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 AX25 Appropriate AX.25 support is enabled. 10 CLK Common clock infrastructure is enabled. 11 CMA Contiguous Memory Area support is enabled. 12 DRM Direct Rendering Management support is enabled. 13 DYNAMIC_DEBUG Build in debug messages and enable them at runtime 14 EARLY Parameter processed too early to be embedded in initrd. 15 EDD BIOS Enhanced Disk Drive Services (EDD) is enabled 16 EFI EFI Partitioning (GPT) is enabled 17 EVM Extended Verification Module 18 FB The frame buffer device is enabled. 19 FTRACE Function tracing enabled. 20 GCOV GCOV profiling is enabled. 21 HIBERNATION HIBERNATION is enabled. 22 HW Appropriate hardware is enabled. 23 HYPER_V HYPERV support is enabled. 24 IMA Integrity measurement architecture is enabled. 25 IP_PNP IP DHCP, BOOTP, or RARP is enabled. 26 IPV6 IPv6 support is enabled. 27 ISAPNP ISA PnP code is enabled. 28 ISDN Appropriate ISDN support is enabled. 29 ISOL CPU Isolation is enabled. 30 JOY Appropriate joystick support is enabled. 31 KGDB Kernel debugger support is enabled. 32 KVM Kernel Virtual Machine support is enabled. 33 LIBATA Libata driver is enabled 34 LOONGARCH LoongArch architecture is enabled. 35 LOOP Loopback device support is enabled. 36 LP Printer support is enabled. 37 M68k M68k architecture is enabled. 38 These options have more detailed description inside of 39 Documentation/arch/m68k/kernel-options.rst. 40 MDA MDA console support is enabled. 41 MIPS MIPS architecture is enabled. 42 MOUSE Appropriate mouse support is enabled. 43 MSI Message Signaled Interrupts (PCI). 44 MTD MTD (Memory Technology Device) support is enabled. 45 NET Appropriate network support is enabled. 46 NFS Appropriate NFS support is enabled. 47 NUMA NUMA support is enabled. 48 OF Devicetree is enabled. 49 PARISC The PA-RISC architecture is enabled. 50 PCI PCI bus support is enabled. 51 PCIE PCI Express support is enabled. 52 PCMCIA The PCMCIA subsystem is enabled. 53 PNP Plug & Play support is enabled. 54 PPC PowerPC architecture is enabled. 55 PPT Parallel port support is enabled. 56 PS2 Appropriate PS/2 support is enabled. 57 PV_OPS A paravirtualized kernel is enabled. 58 RAM RAM disk support is enabled. 59 RDT Intel Resource Director Technology. 60 RISCV RISCV architecture is enabled. 61 S390 S390 architecture is enabled. 62 SCSI Appropriate SCSI support is enabled. 63 A lot of drivers have their options described inside 64 the Documentation/scsi/ sub-directory. 65 SDW SoundWire support is enabled. 66 SECURITY Different security models are enabled. 67 SELINUX SELinux support is enabled. 68 SERIAL Serial support is enabled. 69 SH SuperH architecture is enabled. 70 SMP The kernel is an SMP kernel. 71 SPARC Sparc architecture is enabled. 72 SUSPEND System suspend states are enabled. 73 SWSUSP Software suspend (hibernation) is enabled. 74 TPM TPM drivers are enabled. 75 UMS USB Mass Storage support is enabled. 76 USB USB support is enabled. 77 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_enforce_resources= [ACPI] 193 { strict | lax | no } 194 Check for resource conflicts between native drivers 195 and ACPI OperationRegions (SystemIO and SystemMemory 196 only). IO ports and memory declared in ACPI might be 197 used by the ACPI subsystem in arbitrary AML code and 198 can interfere with legacy drivers. 199 strict (default): access to resources claimed by ACPI 200 is denied; legacy drivers trying to access reserved 201 resources will fail to bind to device using them. 202 lax: access to resources claimed by ACPI is allowed; 203 legacy drivers trying to access reserved resources 204 will bind successfully but a warning message is logged. 205 no: ACPI OperationRegions are not marked as reserved, 206 no further checks are performed. 207 208 acpi_force_table_verification [HW,ACPI,EARLY] 209 Enable table checksum verification during early stage. 210 By default, this is disabled due to x86 early mapping 211 size limitation. 212 213 acpi_irq_balance [HW,ACPI] 214 ACPI will balance active IRQs 215 default in APIC mode 216 217 acpi_irq_nobalance [HW,ACPI] 218 ACPI will not move active IRQs (default) 219 default in PIC mode 220 221 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA 222 Format: <irq>,<irq>... 223 224 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for 225 use by PCI 226 Format: <irq>,<irq>... 227 228 acpi_mask_gpe= [HW,ACPI] 229 Due to the existence of _Lxx/_Exx, some GPEs triggered 230 by unsupported hardware/firmware features can result in 231 GPE floodings that cannot be automatically disabled by 232 the GPE dispatcher. 233 This facility can be used to prevent such uncontrolled 234 GPE floodings. 235 Format: <byte> or <bitmap-list> 236 237 acpi_no_auto_serialize [HW,ACPI] 238 Disable auto-serialization of AML methods 239 AML control methods that contain the opcodes to create 240 named objects will be marked as "Serialized" by the 241 auto-serialization feature. 242 This feature is enabled by default. 243 This option allows to turn off the feature. 244 245 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump 246 kernels. 247 248 acpi_no_static_ssdt [HW,ACPI,EARLY] 249 Disable installation of static SSDTs at early boot time 250 By default, SSDTs contained in the RSDT/XSDT will be 251 installed automatically and they will appear under 252 /sys/firmware/acpi/tables. 253 This option turns off this feature. 254 Note that specifying this option does not affect 255 dynamic table installation which will install SSDT 256 tables to /sys/firmware/acpi/tables/dynamic. 257 258 acpi_no_watchdog [HW,ACPI,WDT] 259 Ignore the ACPI-based watchdog interface (WDAT) and let 260 a native driver control the watchdog device instead. 261 262 acpi_rsdp= [ACPI,EFI,KEXEC,EARLY] 263 Pass the RSDP address to the kernel, mostly used 264 on machines running EFI runtime service to boot the 265 second kernel for kdump. 266 267 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS 268 Format: To spoof as Windows 98: ="Microsoft Windows" 269 270 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead 271 of 2 which is mandated by ACPI 6) as the supported ACPI 272 specification revision (when using this switch, it may 273 be necessary to carry out a cold reboot _twice_ in a 274 row to make it take effect on the platform firmware). 275 276 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings 277 acpi_osi="string1" # add string1 278 acpi_osi="!string2" # remove string2 279 acpi_osi=!* # remove all strings 280 acpi_osi=! # disable all built-in OS vendor 281 strings 282 acpi_osi=!! # enable all built-in OS vendor 283 strings 284 acpi_osi= # disable all strings 285 286 'acpi_osi=!' can be used in combination with single or 287 multiple 'acpi_osi="string1"' to support specific OS 288 vendor string(s). Note that such command can only 289 affect the default state of the OS vendor strings, thus 290 it cannot affect the default state of the feature group 291 strings and the current state of the OS vendor strings, 292 specifying it multiple times through kernel command line 293 is meaningless. This command is useful when one do not 294 care about the state of the feature group strings which 295 should be controlled by the OSPM. 296 Examples: 297 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent 298 to 'acpi_osi="Windows 2000" acpi_osi=!', they all 299 can make '_OSI("Windows 2000")' TRUE. 300 301 'acpi_osi=' cannot be used in combination with other 302 'acpi_osi=' command lines, the _OSI method will not 303 exist in the ACPI namespace. NOTE that such command can 304 only affect the _OSI support state, thus specifying it 305 multiple times through kernel command line is also 306 meaningless. 307 Examples: 308 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)' 309 FALSE. 310 311 'acpi_osi=!*' can be used in combination with single or 312 multiple 'acpi_osi="string1"' to support specific 313 string(s). Note that such command can affect the 314 current state of both the OS vendor strings and the 315 feature group strings, thus specifying it multiple times 316 through kernel command line is meaningful. But it may 317 still not able to affect the final state of a string if 318 there are quirks related to this string. This command 319 is useful when one want to control the state of the 320 feature group strings to debug BIOS issues related to 321 the OSPM features. 322 Examples: 323 1. 'acpi_osi="Module Device" acpi_osi=!*' can make 324 '_OSI("Module Device")' FALSE. 325 2. 'acpi_osi=!* acpi_osi="Module Device"' can make 326 '_OSI("Module Device")' TRUE. 327 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is 328 equivalent to 329 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"' 330 and 331 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!', 332 they all will make '_OSI("Windows 2000")' TRUE. 333 334 acpi_pm_good [X86] 335 Override the pmtimer bug detection: force the kernel 336 to assume that this machine's pmtimer latches its value 337 and always returns good values. 338 339 acpi_sci= [HW,ACPI,EARLY] ACPI System Control Interrupt trigger mode 340 Format: { level | edge | high | low } 341 342 acpi_skip_timer_override [HW,ACPI,EARLY] 343 Recognize and ignore IRQ0/pin2 Interrupt Override. 344 For broken nForce2 BIOS resulting in XT-PIC timer. 345 346 acpi_sleep= [HW,ACPI] Sleep options 347 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig, 348 s4_nohwsig, old_ordering, nonvs, 349 sci_force_enable, nobl } 350 See Documentation/power/video.rst for information on 351 s3_bios and s3_mode. 352 s3_beep is for debugging; it makes the PC's speaker beep 353 as soon as the kernel's real-mode entry point is called. 354 s4_hwsig causes the kernel to check the ACPI hardware 355 signature during resume from hibernation, and gracefully 356 refuse to resume if it has changed. This complies with 357 the ACPI specification but not with reality, since 358 Windows does not do this and many laptops do change it 359 on docking. So the default behaviour is to allow resume 360 and simply warn when the signature changes, unless the 361 s4_hwsig option is enabled. 362 s4_nohwsig prevents ACPI hardware signature from being 363 used (or even warned about) during resume. 364 old_ordering causes the ACPI 1.0 ordering of the _PTS 365 control method, with respect to putting devices into 366 low power states, to be enforced (the ACPI 2.0 ordering 367 of _PTS is used by default). 368 nonvs prevents the kernel from saving/restoring the 369 ACPI NVS memory during suspend/hibernation and resume. 370 sci_force_enable causes the kernel to set SCI_EN directly 371 on resume from S1/S3 (which is against the ACPI spec, 372 but some broken systems don't work without it). 373 nobl causes the internal blacklist of systems known to 374 behave incorrectly in some ways with respect to system 375 suspend and resume to be ignored (use wisely). 376 377 acpi_use_timer_override [HW,ACPI,EARLY] 378 Use timer override. For some broken Nvidia NF5 boards 379 that require a timer override, but don't have HPET 380 381 add_efi_memmap [EFI,X86,EARLY] Include EFI memory map in 382 kernel's map of available physical RAM. 383 384 agp= [AGP] 385 { off | try_unsupported } 386 off: disable AGP support 387 try_unsupported: try to drive unsupported chipsets 388 (may crash computer or cause data corruption) 389 390 ALSA [HW,ALSA] 391 See Documentation/sound/alsa-configuration.rst 392 393 alignment= [KNL,ARM] 394 Allow the default userspace alignment fault handler 395 behaviour to be specified. Bit 0 enables warnings, 396 bit 1 enables fixups, and bit 2 sends a segfault. 397 398 align_va_addr= [X86-64] 399 Align virtual addresses by clearing slice [14:12] when 400 allocating a VMA at process creation time. This option 401 gives you up to 3% performance improvement on AMD F15h 402 machines (where it is enabled by default) for a 403 CPU-intensive style benchmark, and it can vary highly in 404 a microbenchmark depending on workload and compiler. 405 406 32: only for 32-bit processes 407 64: only for 64-bit processes 408 on: enable for both 32- and 64-bit processes 409 off: disable for both 32- and 64-bit processes 410 411 alloc_snapshot [FTRACE] 412 Allocate the ftrace snapshot buffer on boot up when the 413 main buffer is allocated. This is handy if debugging 414 and you need to use tracing_snapshot() on boot up, and 415 do not want to use tracing_snapshot_alloc() as it needs 416 to be done where GFP_KERNEL allocations are allowed. 