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 ISDN Appropriate ISDN support is enabled. 28 ISOL CPU Isolation is enabled. 29 JOY Appropriate joystick support is enabled. 30 KGDB Kernel debugger support is enabled. 31 KVM Kernel Virtual Machine support is enabled. 32 LIBATA Libata driver is enabled 33 LOONGARCH LoongArch architecture is enabled. 34 LOOP Loopback device support is enabled. 35 LP Printer support is enabled. 36 M68k M68k architecture is enabled. 37 These options have more detailed description inside of 38 Documentation/arch/m68k/kernel-options.rst. 39 MDA MDA console support is enabled. 40 MIPS MIPS architecture is enabled. 41 MOUSE Appropriate mouse support is enabled. 42 MSI Message Signaled Interrupts (PCI). 43 MTD MTD (Memory Technology Device) support is enabled. 44 NET Appropriate network support is enabled. 45 NFS Appropriate NFS support is enabled. 46 NUMA NUMA support is enabled. 47 OF Devicetree is enabled. 48 PARISC The PA-RISC architecture is enabled. 49 PCI PCI bus support is enabled. 50 PCIE PCI Express support is enabled. 51 PCMCIA The PCMCIA subsystem is enabled. 52 PNP Plug & Play support is enabled. 53 PPC PowerPC architecture is enabled. 54 PPT Parallel port support is enabled. 55 PS2 Appropriate PS/2 support is enabled. 56 PV_OPS A paravirtualized kernel is enabled. 57 RAM RAM disk support is enabled. 58 RDT Intel Resource Director Technology. 59 RISCV RISCV architecture is enabled. 60 S390 S390 architecture is enabled. 61 SCSI Appropriate SCSI support is enabled. 62 A lot of drivers have their options described inside 63 the Documentation/scsi/ sub-directory. 64 SDW SoundWire support is enabled. 65 SECURITY Different security models are enabled. 66 SELINUX SELinux support is enabled. 67 SERIAL Serial support is enabled. 68 SH SuperH architecture is enabled. 69 SMP The kernel is an SMP kernel. 70 SPARC Sparc architecture is enabled. 71 SUSPEND System suspend states are enabled. 72 SWSUSP Software suspend (hibernation) is enabled. 73 TPM TPM drivers are enabled. 74 UMS USB Mass Storage support is enabled. 75 USB USB support is enabled. 76 NVME NVMe support is enabled 77 USBHID USB Human Interface Device support is enabled. 78 V4L Video For Linux support is enabled. 79 VGA The VGA console has been enabled. 80 VMMIO Driver for memory mapped virtio devices is enabled. 81 VT Virtual terminal support is enabled. 82 WDT Watchdog support is enabled. 83 X86-32 X86-32, aka i386 architecture is enabled. 84 X86-64 X86-64 architecture is enabled. 85 X86 Either 32-bit or 64-bit x86 (same as X86-32+X86-64) 86 X86_UV SGI UV support is enabled. 87 XEN Xen support is enabled 88 XTENSA xtensa architecture is enabled. 89 90In addition, the following text indicates that the option 91 92 BOOT Is a boot loader parameter. 93 BUGS= Relates to possible processor bugs on the said processor. 94 KNL Is a kernel start-up parameter. 95 96 97Kernel parameters 98 99 accept_memory= [MM] 100 Format: { eager | lazy } 101 default: lazy 102 By default, unaccepted memory is accepted lazily to 103 avoid prolonged boot times. The lazy option will add 104 some runtime overhead until all memory is eventually 105 accepted. In most cases the overhead is negligible. 106 For some workloads or for debugging purposes 107 accept_memory=eager can be used to accept all memory 108 at once during boot. 109 110 acpi= [HW,ACPI,X86,ARM64,RISCV64,EARLY] 111 Advanced Configuration and Power Interface 112 Format: { force | on | off | strict | noirq | rsdt | 113 copy_dsdt | nospcr } 114 force -- enable ACPI if default was off 115 on -- enable ACPI but allow fallback to DT [arm64,riscv64] 116 off -- disable ACPI if default was on 117 noirq -- do not use ACPI for IRQ routing 118 strict -- Be less tolerant of platforms that are not 119 strictly ACPI specification compliant. 120 rsdt -- prefer RSDT over (default) XSDT 121 copy_dsdt -- copy DSDT to memory 122 nocmcff -- Disable firmware first mode for corrected 123 errors. This disables parsing the HEST CMC error 124 source to check if firmware has set the FF flag. This 125 may result in duplicate corrected error reports. 126 nospcr -- disable console in ACPI SPCR table as 127 default _serial_ console on ARM64 128 spcr -- enable console in ACPI SPCR table as 129 default _serial_ console on x86 130 For ARM64, ONLY "acpi=off", "acpi=on", "acpi=force" or 131 "acpi=nospcr" are available 132 For RISCV64, ONLY "acpi=off", "acpi=on" or "acpi=force" 133 are available 134 135 See also Documentation/power/runtime_pm.rst, pci=noacpi 136 137 acpi_apic_instance= [ACPI,IOAPIC,EARLY] 138 Format: <int> 139 2: use 2nd APIC table, if available 140 1,0: use 1st APIC table 141 default: 0 142 143 acpi_backlight= [HW,ACPI] 144 { vendor | video | native | none } 145 If set to vendor, prefer vendor-specific driver 146 (e.g. thinkpad_acpi, sony_acpi, etc.) instead 147 of the ACPI video.ko driver. 148 If set to video, use the ACPI video.ko driver. 149 If set to native, use the device's native backlight mode. 150 If set to none, disable the ACPI backlight interface. 151 152 acpi_force_32bit_fadt_addr [ACPI,EARLY] 153 force FADT to use 32 bit addresses rather than the 154 64 bit X_* addresses. Some firmware have broken 64 155 bit addresses for force ACPI ignore these and use 156 the older legacy 32 bit addresses. 157 158 acpica_no_return_repair [HW, ACPI] 159 Disable AML predefined validation mechanism 160 This mechanism can repair the evaluation result to make 161 the return objects more ACPI specification compliant. 162 This option is useful for developers to identify the 163 root cause of an AML interpreter issue when the issue 164 has something to do with the repair mechanism. 165 166 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG] 167 acpi.debug_level= [HW,ACPI,ACPI_DEBUG] 168 Format: <int> 169 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI 170 debug output. Bits in debug_layer correspond to a 171 _COMPONENT in an ACPI source file, e.g., 172 #define _COMPONENT ACPI_EVENTS 173 Bits in debug_level correspond to a level in 174 ACPI_DEBUG_PRINT statements, e.g., 175 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ... 176 The debug_level mask defaults to "info". See 177 Documentation/firmware-guide/acpi/debug.rst for more information about 178 debug layers and levels. 179 180 Enable processor driver info messages: 181 acpi.debug_layer=0x20000000 182 Enable AML "Debug" output, i.e., stores to the Debug 183 object while interpreting AML: 184 acpi.debug_layer=0xffffffff acpi.debug_level=0x2 185 Enable all messages related to ACPI hardware: 186 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff 187 188 Some values produce so much output that the system is 189 unusable. The "log_buf_len" parameter may be useful 190 if you need to capture more output. 191 192 acpi.poweroff_on_fatal= [ACPI] 193 {0 | 1} 194 Causes the system to poweroff when the ACPI bytecode signals 195 a fatal error. The default value of this setting is 1. 196 Overriding this value should only be done for diagnosing 197 ACPI firmware problems, as the system might behave erratically 198 after having encountered a fatal ACPI error. 199 200 acpi_enforce_resources= [ACPI] 201 { strict | lax | no } 202 Check for resource conflicts between native drivers 203 and ACPI OperationRegions (SystemIO and SystemMemory 204 only). IO ports and memory declared in ACPI might be 205 used by the ACPI subsystem in arbitrary AML code and 206 can interfere with legacy drivers. 207 strict (default): access to resources claimed by ACPI 208 is denied; legacy drivers trying to access reserved 209 resources will fail to bind to device using them. 210 lax: access to resources claimed by ACPI is allowed; 211 legacy drivers trying to access reserved resources 212 will bind successfully but a warning message is logged. 213 no: ACPI OperationRegions are not marked as reserved, 214 no further checks are performed. 215 216 acpi_force_table_verification [HW,ACPI,EARLY] 217 Enable table checksum verification during early stage. 218 By default, this is disabled due to x86 early mapping 219 size limitation. 220 221 acpi_irq_balance [HW,ACPI] 222 ACPI will balance active IRQs 223 default in APIC mode 224 225 acpi_irq_nobalance [HW,ACPI] 226 ACPI will not move active IRQs (default) 227 default in PIC mode 228 229 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA 230 Format: <irq>,<irq>... 231 232 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for 233 use by PCI 234 Format: <irq>,<irq>... 235 236 acpi_mask_gpe= [HW,ACPI] 237 Due to the existence of _Lxx/_Exx, some GPEs triggered 238 by unsupported hardware/firmware features can result in 239 GPE floodings that cannot be automatically disabled by 240 the GPE dispatcher. 241 This facility can be used to prevent such uncontrolled 242 GPE floodings. 243 Format: <byte> or <bitmap-list> 244 245 acpi_no_auto_serialize [HW,ACPI] 246 Disable auto-serialization of AML methods 247 AML control methods that contain the opcodes to create 248 named objects will be marked as "Serialized" by the 249 auto-serialization feature. 250 This feature is enabled by default. 251 This option allows to turn off the feature. 252 253 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump 254 kernels. 255 256 acpi_no_static_ssdt [HW,ACPI,EARLY] 257 Disable installation of static SSDTs at early boot time 258 By default, SSDTs contained in the RSDT/XSDT will be 259 installed automatically and they will appear under 260 /sys/firmware/acpi/tables. 261 This option turns off this feature. 262 Note that specifying this option does not affect 263 dynamic table installation which will install SSDT 264 tables to /sys/firmware/acpi/tables/dynamic. 265 266 acpi_no_watchdog [HW,ACPI,WDT] 267 Ignore the ACPI-based watchdog interface (WDAT) and let 268 a native driver control the watchdog device instead. 269 270 acpi_rsdp= [ACPI,EFI,KEXEC,EARLY] 271 Pass the RSDP address to the kernel, mostly used 272 on machines running EFI runtime service to boot the 273 second kernel for kdump. 274 275 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS 276 Format: To spoof as Windows 98: ="Microsoft Windows" 277 278 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead 279 of 2 which is mandated by ACPI 6) as the supported ACPI 280 specification revision (when using this switch, it may 281 be necessary to carry out a cold reboot _twice_ in a 282 row to make it take effect on the platform firmware). 283 284 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings 285 acpi_osi="string1" # add string1 286 acpi_osi="!string2" # remove string2 287 acpi_osi=!* # remove all strings 288 acpi_osi=! # disable all built-in OS vendor 289 strings 290 acpi_osi=!! # enable all built-in OS vendor 291 strings 292 acpi_osi= # disable all strings 293 294 'acpi_osi=!' can be used in combination with single or 295 multiple 'acpi_osi="string1"' to support specific OS 296 vendor string(s). Note that such command can only 297 affect the default state of the OS vendor strings, thus 298 it cannot affect the default state of the feature group 299 strings and the current state of the OS vendor strings, 300 specifying it multiple times through kernel command line 301 is meaningless. This command is useful when one do not 302 care about the state of the feature group strings which 303 should be controlled by the OSPM. 304 Examples: 305 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent 306 to 'acpi_osi="Windows 2000" acpi_osi=!', they all 307 can make '_OSI("Windows 2000")' TRUE. 308 309 'acpi_osi=' cannot be used in combination with other 310 'acpi_osi=' command lines, the _OSI method will not 311 exist in the ACPI namespace. NOTE that such command can 312 only affect the _OSI support state, thus specifying it 313 multiple times through kernel command line is also 314 meaningless. 315 Examples: 316 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)' 317 FALSE. 318 319 'acpi_osi=!*' can be used in combination with single or 320 multiple 'acpi_osi="string1"' to support specific 321 string(s). Note that such command can affect the 322 current state of both the OS vendor strings and the 323 feature group strings, thus specifying it multiple times 324 through kernel command line is meaningful. But it may 325 still not able to affect the final state of a string if 326 there are quirks related to this string. This command 327 is useful when one want to control the state of the 328 feature group strings to debug BIOS issues related to 329 the OSPM features. 330 Examples: 331 1. 'acpi_osi="Module Device" acpi_osi=!*' can make 332 '_OSI("Module Device")' FALSE. 333 2. 'acpi_osi=!* acpi_osi="Module Device"' can make 334 '_OSI("Module Device")' TRUE. 335 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is 336 equivalent to 337 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"' 338 and 339 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!', 340 they all will make '_OSI("Windows 2000")' TRUE. 341 342 acpi_pm_good [X86] 343 Override the pmtimer bug detection: force the kernel 344 to assume that this machine's pmtimer latches its value 345 and always returns good values. 346 347 acpi_sci= [HW,ACPI,EARLY] ACPI System Control Interrupt trigger mode 348 Format: { level | edge | high | low } 349 350 acpi_skip_timer_override [HW,ACPI,EARLY] 351 Recognize and ignore IRQ0/pin2 Interrupt Override. 352 For broken nForce2 BIOS resulting in XT-PIC timer. 353 354 acpi_sleep= [HW,ACPI] Sleep options 355 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig, 356 s4_nohwsig, old_ordering, nonvs, 357 sci_force_enable, nobl } 358 See Documentation/power/video.rst for information on 359 s3_bios and s3_mode. 360 s3_beep is for debugging; it makes the PC's speaker beep 361 as soon as the kernel's real-mode entry point is called. 362 s4_hwsig causes the kernel to check the ACPI hardware 363 signature during resume from hibernation, and gracefully 364 refuse to resume if it has changed. This complies with 365 the ACPI specification but not with reality, since 366 Windows does not do this and many laptops do change it 367 on docking. So the default behaviour is to allow resume 368 and simply warn when the signature changes, unless the 369 s4_hwsig option is enabled. 370 s4_nohwsig prevents ACPI hardware signature from being 371 used (or even warned about) during resume. 372 old_ordering causes the ACPI 1.0 ordering of the _PTS 373 control method, with respect to putting devices into 374 low power states, to be enforced (the ACPI 2.0 ordering 375 of _PTS is used by default). 376 nonvs prevents the kernel from saving/restoring the 377 ACPI NVS memory during suspend/hibernation and resume. 378 sci_force_enable causes the kernel to set SCI_EN directly 379 on resume from S1/S3 (which is against the ACPI spec, 380 but some broken systems don't work without it). 381 nobl causes the internal blacklist of systems known to 382 behave incorrectly in some ways with respect to system 383 suspend and resume to be ignored (use wisely). 384 385 acpi_use_timer_override [HW,ACPI,EARLY] 386 Use timer override. For some broken Nvidia NF5 boards 387 that require a timer override, but don't have HPET 388 389 add_efi_memmap [EFI,X86,EARLY] Include EFI memory map in 390 kernel's map of available physical RAM. 391 392 agp= [AGP] 393 { off | try_unsupported } 394 off: disable AGP support 395 try_unsupported: try to drive unsupported chipsets 396 (may crash computer or cause data corruption) 397 398 ALSA [HW,ALSA] 399 See Documentation/sound/alsa-configuration.rst 400 401 alignment= [KNL,ARM] 402 Allow the default userspace alignment fault handler 403 behaviour to be specified. Bit 0 enables warnings, 404 bit 1 enables fixups, and bit 2 sends a segfault. 405 406 align_va_addr= [X86-64] 407 Align virtual addresses by clearing slice [14:12] when 408 allocating a VMA at process creation time. This option 409 gives you up to 3% performance improvement on AMD F15h 410 machines (where it is enabled by default) for a 411 CPU-intensive style benchmark, and it can vary highly in 412 a microbenchmark depending on workload and compiler. 413 414 32: only for 32-bit processes 415 64: only for 64-bit processes 416 on: enable for both 32- and 64-bit processes 417 off: disable for both 32- and 64-bit processes 418 419 alloc_snapshot [FTRACE] 420 Allocate the ftrace snapshot buffer on boot up when the 421 main buffer is allocated. This is handy if debugging 422 and you need to use tracing_snapshot() on boot up, and 423 do not want to use tracing_snapshot_alloc() as it needs 424 to be done where GFP_KERNEL allocations are allowed. 425 426 allow_mismatched_32bit_el0 [ARM64,EARLY] 427 Allow execve() of 32-bit applications and setting of the 428 PER_LINUX32 personality on systems where only a strict 429 subset of the CPUs support 32-bit EL0. When this 430 parameter is present, the set of CPUs supporting 32-bit 431 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0 432 and hot-unplug operations may be restricted. 433 434 See Documentation/arch/arm64/asymmetric-32bit.rst for more 435 information. 436 437 amd_iommu= [HW,X86-64] 438 Pass parameters to the AMD IOMMU driver in the system. 439 Possible values are: 440 fullflush - Deprecated, equivalent to iommu.strict=1 441 off - do not initialize any AMD IOMMU found in 442 the system 443 force_isolation - Force device isolation for all 444 devices. The IOMMU driver is not 445 allowed anymore to lift isolation 446 requirements as needed. This option 447 does not override iommu=pt 448 force_enable - Force enable the IOMMU on platforms known 449 to be buggy with IOMMU enabled. Use this 450 option with care. 451 pgtbl_v1 - Use v1 page table for DMA-API (Default). 452 pgtbl_v2 - Use v2 page table for DMA-API. 453 irtcachedis - Disable Interrupt Remapping Table (IRT) caching. 454 nohugepages - Limit page-sizes used for v1 page-tables 455 to 4 KiB. 456 v2_pgsizes_only - Limit page-sizes used for v1 page-tables 457 to 4KiB/2Mib/1GiB. 458 459 460 amd_iommu_dump= [HW,X86-64] 461 Enable AMD IOMMU driver option to dump the ACPI table 462 for AMD IOMMU. With this option enabled, AMD IOMMU 463 driver will print ACPI tables for AMD IOMMU during 464 IOMMU initialization. 465 466 amd_iommu_intr= [HW,X86-64] 467 Specifies one of the following AMD IOMMU interrupt 468 remapping modes: 469 legacy - Use legacy interrupt remapping mode. 470 vapic - Use virtual APIC mode, which allows IOMMU 471 to inject interrupts directly into guest. 472 This mode requires kvm-amd.avic=1. 473 (Default when IOMMU HW support is present.) 474 475 amd_pstate= [X86,EARLY] 476 disable 477 Do not enable amd_pstate as the default 478 scaling driver for the supported processors 479 passive 480 Use amd_pstate with passive mode as a scaling driver. 481 In this mode autonomous selection is disabled. 482 Driver requests a desired performance level and platform 483 tries to match the same performance level if it is 484 satisfied by guaranteed performance level. 485 active 486 Use amd_pstate_epp driver instance as the scaling driver, 487 driver provides a hint to the hardware if software wants 488 to bias toward performance (0x0) or energy efficiency (0xff) 489 to the CPPC firmware. then CPPC power algorithm will 490 calculate the runtime workload and adjust the realtime cores 491 frequency. 492 guided 493 Activate guided autonomous mode. Driver requests minimum and 494 maximum performance level and the platform autonomously 495 selects a performance level in this range and appropriate 496 to the current workload. 497 498 amd_prefcore= 499 [X86] 500 disable 501 Disable amd-pstate preferred core. 502 503 amd_dynamic_epp= 504 [X86] 505 disable 506 Disable amd-pstate dynamic EPP. 507 enable 508 Enable amd-pstate dynamic EPP. 509 510 amijoy.map= [HW,JOY] Amiga joystick support 511 Map of devices attached to JOY0DAT and JOY1DAT 512 Format: <a>,<b> 513 See also Documentation/input/joydev/joystick.rst 514 515 analog.map= [HW,JOY] Analog joystick and gamepad support 516 Specifies type or capabilities of an analog joystick 517 connected to one of 16 gameports 518 Format: <type1>,<type2>,..<type16> 519 520 apc= [HW,SPARC] 521 Power management functions (SPARCstation-4/5 + deriv.) 522 Format: noidle 523 Disable APC CPU standby support. SPARCstation-Fox does 524 not play well with APC CPU idle - disable it if you have 525 APC and your system crashes randomly. 526 527 apic [APIC,X86-64] Use IO-APIC. Default. 528 529 apic= [APIC,X86,EARLY] Advanced Programmable Interrupt Controller 530 Change the output verbosity while booting 531 Format: { quiet (default) | verbose | debug } 532 Change the amount of debugging information output 533 when initialising the APIC and IO-APIC components. 534 535 apic_extnmi= [APIC,X86,EARLY] External NMI delivery setting 536 Format: { bsp (default) | all | none } 537 bsp: External NMI is delivered only to CPU 0 538 all: External NMIs are broadcast to all CPUs as a 539 backup of CPU 0 540 none: External NMI is masked for all CPUs. This is 541 useful so that a dump capture kernel won't be 542 shot down by NMI 543 544 apicpmtimer Do APIC timer calibration using the pmtimer. Implies 545 apicmaintimer. Useful when your PIT timer is totally 546 broken. 547 548 autoconf= [IPV6] 549 See Documentation/networking/ipv6.rst. 550 551 apm= [APM] Advanced Power Management 552 See header of arch/x86/kernel/apm_32.c. 553 554 apparmor= [APPARMOR] Disable or enable AppArmor at boot time 555 Format: { "0" | "1" } 556 See security/apparmor/Kconfig help text 557 0 -- disable. 558 1 -- enable. 559 Default value is set via kernel config option. 560 561 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards 562 Format: <io>,<irq>,<nodeID> 563 564 arm64.no32bit_el0 [ARM64] Unconditionally disable the execution of 565 32 bit applications. 566 567 arm64.nobti [ARM64] Unconditionally disable Branch Target 568 Identification support 569 570 arm64.nogcs [ARM64] Unconditionally disable Guarded Control Stack 571 support 572 573 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory 574 Set instructions support 575 576 arm64.nompam [ARM64] Unconditionally disable Memory Partitioning And 577 Monitoring support 578 579 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension 580 support 581 582 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication 583 support 584 585 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix 586 Extension support 587 588 arm64.nosve [ARM64] Unconditionally disable Scalable Vector 589 Extension support 590 591 ataflop= [HW,M68k] 592 593 atarimouse= [HW,MOUSE] Atari Mouse 594 595 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess, 596 EzKey and similar keyboards 597 598 atkbd.reset= [HW] Reset keyboard during initialization 599 600 atkbd.set= [HW] Select keyboard code set 601 Format: <int> (2 = AT (default), 3 = PS/2) 602 603 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar 604 keyboards 605 606 atkbd.softraw= [HW] Choose between synthetic and real raw mode 607 Format: <bool> (0 = real, 1 = synthetic (default)) 608 609 atkbd.softrepeat= [HW] 610 Use software keyboard repeat 611 612 audit= [KNL] Enable the audit sub-system 613 Format: { "0" | "1" | "off" | "on" } 614 0 | off - kernel audit is disabled and can not be 615 enabled until the next reboot 616 unset - kernel audit is initialized but disabled and 617 will be fully enabled by the userspace auditd. 618 1 | on - kernel audit is initialized and partially 619 enabled, storing at most audit_backlog_limit 620 messages in RAM until it is fully enabled by the 621 userspace auditd. 622 Default: unset 623 624 audit_backlog_limit= [KNL] Set the audit queue size limit. 625 Format: <int> (must be >=0) 626 Default: 64 627 628 bau= [X86_UV] Enable the BAU on SGI UV. The default 629 behavior is to disable the BAU (i.e. bau=0). 630 Format: { "0" | "1" } 631 0 - Disable the BAU. 632 1 - Enable the BAU. 633 unset - Disable the BAU. 634 635 bdev_allow_write_mounted= 636 Format: <bool> 637 Control the ability to open a mounted block device 638 for writing, i.e., allow / disallow writes that bypass 639 the FS. This was implemented as a means to prevent 640 fuzzers from crashing the kernel by overwriting the 641 metadata underneath a mounted FS without its awareness. 642 This also prevents destructive formatting of mounted 643 filesystems by naive storage tooling that don't use 644 O_EXCL. Default is Y and can be changed through the 645 Kconfig option CONFIG_BLK_DEV_WRITE_MOUNTED. 646 647 bert_disable [ACPI] 648 Disable BERT OS support on buggy BIOSes. 649 650 bgrt_disable [ACPI,X86,EARLY] 651 Disable BGRT to avoid flickering OEM logo. 652 653 blkdevparts= Manual partition parsing of block device(s) for 654 embedded devices based on command line input. 655 See Documentation/block/cmdline-partition.rst 656 657 boot_delay= [KNL,EARLY] 658 Milliseconds to delay each printk during boot. 659 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled, 660 and you may also have to specify "lpj=". Boot_delay 661 values larger than 10 seconds (10000) are assumed 662 erroneous and ignored. 663 Format: integer 664 665 bootconfig [KNL,EARLY] 666 Extended command line options can be added to an initrd 667 and this will cause the kernel to look for it. 668 669 See Documentation/admin-guide/bootconfig.rst 670 671 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards) 672 bttv.radio= Most important insmod options are available as 673 kernel args too. 674 bttv.pll= See Documentation/admin-guide/media/bttv.rst 675 bttv.tuner= 676 677 bulk_remove=off [PPC] This parameter disables the use of the pSeries 678 firmware feature for flushing multiple hpte entries 679 at a time. 680 681 c101= [NET] Moxa C101 synchronous serial card 682 683 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection. 684 Sometimes CPU hardware bugs make them report the cache 685 size incorrectly. The kernel will attempt work arounds 686 to fix known problems, but for some CPUs it is not 687 possible to determine what the correct size should be. 688 This option provides an override for these situations. 689 690 carrier_timeout= 691 [NET] Specifies amount of time (in seconds) that 692 the kernel should wait for a network carrier. By default 693 it waits 120 seconds. 694 695 ca_keys= [KEYS] This parameter identifies a specific key(s) on 696 the system trusted keyring to be used for certificate 697 trust validation. 698 format: { id:<keyid> | builtin } 699 700 cca= [MIPS,EARLY] Override the kernel pages' cache coherency 701 algorithm. Accepted values range from 0 to 7 702 inclusive. See arch/mips/include/asm/pgtable-bits.h 703 for platform specific values (SB1, Loongson3 and 704 others). 705 706 ccw_timeout_log [S390] 707 See Documentation/arch/s390/common_io.rst for details. 708 709 cfi= [X86-64] Set Control Flow Integrity checking features 710 when CONFIG_FINEIBT is enabled. 711 Format: feature[,feature...] 712 Default: auto 713 714 auto: Use FineIBT if IBT available, otherwise kCFI. 715 Under FineIBT, enable "paranoid" mode when 716 FRED is not available. 717 off: Turn off CFI checking. 718 kcfi: Use kCFI (disable FineIBT). 719 fineibt: Use FineIBT (even if IBT not available). 720 norand: Do not re-randomize CFI hashes. 721 paranoid: Add caller hash checking under FineIBT. 722 bhi: Enable register poisoning to stop speculation 723 across FineIBT. (Disabled by default.) 724 warn: Do not enforce CFI checking: warn only. 725 debug: Report CFI initialization details. 726 727 cgroup_disable= [KNL] Disable a particular controller or optional feature 728 Format: {name of the controller(s) or feature(s) to disable} 729 The effects of cgroup_disable=foo are: 730 - foo isn't auto-mounted if you mount all cgroups in 731 a single hierarchy 732 - foo isn't visible as an individually mountable 733 subsystem 734 - if foo is an optional feature then the feature is 735 disabled and corresponding cgroup files are not 736 created 737 {Currently only "memory" controller deal with this and 738 cut the overhead, others just disable the usage. So 739 only cgroup_disable=memory is actually worthy} 740 Specifying "pressure" disables per-cgroup pressure 741 stall information accounting feature 742 743 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1 744 Format: { { controller | "all" | "named" } 745 [,{ controller | "all" | "named" }...] } 746 Like cgroup_disable, but only applies to cgroup v1; 747 the blacklisted controllers remain available in cgroup2. 748 "all" blacklists all controllers and "named" disables 749 named mounts. Specifying both "all" and "named" disables 750 all v1 hierarchies. 751 752 cgroup_v1_proc= [KNL] Show also missing controllers in /proc/cgroups 753 Format: { "true" | "false" } 754 /proc/cgroups lists only v1 controllers by default. 755 This compatibility option enables listing also v2 756 controllers (whose v1 code is not compiled!), so that 757 semi-legacy software can check this file to decide 758 about usage of v2 (sic) controllers. 759 760 cgroup_favordynmods= [KNL] Enable or Disable favordynmods. 761 Format: { "true" | "false" } 762 Defaults to the value of CONFIG_CGROUP_FAVOR_DYNMODS. 763 764 cgroup.memory= [KNL] Pass options to the cgroup memory controller. 765 Format: <string> 766 nosocket -- Disable socket memory accounting. 767 nokmem -- Disable kernel memory accounting. 768 nobpf -- Disable BPF memory accounting. 769 770 check_pages= [MM,EARLY] Enable sanity checking of pages after 771 allocations / before freeing. This adds checks to catch 772 double-frees, use-after-frees, and other sources of 773 page corruption by inspecting page internals (flags, 774 mapcount/refcount, memcg_data, etc.). 775 Format: { "0" | "1" } 776 Default: 0 (1 if CONFIG_DEBUG_VM is set) 777 778 checkreqprot= [SELINUX] Set initial checkreqprot flag value. 779 Format: { "0" | "1" } 780 See security/selinux/Kconfig help text. 781 0 -- check protection applied by kernel (includes 782 any implied execute protection). 783 1 -- check protection requested by application. 784 Default value is set via a kernel config option. 785 Value can be changed at runtime via 786 /sys/fs/selinux/checkreqprot. 787 Setting checkreqprot to 1 is deprecated. 788 789 cio_ignore= [S390] 790 See Documentation/arch/s390/common_io.rst for details. 791 792 clk_ignore_unused 793 [CLK] 794 Prevents the clock framework from automatically gating 795 clocks that have not been explicitly enabled by a Linux 796 device driver but are enabled in hardware at reset or 797 by the bootloader/firmware. Note that this does not 798 force such clocks to be always-on nor does it reserve 799 those clocks in any way. This parameter is useful for 800 debug and development, but should not be needed on a 801 platform with proper driver support. For more 802 information, see Documentation/driver-api/clk.rst. 803 804 clock= [BUGS=X86-32, HW] gettimeofday clocksource override. 805 [Deprecated] 806 Forces specified clocksource (if available) to be used 807 when calculating gettimeofday(). If specified 808 clocksource is not available, it defaults to PIT. 809 Format: { pit | tsc | cyclone | pmtmr } 810 811 clocksource= Override the default clocksource 812 Format: <string> 813 Override the default clocksource and use the clocksource 814 with the name specified. 815 Some clocksource names to choose from, depending on 816 the platform: 817 [all] jiffies (this is the base, fallback clocksource) 818 [ACPI] acpi_pm 819 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2, 820 pxa_timer,timer3,32k_counter,timer0_1 821 [X86-32] pit,hpet,tsc; 822 scx200_hrt on Geode; cyclone on IBM x440 823 [MIPS] MIPS 824 [PARISC] cr16 825 [S390] tod 826 [SH] SuperH 827 [SPARC64] tick 828 [X86-64] hpet,tsc 829 830 clocksource.arm_arch_timer.evtstrm= 831 [ARM,ARM64,EARLY] 832 Format: <bool> 833 Enable/disable the eventstream feature of the ARM 834 architected timer so that code using WFE-based polling 835 loops can be debugged more effectively on production 836 systems. 837 838 clocksource.verify_n_cpus= [KNL] 839 Limit the number of CPUs checked for clocksources 840 marked with CLOCK_SOURCE_VERIFY_PERCPU that 841 are marked unstable due to excessive skew. 842 A negative value says to check all CPUs, while 843 zero says not to check any. Values larger than 844 nr_cpu_ids are silently truncated to nr_cpu_ids. 845 The actual CPUs are chosen randomly, with 846 no replacement if the same CPU is chosen twice. 847 848 clocksource-wdtest.holdoff= [KNL] 849 Set the time in seconds that the clocksource 850 watchdog test waits before commencing its tests. 851 Defaults to zero when built as a module and to 852 10 seconds when built into the kernel. 853 854 cma=nn[MG]@[start[MG][-end[MG]]] 855 [KNL,CMA,EARLY] 856 Sets the size of kernel global memory area for 857 contiguous memory allocations and optionally the 858 placement constraint by the physical address range of 859 memory allocations. A value of 0 disables CMA 860 altogether. For more information, see 861 kernel/dma/contiguous.c 862 863 cma_pernuma=nn[MG] 864 [KNL,CMA,EARLY] 865 Sets the size of kernel per-numa memory area for 866 contiguous memory allocations. A value of 0 disables 867 per-numa CMA altogether. And If this option is not 868 specified, the default value is 0. 869 With per-numa CMA enabled, DMA users on node nid will 870 first try to allocate buffer from the pernuma area 871 which is located in node nid, if the allocation fails, 872 they will fallback to the global default memory area. 873 874 numa_cma=<node>:nn[MG][,<node>:nn[MG]] 875 [KNL,CMA,EARLY] 876 Sets the size of kernel numa memory area for 877 contiguous memory allocations. It will reserve CMA 878 area for the specified node. 879 880 With numa CMA enabled, DMA users on node nid will 881 first try to allocate buffer from the numa area 882 which is located in node nid, if the allocation fails, 883 they will fallback to the global default memory area. 884 885 cmo_free_hint= [PPC] Format: { yes | no } 886 Specify whether pages are marked as being inactive 887 when they are freed. This is used in CMO environments 888 to determine OS memory pressure for page stealing by 889 a hypervisor. 890 Default: yes 891 892 coherent_pool=nn[KMG] [ARM,KNL,EARLY] 893 Sets the size of memory pool for coherent, atomic dma 894 allocations, by default set to 256K. 895 896 com20020= [HW,NET] ARCnet - COM20020 chipset 897 Format: 898 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]] 899 900 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers) 901 Format: <io>[,<irq>] 902 903 com90xx= [HW,NET] 904 ARCnet - COM90xx chipset (memory-mapped buffers) 905 Format: <io>[,<irq>[,<memstart>]] 906 907 condev= [HW,S390] console device 908 conmode= 909 910 con3215_drop= [S390,EARLY] 3215 console drop mode. 911 Format: y|n|Y|N|1|0 912 When set to true, drop data on the 3215 console when 913 the console buffer is full. In this case the 914 operator using a 3270 terminal emulator (for example 915 x3270) does not have to enter the clear key for the 916 console output to advance and the kernel to continue. 917 This leads to a much faster boot time when a 3270 918 terminal emulator is active. If no 3270 terminal 919 emulator is used, this parameter has no effect. 920 921 console= [KNL] Output console device and options. 922 923 tty<n> Use the virtual console device <n>. 924 925 ttyS<n>[,options] 926 ttyUSB0[,options] 927 Use the specified serial port. The options are of 928 the form "bbbbpnf", where "bbbb" is the baud rate, 929 "p" is parity ("n", "o", or "e"), "n" is number of 930 bits, and "f" is flow control ("r" for RTS or 931 omit it). Default is "9600n8". 932 933 See Documentation/admin-guide/serial-console.rst for more 934 information. See 935 Documentation/networking/netconsole.rst for an 936 alternative. 937 938 <DEVNAME>:<n>.<n>[,options] 939 Use the specified serial port on the serial core bus. 940 The addressing uses DEVNAME of the physical serial port 941 device, followed by the serial core controller instance, 942 and the serial port instance. The options are the same 943 as documented for the ttyS addressing above. 944 945 The mapping of the serial ports to the tty instances 946 can be viewed with: 947 948 $ ls -d /sys/bus/serial-base/devices/*:*.*/tty/* 949 /sys/bus/serial-base/devices/00:04:0.0/tty/ttyS0 950 951 In the above example, the console can be addressed with 952 console=00:04:0.0. Note that a console addressed this 953 way will only get added when the related device driver 954 is ready. The use of an earlycon parameter in addition to 955 the console may be desired for console output early on. 956 957 uart[8250],io,<addr>[,options] 958 uart[8250],mmio,<addr>[,options] 959 uart[8250],mmio16,<addr>[,options] 960 uart[8250],mmio32,<addr>[,options] 961 uart[8250],0x<addr>[,options] 962 Start an early, polled-mode console on the 8250/16550 963 UART at the specified I/O port or MMIO address, 964 switching to the matching ttyS device later. 965 MMIO inter-register address stride is either 8-bit 966 (mmio), 16-bit (mmio16), or 32-bit (mmio32). 967 If none of [io|mmio|mmio16|mmio32], <addr> is assumed 968 to be equivalent to 'mmio'. 'options' are specified in 969 the same format described for ttyS above; if unspecified, 970 the h/w is not re-initialized. 971 972 hvc<n> Use the hypervisor console device <n>. This is for 973 both Xen and PowerPC hypervisors. 974 975 { null | "" } 976 Use to disable console output, i.e., to have kernel 977 console messages discarded. 978 This must be the only console= parameter used on the 979 kernel command line. 980 981 If the device connected to the port is not a TTY but a braille 982 device, prepend "brl," before the device type, for instance 983 console=brl,ttyS0 984 For now, only VisioBraille is supported. 985 986 console_msg_format= 987 [KNL] Change console messages format 988 default 989 By default we print messages on consoles in 990 "[time stamp] text\n" format (time stamp may not be 991 printed, depending on CONFIG_PRINTK_TIME or 992 `printk_time' param). 993 syslog 994 Switch to syslog format: "<%u>[time stamp] text\n" 995 IOW, each message will have a facility and loglevel 996 prefix. The format is similar to one used by syslog() 997 syscall, or to executing "dmesg -S --raw" or to reading 998 from /proc/kmsg. 999 1000 consoleblank= [KNL] The console blank (screen saver) timeout in 1001 seconds. A value of 0 disables the blank timer. 1002 Defaults to 0. 1003 1004 coredump_filter= 1005 [KNL] Change the default value for 1006 /proc/<pid>/coredump_filter. 1007 See also Documentation/filesystems/proc.rst. 1008 1009 coresight_cpu_debug.enable 1010 [ARM,ARM64] 1011 Format: <bool> 1012 Enable/disable the CPU sampling based debugging. 1013 0: default value, disable debugging 1014 1: enable debugging at boot time 1015 1016 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver 1017 Format: 1018 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>] 1019 1020 cpuidle.off=1 [CPU_IDLE] 1021 disable the cpuidle sub-system 1022 1023 cpuidle.governor= 1024 [CPU_IDLE] Name of the cpuidle governor to use. 1025 1026 cpufreq.off=1 [CPU_FREQ] 1027 disable the cpufreq sub-system 1028 1029 cpufreq.default_governor= 1030 [CPU_FREQ] Name of the default cpufreq governor or 1031 policy to use. This governor must be registered in the 1032 kernel before the cpufreq driver probes. 1033 1034 cpu_init_udelay=N 1035 [X86,EARLY] Delay for N microsec between assert and de-assert 1036 of APIC INIT to start processors. This delay occurs 1037 on every CPU online, such as boot, and resume from suspend. 1038 Default: 10000 1039 1040 cpuhp.parallel= 1041 [SMP] Enable/disable parallel bringup of secondary CPUs 1042 Format: <bool> 1043 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise 1044 the parameter has no effect. 1045 1046 crash_kexec_post_notifiers 1047 Only jump to kdump kernel after running the panic 1048 notifiers and dumping kmsg. This option increases 1049 the risks of a kdump failure, since some panic 1050 notifiers can make the crashed kernel more unstable. 1051 In configurations where kdump may not be reliable, 1052 running the panic notifiers could allow collecting 1053 more data on dmesg, like stack traces from other CPUS 1054 or extra data dumped by panic_print. Note that some 1055 configurations enable this option unconditionally, 1056 like Hyper-V, PowerPC (fadump) and AMD SEV-SNP. 1057 1058 crashkernel=size[KMG][@offset[KMG]] 1059 [KNL,EARLY] Using kexec, Linux can switch to a 'crash kernel' 1060 upon panic. This parameter reserves the physical 1061 memory region [offset, offset + size] for that kernel 1062 image. If '@offset' is omitted, then a suitable offset 1063 is selected automatically. 1064 [KNL, X86-64, ARM64, RISCV, LoongArch] Select a region 1065 under 4G first, and fall back to reserve region above 1066 4G when '@offset' hasn't been specified. 1067 See Documentation/admin-guide/kdump/kdump.rst for further details. 1068 1069 crashkernel=range1:size1[,range2:size2,...][@offset] 1070 [KNL] Same as above, but depends on the memory 1071 in the running system. The syntax of range is 1072 start-[end] where start and end are both 1073 a memory unit (amount[KMG]). See also 1074 Documentation/admin-guide/kdump/kdump.rst for an example. 1075 1076 crashkernel=size[KMG],high 1077 [KNL, X86-64, ARM64, RISCV, LoongArch] range could be 1078 above 4G. 1079 Allow kernel to allocate physical memory region from top, 1080 so could be above 4G if system have more than 4G ram 1081 installed. Otherwise memory region will be allocated 1082 below 4G, if available. 1083 It will be ignored if crashkernel=X is specified. 