA node 4755 spanning all memory. 4756 4757 numa=fake=<size>[MG] 4758 [KNL, ARM64, RISCV, X86, EARLY] 4759 If given as a memory unit, fills all system RAM with 4760 nodes of size interleaved over physical nodes. 4761 4762 numa=fake=<N> 4763 [KNL, ARM64, RISCV, X86, EARLY] 4764 If given as an integer, fills all system RAM with N 4765 fake nodes interleaved over physical nodes. 4766 4767 numa=fake=<N>U 4768 [KNL, ARM64, RISCV, X86, EARLY] 4769 If given as an integer followed by 'U', it will 4770 divide each physical node into N emulated nodes. 4771 4772 numa=noacpi [X86] Don't parse the SRAT table for NUMA setup 4773 4774 numa=nohmat [X86] Don't parse the HMAT table for NUMA setup, or 4775 soft-reserved memory partitioning. 4776 4777 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic 4778 NUMA balancing. 4779 Allowed values are enable and disable 4780 4781 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA. 4782 'node', 'default' can be specified 4783 This can be set from sysctl after boot. 4784 See Documentation/admin-guide/sysctl/vm.rst for details. 4785 4786 nvme.quirks= [NVME] A list of quirk entries to augment the built-in 4787 nvme quirk list. List entries are separated by a 4788 '-' character. 4789 Each entry has the form VendorID:ProductID:quirk_names. 4790 The IDs are 4-digits hex numbers and quirk_names is a 4791 list of quirk names separated by commas. A quirk name 4792 can be prefixed by '^', meaning that the specified 4793 quirk must be disabled. 4794 4795 Example: 4796 nvme.quirks=7710:2267:bogus_nid,^identify_cns-9900:7711:broken_msi 4797 4798 ohci1394_dma=early [HW,EARLY] enable debugging via the ohci1394 driver. 4799 See Documentation/core-api/debugging-via-ohci1394.rst for more 4800 info. 4801 4802 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands 4803 Rather than timing out after 20 ms if an EC 4804 command is not properly ACKed, override the length 4805 of the timeout. We have interrupts disabled while 4806 waiting for the ACK, so if this is set too high 4807 interrupts *may* be lost! 4808 4809 omap_mux= [OMAP] Override bootloader pin multiplexing. 4810 Format: <mux_mode0.mode_name=value>... 4811 For example, to override I2C bus2: 4812 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100 4813 4814 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration 4815 4816 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock] 4817 4818 boundary - index of last SLC block on Flex-OneNAND. 4819 The remaining blocks are configured as MLC blocks. 4820 lock - Configure if Flex-OneNAND boundary should be locked. 4821 Once locked, the boundary cannot be changed. 4822 1 indicates lock status, 0 indicates unlock status. 4823 4824 oops=panic [KNL,EARLY] 4825 Always panic on oopses. Default is to just kill the 4826 process, but there is a small probability of 4827 deadlocking the machine. 4828 This will also cause panics on machine check exceptions. 4829 Useful together with panic=30 to trigger a reboot. 4830 4831 page_alloc.shuffle= 4832 [KNL] Boolean flag to control whether the page allocator 4833 should randomize its free lists. This parameter can be 4834 used to enable/disable page randomization. The state of 4835 the flag can be read from sysfs at: 4836 /sys/module/page_alloc/parameters/shuffle. 4837 This parameter is only available if CONFIG_SHUFFLE_PAGE_ALLOCATOR=y. 4838 4839 page_owner= [KNL,EARLY] Boot-time page_owner enabling option. 4840 Storage of the information about who allocated 4841 each page is disabled in default. With this switch, 4842 we can turn it on. 4843 on: enable the feature 4844 4845 page_poison= [KNL,EARLY] Boot-time parameter changing the state of 4846 poisoning on the buddy allocator, available with 4847 CONFIG_PAGE_POISONING=y. 4848 off: turn off poisoning (default) 4849 on: turn on poisoning 4850 4851 page_reporting.page_reporting_order= 4852 [KNL] Minimal page reporting order 4853 Format: <integer> 4854 Adjust the minimal page reporting order. The page 4855 reporting is disabled when it exceeds MAX_PAGE_ORDER. 4856 4857 panic= [KNL] Kernel behaviour on panic: delay <timeout> 4858 timeout > 0: seconds before rebooting 4859 timeout = 0: wait forever 4860 timeout < 0: reboot immediately 4861 Format: <timeout> 4862 4863 panic_on_taint= [KNL,EARLY] 4864 Bitmask for conditionally calling panic() in add_taint() 4865 Format: <hex>[,nousertaint] 4866 Hexadecimal bitmask representing the set of TAINT flags 4867 that will cause the kernel to panic when add_taint() is 4868 called with any of the flags in this set. 4869 The optional switch "nousertaint" can be utilized to 4870 prevent userspace forced crashes by writing to sysctl 4871 /proc/sys/kernel/tainted any flagset matching with the 4872 bitmask set on panic_on_taint. 4873 See Documentation/admin-guide/tainted-kernels.rst for 4874 extra details on the taint flags that users can pick 4875 to compose the bitmask to assign to panic_on_taint. 4876 4877 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump 4878 on a WARN(). 4879 4880 panic_force_cpu= 4881 [KNL,SMP] Force panic handling to execute on a specific CPU. 4882 Format: <cpu number> 4883 Some platforms require panic handling to occur on a 4884 specific CPU for the crash kernel to function correctly. 4885 This can be due to firmware limitations, interrupt routing 4886 constraints, or platform-specific requirements where only 4887 a particular CPU can safely enter the crash kernel. 4888 When set, panic() will redirect execution to the specified 4889 CPU before proceeding with the normal panic and kexec flow. 4890 If the target CPU is offline or unavailable, panic proceeds 4891 on the current CPU. 4892 This option should only be used for systems with the above 4893 constraints as it might cause the panic operation to be less reliable. 4894 4895 panic_print= Bitmask for printing system info when panic happens. 4896 User can chose combination of the following bits: 4897 4898 bit 0: print all tasks info 4899 bit 1: print system memory info 4900 bit 2: print timer info 4901 bit 3: print locks info if CONFIG_LOCKDEP is on 4902 bit 4: print ftrace buffer 4903 bit 5: replay all kernel messages on consoles at the end of panic 4904 bit 6: print all CPUs backtrace (if available in the arch) 4905 bit 7: print only tasks in uninterruptible (blocked) state 4906 4907 *Be aware* that this option may print a _lot_ of lines. 4908 4909 There is an increased risk of losing older messages in 4910 the log. Maybe worth to setup a bigger log buffer with 4911 "log_buf_len" along with this. 4912 4913 Also consider using "printk.console_no_auto_verbose=Y" 4914 if using this along with pstore, to avoid extra delays 4915 due to increased console verbosity during panic. 4916 4917 panic_sys_info= A comma separated list of extra information to be dumped 4918 on panic. 4919 Format: val[,val...] 4920 Where @val can be any of the following: 4921 4922 tasks: print all tasks info 4923 mem: print system memory info 4924 timers: print timers info 4925 locks: print locks info if CONFIG_LOCKDEP is on 4926 ftrace: print ftrace buffer 4927 all_bt: print all CPUs backtrace (if available in the arch) 4928 blocked_tasks: print only tasks in uninterruptible (blocked) state 4929 4930 This is a human readable alternative to the 'panic_print' 4931 option above. Please be aware of the caveats using this, 4932 check the 'panic_print' parameter for more details. 4933 4934 panic_console_replay 4935 When panic happens, replay all kernel messages on 4936 consoles at the end of panic. 4937 4938 parkbd.port= [HW] Parallel port number the keyboard adapter is 4939 connected to, default is 0. 4940 Format: <parport#> 4941 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation, 4942 0 for XT, 1 for AT (default is AT). 4943 Format: <mode> 4944 4945 parport= [HW,PPT] Specify parallel ports. 0 disables. 4946 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] } 4947 Use 'auto' to force the driver to use any 4948 IRQ/DMA settings detected (the default is to 4949 ignore detected IRQ/DMA settings because of 4950 possible conflicts). You can specify the base 4951 address, IRQ, and DMA settings; IRQ and DMA 4952 should be numbers, or 'auto' (for using detected 4953 settings on that particular port), or 'nofifo' 4954 (to avoid using a FIFO even if it is detected). 4955 Parallel ports are assigned in the order they 4956 are specified on the command line, starting 4957 with parport0. 4958 4959 parport_init_mode= [HW,PPT] 4960 Configure VIA parallel port to operate in 4961 a specific mode. This is necessary on Pegasos 4962 computer where firmware has no options for setting 4963 up parallel port mode and sets it to spp. 4964 Currently this function knows 686a and 8231 chips. 4965 Format: [spp|ps2|epp|ecp|ecpepp] 4966 4967 pata_legacy.all= [HW,LIBATA] 4968 Format: <int> 4969 Set to non-zero to probe primary and secondary ISA 4970 port ranges on PCI systems where no PCI PATA device 4971 has been found at either range. Disabled by default. 4972 4973 pata_legacy.autospeed= [HW,LIBATA] 4974 Format: <int> 4975 Set to non-zero if a chip is present that snoops speed 4976 changes. Disabled by default. 4977 4978 pata_legacy.ht6560a= [HW,LIBATA] 4979 Format: <int> 4980 Set to 1, 2, or 3 for HT 6560A on the primary channel, 4981 the secondary channel, or both channels respectively. 4982 Disabled by default. 4983 4984 pata_legacy.ht6560b= [HW,LIBATA] 4985 Format: <int> 4986 Set to 1, 2, or 3 for HT 6560B on the primary channel, 4987 the secondary channel, or both channels respectively. 4988 Disabled by default. 4989 4990 pata_legacy.iordy_mask= [HW,LIBATA] 4991 Format: <int> 4992 IORDY enable mask. Set individual bits to allow IORDY 4993 for the respective channel. Bit 0 is for the first 4994 legacy channel handled by this driver, bit 1 is for 4995 the second channel, and so on. The sequence will often 4996 correspond to the primary legacy channel, the secondary 4997 legacy channel, and so on, but the handling of a PCI 4998 bus and the use of other driver options may interfere 4999 with the sequence. By default IORDY is allowed across 5000 all channels. 5001 5002 pata_legacy.opti82c46x= [HW,LIBATA] 5003 Format: <int> 5004 Set to 1, 2, or 3 for Opti 82c611A on the primary 5005 channel, the secondary channel, or both channels 5006 respectively. Disabled by default. 5007 5008 pata_legacy.opti82c611a= [HW,LIBATA] 5009 Format: <int> 5010 Set to 1, 2, or 3 for Opti 82c465MV on the primary 5011 channel, the secondary channel, or both channels 5012 respectively. Disabled by default. 5013 5014 pata_legacy.pio_mask= [HW,LIBATA] 5015 Format: <int> 5016 PIO mode mask for autospeed devices. Set individual 5017 bits to allow the use of the respective PIO modes. 5018 Bit 0 is for mode 0, bit 1 is for mode 1, and so on. 5019 All modes allowed by default. 5020 5021 pata_legacy.probe_all= [HW,LIBATA] 5022 Format: <int> 5023 Set to non-zero to probe tertiary and further ISA 5024 port ranges on PCI systems. Disabled by default. 5025 5026 pata_legacy.probe_mask= [HW,LIBATA] 5027 Format: <int> 5028 Probe mask for legacy ISA PATA ports. Depending on 5029 platform configuration and the use of other driver 5030 options up to 6 legacy ports are supported: 0x1f0, 5031 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing 5032 of individual ports can be disabled by setting the 5033 corresponding bits in the mask to 1. Bit 0 is for 5034 the first port in the list above (0x1f0), and so on. 5035 By default all supported ports are probed. 5036 5037 pata_legacy.qdi= [HW,LIBATA] 5038 Format: <int> 5039 Set to non-zero to probe QDI controllers. By default 5040 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise. 5041 5042 pata_legacy.winbond= [HW,LIBATA] 5043 Format: <int> 5044 Set to non-zero to probe Winbond controllers. Use 5045 the standard I/O port (0x130) if 1, otherwise the 5046 value given is the I/O port to use (typically 0x1b0). 5047 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE, 5048 0 otherwise. 5049 5050 pata_platform.pio_mask= [HW,LIBATA] 5051 Format: <int> 5052 Supported PIO mode mask. Set individual bits to allow 5053 the use of the respective PIO modes. Bit 0 is for 5054 mode 0, bit 1 is for mode 1, and so on. Mode 0 only 5055 allowed by default. 5056 5057 pause_on_oops=<int> 5058 Halt all CPUs after the first oops has been printed for 5059 the specified number of seconds. This is to be used if 5060 your oopses keep scrolling off the screen. 5061 5062 pci=option[,option...] [PCI,EARLY] various PCI subsystem options. 5063 5064 Some options herein operate on a specific device 5065 or a set of devices (<pci_dev>). These are 5066 specified in one of the following formats: 5067 5068 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]* 5069 pci:<vendor>:<device>[:<subvendor>:<subdevice>] 5070 5071 Note: the first format specifies a PCI 5072 bus/device/function address which may change 5073 if new hardware is inserted, if motherboard 5074 firmware changes, or due to changes caused 5075 by other kernel parameters. If the 5076 domain is left unspecified, it is 5077 taken to be zero. Optionally, a path 5078 to a device through multiple device/function 5079 addresses can be specified after the base 5080 address (this is more robust against 5081 renumbering issues). The second format 5082 selects devices using IDs from the 5083 configuration space which may match multiple 5084 devices in the system. 5085 5086 earlydump dump PCI config space before the kernel 5087 changes anything 5088 off [X86] don't probe for the PCI bus 5089 bios [X86-32] force use of PCI BIOS, don't access 5090 the hardware directly. Use this if your machine 5091 has a non-standard PCI host bridge. 5092 nobios [X86-32] disallow use of PCI BIOS, only direct 5093 hardware access methods are allowed. Use this 5094 if you experience crashes upon bootup and you 5095 suspect they are caused by the BIOS. 5096 conf1 [X86] Force use of PCI Configuration Access 5097 Mechanism 1 (config address in IO port 0xCF8, 5098 data in IO port 0xCFC, both 32-bit). 5099 conf2 [X86] Force use of PCI Configuration Access 5100 Mechanism 2 (IO port 0xCF8 is an 8-bit port for 5101 the function, IO port 0xCFA, also 8-bit, sets 5102 bus number. The config space is then accessed 5103 through ports 0xC000-0xCFFF). 5104 See http://wiki.osdev.org/PCI for more info 5105 on the configuration access mechanisms. 5106 noaer [PCIE] If the PCIEAER kernel config parameter is 5107 enabled, this kernel boot option can be used to 5108 disable the use of PCIE advanced error reporting. 5109 nodomains [PCI] Disable support for multiple PCI 5110 root domains (aka PCI segments, in ACPI-speak). 5111 nommconf [X86] Disable use of MMCONFIG for PCI 5112 Configuration 5113 check_enable_amd_mmconf [X86] check for and enable 5114 properly configured MMIO access to PCI 5115 config space on AMD family 10h CPU 5116 nomsi [MSI] If the PCI_MSI kernel config parameter is 5117 enabled, this kernel boot option can be used to 5118 disable the use of MSI interrupts system-wide. 5119 noioapicquirk [APIC] Disable all boot interrupt quirks. 5120 Safety option to keep boot IRQs enabled. This 5121 should never be necessary. 5122 ioapicreroute [APIC] Enable rerouting of boot IRQs to the 5123 primary IO-APIC for bridges that cannot disable 5124 boot IRQs. This fixes a source of spurious IRQs 5125 when the system masks IRQs. 5126 noioapicreroute [APIC] Disable workaround that uses the 5127 boot IRQ equivalent of an IRQ that connects to 5128 a chipset where boot IRQs cannot be disabled. 5129 The opposite of ioapicreroute. 5130 biosirq [X86-32] Use PCI BIOS calls to get the interrupt 5131 routing table. These calls are known to be buggy 5132 on several machines and they hang the machine 5133 when used, but on other computers it's the only 5134 way to get the interrupt routing table. Try 5135 this option if the kernel is unable to allocate 5136 IRQs or discover secondary PCI buses on your 5137 motherboard. 5138 rom [X86] Assign address space to expansion ROMs. 5139 Use with caution as certain devices share 5140 address decoders between ROMs and other 5141 resources. 5142 norom [X86] Do not assign address space to 5143 expansion ROMs that do not already have 5144 BIOS assigned address ranges. 5145 nobar [X86] Do not assign address space to the 5146 BARs that weren't assigned by the BIOS. 5147 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be 5148 assigned automatically to PCI devices. You can 5149 make the kernel exclude IRQs of your ISA cards 5150 this way. 5151 pirqaddr=0xAAAAA [X86] Specify the physical address 5152 of the PIRQ table (normally generated 5153 by the BIOS) if it is outside the 5154 F0000h-100000h range. 5155 lastbus=N [X86] Scan all buses thru bus #N. Can be 5156 useful if the kernel is unable to find your 5157 secondary buses and you want to tell it 5158 explicitly which ones they are. 5159 assign-busses [X86] Always assign all PCI bus 5160 numbers ourselves, overriding 5161 whatever the firmware may have done. 5162 usepirqmask [X86] Honor the possible IRQ mask stored 5163 in the BIOS $PIR table. This is needed on 5164 some systems with broken BIOSes, notably 5165 some HP Pavilion N5400 and Omnibook XE3 5166 notebooks. This will have no effect if ACPI 5167 IRQ routing is enabled. 5168 noacpi [X86] Do not use ACPI for IRQ routing 5169 or for PCI scanning. 5170 use_crs [X86] Use PCI host bridge window information 5171 from ACPI. On BIOSes from 2008 or later, this 5172 is enabled by default. If you need to use this, 5173 please report a bug. 5174 nocrs [X86] Ignore PCI host bridge windows from ACPI. 5175 If you need to use this, please report a bug. 5176 use_e820 [X86] Use E820 reservations to exclude parts of 5177 PCI host bridge windows. This is a workaround 5178 for BIOS defects in host bridge _CRS methods. 5179 If you need to use this, please report a bug to 5180 <linux-pci@vger.kernel.org>. 5181 no_e820 [X86] Ignore E820 reservations for PCI host 5182 bridge windows. This is the default on modern 5183 hardware. If you need to use this, please report 5184 a bug to <linux-pci@vger.kernel.org>. 5185 routeirq Do IRQ routing for all PCI devices. 5186 This is normally done in pci_enable_device(), 5187 so this option is a temporary workaround 5188 for broken drivers that don't call it. 5189 skip_isa_align [X86] do not align io start addr, so can 5190 handle more pci cards 5191 noearly [X86] Don't do any early type 1 scanning. 5192 This might help on some broken boards which 5193 machine check when some devices' config space 5194 is read. But various workarounds are disabled 5195 and some IOMMU drivers will not work. 5196 bfsort Sort PCI devices into breadth-first order. 5197 This sorting is done to get a device 5198 order compatible with older (<= 2.4) kernels. 5199 nobfsort Don't sort PCI devices into breadth-first order. 5200 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size) 5201 tuning and use the BIOS-configured MPS defaults. 5202 pcie_bus_safe Set every device's MPS to the largest value 5203 supported by all devices below the root complex. 5204 pcie_bus_perf Set device MPS to the largest allowable MPS 5205 based on its parent bus. Also set MRRS (Max 5206 Read Request Size) to the largest supported 5207 value (no larger than the MPS that the device 5208 or bus can support) for best performance. 