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 DRM Direct Rendering Management support is enabled. 12 DYNAMIC_DEBUG Build in debug messages and enable them at runtime 13 EARLY Parameter processed too early to be embedded in initrd. 14 EDD BIOS Enhanced Disk Drive Services (EDD) is enabled 15 EFI EFI Partitioning (GPT) is enabled 16 EVM Extended Verification Module 17 FB The frame buffer device is enabled. 18 FTRACE Function tracing enabled. 19 GCOV GCOV profiling is enabled. 20 HIBERNATION HIBERNATION is enabled. 21 HW Appropriate hardware is enabled. 22 HYPER_V HYPERV support is enabled. 23 IMA Integrity measurement architecture is enabled. 24 IP_PNP IP DHCP, BOOTP, or RARP is enabled. 25 IPV6 IPv6 support is enabled. 26 ISAPNP ISA PnP code is enabled. 27 ISOL CPU Isolation is enabled. 28 JOY Appropriate joystick support is enabled. 29 KGDB Kernel debugger support is enabled. 30 KVM Kernel Virtual Machine support is enabled. 31 LIBATA Libata driver is enabled 32 LOONGARCH LoongArch architecture is enabled. 33 LOOP Loopback device support is enabled. 34 LP Printer support is enabled. 35 M68k M68k architecture is enabled. 36 These options have more detailed description inside of 37 Documentation/arch/m68k/kernel-options.rst. 38 MIPS MIPS architecture is enabled. 39 MOUSE Appropriate mouse support is enabled. 40 MSI Message Signaled Interrupts (PCI). 41 MTD MTD (Memory Technology Device) support is enabled. 42 NET Appropriate network support is enabled. 43 NFS Appropriate NFS support is enabled. 44 NUMA NUMA support is enabled. 45 OF Devicetree is enabled. 46 PARISC The PA-RISC architecture is enabled. 47 PCI PCI bus support is enabled. 48 PCIE PCI Express support is enabled. 49 PCMCIA The PCMCIA subsystem is enabled. 50 PNP Plug & Play support is enabled. 51 PPC PowerPC architecture is enabled. 52 PPT Parallel port support is enabled. 53 PS2 Appropriate PS/2 support is enabled. 54 PV_OPS A paravirtualized kernel is enabled. 55 RAM RAM disk support is enabled. 56 RDT Intel Resource Director Technology. 57 RISCV RISCV architecture is enabled. 58 S390 S390 architecture is enabled. 59 SCSI Appropriate SCSI support is enabled. 60 A lot of drivers have their options described inside 61 the Documentation/scsi/ sub-directory. 62 SDW SoundWire support is enabled. 63 SECURITY Different security models are enabled. 64 SELINUX SELinux support is enabled. 65 SERIAL Serial support is enabled. 66 SH SuperH architecture is enabled. 67 SMP The kernel is an SMP kernel. 68 SPARC Sparc architecture is enabled. 69 SUSPEND System suspend states are enabled. 70 SWSUSP Software suspend (hibernation) is enabled. 71 TPM TPM drivers are enabled. 72 UMS USB Mass Storage support is enabled. 73 USB USB support is enabled. 74 NVME NVMe support is enabled 75 USBHID USB Human Interface Device support is enabled. 76 V4L Video For Linux support is enabled. 77 VGA The VGA console has been enabled. 78 VMMIO Driver for memory mapped virtio devices is enabled. 79 VT Virtual terminal support is enabled. 80 WDT Watchdog support is enabled. 81 X86-32 X86-32, aka i386 architecture is enabled. 82 X86-64 X86-64 architecture is enabled. 83 X86 Either 32-bit or 64-bit x86 (same as X86-32+X86-64) 84 X86_UV SGI UV support is enabled. 85 XEN Xen support is enabled 86 XTENSA xtensa architecture is enabled. 87 88In addition, the following text indicates that the option 89 90 BOOT Is a boot loader parameter. 91 BUGS= Relates to possible processor bugs on the said processor. 92 KNL Is a kernel start-up parameter. 93 94 95Kernel parameters 96 97 accept_memory= [MM] 98 Format: { eager | lazy } 99 default: lazy 100 By default, unaccepted memory is accepted lazily to 101 avoid prolonged boot times. The lazy option will add 102 some runtime overhead until all memory is eventually 103 accepted. In most cases the overhead is negligible. 104 For some workloads or for debugging purposes 105 accept_memory=eager can be used to accept all memory 106 at once during boot. 107 108 acpi= [HW,ACPI,X86,ARM64,RISCV64,EARLY] 109 Advanced Configuration and Power Interface 110 Format: { force | on | off | strict | noirq | rsdt | 111 copy_dsdt | nospcr } 112 force -- enable ACPI if default was off 113 on -- enable ACPI but allow fallback to DT [arm64,riscv64] 114 off -- disable ACPI if default was on 115 noirq -- do not use ACPI for IRQ routing 116 strict -- Be less tolerant of platforms that are not 117 strictly ACPI specification compliant. 118 rsdt -- prefer RSDT over (default) XSDT 119 copy_dsdt -- copy DSDT to memory 120 nocmcff -- Disable firmware first mode for corrected 121 errors. This disables parsing the HEST CMC error 122 source to check if firmware has set the FF flag. This 123 may result in duplicate corrected error reports. 124 nospcr -- disable console in ACPI SPCR table as 125 default _serial_ console on ARM64 126 spcr -- enable console in ACPI SPCR table as 127 default _serial_ console on x86 128 For ARM64, ONLY "acpi=off", "acpi=on", "acpi=force" or 129 "acpi=nospcr" are available 130 For RISCV64, ONLY "acpi=off", "acpi=on" or "acpi=force" 131 are available 132 133 See also Documentation/power/runtime_pm.rst, pci=noacpi 134 135 acpi_apic_instance= [ACPI,IOAPIC,EARLY] 136 Format: <int> 137 2: use 2nd APIC table, if available 138 1,0: use 1st APIC table 139 default: 0 140 141 acpi_backlight= [HW,ACPI] 142 { vendor | video | native | none } 143 If set to vendor, prefer vendor-specific driver 144 (e.g. thinkpad_acpi, sony_acpi, etc.) instead 145 of the ACPI video.ko driver. 146 If set to video, use the ACPI video.ko driver. 147 If set to native, use the device's native backlight mode. 148 If set to none, disable the ACPI backlight interface. 149 150 acpi_force_32bit_fadt_addr [ACPI,EARLY] 151 force FADT to use 32 bit addresses rather than the 152 64 bit X_* addresses. Some firmware have broken 64 153 bit addresses for force ACPI ignore these and use 154 the older legacy 32 bit addresses. 155 156 acpica_no_return_repair [HW, ACPI] 157 Disable AML predefined validation mechanism 158 This mechanism can repair the evaluation result to make 159 the return objects more ACPI specification compliant. 160 This option is useful for developers to identify the 161 root cause of an AML interpreter issue when the issue 162 has something to do with the repair mechanism. 163 164 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG] 165 acpi.debug_level= [HW,ACPI,ACPI_DEBUG] 166 Format: <int> 167 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI 168 debug output. Bits in debug_layer correspond to a 169 _COMPONENT in an ACPI source file, e.g., 170 #define _COMPONENT ACPI_EVENTS 171 Bits in debug_level correspond to a level in 172 ACPI_DEBUG_PRINT statements, e.g., 173 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ... 174 The debug_level mask defaults to "info". See 175 Documentation/firmware-guide/acpi/debug.rst for more information about 176 debug layers and levels. 177 178 Enable processor driver info messages: 179 acpi.debug_layer=0x20000000 180 Enable AML "Debug" output, i.e., stores to the Debug 181 object while interpreting AML: 182 acpi.debug_layer=0xffffffff acpi.debug_level=0x2 183 Enable all messages related to ACPI hardware: 184 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff 185 186 Some values produce so much output that the system is 187 unusable. The "log_buf_len" parameter may be useful 188 if you need to capture more output. 189 190 acpi.poweroff_on_fatal= [ACPI] 191 {0 | 1} 192 Causes the system to poweroff when the ACPI bytecode signals 193 a fatal error. The default value of this setting is 1. 194 Overriding this value should only be done for diagnosing 195 ACPI firmware problems, as the system might behave erratically 196 after having encountered a fatal ACPI error. 197 198 acpi_enforce_resources= [ACPI] 199 { strict | lax | no } 200 Check for resource conflicts between native drivers 201 and ACPI OperationRegions (SystemIO and SystemMemory 202 only). IO ports and memory declared in ACPI might be 203 used by the ACPI subsystem in arbitrary AML code and 204 can interfere with legacy drivers. 205 strict (default): access to resources claimed by ACPI 206 is denied; legacy drivers trying to access reserved 207 resources will fail to bind to device using them. 208 lax: access to resources claimed by ACPI is allowed; 209 legacy drivers trying to access reserved resources 210 will bind successfully but a warning message is logged. 211 no: ACPI OperationRegions are not marked as reserved, 212 no further checks are performed. 213 214 acpi_force_table_verification [HW,ACPI,EARLY] 215 Enable table checksum verification during early stage. 216 By default, this is disabled due to x86 early mapping 217 size limitation. 218 219 acpi_irq_balance [HW,ACPI] 220 ACPI will balance active IRQs 221 default in APIC mode 222 223 acpi_irq_nobalance [HW,ACPI] 224 ACPI will not move active IRQs (default) 225 default in PIC mode 226 227 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA 228 Format: <irq>,<irq>... 229 230 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for 231 use by PCI 232 Format: <irq>,<irq>... 233 234 acpi_mask_gpe= [HW,ACPI] 235 Due to the existence of _Lxx/_Exx, some GPEs triggered 236 by unsupported hardware/firmware features can result in 237 GPE floodings that cannot be automatically disabled by 238 the GPE dispatcher. 239 This facility can be used to prevent such uncontrolled 240 GPE floodings. 241 Format: <byte> or <bitmap-list> 242 243 acpi_no_auto_serialize [HW,ACPI] 244 Disable auto-serialization of AML methods 245 AML control methods that contain the opcodes to create 246 named objects will be marked as "Serialized" by the 247 auto-serialization feature. 248 This feature is enabled by default. 249 This option allows to turn off the feature. 250 251 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump 252 kernels. 253 254 acpi_no_static_ssdt [HW,ACPI,EARLY] 255 Disable installation of static SSDTs at early boot time 256 By default, SSDTs contained in the RSDT/XSDT will be 257 installed automatically and they will appear under 258 /sys/firmware/acpi/tables. 259 This option turns off this feature. 260 Note that specifying this option does not affect 261 dynamic table installation which will install SSDT 262 tables to /sys/firmware/acpi/tables/dynamic. 263 264 acpi_no_watchdog [HW,ACPI,WDT] 265 Ignore the ACPI-based watchdog interface (WDAT) and let 266 a native driver control the watchdog device instead. 267 268 acpi_rsdp= [ACPI,EFI,KEXEC,EARLY] 269 Pass the RSDP address to the kernel, mostly used 270 on machines running EFI runtime service to boot the 271 second kernel for kdump. 272 273 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS 274 Format: To spoof as Windows 98: ="Microsoft Windows" 275 276 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead 277 of 2 which is mandated by ACPI 6) as the supported ACPI 278 specification revision (when using this switch, it may 279 be necessary to carry out a cold reboot _twice_ in a 280 row to make it take effect on the platform firmware). 281 282 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings 283 acpi_osi="string1" # add string1 284 acpi_osi="!string2" # remove string2 285 acpi_osi=!* # remove all strings 286 acpi_osi=! # disable all built-in OS vendor 287 strings 288 acpi_osi=!! # enable all built-in OS vendor 289 strings 290 acpi_osi= # disable all strings 291 292 'acpi_osi=!' can be used in combination with single or 293 multiple 'acpi_osi="string1"' to support specific OS 294 vendor string(s). Note that such command can only 295 affect the default state of the OS vendor strings, thus 296 it cannot affect the default state of the feature group 297 strings and the current state of the OS vendor strings, 298 specifying it multiple times through kernel command line 299 is meaningless. This command is useful when one do not 300 care about the state of the feature group strings which 301 should be controlled by the OSPM. 302 Examples: 303 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent 304 to 'acpi_osi="Windows 2000" acpi_osi=!', they all 305 can make '_OSI("Windows 2000")' TRUE. 306 307 'acpi_osi=' cannot be used in combination with other 308 'acpi_osi=' command lines, the _OSI method will not 309 exist in the ACPI namespace. NOTE that such command can 310 only affect the _OSI support state, thus specifying it 311 multiple times through kernel command line is also 312 meaningless. 313 Examples: 314 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)' 315 FALSE. 316 317 'acpi_osi=!*' can be used in combination with single or 318 multiple 'acpi_osi="string1"' to support specific 319 string(s). Note that such command can affect the 320 current state of both the OS vendor strings and the 321 feature group strings, thus specifying it multiple times 322 through kernel command line is meaningful. But it may 323 still not able to affect the final state of a string if 324 there are quirks related to this string. This command 325 is useful when one want to control the state of the 326 feature group strings to debug BIOS issues related to 327 the OSPM features. 328 Examples: 329 1. 'acpi_osi="Module Device" acpi_osi=!*' can make 330 '_OSI("Module Device")' FALSE. 331 2. 'acpi_osi=!* acpi_osi="Module Device"' can make 332 '_OSI("Module Device")' TRUE. 333 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is 334 equivalent to 335 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"' 336 and 337 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!', 338 they all will make '_OSI("Windows 2000")' TRUE. 339 340 acpi_pm_good [X86] 341 Override the pmtimer bug detection: force the kernel 342 to assume that this machine's pmtimer latches its value 343 and always returns good values. 344 345 acpi_sci= [HW,ACPI,EARLY] ACPI System Control Interrupt trigger mode 346 Format: { level | edge | high | low } 347 348 acpi_skip_timer_override [HW,ACPI,EARLY] 349 Recognize and ignore IRQ0/pin2 Interrupt Override. 350 For broken nForce2 BIOS resulting in XT-PIC timer. 351 352 acpi_sleep= [HW,ACPI] Sleep options 353 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig, 354 s4_nohwsig, old_ordering, nonvs, 355 sci_force_enable, nobl } 356 See Documentation/power/video.rst for information on 357 s3_bios and s3_mode. 358 s3_beep is for debugging; it makes the PC's speaker beep 359 as soon as the kernel's real-mode entry point is called. 360 s4_hwsig causes the kernel to check the ACPI hardware 361 signature during resume from hibernation, and gracefully 362 refuse to resume if it has changed. This complies with 363 the ACPI specification but not with reality, since 364 Windows does not do this and many laptops do change it 365 on docking. So the default behaviour is to allow resume 366 and simply warn when the signature changes, unless the 367 s4_hwsig option is enabled. 368 s4_nohwsig prevents ACPI hardware signature from being 369 used (or even warned about) during resume. 370 old_ordering causes the ACPI 1.0 ordering of the _PTS 371 control method, with respect to putting devices into 372 low power states, to be enforced (the ACPI 2.0 ordering 373 of _PTS is used by default). 374 nonvs prevents the kernel from saving/restoring the 375 ACPI NVS memory during suspend/hibernation and resume. 376 sci_force_enable causes the kernel to set SCI_EN directly 377 on resume from S1/S3 (which is against the ACPI spec, 378 but some broken systems don't work without it). 379 nobl causes the internal blacklist of systems known to 380 behave incorrectly in some ways with respect to system 381 suspend and resume to be ignored (use wisely). 382 383 acpi_use_timer_override [HW,ACPI,EARLY] 384 Use timer override. For some broken Nvidia NF5 boards 385 that require a timer override, but don't have HPET 386 387 add_efi_memmap [EFI,X86,EARLY] Include EFI memory map in 388 kernel's map of available physical RAM. 389 390 agp= [AGP] 391 { off | try_unsupported } 392 off: disable AGP support 393 try_unsupported: try to drive unsupported chipsets 394 (may crash computer or cause data corruption) 395 396 ALSA [HW,ALSA] 397 See Documentation/sound/alsa-configuration.rst 398 399 alignment= [KNL,ARM] 400 Allow the default userspace alignment fault handler 401 behaviour to be specified. Bit 0 enables warnings, 402 bit 1 enables fixups, and bit 2 sends a segfault. 403 404 align_va_addr= [X86-64] 405 Align virtual addresses by clearing slice [14:12] when 406 allocating a VMA at process creation time. This option 407 gives you up to 3% performance improvement on AMD F15h 408 machines (where it is enabled by default) for a 409 CPU-intensive style benchmark, and it can vary highly in 410 a microbenchmark depending on workload and compiler. 411 412 32: only for 32-bit processes 413 64: only for 64-bit processes 414 on: enable for both 32- and 64-bit processes 415 off: disable for both 32- and 64-bit processes 416 417 alloc_snapshot [FTRACE] 418 Allocate the ftrace snapshot buffer on boot up when the 419 main buffer is allocated. This is handy if debugging 420 and you need to use tracing_snapshot() on boot up, and 421 do not want to use tracing_snapshot_alloc() as it needs 422 to be done where GFP_KERNEL allocations are allowed. 423 424 allow_mismatched_32bit_el0 [ARM64,EARLY] 425 Allow execve() of 32-bit applications and setting of the 426 PER_LINUX32 personality on systems where only a strict 427 subset of the CPUs support 32-bit EL0. When this 428 parameter is present, the set of CPUs supporting 32-bit 429 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0 430 and hot-unplug operations may be restricted. 431 432 See Documentation/arch/arm64/asymmetric-32bit.rst for more 433 information. 434 435 amd_iommu= [HW,X86-64] 436 Pass parameters to the AMD IOMMU driver in the system. 437 Possible values are: 438 fullflush - Deprecated, equivalent to iommu.strict=1 439 off - do not initialize any AMD IOMMU found in 440 the system 441 force_isolation - Force device isolation for all 442 devices. The IOMMU driver is not 443 allowed anymore to lift isolation 444 requirements as needed. This option 445 does not override iommu=pt 446 force_enable - Force enable the IOMMU on platforms known 447 to be buggy with IOMMU enabled. Use this 448 option with care. 449 pgtbl_v1 - Use v1 page table for DMA-API (Default). 450 pgtbl_v2 - Use v2 page table for DMA-API. 451 irtcachedis - Disable Interrupt Remapping Table (IRT) caching. 452 nohugepages - Limit page-sizes used for v1 page-tables 453 to 4 KiB. 454 v2_pgsizes_only - Limit page-sizes used for v1 page-tables 455 to 4KiB/2Mib/1GiB. 456 457 458 amd_iommu_dump= [HW,X86-64] 459 Enable AMD IOMMU driver option to dump the ACPI table 460 for AMD IOMMU. With this option enabled, AMD IOMMU 461 driver will print ACPI tables for AMD IOMMU during 462 IOMMU initialization. 463 464 amd_iommu_intr= [HW,X86-64] 465 Specifies one of the following AMD IOMMU interrupt 466 remapping modes: 467 legacy - Use legacy interrupt remapping mode. 468 vapic - Use virtual APIC mode, which allows IOMMU 469 to inject interrupts directly into guest. 470 This mode requires kvm-amd.avic=1. 471 (Default when IOMMU HW support is present.) 472 473 amd_pstate= [X86,EARLY] 474 disable 475 Do not enable amd_pstate as the default 476 scaling driver for the supported processors 477 passive 478 Use amd_pstate with passive mode as a scaling driver. 479 In this mode autonomous selection is disabled. 480 Driver requests a desired performance level and platform 481 tries to match the same performance level if it is 482 satisfied by guaranteed performance level. 483 active 484 Use amd_pstate_epp driver instance as the scaling driver, 485 driver provides a hint to the hardware if software wants 486 to bias toward performance (0x0) or energy efficiency (0xff) 487 to the CPPC firmware. then CPPC power algorithm will 488 calculate the runtime workload and adjust the realtime cores 489 frequency. 490 guided 491 Activate guided autonomous mode. Driver requests minimum and 492 maximum performance level and the platform autonomously 493 selects a performance level in this range and appropriate 494 to the current workload. 495 496 amd_prefcore= 497 [X86] 498 disable 499 Disable amd-pstate preferred core. 500 501 amd_dynamic_epp= 502 [X86] 503 disable 504 Disable amd-pstate dynamic EPP. 505 enable 506 Enable amd-pstate dynamic EPP. 507 508 amijoy.map= [HW,JOY] Amiga joystick support 509 Map of devices attached to JOY0DAT and JOY1DAT 510 Format: <a>,<b> 511 See also Documentation/input/joydev/joystick.rst 512 513 analog.map= [HW,JOY] Analog joystick and gamepad support 514 Specifies type or capabilities of an analog joystick 515 connected to one of 16 gameports 516 Format: <type1>,<type2>,..<type16> 517 518 apc= [HW,SPARC] 519 Power management functions (SPARCstation-4/5 + deriv.) 520 Format: noidle 521 Disable APC CPU standby support. SPARCstation-Fox does 522 not play well with APC CPU idle - disable it if you have 523 APC and your system crashes randomly. 524 525 apic [APIC,X86-64] Use IO-APIC. Default. 526 527 apic= [APIC,X86,EARLY] Advanced Programmable Interrupt Controller 528 Change the output verbosity while booting 529 Format: { quiet (default) | verbose | debug } 530 Change the amount of debugging information output 531 when initialising the APIC and IO-APIC components. 532 533 apic_extnmi= [APIC,X86,EARLY] External NMI delivery setting 534 Format: { bsp (default) | all | none } 535 bsp: External NMI is delivered only to CPU 0 536 all: External NMIs are broadcast to all CPUs as a 537 backup of CPU 0 538 none: External NMI is masked for all CPUs. This is 539 useful so that a dump capture kernel won't be 540 shot down by NMI 541 542 apicpmtimer Do APIC timer calibration using the pmtimer. Implies 543 apicmaintimer. Useful when your PIT timer is totally 544 broken. 545 546 autoconf= [IPV6] 547 See Documentation/networking/ipv6.rst. 548 549 apm= [APM] Advanced Power Management 550 See header of arch/x86/kernel/apm_32.c. 551 552 apparmor= [APPARMOR] Disable or enable AppArmor at boot time 553 Format: { "0" | "1" } 554 See security/apparmor/Kconfig help text 555 0 -- disable. 556 1 -- enable. 557 Default value is set via kernel config option. 558 559 arm64.no32bit_el0 [ARM64] Unconditionally disable the execution of 560 32 bit applications. 561 562 arm64.nobti [ARM64] Unconditionally disable Branch Target 563 Identification support 564 565 arm64.nogcs [ARM64] Unconditionally disable Guarded Control Stack 566 support 567 568 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory 569 Set instructions support 570 571 arm64.nompam [ARM64] Unconditionally disable Memory Partitioning And 572 Monitoring support 573 574 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension 575 support 576 577 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication 578 support 579 580 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix 581 Extension support 582 583 arm64.nosve [ARM64] Unconditionally disable Scalable Vector 584 Extension support 585 586 ataflop= [HW,M68k] 587 588 atarimouse= [HW,MOUSE] Atari Mouse 589 590 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess, 591 EzKey and similar keyboards 592 593 atkbd.reset= [HW] Reset keyboard during initialization 594 595 atkbd.set= [HW] Select keyboard code set 596 Format: <int> (2 = AT (default), 3 = PS/2) 597 598 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar 599 keyboards 600 601 atkbd.softraw= [HW] Choose between synthetic and real raw mode 602 Format: <bool> (0 = real, 1 = synthetic (default)) 603 604 atkbd.softrepeat= [HW] 605 Use software keyboard repeat 606 607 audit= [KNL] Enable the audit sub-system 608 Format: { "0" | "1" | "off" | "on" } 609 0 | off - kernel audit is disabled and can not be 610 enabled until the next reboot 611 unset - kernel audit is initialized but disabled and 612 will be fully enabled by the userspace auditd. 613 1 | on - kernel audit is initialized and partially 614 enabled, storing at most audit_backlog_limit 615 messages in RAM until it is fully enabled by the 616 userspace auditd. 617 Default: unset 618 619 audit_backlog_limit= [KNL] Set the audit queue size limit. 620 Format: <int> (must be >=0) 621 Default: 64 622 623 bau= [X86_UV] Enable the BAU on SGI UV. The default 624 behavior is to disable the BAU (i.e. bau=0). 625 Format: { "0" | "1" } 626 0 - Disable the BAU. 627 1 - Enable the BAU. 628 unset - Disable the BAU. 629 630 bdev_allow_write_mounted= 631 Format: <bool> 632 Control the ability to open a mounted block device 633 for writing, i.e., allow / disallow writes that bypass 634 the FS. This was implemented as a means to prevent 635 fuzzers from crashing the kernel by overwriting the 636 metadata underneath a mounted FS without its awareness. 637 This also prevents destructive formatting of mounted 638 filesystems by naive storage tooling that don't use 639 O_EXCL. Default is Y and can be changed through the 640 Kconfig option CONFIG_BLK_DEV_WRITE_MOUNTED. 641 642 bert_disable [ACPI] 643 Disable BERT OS support on buggy BIOSes. 644 645 bgrt_disable [ACPI,X86,EARLY] 646 Disable BGRT to avoid flickering OEM logo. 647 648 blkdevparts= Manual partition parsing of block device(s) for 649 embedded devices based on command line input. 650 See Documentation/block/cmdline-partition.rst 651 652 boot_delay= [KNL,EARLY] 653 Milliseconds to delay each printk during boot. 654 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled, 655 and you may also have to specify "lpj=". Boot_delay 656 values larger than 10 seconds (10000) are assumed 657 erroneous and ignored. 658 Format: integer 659 660 bootconfig [KNL,EARLY] 661 Extended command line options can be added to an initrd 662 and this will cause the kernel to look for it. 663 664 See Documentation/admin-guide/bootconfig.rst 665 666 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards) 667 bttv.radio= Most important insmod options are available as 668 kernel args too. 669 bttv.pll= See Documentation/admin-guide/media/bttv.rst 670 bttv.tuner= 671 672 bulk_remove=off [PPC] This parameter disables the use of the pSeries 673 firmware feature for flushing multiple hpte entries 674 at a time. 675 676 c101= [NET] Moxa C101 synchronous serial card 677 678 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection. 679 Sometimes CPU hardware bugs make them report the cache 680 size incorrectly. The kernel will attempt work arounds 681 to fix known problems, but for some CPUs it is not 682 possible to determine what the correct size should be. 683 This option provides an override for these situations. 684 685 carrier_timeout= 686 [NET] Specifies amount of time (in seconds) that 687 the kernel should wait for a network carrier. By default 688 it waits 120 seconds. 689 690 ca_keys= [KEYS] This parameter identifies a specific key(s) on 691 the system trusted keyring to be used for certificate 692 trust validation. 693 format: { id:<keyid> | builtin } 694 695 cca= [MIPS,EARLY] Override the kernel pages' cache coherency 696 algorithm. Accepted values range from 0 to 7 697 inclusive. See arch/mips/include/asm/pgtable-bits.h 698 for platform specific values (SB1, Loongson3 and 699 others). 700 701 ccw_timeout_log [S390] 702 See Documentation/arch/s390/common_io.rst for details. 703 704 cfi= [X86-64] Set Control Flow Integrity checking features 705 when CONFIG_FINEIBT is enabled. 706 Format: feature[,feature...] 707 Default: auto 708 709 auto: Use FineIBT if IBT available, otherwise kCFI. 710 Under FineIBT, enable "paranoid" mode when 711 FRED is not available. 712 off: Turn off CFI checking. 713 kcfi: Use kCFI (disable FineIBT). 714 fineibt: Use FineIBT (even if IBT not available). 715 norand: Do not re-randomize CFI hashes. 716 paranoid: Add caller hash checking under FineIBT. 717 bhi: Enable register poisoning to stop speculation 718 across FineIBT. (Disabled by default.) 719 warn: Do not enforce CFI checking: warn only. 720 debug: Report CFI initialization details. 721 722 cgroup_disable= [KNL] Disable a particular controller or optional feature 723 Format: {name of the controller(s) or feature(s) to disable} 724 The effects of cgroup_disable=foo are: 725 - foo isn't auto-mounted if you mount all cgroups in 726 a single hierarchy 727 - foo isn't visible as an individually mountable 728 subsystem 729 - if foo is an optional feature then the feature is 730 disabled and corresponding cgroup files are not 731 created 732 {Currently only "memory" controller deal with this and 733 cut the overhead, others just disable the usage. So 734 only cgroup_disable=memory is actually worthy} 735 Specifying "pressure" disables per-cgroup pressure 736 stall information accounting feature 737 738 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1 739 Format: { { controller | "all" | "named" } 740 [,{ controller | "all" | "named" }...] } 741 Like cgroup_disable, but only applies to cgroup v1; 742 the blacklisted controllers remain available in cgroup2. 743 "all" blacklists all controllers and "named" disables 744 named mounts. Specifying both "all" and "named" disables 745 all v1 hierarchies. 746 747 cgroup_v1_proc= [KNL] Show also missing controllers in /proc/cgroups 748 Format: { "true" | "false" } 749 /proc/cgroups lists only v1 controllers by default. 750 This compatibility option enables listing also v2 751 controllers (whose v1 code is not compiled!), so that 752 semi-legacy software can check this file to decide 753 about usage of v2 (sic) controllers. 754 755 cgroup_favordynmods= [KNL] Enable or Disable favordynmods. 756 Format: { "true" | "false" } 757 Defaults to the value of CONFIG_CGROUP_FAVOR_DYNMODS. 758 759 cgroup.memory= [KNL] Pass options to the cgroup memory controller. 760 Format: <string> 761 nosocket -- Disable socket memory accounting. 762 nokmem -- Disable kernel memory accounting. 763 nobpf -- Disable BPF memory accounting. 764 765 check_pages= [MM,EARLY] Enable sanity checking of pages after 766 allocations / before freeing. This adds checks to catch 767 double-frees, use-after-frees, and other sources of 768 page corruption by inspecting page internals (flags, 769 mapcount/refcount, memcg_data, etc.). 770 Format: { "0" | "1" } 771 Default: 0 (1 if CONFIG_DEBUG_VM is set) 772 773 checkreqprot= [SELINUX] Set initial checkreqprot flag value. 774 Format: { "0" | "1" } 775 See security/selinux/Kconfig help text. 776 0 -- check protection applied by kernel (includes 777 any implied execute protection). 778 1 -- check protection requested by application. 779 Default value is set via a kernel config option. 780 Value can be changed at runtime via 781 /sys/fs/selinux/checkreqprot. 782 Setting checkreqprot to 1 is deprecated. 783 784 cio_ignore= [S390] 785 See Documentation/arch/s390/common_io.rst for details. 786 787 clk_ignore_unused 788 [CLK] 789 Prevents the clock framework from automatically gating 790 clocks that have not been explicitly enabled by a Linux 791 device driver but are enabled in hardware at reset or 792 by the bootloader/firmware. Note that this does not 793 force such clocks to be always-on nor does it reserve 794 those clocks in any way. This parameter is useful for 795 debug and development, but should not be needed on a 796 platform with proper driver support. For more 797 information, see Documentation/driver-api/clk.rst. 798 799 clock= [BUGS=X86-32, HW] gettimeofday clocksource override. 800 [Deprecated] 801 Forces specified clocksource (if available) to be used 802 when calculating gettimeofday(). If specified 803 clocksource is not available, it defaults to PIT. 804 Format: { pit | tsc | cyclone | pmtmr } 805 806 clocksource= Override the default clocksource 807 Format: <string> 808 Override the default clocksource and use the clocksource 809 with the name specified. 810 Some clocksource names to choose from, depending on 811 the platform: 812 [all] jiffies (this is the base, fallback clocksource) 813 [ACPI] acpi_pm 814 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2, 815 pxa_timer,timer3,32k_counter,timer0_1 816 [X86-32] pit,hpet,tsc; 817 scx200_hrt on Geode; cyclone on IBM x440 818 [MIPS] MIPS 819 [PARISC] cr16 820 [S390] tod 821 [SH] SuperH 822 [SPARC64] tick 823 [X86-64] hpet,tsc 824 825 clocksource.arm_arch_timer.evtstrm= 826 [ARM,ARM64,EARLY] 827 Format: <bool> 828 Enable/disable the eventstream feature of the ARM 829 architected timer so that code using WFE-based polling 830 loops can be debugged more effectively on production 831 systems. 832 833 clocksource.verify_n_cpus= [KNL] 834 Limit the number of CPUs checked for clocksources 835 marked with CLOCK_SOURCE_VERIFY_PERCPU that 836 are marked unstable due to excessive skew. 837 A negative value says to check all CPUs, while 838 zero says not to check any. Values larger than 839 nr_cpu_ids are silently truncated to nr_cpu_ids. 840 The actual CPUs are chosen randomly, with 841 no replacement if the same CPU is chosen twice. 842 843 clocksource-wdtest.holdoff= [KNL] 844 Set the time in seconds that the clocksource 845 watchdog test waits before commencing its tests. 846 Defaults to zero when built as a module and to 847 10 seconds when built into the kernel. 848 849 cma=nn[MG]@[start[MG][-end[MG]]] 850 [KNL,CMA,EARLY] 851 Sets the size of kernel global memory area for 852 contiguous memory allocations and optionally the 853 placement constraint by the physical address range of 854 memory allocations. A value of 0 disables CMA 855 altogether. For more information, see 856 kernel/dma/contiguous.c 857 858 cma_pernuma=nn[MG] 859 [KNL,CMA,EARLY] 860 Sets the size of kernel per-numa memory area for 861 contiguous memory allocations. A value of 0 disables 862 per-numa CMA altogether. And If this option is not 863 specified, the default value is 0. 864 With per-numa CMA enabled, DMA users on node nid will 865 first try to allocate buffer from the pernuma area 866 which is located in node nid, if the allocation fails, 867 they will fallback to the global default memory area. 868 869 numa_cma=<node>:nn[MG][,<node>:nn[MG]] 870 [KNL,CMA,EARLY] 871 Sets the size of kernel numa memory area for 872 contiguous memory allocations. It will reserve CMA 873 area for the specified node. 874 875 With numa CMA enabled, DMA users on node nid will 876 first try to allocate buffer from the numa area 877 which is located in node nid, if the allocation fails, 878 they will fallback to the global default memory area. 879 880 cmo_free_hint= [PPC] Format: { yes | no } 881 Specify whether pages are marked as being inactive 882 when they are freed. This is used in CMO environments 883 to determine OS memory pressure for page stealing by 884 a hypervisor. 885 Default: yes 886 887 coherent_pool=nn[KMG] [ARM,KNL,EARLY] 888 Sets the size of memory pool for coherent, atomic dma 889 allocations, by default set to 256K. 890 891 condev= [HW,S390] console device 892 conmode= 893 894 con3215_drop= [S390,EARLY] 3215 console drop mode. 895 Format: y|n|Y|N|1|0 896 When set to true, drop data on the 3215 console when 897 the console buffer is full. In this case the 898 operator using a 3270 terminal emulator (for example 899 x3270) does not have to enter the clear key for the 900 console output to advance and the kernel to continue. 901 This leads to a much faster boot time when a 3270 902 terminal emulator is active. If no 3270 terminal 903 emulator is used, this parameter has no effect. 904 905 console= [KNL] Output console device and options. 906 907 tty<n> Use the virtual console device <n>. 908 909 ttyS<n>[,options] 910 ttyUSB0[,options] 911 Use the specified serial port. The options are of 912 the form "bbbbpnf", where "bbbb" is the baud rate, 913 "p" is parity ("n", "o", or "e"), "n" is number of 914 bits, and "f" is flow control ("r" for RTS or 915 omit it). Default is "9600n8". 916 917 See Documentation/admin-guide/serial-console.rst for more 918 information. See 919 Documentation/networking/netconsole.rst for an 920 alternative. 921 922 <DEVNAME>:<n>.<n>[,options] 923 Use the specified serial port on the serial core bus. 924 The addressing uses DEVNAME of the physical serial port 925 device, followed by the serial core controller instance, 926 and the serial port instance. The options are the same 927 as documented for the ttyS addressing above. 928 929 The mapping of the serial ports to the tty instances 930 can be viewed with: 931 932 $ ls -d /sys/bus/serial-base/devices/*:*.*/tty/* 933 /sys/bus/serial-base/devices/00:04:0.0/tty/ttyS0 934 935 In the above example, the console can be addressed with 936 console=00:04:0.0. Note that a console addressed this 937 way will only get added when the related device driver 938 is ready. The use of an earlycon parameter in addition to 939 the console may be desired for console output early on. 940 941 uart[8250],io,<addr>[,options] 942 uart[8250],mmio,<addr>[,options] 943 uart[8250],mmio16,<addr>[,options] 944 uart[8250],mmio32,<addr>[,options] 945 uart[8250],0x<addr>[,options] 946 Start an early, polled-mode console on the 8250/16550 947 UART at the specified I/O port or MMIO address, 948 switching to the matching ttyS device later. 949 MMIO inter-register address stride is either 8-bit 950 (mmio), 16-bit (mmio16), or 32-bit (mmio32). 951 If none of [io|mmio|mmio16|mmio32], <addr> is assumed 952 to be equivalent to 'mmio'. 'options' are specified in 953 the same format described for ttyS above; if unspecified, 954 the h/w is not re-initialized. 955 956 hvc<n> Use the hypervisor console device <n>. This is for 957 both Xen and PowerPC hypervisors. 958 959 { null | "" } 960 Use to disable console output, i.e., to have kernel 961 console messages discarded. 962 This must be the only console= parameter used on the 963 kernel command line. 964 965 If the device connected to the port is not a TTY but a braille 966 device, prepend "brl," before the device type, for instance 967 console=brl,ttyS0 968 For now, only VisioBraille is supported. 969 970 console_msg_format= 971 [KNL] Change console messages format 972 default 973 By default we print messages on consoles in 974 "[time stamp] text\n" format (time stamp may not be 975 printed, depending on CONFIG_PRINTK_TIME or 976 `printk_time' param). 977 syslog 978 Switch to syslog format: "<%u>[time stamp] text\n" 979 IOW, each message will have a facility and loglevel 980 prefix. The format is similar to one used by syslog() 981 syscall, or to executing "dmesg -S --raw" or to reading 982 from /proc/kmsg. 983 984 consoleblank= [KNL] The console blank (screen saver) timeout in 985 seconds. A value of 0 disables the blank timer. 986 Defaults to 0. 987 988 coredump_filter= 989 [KNL] Change the default value for 990 /proc/<pid>/coredump_filter. 991 See also Documentation/filesystems/proc.rst. 992 993 coresight_cpu_debug.enable 994 [ARM,ARM64] 995 Format: <bool> 996 Enable/disable the CPU sampling based debugging. 997 0: default value, disable debugging 998 1: enable debugging at boot time 999 1000 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver 1001 Format: 1002 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>] 1003 1004 cpuidle.off=1 [CPU_IDLE] 1005 disable the cpuidle sub-system 1006 1007 cpuidle.governor= 1008 [CPU_IDLE] Name of the cpuidle governor to use. 1009 1010 cpufreq.off=1 [CPU_FREQ] 1011 disable the cpufreq sub-system 1012 1013 cpufreq.default_governor= 1014 [CPU_FREQ] Name of the default cpufreq governor or 1015 policy to use. This governor must be registered in the 1016 kernel before the cpufreq driver probes. 1017 1018 cpu_init_udelay=N 1019 [X86,EARLY] Delay for N microsec between assert and de-assert 1020 of APIC INIT to start processors. This delay occurs 1021 on every CPU online, such as boot, and resume from suspend. 1022 Default: 10000 1023 1024 cpuhp.parallel= 1025 [SMP] Enable/disable parallel bringup of secondary CPUs 1026 Format: <bool> 1027 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise 1028 the parameter has no effect. 1029 1030 crash_kexec_post_notifiers 1031 Only jump to kdump kernel after running the panic 1032 notifiers and dumping kmsg. This option increases 1033 the risks of a kdump failure, since some panic 1034 notifiers can make the crashed kernel more unstable. 1035 In configurations where kdump may not be reliable, 1036 running the panic notifiers could allow collecting 1037 more data on dmesg, like stack traces from other CPUS 1038 or extra data dumped by panic_print. Note that some 1039 configurations enable this option unconditionally, 1040 like Hyper-V, PowerPC (fadump) and AMD SEV-SNP. 1041 1042 crashkernel=size[KMG][@offset[KMG]] 1043 [KNL,EARLY] Using kexec, Linux can switch to a 'crash kernel' 1044 upon panic. This parameter reserves the physical 1045 memory region [offset, offset + size] for that kernel 1046 image. If '@offset' is omitted, then a suitable offset 1047 is selected automatically. 1048 [KNL, X86-64, ARM64, RISCV, LoongArch] Select a region 1049 under 4G first, and fall back to reserve region above 1050 4G when '@offset' hasn't been specified. 1051 See Documentation/admin-guide/kdump/kdump.rst for further details. 1052 1053 crashkernel=range1:size1[,range2:size2,...][@offset] 1054 [KNL] Same as above, but depends on the memory 1055 in the running system. The syntax of range is 1056 start-[end] where start and end are both 1057 a memory unit (amount[KMG]). See also 1058 Documentation/admin-guide/kdump/kdump.rst for an example. 1059 1060 crashkernel=size[KMG],high 1061 [KNL, X86-64, ARM64, RISCV, LoongArch] range could be 1062 above 4G. 1063 Allow kernel to allocate physical memory region from top, 1064 so could be above 4G if system have more than 4G ram 1065 installed. Otherwise memory region will be allocated 1066 below 4G, if available. 1067 It will be ignored if crashkernel=X is specified. 