1084 crashkernel=size[KMG],low 1085 [KNL, X86-64, ARM64, RISCV, LoongArch] range under 4G. 1086 When crashkernel=X,high is passed, kernel could allocate 1087 physical memory region above 4G, that cause second kernel 1088 crash on system that require some amount of low memory, 1089 e.g. swiotlb requires at least 64M+32K low memory, also 1090 enough extra low memory is needed to make sure DMA buffers 1091 for 32-bit devices won't run out. Kernel would try to allocate 1092 default size of memory below 4G automatically. The default 1093 size is platform dependent. 1094 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB) 1095 --> arm64: 128MiB 1096 --> riscv: 128MiB 1097 --> loongarch: 128MiB 1098 This one lets the user specify own low range under 4G 1099 for second kernel instead. 1100 0: to disable low allocation. 1101 It will be ignored when crashkernel=X,high is not used 1102 or memory reserved is below 4G. 1103 crashkernel=size[KMG],cma 1104 [KNL, X86, ppc] Reserve additional crash kernel memory from 1105 CMA. This reservation is usable by the first system's 1106 userspace memory and kernel movable allocations (memory 1107 balloon, zswap). Pages allocated from this memory range 1108 will not be included in the vmcore so this should not 1109 be used if dumping of userspace memory is intended and 1110 it has to be expected that some movable kernel pages 1111 may be missing from the dump. 1112 1113 A standard crashkernel reservation, as described above, 1114 is still needed to hold the crash kernel and initrd. 1115 1116 This option increases the risk of a kdump failure: DMA 1117 transfers configured by the first kernel may end up 1118 corrupting the second kernel's memory. 1119 1120 This reservation method is intended for systems that 1121 can't afford to sacrifice enough memory for standard 1122 crashkernel reservation and where less reliable and 1123 possibly incomplete kdump is preferable to no kdump at 1124 all. 1125 1126 cryptomgr.notests 1127 [KNL] Disable crypto self-tests 1128 1129 cs89x0_dma= [HW,NET] 1130 Format: <dma> 1131 1132 cs89x0_media= [HW,NET] 1133 Format: { rj45 | aui | bnc } 1134 1135 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU 1136 function call handling. When switched on, 1137 additional debug data is printed to the console 1138 in case a hanging CPU is detected, and that 1139 CPU is pinged again in order to try to resolve 1140 the hang situation. The default value of this 1141 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT 1142 Kconfig option. 1143 1144 dasd= [HW,NET] 1145 See header of drivers/s390/block/dasd_devmap.c. 1146 1147 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port 1148 (one device per port) 1149 Format: <port#>,<type> 1150 See also Documentation/input/devices/joystick-parport.rst 1151 1152 debug [KNL,EARLY] Enable kernel debugging (events log level). 1153 1154 debug_boot_weak_hash 1155 [KNL,EARLY] Enable printing [hashed] pointers early in the 1156 boot sequence. If enabled, we use a weak hash instead 1157 of siphash to hash pointers. Use this option if you are 1158 seeing instances of '(___ptrval___)') and need to see a 1159 value (hashed pointer) instead. Cryptographically 1160 insecure, please do not use on production kernels. 1161 1162 debug_locks_verbose= 1163 [KNL] verbose locking self-tests 1164 Format: <int> 1165 Print debugging info while doing the locking API 1166 self-tests. 1167 Bitmask for the various LOCKTYPE_ tests. Defaults to 0 1168 (no extra messages), setting it to -1 (all bits set) 1169 will print _a_lot_ more information - normally only 1170 useful to lockdep developers. 1171 1172 debug_objects [KNL,EARLY] Enable object debugging 1173 1174 debug_guardpage_minorder= 1175 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this 1176 parameter allows control of the order of pages that will 1177 be intentionally kept free (and hence protected) by the 1178 buddy allocator. Bigger value increase the probability 1179 of catching random memory corruption, but reduce the 1180 amount of memory for normal system use. The maximum 1181 possible value is MAX_PAGE_ORDER/2. Setting this 1182 parameter to 1 or 2 should be enough to identify most 1183 random memory corruption problems caused by bugs in 1184 kernel or driver code when a CPU writes to (or reads 1185 from) a random memory location. Note that there exists 1186 a class of memory corruptions problems caused by buggy 1187 H/W or F/W or by drivers badly programming DMA 1188 (basically when memory is written at bus level and the 1189 CPU MMU is bypassed) which are not detectable by 1190 CONFIG_DEBUG_PAGEALLOC, hence this option will not 1191 help tracking down these problems. 1192 1193 debug_pagealloc= 1194 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this parameter 1195 enables the feature at boot time. By default, it is 1196 disabled and the system will work mostly the same as a 1197 kernel built without CONFIG_DEBUG_PAGEALLOC. 1198 Note: to get most of debug_pagealloc error reports, it's 1199 useful to also enable the page_owner functionality. 1200 on: enable the feature 1201 1202 debugfs= [KNL,EARLY] This parameter enables what is exposed to 1203 userspace and debugfs internal clients. 1204 Format: { on, off } 1205 on: All functions are enabled. 1206 off: Filesystem is not registered and clients 1207 get a -EPERM as result when trying to register files 1208 or directories within debugfs. 1209 This is equivalent of the runtime functionality if 1210 debugfs was not enabled in the kernel at all. 1211 Default value is set in build-time with a kernel configuration. 1212 1213 debugpat [X86] Enable PAT debugging 1214 1215 default_hugepagesz= 1216 [HW] The size of the default HugeTLB page. This is 1217 the size represented by the legacy /proc/ hugepages 1218 APIs. In addition, this is the default hugetlb size 1219 used for shmget(), mmap() and mounting hugetlbfs 1220 filesystems. If not specified, defaults to the 1221 architecture's default huge page size. Huge page 1222 sizes are architecture dependent. See also 1223 Documentation/admin-guide/mm/hugetlbpage.rst. 1224 Format: size[KMG] 1225 1226 deferred_probe_timeout= 1227 [KNL] Debugging option to set a timeout in seconds for 1228 deferred probe to give up waiting on dependencies to 1229 probe. Only specific dependencies (subsystems or 1230 drivers) that have opted in will be ignored. A timeout 1231 of 0 will timeout at the end of initcalls; a negative 1232 value is treated as an infinite timeout value. If the 1233 timeout hasn't expired, it'll be restarted by each 1234 successful driver registration. This option will also 1235 dump out devices still on the deferred probe list after 1236 retrying. 1237 1238 delayacct [KNL] Enable per-task delay accounting 1239 1240 dell_smm_hwmon.ignore_dmi= 1241 [HW] Continue probing hardware even if DMI data 1242 indicates that the driver is running on unsupported 1243 hardware. 1244 1245 dell_smm_hwmon.force= 1246 [HW] Activate driver even if SMM BIOS signature does 1247 not match list of supported models and enable otherwise 1248 blacklisted features. 1249 1250 dell_smm_hwmon.power_status= 1251 [HW] Report power status in /proc/i8k 1252 (disabled by default). 1253 1254 dell_smm_hwmon.restricted= 1255 [HW] Allow controlling fans only if SYS_ADMIN 1256 capability is set. 1257 1258 dell_smm_hwmon.fan_mult= 1259 [HW] Factor to multiply fan speed with. 1260 1261 dell_smm_hwmon.fan_max= 1262 [HW] Maximum configurable fan speed. 1263 1264 dfltcc= [HW,S390] 1265 Format: { on | off | def_only | inf_only | always } 1266 on: s390 zlib hardware support for compression on 1267 level 1 and decompression (default) 1268 off: No s390 zlib hardware support 1269 def_only: s390 zlib hardware support for deflate 1270 only (compression on level 1) 1271 inf_only: s390 zlib hardware support for inflate 1272 only (decompression) 1273 always: Same as 'on' but ignores the selected compression 1274 level always using hardware support (used for debugging) 1275 1276 dhash_entries= [KNL] 1277 Set number of hash buckets for dentry cache. 1278 1279 disable_1tb_segments [PPC,EARLY] 1280 Disables the use of 1TB hash page table segments. This 1281 causes the kernel to fall back to 256MB segments which 1282 can be useful when debugging issues that require an SLB 1283 miss to occur. 1284 1285 disable= [IPV6] 1286 See Documentation/networking/ipv6.rst. 1287 1288 disable_radix [PPC,EARLY] 1289 Disable RADIX MMU mode on POWER9 1290 1291 disable_tlbie [PPC] 1292 Disable TLBIE instruction. Currently does not work 1293 with KVM, with HASH MMU, or with coherent accelerators. 1294 1295 disable_ddw [PPC/PSERIES,EARLY] 1296 Disable Dynamic DMA Window support. Use this 1297 to workaround buggy firmware. 1298 1299 disable_ipv6= [IPV6] 1300 See Documentation/networking/ipv6.rst. 1301 1302 disable_mtrr_cleanup [X86,EARLY] 1303 The kernel tries to adjust MTRR layout from continuous 1304 to discrete, to make X server driver able to add WB 1305 entry later. This parameter disables that. 1306 1307 disable_mtrr_trim [X86, Intel and AMD only,EARLY] 1308 By default the kernel will trim any uncacheable 1309 memory out of your available memory pool based on 1310 MTRR settings. This parameter disables that behavior, 1311 possibly causing your machine to run very slowly. 1312 1313 disable_timer_pin_1 [X86,EARLY] 1314 Disable PIN 1 of APIC timer 1315 Can be useful to work around chipset bugs. 1316 1317 dis_ucode_ldr [X86] Disable the microcode loader. 1318 1319 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support, 1320 this option disables the debugging code at boot. 1321 1322 dma_debug_entries=<number> 1323 This option allows to tune the number of preallocated 1324 entries for DMA-API debugging code. One entry is 1325 required per DMA-API allocation. Use this if the 1326 DMA-API debugging code disables itself because the 1327 architectural default is too low. 1328 1329 dma_debug_driver=<driver_name> 1330 With this option the DMA-API debugging driver 1331 filter feature can be enabled at boot time. Just 1332 pass the driver to filter for as the parameter. 1333 The filter can be disabled or changed to another 1334 driver later using sysfs. 1335 1336 reg_file_data_sampling= 1337 [X86] Controls mitigation for Register File Data 1338 Sampling (RFDS) vulnerability. RFDS is a CPU 1339 vulnerability which may allow userspace to infer 1340 kernel data values previously stored in floating point 1341 registers, vector registers, or integer registers. 1342 RFDS only affects Intel Atom processors. 1343 1344 on: Turns ON the mitigation. 1345 off: Turns OFF the mitigation. 1346 1347 This parameter overrides the compile time default set 1348 by CONFIG_MITIGATION_RFDS. Mitigation cannot be 1349 disabled when other VERW based mitigations (like MDS) 1350 are enabled. In order to disable RFDS mitigation all 1351 VERW based mitigations need to be disabled. 1352 1353 For details see: 1354 Documentation/admin-guide/hw-vuln/reg-file-data-sampling.rst 1355 1356 dm_verity.keyring_unsealed= 1357 [KNL] When set to 1, leave the dm-verity keyring 1358 unsealed after initialization so userspace can 1359 provision keys. Once the keyring is restricted 1360 it becomes active and is searched during signature 1361 verification. 1362 1363 driver_async_probe= [KNL] 1364 List of driver names to be probed asynchronously. * 1365 matches with all driver names. If * is specified, the 1366 rest of the listed driver names are those that will NOT 1367 match the *. 1368 Format: <driver_name1>,<driver_name2>... 1369 1370 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>] 1371 Broken monitors, graphic adapters, KVMs and EDIDless 1372 panels may send no or incorrect EDID data sets. 1373 This parameter allows to specify an EDID data sets 1374 in the /lib/firmware directory that are used instead. 1375 An EDID data set will only be used for a particular 1376 connector, if its name and a colon are prepended to 1377 the EDID name. Each connector may use a unique EDID 1378 data set by separating the files with a comma. An EDID 1379 data set with no connector name will be used for 1380 any connectors not explicitly specified. 1381 1382 dscc4.setup= [NET] 1383 1384 dt_cpu_ftrs= [PPC,EARLY] 1385 Format: {"off" | "known"} 1386 Control how the dt_cpu_ftrs device-tree binding is 1387 used for CPU feature discovery and setup (if it 1388 exists). 1389 off: Do not use it, fall back to legacy cpu table. 1390 known: Do not pass through unknown features to guests 1391 or userspace, only those that the kernel is aware of. 1392 1393 dump_apple_properties [X86] 1394 Dump name and content of EFI device properties on 1395 x86 Macs. Useful for driver authors to determine 1396 what data is available or for reverse-engineering. 1397 1398 dyndbg[="val"] [KNL,DYNAMIC_DEBUG] 1399 <module>.dyndbg[="val"] 1400 Enable debug messages at boot time. See 1401 Documentation/admin-guide/dynamic-debug-howto.rst 1402 for details. 1403 1404 early_ioremap_debug [KNL,EARLY] 1405 Enable debug messages in early_ioremap support. This 1406 is useful for tracking down temporary early mappings 1407 which are not unmapped. 1408 1409 earlycon= [KNL,EARLY] Output early console device and options. 1410 1411 When used with no options, the early console is 1412 determined by stdout-path property in device tree's 1413 chosen node or the ACPI SPCR table if supported by 1414 the platform. 1415 1416 cdns,<addr>[,options] 1417 Start an early, polled-mode console on a Cadence 1418 (xuartps) serial port at the specified address. Only 1419 supported option is baud rate. If baud rate is not 1420 specified, the serial port must already be setup and 1421 configured. 1422 1423 uart[8250],io,<addr>[,options[,uartclk]] 1424 uart[8250],mmio,<addr>[,options[,uartclk]] 1425 uart[8250],mmio32,<addr>[,options[,uartclk]] 1426 uart[8250],mmio32be,<addr>[,options[,uartclk]] 1427 uart[8250],0x<addr>[,options] 1428 Start an early, polled-mode console on the 8250/16550 1429 UART at the specified I/O port or MMIO address. 1430 MMIO inter-register address stride is either 8-bit 1431 (mmio) or 32-bit (mmio32 or mmio32be). 1432 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed 1433 to be equivalent to 'mmio'. 'options' are specified 1434 in the same format described for "console=ttyS<n>"; if 1435 unspecified, the h/w is not initialized. 'uartclk' is 1436 the uart clock frequency; if unspecified, it is set 1437 to 'BASE_BAUD' * 16. 1438 1439 pl011,<addr> 1440 pl011,mmio32,<addr> 1441 Start an early, polled-mode console on a pl011 serial 1442 port at the specified address. The pl011 serial port 1443 must already be setup and configured. Options are not 1444 yet supported. If 'mmio32' is specified, then only 1445 the driver will use only 32-bit accessors to read/write 1446 the device registers. 1447 1448 liteuart,<addr> 1449 Start an early console on a litex serial port at the 1450 specified address. The serial port must already be 1451 setup and configured. Options are not yet supported. 1452 1453 meson,<addr> 1454 Start an early, polled-mode console on a meson serial 1455 port at the specified address. The serial port must 1456 already be setup and configured. Options are not yet 1457 supported. 1458 1459 msm_serial,<addr> 1460 Start an early, polled-mode console on an msm serial 1461 port at the specified address. The serial port 1462 must already be setup and configured. Options are not 1463 yet supported. 1464 1465 msm_serial_dm,<addr> 1466 Start an early, polled-mode console on an msm serial 1467 dm port at the specified address. The serial port 1468 must already be setup and configured. Options are not 1469 yet supported. 1470 1471 owl,<addr> 1472 Start an early, polled-mode console on a serial port 1473 of an Actions Semi SoC, such as S500 or S900, at the 1474 specified address. The serial port must already be 1475 setup and configured. Options are not yet supported. 1476 1477 rda,<addr> 1478 Start an early, polled-mode console on a serial port 1479 of an RDA Micro SoC, such as RDA8810PL, at the 1480 specified address. The serial port must already be 1481 setup and configured. Options are not yet supported. 1482 1483 sbi 1484 Use RISC-V SBI (Supervisor Binary Interface) for early 1485 console. 1486 1487 smh Use ARM semihosting calls for early console. 1488 1489 s3c2410,<addr> 1490 s3c2412,<addr> 1491 s3c2440,<addr> 1492 s3c6400,<addr> 1493 s5pv210,<addr> 1494 exynos4210,<addr> 1495 Use early console provided by serial driver available 1496 on Samsung SoCs, requires selecting proper type and 1497 a correct base address of the selected UART port. The 1498 serial port must already be setup and configured. 1499 Options are not yet supported. 1500 1501 lantiq,<addr> 1502 Start an early, polled-mode console on a lantiq serial 1503 (lqasc) port at the specified address. The serial port 1504 must already be setup and configured. Options are not 1505 yet supported. 1506 1507 lpuart,<addr> 1508 lpuart32,<addr> 1509 Use early console provided by Freescale LP UART driver 1510 found on Freescale Vybrid and QorIQ LS1021A processors. 1511 A valid base address must be provided, and the serial 1512 port must already be setup and configured. 1513 1514 ec_imx21,<addr> 1515 ec_imx6q,<addr> 1516 Start an early, polled-mode, output-only console on the 1517 Freescale i.MX UART at the specified address. The UART 1518 must already be setup and configured. 1519 1520 ar3700_uart,<addr> 1521 Start an early, polled-mode console on the 1522 Armada 3700 serial port at the specified 1523 address. The serial port must already be setup 1524 and configured. Options are not yet supported. 1525 1526 qcom_geni,<addr> 1527 Start an early, polled-mode console on a Qualcomm 1528 Generic Interface (GENI) based serial port at the 1529 specified address. The serial port must already be 1530 setup and configured. Options are not yet supported. 1531 1532 efifb,[options] 1533 Start an early, unaccelerated console on the EFI 1534 memory mapped framebuffer (if available). On cache 1535 coherent non-x86 systems that use system memory for 1536 the framebuffer, pass the 'ram' option so that it is 1537 mapped with the correct attributes. 1538 1539 linflex,<addr> 1540 Use early console provided by Freescale LINFlexD UART 1541 serial driver for NXP S32V234 SoCs. A valid base 1542 address must be provided, and the serial port must 1543 already be setup and configured. 1544 1545 earlyprintk= [X86,SH,ARM,M68k,S390,UM,EARLY] 1546 earlyprintk=vga 1547 earlyprintk=sclp 1548 earlyprintk=xen 1549 earlyprintk=serial[,ttySn[,baudrate]] 1550 earlyprintk=serial[,0x...[,baudrate]] 1551 earlyprintk=ttySn[,baudrate] 1552 earlyprintk=dbgp[debugController#] 1553 earlyprintk=mmio32,membase[,{nocfg|baudrate}] 1554 earlyprintk=pciserial[,force],bus:device.function[,{nocfg|baudrate}] 1555 earlyprintk=xdbc[xhciController#] 1556 earlyprintk=bios 1557 1558 earlyprintk is useful when the kernel crashes before 1559 the normal console is initialized. It is not enabled by 1560 default because it has some cosmetic problems. 1561 1562 Use "nocfg" to skip UART configuration, assume 1563 BIOS/firmware has configured UART correctly. 1564 1565 Append ",keep" to not disable it when the real console 1566 takes over. 1567 1568 Only one of vga, serial, or usb debug port can 1569 be used at a time. 1570 1571 Currently only ttyS0 and ttyS1 may be specified by 1572 name. Other I/O ports may be explicitly specified 1573 on some architectures (x86 and arm at least) by 1574 replacing ttySn with an I/O port address, like this: 1575 earlyprintk=serial,0x1008,115200 1576 You can find the port for a given device in 1577 /proc/tty/driver/serial: 1578 2: uart:ST16650V2 port:00001008 irq:18 ... 1579 1580 Interaction with the standard serial driver is not 1581 very good. 1582 1583 The VGA output is eventually overwritten by 1584 the real console. 1585 1586 The xen option can only be used in Xen domains. 1587 1588 The sclp output can only be used on s390. 1589 1590 The bios output can only be used on SuperH. 1591 1592 The optional "force" to "pciserial" enables use of a 1593 PCI device even when its classcode is not of the 1594 UART class. 1595 1596 edac_report= [HW,EDAC] Control how to report EDAC event 1597 Format: {"on" | "off" | "force"} 1598 on: enable EDAC to report H/W event. May be overridden 1599 by other higher priority error reporting module. 1600 off: disable H/W event reporting through EDAC. 1601 force: enforce the use of EDAC to report H/W event. 1602 default: on. 1603 1604 edd= [EDD] 1605 Format: {"off" | "on" | "skip[mbr]"} 1606 1607 efi= [EFI,EARLY] 1608 Format: { "debug", "disable_early_pci_dma", 1609 "nochunk", "noruntime", "nosoftreserve", 1610 "novamap", "no_disable_early_pci_dma" } 1611 debug: enable misc debug output. 1612 disable_early_pci_dma: disable the busmaster bit on all 1613 PCI bridges while in the EFI boot stub. 1614 nochunk: disable reading files in "chunks" in the EFI 1615 boot stub, as chunking can cause problems with some 1616 firmware implementations. 1617 noruntime : disable EFI runtime services support 1618 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose) 1619 attribute may cause the kernel to reserve the 1620 memory range for a memory mapping driver to 1621 claim. Specify efi=nosoftreserve to disable this 1622 reservation and treat the memory by its base type 1623 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM"). 1624 novamap: do not call SetVirtualAddressMap(). 1625 no_disable_early_pci_dma: Leave the busmaster bit set 1626 on all PCI bridges while in the EFI boot stub 1627 1628 efi_no_storage_paranoia [EFI,X86,EARLY] 1629 Using this parameter you can use more than 50% of 1630 your efi variable storage. Use this parameter only if 1631 you are really sure that your UEFI does sane gc and 1632 fulfills the spec otherwise your board may brick. 1633 1634 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT 1635 that is to be dynamically loaded by Linux. If there are 1636 multiple variables with the same name but with different 1637 vendor GUIDs, all of them will be loaded. See 1638 Documentation/admin-guide/acpi/ssdt-overlays.rst for details. 1639 1640 1641 eisa_irq_edge= [PARISC,HW] 1642 See header of drivers/parisc/eisa.c. 1643 1644 ekgdboc= [X86,KGDB,EARLY] Allow early kernel console debugging 1645 Format: ekgdboc=kbd 1646 1647 This is designed to be used in conjunction with 1648 the boot argument: earlyprintk=vga 1649 1650 This parameter works in place of the kgdboc parameter 1651 but can only be used if the backing tty is available 1652 very early in the boot process. For early debugging 1653 via a serial port see kgdboc_earlycon instead. 1654 1655 elfcorehdr=[size[KMG]@]offset[KMG] [PPC,SH,X86,S390,EARLY] 1656 Specifies physical address of start of kernel core 1657 image elf header and optionally the size. Generally 1658 kexec loader will pass this option to capture kernel. 1659 See Documentation/admin-guide/kdump/kdump.rst for details. 1660 1661 enable_mtrr_cleanup [X86,EARLY] 1662 The kernel tries to adjust MTRR layout from continuous 1663 to discrete, to make X server driver able to add WB 1664 entry later. This parameter enables that. 1665 1666 enable_timer_pin_1 [X86] 1667 Enable PIN 1 of APIC timer 1668 Can be useful to work around chipset bugs 1669 (in particular on some ATI chipsets). 1670 The kernel tries to set a reasonable default. 1671 1672 enforcing= [SELINUX] Set initial enforcing status. 1673 Format: {"0" | "1"} 1674 See security/selinux/Kconfig help text. 1675 0 -- permissive (log only, no denials). 1676 1 -- enforcing (deny and log). 1677 Default value is 0. 1678 Value can be changed at runtime via 1679 /sys/fs/selinux/enforce. 1680 1681 erst_disable [ACPI] 1682 Disable Error Record Serialization Table (ERST) 1683 support. 1684 1685 ether= [HW,NET] Ethernet cards parameters 1686 This option is obsoleted by the "netdev=" option, which 1687 has equivalent usage. See its documentation for details. 1688 1689 evm= [EVM] 1690 Format: { "fix" } 1691 Permit 'security.evm' to be updated regardless of 1692 current integrity status. 1693 1694 early_page_ext [KNL,EARLY] Enforces page_ext initialization to earlier 1695 stages so cover more early boot allocations. 1696 Please note that as side effect some optimizations 1697 might be disabled to achieve that (e.g. parallelized 1698 memory initialization is disabled) so the boot process 1699 might take longer, especially on systems with a lot of 1700 memory. Available with CONFIG_PAGE_EXTENSION=y. 1701 1702 failslab= 1703 fail_usercopy= 1704 fail_page_alloc= 1705 fail_skb_realloc= 1706 fail_make_request=[KNL] 1707 General fault injection mechanism. 1708 Format: <interval>,<probability>,<space>,<times> 1709 See also Documentation/fault-injection/. 1710 1711 fb_tunnels= [NET] 1712 Format: { initns | none } 1713 See Documentation/admin-guide/sysctl/net.rst for 1714 fb_tunnels_only_for_init_ns 1715 1716 floppy= [HW] 1717 See Documentation/admin-guide/blockdev/floppy.rst. 1718 1719 forcepae [X86-32] 1720 Forcefully enable Physical Address Extension (PAE). 1721 Many Pentium M systems disable PAE but may have a 1722 functionally usable PAE implementation. 1723 Warning: use of this parameter will taint the kernel 1724 and may cause unknown problems. 1725 1726 fred= [X86-64] 1727 Enable/disable Flexible Return and Event Delivery. 1728 Format: { on | off } 1729 on: enable FRED when it's present, the default setting. 1730 off: disable FRED. 1731 1732 ftrace=[tracer] 1733 [FTRACE] will set and start the specified tracer 1734 as early as possible in order to facilitate early 1735 boot debugging. 1736 1737 ftrace_boot_snapshot 1738 [FTRACE] On boot up, a snapshot will be taken of the 1739 ftrace ring buffer that can be read at: 1740 /sys/kernel/tracing/snapshot. 1741 This is useful if you need tracing information from kernel 1742 boot up that is likely to be overridden by user space 1743 start up functionality. 1744 1745 Optionally, the snapshot can also be defined for a tracing 1746 instance that was created by the trace_instance= command 1747 line parameter. 1748 1749 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo 1750 1751 The above will cause the "foo" tracing instance to trigger 1752 a snapshot at the end of boot up. 1753 1754 ftrace_dump_on_oops[=2(orig_cpu) | =<instance>][,<instance> | 1755 ,<instance>=2(orig_cpu)] 1756 [FTRACE] will dump the trace buffers on oops. 1757 If no parameter is passed, ftrace will dump global 1758 buffers of all CPUs, if you pass 2 or orig_cpu, it 1759 will dump only the buffer of the CPU that triggered 1760 the oops, or the specific instance will be dumped if 1761 its name is passed. Multiple instance dump is also 1762 supported, and instances are separated by commas. Each 1763 instance supports only dump on CPU that triggered the 1764 oops by passing 2 or orig_cpu to it. 1765 1766 ftrace_dump_on_oops=foo=orig_cpu 1767 1768 The above will dump only the buffer of "foo" instance 1769 on CPU that triggered the oops. 1770 1771 ftrace_dump_on_oops,foo,bar=orig_cpu 1772 1773 The above will dump global buffer on all CPUs, the 1774 buffer of "foo" instance on all CPUs and the buffer 1775 of "bar" instance on CPU that triggered the oops. 1776 1777 ftrace_filter=[function-list] 1778 [FTRACE] Limit the functions traced by the function 1779 tracer at boot up. function-list is a comma-separated 1780 list of functions. This list can be changed at run 1781 time by the set_ftrace_filter file in the debugfs 1782 tracing directory. 1783 1784 ftrace_notrace=[function-list] 1785 [FTRACE] Do not trace the functions specified in 1786 function-list. This list can be changed at run time 1787 by the set_ftrace_notrace file in the debugfs 1788 tracing directory. 1789 1790 ftrace_graph_filter=[function-list] 1791 [FTRACE] Limit the top level callers functions traced 1792 by the function graph tracer at boot up. 1793 function-list is a comma-separated list of functions 1794 that can be changed at run time by the 1795 set_graph_function file in the debugfs tracing directory. 1796 1797 ftrace_graph_notrace=[function-list] 1798 [FTRACE] Do not trace from the functions specified in 1799 function-list. This list is a comma-separated list of 1800 functions that can be changed at run time by the 1801 set_graph_notrace file in the debugfs tracing directory. 1802 1803 ftrace_graph_max_depth=<uint> 1804 [FTRACE] Used with the function graph tracer. This is 1805 the max depth it will trace into a function. This value 1806 can be changed at run time by the max_graph_depth file 1807 in the tracefs tracing directory. default: 0 (no limit) 1808 1809 fw_devlink= [KNL,EARLY] Create device links between consumer and supplier 1810 devices by scanning the firmware to infer the 1811 consumer/supplier relationships. This feature is 1812 especially useful when drivers are loaded as modules as 1813 it ensures proper ordering of tasks like device probing 1814 (suppliers first, then consumers), supplier boot state 1815 clean up (only after all consumers have probed), 1816 suspend/resume & runtime PM (consumers first, then 1817 suppliers). 1818 Format: { off | permissive | on | rpm } 1819 off -- Don't create device links from firmware info. 1820 permissive -- Create device links from firmware info 1821 but use it only for ordering boot state clean 1822 up (sync_state() calls). 1823 on -- Create device links from firmware info and use it 1824 to enforce probe and suspend/resume ordering. 1825 rpm -- Like "on", but also use to order runtime PM. 1826 1827 fw_devlink.strict=<bool> 1828 [KNL,EARLY] Treat all inferred dependencies as mandatory 1829 dependencies. This only applies for fw_devlink=on|rpm. 1830 Format: <bool> 1831 1832 fw_devlink.sync_state = 1833 [KNL,EARLY] When all devices that could probe have finished 1834 probing, this parameter controls what to do with 1835 devices that haven't yet received their sync_state() 1836 calls. 1837 Format: { strict | timeout } 1838 strict -- Default. Continue waiting on consumers to 1839 probe successfully. 1840 timeout -- Give up waiting on consumers and call 1841 sync_state() on any devices that haven't yet 1842 received their sync_state() calls after 1843 deferred_probe_timeout has expired or by 1844 late_initcall() if !CONFIG_MODULES. 1845 1846 gamecon.map[2|3]= 1847 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad 1848 support via parallel port (up to 5 devices per port) 1849 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5> 1850 See also Documentation/input/devices/joystick-parport.rst 1851 1852 gamma= [HW,DRM] 1853 1854 gart_fix_e820= [X86-64,EARLY] disable the fix e820 for K8 GART 1855 Format: off | on 1856 default: on 1857 1858 gather_data_sampling= 1859 [X86,INTEL,EARLY] Control the Gather Data Sampling (GDS) 1860 mitigation. 1861 1862 Gather Data Sampling is a hardware vulnerability which 1863 allows unprivileged speculative access to data which was 1864 previously stored in vector registers. 1865 1866 This issue is mitigated by default in updated microcode. 1867 The mitigation may have a performance impact but can be 1868 disabled. On systems without the microcode mitigation 1869 disabling AVX serves as a mitigation. 1870 1871 force: Disable AVX to mitigate systems without 1872 microcode mitigation. No effect if the microcode 1873 mitigation is present. Known to cause crashes in 1874 userspace with buggy AVX enumeration. 1875 1876 off: Disable GDS mitigation. 1877 1878 gbpages [X86] Use GB pages for kernel direct mappings. 1879 1880 gcov_persist= [GCOV] When non-zero (default), profiling data for 1881 kernel modules is saved and remains accessible via 1882 debugfs, even when the module is unloaded/reloaded. 1883 When zero, profiling data is discarded and associated 1884 debugfs files are removed at module unload time. 1885 1886 goldfish [X86] Enable the goldfish android emulator platform. 1887 Don't use this when you are not running on the 1888 android emulator 1889 1890 gpio-mockup.gpio_mockup_ranges 1891 [HW] Sets the ranges of gpiochip of for this device. 1892 Format: <start1>,<end1>,<start2>,<end2>... 1893 gpio-mockup.gpio_mockup_named_lines 1894 [HW] Let the driver know GPIO lines should be named. 1895 1896 gpt [EFI] Forces disk with valid GPT signature but 1897 invalid Protective MBR to be treated as GPT. If the 1898 primary GPT is corrupted, it enables the backup/alternate 1899 GPT to be used instead. 1900 1901 grcan.enable0= [HW] Configuration of physical interface 0. Determines 1902 the "Enable 0" bit of the configuration register. 1903 Format: 0 | 1 1904 Default: 0 1905 grcan.enable1= [HW] Configuration of physical interface 1. Determines 1906 the "Enable 0" bit of the configuration register. 1907 Format: 0 | 1 1908 Default: 0 1909 grcan.select= [HW] Select which physical interface to use. 1910 Format: 0 | 1 1911 Default: 0 1912 grcan.txsize= [HW] Sets the size of the tx buffer. 1913 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0. 1914 Default: 1024 1915 grcan.rxsize= [HW] Sets the size of the rx buffer. 1916 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0. 1917 Default: 1024 1918 1919 hardened_usercopy= 1920 [KNL] Under CONFIG_HARDENED_USERCOPY, whether 1921 hardening is enabled for this boot. Hardened 1922 usercopy checking is used to protect the kernel 1923 from reading or writing beyond known memory 1924 allocation boundaries as a proactive defense 1925 against bounds-checking flaws in the kernel's 1926 copy_to_user()/copy_from_user() interface. 1927 The default is determined by 1928 CONFIG_HARDENED_USERCOPY_DEFAULT_ON. 1929 on Perform hardened usercopy checks. 1930 off Disable hardened usercopy checks. 1931 1932 hardlockup_all_cpu_backtrace= 1933 [KNL] Should the hard-lockup detector generate 1934 backtraces on all cpus. 1935 Format: 0 | 1 1936 1937 hash_pointers= 1938 [KNL,EARLY] 1939 By default, when pointers are printed to the console 1940 or buffers via the %p format string, that pointer is 1941 "hashed", i.e. obscured by hashing the pointer value. 1942 This is a security feature that hides actual kernel 1943 addresses from unprivileged users, but it also makes 1944 debugging the kernel more difficult since unequal 1945 pointers can no longer be compared. The choices are: 1946 Format: { auto | always | never } 1947 Default: auto 1948 1949 auto - Hash pointers unless slab_debug is enabled. 1950 always - Always hash pointers (even if slab_debug is 1951 enabled). 1952 never - Never hash pointers. This option should only 1953 be specified when debugging the kernel. Do 1954 not use on production kernels. The boot 1955 param "no_hash_pointers" is an alias for 1956 this mode. 1957 1958 For controlling hashing dynamically at runtime, 1959 use the "kernel.kptr_restrict" sysctl instead. 1960 1961 hashdist= [KNL,NUMA] Large hashes allocated during boot 1962 are distributed across NUMA nodes. Defaults on 1963 for 64-bit NUMA, off otherwise. 1964 Format: 0 | 1 (for off | on) 1965 1966 hd= [EIDE] (E)IDE hard drive subsystem geometry 1967 Format: <cyl>,<head>,<sect> 1968 1969 hest_disable [ACPI] 1970 Disable Hardware Error Source Table (HEST) support; 1971 corresponding firmware-first mode error processing 1972 logic will be disabled. 1973 1974 hibernate= [HIBERNATION] 1975 noresume Don't check if there's a hibernation image 1976 present during boot. 1977 nocompress Don't compress/decompress hibernation images. 1978 no Disable hibernation and resume. 1979 protect_image Turn on image protection during restoration 1980 (that will set all pages holding image data 1981 during restoration read-only). 1982 1983 hibernate.compressor= [HIBERNATION] Compression algorithm to be 1984 used with hibernation. 1985 Format: { lzo | lz4 } 1986 Default: lzo 1987 1988 lzo: Select LZO compression algorithm to 1989 compress/decompress hibernation image. 1990 1991 lz4: Select LZ4 compression algorithm to 1992 compress/decompress hibernation image. 1993 1994 hibernate.pm_test_delay= 1995 [HIBERNATION] 1996 Sets the number of seconds to remain in a hibernation test 1997 mode before resuming the system (see 1998 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 1999 is set. Default value is 5. 2000 2001 hibernate_compression_threads= 2002 [HIBERNATION] 2003 Set the number of threads used for compressing or decompressing 2004 hibernation images. 2005 2006 Format: <integer> 2007 Default: 3 2008 Minimum: 1 2009 Example: hibernate_compression_threads=4 2010 2011 highmem=nn[KMG] [KNL,BOOT,EARLY] forces the highmem zone to have an exact 2012 size of <nn>. This works even on boxes that have no 2013 highmem otherwise. This also works to reduce highmem 2014 size on bigger boxes. 2015 2016 highres= [KNL] Enable/disable high resolution timer mode. 2017 Valid parameters: "on", "off" 2018 Default: "on" 2019 2020 hlt [BUGS=ARM,SH] 2021 2022 hostname= [KNL,EARLY] Set the hostname (aka UTS nodename). 2023 Format: <string> 2024 This allows setting the system's hostname during early 2025 startup. This sets the name returned by gethostname. 2026 Using this parameter to set the hostname makes it 2027 possible to ensure the hostname is correctly set before 2028 any userspace processes run, avoiding the possibility 2029 that a process may call gethostname before the hostname 2030 has been explicitly set, resulting in the calling 2031 process getting an incorrect result. The string must 2032 not exceed the maximum allowed hostname length (usually 2033 64 characters) and will be truncated otherwise. 2034 2035 hpet= [X86-32,HPET] option to control HPET usage 2036 Format: { enable (default) | disable | force | 2037 verbose } 2038 disable: disable HPET and use PIT instead 2039 force: allow force enabled of undocumented chips (ICH4, 2040 VIA, nVidia) 2041 verbose: show contents of HPET registers during setup 2042 2043 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET 2044 registers. Default set by CONFIG_HPET_MMAP_DEFAULT. 2045 2046 hugepages= [HW,EARLY] Number of HugeTLB pages to allocate at boot. 2047 If this follows hugepagesz (below), it specifies 2048 the number of pages of hugepagesz to be allocated. 2049 If this is the first HugeTLB parameter on the command 2050 line, it specifies the number of pages to allocate for 2051 the default huge page size. If using node format, the 2052 number of pages to allocate per-node can be specified. 2053 See also Documentation/admin-guide/mm/hugetlbpage.rst. 2054 Format: <integer> or (node format) 2055 <node>:<integer>[,<node>:<integer>] 2056 2057 hugepagesz= 2058 [HW,EARLY] The size of the HugeTLB pages. This is 2059 used in conjunction with hugepages (above) to 2060 allocate huge pages of a specific size at boot. The 2061 pair hugepagesz=X hugepages=Y can be specified once 2062 for each supported huge page size. Huge page sizes 2063 are architecture dependent. See also 2064 Documentation/admin-guide/mm/hugetlbpage.rst. 2065 Format: size[KMG] 2066 2067 hugepage_alloc_threads= 2068 [HW] The number of threads that should be used to 2069 allocate hugepages during boot. This option can be 2070 used to improve system bootup time when allocating 2071 a large amount of huge pages. 2072 The default value is 25% of the available hardware threads. 2073 2074 Note that this parameter only applies to non-gigantic huge pages. 2075 2076 hugetlb_cma= [HW,CMA,EARLY] The size of a CMA area used for allocation 2077 of gigantic hugepages. Or using node format, the size 2078 of a CMA area per node can be specified. 2079 Format: nn[KMGTPE] or (node format) 2080 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]] 2081 2082 Reserve a CMA area of given size and allocate gigantic 2083 hugepages using the CMA allocator. If enabled, the 2084 boot-time allocation of gigantic hugepages is skipped. 2085 2086 hugetlb_cma_only= 2087 [HW,CMA,EARLY] When allocating new HugeTLB pages, only 2088 try to allocate from the CMA areas. 2089 2090 This option does nothing if hugetlb_cma= is not also 2091 specified. 2092 2093 hugetlb_free_vmemmap= 2094 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP 2095 enabled. 2096 Control if HugeTLB Vmemmap Optimization (HVO) is enabled. 2097 Allows heavy hugetlb users to free up some more 2098 memory (7 * PAGE_SIZE for each 2MB hugetlb page). 2099 Format: { on | off (default) } 2100 2101 on: enable HVO 2102 off: disable HVO 2103 2104 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y, 2105 the default is on. 2106 2107 Note that the vmemmap pages may be allocated from the added 2108 memory block itself when memory_hotplug.memmap_on_memory is 2109 enabled, those vmemmap pages cannot be optimized even if this 2110 feature is enabled. Other vmemmap pages not allocated from 2111 the added memory block itself do not be affected. 2112 2113 hung_task_panic= 2114 [KNL] Number of hung tasks to trigger kernel panic. 2115 Format: <int> 2116 2117 When set to a non-zero value, a kernel panic will be triggered if 2118 the number of detected hung tasks reaches this value. 2119 2120 0: don't panic 2121 1: panic immediately on first hung task 2122 N: panic after N hung tasks are detected in a single scan 2123 2124 The default value is controlled by the 2125 CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time option. The value 2126 selected by this boot parameter can be changed later by the 2127 kernel.hung_task_panic sysctl. 2128 2129 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC) 2130 terminal devices. Valid values: 0..8 2131 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs. 2132 If specified, z/VM IUCV HVC accepts connections 2133 from listed z/VM user IDs only. 2134 2135 hv_nopvspin [X86,HYPER_V,EARLY] 2136 Disables the paravirt spinlock optimizations 2137 which allow the hypervisor to 'idle' the guest 2138 on lock contention. 