5209 pcie_bus_peer2peer Set every device's MPS to 128B, which 5210 every device is guaranteed to support. This 5211 configuration allows peer-to-peer DMA between 5212 any pair of devices, possibly at the cost of 5213 reduced performance. This also guarantees 5214 that hot-added devices will work. 5215 cbiosize=nn[KMG] The fixed amount of bus space which is 5216 reserved for the CardBus bridge's IO window. 5217 The default value is 256 bytes. 5218 cbmemsize=nn[KMG] The fixed amount of bus space which is 5219 reserved for the CardBus bridge's memory 5220 window. The default value is 64 megabytes. 5221 resource_alignment= 5222 Format: 5223 [<order of align>@]<pci_dev>[; ...] 5224 Specifies alignment and device to reassign 5225 aligned memory resources. How to 5226 specify the device is described above. 5227 If <order of align> is not specified, 5228 PAGE_SIZE is used as alignment. 5229 A PCI-PCI bridge can be specified if resource 5230 windows need to be expanded. 5231 To specify the alignment for several 5232 instances of a device, the PCI vendor, 5233 device, subvendor, and subdevice may be 5234 specified, e.g., 12@pci:8086:9c22:103c:198f 5235 for 4096-byte alignment. 5236 ecrc= Enable/disable PCIe ECRC (transaction layer 5237 end-to-end CRC checking). Only effective if 5238 OS has native AER control (either granted by 5239 ACPI _OSC or forced via "pcie_ports=native") 5240 bios: Use BIOS/firmware settings. This is the 5241 the default. 5242 off: Turn ECRC off 5243 on: Turn ECRC on. 5244 hpiosize=nn[KMG] The fixed amount of bus space which is 5245 reserved for hotplug bridge's IO window. 5246 Default size is 256 bytes. 5247 hpmmiosize=nn[KMG] The fixed amount of bus space which is 5248 reserved for hotplug bridge's MMIO window. 5249 Default size is 2 megabytes. 5250 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is 5251 reserved for hotplug bridge's MMIO_PREF window. 5252 Default size is 2 megabytes. 5253 hpmemsize=nn[KMG] The fixed amount of bus space which is 5254 reserved for hotplug bridge's MMIO and 5255 MMIO_PREF window. 5256 Default size is 2 megabytes. 5257 hpbussize=nn The minimum amount of additional bus numbers 5258 reserved for buses below a hotplug bridge. 5259 Default is 1. 5260 realloc= Enable/disable reallocating PCI bridge resources 5261 if allocations done by BIOS are too small to 5262 accommodate resources required by all child 5263 devices. 5264 off: Turn realloc off 5265 on: Turn realloc on 5266 realloc same as realloc=on 5267 noari do not use PCIe ARI. 5268 noats [PCIE, Intel-IOMMU, AMD-IOMMU] 5269 do not use PCIe ATS (and IOMMU device IOTLB). 5270 pcie_scan_all Scan all possible PCIe devices. Otherwise we 5271 only look for one device below a PCIe downstream 5272 port. 5273 big_root_window Try to add a big 64bit memory window to the PCIe 5274 root complex on AMD CPUs. Some GFX hardware 5275 can resize a BAR to allow access to all VRAM. 5276 Adding the window is slightly risky (it may 5277 conflict with unreported devices), so this 5278 taints the kernel. 5279 disable_acs_redir=<pci_dev>[; ...] 5280 Specify one or more PCI devices (in the format 5281 specified above) separated by semicolons. 5282 Each device specified will have the PCI ACS 5283 redirect capabilities forced off which will 5284 allow P2P traffic between devices through 5285 bridges without forcing it upstream. Note: 5286 this removes isolation between devices and 5287 may put more devices in an IOMMU group. 5288 config_acs= 5289 Format: 5290 <ACS flags>@<pci_dev>[; ...] 5291 Specify one or more PCI devices (in the format 5292 specified above) optionally prepended with flags 5293 and separated by semicolons. The respective 5294 capabilities will be enabled, disabled or 5295 unchanged based on what is specified in 5296 flags. 5297 5298 ACS Flags is defined as follows: 5299 bit-0 : ACS Source Validation 5300 bit-1 : ACS Translation Blocking 5301 bit-2 : ACS P2P Request Redirect 5302 bit-3 : ACS P2P Completion Redirect 5303 bit-4 : ACS Upstream Forwarding 5304 bit-5 : ACS P2P Egress Control 5305 bit-6 : ACS Direct Translated P2P 5306 Each bit can be marked as: 5307 '0' – force disabled 5308 '1' – force enabled 5309 'x' – unchanged 5310 For example, 5311 pci=config_acs=10x@pci:0:0 5312 would configure all devices that support 5313 ACS to enable P2P Request Redirect, disable 5314 Translation Blocking, and leave Source 5315 Validation unchanged from whatever power-up 5316 or firmware set it to. 5317 5318 Note: this may remove isolation between devices 5319 and may put more devices in an IOMMU group. 5320 force_floating [S390] Force usage of floating interrupts. 5321 nomio [S390] Do not use MIO instructions. 5322 norid [S390] ignore the RID field and force use of 5323 one PCI domain per PCI function 5324 notph [PCIE] If the PCIE_TPH kernel config parameter 5325 is enabled, this kernel boot option can be used 5326 to disable PCIe TLP Processing Hints support 5327 system-wide. 5328 5329 pcie_aspm= [PCIE] Forcibly enable or ignore PCIe Active State Power 5330 Management. 5331 off Don't touch ASPM configuration at all. Leave any 5332 configuration done by firmware unchanged. 5333 force Enable ASPM even on devices that claim not to support it. 5334 WARNING: Forcing ASPM on may cause system lockups. 5335 5336 pcie_ports= [PCIE] PCIe port services handling: 5337 native Use native PCIe services (PME, AER, DPC, PCIe hotplug) 5338 even if the platform doesn't give the OS permission to 5339 use them. This may cause conflicts if the platform 5340 also tries to use these services. 5341 dpc-native Use native PCIe service for DPC only. May 5342 cause conflicts if firmware uses AER or DPC. 5343 compat Disable native PCIe services (PME, AER, DPC, PCIe 5344 hotplug). 5345 5346 pcie_port_pm= [PCIE] PCIe port power management handling: 5347 off Disable power management of all PCIe ports 5348 force Forcibly enable power management of all PCIe ports 5349 5350 pcie_pme= [PCIE,PM] Native PCIe PME signaling options: 5351 nomsi Do not use MSI for native PCIe PME signaling (this makes 5352 all PCIe root ports use INTx for all services). 5353 5354 pcmv= [HW,PCMCIA] BadgePAD 4 5355 5356 pd_ignore_unused 5357 [PM] 5358 Keep all power-domains already enabled by bootloader on, 5359 even if no driver has claimed them. This is useful 5360 for debug and development, but should not be 5361 needed on a platform with proper driver support. 5362 5363 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at 5364 boot time. 5365 Format: { 0 | 1 } 5366 See arch/parisc/kernel/pdc_chassis.c 5367 5368 percpu_alloc= [MM,EARLY] 5369 Select which percpu first chunk allocator to use. 5370 Currently supported values are "embed" and "page". 5371 Archs may support subset or none of the selections. 5372 See comments in mm/percpu.c for details on each 5373 allocator. This parameter is primarily for debugging 5374 and performance comparison. 5375 5376 pirq= [SMP,APIC] Manual mp-table setup 5377 See Documentation/arch/x86/i386/IO-APIC.rst. 5378 5379 plip= [PPT,NET] Parallel port network link 5380 Format: { parport<nr> | timid | 0 } 5381 See also Documentation/admin-guide/parport.rst. 5382 5383 pmtmr= [X86] Manual setup of pmtmr I/O Port. 5384 Override pmtimer IOPort with a hex value. 5385 e.g. pmtmr=0x508 5386 5387 pmu_override= [PPC] Override the PMU. 5388 This option takes over the PMU facility, so it is no 5389 longer usable by perf. Setting this option starts the 5390 PMU counters by setting MMCR0 to 0 (the FC bit is 5391 cleared). If a number is given, then MMCR1 is set to 5392 that number, otherwise (e.g., 'pmu_override=on'), MMCR1 5393 remains 0. 5394 5395 pm_async= [PM] 5396 Format: off 5397 This parameter sets the initial value of the 5398 /sys/power/pm_async sysfs knob at boot time. 5399 If set to "off", disables asynchronous suspend and 5400 resume of devices during system-wide power transitions. 5401 This can be useful on platforms where device 5402 dependencies are not well-defined, or for debugging 5403 power management issues. Asynchronous operations are 5404 enabled by default. 5405 5406 5407 pm_debug_messages [SUSPEND,KNL] 5408 Enable suspend/resume debug messages during boot up. 5409 5410 pm.dpm_watchdog_enabled= 5411 [PM] Enable or disable the DPM watchdog. Requires 5412 CONFIG_PM_SLEEP and CONFIG_DPM_WATCHDOG enabled. 5413 Format: <bool> 5414 Default value is set by CONFIG_DPM_WATCHDOG_ENABLED. 5415 5416 pnp.debug=1 [PNP] 5417 Enable PNP debug messages (depends on the 5418 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time 5419 via /sys/module/pnp/parameters/debug. We always show 5420 current resource usage; turning this on also shows 5421 possible settings and some assignment information. 5422 5423 pnpacpi= [ACPI] 5424 { off } 5425 5426 pnpbios= [ISAPNP] 5427 { on | off | curr | res | no-curr | no-res } 5428 5429 pnp_reserve_irq= 5430 [ISAPNP] Exclude IRQs for the autoconfiguration 5431 5432 pnp_reserve_dma= 5433 [ISAPNP] Exclude DMAs for the autoconfiguration 5434 5435 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration 5436 Ranges are in pairs (I/O port base and size). 5437 5438 pnp_reserve_mem= 5439 [ISAPNP] Exclude memory regions for the 5440 autoconfiguration. 5441 Ranges are in pairs (memory base and size). 5442 5443 ports= [IP_VS_FTP] IPVS ftp helper module 5444 Default is 21. 5445 Up to 8 (IP_VS_APP_MAX_PORTS) ports 5446 may be specified. 5447 Format: <port>,<port>.... 5448 5449 possible_cpus= [SMP,S390,X86] 5450 Format: <unsigned int> 5451 Set the number of possible CPUs, overriding the 5452 regular discovery mechanisms (such as ACPI/FW, etc). 5453 5454 powersave=off [PPC] This option disables power saving features. 5455 It specifically disables cpuidle and sets the 5456 platform machine description specific power_save 5457 function to NULL. On Idle the CPU just reduces 5458 execution priority. 5459 5460 ppc_strict_facility_enable 5461 [PPC,ENABLE] This option catches any kernel floating point, 5462 Altivec, VSX and SPE outside of regions specifically 5463 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()). 5464 There is some performance impact when enabling this. 5465 5466 ppc_tm= [PPC,EARLY] 5467 Format: {"off"} 5468 Disable Hardware Transactional Memory 5469 5470 preempt= [KNL] 5471 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC 5472 none - Limited to cond_resched() calls 5473 voluntary - Limited to cond_resched() and might_sleep() calls 5474 full - Any section that isn't explicitly preempt disabled 5475 can be preempted anytime. Tasks will also yield 5476 contended spinlocks (if the critical section isn't 5477 explicitly preempt disabled beyond the lock itself). 5478 lazy - Scheduler controlled. Similar to full but instead 5479 of preempting the task immediately, the task gets 5480 one HZ tick time to yield itself before the 5481 preemption will be forced. One preemption is when the 5482 task returns to user space. 5483 5484 print-fatal-signals= 5485 [KNL] debug: print fatal signals 5486 5487 If enabled, warn about various signal handling 5488 related application anomalies: too many signals, 5489 too many POSIX.1 timers, fatal signals causing a 5490 coredump - etc. 5491 5492 If you hit the warning due to signal overflow, 5493 you might want to try "ulimit -i unlimited". 5494 5495 default: off. 5496 5497 printk.always_kmsg_dump= 5498 Trigger kmsg_dump for cases other than kernel oops or 5499 panics 5500 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5501 default: disabled 5502 5503 printk.console_no_auto_verbose= 5504 Disable console loglevel raise on oops, panic 5505 or lockdep-detected issues (only if lock debug is on). 5506 With an exception to setups with low baudrate on 5507 serial console, keeping this 0 is a good choice 5508 in order to provide more debug information. 5509 Format: <bool> 5510 default: 0 (auto_verbose is enabled) 5511 5512 printk.debug_non_panic_cpus= 5513 Allows storing messages from non-panic CPUs into 5514 the printk log buffer during panic(). They are 5515 flushed to consoles by the panic-CPU on 5516 a best-effort basis. 5517 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5518 Default: disabled 5519 5520 printk.devkmsg={on,off,ratelimit} 5521 Control writing to /dev/kmsg. 5522 on - unlimited logging to /dev/kmsg from userspace 5523 off - logging to /dev/kmsg disabled 5524 ratelimit - ratelimit the logging 5525 Default: ratelimit 5526 5527 printk.time= Show timing data prefixed to each printk message line 5528 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5529 5530 proc_mem.force_override= [KNL] 5531 Format: {always | ptrace | never} 5532 Traditionally /proc/pid/mem allows memory permissions to be 5533 overridden without restrictions. This option may be set to 5534 restrict that. Can be one of: 5535 - 'always': traditional behavior always allows mem overrides. 5536 - 'ptrace': only allow mem overrides for active ptracers. 5537 - 'never': never allow mem overrides. 5538 If not specified, default is the CONFIG_PROC_MEM_* choice. 5539 5540 processor.max_cstate= [HW,ACPI] 5541 Limit processor to maximum C-state 5542 max_cstate=9 overrides any DMI blacklist limit. 5543 5544 processor.nocst [HW,ACPI] 5545 Ignore the _CST method to determine C-states, 5546 instead using the legacy FADT method 5547 5548 profile= [KNL] Enable kernel profiling via /proc/profile 5549 Format: [<profiletype>,]<number> 5550 Param: <profiletype>: "schedule" or "kvm" 5551 [defaults to kernel profiling] 5552 Param: "schedule" - profile schedule points. 5553 Param: "kvm" - profile VM exits. 5554 Param: <number> - step/bucket size as a power of 2 for 5555 statistical time based profiling. 5556 5557 prot_virt= [S390] enable hosting protected virtual machines 5558 isolated from the hypervisor (if hardware supports 5559 that). If enabled, the default kernel base address 5560 might be overridden even when Kernel Address Space 5561 Layout Randomization is disabled. 5562 Format: <bool> 5563 5564 psi= [KNL] Enable or disable pressure stall information 5565 tracking. 5566 Format: <bool> 5567 5568 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to 5569 probe for; one of (bare|imps|exps|lifebook|any). 5570 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports 5571 per second. 5572 psmouse.resetafter= [HW,MOUSE] 5573 Try to reset the device after so many bad packets 5574 (0 = never). 5575 psmouse.resolution= 5576 [HW,MOUSE] Set desired mouse resolution, in dpi. 5577 psmouse.smartscroll= 5578 [HW,MOUSE] Controls Logitech smartscroll autorepeat. 5579 0 = disabled, 1 = enabled (default). 5580 5581 pstore.backend= Specify the name of the pstore backend to use 5582 5583 pti= [X86-64] Control Page Table Isolation of user and 5584 kernel address spaces. Disabling this feature 5585 removes hardening, but improves performance of 5586 system calls and interrupts. 5587 5588 on - unconditionally enable 5589 off - unconditionally disable 5590 auto - kernel detects whether your CPU model is 5591 vulnerable to issues that PTI mitigates 5592 5593 Not specifying this option is equivalent to pti=auto. 5594 5595 pty.legacy_count= 5596 [KNL] Number of legacy pty's. Overwrites compiled-in 5597 default number. 5598 5599 quiet [KNL,EARLY] Disable most log messages 5600 5601 r128= [HW,DRM] 5602 5603 radix_hcall_invalidate=on [PPC/PSERIES] 5604 Disable RADIX GTSE feature and use hcall for TLB 5605 invalidate. 5606 5607 raid= [HW,RAID] 5608 See Documentation/admin-guide/md.rst. 5609 5610 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes 5611 See Documentation/admin-guide/blockdev/ramdisk.rst. 5612 5613 ramdisk_start= [Deprecated,RAM] RAM disk image start address 5614 5615 random.trust_cpu=off 5616 [KNL,EARLY] Disable trusting the use of the CPU's 5617 random number generator (if available) to 5618 initialize the kernel's RNG. 5619 5620 random.trust_bootloader=off 5621 [KNL,EARLY] Disable trusting the use of the a seed 5622 passed by the bootloader (if available) to 5623 initialize the kernel's RNG. 5624 5625 randomize_kstack_offset= 5626 [KNL,EARLY] Enable or disable kernel stack offset 5627 randomization, which provides roughly 5 bits of 5628 entropy, frustrating memory corruption attacks 5629 that depend on stack address determinism or 5630 cross-syscall address exposures. This is only 5631 available on architectures that have defined 5632 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET. 5633 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5634 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT. 5635 5636 ras=option[,option,...] [KNL] RAS-specific options 5637 5638 cec_disable [X86] 5639 Disable the Correctable Errors Collector, 5640 see CONFIG_RAS_CEC help text. 5641 5642 rcu_nocbs[=cpu-list] 5643 [KNL] The optional argument is a cpu list, 5644 as described above. 5645 5646 In kernels built with CONFIG_RCU_NOCB_CPU=y, 5647 enable the no-callback CPU mode, which prevents 5648 such CPUs' callbacks from being invoked in 5649 softirq context. Invocation of such CPUs' RCU 5650 callbacks will instead be offloaded to "rcuox/N" 5651 kthreads created for that purpose, where "x" is 5652 "p" for RCU-preempt, "s" for RCU-sched, and "g" 5653 for the kthreads that mediate grace periods; and 5654 "N" is the CPU number. This reduces OS jitter on 5655 the offloaded CPUs, which can be useful for HPC 5656 and real-time workloads. It can also improve 5657 energy efficiency for asymmetric multiprocessors. 5658 5659 If a cpulist is passed as an argument, the specified 5660 list of CPUs is set to no-callback mode from boot. 5661 5662 Otherwise, if the '=' sign and the cpulist 5663 arguments are omitted, no CPU will be set to 5664 no-callback mode from boot but the mode may be 5665 toggled at runtime via cpusets. 5666 5667 Note that this argument takes precedence over 5668 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 5669 5670 rcu_nocb_poll [KNL] 5671 Rather than requiring that offloaded CPUs 5672 (specified by rcu_nocbs= above) explicitly 5673 awaken the corresponding "rcuoN" kthreads, 5674 make these kthreads poll for callbacks. 5675 This improves the real-time response for the 5676 offloaded CPUs by relieving them of the need to 5677 wake up the corresponding kthread, but degrades 5678 energy efficiency by requiring that the kthreads 5679 periodically wake up to do the polling. 5680 5681 rcutree.blimit= [KNL] 5682 Set maximum number of finished RCU callbacks to 5683 process in one batch. 5684 5685 rcutree.csd_lock_suppress_rcu_stall= [KNL] 5686 Do only a one-line RCU CPU stall warning when 5687 there is an ongoing too-long CSD-lock wait. 5688 5689 rcutree.do_rcu_barrier= [KNL] 5690 Request a call to rcu_barrier(). This is 5691 throttled so that userspace tests can safely 5692 hammer on the sysfs variable if they so choose. 5693 If triggered before the RCU grace-period machinery 5694 is fully active, this will error out with EAGAIN. 5695 5696 rcutree.dump_tree= [KNL] 5697 Dump the structure of the rcu_node combining tree 5698 out at early boot. This is used for diagnostic 5699 purposes, to verify correct tree setup. 5700 5701 rcutree.gp_cleanup_delay= [KNL] 5702 Set the number of jiffies to delay each step of 5703 RCU grace-period cleanup. 5704 5705 rcutree.gp_init_delay= [KNL] 5706 Set the number of jiffies to delay each step of 5707 RCU grace-period initialization. 