1068 crashkernel=size[KMG],low 1069 [KNL, X86-64, ARM64, RISCV, LoongArch] range under 4G. 1070 When crashkernel=X,high is passed, kernel could allocate 1071 physical memory region above 4G, that cause second kernel 1072 crash on system that require some amount of low memory, 1073 e.g. swiotlb requires at least 64M+32K low memory, also 1074 enough extra low memory is needed to make sure DMA buffers 1075 for 32-bit devices won't run out. Kernel would try to allocate 1076 default size of memory below 4G automatically. The default 1077 size is platform dependent. 1078 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB) 1079 --> arm64: 128MiB 1080 --> riscv: 128MiB 1081 --> loongarch: 128MiB 1082 This one lets the user specify own low range under 4G 1083 for second kernel instead. 1084 0: to disable low allocation. 1085 It will be ignored when crashkernel=X,high is not used 1086 or memory reserved is below 4G. 1087 crashkernel=size[KMG],cma 1088 [KNL, X86, ppc] Reserve additional crash kernel memory from 1089 CMA. This reservation is usable by the first system's 1090 userspace memory and kernel movable allocations (memory 1091 balloon, zswap). Pages allocated from this memory range 1092 will not be included in the vmcore so this should not 1093 be used if dumping of userspace memory is intended and 1094 it has to be expected that some movable kernel pages 1095 may be missing from the dump. 1096 1097 A standard crashkernel reservation, as described above, 1098 is still needed to hold the crash kernel and initrd. 1099 1100 This option increases the risk of a kdump failure: DMA 1101 transfers configured by the first kernel may end up 1102 corrupting the second kernel's memory. 1103 1104 This reservation method is intended for systems that 1105 can't afford to sacrifice enough memory for standard 1106 crashkernel reservation and where less reliable and 1107 possibly incomplete kdump is preferable to no kdump at 1108 all. 1109 1110 cryptomgr.notests 1111 [KNL] Disable crypto self-tests 1112 1113 cs89x0_dma= [HW,NET] 1114 Format: <dma> 1115 1116 cs89x0_media= [HW,NET] 1117 Format: { rj45 | aui | bnc } 1118 1119 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU 1120 function call handling. When switched on, 1121 additional debug data is printed to the console 1122 in case a hanging CPU is detected, and that 1123 CPU is pinged again in order to try to resolve 1124 the hang situation. The default value of this 1125 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT 1126 Kconfig option. 1127 1128 dasd= [HW,NET] 1129 See header of drivers/s390/block/dasd_devmap.c. 1130 1131 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port 1132 (one device per port) 1133 Format: <port#>,<type> 1134 See also Documentation/input/devices/joystick-parport.rst 1135 1136 debug [KNL,EARLY] Enable kernel debugging (events log level). 1137 1138 debug_boot_weak_hash 1139 [KNL,EARLY] Enable printing [hashed] pointers early in the 1140 boot sequence. If enabled, we use a weak hash instead 1141 of siphash to hash pointers. Use this option if you are 1142 seeing instances of '(___ptrval___)') and need to see a 1143 value (hashed pointer) instead. Cryptographically 1144 insecure, please do not use on production kernels. 1145 1146 debug_locks_verbose= 1147 [KNL] verbose locking self-tests 1148 Format: <int> 1149 Print debugging info while doing the locking API 1150 self-tests. 1151 Bitmask for the various LOCKTYPE_ tests. Defaults to 0 1152 (no extra messages), setting it to -1 (all bits set) 1153 will print _a_lot_ more information - normally only 1154 useful to lockdep developers. 1155 1156 debug_objects [KNL,EARLY] Enable object debugging 1157 1158 debug_guardpage_minorder= 1159 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this 1160 parameter allows control of the order of pages that will 1161 be intentionally kept free (and hence protected) by the 1162 buddy allocator. Bigger value increase the probability 1163 of catching random memory corruption, but reduce the 1164 amount of memory for normal system use. The maximum 1165 possible value is MAX_PAGE_ORDER/2. Setting this 1166 parameter to 1 or 2 should be enough to identify most 1167 random memory corruption problems caused by bugs in 1168 kernel or driver code when a CPU writes to (or reads 1169 from) a random memory location. Note that there exists 1170 a class of memory corruptions problems caused by buggy 1171 H/W or F/W or by drivers badly programming DMA 1172 (basically when memory is written at bus level and the 1173 CPU MMU is bypassed) which are not detectable by 1174 CONFIG_DEBUG_PAGEALLOC, hence this option will not 1175 help tracking down these problems. 1176 1177 debug_pagealloc= 1178 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this parameter 1179 enables the feature at boot time. By default, it is 1180 disabled and the system will work mostly the same as a 1181 kernel built without CONFIG_DEBUG_PAGEALLOC. 1182 Note: to get most of debug_pagealloc error reports, it's 1183 useful to also enable the page_owner functionality. 1184 on: enable the feature 1185 1186 debugfs= [KNL,EARLY] This parameter enables what is exposed to 1187 userspace and debugfs internal clients. 1188 Format: { on, off } 1189 on: All functions are enabled. 1190 off: Filesystem is not registered and clients 1191 get a -EPERM as result when trying to register files 1192 or directories within debugfs. 1193 This is equivalent of the runtime functionality if 1194 debugfs was not enabled in the kernel at all. 1195 Default value is set in build-time with a kernel configuration. 1196 1197 debugpat [X86] Enable PAT debugging 1198 1199 default_hugepagesz= 1200 [HW] The size of the default HugeTLB page. This is 1201 the size represented by the legacy /proc/ hugepages 1202 APIs. In addition, this is the default hugetlb size 1203 used for shmget(), mmap() and mounting hugetlbfs 1204 filesystems. If not specified, defaults to the 1205 architecture's default huge page size. Huge page 1206 sizes are architecture dependent. See also 1207 Documentation/admin-guide/mm/hugetlbpage.rst. 1208 Format: size[KMG] 1209 1210 deferred_probe_timeout= 1211 [KNL] Debugging option to set a timeout in seconds for 1212 deferred probe to give up waiting on dependencies to 1213 probe. Only specific dependencies (subsystems or 1214 drivers) that have opted in will be ignored. A timeout 1215 of 0 will timeout at the end of initcalls; a negative 1216 value is treated as an infinite timeout value. If the 1217 timeout hasn't expired, it'll be restarted by each 1218 successful driver registration. This option will also 1219 dump out devices still on the deferred probe list after 1220 retrying. 1221 1222 delayacct [KNL] Enable per-task delay accounting 1223 1224 dell_smm_hwmon.ignore_dmi= 1225 [HW] Continue probing hardware even if DMI data 1226 indicates that the driver is running on unsupported 1227 hardware. 1228 1229 dell_smm_hwmon.force= 1230 [HW] Activate driver even if SMM BIOS signature does 1231 not match list of supported models and enable otherwise 1232 blacklisted features. 1233 1234 dell_smm_hwmon.power_status= 1235 [HW] Report power status in /proc/i8k 1236 (disabled by default). 1237 1238 dell_smm_hwmon.restricted= 1239 [HW] Allow controlling fans only if SYS_ADMIN 1240 capability is set. 1241 1242 dell_smm_hwmon.fan_mult= 1243 [HW] Factor to multiply fan speed with. 1244 1245 dell_smm_hwmon.fan_max= 1246 [HW] Maximum configurable fan speed. 1247 1248 dfltcc= [HW,S390] 1249 Format: { on | off | def_only | inf_only | always } 1250 on: s390 zlib hardware support for compression on 1251 level 1 and decompression (default) 1252 off: No s390 zlib hardware support 1253 def_only: s390 zlib hardware support for deflate 1254 only (compression on level 1) 1255 inf_only: s390 zlib hardware support for inflate 1256 only (decompression) 1257 always: Same as 'on' but ignores the selected compression 1258 level always using hardware support (used for debugging) 1259 1260 dhash_entries= [KNL] 1261 Set number of hash buckets for dentry cache. 1262 1263 disable_1tb_segments [PPC,EARLY] 1264 Disables the use of 1TB hash page table segments. This 1265 causes the kernel to fall back to 256MB segments which 1266 can be useful when debugging issues that require an SLB 1267 miss to occur. 1268 1269 disable= [IPV6] 1270 See Documentation/networking/ipv6.rst. 1271 1272 disable_radix [PPC,EARLY] 1273 Disable RADIX MMU mode on POWER9 1274 1275 disable_tlbie [PPC] 1276 Disable TLBIE instruction. Currently does not work 1277 with KVM, with HASH MMU, or with coherent accelerators. 1278 1279 disable_ddw [PPC/PSERIES,EARLY] 1280 Disable Dynamic DMA Window support. Use this 1281 to workaround buggy firmware. 1282 1283 disable_ipv6= [IPV6] 1284 See Documentation/networking/ipv6.rst. 1285 1286 disable_mtrr_cleanup [X86,EARLY] 1287 The kernel tries to adjust MTRR layout from continuous 1288 to discrete, to make X server driver able to add WB 1289 entry later. This parameter disables that. 1290 1291 disable_mtrr_trim [X86, Intel and AMD only,EARLY] 1292 By default the kernel will trim any uncacheable 1293 memory out of your available memory pool based on 1294 MTRR settings. This parameter disables that behavior, 1295 possibly causing your machine to run very slowly. 1296 1297 disable_timer_pin_1 [X86,EARLY] 1298 Disable PIN 1 of APIC timer 1299 Can be useful to work around chipset bugs. 1300 1301 dis_ucode_ldr [X86] Disable the microcode loader. 1302 1303 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support, 1304 this option disables the debugging code at boot. 1305 1306 dma_debug_entries=<number> 1307 This option allows to tune the number of preallocated 1308 entries for DMA-API debugging code. One entry is 1309 required per DMA-API allocation. Use this if the 1310 DMA-API debugging code disables itself because the 1311 architectural default is too low. 1312 1313 dma_debug_driver=<driver_name> 1314 With this option the DMA-API debugging driver 1315 filter feature can be enabled at boot time. Just 1316 pass the driver to filter for as the parameter. 1317 The filter can be disabled or changed to another 1318 driver later using sysfs. 1319 1320 reg_file_data_sampling= 1321 [X86] Controls mitigation for Register File Data 1322 Sampling (RFDS) vulnerability. RFDS is a CPU 1323 vulnerability which may allow userspace to infer 1324 kernel data values previously stored in floating point 1325 registers, vector registers, or integer registers. 1326 RFDS only affects Intel Atom processors. 1327 1328 on: Turns ON the mitigation. 1329 off: Turns OFF the mitigation. 1330 1331 This parameter overrides the compile time default set 1332 by CONFIG_MITIGATION_RFDS. Mitigation cannot be 1333 disabled when other VERW based mitigations (like MDS) 1334 are enabled. In order to disable RFDS mitigation all 1335 VERW based mitigations need to be disabled. 1336 1337 For details see: 1338 Documentation/admin-guide/hw-vuln/reg-file-data-sampling.rst 1339 1340 dm_verity.keyring_unsealed= 1341 [KNL] When set to 1, leave the dm-verity keyring 1342 unsealed after initialization so userspace can 1343 provision keys. Once the keyring is restricted 1344 it becomes active and is searched during signature 1345 verification. 1346 1347 driver_async_probe= [KNL] 1348 List of driver names to be probed asynchronously. * 1349 matches with all driver names. If * is specified, the 1350 rest of the listed driver names are those that will NOT 1351 match the *. 1352 Format: <driver_name1>,<driver_name2>... 1353 1354 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>] 1355 Broken monitors, graphic adapters, KVMs and EDIDless 1356 panels may send no or incorrect EDID data sets. 1357 This parameter allows to specify an EDID data sets 1358 in the /lib/firmware directory that are used instead. 1359 An EDID data set will only be used for a particular 1360 connector, if its name and a colon are prepended to 1361 the EDID name. Each connector may use a unique EDID 1362 data set by separating the files with a comma. An EDID 1363 data set with no connector name will be used for 1364 any connectors not explicitly specified. 1365 1366 dscc4.setup= [NET] 1367 1368 dt_cpu_ftrs= [PPC,EARLY] 1369 Format: {"off" | "known"} 1370 Control how the dt_cpu_ftrs device-tree binding is 1371 used for CPU feature discovery and setup (if it 1372 exists). 1373 off: Do not use it, fall back to legacy cpu table. 1374 known: Do not pass through unknown features to guests 1375 or userspace, only those that the kernel is aware of. 1376 1377 dump_apple_properties [X86] 1378 Dump name and content of EFI device properties on 1379 x86 Macs. Useful for driver authors to determine 1380 what data is available or for reverse-engineering. 1381 1382 dyndbg[="val"] [KNL,DYNAMIC_DEBUG] 1383 <module>.dyndbg[="val"] 1384 Enable debug messages at boot time. See 1385 Documentation/admin-guide/dynamic-debug-howto.rst 1386 for details. 1387 1388 early_ioremap_debug [KNL,EARLY] 1389 Enable debug messages in early_ioremap support. This 1390 is useful for tracking down temporary early mappings 1391 which are not unmapped. 1392 1393 earlycon= [KNL,EARLY] Output early console device and options. 1394 1395 When used with no options, the early console is 1396 determined by stdout-path property in device tree's 1397 chosen node or the ACPI SPCR table if supported by 1398 the platform. 1399 1400 cdns,<addr>[,options] 1401 Start an early, polled-mode console on a Cadence 1402 (xuartps) serial port at the specified address. Only 1403 supported option is baud rate. If baud rate is not 1404 specified, the serial port must already be setup and 1405 configured. 1406 1407 uart[8250],io,<addr>[,options[,uartclk]] 1408 uart[8250],mmio,<addr>[,options[,uartclk]] 1409 uart[8250],mmio32,<addr>[,options[,uartclk]] 1410 uart[8250],mmio32be,<addr>[,options[,uartclk]] 1411 uart[8250],0x<addr>[,options] 1412 Start an early, polled-mode console on the 8250/16550 1413 UART at the specified I/O port or MMIO address. 1414 MMIO inter-register address stride is either 8-bit 1415 (mmio) or 32-bit (mmio32 or mmio32be). 1416 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed 1417 to be equivalent to 'mmio'. 'options' are specified 1418 in the same format described for "console=ttyS<n>"; if 1419 unspecified, the h/w is not initialized. 'uartclk' is 1420 the uart clock frequency; if unspecified, it is set 1421 to 'BASE_BAUD' * 16. 1422 1423 pl011,<addr> 1424 pl011,mmio32,<addr> 1425 Start an early, polled-mode console on a pl011 serial 1426 port at the specified address. The pl011 serial port 1427 must already be setup and configured. Options are not 1428 yet supported. If 'mmio32' is specified, then only 1429 the driver will use only 32-bit accessors to read/write 1430 the device registers. 1431 1432 liteuart,<addr> 1433 Start an early console on a litex serial port at the 1434 specified address. The serial port must already be 1435 setup and configured. Options are not yet supported. 1436 1437 meson,<addr> 1438 Start an early, polled-mode console on a meson serial 1439 port at the specified address. The serial port must 1440 already be setup and configured. Options are not yet 1441 supported. 1442 1443 msm_serial,<addr> 1444 Start an early, polled-mode console on an msm serial 1445 port at the specified address. The serial port 1446 must already be setup and configured. Options are not 1447 yet supported. 1448 1449 msm_serial_dm,<addr> 1450 Start an early, polled-mode console on an msm serial 1451 dm port at the specified address. The serial port 1452 must already be setup and configured. Options are not 1453 yet supported. 1454 1455 owl,<addr> 1456 Start an early, polled-mode console on a serial port 1457 of an Actions Semi SoC, such as S500 or S900, at the 1458 specified address. The serial port must already be 1459 setup and configured. Options are not yet supported. 1460 1461 rda,<addr> 1462 Start an early, polled-mode console on a serial port 1463 of an RDA Micro SoC, such as RDA8810PL, at the 1464 specified address. The serial port must already be 1465 setup and configured. Options are not yet supported. 1466 1467 sbi 1468 Use RISC-V SBI (Supervisor Binary Interface) for early 1469 console. 1470 1471 smh Use ARM semihosting calls for early console. 1472 1473 s3c2410,<addr> 1474 s3c2412,<addr> 1475 s3c2440,<addr> 1476 s3c6400,<addr> 1477 s5pv210,<addr> 1478 exynos4210,<addr> 1479 Use early console provided by serial driver available 1480 on Samsung SoCs, requires selecting proper type and 1481 a correct base address of the selected UART port. The 1482 serial port must already be setup and configured. 1483 Options are not yet supported. 1484 1485 lantiq,<addr> 1486 Start an early, polled-mode console on a lantiq serial 1487 (lqasc) port at the specified address. The serial port 1488 must already be setup and configured. Options are not 1489 yet supported. 1490 1491 lpuart,<addr> 1492 lpuart32,<addr> 1493 Use early console provided by Freescale LP UART driver 1494 found on Freescale Vybrid and QorIQ LS1021A processors. 1495 A valid base address must be provided, and the serial 1496 port must already be setup and configured. 1497 1498 ec_imx21,<addr> 1499 ec_imx6q,<addr> 1500 Start an early, polled-mode, output-only console on the 1501 Freescale i.MX UART at the specified address. The UART 1502 must already be setup and configured. 1503 1504 ar3700_uart,<addr> 1505 Start an early, polled-mode console on the 1506 Armada 3700 serial port at the specified 1507 address. The serial port must already be setup 1508 and configured. Options are not yet supported. 1509 1510 qcom_geni,<addr> 1511 Start an early, polled-mode console on a Qualcomm 1512 Generic Interface (GENI) based serial port at the 1513 specified address. The serial port must already be 1514 setup and configured. Options are not yet supported. 1515 1516 efifb,[options] 1517 Start an early, unaccelerated console on the EFI 1518 memory mapped framebuffer (if available). On cache 1519 coherent non-x86 systems that use system memory for 1520 the framebuffer, pass the 'ram' option so that it is 1521 mapped with the correct attributes. 1522 1523 linflex,<addr> 1524 Use early console provided by Freescale LINFlexD UART 1525 serial driver for NXP S32V234 SoCs. A valid base 1526 address must be provided, and the serial port must 1527 already be setup and configured. 1528 1529 earlyprintk= [X86,SH,ARM,M68k,S390,UM,EARLY] 1530 earlyprintk=vga 1531 earlyprintk=sclp 1532 earlyprintk=xen 1533 earlyprintk=serial[,ttySn[,baudrate]] 1534 earlyprintk=serial[,0x...[,baudrate]] 1535 earlyprintk=ttySn[,baudrate] 1536 earlyprintk=dbgp[debugController#] 1537 earlyprintk=mmio32,membase[,{nocfg|baudrate}] 1538 earlyprintk=pciserial[,force],bus:device.function[,{nocfg|baudrate}] 1539 earlyprintk=xdbc[xhciController#] 1540 earlyprintk=bios 1541 1542 earlyprintk is useful when the kernel crashes before 1543 the normal console is initialized. It is not enabled by 1544 default because it has some cosmetic problems. 1545 1546 Use "nocfg" to skip UART configuration, assume 1547 BIOS/firmware has configured UART correctly. 1548 1549 Append ",keep" to not disable it when the real console 1550 takes over. 1551 1552 Only one of vga, serial, or usb debug port can 1553 be used at a time. 1554 1555 Currently only ttyS0 and ttyS1 may be specified by 1556 name. Other I/O ports may be explicitly specified 1557 on some architectures (x86 and arm at least) by 1558 replacing ttySn with an I/O port address, like this: 1559 earlyprintk=serial,0x1008,115200 1560 You can find the port for a given device in 1561 /proc/tty/driver/serial: 1562 2: uart:ST16650V2 port:00001008 irq:18 ... 1563 1564 Interaction with the standard serial driver is not 1565 very good. 1566 1567 The VGA output is eventually overwritten by 1568 the real console. 1569 1570 The xen option can only be used in Xen domains. 1571 1572 The sclp output can only be used on s390. 1573 1574 The bios output can only be used on SuperH. 1575 1576 The optional "force" to "pciserial" enables use of a 1577 PCI device even when its classcode is not of the 1578 UART class. 1579 1580 edac_report= [HW,EDAC] Control how to report EDAC event 1581 Format: {"on" | "off" | "force"} 1582 on: enable EDAC to report H/W event. May be overridden 1583 by other higher priority error reporting module. 1584 off: disable H/W event reporting through EDAC. 1585 force: enforce the use of EDAC to report H/W event. 1586 default: on. 1587 1588 edd= [EDD] 1589 Format: {"off" | "on" | "skip[mbr]"} 1590 1591 efi= [EFI,EARLY] 1592 Format: { "debug", "disable_early_pci_dma", 1593 "nochunk", "noruntime", "nosoftreserve", 1594 "novamap", "no_disable_early_pci_dma" } 1595 debug: enable misc debug output. 1596 disable_early_pci_dma: disable the busmaster bit on all 1597 PCI bridges while in the EFI boot stub. 1598 nochunk: disable reading files in "chunks" in the EFI 1599 boot stub, as chunking can cause problems with some 1600 firmware implementations. 1601 noruntime : disable EFI runtime services support 1602 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose) 1603 attribute may cause the kernel to reserve the 1604 memory range for a memory mapping driver to 1605 claim. Specify efi=nosoftreserve to disable this 1606 reservation and treat the memory by its base type 1607 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM"). 1608 novamap: do not call SetVirtualAddressMap(). 1609 no_disable_early_pci_dma: Leave the busmaster bit set 1610 on all PCI bridges while in the EFI boot stub 1611 1612 efi_no_storage_paranoia [EFI,X86,EARLY] 1613 Using this parameter you can use more than 50% of 1614 your efi variable storage. Use this parameter only if 1615 you are really sure that your UEFI does sane gc and 1616 fulfills the spec otherwise your board may brick. 1617 1618 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT 1619 that is to be dynamically loaded by Linux. If there are 1620 multiple variables with the same name but with different 1621 vendor GUIDs, all of them will be loaded. See 1622 Documentation/admin-guide/acpi/ssdt-overlays.rst for details. 1623 1624 1625 eisa_irq_edge= [PARISC,HW] 1626 See header of drivers/parisc/eisa.c. 1627 1628 ekgdboc= [X86,KGDB,EARLY] Allow early kernel console debugging 1629 Format: ekgdboc=kbd 1630 1631 This is designed to be used in conjunction with 1632 the boot argument: earlyprintk=vga 1633 1634 This parameter works in place of the kgdboc parameter 1635 but can only be used if the backing tty is available 1636 very early in the boot process. For early debugging 1637 via a serial port see kgdboc_earlycon instead. 1638 1639 elfcorehdr=[size[KMG]@]offset[KMG] [PPC,SH,X86,S390,EARLY] 1640 Specifies physical address of start of kernel core 1641 image elf header and optionally the size. Generally 1642 kexec loader will pass this option to capture kernel. 1643 See Documentation/admin-guide/kdump/kdump.rst for details. 1644 1645 enable_mtrr_cleanup [X86,EARLY] 1646 The kernel tries to adjust MTRR layout from continuous 1647 to discrete, to make X server driver able to add WB 1648 entry later. This parameter enables that. 1649 1650 enable_timer_pin_1 [X86] 1651 Enable PIN 1 of APIC timer 1652 Can be useful to work around chipset bugs 1653 (in particular on some ATI chipsets). 1654 The kernel tries to set a reasonable default. 1655 1656 enforcing= [SELINUX] Set initial enforcing status. 1657 Format: {"0" | "1"} 1658 See security/selinux/Kconfig help text. 1659 0 -- permissive (log only, no denials). 1660 1 -- enforcing (deny and log). 1661 Default value is 0. 1662 Value can be changed at runtime via 1663 /sys/fs/selinux/enforce. 1664 1665 erst_disable [ACPI] 1666 Disable Error Record Serialization Table (ERST) 1667 support. 1668 1669 ether= [HW,NET] Ethernet cards parameters 1670 This option is obsoleted by the "netdev=" option, which 1671 has equivalent usage. See its documentation for details. 1672 1673 evm= [EVM] 1674 Format: { "fix" } 1675 Permit 'security.evm' to be updated regardless of 1676 current integrity status. 1677 1678 early_page_ext [KNL,EARLY] Enforces page_ext initialization to earlier 1679 stages so cover more early boot allocations. 1680 Please note that as side effect some optimizations 1681 might be disabled to achieve that (e.g. parallelized 1682 memory initialization is disabled) so the boot process 1683 might take longer, especially on systems with a lot of 1684 memory. Available with CONFIG_PAGE_EXTENSION=y. 1685 1686 failslab= 1687 fail_usercopy= 1688 fail_page_alloc= 1689 fail_skb_realloc= 1690 fail_make_request=[KNL] 1691 General fault injection mechanism. 1692 Format: <interval>,<probability>,<space>,<times> 1693 See also Documentation/fault-injection/. 1694 1695 fb_tunnels= [NET] 1696 Format: { initns | none } 1697 See Documentation/admin-guide/sysctl/net.rst for 1698 fb_tunnels_only_for_init_ns 1699 1700 floppy= [HW] 1701 See Documentation/admin-guide/blockdev/floppy.rst. 1702 1703 forcepae [X86-32] 1704 Forcefully enable Physical Address Extension (PAE). 1705 Many Pentium M systems disable PAE but may have a 1706 functionally usable PAE implementation. 1707 Warning: use of this parameter will taint the kernel 1708 and may cause unknown problems. 1709 1710 fred= [X86-64] 1711 Enable/disable Flexible Return and Event Delivery. 1712 Format: { on | off } 1713 on: enable FRED when it's present, the default setting. 1714 off: disable FRED. 1715 1716 ftrace=[tracer] 1717 [FTRACE] will set and start the specified tracer 1718 as early as possible in order to facilitate early 1719 boot debugging. 1720 1721 ftrace_boot_snapshot 1722 [FTRACE] On boot up, a snapshot will be taken of the 1723 ftrace ring buffer that can be read at: 1724 /sys/kernel/tracing/snapshot. 1725 This is useful if you need tracing information from kernel 1726 boot up that is likely to be overridden by user space 1727 start up functionality. 1728 1729 Optionally, the snapshot can also be defined for a tracing 1730 instance that was created by the trace_instance= command 1731 line parameter. 1732 1733 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo 1734 1735 The above will cause the "foo" tracing instance to trigger 1736 a snapshot at the end of boot up. 1737 1738 ftrace_dump_on_oops[=2(orig_cpu) | =<instance>][,<instance> | 1739 ,<instance>=2(orig_cpu)] 1740 [FTRACE] will dump the trace buffers on oops. 1741 If no parameter is passed, ftrace will dump global 1742 buffers of all CPUs, if you pass 2 or orig_cpu, it 1743 will dump only the buffer of the CPU that triggered 1744 the oops, or the specific instance will be dumped if 1745 its name is passed. Multiple instance dump is also 1746 supported, and instances are separated by commas. Each 1747 instance supports only dump on CPU that triggered the 1748 oops by passing 2 or orig_cpu to it. 1749 1750 ftrace_dump_on_oops=foo=orig_cpu 1751 1752 The above will dump only the buffer of "foo" instance 1753 on CPU that triggered the oops. 1754 1755 ftrace_dump_on_oops,foo,bar=orig_cpu 1756 1757 The above will dump global buffer on all CPUs, the 1758 buffer of "foo" instance on all CPUs and the buffer 1759 of "bar" instance on CPU that triggered the oops. 1760 1761 ftrace_filter=[function-list] 1762 [FTRACE] Limit the functions traced by the function 1763 tracer at boot up. function-list is a comma-separated 1764 list of functions. This list can be changed at run 1765 time by the set_ftrace_filter file in the debugfs 1766 tracing directory. 1767 1768 ftrace_notrace=[function-list] 1769 [FTRACE] Do not trace the functions specified in 1770 function-list. This list can be changed at run time 1771 by the set_ftrace_notrace file in the debugfs 1772 tracing directory. 1773 1774 ftrace_graph_filter=[function-list] 1775 [FTRACE] Limit the top level callers functions traced 1776 by the function graph tracer at boot up. 1777 function-list is a comma-separated list of functions 1778 that can be changed at run time by the 1779 set_graph_function file in the debugfs tracing directory. 1780 1781 ftrace_graph_notrace=[function-list] 1782 [FTRACE] Do not trace from the functions specified in 1783 function-list. This list is a comma-separated list of 1784 functions that can be changed at run time by the 1785 set_graph_notrace file in the debugfs tracing directory. 1786 1787 ftrace_graph_max_depth=<uint> 1788 [FTRACE] Used with the function graph tracer. This is 1789 the max depth it will trace into a function. This value 1790 can be changed at run time by the max_graph_depth file 1791 in the tracefs tracing directory. default: 0 (no limit) 1792 1793 fw_devlink= [KNL,EARLY] Create device links between consumer and supplier 1794 devices by scanning the firmware to infer the 1795 consumer/supplier relationships. This feature is 1796 especially useful when drivers are loaded as modules as 1797 it ensures proper ordering of tasks like device probing 1798 (suppliers first, then consumers), supplier boot state 1799 clean up (only after all consumers have probed), 1800 suspend/resume & runtime PM (consumers first, then 1801 suppliers). 1802 Format: { off | permissive | on | rpm } 1803 off -- Don't create device links from firmware info. 1804 permissive -- Create device links from firmware info 1805 but use it only for ordering boot state clean 1806 up (sync_state() calls). 1807 on -- Create device links from firmware info and use it 1808 to enforce probe and suspend/resume ordering. 1809 rpm -- Like "on", but also use to order runtime PM. 1810 1811 fw_devlink.strict=<bool> 1812 [KNL,EARLY] Treat all inferred dependencies as mandatory 1813 dependencies. This only applies for fw_devlink=on|rpm. 1814 Format: <bool> 1815 1816 fw_devlink.sync_state = 1817 [KNL,EARLY] When all devices that could probe have finished 1818 probing, this parameter controls what to do with 1819 devices that haven't yet received their sync_state() 1820 calls. 1821 Format: { strict | timeout } 1822 strict -- Default. Continue waiting on consumers to 1823 probe successfully. 1824 timeout -- Give up waiting on consumers and call 1825 sync_state() on any devices that haven't yet 1826 received their sync_state() calls after 1827 deferred_probe_timeout has expired or by 1828 late_initcall() if !CONFIG_MODULES. 1829 1830 gamecon.map[2|3]= 1831 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad 1832 support via parallel port (up to 5 devices per port) 1833 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5> 1834 See also Documentation/input/devices/joystick-parport.rst 1835 1836 gamma= [HW,DRM] 1837 1838 gart_fix_e820= [X86-64,EARLY] disable the fix e820 for K8 GART 1839 Format: off | on 1840 default: on 1841 1842 gather_data_sampling= 1843 [X86,INTEL,EARLY] Control the Gather Data Sampling (GDS) 1844 mitigation. 1845 1846 Gather Data Sampling is a hardware vulnerability which 1847 allows unprivileged speculative access to data which was 1848 previously stored in vector registers. 1849 1850 This issue is mitigated by default in updated microcode. 1851 The mitigation may have a performance impact but can be 1852 disabled. On systems without the microcode mitigation 1853 disabling AVX serves as a mitigation. 1854 1855 force: Disable AVX to mitigate systems without 1856 microcode mitigation. No effect if the microcode 1857 mitigation is present. Known to cause crashes in 1858 userspace with buggy AVX enumeration. 1859 1860 off: Disable GDS mitigation. 1861 1862 gbpages [X86] Use GB pages for kernel direct mappings. 1863 1864 gcov_persist= [GCOV] When non-zero (default), profiling data for 1865 kernel modules is saved and remains accessible via 1866 debugfs, even when the module is unloaded/reloaded. 1867 When zero, profiling data is discarded and associated 1868 debugfs files are removed at module unload time. 1869 1870 goldfish [X86] Enable the goldfish android emulator platform. 1871 Don't use this when you are not running on the 1872 android emulator 1873 1874 gpio-mockup.gpio_mockup_ranges 1875 [HW] Sets the ranges of gpiochip of for this device. 1876 Format: <start1>,<end1>,<start2>,<end2>... 1877 gpio-mockup.gpio_mockup_named_lines 1878 [HW] Let the driver know GPIO lines should be named. 1879 1880 gpt [EFI] Forces disk with valid GPT signature but 1881 invalid Protective MBR to be treated as GPT. If the 1882 primary GPT is corrupted, it enables the backup/alternate 1883 GPT to be used instead. 1884 1885 grcan.enable0= [HW] Configuration of physical interface 0. Determines 1886 the "Enable 0" bit of the configuration register. 1887 Format: 0 | 1 1888 Default: 0 1889 grcan.enable1= [HW] Configuration of physical interface 1. Determines 1890 the "Enable 0" bit of the configuration register. 1891 Format: 0 | 1 1892 Default: 0 1893 grcan.select= [HW] Select which physical interface to use. 1894 Format: 0 | 1 1895 Default: 0 1896 grcan.txsize= [HW] Sets the size of the tx buffer. 1897 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0. 1898 Default: 1024 1899 grcan.rxsize= [HW] Sets the size of the rx buffer. 1900 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0. 1901 Default: 1024 1902 1903 hardened_usercopy= 1904 [KNL] Under CONFIG_HARDENED_USERCOPY, whether 1905 hardening is enabled for this boot. Hardened 1906 usercopy checking is used to protect the kernel 1907 from reading or writing beyond known memory 1908 allocation boundaries as a proactive defense 1909 against bounds-checking flaws in the kernel's 1910 copy_to_user()/copy_from_user() interface. 1911 The default is determined by 1912 CONFIG_HARDENED_USERCOPY_DEFAULT_ON. 1913 on Perform hardened usercopy checks. 1914 off Disable hardened usercopy checks. 1915 1916 hardlockup_all_cpu_backtrace= 1917 [KNL] Should the hard-lockup detector generate 1918 backtraces on all cpus. 1919 Format: 0 | 1 1920 1921 hash_pointers= 1922 [KNL,EARLY] 1923 By default, when pointers are printed to the console 1924 or buffers via the %p format string, that pointer is 1925 "hashed", i.e. obscured by hashing the pointer value. 1926 This is a security feature that hides actual kernel 1927 addresses from unprivileged users, but it also makes 1928 debugging the kernel more difficult since unequal 1929 pointers can no longer be compared. The choices are: 1930 Format: { auto | always | never } 1931 Default: auto 1932 1933 auto - Hash pointers unless slab_debug is enabled. 1934 always - Always hash pointers (even if slab_debug is 1935 enabled). 1936 never - Never hash pointers. This option should only 1937 be specified when debugging the kernel. Do 1938 not use on production kernels. The boot 1939 param "no_hash_pointers" is an alias for 1940 this mode. 1941 1942 For controlling hashing dynamically at runtime, 1943 use the "kernel.kptr_restrict" sysctl instead. 1944 1945 hashdist= [KNL,NUMA] Large hashes allocated during boot 1946 are distributed across NUMA nodes. Defaults on 1947 for 64-bit NUMA, off otherwise. 1948 Format: 0 | 1 (for off | on) 1949 1950 hd= [EIDE] (E)IDE hard drive subsystem geometry 1951 Format: <cyl>,<head>,<sect> 1952 1953 hest_disable [ACPI] 1954 Disable Hardware Error Source Table (HEST) support; 1955 corresponding firmware-first mode error processing 1956 logic will be disabled. 1957 1958 hibernate= [HIBERNATION] 1959 noresume Don't check if there's a hibernation image 1960 present during boot. 1961 nocompress Don't compress/decompress hibernation images. 1962 no Disable hibernation and resume. 1963 protect_image Turn on image protection during restoration 1964 (that will set all pages holding image data 1965 during restoration read-only). 1966 1967 hibernate.compressor= [HIBERNATION] Compression algorithm to be 1968 used with hibernation. 1969 Format: { lzo | lz4 } 1970 Default: lzo 1971 1972 lzo: Select LZO compression algorithm to 1973 compress/decompress hibernation image. 1974 1975 lz4: Select LZ4 compression algorithm to 1976 compress/decompress hibernation image. 1977 1978 hibernate.pm_test_delay= 1979 [HIBERNATION] 1980 Sets the number of seconds to remain in a hibernation test 1981 mode before resuming the system (see 1982 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 1983 is set. Default value is 5. 1984 1985 hibernate_compression_threads= 1986 [HIBERNATION] 1987 Set the number of threads used for compressing or decompressing 1988 hibernation images. 1989 1990 Format: <integer> 1991 Default: 3 1992 Minimum: 1 1993 Example: hibernate_compression_threads=4 1994 1995 highmem=nn[KMG] [KNL,BOOT,EARLY] forces the highmem zone to have an exact 1996 size of <nn>. This works even on boxes that have no 1997 highmem otherwise. This also works to reduce highmem 1998 size on bigger boxes. 1999 2000 highres= [KNL] Enable/disable high resolution timer mode. 2001 Valid parameters: "on", "off" 2002 Default: "on" 2003 2004 hlt [BUGS=ARM,SH] 2005 2006 hostname= [KNL,EARLY] Set the hostname (aka UTS nodename). 2007 Format: <string> 2008 This allows setting the system's hostname during early 2009 startup. This sets the name returned by gethostname. 2010 Using this parameter to set the hostname makes it 2011 possible to ensure the hostname is correctly set before 2012 any userspace processes run, avoiding the possibility 2013 that a process may call gethostname before the hostname 2014 has been explicitly set, resulting in the calling 2015 process getting an incorrect result. The string must 2016 not exceed the maximum allowed hostname length (usually 2017 64 characters) and will be truncated otherwise. 2018 2019 hpet= [X86-32,HPET] option to control HPET usage 2020 Format: { enable (default) | disable | force | 2021 verbose } 2022 disable: disable HPET and use PIT instead 2023 force: allow force enabled of undocumented chips (ICH4, 2024 VIA, nVidia) 2025 verbose: show contents of HPET registers during setup 2026 2027 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET 2028 registers. Default set by CONFIG_HPET_MMAP_DEFAULT. 2029 2030 hugepages= [HW,EARLY] Number of HugeTLB pages to allocate at boot. 2031 If this follows hugepagesz (below), it specifies 2032 the number of pages of hugepagesz to be allocated. 2033 If this is the first HugeTLB parameter on the command 2034 line, it specifies the number of pages to allocate for 2035 the default huge page size. If using node format, the 2036 number of pages to allocate per-node can be specified. 2037 See also Documentation/admin-guide/mm/hugetlbpage.rst. 2038 Format: <integer> or (node format) 2039 <node>:<integer>[,<node>:<integer>] 2040 2041 hugepagesz= 2042 [HW,EARLY] The size of the HugeTLB pages. This is 2043 used in conjunction with hugepages (above) to 2044 allocate huge pages of a specific size at boot. The 2045 pair hugepagesz=X hugepages=Y can be specified once 2046 for each supported huge page size. Huge page sizes 2047 are architecture dependent. See also 2048 Documentation/admin-guide/mm/hugetlbpage.rst. 2049 Format: size[KMG] 2050 2051 hugepage_alloc_threads= 2052 [HW] The number of threads that should be used to 2053 allocate hugepages during boot. This option can be 2054 used to improve system bootup time when allocating 2055 a large amount of huge pages. 2056 The default value is 25% of the available hardware threads. 2057 2058 Note that this parameter only applies to non-gigantic huge pages. 2059 2060 hugetlb_cma= [HW,CMA,EARLY] The size of a CMA area used for allocation 2061 of gigantic hugepages. Or using node format, the size 2062 of a CMA area per node can be specified. 2063 Format: nn[KMGTPE] or (node format) 2064 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]] 2065 2066 Reserve a CMA area of given size and allocate gigantic 2067 hugepages using the CMA allocator. If enabled, the 2068 boot-time allocation of gigantic hugepages is skipped. 2069 2070 hugetlb_cma_only= 2071 [HW,CMA,EARLY] When allocating new HugeTLB pages, only 2072 try to allocate from the CMA areas. 2073 2074 This option does nothing if hugetlb_cma= is not also 2075 specified. 2076 2077 hugetlb_free_vmemmap= 2078 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP 2079 enabled. 2080 Control if HugeTLB Vmemmap Optimization (HVO) is enabled. 2081 Allows heavy hugetlb users to free up some more 2082 memory (7 * PAGE_SIZE for each 2MB hugetlb page). 2083 Format: { on | off (default) } 2084 2085 on: enable HVO 2086 off: disable HVO 2087 2088 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y, 2089 the default is on. 2090 2091 Note that the vmemmap pages may be allocated from the added 2092 memory block itself when memory_hotplug.memmap_on_memory is 2093 enabled, those vmemmap pages cannot be optimized even if this 2094 feature is enabled. Other vmemmap pages not allocated from 2095 the added memory block itself do not be affected. 2096 2097 hung_task_panic= 2098 [KNL] Number of hung tasks to trigger kernel panic. 2099 Format: <int> 2100 2101 When set to a non-zero value, a kernel panic will be triggered if 2102 the number of detected hung tasks reaches this value. 2103 2104 0: don't panic 2105 1: panic immediately on first hung task 2106 N: panic after N hung tasks are detected in a single scan 2107 2108 The default value is controlled by the 2109 CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time option. The value 2110 selected by this boot parameter can be changed later by the 2111 kernel.hung_task_panic sysctl. 2112 2113 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC) 2114 terminal devices. Valid values: 0..8 2115 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs. 2116 If specified, z/VM IUCV HVC accepts connections 2117 from listed z/VM user IDs only. 2118 2119 hv_nopvspin [X86,HYPER_V,EARLY] 2120 Disables the paravirt spinlock optimizations 2121 which allow the hypervisor to 'idle' the guest 2122 on lock contention. 