2139 2140 hw_protection= [HW] 2141 Format: reboot | shutdown 2142 2143 Hardware protection action taken on critical events like 2144 overtemperature or imminent voltage loss. 2145 2146 i2c_bus= [HW] Override the default board specific I2C bus speed 2147 or register an additional I2C bus that is not 2148 registered from board initialization code. 2149 Format: 2150 <bus_id>,<clkrate> 2151 2152 i2c_touchscreen_props= [HW,ACPI,X86] 2153 Set device-properties for ACPI-enumerated I2C-attached 2154 touchscreen, to e.g. fix coordinates of upside-down 2155 mounted touchscreens. If you need this option please 2156 submit a drivers/platform/x86/touchscreen_dmi.c patch 2157 adding a DMI quirk for this. 2158 2159 Format: 2160 <ACPI_HW_ID>:<prop_name>=<val>[:prop_name=val][:...] 2161 Where <val> is one of: 2162 Omit "=<val>" entirely Set a boolean device-property 2163 Unsigned number Set a u32 device-property 2164 Anything else Set a string device-property 2165 2166 Examples (split over multiple lines): 2167 i2c_touchscreen_props=GDIX1001:touchscreen-inverted-x: 2168 touchscreen-inverted-y 2169 2170 i2c_touchscreen_props=MSSL1680:touchscreen-size-x=1920: 2171 touchscreen-size-y=1080:touchscreen-inverted-y: 2172 firmware-name=gsl1680-vendor-model.fw:silead,home-button 2173 2174 i8042.debug [HW] Toggle i8042 debug mode 2175 i8042.unmask_kbd_data 2176 [HW] Enable printing of interrupt data from the KBD port 2177 (disabled by default, and as a pre-condition 2178 requires that i8042.debug=1 be enabled) 2179 i8042.direct [HW] Put keyboard port into non-translated mode 2180 i8042.dumbkbd [HW] Pretend that controller can only read data from 2181 keyboard and cannot control its state 2182 (Don't attempt to blink the leds) 2183 i8042.noaux [HW] Don't check for auxiliary (== mouse) port 2184 i8042.nokbd [HW] Don't check/create keyboard port 2185 i8042.noloop [HW] Disable the AUX Loopback command while probing 2186 for the AUX port 2187 i8042.nomux [HW] Don't check presence of an active multiplexing 2188 controller 2189 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX 2190 controllers 2191 i8042.notimeout [HW] Ignore timeout condition signalled by controller 2192 i8042.reset [HW] Reset the controller during init, cleanup and 2193 suspend-to-ram transitions, only during s2r 2194 transitions, or never reset 2195 Format: { 1 | Y | y | 0 | N | n } 2196 1, Y, y: always reset controller 2197 0, N, n: don't ever reset controller 2198 Default: only on s2r transitions on x86; most other 2199 architectures force reset to be always executed 2200 i8042.unlock [HW] Unlock (ignore) the keylock 2201 i8042.kbdreset [HW] Reset device connected to KBD port 2202 i8042.probe_defer 2203 [HW] Allow deferred probing upon i8042 probe errors 2204 2205 i810= [HW,DRM] 2206 2207 i915.invert_brightness= 2208 [DRM] Invert the sense of the variable that is used to 2209 set the brightness of the panel backlight. Normally a 2210 brightness value of 0 indicates backlight switched off, 2211 and the maximum of the brightness value sets the backlight 2212 to maximum brightness. If this parameter is set to 0 2213 (default) and the machine requires it, or this parameter 2214 is set to 1, a brightness value of 0 sets the backlight 2215 to maximum brightness, and the maximum of the brightness 2216 value switches the backlight off. 2217 -1 -- never invert brightness 2218 0 -- machine default 2219 1 -- force brightness inversion 2220 2221 ia32_emulation= [X86-64] 2222 Format: <bool> 2223 When true, allows loading 32-bit programs and executing 32-bit 2224 syscalls, essentially overriding IA32_EMULATION_DEFAULT_DISABLED at 2225 boot time. When false, unconditionally disables IA32 emulation. 2226 2227 icn= [HW,ISDN] 2228 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]] 2229 2230 2231 idle= [X86,EARLY] 2232 Format: idle=poll, idle=halt, idle=nomwait 2233 2234 idle=poll: Don't do power saving in the idle loop 2235 using HLT, but poll for rescheduling event. This will 2236 make the CPUs eat a lot more power, but may be useful 2237 to get slightly better performance in multiprocessor 2238 benchmarks. It also makes some profiling using 2239 performance counters more accurate. Please note that 2240 on systems with MONITOR/MWAIT support (like Intel 2241 EM64T CPUs) this option has no performance advantage 2242 over the normal idle loop. It may also interact badly 2243 with hyperthreading. 2244 2245 idle=halt: Halt is forced to be used for CPU idle. 2246 In such case C2/C3 won't be used again. 2247 2248 idle=nomwait: Disable mwait for CPU C-states 2249 2250 idxd.sva= [HW] 2251 Format: <bool> 2252 Allow force disabling of Shared Virtual Memory (SVA) 2253 support for the idxd driver. By default it is set to 2254 true (1). 2255 2256 idxd.tc_override= [HW] 2257 Format: <bool> 2258 Allow override of default traffic class configuration 2259 for the device. By default it is set to false (0). 2260 2261 ieee754= [MIPS] Select IEEE Std 754 conformance mode 2262 Format: { strict | legacy | 2008 | relaxed | emulated } 2263 Default: strict 2264 2265 Choose which programs will be accepted for execution 2266 based on the IEEE 754 NaN encoding(s) supported by 2267 the FPU and the NaN encoding requested with the value 2268 of an ELF file header flag individually set by each 2269 binary. Hardware implementations are permitted to 2270 support either or both of the legacy and the 2008 NaN 2271 encoding mode. 2272 2273 Available settings are as follows: 2274 strict accept binaries that request a NaN encoding 2275 supported by the FPU 2276 legacy only accept legacy-NaN binaries, if supported 2277 by the FPU 2278 2008 only accept 2008-NaN binaries, if supported 2279 by the FPU 2280 relaxed accept any binaries regardless of whether 2281 supported by the FPU 2282 emulated accept any binaries but enable FPU emulator 2283 if binary mode is unsupported by the FPU. 2284 2285 The FPU emulator is always able to support both NaN 2286 encodings, so if no FPU hardware is present or it has 2287 been disabled with 'nofpu', then the settings of 2288 'legacy' and '2008' strap the emulator accordingly, 2289 'relaxed' straps the emulator for both legacy-NaN and 2290 2008-NaN, whereas 'strict' enables legacy-NaN only on 2291 legacy processors and both NaN encodings on MIPS32 or 2292 MIPS64 CPUs. 2293 2294 The setting for ABS.fmt/NEG.fmt instruction execution 2295 mode generally follows that for the NaN encoding, 2296 except where unsupported by hardware. 2297 2298 ignore_loglevel [KNL,EARLY] 2299 Ignore loglevel setting - this will print /all/ 2300 kernel messages to the console. Useful for debugging. 2301 We also add it as printk module parameter, so users 2302 could change it dynamically, usually by 2303 /sys/module/printk/parameters/ignore_loglevel. 2304 2305 ignore_rlimit_data 2306 Ignore RLIMIT_DATA setting for data mappings, 2307 print warning at first misuse. Can be changed via 2308 /sys/module/kernel/parameters/ignore_rlimit_data. 2309 2310 ihash_entries= [KNL] 2311 Set number of hash buckets for inode cache. 2312 2313 ima_appraise= [IMA] appraise integrity measurements 2314 Format: { "off" | "enforce" | "fix" | "log" } 2315 default: "enforce" 2316 2317 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead. 2318 The builtin appraise policy appraises all files 2319 owned by uid=0. 2320 2321 ima_canonical_fmt [IMA] 2322 Use the canonical format for the binary runtime 2323 measurements, instead of host native format. 2324 2325 ima_hash= [IMA] 2326 Format: { md5 | sha1 | rmd160 | sha256 | sha384 2327 | sha512 | ... } 2328 default: "sha1" 2329 2330 The list of supported hash algorithms is defined 2331 in crypto/hash_info.h. 2332 2333 ima_policy= [IMA] 2334 The builtin policies to load during IMA setup. 2335 Format: "tcb | appraise_tcb | secure_boot | 2336 fail_securely | critical_data" 2337 2338 The "tcb" policy measures all programs exec'd, files 2339 mmap'd for exec, and all files opened with the read 2340 mode bit set by either the effective uid (euid=0) or 2341 uid=0. 2342 2343 The "appraise_tcb" policy appraises the integrity of 2344 all files owned by root. 2345 2346 The "secure_boot" policy appraises the integrity 2347 of files (eg. kexec kernel image, kernel modules, 2348 firmware, policy, etc) based on file signatures. 2349 2350 The "fail_securely" policy forces file signature 2351 verification failure also on privileged mounted 2352 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE 2353 flag. 2354 2355 The "critical_data" policy measures kernel integrity 2356 critical data. 2357 2358 ima_tcb [IMA] Deprecated. Use ima_policy= instead. 2359 Load a policy which meets the needs of the Trusted 2360 Computing Base. This means IMA will measure all 2361 programs exec'd, files mmap'd for exec, and all files 2362 opened for read by uid=0. 2363 2364 ima_template= [IMA] 2365 Select one of defined IMA measurements template formats. 2366 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" | 2367 "ima-sigv2" } 2368 Default: "ima-ng" 2369 2370 ima_template_fmt= 2371 [IMA] Define a custom template format. 2372 Format: { "field1|...|fieldN" } 2373 2374 ima= [IMA] Enable or disable IMA 2375 Format: { "off" | "on" } 2376 Default: "on" 2377 Note that disabling IMA is limited to kdump kernel. 2378 2379 indirect_target_selection= [X86,Intel] Mitigation control for Indirect 2380 Target Selection(ITS) bug in Intel CPUs. Updated 2381 microcode is also required for a fix in IBPB. 2382 2383 on: Enable mitigation (default). 2384 off: Disable mitigation. 2385 force: Force the ITS bug and deploy default 2386 mitigation. 2387 vmexit: Only deploy mitigation if CPU is affected by 2388 guest/host isolation part of ITS. 2389 stuff: Deploy RSB-fill mitigation when retpoline is 2390 also deployed. Otherwise, deploy the default 2391 mitigation. 2392 2393 For details see: 2394 Documentation/admin-guide/hw-vuln/indirect-target-selection.rst 2395 2396 init= [KNL] 2397 Format: <full_path> 2398 Run specified binary instead of /sbin/init as init 2399 process. 2400 2401 initcall_debug [KNL] Trace initcalls as they are executed. Useful 2402 for working out where the kernel is dying during 2403 startup. 2404 2405 initcall_blacklist= [KNL] Do not execute a comma-separated list of 2406 initcall functions. Useful for debugging built-in 2407 modules and initcalls. 2408 2409 initramfs_async= [KNL] 2410 Format: <bool> 2411 Default: 1 2412 This parameter controls whether the initramfs 2413 image is unpacked asynchronously, concurrently 2414 with devices being probed and 2415 initialized. This should normally just work, 2416 but as a debugging aid, one can get the 2417 historical behaviour of the initramfs 2418 unpacking being completed before device_ and 2419 late_ initcalls. 2420 2421 initrd= [BOOT,EARLY] Specify the location of the initial ramdisk 2422 2423 initrdmem= [KNL,EARLY] Specify a physical address and size from which to 2424 load the initrd. If an initrd is compiled in or 2425 specified in the bootparams, it takes priority over this 2426 setting. 2427 Format: ss[KMG],nn[KMG] 2428 Default is 0, 0 2429 2430 init_on_alloc= [MM,EARLY] Fill newly allocated pages and heap objects with 2431 zeroes. 2432 Format: 0 | 1 2433 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON. 2434 2435 init_on_free= [MM,EARLY] Fill freed pages and heap objects with zeroes. 2436 Format: 0 | 1 2437 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON. 2438 2439 init_pkru= [X86] Specify the default memory protection keys rights 2440 register contents for all processes. 0x55555554 by 2441 default (disallow access to all but pkey 0). Can 2442 override in debugfs after boot. 2443 2444 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver 2445 Format: <irq> 2446 2447 int_pln_enable [X86] Enable power limit notification interrupt 2448 2449 integrity_audit=[IMA] 2450 Format: { "0" | "1" } 2451 0 -- basic integrity auditing messages. (Default) 2452 1 -- additional integrity auditing messages. 2453 2454 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option 2455 on 2456 Enable intel iommu driver. 2457 off 2458 Disable intel iommu driver. 2459 igfx_off [Default Off] 2460 By default, gfx is mapped as normal device. If a gfx 2461 device has a dedicated DMAR unit, the DMAR unit is 2462 bypassed by not enabling DMAR with this option. In 2463 this case, gfx device will use physical address for 2464 DMA. 2465 strict [Default Off] 2466 Deprecated, equivalent to iommu.strict=1. 2467 sp_off [Default Off] 2468 By default, super page will be supported if Intel IOMMU 2469 has the capability. With this option, super page will 2470 not be supported. 2471 sm_on 2472 Enable the Intel IOMMU scalable mode if the hardware 2473 advertises that it has support for the scalable mode 2474 translation. 2475 sm_off 2476 Disallow use of the Intel IOMMU scalable mode. 2477 tboot_noforce [Default Off] 2478 Do not force the Intel IOMMU enabled under tboot. 2479 By default, tboot will force Intel IOMMU on, which 2480 could harm performance of some high-throughput 2481 devices like 40GBit network cards, even if identity 2482 mapping is enabled. 2483 Note that using this option lowers the security 2484 provided by tboot because it makes the system 2485 vulnerable to DMA attacks. 2486 2487 intel_idle.max_cstate= [KNL,HW,ACPI,X86] 2488 0 disables intel_idle and fall back on acpi_idle. 2489 1 to 9 specify maximum depth of C-state. 2490 2491 intel_pstate= [X86,EARLY] 2492 disable 2493 Do not enable intel_pstate as the default 2494 scaling driver for the supported processors 2495 active 2496 Use intel_pstate driver to bypass the scaling 2497 governors layer of cpufreq and provides it own 2498 algorithms for p-state selection. There are two 2499 P-state selection algorithms provided by 2500 intel_pstate in the active mode: powersave and 2501 performance. The way they both operate depends 2502 on whether or not the hardware managed P-states 2503 (HWP) feature has been enabled in the processor 2504 and possibly on the processor model. 2505 passive 2506 Use intel_pstate as a scaling driver, but configure it 2507 to work with generic cpufreq governors (instead of 2508 enabling its internal governor). This mode cannot be 2509 used along with the hardware-managed P-states (HWP) 2510 feature. 2511 force 2512 Enable intel_pstate on systems that prohibit it by default 2513 in favor of acpi-cpufreq. Forcing the intel_pstate driver 2514 instead of acpi-cpufreq may disable platform features, such 2515 as thermal controls and power capping, that rely on ACPI 2516 P-States information being indicated to OSPM and therefore 2517 should be used with caution. This option does not work with 2518 processors that aren't supported by the intel_pstate driver 2519 or on platforms that use pcc-cpufreq instead of acpi-cpufreq. 2520 no_hwp 2521 Do not enable hardware P state control (HWP) 2522 if available. 2523 hwp_only 2524 Only load intel_pstate on systems which support 2525 hardware P state control (HWP) if available. 2526 support_acpi_ppc 2527 Enforce ACPI _PPC performance limits. If the Fixed ACPI 2528 Description Table, specifies preferred power management 2529 profile as "Enterprise Server" or "Performance Server", 2530 then this feature is turned on by default. 2531 per_cpu_perf_limits 2532 Allow per-logical-CPU P-State performance control limits using 2533 cpufreq sysfs interface 2534 no_cas 2535 Do not enable capacity-aware scheduling (CAS) on 2536 hybrid systems 2537 2538 intremap= [X86-64,Intel-IOMMU,EARLY] 2539 on enable Interrupt Remapping (default) 2540 off disable Interrupt Remapping 2541 nosid disable Source ID checking 2542 no_x2apic_optout 2543 BIOS x2APIC opt-out request will be ignored 2544 nopost disable Interrupt Posting 2545 posted_msi 2546 enable MSIs delivered as posted interrupts 2547 2548 iomem= Disable strict checking of access to MMIO memory 2549 strict regions from userspace. 2550 relaxed 2551 2552 iommu= [X86,EARLY] 2553 2554 off 2555 Don't initialize and use any kind of IOMMU. 2556 2557 force 2558 Force the use of the hardware IOMMU even when 2559 it is not actually needed (e.g. because < 3 GB 2560 memory). 2561 2562 noforce 2563 Don't force hardware IOMMU usage when it is not 2564 needed. (default). 2565 2566 biomerge 2567 panic 2568 nopanic 2569 merge 2570 nomerge 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 pt 2578 nopt 2579 nobypass [PPC/POWERNV] 2580 Disable IOMMU bypass, using IOMMU for PCI devices. 2581 2582 AMD Gart HW IOMMU-specific options: 2583 2584 <size> 2585 Set the size of the remapping area in bytes. 2586 2587 allowed 2588 Overwrite iommu off workarounds for specific chipsets 2589 2590 fullflush 2591 Flush IOMMU on each allocation (default). 2592 2593 nofullflush 2594 Don't use IOMMU fullflush. 2595 2596 memaper[=<order>] 2597 Allocate an own aperture over RAM with size 2598 32MB<<order. (default: order=1, i.e. 64MB) 2599 2600 merge 2601 Do scatter-gather (SG) merging. Implies "force" 2602 (experimental). 2603 2604 nomerge 2605 Don't do scatter-gather (SG) merging. 2606 2607 noaperture 2608 Ask the IOMMU not to touch the aperture for AGP. 2609 2610 noagp 2611 Don't initialize the AGP driver and use full aperture. 2612 2613 panic 2614 Always panic when IOMMU overflows. 2615 2616 iommu.forcedac= [ARM64,X86,EARLY] Control IOVA allocation for PCI devices. 2617 Format: { "0" | "1" } 2618 0 - Try to allocate a 32-bit DMA address first, before 2619 falling back to the full range if needed. 2620 1 - Allocate directly from the full usable range, 2621 forcing Dual Address Cycle for PCI cards supporting 2622 greater than 32-bit addressing. 2623 2624 iommu.strict= [ARM64,X86,S390,EARLY] Configure TLB invalidation behaviour 2625 Format: { "0" | "1" } 2626 0 - Lazy mode. 2627 Request that DMA unmap operations use deferred 2628 invalidation of hardware TLBs, for increased 2629 throughput at the cost of reduced device isolation. 2630 Will fall back to strict mode if not supported by 2631 the relevant IOMMU driver. 2632 1 - Strict mode. 2633 DMA unmap operations invalidate IOMMU hardware TLBs 2634 synchronously. 2635 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}. 2636 Note: on x86, strict mode specified via one of the 2637 legacy driver-specific options takes precedence. 2638 2639 iommu.passthrough= 2640 [ARM64,X86,EARLY] Configure DMA to bypass the IOMMU by default. 2641 Format: { "0" | "1" } 2642 0 - Use IOMMU translation for DMA. 2643 1 - Bypass the IOMMU for DMA. 2644 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH. 2645 2646 iommu.debug_pagealloc= 2647 [KNL,EARLY] When CONFIG_IOMMU_DEBUG_PAGEALLOC is set, this 2648 parameter enables the feature at boot time. By default, it 2649 is disabled and the system behaves the same way as a kernel 2650 built without CONFIG_IOMMU_DEBUG_PAGEALLOC. 2651 Format: { "0" | "1" } 2652 0 - Sanitizer disabled. 2653 1 - Sanitizer enabled, expect runtime overhead. 2654 2655 io7= [HW] IO7 for Marvel-based Alpha systems 2656 See comment before marvel_specify_io7 in 2657 arch/alpha/kernel/core_marvel.c. 2658 2659 io_delay= [X86,EARLY] I/O delay method 2660 0x80 2661 Standard port 0x80 based delay 2662 0xed 2663 Alternate port 0xed based delay (needed on some systems) 2664 udelay 2665 Simple two microseconds delay 2666 none 2667 No delay 2668 2669 ip= [IP_PNP] 2670 See Documentation/admin-guide/nfs/nfsroot.rst. 2671 2672 ipcmni_extend [KNL,EARLY] Extend the maximum number of unique System V 2673 IPC identifiers from 32,768 to 16,777,216. 2674 2675 ipe.enforce= [IPE] 2676 Format: <bool> 2677 Determine whether IPE starts in permissive (0) or 2678 enforce (1) mode. The default is enforce. 2679 2680 ipe.success_audit= 2681 [IPE] 2682 Format: <bool> 2683 Start IPE with success auditing enabled, emitting 2684 an audit event when a binary is allowed. The default 2685 is 0. 2686 2687 irqaffinity= [SMP] Set the default irq affinity mask 2688 The argument is a cpu list, as described above. 2689 2690 irqchip.gicv2_force_probe= 2691 [ARM,ARM64,EARLY] 2692 Format: <bool> 2693 Force the kernel to look for the second 4kB page 2694 of a GICv2 controller even if the memory range 2695 exposed by the device tree is too small. 2696 2697 irqchip.gicv3_nolpi= 2698 [ARM,ARM64,EARLY] 2699 Force the kernel to ignore the availability of 2700 LPIs (and by consequence ITSs). Intended for system 2701 that use the kernel as a bootloader, and thus want 2702 to let secondary kernels in charge of setting up 2703 LPIs. 2704 2705 irqchip.gicv3_pseudo_nmi= [ARM64,EARLY] 2706 Enables support for pseudo-NMIs in the kernel. This 2707 requires the kernel to be built with 2708 CONFIG_ARM64_PSEUDO_NMI. 2709 2710 irqchip.riscv_imsic_noipi 2711 [RISC-V,EARLY] 2712 Force the kernel to not use IMSIC software injected MSIs 2713 as IPIs. Intended for system where IMSIC is trap-n-emulated, 2714 and thus want to reduce MMIO traps when triggering IPIs 2715 to multiple harts. 2716 2717 irqfixup [HW] 2718 When an interrupt is not handled search all handlers 2719 for it. Intended to get systems with badly broken 2720 firmware running. 2721 2722 irqhandler.duration_warn_us= [KNL] 2723 Warn if an IRQ handler exceeds the specified duration 2724 threshold in microseconds. Useful for identifying 2725 long-running IRQs in the system. 2726 2727 irqpoll [HW] 2728 When an interrupt is not handled search all handlers 2729 for it. Also check all handlers each timer 2730 interrupt. Intended to get systems with badly broken 2731 firmware running. 2732 2733 isapnp= [ISAPNP] 2734 Format: <RDP>,<reset>,<pci_scan>,<verbosity> 2735 2736 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance. 2737 [Deprecated - use cpusets instead] 2738 Format: [flag-list,]<cpu-list> 2739 2740 Specify one or more CPUs to isolate from disturbances 2741 specified in the flag list (default: domain): 2742 2743 nohz 2744 Disable the tick when a single task runs as well as 2745 disabling other kernel noises like having RCU callbacks 2746 offloaded. This is equivalent to the nohz_full parameter. 2747 2748 A residual 1Hz tick is offloaded to workqueues, which you 2749 need to affine to housekeeping through the global 2750 workqueue's affinity configured via the 2751 /sys/devices/virtual/workqueue/cpumask sysfs file, or 2752 by using the 'domain' flag described below. 2753 2754 NOTE: by default the global workqueue runs on all CPUs, 2755 so to protect individual CPUs the 'cpumask' file has to 2756 be configured manually after bootup. 2757 2758 domain 2759 Isolate from the general SMP balancing and scheduling 2760 algorithms. Note that performing domain isolation this way 2761 is irreversible: it's not possible to bring back a CPU to 2762 the domains once isolated through isolcpus. It's strongly 2763 advised to use cpusets instead to disable scheduler load 2764 balancing through the "cpuset.sched_load_balance" file. 2765 It offers a much more flexible interface where CPUs can 2766 move in and out of an isolated set anytime. 2767 2768 You can move a process onto or off an "isolated" CPU via 2769 the CPU affinity syscalls or cpuset. 2770 <cpu number> begins at 0 and the maximum value is 2771 "number of CPUs in system - 1". 2772 2773 managed_irq 2774 2775 Isolate from being targeted by managed interrupts 2776 which have an interrupt mask containing isolated 2777 CPUs. The affinity of managed interrupts is 2778 handled by the kernel and cannot be changed via 2779 the /proc/irq/* interfaces. 2780 2781 This isolation is best effort and only effective 2782 if the automatically assigned interrupt mask of a 2783 device queue contains isolated and housekeeping 2784 CPUs. If housekeeping CPUs are online then such 2785 interrupts are directed to the housekeeping CPU 2786 so that IO submitted on the housekeeping CPU 2787 cannot disturb the isolated CPU. 2788 2789 If a queue's affinity mask contains only isolated 2790 CPUs then this parameter has no effect on the 2791 interrupt routing decision, though interrupts are 2792 only delivered when tasks running on those 2793 isolated CPUs submit IO. IO submitted on 2794 housekeeping CPUs has no influence on those 2795 queues. 2796 2797 The format of <cpu-list> is described above. 2798 2799 iucv= [HW,NET] 2800 2801 ivrs_ioapic [HW,X86-64] 2802 Provide an override to the IOAPIC-ID<->DEVICE-ID 2803 mapping provided in the IVRS ACPI table. 2804 By default, PCI segment is 0, and can be omitted. 2805 2806 For example, to map IOAPIC-ID decimal 10 to 2807 PCI segment 0x1 and PCI device 00:14.0, 2808 write the parameter as: 2809 ivrs_ioapic=10@0001:00:14.0 2810 2811 Deprecated formats: 2812 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0 2813 write the parameter as: 2814 ivrs_ioapic[10]=00:14.0 2815 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and 2816 PCI device 00:14.0 write the parameter as: 2817 ivrs_ioapic[10]=0001:00:14.0 2818 2819 ivrs_hpet [HW,X86-64] 2820 Provide an override to the HPET-ID<->DEVICE-ID 2821 mapping provided in the IVRS ACPI table. 2822 By default, PCI segment is 0, and can be omitted. 2823 2824 For example, to map HPET-ID decimal 10 to 2825 PCI segment 0x1 and PCI device 00:14.0, 2826 write the parameter as: 2827 ivrs_hpet=10@0001:00:14.0 2828 2829 Deprecated formats: 2830 * To map HPET-ID decimal 0 to PCI device 00:14.0 2831 write the parameter as: 2832 ivrs_hpet[0]=00:14.0 2833 * To map HPET-ID decimal 10 to PCI segment 0x1 and 2834 PCI device 00:14.0 write the parameter as: 2835 ivrs_ioapic[10]=0001:00:14.0 2836 2837 ivrs_acpihid [HW,X86-64] 2838 Provide an override to the ACPI-HID:UID<->DEVICE-ID 2839 mapping provided in the IVRS ACPI table. 2840 By default, PCI segment is 0, and can be omitted. 2841 2842 For example, to map UART-HID:UID AMD0020:0 to 2843 PCI segment 0x1 and PCI device ID 00:14.5, 2844 write the parameter as: 2845 ivrs_acpihid=AMD0020:0@0001:00:14.5 2846 2847 Deprecated formats: 2848 * To map UART-HID:UID AMD0020:0 to PCI segment is 0, 2849 PCI device ID 00:14.5, write the parameter as: 2850 ivrs_acpihid[00:14.5]=AMD0020:0 2851 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and 2852 PCI device ID 00:14.5, write the parameter as: 2853 ivrs_acpihid[0001:00:14.5]=AMD0020:0 2854 2855 js= [HW,JOY] Analog joystick 2856 See Documentation/input/joydev/joystick.rst. 2857 2858 kasan_multi_shot 2859 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print 2860 report on every invalid memory access. Without this 2861 parameter KASAN will print report only for the first 2862 invalid access. 2863 2864 keep_bootcon [KNL,EARLY] 2865 Do not unregister boot console at start. This is only 2866 useful for debugging when something happens in the window 2867 between unregistering the boot console and initializing 2868 the real console. 2869 2870 keepinitrd [HW,ARM] See retain_initrd. 2871 2872 kernelcore= [KNL,X86,PPC,EARLY] 2873 Format: nn[KMGTPE] | nn% | "mirror" 2874 This parameter specifies the amount of memory usable by 2875 the kernel for non-movable allocations. The requested 2876 amount is spread evenly throughout all nodes in the 2877 system as ZONE_NORMAL. The remaining memory is used for 2878 movable memory in its own zone, ZONE_MOVABLE. In the 2879 event, a node is too small to have both ZONE_NORMAL and 2880 ZONE_MOVABLE, kernelcore memory will take priority and 2881 other nodes will have a larger ZONE_MOVABLE. 2882 2883 ZONE_MOVABLE is used for the allocation of pages that 2884 may be reclaimed or moved by the page migration 2885 subsystem. Note that allocations like PTEs-from-HighMem 2886 still use the HighMem zone if it exists, and the Normal 2887 zone if it does not. 2888 2889 It is possible to specify the exact amount of memory in 2890 the form of "nn[KMGTPE]", a percentage of total system 2891 memory in the form of "nn%", or "mirror". If "mirror" 2892 option is specified, mirrored (reliable) memory is used 2893 for non-movable allocations and remaining memory is used 2894 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror" 2895 are exclusive, so you cannot specify multiple forms. 2896 2897 kfence.burst= [MM,KFENCE] The number of additional successive 2898 allocations to be attempted through KFENCE for each 2899 sample interval. 2900 Format: <unsigned integer> 2901 Default: 0 2902 2903 kfence.check_on_panic= 2904 [MM,KFENCE] Whether to check all KFENCE-managed objects' 2905 canaries on panic. 2906 Format: <bool> 2907 Default: false 2908 2909 kfence.deferrable= 2910 [MM,KFENCE] Whether to use a deferrable timer to trigger 2911 allocations. This avoids forcing CPU wake-ups if the 2912 system is idle, at the risk of a less predictable 2913 sample interval. 2914 Format: <bool> 2915 Default: CONFIG_KFENCE_DEFERRABLE 2916 2917 kfence.fault= [MM,KFENCE] Controls the behavior when a KFENCE 2918 error is detected. 2919 report - print the error report and continue (default). 2920 oops - print the error report and oops. 2921 panic - print the error report and panic. 2922 2923 kfence.sample_interval= 2924 [MM,KFENCE] KFENCE's sample interval in milliseconds. 2925 Format: <unsigned integer> 2926 0 - Disable KFENCE. 2927 >0 - Enabled KFENCE with given sample interval. 2928 Default: CONFIG_KFENCE_SAMPLE_INTERVAL 2929 2930 kfence.skip_covered_thresh= 2931 [MM,KFENCE] If pool utilization reaches this threshold 2932 (pool usage%), KFENCE limits currently covered 2933 allocations of the same source from further filling 2934 up the pool. 2935 Format: <unsigned integer> 2936 Default: 75 2937 2938 kgdbdbgp= [KGDB,HW,EARLY] kgdb over EHCI usb debug port. 2939 Format: <Controller#>[,poll interval] 2940 The controller # is the number of the ehci usb debug 2941 port as it is probed via PCI. The poll interval is 2942 optional and is the number seconds in between 2943 each poll cycle to the debug port in case you need 2944 the functionality for interrupting the kernel with 2945 gdb or control-c on the dbgp connection. When 2946 not using this parameter you use sysrq-g to break into 2947 the kernel debugger. 2948 2949 kgdboc= [KGDB,HW] kgdb over consoles. 2950 Requires a tty driver that supports console polling, 2951 or a supported polling keyboard driver (non-usb). 2952 Serial only format: <serial_device>[,baud] 2953 keyboard only format: kbd 2954 keyboard and serial format: kbd,<serial_device>[,baud] 2955 Optional Kernel mode setting: 2956 kms, kbd format: kms,kbd 2957 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud] 2958 2959 kgdboc_earlycon= [KGDB,HW,EARLY] 2960 If the boot console provides the ability to read 2961 characters and can work in polling mode, you can use 2962 this parameter to tell kgdb to use it as a backend 2963 until the normal console is registered. Intended to 2964 be used together with the kgdboc parameter which 2965 specifies the normal console to transition to. 2966 2967 The name of the early console should be specified 2968 as the value of this parameter. Note that the name of 2969 the early console might be different than the tty 2970 name passed to kgdboc. It's OK to leave the value 2971 blank and the first boot console that implements 2972 read() will be picked. 2973 2974 kgdbwait [KGDB,EARLY] Stop kernel execution and enter the 2975 kernel debugger at the earliest opportunity. 2976 2977 kho= [KEXEC,EARLY] 2978 Format: { "0" | "1" | "off" | "on" | "y" | "n" } 2979 Enables or disables Kexec HandOver. 2980 "0" | "off" | "n" - kexec handover is disabled 2981 "1" | "on" | "y" - kexec handover is enabled 2982 2983 kho_scratch= [KEXEC,EARLY] 2984 Format: ll[KMG],mm[KMG],nn[KMG] | nn% 2985 Defines the size of the KHO scratch region. The KHO 2986 scratch regions are physically contiguous memory 2987 ranges that can only be used for non-kernel 2988 allocations. That way, even when memory is heavily 2989 fragmented with handed over memory, the kexeced 2990 kernel will always have enough contiguous ranges to 2991 bootstrap itself. 2992 2993 It is possible to specify the exact amount of 2994 memory in the form of "ll[KMG],mm[KMG],nn[KMG]" 2995 where the first parameter defines the size of a low 2996 memory scratch area, the second parameter defines 2997 the size of a global scratch area and the third 2998 parameter defines the size of additional per-node 2999 scratch areas. The form "nn%" defines scale factor 3000 (in percents) of memory that was used during boot. 3001 3002 kmac= [MIPS] Korina ethernet MAC address. 3003 Configure the RouterBoard 532 series on-chip 3004 Ethernet adapter MAC address. 3005 3006 kmemleak= [KNL,EARLY] Boot-time kmemleak enable/disable 3007 Valid arguments: on, off 3008 Default: on 3009 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y, 3010 the default is off. 3011 3012 kprobe_event=[probe-list] 3013 [FTRACE] Add kprobe events and enable at boot time. 3014 The probe-list is a semicolon delimited list of probe 3015 definitions. Each definition is same as kprobe_events 3016 interface, but the parameters are comma delimited. 3017 For example, to add a kprobe event on vfs_read with 3018 arg1 and arg2, add to the command line; 3019 3020 kprobe_event=p,vfs_read,$arg1,$arg2 3021 3022 See also Documentation/trace/kprobetrace.rst "Kernel 3023 Boot Parameter" section. 3024 3025 kpti= [ARM64,EARLY] Control page table isolation of 3026 user and kernel address spaces. 3027 Default: enabled on cores which need mitigation. 3028 0: force disabled 3029 1: force enabled 3030 3031 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires 3032 CONFIG_KUNIT to be set to be fully enabled. The 3033 default value can be overridden via 3034 KUNIT_DEFAULT_ENABLED. 3035 Default is 1 (enabled) 3036 3037 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs. 3038 Default is 0 (don't ignore, but inject #GP) 3039 3040 kvm.eager_page_split= 3041 [KVM,X86] Controls whether or not KVM will try to 3042 proactively split all huge pages during dirty logging. 3043 Eager page splitting reduces interruptions to vCPU 3044 execution by eliminating the write-protection faults 3045 and MMU lock contention that would otherwise be 3046 required to split huge pages lazily. 3047 3048 VM workloads that rarely perform writes or that write 3049 only to a small region of VM memory may benefit from 3050 disabling eager page splitting to allow huge pages to 3051 still be used for reads. 3052 3053 The behavior of eager page splitting depends on whether 3054 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If 3055 disabled, all huge pages in a memslot will be eagerly 3056 split when dirty logging is enabled on that memslot. If 3057 enabled, eager page splitting will be performed during 3058 the KVM_CLEAR_DIRTY ioctl, and only for the pages being 3059 cleared. 3060 3061 Eager page splitting is only supported when kvm.tdp_mmu=Y. 3062 3063 Default is Y (on). 3064 3065 kvm.enable_pmu=[KVM,X86] 3066 If enabled, KVM will virtualize PMU functionality based 3067 on the virtual CPU model defined by userspace. This 3068 can be overridden on a per-VM basis via 3069 KVM_CAP_PMU_CAPABILITY. 3070 3071 If disabled, KVM will not virtualize PMU functionality, 3072 e.g. MSRs, PMCs, PMIs, etc., even if userspace defines 3073 a virtual CPU model that contains PMU assets. 3074 3075 Note, KVM's vPMU support implicitly requires running 3076 with an in-kernel local APIC, e.g. to deliver PMIs to 3077 the guest. Running without an in-kernel local APIC is 3078 not supported, though KVM will allow such a combination 3079 (with severely degraded functionality). 3080 3081 See also enable_mediated_pmu. 3082 3083 Default is Y (on). 3084 3085 kvm.enable_virt_at_load=[KVM,ARM64,LOONGARCH,MIPS,RISCV,X86] 3086 If enabled, KVM will enable virtualization in hardware 3087 when KVM is loaded, and disable virtualization when KVM 3088 is unloaded (if KVM is built as a module). 3089 3090 If disabled, KVM will dynamically enable and disable 3091 virtualization on-demand when creating and destroying 3092 VMs, i.e. on the 0=>1 and 1=>0 transitions of the 3093 number of VMs. 3094 3095 Enabling virtualization at module load avoids potential 3096 latency for creation of the 0=>1 VM, as KVM serializes 3097 virtualization enabling across all online CPUs. The 3098 "cost" of enabling virtualization when KVM is loaded, 3099 is that doing so may interfere with using out-of-tree 3100 hypervisors that want to "own" virtualization hardware. 3101 3102 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface. 3103 Default is false (don't support). 3104 3105 kvm.nx_huge_pages= 3106 [KVM] Controls the software workaround for the 3107 X86_BUG_ITLB_MULTIHIT bug. 3108 force : Always deploy workaround. 3109 off : Never deploy workaround. 3110 auto : Deploy workaround based on the presence of 3111 X86_BUG_ITLB_MULTIHIT. 3112 3113 Default is 'auto'. 3114 3115 If the software workaround is enabled for the host, 3116 guests do need not to enable it for nested guests. 3117 3118 kvm.nx_huge_pages_recovery_ratio= 3119 [KVM] Controls how many 4KiB pages are periodically zapped 3120 back to huge pages. 0 disables the recovery, otherwise if 3121 the value is N KVM will zap 1/Nth of the 4KiB pages every 3122 period (see below). The default is 60. 3123 3124 kvm.nx_huge_pages_recovery_period_ms= 3125 [KVM] Controls the time period at which KVM zaps 4KiB pages 3126 back to huge pages. If the value is a non-zero N, KVM will 3127 zap a portion (see ratio above) of the pages every N msecs. 3128 If the value is 0 (the default), KVM will pick a period based 3129 on the ratio, such that a page is zapped after 1 hour on average. 3130 3131 kvm-{amd,intel}.enable_mediated_pmu=[KVM,AMD,INTEL] 3132 If enabled, KVM will provide a mediated virtual PMU, 3133 instead of the default perf-based virtual PMU (if 3134 kvm.enable_pmu is true and PMU is enumerated via the 3135 virtual CPU model). 3136 3137 With a perf-based vPMU, KVM operates as a user of perf, 3138 i.e. emulates guest PMU counters using perf events. 3139 KVM-created perf events are managed by perf as regular 3140 (guest-only) events, e.g. are scheduled in/out, contend 3141 for hardware resources, etc. Using a perf-based vPMU 3142 allows guest and host usage of the PMU to co-exist, but 3143 incurs non-trivial overhead and can result in silently 3144 dropped guest events (due to resource contention). 3145 3146 With a mediated vPMU, hardware PMU state is context 3147 switched around the world switch to/from the guest. 3148 KVM mediates which events the guest can utilize, but 3149 gives the guest direct access to all other PMU assets 3150 when possible (KVM may intercept some accesses if the 3151 virtual CPU model provides a subset of hardware PMU 3152 functionality). Using a mediated vPMU significantly 3153 reduces PMU virtualization overhead and eliminates lost 3154 guest events, but is mutually exclusive with using perf 3155 to profile KVM guests and adds latency to most VM-Exits 3156 (to context switch PMU state). 3157 3158 Default is N (off). 3159 3160 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in 3161 KVM/SVM. Default is 1 (enabled). 3162 3163 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables, 3164 a.k.a. Two-Dimensional Page Tables. Default is 1 3165 (enabled). Disable by KVM if hardware lacks support 3166 for NPT. 3167 3168 kvm-amd.ciphertext_hiding_asids= 3169 [KVM,AMD] Ciphertext hiding prevents disallowed accesses 3170 to SNP private memory from reading ciphertext. Instead, 3171 reads will see constant default values (0xff). 3172 3173 If ciphertext hiding is enabled, the joint SEV-ES and 3174 SEV-SNP ASID space is partitioned into separate SEV-ES 3175 and SEV-SNP ASID ranges, with the SEV-SNP range being 3176 [1..max_snp_asid] and the SEV-ES range being 3177 (max_snp_asid..min_sev_asid), where min_sev_asid is 3178 enumerated by CPUID.0x.8000_001F[EDX]. 3179 3180 A non-zero value enables SEV-SNP ciphertext hiding and 3181 adjusts the ASID ranges for SEV-ES and SEV-SNP guests. 3182 KVM caps the number of SEV-SNP ASIDs at the maximum 3183 possible value, e.g. specifying -1u will assign all 3184 joint SEV-ES and SEV-SNP ASIDs to SEV-SNP. Note, 3185 assigning all joint ASIDs to SEV-SNP, i.e. configuring 3186 max_snp_asid == min_sev_asid-1, will effectively make 3187 SEV-ES unusable. 3188 3189 kvm-arm.mode= 3190 [KVM,ARM,EARLY] Select one of KVM/arm64's modes of 3191 operation. 3192 3193 none: Forcefully disable KVM. 3194 3195 nvhe: Standard nVHE-based mode, without support for 3196 protected guests. 3197 3198 protected: Mode with support for guests whose state is 3199 kept private from the host, using VHE or 3200 nVHE depending on HW support. 3201 3202 nested: VHE-based mode with support for nested 3203 virtualization. Requires at least ARMv8.4 3204 hardware (with FEAT_NV2). 3205 3206 Defaults to VHE/nVHE based on hardware support. Setting 3207 mode to "protected" will disable kexec and hibernation 3208 for the host. To force nVHE on VHE hardware, add 3209 "arm64_sw.hvhe=0 id_aa64mmfr1.vh=0" to the 3210 command-line. 3211 "nested" and "protected" are experimental and should be 3212 used with extreme caution. 3213 3214 kvm-arm.vgic_v3_group0_trap= 3215 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-0 3216 system registers 3217 3218 kvm-arm.vgic_v3_group1_trap= 3219 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-1 3220 system registers 3221 3222 kvm-arm.vgic_v3_common_trap= 3223 [KVM,ARM,EARLY] Trap guest accesses to GICv3 common 3224 system registers 3225 3226 kvm-arm.vgic_v4_enable= 3227 [KVM,ARM,EARLY] Allow use of GICv4 for direct 3228 injection of LPIs. 3229 3230 kvm-arm.wfe_trap_policy= 3231 [KVM,ARM] Control when to set WFE instruction trap for 3232 KVM VMs. Traps are allowed but not guaranteed by the 3233 CPU architecture. 3234 3235 trap: set WFE instruction trap 3236 3237 notrap: clear WFE instruction trap 3238 3239 kvm-arm.wfi_trap_policy= 3240 [KVM,ARM] Control when to set WFI instruction trap for 3241 KVM VMs. Traps are allowed but not guaranteed by the 3242 CPU architecture. 3243 3244 trap: set WFI instruction trap 3245 3246 notrap: clear WFI instruction trap 3247 3248 kvm_cma_resv_ratio=n [PPC,EARLY] 3249 Reserves given percentage from system memory area for 3250 contiguous memory allocation for KVM hash pagetable 3251 allocation. 