5708 5709 rcutree.gp_preinit_delay= [KNL] 5710 Set the number of jiffies to delay each step of 5711 RCU grace-period pre-initialization, that is, 5712 the propagation of recent CPU-hotplug changes up 5713 the rcu_node combining tree. 5714 5715 rcutree.jiffies_till_first_fqs= [KNL] 5716 Set delay from grace-period initialization to 5717 first attempt to force quiescent states. 5718 Units are jiffies, minimum value is zero, 5719 and maximum value is HZ. 5720 5721 rcutree.jiffies_till_next_fqs= [KNL] 5722 Set delay between subsequent attempts to force 5723 quiescent states. Units are jiffies, minimum 5724 value is one, and maximum value is HZ. 5725 5726 rcutree.jiffies_till_sched_qs= [KNL] 5727 Set required age in jiffies for a 5728 given grace period before RCU starts 5729 soliciting quiescent-state help from 5730 rcu_note_context_switch() and cond_resched(). 5731 If not specified, the kernel will calculate 5732 a value based on the most recent settings 5733 of rcutree.jiffies_till_first_fqs 5734 and rcutree.jiffies_till_next_fqs. 5735 This calculated value may be viewed in 5736 rcutree.jiffies_to_sched_qs. Any attempt to set 5737 rcutree.jiffies_to_sched_qs will be cheerfully 5738 overwritten. 5739 5740 rcutree.kthread_prio= [KNL,BOOT] 5741 Set the SCHED_FIFO priority of the RCU per-CPU 5742 kthreads (rcuc/N). This value is also used for 5743 the priority of the RCU boost threads (rcub/N) 5744 and for the RCU grace-period kthreads (rcu_bh, 5745 rcu_preempt, and rcu_sched). If RCU_BOOST is 5746 set, valid values are 1-99 and the default is 1 5747 (the least-favored priority). Otherwise, when 5748 RCU_BOOST is not set, valid values are 0-99 and 5749 the default is zero (non-realtime operation). 5750 When RCU_NOCB_CPU is set, also adjust the 5751 priority of NOCB callback kthreads. 5752 5753 rcutree.nocb_nobypass_lim_per_jiffy= [KNL] 5754 On callback-offloaded (rcu_nocbs) CPUs, 5755 RCU reduces the lock contention that would 5756 otherwise be caused by callback floods through 5757 use of the ->nocb_bypass list. However, in the 5758 common non-flooded case, RCU queues directly to 5759 the main ->cblist in order to avoid the extra 5760 overhead of the ->nocb_bypass list and its lock. 5761 But if there are too many callbacks queued during 5762 a single jiffy, RCU pre-queues the callbacks into 5763 the ->nocb_bypass queue. The definition of "too 5764 many" is supplied by this kernel boot parameter. 5765 5766 rcutree.nohz_full_patience_delay= [KNL] 5767 On callback-offloaded (rcu_nocbs) CPUs, avoid 5768 disturbing RCU unless the grace period has 5769 reached the specified age in milliseconds. 5770 Defaults to zero. Large values will be capped 5771 at five seconds. All values will be rounded down 5772 to the nearest value representable by jiffies. 5773 5774 rcutree.qhimark= [KNL] 5775 Set threshold of queued RCU callbacks beyond which 5776 batch limiting is disabled. 5777 5778 rcutree.qlowmark= [KNL] 5779 Set threshold of queued RCU callbacks below which 5780 batch limiting is re-enabled. 5781 5782 rcutree.qovld= [KNL] 5783 Set threshold of queued RCU callbacks beyond which 5784 RCU's force-quiescent-state scan will aggressively 5785 enlist help from cond_resched() and sched IPIs to 5786 help CPUs more quickly reach quiescent states. 5787 Set to less than zero to make this be set based 5788 on rcutree.qhimark at boot time and to zero to 5789 disable more aggressive help enlistment. 5790 5791 rcutree.rcu_delay_page_cache_fill_msec= [KNL] 5792 Set the page-cache refill delay (in milliseconds) 5793 in response to low-memory conditions. The range 5794 of permitted values is in the range 0:100000. 5795 5796 rcutree.rcu_divisor= [KNL] 5797 Set the shift-right count to use to compute 5798 the callback-invocation batch limit bl from 5799 the number of callbacks queued on this CPU. 5800 The result will be bounded below by the value of 5801 the rcutree.blimit kernel parameter. Every bl 5802 callbacks, the softirq handler will exit in 5803 order to allow the CPU to do other work. 5804 5805 Please note that this callback-invocation batch 5806 limit applies only to non-offloaded callback 5807 invocation. Offloaded callbacks are instead 5808 invoked in the context of an rcuoc kthread, which 5809 scheduler will preempt as it does any other task. 5810 5811 rcutree.rcu_fanout_exact= [KNL] 5812 Disable autobalancing of the rcu_node combining 5813 tree. This is used by rcutorture, and might 5814 possibly be useful for architectures having high 5815 cache-to-cache transfer latencies. 5816 5817 rcutree.rcu_fanout_leaf= [KNL] 5818 Change the number of CPUs assigned to each 5819 leaf rcu_node structure. Useful for very 5820 large systems, which will choose the value 64, 5821 and for NUMA systems with large remote-access 5822 latencies, which will choose a value aligned 5823 with the appropriate hardware boundaries. 5824 5825 rcutree.rcu_min_cached_objs= [KNL] 5826 Minimum number of objects which are cached and 5827 maintained per one CPU. Object size is equal 5828 to PAGE_SIZE. The cache allows to reduce the 5829 pressure to page allocator, also it makes the 5830 whole algorithm to behave better in low memory 5831 condition. 5832 5833 rcutree.rcu_nocb_gp_stride= [KNL] 5834 Set the number of NOCB callback kthreads in 5835 each group, which defaults to the square root 5836 of the number of CPUs. Larger numbers reduce 5837 the wakeup overhead on the global grace-period 5838 kthread, but increases that same overhead on 5839 each group's NOCB grace-period kthread. 5840 5841 rcutree.rcu_kick_kthreads= [KNL] 5842 Cause the grace-period kthread to get an extra 5843 wake_up() if it sleeps three times longer than 5844 it should at force-quiescent-state time. 5845 This wake_up() will be accompanied by a 5846 WARN_ONCE() splat and an ftrace_dump(). 5847 5848 rcutree.rcu_resched_ns= [KNL] 5849 Limit the time spend invoking a batch of RCU 5850 callbacks to the specified number of nanoseconds. 5851 By default, this limit is checked only once 5852 every 32 callbacks in order to limit the pain 5853 inflicted by local_clock() overhead. 5854 5855 rcutree.rcu_unlock_delay= [KNL] 5856 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels, 5857 this specifies an rcu_read_unlock()-time delay 5858 in microseconds. This defaults to zero. 5859 Larger delays increase the probability of 5860 catching RCU pointer leaks, that is, buggy use 5861 of RCU-protected pointers after the relevant 5862 rcu_read_unlock() has completed. 5863 5864 rcutree.sysrq_rcu= [KNL] 5865 Commandeer a sysrq key to dump out Tree RCU's 5866 rcu_node tree with an eye towards determining 5867 why a new grace period has not yet started. 5868 5869 rcutree.use_softirq= [KNL] 5870 If set to zero, move all RCU_SOFTIRQ processing to 5871 per-CPU rcuc kthreads. Defaults to a non-zero 5872 value, meaning that RCU_SOFTIRQ is used by default. 5873 Specify rcutree.use_softirq=0 to use rcuc kthreads. 5874 5875 But note that CONFIG_PREEMPT_RT=y kernels disable 5876 this kernel boot parameter, forcibly setting it 5877 to zero. 5878 5879 rcutree.enable_rcu_lazy= [KNL] 5880 To save power, batch RCU callbacks and flush after 5881 delay, memory pressure or callback list growing too 5882 big. 5883 5884 rcutree.rcu_normal_wake_from_gp= [KNL] 5885 Reduces a latency of synchronize_rcu() call. This approach 5886 maintains its own track of synchronize_rcu() callers, so it 5887 does not interact with regular callbacks because it does not 5888 use a call_rcu[_hurry]() path. Please note, this is for a 5889 normal grace period. 5890 5891 How to disable it: 5892 5893 echo 0 > /sys/module/rcutree/parameters/rcu_normal_wake_from_gp 5894 or pass a boot parameter "rcutree.rcu_normal_wake_from_gp=0" 5895 5896 Default is 1 if it is not explicitly disabled by the boot parameter 5897 passing 0. 5898 5899 rcuscale.gp_async= [KNL] 5900 Measure performance of asynchronous 5901 grace-period primitives such as call_rcu(). 5902 5903 rcuscale.gp_async_max= [KNL] 5904 Specify the maximum number of outstanding 5905 callbacks per writer thread. When a writer 5906 thread exceeds this limit, it invokes the 5907 corresponding flavor of rcu_barrier() to allow 5908 previously posted callbacks to drain. 5909 5910 rcuscale.gp_exp= [KNL] 5911 Measure performance of expedited synchronous 5912 grace-period primitives. 5913 5914 rcuscale.holdoff= [KNL] 5915 Set test-start holdoff period. The purpose of 5916 this parameter is to delay the start of the 5917 test until boot completes in order to avoid 5918 interference. 5919 5920 rcuscale.kfree_by_call_rcu= [KNL] 5921 In kernels built with CONFIG_RCU_LAZY=y, test 5922 call_rcu() instead of kfree_rcu(). 5923 5924 rcuscale.kfree_mult= [KNL] 5925 Instead of allocating an object of size kfree_obj, 5926 allocate one of kfree_mult * sizeof(kfree_obj). 5927 Defaults to 1. 5928 5929 rcuscale.kfree_rcu_test= [KNL] 5930 Set to measure performance of kfree_rcu() flooding. 5931 5932 rcuscale.kfree_rcu_test_double= [KNL] 5933 Test the double-argument variant of kfree_rcu(). 5934 If this parameter has the same value as 5935 rcuscale.kfree_rcu_test_single, both the single- 5936 and double-argument variants are tested. 5937 5938 rcuscale.kfree_rcu_test_single= [KNL] 5939 Test the single-argument variant of kfree_rcu(). 5940 If this parameter has the same value as 5941 rcuscale.kfree_rcu_test_double, both the single- 5942 and double-argument variants are tested. 5943 5944 rcuscale.kfree_nthreads= [KNL] 5945 The number of threads running loops of kfree_rcu(). 5946 5947 rcuscale.kfree_alloc_num= [KNL] 5948 Number of allocations and frees done in an iteration. 5949 5950 rcuscale.kfree_loops= [KNL] 5951 Number of loops doing rcuscale.kfree_alloc_num number 5952 of allocations and frees. 5953 5954 rcuscale.minruntime= [KNL] 5955 Set the minimum test run time in seconds. This 5956 does not affect the data-collection interval, 5957 but instead allows better measurement of things 5958 like CPU consumption. 5959 5960 rcuscale.nreaders= [KNL] 5961 Set number of RCU readers. The value -1 selects 5962 N, where N is the number of CPUs. A value 5963 "n" less than -1 selects N-n+1, where N is again 5964 the number of CPUs. For example, -2 selects N 5965 (the number of CPUs), -3 selects N+1, and so on. 5966 A value of "n" less than or equal to -N selects 5967 a single reader. 5968 5969 rcuscale.nwriters= [KNL] 5970 Set number of RCU writers. The values operate 5971 the same as for rcuscale.nreaders. 5972 N, where N is the number of CPUs 5973 5974 rcuscale.scale_type= [KNL] 5975 Specify the RCU implementation to test. 5976 5977 rcuscale.shutdown= [KNL] 5978 Shut the system down after performance tests 5979 complete. This is useful for hands-off automated 5980 testing. 5981 5982 rcuscale.verbose= [KNL] 5983 Enable additional printk() statements. 5984 5985 rcuscale.writer_holdoff= [KNL] 5986 Write-side holdoff between grace periods, 5987 in microseconds. The default of zero says 5988 no holdoff. 5989 5990 rcuscale.writer_holdoff_jiffies= [KNL] 5991 Additional write-side holdoff between grace 5992 periods, but in jiffies. The default of zero 5993 says no holdoff. 5994 5995 rcutorture.fqs_duration= [KNL] 5996 Set duration of force_quiescent_state bursts 5997 in microseconds. 5998 5999 rcutorture.fqs_holdoff= [KNL] 6000 Set holdoff time within force_quiescent_state bursts 6001 in microseconds. 6002 6003 rcutorture.fqs_stutter= [KNL] 6004 Set wait time between force_quiescent_state bursts 6005 in seconds. 6006 6007 rcutorture.fwd_progress= [KNL] 6008 Specifies the number of kthreads to be used 6009 for RCU grace-period forward-progress testing 6010 for the types of RCU supporting this notion. 6011 Defaults to 1 kthread, values less than zero or 6012 greater than the number of CPUs cause the number 6013 of CPUs to be used. 6014 6015 rcutorture.fwd_progress_div= [KNL] 6016 Specify the fraction of a CPU-stall-warning 6017 period to do tight-loop forward-progress testing. 6018 6019 rcutorture.fwd_progress_holdoff= [KNL] 6020 Number of seconds to wait between successive 6021 forward-progress tests. 6022 6023 rcutorture.fwd_progress_need_resched= [KNL] 6024 Enclose cond_resched() calls within checks for 6025 need_resched() during tight-loop forward-progress 6026 testing. 6027 6028 rcutorture.gp_cond= [KNL] 6029 Use conditional/asynchronous update-side 6030 normal-grace-period primitives, if available. 6031 6032 rcutorture.gp_cond_exp= [KNL] 6033 Use conditional/asynchronous update-side 6034 expedited-grace-period primitives, if available. 6035 6036 rcutorture.gp_cond_full= [KNL] 6037 Use conditional/asynchronous update-side 6038 normal-grace-period primitives that also take 6039 concurrent expedited grace periods into account, 6040 if available. 6041 6042 rcutorture.gp_cond_exp_full= [KNL] 6043 Use conditional/asynchronous update-side 6044 expedited-grace-period primitives that also take 6045 concurrent normal grace periods into account, 6046 if available. 6047 6048 rcutorture.gp_cond_wi= [KNL] 6049 Nominal wait interval for normal conditional 6050 grace periods (specified by rcutorture's 6051 gp_cond and gp_cond_full module parameters), 6052 in microseconds. The actual wait interval will 6053 be randomly selected to nanosecond granularity up 6054 to this wait interval. Defaults to 16 jiffies, 6055 for example, 16,000 microseconds on a system 6056 with HZ=1000. 6057 6058 rcutorture.gp_cond_wi_exp= [KNL] 6059 Nominal wait interval for expedited conditional 6060 grace periods (specified by rcutorture's 6061 gp_cond_exp and gp_cond_exp_full module 6062 parameters), in microseconds. The actual wait 6063 interval will be randomly selected to nanosecond 6064 granularity up to this wait interval. Defaults to 6065 128 microseconds. 6066 6067 rcutorture.gp_exp= [KNL] 6068 Use expedited update-side primitives, if available. 6069 6070 rcutorture.gp_normal= [KNL] 6071 Use normal (non-expedited) asynchronous 6072 update-side primitives, if available. 6073 6074 rcutorture.gp_poll= [KNL] 6075 Use polled update-side normal-grace-period 6076 primitives, if available. 6077 6078 rcutorture.gp_poll_exp= [KNL] 6079 Use polled update-side expedited-grace-period 6080 primitives, if available. 6081 6082 rcutorture.gp_poll_full= [KNL] 6083 Use polled update-side normal-grace-period 6084 primitives that also take concurrent expedited 6085 grace periods into account, if available. 6086 6087 rcutorture.gp_poll_exp_full= [KNL] 6088 Use polled update-side expedited-grace-period 6089 primitives that also take concurrent normal 6090 grace periods into account, if available. 6091 6092 rcutorture.gp_poll_wi= [KNL] 6093 Nominal wait interval for normal conditional 6094 grace periods (specified by rcutorture's 6095 gp_poll and gp_poll_full module parameters), 6096 in microseconds. The actual wait interval will 6097 be randomly selected to nanosecond granularity up 6098 to this wait interval. Defaults to 16 jiffies, 6099 for example, 16,000 microseconds on a system 6100 with HZ=1000. 6101 6102 rcutorture.gp_poll_wi_exp= [KNL] 6103 Nominal wait interval for expedited conditional 6104 grace periods (specified by rcutorture's 6105 gp_poll_exp and gp_poll_exp_full module 6106 parameters), in microseconds. The actual wait 6107 interval will be randomly selected to nanosecond 6108 granularity up to this wait interval. Defaults to 6109 128 microseconds. 6110 6111 rcutorture.gp_sync= [KNL] 6112 Use normal (non-expedited) synchronous 6113 update-side primitives, if available. If all 6114 of rcutorture.gp_cond=, rcutorture.gp_exp=, 6115 rcutorture.gp_normal=, and rcutorture.gp_sync= 6116 are zero, rcutorture acts as if is interpreted 6117 they are all non-zero. 6118 6119 rcutorture.gpwrap_lag= [KNL] 6120 Enable grace-period wrap lag testing. Setting 6121 to false prevents the gpwrap lag test from 6122 running. Default is true. 6123 6124 rcutorture.gpwrap_lag_gps= [KNL] 6125 Set the value for grace-period wrap lag during 6126 active lag testing periods. This controls how many 6127 grace periods differences we tolerate between 6128 rdp and rnp's gp_seq before setting overflow flag. 6129 The default is always set to 8. 6130 6131 rcutorture.gpwrap_lag_cycle_mins= [KNL] 6132 Set the total cycle duration for gpwrap lag 6133 testing in minutes. This is the total time for 6134 one complete cycle of active and inactive 6135 testing periods. Default is 30 minutes. 6136 6137 rcutorture.gpwrap_lag_active_mins= [KNL] 6138 Set the duration for which gpwrap lag is active 6139 within each cycle, in minutes. During this time, 6140 the grace-period wrap lag will be set to the 6141 value specified by gpwrap_lag_gps. Default is 6142 5 minutes. 6143 6144 rcutorture.irqreader= [KNL] 6145 Run RCU readers from irq handlers, or, more 6146 accurately, from a timer handler. Not all RCU 6147 flavors take kindly to this sort of thing. 6148 6149 rcutorture.leakpointer= [KNL] 6150 Leak an RCU-protected pointer out of the reader. 6151 This can of course result in splats, and is 6152 intended to test the ability of things like 6153 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect 6154 such leaks. 6155 6156 rcutorture.n_barrier_cbs= [KNL] 6157 Set callbacks/threads for rcu_barrier() testing. 6158 6159 rcutorture.nfakewriters= [KNL] 6160 Set number of concurrent RCU writers. These just 6161 stress RCU, they don't participate in the actual 6162 test, hence the "fake". 6163 6164 rcutorture.nocbs_nthreads= [KNL] 6165 Set number of RCU callback-offload togglers. 6166 Zero (the default) disables toggling. 6167 6168 rcutorture.nocbs_toggle= [KNL] 6169 Set the delay in milliseconds between successive 6170 callback-offload toggling attempts. 6171 6172 rcutorture.nreaders= [KNL] 6173 Set number of RCU readers. The value -1 selects 6174 N-1, where N is the number of CPUs. A value 6175 "n" less than -1 selects N-n-2, where N is again 6176 the number of CPUs. For example, -2 selects N 6177 (the number of CPUs), -3 selects N+1, and so on. 6178 6179 rcutorture.nwriters= [KNL] 6180 Set number of RCU writers, which must be either 6181 zero or one. For additional writers, use instead 6182 the rcutorture.nfakewriters parameter. 6183 6184 rcutorture.object_debug= [KNL] 6185 Enable debug-object double-call_rcu() testing. 6186 6187 rcutorture.onoff_holdoff= [KNL] 6188 Set time (s) after boot for CPU-hotplug testing. 6189 6190 rcutorture.onoff_interval= [KNL] 6191 Set time (jiffies) between CPU-hotplug operations, 6192 or zero to disable CPU-hotplug testing. 6193 6194 rcutorture.preempt_duration= [KNL] 6195 Set duration (in milliseconds) of preemptions 6196 by a high-priority FIFO real-time task. Set to 6197 zero (the default) to disable. The CPUs to 6198 preempt are selected randomly from the set that 6199 are online at a given point in time. Races with 6200 CPUs going offline are ignored, with that attempt 6201 at preemption skipped. 6202 6203 rcutorture.preempt_interval= [KNL] 6204 Set interval (in milliseconds, defaulting to one 6205 second) between preemptions by a high-priority 6206 FIFO real-time task. This delay is mediated 6207 by an hrtimer and is further fuzzed to avoid 6208 inadvertent synchronizations. 6209 6210 rcutorture.read_exit_burst= [KNL] 6211 The number of times in a given read-then-exit 6212 episode that a set of read-then-exit kthreads 6213 is spawned. 