417 418 allow_mismatched_32bit_el0 [ARM64,EARLY] 419 Allow execve() of 32-bit applications and setting of the 420 PER_LINUX32 personality on systems where only a strict 421 subset of the CPUs support 32-bit EL0. When this 422 parameter is present, the set of CPUs supporting 32-bit 423 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0 424 and hot-unplug operations may be restricted. 425 426 See Documentation/arch/arm64/asymmetric-32bit.rst for more 427 information. 428 429 amd_iommu= [HW,X86-64] 430 Pass parameters to the AMD IOMMU driver in the system. 431 Possible values are: 432 fullflush - Deprecated, equivalent to iommu.strict=1 433 off - do not initialize any AMD IOMMU found in 434 the system 435 force_isolation - Force device isolation for all 436 devices. The IOMMU driver is not 437 allowed anymore to lift isolation 438 requirements as needed. This option 439 does not override iommu=pt 440 force_enable - Force enable the IOMMU on platforms known 441 to be buggy with IOMMU enabled. Use this 442 option with care. 443 pgtbl_v1 - Use v1 page table for DMA-API (Default). 444 pgtbl_v2 - Use v2 page table for DMA-API. 445 irtcachedis - Disable Interrupt Remapping Table (IRT) caching. 446 nohugepages - Limit page-sizes used for v1 page-tables 447 to 4 KiB. 448 v2_pgsizes_only - Limit page-sizes used for v1 page-tables 449 to 4KiB/2Mib/1GiB. 450 451 452 amd_iommu_dump= [HW,X86-64] 453 Enable AMD IOMMU driver option to dump the ACPI table 454 for AMD IOMMU. With this option enabled, AMD IOMMU 455 driver will print ACPI tables for AMD IOMMU during 456 IOMMU initialization. 457 458 amd_iommu_intr= [HW,X86-64] 459 Specifies one of the following AMD IOMMU interrupt 460 remapping modes: 461 legacy - Use legacy interrupt remapping mode. 462 vapic - Use virtual APIC mode, which allows IOMMU 463 to inject interrupts directly into guest. 464 This mode requires kvm-amd.avic=1. 465 (Default when IOMMU HW support is present.) 466 467 amd_pstate= [X86,EARLY] 468 disable 469 Do not enable amd_pstate as the default 470 scaling driver for the supported processors 471 passive 472 Use amd_pstate with passive mode as a scaling driver. 473 In this mode autonomous selection is disabled. 474 Driver requests a desired performance level and platform 475 tries to match the same performance level if it is 476 satisfied by guaranteed performance level. 477 active 478 Use amd_pstate_epp driver instance as the scaling driver, 479 driver provides a hint to the hardware if software wants 480 to bias toward performance (0x0) or energy efficiency (0xff) 481 to the CPPC firmware. then CPPC power algorithm will 482 calculate the runtime workload and adjust the realtime cores 483 frequency. 484 guided 485 Activate guided autonomous mode. Driver requests minimum and 486 maximum performance level and the platform autonomously 487 selects a performance level in this range and appropriate 488 to the current workload. 489 490 amd_prefcore= 491 [X86] 492 disable 493 Disable amd-pstate preferred core. 494 495 amijoy.map= [HW,JOY] Amiga joystick support 496 Map of devices attached to JOY0DAT and JOY1DAT 497 Format: <a>,<b> 498 See also Documentation/input/joydev/joystick.rst 499 500 analog.map= [HW,JOY] Analog joystick and gamepad support 501 Specifies type or capabilities of an analog joystick 502 connected to one of 16 gameports 503 Format: <type1>,<type2>,..<type16> 504 505 apc= [HW,SPARC] 506 Power management functions (SPARCstation-4/5 + deriv.) 507 Format: noidle 508 Disable APC CPU standby support. SPARCstation-Fox does 509 not play well with APC CPU idle - disable it if you have 510 APC and your system crashes randomly. 511 512 apic [APIC,X86-64] Use IO-APIC. Default. 513 514 apic= [APIC,X86,EARLY] Advanced Programmable Interrupt Controller 515 Change the output verbosity while booting 516 Format: { quiet (default) | verbose | debug } 517 Change the amount of debugging information output 518 when initialising the APIC and IO-APIC components. 519 520 apic_extnmi= [APIC,X86,EARLY] External NMI delivery setting 521 Format: { bsp (default) | all | none } 522 bsp: External NMI is delivered only to CPU 0 523 all: External NMIs are broadcast to all CPUs as a 524 backup of CPU 0 525 none: External NMI is masked for all CPUs. This is 526 useful so that a dump capture kernel won't be 527 shot down by NMI 528 529 apicpmtimer Do APIC timer calibration using the pmtimer. Implies 530 apicmaintimer. Useful when your PIT timer is totally 531 broken. 532 533 autoconf= [IPV6] 534 See Documentation/networking/ipv6.rst. 535 536 apm= [APM] Advanced Power Management 537 See header of arch/x86/kernel/apm_32.c. 538 539 apparmor= [APPARMOR] Disable or enable AppArmor at boot time 540 Format: { "0" | "1" } 541 See security/apparmor/Kconfig help text 542 0 -- disable. 543 1 -- enable. 544 Default value is set via kernel config option. 545 546 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards 547 Format: <io>,<irq>,<nodeID> 548 549 arm64.no32bit_el0 [ARM64] Unconditionally disable the execution of 550 32 bit applications. 551 552 arm64.nobti [ARM64] Unconditionally disable Branch Target 553 Identification support 554 555 arm64.nogcs [ARM64] Unconditionally disable Guarded Control Stack 556 support 557 558 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory 559 Set instructions support 560 561 arm64.nompam [ARM64] Unconditionally disable Memory Partitioning And 562 Monitoring support 563 564 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension 565 support 566 567 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication 568 support 569 570 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix 571 Extension support 572 573 arm64.nosve [ARM64] Unconditionally disable Scalable Vector 574 Extension support 575 576 ataflop= [HW,M68k] 577 578 atarimouse= [HW,MOUSE] Atari Mouse 579 580 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess, 581 EzKey and similar keyboards 582 583 atkbd.reset= [HW] Reset keyboard during initialization 584 585 atkbd.set= [HW] Select keyboard code set 586 Format: <int> (2 = AT (default), 3 = PS/2) 587 588 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar 589 keyboards 590 591 atkbd.softraw= [HW] Choose between synthetic and real raw mode 592 Format: <bool> (0 = real, 1 = synthetic (default)) 593 594 atkbd.softrepeat= [HW] 595 Use software keyboard repeat 596 597 audit= [KNL] Enable the audit sub-system 598 Format: { "0" | "1" | "off" | "on" } 599 0 | off - kernel audit is disabled and can not be 600 enabled until the next reboot 601 unset - kernel audit is initialized but disabled and 602 will be fully enabled by the userspace auditd. 603 1 | on - kernel audit is initialized and partially 604 enabled, storing at most audit_backlog_limit 605 messages in RAM until it is fully enabled by the 606 userspace auditd. 607 Default: unset 608 609 audit_backlog_limit= [KNL] Set the audit queue size limit. 610 Format: <int> (must be >=0) 611 Default: 64 612 613 bau= [X86_UV] Enable the BAU on SGI UV. The default 614 behavior is to disable the BAU (i.e. bau=0). 615 Format: { "0" | "1" } 616 0 - Disable the BAU. 617 1 - Enable the BAU. 618 unset - Disable the BAU. 619 620 baycom_epp= [HW,AX25] 621 Format: <io>,<mode> 622 623 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem 624 Format: <io>,<mode> 625 See header of drivers/net/hamradio/baycom_par.c. 626 627 baycom_ser_fdx= [HW,AX25] 628 BayCom Serial Port AX.25 Modem (Full Duplex Mode) 629 Format: <io>,<irq>,<mode>[,<baud>] 630 See header of drivers/net/hamradio/baycom_ser_fdx.c. 631 632 baycom_ser_hdx= [HW,AX25] 633 BayCom Serial Port AX.25 Modem (Half Duplex Mode) 634 Format: <io>,<irq>,<mode> 635 See header of drivers/net/hamradio/baycom_ser_hdx.c. 636 637 bdev_allow_write_mounted= 638 Format: <bool> 639 Control the ability to open a mounted block device 640 for writing, i.e., allow / disallow writes that bypass 641 the FS. This was implemented as a means to prevent 642 fuzzers from crashing the kernel by overwriting the 643 metadata underneath a mounted FS without its awareness. 644 This also prevents destructive formatting of mounted 645 filesystems by naive storage tooling that don't use 646 O_EXCL. Default is Y and can be changed through the 647 Kconfig option CONFIG_BLK_DEV_WRITE_MOUNTED. 648 649 bert_disable [ACPI] 650 Disable BERT OS support on buggy BIOSes. 651 652 bgrt_disable [ACPI,X86,EARLY] 653 Disable BGRT to avoid flickering OEM logo. 654 655 blkdevparts= Manual partition parsing of block device(s) for 656 embedded devices based on command line input. 657 See Documentation/block/cmdline-partition.rst 658 659 boot_delay= [KNL,EARLY] 660 Milliseconds to delay each printk during boot. 661 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled, 662 and you may also have to specify "lpj=". Boot_delay 663 values larger than 10 seconds (10000) are assumed 664 erroneous and ignored. 665 Format: integer 666 667 bootconfig [KNL,EARLY] 668 Extended command line options can be added to an initrd 669 and this will cause the kernel to look for it. 670 671 See Documentation/admin-guide/bootconfig.rst 672 673 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards) 674 bttv.radio= Most important insmod options are available as 675 kernel args too. 676 bttv.pll= See Documentation/admin-guide/media/bttv.rst 677 bttv.tuner= 678 679 bulk_remove=off [PPC] This parameter disables the use of the pSeries 680 firmware feature for flushing multiple hpte entries 681 at a time. 682 683 c101= [NET] Moxa C101 synchronous serial card 684 685 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection. 686 Sometimes CPU hardware bugs make them report the cache 687 size incorrectly. The kernel will attempt work arounds 688 to fix known problems, but for some CPUs it is not 689 possible to determine what the correct size should be. 690 This option provides an override for these situations. 691 692 carrier_timeout= 693 [NET] Specifies amount of time (in seconds) that 694 the kernel should wait for a network carrier. By default 695 it waits 120 seconds. 696 697 ca_keys= [KEYS] This parameter identifies a specific key(s) on 698 the system trusted keyring to be used for certificate 699 trust validation. 700 format: { id:<keyid> | builtin } 701 702 cca= [MIPS,EARLY] Override the kernel pages' cache coherency 703 algorithm. Accepted values range from 0 to 7 704 inclusive. See arch/mips/include/asm/pgtable-bits.h 705 for platform specific values (SB1, Loongson3 and 706 others). 707 708 ccw_timeout_log [S390] 709 See Documentation/arch/s390/common_io.rst for details. 710 711 cfi= [X86-64] Set Control Flow Integrity checking features 712 when CONFIG_FINEIBT is enabled. 713 Format: feature[,feature...] 714 Default: auto 715 716 auto: Use FineIBT if IBT available, otherwise kCFI. 717 Under FineIBT, enable "paranoid" mode when 718 FRED is not available. 719 off: Turn off CFI checking. 720 kcfi: Use kCFI (disable FineIBT). 721 fineibt: Use FineIBT (even if IBT not available). 722 norand: Do not re-randomize CFI hashes. 723 paranoid: Add caller hash checking under FineIBT. 724 bhi: Enable register poisoning to stop speculation 725 across FineIBT. (Disabled by default.) 726 warn: Do not enforce CFI checking: warn only. 727 debug: Report CFI initialization details. 728 729 cgroup_disable= [KNL] Disable a particular controller or optional feature 730 Format: {name of the controller(s) or feature(s) to disable} 731 The effects of cgroup_disable=foo are: 732 - foo isn't auto-mounted if you mount all cgroups in 733 a single hierarchy 734 - foo isn't visible as an individually mountable 735 subsystem 736 - if foo is an optional feature then the feature is 737 disabled and corresponding cgroup files are not 738 created 739 {Currently only "memory" controller deal with this and 740 cut the overhead, others just disable the usage. So 741 only cgroup_disable=memory is actually worthy} 742 Specifying "pressure" disables per-cgroup pressure 743 stall information accounting feature 744 745 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1 746 Format: { { controller | "all" | "named" } 747 [,{ controller | "all" | "named" }...] } 748 Like cgroup_disable, but only applies to cgroup v1; 749 the blacklisted controllers remain available in cgroup2. 