3252 By default it reserves 5% of total system memory. 3253 Format: <integer> 3254 Default: 5 3255 3256 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables, 3257 a.k.a. Two-Dimensional Page Tables. Default is 1 3258 (enabled). Disable by KVM if hardware lacks support 3259 for EPT. 3260 3261 kvm-intel.emulate_invalid_guest_state= 3262 [KVM,Intel] Control whether to emulate invalid guest 3263 state. Ignored if kvm-intel.enable_unrestricted_guest=1, 3264 as guest state is never invalid for unrestricted 3265 guests. This param doesn't apply to nested guests (L2), 3266 as KVM never emulates invalid L2 guest state. 3267 Default is 1 (enabled). 3268 3269 kvm-intel.flexpriority= 3270 [KVM,Intel] Control KVM's use of FlexPriority feature 3271 (TPR shadow). Default is 1 (enabled). Disable by KVM if 3272 hardware lacks support for it. 3273 3274 kvm-intel.nested= 3275 [KVM,Intel] Control nested virtualization feature in 3276 KVM/VMX. Default is 1 (enabled). 3277 3278 kvm-intel.unrestricted_guest= 3279 [KVM,Intel] Control KVM's use of unrestricted guest 3280 feature (virtualized real and unpaged mode). Default 3281 is 1 (enabled). Disable by KVM if EPT is disabled or 3282 hardware lacks support for it. 3283 3284 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault 3285 CVE-2018-3620. 3286 3287 Valid arguments: never, cond, always 3288 3289 always: L1D cache flush on every VMENTER. 3290 cond: Flush L1D on VMENTER only when the code between 3291 VMEXIT and VMENTER can leak host memory. 3292 never: Disables the mitigation 3293 3294 Default is cond (do L1 cache flush in specific instances) 3295 3296 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor 3297 Identification feature (tagged TLBs). Default is 1 3298 (enabled). Disable by KVM if hardware lacks support 3299 for it. 3300 3301 l1d_flush= [X86,INTEL,EARLY] 3302 Control mitigation for L1D based snooping vulnerability. 3303 3304 Certain CPUs are vulnerable to an exploit against CPU 3305 internal buffers which can forward information to a 3306 disclosure gadget under certain conditions. 3307 3308 In vulnerable processors, the speculatively 3309 forwarded data can be used in a cache side channel 3310 attack, to access data to which the attacker does 3311 not have direct access. 3312 3313 This parameter controls the mitigation. The 3314 options are: 3315 3316 on - enable the interface for the mitigation 3317 3318 l1tf= [X86,EARLY] Control mitigation of the L1TF vulnerability on 3319 affected CPUs 3320 3321 The kernel PTE inversion protection is unconditionally 3322 enabled and cannot be disabled. 3323 3324 full 3325 Provides all available mitigations for the 3326 L1TF vulnerability. Disables SMT and 3327 enables all mitigations in the 3328 hypervisors, i.e. unconditional L1D flush. 3329 3330 SMT control and L1D flush control via the 3331 sysfs interface is still possible after 3332 boot. Hypervisors will issue a warning 3333 when the first VM is started in a 3334 potentially insecure configuration, 3335 i.e. SMT enabled or L1D flush disabled. 3336 3337 full,force 3338 Same as 'full', but disables SMT and L1D 3339 flush runtime control. Implies the 3340 'nosmt=force' command line option. 3341 (i.e. sysfs control of SMT is disabled.) 3342 3343 flush 3344 Leaves SMT enabled and enables the default 3345 hypervisor mitigation, i.e. conditional 3346 L1D flush. 3347 3348 SMT control and L1D flush control via the 3349 sysfs interface is still possible after 3350 boot. Hypervisors will issue a warning 3351 when the first VM is started in a 3352 potentially insecure configuration, 3353 i.e. SMT enabled or L1D flush disabled. 3354 3355 flush,nosmt 3356 3357 Disables SMT and enables the default 3358 hypervisor mitigation. 3359 3360 SMT control and L1D flush control via the 3361 sysfs interface is still possible after 3362 boot. Hypervisors will issue a warning 3363 when the first VM is started in a 3364 potentially insecure configuration, 3365 i.e. SMT enabled or L1D flush disabled. 3366 3367 flush,nowarn 3368 Same as 'flush', but hypervisors will not 3369 warn when a VM is started in a potentially 3370 insecure configuration. 3371 3372 off 3373 Disables hypervisor mitigations and doesn't 3374 emit any warnings. 3375 It also drops the swap size and available 3376 RAM limit restriction on both hypervisor and 3377 bare metal. 3378 3379 Default is 'flush'. 3380 3381 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst 3382 3383 l2cr= [PPC] 3384 3385 l3cr= [PPC] 3386 3387 lapic [X86-32,APIC,EARLY] Enable the local APIC even if BIOS 3388 disabled it. 3389 3390 lapic= [X86,APIC] Do not use TSC deadline 3391 value for LAPIC timer one-shot implementation. Default 3392 back to the programmable timer unit in the LAPIC. 3393 Format: notscdeadline 3394 3395 lapic_timer_c2_ok [X86,APIC,EARLY] trust the local apic timer 3396 in C2 power state. 3397 3398 libata.dma= [LIBATA] DMA control 3399 libata.dma=0 Disable all PATA and SATA DMA 3400 libata.dma=1 PATA and SATA Disk DMA only 3401 libata.dma=2 ATAPI (CDROM) DMA only 3402 libata.dma=4 Compact Flash DMA only 3403 Combinations also work, so libata.dma=3 enables DMA 3404 for disks and CDROMs, but not CFs. 3405 3406 libata.ignore_hpa= [LIBATA] Ignore HPA limit 3407 libata.ignore_hpa=0 keep BIOS limits (default) 3408 libata.ignore_hpa=1 ignore limits, using full disk 3409 3410 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume 3411 when set. 3412 Format: <int> 3413 3414 libata.force= [LIBATA] Force configurations. The format is a comma- 3415 separated list of "[ID:]VAL" where ID is PORT[.DEVICE]. 3416 PORT and DEVICE are decimal numbers matching port, link 3417 or device. Basically, it matches the ATA ID string 3418 printed on console by libata. If the whole ID part is 3419 omitted, the last PORT and DEVICE values are used. If 3420 ID hasn't been specified yet, the configuration applies 3421 to all ports, links and devices. 3422 3423 If only DEVICE is omitted, the parameter applies to 3424 the port and all links and devices behind it. DEVICE 3425 number of 0 either selects the first device or the 3426 first fan-out link behind PMP device. It does not 3427 select the host link. DEVICE number of 15 selects the 3428 host link and device attached to it. 3429 3430 The VAL specifies the configuration to force. As long 3431 as there is no ambiguity, shortcut notation is allowed. 3432 For example, both 1.5 and 1.5G would work for 1.5Gbps. 3433 The following configurations can be forced. 3434 3435 * Cable type: 40c, 80c, short40c, unk, ign or sata. 3436 Any ID with matching PORT is used. 3437 3438 * SATA link speed limit: 1.5Gbps or 3.0Gbps. 3439 3440 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7]. 3441 udma[/][16,25,33,44,66,100,133] notation is also 3442 allowed. 3443 3444 * nohrst, nosrst, norst: suppress hard, soft and both 3445 resets. 3446 3447 * rstonce: only attempt one reset during hot-unplug 3448 link recovery. 3449 3450 * [no]dbdelay: Enable or disable the extra 200ms delay 3451 before debouncing a link PHY and device presence 3452 detection. 3453 3454 * [no]ncq: Turn on or off NCQ. 3455 3456 * [no]ncqtrim: Enable or disable queued DSM TRIM. 3457 3458 * [no]ncqati: Enable or disable NCQ trim on ATI chipset. 3459 3460 * [no]trim: Enable or disable (unqueued) TRIM. 3461 3462 * trim_zero: Indicate that TRIM command zeroes data. 3463 3464 * max_trim_128m: Set 128M maximum trim size limit. 3465 3466 * [no]dma: Turn on or off DMA transfers. 3467 3468 * atapi_dmadir: Enable ATAPI DMADIR bridge support. 3469 3470 * atapi_mod16_dma: Enable the use of ATAPI DMA for 3471 commands that are not a multiple of 16 bytes. 3472 3473 * [no]dmalog: Enable or disable the use of the 3474 READ LOG DMA EXT command to access logs. 3475 3476 * [no]iddevlog: Enable or disable access to the 3477 identify device data log. 3478 3479 * [no]logdir: Enable or disable access to the general 3480 purpose log directory. 3481 3482 * max_sec=<sectors>: Set the transfer size limit, in 3483 number of 512-byte sectors, to the value specified in 3484 <sectors>. The value specified in <sectors> has to be 3485 a non-zero positive integer. 3486 3487 * max_sec_128: Set transfer size limit to 128 sectors. 3488 3489 * max_sec_1024: Set or clear transfer size limit to 3490 1024 sectors. 3491 3492 * max_sec_lba48: Set or clear transfer size limit to 3493 65535 sectors. 3494 3495 * external: Mark port as external (hotplug-capable). 3496 3497 * [no]lpm: Enable or disable link power management. 3498 3499 * [no]setxfer: Indicate if transfer speed mode setting 3500 should be skipped. 3501 3502 * [no]fua: Disable or enable FUA (Force Unit Access) 3503 support for devices supporting this feature. 3504 3505 * dump_id: Dump IDENTIFY data. 3506 3507 * disable: Disable this device. 3508 3509 If there are multiple matching configurations changing 3510 the same attribute, the last one is used. 3511 3512 liveupdate= [KNL,EARLY] 3513 Format: <bool> 3514 Enable Live Update Orchestrator (LUO). 3515 Default: off. 3516 3517 lockd.nlm_grace_period=P [NFS] Assign grace period. 3518 Format: <integer> 3519 3520 lockd.nlm_tcpport=N [NFS] Assign TCP port. 3521 Format: <integer> 3522 3523 lockd.nlm_timeout=T [NFS] Assign timeout value. 3524 Format: <integer> 3525 3526 lockd.nlm_udpport=M [NFS] Assign UDP port. 3527 Format: <integer> 3528 3529 lockdown= [SECURITY,EARLY] 3530 { integrity | confidentiality } 3531 Enable the kernel lockdown feature. If set to 3532 integrity, kernel features that allow userland to 3533 modify the running kernel are disabled. If set to 3534 confidentiality, kernel features that allow userland 3535 to extract confidential information from the kernel 3536 are also disabled. 3537 3538 locktorture.acq_writer_lim= [KNL] 3539 Set the time limit in jiffies for a lock 3540 acquisition. Acquisitions exceeding this limit 3541 will result in a splat once they do complete. 3542 3543 locktorture.bind_readers= [KNL] 3544 Specify the list of CPUs to which the readers are 3545 to be bound. 3546 3547 locktorture.bind_writers= [KNL] 3548 Specify the list of CPUs to which the writers are 3549 to be bound. 3550 3551 locktorture.call_rcu_chains= [KNL] 3552 Specify the number of self-propagating call_rcu() 3553 chains to set up. These are used to ensure that 3554 there is a high probability of an RCU grace period 3555 in progress at any given time. Defaults to 0, 3556 which disables these call_rcu() chains. 3557 3558 locktorture.long_hold= [KNL] 3559 Specify the duration in milliseconds for the 3560 occasional long-duration lock hold time. Defaults 3561 to 100 milliseconds. Select 0 to disable. 3562 3563 locktorture.nested_locks= [KNL] 3564 Specify the maximum lock nesting depth that 3565 locktorture is to exercise, up to a limit of 8 3566 (MAX_NESTED_LOCKS). Specify zero to disable. 3567 Note that this parameter is ineffective on types 3568 of locks that do not support nested acquisition. 3569 3570 locktorture.nreaders_stress= [KNL] 3571 Set the number of locking read-acquisition kthreads. 3572 Defaults to being automatically set based on the 3573 number of online CPUs. 3574 3575 locktorture.nwriters_stress= [KNL] 3576 Set the number of locking write-acquisition kthreads. 3577 3578 locktorture.onoff_holdoff= [KNL] 3579 Set time (s) after boot for CPU-hotplug testing. 3580 3581 locktorture.onoff_interval= [KNL] 3582 Set time (s) between CPU-hotplug operations, or 3583 zero to disable CPU-hotplug testing. 3584 3585 locktorture.rt_boost= [KNL] 3586 Do periodic testing of real-time lock priority 3587 boosting. Select 0 to disable, 1 to boost 3588 only rt_mutex, and 2 to boost unconditionally. 3589 Defaults to 2, which might seem to be an 3590 odd choice, but which should be harmless for 3591 non-real-time spinlocks, due to their disabling 3592 of preemption. Note that non-realtime mutexes 3593 disable boosting. 3594 3595 locktorture.rt_boost_factor= [KNL] 3596 Number that determines how often and for how 3597 long priority boosting is exercised. This is 3598 scaled down by the number of writers, so that the 3599 number of boosts per unit time remains roughly 3600 constant as the number of writers increases. 3601 On the other hand, the duration of each boost 3602 increases with the number of writers. 3603 3604 locktorture.shuffle_interval= [KNL] 3605 Set task-shuffle interval (jiffies). Shuffling 3606 tasks allows some CPUs to go into dyntick-idle 3607 mode during the locktorture test. 3608 3609 locktorture.shutdown_secs= [KNL] 3610 Set time (s) after boot system shutdown. This 3611 is useful for hands-off automated testing. 3612 3613 locktorture.stat_interval= [KNL] 3614 Time (s) between statistics printk()s. 3615 3616 locktorture.stutter= [KNL] 3617 Time (s) to stutter testing, for example, 3618 specifying five seconds causes the test to run for 3619 five seconds, wait for five seconds, and so on. 3620 This tests the locking primitive's ability to 3621 transition abruptly to and from idle. 3622 3623 locktorture.torture_type= [KNL] 3624 Specify the locking implementation to test. 3625 3626 locktorture.verbose= [KNL] 3627 Enable additional printk() statements. 3628 3629 locktorture.writer_fifo= [KNL] 3630 Run the write-side locktorture kthreads at 3631 sched_set_fifo() real-time priority. 3632 3633 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver 3634 Format: <irq> 3635 3636 loglevel= [KNL,EARLY] 3637 All Kernel Messages with a loglevel smaller than the 3638 console loglevel will be printed to the console. It can 3639 also be changed with klogd or other programs. The 3640 loglevels are defined as follows: 3641 3642 0 (KERN_EMERG) system is unusable 3643 1 (KERN_ALERT) action must be taken immediately 3644 2 (KERN_CRIT) critical conditions 3645 3 (KERN_ERR) error conditions 3646 4 (KERN_WARNING) warning conditions 3647 5 (KERN_NOTICE) normal but significant condition 3648 6 (KERN_INFO) informational 3649 7 (KERN_DEBUG) debug-level messages 3650 3651 log_buf_len=n[KMG] [KNL,EARLY] 3652 Sets the size of the printk ring buffer, in bytes. 3653 n must be a power of two and greater than the 3654 minimal size. The minimal size is defined by 3655 LOG_BUF_SHIFT kernel config parameter. There 3656 is also CONFIG_LOG_CPU_MAX_BUF_SHIFT config 3657 parameter that allows to increase the default size 3658 depending on the number of CPUs. See init/Kconfig 3659 for more details. 3660 3661 logo.nologo [FB] Disables display of the built-in Linux logo. 3662 This may be used to provide more screen space for 3663 kernel log messages and is useful when debugging 3664 kernel boot problems. 3665 3666 lp=0 [LP] Specify parallel ports to use, e.g, 3667 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses 3668 lp=reset first parallel port). 'lp=0' disables the 3669 lp=auto printer driver. 'lp=reset' (which can be 3670 specified in addition to the ports) causes 3671 attached printers to be reset. Using 3672 lp=port1,port2,... specifies the parallel ports 3673 to associate lp devices with, starting with 3674 lp0. A port specification may be 'none' to skip 3675 that lp device, or a parport name such as 3676 'parport0'. Specifying 'lp=auto' instead of a 3677 port specification list means that device IDs 3678 from each port should be examined, to see if 3679 an IEEE 1284-compliant printer is attached; if 3680 so, the driver will manage that printer. 3681 See also header of drivers/char/lp.c. 3682 3683 lpj=n [KNL] 3684 Sets loops_per_jiffy to given constant, thus avoiding 3685 time-consuming boot-time autodetection (up to 250 ms per 3686 CPU). 0 enables autodetection (default). To determine 3687 the correct value for your kernel, boot with normal 3688 autodetection and see what value is printed. Note that 3689 on SMP systems the preset will be applied to all CPUs, 3690 which is likely to cause problems if your CPUs need 3691 significantly divergent settings. An incorrect value 3692 will cause delays in the kernel to be wrong, leading to 3693 unpredictable I/O errors and other breakage. Although 3694 unlikely, in the extreme case this might damage your 3695 hardware. 3696 3697 lsm.debug [SECURITY] Enable LSM initialization debugging output. 3698 3699 lsm=lsm1,...,lsmN 3700 [SECURITY] Choose order of LSM initialization. This 3701 overrides CONFIG_LSM, and the "security=" parameter. 3702 3703 machtype= [Loongson] Share the same kernel image file between 3704 different yeeloong laptops. 3705 Example: machtype=lemote-yeeloong-2f-7inch 3706 3707 maxcpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 3708 will bring up during bootup. maxcpus=n : n >= 0 limits 3709 the kernel to bring up 'n' processors. Surely after 3710 bootup you can bring up the other plugged cpu by executing 3711 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus 3712 only takes effect during system bootup. 3713 While n=0 is a special case, it is equivalent to "nosmp", 3714 which also disables the IO APIC. 3715 3716 max_loop= [LOOP] The number of loop block devices that get 3717 (loop.max_loop) unconditionally pre-created at init time. The default 3718 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead 3719 of statically allocating a predefined number, loop 3720 devices can be requested on-demand with the 3721 /dev/loop-control interface. 3722 3723 mce= [X86-{32,64}] 3724 3725 Please see Documentation/arch/x86/x86_64/machinecheck.rst for sysfs runtime tunables. 3726 3727 off 3728 disable machine check 3729 3730 no_cmci 3731 disable CMCI(Corrected Machine Check Interrupt) that 3732 Intel processor supports. Usually this disablement is 3733 not recommended, but it might be handy if your 3734 hardware is misbehaving. 3735 3736 Note that you'll get more problems without CMCI than 3737 with due to the shared banks, i.e. you might get 3738 duplicated error logs. 3739 3740 dont_log_ce 3741 don't make logs for corrected errors. All events 3742 reported as corrected are silently cleared by OS. This 3743 option will be useful if you have no interest in any 3744 of corrected errors. 3745 3746 ignore_ce 3747 disable features for corrected errors, e.g. 3748 polling timer and CMCI. All events reported as 3749 corrected are not cleared by OS and remained in its 3750 error banks. 3751 3752 Usually this disablement is not recommended, however 3753 if there is an agent checking/clearing corrected 3754 errors (e.g. BIOS or hardware monitoring 3755 applications), conflicting with OS's error handling, 3756 and you cannot deactivate the agent, then this option 3757 will be a help. 3758 3759 no_lmce 3760 do not opt-in to Local MCE delivery. Use legacy method 3761 to broadcast MCEs. 3762 3763 bootlog 3764 enable logging of machine checks left over from 3765 booting. Disabled by default on AMD Fam10h and older 3766 because some BIOS leave bogus ones. 3767 3768 If your BIOS doesn't do that it's a good idea to 3769 enable though to make sure you log even machine check 3770 events that result in a reboot. On Intel systems it is 3771 enabled by default. 3772 3773 nobootlog 3774 disable boot machine check logging. 3775 3776 monarchtimeout (number) 3777 sets the time in us to wait for other CPUs on machine 3778 checks. 0 to disable. 3779 3780 bios_cmci_threshold 3781 don't overwrite the bios-set CMCI threshold. This boot 3782 option prevents Linux from overwriting the CMCI 3783 threshold set by the bios. Without this option, Linux 3784 always sets the CMCI threshold to 1. Enabling this may 3785 make memory predictive failure analysis less effective 3786 if the bios sets thresholds for memory errors since we 3787 will not see details for all errors. 3788 3789 recovery 3790 force-enable recoverable machine check code paths 3791 3792 Everything else is in sysfs now. 3793 3794 3795 md= [HW] RAID subsystems devices and level 3796 See Documentation/admin-guide/md.rst. 3797 3798 mdacon= [MDA] 3799 Format: <first>,<last> 3800 Specifies range of consoles to be captured by the MDA. 3801 3802 mds= [X86,INTEL,EARLY] 3803 Control mitigation for the Micro-architectural Data 3804 Sampling (MDS) vulnerability. 3805 3806 Certain CPUs are vulnerable to an exploit against CPU 3807 internal buffers which can forward information to a 3808 disclosure gadget under certain conditions. 3809 3810 In vulnerable processors, the speculatively 3811 forwarded data can be used in a cache side channel 3812 attack, to access data to which the attacker does 3813 not have direct access. 3814 3815 This parameter controls the MDS mitigation. The 3816 options are: 3817 3818 full - Enable MDS mitigation on vulnerable CPUs 3819 full,nosmt - Enable MDS mitigation and disable 3820 SMT on vulnerable CPUs 3821 off - Unconditionally disable MDS mitigation 3822 3823 On TAA-affected machines, mds=off can be prevented by 3824 an active TAA mitigation as both vulnerabilities are 3825 mitigated with the same mechanism so in order to disable 3826 this mitigation, you need to specify tsx_async_abort=off 3827 too. 3828 3829 Not specifying this option is equivalent to 3830 mds=full. 3831 3832 For details see: Documentation/admin-guide/hw-vuln/mds.rst 3833 3834 mem=nn[KMG] [HEXAGON,EARLY] Set the memory size. 3835 Must be specified, otherwise memory size will be 0. 3836 3837 mem=nn[KMG] [KNL,BOOT,EARLY] Force usage of a specific amount 3838 of memory Amount of memory to be used in cases 3839 as follows: 3840 3841 1 for test; 3842 2 when the kernel is not able to see the whole system memory; 3843 3 memory that lies after 'mem=' boundary is excluded from 3844 the hypervisor, then assigned to KVM guests. 3845 4 to limit the memory available for kdump kernel. 3846 3847 [ARC,MICROBLAZE] - the limit applies only to low memory, 3848 high memory is not affected. 3849 3850 [ARM64] - only limits memory covered by the linear 3851 mapping. The NOMAP regions are not affected. 3852 3853 [X86] Work as limiting max address. Use together 3854 with memmap= to avoid physical address space collisions. 3855 Without memmap= PCI devices could be placed at addresses 3856 belonging to unused RAM. 3857 3858 Note that this only takes effects during boot time since 3859 in above case 3, memory may need be hot added after boot 3860 if system memory of hypervisor is not sufficient. 3861 3862 mem=nn[KMG]@ss[KMG] 3863 [ARM,MIPS,EARLY] - override the memory layout 3864 reported by firmware. 3865 Define a memory region of size nn[KMG] starting at 3866 ss[KMG]. 3867 Multiple different regions can be specified with 3868 multiple mem= parameters on the command line. 3869 3870 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel 3871 memory. 3872 3873 memblock=debug [KNL,EARLY] Enable memblock debug messages. 3874 3875 memchunk=nn[KMG] 3876 [KNL,SH] Allow user to override the default size for 3877 per-device physically contiguous DMA buffers. 3878 3879 memhp_default_state=online/offline/online_kernel/online_movable 3880 [KNL] Set the initial state for the memory hotplug 3881 onlining policy. If not specified, the default value is 3882 set according to the 3883 CONFIG_MHP_DEFAULT_ONLINE_TYPE kernel config 3884 options. 3885 See Documentation/admin-guide/mm/memory-hotplug.rst. 3886 3887 memmap=exactmap [KNL,X86,EARLY] Enable setting of an exact 3888 E820 memory map, as specified by the user. 3889 Such memmap=exactmap lines can be constructed based on 3890 BIOS output or other requirements. See the memmap=nn@ss 3891 option description. 3892 3893 memmap=nn[KMG]@ss[KMG] 3894 [KNL, X86,MIPS,XTENSA,EARLY] Force usage of a specific region of memory. 3895 Region of memory to be used is from ss to ss+nn. 3896 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG], 3897 which limits max address to nn[KMG]. 3898 Multiple different regions can be specified, 3899 comma delimited. 3900 Example: 3901 memmap=100M@2G,100M#3G,1G!1024G 3902 3903 memmap=nn[KMG]#ss[KMG] 3904 [KNL,ACPI,EARLY] Mark specific memory as ACPI data. 3905 Region of memory to be marked is from ss to ss+nn. 3906 3907 memmap=nn[KMG]$ss[KMG] 3908 [KNL,ACPI,EARLY] Mark specific memory as reserved. 3909 Region of memory to be reserved is from ss to ss+nn. 3910 Example: Exclude memory from 0x18690000-0x1869ffff 3911 memmap=64K$0x18690000 3912 or 3913 memmap=0x10000$0x18690000 3914 Some bootloaders may need an escape character before '$', 3915 like Grub2, otherwise '$' and the following number 3916 will be eaten. 3917 3918 memmap=nn[KMG]!ss[KMG,EARLY] 3919 [KNL,X86] Mark specific memory as protected. 3920 Region of memory to be used, from ss to ss+nn. 3921 The memory region may be marked as e820 type 12 (0xc) 3922 and is NVDIMM or ADR memory. 3923 3924 memmap=<size>%<offset>-<oldtype>+<newtype> 3925 [KNL,ACPI,EARLY] Convert memory within the specified region 3926 from <oldtype> to <newtype>. If "-<oldtype>" is left 3927 out, the whole region will be marked as <newtype>, 3928 even if previously unavailable. If "+<newtype>" is left 3929 out, matching memory will be removed. Types are 3930 specified as e820 types, e.g., 1 = RAM, 2 = reserved, 3931 3 = ACPI, 12 = PRAM. 3932 3933 memory_corruption_check=0/1 [X86,EARLY] 3934 Some BIOSes seem to corrupt the first 64k of 3935 memory when doing things like suspend/resume. 3936 Setting this option will scan the memory 3937 looking for corruption. Enabling this will 3938 both detect corruption and prevent the kernel 3939 from using the memory being corrupted. 3940 However, it's intended as a diagnostic tool; if 3941 repeatable BIOS-originated corruption always 3942 affects the same memory, you can use memmap= 3943 to prevent the kernel from using that memory. 3944 3945 memory_corruption_check_size=size [X86,EARLY] 3946 By default it checks for corruption in the low 3947 64k, making this memory unavailable for normal 3948 use. Use this parameter to scan for 3949 corruption in more or less memory. 3950 3951 memory_corruption_check_period=seconds [X86,EARLY] 3952 By default it checks for corruption every 60 3953 seconds. Use this parameter to check at some 3954 other rate. 0 disables periodic checking. 3955 3956 memory_hotplug.memmap_on_memory 3957 [KNL,X86,ARM] Boolean flag to enable this feature. 3958 Format: {on | off (default)} 3959 When enabled, runtime hotplugged memory will 3960 allocate its internal metadata (struct pages, 3961 those vmemmap pages cannot be optimized even 3962 if hugetlb_free_vmemmap is enabled) from the 3963 hotadded memory which will allow to hotadd a 3964 lot of memory without requiring additional 3965 memory to do so. 3966 This feature is disabled by default because it 3967 has some implication on large (e.g. GB) 3968 allocations in some configurations (e.g. small 3969 memory blocks). 3970 The state of the flag can be read in 3971 /sys/module/memory_hotplug/parameters/memmap_on_memory. 3972 Note that even when enabled, there are a few cases where 3973 the feature is not effective. 3974 3975 memtest= [KNL,X86,ARM,M68K,PPC,RISCV,EARLY] Enable memtest 3976 Format: <integer> 3977 default : 0 <disable> 3978 Specifies the number of memtest passes to be 3979 performed. Each pass selects another test 3980 pattern from a given set of patterns. Memtest 3981 fills the memory with this pattern, validates 3982 memory contents and reserves bad memory 3983 regions that are detected. 3984 3985 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control 3986 Valid arguments: on, off 3987 Default: off 3988 mem_encrypt=on: Activate SME 3989 mem_encrypt=off: Do not activate SME 3990 3991 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst 3992 for details on when memory encryption can be activated. 3993 3994 mem_sleep_default= [SUSPEND] Default system suspend mode: 3995 s2idle - Suspend-To-Idle 3996 shallow - Power-On Suspend or equivalent (if supported) 3997 deep - Suspend-To-RAM or equivalent (if supported) 3998 See Documentation/admin-guide/pm/sleep-states.rst. 3999 4000 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when 4001 the BIOS has incorrectly applied a workaround. TinyBIOS 4002 version 0.98 is known to be affected, 0.99 fixes the 4003 problem by letting the user disable the workaround. 4004 4005 mga= [HW,DRM] 4006 4007 microcode= [X86] Control the behavior of the microcode loader. 4008 Available options, comma separated: 4009 4010 base_rev=X - with <X> with format: <u32> 4011 Set the base microcode revision of each thread when in 4012 debug mode. 4013 4014 dis_ucode_ldr: disable the microcode loader 4015 4016 force_minrev: 4017 Enable or disable the microcode minimal revision 4018 enforcement for the runtime microcode loader. 4019 4020 mini2440= [ARM,HW,KNL] 4021 Format:[0..2][b][c][t] 4022 Default: "0tb" 4023 MINI2440 configuration specification: 4024 0 - The attached screen is the 3.5" TFT 4025 1 - The attached screen is the 7" TFT 4026 2 - The VGA Shield is attached (1024x768) 4027 Leaving out the screen size parameter will not load 4028 the TFT driver, and the framebuffer will be left 4029 unconfigured. 4030 b - Enable backlight. The TFT backlight pin will be 4031 linked to the kernel VESA blanking code and a GPIO 4032 LED. This parameter is not necessary when using the 4033 VGA shield. 4034 c - Enable the s3c camera interface. 4035 t - Reserved for enabling touchscreen support. The 4036 touchscreen support is not enabled in the mainstream 4037 kernel as of 2.6.30, a preliminary port can be found 4038 in the "bleeding edge" mini2440 support kernel at 4039 https://repo.or.cz/w/linux-2.6/mini2440.git 4040 4041 mitigations= 4042 [X86,PPC,S390,ARM64,EARLY] Control optional mitigations for 4043 CPU vulnerabilities. This is a set of curated, 4044 arch-independent options, each of which is an 4045 aggregation of existing arch-specific options. 4046 4047 Note, "mitigations" is supported if and only if the 4048 kernel was built with CPU_MITIGATIONS=y. 4049 4050 off 4051 Disable all optional CPU mitigations. This 4052 improves system performance, but it may also 4053 expose users to several CPU vulnerabilities. 4054 Equivalent to: if nokaslr then kpti=0 [ARM64] 4055 gather_data_sampling=off [X86] 4056 indirect_target_selection=off [X86] 4057 kvm.nx_huge_pages=off [X86] 4058 l1tf=off [X86] 4059 mds=off [X86] 4060 mmio_stale_data=off [X86] 4061 no_entry_flush [PPC] 4062 no_uaccess_flush [PPC] 4063 nobp=0 [S390] 4064 nopti [X86,PPC] 4065 nospectre_bhb [ARM64] 4066 nospectre_v1 [X86,PPC] 4067 nospectre_v2 [X86,PPC,S390,ARM64] 4068 reg_file_data_sampling=off [X86] 4069 retbleed=off [X86] 4070 spec_rstack_overflow=off [X86] 4071 spec_store_bypass_disable=off [X86,PPC] 4072 spectre_bhi=off [X86] 4073 spectre_v2_user=off [X86] 4074 srbds=off [X86,INTEL] 4075 ssbd=force-off [ARM64] 4076 tsa=off [X86,AMD] 4077 tsx_async_abort=off [X86] 4078 vmscape=off [X86] 4079 4080 Exceptions: 4081 This does not have any effect on 4082 kvm.nx_huge_pages when 4083 kvm.nx_huge_pages=force. 4084 4085 auto (default) 4086 Mitigate all CPU vulnerabilities, but leave SMT 4087 enabled, even if it's vulnerable. This is for 4088 users who don't want to be surprised by SMT 4089 getting disabled across kernel upgrades, or who 4090 have other ways of avoiding SMT-based attacks. 4091 Equivalent to: (default behavior) 4092 4093 auto,nosmt 4094 Mitigate all CPU vulnerabilities, disabling SMT 4095 if needed. This is for users who always want to 4096 be fully mitigated, even if it means losing SMT. 4097 Equivalent to: l1tf=flush,nosmt [X86] 4098 mds=full,nosmt [X86] 4099 tsx_async_abort=full,nosmt [X86] 4100 mmio_stale_data=full,nosmt [X86] 4101 retbleed=auto,nosmt [X86] 4102 4103 [X86] After one of the above options, additionally 4104 supports attack-vector based controls as documented in 4105 Documentation/admin-guide/hw-vuln/attack_vector_controls.rst 4106 4107 mminit_loglevel= 4108 [KNL,EARLY] When CONFIG_DEBUG_MEMORY_INIT is set, this 4109 parameter allows control of the logging verbosity for 4110 the additional memory initialisation checks. A value 4111 of 0 disables mminit logging and a level of 4 will 4112 log everything. Information is printed at KERN_DEBUG 4113 so loglevel=8 may also need to be specified. 4114 4115 mmio_stale_data= 4116 [X86,INTEL,EARLY] Control mitigation for the Processor 4117 MMIO Stale Data vulnerabilities. 4118 4119 Processor MMIO Stale Data is a class of 4120 vulnerabilities that may expose data after an MMIO 4121 operation. Exposed data could originate or end in 4122 the same CPU buffers as affected by MDS and TAA. 4123 Therefore, similar to MDS and TAA, the mitigation 4124 is to clear the affected CPU buffers. 4125 4126 This parameter controls the mitigation. The 4127 options are: 4128 4129 full - Enable mitigation on vulnerable CPUs 4130 4131 full,nosmt - Enable mitigation and disable SMT on 4132 vulnerable CPUs. 4133 4134 off - Unconditionally disable mitigation 4135 4136 On MDS or TAA affected machines, 4137 mmio_stale_data=off can be prevented by an active 4138 MDS or TAA mitigation as these vulnerabilities are 4139 mitigated with the same mechanism so in order to 4140 disable this mitigation, you need to specify 4141 mds=off and tsx_async_abort=off too. 4142 4143 Not specifying this option is equivalent to 4144 mmio_stale_data=full. 4145 4146 For details see: 4147 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst 4148 4149 <module>.async_probe[=<bool>] [KNL] 4150 If no <bool> value is specified or if the value 4151 specified is not a valid <bool>, enable asynchronous 4152 probe on this module. Otherwise, enable/disable 4153 asynchronous probe on this module as indicated by the 4154 <bool> value. See also: module.async_probe 4155 4156 module.async_probe=<bool> 4157 [KNL] When set to true, modules will use async probing 4158 by default. To enable/disable async probing for a 4159 specific module, use the module specific control that 4160 is documented under <module>.async_probe. When both 4161 module.async_probe and <module>.async_probe are 4162 specified, <module>.async_probe takes precedence for 4163 the specific module. 4164 4165 module.enable_dups_trace 4166 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set, 4167 this means that duplicate request_module() calls will 4168 trigger a WARN_ON() instead of a pr_warn(). Note that 4169 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s 4170 will always be issued and this option does nothing. 4171 module.sig_enforce 4172 [KNL] When CONFIG_MODULE_SIG is set, this means that 4173 modules without (valid) signatures will fail to load. 4174 Note that if CONFIG_MODULE_SIG_FORCE is set, that 4175 is always true, so this option does nothing. 4176 4177 module_blacklist= [KNL] Do not load a comma-separated list of 4178 modules. Useful for debugging problem modules. 4179 4180 mousedev.tap_time= 4181 [MOUSE] Maximum time between finger touching and 4182 leaving touchpad surface for touch to be considered 4183 a tap and be reported as a left button click (for 4184 touchpads working in absolute mode only). 4185 Format: <msecs> 4186 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices 4187 reporting absolute coordinates, such as tablets 4188 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices 4189 reporting absolute coordinates, such as tablets 4190 4191 movablecore= [KNL,X86,PPC,EARLY] 4192 Format: nn[KMGTPE] | nn% 4193 This parameter is the complement to kernelcore=, it 4194 specifies the amount of memory used for migratable 4195 allocations. If both kernelcore and movablecore is 4196 specified, then kernelcore will be at *least* the 4197 specified value but may be more. If movablecore on its 4198 own is specified, the administrator must be careful 4199 that the amount of memory usable for all allocations 4200 is not too small. 4201 4202 movable_node [KNL,EARLY] Boot-time switch to make hotplugable memory 4203 NUMA nodes to be movable. This means that the memory 4204 of such nodes will be usable only for movable 4205 allocations which rules out almost all kernel 4206 allocations. Use with caution! 4207 4208 MTD_Partition= [MTD] 4209 Format: <name>,<region-number>,<size>,<offset> 4210 4211 MTD_Region= [MTD] Format: 4212 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>] 4213 4214 mtdparts= [MTD] 4215 See drivers/mtd/parsers/cmdlinepart.c 4216 4217 mtouchusb.raw_coordinates= 4218 [HW] Make the MicroTouch USB driver use raw coordinates 4219 ('y', default) or cooked coordinates ('n') 4220 4221 mtrr=debug [X86,EARLY] 4222 Enable printing debug information related to MTRR 4223 registers at boot time. 4224 4225 mtrr_chunk_size=nn[KMG,X86,EARLY] 4226 used for mtrr cleanup. It is largest continuous chunk 4227 that could hold holes aka. UC entries. 4228 4229 mtrr_gran_size=nn[KMG,X86,EARLY] 4230 Used for mtrr cleanup. It is granularity of mtrr block. 4231 Default is 1. 4232 Large value could prevent small alignment from 4233 using up MTRRs. 4234 4235 mtrr_spare_reg_nr=n [X86,EARLY] 4236 Format: <integer> 4237 Range: 0,7 : spare reg number 4238 Default : 1 4239 Used for mtrr cleanup. It is spare mtrr entries number. 4240 Set to 2 or more if your graphical card needs more. 4241 4242 multitce=off [PPC] This parameter disables the use of the pSeries 4243 firmware feature for updating multiple TCE entries 4244 at a time. 4245 4246 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card 4247 4248 netdev= [NET] Network devices parameters 4249 Format: <irq>,<io>,<mem_start>,<mem_end>,<name> 4250 Note that mem_start is often overloaded to mean 4251 something different and driver-specific. 4252 This usage is only documented in each driver source 4253 file if at all. 4254 4255 netpoll.carrier_timeout= 4256 [NET] Specifies amount of time (in seconds) that 4257 netpoll should wait for a carrier. By default netpoll 4258 waits 4 seconds. 4259 4260 nf_conntrack.acct= 4261 [NETFILTER] Enable connection tracking flow accounting 4262 0 to disable accounting 4263 1 to enable accounting 4264 Default value is 0. 4265 4266 nfs.cache_getent= 4267 [NFS] sets the pathname to the program which is used 4268 to update the NFS client cache entries. 4269 4270 nfs.cache_getent_timeout= 4271 [NFS] sets the timeout after which an attempt to 4272 update a cache entry is deemed to have failed. 4273 4274 nfs.callback_nr_threads= 4275 [NFSv4] set the total number of threads that the 4276 NFS client will assign to service NFSv4 callback 4277 requests. 4278 4279 nfs.callback_tcpport= 4280 [NFS] set the TCP port on which the NFSv4 callback 4281 channel should listen. 4282 4283 nfs.delay_retrans= 4284 [NFS] specifies the number of times the NFSv4 client 4285 retries the request before returning an EAGAIN error, 4286 after a reply of NFS4ERR_DELAY from the server. 4287 Only applies if the softerr mount option is enabled, 4288 and the specified value is >= 0. 4289 4290 nfs.enable_ino64= 4291 [NFS] enable 64-bit inode numbers. 4292 If zero, the NFS client will fake up a 32-bit inode 4293 number for the readdir() and stat() syscalls instead 4294 of returning the full 64-bit number. 4295 The default is to return 64-bit inode numbers. 4296 4297 nfs.idmap_cache_timeout= 4298 [NFS] set the maximum lifetime for idmapper cache 4299 entries. 4300 4301 nfs.max_session_cb_slots= 4302 [NFSv4.1] Sets the maximum number of session 4303 slots the client will assign to the callback 4304 channel. This determines the maximum number of 4305 callbacks the client will process in parallel for 4306 a particular server. 4307 4308 nfs.max_session_slots= 4309 [NFSv4.1] Sets the maximum number of session slots 4310 the client will attempt to negotiate with the server. 4311 This limits the number of simultaneous RPC requests 4312 that the client can send to the NFSv4.1 server. 4313 Note that there is little point in setting this 4314 value higher than the max_tcp_slot_table_limit. 4315 4316 nfs.nfs4_disable_idmapping= 4317 [NFSv4] When set to the default of '1', this option 4318 ensures that both the RPC level authentication 4319 scheme and the NFS level operations agree to use 4320 numeric uids/gids if the mount is using the 4321 'sec=sys' security flavour. In effect it is 4322 disabling idmapping, which can make migration from 4323 legacy NFSv2/v3 systems to NFSv4 easier. 4324 Servers that do not support this mode of operation 4325 will be autodetected by the client, and it will fall 4326 back to using the idmapper. 4327 To turn off this behaviour, set the value to '0'. 4328 4329 nfs.nfs4_unique_id= 4330 [NFS4] Specify an additional fixed unique ident- 4331 ification string that NFSv4 clients can insert into 4332 their nfs_client_id4 string. This is typically a 4333 UUID that is generated at system install time. 4334 4335 nfs.recover_lost_locks= 4336 [NFSv4] Attempt to recover locks that were lost due 4337 to a lease timeout on the server. Please note that 4338 doing this risks data corruption, since there are 4339 no guarantees that the file will remain unchanged 4340 after the locks are lost. 4341 If you want to enable the kernel legacy behaviour of 4342 attempting to recover these locks, then set this 4343 parameter to '1'. 4344 The default parameter value of '0' causes the kernel 4345 not to attempt recovery of lost locks. 4346 4347 nfs.send_implementation_id= 4348 [NFSv4.1] Send client implementation identification 4349 information in exchange_id requests. 4350 If zero, no implementation identification information 4351 will be sent. 4352 The default is to send the implementation identification 4353 information. 4354 4355 nfs4.layoutstats_timer= 4356 [NFSv4.2] Change the rate at which the kernel sends 4357 layoutstats to the pNFS metadata server. 4358 4359 Setting this to value to 0 causes the kernel to use 4360 whatever value is the default set by the layout 4361 driver. A non-zero value sets the minimum interval 4362 in seconds between layoutstats transmissions. 