6214 6215 rcutorture.read_exit_delay= [KNL] 6216 The delay, in seconds, between successive 6217 read-then-exit testing episodes. 6218 6219 rcutorture.reader_flavor= [KNL] 6220 A bit mask indicating which readers to use. 6221 If there is more than one bit set, the readers 6222 are entered from low-order bit up, and are 6223 exited in the opposite order. For SRCU, the 6224 0x1 bit is normal readers, 0x2 NMI-safe readers, 6225 and 0x4 light-weight readers. 6226 6227 rcutorture.shuffle_interval= [KNL] 6228 Set task-shuffle interval (s). Shuffling tasks 6229 allows some CPUs to go into dyntick-idle mode 6230 during the rcutorture test. 6231 6232 rcutorture.shutdown_secs= [KNL] 6233 Set time (s) after boot system shutdown. This 6234 is useful for hands-off automated testing. 6235 6236 rcutorture.stall_cpu= [KNL] 6237 Duration of CPU stall (s) to test RCU CPU stall 6238 warnings, zero to disable. 6239 6240 rcutorture.stall_cpu_block= [KNL] 6241 Sleep while stalling if set. This will result 6242 in warnings from preemptible RCU in addition to 6243 any other stall-related activity. Note that 6244 in kernels built with CONFIG_PREEMPTION=n and 6245 CONFIG_PREEMPT_COUNT=y, this parameter will 6246 cause the CPU to pass through a quiescent state. 6247 Given CONFIG_PREEMPTION=n, this will suppress 6248 RCU CPU stall warnings, but will instead result 6249 in scheduling-while-atomic splats. 6250 6251 Use of this module parameter results in splats. 6252 6253 6254 rcutorture.stall_cpu_holdoff= [KNL] 6255 Time to wait (s) after boot before inducing stall. 6256 6257 rcutorture.stall_cpu_irqsoff= [KNL] 6258 Disable interrupts while stalling if set, but only 6259 on the first stall in the set. 6260 6261 rcutorture.stall_cpu_repeat= [KNL] 6262 Number of times to repeat the stall sequence, 6263 so that rcutorture.stall_cpu_repeat=3 will result 6264 in four stall sequences. 6265 6266 rcutorture.stall_gp_kthread= [KNL] 6267 Duration (s) of forced sleep within RCU 6268 grace-period kthread to test RCU CPU stall 6269 warnings, zero to disable. If both stall_cpu 6270 and stall_gp_kthread are specified, the 6271 kthread is starved first, then the CPU. 6272 6273 rcutorture.stall_only= [KNL] 6274 Shut off all rcutorture kthreads other than the 6275 RCU CPU stall-warning test kthreads. The purpose 6276 of this is to test production applictions' 6277 reactions to CPU stalls, and with minimal 6278 additional overhead. Or you can omit the 6279 stall-warning tests as well and get a heavy 6280 no-op, your choice! 6281 6282 rcutorture.stat_interval= [KNL] 6283 Time (s) between statistics printk()s. 6284 6285 rcutorture.stutter= [KNL] 6286 Time (s) to stutter testing, for example, specifying 6287 five seconds causes the test to run for five seconds, 6288 wait for five seconds, and so on. This tests RCU's 6289 ability to transition abruptly to and from idle. 6290 6291 rcutorture.test_boost= [KNL] 6292 Test RCU priority boosting? 0=no, 1=maybe, 2=yes. 6293 "Maybe" means test if the RCU implementation 6294 under test support RCU priority boosting. 6295 6296 rcutorture.test_boost_duration= [KNL] 6297 Duration (s) of each individual boost test. 6298 6299 rcutorture.test_boost_holdoff= [KNL] 6300 Holdoff time (s) from start of test to the start 6301 of RCU priority-boost testing. Defaults to zero, 6302 that is, no holdoff. 6303 6304 rcutorture.test_boost_interval= [KNL] 6305 Interval (s) between each boost test. 6306 6307 rcutorture.test_no_idle_hz= [KNL] 6308 Test RCU's dyntick-idle handling. See also the 6309 rcutorture.shuffle_interval parameter. 6310 6311 rcutorture.torture_type= [KNL] 6312 Specify the RCU implementation to test. 6313 6314 rcutorture.verbose= [KNL] 6315 Enable additional printk() statements. 6316 6317 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL] 6318 Dump ftrace buffer after reporting RCU CPU 6319 stall warning. 6320 6321 rcupdate.rcu_cpu_stall_notifiers= [KNL] 6322 Provide RCU CPU stall notifiers, but see the 6323 warnings in the RCU_CPU_STALL_NOTIFIER Kconfig 6324 option's help text. TL;DR: You almost certainly 6325 do not want rcupdate.rcu_cpu_stall_notifiers. 6326 6327 rcupdate.rcu_cpu_stall_suppress= [KNL] 6328 Suppress RCU CPU stall warning messages. 6329 6330 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL] 6331 Suppress RCU CPU stall warning messages and 6332 rcutorture writer stall warnings that occur 6333 during early boot, that is, during the time 6334 before the init task is spawned. 6335 6336 rcupdate.rcu_cpu_stall_timeout= [KNL] 6337 Set timeout for RCU CPU stall warning messages. 6338 The value is in seconds and the maximum allowed 6339 value is 300 seconds. 6340 6341 rcupdate.rcu_exp_cpu_stall_timeout= [KNL] 6342 Set timeout for expedited RCU CPU stall warning 6343 messages. The value is in milliseconds 6344 and the maximum allowed value is 21000 6345 milliseconds. Please note that this value is 6346 adjusted to an arch timer tick resolution. 6347 Setting this to zero causes the value from 6348 rcupdate.rcu_cpu_stall_timeout to be used (after 6349 conversion from seconds to milliseconds). 6350 6351 rcupdate.rcu_cpu_stall_cputime= [KNL] 6352 Provide statistics on the cputime and count of 6353 interrupts and tasks during the sampling period. For 6354 multiple continuous RCU stalls, all sampling periods 6355 begin at half of the first RCU stall timeout. 6356 6357 rcupdate.rcu_exp_stall_task_details= [KNL] 6358 Print stack dumps of any tasks blocking the 6359 current expedited RCU grace period during an 6360 expedited RCU CPU stall warning. 6361 6362 rcupdate.rcu_expedited= [KNL] 6363 Use expedited grace-period primitives, for 6364 example, synchronize_rcu_expedited() instead 6365 of synchronize_rcu(). This reduces latency, 6366 but can increase CPU utilization, degrade 6367 real-time latency, and degrade energy efficiency. 6368 No effect on CONFIG_TINY_RCU kernels. 6369 6370 rcupdate.rcu_normal= [KNL] 6371 Use only normal grace-period primitives, 6372 for example, synchronize_rcu() instead of 6373 synchronize_rcu_expedited(). This improves 6374 real-time latency, CPU utilization, and 6375 energy efficiency, but can expose users to 6376 increased grace-period latency. This parameter 6377 overrides rcupdate.rcu_expedited. No effect on 6378 CONFIG_TINY_RCU kernels. 6379 6380 rcupdate.rcu_normal_after_boot= [KNL] 6381 Once boot has completed (that is, after 6382 rcu_end_inkernel_boot() has been invoked), use 6383 only normal grace-period primitives. No effect 6384 on CONFIG_TINY_RCU kernels. 6385 6386 But note that CONFIG_PREEMPT_RT=y kernels enables 6387 this kernel boot parameter, forcibly setting 6388 it to the value one, that is, converting any 6389 post-boot attempt at an expedited RCU grace 6390 period to instead use normal non-expedited 6391 grace-period processing. 6392 6393 rcupdate.rcu_task_collapse_lim= [KNL] 6394 Set the maximum number of callbacks present 6395 at the beginning of a grace period that allows 6396 the RCU Tasks flavors to collapse back to using 6397 a single callback queue. This switching only 6398 occurs when rcupdate.rcu_task_enqueue_lim is 6399 set to the default value of -1. 6400 6401 rcupdate.rcu_task_contend_lim= [KNL] 6402 Set the minimum number of callback-queuing-time 6403 lock-contention events per jiffy required to 6404 cause the RCU Tasks flavors to switch to per-CPU 6405 callback queuing. This switching only occurs 6406 when rcupdate.rcu_task_enqueue_lim is set to 6407 the default value of -1. 6408 6409 rcupdate.rcu_task_enqueue_lim= [KNL] 6410 Set the number of callback queues to use for the 6411 RCU Tasks family of RCU flavors. The default 6412 of -1 allows this to be automatically (and 6413 dynamically) adjusted. This parameter is intended 6414 for use in testing. 6415 6416 rcupdate.rcu_task_lazy_lim= [KNL] 6417 Number of callbacks on a given CPU that will 6418 cancel laziness on that CPU. Use -1 to disable 6419 cancellation of laziness, but be advised that 6420 doing so increases the danger of OOM due to 6421 callback flooding. 6422 6423 rcupdate.rcu_task_stall_info= [KNL] 6424 Set initial timeout in jiffies for RCU task stall 6425 informational messages, which give some indication 6426 of the problem for those not patient enough to 6427 wait for ten minutes. Informational messages are 6428 only printed prior to the stall-warning message 6429 for a given grace period. Disable with a value 6430 less than or equal to zero. Defaults to ten 6431 seconds. A change in value does not take effect 6432 until the beginning of the next grace period. 6433 6434 rcupdate.rcu_task_stall_info_mult= [KNL] 6435 Multiplier for time interval between successive 6436 RCU task stall informational messages for a given 6437 RCU tasks grace period. This value is clamped 6438 to one through ten, inclusive. It defaults to 6439 the value three, so that the first informational 6440 message is printed 10 seconds into the grace 6441 period, the second at 40 seconds, the third at 6442 160 seconds, and then the stall warning at 600 6443 seconds would prevent a fourth at 640 seconds. 6444 6445 rcupdate.rcu_task_stall_timeout= [KNL] 6446 Set timeout in jiffies for RCU task stall 6447 warning messages. Disable with a value less 6448 than or equal to zero. Defaults to ten minutes. 6449 A change in value does not take effect until 6450 the beginning of the next grace period. 6451 6452 rcupdate.rcu_tasks_lazy_ms= [KNL] 6453 Set timeout in milliseconds RCU Tasks asynchronous 6454 callback batching for call_rcu_tasks(). 6455 A negative value will take the default. A value 6456 of zero will disable batching. Batching is 6457 always disabled for synchronize_rcu_tasks(). 6458 6459 rcupdate.rcu_self_test= [KNL] 6460 Run the RCU early boot self tests 6461 6462 rdinit= [KNL] 6463 Format: <full_path> 6464 Run specified binary instead of /init from the ramdisk, 6465 used for early userspace startup. See initrd. 6466 6467 rdrand= [X86,EARLY] 6468 force - Override the decision by the kernel to hide the 6469 advertisement of RDRAND support (this affects 6470 certain AMD processors because of buggy BIOS 6471 support, specifically around the suspend/resume 6472 path). 6473 6474 rdt= [HW,X86,RDT] 6475 Turn on/off individual RDT features. List is: 6476 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp, 6477 mba, smba, bmec, abmc, sdciae, energy[:guid], 6478 perf[:guid]. 6479 E.g. to turn on cmt and turn off mba use: 6480 rdt=cmt,!mba 6481 To turn off all energy telemetry monitoring and ensure that 6482 perf telemetry monitoring associated with guid 0x12345 6483 is enabled use: 6484 rdt=!energy,perf:0x12345 6485 6486 reboot= [KNL] 6487 Format (x86 or x86_64): 6488 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \ 6489 [[,]s[mp]####] \ 6490 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \ 6491 [[,]f[orce]] 6492 Where reboot_mode is one of warm (soft) or cold (hard) or gpio 6493 (prefix with 'panic_' to set mode for panic 6494 reboot only), 6495 reboot_type is one of bios, acpi, kbd, triple, efi, or pci, 6496 reboot_force is either force or not specified, 6497 reboot_cpu is s[mp]#### with #### being the processor 6498 to be used for rebooting. 6499 6500 acpi 6501 Use the ACPI RESET_REG in the FADT. If ACPI is not 6502 configured or the ACPI reset does not work, the reboot 6503 path attempts the reset using the keyboard controller. 6504 6505 bios 6506 Use the CPU reboot vector for warm reset 6507 6508 cold 6509 Set the cold reboot flag 6510 6511 default 6512 There are some built-in platform specific "quirks" 6513 - you may see: "reboot: <name> series board detected. 6514 Selecting <type> for reboots." In the case where you 6515 think the quirk is in error (e.g. you have newer BIOS, 6516 or newer board) using this option will ignore the 6517 built-in quirk table, and use the generic default 6518 reboot actions. 6519 6520 efi 6521 Use efi reset_system runtime service. If EFI is not 6522 configured or the EFI reset does not work, the reboot 6523 path attempts the reset using the keyboard controller. 6524 6525 force 6526 Don't stop other CPUs on reboot. This can make reboot 6527 more reliable in some cases. 6528 6529 kbd 6530 Use the keyboard controller. cold reset (default) 6531 6532 pci 6533 Use a write to the PCI config space register 0xcf9 to 6534 trigger reboot. 6535 6536 triple 6537 Force a triple fault (init) 6538 6539 warm 6540 Don't set the cold reboot flag 6541 6542 Using warm reset will be much faster especially on big 6543 memory systems because the BIOS will not go through 6544 the memory check. Disadvantage is that not all 6545 hardware will be completely reinitialized on reboot so 6546 there may be boot problems on some systems. 6547 6548 6549 refscale.holdoff= [KNL] 6550 Set test-start holdoff period. The purpose of 6551 this parameter is to delay the start of the 6552 test until boot completes in order to avoid 6553 interference. 6554 6555 refscale.lookup_instances= [KNL] 6556 Number of data elements to use for the forms of 6557 SLAB_TYPESAFE_BY_RCU testing. A negative number 6558 is negated and multiplied by nr_cpu_ids, while 6559 zero specifies nr_cpu_ids. 6560 6561 refscale.loops= [KNL] 6562 Set the number of loops over the synchronization 6563 primitive under test. Increasing this number 6564 reduces noise due to loop start/end overhead, 6565 but the default has already reduced the per-pass 6566 noise to a handful of picoseconds on ca. 2020 6567 x86 laptops. 6568 6569 refscale.nreaders= [KNL] 6570 Set number of readers. The default value of -1 6571 selects N, where N is roughly 75% of the number 6572 of CPUs. A value of zero is an interesting choice. 6573 6574 refscale.nruns= [KNL] 6575 Set number of runs, each of which is dumped onto 6576 the console log. 6577 6578 refscale.readdelay= [KNL] 6579 Set the read-side critical-section duration, 6580 measured in microseconds. 6581 6582 refscale.scale_type= [KNL] 6583 Specify the read-protection implementation to test. 6584 6585 refscale.shutdown= [KNL] 6586 Shut down the system at the end of the performance 6587 test. This defaults to 1 (shut it down) when 6588 refscale is built into the kernel and to 0 (leave 6589 it running) when refscale is built as a module. 6590 6591 refscale.verbose= [KNL] 6592 Enable additional printk() statements. 6593 6594 refscale.verbose_batched= [KNL] 6595 Batch the additional printk() statements. If zero 6596 (the default) or negative, print everything. Otherwise, 6597 print every Nth verbose statement, where N is the value 6598 specified. 6599 6600 regulator_ignore_unused 6601 [REGULATOR] 6602 Prevents regulator framework from disabling regulators 6603 that are unused, due no driver claiming them. This may 6604 be useful for debug and development, but should not be 6605 needed on a platform with proper driver support. 6606 6607 relax_domain_level= 6608 [KNL, SMP] Set scheduler's default relax_domain_level. 6609 See Documentation/admin-guide/cgroup-v1/cpusets.rst. 6610 6611 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory 6612 Format: <base1>,<size1>[,<base2>,<size2>,...] 6613 Reserve I/O ports or memory so the kernel won't use 6614 them. If <base> is less than 0x10000, the region 6615 is assumed to be I/O ports; otherwise it is memory. 6616 6617 reserve_mem= [RAM] 6618 Format: nn[KMG]:<align>:<label> 6619 Reserve physical memory and label it with a name that 6620 other subsystems can use to access it. This is typically 6621 used for systems that do not wipe the RAM, and this command 6622 line will try to reserve the same physical memory on 6623 soft reboots. Note, it is not guaranteed to be the same 6624 location. For example, if anything about the system changes 6625 or if booting a different kernel. It can also fail if KASLR 6626 places the kernel at the location of where the RAM reservation 6627 was from a previous boot, the new reservation will be at a 6628 different location. 6629 Any subsystem using this feature must add a way to verify 6630 that the contents of the physical memory is from a previous 6631 boot, as there may be cases where the memory will not be 6632 located at the same location. 6633 6634 The format is size:align:label for example, to request 6635 12 megabytes of 4096 alignment for ramoops: 6636 6637 reserve_mem=12M:4096:oops ramoops.mem_name=oops 6638 6639 reservetop= [X86-32,EARLY] 6640 Format: nn[KMG] 6641 Reserves a hole at the top of the kernel virtual 6642 address space. 6643 6644 reset_devices [KNL] Force drivers to reset the underlying device 6645 during initialization. 6646 6647 resume= [SWSUSP] 6648 Specify the partition device for software suspend 6649 Format: 6650 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>} 6651 6652 resume_offset= [SWSUSP] 6653 Specify the offset from the beginning of the partition 6654 given by "resume=" at which the swap header is located, 6655 in <PAGE_SIZE> units (needed only for swap files). 6656 See Documentation/power/swsusp-and-swap-files.rst 6657 6658 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to 6659 read the resume files 6660 6661 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up. 6662 Useful for devices that are detected asynchronously 6663 (e.g. USB and MMC devices). 6664 6665 retain_initrd [RAM] Keep initrd memory after extraction. After boot, it will 6666 be accessible via /sys/firmware/initrd. 6667 6668 retbleed= [X86] Control mitigation of RETBleed (Arbitrary 6669 Speculative Code Execution with Return Instructions) 6670 vulnerability. 6671 6672 AMD-based UNRET and IBPB mitigations alone do not stop 6673 sibling threads from influencing the predictions of other 6674 sibling threads. For that reason, STIBP is used on pro- 6675 cessors that support it, and mitigate SMT on processors 6676 that don't. 6677 6678 off - no mitigation 6679 auto - automatically select a mitigation 6680 auto,nosmt - automatically select a mitigation, 6681 disabling SMT if necessary for 6682 the full mitigation (only on Zen1 6683 and older without STIBP). 6684 ibpb - On AMD, mitigate short speculation 6685 windows on basic block boundaries too. 6686 Safe, highest perf impact. It also 6687 enables STIBP if present. Not suitable 6688 on Intel. 6689 ibpb,nosmt - Like "ibpb" above but will disable SMT 6690 when STIBP is not available. This is 6691 the alternative for systems which do not 6692 have STIBP. 6693 unret - Force enable untrained return thunks, 6694 only effective on AMD f15h-f17h based 6695 systems. 6696 unret,nosmt - Like unret, but will disable SMT when STIBP 6697 is not available. This is the alternative for 6698 systems which do not have STIBP. 6699 6700 Selecting 'auto' will choose a mitigation method at run 6701 time according to the CPU. 6702 6703 Not specifying this option is equivalent to retbleed=auto. 6704 6705 rfkill.default_state= 6706 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm, 6707 etc. communication is blocked by default. 6708 1 Unblocked. 6709 6710 rfkill.master_switch_mode= 6711 0 The "airplane mode" button does nothing. 6712 1 The "airplane mode" button toggles between everything 6713 blocked and the previous configuration. 6714 2 The "airplane mode" button toggles between everything 6715 blocked and everything unblocked. 