2123 2124 hw_protection= [HW] 2125 Format: reboot | shutdown 2126 2127 Hardware protection action taken on critical events like 2128 overtemperature or imminent voltage loss. 2129 2130 i2c_bus= [HW] Override the default board specific I2C bus speed 2131 or register an additional I2C bus that is not 2132 registered from board initialization code. 2133 Format: 2134 <bus_id>,<clkrate> 2135 2136 i2c_touchscreen_props= [HW,ACPI,X86] 2137 Set device-properties for ACPI-enumerated I2C-attached 2138 touchscreen, to e.g. fix coordinates of upside-down 2139 mounted touchscreens. If you need this option please 2140 submit a drivers/platform/x86/touchscreen_dmi.c patch 2141 adding a DMI quirk for this. 2142 2143 Format: 2144 <ACPI_HW_ID>:<prop_name>=<val>[:prop_name=val][:...] 2145 Where <val> is one of: 2146 Omit "=<val>" entirely Set a boolean device-property 2147 Unsigned number Set a u32 device-property 2148 Anything else Set a string device-property 2149 2150 Examples (split over multiple lines): 2151 i2c_touchscreen_props=GDIX1001:touchscreen-inverted-x: 2152 touchscreen-inverted-y 2153 2154 i2c_touchscreen_props=MSSL1680:touchscreen-size-x=1920: 2155 touchscreen-size-y=1080:touchscreen-inverted-y: 2156 firmware-name=gsl1680-vendor-model.fw:silead,home-button 2157 2158 i8042.debug [HW] Toggle i8042 debug mode 2159 i8042.unmask_kbd_data 2160 [HW] Enable printing of interrupt data from the KBD port 2161 (disabled by default, and as a pre-condition 2162 requires that i8042.debug=1 be enabled) 2163 i8042.direct [HW] Put keyboard port into non-translated mode 2164 i8042.dumbkbd [HW] Pretend that controller can only read data from 2165 keyboard and cannot control its state 2166 (Don't attempt to blink the leds) 2167 i8042.noaux [HW] Don't check for auxiliary (== mouse) port 2168 i8042.nokbd [HW] Don't check/create keyboard port 2169 i8042.noloop [HW] Disable the AUX Loopback command while probing 2170 for the AUX port 2171 i8042.nomux [HW] Don't check presence of an active multiplexing 2172 controller 2173 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX 2174 controllers 2175 i8042.notimeout [HW] Ignore timeout condition signalled by controller 2176 i8042.reset [HW] Reset the controller during init, cleanup and 2177 suspend-to-ram transitions, only during s2r 2178 transitions, or never reset 2179 Format: { 1 | Y | y | 0 | N | n } 2180 1, Y, y: always reset controller 2181 0, N, n: don't ever reset controller 2182 Default: only on s2r transitions on x86; most other 2183 architectures force reset to be always executed 2184 i8042.unlock [HW] Unlock (ignore) the keylock 2185 i8042.kbdreset [HW] Reset device connected to KBD port 2186 i8042.probe_defer 2187 [HW] Allow deferred probing upon i8042 probe errors 2188 2189 i810= [HW,DRM] 2190 2191 i915.invert_brightness= 2192 [DRM] Invert the sense of the variable that is used to 2193 set the brightness of the panel backlight. Normally a 2194 brightness value of 0 indicates backlight switched off, 2195 and the maximum of the brightness value sets the backlight 2196 to maximum brightness. If this parameter is set to 0 2197 (default) and the machine requires it, or this parameter 2198 is set to 1, a brightness value of 0 sets the backlight 2199 to maximum brightness, and the maximum of the brightness 2200 value switches the backlight off. 2201 -1 -- never invert brightness 2202 0 -- machine default 2203 1 -- force brightness inversion 2204 2205 ia32_emulation= [X86-64] 2206 Format: <bool> 2207 When true, allows loading 32-bit programs and executing 32-bit 2208 syscalls, essentially overriding IA32_EMULATION_DEFAULT_DISABLED at 2209 boot time. When false, unconditionally disables IA32 emulation. 2210 2211 2212 idle= [X86,EARLY] 2213 Format: idle=poll, idle=halt, idle=nomwait 2214 2215 idle=poll: Don't do power saving in the idle loop 2216 using HLT, but poll for rescheduling event. This will 2217 make the CPUs eat a lot more power, but may be useful 2218 to get slightly better performance in multiprocessor 2219 benchmarks. It also makes some profiling using 2220 performance counters more accurate. Please note that 2221 on systems with MONITOR/MWAIT support (like Intel 2222 EM64T CPUs) this option has no performance advantage 2223 over the normal idle loop. It may also interact badly 2224 with hyperthreading. 2225 2226 idle=halt: Halt is forced to be used for CPU idle. 2227 In such case C2/C3 won't be used again. 2228 2229 idle=nomwait: Disable mwait for CPU C-states 2230 2231 idxd.sva= [HW] 2232 Format: <bool> 2233 Allow force disabling of Shared Virtual Memory (SVA) 2234 support for the idxd driver. By default it is set to 2235 true (1). 2236 2237 idxd.tc_override= [HW] 2238 Format: <bool> 2239 Allow override of default traffic class configuration 2240 for the device. By default it is set to false (0). 2241 2242 ieee754= [MIPS] Select IEEE Std 754 conformance mode 2243 Format: { strict | legacy | 2008 | relaxed | emulated } 2244 Default: strict 2245 2246 Choose which programs will be accepted for execution 2247 based on the IEEE 754 NaN encoding(s) supported by 2248 the FPU and the NaN encoding requested with the value 2249 of an ELF file header flag individually set by each 2250 binary. Hardware implementations are permitted to 2251 support either or both of the legacy and the 2008 NaN 2252 encoding mode. 2253 2254 Available settings are as follows: 2255 strict accept binaries that request a NaN encoding 2256 supported by the FPU 2257 legacy only accept legacy-NaN binaries, if supported 2258 by the FPU 2259 2008 only accept 2008-NaN binaries, if supported 2260 by the FPU 2261 relaxed accept any binaries regardless of whether 2262 supported by the FPU 2263 emulated accept any binaries but enable FPU emulator 2264 if binary mode is unsupported by the FPU. 2265 2266 The FPU emulator is always able to support both NaN 2267 encodings, so if no FPU hardware is present or it has 2268 been disabled with 'nofpu', then the settings of 2269 'legacy' and '2008' strap the emulator accordingly, 2270 'relaxed' straps the emulator for both legacy-NaN and 2271 2008-NaN, whereas 'strict' enables legacy-NaN only on 2272 legacy processors and both NaN encodings on MIPS32 or 2273 MIPS64 CPUs. 2274 2275 The setting for ABS.fmt/NEG.fmt instruction execution 2276 mode generally follows that for the NaN encoding, 2277 except where unsupported by hardware. 2278 2279 ignore_loglevel [KNL,EARLY] 2280 Ignore loglevel setting - this will print /all/ 2281 kernel messages to the console. Useful for debugging. 2282 We also add it as printk module parameter, so users 2283 could change it dynamically, usually by 2284 /sys/module/printk/parameters/ignore_loglevel. 2285 2286 ignore_rlimit_data 2287 Ignore RLIMIT_DATA setting for data mappings, 2288 print warning at first misuse. Can be changed via 2289 /sys/module/kernel/parameters/ignore_rlimit_data. 2290 2291 ihash_entries= [KNL] 2292 Set number of hash buckets for inode cache. 2293 2294 ima_appraise= [IMA] appraise integrity measurements 2295 Format: { "off" | "enforce" | "fix" | "log" } 2296 default: "enforce" 2297 2298 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead. 2299 The builtin appraise policy appraises all files 2300 owned by uid=0. 2301 2302 ima_canonical_fmt [IMA] 2303 Use the canonical format for the binary runtime 2304 measurements, instead of host native format. 2305 2306 ima_hash= [IMA] 2307 Format: { md5 | sha1 | rmd160 | sha256 | sha384 2308 | sha512 | ... } 2309 default: "sha1" 2310 2311 The list of supported hash algorithms is defined 2312 in crypto/hash_info.h. 2313 2314 ima_policy= [IMA] 2315 The builtin policies to load during IMA setup. 2316 Format: "tcb | appraise_tcb | secure_boot | 2317 fail_securely | critical_data" 2318 2319 The "tcb" policy measures all programs exec'd, files 2320 mmap'd for exec, and all files opened with the read 2321 mode bit set by either the effective uid (euid=0) or 2322 uid=0. 2323 2324 The "appraise_tcb" policy appraises the integrity of 2325 all files owned by root. 2326 2327 The "secure_boot" policy appraises the integrity 2328 of files (eg. kexec kernel image, kernel modules, 2329 firmware, policy, etc) based on file signatures. 2330 2331 The "fail_securely" policy forces file signature 2332 verification failure also on privileged mounted 2333 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE 2334 flag. 2335 2336 The "critical_data" policy measures kernel integrity 2337 critical data. 2338 2339 ima_tcb [IMA] Deprecated. Use ima_policy= instead. 2340 Load a policy which meets the needs of the Trusted 2341 Computing Base. This means IMA will measure all 2342 programs exec'd, files mmap'd for exec, and all files 2343 opened for read by uid=0. 2344 2345 ima_template= [IMA] 2346 Select one of defined IMA measurements template formats. 2347 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" | 2348 "ima-sigv2" } 2349 Default: "ima-ng" 2350 2351 ima_template_fmt= 2352 [IMA] Define a custom template format. 2353 Format: { "field1|...|fieldN" } 2354 2355 ima= [IMA] Enable or disable IMA 2356 Format: { "off" | "on" } 2357 Default: "on" 2358 Note that disabling IMA is limited to kdump kernel. 2359 2360 indirect_target_selection= [X86,Intel] Mitigation control for Indirect 2361 Target Selection(ITS) bug in Intel CPUs. Updated 2362 microcode is also required for a fix in IBPB. 2363 2364 on: Enable mitigation (default). 2365 off: Disable mitigation. 2366 force: Force the ITS bug and deploy default 2367 mitigation. 2368 vmexit: Only deploy mitigation if CPU is affected by 2369 guest/host isolation part of ITS. 2370 stuff: Deploy RSB-fill mitigation when retpoline is 2371 also deployed. Otherwise, deploy the default 2372 mitigation. 2373 2374 For details see: 2375 Documentation/admin-guide/hw-vuln/indirect-target-selection.rst 2376 2377 init= [KNL] 2378 Format: <full_path> 2379 Run specified binary instead of /sbin/init as init 2380 process. 2381 2382 initcall_debug [KNL] Trace initcalls as they are executed. Useful 2383 for working out where the kernel is dying during 2384 startup. 2385 2386 initcall_blacklist= [KNL] Do not execute a comma-separated list of 2387 initcall functions. Useful for debugging built-in 2388 modules and initcalls. 2389 2390 initramfs_async= [KNL] 2391 Format: <bool> 2392 Default: 1 2393 This parameter controls whether the initramfs 2394 image is unpacked asynchronously, concurrently 2395 with devices being probed and 2396 initialized. This should normally just work, 2397 but as a debugging aid, one can get the 2398 historical behaviour of the initramfs 2399 unpacking being completed before device_ and 2400 late_ initcalls. 2401 2402 initrd= [BOOT,EARLY] Specify the location of the initial ramdisk 2403 2404 initrdmem= [KNL,EARLY] Specify a physical address and size from which to 2405 load the initrd. If an initrd is compiled in or 2406 specified in the bootparams, it takes priority over this 2407 setting. 2408 Format: ss[KMG],nn[KMG] 2409 Default is 0, 0 2410 2411 init_on_alloc= [MM,EARLY] Fill newly allocated pages and heap objects with 2412 zeroes. 2413 Format: 0 | 1 2414 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON. 2415 2416 init_on_free= [MM,EARLY] Fill freed pages and heap objects with zeroes. 2417 Format: 0 | 1 2418 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON. 2419 2420 init_pkru= [X86] Specify the default memory protection keys rights 2421 register contents for all processes. 0x55555554 by 2422 default (disallow access to all but pkey 0). Can 2423 override in debugfs after boot. 2424 2425 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver 2426 Format: <irq> 2427 2428 int_pln_enable [X86] Enable power limit notification interrupt 2429 2430 integrity_audit=[IMA] 2431 Format: { "0" | "1" } 2432 0 -- basic integrity auditing messages. (Default) 2433 1 -- additional integrity auditing messages. 2434 2435 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option 2436 on 2437 Enable intel iommu driver. 2438 off 2439 Disable intel iommu driver. 2440 igfx_off [Default Off] 2441 By default, gfx is mapped as normal device. If a gfx 2442 device has a dedicated DMAR unit, the DMAR unit is 2443 bypassed by not enabling DMAR with this option. In 2444 this case, gfx device will use physical address for 2445 DMA. 2446 strict [Default Off] 2447 Deprecated, equivalent to iommu.strict=1. 2448 sp_off [Default Off] 2449 By default, super page will be supported if Intel IOMMU 2450 has the capability. With this option, super page will 2451 not be supported. 2452 sm_on 2453 Enable the Intel IOMMU scalable mode if the hardware 2454 advertises that it has support for the scalable mode 2455 translation. 2456 sm_off 2457 Disallow use of the Intel IOMMU scalable mode. 2458 tboot_noforce [Default Off] 2459 Do not force the Intel IOMMU enabled under tboot. 2460 By default, tboot will force Intel IOMMU on, which 2461 could harm performance of some high-throughput 2462 devices like 40GBit network cards, even if identity 2463 mapping is enabled. 2464 Note that using this option lowers the security 2465 provided by tboot because it makes the system 2466 vulnerable to DMA attacks. 2467 2468 intel_idle.max_cstate= [KNL,HW,ACPI,X86] 2469 0 disables intel_idle and fall back on acpi_idle. 2470 1 to 9 specify maximum depth of C-state. 2471 2472 intel_pstate= [X86,EARLY] 2473 disable 2474 Do not enable intel_pstate as the default 2475 scaling driver for the supported processors 2476 active 2477 Use intel_pstate driver to bypass the scaling 2478 governors layer of cpufreq and provides it own 2479 algorithms for p-state selection. There are two 2480 P-state selection algorithms provided by 2481 intel_pstate in the active mode: powersave and 2482 performance. The way they both operate depends 2483 on whether or not the hardware managed P-states 2484 (HWP) feature has been enabled in the processor 2485 and possibly on the processor model. 2486 passive 2487 Use intel_pstate as a scaling driver, but configure it 2488 to work with generic cpufreq governors (instead of 2489 enabling its internal governor). This mode cannot be 2490 used along with the hardware-managed P-states (HWP) 2491 feature. 2492 force 2493 Enable intel_pstate on systems that prohibit it by default 2494 in favor of acpi-cpufreq. Forcing the intel_pstate driver 2495 instead of acpi-cpufreq may disable platform features, such 2496 as thermal controls and power capping, that rely on ACPI 2497 P-States information being indicated to OSPM and therefore 2498 should be used with caution. This option does not work with 2499 processors that aren't supported by the intel_pstate driver 2500 or on platforms that use pcc-cpufreq instead of acpi-cpufreq. 2501 no_hwp 2502 Do not enable hardware P state control (HWP) 2503 if available. 2504 hwp_only 2505 Only load intel_pstate on systems which support 2506 hardware P state control (HWP) if available. 2507 support_acpi_ppc 2508 Enforce ACPI _PPC performance limits. If the Fixed ACPI 2509 Description Table, specifies preferred power management 2510 profile as "Enterprise Server" or "Performance Server", 2511 then this feature is turned on by default. 2512 per_cpu_perf_limits 2513 Allow per-logical-CPU P-State performance control limits using 2514 cpufreq sysfs interface 2515 no_cas 2516 Do not enable capacity-aware scheduling (CAS) on 2517 hybrid systems 2518 2519 intremap= [X86-64,Intel-IOMMU,EARLY] 2520 on enable Interrupt Remapping (default) 2521 off disable Interrupt Remapping 2522 nosid disable Source ID checking 2523 no_x2apic_optout 2524 BIOS x2APIC opt-out request will be ignored 2525 nopost disable Interrupt Posting 2526 posted_msi 2527 enable MSIs delivered as posted interrupts 2528 2529 iomem= Disable strict checking of access to MMIO memory 2530 strict regions from userspace. 2531 relaxed 2532 2533 iommu= [X86,EARLY] 2534 2535 off 2536 Don't initialize and use any kind of IOMMU. 2537 2538 force 2539 Force the use of the hardware IOMMU even when 2540 it is not actually needed (e.g. because < 3 GB 2541 memory). 2542 2543 noforce 2544 Don't force hardware IOMMU usage when it is not 2545 needed. (default). 2546 2547 merge 2548 Do scatter-gather (SG) merging. Implies "force" 2549 (experimental). 2550 2551 nomerge 2552 Don't do scatter-gather (SG) merging. 2553 2554 biomerge 2555 Do scatter-gather (SG) merging. Implies "force" 2556 (experimental). [same as "merge"] 2557 2558 panic 2559 Always panic when IOMMU overflows. 2560 2561 nopanic 2562 Don't panic on IOMMU overflows. 2563 2564 pt 2565 Use passththrough mode by default 2566 (Equivalent to iommu.passthrough=1) 2567 2568 nopt 2569 Use translated mode for DMA by default 2570 (Equivalent to iommu.passthrough=0) 2571 2572 soft 2573 Use software bounce buffering (SWIOTLB) (default for 2574 Intel machines). This can be used to prevent the usage 2575 of an available hardware IOMMU. 2576 2577 usedac 2578 Use the DAC on VIA PCI bridge 2579 (default: disable the VIA PCI bridge DAC) 2580 2581 AMD Gart HW IOMMU-specific options: (CONFIG_GART_IOMMU) 2582 2583 <size> 2584 Set the size of the remapping area in bytes. 2585 2586 allowed 2587 Overwrite iommu off workarounds for specific chipsets 2588 2589 force 2590 Overwrite iommu off workarounds for specific chipsets 2591 2592 fullflush 2593 Flush IOMMU on each allocation (default). 2594 2595 nofullflush 2596 Don't use IOMMU fullflush. 2597 2598 memaper[=<order>] 2599 Allocate an own aperture over RAM with size 2600 32MB<<order. (default: order=1, i.e. 64MB) 2601 2602 noaperture 2603 Ask the IOMMU not to touch the aperture for AGP. 2604 2605 noagp 2606 Don't initialize the AGP driver and use full aperture. 2607 2608 iommu= [PPC/POWERNV] 2609 2610 nobypass 2611 Disable IOMMU bypass, using IOMMU for PCI devices. 2612 2613 iommu.forcedac= [ARM64,X86,EARLY] Control IOVA allocation for PCI devices. 2614 Format: { "0" | "1" } 2615 0 - Try to allocate a 32-bit DMA address first, before 2616 falling back to the full range if needed. 2617 1 - Allocate directly from the full usable range, 2618 forcing Dual Address Cycle for PCI cards supporting 2619 greater than 32-bit addressing. 2620 2621 iommu.strict= [ARM64,X86,S390,EARLY] Configure TLB invalidation behaviour 2622 Format: { "0" | "1" } 2623 0 - Lazy mode. 2624 Request that DMA unmap operations use deferred 2625 invalidation of hardware TLBs, for increased 2626 throughput at the cost of reduced device isolation. 2627 Will fall back to strict mode if not supported by 2628 the relevant IOMMU driver. 2629 1 - Strict mode. 2630 DMA unmap operations invalidate IOMMU hardware TLBs 2631 synchronously. 2632 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}. 2633 Note: on x86, strict mode specified via one of the 2634 legacy driver-specific options takes precedence. 2635 2636 iommu.passthrough= 2637 [ARM64,X86,EARLY] Configure DMA to bypass the IOMMU by default. 2638 Format: { "0" | "1" } 2639 0 - Use IOMMU translation for DMA. 2640 1 - Bypass the IOMMU for DMA. 2641 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH. 2642 2643 iommu.debug_pagealloc= 2644 [KNL,EARLY] When CONFIG_IOMMU_DEBUG_PAGEALLOC is set, this 2645 parameter enables the feature at boot time. By default, it 2646 is disabled and the system behaves the same way as a kernel 2647 built without CONFIG_IOMMU_DEBUG_PAGEALLOC. 2648 Format: { "0" | "1" } 2649 0 - Sanitizer disabled. 2650 1 - Sanitizer enabled, expect runtime overhead. 2651 2652 io7= [HW] IO7 for Marvel-based Alpha systems 2653 See comment before marvel_specify_io7 in 2654 arch/alpha/kernel/core_marvel.c. 2655 2656 io_delay= [X86,EARLY] I/O delay method 2657 0x80 2658 Standard port 0x80 based delay 2659 0xed 2660 Alternate port 0xed based delay (needed on some systems) 2661 udelay 2662 Simple two microseconds delay 2663 none 2664 No delay 2665 2666 ip= [IP_PNP] 2667 See Documentation/admin-guide/nfs/nfsroot.rst. 2668 2669 ipcmni_extend [KNL,EARLY] Extend the maximum number of unique System V 2670 IPC identifiers from 32,768 to 16,777,216. 2671 2672 ipe.enforce= [IPE] 2673 Format: <bool> 2674 Determine whether IPE starts in permissive (0) or 2675 enforce (1) mode. The default is enforce. 2676 2677 ipe.success_audit= 2678 [IPE] 2679 Format: <bool> 2680 Start IPE with success auditing enabled, emitting 2681 an audit event when a binary is allowed. The default 2682 is 0. 2683 2684 irqaffinity= [SMP] Set the default irq affinity mask 2685 The argument is a cpu list, as described above. 2686 2687 irqchip.gicv2_force_probe= 2688 [ARM,ARM64,EARLY] 2689 Format: <bool> 2690 Force the kernel to look for the second 4kB page 2691 of a GICv2 controller even if the memory range 2692 exposed by the device tree is too small. 2693 2694 irqchip.gicv3_nolpi= 2695 [ARM,ARM64,EARLY] 2696 Force the kernel to ignore the availability of 2697 LPIs (and by consequence ITSs). Intended for system 2698 that use the kernel as a bootloader, and thus want 2699 to let secondary kernels in charge of setting up 2700 LPIs. 2701 2702 irqchip.gicv3_pseudo_nmi= [ARM64,EARLY] 2703 Enables support for pseudo-NMIs in the kernel. This 2704 requires the kernel to be built with 2705 CONFIG_ARM64_PSEUDO_NMI. 2706 2707 irqchip.riscv_imsic_noipi 2708 [RISC-V,EARLY] 2709 Force the kernel to not use IMSIC software injected MSIs 2710 as IPIs. Intended for system where IMSIC is trap-n-emulated, 2711 and thus want to reduce MMIO traps when triggering IPIs 2712 to multiple harts. 2713 2714 irqfixup [HW] 2715 When an interrupt is not handled search all handlers 2716 for it. Intended to get systems with badly broken 2717 firmware running. 2718 2719 irqhandler.duration_warn_us= [KNL] 2720 Warn if an IRQ handler exceeds the specified duration 2721 threshold in microseconds. Useful for identifying 2722 long-running IRQs in the system. 2723 2724 irqpoll [HW] 2725 When an interrupt is not handled search all handlers 2726 for it. Also check all handlers each timer 2727 interrupt. Intended to get systems with badly broken 2728 firmware running. 2729 2730 isapnp= [ISAPNP] 2731 Format: <RDP>,<reset>,<pci_scan>,<verbosity> 2732 2733 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance. 2734 Format: [flag-list,]<cpu-list> 2735 2736 Specify one or more CPUs to isolate from disturbances 2737 specified in the flag list (default: domain): 2738 2739 nohz 2740 Disable the tick when a single task runs as well as 2741 disabling other kernel noises like having RCU callbacks 2742 offloaded. This is equivalent to the nohz_full parameter. 2743 2744 A residual 1Hz tick is offloaded to workqueues, which you 2745 need to affine to housekeeping through the global 2746 workqueue's affinity configured via the 2747 /sys/devices/virtual/workqueue/cpumask sysfs file, or 2748 by using the 'domain' flag described below. 2749 2750 NOTE: by default the global workqueue runs on all CPUs, 2751 so to protect individual CPUs the 'cpumask' file has to 2752 be configured manually after bootup. 2753 2754 domain 2755 Isolate from the general SMP balancing and scheduling 2756 algorithms. Note that performing domain isolation this way 2757 is irreversible: it's not possible to bring back a CPU to 2758 the domains once isolated through this boot time 2759 configuration. Use cpusets for a dynamic configuration 2760 which can be altered at runtime. For details see 2761 Documentation/admin-guide/cpu-isolation.rst. 2762 2763 You can move a process onto or off an "isolated" CPU via 2764 the CPU affinity syscalls or cpuset. 2765 <cpu number> begins at 0 and the maximum value is 2766 "number of CPUs in system - 1". 2767 2768 managed_irq 2769 2770 Isolate from being targeted by managed interrupts 2771 which have an interrupt mask containing isolated 2772 CPUs. The affinity of managed interrupts is 2773 handled by the kernel and cannot be changed via 2774 the /proc/irq/* interfaces. 2775 2776 This isolation is best effort and only effective 2777 if the automatically assigned interrupt mask of a 2778 device queue contains isolated and housekeeping 2779 CPUs. If housekeeping CPUs are online then such 2780 interrupts are directed to the housekeeping CPU 2781 so that IO submitted on the housekeeping CPU 2782 cannot disturb the isolated CPU. 2783 2784 If a queue's affinity mask contains only isolated 2785 CPUs then this parameter has no effect on the 2786 interrupt routing decision, though interrupts are 2787 only delivered when tasks running on those 2788 isolated CPUs submit IO. IO submitted on 2789 housekeeping CPUs has no influence on those 2790 queues. 2791 2792 The format of <cpu-list> is described above. 2793 2794 iucv= [HW,NET] 2795 2796 ivrs_ioapic [HW,X86-64] 2797 Provide an override to the IOAPIC-ID<->DEVICE-ID 2798 mapping provided in the IVRS ACPI table. 2799 By default, PCI segment is 0, and can be omitted. 2800 2801 For example, to map IOAPIC-ID decimal 10 to 2802 PCI segment 0x1 and PCI device 00:14.0, 2803 write the parameter as: 2804 ivrs_ioapic=10@0001:00:14.0 2805 2806 Deprecated formats: 2807 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0 2808 write the parameter as: 2809 ivrs_ioapic[10]=00:14.0 2810 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and 2811 PCI device 00:14.0 write the parameter as: 2812 ivrs_ioapic[10]=0001:00:14.0 2813 2814 ivrs_hpet [HW,X86-64] 2815 Provide an override to the HPET-ID<->DEVICE-ID 2816 mapping provided in the IVRS ACPI table. 2817 By default, PCI segment is 0, and can be omitted. 2818 2819 For example, to map HPET-ID decimal 10 to 2820 PCI segment 0x1 and PCI device 00:14.0, 2821 write the parameter as: 2822 ivrs_hpet=10@0001:00:14.0 2823 2824 Deprecated formats: 2825 * To map HPET-ID decimal 0 to PCI device 00:14.0 2826 write the parameter as: 2827 ivrs_hpet[0]=00:14.0 2828 * To map HPET-ID decimal 10 to PCI segment 0x1 and 2829 PCI device 00:14.0 write the parameter as: 2830 ivrs_ioapic[10]=0001:00:14.0 2831 2832 ivrs_acpihid [HW,X86-64] 2833 Provide an override to the ACPI-HID:UID<->DEVICE-ID 2834 mapping provided in the IVRS ACPI table. 2835 By default, PCI segment is 0, and can be omitted. 2836 2837 For example, to map UART-HID:UID AMD0020:0 to 2838 PCI segment 0x1 and PCI device ID 00:14.5, 2839 write the parameter as: 2840 ivrs_acpihid=AMD0020:0@0001:00:14.5 2841 2842 Deprecated formats: 2843 * To map UART-HID:UID AMD0020:0 to PCI segment is 0, 2844 PCI device ID 00:14.5, write the parameter as: 2845 ivrs_acpihid[00:14.5]=AMD0020:0 2846 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and 2847 PCI device ID 00:14.5, write the parameter as: 2848 ivrs_acpihid[0001:00:14.5]=AMD0020:0 2849 2850 js= [HW,JOY] Analog joystick 2851 See Documentation/input/joydev/joystick.rst. 2852 2853 kasan_multi_shot 2854 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print 2855 report on every invalid memory access. Without this 2856 parameter KASAN will print report only for the first 2857 invalid access. 2858 2859 keep_bootcon [KNL,EARLY] 2860 Do not unregister boot console at start. This is only 2861 useful for debugging when something happens in the window 2862 between unregistering the boot console and initializing 2863 the real console. 2864 2865 keepinitrd [HW,ARM] See retain_initrd. 2866 2867 kernelcore= [KNL,X86,PPC,EARLY] 2868 Format: nn[KMGTPE] | nn% | "mirror" 2869 This parameter specifies the amount of memory usable by 2870 the kernel for non-movable allocations. The requested 2871 amount is spread evenly throughout all nodes in the 2872 system as ZONE_NORMAL. The remaining memory is used for 2873 movable memory in its own zone, ZONE_MOVABLE. In the 2874 event, a node is too small to have both ZONE_NORMAL and 2875 ZONE_MOVABLE, kernelcore memory will take priority and 2876 other nodes will have a larger ZONE_MOVABLE. 2877 2878 ZONE_MOVABLE is used for the allocation of pages that 2879 may be reclaimed or moved by the page migration 2880 subsystem. Note that allocations like PTEs-from-HighMem 2881 still use the HighMem zone if it exists, and the Normal 2882 zone if it does not. 2883 2884 It is possible to specify the exact amount of memory in 2885 the form of "nn[KMGTPE]", a percentage of total system 2886 memory in the form of "nn%", or "mirror". If "mirror" 2887 option is specified, mirrored (reliable) memory is used 2888 for non-movable allocations and remaining memory is used 2889 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror" 2890 are exclusive, so you cannot specify multiple forms. 2891 2892 kfence.burst= [MM,KFENCE] The number of additional successive 2893 allocations to be attempted through KFENCE for each 2894 sample interval. 2895 Format: <unsigned integer> 2896 Default: 0 2897 2898 kfence.check_on_panic= 2899 [MM,KFENCE] Whether to check all KFENCE-managed objects' 2900 canaries on panic. 2901 Format: <bool> 2902 Default: false 2903 2904 kfence.deferrable= 2905 [MM,KFENCE] Whether to use a deferrable timer to trigger 2906 allocations. This avoids forcing CPU wake-ups if the 2907 system is idle, at the risk of a less predictable 2908 sample interval. 2909 Format: <bool> 2910 Default: CONFIG_KFENCE_DEFERRABLE 2911 2912 kfence.fault= [MM,KFENCE] Controls the behavior when a KFENCE 2913 error is detected. 2914 report - print the error report and continue (default). 2915 oops - print the error report and oops. 2916 panic - print the error report and panic. 2917 2918 kfence.sample_interval= 2919 [MM,KFENCE] KFENCE's sample interval in milliseconds. 2920 Format: <unsigned integer> 2921 0 - Disable KFENCE. 2922 >0 - Enabled KFENCE with given sample interval. 2923 Default: CONFIG_KFENCE_SAMPLE_INTERVAL 2924 2925 kfence.skip_covered_thresh= 2926 [MM,KFENCE] If pool utilization reaches this threshold 2927 (pool usage%), KFENCE limits currently covered 2928 allocations of the same source from further filling 2929 up the pool. 2930 Format: <unsigned integer> 2931 Default: 75 2932 2933 kgdbdbgp= [KGDB,HW,EARLY] kgdb over EHCI usb debug port. 2934 Format: <Controller#>[,poll interval] 2935 The controller # is the number of the ehci usb debug 2936 port as it is probed via PCI. The poll interval is 2937 optional and is the number seconds in between 2938 each poll cycle to the debug port in case you need 2939 the functionality for interrupting the kernel with 2940 gdb or control-c on the dbgp connection. When 2941 not using this parameter you use sysrq-g to break into 2942 the kernel debugger. 2943 2944 kgdboc= [KGDB,HW] kgdb over consoles. 2945 Requires a tty driver that supports console polling, 2946 or a supported polling keyboard driver (non-usb). 2947 Serial only format: <serial_device>[,baud] 2948 keyboard only format: kbd 2949 keyboard and serial format: kbd,<serial_device>[,baud] 2950 Optional Kernel mode setting: 2951 kms, kbd format: kms,kbd 2952 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud] 2953 2954 kgdboc_earlycon= [KGDB,HW,EARLY] 2955 If the boot console provides the ability to read 2956 characters and can work in polling mode, you can use 2957 this parameter to tell kgdb to use it as a backend 2958 until the normal console is registered. Intended to 2959 be used together with the kgdboc parameter which 2960 specifies the normal console to transition to. 2961 2962 The name of the early console should be specified 2963 as the value of this parameter. Note that the name of 2964 the early console might be different than the tty 2965 name passed to kgdboc. It's OK to leave the value 2966 blank and the first boot console that implements 2967 read() will be picked. 2968 2969 kgdbwait [KGDB,EARLY] Stop kernel execution and enter the 2970 kernel debugger at the earliest opportunity. 2971 2972 kho= [KEXEC,EARLY] 2973 Format: { "0" | "1" | "off" | "on" | "y" | "n" } 2974 Enables or disables Kexec HandOver. 2975 "0" | "off" | "n" - kexec handover is disabled 2976 "1" | "on" | "y" - kexec handover is enabled 2977 2978 kho_scratch= [KEXEC,EARLY] 2979 Format: ll[KMG],mm[KMG],nn[KMG] | nn% 2980 Defines the size of the KHO scratch region. The KHO 2981 scratch regions are physically contiguous memory 2982 ranges that can only be used for non-kernel 2983 allocations. That way, even when memory is heavily 2984 fragmented with handed over memory, the kexeced 2985 kernel will always have enough contiguous ranges to 2986 bootstrap itself. 2987 2988 It is possible to specify the exact amount of 2989 memory in the form of "ll[KMG],mm[KMG],nn[KMG]" 2990 where the first parameter defines the size of a low 2991 memory scratch area, the second parameter defines 2992 the size of a global scratch area and the third 2993 parameter defines the size of additional per-node 2994 scratch areas. The form "nn%" defines scale factor 2995 (in percents) of memory that was used during boot. 2996 2997 kmac= [MIPS] Korina ethernet MAC address. 2998 Configure the RouterBoard 532 series on-chip 2999 Ethernet adapter MAC address. 3000 3001 kmemleak= [KNL,EARLY] Boot-time kmemleak enable/disable 3002 Valid arguments: on, off 3003 Default: on 3004 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y, 3005 the default is off. 3006 3007 kprobe_event=[probe-list] 3008 [FTRACE] Add kprobe events and enable at boot time. 3009 The probe-list is a semicolon delimited list of probe 3010 definitions. Each definition is same as kprobe_events 3011 interface, but the parameters are comma delimited. 3012 For example, to add a kprobe event on vfs_read with 3013 arg1 and arg2, add to the command line; 3014 3015 kprobe_event=p,vfs_read,$arg1,$arg2 3016 3017 See also Documentation/trace/kprobetrace.rst "Kernel 3018 Boot Parameter" section. 3019 3020 kpti= [ARM64,EARLY] Control page table isolation of 3021 user and kernel address spaces. 3022 Default: enabled on cores which need mitigation. 3023 0: force disabled 3024 1: force enabled 3025 3026 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires 3027 CONFIG_KUNIT to be set to be fully enabled. The 3028 default value can be overridden via 3029 KUNIT_DEFAULT_ENABLED. 3030 Default is 1 (enabled) 3031 3032 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs. 3033 Default is 0 (don't ignore, but inject #GP) 3034 3035 kvm.eager_page_split= 3036 [KVM,X86] Controls whether or not KVM will try to 3037 proactively split all huge pages during dirty logging. 3038 Eager page splitting reduces interruptions to vCPU 3039 execution by eliminating the write-protection faults 3040 and MMU lock contention that would otherwise be 3041 required to split huge pages lazily. 3042 3043 VM workloads that rarely perform writes or that write 3044 only to a small region of VM memory may benefit from 3045 disabling eager page splitting to allow huge pages to 3046 still be used for reads. 3047 3048 The behavior of eager page splitting depends on whether 3049 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If 3050 disabled, all huge pages in a memslot will be eagerly 3051 split when dirty logging is enabled on that memslot. If 3052 enabled, eager page splitting will be performed during 3053 the KVM_CLEAR_DIRTY ioctl, and only for the pages being 3054 cleared. 3055 3056 Eager page splitting is only supported when kvm.tdp_mmu=Y. 3057 3058 Default is Y (on). 3059 3060 kvm.enable_pmu=[KVM,X86] 3061 If enabled, KVM will virtualize PMU functionality based 3062 on the virtual CPU model defined by userspace. This 3063 can be overridden on a per-VM basis via 3064 KVM_CAP_PMU_CAPABILITY. 3065 3066 If disabled, KVM will not virtualize PMU functionality, 3067 e.g. MSRs, PMCs, PMIs, etc., even if userspace defines 3068 a virtual CPU model that contains PMU assets. 3069 3070 Note, KVM's vPMU support implicitly requires running 3071 with an in-kernel local APIC, e.g. to deliver PMIs to 3072 the guest. Running without an in-kernel local APIC is 3073 not supported, though KVM will allow such a combination 3074 (with severely degraded functionality). 3075 3076 See also enable_mediated_pmu. 3077 3078 Default is Y (on). 3079 3080 kvm.enable_virt_at_load=[KVM,ARM64,LOONGARCH,MIPS,RISCV,X86] 3081 If enabled, KVM will enable virtualization in hardware 3082 when KVM is loaded, and disable virtualization when KVM 3083 is unloaded (if KVM is built as a module). 3084 3085 If disabled, KVM will dynamically enable and disable 3086 virtualization on-demand when creating and destroying 3087 VMs, i.e. on the 0=>1 and 1=>0 transitions of the 3088 number of VMs. 3089 3090 Enabling virtualization at module load avoids potential 3091 latency for creation of the 0=>1 VM, as KVM serializes 3092 virtualization enabling across all online CPUs. The 3093 "cost" of enabling virtualization when KVM is loaded, 3094 is that doing so may interfere with using out-of-tree 3095 hypervisors that want to "own" virtualization hardware. 3096 3097 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface. 3098 Default is false (don't support). 3099 3100 kvm.nx_huge_pages= 3101 [KVM] Controls the software workaround for the 3102 X86_BUG_ITLB_MULTIHIT bug. 3103 force : Always deploy workaround. 3104 off : Never deploy workaround. 3105 auto : Deploy workaround based on the presence of 3106 X86_BUG_ITLB_MULTIHIT. 3107 3108 Default is 'auto'. 3109 3110 If the software workaround is enabled for the host, 3111 guests do need not to enable it for nested guests. 3112 3113 kvm.nx_huge_pages_recovery_ratio= 3114 [KVM] Controls how many 4KiB pages are periodically zapped 3115 back to huge pages. 0 disables the recovery, otherwise if 3116 the value is N KVM will zap 1/Nth of the 4KiB pages every 3117 period (see below). The default is 60. 3118 3119 kvm.nx_huge_pages_recovery_period_ms= 3120 [KVM] Controls the time period at which KVM zaps 4KiB pages 3121 back to huge pages. If the value is a non-zero N, KVM will 3122 zap a portion (see ratio above) of the pages every N msecs. 3123 If the value is 0 (the default), KVM will pick a period based 3124 on the ratio, such that a page is zapped after 1 hour on average. 3125 3126 kvm-{amd,intel}.enable_mediated_pmu=[KVM,AMD,INTEL] 3127 If enabled, KVM will provide a mediated virtual PMU, 3128 instead of the default perf-based virtual PMU (if 3129 kvm.enable_pmu is true and PMU is enumerated via the 3130 virtual CPU model). 3131 3132 With a perf-based vPMU, KVM operates as a user of perf, 3133 i.e. emulates guest PMU counters using perf events. 3134 KVM-created perf events are managed by perf as regular 3135 (guest-only) events, e.g. are scheduled in/out, contend 3136 for hardware resources, etc. Using a perf-based vPMU 3137 allows guest and host usage of the PMU to co-exist, but 3138 incurs non-trivial overhead and can result in silently 3139 dropped guest events (due to resource contention). 3140 3141 With a mediated vPMU, hardware PMU state is context 3142 switched around the world switch to/from the guest. 3143 KVM mediates which events the guest can utilize, but 3144 gives the guest direct access to all other PMU assets 3145 when possible (KVM may intercept some accesses if the 3146 virtual CPU model provides a subset of hardware PMU 3147 functionality). Using a mediated vPMU significantly 3148 reduces PMU virtualization overhead and eliminates lost 3149 guest events, but is mutually exclusive with using perf 3150 to profile KVM guests and adds latency to most VM-Exits 3151 (to context switch PMU state). 3152 3153 Default is N (off). 3154 3155 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in 3156 KVM/SVM. Default is 1 (enabled). 3157 3158 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables, 3159 a.k.a. Two-Dimensional Page Tables. Default is 1 3160 (enabled). Disable by KVM if hardware lacks support 3161 for NPT. 3162 3163 kvm-amd.ciphertext_hiding_asids= 3164 [KVM,AMD] Ciphertext hiding prevents disallowed accesses 3165 to SNP private memory from reading ciphertext. Instead, 3166 reads will see constant default values (0xff). 3167 3168 If ciphertext hiding is enabled, the joint SEV-ES and 3169 SEV-SNP ASID space is partitioned into separate SEV-ES 3170 and SEV-SNP ASID ranges, with the SEV-SNP range being 3171 [1..max_snp_asid] and the SEV-ES range being 3172 (max_snp_asid..min_sev_asid), where min_sev_asid is 3173 enumerated by CPUID.0x.8000_001F[EDX]. 3174 3175 A non-zero value enables SEV-SNP ciphertext hiding and 3176 adjusts the ASID ranges for SEV-ES and SEV-SNP guests. 3177 KVM caps the number of SEV-SNP ASIDs at the maximum 3178 possible value, e.g. specifying -1u will assign all 3179 joint SEV-ES and SEV-SNP ASIDs to SEV-SNP. Note, 3180 assigning all joint ASIDs to SEV-SNP, i.e. configuring 3181 max_snp_asid == min_sev_asid-1, will effectively make 3182 SEV-ES unusable. 3183 3184 kvm-arm.mode= 3185 [KVM,ARM,EARLY] Select one of KVM/arm64's modes of 3186 operation. 3187 3188 none: Forcefully disable KVM. 3189 3190 nvhe: Standard nVHE-based mode, without support for 3191 protected guests. 3192 3193 protected: Mode with support for guests whose state is 3194 kept private from the host, using VHE or 3195 nVHE depending on HW support. 3196 3197 nested: VHE-based mode with support for nested 3198 virtualization. Requires at least ARMv8.4 3199 hardware (with FEAT_NV2). 3200 3201 Defaults to VHE/nVHE based on hardware support. Setting 3202 mode to "protected" will disable kexec and hibernation 3203 for the host. To force nVHE on VHE hardware, add 3204 "arm64_sw.hvhe=0 id_aa64mmfr1.vh=0" to the 3205 command-line. 3206 "nested" and "protected" are experimental and should be 3207 used with extreme caution. 3208 3209 kvm-arm.vgic_v3_group0_trap= 3210 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-0 3211 system registers 3212 3213 kvm-arm.vgic_v3_group1_trap= 3214 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-1 3215 system registers 3216 3217 kvm-arm.vgic_v3_common_trap= 3218 [KVM,ARM,EARLY] Trap guest accesses to GICv3 common 3219 system registers 3220 3221 kvm-arm.vgic_v4_enable= 3222 [KVM,ARM,EARLY] Allow use of GICv4 for direct 3223 injection of LPIs. 3224 3225 kvm-arm.wfe_trap_policy= 3226 [KVM,ARM] Control when to set WFE instruction trap for 3227 KVM VMs. Traps are allowed but not guaranteed by the 3228 CPU architecture. 