4363 4364 nfsd.inter_copy_offload_enable= 4365 [NFSv4.2] When set to 1, the server will support 4366 server-to-server copies for which this server is 4367 the destination of the copy. 4368 4369 nfsd.nfs4_disable_idmapping= 4370 [NFSv4] When set to the default of '1', the NFSv4 4371 server will return only numeric uids and gids to 4372 clients using auth_sys, and will accept numeric uids 4373 and gids from such clients. This is intended to ease 4374 migration from NFSv2/v3. 4375 4376 nfsd.nfsd4_ssc_umount_timeout= 4377 [NFSv4.2] When used as the destination of a 4378 server-to-server copy, knfsd temporarily mounts 4379 the source server. It caches the mount in case 4380 it will be needed again, and discards it if not 4381 used for the number of milliseconds specified by 4382 this parameter. 4383 4384 nfsaddrs= [NFS] Deprecated. Use ip= instead. 4385 See Documentation/admin-guide/nfs/nfsroot.rst. 4386 4387 nfsroot= [NFS] nfs root filesystem for disk-less boxes. 4388 See Documentation/admin-guide/nfs/nfsroot.rst. 4389 4390 nfsrootdebug [NFS] enable nfsroot debugging messages. 4391 See Documentation/admin-guide/nfs/nfsroot.rst. 4392 4393 nmi_backtrace.backtrace_idle [KNL] 4394 Dump stacks even of idle CPUs in response to an 4395 NMI stack-backtrace request. 4396 4397 nmi_debug= [KNL,SH] Specify one or more actions to take 4398 when a NMI is triggered. 4399 Format: [state][,regs][,debounce][,die] 4400 4401 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels 4402 Format: [panic,][nopanic,][rNNN,][num] 4403 Valid num: 0 or 1 4404 0 - turn hardlockup detector in nmi_watchdog off 4405 1 - turn hardlockup detector in nmi_watchdog on 4406 rNNN - configure the watchdog with raw perf event 0xNNN 4407 4408 When panic is specified, panic when an NMI watchdog 4409 timeout occurs (or 'nopanic' to not panic on an NMI 4410 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set) 4411 To disable both hard and soft lockup detectors, 4412 please see 'nowatchdog'. 4413 This is useful when you use a panic=... timeout and 4414 need the box quickly up again. 4415 4416 These settings can be accessed at runtime via 4417 the nmi_watchdog and hardlockup_panic sysctls. 4418 4419 no4lvl [RISCV,EARLY] Disable 4-level and 5-level paging modes. 4420 Forces kernel to use 3-level paging instead. 4421 4422 no5lvl [X86-64,RISCV,EARLY] Disable 5-level paging mode. Forces 4423 kernel to use 4-level paging instead. 4424 4425 noalign [KNL,ARM] 4426 4427 noapic [SMP,APIC,EARLY] Tells the kernel to not make use of any 4428 IOAPICs that may be present in the system. 4429 4430 noapictimer [APIC,X86] Don't set up the APIC timer 4431 4432 noautogroup Disable scheduler automatic task group creation. 4433 4434 nocache [ARM,EARLY] 4435 4436 no_console_suspend 4437 [HW] Never suspend the console 4438 Disable suspending of consoles during suspend and 4439 hibernate operations. Once disabled, debugging 4440 messages can reach various consoles while the rest 4441 of the system is being put to sleep (ie, while 4442 debugging driver suspend/resume hooks). This may 4443 not work reliably with all consoles, but is known 4444 to work with serial and VGA consoles. 4445 To facilitate more flexible debugging, we also add 4446 console_suspend, a printk module parameter to control 4447 it. Users could use console_suspend (usually 4448 /sys/module/printk/parameters/console_suspend) to 4449 turn on/off it dynamically. 4450 4451 no_debug_objects 4452 [KNL,EARLY] Disable object debugging 4453 4454 nodsp [SH] Disable hardware DSP at boot time. 4455 4456 noefi [EFI,EARLY] Disable EFI runtime services support. 4457 4458 no_entry_flush [PPC,EARLY] Don't flush the L1-D cache when entering the kernel. 4459 4460 noexec32 [X86-64] 4461 This affects only 32-bit executables. 4462 noexec32=on: enable non-executable mappings (default) 4463 read doesn't imply executable mappings 4464 noexec32=off: disable non-executable mappings 4465 read implies executable mappings 4466 4467 no_file_caps Tells the kernel not to honor file capabilities. The 4468 only way then for a file to be executed with privilege 4469 is to be setuid root or executed by root. 4470 4471 nofpu [MIPS,SH] Disable hardware FPU at boot time. 4472 4473 nofsgsbase [X86] Disables FSGSBASE instructions. 4474 4475 nofxsr [BUGS=X86-32] Disables x86 floating point extended 4476 register save and restore. The kernel will only save 4477 legacy floating-point registers on task switch. 4478 4479 nogbpages [X86] Do not use GB pages for kernel direct mappings. 4480 4481 no_hash_pointers 4482 [KNL,EARLY] 4483 Alias for "hash_pointers=never". 4484 4485 nohibernate [HIBERNATION] Disable hibernation and resume. 4486 4487 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,RISCV,SH] Forces the kernel to 4488 busy wait in do_idle() and not use the arch_cpu_idle() 4489 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP 4490 to be effective. This is useful on platforms where the 4491 sleep(SH) or wfi(ARM,ARM64) instructions do not work 4492 correctly or when doing power measurements to evaluate 4493 the impact of the sleep instructions. This is also 4494 useful when using JTAG debugger. 4495 4496 nohpet [X86] Don't use the HPET timer. 4497 4498 nohugeiomap [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge I/O mappings. 4499 4500 nohugevmalloc [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge vmalloc mappings. 4501 4502 nohz= [KNL] Boottime enable/disable dynamic ticks 4503 Valid arguments: on, off 4504 Default: on 4505 4506 nohz_full= [KNL,BOOT,SMP,ISOL] 4507 The argument is a cpu list, as described above. 4508 In kernels built with CONFIG_NO_HZ_FULL=y, set 4509 the specified list of CPUs whose tick will be stopped 4510 whenever possible. The boot CPU will be forced outside 4511 the range to maintain the timekeeping. Any CPUs 4512 in this list will have their RCU callbacks offloaded, 4513 just as if they had also been called out in the 4514 rcu_nocbs= boot parameter. 4515 4516 Note that this argument takes precedence over 4517 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 4518 4519 noinitrd [Deprecated,RAM] Tells the kernel not to load any configured 4520 initial RAM disk. Currently this parameter applies to 4521 initrd only, not to initramfs. But it applies to both 4522 in EFI mode. 4523 4524 nointremap [X86-64,Intel-IOMMU,EARLY] Do not enable interrupt 4525 remapping. 4526 [Deprecated - use intremap=off] 4527 4528 noinvpcid [X86,EARLY] Disable the INVPCID cpu feature. 4529 4530 noiotrap [SH] Disables trapped I/O port accesses. 4531 4532 noirqdebug [X86-32] Disables the code which attempts to detect and 4533 disable unhandled interrupt sources. 4534 4535 noisapnp [ISAPNP] Disables ISA PnP code. 4536 4537 nokaslr [KNL,EARLY] 4538 When CONFIG_RANDOMIZE_BASE is set, this disables 4539 kernel and module base offset ASLR (Address Space 4540 Layout Randomization). 4541 4542 no-kvmapf [X86,KVM,EARLY] Disable paravirtualized asynchronous page 4543 fault handling. 4544 4545 no-kvmclock [X86,KVM,EARLY] Disable paravirtualized KVM clock driver 4546 4547 nolapic [X86-32,APIC,EARLY] Do not enable or use the local APIC. 4548 4549 nolapic_timer [X86-32,APIC,EARLY] Do not use the local APIC timer. 4550 4551 nomce [X86-32] Disable Machine Check Exception 4552 4553 nomfgpt [X86-32] Disable Multi-Function General Purpose 4554 Timer usage (for AMD Geode machines). 4555 4556 nomodeset Disable kernel modesetting. Most systems' firmware 4557 sets up a display mode and provides framebuffer memory 4558 for output. With nomodeset, DRM and fbdev drivers will 4559 not load if they could possibly displace the pre- 4560 initialized output. Only the system framebuffer will 4561 be available for use. The respective drivers will not 4562 perform display-mode changes or accelerated rendering. 4563 4564 Useful as error fallback, or for testing and debugging. 4565 4566 nomodule Disable module load 4567 4568 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to 4569 shutdown the other cpus. Instead use the REBOOT_VECTOR 4570 irq. 4571 4572 nopat [X86,EARLY] Disable PAT (page attribute table extension of 4573 pagetables) support. 4574 4575 nopcid [X86-64,EARLY] Disable the PCID cpu feature. 4576 4577 nopku [X86] Disable Memory Protection Keys CPU feature found 4578 in some Intel CPUs. 4579 4580 nopti [X86-64,EARLY] 4581 Equivalent to pti=off 4582 4583 nopv= [X86,XEN,KVM,HYPER_V,VMWARE,EARLY] 4584 Disables the PV optimizations forcing the guest to run 4585 as generic guest with no PV drivers. Currently support 4586 XEN HVM, KVM, HYPER_V and VMWARE guest. 4587 4588 nopvspin [X86,XEN,KVM,EARLY] 4589 Disables the qspinlock slow path using PV optimizations 4590 which allow the hypervisor to 'idle' the guest on lock 4591 contention. 4592 4593 norandmaps Don't use address space randomization. Equivalent to 4594 echo 0 > /proc/sys/kernel/randomize_va_space 4595 4596 noreplace-smp [X86-32,SMP] Don't replace SMP instructions 4597 with UP alternatives 4598 4599 noresume [SWSUSP] Disables resume and restores original swap 4600 space. 4601 4602 no-scroll [VGA] Disables scrollback. 4603 This is required for the Braillex ib80-piezo Braille 4604 reader made by F.H. Papenmeier (Germany). 4605 4606 nosgx [X86-64,SGX,EARLY] Disables Intel SGX kernel support. 4607 4608 nosmap [PPC,EARLY] 4609 Disable SMAP (Supervisor Mode Access Prevention) 4610 even if it is supported by processor. 4611 4612 nosmep [PPC64s,EARLY] 4613 Disable SMEP (Supervisor Mode Execution Prevention) 4614 even if it is supported by processor. 4615 4616 nosmp [SMP,EARLY] Tells an SMP kernel to act as a UP kernel, 4617 and disable the IO APIC. legacy for "maxcpus=0". 4618 4619 nosmt [KNL,MIPS,PPC,EARLY] Disable symmetric multithreading (SMT). 4620 Equivalent to smt=1. 4621 4622 [KNL,LOONGARCH,X86,PPC,S390] Disable symmetric multithreading (SMT). 4623 nosmt=force: Force disable SMT, cannot be undone 4624 via the sysfs control file. 4625 4626 nosoftlockup [KNL] Disable the soft-lockup detector. 4627 4628 nospec_store_bypass_disable 4629 [HW,EARLY] Disable all mitigations for the Speculative 4630 Store Bypass vulnerability 4631 4632 nospectre_bhb [ARM64,EARLY] Disable all mitigations for Spectre-BHB (branch 4633 history injection) vulnerability. System may allow data leaks 4634 with this option. 4635 4636 nospectre_v1 [X86,PPC,EARLY] Disable mitigations for Spectre Variant 1 4637 (bounds check bypass). With this option data leaks are 4638 possible in the system. 4639 4640 nospectre_v2 [X86,PPC_E500,ARM64,EARLY] Disable all mitigations 4641 for the Spectre variant 2 (indirect branch 4642 prediction) vulnerability. System may allow data 4643 leaks with this option. 4644 4645 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES,RISCV,LOONGARCH,EARLY] 4646 Disable paravirtualized steal time accounting. steal time 4647 is computed, but won't influence scheduler behaviour 4648 4649 nosync [HW,M68K] Disables sync negotiation for all devices. 4650 4651 no_timer_check [X86,APIC] Disables the code which tests for broken 4652 timer IRQ sources, i.e., the IO-APIC timer. This can 4653 work around problems with incorrect timer 4654 initialization on some boards. 4655 4656 no_uaccess_flush 4657 [PPC,EARLY] Don't flush the L1-D cache after accessing user data. 4658 4659 novmcoredd [KNL,KDUMP] 4660 Disable device dump. Device dump allows drivers to 4661 append dump data to vmcore so you can collect driver 4662 specified debug info. Drivers can append the data 4663 without any limit and this data is stored in memory, 4664 so this may cause significant memory stress. Disabling 4665 device dump can help save memory but the driver debug 4666 data will be no longer available. This parameter 4667 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP 4668 is set. 4669 4670 no-vmw-sched-clock 4671 [X86,PV_OPS,EARLY] Disable paravirtualized VMware 4672 scheduler clock and use the default one. 4673 4674 nowatchdog [KNL] Disable both lockup detectors, i.e. 4675 soft-lockup and NMI watchdog (hard-lockup). 4676 4677 nowb [ARM,EARLY] 4678 4679 nox2apic [X86-64,APIC,EARLY] Do not enable x2APIC mode. 4680 4681 NOTE: this parameter will be ignored on systems with the 4682 LEGACY_XAPIC_DISABLED bit set in the 4683 IA32_XAPIC_DISABLE_STATUS MSR. 4684 4685 noxsave [BUGS=X86] Disables x86 extended register state save 4686 and restore using xsave. The kernel will fallback to 4687 enabling legacy floating-point and sse state. 4688 4689 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended 4690 register states. The kernel will fall back to use 4691 xsave to save the states. By using this parameter, 4692 performance of saving the states is degraded because 4693 xsave doesn't support modified optimization while 4694 xsaveopt supports it on xsaveopt enabled systems. 4695 4696 noxsaves [X86] Disables xsaves and xrstors used in saving and 4697 restoring x86 extended register state in compacted 4698 form of xsave area. The kernel will fall back to use 4699 xsaveopt and xrstor to save and restore the states 4700 in standard form of xsave area. By using this 4701 parameter, xsave area per process might occupy more 4702 memory on xsaves enabled systems. 4703 4704 nr_cpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 4705 could support. nr_cpus=n : n >= 1 limits the kernel to 4706 support 'n' processors. It could be larger than the 4707 number of already plugged CPU during bootup, later in 4708 runtime you can physically add extra cpu until it reaches 4709 n. So during boot up some boot time memory for per-cpu 4710 variables need be pre-allocated for later physical cpu 4711 hot plugging. 4712 4713 nr_uarts= [SERIAL] maximum number of UARTs to be registered. 4714 4715 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86, EARLY] 4716 Disable NUMA, Only set up a single NUMA node 4717 spanning all memory. 4718 4719 numa=fake=<size>[MG] 4720 [KNL, ARM64, RISCV, X86, EARLY] 4721 If given as a memory unit, fills all system RAM with 4722 nodes of size interleaved over physical nodes. 4723 4724 numa=fake=<N> 4725 [KNL, ARM64, RISCV, X86, EARLY] 4726 If given as an integer, fills all system RAM with N 4727 fake nodes interleaved over physical nodes. 4728 4729 numa=fake=<N>U 4730 [KNL, ARM64, RISCV, X86, EARLY] 4731 If given as an integer followed by 'U', it will 4732 divide each physical node into N emulated nodes. 4733 4734 numa=noacpi [X86] Don't parse the SRAT table for NUMA setup 4735 4736 numa=nohmat [X86] Don't parse the HMAT table for NUMA setup, or 4737 soft-reserved memory partitioning. 4738 4739 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic 4740 NUMA balancing. 4741 Allowed values are enable and disable 4742 4743 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA. 4744 'node', 'default' can be specified 4745 This can be set from sysctl after boot. 4746 See Documentation/admin-guide/sysctl/vm.rst for details. 4747 4748 nvme.quirks= [NVME] A list of quirk entries to augment the built-in 4749 nvme quirk list. List entries are separated by a 4750 '-' character. 4751 Each entry has the form VendorID:ProductID:quirk_names. 4752 The IDs are 4-digits hex numbers and quirk_names is a 4753 list of quirk names separated by commas. A quirk name 4754 can be prefixed by '^', meaning that the specified 4755 quirk must be disabled. 4756 4757 Example: 4758 nvme.quirks=7710:2267:bogus_nid,^identify_cns-9900:7711:broken_msi 4759 4760 ohci1394_dma=early [HW,EARLY] enable debugging via the ohci1394 driver. 4761 See Documentation/core-api/debugging-via-ohci1394.rst for more 4762 info. 4763 4764 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands 4765 Rather than timing out after 20 ms if an EC 4766 command is not properly ACKed, override the length 4767 of the timeout. We have interrupts disabled while 4768 waiting for the ACK, so if this is set too high 4769 interrupts *may* be lost! 4770 4771 omap_mux= [OMAP] Override bootloader pin multiplexing. 4772 Format: <mux_mode0.mode_name=value>... 4773 For example, to override I2C bus2: 4774 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100 4775 4776 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration 4777 4778 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock] 4779 4780 boundary - index of last SLC block on Flex-OneNAND. 4781 The remaining blocks are configured as MLC blocks. 4782 lock - Configure if Flex-OneNAND boundary should be locked. 4783 Once locked, the boundary cannot be changed. 4784 1 indicates lock status, 0 indicates unlock status. 4785 4786 oops=panic [KNL,EARLY] 4787 Always panic on oopses. Default is to just kill the 4788 process, but there is a small probability of 4789 deadlocking the machine. 4790 This will also cause panics on machine check exceptions. 4791 Useful together with panic=30 to trigger a reboot. 4792 4793 page_alloc.shuffle= 4794 [KNL] Boolean flag to control whether the page allocator 4795 should randomize its free lists. This parameter can be 4796 used to enable/disable page randomization. The state of 4797 the flag can be read from sysfs at: 4798 /sys/module/page_alloc/parameters/shuffle. 4799 This parameter is only available if CONFIG_SHUFFLE_PAGE_ALLOCATOR=y. 4800 4801 page_owner= [KNL,EARLY] Boot-time page_owner enabling option. 4802 Storage of the information about who allocated 4803 each page is disabled in default. With this switch, 4804 we can turn it on. 4805 on: enable the feature 4806 4807 page_poison= [KNL,EARLY] Boot-time parameter changing the state of 4808 poisoning on the buddy allocator, available with 4809 CONFIG_PAGE_POISONING=y. 4810 off: turn off poisoning (default) 4811 on: turn on poisoning 4812 4813 page_reporting.page_reporting_order= 4814 [KNL] Minimal page reporting order 4815 Format: <integer> 4816 Adjust the minimal page reporting order. The page 4817 reporting is disabled when it exceeds MAX_PAGE_ORDER. 4818 4819 panic= [KNL] Kernel behaviour on panic: delay <timeout> 4820 timeout > 0: seconds before rebooting 4821 timeout = 0: wait forever 4822 timeout < 0: reboot immediately 4823 Format: <timeout> 4824 4825 panic_on_taint= [KNL,EARLY] 4826 Bitmask for conditionally calling panic() in add_taint() 4827 Format: <hex>[,nousertaint] 4828 Hexadecimal bitmask representing the set of TAINT flags 4829 that will cause the kernel to panic when add_taint() is 4830 called with any of the flags in this set. 4831 The optional switch "nousertaint" can be utilized to 4832 prevent userspace forced crashes by writing to sysctl 4833 /proc/sys/kernel/tainted any flagset matching with the 4834 bitmask set on panic_on_taint. 4835 See Documentation/admin-guide/tainted-kernels.rst for 4836 extra details on the taint flags that users can pick 4837 to compose the bitmask to assign to panic_on_taint. 4838 4839 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump 4840 on a WARN(). 4841 4842 panic_force_cpu= 4843 [KNL,SMP] Force panic handling to execute on a specific CPU. 4844 Format: <cpu number> 4845 Some platforms require panic handling to occur on a 4846 specific CPU for the crash kernel to function correctly. 4847 This can be due to firmware limitations, interrupt routing 4848 constraints, or platform-specific requirements where only 4849 a particular CPU can safely enter the crash kernel. 4850 When set, panic() will redirect execution to the specified 4851 CPU before proceeding with the normal panic and kexec flow. 4852 If the target CPU is offline or unavailable, panic proceeds 4853 on the current CPU. 4854 This option should only be used for systems with the above 4855 constraints as it might cause the panic operation to be less reliable. 4856 4857 panic_print= Bitmask for printing system info when panic happens. 4858 User can chose combination of the following bits: 4859 bit 0: print all tasks info 4860 bit 1: print system memory info 4861 bit 2: print timer info 4862 bit 3: print locks info if CONFIG_LOCKDEP is on 4863 bit 4: print ftrace buffer 4864 bit 5: replay all kernel messages on consoles at the end of panic 4865 bit 6: print all CPUs backtrace (if available in the arch) 4866 bit 7: print only tasks in uninterruptible (blocked) state 4867 *Be aware* that this option may print a _lot_ of lines, 4868 so there are risks of losing older messages in the log. 4869 Use this option carefully, maybe worth to setup a 4870 bigger log buffer with "log_buf_len" along with this. 4871 4872 panic_sys_info= A comma separated list of extra information to be dumped 4873 on panic. 4874 Format: val[,val...] 4875 Where @val can be any of the following: 4876 4877 tasks: print all tasks info 4878 mem: print system memory info 4879 timers: print timers info 4880 locks: print locks info if CONFIG_LOCKDEP is on 4881 ftrace: print ftrace buffer 4882 all_bt: print all CPUs backtrace (if available in the arch) 4883 blocked_tasks: print only tasks in uninterruptible (blocked) state 4884 4885 This is a human readable alternative to the 'panic_print' option. 4886 4887 panic_console_replay 4888 When panic happens, replay all kernel messages on 4889 consoles at the end of panic. 4890 4891 parkbd.port= [HW] Parallel port number the keyboard adapter is 4892 connected to, default is 0. 4893 Format: <parport#> 4894 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation, 4895 0 for XT, 1 for AT (default is AT). 4896 Format: <mode> 4897 4898 parport= [HW,PPT] Specify parallel ports. 0 disables. 4899 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] } 4900 Use 'auto' to force the driver to use any 4901 IRQ/DMA settings detected (the default is to 4902 ignore detected IRQ/DMA settings because of 4903 possible conflicts). You can specify the base 4904 address, IRQ, and DMA settings; IRQ and DMA 4905 should be numbers, or 'auto' (for using detected 4906 settings on that particular port), or 'nofifo' 4907 (to avoid using a FIFO even if it is detected). 4908 Parallel ports are assigned in the order they 4909 are specified on the command line, starting 4910 with parport0. 4911 4912 parport_init_mode= [HW,PPT] 4913 Configure VIA parallel port to operate in 4914 a specific mode. This is necessary on Pegasos 4915 computer where firmware has no options for setting 4916 up parallel port mode and sets it to spp. 4917 Currently this function knows 686a and 8231 chips. 4918 Format: [spp|ps2|epp|ecp|ecpepp] 4919 4920 pata_legacy.all= [HW,LIBATA] 4921 Format: <int> 4922 Set to non-zero to probe primary and secondary ISA 4923 port ranges on PCI systems where no PCI PATA device 4924 has been found at either range. Disabled by default. 4925 4926 pata_legacy.autospeed= [HW,LIBATA] 4927 Format: <int> 4928 Set to non-zero if a chip is present that snoops speed 4929 changes. Disabled by default. 4930 4931 pata_legacy.ht6560a= [HW,LIBATA] 4932 Format: <int> 4933 Set to 1, 2, or 3 for HT 6560A on the primary channel, 4934 the secondary channel, or both channels respectively. 4935 Disabled by default. 4936 4937 pata_legacy.ht6560b= [HW,LIBATA] 4938 Format: <int> 4939 Set to 1, 2, or 3 for HT 6560B on the primary channel, 4940 the secondary channel, or both channels respectively. 4941 Disabled by default. 4942 4943 pata_legacy.iordy_mask= [HW,LIBATA] 4944 Format: <int> 4945 IORDY enable mask. Set individual bits to allow IORDY 4946 for the respective channel. Bit 0 is for the first 4947 legacy channel handled by this driver, bit 1 is for 4948 the second channel, and so on. The sequence will often 4949 correspond to the primary legacy channel, the secondary 4950 legacy channel, and so on, but the handling of a PCI 4951 bus and the use of other driver options may interfere 4952 with the sequence. By default IORDY is allowed across 4953 all channels. 4954 4955 pata_legacy.opti82c46x= [HW,LIBATA] 4956 Format: <int> 4957 Set to 1, 2, or 3 for Opti 82c611A on the primary 4958 channel, the secondary channel, or both channels 4959 respectively. Disabled by default. 4960 4961 pata_legacy.opti82c611a= [HW,LIBATA] 4962 Format: <int> 4963 Set to 1, 2, or 3 for Opti 82c465MV on the primary 4964 channel, the secondary channel, or both channels 4965 respectively. Disabled by default. 4966 4967 pata_legacy.pio_mask= [HW,LIBATA] 4968 Format: <int> 4969 PIO mode mask for autospeed devices. Set individual 4970 bits to allow the use of the respective PIO modes. 4971 Bit 0 is for mode 0, bit 1 is for mode 1, and so on. 4972 All modes allowed by default. 4973 4974 pata_legacy.probe_all= [HW,LIBATA] 4975 Format: <int> 4976 Set to non-zero to probe tertiary and further ISA 4977 port ranges on PCI systems. Disabled by default. 4978 4979 pata_legacy.probe_mask= [HW,LIBATA] 4980 Format: <int> 4981 Probe mask for legacy ISA PATA ports. Depending on 4982 platform configuration and the use of other driver 4983 options up to 6 legacy ports are supported: 0x1f0, 4984 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing 4985 of individual ports can be disabled by setting the 4986 corresponding bits in the mask to 1. Bit 0 is for 4987 the first port in the list above (0x1f0), and so on. 4988 By default all supported ports are probed. 4989 4990 pata_legacy.qdi= [HW,LIBATA] 4991 Format: <int> 4992 Set to non-zero to probe QDI controllers. By default 4993 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise. 4994 4995 pata_legacy.winbond= [HW,LIBATA] 4996 Format: <int> 4997 Set to non-zero to probe Winbond controllers. Use 4998 the standard I/O port (0x130) if 1, otherwise the 4999 value given is the I/O port to use (typically 0x1b0). 5000 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE, 5001 0 otherwise. 5002 5003 pata_platform.pio_mask= [HW,LIBATA] 5004 Format: <int> 5005 Supported PIO mode mask. Set individual bits to allow 5006 the use of the respective PIO modes. Bit 0 is for 5007 mode 0, bit 1 is for mode 1, and so on. Mode 0 only 5008 allowed by default. 5009 5010 pause_on_oops=<int> 5011 Halt all CPUs after the first oops has been printed for 5012 the specified number of seconds. This is to be used if 5013 your oopses keep scrolling off the screen. 5014 5015 pcbit= [HW,ISDN] 5016 5017 pci=option[,option...] [PCI,EARLY] various PCI subsystem options. 5018 5019 Some options herein operate on a specific device 5020 or a set of devices (<pci_dev>). These are 5021 specified in one of the following formats: 5022 5023 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]* 5024 pci:<vendor>:<device>[:<subvendor>:<subdevice>] 5025 5026 Note: the first format specifies a PCI 5027 bus/device/function address which may change 5028 if new hardware is inserted, if motherboard 5029 firmware changes, or due to changes caused 5030 by other kernel parameters. If the 5031 domain is left unspecified, it is 5032 taken to be zero. Optionally, a path 5033 to a device through multiple device/function 5034 addresses can be specified after the base 5035 address (this is more robust against 5036 renumbering issues). The second format 5037 selects devices using IDs from the 5038 configuration space which may match multiple 5039 devices in the system. 5040 5041 earlydump dump PCI config space before the kernel 5042 changes anything 5043 off [X86] don't probe for the PCI bus 5044 bios [X86-32] force use of PCI BIOS, don't access 5045 the hardware directly. Use this if your machine 5046 has a non-standard PCI host bridge. 5047 nobios [X86-32] disallow use of PCI BIOS, only direct 5048 hardware access methods are allowed. Use this 5049 if you experience crashes upon bootup and you 5050 suspect they are caused by the BIOS. 5051 conf1 [X86] Force use of PCI Configuration Access 5052 Mechanism 1 (config address in IO port 0xCF8, 5053 data in IO port 0xCFC, both 32-bit). 5054 conf2 [X86] Force use of PCI Configuration Access 5055 Mechanism 2 (IO port 0xCF8 is an 8-bit port for 5056 the function, IO port 0xCFA, also 8-bit, sets 5057 bus number. The config space is then accessed 5058 through ports 0xC000-0xCFFF). 5059 See http://wiki.osdev.org/PCI for more info 5060 on the configuration access mechanisms. 5061 noaer [PCIE] If the PCIEAER kernel config parameter is 5062 enabled, this kernel boot option can be used to 5063 disable the use of PCIE advanced error reporting. 5064 nodomains [PCI] Disable support for multiple PCI 5065 root domains (aka PCI segments, in ACPI-speak). 5066 nommconf [X86] Disable use of MMCONFIG for PCI 5067 Configuration 5068 check_enable_amd_mmconf [X86] check for and enable 5069 properly configured MMIO access to PCI 5070 config space on AMD family 10h CPU 5071 nomsi [MSI] If the PCI_MSI kernel config parameter is 5072 enabled, this kernel boot option can be used to 5073 disable the use of MSI interrupts system-wide. 5074 noioapicquirk [APIC] Disable all boot interrupt quirks. 5075 Safety option to keep boot IRQs enabled. This 5076 should never be necessary. 5077 ioapicreroute [APIC] Enable rerouting of boot IRQs to the 5078 primary IO-APIC for bridges that cannot disable 5079 boot IRQs. This fixes a source of spurious IRQs 5080 when the system masks IRQs. 5081 noioapicreroute [APIC] Disable workaround that uses the 5082 boot IRQ equivalent of an IRQ that connects to 5083 a chipset where boot IRQs cannot be disabled. 5084 The opposite of ioapicreroute. 5085 biosirq [X86-32] Use PCI BIOS calls to get the interrupt 5086 routing table. These calls are known to be buggy 5087 on several machines and they hang the machine 5088 when used, but on other computers it's the only 5089 way to get the interrupt routing table. Try 5090 this option if the kernel is unable to allocate 5091 IRQs or discover secondary PCI buses on your 5092 motherboard. 5093 rom [X86] Assign address space to expansion ROMs. 5094 Use with caution as certain devices share 5095 address decoders between ROMs and other 5096 resources. 5097 norom [X86] Do not assign address space to 5098 expansion ROMs that do not already have 5099 BIOS assigned address ranges. 5100 nobar [X86] Do not assign address space to the 5101 BARs that weren't assigned by the BIOS. 5102 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be 5103 assigned automatically to PCI devices. You can 5104 make the kernel exclude IRQs of your ISA cards 5105 this way. 5106 pirqaddr=0xAAAAA [X86] Specify the physical address 5107 of the PIRQ table (normally generated 5108 by the BIOS) if it is outside the 5109 F0000h-100000h range. 5110 lastbus=N [X86] Scan all buses thru bus #N. Can be 5111 useful if the kernel is unable to find your 5112 secondary buses and you want to tell it 5113 explicitly which ones they are. 5114 assign-busses [X86] Always assign all PCI bus 5115 numbers ourselves, overriding 5116 whatever the firmware may have done. 5117 usepirqmask [X86] Honor the possible IRQ mask stored 5118 in the BIOS $PIR table. This is needed on 5119 some systems with broken BIOSes, notably 5120 some HP Pavilion N5400 and Omnibook XE3 5121 notebooks. This will have no effect if ACPI 5122 IRQ routing is enabled. 5123 noacpi [X86] Do not use ACPI for IRQ routing 5124 or for PCI scanning. 5125 use_crs [X86] Use PCI host bridge window information 5126 from ACPI. On BIOSes from 2008 or later, this 5127 is enabled by default. If you need to use this, 5128 please report a bug. 5129 nocrs [X86] Ignore PCI host bridge windows from ACPI. 5130 If you need to use this, please report a bug. 5131 use_e820 [X86] Use E820 reservations to exclude parts of 5132 PCI host bridge windows. This is a workaround 5133 for BIOS defects in host bridge _CRS methods. 5134 If you need to use this, please report a bug to 5135 <linux-pci@vger.kernel.org>. 5136 no_e820 [X86] Ignore E820 reservations for PCI host 5137 bridge windows. This is the default on modern 5138 hardware. If you need to use this, please report 5139 a bug to <linux-pci@vger.kernel.org>. 5140 routeirq Do IRQ routing for all PCI devices. 5141 This is normally done in pci_enable_device(), 5142 so this option is a temporary workaround 5143 for broken drivers that don't call it. 5144 skip_isa_align [X86] do not align io start addr, so can 5145 handle more pci cards 5146 noearly [X86] Don't do any early type 1 scanning. 5147 This might help on some broken boards which 5148 machine check when some devices' config space 5149 is read. But various workarounds are disabled 5150 and some IOMMU drivers will not work. 5151 bfsort Sort PCI devices into breadth-first order. 5152 This sorting is done to get a device 5153 order compatible with older (<= 2.4) kernels. 5154 nobfsort Don't sort PCI devices into breadth-first order. 5155 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size) 5156 tuning and use the BIOS-configured MPS defaults. 5157 pcie_bus_safe Set every device's MPS to the largest value 5158 supported by all devices below the root complex. 5159 pcie_bus_perf Set device MPS to the largest allowable MPS 5160 based on its parent bus. Also set MRRS (Max 5161 Read Request Size) to the largest supported 5162 value (no larger than the MPS that the device 5163 or bus can support) for best performance. 5164 pcie_bus_peer2peer Set every device's MPS to 128B, which 5165 every device is guaranteed to support. This 5166 configuration allows peer-to-peer DMA between 5167 any pair of devices, possibly at the cost of 5168 reduced performance. This also guarantees 5169 that hot-added devices will work. 5170 cbiosize=nn[KMG] The fixed amount of bus space which is 5171 reserved for the CardBus bridge's IO window. 5172 The default value is 256 bytes. 5173 cbmemsize=nn[KMG] The fixed amount of bus space which is 5174 reserved for the CardBus bridge's memory 5175 window. The default value is 64 megabytes. 5176 resource_alignment= 5177 Format: 5178 [<order of align>@]<pci_dev>[; ...] 5179 Specifies alignment and device to reassign 5180 aligned memory resources. How to 5181 specify the device is described above. 5182 If <order of align> is not specified, 5183 PAGE_SIZE is used as alignment. 5184 A PCI-PCI bridge can be specified if resource 5185 windows need to be expanded. 5186 To specify the alignment for several 5187 instances of a device, the PCI vendor, 5188 device, subvendor, and subdevice may be 5189 specified, e.g., 12@pci:8086:9c22:103c:198f 5190 for 4096-byte alignment. 5191 ecrc= Enable/disable PCIe ECRC (transaction layer 5192 end-to-end CRC checking). Only effective if 5193 OS has native AER control (either granted by 5194 ACPI _OSC or forced via "pcie_ports=native") 5195 bios: Use BIOS/firmware settings. This is the 5196 the default. 5197 off: Turn ECRC off 5198 on: Turn ECRC on. 5199 hpiosize=nn[KMG] The fixed amount of bus space which is 5200 reserved for hotplug bridge's IO window. 5201 Default size is 256 bytes. 5202 hpmmiosize=nn[KMG] The fixed amount of bus space which is 5203 reserved for hotplug bridge's MMIO window. 5204 Default size is 2 megabytes. 5205 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is 5206 reserved for hotplug bridge's MMIO_PREF window. 5207 Default size is 2 megabytes. 5208 hpmemsize=nn[KMG] The fixed amount of bus space which is 5209 reserved for hotplug bridge's MMIO and 5210 MMIO_PREF window. 5211 Default size is 2 megabytes. 5212 hpbussize=nn The minimum amount of additional bus numbers 5213 reserved for buses below a hotplug bridge. 5214 Default is 1. 5215 realloc= Enable/disable reallocating PCI bridge resources 5216 if allocations done by BIOS are too small to 5217 accommodate resources required by all child 5218 devices. 5219 off: Turn realloc off 5220 on: Turn realloc on 5221 realloc same as realloc=on 5222 noari do not use PCIe ARI. 5223 noats [PCIE, Intel-IOMMU, AMD-IOMMU] 5224 do not use PCIe ATS (and IOMMU device IOTLB). 5225 pcie_scan_all Scan all possible PCIe devices. Otherwise we 5226 only look for one device below a PCIe downstream 5227 port. 5228 big_root_window Try to add a big 64bit memory window to the PCIe 5229 root complex on AMD CPUs. Some GFX hardware 5230 can resize a BAR to allow access to all VRAM. 5231 Adding the window is slightly risky (it may 5232 conflict with unreported devices), so this 5233 taints the kernel. 5234 disable_acs_redir=<pci_dev>[; ...] 5235 Specify one or more PCI devices (in the format 5236 specified above) separated by semicolons. 5237 Each device specified will have the PCI ACS 5238 redirect capabilities forced off which will 5239 allow P2P traffic between devices through 5240 bridges without forcing it upstream. Note: 5241 this removes isolation between devices and 5242 may put more devices in an IOMMU group. 5243 config_acs= 5244 Format: 5245 <ACS flags>@<pci_dev>[; ...] 5246 Specify one or more PCI devices (in the format 5247 specified above) optionally prepended with flags 5248 and separated by semicolons. The respective 5249 capabilities will be enabled, disabled or 5250 unchanged based on what is specified in 5251 flags. 5252 5253 ACS Flags is defined as follows: 5254 bit-0 : ACS Source Validation 5255 bit-1 : ACS Translation Blocking 5256 bit-2 : ACS P2P Request Redirect 5257 bit-3 : ACS P2P Completion Redirect 5258 bit-4 : ACS Upstream Forwarding 5259 bit-5 : ACS P2P Egress Control 5260 bit-6 : ACS Direct Translated P2P 5261 Each bit can be marked as: 5262 '0' – force disabled 5263 '1' – force enabled 5264 'x' – unchanged 5265 For example, 5266 pci=config_acs=10x@pci:0:0 5267 would configure all devices that support 5268 ACS to enable P2P Request Redirect, disable 5269 Translation Blocking, and leave Source 5270 Validation unchanged from whatever power-up 5271 or firmware set it to. 5272 5273 Note: this may remove isolation between devices 5274 and may put more devices in an IOMMU group. 5275 force_floating [S390] Force usage of floating interrupts. 5276 nomio [S390] Do not use MIO instructions. 5277 norid [S390] ignore the RID field and force use of 5278 one PCI domain per PCI function 5279 notph [PCIE] If the PCIE_TPH kernel config parameter 5280 is enabled, this kernel boot option can be used 5281 to disable PCIe TLP Processing Hints support 5282 system-wide. 5283 5284 pcie_aspm= [PCIE] Forcibly enable or ignore PCIe Active State Power 5285 Management. 5286 off Don't touch ASPM configuration at all. Leave any 5287 configuration done by firmware unchanged. 5288 force Enable ASPM even on devices that claim not to support it. 5289 WARNING: Forcing ASPM on may cause system lockups. 5290 5291 pcie_ports= [PCIE] PCIe port services handling: 5292 native Use native PCIe services (PME, AER, DPC, PCIe hotplug) 5293 even if the platform doesn't give the OS permission to 5294 use them. This may cause conflicts if the platform 5295 also tries to use these services. 5296 dpc-native Use native PCIe service for DPC only. May 5297 cause conflicts if firmware uses AER or DPC. 5298 compat Disable native PCIe services (PME, AER, DPC, PCIe 5299 hotplug). 5300 5301 pcie_port_pm= [PCIE] PCIe port power management handling: 5302 off Disable power management of all PCIe ports 5303 force Forcibly enable power management of all PCIe ports 5304 5305 pcie_pme= [PCIE,PM] Native PCIe PME signaling options: 5306 nomsi Do not use MSI for native PCIe PME signaling (this makes 5307 all PCIe root ports use INTx for all services). 5308 5309 pcmv= [HW,PCMCIA] BadgePAD 4 5310 5311 pd_ignore_unused 5312 [PM] 5313 Keep all power-domains already enabled by bootloader on, 5314 even if no driver has claimed them. This is useful 5315 for debug and development, but should not be 5316 needed on a platform with proper driver support. 5317 5318 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at 5319 boot time. 5320 Format: { 0 | 1 } 5321 See arch/parisc/kernel/pdc_chassis.c 5322 5323 percpu_alloc= [MM,EARLY] 5324 Select which percpu first chunk allocator to use. 5325 Currently supported values are "embed" and "page". 5326 Archs may support subset or none of the selections. 5327 See comments in mm/percpu.c for details on each 5328 allocator. This parameter is primarily for debugging 5329 and performance comparison. 5330 5331 pirq= [SMP,APIC] Manual mp-table setup 5332 See Documentation/arch/x86/i386/IO-APIC.rst. 5333 5334 plip= [PPT,NET] Parallel port network link 5335 Format: { parport<nr> | timid | 0 } 5336 See also Documentation/admin-guide/parport.rst. 5337 5338 pmtmr= [X86] Manual setup of pmtmr I/O Port. 5339 Override pmtimer IOPort with a hex value. 5340 e.g. pmtmr=0x508 5341 5342 pmu_override= [PPC] Override the PMU. 5343 This option takes over the PMU facility, so it is no 5344 longer usable by perf. Setting this option starts the 5345 PMU counters by setting MMCR0 to 0 (the FC bit is 5346 cleared). If a number is given, then MMCR1 is set to 5347 that number, otherwise (e.g., 'pmu_override=on'), MMCR1 5348 remains 0. 5349 5350 pm_async= [PM] 5351 Format: off 5352 This parameter sets the initial value of the 5353 /sys/power/pm_async sysfs knob at boot time. 5354 If set to "off", disables asynchronous suspend and 5355 resume of devices during system-wide power transitions. 5356 This can be useful on platforms where device 5357 dependencies are not well-defined, or for debugging 5358 power management issues. Asynchronous operations are 5359 enabled by default. 5360 5361 5362 pm_debug_messages [SUSPEND,KNL] 5363 Enable suspend/resume debug messages during boot up. 5364 5365 pnp.debug=1 [PNP] 5366 Enable PNP debug messages (depends on the 5367 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time 5368 via /sys/module/pnp/parameters/debug. We always show 5369 current resource usage; turning this on also shows 5370 possible settings and some assignment information. 5371 5372 pnpacpi= [ACPI] 5373 { off } 5374 5375 pnpbios= [ISAPNP] 5376 { on | off | curr | res | no-curr | no-res } 5377 5378 pnp_reserve_irq= 5379 [ISAPNP] Exclude IRQs for the autoconfiguration 5380 5381 pnp_reserve_dma= 5382 [ISAPNP] Exclude DMAs for the autoconfiguration 5383 5384 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration 5385 Ranges are in pairs (I/O port base and size). 5386 5387 pnp_reserve_mem= 5388 [ISAPNP] Exclude memory regions for the 5389 autoconfiguration. 