6716 6717 ring3mwait=disable 6718 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported 6719 CPUs. 6720 6721 riscv_isa_fallback [RISCV,EARLY] 6722 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit 6723 falling back to detecting extension support by parsing 6724 "riscv,isa" property on devicetree systems when the 6725 replacement properties are not found. See the Kconfig 6726 entry for RISCV_ISA_FALLBACK. 6727 6728 riscv_nousercfi= 6729 all Disable user CFI ABI to userspace even if cpu extension 6730 are available. 6731 bcfi Disable user backward CFI ABI to userspace even if 6732 the shadow stack extension is available. 6733 fcfi Disable user forward CFI ABI to userspace even if the 6734 landing pad extension is available. 6735 6736 ro [KNL] Mount root device read-only on boot 6737 6738 rodata= [KNL,EARLY] 6739 on Mark read-only kernel memory as read-only (default). 6740 off Leave read-only kernel memory writable for debugging. 6741 noalias Mark read-only kernel memory as read-only but retain 6742 writable aliases in the direct map for regions outside 6743 of the kernel image. [arm64] 6744 6745 rockchip.usb_uart 6746 [EARLY] 6747 Enable the uart passthrough on the designated usb port 6748 on Rockchip SoCs. When active, the signals of the 6749 debug-uart get routed to the D+ and D- pins of the usb 6750 port and the regular usb controller gets disabled. 6751 6752 root= [KNL] Root filesystem 6753 Usually this is a block device specifier of some kind, 6754 see the early_lookup_bdev comment in 6755 block/early-lookup.c for details. 6756 Alternatively this can be "ram" for the legacy initial 6757 ramdisk, "nfs" and "cifs" for root on a network file 6758 system, or "mtd" and "ubi" for mounting from raw flash. 6759 6760 rootdelay= [KNL] Delay (in seconds) to pause before attempting to 6761 mount the root filesystem 6762 6763 rootflags= [KNL] Set root filesystem mount option string 6764 6765 rseq_slice_ext= [KNL] RSEQ based time slice extension 6766 Format: boolean 6767 Control enablement of RSEQ based time slice extension. 6768 Default is 'on'. 6769 6770 initramfs_options= [KNL] 6771 Specify mount options for the initramfs mount. 6772 6773 rootfstype= [KNL] Set root filesystem type 6774 6775 rootwait [KNL] Wait (indefinitely) for root device to show up. 6776 Useful for devices that are detected asynchronously 6777 (e.g. USB and MMC devices). 6778 6779 rootwait= [KNL] Maximum time (in seconds) to wait for root device 6780 to show up before attempting to mount the root 6781 filesystem. 6782 6783 rproc_mem=nn[KMG][@address] 6784 [KNL,ARM,CMA] Remoteproc physical memory block. 6785 Memory area to be used by remote processor image, 6786 managed by CMA. 6787 6788 rseq_debug= [KNL] Enable or disable restartable sequence 6789 debug mode. Defaults to CONFIG_RSEQ_DEBUG_DEFAULT_ENABLE. 6790 Format: <bool> 6791 6792 rt_group_sched= [KNL] Enable or disable SCHED_RR/FIFO group scheduling 6793 when CONFIG_RT_GROUP_SCHED=y. Defaults to 6794 !CONFIG_RT_GROUP_SCHED_DEFAULT_DISABLED. 6795 Format: <bool> 6796 6797 rw [KNL] Mount root device read-write on boot 6798 6799 S [KNL] Run init in single mode 6800 6801 s390_iommu= [HW,S390] 6802 Set s390 IOTLB flushing mode 6803 strict 6804 With strict flushing every unmap operation will result 6805 in an IOTLB flush. Default is lazy flushing before 6806 reuse, which is faster. Deprecated, equivalent to 6807 iommu.strict=1. 6808 6809 s390_iommu_aperture= [KNL,S390] 6810 Specifies the size of the per device DMA address space 6811 accessible through the DMA and IOMMU APIs as a decimal 6812 factor of the size of main memory. 6813 The default is 1 meaning that one can concurrently use 6814 as many DMA addresses as physical memory is installed, 6815 if supported by hardware, and thus map all of memory 6816 once. With a value of 2 one can map all of memory twice 6817 and so on. As a special case a factor of 0 imposes no 6818 restrictions other than those given by hardware at the 6819 cost of significant additional memory use for tables. 6820 6821 6822 sched_proxy_exec= [KNL] 6823 Enables or disables "proxy execution" style 6824 solution to mutex-based priority inversion. 6825 Format: <bool> 6826 6827 sched_verbose [KNL,EARLY] Enables verbose scheduler debug messages. 6828 6829 schedstats= [KNL,X86] Enable or disable scheduled statistics. 6830 Allowed values are enable and disable. This feature 6831 incurs a small amount of overhead in the scheduler 6832 but is useful for debugging and performance tuning. 6833 6834 sched_thermal_decay_shift= 6835 [Deprecated] 6836 [KNL, SMP] Set a decay shift for scheduler thermal 6837 pressure signal. Thermal pressure signal follows the 6838 default decay period of other scheduler pelt 6839 signals(usually 32 ms but configurable). Setting 6840 sched_thermal_decay_shift will left shift the decay 6841 period for the thermal pressure signal by the shift 6842 value. 6843 i.e. with the default pelt decay period of 32 ms 6844 sched_thermal_decay_shift thermal pressure decay pr 6845 1 64 ms 6846 2 128 ms 6847 and so on. 6848 Format: integer between 0 and 10 6849 Default is 0. 6850 6851 scftorture.holdoff= [KNL] 6852 Number of seconds to hold off before starting 6853 test. Defaults to zero for module insertion and 6854 to 10 seconds for built-in smp_call_function() 6855 tests. 6856 6857 scftorture.longwait= [KNL] 6858 Request ridiculously long waits randomly selected 6859 up to the chosen limit in seconds. Zero (the 6860 default) disables this feature. Please note 6861 that requesting even small non-zero numbers of 6862 seconds can result in RCU CPU stall warnings, 6863 softlockup complaints, and so on. 6864 6865 scftorture.nthreads= [KNL] 6866 Number of kthreads to spawn to invoke the 6867 smp_call_function() family of functions. 6868 The default of -1 specifies a number of kthreads 6869 equal to the number of CPUs. 6870 6871 scftorture.onoff_holdoff= [KNL] 6872 Number seconds to wait after the start of the 6873 test before initiating CPU-hotplug operations. 6874 6875 scftorture.onoff_interval= [KNL] 6876 Number seconds to wait between successive 6877 CPU-hotplug operations. Specifying zero (which 6878 is the default) disables CPU-hotplug operations. 6879 6880 scftorture.shutdown_secs= [KNL] 6881 The number of seconds following the start of the 6882 test after which to shut down the system. The 6883 default of zero avoids shutting down the system. 6884 Non-zero values are useful for automated tests. 6885 6886 scftorture.stat_interval= [KNL] 6887 The number of seconds between outputting the 6888 current test statistics to the console. A value 6889 of zero disables statistics output. 6890 6891 scftorture.stutter_cpus= [KNL] 6892 The number of jiffies to wait between each change 6893 to the set of CPUs under test. 6894 6895 scftorture.use_cpus_read_lock= [KNL] 6896 Use use_cpus_read_lock() instead of the default 6897 preempt_disable() to disable CPU hotplug 6898 while invoking one of the smp_call_function*() 6899 functions. 6900 6901 scftorture.verbose= [KNL] 6902 Enable additional printk() statements. 6903 6904 scftorture.weight_single= [KNL] 6905 The probability weighting to use for the 6906 smp_call_function_single() function with a zero 6907 "wait" parameter. A value of -1 selects the 6908 default if all other weights are -1. However, 6909 if at least one weight has some other value, a 6910 value of -1 will instead select a weight of zero. 6911 6912 scftorture.weight_single_wait= [KNL] 6913 The probability weighting to use for the 6914 smp_call_function_single() function with a 6915 non-zero "wait" parameter. See weight_single. 6916 6917 scftorture.weight_many= [KNL] 6918 The probability weighting to use for the 6919 smp_call_function_many() function with a zero 6920 "wait" parameter. See weight_single. 6921 Note well that setting a high probability for 6922 this weighting can place serious IPI load 6923 on the system. 6924 6925 scftorture.weight_many_wait= [KNL] 6926 The probability weighting to use for the 6927 smp_call_function_many() function with a 6928 non-zero "wait" parameter. See weight_single 6929 and weight_many. 6930 6931 scftorture.weight_all= [KNL] 6932 The probability weighting to use for the 6933 smp_call_function_all() function with a zero 6934 "wait" parameter. See weight_single and 6935 weight_many. 6936 6937 scftorture.weight_all_wait= [KNL] 6938 The probability weighting to use for the 6939 smp_call_function_all() function with a 6940 non-zero "wait" parameter. See weight_single 6941 and weight_many. 6942 6943 sdw_mclk_divider=[SDW] 6944 Specify the MCLK divider for Intel SoundWire buses in 6945 case the BIOS does not provide the clock rate properly. 6946 6947 skew_tick= [KNL,EARLY] Offset the periodic timer tick per cpu to mitigate 6948 xtime_lock contention on larger systems, and/or RCU lock 6949 contention on all systems with CONFIG_MAXSMP set. 6950 Format: { "0" | "1" } 6951 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1") 6952 1 -- enable. 6953 Note: increases power consumption, thus should only be 6954 enabled if running jitter sensitive (HPC/RT) workloads. 6955 6956 security= [SECURITY] Choose a legacy "major" security module to 6957 enable at boot. This has been deprecated by the 6958 "lsm=" parameter. 6959 6960 selinux= [SELINUX] Disable or enable SELinux at boot time. 6961 Format: { "0" | "1" } 6962 See security/selinux/Kconfig help text. 6963 0 -- disable. 6964 1 -- enable. 6965 Default value is 1. 6966 6967 serialnumber [BUGS=X86-32] 6968 6969 sev=option[,option...] [X86-64] 6970 6971 debug 6972 Enable debug messages. 6973 6974 nosnp 6975 Do not enable SEV-SNP (applies to host/hypervisor 6976 only). Setting 'nosnp' avoids the RMP check overhead 6977 in memory accesses when users do not want to run 6978 SEV-SNP guests. 6979 6980 shapers= [NET] 6981 Maximal number of shapers. 6982 6983 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller 6984 Limit apic dumping. The parameter defines the maximal 6985 number of local apics being dumped. Also it is possible 6986 to set it to "all" by meaning -- no limit here. 6987 Format: { 1 (default) | 2 | ... | all }. 6988 The parameter valid if only apic=debug or 6989 apic=verbose is specified. 6990 Example: apic=debug show_lapic=all 6991 6992 slab_debug[=options[,slabs][;[options[,slabs]]...]] [MM] 6993 Enabling slab_debug allows one to determine the 6994 culprit if slab objects become corrupted. Enabling 6995 slab_debug can create guard zones around objects and 6996 may poison objects when not in use. Also tracks the 6997 last alloc / free. For more information see 6998 Documentation/admin-guide/mm/slab.rst. 6999 (slub_debug legacy name also accepted for now) 7000 7001 Using this option implies the "no_hash_pointers" 7002 option which can be undone by adding the 7003 "hash_pointers=always" option. 7004 7005 slab_max_order= [MM] 7006 Determines the maximum allowed order for slabs. 7007 A high setting may cause OOMs due to memory 7008 fragmentation. For more information see 7009 Documentation/admin-guide/mm/slab.rst. 7010 (slub_max_order legacy name also accepted for now) 7011 7012 slab_merge [MM] 7013 Enable merging of slabs with similar size when the 7014 kernel is built without CONFIG_SLAB_MERGE_DEFAULT. 7015 (slub_merge legacy name also accepted for now) 7016 7017 slab_min_objects= [MM] 7018 The minimum number of objects per slab. SLUB will 7019 increase the slab order up to slab_max_order to 7020 generate a sufficiently large slab able to contain 7021 the number of objects indicated. The higher the number 7022 of objects the smaller the overhead of tracking slabs 7023 and the less frequently locks need to be acquired. 7024 For more information see 7025 Documentation/admin-guide/mm/slab.rst. 7026 (slub_min_objects legacy name also accepted for now) 7027 7028 slab_min_order= [MM] 7029 Determines the minimum page order for slabs. Must be 7030 lower or equal to slab_max_order. For more information see 7031 Documentation/admin-guide/mm/slab.rst. 7032 (slub_min_order legacy name also accepted for now) 7033 7034 slab_nomerge [MM] 7035 Disable merging of slabs with similar size. May be 7036 necessary if there is some reason to distinguish 7037 allocs to different slabs, especially in hardened 7038 environments where the risk of heap overflows and 7039 layout control by attackers can usually be 7040 frustrated by disabling merging. This will reduce 7041 most of the exposure of a heap attack to a single 7042 cache (risks via metadata attacks are mostly 7043 unchanged). Debug options disable merging on their 7044 own. 7045 For more information see 7046 Documentation/admin-guide/mm/slab.rst. 7047 (slub_nomerge legacy name also accepted for now) 7048 7049 slab_strict_numa [MM] 7050 Support memory policies on a per object level 7051 in the slab allocator. The default is for memory 7052 policies to be applied at the folio level when 7053 a new folio is needed or a partial folio is 7054 retrieved from the lists. Increases overhead 7055 in the slab fastpaths but gains more accurate 7056 NUMA kernel object placement which helps with slow 7057 interconnects in NUMA systems. 7058 7059 slram= [HW,MTD] 7060 7061 smart2= [HW] 7062 Format: <io1>[,<io2>[,...,<io8>]] 7063 7064 smp.csd_lock_timeout= [KNL] 7065 Specify the period of time in milliseconds 7066 that smp_call_function() and friends will wait 7067 for a CPU to release the CSD lock. This is 7068 useful when diagnosing bugs involving CPUs 7069 disabling interrupts for extended periods 7070 of time. Defaults to 5,000 milliseconds, and 7071 setting a value of zero disables this feature. 7072 This feature may be more efficiently disabled 7073 using the csdlock_debug- kernel parameter. 7074 7075 smp.panic_on_ipistall= [KNL] 7076 If a csd_lock_timeout extends for more than 7077 the specified number of milliseconds, panic the 7078 system. By default, let CSD-lock acquisition 7079 take as long as they take. Specifying 300,000 7080 for this value provides a 5-minute timeout. 7081 7082 smt= [KNL,MIPS,S390,EARLY] Set the maximum number of threads 7083 (logical CPUs) to use per physical CPU on systems 7084 capable of symmetric multithreading (SMT). Will 7085 be capped to the actual hardware limit. 7086 Format: <integer> 7087 Default: -1 (no limit) 7088 7089 softlockup_panic= 7090 [KNL] Should the soft-lockup detector generate panics. 7091 Format: <int> 7092 7093 A value of non-zero instructs the soft-lockup detector 7094 to panic the machine when a soft-lockup duration exceeds 7095 N thresholds. It is also controlled by the kernel.softlockup_panic 7096 sysctl and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the 7097 respective build-time switch to that functionality. 7098 7099 softlockup_all_cpu_backtrace= 7100 [KNL] Should the soft-lockup detector generate 7101 backtraces on all cpus. 7102 Format: 0 | 1 7103 7104 sonypi.*= [HW] Sony Programmable I/O Control Device driver 7105 See Documentation/admin-guide/laptops/sonypi.rst 7106 7107 spectre_bhi= [X86] Control mitigation of Branch History Injection 7108 (BHI) vulnerability. This setting affects the 7109 deployment of the HW BHI control and the SW BHB 7110 clearing sequence. 7111 7112 on - (default) Enable the HW or SW mitigation as 7113 needed. This protects the kernel from 7114 both syscalls and VMs. 7115 vmexit - On systems which don't have the HW mitigation 7116 available, enable the SW mitigation on vmexit 7117 ONLY. On such systems, the host kernel is 7118 protected from VM-originated BHI attacks, but 7119 may still be vulnerable to syscall attacks. 7120 off - Disable the mitigation. 7121 7122 spectre_v2= [X86,EARLY] Control mitigation of Spectre variant 2 7123 (indirect branch speculation) vulnerability. 7124 The default operation protects the kernel from 7125 user space attacks. 7126 7127 on - unconditionally enable, implies 7128 spectre_v2_user=on 7129 off - unconditionally disable, implies 7130 spectre_v2_user=off 7131 auto - kernel detects whether your CPU model is 7132 vulnerable 7133 7134 Selecting 'on' will, and 'auto' may, choose a 7135 mitigation method at run time according to the 7136 CPU, the available microcode, the setting of the 7137 CONFIG_MITIGATION_RETPOLINE configuration option, 7138 and the compiler with which the kernel was built. 7139 7140 Selecting 'on' will also enable the mitigation 7141 against user space to user space task attacks. 7142 Selecting specific mitigation does not force enable 7143 user mitigations. 7144 7145 Selecting 'off' will disable both the kernel and 7146 the user space protections. 7147 7148 Specific mitigations can also be selected manually: 7149 7150 retpoline - replace indirect branches 7151 retpoline,generic - Retpolines 7152 retpoline,lfence - LFENCE; indirect branch 7153 retpoline,amd - alias for retpoline,lfence 7154 eibrs - Enhanced/Auto IBRS 7155 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines 7156 eibrs,lfence - Enhanced/Auto IBRS + LFENCE 7157 ibrs - use IBRS to protect kernel 7158 7159 Not specifying this option is equivalent to 7160 spectre_v2=auto. 7161 7162 spectre_v2_user= 7163 [X86] Control mitigation of Spectre variant 2 7164 (indirect branch speculation) vulnerability between 7165 user space tasks 7166 7167 on - Unconditionally enable mitigations. Is 7168 enforced by spectre_v2=on 7169 7170 off - Unconditionally disable mitigations. Is 7171 enforced by spectre_v2=off 7172 7173 prctl - Indirect branch speculation is enabled, 7174 but mitigation can be enabled via prctl 7175 per thread. The mitigation control state 7176 is inherited on fork. 7177 7178 prctl,ibpb 7179 - Like "prctl" above, but only STIBP is 7180 controlled per thread. IBPB is issued 7181 always when switching between different user 7182 space processes. 7183 7184 seccomp 7185 - Same as "prctl" above, but all seccomp 7186 threads will enable the mitigation unless 7187 they explicitly opt out. 7188 7189 seccomp,ibpb 7190 - Like "seccomp" above, but only STIBP is 7191 controlled per thread. IBPB is issued 7192 always when switching between different 7193 user space processes. 7194 7195 auto - Kernel selects the mitigation depending on 7196 the available CPU features and vulnerability. 7197 7198 Default mitigation: "prctl" 7199 7200 Not specifying this option is equivalent to 7201 spectre_v2_user=auto. 7202 7203 spec_rstack_overflow= 7204 [X86,EARLY] Control RAS overflow mitigation on AMD Zen CPUs 7205 7206 off - Disable mitigation 7207 microcode - Enable microcode mitigation only 7208 safe-ret - Enable sw-only safe RET mitigation (default) 7209 ibpb - Enable mitigation by issuing IBPB on 7210 kernel entry 7211 ibpb-vmexit - Issue IBPB only on VMEXIT 7212 (cloud-specific mitigation) 7213 7214 spec_store_bypass_disable= 7215 [HW,EARLY] Control Speculative Store Bypass (SSB) Disable mitigation 7216 (Speculative Store Bypass vulnerability) 7217 7218 Certain CPUs are vulnerable to an exploit against a 7219 a common industry wide performance optimization known 7220 as "Speculative Store Bypass" in which recent stores 7221 to the same memory location may not be observed by 7222 later loads during speculative execution. The idea 7223 is that such stores are unlikely and that they can 7224 be detected prior to instruction retirement at the 7225 end of a particular speculation execution window. 7226 7227 In vulnerable processors, the speculatively forwarded 7228 store can be used in a cache side channel attack, for 7229 example to read memory to which the attacker does not 7230 directly have access (e.g. inside sandboxed code). 