3229 3230 trap: set WFE instruction trap 3231 3232 notrap: clear WFE instruction trap 3233 3234 kvm-arm.wfi_trap_policy= 3235 [KVM,ARM] Control when to set WFI instruction trap for 3236 KVM VMs. Traps are allowed but not guaranteed by the 3237 CPU architecture. 3238 3239 trap: set WFI instruction trap 3240 3241 notrap: clear WFI instruction trap 3242 3243 kvm_cma_resv_ratio=n [PPC,EARLY] 3244 Reserves given percentage from system memory area for 3245 contiguous memory allocation for KVM hash pagetable 3246 allocation. 3247 By default it reserves 5% of total system memory. 3248 Format: <integer> 3249 Default: 5 3250 3251 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables, 3252 a.k.a. Two-Dimensional Page Tables. Default is 1 3253 (enabled). Disable by KVM if hardware lacks support 3254 for EPT. 3255 3256 kvm-intel.emulate_invalid_guest_state= 3257 [KVM,Intel] Control whether to emulate invalid guest 3258 state. Ignored if kvm-intel.enable_unrestricted_guest=1, 3259 as guest state is never invalid for unrestricted 3260 guests. This param doesn't apply to nested guests (L2), 3261 as KVM never emulates invalid L2 guest state. 3262 Default is 1 (enabled). 3263 3264 kvm-intel.flexpriority= 3265 [KVM,Intel] Control KVM's use of FlexPriority feature 3266 (TPR shadow). Default is 1 (enabled). Disable by KVM if 3267 hardware lacks support for it. 3268 3269 kvm-intel.nested= 3270 [KVM,Intel] Control nested virtualization feature in 3271 KVM/VMX. Default is 1 (enabled). 3272 3273 kvm-intel.unrestricted_guest= 3274 [KVM,Intel] Control KVM's use of unrestricted guest 3275 feature (virtualized real and unpaged mode). Default 3276 is 1 (enabled). Disable by KVM if EPT is disabled or 3277 hardware lacks support for it. 3278 3279 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault 3280 CVE-2018-3620. 3281 3282 Valid arguments: never, cond, always 3283 3284 always: L1D cache flush on every VMENTER. 3285 cond: Flush L1D on VMENTER only when the code between 3286 VMEXIT and VMENTER can leak host memory. 3287 never: Disables the mitigation 3288 3289 Default is cond (do L1 cache flush in specific instances) 3290 3291 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor 3292 Identification feature (tagged TLBs). Default is 1 3293 (enabled). Disable by KVM if hardware lacks support 3294 for it. 3295 3296 l1d_flush= [X86,INTEL,EARLY] 3297 Control mitigation for L1D based snooping vulnerability. 3298 3299 Certain CPUs are vulnerable to an exploit against CPU 3300 internal buffers which can forward information to a 3301 disclosure gadget under certain conditions. 3302 3303 In vulnerable processors, the speculatively 3304 forwarded data can be used in a cache side channel 3305 attack, to access data to which the attacker does 3306 not have direct access. 3307 3308 This parameter controls the mitigation. The 3309 options are: 3310 3311 on - enable the interface for the mitigation 3312 3313 l1tf= [X86,EARLY] Control mitigation of the L1TF vulnerability on 3314 affected CPUs 3315 3316 The kernel PTE inversion protection is unconditionally 3317 enabled and cannot be disabled. 3318 3319 full 3320 Provides all available mitigations for the 3321 L1TF vulnerability. Disables SMT and 3322 enables all mitigations in the 3323 hypervisors, i.e. unconditional L1D flush. 3324 3325 SMT control and L1D flush control via the 3326 sysfs interface is still possible after 3327 boot. Hypervisors will issue a warning 3328 when the first VM is started in a 3329 potentially insecure configuration, 3330 i.e. SMT enabled or L1D flush disabled. 3331 3332 full,force 3333 Same as 'full', but disables SMT and L1D 3334 flush runtime control. Implies the 3335 'nosmt=force' command line option. 3336 (i.e. sysfs control of SMT is disabled.) 3337 3338 flush 3339 Leaves SMT enabled and enables the default 3340 hypervisor mitigation, i.e. conditional 3341 L1D flush. 3342 3343 SMT control and L1D flush control via the 3344 sysfs interface is still possible after 3345 boot. Hypervisors will issue a warning 3346 when the first VM is started in a 3347 potentially insecure configuration, 3348 i.e. SMT enabled or L1D flush disabled. 3349 3350 flush,nosmt 3351 3352 Disables SMT and enables the default 3353 hypervisor mitigation. 3354 3355 SMT control and L1D flush control via the 3356 sysfs interface is still possible after 3357 boot. Hypervisors will issue a warning 3358 when the first VM is started in a 3359 potentially insecure configuration, 3360 i.e. SMT enabled or L1D flush disabled. 3361 3362 flush,nowarn 3363 Same as 'flush', but hypervisors will not 3364 warn when a VM is started in a potentially 3365 insecure configuration. 3366 3367 off 3368 Disables hypervisor mitigations and doesn't 3369 emit any warnings. 3370 It also drops the swap size and available 3371 RAM limit restriction on both hypervisor and 3372 bare metal. 3373 3374 Default is 'flush'. 3375 3376 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst 3377 3378 l2cr= [PPC] 3379 3380 l3cr= [PPC] 3381 3382 lapic [X86-32,APIC,EARLY] Enable the local APIC even if BIOS 3383 disabled it. 3384 3385 lapic= [X86,APIC] Do not use TSC deadline 3386 value for LAPIC timer one-shot implementation. Default 3387 back to the programmable timer unit in the LAPIC. 3388 Format: notscdeadline 3389 3390 lapic_timer_c2_ok [X86,APIC,EARLY] trust the local apic timer 3391 in C2 power state. 3392 3393 libata.dma= [LIBATA] DMA control 3394 libata.dma=0 Disable all PATA and SATA DMA 3395 libata.dma=1 PATA and SATA Disk DMA only 3396 libata.dma=2 ATAPI (CDROM) DMA only 3397 libata.dma=4 Compact Flash DMA only 3398 Combinations also work, so libata.dma=3 enables DMA 3399 for disks and CDROMs, but not CFs. 3400 3401 libata.ignore_hpa= [LIBATA] Ignore HPA limit 3402 libata.ignore_hpa=0 keep BIOS limits (default) 3403 libata.ignore_hpa=1 ignore limits, using full disk 3404 3405 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume 3406 when set. 3407 Format: <int> 3408 3409 libata.force= [LIBATA] Force configurations. The format is a comma- 3410 separated list of "[ID:]VAL" where ID is PORT[.DEVICE]. 3411 PORT and DEVICE are decimal numbers matching port, link 3412 or device. Basically, it matches the ATA ID string 3413 printed on console by libata. If the whole ID part is 3414 omitted, the last PORT and DEVICE values are used. If 3415 ID hasn't been specified yet, the configuration applies 3416 to all ports, links and devices. 3417 3418 If only DEVICE is omitted, the parameter applies to 3419 the port and all links and devices behind it. DEVICE 3420 number of 0 either selects the first device or the 3421 first fan-out link behind PMP device. It does not 3422 select the host link. DEVICE number of 15 selects the 3423 host link and device attached to it. 3424 3425 The VAL specifies the configuration to force. As long 3426 as there is no ambiguity, shortcut notation is allowed. 3427 For example, both 1.5 and 1.5G would work for 1.5Gbps. 3428 The following configurations can be forced. 3429 3430 * Cable type: 40c, 80c, short40c, unk, ign or sata. 3431 Any ID with matching PORT is used. 3432 3433 * SATA link speed limit: 1.5Gbps or 3.0Gbps. 3434 3435 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7]. 3436 udma[/][16,25,33,44,66,100,133] notation is also 3437 allowed. 3438 3439 * nohrst, nosrst, norst: suppress hard, soft and both 3440 resets. 3441 3442 * rstonce: only attempt one reset during hot-unplug 3443 link recovery. 3444 3445 * [no]dbdelay: Enable or disable the extra 200ms delay 3446 before debouncing a link PHY and device presence 3447 detection. 3448 3449 * [no]ncq: Turn on or off NCQ. 3450 3451 * [no]ncqtrim: Enable or disable queued DSM TRIM. 3452 3453 * [no]ncqati: Enable or disable NCQ trim on ATI chipset. 3454 3455 * [no]trim: Enable or disable (unqueued) TRIM. 3456 3457 * trim_zero: Indicate that TRIM command zeroes data. 3458 3459 * max_trim_128m: Set 128M maximum trim size limit. 3460 3461 * [no]dma: Turn on or off DMA transfers. 3462 3463 * atapi_dmadir: Enable ATAPI DMADIR bridge support. 3464 3465 * atapi_mod16_dma: Enable the use of ATAPI DMA for 3466 commands that are not a multiple of 16 bytes. 3467 3468 * [no]dmalog: Enable or disable the use of the 3469 READ LOG DMA EXT command to access logs. 3470 3471 * [no]iddevlog: Enable or disable access to the 3472 identify device data log. 3473 3474 * [no]logdir: Enable or disable access to the general 3475 purpose log directory. 3476 3477 * max_sec=<sectors>: Set the transfer size limit, in 3478 number of 512-byte sectors, to the value specified in 3479 <sectors>. The value specified in <sectors> has to be 3480 a non-zero positive integer. 3481 3482 * max_sec_128: Set transfer size limit to 128 sectors. 3483 3484 * max_sec_1024: Set or clear transfer size limit to 3485 1024 sectors. 3486 3487 * max_sec_lba48: Set or clear transfer size limit to 3488 65535 sectors. 3489 3490 * external: Mark port as external (hotplug-capable). 3491 3492 * [no]lpm: Enable or disable link power management. 3493 3494 * [no]setxfer: Indicate if transfer speed mode setting 3495 should be skipped. 3496 3497 * [no]fua: Disable or enable FUA (Force Unit Access) 3498 support for devices supporting this feature. 3499 3500 * dump_id: Dump IDENTIFY data. 3501 3502 * disable: Disable this device. 3503 3504 If there are multiple matching configurations changing 3505 the same attribute, the last one is used. 3506 3507 liveupdate= [KNL,EARLY] 3508 Format: <bool> 3509 Enable Live Update Orchestrator (LUO). 3510 Default: off. 3511 3512 lockd.nlm_grace_period=P [NFS] Assign grace period. 3513 Format: <integer> 3514 3515 lockd.nlm_tcpport=N [NFS] Assign TCP port. 3516 Format: <integer> 3517 3518 lockd.nlm_timeout=T [NFS] Assign timeout value. 3519 Format: <integer> 3520 3521 lockd.nlm_udpport=M [NFS] Assign UDP port. 3522 Format: <integer> 3523 3524 lockdown= [SECURITY,EARLY] 3525 { integrity | confidentiality } 3526 Enable the kernel lockdown feature. If set to 3527 integrity, kernel features that allow userland to 3528 modify the running kernel are disabled. If set to 3529 confidentiality, kernel features that allow userland 3530 to extract confidential information from the kernel 3531 are also disabled. 3532 3533 locktorture.acq_writer_lim= [KNL] 3534 Set the time limit in jiffies for a lock 3535 acquisition. Acquisitions exceeding this limit 3536 will result in a splat once they do complete. 3537 3538 locktorture.bind_readers= [KNL] 3539 Specify the list of CPUs to which the readers are 3540 to be bound. 3541 3542 locktorture.bind_writers= [KNL] 3543 Specify the list of CPUs to which the writers are 3544 to be bound. 3545 3546 locktorture.call_rcu_chains= [KNL] 3547 Specify the number of self-propagating call_rcu() 3548 chains to set up. These are used to ensure that 3549 there is a high probability of an RCU grace period 3550 in progress at any given time. Defaults to 0, 3551 which disables these call_rcu() chains. 3552 3553 locktorture.long_hold= [KNL] 3554 Specify the duration in milliseconds for the 3555 occasional long-duration lock hold time. Defaults 3556 to 100 milliseconds. Select 0 to disable. 3557 3558 locktorture.nested_locks= [KNL] 3559 Specify the maximum lock nesting depth that 3560 locktorture is to exercise, up to a limit of 8 3561 (MAX_NESTED_LOCKS). Specify zero to disable. 3562 Note that this parameter is ineffective on types 3563 of locks that do not support nested acquisition. 3564 3565 locktorture.nreaders_stress= [KNL] 3566 Set the number of locking read-acquisition kthreads. 3567 Defaults to being automatically set based on the 3568 number of online CPUs. 3569 3570 locktorture.nwriters_stress= [KNL] 3571 Set the number of locking write-acquisition kthreads. 3572 3573 locktorture.onoff_holdoff= [KNL] 3574 Set time (s) after boot for CPU-hotplug testing. 3575 3576 locktorture.onoff_interval= [KNL] 3577 Set time (s) between CPU-hotplug operations, or 3578 zero to disable CPU-hotplug testing. 3579 3580 locktorture.rt_boost= [KNL] 3581 Do periodic testing of real-time lock priority 3582 boosting. Select 0 to disable, 1 to boost 3583 only rt_mutex, and 2 to boost unconditionally. 3584 Defaults to 2, which might seem to be an 3585 odd choice, but which should be harmless for 3586 non-real-time spinlocks, due to their disabling 3587 of preemption. Note that non-realtime mutexes 3588 disable boosting. 3589 3590 locktorture.rt_boost_factor= [KNL] 3591 Number that determines how often and for how 3592 long priority boosting is exercised. This is 3593 scaled down by the number of writers, so that the 3594 number of boosts per unit time remains roughly 3595 constant as the number of writers increases. 3596 On the other hand, the duration of each boost 3597 increases with the number of writers. 3598 3599 locktorture.shuffle_interval= [KNL] 3600 Set task-shuffle interval (jiffies). Shuffling 3601 tasks allows some CPUs to go into dyntick-idle 3602 mode during the locktorture test. 3603 3604 locktorture.shutdown_secs= [KNL] 3605 Set time (s) after boot system shutdown. This 3606 is useful for hands-off automated testing. 3607 3608 locktorture.stat_interval= [KNL] 3609 Time (s) between statistics printk()s. 3610 3611 locktorture.stutter= [KNL] 3612 Time (s) to stutter testing, for example, 3613 specifying five seconds causes the test to run for 3614 five seconds, wait for five seconds, and so on. 3615 This tests the locking primitive's ability to 3616 transition abruptly to and from idle. 3617 3618 locktorture.torture_type= [KNL] 3619 Specify the locking implementation to test. 3620 3621 locktorture.verbose= [KNL] 3622 Enable additional printk() statements. 3623 3624 locktorture.writer_fifo= [KNL] 3625 Run the write-side locktorture kthreads at 3626 sched_set_fifo() real-time priority. 3627 3628 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver 3629 Format: <irq> 3630 3631 loglevel= [KNL,EARLY] 3632 All Kernel Messages with a loglevel smaller than the 3633 console loglevel will be printed to the console. It can 3634 also be changed with klogd or other programs. The 3635 loglevels are defined as follows: 3636 3637 0 (KERN_EMERG) system is unusable 3638 1 (KERN_ALERT) action must be taken immediately 3639 2 (KERN_CRIT) critical conditions 3640 3 (KERN_ERR) error conditions 3641 4 (KERN_WARNING) warning conditions 3642 5 (KERN_NOTICE) normal but significant condition 3643 6 (KERN_INFO) informational 3644 7 (KERN_DEBUG) debug-level messages 3645 3646 log_buf_len=n[KMG] [KNL,EARLY] 3647 Sets the size of the printk ring buffer, in bytes. 3648 n must be a power of two and greater than the 3649 minimal size. The minimal size is defined by 3650 LOG_BUF_SHIFT kernel config parameter. There 3651 is also CONFIG_LOG_CPU_MAX_BUF_SHIFT config 3652 parameter that allows to increase the default size 3653 depending on the number of CPUs. See init/Kconfig 3654 for more details. 3655 3656 logo.nologo [FB] Disables display of the built-in Linux logo. 3657 This may be used to provide more screen space for 3658 kernel log messages and is useful when debugging 3659 kernel boot problems. 3660 3661 lp=0 [LP] Specify parallel ports to use, e.g, 3662 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses 3663 lp=reset first parallel port). 'lp=0' disables the 3664 lp=auto printer driver. 'lp=reset' (which can be 3665 specified in addition to the ports) causes 3666 attached printers to be reset. Using 3667 lp=port1,port2,... specifies the parallel ports 3668 to associate lp devices with, starting with 3669 lp0. A port specification may be 'none' to skip 3670 that lp device, or a parport name such as 3671 'parport0'. Specifying 'lp=auto' instead of a 3672 port specification list means that device IDs 3673 from each port should be examined, to see if 3674 an IEEE 1284-compliant printer is attached; if 3675 so, the driver will manage that printer. 3676 See also header of drivers/char/lp.c. 3677 3678 lpj=n [KNL] 3679 Sets loops_per_jiffy to given constant, thus avoiding 3680 time-consuming boot-time autodetection (up to 250 ms per 3681 CPU). 0 enables autodetection (default). To determine 3682 the correct value for your kernel, boot with normal 3683 autodetection and see what value is printed. Note that 3684 on SMP systems the preset will be applied to all CPUs, 3685 which is likely to cause problems if your CPUs need 3686 significantly divergent settings. An incorrect value 3687 will cause delays in the kernel to be wrong, leading to 3688 unpredictable I/O errors and other breakage. Although 3689 unlikely, in the extreme case this might damage your 3690 hardware. 3691 3692 lsm.debug [SECURITY] Enable LSM initialization debugging output. 3693 3694 lsm=lsm1,...,lsmN 3695 [SECURITY] Choose order of LSM initialization. This 3696 overrides CONFIG_LSM, and the "security=" parameter. 3697 3698 machtype= [Loongson] Share the same kernel image file between 3699 different yeeloong laptops. 3700 Example: machtype=lemote-yeeloong-2f-7inch 3701 3702 maxcpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 3703 will bring up during bootup. maxcpus=n : n >= 0 limits 3704 the kernel to bring up 'n' processors. Surely after 3705 bootup you can bring up the other plugged cpu by executing 3706 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus 3707 only takes effect during system bootup. 3708 While n=0 is a special case, it is equivalent to "nosmp", 3709 which also disables the IO APIC. 3710 3711 max_loop= [LOOP] The number of loop block devices that get 3712 (loop.max_loop) unconditionally pre-created at init time. The default 3713 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead 3714 of statically allocating a predefined number, loop 3715 devices can be requested on-demand with the 3716 /dev/loop-control interface. 3717 3718 mce= [X86-{32,64}] 3719 3720 Please see Documentation/arch/x86/x86_64/machinecheck.rst for sysfs runtime tunables. 3721 3722 off 3723 disable machine check 3724 3725 no_cmci 3726 disable CMCI(Corrected Machine Check Interrupt) that 3727 Intel processor supports. Usually this disablement is 3728 not recommended, but it might be handy if your 3729 hardware is misbehaving. 3730 3731 Note that you'll get more problems without CMCI than 3732 with due to the shared banks, i.e. you might get 3733 duplicated error logs. 3734 3735 dont_log_ce 3736 don't make logs for corrected errors. All events 3737 reported as corrected are silently cleared by OS. This 3738 option will be useful if you have no interest in any 3739 of corrected errors. 3740 3741 ignore_ce 3742 disable features for corrected errors, e.g. 3743 polling timer and CMCI. All events reported as 3744 corrected are not cleared by OS and remained in its 3745 error banks. 3746 3747 Usually this disablement is not recommended, however 3748 if there is an agent checking/clearing corrected 3749 errors (e.g. BIOS or hardware monitoring 3750 applications), conflicting with OS's error handling, 3751 and you cannot deactivate the agent, then this option 3752 will be a help. 3753 3754 no_lmce 3755 do not opt-in to Local MCE delivery. Use legacy method 3756 to broadcast MCEs. 3757 3758 bootlog 3759 enable logging of machine checks left over from 3760 booting. Disabled by default on AMD Fam10h and older 3761 because some BIOS leave bogus ones. 3762 3763 If your BIOS doesn't do that it's a good idea to 3764 enable though to make sure you log even machine check 3765 events that result in a reboot. On Intel systems it is 3766 enabled by default. 3767 3768 nobootlog 3769 disable boot machine check logging. 3770 3771 monarchtimeout (number) 3772 sets the time in us to wait for other CPUs on machine 3773 checks. 0 to disable. 3774 3775 bios_cmci_threshold 3776 don't overwrite the bios-set CMCI threshold. This boot 3777 option prevents Linux from overwriting the CMCI 3778 threshold set by the bios. Without this option, Linux 3779 always sets the CMCI threshold to 1. Enabling this may 3780 make memory predictive failure analysis less effective 3781 if the bios sets thresholds for memory errors since we 3782 will not see details for all errors. 3783 3784 recovery 3785 force-enable recoverable machine check code paths 3786 3787 Everything else is in sysfs now. 3788 3789 3790 md= [HW] RAID subsystems devices and level 3791 See Documentation/admin-guide/md.rst. 3792 3793 mds= [X86,INTEL,EARLY] 3794 Control mitigation for the Micro-architectural Data 3795 Sampling (MDS) vulnerability. 3796 3797 Certain CPUs are vulnerable to an exploit against CPU 3798 internal buffers which can forward information to a 3799 disclosure gadget under certain conditions. 3800 3801 In vulnerable processors, the speculatively 3802 forwarded data can be used in a cache side channel 3803 attack, to access data to which the attacker does 3804 not have direct access. 3805 3806 This parameter controls the MDS mitigation. The 3807 options are: 3808 3809 full - Enable MDS mitigation on vulnerable CPUs 3810 full,nosmt - Enable MDS mitigation and disable 3811 SMT on vulnerable CPUs 3812 off - Unconditionally disable MDS mitigation 3813 3814 On TAA-affected machines, mds=off can be prevented by 3815 an active TAA mitigation as both vulnerabilities are 3816 mitigated with the same mechanism so in order to disable 3817 this mitigation, you need to specify tsx_async_abort=off 3818 too. 3819 3820 Not specifying this option is equivalent to 3821 mds=full. 3822 3823 For details see: Documentation/admin-guide/hw-vuln/mds.rst 3824 3825 mem=nn[KMG] [HEXAGON,EARLY] Set the memory size. 3826 Must be specified, otherwise memory size will be 0. 3827 3828 mem=nn[KMG] [KNL,BOOT,EARLY] Force usage of a specific amount 3829 of memory Amount of memory to be used in cases 3830 as follows: 3831 3832 1 for test; 3833 2 when the kernel is not able to see the whole system memory; 3834 3 memory that lies after 'mem=' boundary is excluded from 3835 the hypervisor, then assigned to KVM guests. 3836 4 to limit the memory available for kdump kernel. 3837 3838 [ARC,MICROBLAZE] - the limit applies only to low memory, 3839 high memory is not affected. 3840 3841 [ARM64] - only limits memory covered by the linear 3842 mapping. The NOMAP regions are not affected. 3843 3844 [X86] Work as limiting max address. Use together 3845 with memmap= to avoid physical address space collisions. 3846 Without memmap= PCI devices could be placed at addresses 3847 belonging to unused RAM. 3848 3849 Note that this only takes effects during boot time since 3850 in above case 3, memory may need be hot added after boot 3851 if system memory of hypervisor is not sufficient. 3852 3853 mem=nn[KMG]@ss[KMG] 3854 [ARM,MIPS,EARLY] - override the memory layout 3855 reported by firmware. 3856 Define a memory region of size nn[KMG] starting at 3857 ss[KMG]. 3858 Multiple different regions can be specified with 3859 multiple mem= parameters on the command line. 3860 3861 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel 3862 memory. 3863 3864 memblock=debug [KNL,EARLY] Enable memblock debug messages. 3865 3866 memchunk=nn[KMG] 3867 [KNL,SH] Allow user to override the default size for 3868 per-device physically contiguous DMA buffers. 3869 3870 memhp_default_state=online/offline/online_kernel/online_movable 3871 [KNL] Set the initial state for the memory hotplug 3872 onlining policy. If not specified, the default value is 3873 set according to the 3874 CONFIG_MHP_DEFAULT_ONLINE_TYPE kernel config 3875 options. 3876 See Documentation/admin-guide/mm/memory-hotplug.rst. 3877 3878 memmap=exactmap [KNL,X86,EARLY] Enable setting of an exact 3879 E820 memory map, as specified by the user. 3880 Such memmap=exactmap lines can be constructed based on 3881 BIOS output or other requirements. See the memmap=nn@ss 3882 option description. 3883 3884 memmap=nn[KMG]@ss[KMG] 3885 [KNL, X86,MIPS,XTENSA,EARLY] Force usage of a specific region of memory. 3886 Region of memory to be used is from ss to ss+nn. 3887 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG], 3888 which limits max address to nn[KMG]. 3889 Multiple different regions can be specified, 3890 comma delimited. 3891 Example: 3892 memmap=100M@2G,100M#3G,1G!1024G 3893 3894 memmap=nn[KMG]#ss[KMG] 3895 [KNL,ACPI,EARLY] Mark specific memory as ACPI data. 3896 Region of memory to be marked is from ss to ss+nn. 3897 3898 memmap=nn[KMG]$ss[KMG] 3899 [KNL,ACPI,EARLY] Mark specific memory as reserved. 3900 Region of memory to be reserved is from ss to ss+nn. 3901 Example: Exclude memory from 0x18690000-0x1869ffff 3902 memmap=64K$0x18690000 3903 or 3904 memmap=0x10000$0x18690000 3905 Some bootloaders may need an escape character before '$', 3906 like Grub2, otherwise '$' and the following number 3907 will be eaten. 3908 3909 memmap=nn[KMG]!ss[KMG,EARLY] 3910 [KNL,X86] Mark specific memory as protected. 3911 Region of memory to be used, from ss to ss+nn. 3912 The memory region may be marked as e820 type 12 (0xc) 3913 and is NVDIMM or ADR memory. 3914 3915 memmap=<size>%<offset>-<oldtype>+<newtype> 3916 [KNL,ACPI,EARLY] Convert memory within the specified region 3917 from <oldtype> to <newtype>. If "-<oldtype>" is left 3918 out, the whole region will be marked as <newtype>, 3919 even if previously unavailable. If "+<newtype>" is left 3920 out, matching memory will be removed. Types are 3921 specified as e820 types, e.g., 1 = RAM, 2 = reserved, 3922 3 = ACPI, 12 = PRAM. 3923 3924 memory_corruption_check=0/1 [X86,EARLY] 3925 Some BIOSes seem to corrupt the first 64k of 3926 memory when doing things like suspend/resume. 3927 Setting this option will scan the memory 3928 looking for corruption. Enabling this will 3929 both detect corruption and prevent the kernel 3930 from using the memory being corrupted. 3931 However, it's intended as a diagnostic tool; if 3932 repeatable BIOS-originated corruption always 3933 affects the same memory, you can use memmap= 3934 to prevent the kernel from using that memory. 3935 3936 memory_corruption_check_size=size [X86,EARLY] 3937 By default it checks for corruption in the low 3938 64k, making this memory unavailable for normal 3939 use. Use this parameter to scan for 3940 corruption in more or less memory. 3941 3942 memory_corruption_check_period=seconds [X86,EARLY] 3943 By default it checks for corruption every 60 3944 seconds. Use this parameter to check at some 3945 other rate. 0 disables periodic checking. 3946 3947 memory_hotplug.memmap_on_memory 3948 [KNL,X86,ARM] Boolean flag to enable this feature. 3949 Format: {on | off (default)} 3950 When enabled, runtime hotplugged memory will 3951 allocate its internal metadata (struct pages, 3952 those vmemmap pages cannot be optimized even 3953 if hugetlb_free_vmemmap is enabled) from the 3954 hotadded memory which will allow to hotadd a 3955 lot of memory without requiring additional 3956 memory to do so. 3957 This feature is disabled by default because it 3958 has some implication on large (e.g. GB) 3959 allocations in some configurations (e.g. small 3960 memory blocks). 3961 The state of the flag can be read in 3962 /sys/module/memory_hotplug/parameters/memmap_on_memory. 3963 Note that even when enabled, there are a few cases where 3964 the feature is not effective. 3965 3966 memtest= [KNL,X86,ARM,M68K,PPC,RISCV,EARLY] Enable memtest 3967 Format: <integer> 3968 default : 0 <disable> 3969 Specifies the number of memtest passes to be 3970 performed. Each pass selects another test 3971 pattern from a given set of patterns. Memtest 3972 fills the memory with this pattern, validates 3973 memory contents and reserves bad memory 3974 regions that are detected. 3975 3976 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control 3977 Valid arguments: on, off 3978 Default: off 3979 mem_encrypt=on: Activate SME 3980 mem_encrypt=off: Do not activate SME 3981 3982 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst 3983 for details on when memory encryption can be activated. 3984 3985 mem_sleep_default= [SUSPEND] Default system suspend mode: 3986 s2idle - Suspend-To-Idle 3987 shallow - Power-On Suspend or equivalent (if supported) 3988 deep - Suspend-To-RAM or equivalent (if supported) 3989 See Documentation/admin-guide/pm/sleep-states.rst. 3990 3991 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when 3992 the BIOS has incorrectly applied a workaround. TinyBIOS 3993 version 0.98 is known to be affected, 0.99 fixes the 3994 problem by letting the user disable the workaround. 3995 3996 mga= [HW,DRM] 3997 3998 microcode= [X86] Control the behavior of the microcode loader. 3999 Available options, comma separated: 4000 4001 base_rev=X - with <X> with format: <u32> 4002 Set the base microcode revision of each thread when in 4003 debug mode. 4004 4005 dis_ucode_ldr: disable the microcode loader 4006 4007 force_minrev: 4008 Enable or disable the microcode minimal revision 4009 enforcement for the runtime microcode loader. 4010 4011 mini2440= [ARM,HW,KNL] 4012 Format:[0..2][b][c][t] 4013 Default: "0tb" 4014 MINI2440 configuration specification: 4015 0 - The attached screen is the 3.5" TFT 4016 1 - The attached screen is the 7" TFT 4017 2 - The VGA Shield is attached (1024x768) 4018 Leaving out the screen size parameter will not load 4019 the TFT driver, and the framebuffer will be left 4020 unconfigured. 4021 b - Enable backlight. The TFT backlight pin will be 4022 linked to the kernel VESA blanking code and a GPIO 4023 LED. This parameter is not necessary when using the 4024 VGA shield. 4025 c - Enable the s3c camera interface. 4026 t - Reserved for enabling touchscreen support. The 4027 touchscreen support is not enabled in the mainstream 4028 kernel as of 2.6.30, a preliminary port can be found 4029 in the "bleeding edge" mini2440 support kernel at 4030 https://repo.or.cz/w/linux-2.6/mini2440.git 4031 4032 mitigations= 4033 [X86,PPC,S390,ARM64,EARLY] Control optional mitigations for 4034 CPU vulnerabilities. This is a set of curated, 4035 arch-independent options, each of which is an 4036 aggregation of existing arch-specific options. 4037 4038 Note, "mitigations" is supported if and only if the 4039 kernel was built with CPU_MITIGATIONS=y. 4040 4041 off 4042 Disable all optional CPU mitigations. This 4043 improves system performance, but it may also 4044 expose users to several CPU vulnerabilities. 4045 Equivalent to: if nokaslr then kpti=0 [ARM64] 4046 gather_data_sampling=off [X86] 4047 indirect_target_selection=off [X86] 4048 kvm.nx_huge_pages=off [X86] 4049 l1tf=off [X86] 4050 mds=off [X86] 4051 mmio_stale_data=off [X86] 4052 no_entry_flush [PPC] 4053 no_uaccess_flush [PPC] 4054 nobp=0 [S390] 4055 nopti [X86,PPC] 4056 nospectre_bhb [ARM64] 4057 nospectre_v1 [X86,PPC] 4058 nospectre_v2 [X86,PPC,S390,ARM64] 4059 reg_file_data_sampling=off [X86] 4060 retbleed=off [X86] 4061 spec_rstack_overflow=off [X86] 4062 spec_store_bypass_disable=off [X86,PPC] 4063 spectre_bhi=off [X86] 4064 spectre_v2_user=off [X86] 4065 srbds=off [X86,INTEL] 4066 ssbd=force-off [ARM64] 4067 tsa=off [X86,AMD] 4068 tsx_async_abort=off [X86] 4069 vmscape=off [X86] 4070 4071 Exceptions: 4072 This does not have any effect on 4073 kvm.nx_huge_pages when 4074 kvm.nx_huge_pages=force. 4075 4076 auto (default) 4077 Mitigate all CPU vulnerabilities, but leave SMT 4078 enabled, even if it's vulnerable. This is for 4079 users who don't want to be surprised by SMT 4080 getting disabled across kernel upgrades, or who 4081 have other ways of avoiding SMT-based attacks. 4082 Equivalent to: (default behavior) 4083 4084 auto,nosmt 4085 Mitigate all CPU vulnerabilities, disabling SMT 4086 if needed. This is for users who always want to 4087 be fully mitigated, even if it means losing SMT. 4088 Equivalent to: l1tf=flush,nosmt [X86] 4089 mds=full,nosmt [X86] 4090 tsx_async_abort=full,nosmt [X86] 4091 mmio_stale_data=full,nosmt [X86] 4092 retbleed=auto,nosmt [X86] 4093 4094 [X86] After one of the above options, additionally 4095 supports attack-vector based controls as documented in 4096 Documentation/admin-guide/hw-vuln/attack_vector_controls.rst 4097 4098 mminit_loglevel= 4099 [KNL,EARLY] When CONFIG_DEBUG_MEMORY_INIT is set, this 4100 parameter allows control of the logging verbosity for 4101 the additional memory initialisation checks. A value 4102 of 0 disables mminit logging and a level of 4 will 4103 log everything. Information is printed at KERN_DEBUG 4104 so loglevel=8 may also need to be specified. 4105 4106 mmio_stale_data= 4107 [X86,INTEL,EARLY] Control mitigation for the Processor 4108 MMIO Stale Data vulnerabilities. 4109 4110 Processor MMIO Stale Data is a class of 4111 vulnerabilities that may expose data after an MMIO 4112 operation. Exposed data could originate or end in 4113 the same CPU buffers as affected by MDS and TAA. 4114 Therefore, similar to MDS and TAA, the mitigation 4115 is to clear the affected CPU buffers. 4116 4117 This parameter controls the mitigation. The 4118 options are: 4119 4120 full - Enable mitigation on vulnerable CPUs 4121 4122 full,nosmt - Enable mitigation and disable SMT on 4123 vulnerable CPUs. 4124 4125 off - Unconditionally disable mitigation 4126 4127 On MDS or TAA affected machines, 4128 mmio_stale_data=off can be prevented by an active 4129 MDS or TAA mitigation as these vulnerabilities are 4130 mitigated with the same mechanism so in order to 4131 disable this mitigation, you need to specify 4132 mds=off and tsx_async_abort=off too. 4133 4134 Not specifying this option is equivalent to 4135 mmio_stale_data=full. 4136 4137 For details see: 4138 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst 4139 4140 <module>.async_probe[=<bool>] [KNL] 4141 If no <bool> value is specified or if the value 4142 specified is not a valid <bool>, enable asynchronous 4143 probe on this module. Otherwise, enable/disable 4144 asynchronous probe on this module as indicated by the 4145 <bool> value. See also: module.async_probe 4146 4147 module.async_probe=<bool> 4148 [KNL] When set to true, modules will use async probing 4149 by default. To enable/disable async probing for a 4150 specific module, use the module specific control that 4151 is documented under <module>.async_probe. When both 4152 module.async_probe and <module>.async_probe are 4153 specified, <module>.async_probe takes precedence for 4154 the specific module. 4155 4156 module.enable_dups_trace 4157 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set, 4158 this means that duplicate request_module() calls will 4159 trigger a WARN_ON() instead of a pr_warn(). Note that 4160 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s 4161 will always be issued and this option does nothing. 4162 module.sig_enforce 4163 [KNL] When CONFIG_MODULE_SIG is set, this means that 4164 modules without (valid) signatures will fail to load. 4165 Note that if CONFIG_MODULE_SIG_FORCE is set, that 4166 is always true, so this option does nothing. 4167 4168 module_blacklist= [KNL] Do not load a comma-separated list of 4169 modules. Useful for debugging problem modules. 4170 4171 mousedev.tap_time= 4172 [MOUSE] Maximum time between finger touching and 4173 leaving touchpad surface for touch to be considered 4174 a tap and be reported as a left button click (for 4175 touchpads working in absolute mode only). 4176 Format: <msecs> 4177 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices 4178 reporting absolute coordinates, such as tablets 4179 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices 4180 reporting absolute coordinates, such as tablets 4181 4182 movablecore= [KNL,X86,PPC,EARLY] 4183 Format: nn[KMGTPE] | nn% 4184 This parameter is the complement to kernelcore=, it 4185 specifies the amount of memory used for migratable 4186 allocations. If both kernelcore and movablecore is 4187 specified, then kernelcore will be at *least* the 4188 specified value but may be more. If movablecore on its 4189 own is specified, the administrator must be careful 4190 that the amount of memory usable for all allocations 4191 is not too small. 4192 4193 movable_node [KNL,EARLY] Boot-time switch to make hotplugable memory 4194 NUMA nodes to be movable. This means that the memory 4195 of such nodes will be usable only for movable 4196 allocations which rules out almost all kernel 4197 allocations. Use with caution! 4198 4199 MTD_Partition= [MTD] 4200 Format: <name>,<region-number>,<size>,<offset> 4201 4202 MTD_Region= [MTD] Format: 4203 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>] 4204 4205 mtdparts= [MTD] 4206 See drivers/mtd/parsers/cmdlinepart.c 4207 4208 mtouchusb.raw_coordinates= 4209 [HW] Make the MicroTouch USB driver use raw coordinates 4210 ('y', default) or cooked coordinates ('n') 4211 4212 mtrr=debug [X86,EARLY] 4213 Enable printing debug information related to MTRR 4214 registers at boot time. 4215 4216 mtrr_chunk_size=nn[KMG,X86,EARLY] 4217 used for mtrr cleanup. It is largest continuous chunk 4218 that could hold holes aka. UC entries. 4219 4220 mtrr_gran_size=nn[KMG,X86,EARLY] 4221 Used for mtrr cleanup. It is granularity of mtrr block. 4222 Default is 1. 4223 Large value could prevent small alignment from 4224 using up MTRRs. 4225 4226 mtrr_spare_reg_nr=n [X86,EARLY] 4227 Format: <integer> 4228 Range: 0,7 : spare reg number 4229 Default : 1 4230 Used for mtrr cleanup. It is spare mtrr entries number. 4231 Set to 2 or more if your graphical card needs more. 4232 4233 multitce=off [PPC] This parameter disables the use of the pSeries 4234 firmware feature for updating multiple TCE entries 4235 at a time. 4236 4237 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card 4238 4239 netdev= [NET] NE2000 ISA network devices parameters 4240 Format: <irq>,<io>,<mem_start>,<mem_end>,<name> 4241 4242 netpoll.carrier_timeout= 4243 [NET] Specifies amount of time (in seconds) that 4244 netpoll should wait for a carrier. By default netpoll 4245 waits 4 seconds. 4246 4247 nf_conntrack.acct= 4248 [NETFILTER] Enable connection tracking flow accounting 4249 0 to disable accounting 4250 1 to enable accounting 4251 Default value is 0. 4252 4253 nfs.cache_getent= 4254 [NFS] sets the pathname to the program which is used 4255 to update the NFS client cache entries. 4256 4257 nfs.cache_getent_timeout= 4258 [NFS] sets the timeout after which an attempt to 4259 update a cache entry is deemed to have failed. 4260 4261 nfs.callback_nr_threads= 4262 [NFSv4] set the total number of threads that the 4263 NFS client will assign to service NFSv4 callback 4264 requests. 4265 4266 nfs.callback_tcpport= 4267 [NFS] set the TCP port on which the NFSv4 callback 4268 channel should listen. 4269 4270 nfs.delay_retrans= 4271 [NFS] specifies the number of times the NFSv4 client 4272 retries the request before returning an EAGAIN error, 4273 after a reply of NFS4ERR_DELAY from the server. 4274 Only applies if the softerr mount option is enabled, 4275 and the specified value is >= 0. 4276 4277 nfs.enable_ino64= 4278 [NFS] enable 64-bit inode numbers. 4279 If zero, the NFS client will fake up a 32-bit inode 4280 number for the readdir() and stat() syscalls instead 4281 of returning the full 64-bit number. 4282 The default is to return 64-bit inode numbers. 4283 4284 nfs.idmap_cache_timeout= 4285 [NFS] set the maximum lifetime for idmapper cache 4286 entries. 4287 4288 nfs.max_session_cb_slots= 4289 [NFSv4.1] Sets the maximum number of session 4290 slots the client will assign to the callback 4291 channel. This determines the maximum number of 4292 callbacks the client will process in parallel for 4293 a particular server. 4294 4295 nfs.max_session_slots= 4296 [NFSv4.1] Sets the maximum number of session slots 4297 the client will attempt to negotiate with the server. 4298 This limits the number of simultaneous RPC requests 4299 that the client can send to the NFSv4.1 server. 4300 Note that there is little point in setting this 4301 value higher than the max_tcp_slot_table_limit. 4302 4303 nfs.nfs4_disable_idmapping= 4304 [NFSv4] When set to the default of '1', this option 4305 ensures that both the RPC level authentication 4306 scheme and the NFS level operations agree to use 4307 numeric uids/gids if the mount is using the 4308 'sec=sys' security flavour. In effect it is 4309 disabling idmapping, which can make migration from 4310 legacy NFSv2/v3 systems to NFSv4 easier. 4311 Servers that do not support this mode of operation 4312 will be autodetected by the client, and it will fall 4313 back to using the idmapper. 4314 To turn off this behaviour, set the value to '0'. 4315 4316 nfs.nfs4_unique_id= 4317 [NFS4] Specify an additional fixed unique ident- 4318 ification string that NFSv4 clients can insert into 4319 their nfs_client_id4 string. This is typically a 4320 UUID that is generated at system install time. 4321 4322 nfs.recover_lost_locks= 4323 [NFSv4] Attempt to recover locks that were lost due 4324 to a lease timeout on the server. Please note that 4325 doing this risks data corruption, since there are 4326 no guarantees that the file will remain unchanged 4327 after the locks are lost. 4328 If you want to enable the kernel legacy behaviour of 4329 attempting to recover these locks, then set this 4330 parameter to '1'. 4331 The default parameter value of '0' causes the kernel 4332 not to attempt recovery of lost locks. 4333 4334 nfs.send_implementation_id= 4335 [NFSv4.1] Send client implementation identification 4336 information in exchange_id requests. 4337 If zero, no implementation identification information 4338 will be sent. 4339 The default is to send the implementation identification 4340 information. 4341 4342 nfs4.layoutstats_timer= 4343 [NFSv4.2] Change the rate at which the kernel sends 4344 layoutstats to the pNFS metadata server. 