5390 Ranges are in pairs (memory base and size). 5391 5392 ports= [IP_VS_FTP] IPVS ftp helper module 5393 Default is 21. 5394 Up to 8 (IP_VS_APP_MAX_PORTS) ports 5395 may be specified. 5396 Format: <port>,<port>.... 5397 5398 possible_cpus= [SMP,S390,X86] 5399 Format: <unsigned int> 5400 Set the number of possible CPUs, overriding the 5401 regular discovery mechanisms (such as ACPI/FW, etc). 5402 5403 powersave=off [PPC] This option disables power saving features. 5404 It specifically disables cpuidle and sets the 5405 platform machine description specific power_save 5406 function to NULL. On Idle the CPU just reduces 5407 execution priority. 5408 5409 ppc_strict_facility_enable 5410 [PPC,ENABLE] This option catches any kernel floating point, 5411 Altivec, VSX and SPE outside of regions specifically 5412 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()). 5413 There is some performance impact when enabling this. 5414 5415 ppc_tm= [PPC,EARLY] 5416 Format: {"off"} 5417 Disable Hardware Transactional Memory 5418 5419 preempt= [KNL] 5420 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC 5421 none - Limited to cond_resched() calls 5422 voluntary - Limited to cond_resched() and might_sleep() calls 5423 full - Any section that isn't explicitly preempt disabled 5424 can be preempted anytime. Tasks will also yield 5425 contended spinlocks (if the critical section isn't 5426 explicitly preempt disabled beyond the lock itself). 5427 lazy - Scheduler controlled. Similar to full but instead 5428 of preempting the task immediately, the task gets 5429 one HZ tick time to yield itself before the 5430 preemption will be forced. One preemption is when the 5431 task returns to user space. 5432 5433 print-fatal-signals= 5434 [KNL] debug: print fatal signals 5435 5436 If enabled, warn about various signal handling 5437 related application anomalies: too many signals, 5438 too many POSIX.1 timers, fatal signals causing a 5439 coredump - etc. 5440 5441 If you hit the warning due to signal overflow, 5442 you might want to try "ulimit -i unlimited". 5443 5444 default: off. 5445 5446 printk.always_kmsg_dump= 5447 Trigger kmsg_dump for cases other than kernel oops or 5448 panics 5449 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5450 default: disabled 5451 5452 printk.console_no_auto_verbose= 5453 Disable console loglevel raise on oops, panic 5454 or lockdep-detected issues (only if lock debug is on). 5455 With an exception to setups with low baudrate on 5456 serial console, keeping this 0 is a good choice 5457 in order to provide more debug information. 5458 Format: <bool> 5459 default: 0 (auto_verbose is enabled) 5460 5461 printk.debug_non_panic_cpus= 5462 Allows storing messages from non-panic CPUs into 5463 the printk log buffer during panic(). They are 5464 flushed to consoles by the panic-CPU on 5465 a best-effort basis. 5466 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5467 Default: disabled 5468 5469 printk.devkmsg={on,off,ratelimit} 5470 Control writing to /dev/kmsg. 5471 on - unlimited logging to /dev/kmsg from userspace 5472 off - logging to /dev/kmsg disabled 5473 ratelimit - ratelimit the logging 5474 Default: ratelimit 5475 5476 printk.time= Show timing data prefixed to each printk message line 5477 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5478 5479 proc_mem.force_override= [KNL] 5480 Format: {always | ptrace | never} 5481 Traditionally /proc/pid/mem allows memory permissions to be 5482 overridden without restrictions. This option may be set to 5483 restrict that. Can be one of: 5484 - 'always': traditional behavior always allows mem overrides. 5485 - 'ptrace': only allow mem overrides for active ptracers. 5486 - 'never': never allow mem overrides. 5487 If not specified, default is the CONFIG_PROC_MEM_* choice. 5488 5489 processor.max_cstate= [HW,ACPI] 5490 Limit processor to maximum C-state 5491 max_cstate=9 overrides any DMI blacklist limit. 5492 5493 processor.nocst [HW,ACPI] 5494 Ignore the _CST method to determine C-states, 5495 instead using the legacy FADT method 5496 5497 profile= [KNL] Enable kernel profiling via /proc/profile 5498 Format: [<profiletype>,]<number> 5499 Param: <profiletype>: "schedule" or "kvm" 5500 [defaults to kernel profiling] 5501 Param: "schedule" - profile schedule points. 5502 Param: "kvm" - profile VM exits. 5503 Param: <number> - step/bucket size as a power of 2 for 5504 statistical time based profiling. 5505 5506 prot_virt= [S390] enable hosting protected virtual machines 5507 isolated from the hypervisor (if hardware supports 5508 that). If enabled, the default kernel base address 5509 might be overridden even when Kernel Address Space 5510 Layout Randomization is disabled. 5511 Format: <bool> 5512 5513 psi= [KNL] Enable or disable pressure stall information 5514 tracking. 5515 Format: <bool> 5516 5517 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to 5518 probe for; one of (bare|imps|exps|lifebook|any). 5519 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports 5520 per second. 5521 psmouse.resetafter= [HW,MOUSE] 5522 Try to reset the device after so many bad packets 5523 (0 = never). 5524 psmouse.resolution= 5525 [HW,MOUSE] Set desired mouse resolution, in dpi. 5526 psmouse.smartscroll= 5527 [HW,MOUSE] Controls Logitech smartscroll autorepeat. 5528 0 = disabled, 1 = enabled (default). 5529 5530 pstore.backend= Specify the name of the pstore backend to use 5531 5532 pti= [X86-64] Control Page Table Isolation of user and 5533 kernel address spaces. Disabling this feature 5534 removes hardening, but improves performance of 5535 system calls and interrupts. 5536 5537 on - unconditionally enable 5538 off - unconditionally disable 5539 auto - kernel detects whether your CPU model is 5540 vulnerable to issues that PTI mitigates 5541 5542 Not specifying this option is equivalent to pti=auto. 5543 5544 pty.legacy_count= 5545 [KNL] Number of legacy pty's. Overwrites compiled-in 5546 default number. 5547 5548 quiet [KNL,EARLY] Disable most log messages 5549 5550 r128= [HW,DRM] 5551 5552 radix_hcall_invalidate=on [PPC/PSERIES] 5553 Disable RADIX GTSE feature and use hcall for TLB 5554 invalidate. 5555 5556 raid= [HW,RAID] 5557 See Documentation/admin-guide/md.rst. 5558 5559 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes 5560 See Documentation/admin-guide/blockdev/ramdisk.rst. 5561 5562 ramdisk_start= [Deprecated,RAM] RAM disk image start address 5563 5564 random.trust_cpu=off 5565 [KNL,EARLY] Disable trusting the use of the CPU's 5566 random number generator (if available) to 5567 initialize the kernel's RNG. 5568 5569 random.trust_bootloader=off 5570 [KNL,EARLY] Disable trusting the use of the a seed 5571 passed by the bootloader (if available) to 5572 initialize the kernel's RNG. 5573 5574 randomize_kstack_offset= 5575 [KNL,EARLY] Enable or disable kernel stack offset 5576 randomization, which provides roughly 5 bits of 5577 entropy, frustrating memory corruption attacks 5578 that depend on stack address determinism or 5579 cross-syscall address exposures. This is only 5580 available on architectures that have defined 5581 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET. 5582 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5583 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT. 5584 5585 ras=option[,option,...] [KNL] RAS-specific options 5586 5587 cec_disable [X86] 5588 Disable the Correctable Errors Collector, 5589 see CONFIG_RAS_CEC help text. 5590 5591 rcu_nocbs[=cpu-list] 5592 [KNL] The optional argument is a cpu list, 5593 as described above. 5594 5595 In kernels built with CONFIG_RCU_NOCB_CPU=y, 5596 enable the no-callback CPU mode, which prevents 5597 such CPUs' callbacks from being invoked in 5598 softirq context. Invocation of such CPUs' RCU 5599 callbacks will instead be offloaded to "rcuox/N" 5600 kthreads created for that purpose, where "x" is 5601 "p" for RCU-preempt, "s" for RCU-sched, and "g" 5602 for the kthreads that mediate grace periods; and 5603 "N" is the CPU number. This reduces OS jitter on 5604 the offloaded CPUs, which can be useful for HPC 5605 and real-time workloads. It can also improve 5606 energy efficiency for asymmetric multiprocessors. 5607 5608 If a cpulist is passed as an argument, the specified 5609 list of CPUs is set to no-callback mode from boot. 5610 5611 Otherwise, if the '=' sign and the cpulist 5612 arguments are omitted, no CPU will be set to 5613 no-callback mode from boot but the mode may be 5614 toggled at runtime via cpusets. 5615 5616 Note that this argument takes precedence over 5617 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 5618 5619 rcu_nocb_poll [KNL] 5620 Rather than requiring that offloaded CPUs 5621 (specified by rcu_nocbs= above) explicitly 5622 awaken the corresponding "rcuoN" kthreads, 5623 make these kthreads poll for callbacks. 5624 This improves the real-time response for the 5625 offloaded CPUs by relieving them of the need to 5626 wake up the corresponding kthread, but degrades 5627 energy efficiency by requiring that the kthreads 5628 periodically wake up to do the polling. 5629 5630 rcutree.blimit= [KNL] 5631 Set maximum number of finished RCU callbacks to 5632 process in one batch. 5633 5634 rcutree.csd_lock_suppress_rcu_stall= [KNL] 5635 Do only a one-line RCU CPU stall warning when 5636 there is an ongoing too-long CSD-lock wait. 5637 5638 rcutree.do_rcu_barrier= [KNL] 5639 Request a call to rcu_barrier(). This is 5640 throttled so that userspace tests can safely 5641 hammer on the sysfs variable if they so choose. 5642 If triggered before the RCU grace-period machinery 5643 is fully active, this will error out with EAGAIN. 5644 5645 rcutree.dump_tree= [KNL] 5646 Dump the structure of the rcu_node combining tree 5647 out at early boot. This is used for diagnostic 5648 purposes, to verify correct tree setup. 5649 5650 rcutree.gp_cleanup_delay= [KNL] 5651 Set the number of jiffies to delay each step of 5652 RCU grace-period cleanup. 5653 5654 rcutree.gp_init_delay= [KNL] 5655 Set the number of jiffies to delay each step of 5656 RCU grace-period initialization. 5657 5658 rcutree.gp_preinit_delay= [KNL] 5659 Set the number of jiffies to delay each step of 5660 RCU grace-period pre-initialization, that is, 5661 the propagation of recent CPU-hotplug changes up 5662 the rcu_node combining tree. 5663 5664 rcutree.jiffies_till_first_fqs= [KNL] 5665 Set delay from grace-period initialization to 5666 first attempt to force quiescent states. 5667 Units are jiffies, minimum value is zero, 5668 and maximum value is HZ. 5669 5670 rcutree.jiffies_till_next_fqs= [KNL] 5671 Set delay between subsequent attempts to force 5672 quiescent states. Units are jiffies, minimum 5673 value is one, and maximum value is HZ. 5674 5675 rcutree.jiffies_till_sched_qs= [KNL] 5676 Set required age in jiffies for a 5677 given grace period before RCU starts 5678 soliciting quiescent-state help from 5679 rcu_note_context_switch() and cond_resched(). 5680 If not specified, the kernel will calculate 5681 a value based on the most recent settings 5682 of rcutree.jiffies_till_first_fqs 5683 and rcutree.jiffies_till_next_fqs. 5684 This calculated value may be viewed in 5685 rcutree.jiffies_to_sched_qs. Any attempt to set 5686 rcutree.jiffies_to_sched_qs will be cheerfully 5687 overwritten. 5688 5689 rcutree.kthread_prio= [KNL,BOOT] 5690 Set the SCHED_FIFO priority of the RCU per-CPU 5691 kthreads (rcuc/N). This value is also used for 5692 the priority of the RCU boost threads (rcub/N) 5693 and for the RCU grace-period kthreads (rcu_bh, 5694 rcu_preempt, and rcu_sched). If RCU_BOOST is 5695 set, valid values are 1-99 and the default is 1 5696 (the least-favored priority). Otherwise, when 5697 RCU_BOOST is not set, valid values are 0-99 and 5698 the default is zero (non-realtime operation). 5699 When RCU_NOCB_CPU is set, also adjust the 5700 priority of NOCB callback kthreads. 5701 5702 rcutree.nocb_nobypass_lim_per_jiffy= [KNL] 5703 On callback-offloaded (rcu_nocbs) CPUs, 5704 RCU reduces the lock contention that would 5705 otherwise be caused by callback floods through 5706 use of the ->nocb_bypass list. However, in the 5707 common non-flooded case, RCU queues directly to 5708 the main ->cblist in order to avoid the extra 5709 overhead of the ->nocb_bypass list and its lock. 5710 But if there are too many callbacks queued during 5711 a single jiffy, RCU pre-queues the callbacks into 5712 the ->nocb_bypass queue. The definition of "too 5713 many" is supplied by this kernel boot parameter. 5714 5715 rcutree.nohz_full_patience_delay= [KNL] 5716 On callback-offloaded (rcu_nocbs) CPUs, avoid 5717 disturbing RCU unless the grace period has 5718 reached the specified age in milliseconds. 5719 Defaults to zero. Large values will be capped 5720 at five seconds. All values will be rounded down 5721 to the nearest value representable by jiffies. 5722 5723 rcutree.qhimark= [KNL] 5724 Set threshold of queued RCU callbacks beyond which 5725 batch limiting is disabled. 5726 5727 rcutree.qlowmark= [KNL] 5728 Set threshold of queued RCU callbacks below which 5729 batch limiting is re-enabled. 5730 5731 rcutree.qovld= [KNL] 5732 Set threshold of queued RCU callbacks beyond which 5733 RCU's force-quiescent-state scan will aggressively 5734 enlist help from cond_resched() and sched IPIs to 5735 help CPUs more quickly reach quiescent states. 5736 Set to less than zero to make this be set based 5737 on rcutree.qhimark at boot time and to zero to 5738 disable more aggressive help enlistment. 5739 5740 rcutree.rcu_delay_page_cache_fill_msec= [KNL] 5741 Set the page-cache refill delay (in milliseconds) 5742 in response to low-memory conditions. The range 5743 of permitted values is in the range 0:100000. 5744 5745 rcutree.rcu_divisor= [KNL] 5746 Set the shift-right count to use to compute 5747 the callback-invocation batch limit bl from 5748 the number of callbacks queued on this CPU. 5749 The result will be bounded below by the value of 5750 the rcutree.blimit kernel parameter. Every bl 5751 callbacks, the softirq handler will exit in 5752 order to allow the CPU to do other work. 5753 5754 Please note that this callback-invocation batch 5755 limit applies only to non-offloaded callback 5756 invocation. Offloaded callbacks are instead 5757 invoked in the context of an rcuoc kthread, which 5758 scheduler will preempt as it does any other task. 5759 5760 rcutree.rcu_fanout_exact= [KNL] 5761 Disable autobalancing of the rcu_node combining 5762 tree. This is used by rcutorture, and might 5763 possibly be useful for architectures having high 5764 cache-to-cache transfer latencies. 5765 5766 rcutree.rcu_fanout_leaf= [KNL] 5767 Change the number of CPUs assigned to each 5768 leaf rcu_node structure. Useful for very 5769 large systems, which will choose the value 64, 5770 and for NUMA systems with large remote-access 5771 latencies, which will choose a value aligned 5772 with the appropriate hardware boundaries. 5773 5774 rcutree.rcu_min_cached_objs= [KNL] 5775 Minimum number of objects which are cached and 5776 maintained per one CPU. Object size is equal 5777 to PAGE_SIZE. The cache allows to reduce the 5778 pressure to page allocator, also it makes the 5779 whole algorithm to behave better in low memory 5780 condition. 5781 5782 rcutree.rcu_nocb_gp_stride= [KNL] 5783 Set the number of NOCB callback kthreads in 5784 each group, which defaults to the square root 5785 of the number of CPUs. Larger numbers reduce 5786 the wakeup overhead on the global grace-period 5787 kthread, but increases that same overhead on 5788 each group's NOCB grace-period kthread. 5789 5790 rcutree.rcu_kick_kthreads= [KNL] 5791 Cause the grace-period kthread to get an extra 5792 wake_up() if it sleeps three times longer than 5793 it should at force-quiescent-state time. 5794 This wake_up() will be accompanied by a 5795 WARN_ONCE() splat and an ftrace_dump(). 5796 5797 rcutree.rcu_resched_ns= [KNL] 5798 Limit the time spend invoking a batch of RCU 5799 callbacks to the specified number of nanoseconds. 5800 By default, this limit is checked only once 5801 every 32 callbacks in order to limit the pain 5802 inflicted by local_clock() overhead. 5803 5804 rcutree.rcu_unlock_delay= [KNL] 5805 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels, 5806 this specifies an rcu_read_unlock()-time delay 5807 in microseconds. This defaults to zero. 5808 Larger delays increase the probability of 5809 catching RCU pointer leaks, that is, buggy use 5810 of RCU-protected pointers after the relevant 5811 rcu_read_unlock() has completed. 5812 5813 rcutree.sysrq_rcu= [KNL] 5814 Commandeer a sysrq key to dump out Tree RCU's 5815 rcu_node tree with an eye towards determining 5816 why a new grace period has not yet started. 5817 5818 rcutree.use_softirq= [KNL] 5819 If set to zero, move all RCU_SOFTIRQ processing to 5820 per-CPU rcuc kthreads. Defaults to a non-zero 5821 value, meaning that RCU_SOFTIRQ is used by default. 5822 Specify rcutree.use_softirq=0 to use rcuc kthreads. 5823 5824 But note that CONFIG_PREEMPT_RT=y kernels disable 5825 this kernel boot parameter, forcibly setting it 5826 to zero. 5827 5828 rcutree.enable_rcu_lazy= [KNL] 5829 To save power, batch RCU callbacks and flush after 5830 delay, memory pressure or callback list growing too 5831 big. 5832 5833 rcutree.rcu_normal_wake_from_gp= [KNL] 5834 Reduces a latency of synchronize_rcu() call. This approach 5835 maintains its own track of synchronize_rcu() callers, so it 5836 does not interact with regular callbacks because it does not 5837 use a call_rcu[_hurry]() path. Please note, this is for a 5838 normal grace period. 5839 5840 How to disable it: 5841 5842 echo 0 > /sys/module/rcutree/parameters/rcu_normal_wake_from_gp 5843 or pass a boot parameter "rcutree.rcu_normal_wake_from_gp=0" 5844 5845 Default is 1 if it is not explicitly disabled by the boot parameter 5846 passing 0. 5847 5848 rcuscale.gp_async= [KNL] 5849 Measure performance of asynchronous 5850 grace-period primitives such as call_rcu(). 5851 5852 rcuscale.gp_async_max= [KNL] 5853 Specify the maximum number of outstanding 5854 callbacks per writer thread. When a writer 5855 thread exceeds this limit, it invokes the 5856 corresponding flavor of rcu_barrier() to allow 5857 previously posted callbacks to drain. 5858 5859 rcuscale.gp_exp= [KNL] 5860 Measure performance of expedited synchronous 5861 grace-period primitives. 5862 5863 rcuscale.holdoff= [KNL] 5864 Set test-start holdoff period. The purpose of 5865 this parameter is to delay the start of the 5866 test until boot completes in order to avoid 5867 interference. 5868 5869 rcuscale.kfree_by_call_rcu= [KNL] 5870 In kernels built with CONFIG_RCU_LAZY=y, test 5871 call_rcu() instead of kfree_rcu(). 5872 5873 rcuscale.kfree_mult= [KNL] 5874 Instead of allocating an object of size kfree_obj, 5875 allocate one of kfree_mult * sizeof(kfree_obj). 5876 Defaults to 1. 5877 5878 rcuscale.kfree_rcu_test= [KNL] 5879 Set to measure performance of kfree_rcu() flooding. 5880 5881 rcuscale.kfree_rcu_test_double= [KNL] 5882 Test the double-argument variant of kfree_rcu(). 5883 If this parameter has the same value as 5884 rcuscale.kfree_rcu_test_single, both the single- 5885 and double-argument variants are tested. 5886 5887 rcuscale.kfree_rcu_test_single= [KNL] 5888 Test the single-argument variant of kfree_rcu(). 5889 If this parameter has the same value as 5890 rcuscale.kfree_rcu_test_double, both the single- 5891 and double-argument variants are tested. 5892 5893 rcuscale.kfree_nthreads= [KNL] 5894 The number of threads running loops of kfree_rcu(). 5895 5896 rcuscale.kfree_alloc_num= [KNL] 5897 Number of allocations and frees done in an iteration. 5898 5899 rcuscale.kfree_loops= [KNL] 5900 Number of loops doing rcuscale.kfree_alloc_num number 5901 of allocations and frees. 5902 5903 rcuscale.minruntime= [KNL] 5904 Set the minimum test run time in seconds. This 5905 does not affect the data-collection interval, 5906 but instead allows better measurement of things 5907 like CPU consumption. 5908 5909 rcuscale.nreaders= [KNL] 5910 Set number of RCU readers. The value -1 selects 5911 N, where N is the number of CPUs. A value 5912 "n" less than -1 selects N-n+1, where N is again 5913 the number of CPUs. For example, -2 selects N 5914 (the number of CPUs), -3 selects N+1, and so on. 5915 A value of "n" less than or equal to -N selects 5916 a single reader. 5917 5918 rcuscale.nwriters= [KNL] 5919 Set number of RCU writers. The values operate 5920 the same as for rcuscale.nreaders. 5921 N, where N is the number of CPUs 5922 5923 rcuscale.scale_type= [KNL] 5924 Specify the RCU implementation to test. 5925 5926 rcuscale.shutdown= [KNL] 5927 Shut the system down after performance tests 5928 complete. This is useful for hands-off automated 5929 testing. 5930 5931 rcuscale.verbose= [KNL] 5932 Enable additional printk() statements. 5933 5934 rcuscale.writer_holdoff= [KNL] 5935 Write-side holdoff between grace periods, 5936 in microseconds. The default of zero says 5937 no holdoff. 5938 5939 rcuscale.writer_holdoff_jiffies= [KNL] 5940 Additional write-side holdoff between grace 5941 periods, but in jiffies. The default of zero 5942 says no holdoff. 5943 5944 rcutorture.fqs_duration= [KNL] 5945 Set duration of force_quiescent_state bursts 5946 in microseconds. 5947 5948 rcutorture.fqs_holdoff= [KNL] 5949 Set holdoff time within force_quiescent_state bursts 5950 in microseconds. 5951 5952 rcutorture.fqs_stutter= [KNL] 5953 Set wait time between force_quiescent_state bursts 5954 in seconds. 5955 5956 rcutorture.fwd_progress= [KNL] 5957 Specifies the number of kthreads to be used 5958 for RCU grace-period forward-progress testing 5959 for the types of RCU supporting this notion. 5960 Defaults to 1 kthread, values less than zero or 5961 greater than the number of CPUs cause the number 5962 of CPUs to be used. 5963 5964 rcutorture.fwd_progress_div= [KNL] 5965 Specify the fraction of a CPU-stall-warning 5966 period to do tight-loop forward-progress testing. 5967 5968 rcutorture.fwd_progress_holdoff= [KNL] 5969 Number of seconds to wait between successive 5970 forward-progress tests. 5971 5972 rcutorture.fwd_progress_need_resched= [KNL] 5973 Enclose cond_resched() calls within checks for 5974 need_resched() during tight-loop forward-progress 5975 testing. 5976 5977 rcutorture.gp_cond= [KNL] 5978 Use conditional/asynchronous update-side 5979 normal-grace-period primitives, if available. 5980 5981 rcutorture.gp_cond_exp= [KNL] 5982 Use conditional/asynchronous update-side 5983 expedited-grace-period primitives, if available. 5984 5985 rcutorture.gp_cond_full= [KNL] 5986 Use conditional/asynchronous update-side 5987 normal-grace-period primitives that also take 5988 concurrent expedited grace periods into account, 5989 if available. 5990 5991 rcutorture.gp_cond_exp_full= [KNL] 5992 Use conditional/asynchronous update-side 5993 expedited-grace-period primitives that also take 5994 concurrent normal grace periods into account, 5995 if available. 5996 5997 rcutorture.gp_cond_wi= [KNL] 5998 Nominal wait interval for normal conditional 5999 grace periods (specified by rcutorture's 6000 gp_cond and gp_cond_full module parameters), 6001 in microseconds. The actual wait interval will 6002 be randomly selected to nanosecond granularity up 6003 to this wait interval. Defaults to 16 jiffies, 6004 for example, 16,000 microseconds on a system 6005 with HZ=1000. 6006 6007 rcutorture.gp_cond_wi_exp= [KNL] 6008 Nominal wait interval for expedited conditional 6009 grace periods (specified by rcutorture's 6010 gp_cond_exp and gp_cond_exp_full module 6011 parameters), in microseconds. The actual wait 6012 interval will be randomly selected to nanosecond 6013 granularity up to this wait interval. Defaults to 6014 128 microseconds. 6015 6016 rcutorture.gp_exp= [KNL] 6017 Use expedited update-side primitives, if available. 6018 6019 rcutorture.gp_normal= [KNL] 6020 Use normal (non-expedited) asynchronous 6021 update-side primitives, if available. 6022 6023 rcutorture.gp_poll= [KNL] 6024 Use polled update-side normal-grace-period 6025 primitives, if available. 6026 6027 rcutorture.gp_poll_exp= [KNL] 6028 Use polled update-side expedited-grace-period 6029 primitives, if available. 6030 6031 rcutorture.gp_poll_full= [KNL] 6032 Use polled update-side normal-grace-period 6033 primitives that also take concurrent expedited 6034 grace periods into account, if available. 6035 6036 rcutorture.gp_poll_exp_full= [KNL] 6037 Use polled update-side expedited-grace-period 6038 primitives that also take concurrent normal 6039 grace periods into account, if available. 6040 6041 rcutorture.gp_poll_wi= [KNL] 6042 Nominal wait interval for normal conditional 6043 grace periods (specified by rcutorture's 6044 gp_poll and gp_poll_full module parameters), 6045 in microseconds. The actual wait interval will 6046 be randomly selected to nanosecond granularity up 6047 to this wait interval. Defaults to 16 jiffies, 6048 for example, 16,000 microseconds on a system 6049 with HZ=1000. 6050 6051 rcutorture.gp_poll_wi_exp= [KNL] 6052 Nominal wait interval for expedited conditional 6053 grace periods (specified by rcutorture's 6054 gp_poll_exp and gp_poll_exp_full module 6055 parameters), in microseconds. The actual wait 6056 interval will be randomly selected to nanosecond 6057 granularity up to this wait interval. Defaults to 6058 128 microseconds. 6059 6060 rcutorture.gp_sync= [KNL] 6061 Use normal (non-expedited) synchronous 6062 update-side primitives, if available. If all 6063 of rcutorture.gp_cond=, rcutorture.gp_exp=, 6064 rcutorture.gp_normal=, and rcutorture.gp_sync= 6065 are zero, rcutorture acts as if is interpreted 6066 they are all non-zero. 6067 6068 rcutorture.gpwrap_lag= [KNL] 6069 Enable grace-period wrap lag testing. Setting 6070 to false prevents the gpwrap lag test from 6071 running. Default is true. 6072 6073 rcutorture.gpwrap_lag_gps= [KNL] 6074 Set the value for grace-period wrap lag during 6075 active lag testing periods. This controls how many 6076 grace periods differences we tolerate between 6077 rdp and rnp's gp_seq before setting overflow flag. 6078 The default is always set to 8. 6079 6080 rcutorture.gpwrap_lag_cycle_mins= [KNL] 6081 Set the total cycle duration for gpwrap lag 6082 testing in minutes. This is the total time for 6083 one complete cycle of active and inactive 6084 testing periods. Default is 30 minutes. 6085 6086 rcutorture.gpwrap_lag_active_mins= [KNL] 6087 Set the duration for which gpwrap lag is active 6088 within each cycle, in minutes. During this time, 6089 the grace-period wrap lag will be set to the 6090 value specified by gpwrap_lag_gps. Default is 6091 5 minutes. 6092 6093 rcutorture.irqreader= [KNL] 6094 Run RCU readers from irq handlers, or, more 6095 accurately, from a timer handler. Not all RCU 6096 flavors take kindly to this sort of thing. 6097 6098 rcutorture.leakpointer= [KNL] 6099 Leak an RCU-protected pointer out of the reader. 6100 This can of course result in splats, and is 6101 intended to test the ability of things like 6102 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect 6103 such leaks. 6104 6105 rcutorture.n_barrier_cbs= [KNL] 6106 Set callbacks/threads for rcu_barrier() testing. 6107 6108 rcutorture.nfakewriters= [KNL] 6109 Set number of concurrent RCU writers. These just 6110 stress RCU, they don't participate in the actual 6111 test, hence the "fake". 6112 6113 rcutorture.nocbs_nthreads= [KNL] 6114 Set number of RCU callback-offload togglers. 6115 Zero (the default) disables toggling. 6116 6117 rcutorture.nocbs_toggle= [KNL] 6118 Set the delay in milliseconds between successive 6119 callback-offload toggling attempts. 6120 6121 rcutorture.nreaders= [KNL] 6122 Set number of RCU readers. The value -1 selects 6123 N-1, where N is the number of CPUs. A value 6124 "n" less than -1 selects N-n-2, where N is again 6125 the number of CPUs. For example, -2 selects N 6126 (the number of CPUs), -3 selects N+1, and so on. 6127 6128 rcutorture.object_debug= [KNL] 6129 Enable debug-object double-call_rcu() testing. 6130 6131 rcutorture.onoff_holdoff= [KNL] 6132 Set time (s) after boot for CPU-hotplug testing. 6133 6134 rcutorture.onoff_interval= [KNL] 6135 Set time (jiffies) between CPU-hotplug operations, 6136 or zero to disable CPU-hotplug testing. 6137 6138 rcutorture.preempt_duration= [KNL] 6139 Set duration (in milliseconds) of preemptions 6140 by a high-priority FIFO real-time task. Set to 6141 zero (the default) to disable. The CPUs to 6142 preempt are selected randomly from the set that 6143 are online at a given point in time. Races with 6144 CPUs going offline are ignored, with that attempt 6145 at preemption skipped. 6146 6147 rcutorture.preempt_interval= [KNL] 6148 Set interval (in milliseconds, defaulting to one 6149 second) between preemptions by a high-priority 6150 FIFO real-time task. This delay is mediated 6151 by an hrtimer and is further fuzzed to avoid 6152 inadvertent synchronizations. 6153 6154 rcutorture.read_exit_burst= [KNL] 6155 The number of times in a given read-then-exit 6156 episode that a set of read-then-exit kthreads 6157 is spawned. 6158 6159 rcutorture.read_exit_delay= [KNL] 6160 The delay, in seconds, between successive 6161 read-then-exit testing episodes. 6162 6163 rcutorture.reader_flavor= [KNL] 6164 A bit mask indicating which readers to use. 6165 If there is more than one bit set, the readers 6166 are entered from low-order bit up, and are 6167 exited in the opposite order. For SRCU, the 6168 0x1 bit is normal readers, 0x2 NMI-safe readers, 6169 and 0x4 light-weight readers. 6170 6171 rcutorture.shuffle_interval= [KNL] 6172 Set task-shuffle interval (s). Shuffling tasks 6173 allows some CPUs to go into dyntick-idle mode 6174 during the rcutorture test. 6175 6176 rcutorture.shutdown_secs= [KNL] 6177 Set time (s) after boot system shutdown. This 6178 is useful for hands-off automated testing. 6179 6180 rcutorture.stall_cpu= [KNL] 6181 Duration of CPU stall (s) to test RCU CPU stall 6182 warnings, zero to disable. 6183 6184 rcutorture.stall_cpu_block= [KNL] 6185 Sleep while stalling if set. This will result 6186 in warnings from preemptible RCU in addition to 6187 any other stall-related activity. Note that 6188 in kernels built with CONFIG_PREEMPTION=n and 6189 CONFIG_PREEMPT_COUNT=y, this parameter will 6190 cause the CPU to pass through a quiescent state. 6191 Given CONFIG_PREEMPTION=n, this will suppress 6192 RCU CPU stall warnings, but will instead result 6193 in scheduling-while-atomic splats. 6194 6195 Use of this module parameter results in splats. 6196 6197 6198 rcutorture.stall_cpu_holdoff= [KNL] 6199 Time to wait (s) after boot before inducing stall. 6200 6201 rcutorture.stall_cpu_irqsoff= [KNL] 6202 Disable interrupts while stalling if set, but only 6203 on the first stall in the set. 6204 6205 rcutorture.stall_cpu_repeat= [KNL] 6206 Number of times to repeat the stall sequence, 6207 so that rcutorture.stall_cpu_repeat=3 will result 6208 in four stall sequences. 6209 6210 rcutorture.stall_gp_kthread= [KNL] 6211 Duration (s) of forced sleep within RCU 6212 grace-period kthread to test RCU CPU stall 6213 warnings, zero to disable. If both stall_cpu 6214 and stall_gp_kthread are specified, the 6215 kthread is starved first, then the CPU. 6216 6217 rcutorture.stat_interval= [KNL] 6218 Time (s) between statistics printk()s. 6219 6220 rcutorture.stutter= [KNL] 6221 Time (s) to stutter testing, for example, specifying 6222 five seconds causes the test to run for five seconds, 6223 wait for five seconds, and so on. This tests RCU's 6224 ability to transition abruptly to and from idle. 6225 6226 rcutorture.test_boost= [KNL] 6227 Test RCU priority boosting? 0=no, 1=maybe, 2=yes. 6228 "Maybe" means test if the RCU implementation 6229 under test support RCU priority boosting. 6230 6231 rcutorture.test_boost_duration= [KNL] 6232 Duration (s) of each individual boost test. 6233 6234 rcutorture.test_boost_holdoff= [KNL] 6235 Holdoff time (s) from start of test to the start 6236 of RCU priority-boost testing. Defaults to zero, 6237 that is, no holdoff. 6238 6239 rcutorture.test_boost_interval= [KNL] 6240 Interval (s) between each boost test. 6241 6242 rcutorture.test_no_idle_hz= [KNL] 6243 Test RCU's dyntick-idle handling. See also the 6244 rcutorture.shuffle_interval parameter. 6245 6246 rcutorture.torture_type= [KNL] 6247 Specify the RCU implementation to test. 6248 6249 rcutorture.verbose= [KNL] 6250 Enable additional printk() statements. 6251 6252 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL] 6253 Dump ftrace buffer after reporting RCU CPU 6254 stall warning. 6255 6256 rcupdate.rcu_cpu_stall_notifiers= [KNL] 6257 Provide RCU CPU stall notifiers, but see the 6258 warnings in the RCU_CPU_STALL_NOTIFIER Kconfig 6259 option's help text. TL;DR: You almost certainly 6260 do not want rcupdate.rcu_cpu_stall_notifiers. 6261 6262 rcupdate.rcu_cpu_stall_suppress= [KNL] 6263 Suppress RCU CPU stall warning messages. 6264 6265 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL] 6266 Suppress RCU CPU stall warning messages and 6267 rcutorture writer stall warnings that occur 6268 during early boot, that is, during the time 6269 before the init task is spawned. 6270 6271 rcupdate.rcu_cpu_stall_timeout= [KNL] 6272 Set timeout for RCU CPU stall warning messages. 6273 The value is in seconds and the maximum allowed 6274 value is 300 seconds. 6275 6276 rcupdate.rcu_exp_cpu_stall_timeout= [KNL] 6277 Set timeout for expedited RCU CPU stall warning 6278 messages. The value is in milliseconds 6279 and the maximum allowed value is 21000 6280 milliseconds. Please note that this value is 6281 adjusted to an arch timer tick resolution. 6282 Setting this to zero causes the value from 6283 rcupdate.rcu_cpu_stall_timeout to be used (after 6284 conversion from seconds to milliseconds). 6285 6286 rcupdate.rcu_cpu_stall_cputime= [KNL] 6287 Provide statistics on the cputime and count of 6288 interrupts and tasks during the sampling period. For 6289 multiple continuous RCU stalls, all sampling periods 6290 begin at half of the first RCU stall timeout. 6291 6292 rcupdate.rcu_exp_stall_task_details= [KNL] 6293 Print stack dumps of any tasks blocking the 6294 current expedited RCU grace period during an 6295 expedited RCU CPU stall warning. 6296 6297 rcupdate.rcu_expedited= [KNL] 6298 Use expedited grace-period primitives, for 6299 example, synchronize_rcu_expedited() instead 6300 of synchronize_rcu(). This reduces latency, 6301 but can increase CPU utilization, degrade 6302 real-time latency, and degrade energy efficiency. 6303 No effect on CONFIG_TINY_RCU kernels. 6304 6305 rcupdate.rcu_normal= [KNL] 6306 Use only normal grace-period primitives, 6307 for example, synchronize_rcu() instead of 6308 synchronize_rcu_expedited(). This improves 6309 real-time latency, CPU utilization, and 6310 energy efficiency, but can expose users to 6311 increased grace-period latency. This parameter 6312 overrides rcupdate.rcu_expedited. No effect on 6313 CONFIG_TINY_RCU kernels. 6314 6315 rcupdate.rcu_normal_after_boot= [KNL] 6316 Once boot has completed (that is, after 6317 rcu_end_inkernel_boot() has been invoked), use 6318 only normal grace-period primitives. No effect 6319 on CONFIG_TINY_RCU kernels. 6320 6321 But note that CONFIG_PREEMPT_RT=y kernels enables 6322 this kernel boot parameter, forcibly setting 6323 it to the value one, that is, converting any 6324 post-boot attempt at an expedited RCU grace 6325 period to instead use normal non-expedited 6326 grace-period processing. 6327 6328 rcupdate.rcu_task_collapse_lim= [KNL] 6329 Set the maximum number of callbacks present 6330 at the beginning of a grace period that allows 6331 the RCU Tasks flavors to collapse back to using 6332 a single callback queue. This switching only 6333 occurs when rcupdate.rcu_task_enqueue_lim is 6334 set to the default value of -1. 6335 6336 rcupdate.rcu_task_contend_lim= [KNL] 6337 Set the minimum number of callback-queuing-time 6338 lock-contention events per jiffy required to 6339 cause the RCU Tasks flavors to switch to per-CPU 6340 callback queuing. This switching only occurs 6341 when rcupdate.rcu_task_enqueue_lim is set to 6342 the default value of -1. 6343 6344 rcupdate.rcu_task_enqueue_lim= [KNL] 6345 Set the number of callback queues to use for the 6346 RCU Tasks family of RCU flavors. The default 6347 of -1 allows this to be automatically (and 6348 dynamically) adjusted. This parameter is intended 6349 for use in testing. 6350 6351 rcupdate.rcu_task_lazy_lim= [KNL] 6352 Number of callbacks on a given CPU that will 6353 cancel laziness on that CPU. Use -1 to disable 6354 cancellation of laziness, but be advised that 6355 doing so increases the danger of OOM due to 6356 callback flooding. 6357 6358 rcupdate.rcu_task_stall_info= [KNL] 6359 Set initial timeout in jiffies for RCU task stall 6360 informational messages, which give some indication 6361 of the problem for those not patient enough to 6362 wait for ten minutes. Informational messages are 6363 only printed prior to the stall-warning message 6364 for a given grace period. Disable with a value 6365 less than or equal to zero. Defaults to ten 6366 seconds. A change in value does not take effect 6367 until the beginning of the next grace period. 6368 6369 rcupdate.rcu_task_stall_info_mult= [KNL] 6370 Multiplier for time interval between successive 6371 RCU task stall informational messages for a given 6372 RCU tasks grace period. This value is clamped 6373 to one through ten, inclusive. It defaults to 6374 the value three, so that the first informational 6375 message is printed 10 seconds into the grace 6376 period, the second at 40 seconds, the third at 6377 160 seconds, and then the stall warning at 600 6378 seconds would prevent a fourth at 640 seconds. 6379 6380 rcupdate.rcu_task_stall_timeout= [KNL] 6381 Set timeout in jiffies for RCU task stall 6382 warning messages. Disable with a value less 6383 than or equal to zero. Defaults to ten minutes. 6384 A change in value does not take effect until 6385 the beginning of the next grace period. 6386 6387 rcupdate.rcu_tasks_lazy_ms= [KNL] 6388 Set timeout in milliseconds RCU Tasks asynchronous 6389 callback batching for call_rcu_tasks(). 6390 A negative value will take the default. A value 6391 of zero will disable batching. Batching is 6392 always disabled for synchronize_rcu_tasks(). 6393 6394 rcupdate.rcu_self_test= [KNL] 6395 Run the RCU early boot self tests 6396 6397 rdinit= [KNL] 6398 Format: <full_path> 6399 Run specified binary instead of /init from the ramdisk, 6400 used for early userspace startup. See initrd. 6401 6402 rdrand= [X86,EARLY] 6403 force - Override the decision by the kernel to hide the 6404 advertisement of RDRAND support (this affects 6405 certain AMD processors because of buggy BIOS 6406 support, specifically around the suspend/resume 6407 path). 6408 6409 rdt= [HW,X86,RDT] 6410 Turn on/off individual RDT features. List is: 6411 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp, 6412 mba, smba, bmec, abmc, sdciae, energy[:guid], 6413 perf[:guid]. 6414 E.g. to turn on cmt and turn off mba use: 6415 rdt=cmt,!mba 6416 To turn off all energy telemetry monitoring and ensure that 6417 perf telemetry monitoring associated with guid 0x12345 6418 is enabled use: 6419 rdt=!energy,perf:0x12345 6420 6421 reboot= [KNL] 6422 Format (x86 or x86_64): 6423 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \ 6424 [[,]s[mp]#### \ 6425 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \ 6426 [[,]f[orce] 6427 Where reboot_mode is one of warm (soft) or cold (hard) or gpio 6428 (prefix with 'panic_' to set mode for panic 6429 reboot only), 6430 reboot_type is one of bios, acpi, kbd, triple, efi, or pci, 6431 reboot_force is either force or not specified, 6432 reboot_cpu is s[mp]#### with #### being the processor 6433 to be used for rebooting. 6434 6435 acpi 6436 Use the ACPI RESET_REG in the FADT. If ACPI is not 6437 configured or the ACPI reset does not work, the reboot 6438 path attempts the reset using the keyboard controller. 6439 6440 bios 6441 Use the CPU reboot vector for warm reset 6442 6443 cold 6444 Set the cold reboot flag 6445 6446 default 6447 There are some built-in platform specific "quirks" 6448 - you may see: "reboot: <name> series board detected. 6449 Selecting <type> for reboots." In the case where you 6450 think the quirk is in error (e.g. you have newer BIOS, 6451 or newer board) using this option will ignore the 6452 built-in quirk table, and use the generic default 6453 reboot actions. 6454 6455 efi 6456 Use efi reset_system runtime service. If EFI is not 6457 configured or the EFI reset does not work, the reboot 6458 path attempts the reset using the keyboard controller. 6459 6460 force 6461 Don't stop other CPUs on reboot. This can make reboot 6462 more reliable in some cases. 6463 6464 kbd 6465 Use the keyboard controller. cold reset (default) 6466 6467 pci 6468 Use a write to the PCI config space register 0xcf9 to 6469 trigger reboot. 6470 6471 triple 6472 Force a triple fault (init) 6473 6474 warm 6475 Don't set the cold reboot flag 6476 6477 Using warm reset will be much faster especially on big 6478 memory systems because the BIOS will not go through 6479 the memory check. Disadvantage is that not all 6480 hardware will be completely reinitialized on reboot so 6481 there may be boot problems on some systems. 6482 6483 6484 refscale.holdoff= [KNL] 6485 Set test-start holdoff period. The purpose of 6486 this parameter is to delay the start of the 6487 test until boot completes in order to avoid 6488 interference. 