7231 7232 This parameter controls whether the Speculative Store 7233 Bypass optimization is used. 7234 7235 On x86 the options are: 7236 7237 on - Unconditionally disable Speculative Store Bypass 7238 off - Unconditionally enable Speculative Store Bypass 7239 auto - Kernel detects whether the CPU model contains an 7240 implementation of Speculative Store Bypass and 7241 picks the most appropriate mitigation. If the 7242 CPU is not vulnerable, "off" is selected. If the 7243 CPU is vulnerable the default mitigation is 7244 architecture and Kconfig dependent. See below. 7245 prctl - Control Speculative Store Bypass per thread 7246 via prctl. Speculative Store Bypass is enabled 7247 for a process by default. The state of the control 7248 is inherited on fork. 7249 seccomp - Same as "prctl" above, but all seccomp threads 7250 will disable SSB unless they explicitly opt out. 7251 7252 Default mitigations: 7253 X86: "prctl" 7254 7255 On powerpc the options are: 7256 7257 on,auto - On Power8 and Power9 insert a store-forwarding 7258 barrier on kernel entry and exit. On Power7 7259 perform a software flush on kernel entry and 7260 exit. 7261 off - No action. 7262 7263 Not specifying this option is equivalent to 7264 spec_store_bypass_disable=auto. 7265 7266 split_llc= 7267 [X86,EARLY] Split the LLC N-ways 7268 7269 When set, the LLC is split this many ways by matching 7270 'core_id % n'. This is setup before SMP bringup and 7271 used during SMP bringup before it knows the full 7272 topology. If your core count doesn't nicely divide by 7273 the number given, you get to keep the pieces. 7274 7275 This is mostly a debug feature to emulate multiple LLCs 7276 on hardware that only have a single LLC. 7277 7278 split_lock_detect= 7279 [X86] Enable split lock detection or bus lock detection 7280 7281 When enabled (and if hardware support is present), atomic 7282 instructions that access data across cache line 7283 boundaries will result in an alignment check exception 7284 for split lock detection or a debug exception for 7285 bus lock detection. 7286 7287 off - not enabled 7288 7289 warn - the kernel will emit rate-limited warnings 7290 about applications triggering the #AC 7291 exception or the #DB exception. This mode is 7292 the default on CPUs that support split lock 7293 detection or bus lock detection. Default 7294 behavior is by #AC if both features are 7295 enabled in hardware. 7296 7297 fatal - the kernel will send SIGBUS to applications 7298 that trigger the #AC exception or the #DB 7299 exception. Default behavior is by #AC if 7300 both features are enabled in hardware. 7301 7302 ratelimit:N - 7303 Set system wide rate limit to N bus locks 7304 per second for bus lock detection. 7305 0 < N <= 1000. 7306 7307 N/A for split lock detection. 7308 7309 7310 If an #AC exception is hit in the kernel or in 7311 firmware (i.e. not while executing in user mode) 7312 the kernel will oops in either "warn" or "fatal" 7313 mode. 7314 7315 #DB exception for bus lock is triggered only when 7316 CPL > 0. 7317 7318 srbds= [X86,INTEL,EARLY] 7319 Control the Special Register Buffer Data Sampling 7320 (SRBDS) mitigation. 7321 7322 Certain CPUs are vulnerable to an MDS-like 7323 exploit which can leak bits from the random 7324 number generator. 7325 7326 By default, this issue is mitigated by 7327 microcode. However, the microcode fix can cause 7328 the RDRAND and RDSEED instructions to become 7329 much slower. Among other effects, this will 7330 result in reduced throughput from /dev/urandom. 7331 7332 The microcode mitigation can be disabled with 7333 the following option: 7334 7335 off: Disable mitigation and remove 7336 performance impact to RDRAND and RDSEED 7337 7338 srcutree.big_cpu_lim [KNL] 7339 Specifies the number of CPUs constituting a 7340 large system, such that srcu_struct structures 7341 should immediately allocate an srcu_node array. 7342 This kernel-boot parameter defaults to 128, 7343 but takes effect only when the low-order four 7344 bits of srcutree.convert_to_big is equal to 3 7345 (decide at boot). 7346 7347 srcutree.convert_to_big [KNL] 7348 Specifies under what conditions an SRCU tree 7349 srcu_struct structure will be converted to big 7350 form, that is, with an rcu_node tree: 7351 7352 0: Never. 7353 1: At init_srcu_struct() time. 7354 2: When rcutorture decides to. 7355 3: Decide at boot time (default). 7356 0x1X: Above plus if high contention. 7357 7358 Either way, the srcu_node tree will be sized based 7359 on the actual runtime number of CPUs (nr_cpu_ids) 7360 instead of the compile-time CONFIG_NR_CPUS. 7361 7362 srcutree.counter_wrap_check [KNL] 7363 Specifies how frequently to check for 7364 grace-period sequence counter wrap for the 7365 srcu_data structure's ->srcu_gp_seq_needed field. 7366 The greater the number of bits set in this kernel 7367 parameter, the less frequently counter wrap will 7368 be checked for. Note that the bottom two bits 7369 are ignored. 7370 7371 srcutree.exp_holdoff [KNL] 7372 Specifies how many nanoseconds must elapse 7373 since the end of the last SRCU grace period for 7374 a given srcu_struct until the next normal SRCU 7375 grace period will be considered for automatic 7376 expediting. Set to zero to disable automatic 7377 expediting. 7378 7379 srcutree.srcu_max_nodelay [KNL] 7380 Specifies the number of no-delay instances 7381 per jiffy for which the SRCU grace period 7382 worker thread will be rescheduled with zero 7383 delay. Beyond this limit, worker thread will 7384 be rescheduled with a sleep delay of one jiffy. 7385 7386 srcutree.srcu_max_nodelay_phase [KNL] 7387 Specifies the per-grace-period phase, number of 7388 non-sleeping polls of readers. Beyond this limit, 7389 grace period worker thread will be rescheduled 7390 with a sleep delay of one jiffy, between each 7391 rescan of the readers, for a grace period phase. 7392 7393 srcutree.srcu_retry_check_delay [KNL] 7394 Specifies number of microseconds of non-sleeping 7395 delay between each non-sleeping poll of readers. 7396 7397 srcutree.small_contention_lim [KNL] 7398 Specifies the number of update-side contention 7399 events per jiffy will be tolerated before 7400 initiating a conversion of an srcu_struct 7401 structure to big form. Note that the value of 7402 srcutree.convert_to_big must have the 0x10 bit 7403 set for contention-based conversions to occur. 7404 7405 ssbd= [ARM64,HW,EARLY] 7406 Speculative Store Bypass Disable control 7407 7408 On CPUs that are vulnerable to the Speculative 7409 Store Bypass vulnerability and offer a 7410 firmware based mitigation, this parameter 7411 indicates how the mitigation should be used: 7412 7413 force-on: Unconditionally enable mitigation for 7414 for both kernel and userspace 7415 force-off: Unconditionally disable mitigation for 7416 for both kernel and userspace 7417 kernel: Always enable mitigation in the 7418 kernel, and offer a prctl interface 7419 to allow userspace to register its 7420 interest in being mitigated too. 7421 7422 stack_guard_gap= [MM] 7423 override the default stack gap protection. The value 7424 is in page units and it defines how many pages prior 7425 to (for stacks growing down) resp. after (for stacks 7426 growing up) the main stack are reserved for no other 7427 mapping. Default value is 256 pages. 7428 7429 stack_depot_disable= [KNL,EARLY] 7430 Setting this to true through kernel command line will 7431 disable the stack depot thereby saving the static memory 7432 consumed by the stack hash table. By default this is set 7433 to false. 7434 7435 stack_depot_max_pools= [KNL,EARLY] 7436 Specify the maximum number of pools to use for storing 7437 stack traces. Pools are allocated on-demand up to this 7438 limit. Default value is 8191 pools. 7439 7440 stacktrace [FTRACE] 7441 Enable the stack tracer on boot up. 7442 7443 stacktrace_filter=[function-list] 7444 [FTRACE] Limit the functions that the stack tracer 7445 will trace at boot up. function-list is a comma-separated 7446 list of functions. This list can be changed at run 7447 time by the stack_trace_filter file in the debugfs 7448 tracing directory. Note, this enables stack tracing 7449 and the stacktrace above is not needed. 7450 7451 sti= [PARISC,HW] 7452 Format: <num> 7453 Set the STI (builtin display/keyboard on the HP-PARISC 7454 machines) console (graphic card) which should be used 7455 as the initial boot-console. 7456 See also comment in drivers/video/sticore.c. 7457 7458 sti_font= [HW] 7459 See comment in drivers/video/sticore.c. 7460 7461 stifb= [HW] 7462 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]] 7463 7464 strict_sas_size= 7465 [X86] 7466 Format: <bool> 7467 Enable or disable strict sigaltstack size checks 7468 against the required signal frame size which 7469 depends on the supported FPU features. This can 7470 be used to filter out binaries which have 7471 not yet been made aware of AT_MINSIGSTKSZ. 7472 7473 stress_hpt [PPC,EARLY] 7474 Limits the number of kernel HPT entries in the hash 7475 page table to increase the rate of hash page table 7476 faults on kernel addresses. 7477 7478 stress_slb [PPC,EARLY] 7479 Limits the number of kernel SLB entries, and flushes 7480 them frequently to increase the rate of SLB faults 7481 on kernel addresses. 7482 7483 no_slb_preload [PPC,EARLY] 7484 Disables slb preloading for userspace. 7485 7486 sunrpc.min_resvport= 7487 sunrpc.max_resvport= 7488 [NFS,SUNRPC] 7489 SunRPC servers often require that client requests 7490 originate from a privileged port (i.e. a port in the 7491 range 0 < portnr < 1024). 7492 An administrator who wishes to reserve some of these 7493 ports for other uses may adjust the range that the 7494 kernel's sunrpc client considers to be privileged 7495 using these two parameters to set the minimum and 7496 maximum port values. 7497 7498 sunrpc.svc_rpc_per_connection_limit= 7499 [NFS,SUNRPC] 7500 Limit the number of requests that the server will 7501 process in parallel from a single connection. 7502 The default value is 0 (no limit). 7503 7504 sunrpc.pool_mode= 7505 [NFS] 7506 Deprecated. The NFS server now always uses one 7507 service thread pool per NUMA node (equivalent to a 7508 single global pool on non-NUMA machines). All of 7509 the previously accepted values (auto, global, 7510 percpu, pernode) are still accepted for backward 7511 compatibility but are ignored: the mode is always 7512 pernode, and reads always return "pernode". 7513 7514 sunrpc.tcp_slot_table_entries= 7515 sunrpc.udp_slot_table_entries= 7516 [NFS,SUNRPC] 7517 Sets the upper limit on the number of simultaneous 7518 RPC calls that can be sent from the client to a 7519 server. Increasing these values may allow you to 7520 improve throughput, but will also increase the 7521 amount of memory reserved for use by the client. 7522 7523 suspend.pm_test_delay= 7524 [SUSPEND] 7525 Sets the number of seconds to remain in a suspend test 7526 mode before resuming the system (see 7527 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 7528 is set. Default value is 5. 7529 7530 svm= [PPC] 7531 Format: { on | off | y | n | 1 | 0 } 7532 This parameter controls use of the Protected 7533 Execution Facility on pSeries. 7534 7535 swiotlb= [ARM,PPC,MIPS,X86,S390,EARLY] 7536 Format: { <int> [,<int>] | force | noforce | track_hiwater} 7537 <int> -- Number of I/O TLB slabs 7538 <int> -- Second integer after comma. Number of swiotlb 7539 areas with their own lock. Will be rounded up 7540 to a power of 2. 7541 force -- force using of bounce buffers even if they 7542 wouldn't be automatically used by the kernel 7543 noforce -- Never use bounce buffers (for debugging) 7544 track_hiwater -- Track high watermark of swiotlb buffers. 7545 Only available when CONFIG_DEBUG_FS is set. 7546 7547 switches= [HW,M68k,EARLY] 7548 7549 sysctl.*= [KNL] 7550 Set a sysctl parameter, right before loading the init 7551 process, as if the value was written to the respective 7552 /proc/sys/... file. Both '.' and '/' are recognized as 7553 separators. Unrecognized parameters and invalid values 7554 are reported in the kernel log. Sysctls registered 7555 later by a loaded module cannot be set this way. 7556 Example: sysctl.vm.swappiness=40 7557 7558 sysrq_always_enabled 7559 [KNL] 7560 Ignore sysrq setting - this boot parameter will 7561 neutralize any effect of /proc/sys/kernel/sysrq. 7562 Useful for debugging. 7563 7564 tcpmhash_entries= [KNL,NET] 7565 Set the number of tcp_metrics_hash slots. 7566 Default value is 8192 or 16384 depending on total 7567 ram pages. This is used to specify the TCP metrics 7568 cache size. See Documentation/networking/ip-sysctl.rst 7569 "tcp_no_metrics_save" section for more details. 7570 7571 tdfx= [HW,DRM] 7572 7573 test_suspend= [SUSPEND] 7574 Format: { "mem" | "standby" | "freeze" }[,N] 7575 Specify "mem" (for Suspend-to-RAM) or "standby" (for 7576 standby suspend) or "freeze" (for suspend type freeze) 7577 as the system sleep state during system startup with 7578 the optional capability to repeat N number of times. 7579 The system is woken from this state using a 7580 wakeup-capable RTC alarm. 7581 7582 thash_entries= [KNL,NET] 7583 Set number of hash buckets for TCP connection 7584 7585 thermal.act= [HW,ACPI] 7586 -1: disable all active trip points in all thermal zones 7587 <degrees C>: override all lowest active trip points 7588 7589 thermal.crt= [HW,ACPI] 7590 -1: disable all critical trip points in all thermal zones 7591 <degrees C>: override all critical trip points 7592 7593 thermal.off= [HW,ACPI] 7594 1: disable ACPI thermal control 7595 7596 thermal.psv= [HW,ACPI] 7597 -1: disable all passive trip points 7598 <degrees C>: override all passive trip points to this 7599 value 7600 7601 thermal.tzp= [HW,ACPI] 7602 Specify global default ACPI thermal zone polling rate 7603 <deci-seconds>: poll all this frequency 7604 0: no polling (default) 7605 7606 thp_anon= [KNL] 7607 Format: <size>[KMG],<size>[KMG]:<state>;<size>[KMG]-<size>[KMG]:<state> 7608 state is one of "always", "madvise", "never" or "inherit". 7609 Control the default behavior of the system with respect 7610 to anonymous transparent hugepages. 7611 Can be used multiple times for multiple anon THP sizes. 7612 See Documentation/admin-guide/mm/transhuge.rst for more 7613 details. 7614 7615 threadirqs [KNL,EARLY] 7616 Force threading of all interrupt handlers except those 7617 marked explicitly IRQF_NO_THREAD. 7618 7619 thp_shmem= [KNL] 7620 Format: <size>[KMG],<size>[KMG]:<policy>;<size>[KMG]-<size>[KMG]:<policy> 7621 Control the default policy of each hugepage size for the 7622 internal shmem mount. <policy> is one of policies available 7623 for the shmem mount ("always", "inherit", "never", "within_size", 7624 and "advise"). 7625 It can be used multiple times for multiple shmem THP sizes. 7626 See Documentation/admin-guide/mm/transhuge.rst for more 7627 details. 7628 7629 tlbi= [X86-64] 7630 Format: {ipi} 7631 ipi: switch to IPI-based TLB flushing 7632 7633 topology= [S390,EARLY] 7634 Format: {off | on} 7635 Specify if the kernel should make use of the cpu 7636 topology information if the hardware supports this. 7637 The scheduler will make use of this information and 7638 e.g. base its process migration decisions on it. 7639 Default is on. 7640 7641 torture.disable_onoff_at_boot= [KNL] 7642 Prevent the CPU-hotplug component of torturing 7643 until after init has spawned. 7644 7645 torture.ftrace_dump_at_shutdown= [KNL] 7646 Dump the ftrace buffer at torture-test shutdown, 7647 even if there were no errors. This can be a 7648 very costly operation when many torture tests 7649 are running concurrently, especially on systems 7650 with rotating-rust storage. 7651 7652 torture.verbose_sleep_frequency= [KNL] 7653 Specifies how many verbose printk()s should be 7654 emitted between each sleep. The default of zero 7655 disables verbose-printk() sleeping. 7656 7657 torture.verbose_sleep_duration= [KNL] 7658 Duration of each verbose-printk() sleep in jiffies. 7659 7660 tpm.disable_pcr_integrity= [HW,TPM] 7661 Do not protect PCR registers from unintended physical 7662 access, or interposers in the bus by the means of 7663 having an integrity protected session wrapped around 7664 TPM2_PCR_Extend command. Consider this in a situation 7665 where TPM is heavily utilized by IMA, thus protection 7666 causing a major performance hit, and the space where 7667 machines are deployed is by other means guarded. 7668 7669 tpm_crb_ffa.busy_timeout_ms= [ARM64,TPM] 7670 Maximum time in milliseconds to retry sending a message 7671 to the TPM service before giving up. This parameter controls 7672 how long the system will continue retrying when the TPM 7673 service is busy. 7674 Format: <unsigned int> 7675 Default: 2000 (2 seconds) 7676 7677 tpm_suspend_pcr=[HW,TPM] 7678 Format: integer pcr id 7679 Specify that at suspend time, the tpm driver 7680 should extend the specified pcr with zeros, 7681 as a workaround for some chips which fail to 7682 flush the last written pcr on TPM_SaveState. 7683 This will guarantee that all the other pcrs 7684 are saved. 7685 7686 tpm_tis.interrupts= [HW,TPM] 7687 Enable interrupts for the MMIO based physical layer 7688 for the FIFO interface. By default it is set to false 7689 (0). For more information about TPM hardware interfaces 7690 defined by Trusted Computing Group (TCG) see 7691 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/ 7692 7693 tp_printk [FTRACE] 7694 Have the tracepoints sent to printk as well as the 7695 tracing ring buffer. This is useful for early boot up 7696 where the system hangs or reboots and does not give the 7697 option for reading the tracing buffer or performing a 7698 ftrace_dump_on_oops. 7699 7700 To turn off having tracepoints sent to printk, 7701 echo 0 > /proc/sys/kernel/tracepoint_printk 7702 Note, echoing 1 into this file without the 7703 tp_printk kernel cmdline option has no effect. 7704 7705 The tp_printk_stop_on_boot (see below) can also be used 7706 to stop the printing of events to console at 7707 late_initcall_sync. 7708 7709 ** CAUTION ** 7710 7711 Having tracepoints sent to printk() and activating high 7712 frequency tracepoints such as irq or sched, can cause 7713 the system to live lock. 7714 7715 tp_printk_stop_on_boot [FTRACE] 7716 When tp_printk (above) is set, it can cause a lot of noise 7717 on the console. It may be useful to only include the 7718 printing of events during boot up, as user space may 7719 make the system inoperable. 7720 7721 This command line option will stop the printing of events 7722 to console at the late_initcall_sync() time frame. 7723 7724 trace_buf_size=nn[KMG] 7725 [FTRACE] will set tracing buffer size on each cpu. 7726 7727 trace_clock= [FTRACE] Set the clock used for tracing events 7728 at boot up. 7729 local - Use the per CPU time stamp counter 7730 (converted into nanoseconds). Fast, but 7731 depending on the architecture, may not be 7732 in sync between CPUs. 7733 global - Event time stamps are synchronized across 7734 CPUs. May be slower than the local clock, 7735 but better for some race conditions. 7736 counter - Simple counting of events (1, 2, ..) 7737 note, some counts may be skipped due to the 7738 infrastructure grabbing the clock more than 7739 once per event. 7740 uptime - Use jiffies as the time stamp. 