4345 4346 Setting this to value to 0 causes the kernel to use 4347 whatever value is the default set by the layout 4348 driver. A non-zero value sets the minimum interval 4349 in seconds between layoutstats transmissions. 4350 4351 nfsd.inter_copy_offload_enable= 4352 [NFSv4.2] When set to 1, the server will support 4353 server-to-server copies for which this server is 4354 the destination of the copy. 4355 4356 nfsd.nfs4_disable_idmapping= 4357 [NFSv4] When set to the default of '1', the NFSv4 4358 server will return only numeric uids and gids to 4359 clients using auth_sys, and will accept numeric uids 4360 and gids from such clients. This is intended to ease 4361 migration from NFSv2/v3. 4362 4363 nfsd.nfsd4_ssc_umount_timeout= 4364 [NFSv4.2] When used as the destination of a 4365 server-to-server copy, knfsd temporarily mounts 4366 the source server. It caches the mount in case 4367 it will be needed again, and discards it if not 4368 used for the number of milliseconds specified by 4369 this parameter. 4370 4371 nfsaddrs= [NFS] Deprecated. Use ip= instead. 4372 See Documentation/admin-guide/nfs/nfsroot.rst. 4373 4374 nfsroot= [NFS] nfs root filesystem for disk-less boxes. 4375 See Documentation/admin-guide/nfs/nfsroot.rst. 4376 4377 nfsrootdebug [NFS] enable nfsroot debugging messages. 4378 See Documentation/admin-guide/nfs/nfsroot.rst. 4379 4380 nmi_backtrace.backtrace_idle [KNL] 4381 Dump stacks even of idle CPUs in response to an 4382 NMI stack-backtrace request. 4383 4384 nmi_debug= [KNL,SH] Specify one or more actions to take 4385 when a NMI is triggered. 4386 Format: [state][,regs][,debounce][,die] 4387 4388 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels 4389 Format: [panic,][nopanic,][rNNN,][num] 4390 Valid num: 0 or 1 4391 0 - turn hardlockup detector in nmi_watchdog off 4392 1 - turn hardlockup detector in nmi_watchdog on 4393 rNNN - configure the watchdog with raw perf event 0xNNN 4394 4395 When panic is specified, panic when an NMI watchdog 4396 timeout occurs (or 'nopanic' to not panic on an NMI 4397 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set) 4398 To disable both hard and soft lockup detectors, 4399 please see 'nowatchdog'. 4400 This is useful when you use a panic=... timeout and 4401 need the box quickly up again. 4402 4403 These settings can be accessed at runtime via 4404 the nmi_watchdog and hardlockup_panic sysctls. 4405 4406 no4lvl [RISCV,EARLY] Disable 4-level and 5-level paging modes. 4407 Forces kernel to use 3-level paging instead. 4408 4409 no5lvl [X86-64,RISCV,EARLY] Disable 5-level paging mode. Forces 4410 kernel to use 4-level paging instead. 4411 4412 noalign [KNL,ARM] 4413 4414 noapic [SMP,APIC,EARLY] Tells the kernel to not make use of any 4415 IOAPICs that may be present in the system. 4416 4417 noapictimer [APIC,X86] Don't set up the APIC timer 4418 4419 noautogroup Disable scheduler automatic task group creation. 4420 4421 nocache [ARM,EARLY] 4422 4423 no_console_suspend 4424 [HW] Never suspend the console 4425 Disable suspending of consoles during suspend and 4426 hibernate operations. Once disabled, debugging 4427 messages can reach various consoles while the rest 4428 of the system is being put to sleep (ie, while 4429 debugging driver suspend/resume hooks). This may 4430 not work reliably with all consoles, but is known 4431 to work with serial and VGA consoles. 4432 To facilitate more flexible debugging, we also add 4433 console_suspend, a printk module parameter to control 4434 it. Users could use console_suspend (usually 4435 /sys/module/printk/parameters/console_suspend) to 4436 turn on/off it dynamically. 4437 4438 no_debug_objects 4439 [KNL,EARLY] Disable object debugging 4440 4441 nodsp [SH] Disable hardware DSP at boot time. 4442 4443 noefi [EFI,EARLY] Disable EFI runtime services support. 4444 4445 no_entry_flush [PPC,EARLY] Don't flush the L1-D cache when entering the kernel. 4446 4447 noexec32 [X86-64] 4448 This affects only 32-bit executables. 4449 noexec32=on: enable non-executable mappings (default) 4450 read doesn't imply executable mappings 4451 noexec32=off: disable non-executable mappings 4452 read implies executable mappings 4453 4454 no_file_caps Tells the kernel not to honor file capabilities. The 4455 only way then for a file to be executed with privilege 4456 is to be setuid root or executed by root. 4457 4458 nofpu [MIPS,SH] Disable hardware FPU at boot time. 4459 4460 nofsgsbase [X86] Disables FSGSBASE instructions. 4461 4462 nofxsr [BUGS=X86-32] Disables x86 floating point extended 4463 register save and restore. The kernel will only save 4464 legacy floating-point registers on task switch. 4465 4466 nogbpages [X86] Do not use GB pages for kernel direct mappings. 4467 4468 no_hash_pointers 4469 [KNL,EARLY] 4470 Alias for "hash_pointers=never". 4471 4472 nohibernate [HIBERNATION] Disable hibernation and resume. 4473 4474 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,RISCV,SH] Forces the kernel to 4475 busy wait in do_idle() and not use the arch_cpu_idle() 4476 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP 4477 to be effective. This is useful on platforms where the 4478 sleep(SH) or wfi(ARM,ARM64) instructions do not work 4479 correctly or when doing power measurements to evaluate 4480 the impact of the sleep instructions. This is also 4481 useful when using JTAG debugger. 4482 4483 nohpet [X86] Don't use the HPET timer. 4484 4485 nohugeiomap [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge I/O mappings. 4486 4487 nohugevmalloc [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge vmalloc mappings. 4488 4489 nohz= [KNL] Boottime enable/disable dynamic ticks 4490 Valid arguments: on, off 4491 Default: on 4492 4493 nohz_full= [KNL,BOOT,SMP,ISOL] 4494 The argument is a cpu list, as described above. 4495 In kernels built with CONFIG_NO_HZ_FULL=y, set 4496 the specified list of CPUs whose tick will be stopped 4497 whenever possible. The boot CPU will be forced outside 4498 the range to maintain the timekeeping. Any CPUs 4499 in this list will have their RCU callbacks offloaded, 4500 just as if they had also been called out in the 4501 rcu_nocbs= boot parameter. 4502 4503 Note that this argument takes precedence over 4504 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 4505 4506 noinitrd [Deprecated,RAM] Tells the kernel not to load any configured 4507 initial RAM disk. Currently this parameter applies to 4508 initrd only, not to initramfs. But it applies to both 4509 in EFI mode. 4510 4511 nointremap [X86-64,Intel-IOMMU,EARLY] Do not enable interrupt 4512 remapping. 4513 [Deprecated - use intremap=off] 4514 4515 noinvpcid [X86,EARLY] Disable the INVPCID cpu feature. 4516 4517 noiotrap [SH] Disables trapped I/O port accesses. 4518 4519 noirqdebug [X86-32] Disables the code which attempts to detect and 4520 disable unhandled interrupt sources. 4521 4522 noisapnp [ISAPNP] Disables ISA PnP code. 4523 4524 nokaslr [KNL,EARLY] 4525 When CONFIG_RANDOMIZE_BASE is set, this disables 4526 kernel and module base offset ASLR (Address Space 4527 Layout Randomization). 4528 4529 no-kvmapf [X86,KVM,EARLY] Disable paravirtualized asynchronous page 4530 fault handling. 4531 4532 no-kvmclock [X86,KVM,EARLY] Disable paravirtualized KVM clock driver 4533 4534 nolapic [X86-32,APIC,EARLY] Do not enable or use the local APIC. 4535 4536 nolapic_timer [X86-32,APIC,EARLY] Do not use the local APIC timer. 4537 4538 nomce [X86-32] Disable Machine Check Exception 4539 4540 nomfgpt [X86-32] Disable Multi-Function General Purpose 4541 Timer usage (for AMD Geode machines). 4542 4543 nomodeset Disable kernel modesetting. Most systems' firmware 4544 sets up a display mode and provides framebuffer memory 4545 for output. With nomodeset, DRM and fbdev drivers will 4546 not load if they could possibly displace the pre- 4547 initialized output. Only the system framebuffer will 4548 be available for use. The respective drivers will not 4549 perform display-mode changes or accelerated rendering. 4550 4551 Useful as error fallback, or for testing and debugging. 4552 4553 nomodule Disable module load 4554 4555 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to 4556 shutdown the other cpus. Instead use the REBOOT_VECTOR 4557 irq. 4558 4559 nopat [X86,EARLY] Disable PAT (page attribute table extension of 4560 pagetables) support. 4561 4562 nopcid [X86-64,EARLY] Disable the PCID cpu feature. 4563 4564 nopku [X86] Disable Memory Protection Keys CPU feature found 4565 in some Intel CPUs. 4566 4567 nopti [X86-64,EARLY] 4568 Equivalent to pti=off 4569 4570 nopv= [X86,XEN,KVM,HYPER_V,VMWARE,EARLY] 4571 Disables the PV optimizations forcing the guest to run 4572 as generic guest with no PV drivers. Currently support 4573 XEN HVM, KVM, HYPER_V and VMWARE guest. 4574 4575 nopvspin [X86,XEN,KVM,EARLY] 4576 Disables the qspinlock slow path using PV optimizations 4577 which allow the hypervisor to 'idle' the guest on lock 4578 contention. 4579 4580 norandmaps Don't use address space randomization. Equivalent to 4581 echo 0 > /proc/sys/kernel/randomize_va_space 4582 4583 noreplace-smp [X86-32,SMP] Don't replace SMP instructions 4584 with UP alternatives 4585 4586 noresume [SWSUSP] Disables resume and restores original swap 4587 space. 4588 4589 no-scroll [VGA] Disables scrollback. 4590 This is required for the Braillex ib80-piezo Braille 4591 reader made by F.H. Papenmeier (Germany). 4592 4593 nosgx [X86-64,SGX,EARLY] Disables Intel SGX kernel support. 4594 4595 nosmap [PPC,EARLY] 4596 Disable SMAP (Supervisor Mode Access Prevention) 4597 even if it is supported by processor. 4598 4599 nosmep [PPC64s,EARLY] 4600 Disable SMEP (Supervisor Mode Execution Prevention) 4601 even if it is supported by processor. 4602 4603 nosmp [SMP,EARLY] Tells an SMP kernel to act as a UP kernel, 4604 and disable the IO APIC. legacy for "maxcpus=0". 4605 4606 nosmt [KNL,MIPS,PPC,EARLY] Disable symmetric multithreading (SMT). 4607 Equivalent to smt=1. 4608 4609 [KNL,LOONGARCH,X86,ARM64,PPC,S390] Disable symmetric multithreading (SMT). 4610 nosmt=force: Force disable SMT, cannot be undone 4611 via the sysfs control file. 4612 4613 nosoftlockup [KNL] Disable the soft-lockup detector. 4614 4615 nospec_store_bypass_disable 4616 [HW,EARLY] Disable all mitigations for the Speculative 4617 Store Bypass vulnerability 4618 4619 nospectre_bhb [ARM64,EARLY] Disable all mitigations for Spectre-BHB (branch 4620 history injection) vulnerability. System may allow data leaks 4621 with this option. 4622 4623 nospectre_v1 [X86,PPC,EARLY] Disable mitigations for Spectre Variant 1 4624 (bounds check bypass). With this option data leaks are 4625 possible in the system. 4626 4627 nospectre_v2 [X86,PPC_E500,ARM64,EARLY] Disable all mitigations 4628 for the Spectre variant 2 (indirect branch 4629 prediction) vulnerability. System may allow data 4630 leaks with this option. 4631 4632 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES,RISCV,LOONGARCH,EARLY] 4633 Disable paravirtualized steal time accounting. steal time 4634 is computed, but won't influence scheduler behaviour 4635 4636 nosync [HW,M68K] Disables sync negotiation for all devices. 4637 4638 no_timer_check [X86,APIC] Disables the code which tests for broken 4639 timer IRQ sources, i.e., the IO-APIC timer. This can 4640 work around problems with incorrect timer 4641 initialization on some boards. 4642 4643 no_uaccess_flush 4644 [PPC,EARLY] Don't flush the L1-D cache after accessing user data. 4645 4646 novmcoredd [KNL,KDUMP] 4647 Disable device dump. Device dump allows drivers to 4648 append dump data to vmcore so you can collect driver 4649 specified debug info. Drivers can append the data 4650 without any limit and this data is stored in memory, 4651 so this may cause significant memory stress. Disabling 4652 device dump can help save memory but the driver debug 4653 data will be no longer available. This parameter 4654 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP 4655 is set. 4656 4657 no-vmw-sched-clock 4658 [X86,PV_OPS,EARLY] Disable paravirtualized VMware 4659 scheduler clock and use the default one. 4660 4661 nowatchdog [KNL] Disable both lockup detectors, i.e. 4662 soft-lockup and NMI watchdog (hard-lockup). 4663 4664 nowb [ARM,EARLY] 4665 4666 nox2apic [X86-64,APIC,EARLY] Do not enable x2APIC mode. 4667 4668 NOTE: this parameter will be ignored on systems with the 4669 LEGACY_XAPIC_DISABLED bit set in the 4670 IA32_XAPIC_DISABLE_STATUS MSR. 4671 4672 noxsave [BUGS=X86] Disables x86 extended register state save 4673 and restore using xsave. The kernel will fallback to 4674 enabling legacy floating-point and sse state. 4675 4676 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended 4677 register states. The kernel will fall back to use 4678 xsave to save the states. By using this parameter, 4679 performance of saving the states is degraded because 4680 xsave doesn't support modified optimization while 4681 xsaveopt supports it on xsaveopt enabled systems. 4682 4683 noxsaves [X86] Disables xsaves and xrstors used in saving and 4684 restoring x86 extended register state in compacted 4685 form of xsave area. The kernel will fall back to use 4686 xsaveopt and xrstor to save and restore the states 4687 in standard form of xsave area. By using this 4688 parameter, xsave area per process might occupy more 4689 memory on xsaves enabled systems. 4690 4691 nr_cpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 4692 could support. nr_cpus=n : n >= 1 limits the kernel to 4693 support 'n' processors. It could be larger than the 4694 number of already plugged CPU during bootup, later in 4695 runtime you can physically add extra cpu until it reaches 4696 n. So during boot up some boot time memory for per-cpu 4697 variables need be pre-allocated for later physical cpu 4698 hot plugging. 4699 4700 nr_uarts= [SERIAL] maximum number of UARTs to be registered. 4701 4702 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86, EARLY] 4703 Disable NUMA, Only set up a single NUMA node 4704 spanning all memory. 4705 4706 numa=fake=<size>[MG] 4707 [KNL, ARM64, RISCV, X86, EARLY] 4708 If given as a memory unit, fills all system RAM with 4709 nodes of size interleaved over physical nodes. 4710 4711 numa=fake=<N> 4712 [KNL, ARM64, RISCV, X86, EARLY] 4713 If given as an integer, fills all system RAM with N 4714 fake nodes interleaved over physical nodes. 4715 4716 numa=fake=<N>U 4717 [KNL, ARM64, RISCV, X86, EARLY] 4718 If given as an integer followed by 'U', it will 4719 divide each physical node into N emulated nodes. 4720 4721 numa=noacpi [X86] Don't parse the SRAT table for NUMA setup 4722 4723 numa=nohmat [X86] Don't parse the HMAT table for NUMA setup, or 4724 soft-reserved memory partitioning. 4725 4726 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic 4727 NUMA balancing. 4728 Allowed values are enable and disable 4729 4730 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA. 4731 'node', 'default' can be specified 4732 This can be set from sysctl after boot. 4733 See Documentation/admin-guide/sysctl/vm.rst for details. 4734 4735 nvme.quirks= [NVME] A list of quirk entries to augment the built-in 4736 nvme quirk list. List entries are separated by a 4737 '-' character. 4738 Each entry has the form VendorID:ProductID:quirk_names. 4739 The IDs are 4-digits hex numbers and quirk_names is a 4740 list of quirk names separated by commas. A quirk name 4741 can be prefixed by '^', meaning that the specified 4742 quirk must be disabled. 4743 4744 Example: 4745 nvme.quirks=7710:2267:bogus_nid,^identify_cns-9900:7711:broken_msi 4746 4747 ohci1394_dma=early [HW,EARLY] enable debugging via the ohci1394 driver. 4748 See Documentation/core-api/debugging-via-ohci1394.rst for more 4749 info. 4750 4751 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands 4752 Rather than timing out after 20 ms if an EC 4753 command is not properly ACKed, override the length 4754 of the timeout. We have interrupts disabled while 4755 waiting for the ACK, so if this is set too high 4756 interrupts *may* be lost! 4757 4758 omap_mux= [OMAP] Override bootloader pin multiplexing. 4759 Format: <mux_mode0.mode_name=value>... 4760 For example, to override I2C bus2: 4761 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100 4762 4763 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration 4764 4765 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock] 4766 4767 boundary - index of last SLC block on Flex-OneNAND. 4768 The remaining blocks are configured as MLC blocks. 4769 lock - Configure if Flex-OneNAND boundary should be locked. 4770 Once locked, the boundary cannot be changed. 4771 1 indicates lock status, 0 indicates unlock status. 4772 4773 oops=panic [KNL,EARLY] 4774 Always panic on oopses. Default is to just kill the 4775 process, but there is a small probability of 4776 deadlocking the machine. 4777 This will also cause panics on machine check exceptions. 4778 Useful together with panic=30 to trigger a reboot. 4779 4780 page_alloc.shuffle= 4781 [KNL] Boolean flag to control whether the page allocator 4782 should randomize its free lists. This parameter can be 4783 used to enable/disable page randomization. The state of 4784 the flag can be read from sysfs at: 4785 /sys/module/page_alloc/parameters/shuffle. 4786 This parameter is only available if CONFIG_SHUFFLE_PAGE_ALLOCATOR=y. 4787 4788 page_owner= [KNL,EARLY] Boot-time page_owner enabling option. 4789 Storage of the information about who allocated 4790 each page is disabled in default. With this switch, 4791 we can turn it on. 4792 on: enable the feature 4793 4794 page_poison= [KNL,EARLY] Boot-time parameter changing the state of 4795 poisoning on the buddy allocator, available with 4796 CONFIG_PAGE_POISONING=y. 4797 off: turn off poisoning (default) 4798 on: turn on poisoning 4799 4800 page_reporting.page_reporting_order= 4801 [KNL] Minimal page reporting order 4802 Format: <integer> 4803 Adjust the minimal page reporting order. The page 4804 reporting is disabled when it exceeds MAX_PAGE_ORDER. 4805 4806 panic= [KNL] Kernel behaviour on panic: delay <timeout> 4807 timeout > 0: seconds before rebooting 4808 timeout = 0: wait forever 4809 timeout < 0: reboot immediately 4810 Format: <timeout> 4811 4812 panic_on_taint= [KNL,EARLY] 4813 Bitmask for conditionally calling panic() in add_taint() 4814 Format: <hex>[,nousertaint] 4815 Hexadecimal bitmask representing the set of TAINT flags 4816 that will cause the kernel to panic when add_taint() is 4817 called with any of the flags in this set. 4818 The optional switch "nousertaint" can be utilized to 4819 prevent userspace forced crashes by writing to sysctl 4820 /proc/sys/kernel/tainted any flagset matching with the 4821 bitmask set on panic_on_taint. 4822 See Documentation/admin-guide/tainted-kernels.rst for 4823 extra details on the taint flags that users can pick 4824 to compose the bitmask to assign to panic_on_taint. 4825 4826 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump 4827 on a WARN(). 4828 4829 panic_force_cpu= 4830 [KNL,SMP] Force panic handling to execute on a specific CPU. 4831 Format: <cpu number> 4832 Some platforms require panic handling to occur on a 4833 specific CPU for the crash kernel to function correctly. 4834 This can be due to firmware limitations, interrupt routing 4835 constraints, or platform-specific requirements where only 4836 a particular CPU can safely enter the crash kernel. 4837 When set, panic() will redirect execution to the specified 4838 CPU before proceeding with the normal panic and kexec flow. 4839 If the target CPU is offline or unavailable, panic proceeds 4840 on the current CPU. 4841 This option should only be used for systems with the above 4842 constraints as it might cause the panic operation to be less reliable. 4843 4844 panic_print= Bitmask for printing system info when panic happens. 4845 User can chose combination of the following bits: 4846 bit 0: print all tasks info 4847 bit 1: print system memory info 4848 bit 2: print timer info 4849 bit 3: print locks info if CONFIG_LOCKDEP is on 4850 bit 4: print ftrace buffer 4851 bit 5: replay all kernel messages on consoles at the end of panic 4852 bit 6: print all CPUs backtrace (if available in the arch) 4853 bit 7: print only tasks in uninterruptible (blocked) state 4854 *Be aware* that this option may print a _lot_ of lines, 4855 so there are risks of losing older messages in the log. 4856 Use this option carefully, maybe worth to setup a 4857 bigger log buffer with "log_buf_len" along with this. 4858 4859 panic_sys_info= A comma separated list of extra information to be dumped 4860 on panic. 4861 Format: val[,val...] 4862 Where @val can be any of the following: 4863 4864 tasks: print all tasks info 4865 mem: print system memory info 4866 timers: print timers info 4867 locks: print locks info if CONFIG_LOCKDEP is on 4868 ftrace: print ftrace buffer 4869 all_bt: print all CPUs backtrace (if available in the arch) 4870 blocked_tasks: print only tasks in uninterruptible (blocked) state 4871 4872 This is a human readable alternative to the 'panic_print' option. 4873 4874 panic_console_replay 4875 When panic happens, replay all kernel messages on 4876 consoles at the end of panic. 4877 4878 parkbd.port= [HW] Parallel port number the keyboard adapter is 4879 connected to, default is 0. 4880 Format: <parport#> 4881 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation, 4882 0 for XT, 1 for AT (default is AT). 4883 Format: <mode> 4884 4885 parport= [HW,PPT] Specify parallel ports. 0 disables. 4886 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] } 4887 Use 'auto' to force the driver to use any 4888 IRQ/DMA settings detected (the default is to 4889 ignore detected IRQ/DMA settings because of 4890 possible conflicts). You can specify the base 4891 address, IRQ, and DMA settings; IRQ and DMA 4892 should be numbers, or 'auto' (for using detected 4893 settings on that particular port), or 'nofifo' 4894 (to avoid using a FIFO even if it is detected). 4895 Parallel ports are assigned in the order they 4896 are specified on the command line, starting 4897 with parport0. 4898 4899 parport_init_mode= [HW,PPT] 4900 Configure VIA parallel port to operate in 4901 a specific mode. This is necessary on Pegasos 4902 computer where firmware has no options for setting 4903 up parallel port mode and sets it to spp. 4904 Currently this function knows 686a and 8231 chips. 4905 Format: [spp|ps2|epp|ecp|ecpepp] 4906 4907 pata_legacy.all= [HW,LIBATA] 4908 Format: <int> 4909 Set to non-zero to probe primary and secondary ISA 4910 port ranges on PCI systems where no PCI PATA device 4911 has been found at either range. Disabled by default. 4912 4913 pata_legacy.autospeed= [HW,LIBATA] 4914 Format: <int> 4915 Set to non-zero if a chip is present that snoops speed 4916 changes. Disabled by default. 4917 4918 pata_legacy.ht6560a= [HW,LIBATA] 4919 Format: <int> 4920 Set to 1, 2, or 3 for HT 6560A on the primary channel, 4921 the secondary channel, or both channels respectively. 4922 Disabled by default. 4923 4924 pata_legacy.ht6560b= [HW,LIBATA] 4925 Format: <int> 4926 Set to 1, 2, or 3 for HT 6560B on the primary channel, 4927 the secondary channel, or both channels respectively. 4928 Disabled by default. 4929 4930 pata_legacy.iordy_mask= [HW,LIBATA] 4931 Format: <int> 4932 IORDY enable mask. Set individual bits to allow IORDY 4933 for the respective channel. Bit 0 is for the first 4934 legacy channel handled by this driver, bit 1 is for 4935 the second channel, and so on. The sequence will often 4936 correspond to the primary legacy channel, the secondary 4937 legacy channel, and so on, but the handling of a PCI 4938 bus and the use of other driver options may interfere 4939 with the sequence. By default IORDY is allowed across 4940 all channels. 4941 4942 pata_legacy.opti82c46x= [HW,LIBATA] 4943 Format: <int> 4944 Set to 1, 2, or 3 for Opti 82c611A on the primary 4945 channel, the secondary channel, or both channels 4946 respectively. Disabled by default. 4947 4948 pata_legacy.opti82c611a= [HW,LIBATA] 4949 Format: <int> 4950 Set to 1, 2, or 3 for Opti 82c465MV on the primary 4951 channel, the secondary channel, or both channels 4952 respectively. Disabled by default. 4953 4954 pata_legacy.pio_mask= [HW,LIBATA] 4955 Format: <int> 4956 PIO mode mask for autospeed devices. Set individual 4957 bits to allow the use of the respective PIO modes. 4958 Bit 0 is for mode 0, bit 1 is for mode 1, and so on. 4959 All modes allowed by default. 4960 4961 pata_legacy.probe_all= [HW,LIBATA] 4962 Format: <int> 4963 Set to non-zero to probe tertiary and further ISA 4964 port ranges on PCI systems. Disabled by default. 4965 4966 pata_legacy.probe_mask= [HW,LIBATA] 4967 Format: <int> 4968 Probe mask for legacy ISA PATA ports. Depending on 4969 platform configuration and the use of other driver 4970 options up to 6 legacy ports are supported: 0x1f0, 4971 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing 4972 of individual ports can be disabled by setting the 4973 corresponding bits in the mask to 1. Bit 0 is for 4974 the first port in the list above (0x1f0), and so on. 4975 By default all supported ports are probed. 4976 4977 pata_legacy.qdi= [HW,LIBATA] 4978 Format: <int> 4979 Set to non-zero to probe QDI controllers. By default 4980 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise. 4981 4982 pata_legacy.winbond= [HW,LIBATA] 4983 Format: <int> 4984 Set to non-zero to probe Winbond controllers. Use 4985 the standard I/O port (0x130) if 1, otherwise the 4986 value given is the I/O port to use (typically 0x1b0). 4987 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE, 4988 0 otherwise. 4989 4990 pata_platform.pio_mask= [HW,LIBATA] 4991 Format: <int> 4992 Supported PIO mode mask. Set individual bits to allow 4993 the use of the respective PIO modes. Bit 0 is for 4994 mode 0, bit 1 is for mode 1, and so on. Mode 0 only 4995 allowed by default. 4996 4997 pause_on_oops=<int> 4998 Halt all CPUs after the first oops has been printed for 4999 the specified number of seconds. This is to be used if 5000 your oopses keep scrolling off the screen. 5001 5002 pci=option[,option...] [PCI,EARLY] various PCI subsystem options. 5003 5004 Some options herein operate on a specific device 5005 or a set of devices (<pci_dev>). These are 5006 specified in one of the following formats: 5007 5008 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]* 5009 pci:<vendor>:<device>[:<subvendor>:<subdevice>] 5010 5011 Note: the first format specifies a PCI 5012 bus/device/function address which may change 5013 if new hardware is inserted, if motherboard 5014 firmware changes, or due to changes caused 5015 by other kernel parameters. If the 5016 domain is left unspecified, it is 5017 taken to be zero. Optionally, a path 5018 to a device through multiple device/function 5019 addresses can be specified after the base 5020 address (this is more robust against 5021 renumbering issues). The second format 5022 selects devices using IDs from the 5023 configuration space which may match multiple 5024 devices in the system. 5025 5026 earlydump dump PCI config space before the kernel 5027 changes anything 5028 off [X86] don't probe for the PCI bus 5029 bios [X86-32] force use of PCI BIOS, don't access 5030 the hardware directly. Use this if your machine 5031 has a non-standard PCI host bridge. 5032 nobios [X86-32] disallow use of PCI BIOS, only direct 5033 hardware access methods are allowed. Use this 5034 if you experience crashes upon bootup and you 5035 suspect they are caused by the BIOS. 5036 conf1 [X86] Force use of PCI Configuration Access 5037 Mechanism 1 (config address in IO port 0xCF8, 5038 data in IO port 0xCFC, both 32-bit). 5039 conf2 [X86] Force use of PCI Configuration Access 5040 Mechanism 2 (IO port 0xCF8 is an 8-bit port for 5041 the function, IO port 0xCFA, also 8-bit, sets 5042 bus number. The config space is then accessed 5043 through ports 0xC000-0xCFFF). 5044 See http://wiki.osdev.org/PCI for more info 5045 on the configuration access mechanisms. 5046 noaer [PCIE] If the PCIEAER kernel config parameter is 5047 enabled, this kernel boot option can be used to 5048 disable the use of PCIE advanced error reporting. 5049 nodomains [PCI] Disable support for multiple PCI 5050 root domains (aka PCI segments, in ACPI-speak). 5051 nommconf [X86] Disable use of MMCONFIG for PCI 5052 Configuration 5053 check_enable_amd_mmconf [X86] check for and enable 5054 properly configured MMIO access to PCI 5055 config space on AMD family 10h CPU 5056 nomsi [MSI] If the PCI_MSI kernel config parameter is 5057 enabled, this kernel boot option can be used to 5058 disable the use of MSI interrupts system-wide. 5059 noioapicquirk [APIC] Disable all boot interrupt quirks. 5060 Safety option to keep boot IRQs enabled. This 5061 should never be necessary. 5062 ioapicreroute [APIC] Enable rerouting of boot IRQs to the 5063 primary IO-APIC for bridges that cannot disable 5064 boot IRQs. This fixes a source of spurious IRQs 5065 when the system masks IRQs. 5066 noioapicreroute [APIC] Disable workaround that uses the 5067 boot IRQ equivalent of an IRQ that connects to 5068 a chipset where boot IRQs cannot be disabled. 5069 The opposite of ioapicreroute. 5070 biosirq [X86-32] Use PCI BIOS calls to get the interrupt 5071 routing table. These calls are known to be buggy 5072 on several machines and they hang the machine 5073 when used, but on other computers it's the only 5074 way to get the interrupt routing table. Try 5075 this option if the kernel is unable to allocate 5076 IRQs or discover secondary PCI buses on your 5077 motherboard. 5078 rom [X86] Assign address space to expansion ROMs. 5079 Use with caution as certain devices share 5080 address decoders between ROMs and other 5081 resources. 5082 norom [X86] Do not assign address space to 5083 expansion ROMs that do not already have 5084 BIOS assigned address ranges. 5085 nobar [X86] Do not assign address space to the 5086 BARs that weren't assigned by the BIOS. 5087 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be 5088 assigned automatically to PCI devices. You can 5089 make the kernel exclude IRQs of your ISA cards 5090 this way. 5091 pirqaddr=0xAAAAA [X86] Specify the physical address 5092 of the PIRQ table (normally generated 5093 by the BIOS) if it is outside the 5094 F0000h-100000h range. 5095 lastbus=N [X86] Scan all buses thru bus #N. Can be 5096 useful if the kernel is unable to find your 5097 secondary buses and you want to tell it 5098 explicitly which ones they are. 5099 assign-busses [X86] Always assign all PCI bus 5100 numbers ourselves, overriding 5101 whatever the firmware may have done. 5102 usepirqmask [X86] Honor the possible IRQ mask stored 5103 in the BIOS $PIR table. This is needed on 5104 some systems with broken BIOSes, notably 5105 some HP Pavilion N5400 and Omnibook XE3 5106 notebooks. This will have no effect if ACPI 5107 IRQ routing is enabled. 5108 noacpi [X86] Do not use ACPI for IRQ routing 5109 or for PCI scanning. 5110 use_crs [X86] Use PCI host bridge window information 5111 from ACPI. On BIOSes from 2008 or later, this 5112 is enabled by default. If you need to use this, 5113 please report a bug. 5114 nocrs [X86] Ignore PCI host bridge windows from ACPI. 5115 If you need to use this, please report a bug. 5116 use_e820 [X86] Use E820 reservations to exclude parts of 5117 PCI host bridge windows. This is a workaround 5118 for BIOS defects in host bridge _CRS methods. 5119 If you need to use this, please report a bug to 5120 <linux-pci@vger.kernel.org>. 5121 no_e820 [X86] Ignore E820 reservations for PCI host 5122 bridge windows. This is the default on modern 5123 hardware. If you need to use this, please report 5124 a bug to <linux-pci@vger.kernel.org>. 5125 routeirq Do IRQ routing for all PCI devices. 5126 This is normally done in pci_enable_device(), 5127 so this option is a temporary workaround 5128 for broken drivers that don't call it. 5129 skip_isa_align [X86] do not align io start addr, so can 5130 handle more pci cards 5131 noearly [X86] Don't do any early type 1 scanning. 5132 This might help on some broken boards which 5133 machine check when some devices' config space 5134 is read. But various workarounds are disabled 5135 and some IOMMU drivers will not work. 5136 bfsort Sort PCI devices into breadth-first order. 5137 This sorting is done to get a device 5138 order compatible with older (<= 2.4) kernels. 5139 nobfsort Don't sort PCI devices into breadth-first order. 5140 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size) 5141 tuning and use the BIOS-configured MPS defaults. 5142 pcie_bus_safe Set every device's MPS to the largest value 5143 supported by all devices below the root complex. 5144 pcie_bus_perf Set device MPS to the largest allowable MPS 5145 based on its parent bus. Also set MRRS (Max 5146 Read Request Size) to the largest supported 5147 value (no larger than the MPS that the device 5148 or bus can support) for best performance. 5149 pcie_bus_peer2peer Set every device's MPS to 128B, which 5150 every device is guaranteed to support. This 5151 configuration allows peer-to-peer DMA between 5152 any pair of devices, possibly at the cost of 5153 reduced performance. This also guarantees 5154 that hot-added devices will work. 5155 cbiosize=nn[KMG] The fixed amount of bus space which is 5156 reserved for the CardBus bridge's IO window. 5157 The default value is 256 bytes. 5158 cbmemsize=nn[KMG] The fixed amount of bus space which is 5159 reserved for the CardBus bridge's memory 5160 window. The default value is 64 megabytes. 5161 resource_alignment= 5162 Format: 5163 [<order of align>@]<pci_dev>[; ...] 5164 Specifies alignment and device to reassign 5165 aligned memory resources. How to 5166 specify the device is described above. 5167 If <order of align> is not specified, 5168 PAGE_SIZE is used as alignment. 5169 A PCI-PCI bridge can be specified if resource 5170 windows need to be expanded. 5171 To specify the alignment for several 5172 instances of a device, the PCI vendor, 5173 device, subvendor, and subdevice may be 5174 specified, e.g., 12@pci:8086:9c22:103c:198f 5175 for 4096-byte alignment. 5176 ecrc= Enable/disable PCIe ECRC (transaction layer 5177 end-to-end CRC checking). Only effective if 5178 OS has native AER control (either granted by 5179 ACPI _OSC or forced via "pcie_ports=native") 5180 bios: Use BIOS/firmware settings. This is the 5181 the default. 5182 off: Turn ECRC off 5183 on: Turn ECRC on. 5184 hpiosize=nn[KMG] The fixed amount of bus space which is 5185 reserved for hotplug bridge's IO window. 5186 Default size is 256 bytes. 5187 hpmmiosize=nn[KMG] The fixed amount of bus space which is 5188 reserved for hotplug bridge's MMIO window. 5189 Default size is 2 megabytes. 5190 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is 5191 reserved for hotplug bridge's MMIO_PREF window. 5192 Default size is 2 megabytes. 5193 hpmemsize=nn[KMG] The fixed amount of bus space which is 5194 reserved for hotplug bridge's MMIO and 5195 MMIO_PREF window. 5196 Default size is 2 megabytes. 5197 hpbussize=nn The minimum amount of additional bus numbers 5198 reserved for buses below a hotplug bridge. 5199 Default is 1. 5200 realloc= Enable/disable reallocating PCI bridge resources 5201 if allocations done by BIOS are too small to 5202 accommodate resources required by all child 5203 devices. 5204 off: Turn realloc off 5205 on: Turn realloc on 5206 realloc same as realloc=on 5207 noari do not use PCIe ARI. 5208 noats [PCIE, Intel-IOMMU, AMD-IOMMU] 5209 do not use PCIe ATS (and IOMMU device IOTLB). 5210 pcie_scan_all Scan all possible PCIe devices. Otherwise we 5211 only look for one device below a PCIe downstream 5212 port. 5213 big_root_window Try to add a big 64bit memory window to the PCIe 5214 root complex on AMD CPUs. Some GFX hardware 5215 can resize a BAR to allow access to all VRAM. 5216 Adding the window is slightly risky (it may 5217 conflict with unreported devices), so this 5218 taints the kernel. 5219 disable_acs_redir=<pci_dev>[; ...] 5220 Specify one or more PCI devices (in the format 5221 specified above) separated by semicolons. 5222 Each device specified will have the PCI ACS 5223 redirect capabilities forced off which will 5224 allow P2P traffic between devices through 5225 bridges without forcing it upstream. Note: 5226 this removes isolation between devices and 5227 may put more devices in an IOMMU group. 5228 config_acs= 5229 Format: 5230 <ACS flags>@<pci_dev>[; ...] 5231 Specify one or more PCI devices (in the format 5232 specified above) optionally prepended with flags 5233 and separated by semicolons. The respective 5234 capabilities will be enabled, disabled or 5235 unchanged based on what is specified in 5236 flags. 5237 5238 ACS Flags is defined as follows: 5239 bit-0 : ACS Source Validation 5240 bit-1 : ACS Translation Blocking 5241 bit-2 : ACS P2P Request Redirect 5242 bit-3 : ACS P2P Completion Redirect 5243 bit-4 : ACS Upstream Forwarding 5244 bit-5 : ACS P2P Egress Control 5245 bit-6 : ACS Direct Translated P2P 5246 Each bit can be marked as: 5247 '0' – force disabled 5248 '1' – force enabled 5249 'x' – unchanged 5250 For example, 5251 pci=config_acs=10x@pci:0:0 5252 would configure all devices that support 5253 ACS to enable P2P Request Redirect, disable 5254 Translation Blocking, and leave Source 5255 Validation unchanged from whatever power-up 5256 or firmware set it to. 5257 5258 Note: this may remove isolation between devices 5259 and may put more devices in an IOMMU group. 5260 force_floating [S390] Force usage of floating interrupts. 5261 nomio [S390] Do not use MIO instructions. 5262 norid [S390] ignore the RID field and force use of 5263 one PCI domain per PCI function 5264 notph [PCIE] If the PCIE_TPH kernel config parameter 5265 is enabled, this kernel boot option can be used 5266 to disable PCIe TLP Processing Hints support 5267 system-wide. 5268 5269 pcie_aspm= [PCIE] Forcibly enable or ignore PCIe Active State Power 5270 Management. 5271 off Don't touch ASPM configuration at all. Leave any 5272 configuration done by firmware unchanged. 5273 force Enable ASPM even on devices that claim not to support it. 5274 WARNING: Forcing ASPM on may cause system lockups. 5275 5276 pcie_ports= [PCIE] PCIe port services handling: 5277 native Use native PCIe services (PME, AER, DPC, PCIe hotplug) 5278 even if the platform doesn't give the OS permission to 5279 use them. This may cause conflicts if the platform 5280 also tries to use these services. 5281 dpc-native Use native PCIe service for DPC only. May 5282 cause conflicts if firmware uses AER or DPC. 5283 compat Disable native PCIe services (PME, AER, DPC, PCIe 5284 hotplug). 5285 5286 pcie_port_pm= [PCIE] PCIe port power management handling: 5287 off Disable power management of all PCIe ports 5288 force Forcibly enable power management of all PCIe ports 5289 5290 pcie_pme= [PCIE,PM] Native PCIe PME signaling options: 5291 nomsi Do not use MSI for native PCIe PME signaling (this makes 5292 all PCIe root ports use INTx for all services). 5293 5294 pcmv= [HW,PCMCIA] BadgePAD 4 5295 5296 pd_ignore_unused 5297 [PM] 5298 Keep all power-domains already enabled by bootloader on, 5299 even if no driver has claimed them. This is useful 5300 for debug and development, but should not be 5301 needed on a platform with proper driver support. 5302 5303 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at 5304 boot time. 5305 Format: { 0 | 1 } 5306 See arch/parisc/kernel/pdc_chassis.c 5307 5308 percpu_alloc= [MM,EARLY] 5309 Select which percpu first chunk allocator to use. 5310 Currently supported values are "embed" and "page". 5311 Archs may support subset or none of the selections. 5312 See comments in mm/percpu.c for details on each 5313 allocator. This parameter is primarily for debugging 5314 and performance comparison. 5315 5316 pirq= [SMP,APIC] Manual mp-table setup 5317 See Documentation/arch/x86/i386/IO-APIC.rst. 5318 5319 plip= [PPT,NET] Parallel port network link 5320 Format: { parport<nr> | timid | 0 } 5321 See also Documentation/admin-guide/parport.rst. 5322 5323 pmtmr= [X86] Manual setup of pmtmr I/O Port. 5324 Override pmtimer IOPort with a hex value. 5325 e.g. pmtmr=0x508 5326 5327 pmu_override= [PPC] Override the PMU. 5328 This option takes over the PMU facility, so it is no 5329 longer usable by perf. Setting this option starts the 5330 PMU counters by setting MMCR0 to 0 (the FC bit is 5331 cleared). If a number is given, then MMCR1 is set to 5332 that number, otherwise (e.g., 'pmu_override=on'), MMCR1 5333 remains 0. 5334 5335 pm_async= [PM] 5336 Format: off 5337 This parameter sets the initial value of the 5338 /sys/power/pm_async sysfs knob at boot time. 5339 If set to "off", disables asynchronous suspend and 5340 resume of devices during system-wide power transitions. 5341 This can be useful on platforms where device 5342 dependencies are not well-defined, or for debugging 5343 power management issues. Asynchronous operations are 5344 enabled by default. 5345 5346 5347 pm_debug_messages [SUSPEND,KNL] 5348 Enable suspend/resume debug messages during boot up. 5349 5350 pnp.debug=1 [PNP] 5351 Enable PNP debug messages (depends on the 5352 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time 5353 via /sys/module/pnp/parameters/debug. We always show 5354 current resource usage; turning this on also shows 5355 possible settings and some assignment information. 5356 5357 pnpacpi= [ACPI] 5358 { off } 5359 5360 pnpbios= [ISAPNP] 5361 { on | off | curr | res | no-curr | no-res } 5362 5363 pnp_reserve_irq= 5364 [ISAPNP] Exclude IRQs for the autoconfiguration 5365 5366 pnp_reserve_dma= 5367 [ISAPNP] Exclude DMAs for the autoconfiguration 5368 5369 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration 5370 Ranges are in pairs (I/O port base and size). 