6489 6490 refscale.lookup_instances= [KNL] 6491 Number of data elements to use for the forms of 6492 SLAB_TYPESAFE_BY_RCU testing. A negative number 6493 is negated and multiplied by nr_cpu_ids, while 6494 zero specifies nr_cpu_ids. 6495 6496 refscale.loops= [KNL] 6497 Set the number of loops over the synchronization 6498 primitive under test. Increasing this number 6499 reduces noise due to loop start/end overhead, 6500 but the default has already reduced the per-pass 6501 noise to a handful of picoseconds on ca. 2020 6502 x86 laptops. 6503 6504 refscale.nreaders= [KNL] 6505 Set number of readers. The default value of -1 6506 selects N, where N is roughly 75% of the number 6507 of CPUs. A value of zero is an interesting choice. 6508 6509 refscale.nruns= [KNL] 6510 Set number of runs, each of which is dumped onto 6511 the console log. 6512 6513 refscale.readdelay= [KNL] 6514 Set the read-side critical-section duration, 6515 measured in microseconds. 6516 6517 refscale.scale_type= [KNL] 6518 Specify the read-protection implementation to test. 6519 6520 refscale.shutdown= [KNL] 6521 Shut down the system at the end of the performance 6522 test. This defaults to 1 (shut it down) when 6523 refscale is built into the kernel and to 0 (leave 6524 it running) when refscale is built as a module. 6525 6526 refscale.verbose= [KNL] 6527 Enable additional printk() statements. 6528 6529 refscale.verbose_batched= [KNL] 6530 Batch the additional printk() statements. If zero 6531 (the default) or negative, print everything. Otherwise, 6532 print every Nth verbose statement, where N is the value 6533 specified. 6534 6535 regulator_ignore_unused 6536 [REGULATOR] 6537 Prevents regulator framework from disabling regulators 6538 that are unused, due no driver claiming them. This may 6539 be useful for debug and development, but should not be 6540 needed on a platform with proper driver support. 6541 6542 relax_domain_level= 6543 [KNL, SMP] Set scheduler's default relax_domain_level. 6544 See Documentation/admin-guide/cgroup-v1/cpusets.rst. 6545 6546 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory 6547 Format: <base1>,<size1>[,<base2>,<size2>,...] 6548 Reserve I/O ports or memory so the kernel won't use 6549 them. If <base> is less than 0x10000, the region 6550 is assumed to be I/O ports; otherwise it is memory. 6551 6552 reserve_mem= [RAM] 6553 Format: nn[KMG]:<align>:<label> 6554 Reserve physical memory and label it with a name that 6555 other subsystems can use to access it. This is typically 6556 used for systems that do not wipe the RAM, and this command 6557 line will try to reserve the same physical memory on 6558 soft reboots. Note, it is not guaranteed to be the same 6559 location. For example, if anything about the system changes 6560 or if booting a different kernel. It can also fail if KASLR 6561 places the kernel at the location of where the RAM reservation 6562 was from a previous boot, the new reservation will be at a 6563 different location. 6564 Any subsystem using this feature must add a way to verify 6565 that the contents of the physical memory is from a previous 6566 boot, as there may be cases where the memory will not be 6567 located at the same location. 6568 6569 The format is size:align:label for example, to request 6570 12 megabytes of 4096 alignment for ramoops: 6571 6572 reserve_mem=12M:4096:oops ramoops.mem_name=oops 6573 6574 reservetop= [X86-32,EARLY] 6575 Format: nn[KMG] 6576 Reserves a hole at the top of the kernel virtual 6577 address space. 6578 6579 reset_devices [KNL] Force drivers to reset the underlying device 6580 during initialization. 6581 6582 resume= [SWSUSP] 6583 Specify the partition device for software suspend 6584 Format: 6585 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>} 6586 6587 resume_offset= [SWSUSP] 6588 Specify the offset from the beginning of the partition 6589 given by "resume=" at which the swap header is located, 6590 in <PAGE_SIZE> units (needed only for swap files). 6591 See Documentation/power/swsusp-and-swap-files.rst 6592 6593 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to 6594 read the resume files 6595 6596 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up. 6597 Useful for devices that are detected asynchronously 6598 (e.g. USB and MMC devices). 6599 6600 retain_initrd [RAM] Keep initrd memory after extraction. After boot, it will 6601 be accessible via /sys/firmware/initrd. 6602 6603 retbleed= [X86] Control mitigation of RETBleed (Arbitrary 6604 Speculative Code Execution with Return Instructions) 6605 vulnerability. 6606 6607 AMD-based UNRET and IBPB mitigations alone do not stop 6608 sibling threads from influencing the predictions of other 6609 sibling threads. For that reason, STIBP is used on pro- 6610 cessors that support it, and mitigate SMT on processors 6611 that don't. 6612 6613 off - no mitigation 6614 auto - automatically select a mitigation 6615 auto,nosmt - automatically select a mitigation, 6616 disabling SMT if necessary for 6617 the full mitigation (only on Zen1 6618 and older without STIBP). 6619 ibpb - On AMD, mitigate short speculation 6620 windows on basic block boundaries too. 6621 Safe, highest perf impact. It also 6622 enables STIBP if present. Not suitable 6623 on Intel. 6624 ibpb,nosmt - Like "ibpb" above but will disable SMT 6625 when STIBP is not available. This is 6626 the alternative for systems which do not 6627 have STIBP. 6628 unret - Force enable untrained return thunks, 6629 only effective on AMD f15h-f17h based 6630 systems. 6631 unret,nosmt - Like unret, but will disable SMT when STIBP 6632 is not available. This is the alternative for 6633 systems which do not have STIBP. 6634 6635 Selecting 'auto' will choose a mitigation method at run 6636 time according to the CPU. 6637 6638 Not specifying this option is equivalent to retbleed=auto. 6639 6640 rfkill.default_state= 6641 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm, 6642 etc. communication is blocked by default. 6643 1 Unblocked. 6644 6645 rfkill.master_switch_mode= 6646 0 The "airplane mode" button does nothing. 6647 1 The "airplane mode" button toggles between everything 6648 blocked and the previous configuration. 6649 2 The "airplane mode" button toggles between everything 6650 blocked and everything unblocked. 6651 6652 ring3mwait=disable 6653 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported 6654 CPUs. 6655 6656 riscv_isa_fallback [RISCV,EARLY] 6657 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit 6658 falling back to detecting extension support by parsing 6659 "riscv,isa" property on devicetree systems when the 6660 replacement properties are not found. See the Kconfig 6661 entry for RISCV_ISA_FALLBACK. 6662 6663 riscv_nousercfi= 6664 all Disable user CFI ABI to userspace even if cpu extension 6665 are available. 6666 bcfi Disable user backward CFI ABI to userspace even if 6667 the shadow stack extension is available. 6668 fcfi Disable user forward CFI ABI to userspace even if the 6669 landing pad extension is available. 6670 6671 ro [KNL] Mount root device read-only on boot 6672 6673 rodata= [KNL,EARLY] 6674 on Mark read-only kernel memory as read-only (default). 6675 off Leave read-only kernel memory writable for debugging. 6676 noalias Mark read-only kernel memory as read-only but retain 6677 writable aliases in the direct map for regions outside 6678 of the kernel image. [arm64] 6679 6680 rockchip.usb_uart 6681 [EARLY] 6682 Enable the uart passthrough on the designated usb port 6683 on Rockchip SoCs. When active, the signals of the 6684 debug-uart get routed to the D+ and D- pins of the usb 6685 port and the regular usb controller gets disabled. 6686 6687 root= [KNL] Root filesystem 6688 Usually this is a block device specifier of some kind, 6689 see the early_lookup_bdev comment in 6690 block/early-lookup.c for details. 6691 Alternatively this can be "ram" for the legacy initial 6692 ramdisk, "nfs" and "cifs" for root on a network file 6693 system, or "mtd" and "ubi" for mounting from raw flash. 6694 6695 rootdelay= [KNL] Delay (in seconds) to pause before attempting to 6696 mount the root filesystem 6697 6698 rootflags= [KNL] Set root filesystem mount option string 6699 6700 rseq_slice_ext= [KNL] RSEQ based time slice extension 6701 Format: boolean 6702 Control enablement of RSEQ based time slice extension. 6703 Default is 'on'. 6704 6705 initramfs_options= [KNL] 6706 Specify mount options for the initramfs mount. 6707 6708 rootfstype= [KNL] Set root filesystem type 6709 6710 rootwait [KNL] Wait (indefinitely) for root device to show up. 6711 Useful for devices that are detected asynchronously 6712 (e.g. USB and MMC devices). 6713 6714 rootwait= [KNL] Maximum time (in seconds) to wait for root device 6715 to show up before attempting to mount the root 6716 filesystem. 6717 6718 rproc_mem=nn[KMG][@address] 6719 [KNL,ARM,CMA] Remoteproc physical memory block. 6720 Memory area to be used by remote processor image, 6721 managed by CMA. 6722 6723 rseq_debug= [KNL] Enable or disable restartable sequence 6724 debug mode. Defaults to CONFIG_RSEQ_DEBUG_DEFAULT_ENABLE. 6725 Format: <bool> 6726 6727 rt_group_sched= [KNL] Enable or disable SCHED_RR/FIFO group scheduling 6728 when CONFIG_RT_GROUP_SCHED=y. Defaults to 6729 !CONFIG_RT_GROUP_SCHED_DEFAULT_DISABLED. 6730 Format: <bool> 6731 6732 rw [KNL] Mount root device read-write on boot 6733 6734 S [KNL] Run init in single mode 6735 6736 s390_iommu= [HW,S390] 6737 Set s390 IOTLB flushing mode 6738 strict 6739 With strict flushing every unmap operation will result 6740 in an IOTLB flush. Default is lazy flushing before 6741 reuse, which is faster. Deprecated, equivalent to 6742 iommu.strict=1. 6743 6744 s390_iommu_aperture= [KNL,S390] 6745 Specifies the size of the per device DMA address space 6746 accessible through the DMA and IOMMU APIs as a decimal 6747 factor of the size of main memory. 6748 The default is 1 meaning that one can concurrently use 6749 as many DMA addresses as physical memory is installed, 6750 if supported by hardware, and thus map all of memory 6751 once. With a value of 2 one can map all of memory twice 6752 and so on. As a special case a factor of 0 imposes no 6753 restrictions other than those given by hardware at the 6754 cost of significant additional memory use for tables. 6755 6756 sa1100ir [NET] 6757 See drivers/net/irda/sa1100_ir.c. 6758 6759 sched_proxy_exec= [KNL] 6760 Enables or disables "proxy execution" style 6761 solution to mutex-based priority inversion. 6762 Format: <bool> 6763 6764 sched_verbose [KNL,EARLY] Enables verbose scheduler debug messages. 6765 6766 schedstats= [KNL,X86] Enable or disable scheduled statistics. 6767 Allowed values are enable and disable. This feature 6768 incurs a small amount of overhead in the scheduler 6769 but is useful for debugging and performance tuning. 6770 6771 sched_thermal_decay_shift= 6772 [Deprecated] 6773 [KNL, SMP] Set a decay shift for scheduler thermal 6774 pressure signal. Thermal pressure signal follows the 6775 default decay period of other scheduler pelt 6776 signals(usually 32 ms but configurable). Setting 6777 sched_thermal_decay_shift will left shift the decay 6778 period for the thermal pressure signal by the shift 6779 value. 6780 i.e. with the default pelt decay period of 32 ms 6781 sched_thermal_decay_shift thermal pressure decay pr 6782 1 64 ms 6783 2 128 ms 6784 and so on. 6785 Format: integer between 0 and 10 6786 Default is 0. 6787 6788 scftorture.holdoff= [KNL] 6789 Number of seconds to hold off before starting 6790 test. Defaults to zero for module insertion and 6791 to 10 seconds for built-in smp_call_function() 6792 tests. 6793 6794 scftorture.longwait= [KNL] 6795 Request ridiculously long waits randomly selected 6796 up to the chosen limit in seconds. Zero (the 6797 default) disables this feature. Please note 6798 that requesting even small non-zero numbers of 6799 seconds can result in RCU CPU stall warnings, 6800 softlockup complaints, and so on. 6801 6802 scftorture.nthreads= [KNL] 6803 Number of kthreads to spawn to invoke the 6804 smp_call_function() family of functions. 6805 The default of -1 specifies a number of kthreads 6806 equal to the number of CPUs. 6807 6808 scftorture.onoff_holdoff= [KNL] 6809 Number seconds to wait after the start of the 6810 test before initiating CPU-hotplug operations. 6811 6812 scftorture.onoff_interval= [KNL] 6813 Number seconds to wait between successive 6814 CPU-hotplug operations. Specifying zero (which 6815 is the default) disables CPU-hotplug operations. 6816 6817 scftorture.shutdown_secs= [KNL] 6818 The number of seconds following the start of the 6819 test after which to shut down the system. The 6820 default of zero avoids shutting down the system. 6821 Non-zero values are useful for automated tests. 6822 6823 scftorture.stat_interval= [KNL] 6824 The number of seconds between outputting the 6825 current test statistics to the console. A value 6826 of zero disables statistics output. 6827 6828 scftorture.stutter_cpus= [KNL] 6829 The number of jiffies to wait between each change 6830 to the set of CPUs under test. 6831 6832 scftorture.use_cpus_read_lock= [KNL] 6833 Use use_cpus_read_lock() instead of the default 6834 preempt_disable() to disable CPU hotplug 6835 while invoking one of the smp_call_function*() 6836 functions. 6837 6838 scftorture.verbose= [KNL] 6839 Enable additional printk() statements. 6840 6841 scftorture.weight_single= [KNL] 6842 The probability weighting to use for the 6843 smp_call_function_single() function with a zero 6844 "wait" parameter. A value of -1 selects the 6845 default if all other weights are -1. However, 6846 if at least one weight has some other value, a 6847 value of -1 will instead select a weight of zero. 6848 6849 scftorture.weight_single_wait= [KNL] 6850 The probability weighting to use for the 6851 smp_call_function_single() function with a 6852 non-zero "wait" parameter. See weight_single. 6853 6854 scftorture.weight_many= [KNL] 6855 The probability weighting to use for the 6856 smp_call_function_many() function with a zero 6857 "wait" parameter. See weight_single. 6858 Note well that setting a high probability for 6859 this weighting can place serious IPI load 6860 on the system. 6861 6862 scftorture.weight_many_wait= [KNL] 6863 The probability weighting to use for the 6864 smp_call_function_many() function with a 6865 non-zero "wait" parameter. See weight_single 6866 and weight_many. 6867 6868 scftorture.weight_all= [KNL] 6869 The probability weighting to use for the 6870 smp_call_function_all() function with a zero 6871 "wait" parameter. See weight_single and 6872 weight_many. 6873 6874 scftorture.weight_all_wait= [KNL] 6875 The probability weighting to use for the 6876 smp_call_function_all() function with a 6877 non-zero "wait" parameter. See weight_single 6878 and weight_many. 6879 6880 sdw_mclk_divider=[SDW] 6881 Specify the MCLK divider for Intel SoundWire buses in 6882 case the BIOS does not provide the clock rate properly. 6883 6884 skew_tick= [KNL,EARLY] Offset the periodic timer tick per cpu to mitigate 6885 xtime_lock contention on larger systems, and/or RCU lock 6886 contention on all systems with CONFIG_MAXSMP set. 6887 Format: { "0" | "1" } 6888 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1" 6889 1 -- enable. 6890 Note: increases power consumption, thus should only be 6891 enabled if running jitter sensitive (HPC/RT) workloads. 6892 6893 security= [SECURITY] Choose a legacy "major" security module to 6894 enable at boot. This has been deprecated by the 6895 "lsm=" parameter. 6896 6897 selinux= [SELINUX] Disable or enable SELinux at boot time. 6898 Format: { "0" | "1" } 6899 See security/selinux/Kconfig help text. 6900 0 -- disable. 6901 1 -- enable. 6902 Default value is 1. 6903 6904 serialnumber [BUGS=X86-32] 6905 6906 sev=option[,option...] [X86-64] 6907 6908 debug 6909 Enable debug messages. 6910 6911 nosnp 6912 Do not enable SEV-SNP (applies to host/hypervisor 6913 only). Setting 'nosnp' avoids the RMP check overhead 6914 in memory accesses when users do not want to run 6915 SEV-SNP guests. 6916 6917 shapers= [NET] 6918 Maximal number of shapers. 6919 6920 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller 6921 Limit apic dumping. The parameter defines the maximal 6922 number of local apics being dumped. Also it is possible 6923 to set it to "all" by meaning -- no limit here. 6924 Format: { 1 (default) | 2 | ... | all }. 6925 The parameter valid if only apic=debug or 6926 apic=verbose is specified. 6927 Example: apic=debug show_lapic=all 6928 6929 slab_debug[=options[,slabs][;[options[,slabs]]...] [MM] 6930 Enabling slab_debug allows one to determine the 6931 culprit if slab objects become corrupted. Enabling 6932 slab_debug can create guard zones around objects and 6933 may poison objects when not in use. Also tracks the 6934 last alloc / free. For more information see 6935 Documentation/admin-guide/mm/slab.rst. 6936 (slub_debug legacy name also accepted for now) 6937 6938 Using this option implies the "no_hash_pointers" 6939 option which can be undone by adding the 6940 "hash_pointers=always" option. 6941 6942 slab_max_order= [MM] 6943 Determines the maximum allowed order for slabs. 6944 A high setting may cause OOMs due to memory 6945 fragmentation. For more information see 6946 Documentation/admin-guide/mm/slab.rst. 6947 (slub_max_order legacy name also accepted for now) 6948 6949 slab_merge [MM] 6950 Enable merging of slabs with similar size when the 6951 kernel is built without CONFIG_SLAB_MERGE_DEFAULT. 6952 (slub_merge legacy name also accepted for now) 6953 6954 slab_min_objects= [MM] 6955 The minimum number of objects per slab. SLUB will 6956 increase the slab order up to slab_max_order to 6957 generate a sufficiently large slab able to contain 6958 the number of objects indicated. The higher the number 6959 of objects the smaller the overhead of tracking slabs 6960 and the less frequently locks need to be acquired. 6961 For more information see 6962 Documentation/admin-guide/mm/slab.rst. 6963 (slub_min_objects legacy name also accepted for now) 6964 6965 slab_min_order= [MM] 6966 Determines the minimum page order for slabs. Must be 6967 lower or equal to slab_max_order. For more information see 6968 Documentation/admin-guide/mm/slab.rst. 6969 (slub_min_order legacy name also accepted for now) 6970 6971 slab_nomerge [MM] 6972 Disable merging of slabs with similar size. May be 6973 necessary if there is some reason to distinguish 6974 allocs to different slabs, especially in hardened 6975 environments where the risk of heap overflows and 6976 layout control by attackers can usually be 6977 frustrated by disabling merging. This will reduce 6978 most of the exposure of a heap attack to a single 6979 cache (risks via metadata attacks are mostly 6980 unchanged). Debug options disable merging on their 6981 own. 6982 For more information see 6983 Documentation/admin-guide/mm/slab.rst. 6984 (slub_nomerge legacy name also accepted for now) 6985 6986 slab_strict_numa [MM] 6987 Support memory policies on a per object level 6988 in the slab allocator. The default is for memory 6989 policies to be applied at the folio level when 6990 a new folio is needed or a partial folio is 6991 retrieved from the lists. Increases overhead 6992 in the slab fastpaths but gains more accurate 6993 NUMA kernel object placement which helps with slow 6994 interconnects in NUMA systems. 6995 6996 slram= [HW,MTD] 6997 6998 smart2= [HW] 6999 Format: <io1>[,<io2>[,...,<io8>]] 7000 7001 smp.csd_lock_timeout= [KNL] 7002 Specify the period of time in milliseconds 7003 that smp_call_function() and friends will wait 7004 for a CPU to release the CSD lock. This is 7005 useful when diagnosing bugs involving CPUs 7006 disabling interrupts for extended periods 7007 of time. Defaults to 5,000 milliseconds, and 7008 setting a value of zero disables this feature. 7009 This feature may be more efficiently disabled 7010 using the csdlock_debug- kernel parameter. 7011 7012 smp.panic_on_ipistall= [KNL] 7013 If a csd_lock_timeout extends for more than 7014 the specified number of milliseconds, panic the 7015 system. By default, let CSD-lock acquisition 7016 take as long as they take. Specifying 300,000 7017 for this value provides a 5-minute timeout. 7018 7019 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices 7020 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port 7021 smsc-ircc2.ircc_sir= [HW] SIR base I/O port 7022 smsc-ircc2.ircc_fir= [HW] FIR base I/O port 7023 smsc-ircc2.ircc_irq= [HW] IRQ line 7024 smsc-ircc2.ircc_dma= [HW] DMA channel 7025 smsc-ircc2.ircc_transceiver= [HW] Transceiver type: 7026 0: Toshiba Satellite 1800 (GP data pin select) 7027 1: Fast pin select (default) 7028 2: ATC IRMode 7029 7030 smt= [KNL,MIPS,S390,EARLY] Set the maximum number of threads 7031 (logical CPUs) to use per physical CPU on systems 7032 capable of symmetric multithreading (SMT). Will 7033 be capped to the actual hardware limit. 7034 Format: <integer> 7035 Default: -1 (no limit) 7036 7037 softlockup_panic= 7038 [KNL] Should the soft-lockup detector generate panics. 7039 Format: <int> 7040 7041 A value of non-zero instructs the soft-lockup detector 7042 to panic the machine when a soft-lockup duration exceeds 7043 N thresholds. It is also controlled by the kernel.softlockup_panic 7044 sysctl and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the 7045 respective build-time switch to that functionality. 7046 7047 softlockup_all_cpu_backtrace= 7048 [KNL] Should the soft-lockup detector generate 7049 backtraces on all cpus. 7050 Format: 0 | 1 7051 7052 sonypi.*= [HW] Sony Programmable I/O Control Device driver 7053 See Documentation/admin-guide/laptops/sonypi.rst 7054 7055 spectre_bhi= [X86] Control mitigation of Branch History Injection 7056 (BHI) vulnerability. This setting affects the 7057 deployment of the HW BHI control and the SW BHB 7058 clearing sequence. 7059 7060 on - (default) Enable the HW or SW mitigation as 7061 needed. This protects the kernel from 7062 both syscalls and VMs. 7063 vmexit - On systems which don't have the HW mitigation 7064 available, enable the SW mitigation on vmexit 7065 ONLY. On such systems, the host kernel is 7066 protected from VM-originated BHI attacks, but 7067 may still be vulnerable to syscall attacks. 7068 off - Disable the mitigation. 7069 7070 spectre_v2= [X86,EARLY] Control mitigation of Spectre variant 2 7071 (indirect branch speculation) vulnerability. 7072 The default operation protects the kernel from 7073 user space attacks. 7074 7075 on - unconditionally enable, implies 7076 spectre_v2_user=on 7077 off - unconditionally disable, implies 7078 spectre_v2_user=off 7079 auto - kernel detects whether your CPU model is 7080 vulnerable 7081 7082 Selecting 'on' will, and 'auto' may, choose a 7083 mitigation method at run time according to the 7084 CPU, the available microcode, the setting of the 7085 CONFIG_MITIGATION_RETPOLINE configuration option, 7086 and the compiler with which the kernel was built. 7087 7088 Selecting 'on' will also enable the mitigation 7089 against user space to user space task attacks. 7090 Selecting specific mitigation does not force enable 7091 user mitigations. 7092 7093 Selecting 'off' will disable both the kernel and 7094 the user space protections. 7095 7096 Specific mitigations can also be selected manually: 7097 7098 retpoline - replace indirect branches 7099 retpoline,generic - Retpolines 7100 retpoline,lfence - LFENCE; indirect branch 7101 retpoline,amd - alias for retpoline,lfence 7102 eibrs - Enhanced/Auto IBRS 7103 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines 7104 eibrs,lfence - Enhanced/Auto IBRS + LFENCE 7105 ibrs - use IBRS to protect kernel 7106 7107 Not specifying this option is equivalent to 7108 spectre_v2=auto. 7109 7110 spectre_v2_user= 7111 [X86] Control mitigation of Spectre variant 2 7112 (indirect branch speculation) vulnerability between 7113 user space tasks 7114 7115 on - Unconditionally enable mitigations. Is 7116 enforced by spectre_v2=on 7117 7118 off - Unconditionally disable mitigations. Is 7119 enforced by spectre_v2=off 7120 7121 prctl - Indirect branch speculation is enabled, 7122 but mitigation can be enabled via prctl 7123 per thread. The mitigation control state 7124 is inherited on fork. 7125 7126 prctl,ibpb 7127 - Like "prctl" above, but only STIBP is 7128 controlled per thread. IBPB is issued 7129 always when switching between different user 7130 space processes. 7131 7132 seccomp 7133 - Same as "prctl" above, but all seccomp 7134 threads will enable the mitigation unless 7135 they explicitly opt out. 7136 7137 seccomp,ibpb 7138 - Like "seccomp" above, but only STIBP is 7139 controlled per thread. IBPB is issued 7140 always when switching between different 7141 user space processes. 7142 7143 auto - Kernel selects the mitigation depending on 7144 the available CPU features and vulnerability. 7145 7146 Default mitigation: "prctl" 7147 7148 Not specifying this option is equivalent to 7149 spectre_v2_user=auto. 7150 7151 spec_rstack_overflow= 7152 [X86,EARLY] Control RAS overflow mitigation on AMD Zen CPUs 7153 7154 off - Disable mitigation 7155 microcode - Enable microcode mitigation only 7156 safe-ret - Enable sw-only safe RET mitigation (default) 7157 ibpb - Enable mitigation by issuing IBPB on 7158 kernel entry 7159 ibpb-vmexit - Issue IBPB only on VMEXIT 7160 (cloud-specific mitigation) 7161 7162 spec_store_bypass_disable= 7163 [HW,EARLY] Control Speculative Store Bypass (SSB) Disable mitigation 7164 (Speculative Store Bypass vulnerability) 7165 7166 Certain CPUs are vulnerable to an exploit against a 7167 a common industry wide performance optimization known 7168 as "Speculative Store Bypass" in which recent stores 7169 to the same memory location may not be observed by 7170 later loads during speculative execution. The idea 7171 is that such stores are unlikely and that they can 7172 be detected prior to instruction retirement at the 7173 end of a particular speculation execution window. 7174 7175 In vulnerable processors, the speculatively forwarded 7176 store can be used in a cache side channel attack, for 7177 example to read memory to which the attacker does not 7178 directly have access (e.g. inside sandboxed code). 7179 7180 This parameter controls whether the Speculative Store 7181 Bypass optimization is used. 7182 7183 On x86 the options are: 7184 7185 on - Unconditionally disable Speculative Store Bypass 7186 off - Unconditionally enable Speculative Store Bypass 7187 auto - Kernel detects whether the CPU model contains an 7188 implementation of Speculative Store Bypass and 7189 picks the most appropriate mitigation. If the 7190 CPU is not vulnerable, "off" is selected. If the 7191 CPU is vulnerable the default mitigation is 7192 architecture and Kconfig dependent. See below. 7193 prctl - Control Speculative Store Bypass per thread 7194 via prctl. Speculative Store Bypass is enabled 7195 for a process by default. The state of the control 7196 is inherited on fork. 7197 seccomp - Same as "prctl" above, but all seccomp threads 7198 will disable SSB unless they explicitly opt out. 7199 7200 Default mitigations: 7201 X86: "prctl" 7202 7203 On powerpc the options are: 7204 7205 on,auto - On Power8 and Power9 insert a store-forwarding 7206 barrier on kernel entry and exit. On Power7 7207 perform a software flush on kernel entry and 7208 exit. 7209 off - No action. 7210 7211 Not specifying this option is equivalent to 7212 spec_store_bypass_disable=auto. 7213 7214 split_llc= 7215 [X86,EARLY] Split the LLC N-ways 7216 7217 When set, the LLC is split this many ways by matching 7218 'core_id % n'. This is setup before SMP bringup and 7219 used during SMP bringup before it knows the full 7220 topology. If your core count doesn't nicely divide by 7221 the number given, you get to keep the pieces. 7222 7223 This is mostly a debug feature to emulate multiple LLCs 7224 on hardware that only have a single LLC. 7225 7226 split_lock_detect= 7227 [X86] Enable split lock detection or bus lock detection 7228 7229 When enabled (and if hardware support is present), atomic 7230 instructions that access data across cache line 7231 boundaries will result in an alignment check exception 7232 for split lock detection or a debug exception for 7233 bus lock detection. 7234 7235 off - not enabled 7236 7237 warn - the kernel will emit rate-limited warnings 7238 about applications triggering the #AC 7239 exception or the #DB exception. This mode is 7240 the default on CPUs that support split lock 7241 detection or bus lock detection. Default 7242 behavior is by #AC if both features are 7243 enabled in hardware. 7244 7245 fatal - the kernel will send SIGBUS to applications 7246 that trigger the #AC exception or the #DB 7247 exception. Default behavior is by #AC if 7248 both features are enabled in hardware. 7249 7250 ratelimit:N - 7251 Set system wide rate limit to N bus locks 7252 per second for bus lock detection. 7253 0 < N <= 1000. 7254 7255 N/A for split lock detection. 7256 7257 7258 If an #AC exception is hit in the kernel or in 7259 firmware (i.e. not while executing in user mode) 7260 the kernel will oops in either "warn" or "fatal" 7261 mode. 7262 7263 #DB exception for bus lock is triggered only when 7264 CPL > 0. 7265 7266 srbds= [X86,INTEL,EARLY] 7267 Control the Special Register Buffer Data Sampling 7268 (SRBDS) mitigation. 7269 7270 Certain CPUs are vulnerable to an MDS-like 7271 exploit which can leak bits from the random 7272 number generator. 7273 7274 By default, this issue is mitigated by 7275 microcode. However, the microcode fix can cause 7276 the RDRAND and RDSEED instructions to become 7277 much slower. Among other effects, this will 7278 result in reduced throughput from /dev/urandom. 7279 7280 The microcode mitigation can be disabled with 7281 the following option: 7282 7283 off: Disable mitigation and remove 7284 performance impact to RDRAND and RDSEED 7285 7286 srcutree.big_cpu_lim [KNL] 7287 Specifies the number of CPUs constituting a 7288 large system, such that srcu_struct structures 7289 should immediately allocate an srcu_node array. 7290 This kernel-boot parameter defaults to 128, 7291 but takes effect only when the low-order four 7292 bits of srcutree.convert_to_big is equal to 3 7293 (decide at boot). 7294 7295 srcutree.convert_to_big [KNL] 7296 Specifies under what conditions an SRCU tree 7297 srcu_struct structure will be converted to big 7298 form, that is, with an rcu_node tree: 7299 7300 0: Never. 7301 1: At init_srcu_struct() time. 7302 2: When rcutorture decides to. 7303 3: Decide at boot time (default). 7304 0x1X: Above plus if high contention. 7305 7306 Either way, the srcu_node tree will be sized based 7307 on the actual runtime number of CPUs (nr_cpu_ids) 7308 instead of the compile-time CONFIG_NR_CPUS. 7309 7310 srcutree.counter_wrap_check [KNL] 7311 Specifies how frequently to check for 7312 grace-period sequence counter wrap for the 7313 srcu_data structure's ->srcu_gp_seq_needed field. 7314 The greater the number of bits set in this kernel 7315 parameter, the less frequently counter wrap will 7316 be checked for. Note that the bottom two bits 7317 are ignored. 7318 7319 srcutree.exp_holdoff [KNL] 7320 Specifies how many nanoseconds must elapse 7321 since the end of the last SRCU grace period for 7322 a given srcu_struct until the next normal SRCU 7323 grace period will be considered for automatic 7324 expediting. Set to zero to disable automatic 7325 expediting. 7326 7327 srcutree.srcu_max_nodelay [KNL] 7328 Specifies the number of no-delay instances 7329 per jiffy for which the SRCU grace period 7330 worker thread will be rescheduled with zero 7331 delay. Beyond this limit, worker thread will 7332 be rescheduled with a sleep delay of one jiffy. 7333 7334 srcutree.srcu_max_nodelay_phase [KNL] 7335 Specifies the per-grace-period phase, number of 7336 non-sleeping polls of readers. Beyond this limit, 7337 grace period worker thread will be rescheduled 7338 with a sleep delay of one jiffy, between each 7339 rescan of the readers, for a grace period phase. 7340 7341 srcutree.srcu_retry_check_delay [KNL] 7342 Specifies number of microseconds of non-sleeping 7343 delay between each non-sleeping poll of readers. 7344 7345 srcutree.small_contention_lim [KNL] 7346 Specifies the number of update-side contention 7347 events per jiffy will be tolerated before 7348 initiating a conversion of an srcu_struct 7349 structure to big form. Note that the value of 7350 srcutree.convert_to_big must have the 0x10 bit 7351 set for contention-based conversions to occur. 7352 7353 ssbd= [ARM64,HW,EARLY] 7354 Speculative Store Bypass Disable control 7355 7356 On CPUs that are vulnerable to the Speculative 7357 Store Bypass vulnerability and offer a 7358 firmware based mitigation, this parameter 7359 indicates how the mitigation should be used: 7360 7361 force-on: Unconditionally enable mitigation for 7362 for both kernel and userspace 7363 force-off: Unconditionally disable mitigation for 7364 for both kernel and userspace 7365 kernel: Always enable mitigation in the 7366 kernel, and offer a prctl interface 7367 to allow userspace to register its 7368 interest in being mitigated too. 7369 7370 stack_guard_gap= [MM] 7371 override the default stack gap protection. The value 7372 is in page units and it defines how many pages prior 7373 to (for stacks growing down) resp. after (for stacks 7374 growing up) the main stack are reserved for no other 7375 mapping. Default value is 256 pages. 7376 7377 stack_depot_disable= [KNL,EARLY] 7378 Setting this to true through kernel command line will 7379 disable the stack depot thereby saving the static memory 7380 consumed by the stack hash table. By default this is set 7381 to false. 7382 7383 stack_depot_max_pools= [KNL,EARLY] 7384 Specify the maximum number of pools to use for storing 7385 stack traces. Pools are allocated on-demand up to this 7386 limit. Default value is 8191 pools. 7387 7388 stacktrace [FTRACE] 7389 Enable the stack tracer on boot up. 7390 7391 stacktrace_filter=[function-list] 7392 [FTRACE] Limit the functions that the stack tracer 7393 will trace at boot up. function-list is a comma-separated 7394 list of functions. This list can be changed at run 7395 time by the stack_trace_filter file in the debugfs 7396 tracing directory. Note, this enables stack tracing 7397 and the stacktrace above is not needed. 7398 7399 sti= [PARISC,HW] 7400 Format: <num> 7401 Set the STI (builtin display/keyboard on the HP-PARISC 7402 machines) console (graphic card) which should be used 7403 as the initial boot-console. 7404 See also comment in drivers/video/console/sticore.c. 7405 7406 sti_font= [HW] 7407 See comment in drivers/video/console/sticore.c. 7408 7409 stifb= [HW] 7410 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]] 7411 7412 strict_sas_size= 7413 [X86] 7414 Format: <bool> 7415 Enable or disable strict sigaltstack size checks 7416 against the required signal frame size which 7417 depends on the supported FPU features. This can 7418 be used to filter out binaries which have 7419 not yet been made aware of AT_MINSIGSTKSZ. 7420 7421 stress_hpt [PPC,EARLY] 7422 Limits the number of kernel HPT entries in the hash 7423 page table to increase the rate of hash page table 7424 faults on kernel addresses. 7425 7426 stress_slb [PPC,EARLY] 7427 Limits the number of kernel SLB entries, and flushes 7428 them frequently to increase the rate of SLB faults 7429 on kernel addresses. 7430 7431 no_slb_preload [PPC,EARLY] 7432 Disables slb preloading for userspace. 7433 7434 sunrpc.min_resvport= 7435 sunrpc.max_resvport= 7436 [NFS,SUNRPC] 7437 SunRPC servers often require that client requests 7438 originate from a privileged port (i.e. a port in the 7439 range 0 < portnr < 1024). 7440 An administrator who wishes to reserve some of these 7441 ports for other uses may adjust the range that the 7442 kernel's sunrpc client considers to be privileged 7443 using these two parameters to set the minimum and 7444 maximum port values. 7445 7446 sunrpc.svc_rpc_per_connection_limit= 7447 [NFS,SUNRPC] 7448 Limit the number of requests that the server will 7449 process in parallel from a single connection. 7450 The default value is 0 (no limit). 7451 7452 sunrpc.pool_mode= 7453 [NFS] 7454 Control how the NFS server code allocates CPUs to 7455 service thread pools. Depending on how many NICs 7456 you have and where their interrupts are bound, this 7457 option will affect which CPUs will do NFS serving. 7458 Note: this parameter cannot be changed while the 7459 NFS server is running. 7460 7461 auto the server chooses an appropriate mode 7462 automatically using heuristics 7463 global a single global pool contains all CPUs 7464 percpu one pool for each CPU 7465 pernode one pool for each NUMA node (equivalent 7466 to global on non-NUMA machines) 7467 7468 sunrpc.tcp_slot_table_entries= 7469 sunrpc.udp_slot_table_entries= 7470 [NFS,SUNRPC] 7471 Sets the upper limit on the number of simultaneous 7472 RPC calls that can be sent from the client to a 7473 server. Increasing these values may allow you to 7474 improve throughput, but will also increase the 7475 amount of memory reserved for use by the client. 7476 7477 suspend.pm_test_delay= 7478 [SUSPEND] 7479 Sets the number of seconds to remain in a suspend test 7480 mode before resuming the system (see 7481 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 7482 is set. Default value is 5. 7483 7484 svm= [PPC] 7485 Format: { on | off | y | n | 1 | 0 } 7486 This parameter controls use of the Protected 7487 Execution Facility on pSeries. 7488 7489 swiotlb= [ARM,PPC,MIPS,X86,S390,EARLY] 7490 Format: { <int> [,<int>] | force | noforce } 7491 <int> -- Number of I/O TLB slabs 7492 <int> -- Second integer after comma. Number of swiotlb 7493 areas with their own lock. Will be rounded up 7494 to a power of 2. 7495 force -- force using of bounce buffers even if they 7496 wouldn't be automatically used by the kernel 7497 noforce -- Never use bounce buffers (for debugging) 7498 7499 switches= [HW,M68k,EARLY] 7500 7501 sysctl.*= [KNL] 7502 Set a sysctl parameter, right before loading the init 7503 process, as if the value was written to the respective 7504 /proc/sys/... file. Both '.' and '/' are recognized as 7505 separators. Unrecognized parameters and invalid values 7506 are reported in the kernel log. Sysctls registered 7507 later by a loaded module cannot be set this way. 7508 Example: sysctl.vm.swappiness=40 7509 7510 sysrq_always_enabled 7511 [KNL] 7512 Ignore sysrq setting - this boot parameter will 7513 neutralize any effect of /proc/sys/kernel/sysrq. 7514 Useful for debugging. 7515 7516 tcpmhash_entries= [KNL,NET] 7517 Set the number of tcp_metrics_hash slots. 7518 Default value is 8192 or 16384 depending on total 7519 ram pages. This is used to specify the TCP metrics 7520 cache size. See Documentation/networking/ip-sysctl.rst 7521 "tcp_no_metrics_save" section for more details. 7522 7523 tdfx= [HW,DRM] 7524 7525 test_suspend= [SUSPEND] 7526 Format: { "mem" | "standby" | "freeze" }[,N] 7527 Specify "mem" (for Suspend-to-RAM) or "standby" (for 7528 standby suspend) or "freeze" (for suspend type freeze) 7529 as the system sleep state during system startup with 7530 the optional capability to repeat N number of times. 7531 The system is woken from this state using a 7532 wakeup-capable RTC alarm. 7533 7534 thash_entries= [KNL,NET] 7535 Set number of hash buckets for TCP connection 7536 7537 thermal.act= [HW,ACPI] 7538 -1: disable all active trip points in all thermal zones 7539 <degrees C>: override all lowest active trip points 7540 7541 thermal.crt= [HW,ACPI] 7542 -1: disable all critical trip points in all thermal zones 7543 <degrees C>: override all critical trip points 7544 7545 thermal.off= [HW,ACPI] 7546 1: disable ACPI thermal control 7547 7548 thermal.psv= [HW,ACPI] 7549 -1: disable all passive trip points 7550 <degrees C>: override all passive trip points to this 7551 value 7552 7553 thermal.tzp= [HW,ACPI] 7554 Specify global default ACPI thermal zone polling rate 7555 <deci-seconds>: poll all this frequency 7556 0: no polling (default) 7557 7558 thp_anon= [KNL] 7559 Format: <size>[KMG],<size>[KMG]:<state>;<size>[KMG]-<size>[KMG]:<state> 7560 state is one of "always", "madvise", "never" or "inherit". 7561 Control the default behavior of the system with respect 7562 to anonymous transparent hugepages. 7563 Can be used multiple times for multiple anon THP sizes. 7564 See Documentation/admin-guide/mm/transhuge.rst for more 7565 details. 7566 7567 threadirqs [KNL,EARLY] 7568 Force threading of all interrupt handlers except those 7569 marked explicitly IRQF_NO_THREAD. 7570 7571 thp_shmem= [KNL] 7572 Format: <size>[KMG],<size>[KMG]:<policy>;<size>[KMG]-<size>[KMG]:<policy> 7573 Control the default policy of each hugepage size for the 7574 internal shmem mount. <policy> is one of policies available 7575 for the shmem mount ("always", "inherit", "never", "within_size", 7576 and "advise"). 7577 It can be used multiple times for multiple shmem THP sizes. 7578 See Documentation/admin-guide/mm/transhuge.rst for more 7579 details. 7580 7581 topology= [S390,EARLY] 7582 Format: {off | on} 7583 Specify if the kernel should make use of the cpu 7584 topology information if the hardware supports this. 7585 The scheduler will make use of this information and 7586 e.g. base its process migration decisions on it. 7587 Default is on. 7588 7589 torture.disable_onoff_at_boot= [KNL] 7590 Prevent the CPU-hotplug component of torturing 7591 until after init has spawned. 