7741 perf - Use the same clock that perf uses. 7742 mono - Use ktime_get_mono_fast_ns() for time stamps. 7743 mono_raw - Use ktime_get_raw_fast_ns() for time 7744 stamps. 7745 boot - Use ktime_get_boot_fast_ns() for time stamps. 7746 Architectures may add more clocks. See 7747 Documentation/trace/ftrace.rst for more details. 7748 7749 trace_event=[event-list] 7750 [FTRACE] Set and start specified trace events in order 7751 to facilitate early boot debugging. The event-list is a 7752 comma-separated list of trace events to enable. See 7753 also Documentation/trace/events.rst 7754 7755 To enable modules, use :mod: keyword: 7756 7757 trace_event=:mod:<module> 7758 7759 The value before :mod: will only enable specific events 7760 that are part of the module. See the above mentioned 7761 document for more information. 7762 7763 trace_instance=[instance-info] 7764 [FTRACE] Create a ring buffer instance early in boot up. 7765 This will be listed in: 7766 7767 /sys/kernel/tracing/instances 7768 7769 Events can be enabled at the time the instance is created 7770 via: 7771 7772 trace_instance=<name>,<system1>:<event1>,<system2>:<event2> 7773 7774 Note, the "<system*>:" portion is optional if the event is 7775 unique. 7776 7777 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall 7778 7779 will enable the "sched_switch" event (note, the "sched:" is optional, and 7780 the same thing would happen if it was left off). The irq_handler_entry 7781 event, and all events under the "initcall" system. 7782 7783 Flags can be added to the instance to modify its behavior when it is 7784 created. The flags are separated by '^'. 7785 7786 The available flags are: 7787 7788 traceoff - Have the tracing instance tracing disabled after it is created. 7789 traceprintk - Have trace_printk() write into this trace instance 7790 (note, "printk" and "trace_printk" can also be used) 7791 7792 trace_instance=foo^traceoff^traceprintk,sched,irq 7793 7794 The flags must come before the defined events. 7795 7796 If memory has been reserved (see memmap for x86), the instance 7797 can use that memory: 7798 7799 memmap=12M$0x284500000 trace_instance=boot_map@0x284500000:12M 7800 7801 The above will create a "boot_map" instance that uses the physical 7802 memory at 0x284500000 that is 12Megs. The per CPU buffers of that 7803 instance will be split up accordingly. 7804 7805 Alternatively, the memory can be reserved by the reserve_mem option: 7806 7807 reserve_mem=12M:4096:trace trace_instance=boot_map@trace 7808 7809 This will reserve 12 megabytes at boot up with a 4096 byte alignment 7810 and place the ring buffer in this memory. Note that due to KASLR, the 7811 memory may not be the same location each time, which will not preserve 7812 the buffer content. 7813 7814 Also note that the layout of the ring buffer data may change between 7815 kernel versions where the validator will fail and reset the ring buffer 7816 if the layout is not the same as the previous kernel. 7817 7818 If the ring buffer is used for persistent bootups and has events enabled, 7819 it is recommend to disable tracing so that events from a previous boot do not 7820 mix with events of the current boot (unless you are debugging a random crash 7821 at boot up). 7822 7823 reserve_mem=12M:4096:trace trace_instance=boot_map^traceoff^traceprintk@trace,sched,irq 7824 7825 Note, saving the trace buffer across reboots does require that the system 7826 is set up to not wipe memory. For instance, CONFIG_RESET_ATTACK_MITIGATION 7827 can force a memory reset on boot which will clear any trace that was stored. 7828 This is just one of many ways that can clear memory. Make sure your system 7829 keeps the content of memory across reboots before relying on this option. 7830 7831 NB: Both the mapped address and size must be page aligned for the architecture. 7832 7833 See also Documentation/trace/debugging.rst 7834 7835 7836 trace_options=[option-list] 7837 [FTRACE] Enable or disable tracer options at boot. 7838 The option-list is a comma delimited list of options 7839 that can be enabled or disabled just as if you were 7840 to echo the option name into 7841 7842 /sys/kernel/tracing/trace_options 7843 7844 For example, to enable stacktrace option (to dump the 7845 stack trace of each event), add to the command line: 7846 7847 trace_options=stacktrace 7848 7849 See also Documentation/trace/ftrace.rst "trace options" 7850 section. 7851 7852 trace_trigger=[trigger-list] 7853 [FTRACE] Add an event trigger on specific events. 7854 Set a trigger on top of a specific event, with an optional 7855 filter. 7856 7857 The format is "trace_trigger=<event>.<trigger>[ if <filter>],..." 7858 Where more than one trigger may be specified that are comma delimited. 7859 7860 For example: 7861 7862 trace_trigger="sched_switch.stacktrace if prev_state == 2" 7863 7864 The above will enable the "stacktrace" trigger on the "sched_switch" 7865 event but only trigger it if the "prev_state" of the "sched_switch" 7866 event is "2" (TASK_UNINTERRUPTIBLE). 7867 7868 See also "Event triggers" in Documentation/trace/events.rst 7869 7870 7871 traceoff_after_boot 7872 [FTRACE] Sometimes tracing is used to debug issues 7873 during the boot process. Since the trace buffer has a 7874 limited amount of storage, it may be prudent to 7875 disable tracing after the boot is finished, otherwise 7876 the critical information may be overwritten. With this 7877 option, the main tracing buffer will be turned off at 7878 the end of the boot process. 7879 7880 traceoff_on_warning 7881 [FTRACE] enable this option to disable tracing when a 7882 warning is hit. This turns off "tracing_on". Tracing can 7883 be enabled again by echoing '1' into the "tracing_on" 7884 file located in /sys/kernel/tracing/ 7885 7886 This option is useful, as it disables the trace before 7887 the WARNING dump is called, which prevents the trace to 7888 be filled with content caused by the warning output. 7889 7890 This option can also be set at run time via the sysctl 7891 option: kernel/traceoff_on_warning 7892 7893 transparent_hugepage= 7894 [KNL] 7895 Format: [always|madvise|never] 7896 Can be used to control the default behavior of the system 7897 with respect to transparent hugepages. 7898 See Documentation/admin-guide/mm/transhuge.rst 7899 for more details. 7900 7901 transparent_hugepage_shmem= [KNL] 7902 Format: [always|within_size|advise|never|deny|force] 7903 Can be used to control the hugepage allocation policy for 7904 the internal shmem mount. 7905 See Documentation/admin-guide/mm/transhuge.rst 7906 for more details. 7907 7908 transparent_hugepage_tmpfs= [KNL] 7909 Format: [always|within_size|advise|never] 7910 Can be used to control the default hugepage allocation policy 7911 for the tmpfs mount. 7912 See Documentation/admin-guide/mm/transhuge.rst 7913 for more details. 7914 7915 trusted.source= [KEYS] 7916 Format: <string> 7917 This parameter identifies the trust source as a backend 7918 for trusted keys implementation. Supported trust 7919 sources: 7920 - "tpm" 7921 - "tee" 7922 - "caam" 7923 - "dcp" 7924 - "pkwm" 7925 If not specified then it defaults to iterating through 7926 the trust source list starting with TPM and assigns the 7927 first trust source as a backend which is initialized 7928 successfully during iteration. 7929 7930 trusted.debug= [KEYS] 7931 Format: <bool> 7932 Enable trusted keys debug traces at runtime when 7933 CONFIG_TRUSTED_KEYS_DEBUG=y. 7934 7935 To make the traces visible after enabling the option, 7936 use trusted.dyndbg='+p' as needed. By convention, 7937 the subsystem uses pr_debug() for these traces. 7938 7939 SAFETY: The traces can leak sensitive data, so be 7940 cautious before enabling this. They remain inactive 7941 unless this parameter is set this option to a true 7942 value. 7943 7944 Default: false 7945 7946 trusted.rng= [KEYS] 7947 Format: <string> 7948 The RNG used to generate key material for trusted keys. 7949 Can be one of: 7950 - "kernel" 7951 - the same value as trusted.source: "tpm" or "tee" 7952 - "default" 7953 If not specified, "default" is used. In this case, 7954 the RNG's choice is left to each individual trust source. 7955 7956 trusted.dcp_use_otp_key 7957 This is intended to be used in combination with 7958 trusted.source=dcp and will select the DCP OTP key 7959 instead of the DCP UNIQUE key blob encryption. 7960 7961 trusted.dcp_skip_zk_test 7962 This is intended to be used in combination with 7963 trusted.source=dcp and will disable the check if the 7964 blob key is all zeros. This is helpful for situations where 7965 having this key zero'ed is acceptable. E.g. in testing 7966 scenarios. 7967 7968 tsa= [X86] Control mitigation for Transient Scheduler 7969 Attacks on AMD CPUs. Search the following in your 7970 favourite search engine for more details: 7971 7972 "Technical guidance for mitigating transient scheduler 7973 attacks". 7974 7975 off - disable the mitigation 7976 on - enable the mitigation (default) 7977 user - mitigate only user/kernel transitions 7978 vm - mitigate only guest/host transitions 7979 7980 7981 tsc= Disable clocksource stability checks for TSC. 7982 Format: <string> 7983 [x86] reliable: mark tsc clocksource as reliable, this 7984 disables clocksource verification at runtime, as well 7985 as the stability checks done at bootup. Used to enable 7986 high-resolution timer mode on older hardware, and in 7987 virtualized environment. 7988 [x86] noirqtime: Do not use TSC to do irq accounting. 7989 Used to run time disable IRQ_TIME_ACCOUNTING on any 7990 platforms where RDTSC is slow and this accounting 7991 can add overhead. 7992 [x86] unstable: mark the TSC clocksource as unstable, this 7993 marks the TSC unconditionally unstable at bootup and 7994 avoids any further wobbles once the TSC watchdog notices. 7995 [x86] nowatchdog: disable clocksource watchdog. Used 7996 in situations with strict latency requirements (where 7997 interruptions from clocksource watchdog are not 7998 acceptable). 7999 [x86] recalibrate: force recalibration against a HW timer 8000 (HPET or PM timer) on systems whose TSC frequency was 8001 obtained from HW or FW using either an MSR or CPUID(0x15). 8002 Warn if the difference is more than 500 ppm. 8003 [x86] watchdog: Enforce the clocksource watchdog on TSC 8004 8005 tsc_early_khz= [X86,EARLY] Skip early TSC calibration and use the given 8006 value instead. Useful when the early TSC frequency discovery 8007 procedure is not reliable, such as on overclocked systems 8008 with CPUID.16h support and partial CPUID.15h support. 8009 Format: <unsigned int> 8010 8011 tsx= [X86] Control Transactional Synchronization 8012 Extensions (TSX) feature in Intel processors that 8013 support TSX control. 8014 8015 This parameter controls the TSX feature. The options are: 8016 8017 on - Enable TSX on the system. Although there are 8018 mitigations for all known security vulnerabilities, 8019 TSX has been known to be an accelerator for 8020 several previous speculation-related CVEs, and 8021 so there may be unknown security risks associated 8022 with leaving it enabled. 8023 8024 off - Disable TSX on the system. (Note that this 8025 option takes effect only on newer CPUs which are 8026 not vulnerable to MDS, i.e., have 8027 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get 8028 the new IA32_TSX_CTRL MSR through a microcode 8029 update. This new MSR allows for the reliable 8030 deactivation of the TSX functionality.) 8031 8032 auto - Disable TSX if X86_BUG_TAA is present, 8033 otherwise enable TSX on the system. 8034 8035 Not specifying this option is equivalent to tsx=off. 8036 8037 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 8038 for more details. 8039 8040 tsx_async_abort= [X86,INTEL,EARLY] Control mitigation for the TSX Async 8041 Abort (TAA) vulnerability. 8042 8043 Similar to Micro-architectural Data Sampling (MDS) 8044 certain CPUs that support Transactional 8045 Synchronization Extensions (TSX) are vulnerable to an 8046 exploit against CPU internal buffers which can forward 8047 information to a disclosure gadget under certain 8048 conditions. 8049 8050 In vulnerable processors, the speculatively forwarded 8051 data can be used in a cache side channel attack, to 8052 access data to which the attacker does not have direct 8053 access. 8054 8055 This parameter controls the TAA mitigation. The 8056 options are: 8057 8058 full - Enable TAA mitigation on vulnerable CPUs 8059 if TSX is enabled. 8060 8061 full,nosmt - Enable TAA mitigation and disable SMT on 8062 vulnerable CPUs. If TSX is disabled, SMT 8063 is not disabled because CPU is not 8064 vulnerable to cross-thread TAA attacks. 8065 off - Unconditionally disable TAA mitigation 8066 8067 On MDS-affected machines, tsx_async_abort=off can be 8068 prevented by an active MDS mitigation as both vulnerabilities 8069 are mitigated with the same mechanism so in order to disable 8070 this mitigation, you need to specify mds=off too. 8071 8072 Not specifying this option is equivalent to 8073 tsx_async_abort=full. On CPUs which are MDS affected 8074 and deploy MDS mitigation, TAA mitigation is not 8075 required and doesn't provide any additional 8076 mitigation. 8077 8078 For details see: 8079 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 8080 8081 turbografx.map[2|3]= [HW,JOY] 8082 TurboGraFX parallel port interface 8083 Format: 8084 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7> 8085 See also Documentation/input/devices/joystick-parport.rst 8086 8087 udbg-immortal [PPC] When debugging early kernel crashes that 8088 happen after console_init() and before a proper 8089 console driver takes over, this boot options might 8090 help "seeing" what's going on. 8091 8092 uhash_entries= [KNL,NET] 8093 Set number of hash buckets for UDP/UDP-Lite connections 8094 8095 uhci-hcd.ignore_oc= 8096 [USB] Ignore overcurrent events (default N). 8097 Some badly-designed motherboards generate lots of 8098 bogus events, for ports that aren't wired to 8099 anything. Set this parameter to avoid log spamming. 8100 Note that genuine overcurrent events won't be 8101 reported either. 8102 8103 unaligned_scalar_speed= 8104 [RISCV] 8105 Format: {slow | fast | unsupported} 8106 Allow skipping scalar unaligned access speed tests. This 8107 is useful for testing alternative code paths and to skip 8108 the tests in environments where they run too slowly. All 8109 CPUs must have the same scalar unaligned access speed. 8110 8111 unaligned_vector_speed= 8112 [RISCV] 8113 Format: {slow | fast | unsupported} 8114 Allow skipping vector unaligned access speed tests. This 8115 is useful for testing alternative code paths and to skip 8116 the tests in environments where they run too slowly. All 8117 CPUs must have the same vector unaligned access speed. 8118 8119 unknown_nmi_panic 8120 [X86] Cause panic on unknown NMI. 8121 8122 unwind_debug [X86-64,EARLY] 8123 Enable unwinder debug output. This can be 8124 useful for debugging certain unwinder error 8125 conditions, including corrupt stacks and 8126 bad/missing unwinder metadata. 8127 8128 usbcore.authorized_default= 8129 [USB] Default USB device authorization: 8130 (default -1 = authorized (same as 1), 8131 0 = not authorized, 1 = authorized, 2 = authorized 8132 if device connected to internal port) 8133 8134 usbcore.autosuspend= 8135 [USB] The autosuspend time delay (in seconds) used 8136 for newly-detected USB devices (default 2). This 8137 is the time required before an idle device will be 8138 autosuspended. Devices for which the delay is set 8139 to a negative value won't be autosuspended at all. 8140 8141 usbcore.usbfs_snoop= 8142 [USB] Set to log all usbfs traffic (default 0 = off). 8143 8144 usbcore.usbfs_snoop_max= 8145 [USB] Maximum number of bytes to snoop in each URB 8146 (default = 65536). 8147 8148 usbcore.blinkenlights= 8149 [USB] Set to cycle leds on hubs (default 0 = off). 8150 8151 usbcore.old_scheme_first= 8152 [USB] Start with the old device initialization 8153 scheme (default 0 = off). 8154 8155 usbcore.usbfs_memory_mb= 8156 [USB] Memory limit (in MB) for buffers allocated by 8157 usbfs (default = 16, 0 = max = 2047). 8158 8159 usbcore.use_both_schemes= 8160 [USB] Try the other device initialization scheme 8161 if the first one fails (default 1 = enabled). 8162 8163 usbcore.initial_descriptor_timeout= 8164 [USB] Specifies timeout for the initial 64-byte 8165 USB_REQ_GET_DESCRIPTOR request in milliseconds 8166 (default 5000 = 5.0 seconds). 8167 8168 usbcore.nousb [USB] Disable the USB subsystem 8169 8170 usbcore.quirks= 8171 [USB] A list of quirk entries to augment the built-in 8172 usb core quirk list. List entries are separated by 8173 commas. Each entry has the form 8174 VendorID:ProductID:Flags. The IDs are 4-digit hex 8175 numbers and Flags is a set of letters. Each letter 8176 will change the built-in quirk; setting it if it is 8177 clear and clearing it if it is set. The letters have 8178 the following meanings: 8179 a = USB_QUIRK_STRING_FETCH_255 (string 8180 descriptors must not be fetched using 8181 a 255-byte read); 8182 b = USB_QUIRK_RESET_RESUME (device can't resume 8183 correctly so reset it instead); 8184 c = USB_QUIRK_NO_SET_INTF (device can't handle 8185 Set-Interface requests); 8186 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't 8187 handle its Configuration or Interface 8188 strings); 8189 e = USB_QUIRK_RESET (device can't be reset 8190 (e.g morph devices), don't use reset); 8191 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has 8192 more interface descriptions than the 8193 bNumInterfaces count, and can't handle 8194 talking to these interfaces); 8195 g = USB_QUIRK_DELAY_INIT (device needs a pause 8196 during initialization, after we read 8197 the device descriptor); 8198 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For 8199 high speed and super speed interrupt 8200 endpoints, the USB 2.0 and USB 3.0 spec 8201 require the interval in microframes (1 8202 microframe = 125 microseconds) to be 8203 calculated as interval = 2 ^ 8204 (bInterval-1). 8205 Devices with this quirk report their 8206 bInterval as the result of this 8207 calculation instead of the exponent 8208 variable used in the calculation); 8209 i = USB_QUIRK_DEVICE_QUALIFIER (device can't 8210 handle device_qualifier descriptor 8211 requests); 8212 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device 8213 generates spurious wakeup, ignore 8214 remote wakeup capability); 8215 k = USB_QUIRK_NO_LPM (device can't handle Link 8216 Power Management); 8217 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL 8218 (Device reports its bInterval as linear 8219 frames instead of the USB 2.0 8220 calculation); 8221 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs 8222 to be disconnected before suspend to 8223 prevent spurious wakeup); 8224 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a 8225 pause after every control message); 8226 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra 8227 delay after resetting its port); 8228 p = USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT 8229 (Reduce timeout of the SET_ADDRESS 8230 request from 5000 ms to 500 ms); 8231 q = USB_QUIRK_FORCE_ONE_CONFIG (Device 8232 claims zero configurations, 8233 forcing to 1); 8234 r = USB_QUIRK_WINDOWS_CONFIG_REQ_SIZE (Device 8235 fails during initialization when asked for 8236 9-bytes configuration descriptor request. 