750 "all" blacklists all controllers and "named" disables 751 named mounts. Specifying both "all" and "named" disables 752 all v1 hierarchies. 753 754 cgroup_v1_proc= [KNL] Show also missing controllers in /proc/cgroups 755 Format: { "true" | "false" } 756 /proc/cgroups lists only v1 controllers by default. 757 This compatibility option enables listing also v2 758 controllers (whose v1 code is not compiled!), so that 759 semi-legacy software can check this file to decide 760 about usage of v2 (sic) controllers. 761 762 cgroup_favordynmods= [KNL] Enable or Disable favordynmods. 763 Format: { "true" | "false" } 764 Defaults to the value of CONFIG_CGROUP_FAVOR_DYNMODS. 765 766 cgroup.memory= [KNL] Pass options to the cgroup memory controller. 767 Format: <string> 768 nosocket -- Disable socket memory accounting. 769 nokmem -- Disable kernel memory accounting. 770 nobpf -- Disable BPF memory accounting. 771 772 check_pages= [MM,EARLY] Enable sanity checking of pages after 773 allocations / before freeing. This adds checks to catch 774 double-frees, use-after-frees, and other sources of 775 page corruption by inspecting page internals (flags, 776 mapcount/refcount, memcg_data, etc.). 777 Format: { "0" | "1" } 778 Default: 0 (1 if CONFIG_DEBUG_VM is set) 779 780 checkreqprot= [SELINUX] Set initial checkreqprot flag value. 781 Format: { "0" | "1" } 782 See security/selinux/Kconfig help text. 783 0 -- check protection applied by kernel (includes 784 any implied execute protection). 785 1 -- check protection requested by application. 786 Default value is set via a kernel config option. 787 Value can be changed at runtime via 788 /sys/fs/selinux/checkreqprot. 789 Setting checkreqprot to 1 is deprecated. 790 791 cio_ignore= [S390] 792 See Documentation/arch/s390/common_io.rst for details. 793 794 clearcpuid=X[,X...] [X86] 795 Disable CPUID feature X for the kernel. See 796 arch/x86/include/asm/cpufeatures.h for the valid bit 797 numbers X. Note the Linux-specific bits are not necessarily 798 stable over kernel options, but the vendor-specific 799 ones should be. 800 X can also be a string as appearing in the flags: line 801 in /proc/cpuinfo which does not have the above 802 instability issue. However, not all features have names 803 in /proc/cpuinfo. 804 Note that using this option will taint your kernel. 805 Also note that user programs calling CPUID directly 806 or using the feature without checking anything 807 will still see it. This just prevents it from 808 being used by the kernel or shown in /proc/cpuinfo. 809 Also note the kernel might malfunction if you disable 810 some critical bits. 811 812 clk_ignore_unused 813 [CLK] 814 Prevents the clock framework from automatically gating 815 clocks that have not been explicitly enabled by a Linux 816 device driver but are enabled in hardware at reset or 817 by the bootloader/firmware. Note that this does not 818 force such clocks to be always-on nor does it reserve 819 those clocks in any way. This parameter is useful for 820 debug and development, but should not be needed on a 821 platform with proper driver support. For more 822 information, see Documentation/driver-api/clk.rst. 823 824 clock= [BUGS=X86-32, HW] gettimeofday clocksource override. 825 [Deprecated] 826 Forces specified clocksource (if available) to be used 827 when calculating gettimeofday(). If specified 828 clocksource is not available, it defaults to PIT. 829 Format: { pit | tsc | cyclone | pmtmr } 830 831 clocksource= Override the default clocksource 832 Format: <string> 833 Override the default clocksource and use the clocksource 834 with the name specified. 835 Some clocksource names to choose from, depending on 836 the platform: 837 [all] jiffies (this is the base, fallback clocksource) 838 [ACPI] acpi_pm 839 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2, 840 pxa_timer,timer3,32k_counter,timer0_1 841 [X86-32] pit,hpet,tsc; 842 scx200_hrt on Geode; cyclone on IBM x440 843 [MIPS] MIPS 844 [PARISC] cr16 845 [S390] tod 846 [SH] SuperH 847 [SPARC64] tick 848 [X86-64] hpet,tsc 849 850 clocksource.arm_arch_timer.evtstrm= 851 [ARM,ARM64,EARLY] 852 Format: <bool> 853 Enable/disable the eventstream feature of the ARM 854 architected timer so that code using WFE-based polling 855 loops can be debugged more effectively on production 856 systems. 857 858 clocksource.verify_n_cpus= [KNL] 859 Limit the number of CPUs checked for clocksources 860 marked with CLOCK_SOURCE_VERIFY_PERCPU that 861 are marked unstable due to excessive skew. 862 A negative value says to check all CPUs, while 863 zero says not to check any. Values larger than 864 nr_cpu_ids are silently truncated to nr_cpu_ids. 865 The actual CPUs are chosen randomly, with 866 no replacement if the same CPU is chosen twice. 867 868 clocksource-wdtest.holdoff= [KNL] 869 Set the time in seconds that the clocksource 870 watchdog test waits before commencing its tests. 871 Defaults to zero when built as a module and to 872 10 seconds when built into the kernel. 873 874 cma=nn[MG]@[start[MG][-end[MG]]] 875 [KNL,CMA,EARLY] 876 Sets the size of kernel global memory area for 877 contiguous memory allocations and optionally the 878 placement constraint by the physical address range of 879 memory allocations. A value of 0 disables CMA 880 altogether. For more information, see 881 kernel/dma/contiguous.c 882 883 cma_pernuma=nn[MG] 884 [KNL,CMA,EARLY] 885 Sets the size of kernel per-numa memory area for 886 contiguous memory allocations. A value of 0 disables 887 per-numa CMA altogether. And If this option is not 888 specified, the default value is 0. 889 With per-numa CMA enabled, DMA users on node nid will 890 first try to allocate buffer from the pernuma area 891 which is located in node nid, if the allocation fails, 892 they will fallback to the global default memory area. 893 894 numa_cma=<node>:nn[MG][,<node>:nn[MG]] 895 [KNL,CMA,EARLY] 896 Sets the size of kernel numa memory area for 897 contiguous memory allocations. It will reserve CMA 898 area for the specified node. 899 900 With numa CMA enabled, DMA users on node nid will 901 first try to allocate buffer from the numa area 902 which is located in node nid, if the allocation fails, 903 they will fallback to the global default memory area. 904 905 cmo_free_hint= [PPC] Format: { yes | no } 906 Specify whether pages are marked as being inactive 907 when they are freed. This is used in CMO environments 908 to determine OS memory pressure for page stealing by 909 a hypervisor. 910 Default: yes 911 912 coherent_pool=nn[KMG] [ARM,KNL,EARLY] 913 Sets the size of memory pool for coherent, atomic dma 914 allocations, by default set to 256K. 915 916 com20020= [HW,NET] ARCnet - COM20020 chipset 917 Format: 918 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]] 919 920 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers) 921 Format: <io>[,<irq>] 922 923 com90xx= [HW,NET] 924 ARCnet - COM90xx chipset (memory-mapped buffers) 925 Format: <io>[,<irq>[,<memstart>]] 926 927 condev= [HW,S390] console device 928 conmode= 929 930 con3215_drop= [S390,EARLY] 3215 console drop mode. 931 Format: y|n|Y|N|1|0 932 When set to true, drop data on the 3215 console when 933 the console buffer is full. In this case the 934 operator using a 3270 terminal emulator (for example 935 x3270) does not have to enter the clear key for the 936 console output to advance and the kernel to continue. 937 This leads to a much faster boot time when a 3270 938 terminal emulator is active. If no 3270 terminal 939 emulator is used, this parameter has no effect. 940 941 console= [KNL] Output console device and options. 942 943 tty<n> Use the virtual console device <n>. 944 945 ttyS<n>[,options] 946 ttyUSB0[,options] 947 Use the specified serial port. The options are of 948 the form "bbbbpnf", where "bbbb" is the baud rate, 949 "p" is parity ("n", "o", or "e"), "n" is number of 950 bits, and "f" is flow control ("r" for RTS or 951 omit it). Default is "9600n8". 952 953 See Documentation/admin-guide/serial-console.rst for more 954 information. See 955 Documentation/networking/netconsole.rst for an 956 alternative. 957 958 <DEVNAME>:<n>.<n>[,options] 959 Use the specified serial port on the serial core bus. 960 The addressing uses DEVNAME of the physical serial port 961 device, followed by the serial core controller instance, 962 and the serial port instance. The options are the same 963 as documented for the ttyS addressing above. 964 965 The mapping of the serial ports to the tty instances 966 can be viewed with: 967 968 $ ls -d /sys/bus/serial-base/devices/*:*.*/tty/* 969 /sys/bus/serial-base/devices/00:04:0.0/tty/ttyS0 970 971 In the above example, the console can be addressed with 972 console=00:04:0.0. Note that a console addressed this 973 way will only get added when the related device driver 974 is ready. The use of an earlycon parameter in addition to 975 the console may be desired for console output early on. 976 977 uart[8250],io,<addr>[,options] 978 uart[8250],mmio,<addr>[,options] 979 uart[8250],mmio16,<addr>[,options] 980 uart[8250],mmio32,<addr>[,options] 981 uart[8250],0x<addr>[,options] 982 Start an early, polled-mode console on the 8250/16550 983 UART at the specified I/O port or MMIO address, 984 switching to the matching ttyS device later. 985 MMIO inter-register address stride is either 8-bit 986 (mmio), 16-bit (mmio16), or 32-bit (mmio32). 987 If none of [io|mmio|mmio16|mmio32], <addr> is assumed 988 to be equivalent to 'mmio'. 'options' are specified in 989 the same format described for ttyS above; if unspecified, 990 the h/w is not re-initialized. 991 992 hvc<n> Use the hypervisor console device <n>. This is for 993 both Xen and PowerPC hypervisors. 994 995 { null | "" } 996 Use to disable console output, i.e., to have kernel 997 console messages discarded. 998 This must be the only console= parameter used on the 999 kernel command line. 1000 1001 If the device connected to the port is not a TTY but a braille 1002 device, prepend "brl," before the device type, for instance 1003 console=brl,ttyS0 1004 For now, only VisioBraille is supported. 1005 1006 console_msg_format= 1007 [KNL] Change console messages format 1008 default 1009 By default we print messages on consoles in 1010 "[time stamp] text\n" format (time stamp may not be 1011 printed, depending on CONFIG_PRINTK_TIME or 1012 `printk_time' param). 1013 syslog 1014 Switch to syslog format: "<%u>[time stamp] text\n" 1015 IOW, each message will have a facility and loglevel 1016 prefix. The format is similar to one used by syslog() 1017 syscall, or to executing "dmesg -S --raw" or to reading 1018 from /proc/kmsg. 1019 1020 consoleblank= [KNL] The console blank (screen saver) timeout in 1021 seconds. A value of 0 disables the blank timer. 1022 Defaults to 0. 1023 1024 coredump_filter= 1025 [KNL] Change the default value for 1026 /proc/<pid>/coredump_filter. 1027 See also Documentation/filesystems/proc.rst. 1028 1029 coresight_cpu_debug.enable 1030 [ARM,ARM64] 1031 Format: <bool> 1032 Enable/disable the CPU sampling based debugging. 1033 0: default value, disable debugging 1034 1: enable debugging at boot time 1035 1036 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver 1037 Format: 1038 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>] 1039 1040 cpuidle.off=1 [CPU_IDLE] 1041 disable the cpuidle sub-system 1042 1043 cpuidle.governor= 1044 [CPU_IDLE] Name of the cpuidle governor to use. 1045 1046 cpufreq.off=1 [CPU_FREQ] 1047 disable the cpufreq sub-system 1048 1049 cpufreq.default_governor= 1050 [CPU_FREQ] Name of the default cpufreq governor or 1051 policy to use. This governor must be registered in the 1052 kernel before the cpufreq driver probes. 1053 1054 cpu_init_udelay=N 1055 [X86,EARLY] Delay for N microsec between assert and de-assert 1056 of APIC INIT to start processors. This delay occurs 1057 on every CPU online, such as boot, and resume from suspend. 1058 Default: 10000 1059 1060 cpuhp.parallel= 1061 [SMP] Enable/disable parallel bringup of secondary CPUs 1062 Format: <bool> 1063 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise 1064 the parameter has no effect. 1065 1066 crash_kexec_post_notifiers 1067 Only jump to kdump kernel after running the panic 1068 notifiers and dumping kmsg. This option increases 1069 the risks of a kdump failure, since some panic 1070 notifiers can make the crashed kernel more unstable. 