5371 5372 pnp_reserve_mem= 5373 [ISAPNP] Exclude memory regions for the 5374 autoconfiguration. 5375 Ranges are in pairs (memory base and size). 5376 5377 ports= [IP_VS_FTP] IPVS ftp helper module 5378 Default is 21. 5379 Up to 8 (IP_VS_APP_MAX_PORTS) ports 5380 may be specified. 5381 Format: <port>,<port>.... 5382 5383 possible_cpus= [SMP,S390,X86] 5384 Format: <unsigned int> 5385 Set the number of possible CPUs, overriding the 5386 regular discovery mechanisms (such as ACPI/FW, etc). 5387 5388 powersave=off [PPC] This option disables power saving features. 5389 It specifically disables cpuidle and sets the 5390 platform machine description specific power_save 5391 function to NULL. On Idle the CPU just reduces 5392 execution priority. 5393 5394 ppc_strict_facility_enable 5395 [PPC,ENABLE] This option catches any kernel floating point, 5396 Altivec, VSX and SPE outside of regions specifically 5397 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()). 5398 There is some performance impact when enabling this. 5399 5400 ppc_tm= [PPC,EARLY] 5401 Format: {"off"} 5402 Disable Hardware Transactional Memory 5403 5404 preempt= [KNL] 5405 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC 5406 none - Limited to cond_resched() calls 5407 voluntary - Limited to cond_resched() and might_sleep() calls 5408 full - Any section that isn't explicitly preempt disabled 5409 can be preempted anytime. Tasks will also yield 5410 contended spinlocks (if the critical section isn't 5411 explicitly preempt disabled beyond the lock itself). 5412 lazy - Scheduler controlled. Similar to full but instead 5413 of preempting the task immediately, the task gets 5414 one HZ tick time to yield itself before the 5415 preemption will be forced. One preemption is when the 5416 task returns to user space. 5417 5418 print-fatal-signals= 5419 [KNL] debug: print fatal signals 5420 5421 If enabled, warn about various signal handling 5422 related application anomalies: too many signals, 5423 too many POSIX.1 timers, fatal signals causing a 5424 coredump - etc. 5425 5426 If you hit the warning due to signal overflow, 5427 you might want to try "ulimit -i unlimited". 5428 5429 default: off. 5430 5431 printk.always_kmsg_dump= 5432 Trigger kmsg_dump for cases other than kernel oops or 5433 panics 5434 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5435 default: disabled 5436 5437 printk.console_no_auto_verbose= 5438 Disable console loglevel raise on oops, panic 5439 or lockdep-detected issues (only if lock debug is on). 5440 With an exception to setups with low baudrate on 5441 serial console, keeping this 0 is a good choice 5442 in order to provide more debug information. 5443 Format: <bool> 5444 default: 0 (auto_verbose is enabled) 5445 5446 printk.debug_non_panic_cpus= 5447 Allows storing messages from non-panic CPUs into 5448 the printk log buffer during panic(). They are 5449 flushed to consoles by the panic-CPU on 5450 a best-effort basis. 5451 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5452 Default: disabled 5453 5454 printk.devkmsg={on,off,ratelimit} 5455 Control writing to /dev/kmsg. 5456 on - unlimited logging to /dev/kmsg from userspace 5457 off - logging to /dev/kmsg disabled 5458 ratelimit - ratelimit the logging 5459 Default: ratelimit 5460 5461 printk.time= Show timing data prefixed to each printk message line 5462 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5463 5464 proc_mem.force_override= [KNL] 5465 Format: {always | ptrace | never} 5466 Traditionally /proc/pid/mem allows memory permissions to be 5467 overridden without restrictions. This option may be set to 5468 restrict that. Can be one of: 5469 - 'always': traditional behavior always allows mem overrides. 5470 - 'ptrace': only allow mem overrides for active ptracers. 5471 - 'never': never allow mem overrides. 5472 If not specified, default is the CONFIG_PROC_MEM_* choice. 5473 5474 processor.max_cstate= [HW,ACPI] 5475 Limit processor to maximum C-state 5476 max_cstate=9 overrides any DMI blacklist limit. 5477 5478 processor.nocst [HW,ACPI] 5479 Ignore the _CST method to determine C-states, 5480 instead using the legacy FADT method 5481 5482 profile= [KNL] Enable kernel profiling via /proc/profile 5483 Format: [<profiletype>,]<number> 5484 Param: <profiletype>: "schedule" or "kvm" 5485 [defaults to kernel profiling] 5486 Param: "schedule" - profile schedule points. 5487 Param: "kvm" - profile VM exits. 5488 Param: <number> - step/bucket size as a power of 2 for 5489 statistical time based profiling. 5490 5491 prot_virt= [S390] enable hosting protected virtual machines 5492 isolated from the hypervisor (if hardware supports 5493 that). If enabled, the default kernel base address 5494 might be overridden even when Kernel Address Space 5495 Layout Randomization is disabled. 5496 Format: <bool> 5497 5498 psi= [KNL] Enable or disable pressure stall information 5499 tracking. 5500 Format: <bool> 5501 5502 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to 5503 probe for; one of (bare|imps|exps|lifebook|any). 5504 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports 5505 per second. 5506 psmouse.resetafter= [HW,MOUSE] 5507 Try to reset the device after so many bad packets 5508 (0 = never). 5509 psmouse.resolution= 5510 [HW,MOUSE] Set desired mouse resolution, in dpi. 5511 psmouse.smartscroll= 5512 [HW,MOUSE] Controls Logitech smartscroll autorepeat. 5513 0 = disabled, 1 = enabled (default). 5514 5515 pstore.backend= Specify the name of the pstore backend to use 5516 5517 pti= [X86-64] Control Page Table Isolation of user and 5518 kernel address spaces. Disabling this feature 5519 removes hardening, but improves performance of 5520 system calls and interrupts. 5521 5522 on - unconditionally enable 5523 off - unconditionally disable 5524 auto - kernel detects whether your CPU model is 5525 vulnerable to issues that PTI mitigates 5526 5527 Not specifying this option is equivalent to pti=auto. 5528 5529 pty.legacy_count= 5530 [KNL] Number of legacy pty's. Overwrites compiled-in 5531 default number. 5532 5533 quiet [KNL,EARLY] Disable most log messages 5534 5535 r128= [HW,DRM] 5536 5537 radix_hcall_invalidate=on [PPC/PSERIES] 5538 Disable RADIX GTSE feature and use hcall for TLB 5539 invalidate. 5540 5541 raid= [HW,RAID] 5542 See Documentation/admin-guide/md.rst. 5543 5544 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes 5545 See Documentation/admin-guide/blockdev/ramdisk.rst. 5546 5547 ramdisk_start= [Deprecated,RAM] RAM disk image start address 5548 5549 random.trust_cpu=off 5550 [KNL,EARLY] Disable trusting the use of the CPU's 5551 random number generator (if available) to 5552 initialize the kernel's RNG. 5553 5554 random.trust_bootloader=off 5555 [KNL,EARLY] Disable trusting the use of the a seed 5556 passed by the bootloader (if available) to 5557 initialize the kernel's RNG. 5558 5559 randomize_kstack_offset= 5560 [KNL,EARLY] Enable or disable kernel stack offset 5561 randomization, which provides roughly 5 bits of 5562 entropy, frustrating memory corruption attacks 5563 that depend on stack address determinism or 5564 cross-syscall address exposures. This is only 5565 available on architectures that have defined 5566 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET. 5567 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5568 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT. 5569 5570 ras=option[,option,...] [KNL] RAS-specific options 5571 5572 cec_disable [X86] 5573 Disable the Correctable Errors Collector, 5574 see CONFIG_RAS_CEC help text. 5575 5576 rcu_nocbs[=cpu-list] 5577 [KNL] The optional argument is a cpu list, 5578 as described above. 5579 5580 In kernels built with CONFIG_RCU_NOCB_CPU=y, 5581 enable the no-callback CPU mode, which prevents 5582 such CPUs' callbacks from being invoked in 5583 softirq context. Invocation of such CPUs' RCU 5584 callbacks will instead be offloaded to "rcuox/N" 5585 kthreads created for that purpose, where "x" is 5586 "p" for RCU-preempt, "s" for RCU-sched, and "g" 5587 for the kthreads that mediate grace periods; and 5588 "N" is the CPU number. This reduces OS jitter on 5589 the offloaded CPUs, which can be useful for HPC 5590 and real-time workloads. It can also improve 5591 energy efficiency for asymmetric multiprocessors. 5592 5593 If a cpulist is passed as an argument, the specified 5594 list of CPUs is set to no-callback mode from boot. 5595 5596 Otherwise, if the '=' sign and the cpulist 5597 arguments are omitted, no CPU will be set to 5598 no-callback mode from boot but the mode may be 5599 toggled at runtime via cpusets. 5600 5601 Note that this argument takes precedence over 5602 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 5603 5604 rcu_nocb_poll [KNL] 5605 Rather than requiring that offloaded CPUs 5606 (specified by rcu_nocbs= above) explicitly 5607 awaken the corresponding "rcuoN" kthreads, 5608 make these kthreads poll for callbacks. 5609 This improves the real-time response for the 5610 offloaded CPUs by relieving them of the need to 5611 wake up the corresponding kthread, but degrades 5612 energy efficiency by requiring that the kthreads 5613 periodically wake up to do the polling. 5614 5615 rcutree.blimit= [KNL] 5616 Set maximum number of finished RCU callbacks to 5617 process in one batch. 5618 5619 rcutree.csd_lock_suppress_rcu_stall= [KNL] 5620 Do only a one-line RCU CPU stall warning when 5621 there is an ongoing too-long CSD-lock wait. 5622 5623 rcutree.do_rcu_barrier= [KNL] 5624 Request a call to rcu_barrier(). This is 5625 throttled so that userspace tests can safely 5626 hammer on the sysfs variable if they so choose. 5627 If triggered before the RCU grace-period machinery 5628 is fully active, this will error out with EAGAIN. 5629 5630 rcutree.dump_tree= [KNL] 5631 Dump the structure of the rcu_node combining tree 5632 out at early boot. This is used for diagnostic 5633 purposes, to verify correct tree setup. 5634 5635 rcutree.gp_cleanup_delay= [KNL] 5636 Set the number of jiffies to delay each step of 5637 RCU grace-period cleanup. 5638 5639 rcutree.gp_init_delay= [KNL] 5640 Set the number of jiffies to delay each step of 5641 RCU grace-period initialization. 5642 5643 rcutree.gp_preinit_delay= [KNL] 5644 Set the number of jiffies to delay each step of 5645 RCU grace-period pre-initialization, that is, 5646 the propagation of recent CPU-hotplug changes up 5647 the rcu_node combining tree. 5648 5649 rcutree.jiffies_till_first_fqs= [KNL] 5650 Set delay from grace-period initialization to 5651 first attempt to force quiescent states. 5652 Units are jiffies, minimum value is zero, 5653 and maximum value is HZ. 5654 5655 rcutree.jiffies_till_next_fqs= [KNL] 5656 Set delay between subsequent attempts to force 5657 quiescent states. Units are jiffies, minimum 5658 value is one, and maximum value is HZ. 5659 5660 rcutree.jiffies_till_sched_qs= [KNL] 5661 Set required age in jiffies for a 5662 given grace period before RCU starts 5663 soliciting quiescent-state help from 5664 rcu_note_context_switch() and cond_resched(). 5665 If not specified, the kernel will calculate 5666 a value based on the most recent settings 5667 of rcutree.jiffies_till_first_fqs 5668 and rcutree.jiffies_till_next_fqs. 5669 This calculated value may be viewed in 5670 rcutree.jiffies_to_sched_qs. Any attempt to set 5671 rcutree.jiffies_to_sched_qs will be cheerfully 5672 overwritten. 5673 5674 rcutree.kthread_prio= [KNL,BOOT] 5675 Set the SCHED_FIFO priority of the RCU per-CPU 5676 kthreads (rcuc/N). This value is also used for 5677 the priority of the RCU boost threads (rcub/N) 5678 and for the RCU grace-period kthreads (rcu_bh, 5679 rcu_preempt, and rcu_sched). If RCU_BOOST is 5680 set, valid values are 1-99 and the default is 1 5681 (the least-favored priority). Otherwise, when 5682 RCU_BOOST is not set, valid values are 0-99 and 5683 the default is zero (non-realtime operation). 5684 When RCU_NOCB_CPU is set, also adjust the 5685 priority of NOCB callback kthreads. 5686 5687 rcutree.nocb_nobypass_lim_per_jiffy= [KNL] 5688 On callback-offloaded (rcu_nocbs) CPUs, 5689 RCU reduces the lock contention that would 5690 otherwise be caused by callback floods through 5691 use of the ->nocb_bypass list. However, in the 5692 common non-flooded case, RCU queues directly to 5693 the main ->cblist in order to avoid the extra 5694 overhead of the ->nocb_bypass list and its lock. 5695 But if there are too many callbacks queued during 5696 a single jiffy, RCU pre-queues the callbacks into 5697 the ->nocb_bypass queue. The definition of "too 5698 many" is supplied by this kernel boot parameter. 5699 5700 rcutree.nohz_full_patience_delay= [KNL] 5701 On callback-offloaded (rcu_nocbs) CPUs, avoid 5702 disturbing RCU unless the grace period has 5703 reached the specified age in milliseconds. 5704 Defaults to zero. Large values will be capped 5705 at five seconds. All values will be rounded down 5706 to the nearest value representable by jiffies. 5707 5708 rcutree.qhimark= [KNL] 5709 Set threshold of queued RCU callbacks beyond which 5710 batch limiting is disabled. 5711 5712 rcutree.qlowmark= [KNL] 5713 Set threshold of queued RCU callbacks below which 5714 batch limiting is re-enabled. 5715 5716 rcutree.qovld= [KNL] 5717 Set threshold of queued RCU callbacks beyond which 5718 RCU's force-quiescent-state scan will aggressively 5719 enlist help from cond_resched() and sched IPIs to 5720 help CPUs more quickly reach quiescent states. 5721 Set to less than zero to make this be set based 5722 on rcutree.qhimark at boot time and to zero to 5723 disable more aggressive help enlistment. 5724 5725 rcutree.rcu_delay_page_cache_fill_msec= [KNL] 5726 Set the page-cache refill delay (in milliseconds) 5727 in response to low-memory conditions. The range 5728 of permitted values is in the range 0:100000. 5729 5730 rcutree.rcu_divisor= [KNL] 5731 Set the shift-right count to use to compute 5732 the callback-invocation batch limit bl from 5733 the number of callbacks queued on this CPU. 5734 The result will be bounded below by the value of 5735 the rcutree.blimit kernel parameter. Every bl 5736 callbacks, the softirq handler will exit in 5737 order to allow the CPU to do other work. 5738 5739 Please note that this callback-invocation batch 5740 limit applies only to non-offloaded callback 5741 invocation. Offloaded callbacks are instead 5742 invoked in the context of an rcuoc kthread, which 5743 scheduler will preempt as it does any other task. 5744 5745 rcutree.rcu_fanout_exact= [KNL] 5746 Disable autobalancing of the rcu_node combining 5747 tree. This is used by rcutorture, and might 5748 possibly be useful for architectures having high 5749 cache-to-cache transfer latencies. 5750 5751 rcutree.rcu_fanout_leaf= [KNL] 5752 Change the number of CPUs assigned to each 5753 leaf rcu_node structure. Useful for very 5754 large systems, which will choose the value 64, 5755 and for NUMA systems with large remote-access 5756 latencies, which will choose a value aligned 5757 with the appropriate hardware boundaries. 5758 5759 rcutree.rcu_min_cached_objs= [KNL] 5760 Minimum number of objects which are cached and 5761 maintained per one CPU. Object size is equal 5762 to PAGE_SIZE. The cache allows to reduce the 5763 pressure to page allocator, also it makes the 5764 whole algorithm to behave better in low memory 5765 condition. 5766 5767 rcutree.rcu_nocb_gp_stride= [KNL] 5768 Set the number of NOCB callback kthreads in 5769 each group, which defaults to the square root 5770 of the number of CPUs. Larger numbers reduce 5771 the wakeup overhead on the global grace-period 5772 kthread, but increases that same overhead on 5773 each group's NOCB grace-period kthread. 5774 5775 rcutree.rcu_kick_kthreads= [KNL] 5776 Cause the grace-period kthread to get an extra 5777 wake_up() if it sleeps three times longer than 5778 it should at force-quiescent-state time. 5779 This wake_up() will be accompanied by a 5780 WARN_ONCE() splat and an ftrace_dump(). 5781 5782 rcutree.rcu_resched_ns= [KNL] 5783 Limit the time spend invoking a batch of RCU 5784 callbacks to the specified number of nanoseconds. 5785 By default, this limit is checked only once 5786 every 32 callbacks in order to limit the pain 5787 inflicted by local_clock() overhead. 5788 5789 rcutree.rcu_unlock_delay= [KNL] 5790 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels, 5791 this specifies an rcu_read_unlock()-time delay 5792 in microseconds. This defaults to zero. 5793 Larger delays increase the probability of 5794 catching RCU pointer leaks, that is, buggy use 5795 of RCU-protected pointers after the relevant 5796 rcu_read_unlock() has completed. 5797 5798 rcutree.sysrq_rcu= [KNL] 5799 Commandeer a sysrq key to dump out Tree RCU's 5800 rcu_node tree with an eye towards determining 5801 why a new grace period has not yet started. 5802 5803 rcutree.use_softirq= [KNL] 5804 If set to zero, move all RCU_SOFTIRQ processing to 5805 per-CPU rcuc kthreads. Defaults to a non-zero 5806 value, meaning that RCU_SOFTIRQ is used by default. 5807 Specify rcutree.use_softirq=0 to use rcuc kthreads. 5808 5809 But note that CONFIG_PREEMPT_RT=y kernels disable 5810 this kernel boot parameter, forcibly setting it 5811 to zero. 5812 5813 rcutree.enable_rcu_lazy= [KNL] 5814 To save power, batch RCU callbacks and flush after 5815 delay, memory pressure or callback list growing too 5816 big. 5817 5818 rcutree.rcu_normal_wake_from_gp= [KNL] 5819 Reduces a latency of synchronize_rcu() call. This approach 5820 maintains its own track of synchronize_rcu() callers, so it 5821 does not interact with regular callbacks because it does not 5822 use a call_rcu[_hurry]() path. Please note, this is for a 5823 normal grace period. 5824 5825 How to disable it: 5826 5827 echo 0 > /sys/module/rcutree/parameters/rcu_normal_wake_from_gp 5828 or pass a boot parameter "rcutree.rcu_normal_wake_from_gp=0" 5829 5830 Default is 1 if it is not explicitly disabled by the boot parameter 5831 passing 0. 5832 5833 rcuscale.gp_async= [KNL] 5834 Measure performance of asynchronous 5835 grace-period primitives such as call_rcu(). 5836 5837 rcuscale.gp_async_max= [KNL] 5838 Specify the maximum number of outstanding 5839 callbacks per writer thread. When a writer 5840 thread exceeds this limit, it invokes the 5841 corresponding flavor of rcu_barrier() to allow 5842 previously posted callbacks to drain. 5843 5844 rcuscale.gp_exp= [KNL] 5845 Measure performance of expedited synchronous 5846 grace-period primitives. 5847 5848 rcuscale.holdoff= [KNL] 5849 Set test-start holdoff period. The purpose of 5850 this parameter is to delay the start of the 5851 test until boot completes in order to avoid 5852 interference. 5853 5854 rcuscale.kfree_by_call_rcu= [KNL] 5855 In kernels built with CONFIG_RCU_LAZY=y, test 5856 call_rcu() instead of kfree_rcu(). 5857 5858 rcuscale.kfree_mult= [KNL] 5859 Instead of allocating an object of size kfree_obj, 5860 allocate one of kfree_mult * sizeof(kfree_obj). 5861 Defaults to 1. 5862 5863 rcuscale.kfree_rcu_test= [KNL] 5864 Set to measure performance of kfree_rcu() flooding. 5865 5866 rcuscale.kfree_rcu_test_double= [KNL] 5867 Test the double-argument variant of kfree_rcu(). 5868 If this parameter has the same value as 5869 rcuscale.kfree_rcu_test_single, both the single- 5870 and double-argument variants are tested. 5871 5872 rcuscale.kfree_rcu_test_single= [KNL] 5873 Test the single-argument variant of kfree_rcu(). 5874 If this parameter has the same value as 5875 rcuscale.kfree_rcu_test_double, both the single- 5876 and double-argument variants are tested. 5877 5878 rcuscale.kfree_nthreads= [KNL] 5879 The number of threads running loops of kfree_rcu(). 5880 5881 rcuscale.kfree_alloc_num= [KNL] 5882 Number of allocations and frees done in an iteration. 5883 5884 rcuscale.kfree_loops= [KNL] 5885 Number of loops doing rcuscale.kfree_alloc_num number 5886 of allocations and frees. 5887 5888 rcuscale.minruntime= [KNL] 5889 Set the minimum test run time in seconds. This 5890 does not affect the data-collection interval, 5891 but instead allows better measurement of things 5892 like CPU consumption. 5893 5894 rcuscale.nreaders= [KNL] 5895 Set number of RCU readers. The value -1 selects 5896 N, where N is the number of CPUs. A value 5897 "n" less than -1 selects N-n+1, where N is again 5898 the number of CPUs. For example, -2 selects N 5899 (the number of CPUs), -3 selects N+1, and so on. 5900 A value of "n" less than or equal to -N selects 5901 a single reader. 5902 5903 rcuscale.nwriters= [KNL] 5904 Set number of RCU writers. The values operate 5905 the same as for rcuscale.nreaders. 5906 N, where N is the number of CPUs 5907 5908 rcuscale.scale_type= [KNL] 5909 Specify the RCU implementation to test. 5910 5911 rcuscale.shutdown= [KNL] 5912 Shut the system down after performance tests 5913 complete. This is useful for hands-off automated 5914 testing. 5915 5916 rcuscale.verbose= [KNL] 5917 Enable additional printk() statements. 5918 5919 rcuscale.writer_holdoff= [KNL] 5920 Write-side holdoff between grace periods, 5921 in microseconds. The default of zero says 5922 no holdoff. 5923 5924 rcuscale.writer_holdoff_jiffies= [KNL] 5925 Additional write-side holdoff between grace 5926 periods, but in jiffies. The default of zero 5927 says no holdoff. 5928 5929 rcutorture.fqs_duration= [KNL] 5930 Set duration of force_quiescent_state bursts 5931 in microseconds. 5932 5933 rcutorture.fqs_holdoff= [KNL] 5934 Set holdoff time within force_quiescent_state bursts 5935 in microseconds. 5936 5937 rcutorture.fqs_stutter= [KNL] 5938 Set wait time between force_quiescent_state bursts 5939 in seconds. 5940 5941 rcutorture.fwd_progress= [KNL] 5942 Specifies the number of kthreads to be used 5943 for RCU grace-period forward-progress testing 5944 for the types of RCU supporting this notion. 5945 Defaults to 1 kthread, values less than zero or 5946 greater than the number of CPUs cause the number 5947 of CPUs to be used. 5948 5949 rcutorture.fwd_progress_div= [KNL] 5950 Specify the fraction of a CPU-stall-warning 5951 period to do tight-loop forward-progress testing. 5952 5953 rcutorture.fwd_progress_holdoff= [KNL] 5954 Number of seconds to wait between successive 5955 forward-progress tests. 5956 5957 rcutorture.fwd_progress_need_resched= [KNL] 5958 Enclose cond_resched() calls within checks for 5959 need_resched() during tight-loop forward-progress 5960 testing. 5961 5962 rcutorture.gp_cond= [KNL] 5963 Use conditional/asynchronous update-side 5964 normal-grace-period primitives, if available. 5965 5966 rcutorture.gp_cond_exp= [KNL] 5967 Use conditional/asynchronous update-side 5968 expedited-grace-period primitives, if available. 5969 5970 rcutorture.gp_cond_full= [KNL] 5971 Use conditional/asynchronous update-side 5972 normal-grace-period primitives that also take 5973 concurrent expedited grace periods into account, 5974 if available. 5975 5976 rcutorture.gp_cond_exp_full= [KNL] 5977 Use conditional/asynchronous update-side 5978 expedited-grace-period primitives that also take 5979 concurrent normal grace periods into account, 5980 if available. 5981 5982 rcutorture.gp_cond_wi= [KNL] 5983 Nominal wait interval for normal conditional 5984 grace periods (specified by rcutorture's 5985 gp_cond and gp_cond_full module parameters), 5986 in microseconds. The actual wait interval will 5987 be randomly selected to nanosecond granularity up 5988 to this wait interval. Defaults to 16 jiffies, 5989 for example, 16,000 microseconds on a system 5990 with HZ=1000. 5991 5992 rcutorture.gp_cond_wi_exp= [KNL] 5993 Nominal wait interval for expedited conditional 5994 grace periods (specified by rcutorture's 5995 gp_cond_exp and gp_cond_exp_full module 5996 parameters), in microseconds. The actual wait 5997 interval will be randomly selected to nanosecond 5998 granularity up to this wait interval. Defaults to 5999 128 microseconds. 6000 6001 rcutorture.gp_exp= [KNL] 6002 Use expedited update-side primitives, if available. 6003 6004 rcutorture.gp_normal= [KNL] 6005 Use normal (non-expedited) asynchronous 6006 update-side primitives, if available. 6007 6008 rcutorture.gp_poll= [KNL] 6009 Use polled update-side normal-grace-period 6010 primitives, if available. 6011 6012 rcutorture.gp_poll_exp= [KNL] 6013 Use polled update-side expedited-grace-period 6014 primitives, if available. 6015 6016 rcutorture.gp_poll_full= [KNL] 6017 Use polled update-side normal-grace-period 6018 primitives that also take concurrent expedited 6019 grace periods into account, if available. 6020 6021 rcutorture.gp_poll_exp_full= [KNL] 6022 Use polled update-side expedited-grace-period 6023 primitives that also take concurrent normal 6024 grace periods into account, if available. 6025 6026 rcutorture.gp_poll_wi= [KNL] 6027 Nominal wait interval for normal conditional 6028 grace periods (specified by rcutorture's 6029 gp_poll and gp_poll_full module parameters), 6030 in microseconds. The actual wait interval will 6031 be randomly selected to nanosecond granularity up 6032 to this wait interval. Defaults to 16 jiffies, 6033 for example, 16,000 microseconds on a system 6034 with HZ=1000. 6035 6036 rcutorture.gp_poll_wi_exp= [KNL] 6037 Nominal wait interval for expedited conditional 6038 grace periods (specified by rcutorture's 6039 gp_poll_exp and gp_poll_exp_full module 6040 parameters), in microseconds. The actual wait 6041 interval will be randomly selected to nanosecond 6042 granularity up to this wait interval. Defaults to 6043 128 microseconds. 6044 6045 rcutorture.gp_sync= [KNL] 6046 Use normal (non-expedited) synchronous 6047 update-side primitives, if available. If all 6048 of rcutorture.gp_cond=, rcutorture.gp_exp=, 6049 rcutorture.gp_normal=, and rcutorture.gp_sync= 6050 are zero, rcutorture acts as if is interpreted 6051 they are all non-zero. 6052 6053 rcutorture.gpwrap_lag= [KNL] 6054 Enable grace-period wrap lag testing. Setting 6055 to false prevents the gpwrap lag test from 6056 running. Default is true. 6057 6058 rcutorture.gpwrap_lag_gps= [KNL] 6059 Set the value for grace-period wrap lag during 6060 active lag testing periods. This controls how many 6061 grace periods differences we tolerate between 6062 rdp and rnp's gp_seq before setting overflow flag. 6063 The default is always set to 8. 6064 6065 rcutorture.gpwrap_lag_cycle_mins= [KNL] 6066 Set the total cycle duration for gpwrap lag 6067 testing in minutes. This is the total time for 6068 one complete cycle of active and inactive 6069 testing periods. Default is 30 minutes. 6070 6071 rcutorture.gpwrap_lag_active_mins= [KNL] 6072 Set the duration for which gpwrap lag is active 6073 within each cycle, in minutes. During this time, 6074 the grace-period wrap lag will be set to the 6075 value specified by gpwrap_lag_gps. Default is 6076 5 minutes. 6077 6078 rcutorture.irqreader= [KNL] 6079 Run RCU readers from irq handlers, or, more 6080 accurately, from a timer handler. Not all RCU 6081 flavors take kindly to this sort of thing. 6082 6083 rcutorture.leakpointer= [KNL] 6084 Leak an RCU-protected pointer out of the reader. 6085 This can of course result in splats, and is 6086 intended to test the ability of things like 6087 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect 6088 such leaks. 6089 6090 rcutorture.n_barrier_cbs= [KNL] 6091 Set callbacks/threads for rcu_barrier() testing. 6092 6093 rcutorture.nfakewriters= [KNL] 6094 Set number of concurrent RCU writers. These just 6095 stress RCU, they don't participate in the actual 6096 test, hence the "fake". 6097 6098 rcutorture.nocbs_nthreads= [KNL] 6099 Set number of RCU callback-offload togglers. 6100 Zero (the default) disables toggling. 6101 6102 rcutorture.nocbs_toggle= [KNL] 6103 Set the delay in milliseconds between successive 6104 callback-offload toggling attempts. 6105 6106 rcutorture.nreaders= [KNL] 6107 Set number of RCU readers. The value -1 selects 6108 N-1, where N is the number of CPUs. A value 6109 "n" less than -1 selects N-n-2, where N is again 6110 the number of CPUs. For example, -2 selects N 6111 (the number of CPUs), -3 selects N+1, and so on. 6112 6113 rcutorture.object_debug= [KNL] 6114 Enable debug-object double-call_rcu() testing. 6115 6116 rcutorture.onoff_holdoff= [KNL] 6117 Set time (s) after boot for CPU-hotplug testing. 6118 6119 rcutorture.onoff_interval= [KNL] 6120 Set time (jiffies) between CPU-hotplug operations, 6121 or zero to disable CPU-hotplug testing. 6122 6123 rcutorture.preempt_duration= [KNL] 6124 Set duration (in milliseconds) of preemptions 6125 by a high-priority FIFO real-time task. Set to 6126 zero (the default) to disable. The CPUs to 6127 preempt are selected randomly from the set that 6128 are online at a given point in time. Races with 6129 CPUs going offline are ignored, with that attempt 6130 at preemption skipped. 6131 6132 rcutorture.preempt_interval= [KNL] 6133 Set interval (in milliseconds, defaulting to one 6134 second) between preemptions by a high-priority 6135 FIFO real-time task. This delay is mediated 6136 by an hrtimer and is further fuzzed to avoid 6137 inadvertent synchronizations. 6138 6139 rcutorture.read_exit_burst= [KNL] 6140 The number of times in a given read-then-exit 6141 episode that a set of read-then-exit kthreads 6142 is spawned. 6143 6144 rcutorture.read_exit_delay= [KNL] 6145 The delay, in seconds, between successive 6146 read-then-exit testing episodes. 6147 6148 rcutorture.reader_flavor= [KNL] 6149 A bit mask indicating which readers to use. 6150 If there is more than one bit set, the readers 6151 are entered from low-order bit up, and are 6152 exited in the opposite order. For SRCU, the 6153 0x1 bit is normal readers, 0x2 NMI-safe readers, 6154 and 0x4 light-weight readers. 6155 6156 rcutorture.shuffle_interval= [KNL] 6157 Set task-shuffle interval (s). Shuffling tasks 6158 allows some CPUs to go into dyntick-idle mode 6159 during the rcutorture test. 6160 6161 rcutorture.shutdown_secs= [KNL] 6162 Set time (s) after boot system shutdown. This 6163 is useful for hands-off automated testing. 6164 6165 rcutorture.stall_cpu= [KNL] 6166 Duration of CPU stall (s) to test RCU CPU stall 6167 warnings, zero to disable. 6168 6169 rcutorture.stall_cpu_block= [KNL] 6170 Sleep while stalling if set. This will result 6171 in warnings from preemptible RCU in addition to 6172 any other stall-related activity. Note that 6173 in kernels built with CONFIG_PREEMPTION=n and 6174 CONFIG_PREEMPT_COUNT=y, this parameter will 6175 cause the CPU to pass through a quiescent state. 6176 Given CONFIG_PREEMPTION=n, this will suppress 6177 RCU CPU stall warnings, but will instead result 6178 in scheduling-while-atomic splats. 6179 6180 Use of this module parameter results in splats. 6181 6182 6183 rcutorture.stall_cpu_holdoff= [KNL] 6184 Time to wait (s) after boot before inducing stall. 6185 6186 rcutorture.stall_cpu_irqsoff= [KNL] 6187 Disable interrupts while stalling if set, but only 6188 on the first stall in the set. 6189 6190 rcutorture.stall_cpu_repeat= [KNL] 6191 Number of times to repeat the stall sequence, 6192 so that rcutorture.stall_cpu_repeat=3 will result 6193 in four stall sequences. 6194 6195 rcutorture.stall_gp_kthread= [KNL] 6196 Duration (s) of forced sleep within RCU 6197 grace-period kthread to test RCU CPU stall 6198 warnings, zero to disable. If both stall_cpu 6199 and stall_gp_kthread are specified, the 6200 kthread is starved first, then the CPU. 6201 6202 rcutorture.stat_interval= [KNL] 6203 Time (s) between statistics printk()s. 6204 6205 rcutorture.stutter= [KNL] 6206 Time (s) to stutter testing, for example, specifying 6207 five seconds causes the test to run for five seconds, 6208 wait for five seconds, and so on. This tests RCU's 6209 ability to transition abruptly to and from idle. 6210 6211 rcutorture.test_boost= [KNL] 6212 Test RCU priority boosting? 0=no, 1=maybe, 2=yes. 6213 "Maybe" means test if the RCU implementation 6214 under test support RCU priority boosting. 6215 6216 rcutorture.test_boost_duration= [KNL] 6217 Duration (s) of each individual boost test. 6218 6219 rcutorture.test_boost_holdoff= [KNL] 6220 Holdoff time (s) from start of test to the start 6221 of RCU priority-boost testing. Defaults to zero, 6222 that is, no holdoff. 6223 6224 rcutorture.test_boost_interval= [KNL] 6225 Interval (s) between each boost test. 6226 6227 rcutorture.test_no_idle_hz= [KNL] 6228 Test RCU's dyntick-idle handling. See also the 6229 rcutorture.shuffle_interval parameter. 6230 6231 rcutorture.torture_type= [KNL] 6232 Specify the RCU implementation to test. 6233 6234 rcutorture.verbose= [KNL] 6235 Enable additional printk() statements. 6236 6237 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL] 6238 Dump ftrace buffer after reporting RCU CPU 6239 stall warning. 6240 6241 rcupdate.rcu_cpu_stall_notifiers= [KNL] 6242 Provide RCU CPU stall notifiers, but see the 6243 warnings in the RCU_CPU_STALL_NOTIFIER Kconfig 6244 option's help text. TL;DR: You almost certainly 6245 do not want rcupdate.rcu_cpu_stall_notifiers. 6246 6247 rcupdate.rcu_cpu_stall_suppress= [KNL] 6248 Suppress RCU CPU stall warning messages. 6249 6250 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL] 6251 Suppress RCU CPU stall warning messages and 6252 rcutorture writer stall warnings that occur 6253 during early boot, that is, during the time 6254 before the init task is spawned. 6255 6256 rcupdate.rcu_cpu_stall_timeout= [KNL] 6257 Set timeout for RCU CPU stall warning messages. 6258 The value is in seconds and the maximum allowed 6259 value is 300 seconds. 6260 6261 rcupdate.rcu_exp_cpu_stall_timeout= [KNL] 6262 Set timeout for expedited RCU CPU stall warning 6263 messages. The value is in milliseconds 6264 and the maximum allowed value is 21000 6265 milliseconds. Please note that this value is 6266 adjusted to an arch timer tick resolution. 6267 Setting this to zero causes the value from 6268 rcupdate.rcu_cpu_stall_timeout to be used (after 6269 conversion from seconds to milliseconds). 6270 6271 rcupdate.rcu_cpu_stall_cputime= [KNL] 6272 Provide statistics on the cputime and count of 6273 interrupts and tasks during the sampling period. For 6274 multiple continuous RCU stalls, all sampling periods 6275 begin at half of the first RCU stall timeout. 6276 6277 rcupdate.rcu_exp_stall_task_details= [KNL] 6278 Print stack dumps of any tasks blocking the 6279 current expedited RCU grace period during an 6280 expedited RCU CPU stall warning. 6281 6282 rcupdate.rcu_expedited= [KNL] 6283 Use expedited grace-period primitives, for 6284 example, synchronize_rcu_expedited() instead 6285 of synchronize_rcu(). This reduces latency, 6286 but can increase CPU utilization, degrade 6287 real-time latency, and degrade energy efficiency. 6288 No effect on CONFIG_TINY_RCU kernels. 6289 6290 rcupdate.rcu_normal= [KNL] 6291 Use only normal grace-period primitives, 6292 for example, synchronize_rcu() instead of 6293 synchronize_rcu_expedited(). This improves 6294 real-time latency, CPU utilization, and 6295 energy efficiency, but can expose users to 6296 increased grace-period latency. This parameter 6297 overrides rcupdate.rcu_expedited. No effect on 6298 CONFIG_TINY_RCU kernels. 6299 6300 rcupdate.rcu_normal_after_boot= [KNL] 6301 Once boot has completed (that is, after 6302 rcu_end_inkernel_boot() has been invoked), use 6303 only normal grace-period primitives. No effect 6304 on CONFIG_TINY_RCU kernels. 6305 6306 But note that CONFIG_PREEMPT_RT=y kernels enables 6307 this kernel boot parameter, forcibly setting 6308 it to the value one, that is, converting any 6309 post-boot attempt at an expedited RCU grace 6310 period to instead use normal non-expedited 6311 grace-period processing. 6312 6313 rcupdate.rcu_task_collapse_lim= [KNL] 6314 Set the maximum number of callbacks present 6315 at the beginning of a grace period that allows 6316 the RCU Tasks flavors to collapse back to using 6317 a single callback queue. This switching only 6318 occurs when rcupdate.rcu_task_enqueue_lim is 6319 set to the default value of -1. 6320 6321 rcupdate.rcu_task_contend_lim= [KNL] 6322 Set the minimum number of callback-queuing-time 6323 lock-contention events per jiffy required to 6324 cause the RCU Tasks flavors to switch to per-CPU 6325 callback queuing. This switching only occurs 6326 when rcupdate.rcu_task_enqueue_lim is set to 6327 the default value of -1. 6328 6329 rcupdate.rcu_task_enqueue_lim= [KNL] 6330 Set the number of callback queues to use for the 6331 RCU Tasks family of RCU flavors. The default 6332 of -1 allows this to be automatically (and 6333 dynamically) adjusted. This parameter is intended 6334 for use in testing. 6335 6336 rcupdate.rcu_task_lazy_lim= [KNL] 6337 Number of callbacks on a given CPU that will 6338 cancel laziness on that CPU. Use -1 to disable 6339 cancellation of laziness, but be advised that 6340 doing so increases the danger of OOM due to 6341 callback flooding. 6342 6343 rcupdate.rcu_task_stall_info= [KNL] 6344 Set initial timeout in jiffies for RCU task stall 6345 informational messages, which give some indication 6346 of the problem for those not patient enough to 6347 wait for ten minutes. Informational messages are 6348 only printed prior to the stall-warning message 6349 for a given grace period. Disable with a value 6350 less than or equal to zero. Defaults to ten 6351 seconds. A change in value does not take effect 6352 until the beginning of the next grace period. 6353 6354 rcupdate.rcu_task_stall_info_mult= [KNL] 6355 Multiplier for time interval between successive 6356 RCU task stall informational messages for a given 6357 RCU tasks grace period. This value is clamped 6358 to one through ten, inclusive. It defaults to 6359 the value three, so that the first informational 6360 message is printed 10 seconds into the grace 6361 period, the second at 40 seconds, the third at 6362 160 seconds, and then the stall warning at 600 6363 seconds would prevent a fourth at 640 seconds. 6364 6365 rcupdate.rcu_task_stall_timeout= [KNL] 6366 Set timeout in jiffies for RCU task stall 6367 warning messages. Disable with a value less 6368 than or equal to zero. Defaults to ten minutes. 6369 A change in value does not take effect until 6370 the beginning of the next grace period. 6371 6372 rcupdate.rcu_tasks_lazy_ms= [KNL] 6373 Set timeout in milliseconds RCU Tasks asynchronous 6374 callback batching for call_rcu_tasks(). 6375 A negative value will take the default. A value 6376 of zero will disable batching. Batching is 6377 always disabled for synchronize_rcu_tasks(). 6378 6379 rcupdate.rcu_self_test= [KNL] 6380 Run the RCU early boot self tests 6381 6382 rdinit= [KNL] 6383 Format: <full_path> 6384 Run specified binary instead of /init from the ramdisk, 6385 used for early userspace startup. See initrd. 6386 6387 rdrand= [X86,EARLY] 6388 force - Override the decision by the kernel to hide the 6389 advertisement of RDRAND support (this affects 6390 certain AMD processors because of buggy BIOS 6391 support, specifically around the suspend/resume 6392 path). 6393 6394 rdt= [HW,X86,RDT] 6395 Turn on/off individual RDT features. List is: 6396 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp, 6397 mba, smba, bmec, abmc, sdciae, energy[:guid], 6398 perf[:guid]. 6399 E.g. to turn on cmt and turn off mba use: 6400 rdt=cmt,!mba 6401 To turn off all energy telemetry monitoring and ensure that 6402 perf telemetry monitoring associated with guid 0x12345 6403 is enabled use: 6404 rdt=!energy,perf:0x12345 6405 6406 reboot= [KNL] 6407 Format (x86 or x86_64): 6408 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \ 6409 [[,]s[mp]#### \ 6410 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \ 6411 [[,]f[orce] 6412 Where reboot_mode is one of warm (soft) or cold (hard) or gpio 6413 (prefix with 'panic_' to set mode for panic 6414 reboot only), 6415 reboot_type is one of bios, acpi, kbd, triple, efi, or pci, 6416 reboot_force is either force or not specified, 6417 reboot_cpu is s[mp]#### with #### being the processor 6418 to be used for rebooting. 6419 6420 acpi 6421 Use the ACPI RESET_REG in the FADT. If ACPI is not 6422 configured or the ACPI reset does not work, the reboot 6423 path attempts the reset using the keyboard controller. 6424 6425 bios 6426 Use the CPU reboot vector for warm reset 6427 6428 cold 6429 Set the cold reboot flag 6430 6431 default 6432 There are some built-in platform specific "quirks" 6433 - you may see: "reboot: <name> series board detected. 6434 Selecting <type> for reboots." In the case where you 6435 think the quirk is in error (e.g. you have newer BIOS, 6436 or newer board) using this option will ignore the 6437 built-in quirk table, and use the generic default 6438 reboot actions. 6439 6440 efi 6441 Use efi reset_system runtime service. If EFI is not 6442 configured or the EFI reset does not work, the reboot 6443 path attempts the reset using the keyboard controller. 6444 6445 force 6446 Don't stop other CPUs on reboot. This can make reboot 6447 more reliable in some cases. 6448 6449 kbd 6450 Use the keyboard controller. cold reset (default) 6451 6452 pci 6453 Use a write to the PCI config space register 0xcf9 to 6454 trigger reboot. 6455 6456 triple 6457 Force a triple fault (init) 6458 6459 warm 6460 Don't set the cold reboot flag 6461 6462 Using warm reset will be much faster especially on big 6463 memory systems because the BIOS will not go through 6464 the memory check. Disadvantage is that not all 6465 hardware will be completely reinitialized on reboot so 6466 there may be boot problems on some systems. 