8237 Ask for 255-bytes request instead to mirror 8238 Windows' behavior); 8239 Example: quirks=0781:5580:bk,0a5c:5834:gij 8240 8241 usbhid.mousepoll= 8242 [USBHID] The interval which mice are to be polled at. 8243 8244 usbhid.jspoll= 8245 [USBHID] The interval which joysticks are to be polled at. 8246 8247 usbhid.kbpoll= 8248 [USBHID] The interval which keyboards are to be polled at. 8249 8250 usb-storage.delay_use= 8251 [UMS] The delay in seconds before a new device is 8252 scanned for Logical Units (default 1). 8253 Optionally the delay in milliseconds if the value has 8254 suffix with "ms". 8255 Example: delay_use=2567ms 8256 8257 usb-storage.quirks= 8258 [UMS] A list of quirks entries to supplement or 8259 override the built-in unusual_devs list. List 8260 entries are separated by commas. Each entry has 8261 the form VID:PID:Flags where VID and PID are Vendor 8262 and Product ID values (4-digit hex numbers) and 8263 Flags is a set of characters, each corresponding 8264 to a common usb-storage quirk flag as follows: 8265 a = SANE_SENSE (collect more than 18 bytes 8266 of sense data, not on uas); 8267 b = BAD_SENSE (don't collect more than 18 8268 bytes of sense data, not on uas); 8269 c = FIX_CAPACITY (decrease the reported 8270 device capacity by one sector); 8271 d = NO_READ_DISC_INFO (don't use 8272 READ_DISC_INFO command, not on uas); 8273 e = NO_READ_CAPACITY_16 (don't use 8274 READ_CAPACITY_16 command); 8275 f = NO_REPORT_OPCODES (don't use report opcodes 8276 command, uas only); 8277 g = MAX_SECTORS_240 (don't transfer more than 8278 240 sectors at a time, uas only); 8279 h = CAPACITY_HEURISTICS (decrease the 8280 reported device capacity by one 8281 sector if the number is odd); 8282 i = IGNORE_DEVICE (don't bind to this 8283 device); 8284 j = NO_REPORT_LUNS (don't use report luns 8285 command, uas only); 8286 k = NO_SAME (do not use WRITE_SAME, uas only) 8287 l = NOT_LOCKABLE (don't try to lock and 8288 unlock ejectable media, not on uas); 8289 m = MAX_SECTORS_64 (don't transfer more 8290 than 64 sectors = 32 KB at a time, 8291 not on uas); 8292 n = INITIAL_READ10 (force a retry of the 8293 initial READ(10) command, not on uas); 8294 o = CAPACITY_OK (accept the capacity 8295 reported by the device, not on uas); 8296 p = WRITE_CACHE (the device cache is ON 8297 by default, not on uas); 8298 r = IGNORE_RESIDUE (the device reports 8299 bogus residue values, not on uas); 8300 s = SINGLE_LUN (the device has only one 8301 Logical Unit); 8302 t = NO_ATA_1X (don't allow ATA(12) and ATA(16) 8303 commands, uas only); 8304 u = IGNORE_UAS (don't bind to the uas driver); 8305 w = NO_WP_DETECT (don't test whether the 8306 medium is write-protected). 8307 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE 8308 even if the device claims no cache, 8309 not on uas) 8310 Example: quirks=0419:aaf5:rl,0421:0433:rc 8311 8312 user_debug= [KNL,ARM] 8313 Format: <int> 8314 See arch/arm/Kconfig.debug help text. 8315 1 - undefined instruction events 8316 2 - system calls 8317 4 - invalid data aborts 8318 8 - SIGSEGV faults 8319 16 - SIGBUS faults 8320 Example: user_debug=31 8321 8322 vdso= [X86,SH,SPARC] 8323 On X86_32, this is an alias for vdso32=. Otherwise: 8324 8325 vdso=1: enable VDSO (the default) 8326 vdso=0: disable VDSO mapping 8327 8328 vdso32= [X86] Control the 32-bit vDSO 8329 vdso32=1: enable 32-bit VDSO 8330 vdso32=0 or vdso32=2: disable 32-bit VDSO 8331 8332 See the help text for CONFIG_COMPAT_VDSO for more 8333 details. If CONFIG_COMPAT_VDSO is set, the default is 8334 vdso32=0; otherwise, the default is vdso32=1. 8335 8336 For compatibility with older kernels, vdso32=2 is an 8337 alias for vdso32=0. 8338 8339 Try vdso32=0 if you encounter an error that says: 8340 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed! 8341 8342 video= [FB,EARLY] Frame buffer configuration 8343 See Documentation/fb/modedb.rst. 8344 8345 video.brightness_switch_enabled= [ACPI] 8346 Format: [0|1] 8347 If set to 1, on receiving an ACPI notify event 8348 generated by hotkey, video driver will adjust brightness 8349 level and then send out the event to user space through 8350 the allocated input device. If set to 0, video driver 8351 will only send out the event without touching backlight 8352 brightness level. 8353 default: 1 8354 8355 virtio_mmio.device= 8356 [VMMIO] Memory mapped virtio (platform) device. 8357 8358 <size>@<baseaddr>:<irq>[:<id>] 8359 where: 8360 <size> := size (can use standard suffixes 8361 like K, M and G) 8362 <baseaddr> := physical base address 8363 <irq> := interrupt number (as passed to 8364 request_irq()) 8365 <id> := (optional) platform device id 8366 example: 8367 virtio_mmio.device=1K@0x100b0000:48:7 8368 8369 Can be used multiple times for multiple devices. 8370 8371 vga= [BOOT,X86-32] Select a particular video mode 8372 See Documentation/arch/x86/boot.rst and 8373 Documentation/admin-guide/svga.rst. 8374 Use vga=ask for menu. 8375 This is actually a boot loader parameter; the value is 8376 passed to the kernel using a special protocol. 8377 8378 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y. 8379 May slow down system boot speed, especially when 8380 enabled on systems with a large amount of memory. 8381 All options are enabled by default, and this 8382 interface is meant to allow for selectively 8383 enabling or disabling specific virtual memory 8384 debugging features. 8385 8386 Available options are: 8387 P Enable page structure init time poisoning 8388 - Disable all of the above options 8389 8390 vmalloc=nn[KMG] [KNL,BOOT,EARLY] Forces the vmalloc area to have an 8391 exact size of <nn>. This can be used to increase 8392 the minimum size (128MB on x86, arm32 platforms). 8393 It can also be used to decrease the size and leave more room 8394 for directly mapped kernel RAM. Note that this parameter does 8395 not exist on many other platforms (including arm64, alpha, 8396 loongarch, arc, csky, hexagon, microblaze, mips, nios2, openrisc, 8397 parisc, m64k, powerpc, riscv, sh, um, xtensa, s390, sparc). 8398 8399 vmcp_cma=nn[MG] [KNL,S390,EARLY] 8400 Sets the memory size reserved for contiguous memory 8401 allocations for the vmcp device driver. 8402 8403 vmhalt= [KNL,S390] Perform z/VM CP command after system halt. 8404 Format: <command> 8405 8406 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic. 8407 Format: <command> 8408 8409 vmpoff= [KNL,S390] Perform z/VM CP command after power off. 8410 Format: <command> 8411 8412 vmscape= [X86] Controls mitigation for VMscape attacks. 8413 VMscape attacks can leak information from a userspace 8414 hypervisor to a guest via speculative side-channels. 8415 8416 off - disable the mitigation 8417 ibpb - use Indirect Branch Prediction Barrier 8418 (IBPB) mitigation (default) 8419 force - force vulnerability detection even on 8420 unaffected processors 8421 8422 vsyscall= [X86-64,EARLY] 8423 Controls the behavior of vsyscalls (i.e. calls to 8424 fixed addresses of 0xffffffffff600x00 from legacy 8425 code). Most statically-linked binaries and older 8426 versions of glibc use these calls. Because these 8427 functions are at fixed addresses, they make nice 8428 targets for exploits that can control RIP. 8429 8430 emulate Vsyscalls turn into traps and are emulated 8431 reasonably safely. The vsyscall page is 8432 readable. This disables the Linear 8433 Address Space Separation (LASS) security 8434 feature and makes the system less secure. 8435 8436 xonly [default] Vsyscalls turn into traps and are 8437 emulated reasonably safely. The vsyscall 8438 page is not readable. 8439 8440 none Vsyscalls don't work at all. This makes 8441 them quite hard to use for exploits but 8442 might break your system. 8443 8444 vt.color= [VT] Default text color. 8445 Format: 0xYX, X = foreground, Y = background. 8446 Default: 0x07 = light gray on black. 8447 8448 vt.cur_default= [VT] Default cursor shape. 8449 Format: 0xCCBBAA, where AA, BB, and CC are the same as 8450 the parameters of the <Esc>[?A;B;Cc escape sequence; 8451 see vga-softcursor.rst. Default: 2 = underline. 8452 8453 vt.default_blu= [VT] 8454 Format: <blue0>,<blue1>,<blue2>,...,<blue15> 8455 Change the default blue palette of the console. 8456 This is a 16-member array composed of values 8457 ranging from 0-255. 8458 8459 vt.default_grn= [VT] 8460 Format: <green0>,<green1>,<green2>,...,<green15> 8461 Change the default green palette of the console. 8462 This is a 16-member array composed of values 8463 ranging from 0-255. 8464 8465 vt.default_red= [VT] 8466 Format: <red0>,<red1>,<red2>,...,<red15> 8467 Change the default red palette of the console. 8468 This is a 16-member array composed of values 8469 ranging from 0-255. 8470 8471 vt.default_utf8= 8472 [VT] 8473 Format=<0|1> 8474 Set system-wide default UTF-8 mode for all tty's. 8475 Default is 1, i.e. UTF-8 mode is enabled for all 8476 newly opened terminals. 8477 8478 vt.global_cursor_default= 8479 [VT] 8480 Format=<-1|0|1> 8481 Set system-wide default for whether a cursor 8482 is shown on new VTs. Default is -1, 8483 i.e. cursors will be created by default unless 8484 overridden by individual drivers. 0 will hide 8485 cursors, 1 will display them. 8486 8487 vt.italic= [VT] Default color for italic text; 0-15. 8488 Default: 2 = green. 8489 8490 vt.underline= [VT] Default color for underlined text; 0-15. 8491 Default: 3 = cyan. 8492 8493 watchdog timers [HW,WDT] For information on watchdog timers, 8494 see Documentation/watchdog/watchdog-parameters.rst 8495 or other driver-specific files in the 8496 Documentation/watchdog/ directory. 8497 8498 watchdog_thresh= 8499 [KNL] 8500 Set the hard lockup detector stall duration 8501 threshold in seconds. The soft lockup detector 8502 threshold is set to twice the value. A value of 0 8503 disables both lockup detectors. Default is 10 8504 seconds. 8505 8506 workqueue.unbound_cpus= 8507 [KNL,SMP] Specify to constrain one or some CPUs 8508 to use in unbound workqueues. 8509 Format: <cpu-list> 8510 By default, all online CPUs are available for 8511 unbound workqueues. 8512 8513 workqueue.watchdog_thresh= 8514 If CONFIG_WQ_WATCHDOG is configured, workqueue can 8515 warn stall conditions and dump internal state to 8516 help debugging. 0 disables workqueue stall 8517 detection; otherwise, it's the stall threshold 8518 duration in seconds. The default value is 30 and 8519 it can be updated at runtime by writing to the 8520 corresponding sysfs file. 8521 8522 workqueue.panic_on_stall=<uint> 8523 Panic when workqueue stall is detected by 8524 CONFIG_WQ_WATCHDOG. It sets the number times of the 8525 stall to trigger panic. 8526 8527 The default is set by CONFIG_BOOTPARAM_WQ_STALL_PANIC, 8528 which is 0 (disabled) if not configured. 8529 8530 workqueue.panic_on_stall_time=<uint> 8531 Panic when a workqueue stall has been continuous for 8532 the specified number of seconds. Unlike panic_on_stall 8533 which counts accumulated stall events, this triggers 8534 based on the duration of a single continuous stall. 8535 8536 The default is 0, which disables the time-based panic. 8537 8538 workqueue.cpu_intensive_thresh_us= 8539 Per-cpu work items which run for longer than this 8540 threshold are automatically considered CPU intensive 8541 and excluded from concurrency management to prevent 8542 them from noticeably delaying other per-cpu work 8543 items. Default is 10000 (10ms). 8544 8545 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 8546 will report the work functions which violate this 8547 threshold repeatedly. They are likely good 8548 candidates for using WQ_UNBOUND workqueues instead. 8549 8550 workqueue.cpu_intensive_warning_thresh=<uint> 8551 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 8552 will report the work functions which violate the 8553 intensive_threshold_us repeatedly. In order to prevent 8554 spurious warnings, start printing only after a work 8555 function has violated this threshold number of times. 8556 8557 The default is 4 times. 0 disables the warning. 8558 8559 workqueue.power_efficient 8560 Per-cpu workqueues are generally preferred because 8561 they show better performance thanks to cache 8562 locality; unfortunately, per-cpu workqueues tend to 8563 be more power hungry than unbound workqueues. 8564 8565 Enabling this makes the per-cpu workqueues which 8566 were observed to contribute significantly to power 8567 consumption unbound, leading to measurably lower 8568 power usage at the cost of small performance 8569 overhead. 8570 8571 The default value of this parameter is determined by 8572 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT. 8573 8574 workqueue.default_affinity_scope= 8575 Select the default affinity scope to use for unbound 8576 workqueues. Can be one of "cpu", "smt", "cache", 8577 "cache_shard", "numa" and "system". Default is 8578 "cache_shard". For more 8579 information, see the Affinity Scopes section in 8580 Documentation/core-api/workqueue.rst. 8581 8582 This can be changed after boot by writing to the 8583 matching /sys/module/workqueue/parameters file. All 8584 workqueues with the "default" affinity scope will be 8585 updated accordingly. 8586 8587 workqueue.debug_force_rr_cpu 8588 Workqueue used to implicitly guarantee that work 8589 items queued without explicit CPU specified are put 8590 on the local CPU. This guarantee is no longer true 8591 and while local CPU is still preferred work items 8592 may be put on foreign CPUs. This debug option 8593 forces round-robin CPU selection to flush out 8594 usages which depend on the now broken guarantee. 8595 When enabled, memory and cache locality will be 8596 impacted. 8597 8598 writecombine= [LOONGARCH,EARLY] Control the MAT (Memory Access 8599 Type) of ioremap_wc(). 8600 8601 on - Enable writecombine, use WUC for ioremap_wc() 8602 off - Disable writecombine, use SUC for ioremap_wc() 8603 8604 x2apic_phys [X86-64,APIC,EARLY] Use x2apic physical mode instead of 8605 default x2apic cluster mode on platforms 8606 supporting x2apic. 8607 8608 xen_512gb_limit [KNL,X86-64,XEN] 8609 Restricts the kernel running paravirtualized under Xen 8610 to use only up to 512 GB of RAM. The reason to do so is 8611 crash analysis tools and Xen tools for doing domain 8612 save/restore/migration must be enabled to handle larger 8613 domains. 8614 8615 xen_console_io [XEN,EARLY] 8616 Boolean option to enable/disable the usage of the Xen 8617 console_io hypercalls to read and write to the console. 8618 Mostly useful for debugging and development. 8619 8620 xen_emul_unplug= [HW,X86,XEN,EARLY] 8621 Unplug Xen emulated devices 8622 Format: [unplug0,][unplug1] 8623 ide-disks -- unplug primary master IDE devices 8624 aux-ide-disks -- unplug non-primary-master IDE devices 8625 nics -- unplug network devices 8626 all -- unplug all emulated devices (NICs and IDE disks) 8627 unnecessary -- unplugging emulated devices is 8628 unnecessary even if the host did not respond to 8629 the unplug protocol 8630 never -- do not unplug even if version check succeeds 8631 8632 xen_legacy_crash [X86,XEN,EARLY] 8633 Crash from Xen panic notifier, without executing late 8634 panic() code such as dumping handler. 8635 8636 xen_mc_debug [X86,XEN,EARLY] 8637 Enable multicall debugging when running as a Xen PV guest. 8638 Enabling this feature will reduce performance a little 8639 bit, so it should only be enabled for obtaining extended 8640 debug data in case of multicall errors. 8641 8642 xen_msr_safe= [X86,XEN,EARLY] 8643 Format: <bool> 8644 Select whether to always use non-faulting (safe) MSR 8645 access functions when running as Xen PV guest. The 8646 default value is controlled by CONFIG_XEN_PV_MSR_SAFE. 8647 8648 xen_nopv [X86] 8649 Disables the PV optimizations forcing the HVM guest to 8650 run as generic HVM guest with no PV drivers. 8651 This option is obsoleted by the "nopv" option, which 8652 has equivalent effect for XEN platform. 8653 8654 xen_no_vector_callback 8655 [KNL,X86,XEN,EARLY] Disable the vector callback for Xen 8656 event channel interrupts. 8657 8658 xen_scrub_pages= [XEN] 8659 Boolean option to control scrubbing pages before giving them back 8660 to Xen, for use by other domains. Can be also changed at runtime 8661 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages. 8662 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT. 8663 8664 xen_timer_slop= [X86-64,XEN,EARLY] 8665 Set the timer slop (in nanoseconds) for the virtual Xen 8666 timers (default is 100000). This adjusts the minimum 8667 delta of virtualized Xen timers, where lower values 8668 improve timer resolution at the expense of processing 8669 more timer interrupts. 8670 8671 xen.balloon_boot_timeout= [XEN] 8672 The time (in seconds) to wait before giving up to boot 8673 in case initial ballooning fails to free enough memory. 8674 Applies only when running as HVM or PVH guest and 8675 started with less memory configured than allowed at 8676 max. Default is 180. 8677 8678 xen.event_eoi_delay= [XEN] 8679 How long to delay EOI handling in case of event 8680 storms (jiffies). Default is 10. 8681 8682 xen.event_loop_timeout= [XEN] 8683 After which time (jiffies) the event handling loop 8684 should start to delay EOI handling. Default is 2. 8685 8686 xen.fifo_events= [XEN] 8687 Boolean parameter to disable using fifo event handling 8688 even if available. Normally fifo event handling is 8689 preferred over the 2-level event handling, as it is 8690 fairer and the number of possible event channels is 8691 much higher. Default is on (use fifo events). 8692 8693 xirc2ps_cs= [NET,PCMCIA] 8694 Format: 8695 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]] 8696 8697 xive= [PPC] 8698 By default on POWER9 and above, the kernel will 8699 natively use the XIVE interrupt controller. This option 8700 allows the fallback firmware mode to be used: 8701 8702 off Fallback to firmware control of XIVE interrupt 8703 controller on both pseries and powernv 8704 platforms. Only useful on POWER9 and above. 8705 8706 xive.store-eoi=off [PPC] 8707 By default on POWER10 and above, the kernel will use 8708 stores for EOI handling when the XIVE interrupt mode 8709 is active. This option allows the XIVE driver to use 8710 loads instead, as on POWER9. 8711 8712 xhci-hcd.quirks [USB,KNL] 8713 A hex value specifying bitmask with supplemental xhci 8714 host controller quirks. Meaning of each bit can be 8715 consulted in header drivers/usb/host/xhci.h. 8716 8717 xmon [PPC,EARLY] 8718 Format: { early | on | rw | ro | off } 8719 Controls if xmon debugger is enabled. Default is off. 8720 Passing only "xmon" is equivalent to "xmon=early". 8721 early Call xmon as early as possible on boot; xmon 8722 debugger is called from setup_arch(). 8723 on xmon debugger hooks will be installed so xmon 8724 is only called on a kernel crash. Default mode, 8725 i.e. either "ro" or "rw" mode, is controlled 8726 with CONFIG_XMON_DEFAULT_RO_MODE. 8727 rw xmon debugger hooks will be installed so xmon 8728 is called only on a kernel crash, mode is write, 8729 meaning SPR registers, memory and, other data 8730 can be written using xmon commands. 8731 ro same as "rw" option above but SPR registers, 8732 memory, and other data can't be written using 8733 xmon commands. 8734 off xmon is disabled. 8735