1071 In configurations where kdump may not be reliable, 1072 running the panic notifiers could allow collecting 1073 more data on dmesg, like stack traces from other CPUS 1074 or extra data dumped by panic_print. Note that some 1075 configurations enable this option unconditionally, 1076 like Hyper-V, PowerPC (fadump) and AMD SEV-SNP. 1077 1078 crashkernel=size[KMG][@offset[KMG]] 1079 [KNL,EARLY] Using kexec, Linux can switch to a 'crash kernel' 1080 upon panic. This parameter reserves the physical 1081 memory region [offset, offset + size] for that kernel 1082 image. If '@offset' is omitted, then a suitable offset 1083 is selected automatically. 1084 [KNL, X86-64, ARM64, RISCV, LoongArch] Select a region 1085 under 4G first, and fall back to reserve region above 1086 4G when '@offset' hasn't been specified. 1087 See Documentation/admin-guide/kdump/kdump.rst for further details. 1088 1089 crashkernel=range1:size1[,range2:size2,...][@offset] 1090 [KNL] Same as above, but depends on the memory 1091 in the running system. The syntax of range is 1092 start-[end] where start and end are both 1093 a memory unit (amount[KMG]). See also 1094 Documentation/admin-guide/kdump/kdump.rst for an example. 1095 1096 crashkernel=size[KMG],high 1097 [KNL, X86-64, ARM64, RISCV, LoongArch] range could be 1098 above 4G. 1099 Allow kernel to allocate physical memory region from top, 1100 so could be above 4G if system have more than 4G ram 1101 installed. Otherwise memory region will be allocated 1102 below 4G, if available. 1103 It will be ignored if crashkernel=X is specified. 1104 crashkernel=size[KMG],low 1105 [KNL, X86-64, ARM64, RISCV, LoongArch] range under 4G. 1106 When crashkernel=X,high is passed, kernel could allocate 1107 physical memory region above 4G, that cause second kernel 1108 crash on system that require some amount of low memory, 1109 e.g. swiotlb requires at least 64M+32K low memory, also 1110 enough extra low memory is needed to make sure DMA buffers 1111 for 32-bit devices won't run out. Kernel would try to allocate 1112 default size of memory below 4G automatically. The default 1113 size is platform dependent. 1114 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB) 1115 --> arm64: 128MiB 1116 --> riscv: 128MiB 1117 --> loongarch: 128MiB 1118 This one lets the user specify own low range under 4G 1119 for second kernel instead. 1120 0: to disable low allocation. 1121 It will be ignored when crashkernel=X,high is not used 1122 or memory reserved is below 4G. 1123 crashkernel=size[KMG],cma 1124 [KNL, X86, ppc] Reserve additional crash kernel memory from 1125 CMA. This reservation is usable by the first system's 1126 userspace memory and kernel movable allocations (memory 1127 balloon, zswap). Pages allocated from this memory range 1128 will not be included in the vmcore so this should not 1129 be used if dumping of userspace memory is intended and 1130 it has to be expected that some movable kernel pages 1131 may be missing from the dump. 1132 1133 A standard crashkernel reservation, as described above, 1134 is still needed to hold the crash kernel and initrd. 1135 1136 This option increases the risk of a kdump failure: DMA 1137 transfers configured by the first kernel may end up 1138 corrupting the second kernel's memory. 1139 1140 This reservation method is intended for systems that 1141 can't afford to sacrifice enough memory for standard 1142 crashkernel reservation and where less reliable and 1143 possibly incomplete kdump is preferable to no kdump at 1144 all. 1145 1146 cryptomgr.notests 1147 [KNL] Disable crypto self-tests 1148 1149 cs89x0_dma= [HW,NET] 1150 Format: <dma> 1151 1152 cs89x0_media= [HW,NET] 1153 Format: { rj45 | aui | bnc } 1154 1155 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU 1156 function call handling. When switched on, 1157 additional debug data is printed to the console 1158 in case a hanging CPU is detected, and that 1159 CPU is pinged again in order to try to resolve 1160 the hang situation. The default value of this 1161 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT 1162 Kconfig option. 1163 1164 dasd= [HW,NET] 1165 See header of drivers/s390/block/dasd_devmap.c. 1166 1167 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port 1168 (one device per port) 1169 Format: <port#>,<type> 1170 See also Documentation/input/devices/joystick-parport.rst 1171 1172 debug [KNL,EARLY] Enable kernel debugging (events log level). 1173 1174 debug_boot_weak_hash 1175 [KNL,EARLY] Enable printing [hashed] pointers early in the 1176 boot sequence. If enabled, we use a weak hash instead 1177 of siphash to hash pointers. Use this option if you are 1178 seeing instances of '(___ptrval___)') and need to see a 1179 value (hashed pointer) instead. Cryptographically 1180 insecure, please do not use on production kernels. 1181 1182 debug_locks_verbose= 1183 [KNL] verbose locking self-tests 1184 Format: <int> 1185 Print debugging info while doing the locking API 1186 self-tests. 1187 Bitmask for the various LOCKTYPE_ tests. Defaults to 0 1188 (no extra messages), setting it to -1 (all bits set) 1189 will print _a_lot_ more information - normally only 1190 useful to lockdep developers. 1191 1192 debug_objects [KNL,EARLY] Enable object debugging 1193 1194 debug_guardpage_minorder= 1195 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this 1196 parameter allows control of the order of pages that will 1197 be intentionally kept free (and hence protected) by the 1198 buddy allocator. Bigger value increase the probability 1199 of catching random memory corruption, but reduce the 1200 amount of memory for normal system use. The maximum 1201 possible value is MAX_PAGE_ORDER/2. Setting this 1202 parameter to 1 or 2 should be enough to identify most 1203 random memory corruption problems caused by bugs in 1204 kernel or driver code when a CPU writes to (or reads 1205 from) a random memory location. Note that there exists 1206 a class of memory corruptions problems caused by buggy 1207 H/W or F/W or by drivers badly programming DMA 1208 (basically when memory is written at bus level and the 1209 CPU MMU is bypassed) which are not detectable by 1210 CONFIG_DEBUG_PAGEALLOC, hence this option will not 1211 help tracking down these problems. 1212 1213 debug_pagealloc= 1214 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this parameter 1215 enables the feature at boot time. By default, it is 1216 disabled and the system will work mostly the same as a 1217 kernel built without CONFIG_DEBUG_PAGEALLOC. 1218 Note: to get most of debug_pagealloc error reports, it's 1219 useful to also enable the page_owner functionality. 1220 on: enable the feature 1221 1222 debugfs= [KNL,EARLY] This parameter enables what is exposed to 1223 userspace and debugfs internal clients. 1224 Format: { on, off } 1225 on: All functions are enabled. 1226 off: Filesystem is not registered and clients 1227 get a -EPERM as result when trying to register files 1228 or directories within debugfs. 1229 This is equivalent of the runtime functionality if 1230 debugfs was not enabled in the kernel at all. 1231 Default value is set in build-time with a kernel configuration. 1232 1233 debugpat [X86] Enable PAT debugging 1234 1235 default_hugepagesz= 1236 [HW] The size of the default HugeTLB page. This is 1237 the size represented by the legacy /proc/ hugepages 1238 APIs. In addition, this is the default hugetlb size 1239 used for shmget(), mmap() and mounting hugetlbfs 1240 filesystems. If not specified, defaults to the 1241 architecture's default huge page size. Huge page 1242 sizes are architecture dependent. See also 1243 Documentation/admin-guide/mm/hugetlbpage.rst. 1244 Format: size[KMG] 1245 1246 deferred_probe_timeout= 1247 [KNL] Debugging option to set a timeout in seconds for 1248 deferred probe to give up waiting on dependencies to 1249 probe. Only specific dependencies (subsystems or 1250 drivers) that have opted in will be ignored. A timeout 1251 of 0 will timeout at the end of initcalls. If the time 1252 out hasn't expired, it'll be restarted by each 1253 successful driver registration. This option will also 1254 dump out devices still on the deferred probe list after 1255 retrying. 1256 1257 delayacct [KNL] Enable per-task delay accounting 1258 1259 dell_smm_hwmon.ignore_dmi= 1260 [HW] Continue probing hardware even if DMI data 1261 indicates that the driver is running on unsupported 1262 hardware. 1263 1264 dell_smm_hwmon.force= 1265 [HW] Activate driver even if SMM BIOS signature does 1266 not match list of supported models and enable otherwise 1267 blacklisted features. 1268 1269 dell_smm_hwmon.power_status= 1270 [HW] Report power status in /proc/i8k 1271 (disabled by default). 1272 1273 dell_smm_hwmon.restricted= 1274 [HW] Allow controlling fans only if SYS_ADMIN 1275 capability is set. 1276 1277 dell_smm_hwmon.fan_mult= 1278 [HW] Factor to multiply fan speed with. 1279 1280 dell_smm_hwmon.fan_max= 1281 [HW] Maximum configurable fan speed. 1282 1283 dfltcc= [HW,S390] 1284 Format: { on | off | def_only | inf_only | always } 1285 on: s390 zlib hardware support for compression on 1286 level 1 and decompression (default) 1287 off: No s390 zlib hardware support 1288 def_only: s390 zlib hardware support for deflate 1289 only (compression on level 1) 1290 inf_only: s390 zlib hardware support for inflate 1291 only (decompression) 1292 always: Same as 'on' but ignores the selected compression 1293 level always using hardware support (used for debugging) 1294 1295 dhash_entries= [KNL] 1296 Set number of hash buckets for dentry cache. 1297 1298 disable_1tb_segments [PPC,EARLY] 1299 Disables the use of 1TB hash page table segments. This 1300 causes the kernel to fall back to 256MB segments which 1301 can be useful when debugging issues that require an SLB 1302 miss to occur. 1303 1304 disable= [IPV6] 1305 See Documentation/networking/ipv6.rst. 1306 1307 disable_radix [PPC,EARLY] 1308 Disable RADIX MMU mode on POWER9 1309 1310 disable_tlbie [PPC] 1311 Disable TLBIE instruction. Currently does not work 1312 with KVM, with HASH MMU, or with coherent accelerators. 1313 1314 disable_ddw [PPC/PSERIES,EARLY] 1315 Disable Dynamic DMA Window support. Use this 1316 to workaround buggy firmware. 1317 1318 disable_ipv6= [IPV6] 1319 See Documentation/networking/ipv6.rst. 1320 1321 disable_mtrr_cleanup [X86,EARLY] 1322 The kernel tries to adjust MTRR layout from continuous 1323 to discrete, to make X server driver able to add WB 1324 entry later. This parameter disables that. 1325 1326 disable_mtrr_trim [X86, Intel and AMD only,EARLY] 1327 By default the kernel will trim any uncacheable 1328 memory out of your available memory pool based on 1329 MTRR settings. This parameter disables that behavior, 1330 possibly causing your machine to run very slowly. 1331 1332 disable_timer_pin_1 [X86,EARLY] 1333 Disable PIN 1 of APIC timer 1334 Can be useful to work around chipset bugs. 1335 1336 dis_ucode_ldr [X86] Disable the microcode loader. 1337 1338 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support, 1339 this option disables the debugging code at boot. 1340 1341 dma_debug_entries=<number> 1342 This option allows to tune the number of preallocated 1343 entries for DMA-API debugging code. One entry is 1344 required per DMA-API allocation. Use this if the 1345 DMA-API debugging code disables itself because the 1346 architectural default is too low. 1347 1348 dma_debug_driver=<driver_name> 1349 With this option the DMA-API debugging driver 1350 filter feature can be enabled at boot time. Just 1351 pass the driver to filter for as the parameter. 1352 The filter can be disabled or changed to another 1353 driver later using sysfs. 1354 1355 reg_file_data_sampling= 1356 [X86] Controls mitigation for Register File Data 1357 Sampling (RFDS) vulnerability. RFDS is a CPU 1358 vulnerability which may allow userspace to infer 1359 kernel data values previously stored in floating point 1360 registers, vector registers, or integer registers. 1361 RFDS only affects Intel Atom processors. 1362 1363 on: Turns ON the mitigation. 1364 off: Turns OFF the mitigation. 1365 1366 This parameter overrides the compile time default set 1367 by CONFIG_MITIGATION_RFDS. Mitigation cannot be 1368 disabled when other VERW based mitigations (like MDS) 1369 are enabled. In order to disable RFDS mitigation all 1370 VERW based mitigations need to be disabled. 