7592 7593 torture.ftrace_dump_at_shutdown= [KNL] 7594 Dump the ftrace buffer at torture-test shutdown, 7595 even if there were no errors. This can be a 7596 very costly operation when many torture tests 7597 are running concurrently, especially on systems 7598 with rotating-rust storage. 7599 7600 torture.verbose_sleep_frequency= [KNL] 7601 Specifies how many verbose printk()s should be 7602 emitted between each sleep. The default of zero 7603 disables verbose-printk() sleeping. 7604 7605 torture.verbose_sleep_duration= [KNL] 7606 Duration of each verbose-printk() sleep in jiffies. 7607 7608 tpm.disable_pcr_integrity= [HW,TPM] 7609 Do not protect PCR registers from unintended physical 7610 access, or interposers in the bus by the means of 7611 having an integrity protected session wrapped around 7612 TPM2_PCR_Extend command. Consider this in a situation 7613 where TPM is heavily utilized by IMA, thus protection 7614 causing a major performance hit, and the space where 7615 machines are deployed is by other means guarded. 7616 7617 tpm_crb_ffa.busy_timeout_ms= [ARM64,TPM] 7618 Maximum time in milliseconds to retry sending a message 7619 to the TPM service before giving up. This parameter controls 7620 how long the system will continue retrying when the TPM 7621 service is busy. 7622 Format: <unsigned int> 7623 Default: 2000 (2 seconds) 7624 7625 tpm_suspend_pcr=[HW,TPM] 7626 Format: integer pcr id 7627 Specify that at suspend time, the tpm driver 7628 should extend the specified pcr with zeros, 7629 as a workaround for some chips which fail to 7630 flush the last written pcr on TPM_SaveState. 7631 This will guarantee that all the other pcrs 7632 are saved. 7633 7634 tpm_tis.interrupts= [HW,TPM] 7635 Enable interrupts for the MMIO based physical layer 7636 for the FIFO interface. By default it is set to false 7637 (0). For more information about TPM hardware interfaces 7638 defined by Trusted Computing Group (TCG) see 7639 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/ 7640 7641 tp_printk [FTRACE] 7642 Have the tracepoints sent to printk as well as the 7643 tracing ring buffer. This is useful for early boot up 7644 where the system hangs or reboots and does not give the 7645 option for reading the tracing buffer or performing a 7646 ftrace_dump_on_oops. 7647 7648 To turn off having tracepoints sent to printk, 7649 echo 0 > /proc/sys/kernel/tracepoint_printk 7650 Note, echoing 1 into this file without the 7651 tp_printk kernel cmdline option has no effect. 7652 7653 The tp_printk_stop_on_boot (see below) can also be used 7654 to stop the printing of events to console at 7655 late_initcall_sync. 7656 7657 ** CAUTION ** 7658 7659 Having tracepoints sent to printk() and activating high 7660 frequency tracepoints such as irq or sched, can cause 7661 the system to live lock. 7662 7663 tp_printk_stop_on_boot [FTRACE] 7664 When tp_printk (above) is set, it can cause a lot of noise 7665 on the console. It may be useful to only include the 7666 printing of events during boot up, as user space may 7667 make the system inoperable. 7668 7669 This command line option will stop the printing of events 7670 to console at the late_initcall_sync() time frame. 7671 7672 trace_buf_size=nn[KMG] 7673 [FTRACE] will set tracing buffer size on each cpu. 7674 7675 trace_clock= [FTRACE] Set the clock used for tracing events 7676 at boot up. 7677 local - Use the per CPU time stamp counter 7678 (converted into nanoseconds). Fast, but 7679 depending on the architecture, may not be 7680 in sync between CPUs. 7681 global - Event time stamps are synchronized across 7682 CPUs. May be slower than the local clock, 7683 but better for some race conditions. 7684 counter - Simple counting of events (1, 2, ..) 7685 note, some counts may be skipped due to the 7686 infrastructure grabbing the clock more than 7687 once per event. 7688 uptime - Use jiffies as the time stamp. 7689 perf - Use the same clock that perf uses. 7690 mono - Use ktime_get_mono_fast_ns() for time stamps. 7691 mono_raw - Use ktime_get_raw_fast_ns() for time 7692 stamps. 7693 boot - Use ktime_get_boot_fast_ns() for time stamps. 7694 Architectures may add more clocks. See 7695 Documentation/trace/ftrace.rst for more details. 7696 7697 trace_event=[event-list] 7698 [FTRACE] Set and start specified trace events in order 7699 to facilitate early boot debugging. The event-list is a 7700 comma-separated list of trace events to enable. See 7701 also Documentation/trace/events.rst 7702 7703 To enable modules, use :mod: keyword: 7704 7705 trace_event=:mod:<module> 7706 7707 The value before :mod: will only enable specific events 7708 that are part of the module. See the above mentioned 7709 document for more information. 7710 7711 trace_instance=[instance-info] 7712 [FTRACE] Create a ring buffer instance early in boot up. 7713 This will be listed in: 7714 7715 /sys/kernel/tracing/instances 7716 7717 Events can be enabled at the time the instance is created 7718 via: 7719 7720 trace_instance=<name>,<system1>:<event1>,<system2>:<event2> 7721 7722 Note, the "<system*>:" portion is optional if the event is 7723 unique. 7724 7725 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall 7726 7727 will enable the "sched_switch" event (note, the "sched:" is optional, and 7728 the same thing would happen if it was left off). The irq_handler_entry 7729 event, and all events under the "initcall" system. 7730 7731 Flags can be added to the instance to modify its behavior when it is 7732 created. The flags are separated by '^'. 7733 7734 The available flags are: 7735 7736 traceoff - Have the tracing instance tracing disabled after it is created. 7737 traceprintk - Have trace_printk() write into this trace instance 7738 (note, "printk" and "trace_printk" can also be used) 7739 7740 trace_instance=foo^traceoff^traceprintk,sched,irq 7741 7742 The flags must come before the defined events. 7743 7744 If memory has been reserved (see memmap for x86), the instance 7745 can use that memory: 7746 7747 memmap=12M$0x284500000 trace_instance=boot_map@0x284500000:12M 7748 7749 The above will create a "boot_map" instance that uses the physical 7750 memory at 0x284500000 that is 12Megs. The per CPU buffers of that 7751 instance will be split up accordingly. 7752 7753 Alternatively, the memory can be reserved by the reserve_mem option: 7754 7755 reserve_mem=12M:4096:trace trace_instance=boot_map@trace 7756 7757 This will reserve 12 megabytes at boot up with a 4096 byte alignment 7758 and place the ring buffer in this memory. Note that due to KASLR, the 7759 memory may not be the same location each time, which will not preserve 7760 the buffer content. 7761 7762 Also note that the layout of the ring buffer data may change between 7763 kernel versions where the validator will fail and reset the ring buffer 7764 if the layout is not the same as the previous kernel. 7765 7766 If the ring buffer is used for persistent bootups and has events enabled, 7767 it is recommend to disable tracing so that events from a previous boot do not 7768 mix with events of the current boot (unless you are debugging a random crash 7769 at boot up). 7770 7771 reserve_mem=12M:4096:trace trace_instance=boot_map^traceoff^traceprintk@trace,sched,irq 7772 7773 Note, saving the trace buffer across reboots does require that the system 7774 is set up to not wipe memory. For instance, CONFIG_RESET_ATTACK_MITIGATION 7775 can force a memory reset on boot which will clear any trace that was stored. 7776 This is just one of many ways that can clear memory. Make sure your system 7777 keeps the content of memory across reboots before relying on this option. 7778 7779 NB: Both the mapped address and size must be page aligned for the architecture. 7780 7781 See also Documentation/trace/debugging.rst 7782 7783 7784 trace_options=[option-list] 7785 [FTRACE] Enable or disable tracer options at boot. 7786 The option-list is a comma delimited list of options 7787 that can be enabled or disabled just as if you were 7788 to echo the option name into 7789 7790 /sys/kernel/tracing/trace_options 7791 7792 For example, to enable stacktrace option (to dump the 7793 stack trace of each event), add to the command line: 7794 7795 trace_options=stacktrace 7796 7797 See also Documentation/trace/ftrace.rst "trace options" 7798 section. 7799 7800 trace_trigger=[trigger-list] 7801 [FTRACE] Add an event trigger on specific events. 7802 Set a trigger on top of a specific event, with an optional 7803 filter. 7804 7805 The format is "trace_trigger=<event>.<trigger>[ if <filter>],..." 7806 Where more than one trigger may be specified that are comma delimited. 7807 7808 For example: 7809 7810 trace_trigger="sched_switch.stacktrace if prev_state == 2" 7811 7812 The above will enable the "stacktrace" trigger on the "sched_switch" 7813 event but only trigger it if the "prev_state" of the "sched_switch" 7814 event is "2" (TASK_UNINTERRUPTIBLE). 7815 7816 See also "Event triggers" in Documentation/trace/events.rst 7817 7818 7819 traceoff_after_boot 7820 [FTRACE] Sometimes tracing is used to debug issues 7821 during the boot process. Since the trace buffer has a 7822 limited amount of storage, it may be prudent to 7823 disable tracing after the boot is finished, otherwise 7824 the critical information may be overwritten. With this 7825 option, the main tracing buffer will be turned off at 7826 the end of the boot process. 7827 7828 traceoff_on_warning 7829 [FTRACE] enable this option to disable tracing when a 7830 warning is hit. This turns off "tracing_on". Tracing can 7831 be enabled again by echoing '1' into the "tracing_on" 7832 file located in /sys/kernel/tracing/ 7833 7834 This option is useful, as it disables the trace before 7835 the WARNING dump is called, which prevents the trace to 7836 be filled with content caused by the warning output. 7837 7838 This option can also be set at run time via the sysctl 7839 option: kernel/traceoff_on_warning 7840 7841 transparent_hugepage= 7842 [KNL] 7843 Format: [always|madvise|never] 7844 Can be used to control the default behavior of the system 7845 with respect to transparent hugepages. 7846 See Documentation/admin-guide/mm/transhuge.rst 7847 for more details. 7848 7849 transparent_hugepage_shmem= [KNL] 7850 Format: [always|within_size|advise|never|deny|force] 7851 Can be used to control the hugepage allocation policy for 7852 the internal shmem mount. 7853 See Documentation/admin-guide/mm/transhuge.rst 7854 for more details. 7855 7856 transparent_hugepage_tmpfs= [KNL] 7857 Format: [always|within_size|advise|never] 7858 Can be used to control the default hugepage allocation policy 7859 for the tmpfs mount. 7860 See Documentation/admin-guide/mm/transhuge.rst 7861 for more details. 7862 7863 trusted.source= [KEYS] 7864 Format: <string> 7865 This parameter identifies the trust source as a backend 7866 for trusted keys implementation. Supported trust 7867 sources: 7868 - "tpm" 7869 - "tee" 7870 - "caam" 7871 - "dcp" 7872 - "pkwm" 7873 If not specified then it defaults to iterating through 7874 the trust source list starting with TPM and assigns the 7875 first trust source as a backend which is initialized 7876 successfully during iteration. 7877 7878 trusted.rng= [KEYS] 7879 Format: <string> 7880 The RNG used to generate key material for trusted keys. 7881 Can be one of: 7882 - "kernel" 7883 - the same value as trusted.source: "tpm" or "tee" 7884 - "default" 7885 If not specified, "default" is used. In this case, 7886 the RNG's choice is left to each individual trust source. 7887 7888 trusted.dcp_use_otp_key 7889 This is intended to be used in combination with 7890 trusted.source=dcp and will select the DCP OTP key 7891 instead of the DCP UNIQUE key blob encryption. 7892 7893 trusted.dcp_skip_zk_test 7894 This is intended to be used in combination with 7895 trusted.source=dcp and will disable the check if the 7896 blob key is all zeros. This is helpful for situations where 7897 having this key zero'ed is acceptable. E.g. in testing 7898 scenarios. 7899 7900 tsa= [X86] Control mitigation for Transient Scheduler 7901 Attacks on AMD CPUs. Search the following in your 7902 favourite search engine for more details: 7903 7904 "Technical guidance for mitigating transient scheduler 7905 attacks". 7906 7907 off - disable the mitigation 7908 on - enable the mitigation (default) 7909 user - mitigate only user/kernel transitions 7910 vm - mitigate only guest/host transitions 7911 7912 7913 tsc= Disable clocksource stability checks for TSC. 7914 Format: <string> 7915 [x86] reliable: mark tsc clocksource as reliable, this 7916 disables clocksource verification at runtime, as well 7917 as the stability checks done at bootup. Used to enable 7918 high-resolution timer mode on older hardware, and in 7919 virtualized environment. 7920 [x86] noirqtime: Do not use TSC to do irq accounting. 7921 Used to run time disable IRQ_TIME_ACCOUNTING on any 7922 platforms where RDTSC is slow and this accounting 7923 can add overhead. 7924 [x86] unstable: mark the TSC clocksource as unstable, this 7925 marks the TSC unconditionally unstable at bootup and 7926 avoids any further wobbles once the TSC watchdog notices. 7927 [x86] nowatchdog: disable clocksource watchdog. Used 7928 in situations with strict latency requirements (where 7929 interruptions from clocksource watchdog are not 7930 acceptable). 7931 [x86] recalibrate: force recalibration against a HW timer 7932 (HPET or PM timer) on systems whose TSC frequency was 7933 obtained from HW or FW using either an MSR or CPUID(0x15). 7934 Warn if the difference is more than 500 ppm. 7935 [x86] watchdog: Enforce the clocksource watchdog on TSC 7936 7937 tsc_early_khz= [X86,EARLY] Skip early TSC calibration and use the given 7938 value instead. Useful when the early TSC frequency discovery 7939 procedure is not reliable, such as on overclocked systems 7940 with CPUID.16h support and partial CPUID.15h support. 7941 Format: <unsigned int> 7942 7943 tsx= [X86] Control Transactional Synchronization 7944 Extensions (TSX) feature in Intel processors that 7945 support TSX control. 7946 7947 This parameter controls the TSX feature. The options are: 7948 7949 on - Enable TSX on the system. Although there are 7950 mitigations for all known security vulnerabilities, 7951 TSX has been known to be an accelerator for 7952 several previous speculation-related CVEs, and 7953 so there may be unknown security risks associated 7954 with leaving it enabled. 7955 7956 off - Disable TSX on the system. (Note that this 7957 option takes effect only on newer CPUs which are 7958 not vulnerable to MDS, i.e., have 7959 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get 7960 the new IA32_TSX_CTRL MSR through a microcode 7961 update. This new MSR allows for the reliable 7962 deactivation of the TSX functionality.) 7963 7964 auto - Disable TSX if X86_BUG_TAA is present, 7965 otherwise enable TSX on the system. 7966 7967 Not specifying this option is equivalent to tsx=off. 7968 7969 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 7970 for more details. 7971 7972 tsx_async_abort= [X86,INTEL,EARLY] Control mitigation for the TSX Async 7973 Abort (TAA) vulnerability. 7974 7975 Similar to Micro-architectural Data Sampling (MDS) 7976 certain CPUs that support Transactional 7977 Synchronization Extensions (TSX) are vulnerable to an 7978 exploit against CPU internal buffers which can forward 7979 information to a disclosure gadget under certain 7980 conditions. 7981 7982 In vulnerable processors, the speculatively forwarded 7983 data can be used in a cache side channel attack, to 7984 access data to which the attacker does not have direct 7985 access. 7986 7987 This parameter controls the TAA mitigation. The 7988 options are: 7989 7990 full - Enable TAA mitigation on vulnerable CPUs 7991 if TSX is enabled. 7992 7993 full,nosmt - Enable TAA mitigation and disable SMT on 7994 vulnerable CPUs. If TSX is disabled, SMT 7995 is not disabled because CPU is not 7996 vulnerable to cross-thread TAA attacks. 7997 off - Unconditionally disable TAA mitigation 7998 7999 On MDS-affected machines, tsx_async_abort=off can be 8000 prevented by an active MDS mitigation as both vulnerabilities 8001 are mitigated with the same mechanism so in order to disable 8002 this mitigation, you need to specify mds=off too. 8003 8004 Not specifying this option is equivalent to 8005 tsx_async_abort=full. On CPUs which are MDS affected 8006 and deploy MDS mitigation, TAA mitigation is not 8007 required and doesn't provide any additional 8008 mitigation. 8009 8010 For details see: 8011 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 8012 8013 turbografx.map[2|3]= [HW,JOY] 8014 TurboGraFX parallel port interface 8015 Format: 8016 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7> 8017 See also Documentation/input/devices/joystick-parport.rst 8018 8019 udbg-immortal [PPC] When debugging early kernel crashes that 8020 happen after console_init() and before a proper 8021 console driver takes over, this boot options might 8022 help "seeing" what's going on. 8023 8024 uhash_entries= [KNL,NET] 8025 Set number of hash buckets for UDP/UDP-Lite connections 8026 8027 uhci-hcd.ignore_oc= 8028 [USB] Ignore overcurrent events (default N). 8029 Some badly-designed motherboards generate lots of 8030 bogus events, for ports that aren't wired to 8031 anything. Set this parameter to avoid log spamming. 8032 Note that genuine overcurrent events won't be 8033 reported either. 8034 8035 unaligned_scalar_speed= 8036 [RISCV] 8037 Format: {slow | fast | unsupported} 8038 Allow skipping scalar unaligned access speed tests. This 8039 is useful for testing alternative code paths and to skip 8040 the tests in environments where they run too slowly. All 8041 CPUs must have the same scalar unaligned access speed. 8042 8043 unaligned_vector_speed= 8044 [RISCV] 8045 Format: {slow | fast | unsupported} 8046 Allow skipping vector unaligned access speed tests. This 8047 is useful for testing alternative code paths and to skip 8048 the tests in environments where they run too slowly. All 8049 CPUs must have the same vector unaligned access speed. 8050 8051 unknown_nmi_panic 8052 [X86] Cause panic on unknown NMI. 8053 8054 unwind_debug [X86-64,EARLY] 8055 Enable unwinder debug output. This can be 8056 useful for debugging certain unwinder error 8057 conditions, including corrupt stacks and 8058 bad/missing unwinder metadata. 8059 8060 usbcore.authorized_default= 8061 [USB] Default USB device authorization: 8062 (default -1 = authorized (same as 1), 8063 0 = not authorized, 1 = authorized, 2 = authorized 8064 if device connected to internal port) 8065 8066 usbcore.autosuspend= 8067 [USB] The autosuspend time delay (in seconds) used 8068 for newly-detected USB devices (default 2). This 8069 is the time required before an idle device will be 8070 autosuspended. Devices for which the delay is set 8071 to a negative value won't be autosuspended at all. 8072 8073 usbcore.usbfs_snoop= 8074 [USB] Set to log all usbfs traffic (default 0 = off). 8075 8076 usbcore.usbfs_snoop_max= 8077 [USB] Maximum number of bytes to snoop in each URB 8078 (default = 65536). 8079 8080 usbcore.blinkenlights= 8081 [USB] Set to cycle leds on hubs (default 0 = off). 8082 8083 usbcore.old_scheme_first= 8084 [USB] Start with the old device initialization 8085 scheme (default 0 = off). 8086 8087 usbcore.usbfs_memory_mb= 8088 [USB] Memory limit (in MB) for buffers allocated by 8089 usbfs (default = 16, 0 = max = 2047). 8090 8091 usbcore.use_both_schemes= 8092 [USB] Try the other device initialization scheme 8093 if the first one fails (default 1 = enabled). 8094 8095 usbcore.initial_descriptor_timeout= 8096 [USB] Specifies timeout for the initial 64-byte 8097 USB_REQ_GET_DESCRIPTOR request in milliseconds 8098 (default 5000 = 5.0 seconds). 8099 8100 usbcore.nousb [USB] Disable the USB subsystem 8101 8102 usbcore.quirks= 8103 [USB] A list of quirk entries to augment the built-in 8104 usb core quirk list. List entries are separated by 8105 commas. Each entry has the form 8106 VendorID:ProductID:Flags. The IDs are 4-digit hex 8107 numbers and Flags is a set of letters. Each letter 8108 will change the built-in quirk; setting it if it is 8109 clear and clearing it if it is set. The letters have 8110 the following meanings: 8111 a = USB_QUIRK_STRING_FETCH_255 (string 8112 descriptors must not be fetched using 8113 a 255-byte read); 8114 b = USB_QUIRK_RESET_RESUME (device can't resume 8115 correctly so reset it instead); 8116 c = USB_QUIRK_NO_SET_INTF (device can't handle 8117 Set-Interface requests); 8118 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't 8119 handle its Configuration or Interface 8120 strings); 8121 e = USB_QUIRK_RESET (device can't be reset 8122 (e.g morph devices), don't use reset); 8123 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has 8124 more interface descriptions than the 8125 bNumInterfaces count, and can't handle 8126 talking to these interfaces); 8127 g = USB_QUIRK_DELAY_INIT (device needs a pause 8128 during initialization, after we read 8129 the device descriptor); 8130 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For 8131 high speed and super speed interrupt 8132 endpoints, the USB 2.0 and USB 3.0 spec 8133 require the interval in microframes (1 8134 microframe = 125 microseconds) to be 8135 calculated as interval = 2 ^ 8136 (bInterval-1). 8137 Devices with this quirk report their 8138 bInterval as the result of this 8139 calculation instead of the exponent 8140 variable used in the calculation); 8141 i = USB_QUIRK_DEVICE_QUALIFIER (device can't 8142 handle device_qualifier descriptor 8143 requests); 8144 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device 8145 generates spurious wakeup, ignore 8146 remote wakeup capability); 8147 k = USB_QUIRK_NO_LPM (device can't handle Link 8148 Power Management); 8149 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL 8150 (Device reports its bInterval as linear 8151 frames instead of the USB 2.0 8152 calculation); 8153 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs 8154 to be disconnected before suspend to 8155 prevent spurious wakeup); 8156 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a 8157 pause after every control message); 8158 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra 8159 delay after resetting its port); 8160 p = USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT 8161 (Reduce timeout of the SET_ADDRESS 8162 request from 5000 ms to 500 ms); 8163 q = USB_QUIRK_FORCE_ONE_CONFIG (Device 8164 claims zero configurations, 8165 forcing to 1); 8166 Example: quirks=0781:5580:bk,0a5c:5834:gij 8167 8168 usbhid.mousepoll= 8169 [USBHID] The interval which mice are to be polled at. 8170 8171 usbhid.jspoll= 8172 [USBHID] The interval which joysticks are to be polled at. 8173 8174 usbhid.kbpoll= 8175 [USBHID] The interval which keyboards are to be polled at. 8176 8177 usb-storage.delay_use= 8178 [UMS] The delay in seconds before a new device is 8179 scanned for Logical Units (default 1). 8180 Optionally the delay in milliseconds if the value has 8181 suffix with "ms". 8182 Example: delay_use=2567ms 8183 8184 usb-storage.quirks= 8185 [UMS] A list of quirks entries to supplement or 8186 override the built-in unusual_devs list. List 8187 entries are separated by commas. Each entry has 8188 the form VID:PID:Flags where VID and PID are Vendor 8189 and Product ID values (4-digit hex numbers) and 8190 Flags is a set of characters, each corresponding 8191 to a common usb-storage quirk flag as follows: 8192 a = SANE_SENSE (collect more than 18 bytes 8193 of sense data, not on uas); 8194 b = BAD_SENSE (don't collect more than 18 8195 bytes of sense data, not on uas); 8196 c = FIX_CAPACITY (decrease the reported 8197 device capacity by one sector); 8198 d = NO_READ_DISC_INFO (don't use 8199 READ_DISC_INFO command, not on uas); 8200 e = NO_READ_CAPACITY_16 (don't use 8201 READ_CAPACITY_16 command); 8202 f = NO_REPORT_OPCODES (don't use report opcodes 8203 command, uas only); 8204 g = MAX_SECTORS_240 (don't transfer more than 8205 240 sectors at a time, uas only); 8206 h = CAPACITY_HEURISTICS (decrease the 8207 reported device capacity by one 8208 sector if the number is odd); 8209 i = IGNORE_DEVICE (don't bind to this 8210 device); 8211 j = NO_REPORT_LUNS (don't use report luns 8212 command, uas only); 8213 k = NO_SAME (do not use WRITE_SAME, uas only) 8214 l = NOT_LOCKABLE (don't try to lock and 8215 unlock ejectable media, not on uas); 8216 m = MAX_SECTORS_64 (don't transfer more 8217 than 64 sectors = 32 KB at a time, 8218 not on uas); 8219 n = INITIAL_READ10 (force a retry of the 8220 initial READ(10) command, not on uas); 8221 o = CAPACITY_OK (accept the capacity 8222 reported by the device, not on uas); 8223 p = WRITE_CACHE (the device cache is ON 8224 by default, not on uas); 8225 r = IGNORE_RESIDUE (the device reports 8226 bogus residue values, not on uas); 8227 s = SINGLE_LUN (the device has only one 8228 Logical Unit); 8229 t = NO_ATA_1X (don't allow ATA(12) and ATA(16) 8230 commands, uas only); 8231 u = IGNORE_UAS (don't bind to the uas driver); 8232 w = NO_WP_DETECT (don't test whether the 8233 medium is write-protected). 8234 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE 8235 even if the device claims no cache, 8236 not on uas) 8237 Example: quirks=0419:aaf5:rl,0421:0433:rc 8238 8239 user_debug= [KNL,ARM] 8240 Format: <int> 8241 See arch/arm/Kconfig.debug help text. 8242 1 - undefined instruction events 8243 2 - system calls 8244 4 - invalid data aborts 8245 8 - SIGSEGV faults 8246 16 - SIGBUS faults 8247 Example: user_debug=31 8248 8249 vdso= [X86,SH,SPARC] 8250 On X86_32, this is an alias for vdso32=. Otherwise: 8251 8252 vdso=1: enable VDSO (the default) 8253 vdso=0: disable VDSO mapping 8254 8255 vdso32= [X86] Control the 32-bit vDSO 8256 vdso32=1: enable 32-bit VDSO 8257 vdso32=0 or vdso32=2: disable 32-bit VDSO 8258 8259 See the help text for CONFIG_COMPAT_VDSO for more 8260 details. If CONFIG_COMPAT_VDSO is set, the default is 8261 vdso32=0; otherwise, the default is vdso32=1. 8262 8263 For compatibility with older kernels, vdso32=2 is an 8264 alias for vdso32=0. 8265 8266 Try vdso32=0 if you encounter an error that says: 8267 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed! 8268 8269 video= [FB,EARLY] Frame buffer configuration 8270 See Documentation/fb/modedb.rst. 8271 8272 video.brightness_switch_enabled= [ACPI] 8273 Format: [0|1] 8274 If set to 1, on receiving an ACPI notify event 8275 generated by hotkey, video driver will adjust brightness 8276 level and then send out the event to user space through 8277 the allocated input device. If set to 0, video driver 8278 will only send out the event without touching backlight 8279 brightness level. 8280 default: 1 8281 8282 virtio_mmio.device= 8283 [VMMIO] Memory mapped virtio (platform) device. 8284 8285 <size>@<baseaddr>:<irq>[:<id>] 8286 where: 8287 <size> := size (can use standard suffixes 8288 like K, M and G) 8289 <baseaddr> := physical base address 8290 <irq> := interrupt number (as passed to 8291 request_irq()) 8292 <id> := (optional) platform device id 8293 example: 8294 virtio_mmio.device=1K@0x100b0000:48:7 8295 8296 Can be used multiple times for multiple devices. 8297 8298 vga= [BOOT,X86-32] Select a particular video mode 8299 See Documentation/arch/x86/boot.rst and 8300 Documentation/admin-guide/svga.rst. 8301 Use vga=ask for menu. 8302 This is actually a boot loader parameter; the value is 8303 passed to the kernel using a special protocol. 8304 8305 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y. 8306 May slow down system boot speed, especially when 8307 enabled on systems with a large amount of memory. 8308 All options are enabled by default, and this 8309 interface is meant to allow for selectively 8310 enabling or disabling specific virtual memory 8311 debugging features. 8312 8313 Available options are: 8314 P Enable page structure init time poisoning 8315 - Disable all of the above options 8316 8317 vmalloc=nn[KMG] [KNL,BOOT,EARLY] Forces the vmalloc area to have an 8318 exact size of <nn>. This can be used to increase 8319 the minimum size (128MB on x86, arm32 platforms). 8320 It can also be used to decrease the size and leave more room 8321 for directly mapped kernel RAM. Note that this parameter does 8322 not exist on many other platforms (including arm64, alpha, 8323 loongarch, arc, csky, hexagon, microblaze, mips, nios2, openrisc, 8324 parisc, m64k, powerpc, riscv, sh, um, xtensa, s390, sparc). 8325 8326 vmcp_cma=nn[MG] [KNL,S390,EARLY] 8327 Sets the memory size reserved for contiguous memory 8328 allocations for the vmcp device driver. 8329 8330 vmhalt= [KNL,S390] Perform z/VM CP command after system halt. 8331 Format: <command> 8332 8333 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic. 8334 Format: <command> 8335 8336 vmpoff= [KNL,S390] Perform z/VM CP command after power off. 8337 Format: <command> 8338 8339 vmscape= [X86] Controls mitigation for VMscape attacks. 8340 VMscape attacks can leak information from a userspace 8341 hypervisor to a guest via speculative side-channels. 8342 8343 off - disable the mitigation 8344 ibpb - use Indirect Branch Prediction Barrier 8345 (IBPB) mitigation (default) 8346 force - force vulnerability detection even on 8347 unaffected processors 8348 8349 vsyscall= [X86-64,EARLY] 8350 Controls the behavior of vsyscalls (i.e. calls to 8351 fixed addresses of 0xffffffffff600x00 from legacy 8352 code). Most statically-linked binaries and older 8353 versions of glibc use these calls. Because these 8354 functions are at fixed addresses, they make nice 8355 targets for exploits that can control RIP. 8356 8357 emulate Vsyscalls turn into traps and are emulated 8358 reasonably safely. The vsyscall page is 8359 readable. This disables the Linear 8360 Address Space Separation (LASS) security 8361 feature and makes the system less secure. 8362 8363 xonly [default] Vsyscalls turn into traps and are 8364 emulated reasonably safely. The vsyscall 8365 page is not readable. 8366 8367 none Vsyscalls don't work at all. This makes 8368 them quite hard to use for exploits but 8369 might break your system. 8370 8371 vt.color= [VT] Default text color. 8372 Format: 0xYX, X = foreground, Y = background. 8373 Default: 0x07 = light gray on black. 8374 8375 vt.cur_default= [VT] Default cursor shape. 8376 Format: 0xCCBBAA, where AA, BB, and CC are the same as 8377 the parameters of the <Esc>[?A;B;Cc escape sequence; 8378 see vga-softcursor.rst. Default: 2 = underline. 8379 8380 vt.default_blu= [VT] 8381 Format: <blue0>,<blue1>,<blue2>,...,<blue15> 8382 Change the default blue palette of the console. 8383 This is a 16-member array composed of values 8384 ranging from 0-255. 8385 8386 vt.default_grn= [VT] 8387 Format: <green0>,<green1>,<green2>,...,<green15> 8388 Change the default green palette of the console. 8389 This is a 16-member array composed of values 8390 ranging from 0-255. 8391 8392 vt.default_red= [VT] 8393 Format: <red0>,<red1>,<red2>,...,<red15> 8394 Change the default red palette of the console. 8395 This is a 16-member array composed of values 8396 ranging from 0-255. 8397 8398 vt.default_utf8= 8399 [VT] 8400 Format=<0|1> 8401 Set system-wide default UTF-8 mode for all tty's. 8402 Default is 1, i.e. UTF-8 mode is enabled for all 8403 newly opened terminals. 8404 8405 vt.global_cursor_default= 8406 [VT] 8407 Format=<-1|0|1> 8408 Set system-wide default for whether a cursor 8409 is shown on new VTs. Default is -1, 8410 i.e. cursors will be created by default unless 8411 overridden by individual drivers. 0 will hide 8412 cursors, 1 will display them. 8413 8414 vt.italic= [VT] Default color for italic text; 0-15. 8415 Default: 2 = green. 8416 8417 vt.underline= [VT] Default color for underlined text; 0-15. 8418 Default: 3 = cyan. 8419 8420 watchdog timers [HW,WDT] For information on watchdog timers, 8421 see Documentation/watchdog/watchdog-parameters.rst 8422 or other driver-specific files in the 8423 Documentation/watchdog/ directory. 8424 8425 watchdog_thresh= 8426 [KNL] 8427 Set the hard lockup detector stall duration 8428 threshold in seconds. The soft lockup detector 8429 threshold is set to twice the value. A value of 0 8430 disables both lockup detectors. Default is 10 8431 seconds. 8432 8433 workqueue.unbound_cpus= 8434 [KNL,SMP] Specify to constrain one or some CPUs 8435 to use in unbound workqueues. 8436 Format: <cpu-list> 8437 By default, all online CPUs are available for 8438 unbound workqueues. 8439 8440 workqueue.watchdog_thresh= 8441 If CONFIG_WQ_WATCHDOG is configured, workqueue can 8442 warn stall conditions and dump internal state to 8443 help debugging. 0 disables workqueue stall 8444 detection; otherwise, it's the stall threshold 8445 duration in seconds. The default value is 30 and 8446 it can be updated at runtime by writing to the 8447 corresponding sysfs file. 8448 8449 workqueue.panic_on_stall=<uint> 8450 Panic when workqueue stall is detected by 8451 CONFIG_WQ_WATCHDOG. It sets the number times of the 8452 stall to trigger panic. 8453 8454 The default is set by CONFIG_BOOTPARAM_WQ_STALL_PANIC, 8455 which is 0 (disabled) if not configured. 8456 8457 workqueue.panic_on_stall_time=<uint> 8458 Panic when a workqueue stall has been continuous for 8459 the specified number of seconds. Unlike panic_on_stall 8460 which counts accumulated stall events, this triggers 8461 based on the duration of a single continuous stall. 8462 8463 The default is 0, which disables the time-based panic. 8464 8465 workqueue.cpu_intensive_thresh_us= 8466 Per-cpu work items which run for longer than this 8467 threshold are automatically considered CPU intensive 8468 and excluded from concurrency management to prevent 8469 them from noticeably delaying other per-cpu work 8470 items. Default is 10000 (10ms). 8471 8472 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 8473 will report the work functions which violate this 8474 threshold repeatedly. They are likely good 8475 candidates for using WQ_UNBOUND workqueues instead. 8476 8477 workqueue.cpu_intensive_warning_thresh=<uint> 8478 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 8479 will report the work functions which violate the 8480 intensive_threshold_us repeatedly. In order to prevent 8481 spurious warnings, start printing only after a work 8482 function has violated this threshold number of times. 8483 8484 The default is 4 times. 0 disables the warning. 8485 8486 workqueue.power_efficient 8487 Per-cpu workqueues are generally preferred because 8488 they show better performance thanks to cache 8489 locality; unfortunately, per-cpu workqueues tend to 8490 be more power hungry than unbound workqueues. 8491 8492 Enabling this makes the per-cpu workqueues which 8493 were observed to contribute significantly to power 8494 consumption unbound, leading to measurably lower 8495 power usage at the cost of small performance 8496 overhead. 8497 8498 The default value of this parameter is determined by 8499 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT. 8500 8501 workqueue.default_affinity_scope= 8502 Select the default affinity scope to use for unbound 8503 workqueues. Can be one of "cpu", "smt", "cache", 8504 "cache_shard", "numa" and "system". Default is 8505 "cache_shard". For more 8506 information, see the Affinity Scopes section in 8507 Documentation/core-api/workqueue.rst. 8508 8509 This can be changed after boot by writing to the 8510 matching /sys/module/workqueue/parameters file. All 8511 workqueues with the "default" affinity scope will be 8512 updated accordingly. 8513 8514 workqueue.debug_force_rr_cpu 8515 Workqueue used to implicitly guarantee that work 8516 items queued without explicit CPU specified are put 8517 on the local CPU. This guarantee is no longer true 8518 and while local CPU is still preferred work items 8519 may be put on foreign CPUs. This debug option 8520 forces round-robin CPU selection to flush out 8521 usages which depend on the now broken guarantee. 8522 When enabled, memory and cache locality will be 8523 impacted. 8524 8525 writecombine= [LOONGARCH,EARLY] Control the MAT (Memory Access 8526 Type) of ioremap_wc(). 8527 8528 on - Enable writecombine, use WUC for ioremap_wc() 8529 off - Disable writecombine, use SUC for ioremap_wc() 8530 8531 x2apic_phys [X86-64,APIC,EARLY] Use x2apic physical mode instead of 8532 default x2apic cluster mode on platforms 8533 supporting x2apic. 8534 8535 xen_512gb_limit [KNL,X86-64,XEN] 8536 Restricts the kernel running paravirtualized under Xen 8537 to use only up to 512 GB of RAM. The reason to do so is 8538 crash analysis tools and Xen tools for doing domain 8539 save/restore/migration must be enabled to handle larger 8540 domains. 8541 8542 xen_console_io [XEN,EARLY] 8543 Boolean option to enable/disable the usage of the Xen 8544 console_io hypercalls to read and write to the console. 8545 Mostly useful for debugging and development. 8546 8547 xen_emul_unplug= [HW,X86,XEN,EARLY] 8548 Unplug Xen emulated devices 8549 Format: [unplug0,][unplug1] 8550 ide-disks -- unplug primary master IDE devices 8551 aux-ide-disks -- unplug non-primary-master IDE devices 8552 nics -- unplug network devices 8553 all -- unplug all emulated devices (NICs and IDE disks) 8554 unnecessary -- unplugging emulated devices is 8555 unnecessary even if the host did not respond to 8556 the unplug protocol 8557 never -- do not unplug even if version check succeeds 8558 8559 xen_legacy_crash [X86,XEN,EARLY] 8560 Crash from Xen panic notifier, without executing late 8561 panic() code such as dumping handler. 8562 8563 xen_mc_debug [X86,XEN,EARLY] 8564 Enable multicall debugging when running as a Xen PV guest. 8565 Enabling this feature will reduce performance a little 8566 bit, so it should only be enabled for obtaining extended 8567 debug data in case of multicall errors. 8568 8569 xen_msr_safe= [X86,XEN,EARLY] 8570 Format: <bool> 8571 Select whether to always use non-faulting (safe) MSR 8572 access functions when running as Xen PV guest. The 8573 default value is controlled by CONFIG_XEN_PV_MSR_SAFE. 8574 8575 xen_nopv [X86] 8576 Disables the PV optimizations forcing the HVM guest to 8577 run as generic HVM guest with no PV drivers. 8578 This option is obsoleted by the "nopv" option, which 8579 has equivalent effect for XEN platform. 8580 8581 xen_no_vector_callback 8582 [KNL,X86,XEN,EARLY] Disable the vector callback for Xen 8583 event channel interrupts. 8584 8585 xen_scrub_pages= [XEN] 8586 Boolean option to control scrubbing pages before giving them back 8587 to Xen, for use by other domains. Can be also changed at runtime 8588 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages. 8589 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT. 8590 8591 xen_timer_slop= [X86-64,XEN,EARLY] 8592 Set the timer slop (in nanoseconds) for the virtual Xen 8593 timers (default is 100000). This adjusts the minimum 8594 delta of virtualized Xen timers, where lower values 8595 improve timer resolution at the expense of processing 8596 more timer interrupts. 8597 8598 xen.balloon_boot_timeout= [XEN] 8599 The time (in seconds) to wait before giving up to boot 8600 in case initial ballooning fails to free enough memory. 8601 Applies only when running as HVM or PVH guest and 8602 started with less memory configured than allowed at 8603 max. Default is 180. 8604 8605 xen.event_eoi_delay= [XEN] 8606 How long to delay EOI handling in case of event 8607 storms (jiffies). Default is 10. 8608 8609 xen.event_loop_timeout= [XEN] 8610 After which time (jiffies) the event handling loop 8611 should start to delay EOI handling. Default is 2. 8612 8613 xen.fifo_events= [XEN] 8614 Boolean parameter to disable using fifo event handling 8615 even if available. Normally fifo event handling is 8616 preferred over the 2-level event handling, as it is 8617 fairer and the number of possible event channels is 8618 much higher. Default is on (use fifo events). 8619 8620 xirc2ps_cs= [NET,PCMCIA] 8621 Format: 8622 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]] 8623 8624 xive= [PPC] 8625 By default on POWER9 and above, the kernel will 8626 natively use the XIVE interrupt controller. This option 8627 allows the fallback firmware mode to be used: 8628 8629 off Fallback to firmware control of XIVE interrupt 8630 controller on both pseries and powernv 8631 platforms. Only useful on POWER9 and above. 8632 8633 xive.store-eoi=off [PPC] 8634 By default on POWER10 and above, the kernel will use 8635 stores for EOI handling when the XIVE interrupt mode 8636 is active. This option allows the XIVE driver to use 8637 loads instead, as on POWER9. 8638 8639 xhci-hcd.quirks [USB,KNL] 8640 A hex value specifying bitmask with supplemental xhci 8641 host controller quirks. Meaning of each bit can be 8642 consulted in header drivers/usb/host/xhci.h. 8643 8644 xmon [PPC,EARLY] 8645 Format: { early | on | rw | ro | off } 8646 Controls if xmon debugger is enabled. Default is off. 8647 Passing only "xmon" is equivalent to "xmon=early". 8648 early Call xmon as early as possible on boot; xmon 8649 debugger is called from setup_arch(). 8650 on xmon debugger hooks will be installed so xmon 8651 is only called on a kernel crash. Default mode, 8652 i.e. either "ro" or "rw" mode, is controlled 8653 with CONFIG_XMON_DEFAULT_RO_MODE. 8654 rw xmon debugger hooks will be installed so xmon 8655 is called only on a kernel crash, mode is write, 8656 meaning SPR registers, memory and, other data 8657 can be written using xmon commands. 8658 ro same as "rw" option above but SPR registers, 8659 memory, and other data can't be written using 8660 xmon commands. 8661 off xmon is disabled. 8662