6467 6468 6469 refscale.holdoff= [KNL] 6470 Set test-start holdoff period. The purpose of 6471 this parameter is to delay the start of the 6472 test until boot completes in order to avoid 6473 interference. 6474 6475 refscale.lookup_instances= [KNL] 6476 Number of data elements to use for the forms of 6477 SLAB_TYPESAFE_BY_RCU testing. A negative number 6478 is negated and multiplied by nr_cpu_ids, while 6479 zero specifies nr_cpu_ids. 6480 6481 refscale.loops= [KNL] 6482 Set the number of loops over the synchronization 6483 primitive under test. Increasing this number 6484 reduces noise due to loop start/end overhead, 6485 but the default has already reduced the per-pass 6486 noise to a handful of picoseconds on ca. 2020 6487 x86 laptops. 6488 6489 refscale.nreaders= [KNL] 6490 Set number of readers. The default value of -1 6491 selects N, where N is roughly 75% of the number 6492 of CPUs. A value of zero is an interesting choice. 6493 6494 refscale.nruns= [KNL] 6495 Set number of runs, each of which is dumped onto 6496 the console log. 6497 6498 refscale.readdelay= [KNL] 6499 Set the read-side critical-section duration, 6500 measured in microseconds. 6501 6502 refscale.scale_type= [KNL] 6503 Specify the read-protection implementation to test. 6504 6505 refscale.shutdown= [KNL] 6506 Shut down the system at the end of the performance 6507 test. This defaults to 1 (shut it down) when 6508 refscale is built into the kernel and to 0 (leave 6509 it running) when refscale is built as a module. 6510 6511 refscale.verbose= [KNL] 6512 Enable additional printk() statements. 6513 6514 refscale.verbose_batched= [KNL] 6515 Batch the additional printk() statements. If zero 6516 (the default) or negative, print everything. Otherwise, 6517 print every Nth verbose statement, where N is the value 6518 specified. 6519 6520 regulator_ignore_unused 6521 [REGULATOR] 6522 Prevents regulator framework from disabling regulators 6523 that are unused, due no driver claiming them. This may 6524 be useful for debug and development, but should not be 6525 needed on a platform with proper driver support. 6526 6527 relax_domain_level= 6528 [KNL, SMP] Set scheduler's default relax_domain_level. 6529 See Documentation/admin-guide/cgroup-v1/cpusets.rst. 6530 6531 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory 6532 Format: <base1>,<size1>[,<base2>,<size2>,...] 6533 Reserve I/O ports or memory so the kernel won't use 6534 them. If <base> is less than 0x10000, the region 6535 is assumed to be I/O ports; otherwise it is memory. 6536 6537 reserve_mem= [RAM] 6538 Format: nn[KMG]:<align>:<label> 6539 Reserve physical memory and label it with a name that 6540 other subsystems can use to access it. This is typically 6541 used for systems that do not wipe the RAM, and this command 6542 line will try to reserve the same physical memory on 6543 soft reboots. Note, it is not guaranteed to be the same 6544 location. For example, if anything about the system changes 6545 or if booting a different kernel. It can also fail if KASLR 6546 places the kernel at the location of where the RAM reservation 6547 was from a previous boot, the new reservation will be at a 6548 different location. 6549 Any subsystem using this feature must add a way to verify 6550 that the contents of the physical memory is from a previous 6551 boot, as there may be cases where the memory will not be 6552 located at the same location. 6553 6554 The format is size:align:label for example, to request 6555 12 megabytes of 4096 alignment for ramoops: 6556 6557 reserve_mem=12M:4096:oops ramoops.mem_name=oops 6558 6559 reservetop= [X86-32,EARLY] 6560 Format: nn[KMG] 6561 Reserves a hole at the top of the kernel virtual 6562 address space. 6563 6564 reset_devices [KNL] Force drivers to reset the underlying device 6565 during initialization. 6566 6567 resume= [SWSUSP] 6568 Specify the partition device for software suspend 6569 Format: 6570 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>} 6571 6572 resume_offset= [SWSUSP] 6573 Specify the offset from the beginning of the partition 6574 given by "resume=" at which the swap header is located, 6575 in <PAGE_SIZE> units (needed only for swap files). 6576 See Documentation/power/swsusp-and-swap-files.rst 6577 6578 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to 6579 read the resume files 6580 6581 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up. 6582 Useful for devices that are detected asynchronously 6583 (e.g. USB and MMC devices). 6584 6585 retain_initrd [RAM] Keep initrd memory after extraction. After boot, it will 6586 be accessible via /sys/firmware/initrd. 6587 6588 retbleed= [X86] Control mitigation of RETBleed (Arbitrary 6589 Speculative Code Execution with Return Instructions) 6590 vulnerability. 6591 6592 AMD-based UNRET and IBPB mitigations alone do not stop 6593 sibling threads from influencing the predictions of other 6594 sibling threads. For that reason, STIBP is used on pro- 6595 cessors that support it, and mitigate SMT on processors 6596 that don't. 6597 6598 off - no mitigation 6599 auto - automatically select a mitigation 6600 auto,nosmt - automatically select a mitigation, 6601 disabling SMT if necessary for 6602 the full mitigation (only on Zen1 6603 and older without STIBP). 6604 ibpb - On AMD, mitigate short speculation 6605 windows on basic block boundaries too. 6606 Safe, highest perf impact. It also 6607 enables STIBP if present. Not suitable 6608 on Intel. 6609 ibpb,nosmt - Like "ibpb" above but will disable SMT 6610 when STIBP is not available. This is 6611 the alternative for systems which do not 6612 have STIBP. 6613 unret - Force enable untrained return thunks, 6614 only effective on AMD f15h-f17h based 6615 systems. 6616 unret,nosmt - Like unret, but will disable SMT when STIBP 6617 is not available. This is the alternative for 6618 systems which do not have STIBP. 6619 6620 Selecting 'auto' will choose a mitigation method at run 6621 time according to the CPU. 6622 6623 Not specifying this option is equivalent to retbleed=auto. 6624 6625 rfkill.default_state= 6626 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm, 6627 etc. communication is blocked by default. 6628 1 Unblocked. 6629 6630 rfkill.master_switch_mode= 6631 0 The "airplane mode" button does nothing. 6632 1 The "airplane mode" button toggles between everything 6633 blocked and the previous configuration. 6634 2 The "airplane mode" button toggles between everything 6635 blocked and everything unblocked. 6636 6637 ring3mwait=disable 6638 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported 6639 CPUs. 6640 6641 riscv_isa_fallback [RISCV,EARLY] 6642 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit 6643 falling back to detecting extension support by parsing 6644 "riscv,isa" property on devicetree systems when the 6645 replacement properties are not found. See the Kconfig 6646 entry for RISCV_ISA_FALLBACK. 6647 6648 riscv_nousercfi= 6649 all Disable user CFI ABI to userspace even if cpu extension 6650 are available. 6651 bcfi Disable user backward CFI ABI to userspace even if 6652 the shadow stack extension is available. 6653 fcfi Disable user forward CFI ABI to userspace even if the 6654 landing pad extension is available. 6655 6656 ro [KNL] Mount root device read-only on boot 6657 6658 rodata= [KNL,EARLY] 6659 on Mark read-only kernel memory as read-only (default). 6660 off Leave read-only kernel memory writable for debugging. 6661 noalias Mark read-only kernel memory as read-only but retain 6662 writable aliases in the direct map for regions outside 6663 of the kernel image. [arm64] 6664 6665 rockchip.usb_uart 6666 [EARLY] 6667 Enable the uart passthrough on the designated usb port 6668 on Rockchip SoCs. When active, the signals of the 6669 debug-uart get routed to the D+ and D- pins of the usb 6670 port and the regular usb controller gets disabled. 6671 6672 root= [KNL] Root filesystem 6673 Usually this is a block device specifier of some kind, 6674 see the early_lookup_bdev comment in 6675 block/early-lookup.c for details. 6676 Alternatively this can be "ram" for the legacy initial 6677 ramdisk, "nfs" and "cifs" for root on a network file 6678 system, or "mtd" and "ubi" for mounting from raw flash. 6679 6680 rootdelay= [KNL] Delay (in seconds) to pause before attempting to 6681 mount the root filesystem 6682 6683 rootflags= [KNL] Set root filesystem mount option string 6684 6685 rseq_slice_ext= [KNL] RSEQ based time slice extension 6686 Format: boolean 6687 Control enablement of RSEQ based time slice extension. 6688 Default is 'on'. 6689 6690 initramfs_options= [KNL] 6691 Specify mount options for the initramfs mount. 6692 6693 rootfstype= [KNL] Set root filesystem type 6694 6695 rootwait [KNL] Wait (indefinitely) for root device to show up. 6696 Useful for devices that are detected asynchronously 6697 (e.g. USB and MMC devices). 6698 6699 rootwait= [KNL] Maximum time (in seconds) to wait for root device 6700 to show up before attempting to mount the root 6701 filesystem. 6702 6703 rproc_mem=nn[KMG][@address] 6704 [KNL,ARM,CMA] Remoteproc physical memory block. 6705 Memory area to be used by remote processor image, 6706 managed by CMA. 6707 6708 rseq_debug= [KNL] Enable or disable restartable sequence 6709 debug mode. Defaults to CONFIG_RSEQ_DEBUG_DEFAULT_ENABLE. 6710 Format: <bool> 6711 6712 rt_group_sched= [KNL] Enable or disable SCHED_RR/FIFO group scheduling 6713 when CONFIG_RT_GROUP_SCHED=y. Defaults to 6714 !CONFIG_RT_GROUP_SCHED_DEFAULT_DISABLED. 6715 Format: <bool> 6716 6717 rw [KNL] Mount root device read-write on boot 6718 6719 S [KNL] Run init in single mode 6720 6721 s390_iommu= [HW,S390] 6722 Set s390 IOTLB flushing mode 6723 strict 6724 With strict flushing every unmap operation will result 6725 in an IOTLB flush. Default is lazy flushing before 6726 reuse, which is faster. Deprecated, equivalent to 6727 iommu.strict=1. 6728 6729 s390_iommu_aperture= [KNL,S390] 6730 Specifies the size of the per device DMA address space 6731 accessible through the DMA and IOMMU APIs as a decimal 6732 factor of the size of main memory. 6733 The default is 1 meaning that one can concurrently use 6734 as many DMA addresses as physical memory is installed, 6735 if supported by hardware, and thus map all of memory 6736 once. With a value of 2 one can map all of memory twice 6737 and so on. As a special case a factor of 0 imposes no 6738 restrictions other than those given by hardware at the 6739 cost of significant additional memory use for tables. 6740 6741 6742 sched_proxy_exec= [KNL] 6743 Enables or disables "proxy execution" style 6744 solution to mutex-based priority inversion. 6745 Format: <bool> 6746 6747 sched_verbose [KNL,EARLY] Enables verbose scheduler debug messages. 6748 6749 schedstats= [KNL,X86] Enable or disable scheduled statistics. 6750 Allowed values are enable and disable. This feature 6751 incurs a small amount of overhead in the scheduler 6752 but is useful for debugging and performance tuning. 6753 6754 sched_thermal_decay_shift= 6755 [Deprecated] 6756 [KNL, SMP] Set a decay shift for scheduler thermal 6757 pressure signal. Thermal pressure signal follows the 6758 default decay period of other scheduler pelt 6759 signals(usually 32 ms but configurable). Setting 6760 sched_thermal_decay_shift will left shift the decay 6761 period for the thermal pressure signal by the shift 6762 value. 6763 i.e. with the default pelt decay period of 32 ms 6764 sched_thermal_decay_shift thermal pressure decay pr 6765 1 64 ms 6766 2 128 ms 6767 and so on. 6768 Format: integer between 0 and 10 6769 Default is 0. 6770 6771 scftorture.holdoff= [KNL] 6772 Number of seconds to hold off before starting 6773 test. Defaults to zero for module insertion and 6774 to 10 seconds for built-in smp_call_function() 6775 tests. 6776 6777 scftorture.longwait= [KNL] 6778 Request ridiculously long waits randomly selected 6779 up to the chosen limit in seconds. Zero (the 6780 default) disables this feature. Please note 6781 that requesting even small non-zero numbers of 6782 seconds can result in RCU CPU stall warnings, 6783 softlockup complaints, and so on. 6784 6785 scftorture.nthreads= [KNL] 6786 Number of kthreads to spawn to invoke the 6787 smp_call_function() family of functions. 6788 The default of -1 specifies a number of kthreads 6789 equal to the number of CPUs. 6790 6791 scftorture.onoff_holdoff= [KNL] 6792 Number seconds to wait after the start of the 6793 test before initiating CPU-hotplug operations. 6794 6795 scftorture.onoff_interval= [KNL] 6796 Number seconds to wait between successive 6797 CPU-hotplug operations. Specifying zero (which 6798 is the default) disables CPU-hotplug operations. 6799 6800 scftorture.shutdown_secs= [KNL] 6801 The number of seconds following the start of the 6802 test after which to shut down the system. The 6803 default of zero avoids shutting down the system. 6804 Non-zero values are useful for automated tests. 6805 6806 scftorture.stat_interval= [KNL] 6807 The number of seconds between outputting the 6808 current test statistics to the console. A value 6809 of zero disables statistics output. 6810 6811 scftorture.stutter_cpus= [KNL] 6812 The number of jiffies to wait between each change 6813 to the set of CPUs under test. 6814 6815 scftorture.use_cpus_read_lock= [KNL] 6816 Use use_cpus_read_lock() instead of the default 6817 preempt_disable() to disable CPU hotplug 6818 while invoking one of the smp_call_function*() 6819 functions. 6820 6821 scftorture.verbose= [KNL] 6822 Enable additional printk() statements. 6823 6824 scftorture.weight_single= [KNL] 6825 The probability weighting to use for the 6826 smp_call_function_single() function with a zero 6827 "wait" parameter. A value of -1 selects the 6828 default if all other weights are -1. However, 6829 if at least one weight has some other value, a 6830 value of -1 will instead select a weight of zero. 6831 6832 scftorture.weight_single_wait= [KNL] 6833 The probability weighting to use for the 6834 smp_call_function_single() function with a 6835 non-zero "wait" parameter. See weight_single. 6836 6837 scftorture.weight_many= [KNL] 6838 The probability weighting to use for the 6839 smp_call_function_many() function with a zero 6840 "wait" parameter. See weight_single. 6841 Note well that setting a high probability for 6842 this weighting can place serious IPI load 6843 on the system. 6844 6845 scftorture.weight_many_wait= [KNL] 6846 The probability weighting to use for the 6847 smp_call_function_many() function with a 6848 non-zero "wait" parameter. See weight_single 6849 and weight_many. 6850 6851 scftorture.weight_all= [KNL] 6852 The probability weighting to use for the 6853 smp_call_function_all() function with a zero 6854 "wait" parameter. See weight_single and 6855 weight_many. 6856 6857 scftorture.weight_all_wait= [KNL] 6858 The probability weighting to use for the 6859 smp_call_function_all() function with a 6860 non-zero "wait" parameter. See weight_single 6861 and weight_many. 6862 6863 sdw_mclk_divider=[SDW] 6864 Specify the MCLK divider for Intel SoundWire buses in 6865 case the BIOS does not provide the clock rate properly. 6866 6867 skew_tick= [KNL,EARLY] Offset the periodic timer tick per cpu to mitigate 6868 xtime_lock contention on larger systems, and/or RCU lock 6869 contention on all systems with CONFIG_MAXSMP set. 6870 Format: { "0" | "1" } 6871 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1" 6872 1 -- enable. 6873 Note: increases power consumption, thus should only be 6874 enabled if running jitter sensitive (HPC/RT) workloads. 6875 6876 security= [SECURITY] Choose a legacy "major" security module to 6877 enable at boot. This has been deprecated by the 6878 "lsm=" parameter. 6879 6880 selinux= [SELINUX] Disable or enable SELinux at boot time. 6881 Format: { "0" | "1" } 6882 See security/selinux/Kconfig help text. 6883 0 -- disable. 6884 1 -- enable. 6885 Default value is 1. 6886 6887 serialnumber [BUGS=X86-32] 6888 6889 sev=option[,option...] [X86-64] 6890 6891 debug 6892 Enable debug messages. 6893 6894 nosnp 6895 Do not enable SEV-SNP (applies to host/hypervisor 6896 only). Setting 'nosnp' avoids the RMP check overhead 6897 in memory accesses when users do not want to run 6898 SEV-SNP guests. 6899 6900 shapers= [NET] 6901 Maximal number of shapers. 6902 6903 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller 6904 Limit apic dumping. The parameter defines the maximal 6905 number of local apics being dumped. Also it is possible 6906 to set it to "all" by meaning -- no limit here. 6907 Format: { 1 (default) | 2 | ... | all }. 6908 The parameter valid if only apic=debug or 6909 apic=verbose is specified. 6910 Example: apic=debug show_lapic=all 6911 6912 slab_debug[=options[,slabs][;[options[,slabs]]...] [MM] 6913 Enabling slab_debug allows one to determine the 6914 culprit if slab objects become corrupted. Enabling 6915 slab_debug can create guard zones around objects and 6916 may poison objects when not in use. Also tracks the 6917 last alloc / free. For more information see 6918 Documentation/admin-guide/mm/slab.rst. 6919 (slub_debug legacy name also accepted for now) 6920 6921 Using this option implies the "no_hash_pointers" 6922 option which can be undone by adding the 6923 "hash_pointers=always" option. 6924 6925 slab_max_order= [MM] 6926 Determines the maximum allowed order for slabs. 6927 A high setting may cause OOMs due to memory 6928 fragmentation. For more information see 6929 Documentation/admin-guide/mm/slab.rst. 6930 (slub_max_order legacy name also accepted for now) 6931 6932 slab_merge [MM] 6933 Enable merging of slabs with similar size when the 6934 kernel is built without CONFIG_SLAB_MERGE_DEFAULT. 6935 (slub_merge legacy name also accepted for now) 6936 6937 slab_min_objects= [MM] 6938 The minimum number of objects per slab. SLUB will 6939 increase the slab order up to slab_max_order to 6940 generate a sufficiently large slab able to contain 6941 the number of objects indicated. The higher the number 6942 of objects the smaller the overhead of tracking slabs 6943 and the less frequently locks need to be acquired. 6944 For more information see 6945 Documentation/admin-guide/mm/slab.rst. 6946 (slub_min_objects legacy name also accepted for now) 6947 6948 slab_min_order= [MM] 6949 Determines the minimum page order for slabs. Must be 6950 lower or equal to slab_max_order. For more information see 6951 Documentation/admin-guide/mm/slab.rst. 6952 (slub_min_order legacy name also accepted for now) 6953 6954 slab_nomerge [MM] 6955 Disable merging of slabs with similar size. May be 6956 necessary if there is some reason to distinguish 6957 allocs to different slabs, especially in hardened 6958 environments where the risk of heap overflows and 6959 layout control by attackers can usually be 6960 frustrated by disabling merging. This will reduce 6961 most of the exposure of a heap attack to a single 6962 cache (risks via metadata attacks are mostly 6963 unchanged). Debug options disable merging on their 6964 own. 6965 For more information see 6966 Documentation/admin-guide/mm/slab.rst. 6967 (slub_nomerge legacy name also accepted for now) 6968 6969 slab_strict_numa [MM] 6970 Support memory policies on a per object level 6971 in the slab allocator. The default is for memory 6972 policies to be applied at the folio level when 6973 a new folio is needed or a partial folio is 6974 retrieved from the lists. Increases overhead 6975 in the slab fastpaths but gains more accurate 6976 NUMA kernel object placement which helps with slow 6977 interconnects in NUMA systems. 6978 6979 slram= [HW,MTD] 6980 6981 smart2= [HW] 6982 Format: <io1>[,<io2>[,...,<io8>]] 6983 6984 smp.csd_lock_timeout= [KNL] 6985 Specify the period of time in milliseconds 6986 that smp_call_function() and friends will wait 6987 for a CPU to release the CSD lock. This is 6988 useful when diagnosing bugs involving CPUs 6989 disabling interrupts for extended periods 6990 of time. Defaults to 5,000 milliseconds, and 6991 setting a value of zero disables this feature. 6992 This feature may be more efficiently disabled 6993 using the csdlock_debug- kernel parameter. 6994 6995 smp.panic_on_ipistall= [KNL] 6996 If a csd_lock_timeout extends for more than 6997 the specified number of milliseconds, panic the 6998 system. By default, let CSD-lock acquisition 6999 take as long as they take. Specifying 300,000 7000 for this value provides a 5-minute timeout. 7001 7002 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices 7003 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port 7004 smsc-ircc2.ircc_sir= [HW] SIR base I/O port 7005 smsc-ircc2.ircc_fir= [HW] FIR base I/O port 7006 smsc-ircc2.ircc_irq= [HW] IRQ line 7007 smsc-ircc2.ircc_dma= [HW] DMA channel 7008 smsc-ircc2.ircc_transceiver= [HW] Transceiver type: 7009 0: Toshiba Satellite 1800 (GP data pin select) 7010 1: Fast pin select (default) 7011 2: ATC IRMode 7012 7013 smt= [KNL,MIPS,S390,EARLY] Set the maximum number of threads 7014 (logical CPUs) to use per physical CPU on systems 7015 capable of symmetric multithreading (SMT). Will 7016 be capped to the actual hardware limit. 7017 Format: <integer> 7018 Default: -1 (no limit) 7019 7020 softlockup_panic= 7021 [KNL] Should the soft-lockup detector generate panics. 7022 Format: <int> 7023 7024 A value of non-zero instructs the soft-lockup detector 7025 to panic the machine when a soft-lockup duration exceeds 7026 N thresholds. It is also controlled by the kernel.softlockup_panic 7027 sysctl and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the 7028 respective build-time switch to that functionality. 7029 7030 softlockup_all_cpu_backtrace= 7031 [KNL] Should the soft-lockup detector generate 7032 backtraces on all cpus. 7033 Format: 0 | 1 7034 7035 sonypi.*= [HW] Sony Programmable I/O Control Device driver 7036 See Documentation/admin-guide/laptops/sonypi.rst 7037 7038 spectre_bhi= [X86] Control mitigation of Branch History Injection 7039 (BHI) vulnerability. This setting affects the 7040 deployment of the HW BHI control and the SW BHB 7041 clearing sequence. 7042 7043 on - (default) Enable the HW or SW mitigation as 7044 needed. This protects the kernel from 7045 both syscalls and VMs. 7046 vmexit - On systems which don't have the HW mitigation 7047 available, enable the SW mitigation on vmexit 7048 ONLY. On such systems, the host kernel is 7049 protected from VM-originated BHI attacks, but 7050 may still be vulnerable to syscall attacks. 7051 off - Disable the mitigation. 7052 7053 spectre_v2= [X86,EARLY] Control mitigation of Spectre variant 2 7054 (indirect branch speculation) vulnerability. 7055 The default operation protects the kernel from 7056 user space attacks. 7057 7058 on - unconditionally enable, implies 7059 spectre_v2_user=on 7060 off - unconditionally disable, implies 7061 spectre_v2_user=off 7062 auto - kernel detects whether your CPU model is 7063 vulnerable 7064 7065 Selecting 'on' will, and 'auto' may, choose a 7066 mitigation method at run time according to the 7067 CPU, the available microcode, the setting of the 7068 CONFIG_MITIGATION_RETPOLINE configuration option, 7069 and the compiler with which the kernel was built. 7070 7071 Selecting 'on' will also enable the mitigation 7072 against user space to user space task attacks. 7073 Selecting specific mitigation does not force enable 7074 user mitigations. 7075 7076 Selecting 'off' will disable both the kernel and 7077 the user space protections. 7078 7079 Specific mitigations can also be selected manually: 7080 7081 retpoline - replace indirect branches 7082 retpoline,generic - Retpolines 7083 retpoline,lfence - LFENCE; indirect branch 7084 retpoline,amd - alias for retpoline,lfence 7085 eibrs - Enhanced/Auto IBRS 7086 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines 7087 eibrs,lfence - Enhanced/Auto IBRS + LFENCE 7088 ibrs - use IBRS to protect kernel 7089 7090 Not specifying this option is equivalent to 7091 spectre_v2=auto. 7092 7093 spectre_v2_user= 7094 [X86] Control mitigation of Spectre variant 2 7095 (indirect branch speculation) vulnerability between 7096 user space tasks 7097 7098 on - Unconditionally enable mitigations. Is 7099 enforced by spectre_v2=on 7100 7101 off - Unconditionally disable mitigations. Is 7102 enforced by spectre_v2=off 7103 7104 prctl - Indirect branch speculation is enabled, 7105 but mitigation can be enabled via prctl 7106 per thread. The mitigation control state 7107 is inherited on fork. 7108 7109 prctl,ibpb 7110 - Like "prctl" above, but only STIBP is 7111 controlled per thread. IBPB is issued 7112 always when switching between different user 7113 space processes. 7114 7115 seccomp 7116 - Same as "prctl" above, but all seccomp 7117 threads will enable the mitigation unless 7118 they explicitly opt out. 7119 7120 seccomp,ibpb 7121 - Like "seccomp" above, but only STIBP is 7122 controlled per thread. IBPB is issued 7123 always when switching between different 7124 user space processes. 7125 7126 auto - Kernel selects the mitigation depending on 7127 the available CPU features and vulnerability. 7128 7129 Default mitigation: "prctl" 7130 7131 Not specifying this option is equivalent to 7132 spectre_v2_user=auto. 7133 7134 spec_rstack_overflow= 7135 [X86,EARLY] Control RAS overflow mitigation on AMD Zen CPUs 7136 7137 off - Disable mitigation 7138 microcode - Enable microcode mitigation only 7139 safe-ret - Enable sw-only safe RET mitigation (default) 7140 ibpb - Enable mitigation by issuing IBPB on 7141 kernel entry 7142 ibpb-vmexit - Issue IBPB only on VMEXIT 7143 (cloud-specific mitigation) 7144 7145 spec_store_bypass_disable= 7146 [HW,EARLY] Control Speculative Store Bypass (SSB) Disable mitigation 7147 (Speculative Store Bypass vulnerability) 7148 7149 Certain CPUs are vulnerable to an exploit against a 7150 a common industry wide performance optimization known 7151 as "Speculative Store Bypass" in which recent stores 7152 to the same memory location may not be observed by 7153 later loads during speculative execution. The idea 7154 is that such stores are unlikely and that they can 7155 be detected prior to instruction retirement at the 7156 end of a particular speculation execution window. 7157 7158 In vulnerable processors, the speculatively forwarded 7159 store can be used in a cache side channel attack, for 7160 example to read memory to which the attacker does not 7161 directly have access (e.g. inside sandboxed code). 7162 7163 This parameter controls whether the Speculative Store 7164 Bypass optimization is used. 7165 7166 On x86 the options are: 7167 7168 on - Unconditionally disable Speculative Store Bypass 7169 off - Unconditionally enable Speculative Store Bypass 7170 auto - Kernel detects whether the CPU model contains an 7171 implementation of Speculative Store Bypass and 7172 picks the most appropriate mitigation. If the 7173 CPU is not vulnerable, "off" is selected. If the 7174 CPU is vulnerable the default mitigation is 7175 architecture and Kconfig dependent. See below. 7176 prctl - Control Speculative Store Bypass per thread 7177 via prctl. Speculative Store Bypass is enabled 7178 for a process by default. The state of the control 7179 is inherited on fork. 7180 seccomp - Same as "prctl" above, but all seccomp threads 7181 will disable SSB unless they explicitly opt out. 7182 7183 Default mitigations: 7184 X86: "prctl" 7185 7186 On powerpc the options are: 7187 7188 on,auto - On Power8 and Power9 insert a store-forwarding 7189 barrier on kernel entry and exit. On Power7 7190 perform a software flush on kernel entry and 7191 exit. 7192 off - No action. 7193 7194 Not specifying this option is equivalent to 7195 spec_store_bypass_disable=auto. 7196 7197 split_llc= 7198 [X86,EARLY] Split the LLC N-ways 7199 7200 When set, the LLC is split this many ways by matching 7201 'core_id % n'. This is setup before SMP bringup and 7202 used during SMP bringup before it knows the full 7203 topology. If your core count doesn't nicely divide by 7204 the number given, you get to keep the pieces. 7205 7206 This is mostly a debug feature to emulate multiple LLCs 7207 on hardware that only have a single LLC. 7208 7209 split_lock_detect= 7210 [X86] Enable split lock detection or bus lock detection 7211 7212 When enabled (and if hardware support is present), atomic 7213 instructions that access data across cache line 7214 boundaries will result in an alignment check exception 7215 for split lock detection or a debug exception for 7216 bus lock detection. 7217 7218 off - not enabled 7219 7220 warn - the kernel will emit rate-limited warnings 7221 about applications triggering the #AC 7222 exception or the #DB exception. This mode is 7223 the default on CPUs that support split lock 7224 detection or bus lock detection. Default 7225 behavior is by #AC if both features are 7226 enabled in hardware. 7227 7228 fatal - the kernel will send SIGBUS to applications 7229 that trigger the #AC exception or the #DB 7230 exception. Default behavior is by #AC if 7231 both features are enabled in hardware. 7232 7233 ratelimit:N - 7234 Set system wide rate limit to N bus locks 7235 per second for bus lock detection. 7236 0 < N <= 1000. 7237 7238 N/A for split lock detection. 7239 7240 7241 If an #AC exception is hit in the kernel or in 7242 firmware (i.e. not while executing in user mode) 7243 the kernel will oops in either "warn" or "fatal" 7244 mode. 7245 7246 #DB exception for bus lock is triggered only when 7247 CPL > 0. 7248 7249 srbds= [X86,INTEL,EARLY] 7250 Control the Special Register Buffer Data Sampling 7251 (SRBDS) mitigation. 7252 7253 Certain CPUs are vulnerable to an MDS-like 7254 exploit which can leak bits from the random 7255 number generator. 7256 7257 By default, this issue is mitigated by 7258 microcode. However, the microcode fix can cause 7259 the RDRAND and RDSEED instructions to become 7260 much slower. Among other effects, this will 7261 result in reduced throughput from /dev/urandom. 7262 7263 The microcode mitigation can be disabled with 7264 the following option: 7265 7266 off: Disable mitigation and remove 7267 performance impact to RDRAND and RDSEED 7268 7269 srcutree.big_cpu_lim [KNL] 7270 Specifies the number of CPUs constituting a 7271 large system, such that srcu_struct structures 7272 should immediately allocate an srcu_node array. 7273 This kernel-boot parameter defaults to 128, 7274 but takes effect only when the low-order four 7275 bits of srcutree.convert_to_big is equal to 3 7276 (decide at boot). 7277 7278 srcutree.convert_to_big [KNL] 7279 Specifies under what conditions an SRCU tree 7280 srcu_struct structure will be converted to big 7281 form, that is, with an rcu_node tree: 7282 7283 0: Never. 7284 1: At init_srcu_struct() time. 7285 2: When rcutorture decides to. 7286 3: Decide at boot time (default). 7287 0x1X: Above plus if high contention. 7288 7289 Either way, the srcu_node tree will be sized based 7290 on the actual runtime number of CPUs (nr_cpu_ids) 7291 instead of the compile-time CONFIG_NR_CPUS. 7292 7293 srcutree.counter_wrap_check [KNL] 7294 Specifies how frequently to check for 7295 grace-period sequence counter wrap for the 7296 srcu_data structure's ->srcu_gp_seq_needed field. 7297 The greater the number of bits set in this kernel 7298 parameter, the less frequently counter wrap will 7299 be checked for. Note that the bottom two bits 7300 are ignored. 7301 7302 srcutree.exp_holdoff [KNL] 7303 Specifies how many nanoseconds must elapse 7304 since the end of the last SRCU grace period for 7305 a given srcu_struct until the next normal SRCU 7306 grace period will be considered for automatic 7307 expediting. Set to zero to disable automatic 7308 expediting. 7309 7310 srcutree.srcu_max_nodelay [KNL] 7311 Specifies the number of no-delay instances 7312 per jiffy for which the SRCU grace period 7313 worker thread will be rescheduled with zero 7314 delay. Beyond this limit, worker thread will 7315 be rescheduled with a sleep delay of one jiffy. 7316 7317 srcutree.srcu_max_nodelay_phase [KNL] 7318 Specifies the per-grace-period phase, number of 7319 non-sleeping polls of readers. Beyond this limit, 7320 grace period worker thread will be rescheduled 7321 with a sleep delay of one jiffy, between each 7322 rescan of the readers, for a grace period phase. 7323 7324 srcutree.srcu_retry_check_delay [KNL] 7325 Specifies number of microseconds of non-sleeping 7326 delay between each non-sleeping poll of readers. 7327 7328 srcutree.small_contention_lim [KNL] 7329 Specifies the number of update-side contention 7330 events per jiffy will be tolerated before 7331 initiating a conversion of an srcu_struct 7332 structure to big form. Note that the value of 7333 srcutree.convert_to_big must have the 0x10 bit 7334 set for contention-based conversions to occur. 7335 7336 ssbd= [ARM64,HW,EARLY] 7337 Speculative Store Bypass Disable control 7338 7339 On CPUs that are vulnerable to the Speculative 7340 Store Bypass vulnerability and offer a 7341 firmware based mitigation, this parameter 7342 indicates how the mitigation should be used: 7343 7344 force-on: Unconditionally enable mitigation for 7345 for both kernel and userspace 7346 force-off: Unconditionally disable mitigation for 7347 for both kernel and userspace 7348 kernel: Always enable mitigation in the 7349 kernel, and offer a prctl interface 7350 to allow userspace to register its 7351 interest in being mitigated too. 7352 7353 stack_guard_gap= [MM] 7354 override the default stack gap protection. The value 7355 is in page units and it defines how many pages prior 7356 to (for stacks growing down) resp. after (for stacks 7357 growing up) the main stack are reserved for no other 7358 mapping. Default value is 256 pages. 7359 7360 stack_depot_disable= [KNL,EARLY] 7361 Setting this to true through kernel command line will 7362 disable the stack depot thereby saving the static memory 7363 consumed by the stack hash table. By default this is set 7364 to false. 7365 7366 stack_depot_max_pools= [KNL,EARLY] 7367 Specify the maximum number of pools to use for storing 7368 stack traces. Pools are allocated on-demand up to this 7369 limit. Default value is 8191 pools. 7370 7371 stacktrace [FTRACE] 7372 Enable the stack tracer on boot up. 7373 7374 stacktrace_filter=[function-list] 7375 [FTRACE] Limit the functions that the stack tracer 7376 will trace at boot up. function-list is a comma-separated 7377 list of functions. This list can be changed at run 7378 time by the stack_trace_filter file in the debugfs 7379 tracing directory. Note, this enables stack tracing 7380 and the stacktrace above is not needed. 7381 7382 sti= [PARISC,HW] 7383 Format: <num> 7384 Set the STI (builtin display/keyboard on the HP-PARISC 7385 machines) console (graphic card) which should be used 7386 as the initial boot-console. 7387 See also comment in drivers/video/sticore.c. 7388 7389 sti_font= [HW] 7390 See comment in drivers/video/sticore.c. 7391 7392 stifb= [HW] 7393 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]] 7394 7395 strict_sas_size= 7396 [X86] 7397 Format: <bool> 7398 Enable or disable strict sigaltstack size checks 7399 against the required signal frame size which 7400 depends on the supported FPU features. This can 7401 be used to filter out binaries which have 7402 not yet been made aware of AT_MINSIGSTKSZ. 7403 7404 stress_hpt [PPC,EARLY] 7405 Limits the number of kernel HPT entries in the hash 7406 page table to increase the rate of hash page table 7407 faults on kernel addresses. 7408 7409 stress_slb [PPC,EARLY] 7410 Limits the number of kernel SLB entries, and flushes 7411 them frequently to increase the rate of SLB faults 7412 on kernel addresses. 7413 7414 no_slb_preload [PPC,EARLY] 7415 Disables slb preloading for userspace. 7416 7417 sunrpc.min_resvport= 7418 sunrpc.max_resvport= 7419 [NFS,SUNRPC] 7420 SunRPC servers often require that client requests 7421 originate from a privileged port (i.e. a port in the 7422 range 0 < portnr < 1024). 7423 An administrator who wishes to reserve some of these 7424 ports for other uses may adjust the range that the 7425 kernel's sunrpc client considers to be privileged 7426 using these two parameters to set the minimum and 7427 maximum port values. 7428 7429 sunrpc.svc_rpc_per_connection_limit= 7430 [NFS,SUNRPC] 7431 Limit the number of requests that the server will 7432 process in parallel from a single connection. 7433 The default value is 0 (no limit). 7434 7435 sunrpc.pool_mode= 7436 [NFS] 7437 Control how the NFS server code allocates CPUs to 7438 service thread pools. Depending on how many NICs 7439 you have and where their interrupts are bound, this 7440 option will affect which CPUs will do NFS serving. 7441 Note: this parameter cannot be changed while the 7442 NFS server is running. 7443 7444 auto the server chooses an appropriate mode 7445 automatically using heuristics 7446 global a single global pool contains all CPUs 7447 percpu one pool for each CPU 7448 pernode one pool for each NUMA node (equivalent 7449 to global on non-NUMA machines) 7450 7451 sunrpc.tcp_slot_table_entries= 7452 sunrpc.udp_slot_table_entries= 7453 [NFS,SUNRPC] 7454 Sets the upper limit on the number of simultaneous 7455 RPC calls that can be sent from the client to a 7456 server. Increasing these values may allow you to 7457 improve throughput, but will also increase the 7458 amount of memory reserved for use by the client. 7459 7460 suspend.pm_test_delay= 7461 [SUSPEND] 7462 Sets the number of seconds to remain in a suspend test 7463 mode before resuming the system (see 7464 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 7465 is set. Default value is 5. 7466 7467 svm= [PPC] 7468 Format: { on | off | y | n | 1 | 0 } 7469 This parameter controls use of the Protected 7470 Execution Facility on pSeries. 7471 7472 swiotlb= [ARM,PPC,MIPS,X86,S390,EARLY] 7473 Format: { <int> [,<int>] | force | noforce } 7474 <int> -- Number of I/O TLB slabs 7475 <int> -- Second integer after comma. Number of swiotlb 7476 areas with their own lock. Will be rounded up 7477 to a power of 2. 7478 force -- force using of bounce buffers even if they 7479 wouldn't be automatically used by the kernel 7480 noforce -- Never use bounce buffers (for debugging) 7481 7482 switches= [HW,M68k,EARLY] 7483 7484 sysctl.*= [KNL] 7485 Set a sysctl parameter, right before loading the init 7486 process, as if the value was written to the respective 7487 /proc/sys/... file. Both '.' and '/' are recognized as 7488 separators. Unrecognized parameters and invalid values 7489 are reported in the kernel log. Sysctls registered 7490 later by a loaded module cannot be set this way. 7491 Example: sysctl.vm.swappiness=40 7492 7493 sysrq_always_enabled 7494 [KNL] 7495 Ignore sysrq setting - this boot parameter will 7496 neutralize any effect of /proc/sys/kernel/sysrq. 7497 Useful for debugging. 7498 7499 tcpmhash_entries= [KNL,NET] 7500 Set the number of tcp_metrics_hash slots. 7501 Default value is 8192 or 16384 depending on total 7502 ram pages. This is used to specify the TCP metrics 7503 cache size. See Documentation/networking/ip-sysctl.rst 7504 "tcp_no_metrics_save" section for more details. 7505 7506 tdfx= [HW,DRM] 7507 7508 test_suspend= [SUSPEND] 7509 Format: { "mem" | "standby" | "freeze" }[,N] 7510 Specify "mem" (for Suspend-to-RAM) or "standby" (for 7511 standby suspend) or "freeze" (for suspend type freeze) 7512 as the system sleep state during system startup with 7513 the optional capability to repeat N number of times. 7514 The system is woken from this state using a 7515 wakeup-capable RTC alarm. 7516 7517 thash_entries= [KNL,NET] 7518 Set number of hash buckets for TCP connection 7519 7520 thermal.act= [HW,ACPI] 7521 -1: disable all active trip points in all thermal zones 7522 <degrees C>: override all lowest active trip points 7523 7524 thermal.crt= [HW,ACPI] 7525 -1: disable all critical trip points in all thermal zones 7526 <degrees C>: override all critical trip points 7527 7528 thermal.off= [HW,ACPI] 7529 1: disable ACPI thermal control 7530 7531 thermal.psv= [HW,ACPI] 7532 -1: disable all passive trip points 7533 <degrees C>: override all passive trip points to this 7534 value 7535 7536 thermal.tzp= [HW,ACPI] 7537 Specify global default ACPI thermal zone polling rate 7538 <deci-seconds>: poll all this frequency 7539 0: no polling (default) 7540 7541 thp_anon= [KNL] 7542 Format: <size>[KMG],<size>[KMG]:<state>;<size>[KMG]-<size>[KMG]:<state> 7543 state is one of "always", "madvise", "never" or "inherit". 7544 Control the default behavior of the system with respect 7545 to anonymous transparent hugepages. 7546 Can be used multiple times for multiple anon THP sizes. 7547 See Documentation/admin-guide/mm/transhuge.rst for more 7548 details. 7549 7550 threadirqs [KNL,EARLY] 7551 Force threading of all interrupt handlers except those 7552 marked explicitly IRQF_NO_THREAD. 7553 7554 thp_shmem= [KNL] 7555 Format: <size>[KMG],<size>[KMG]:<policy>;<size>[KMG]-<size>[KMG]:<policy> 7556 Control the default policy of each hugepage size for the 7557 internal shmem mount. <policy> is one of policies available 7558 for the shmem mount ("always", "inherit", "never", "within_size", 7559 and "advise"). 7560 It can be used multiple times for multiple shmem THP sizes. 7561 See Documentation/admin-guide/mm/transhuge.rst for more 7562 details. 7563 7564 topology= [S390,EARLY] 7565 Format: {off | on} 7566 Specify if the kernel should make use of the cpu 7567 topology information if the hardware supports this. 7568 The scheduler will make use of this information and 7569 e.g. base its process migration decisions on it. 7570 Default is on. 7571 7572 torture.disable_onoff_at_boot= [KNL] 7573 Prevent the CPU-hotplug component of torturing 7574 until after init has spawned. 