1371 1372 For details see: 1373 Documentation/admin-guide/hw-vuln/reg-file-data-sampling.rst 1374 1375 dm_verity.keyring_unsealed= 1376 [KNL] When set to 1, leave the dm-verity keyring 1377 unsealed after initialization so userspace can 1378 provision keys. Once the keyring is restricted 1379 it becomes active and is searched during signature 1380 verification. 1381 1382 driver_async_probe= [KNL] 1383 List of driver names to be probed asynchronously. * 1384 matches with all driver names. If * is specified, the 1385 rest of the listed driver names are those that will NOT 1386 match the *. 1387 Format: <driver_name1>,<driver_name2>... 1388 1389 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>] 1390 Broken monitors, graphic adapters, KVMs and EDIDless 1391 panels may send no or incorrect EDID data sets. 1392 This parameter allows to specify an EDID data sets 1393 in the /lib/firmware directory that are used instead. 1394 An EDID data set will only be used for a particular 1395 connector, if its name and a colon are prepended to 1396 the EDID name. Each connector may use a unique EDID 1397 data set by separating the files with a comma. An EDID 1398 data set with no connector name will be used for 1399 any connectors not explicitly specified. 1400 1401 dscc4.setup= [NET] 1402 1403 dt_cpu_ftrs= [PPC,EARLY] 1404 Format: {"off" | "known"} 1405 Control how the dt_cpu_ftrs device-tree binding is 1406 used for CPU feature discovery and setup (if it 1407 exists). 1408 off: Do not use it, fall back to legacy cpu table. 1409 known: Do not pass through unknown features to guests 1410 or userspace, only those that the kernel is aware of. 1411 1412 dump_apple_properties [X86] 1413 Dump name and content of EFI device properties on 1414 x86 Macs. Useful for driver authors to determine 1415 what data is available or for reverse-engineering. 1416 1417 dyndbg[="val"] [KNL,DYNAMIC_DEBUG] 1418 <module>.dyndbg[="val"] 1419 Enable debug messages at boot time. See 1420 Documentation/admin-guide/dynamic-debug-howto.rst 1421 for details. 1422 1423 early_ioremap_debug [KNL,EARLY] 1424 Enable debug messages in early_ioremap support. This 1425 is useful for tracking down temporary early mappings 1426 which are not unmapped. 1427 1428 earlycon= [KNL,EARLY] Output early console device and options. 1429 1430 When used with no options, the early console is 1431 determined by stdout-path property in device tree's 1432 chosen node or the ACPI SPCR table if supported by 1433 the platform. 1434 1435 cdns,<addr>[,options] 1436 Start an early, polled-mode console on a Cadence 1437 (xuartps) serial port at the specified address. Only 1438 supported option is baud rate. If baud rate is not 1439 specified, the serial port must already be setup and 1440 configured. 1441 1442 uart[8250],io,<addr>[,options[,uartclk]] 1443 uart[8250],mmio,<addr>[,options[,uartclk]] 1444 uart[8250],mmio32,<addr>[,options[,uartclk]] 1445 uart[8250],mmio32be,<addr>[,options[,uartclk]] 1446 uart[8250],0x<addr>[,options] 1447 Start an early, polled-mode console on the 8250/16550 1448 UART at the specified I/O port or MMIO address. 1449 MMIO inter-register address stride is either 8-bit 1450 (mmio) or 32-bit (mmio32 or mmio32be). 1451 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed 1452 to be equivalent to 'mmio'. 'options' are specified 1453 in the same format described for "console=ttyS<n>"; if 1454 unspecified, the h/w is not initialized. 'uartclk' is 1455 the uart clock frequency; if unspecified, it is set 1456 to 'BASE_BAUD' * 16. 1457 1458 pl011,<addr> 1459 pl011,mmio32,<addr> 1460 Start an early, polled-mode console on a pl011 serial 1461 port at the specified address. The pl011 serial port 1462 must already be setup and configured. Options are not 1463 yet supported. If 'mmio32' is specified, then only 1464 the driver will use only 32-bit accessors to read/write 1465 the device registers. 1466 1467 liteuart,<addr> 1468 Start an early console on a litex serial port at the 1469 specified address. The serial port must already be 1470 setup and configured. Options are not yet supported. 1471 1472 meson,<addr> 1473 Start an early, polled-mode console on a meson serial 1474 port at the specified address. The serial port must 1475 already be setup and configured. Options are not yet 1476 supported. 1477 1478 msm_serial,<addr> 1479 Start an early, polled-mode console on an msm serial 1480 port at the specified address. The serial port 1481 must already be setup and configured. Options are not 1482 yet supported. 1483 1484 msm_serial_dm,<addr> 1485 Start an early, polled-mode console on an msm serial 1486 dm port at the specified address. The serial port 1487 must already be setup and configured. Options are not 1488 yet supported. 1489 1490 owl,<addr> 1491 Start an early, polled-mode console on a serial port 1492 of an Actions Semi SoC, such as S500 or S900, at the 1493 specified address. The serial port must already be 1494 setup and configured. Options are not yet supported. 1495 1496 rda,<addr> 1497 Start an early, polled-mode console on a serial port 1498 of an RDA Micro SoC, such as RDA8810PL, at the 1499 specified address. The serial port must already be 1500 setup and configured. Options are not yet supported. 1501 1502 sbi 1503 Use RISC-V SBI (Supervisor Binary Interface) for early 1504 console. 1505 1506 smh Use ARM semihosting calls for early console. 1507 1508 s3c2410,<addr> 1509 s3c2412,<addr> 1510 s3c2440,<addr> 1511 s3c6400,<addr> 1512 s5pv210,<addr> 1513 exynos4210,<addr> 1514 Use early console provided by serial driver available 1515 on Samsung SoCs, requires selecting proper type and 1516 a correct base address of the selected UART port. The 1517 serial port must already be setup and configured. 1518 Options are not yet supported. 1519 1520 lantiq,<addr> 1521 Start an early, polled-mode console on a lantiq serial 1522 (lqasc) port at the specified address. The serial port 1523 must already be setup and configured. Options are not 1524 yet supported. 1525 1526 lpuart,<addr> 1527 lpuart32,<addr> 1528 Use early console provided by Freescale LP UART driver 1529 found on Freescale Vybrid and QorIQ LS1021A processors. 1530 A valid base address must be provided, and the serial 1531 port must already be setup and configured. 1532 1533 ec_imx21,<addr> 1534 ec_imx6q,<addr> 1535 Start an early, polled-mode, output-only console on the 1536 Freescale i.MX UART at the specified address. The UART 1537 must already be setup and configured. 1538 1539 ar3700_uart,<addr> 1540 Start an early, polled-mode console on the 1541 Armada 3700 serial port at the specified 1542 address. The serial port must already be setup 1543 and configured. Options are not yet supported. 1544 1545 qcom_geni,<addr> 1546 Start an early, polled-mode console on a Qualcomm 1547 Generic Interface (GENI) based serial port at the 1548 specified address. The serial port must already be 1549 setup and configured. Options are not yet supported. 1550 1551 efifb,[options] 1552 Start an early, unaccelerated console on the EFI 1553 memory mapped framebuffer (if available). On cache 1554 coherent non-x86 systems that use system memory for 1555 the framebuffer, pass the 'ram' option so that it is 1556 mapped with the correct attributes. 1557 1558 linflex,<addr> 1559 Use early console provided by Freescale LINFlexD UART 1560 serial driver for NXP S32V234 SoCs. A valid base 1561 address must be provided, and the serial port must 1562 already be setup and configured. 1563 1564 earlyprintk= [X86,SH,ARM,M68k,S390,UM,EARLY] 1565 earlyprintk=vga 1566 earlyprintk=sclp 1567 earlyprintk=xen 1568 earlyprintk=serial[,ttySn[,baudrate]] 1569 earlyprintk=serial[,0x...[,baudrate]] 1570 earlyprintk=ttySn[,baudrate] 1571 earlyprintk=dbgp[debugController#] 1572 earlyprintk=mmio32,membase[,{nocfg|baudrate}] 1573 earlyprintk=pciserial[,force],bus:device.function[,{nocfg|baudrate}] 1574 earlyprintk=xdbc[xhciController#] 1575 earlyprintk=bios 1576 1577 earlyprintk is useful when the kernel crashes before 1578 the normal console is initialized. It is not enabled by 1579 default because it has some cosmetic problems. 1580 1581 Use "nocfg" to skip UART configuration, assume 1582 BIOS/firmware has configured UART correctly. 1583 1584 Append ",keep" to not disable it when the real console 1585 takes over. 1586 1587 Only one of vga, serial, or usb debug port can 1588 be used at a time. 1589 1590 Currently only ttyS0 and ttyS1 may be specified by 1591 name. Other I/O ports may be explicitly specified 1592 on some architectures (x86 and arm at least) by 1593 replacing ttySn with an I/O port address, like this: 1594 earlyprintk=serial,0x1008,115200 1595 You can find the port for a given device in 1596 /proc/tty/driver/serial: 1597 2: uart:ST16650V2 port:00001008 irq:18 ... 1598 1599 Interaction with the standard serial driver is not 1600 very good. 1601 1602 The VGA output is eventually overwritten by 1603 the real console. 1604 1605 The xen option can only be used in Xen domains. 1606 1607 The sclp output can only be used on s390. 1608 1609 The bios output can only be used on SuperH. 1610 1611 The optional "force" to "pciserial" enables use of a 1612 PCI device even when its classcode is not of the 1613 UART class. 1614 1615 edac_report= [HW,EDAC] Control how to report EDAC event 1616 Format: {"on" | "off" | "force"} 1617 on: enable EDAC to report H/W event. May be overridden 1618 by other higher priority error reporting module. 1619 off: disable H/W event reporting through EDAC. 1620 force: enforce the use of EDAC to report H/W event. 1621 default: on. 1622 1623 edd= [EDD] 1624 Format: {"off" | "on" | "skip[mbr]"} 1625 1626 efi= [EFI,EARLY] 1627 Format: { "debug", "disable_early_pci_dma", 1628 "nochunk", "noruntime", "nosoftreserve", 1629 "novamap", "no_disable_early_pci_dma" } 1630 debug: enable misc debug output. 1631 disable_early_pci_dma: disable the busmaster bit on all 1632 PCI bridges while in the EFI boot stub. 1633 nochunk: disable reading files in "chunks" in the EFI 1634 boot stub, as chunking can cause problems with some 1635 firmware implementations. 1636 noruntime : disable EFI runtime services support 1637 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose) 1638 attribute may cause the kernel to reserve the 1639 memory range for a memory mapping driver to 1640 claim. Specify efi=nosoftreserve to disable this 1641 reservation and treat the memory by its base type 1642 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM"). 1643 novamap: do not call SetVirtualAddressMap(). 1644 no_disable_early_pci_dma: Leave the busmaster bit set 1645 on all PCI bridges while in the EFI boot stub 1646 1647 efi_no_storage_paranoia [EFI,X86,EARLY] 1648 Using this parameter you can use more than 50% of 1649 your efi variable storage. Use this parameter only if 1650 you are really sure that your UEFI does sane gc and 1651 fulfills the spec otherwise your board may brick. 1652 1653 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT 1654 that is to be dynamically loaded by Linux. If there are 1655 multiple variables with the same name but with different 1656 vendor GUIDs, all of them will be loaded. See 1657 Documentation/admin-guide/acpi/ssdt-overlays.rst for details. 1658 1659 1660 eisa_irq_edge= [PARISC,HW] 1661 See header of drivers/parisc/eisa.c. 1662 1663 ekgdboc= [X86,KGDB,EARLY] Allow early kernel console debugging 1664 Format: ekgdboc=kbd 1665 1666 This is designed to be used in conjunction with 1667 the boot argument: earlyprintk=vga 1668 1669 This parameter works in place of the kgdboc parameter 1670 but can only be used if the backing tty is available 1671 very early in the boot process. For early debugging 1672 via a serial port see kgdboc_earlycon instead. 1673 1674 elanfreq= [X86-32] 1675 See comment before function elanfreq_setup() in 1676 arch/x86/kernel/cpu/cpufreq/elanfreq.c. 1677 1678 elfcorehdr=[size[KMG]@]offset[KMG] [PPC,SH,X86,S390,EARLY] 1679 Specifies physical address of start of kernel core 1680 image elf header and optionally the size. Generally 1681 kexec loader will pass this option to capture kernel. 1682 See Documentation/admin-guide/kdump/kdump.rst for details. 1683 1684 enable_mtrr_cleanup [X86,EARLY] 1685 The kernel tries to adjust MTRR layout from continuous 1686 to discrete, to make X server driver able to add WB 1687 entry later. This parameter enables that. 1688 1689 enable_timer_pin_1 [X86] 1690 Enable PIN 1 of APIC timer 1691 Can be useful to work around chipset bugs 1692 (in particular on some ATI chipsets). 1693 The kernel tries to set a reasonable default. 1694 1695 enforcing= [SELINUX] Set initial enforcing status. 1696 Format: {"0" | "1"} 1697 See security/selinux/Kconfig help text. 1698 0 -- permissive (log only, no denials). 1699 1 -- enforcing (deny and log). 1700 Default value is 0. 1701 Value can be changed at runtime via 1702 /sys/fs/selinux/enforce. 1703 1704 erst_disable [ACPI] 1705 Disable Error Record Serialization Table (ERST) 1706 support. 1707 1708 ether= [HW,NET] Ethernet cards parameters 1709 This option is obsoleted by the "netdev=" option, which 1710 has equivalent usage. See its documentation for details. 1711 1712 evm= [EVM] 1713 Format: { "fix" } 1714 Permit 'security.evm' to be updated regardless of 1715 current integrity status. 1716 1717 early_page_ext [KNL,EARLY] Enforces page_ext initialization to earlier 1718 stages so cover more early boot allocations. 1719 Please note that as side effect some optimizations 1720 might be disabled to achieve that (e.g. parallelized 1721 memory initialization is disabled) so the boot process 1722 might take longer, especially on systems with a lot of 1723 memory. Available with CONFIG_PAGE_EXTENSION=y. 1724 1725 failslab= 1726 fail_usercopy= 1727 fail_page_alloc= 1728 fail_skb_realloc= 1729 fail_make_request=[KNL] 1730 General fault injection mechanism. 1731 Format: <interval>,<probability>,<space>,<times> 1732 See also Documentation/fault-injection/. 1733 1734 fb_tunnels= [NET] 1735 Format: { initns | none } 1736 See Documentation/admin-guide/sysctl/net.rst for 1737 fb_tunnels_only_for_init_ns 1738 1739 floppy= [HW] 1740 See Documentation/admin-guide/blockdev/floppy.rst. 1741 1742 forcepae [X86-32] 1743 Forcefully enable Physical Address Extension (PAE). 1744 Many Pentium M systems disable PAE but may have a 1745 functionally usable PAE implementation. 1746 Warning: use of this parameter will taint the kernel 1747 and may cause unknown problems. 1748 1749 fred= [X86-64] 1750 Enable/disable Flexible Return and Event Delivery. 1751 Format: { on | off } 1752 on: enable FRED when it's present. 1753 off: disable FRED, the default setting. 1754 1755 ftrace=[tracer] 1756 [FTRACE] will set and start the specified tracer 1757 as early as possible in order to facilitate early 1758 boot debugging. 1759 1760 ftrace_boot_snapshot 1761 [FTRACE] On boot up, a snapshot will be taken of the 1762 ftrace ring buffer that can be read at: 1763 /sys/kernel/tracing/snapshot. 1764 This is useful if you need tracing information from kernel 1765 boot up that is likely to be overridden by user space 1766 start up functionality. 1767 1768 Optionally, the snapshot can also be defined for a tracing 1769 instance that was created by the trace_instance= command 1770 line parameter. 1771 1772 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo 1773 1774 The above will cause the "foo" tracing instance to trigger 1775 a snapshot at the end of boot up. 1776 1777 ftrace_dump_on_oops[=2(orig_cpu) | =<instance>][,<instance> | 1778 ,<instance>=2(orig_cpu)] 1779 [FTRACE] will dump the trace buffers on oops. 1780 If no parameter is passed, ftrace will dump global 1781 buffers of all CPUs, if you pass 2 or orig_cpu, it 1782 will dump only the buffer of the CPU that triggered 1783 the oops, or the specific instance will be dumped if 1784 its name is passed. Multiple instance dump is also 1785 supported, and instances are separated by commas. Each 1786 instance supports only dump on CPU that triggered the 1787 oops by passing 2 or orig_cpu to it. 1788 1789 ftrace_dump_on_oops=foo=orig_cpu 1790 1791 The above will dump only the buffer of "foo" instance 1792 on CPU that triggered the oops. 1793 1794 ftrace_dump_on_oops,foo,bar=orig_cpu 1795 1796 The above will dump global buffer on all CPUs, the 1797 buffer of "foo" instance on all CPUs and the buffer 1798 of "bar" instance on CPU that triggered the oops. 1799 1800 ftrace_filter=[function-list] 1801 [FTRACE] Limit the functions traced by the function 1802 tracer at boot up. function-list is a comma-separated 1803 list of functions. This list can be changed at run 1804 time by the set_ftrace_filter file in the debugfs 1805 tracing directory. 1806 1807 ftrace_notrace=[function-list] 1808 [FTRACE] Do not trace the functions specified in 1809 function-list. This list can be changed at run time 1810 by the set_ftrace_notrace file in the debugfs 1811 tracing directory. 1812 1813 ftrace_graph_filter=[function-list] 1814 [FTRACE] Limit the top level callers functions traced 1815 by the function graph tracer at boot up. 1816 function-list is a comma-separated list of functions 1817 that can be changed at run time by the 1818 set_graph_function file in the debugfs tracing directory. 1819 1820 ftrace_graph_notrace=[function-list] 1821 [FTRACE] Do not trace from the functions specified in 1822 function-list. This list is a comma-separated list of 1823 functions that can be changed at run time by the 1824 set_graph_notrace file in the debugfs tracing directory. 1825 1826 ftrace_graph_max_depth=<uint> 1827 [FTRACE] Used with the function graph tracer. This is 1828 the max depth it will trace into a function. This value 1829 can be changed at run time by the max_graph_depth file 1830 in the tracefs tracing directory. default: 0 (no limit) 1831 1832 fw_devlink= [KNL,EARLY] Create device links between consumer and supplier 1833 devices by scanning the firmware to infer the 1834 consumer/supplier relationships. This feature is 1835 especially useful when drivers are loaded as modules as 1836 it ensures proper ordering of tasks like device probing 1837 (suppliers first, then consumers), supplier boot state 1838 clean up (only after all consumers have probed), 1839 suspend/resume & runtime PM (consumers first, then 1840 suppliers). 1841 Format: { off | permissive | on | rpm } 1842 off -- Don't create device links from firmware info. 1843 permissive -- Create device links from firmware info 1844 but use it only for ordering boot state clean 1845 up (sync_state() calls). 1846 on -- Create device links from firmware info and use it 1847 to enforce probe and suspend/resume ordering. 1848 rpm -- Like "on", but also use to order runtime PM. 1849 1850 fw_devlink.strict=<bool> 1851 [KNL,EARLY] Treat all inferred dependencies as mandatory 1852 dependencies. This only applies for fw_devlink=on|rpm. 1853 Format: <bool> 1854 1855 fw_devlink.sync_state = 1856 [KNL,EARLY] When all devices that could probe have finished 1857 probing, this parameter controls what to do with 1858 devices that haven't yet received their sync_state() 1859 calls. 1860 Format: { strict | timeout } 1861 strict -- Default. Continue waiting on consumers to 1862 probe successfully. 1863 timeout -- Give up waiting on consumers and call 1864 sync_state() on any devices that haven't yet 1865 received their sync_state() calls after 1866 deferred_probe_timeout has expired or by 1867 late_initcall() if !CONFIG_MODULES. 1868 1869 gamecon.map[2|3]= 1870 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad 1871 support via parallel port (up to 5 devices per port) 1872 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5> 1873 See also Documentation/input/devices/joystick-parport.rst 1874 1875 gamma= [HW,DRM] 1876 1877 gart_fix_e820= [X86-64,EARLY] disable the fix e820 for K8 GART 1878 Format: off | on 1879 default: on 1880 1881 gather_data_sampling= 1882 [X86,INTEL,EARLY] Control the Gather Data Sampling (GDS) 1883 mitigation. 1884 1885 Gather Data Sampling is a hardware vulnerability which 1886 allows unprivileged speculative access to data which was 1887 previously stored in vector registers. 1888 1889 This issue is mitigated by default in updated microcode. 1890 The mitigation may have a performance impact but can be 1891 disabled. On systems without the microcode mitigation 1892 disabling AVX serves as a mitigation. 1893 1894 force: Disable AVX to mitigate systems without 1895 microcode mitigation. No effect if the microcode 1896 mitigation is present. Known to cause crashes in 1897 userspace with buggy AVX enumeration. 1898 1899 off: Disable GDS mitigation. 1900 1901 gbpages [X86] Use GB pages for kernel direct mappings. 1902 1903 gcov_persist= [GCOV] When non-zero (default), profiling data for 1904 kernel modules is saved and remains accessible via 1905 debugfs, even when the module is unloaded/reloaded. 1906 When zero, profiling data is discarded and associated 1907 debugfs files are removed at module unload time. 1908 1909 goldfish [X86] Enable the goldfish android emulator platform. 1910 Don't use this when you are not running on the 1911 android emulator 1912 1913 gpio-mockup.gpio_mockup_ranges 1914 [HW] Sets the ranges of gpiochip of for this device. 1915 Format: <start1>,<end1>,<start2>,<end2>... 1916 gpio-mockup.gpio_mockup_named_lines 1917 [HW] Let the driver know GPIO lines should be named. 1918 1919 gpt [EFI] Forces disk with valid GPT signature but 1920 invalid Protective MBR to be treated as GPT. If the 1921 primary GPT is corrupted, it enables the backup/alternate 1922 GPT to be used instead. 1923 1924 grcan.enable0= [HW] Configuration of physical interface 0. Determines 1925 the "Enable 0" bit of the configuration register. 1926 Format: 0 | 1 1927 Default: 0 1928 grcan.enable1= [HW] Configuration of physical interface 1. Determines 1929 the "Enable 0" bit of the configuration register. 1930 Format: 0 | 1 1931 Default: 0 1932 grcan.select= [HW] Select which physical interface to use. 1933 Format: 0 | 1 1934 Default: 0 1935 grcan.txsize= [HW] Sets the size of the tx buffer. 1936 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0. 1937 Default: 1024 1938 grcan.rxsize= [HW] Sets the size of the rx buffer. 1939 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0. 1940 Default: 1024 1941 1942 hardened_usercopy= 1943 [KNL] Under CONFIG_HARDENED_USERCOPY, whether 1944 hardening is enabled for this boot. Hardened 1945 usercopy checking is used to protect the kernel 1946 from reading or writing beyond known memory 1947 allocation boundaries as a proactive defense 1948 against bounds-checking flaws in the kernel's 1949 copy_to_user()/copy_from_user() interface. 1950 The default is determined by 1951 CONFIG_HARDENED_USERCOPY_DEFAULT_ON. 1952 on Perform hardened usercopy checks. 1953 off Disable hardened usercopy checks. 1954 1955 hardlockup_all_cpu_backtrace= 1956 [KNL] Should the hard-lockup detector generate 1957 backtraces on all cpus. 1958 Format: 0 | 1 1959 1960 hash_pointers= 1961 [KNL,EARLY] 1962 By default, when pointers are printed to the console 1963 or buffers via the %p format string, that pointer is 1964 "hashed", i.e. obscured by hashing the pointer value. 1965 This is a security feature that hides actual kernel 1966 addresses from unprivileged users, but it also makes 1967 debugging the kernel more difficult since unequal 1968 pointers can no longer be compared. The choices are: 1969 Format: { auto | always | never } 1970 Default: auto 1971 1972 auto - Hash pointers unless slab_debug is enabled. 1973 always - Always hash pointers (even if slab_debug is 1974 enabled). 1975 never - Never hash pointers. This option should only 1976 be specified when debugging the kernel. Do 1977 not use on production kernels. The boot 1978 param "no_hash_pointers" is an alias for 1979 this mode. 1980 1981 For controlling hashing dynamically at runtime, 1982 use the "kernel.kptr_restrict" sysctl instead. 1983 1984 hashdist= [KNL,NUMA] Large hashes allocated during boot 1985 are distributed across NUMA nodes. Defaults on 1986 for 64-bit NUMA, off otherwise. 1987 Format: 0 | 1 (for off | on) 1988 1989 hd= [EIDE] (E)IDE hard drive subsystem geometry 1990 Format: <cyl>,<head>,<sect> 1991 1992 hest_disable [ACPI] 1993 Disable Hardware Error Source Table (HEST) support; 1994 corresponding firmware-first mode error processing 1995 logic will be disabled. 1996 1997 hibernate= [HIBERNATION] 1998 noresume Don't check if there's a hibernation image 1999 present during boot. 2000 nocompress Don't compress/decompress hibernation images. 