7575 7576 torture.ftrace_dump_at_shutdown= [KNL] 7577 Dump the ftrace buffer at torture-test shutdown, 7578 even if there were no errors. This can be a 7579 very costly operation when many torture tests 7580 are running concurrently, especially on systems 7581 with rotating-rust storage. 7582 7583 torture.verbose_sleep_frequency= [KNL] 7584 Specifies how many verbose printk()s should be 7585 emitted between each sleep. The default of zero 7586 disables verbose-printk() sleeping. 7587 7588 torture.verbose_sleep_duration= [KNL] 7589 Duration of each verbose-printk() sleep in jiffies. 7590 7591 tpm.disable_pcr_integrity= [HW,TPM] 7592 Do not protect PCR registers from unintended physical 7593 access, or interposers in the bus by the means of 7594 having an integrity protected session wrapped around 7595 TPM2_PCR_Extend command. Consider this in a situation 7596 where TPM is heavily utilized by IMA, thus protection 7597 causing a major performance hit, and the space where 7598 machines are deployed is by other means guarded. 7599 7600 tpm_crb_ffa.busy_timeout_ms= [ARM64,TPM] 7601 Maximum time in milliseconds to retry sending a message 7602 to the TPM service before giving up. This parameter controls 7603 how long the system will continue retrying when the TPM 7604 service is busy. 7605 Format: <unsigned int> 7606 Default: 2000 (2 seconds) 7607 7608 tpm_suspend_pcr=[HW,TPM] 7609 Format: integer pcr id 7610 Specify that at suspend time, the tpm driver 7611 should extend the specified pcr with zeros, 7612 as a workaround for some chips which fail to 7613 flush the last written pcr on TPM_SaveState. 7614 This will guarantee that all the other pcrs 7615 are saved. 7616 7617 tpm_tis.interrupts= [HW,TPM] 7618 Enable interrupts for the MMIO based physical layer 7619 for the FIFO interface. By default it is set to false 7620 (0). For more information about TPM hardware interfaces 7621 defined by Trusted Computing Group (TCG) see 7622 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/ 7623 7624 tp_printk [FTRACE] 7625 Have the tracepoints sent to printk as well as the 7626 tracing ring buffer. This is useful for early boot up 7627 where the system hangs or reboots and does not give the 7628 option for reading the tracing buffer or performing a 7629 ftrace_dump_on_oops. 7630 7631 To turn off having tracepoints sent to printk, 7632 echo 0 > /proc/sys/kernel/tracepoint_printk 7633 Note, echoing 1 into this file without the 7634 tp_printk kernel cmdline option has no effect. 7635 7636 The tp_printk_stop_on_boot (see below) can also be used 7637 to stop the printing of events to console at 7638 late_initcall_sync. 7639 7640 ** CAUTION ** 7641 7642 Having tracepoints sent to printk() and activating high 7643 frequency tracepoints such as irq or sched, can cause 7644 the system to live lock. 7645 7646 tp_printk_stop_on_boot [FTRACE] 7647 When tp_printk (above) is set, it can cause a lot of noise 7648 on the console. It may be useful to only include the 7649 printing of events during boot up, as user space may 7650 make the system inoperable. 7651 7652 This command line option will stop the printing of events 7653 to console at the late_initcall_sync() time frame. 7654 7655 trace_buf_size=nn[KMG] 7656 [FTRACE] will set tracing buffer size on each cpu. 7657 7658 trace_clock= [FTRACE] Set the clock used for tracing events 7659 at boot up. 7660 local - Use the per CPU time stamp counter 7661 (converted into nanoseconds). Fast, but 7662 depending on the architecture, may not be 7663 in sync between CPUs. 7664 global - Event time stamps are synchronized across 7665 CPUs. May be slower than the local clock, 7666 but better for some race conditions. 7667 counter - Simple counting of events (1, 2, ..) 7668 note, some counts may be skipped due to the 7669 infrastructure grabbing the clock more than 7670 once per event. 7671 uptime - Use jiffies as the time stamp. 7672 perf - Use the same clock that perf uses. 7673 mono - Use ktime_get_mono_fast_ns() for time stamps. 7674 mono_raw - Use ktime_get_raw_fast_ns() for time 7675 stamps. 7676 boot - Use ktime_get_boot_fast_ns() for time stamps. 7677 Architectures may add more clocks. See 7678 Documentation/trace/ftrace.rst for more details. 7679 7680 trace_event=[event-list] 7681 [FTRACE] Set and start specified trace events in order 7682 to facilitate early boot debugging. The event-list is a 7683 comma-separated list of trace events to enable. See 7684 also Documentation/trace/events.rst 7685 7686 To enable modules, use :mod: keyword: 7687 7688 trace_event=:mod:<module> 7689 7690 The value before :mod: will only enable specific events 7691 that are part of the module. See the above mentioned 7692 document for more information. 7693 7694 trace_instance=[instance-info] 7695 [FTRACE] Create a ring buffer instance early in boot up. 7696 This will be listed in: 7697 7698 /sys/kernel/tracing/instances 7699 7700 Events can be enabled at the time the instance is created 7701 via: 7702 7703 trace_instance=<name>,<system1>:<event1>,<system2>:<event2> 7704 7705 Note, the "<system*>:" portion is optional if the event is 7706 unique. 7707 7708 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall 7709 7710 will enable the "sched_switch" event (note, the "sched:" is optional, and 7711 the same thing would happen if it was left off). The irq_handler_entry 7712 event, and all events under the "initcall" system. 7713 7714 Flags can be added to the instance to modify its behavior when it is 7715 created. The flags are separated by '^'. 7716 7717 The available flags are: 7718 7719 traceoff - Have the tracing instance tracing disabled after it is created. 7720 traceprintk - Have trace_printk() write into this trace instance 7721 (note, "printk" and "trace_printk" can also be used) 7722 7723 trace_instance=foo^traceoff^traceprintk,sched,irq 7724 7725 The flags must come before the defined events. 7726 7727 If memory has been reserved (see memmap for x86), the instance 7728 can use that memory: 7729 7730 memmap=12M$0x284500000 trace_instance=boot_map@0x284500000:12M 7731 7732 The above will create a "boot_map" instance that uses the physical 7733 memory at 0x284500000 that is 12Megs. The per CPU buffers of that 7734 instance will be split up accordingly. 7735 7736 Alternatively, the memory can be reserved by the reserve_mem option: 7737 7738 reserve_mem=12M:4096:trace trace_instance=boot_map@trace 7739 7740 This will reserve 12 megabytes at boot up with a 4096 byte alignment 7741 and place the ring buffer in this memory. Note that due to KASLR, the 7742 memory may not be the same location each time, which will not preserve 7743 the buffer content. 7744 7745 Also note that the layout of the ring buffer data may change between 7746 kernel versions where the validator will fail and reset the ring buffer 7747 if the layout is not the same as the previous kernel. 7748 7749 If the ring buffer is used for persistent bootups and has events enabled, 7750 it is recommend to disable tracing so that events from a previous boot do not 7751 mix with events of the current boot (unless you are debugging a random crash 7752 at boot up). 7753 7754 reserve_mem=12M:4096:trace trace_instance=boot_map^traceoff^traceprintk@trace,sched,irq 7755 7756 Note, saving the trace buffer across reboots does require that the system 7757 is set up to not wipe memory. For instance, CONFIG_RESET_ATTACK_MITIGATION 7758 can force a memory reset on boot which will clear any trace that was stored. 7759 This is just one of many ways that can clear memory. Make sure your system 7760 keeps the content of memory across reboots before relying on this option. 7761 7762 NB: Both the mapped address and size must be page aligned for the architecture. 7763 7764 See also Documentation/trace/debugging.rst 7765 7766 7767 trace_options=[option-list] 7768 [FTRACE] Enable or disable tracer options at boot. 7769 The option-list is a comma delimited list of options 7770 that can be enabled or disabled just as if you were 7771 to echo the option name into 7772 7773 /sys/kernel/tracing/trace_options 7774 7775 For example, to enable stacktrace option (to dump the 7776 stack trace of each event), add to the command line: 7777 7778 trace_options=stacktrace 7779 7780 See also Documentation/trace/ftrace.rst "trace options" 7781 section. 7782 7783 trace_trigger=[trigger-list] 7784 [FTRACE] Add an event trigger on specific events. 7785 Set a trigger on top of a specific event, with an optional 7786 filter. 7787 7788 The format is "trace_trigger=<event>.<trigger>[ if <filter>],..." 7789 Where more than one trigger may be specified that are comma delimited. 7790 7791 For example: 7792 7793 trace_trigger="sched_switch.stacktrace if prev_state == 2" 7794 7795 The above will enable the "stacktrace" trigger on the "sched_switch" 7796 event but only trigger it if the "prev_state" of the "sched_switch" 7797 event is "2" (TASK_UNINTERRUPTIBLE). 7798 7799 See also "Event triggers" in Documentation/trace/events.rst 7800 7801 7802 traceoff_after_boot 7803 [FTRACE] Sometimes tracing is used to debug issues 7804 during the boot process. Since the trace buffer has a 7805 limited amount of storage, it may be prudent to 7806 disable tracing after the boot is finished, otherwise 7807 the critical information may be overwritten. With this 7808 option, the main tracing buffer will be turned off at 7809 the end of the boot process. 7810 7811 traceoff_on_warning 7812 [FTRACE] enable this option to disable tracing when a 7813 warning is hit. This turns off "tracing_on". Tracing can 7814 be enabled again by echoing '1' into the "tracing_on" 7815 file located in /sys/kernel/tracing/ 7816 7817 This option is useful, as it disables the trace before 7818 the WARNING dump is called, which prevents the trace to 7819 be filled with content caused by the warning output. 7820 7821 This option can also be set at run time via the sysctl 7822 option: kernel/traceoff_on_warning 7823 7824 transparent_hugepage= 7825 [KNL] 7826 Format: [always|madvise|never] 7827 Can be used to control the default behavior of the system 7828 with respect to transparent hugepages. 7829 See Documentation/admin-guide/mm/transhuge.rst 7830 for more details. 7831 7832 transparent_hugepage_shmem= [KNL] 7833 Format: [always|within_size|advise|never|deny|force] 7834 Can be used to control the hugepage allocation policy for 7835 the internal shmem mount. 7836 See Documentation/admin-guide/mm/transhuge.rst 7837 for more details. 7838 7839 transparent_hugepage_tmpfs= [KNL] 7840 Format: [always|within_size|advise|never] 7841 Can be used to control the default hugepage allocation policy 7842 for the tmpfs mount. 7843 See Documentation/admin-guide/mm/transhuge.rst 7844 for more details. 7845 7846 trusted.source= [KEYS] 7847 Format: <string> 7848 This parameter identifies the trust source as a backend 7849 for trusted keys implementation. Supported trust 7850 sources: 7851 - "tpm" 7852 - "tee" 7853 - "caam" 7854 - "dcp" 7855 - "pkwm" 7856 If not specified then it defaults to iterating through 7857 the trust source list starting with TPM and assigns the 7858 first trust source as a backend which is initialized 7859 successfully during iteration. 7860 7861 trusted.rng= [KEYS] 7862 Format: <string> 7863 The RNG used to generate key material for trusted keys. 7864 Can be one of: 7865 - "kernel" 7866 - the same value as trusted.source: "tpm" or "tee" 7867 - "default" 7868 If not specified, "default" is used. In this case, 7869 the RNG's choice is left to each individual trust source. 7870 7871 trusted.dcp_use_otp_key 7872 This is intended to be used in combination with 7873 trusted.source=dcp and will select the DCP OTP key 7874 instead of the DCP UNIQUE key blob encryption. 7875 7876 trusted.dcp_skip_zk_test 7877 This is intended to be used in combination with 7878 trusted.source=dcp and will disable the check if the 7879 blob key is all zeros. This is helpful for situations where 7880 having this key zero'ed is acceptable. E.g. in testing 7881 scenarios. 7882 7883 tsa= [X86] Control mitigation for Transient Scheduler 7884 Attacks on AMD CPUs. Search the following in your 7885 favourite search engine for more details: 7886 7887 "Technical guidance for mitigating transient scheduler 7888 attacks". 7889 7890 off - disable the mitigation 7891 on - enable the mitigation (default) 7892 user - mitigate only user/kernel transitions 7893 vm - mitigate only guest/host transitions 7894 7895 7896 tsc= Disable clocksource stability checks for TSC. 7897 Format: <string> 7898 [x86] reliable: mark tsc clocksource as reliable, this 7899 disables clocksource verification at runtime, as well 7900 as the stability checks done at bootup. Used to enable 7901 high-resolution timer mode on older hardware, and in 7902 virtualized environment. 7903 [x86] noirqtime: Do not use TSC to do irq accounting. 7904 Used to run time disable IRQ_TIME_ACCOUNTING on any 7905 platforms where RDTSC is slow and this accounting 7906 can add overhead. 7907 [x86] unstable: mark the TSC clocksource as unstable, this 7908 marks the TSC unconditionally unstable at bootup and 7909 avoids any further wobbles once the TSC watchdog notices. 7910 [x86] nowatchdog: disable clocksource watchdog. Used 7911 in situations with strict latency requirements (where 7912 interruptions from clocksource watchdog are not 7913 acceptable). 7914 [x86] recalibrate: force recalibration against a HW timer 7915 (HPET or PM timer) on systems whose TSC frequency was 7916 obtained from HW or FW using either an MSR or CPUID(0x15). 7917 Warn if the difference is more than 500 ppm. 7918 [x86] watchdog: Enforce the clocksource watchdog on TSC 7919 7920 tsc_early_khz= [X86,EARLY] Skip early TSC calibration and use the given 7921 value instead. Useful when the early TSC frequency discovery 7922 procedure is not reliable, such as on overclocked systems 7923 with CPUID.16h support and partial CPUID.15h support. 7924 Format: <unsigned int> 7925 7926 tsx= [X86] Control Transactional Synchronization 7927 Extensions (TSX) feature in Intel processors that 7928 support TSX control. 7929 7930 This parameter controls the TSX feature. The options are: 7931 7932 on - Enable TSX on the system. Although there are 7933 mitigations for all known security vulnerabilities, 7934 TSX has been known to be an accelerator for 7935 several previous speculation-related CVEs, and 7936 so there may be unknown security risks associated 7937 with leaving it enabled. 7938 7939 off - Disable TSX on the system. (Note that this 7940 option takes effect only on newer CPUs which are 7941 not vulnerable to MDS, i.e., have 7942 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get 7943 the new IA32_TSX_CTRL MSR through a microcode 7944 update. This new MSR allows for the reliable 7945 deactivation of the TSX functionality.) 7946 7947 auto - Disable TSX if X86_BUG_TAA is present, 7948 otherwise enable TSX on the system. 7949 7950 Not specifying this option is equivalent to tsx=off. 7951 7952 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 7953 for more details. 7954 7955 tsx_async_abort= [X86,INTEL,EARLY] Control mitigation for the TSX Async 7956 Abort (TAA) vulnerability. 7957 7958 Similar to Micro-architectural Data Sampling (MDS) 7959 certain CPUs that support Transactional 7960 Synchronization Extensions (TSX) are vulnerable to an 7961 exploit against CPU internal buffers which can forward 7962 information to a disclosure gadget under certain 7963 conditions. 7964 7965 In vulnerable processors, the speculatively forwarded 7966 data can be used in a cache side channel attack, to 7967 access data to which the attacker does not have direct 7968 access. 7969 7970 This parameter controls the TAA mitigation. The 7971 options are: 7972 7973 full - Enable TAA mitigation on vulnerable CPUs 7974 if TSX is enabled. 7975 7976 full,nosmt - Enable TAA mitigation and disable SMT on 7977 vulnerable CPUs. If TSX is disabled, SMT 7978 is not disabled because CPU is not 7979 vulnerable to cross-thread TAA attacks. 7980 off - Unconditionally disable TAA mitigation 7981 7982 On MDS-affected machines, tsx_async_abort=off can be 7983 prevented by an active MDS mitigation as both vulnerabilities 7984 are mitigated with the same mechanism so in order to disable 7985 this mitigation, you need to specify mds=off too. 7986 7987 Not specifying this option is equivalent to 7988 tsx_async_abort=full. On CPUs which are MDS affected 7989 and deploy MDS mitigation, TAA mitigation is not 7990 required and doesn't provide any additional 7991 mitigation. 7992 7993 For details see: 7994 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 7995 7996 turbografx.map[2|3]= [HW,JOY] 7997 TurboGraFX parallel port interface 7998 Format: 7999 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7> 8000 See also Documentation/input/devices/joystick-parport.rst 8001 8002 udbg-immortal [PPC] When debugging early kernel crashes that 8003 happen after console_init() and before a proper 8004 console driver takes over, this boot options might 8005 help "seeing" what's going on. 8006 8007 uhash_entries= [KNL,NET] 8008 Set number of hash buckets for UDP/UDP-Lite connections 8009 8010 uhci-hcd.ignore_oc= 8011 [USB] Ignore overcurrent events (default N). 8012 Some badly-designed motherboards generate lots of 8013 bogus events, for ports that aren't wired to 8014 anything. Set this parameter to avoid log spamming. 8015 Note that genuine overcurrent events won't be 8016 reported either. 8017 8018 unaligned_scalar_speed= 8019 [RISCV] 8020 Format: {slow | fast | unsupported} 8021 Allow skipping scalar unaligned access speed tests. This 8022 is useful for testing alternative code paths and to skip 8023 the tests in environments where they run too slowly. All 8024 CPUs must have the same scalar unaligned access speed. 8025 8026 unaligned_vector_speed= 8027 [RISCV] 8028 Format: {slow | fast | unsupported} 8029 Allow skipping vector unaligned access speed tests. This 8030 is useful for testing alternative code paths and to skip 8031 the tests in environments where they run too slowly. All 8032 CPUs must have the same vector unaligned access speed. 8033 8034 unknown_nmi_panic 8035 [X86] Cause panic on unknown NMI. 8036 8037 unwind_debug [X86-64,EARLY] 8038 Enable unwinder debug output. This can be 8039 useful for debugging certain unwinder error 8040 conditions, including corrupt stacks and 8041 bad/missing unwinder metadata. 8042 8043 usbcore.authorized_default= 8044 [USB] Default USB device authorization: 8045 (default -1 = authorized (same as 1), 8046 0 = not authorized, 1 = authorized, 2 = authorized 8047 if device connected to internal port) 8048 8049 usbcore.autosuspend= 8050 [USB] The autosuspend time delay (in seconds) used 8051 for newly-detected USB devices (default 2). This 8052 is the time required before an idle device will be 8053 autosuspended. Devices for which the delay is set 8054 to a negative value won't be autosuspended at all. 8055 8056 usbcore.usbfs_snoop= 8057 [USB] Set to log all usbfs traffic (default 0 = off). 8058 8059 usbcore.usbfs_snoop_max= 8060 [USB] Maximum number of bytes to snoop in each URB 8061 (default = 65536). 8062 8063 usbcore.blinkenlights= 8064 [USB] Set to cycle leds on hubs (default 0 = off). 8065 8066 usbcore.old_scheme_first= 8067 [USB] Start with the old device initialization 8068 scheme (default 0 = off). 8069 8070 usbcore.usbfs_memory_mb= 8071 [USB] Memory limit (in MB) for buffers allocated by 8072 usbfs (default = 16, 0 = max = 2047). 8073 8074 usbcore.use_both_schemes= 8075 [USB] Try the other device initialization scheme 8076 if the first one fails (default 1 = enabled). 8077 8078 usbcore.initial_descriptor_timeout= 8079 [USB] Specifies timeout for the initial 64-byte 8080 USB_REQ_GET_DESCRIPTOR request in milliseconds 8081 (default 5000 = 5.0 seconds). 8082 8083 usbcore.nousb [USB] Disable the USB subsystem 8084 8085 usbcore.quirks= 8086 [USB] A list of quirk entries to augment the built-in 8087 usb core quirk list. List entries are separated by 8088 commas. Each entry has the form 8089 VendorID:ProductID:Flags. The IDs are 4-digit hex 8090 numbers and Flags is a set of letters. Each letter 8091 will change the built-in quirk; setting it if it is 8092 clear and clearing it if it is set. The letters have 8093 the following meanings: 8094 a = USB_QUIRK_STRING_FETCH_255 (string 8095 descriptors must not be fetched using 8096 a 255-byte read); 8097 b = USB_QUIRK_RESET_RESUME (device can't resume 8098 correctly so reset it instead); 8099 c = USB_QUIRK_NO_SET_INTF (device can't handle 8100 Set-Interface requests); 8101 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't 8102 handle its Configuration or Interface 8103 strings); 8104 e = USB_QUIRK_RESET (device can't be reset 8105 (e.g morph devices), don't use reset); 8106 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has 8107 more interface descriptions than the 8108 bNumInterfaces count, and can't handle 8109 talking to these interfaces); 8110 g = USB_QUIRK_DELAY_INIT (device needs a pause 8111 during initialization, after we read 8112 the device descriptor); 8113 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For 8114 high speed and super speed interrupt 8115 endpoints, the USB 2.0 and USB 3.0 spec 8116 require the interval in microframes (1 8117 microframe = 125 microseconds) to be 8118 calculated as interval = 2 ^ 8119 (bInterval-1). 8120 Devices with this quirk report their 8121 bInterval as the result of this 8122 calculation instead of the exponent 8123 variable used in the calculation); 8124 i = USB_QUIRK_DEVICE_QUALIFIER (device can't 8125 handle device_qualifier descriptor 8126 requests); 8127 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device 8128 generates spurious wakeup, ignore 8129 remote wakeup capability); 8130 k = USB_QUIRK_NO_LPM (device can't handle Link 8131 Power Management); 8132 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL 8133 (Device reports its bInterval as linear 8134 frames instead of the USB 2.0 8135 calculation); 8136 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs 8137 to be disconnected before suspend to 8138 prevent spurious wakeup); 8139 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a 8140 pause after every control message); 8141 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra 8142 delay after resetting its port); 8143 p = USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT 8144 (Reduce timeout of the SET_ADDRESS 8145 request from 5000 ms to 500 ms); 8146 q = USB_QUIRK_FORCE_ONE_CONFIG (Device 8147 claims zero configurations, 8148 forcing to 1); 8149 Example: quirks=0781:5580:bk,0a5c:5834:gij 8150 8151 usbhid.mousepoll= 8152 [USBHID] The interval which mice are to be polled at. 8153 8154 usbhid.jspoll= 8155 [USBHID] The interval which joysticks are to be polled at. 8156 8157 usbhid.kbpoll= 8158 [USBHID] The interval which keyboards are to be polled at. 8159 8160 usb-storage.delay_use= 8161 [UMS] The delay in seconds before a new device is 8162 scanned for Logical Units (default 1). 8163 Optionally the delay in milliseconds if the value has 8164 suffix with "ms". 8165 Example: delay_use=2567ms 8166 8167 usb-storage.quirks= 8168 [UMS] A list of quirks entries to supplement or 8169 override the built-in unusual_devs list. List 8170 entries are separated by commas. Each entry has 8171 the form VID:PID:Flags where VID and PID are Vendor 8172 and Product ID values (4-digit hex numbers) and 8173 Flags is a set of characters, each corresponding 8174 to a common usb-storage quirk flag as follows: 8175 a = SANE_SENSE (collect more than 18 bytes 8176 of sense data, not on uas); 8177 b = BAD_SENSE (don't collect more than 18 8178 bytes of sense data, not on uas); 8179 c = FIX_CAPACITY (decrease the reported 8180 device capacity by one sector); 8181 d = NO_READ_DISC_INFO (don't use 8182 READ_DISC_INFO command, not on uas); 8183 e = NO_READ_CAPACITY_16 (don't use 8184 READ_CAPACITY_16 command); 8185 f = NO_REPORT_OPCODES (don't use report opcodes 8186 command, uas only); 8187 g = MAX_SECTORS_240 (don't transfer more than 8188 240 sectors at a time, uas only); 8189 h = CAPACITY_HEURISTICS (decrease the 8190 reported device capacity by one 8191 sector if the number is odd); 8192 i = IGNORE_DEVICE (don't bind to this 8193 device); 8194 j = NO_REPORT_LUNS (don't use report luns 8195 command, uas only); 8196 k = NO_SAME (do not use WRITE_SAME, uas only) 8197 l = NOT_LOCKABLE (don't try to lock and 8198 unlock ejectable media, not on uas); 8199 m = MAX_SECTORS_64 (don't transfer more 8200 than 64 sectors = 32 KB at a time, 8201 not on uas); 8202 n = INITIAL_READ10 (force a retry of the 8203 initial READ(10) command, not on uas); 8204 o = CAPACITY_OK (accept the capacity 8205 reported by the device, not on uas); 8206 p = WRITE_CACHE (the device cache is ON 8207 by default, not on uas); 8208 r = IGNORE_RESIDUE (the device reports 8209 bogus residue values, not on uas); 8210 s = SINGLE_LUN (the device has only one 8211 Logical Unit); 8212 t = NO_ATA_1X (don't allow ATA(12) and ATA(16) 8213 commands, uas only); 8214 u = IGNORE_UAS (don't bind to the uas driver); 8215 w = NO_WP_DETECT (don't test whether the 8216 medium is write-protected). 8217 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE 8218 even if the device claims no cache, 8219 not on uas) 8220 Example: quirks=0419:aaf5:rl,0421:0433:rc 8221 8222 user_debug= [KNL,ARM] 8223 Format: <int> 8224 See arch/arm/Kconfig.debug help text. 8225 1 - undefined instruction events 8226 2 - system calls 8227 4 - invalid data aborts 8228 8 - SIGSEGV faults 8229 16 - SIGBUS faults 8230 Example: user_debug=31 8231 8232 vdso= [X86,SH,SPARC] 8233 On X86_32, this is an alias for vdso32=. Otherwise: 8234 8235 vdso=1: enable VDSO (the default) 8236 vdso=0: disable VDSO mapping 8237 8238 vdso32= [X86] Control the 32-bit vDSO 8239 vdso32=1: enable 32-bit VDSO 8240 vdso32=0 or vdso32=2: disable 32-bit VDSO 8241 8242 See the help text for CONFIG_COMPAT_VDSO for more 8243 details. If CONFIG_COMPAT_VDSO is set, the default is 8244 vdso32=0; otherwise, the default is vdso32=1. 8245 8246 For compatibility with older kernels, vdso32=2 is an 8247 alias for vdso32=0. 8248 8249 Try vdso32=0 if you encounter an error that says: 8250 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed! 8251 8252 video= [FB,EARLY] Frame buffer configuration 8253 See Documentation/fb/modedb.rst. 8254 8255 video.brightness_switch_enabled= [ACPI] 8256 Format: [0|1] 8257 If set to 1, on receiving an ACPI notify event 8258 generated by hotkey, video driver will adjust brightness 8259 level and then send out the event to user space through 8260 the allocated input device. If set to 0, video driver 8261 will only send out the event without touching backlight 8262 brightness level. 8263 default: 1 8264 8265 virtio_mmio.device= 8266 [VMMIO] Memory mapped virtio (platform) device. 8267 8268 <size>@<baseaddr>:<irq>[:<id>] 8269 where: 8270 <size> := size (can use standard suffixes 8271 like K, M and G) 8272 <baseaddr> := physical base address 8273 <irq> := interrupt number (as passed to 8274 request_irq()) 8275 <id> := (optional) platform device id 8276 example: 8277 virtio_mmio.device=1K@0x100b0000:48:7 8278 8279 Can be used multiple times for multiple devices. 8280 8281 vga= [BOOT,X86-32] Select a particular video mode 8282 See Documentation/arch/x86/boot.rst and 8283 Documentation/admin-guide/svga.rst. 8284 Use vga=ask for menu. 8285 This is actually a boot loader parameter; the value is 8286 passed to the kernel using a special protocol. 8287 8288 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y. 8289 May slow down system boot speed, especially when 8290 enabled on systems with a large amount of memory. 8291 All options are enabled by default, and this 8292 interface is meant to allow for selectively 8293 enabling or disabling specific virtual memory 8294 debugging features. 8295 8296 Available options are: 8297 P Enable page structure init time poisoning 8298 - Disable all of the above options 8299 8300 vmalloc=nn[KMG] [KNL,BOOT,EARLY] Forces the vmalloc area to have an 8301 exact size of <nn>. This can be used to increase 8302 the minimum size (128MB on x86, arm32 platforms). 8303 It can also be used to decrease the size and leave more room 8304 for directly mapped kernel RAM. Note that this parameter does 8305 not exist on many other platforms (including arm64, alpha, 8306 loongarch, arc, csky, hexagon, microblaze, mips, nios2, openrisc, 8307 parisc, m64k, powerpc, riscv, sh, um, xtensa, s390, sparc). 8308 8309 vmcp_cma=nn[MG] [KNL,S390,EARLY] 8310 Sets the memory size reserved for contiguous memory 8311 allocations for the vmcp device driver. 8312 8313 vmhalt= [KNL,S390] Perform z/VM CP command after system halt. 8314 Format: <command> 8315 8316 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic. 8317 Format: <command> 8318 8319 vmpoff= [KNL,S390] Perform z/VM CP command after power off. 8320 Format: <command> 8321 8322 vmscape= [X86] Controls mitigation for VMscape attacks. 8323 VMscape attacks can leak information from a userspace 8324 hypervisor to a guest via speculative side-channels. 8325 8326 off - disable the mitigation 8327 ibpb - use Indirect Branch Prediction Barrier 8328 (IBPB) mitigation (default) 8329 force - force vulnerability detection even on 8330 unaffected processors 8331 8332 vsyscall= [X86-64,EARLY] 8333 Controls the behavior of vsyscalls (i.e. calls to 8334 fixed addresses of 0xffffffffff600x00 from legacy 8335 code). Most statically-linked binaries and older 8336 versions of glibc use these calls. Because these 8337 functions are at fixed addresses, they make nice 8338 targets for exploits that can control RIP. 8339 8340 emulate Vsyscalls turn into traps and are emulated 8341 reasonably safely. The vsyscall page is 8342 readable. This disables the Linear 8343 Address Space Separation (LASS) security 8344 feature and makes the system less secure. 8345 8346 xonly [default] Vsyscalls turn into traps and are 8347 emulated reasonably safely. The vsyscall 8348 page is not readable. 8349 8350 none Vsyscalls don't work at all. This makes 8351 them quite hard to use for exploits but 8352 might break your system. 8353 8354 vt.color= [VT] Default text color. 8355 Format: 0xYX, X = foreground, Y = background. 8356 Default: 0x07 = light gray on black. 8357 8358 vt.cur_default= [VT] Default cursor shape. 8359 Format: 0xCCBBAA, where AA, BB, and CC are the same as 8360 the parameters of the <Esc>[?A;B;Cc escape sequence; 8361 see vga-softcursor.rst. Default: 2 = underline. 8362 8363 vt.default_blu= [VT] 8364 Format: <blue0>,<blue1>,<blue2>,...,<blue15> 8365 Change the default blue palette of the console. 8366 This is a 16-member array composed of values 8367 ranging from 0-255. 8368 8369 vt.default_grn= [VT] 8370 Format: <green0>,<green1>,<green2>,...,<green15> 8371 Change the default green palette of the console. 8372 This is a 16-member array composed of values 8373 ranging from 0-255. 8374 8375 vt.default_red= [VT] 8376 Format: <red0>,<red1>,<red2>,...,<red15> 8377 Change the default red palette of the console. 8378 This is a 16-member array composed of values 8379 ranging from 0-255. 8380 8381 vt.default_utf8= 8382 [VT] 8383 Format=<0|1> 8384 Set system-wide default UTF-8 mode for all tty's. 8385 Default is 1, i.e. UTF-8 mode is enabled for all 8386 newly opened terminals. 8387 8388 vt.global_cursor_default= 8389 [VT] 8390 Format=<-1|0|1> 8391 Set system-wide default for whether a cursor 8392 is shown on new VTs. Default is -1, 8393 i.e. cursors will be created by default unless 8394 overridden by individual drivers. 0 will hide 8395 cursors, 1 will display them. 8396 8397 vt.italic= [VT] Default color for italic text; 0-15. 8398 Default: 2 = green. 8399 8400 vt.underline= [VT] Default color for underlined text; 0-15. 8401 Default: 3 = cyan. 8402 8403 watchdog timers [HW,WDT] For information on watchdog timers, 8404 see Documentation/watchdog/watchdog-parameters.rst 8405 or other driver-specific files in the 8406 Documentation/watchdog/ directory. 8407 8408 watchdog_thresh= 8409 [KNL] 8410 Set the hard lockup detector stall duration 8411 threshold in seconds. The soft lockup detector 8412 threshold is set to twice the value. A value of 0 8413 disables both lockup detectors. Default is 10 8414 seconds. 8415 8416 workqueue.unbound_cpus= 8417 [KNL,SMP] Specify to constrain one or some CPUs 8418 to use in unbound workqueues. 8419 Format: <cpu-list> 8420 By default, all online CPUs are available for 8421 unbound workqueues. 8422 8423 workqueue.watchdog_thresh= 8424 If CONFIG_WQ_WATCHDOG is configured, workqueue can 8425 warn stall conditions and dump internal state to 8426 help debugging. 0 disables workqueue stall 8427 detection; otherwise, it's the stall threshold 8428 duration in seconds. The default value is 30 and 8429 it can be updated at runtime by writing to the 8430 corresponding sysfs file. 8431 8432 workqueue.panic_on_stall=<uint> 8433 Panic when workqueue stall is detected by 8434 CONFIG_WQ_WATCHDOG. It sets the number times of the 8435 stall to trigger panic. 8436 8437 The default is set by CONFIG_BOOTPARAM_WQ_STALL_PANIC, 8438 which is 0 (disabled) if not configured. 8439 8440 workqueue.panic_on_stall_time=<uint> 8441 Panic when a workqueue stall has been continuous for 8442 the specified number of seconds. Unlike panic_on_stall 8443 which counts accumulated stall events, this triggers 8444 based on the duration of a single continuous stall. 8445 8446 The default is 0, which disables the time-based panic. 8447 8448 workqueue.cpu_intensive_thresh_us= 8449 Per-cpu work items which run for longer than this 8450 threshold are automatically considered CPU intensive 8451 and excluded from concurrency management to prevent 8452 them from noticeably delaying other per-cpu work 8453 items. Default is 10000 (10ms). 8454 8455 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 8456 will report the work functions which violate this 8457 threshold repeatedly. They are likely good 8458 candidates for using WQ_UNBOUND workqueues instead. 8459 8460 workqueue.cpu_intensive_warning_thresh=<uint> 8461 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 8462 will report the work functions which violate the 8463 intensive_threshold_us repeatedly. In order to prevent 8464 spurious warnings, start printing only after a work 8465 function has violated this threshold number of times. 8466 8467 The default is 4 times. 0 disables the warning. 8468 8469 workqueue.power_efficient 8470 Per-cpu workqueues are generally preferred because 8471 they show better performance thanks to cache 8472 locality; unfortunately, per-cpu workqueues tend to 8473 be more power hungry than unbound workqueues. 8474 8475 Enabling this makes the per-cpu workqueues which 8476 were observed to contribute significantly to power 8477 consumption unbound, leading to measurably lower 8478 power usage at the cost of small performance 8479 overhead. 8480 8481 The default value of this parameter is determined by 8482 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT. 8483 8484 workqueue.default_affinity_scope= 8485 Select the default affinity scope to use for unbound 8486 workqueues. Can be one of "cpu", "smt", "cache", 8487 "cache_shard", "numa" and "system". Default is 8488 "cache_shard". For more 8489 information, see the Affinity Scopes section in 8490 Documentation/core-api/workqueue.rst. 8491 8492 This can be changed after boot by writing to the 8493 matching /sys/module/workqueue/parameters file. All 8494 workqueues with the "default" affinity scope will be 8495 updated accordingly. 8496 8497 workqueue.debug_force_rr_cpu 8498 Workqueue used to implicitly guarantee that work 8499 items queued without explicit CPU specified are put 8500 on the local CPU. This guarantee is no longer true 8501 and while local CPU is still preferred work items 8502 may be put on foreign CPUs. This debug option 8503 forces round-robin CPU selection to flush out 8504 usages which depend on the now broken guarantee. 8505 When enabled, memory and cache locality will be 8506 impacted. 8507 8508 writecombine= [LOONGARCH,EARLY] Control the MAT (Memory Access 8509 Type) of ioremap_wc(). 8510 8511 on - Enable writecombine, use WUC for ioremap_wc() 8512 off - Disable writecombine, use SUC for ioremap_wc() 8513 8514 x2apic_phys [X86-64,APIC,EARLY] Use x2apic physical mode instead of 8515 default x2apic cluster mode on platforms 8516 supporting x2apic. 8517 8518 xen_512gb_limit [KNL,X86-64,XEN] 8519 Restricts the kernel running paravirtualized under Xen 8520 to use only up to 512 GB of RAM. The reason to do so is 8521 crash analysis tools and Xen tools for doing domain 8522 save/restore/migration must be enabled to handle larger 8523 domains. 8524 8525 xen_console_io [XEN,EARLY] 8526 Boolean option to enable/disable the usage of the Xen 8527 console_io hypercalls to read and write to the console. 8528 Mostly useful for debugging and development. 8529 8530 xen_emul_unplug= [HW,X86,XEN,EARLY] 8531 Unplug Xen emulated devices 8532 Format: [unplug0,][unplug1] 8533 ide-disks -- unplug primary master IDE devices 8534 aux-ide-disks -- unplug non-primary-master IDE devices 8535 nics -- unplug network devices 8536 all -- unplug all emulated devices (NICs and IDE disks) 8537 unnecessary -- unplugging emulated devices is 8538 unnecessary even if the host did not respond to 8539 the unplug protocol 8540 never -- do not unplug even if version check succeeds 8541 8542 xen_legacy_crash [X86,XEN,EARLY] 8543 Crash from Xen panic notifier, without executing late 8544 panic() code such as dumping handler. 8545 8546 xen_mc_debug [X86,XEN,EARLY] 8547 Enable multicall debugging when running as a Xen PV guest. 8548 Enabling this feature will reduce performance a little 8549 bit, so it should only be enabled for obtaining extended 8550 debug data in case of multicall errors. 8551 8552 xen_msr_safe= [X86,XEN,EARLY] 8553 Format: <bool> 8554 Select whether to always use non-faulting (safe) MSR 8555 access functions when running as Xen PV guest. The 8556 default value is controlled by CONFIG_XEN_PV_MSR_SAFE. 8557 8558 xen_nopv [X86] 8559 Disables the PV optimizations forcing the HVM guest to 8560 run as generic HVM guest with no PV drivers. 8561 This option is obsoleted by the "nopv" option, which 8562 has equivalent effect for XEN platform. 8563 8564 xen_no_vector_callback 8565 [KNL,X86,XEN,EARLY] Disable the vector callback for Xen 8566 event channel interrupts. 8567 8568 xen_scrub_pages= [XEN] 8569 Boolean option to control scrubbing pages before giving them back 8570 to Xen, for use by other domains. Can be also changed at runtime 8571 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages. 8572 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT. 8573 8574 xen_timer_slop= [X86-64,XEN,EARLY] 8575 Set the timer slop (in nanoseconds) for the virtual Xen 8576 timers (default is 100000). This adjusts the minimum 8577 delta of virtualized Xen timers, where lower values 8578 improve timer resolution at the expense of processing 8579 more timer interrupts. 8580 8581 xen.balloon_boot_timeout= [XEN] 8582 The time (in seconds) to wait before giving up to boot 8583 in case initial ballooning fails to free enough memory. 8584 Applies only when running as HVM or PVH guest and 8585 started with less memory configured than allowed at 8586 max. Default is 180. 8587 8588 xen.event_eoi_delay= [XEN] 8589 How long to delay EOI handling in case of event 8590 storms (jiffies). Default is 10. 8591 8592 xen.event_loop_timeout= [XEN] 8593 After which time (jiffies) the event handling loop 8594 should start to delay EOI handling. Default is 2. 8595 8596 xen.fifo_events= [XEN] 8597 Boolean parameter to disable using fifo event handling 8598 even if available. Normally fifo event handling is 8599 preferred over the 2-level event handling, as it is 8600 fairer and the number of possible event channels is 8601 much higher. Default is on (use fifo events). 8602 8603 xirc2ps_cs= [NET,PCMCIA] 8604 Format: 8605 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]] 8606 8607 xive= [PPC] 8608 By default on POWER9 and above, the kernel will 8609 natively use the XIVE interrupt controller. This option 8610 allows the fallback firmware mode to be used: 8611 8612 off Fallback to firmware control of XIVE interrupt 8613 controller on both pseries and powernv 8614 platforms. Only useful on POWER9 and above. 8615 8616 xive.store-eoi=off [PPC] 8617 By default on POWER10 and above, the kernel will use 8618 stores for EOI handling when the XIVE interrupt mode 8619 is active. This option allows the XIVE driver to use 8620 loads instead, as on POWER9. 8621 8622 xhci-hcd.quirks [USB,KNL] 8623 A hex value specifying bitmask with supplemental xhci 8624 host controller quirks. Meaning of each bit can be 8625 consulted in header drivers/usb/host/xhci.h. 8626 8627 xmon [PPC,EARLY] 8628 Format: { early | on | rw | ro | off } 8629 Controls if xmon debugger is enabled. Default is off. 8630 Passing only "xmon" is equivalent to "xmon=early". 8631 early Call xmon as early as possible on boot; xmon 8632 debugger is called from setup_arch(). 8633 on xmon debugger hooks will be installed so xmon 8634 is only called on a kernel crash. Default mode, 8635 i.e. either "ro" or "rw" mode, is controlled 8636 with CONFIG_XMON_DEFAULT_RO_MODE. 8637 rw xmon debugger hooks will be installed so xmon 8638 is called only on a kernel crash, mode is write, 8639 meaning SPR registers, memory and, other data 8640 can be written using xmon commands. 8641 ro same as "rw" option above but SPR registers, 8642 memory, and other data can't be written using 8643 xmon commands. 8644 off xmon is disabled. 8645