2001 no Disable hibernation and resume. 2002 protect_image Turn on image protection during restoration 2003 (that will set all pages holding image data 2004 during restoration read-only). 2005 2006 hibernate.compressor= [HIBERNATION] Compression algorithm to be 2007 used with hibernation. 2008 Format: { lzo | lz4 } 2009 Default: lzo 2010 2011 lzo: Select LZO compression algorithm to 2012 compress/decompress hibernation image. 2013 2014 lz4: Select LZ4 compression algorithm to 2015 compress/decompress hibernation image. 2016 2017 hibernate.pm_test_delay= 2018 [HIBERNATION] 2019 Sets the number of seconds to remain in a hibernation test 2020 mode before resuming the system (see 2021 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 2022 is set. Default value is 5. 2023 2024 hibernate_compression_threads= 2025 [HIBERNATION] 2026 Set the number of threads used for compressing or decompressing 2027 hibernation images. 2028 2029 Format: <integer> 2030 Default: 3 2031 Minimum: 1 2032 Example: hibernate_compression_threads=4 2033 2034 highmem=nn[KMG] [KNL,BOOT,EARLY] forces the highmem zone to have an exact 2035 size of <nn>. This works even on boxes that have no 2036 highmem otherwise. This also works to reduce highmem 2037 size on bigger boxes. 2038 2039 highres= [KNL] Enable/disable high resolution timer mode. 2040 Valid parameters: "on", "off" 2041 Default: "on" 2042 2043 hlt [BUGS=ARM,SH] 2044 2045 hostname= [KNL,EARLY] Set the hostname (aka UTS nodename). 2046 Format: <string> 2047 This allows setting the system's hostname during early 2048 startup. This sets the name returned by gethostname. 2049 Using this parameter to set the hostname makes it 2050 possible to ensure the hostname is correctly set before 2051 any userspace processes run, avoiding the possibility 2052 that a process may call gethostname before the hostname 2053 has been explicitly set, resulting in the calling 2054 process getting an incorrect result. The string must 2055 not exceed the maximum allowed hostname length (usually 2056 64 characters) and will be truncated otherwise. 2057 2058 hpet= [X86-32,HPET] option to control HPET usage 2059 Format: { enable (default) | disable | force | 2060 verbose } 2061 disable: disable HPET and use PIT instead 2062 force: allow force enabled of undocumented chips (ICH4, 2063 VIA, nVidia) 2064 verbose: show contents of HPET registers during setup 2065 2066 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET 2067 registers. Default set by CONFIG_HPET_MMAP_DEFAULT. 2068 2069 hugepages= [HW,EARLY] Number of HugeTLB pages to allocate at boot. 2070 If this follows hugepagesz (below), it specifies 2071 the number of pages of hugepagesz to be allocated. 2072 If this is the first HugeTLB parameter on the command 2073 line, it specifies the number of pages to allocate for 2074 the default huge page size. If using node format, the 2075 number of pages to allocate per-node can be specified. 2076 See also Documentation/admin-guide/mm/hugetlbpage.rst. 2077 Format: <integer> or (node format) 2078 <node>:<integer>[,<node>:<integer>] 2079 2080 hugepagesz= 2081 [HW,EARLY] The size of the HugeTLB pages. This is 2082 used in conjunction with hugepages (above) to 2083 allocate huge pages of a specific size at boot. The 2084 pair hugepagesz=X hugepages=Y can be specified once 2085 for each supported huge page size. Huge page sizes 2086 are architecture dependent. See also 2087 Documentation/admin-guide/mm/hugetlbpage.rst. 2088 Format: size[KMG] 2089 2090 hugepage_alloc_threads= 2091 [HW] The number of threads that should be used to 2092 allocate hugepages during boot. This option can be 2093 used to improve system bootup time when allocating 2094 a large amount of huge pages. 2095 The default value is 25% of the available hardware threads. 2096 2097 Note that this parameter only applies to non-gigantic huge pages. 2098 2099 hugetlb_cma= [HW,CMA,EARLY] The size of a CMA area used for allocation 2100 of gigantic hugepages. Or using node format, the size 2101 of a CMA area per node can be specified. 2102 Format: nn[KMGTPE] or (node format) 2103 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]] 2104 2105 Reserve a CMA area of given size and allocate gigantic 2106 hugepages using the CMA allocator. If enabled, the 2107 boot-time allocation of gigantic hugepages is skipped. 2108 2109 hugetlb_cma_only= 2110 [HW,CMA,EARLY] When allocating new HugeTLB pages, only 2111 try to allocate from the CMA areas. 2112 2113 This option does nothing if hugetlb_cma= is not also 2114 specified. 2115 2116 hugetlb_free_vmemmap= 2117 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP 2118 enabled. 2119 Control if HugeTLB Vmemmap Optimization (HVO) is enabled. 2120 Allows heavy hugetlb users to free up some more 2121 memory (7 * PAGE_SIZE for each 2MB hugetlb page). 2122 Format: { on | off (default) } 2123 2124 on: enable HVO 2125 off: disable HVO 2126 2127 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y, 2128 the default is on. 2129 2130 Note that the vmemmap pages may be allocated from the added 2131 memory block itself when memory_hotplug.memmap_on_memory is 2132 enabled, those vmemmap pages cannot be optimized even if this 2133 feature is enabled. Other vmemmap pages not allocated from 2134 the added memory block itself do not be affected. 2135 2136 hung_task_panic= 2137 [KNL] Number of hung tasks to trigger kernel panic. 2138 Format: <int> 2139 2140 When set to a non-zero value, a kernel panic will be triggered if 2141 the number of detected hung tasks reaches this value. 2142 2143 0: don't panic 2144 1: panic immediately on first hung task 2145 N: panic after N hung tasks are detected in a single scan 2146 2147 The default value is controlled by the 2148 CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time option. The value 2149 selected by this boot parameter can be changed later by the 2150 kernel.hung_task_panic sysctl. 2151 2152 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC) 2153 terminal devices. Valid values: 0..8 2154 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs. 2155 If specified, z/VM IUCV HVC accepts connections 2156 from listed z/VM user IDs only. 2157 2158 hv_nopvspin [X86,HYPER_V,EARLY] 2159 Disables the paravirt spinlock optimizations 2160 which allow the hypervisor to 'idle' the guest 2161 on lock contention. 2162 2163 hw_protection= [HW] 2164 Format: reboot | shutdown 2165 2166 Hardware protection action taken on critical events like 2167 overtemperature or imminent voltage loss. 2168 2169 i2c_bus= [HW] Override the default board specific I2C bus speed 2170 or register an additional I2C bus that is not 2171 registered from board initialization code. 2172 Format: 2173 <bus_id>,<clkrate> 2174 2175 i2c_touchscreen_props= [HW,ACPI,X86] 2176 Set device-properties for ACPI-enumerated I2C-attached 2177 touchscreen, to e.g. fix coordinates of upside-down 2178 mounted touchscreens. If you need this option please 2179 submit a drivers/platform/x86/touchscreen_dmi.c patch 2180 adding a DMI quirk for this. 2181 2182 Format: 2183 <ACPI_HW_ID>:<prop_name>=<val>[:prop_name=val][:...] 2184 Where <val> is one of: 2185 Omit "=<val>" entirely Set a boolean device-property 2186 Unsigned number Set a u32 device-property 2187 Anything else Set a string device-property 2188 2189 Examples (split over multiple lines): 2190 i2c_touchscreen_props=GDIX1001:touchscreen-inverted-x: 2191 touchscreen-inverted-y 2192 2193 i2c_touchscreen_props=MSSL1680:touchscreen-size-x=1920: 2194 touchscreen-size-y=1080:touchscreen-inverted-y: 2195 firmware-name=gsl1680-vendor-model.fw:silead,home-button 2196 2197 i8042.debug [HW] Toggle i8042 debug mode 2198 i8042.unmask_kbd_data 2199 [HW] Enable printing of interrupt data from the KBD port 2200 (disabled by default, and as a pre-condition 2201 requires that i8042.debug=1 be enabled) 2202 i8042.direct [HW] Put keyboard port into non-translated mode 2203 i8042.dumbkbd [HW] Pretend that controller can only read data from 2204 keyboard and cannot control its state 2205 (Don't attempt to blink the leds) 2206 i8042.noaux [HW] Don't check for auxiliary (== mouse) port 2207 i8042.nokbd [HW] Don't check/create keyboard port 2208 i8042.noloop [HW] Disable the AUX Loopback command while probing 2209 for the AUX port 2210 i8042.nomux [HW] Don't check presence of an active multiplexing 2211 controller 2212 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX 2213 controllers 2214 i8042.notimeout [HW] Ignore timeout condition signalled by controller 2215 i8042.reset [HW] Reset the controller during init, cleanup and 2216 suspend-to-ram transitions, only during s2r 2217 transitions, or never reset 2218 Format: { 1 | Y | y | 0 | N | n } 2219 1, Y, y: always reset controller 2220 0, N, n: don't ever reset controller 2221 Default: only on s2r transitions on x86; most other 2222 architectures force reset to be always executed 2223 i8042.unlock [HW] Unlock (ignore) the keylock 2224 i8042.kbdreset [HW] Reset device connected to KBD port 2225 i8042.probe_defer 2226 [HW] Allow deferred probing upon i8042 probe errors 2227 2228 i810= [HW,DRM] 2229 2230 i915.invert_brightness= 2231 [DRM] Invert the sense of the variable that is used to 2232 set the brightness of the panel backlight. Normally a 2233 brightness value of 0 indicates backlight switched off, 2234 and the maximum of the brightness value sets the backlight 2235 to maximum brightness. If this parameter is set to 0 2236 (default) and the machine requires it, or this parameter 2237 is set to 1, a brightness value of 0 sets the backlight 2238 to maximum brightness, and the maximum of the brightness 2239 value switches the backlight off. 2240 -1 -- never invert brightness 2241 0 -- machine default 2242 1 -- force brightness inversion 2243 2244 ia32_emulation= [X86-64] 2245 Format: <bool> 2246 When true, allows loading 32-bit programs and executing 32-bit 2247 syscalls, essentially overriding IA32_EMULATION_DEFAULT_DISABLED at 2248 boot time. When false, unconditionally disables IA32 emulation. 2249 2250 icn= [HW,ISDN] 2251 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]] 2252 2253 2254 idle= [X86,EARLY] 2255 Format: idle=poll, idle=halt, idle=nomwait 2256 2257 idle=poll: Don't do power saving in the idle loop 2258 using HLT, but poll for rescheduling event. This will 2259 make the CPUs eat a lot more power, but may be useful 2260 to get slightly better performance in multiprocessor 2261 benchmarks. It also makes some profiling using 2262 performance counters more accurate. Please note that 2263 on systems with MONITOR/MWAIT support (like Intel 2264 EM64T CPUs) this option has no performance advantage 2265 over the normal idle loop. It may also interact badly 2266 with hyperthreading. 2267 2268 idle=halt: Halt is forced to be used for CPU idle. 2269 In such case C2/C3 won't be used again. 2270 2271 idle=nomwait: Disable mwait for CPU C-states 2272 2273 idxd.sva= [HW] 2274 Format: <bool> 2275 Allow force disabling of Shared Virtual Memory (SVA) 2276 support for the idxd driver. By default it is set to 2277 true (1). 2278 2279 idxd.tc_override= [HW] 2280 Format: <bool> 2281 Allow override of default traffic class configuration 2282 for the device. By default it is set to false (0). 2283 2284 ieee754= [MIPS] Select IEEE Std 754 conformance mode 2285 Format: { strict | legacy | 2008 | relaxed | emulated } 2286 Default: strict 2287 2288 Choose which programs will be accepted for execution 2289 based on the IEEE 754 NaN encoding(s) supported by 2290 the FPU and the NaN encoding requested with the value 2291 of an ELF file header flag individually set by each 2292 binary. Hardware implementations are permitted to 2293 support either or both of the legacy and the 2008 NaN 2294 encoding mode. 2295 2296 Available settings are as follows: 2297 strict accept binaries that request a NaN encoding 2298 supported by the FPU 2299 legacy only accept legacy-NaN binaries, if supported 2300 by the FPU 2301 2008 only accept 2008-NaN binaries, if supported 2302 by the FPU 2303 relaxed accept any binaries regardless of whether 2304 supported by the FPU 2305 emulated accept any binaries but enable FPU emulator 2306 if binary mode is unsupported by the FPU. 2307 2308 The FPU emulator is always able to support both NaN 2309 encodings, so if no FPU hardware is present or it has 2310 been disabled with 'nofpu', then the settings of 2311 'legacy' and '2008' strap the emulator accordingly, 2312 'relaxed' straps the emulator for both legacy-NaN and 2313 2008-NaN, whereas 'strict' enables legacy-NaN only on 2314 legacy processors and both NaN encodings on MIPS32 or 2315 MIPS64 CPUs. 2316 2317 The setting for ABS.fmt/NEG.fmt instruction execution 2318 mode generally follows that for the NaN encoding, 2319 except where unsupported by hardware. 2320 2321 ignore_loglevel [KNL,EARLY] 2322 Ignore loglevel setting - this will print /all/ 2323 kernel messages to the console. Useful for debugging. 2324 We also add it as printk module parameter, so users 2325 could change it dynamically, usually by 2326 /sys/module/printk/parameters/ignore_loglevel. 2327 2328 ignore_rlimit_data 2329 Ignore RLIMIT_DATA setting for data mappings, 2330 print warning at first misuse. Can be changed via 2331 /sys/module/kernel/parameters/ignore_rlimit_data. 2332 2333 ihash_entries= [KNL] 2334 Set number of hash buckets for inode cache. 2335 2336 ima_appraise= [IMA] appraise integrity measurements 2337 Format: { "off" | "enforce" | "fix" | "log" } 2338 default: "enforce" 2339 2340 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead. 2341 The builtin appraise policy appraises all files 2342 owned by uid=0. 2343 2344 ima_canonical_fmt [IMA] 2345 Use the canonical format for the binary runtime 2346 measurements, instead of host native format. 2347 2348 ima_hash= [IMA] 2349 Format: { md5 | sha1 | rmd160 | sha256 | sha384 2350 | sha512 | ... } 2351 default: "sha1" 2352 2353 The list of supported hash algorithms is defined 2354 in crypto/hash_info.h. 2355 2356 ima_policy= [IMA] 2357 The builtin policies to load during IMA setup. 2358 Format: "tcb | appraise_tcb | secure_boot | 2359 fail_securely | critical_data" 2360 2361 The "tcb" policy measures all programs exec'd, files 2362 mmap'd for exec, and all files opened with the read 2363 mode bit set by either the effective uid (euid=0) or 2364 uid=0. 2365 2366 The "appraise_tcb" policy appraises the integrity of 2367 all files owned by root. 2368 2369 The "secure_boot" policy appraises the integrity 2370 of files (eg. kexec kernel image, kernel modules, 2371 firmware, policy, etc) based on file signatures. 2372 2373 The "fail_securely" policy forces file signature 2374 verification failure also on privileged mounted 2375 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE 2376 flag. 2377 2378 The "critical_data" policy measures kernel integrity 2379 critical data. 2380 2381 ima_tcb [IMA] Deprecated. Use ima_policy= instead. 2382 Load a policy which meets the needs of the Trusted 2383 Computing Base. This means IMA will measure all 2384 programs exec'd, files mmap'd for exec, and all files 2385 opened for read by uid=0. 2386 2387 ima_template= [IMA] 2388 Select one of defined IMA measurements template formats. 2389 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" | 2390 "ima-sigv2" } 2391 Default: "ima-ng" 2392 2393 ima_template_fmt= 2394 [IMA] Define a custom template format. 2395 Format: { "field1|...|fieldN" } 2396 2397 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage 2398 Format: <min_file_size> 2399 Set the minimal file size for using asynchronous hash. 2400 If left unspecified, ahash usage is disabled. 2401 2402 ahash performance varies for different data sizes on 2403 different crypto accelerators. This option can be used 2404 to achieve the best performance for a particular HW. 2405 2406 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size 2407 Format: <bufsize> 2408 Set hashing buffer size. Default: 4k. 2409 2410 ahash performance varies for different chunk sizes on 2411 different crypto accelerators. This option can be used 2412 to achieve best performance for particular HW. 2413 2414 ima= [IMA] Enable or disable IMA 2415 Format: { "off" | "on" } 2416 Default: "on" 2417 Note that disabling IMA is limited to kdump kernel. 2418 2419 indirect_target_selection= [X86,Intel] Mitigation control for Indirect 2420 Target Selection(ITS) bug in Intel CPUs. Updated 2421 microcode is also required for a fix in IBPB. 2422 2423 on: Enable mitigation (default). 2424 off: Disable mitigation. 2425 force: Force the ITS bug and deploy default 2426 mitigation. 2427 vmexit: Only deploy mitigation if CPU is affected by 2428 guest/host isolation part of ITS. 2429 stuff: Deploy RSB-fill mitigation when retpoline is 2430 also deployed. Otherwise, deploy the default 2431 mitigation. 2432 2433 For details see: 2434 Documentation/admin-guide/hw-vuln/indirect-target-selection.rst 2435 2436 init= [KNL] 2437 Format: <full_path> 2438 Run specified binary instead of /sbin/init as init 2439 process. 2440 2441 initcall_debug [KNL] Trace initcalls as they are executed. Useful 2442 for working out where the kernel is dying during 2443 startup. 2444 2445 initcall_blacklist= [KNL] Do not execute a comma-separated list of 2446 initcall functions. Useful for debugging built-in 2447 modules and initcalls. 2448 2449 initramfs_async= [KNL] 2450 Format: <bool> 2451 Default: 1 2452 This parameter controls whether the initramfs 2453 image is unpacked asynchronously, concurrently 2454 with devices being probed and 2455 initialized. This should normally just work, 2456 but as a debugging aid, one can get the 2457 historical behaviour of the initramfs 2458 unpacking being completed before device_ and 2459 late_ initcalls. 2460 2461 initrd= [BOOT,EARLY] Specify the location of the initial ramdisk 2462 2463 initrdmem= [KNL,EARLY] Specify a physical address and size from which to 2464 load the initrd. If an initrd is compiled in or 2465 specified in the bootparams, it takes priority over this 2466 setting. 2467 Format: ss[KMG],nn[KMG] 2468 Default is 0, 0 2469 2470 init_on_alloc= [MM,EARLY] Fill newly allocated pages and heap objects with 2471 zeroes. 2472 Format: 0 | 1 2473 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON. 2474 2475 init_on_free= [MM,EARLY] Fill freed pages and heap objects with zeroes. 2476 Format: 0 | 1 2477 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON. 2478 2479 init_pkru= [X86] Specify the default memory protection keys rights 2480 register contents for all processes. 0x55555554 by 2481 default (disallow access to all but pkey 0). Can 2482 override in debugfs after boot. 2483 2484 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver 2485 Format: <irq> 2486 2487 int_pln_enable [X86] Enable power limit notification interrupt 2488 2489 integrity_audit=[IMA] 2490 Format: { "0" | "1" } 2491 0 -- basic integrity auditing messages. (Default) 2492 1 -- additional integrity auditing messages. 2493 2494 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option 2495 on 2496 Enable intel iommu driver. 2497 off 2498 Disable intel iommu driver. 2499 igfx_off [Default Off] 2500 By default, gfx is mapped as normal device. If a gfx 2501 device has a dedicated DMAR unit, the DMAR unit is 2502 bypassed by not enabling DMAR with this option. In 2503 this case, gfx device will use physical address for 2504 DMA. 2505 strict [Default Off] 2506 Deprecated, equivalent to iommu.strict=1. 2507 sp_off [Default Off] 2508 By default, super page will be supported if Intel IOMMU 2509 has the capability. With this option, super page will 2510 not be supported. 2511 sm_on 2512 Enable the Intel IOMMU scalable mode if the hardware 2513 advertises that it has support for the scalable mode 2514 translation. 2515 sm_off 2516 Disallow use of the Intel IOMMU scalable mode. 2517 tboot_noforce [Default Off] 2518 Do not force the Intel IOMMU enabled under tboot. 2519 By default, tboot will force Intel IOMMU on, which 2520 could harm performance of some high-throughput 2521 devices like 40GBit network cards, even if identity 2522 mapping is enabled. 2523 Note that using this option lowers the security 2524 provided by tboot because it makes the system 2525 vulnerable to DMA attacks. 2526 2527 intel_idle.max_cstate= [KNL,HW,ACPI,X86] 2528 0 disables intel_idle and fall back on acpi_idle. 2529 1 to 9 specify maximum depth of C-state. 2530 2531 intel_pstate= [X86,EARLY] 2532 disable 2533 Do not enable intel_pstate as the default 2534 scaling driver for the supported processors 2535 active 2536 Use intel_pstate driver to bypass the scaling 2537 governors layer of cpufreq and provides it own 2538 algorithms for p-state selection. There are two 2539 P-state selection algorithms provided by 2540 intel_pstate in the active mode: powersave and 2541 performance. The way they both operate depends 2542 on whether or not the hardware managed P-states 2543 (HWP) feature has been enabled in the processor 2544 and possibly on the processor model. 2545 passive 2546 Use intel_pstate as a scaling driver, but configure it 2547 to work with generic cpufreq governors (instead of 2548 enabling its internal governor). This mode cannot be 2549 used along with the hardware-managed P-states (HWP) 2550 feature. 2551 force 2552 Enable intel_pstate on systems that prohibit it by default 2553 in favor of acpi-cpufreq. Forcing the intel_pstate driver 2554 instead of acpi-cpufreq may disable platform features, such 2555 as thermal controls and power capping, that rely on ACPI 2556 P-States information being indicated to OSPM and therefore 2557 should be used with caution. This option does not work with 2558 processors that aren't supported by the intel_pstate driver 2559 or on platforms that use pcc-cpufreq instead of acpi-cpufreq. 2560 no_hwp 2561 Do not enable hardware P state control (HWP) 2562 if available. 2563 hwp_only 2564 Only load intel_pstate on systems which support 2565 hardware P state control (HWP) if available. 2566 support_acpi_ppc 2567 Enforce ACPI _PPC performance limits. If the Fixed ACPI 2568 Description Table, specifies preferred power management 2569 profile as "Enterprise Server" or "Performance Server", 2570 then this feature is turned on by default. 2571 per_cpu_perf_limits 2572 Allow per-logical-CPU P-State performance control limits using 2573 cpufreq sysfs interface 2574 no_cas 2575 Do not enable capacity-aware scheduling (CAS) on 2576 hybrid systems 2577 2578 intremap= [X86-64,Intel-IOMMU,EARLY] 2579 on enable Interrupt Remapping (default) 2580 off disable Interrupt Remapping 2581 nosid disable Source ID checking 2582 no_x2apic_optout 2583 BIOS x2APIC opt-out request will be ignored 2584 nopost disable Interrupt Posting 2585 posted_msi 2586 enable MSIs delivered as posted interrupts 2587 2588 iomem= Disable strict checking of access to MMIO memory 2589 strict regions from userspace. 2590 relaxed 2591 2592 iommu= [X86,EARLY] 2593 2594 off 2595 Don't initialize and use any kind of IOMMU. 2596 2597 force 2598 Force the use of the hardware IOMMU even when 2599 it is not actually needed (e.g. because < 3 GB 2600 memory). 2601 2602 noforce 2603 Don't force hardware IOMMU usage when it is not 2604 needed. (default). 2605 2606 biomerge 2607 panic 2608 nopanic 2609 merge 2610 nomerge 2611 2612 soft 2613 Use software bounce buffering (SWIOTLB) (default for 2614 Intel machines). This can be used to prevent the usage 2615 of an available hardware IOMMU. 2616 2617 pt 2618 nopt 2619 nobypass [PPC/POWERNV] 2620 Disable IOMMU bypass, using IOMMU for PCI devices. 2621 2622 AMD Gart HW IOMMU-specific options: 2623 2624 <size> 2625 Set the size of the remapping area in bytes. 2626 2627 allowed 2628 Overwrite iommu off workarounds for specific chipsets 2629 2630 fullflush 2631 Flush IOMMU on each allocation (default). 2632 2633 nofullflush 2634 Don't use IOMMU fullflush. 2635 2636 memaper[=<order>] 2637 Allocate an own aperture over RAM with size 2638 32MB<<order. (default: order=1, i.e. 64MB) 2639 2640 merge 2641 Do scatter-gather (SG) merging. Implies "force" 2642 (experimental). 2643 2644 nomerge 2645 Don't do scatter-gather (SG) merging. 2646 2647 noaperture 2648 Ask the IOMMU not to touch the aperture for AGP. 2649 2650 noagp 2651 Don't initialize the AGP driver and use full aperture. 2652 2653 panic 2654 Always panic when IOMMU overflows. 2655 2656 iommu.forcedac= [ARM64,X86,EARLY] Control IOVA allocation for PCI devices. 2657 Format: { "0" | "1" } 2658 0 - Try to allocate a 32-bit DMA address first, before 2659 falling back to the full range if needed. 2660 1 - Allocate directly from the full usable range, 2661 forcing Dual Address Cycle for PCI cards supporting 2662 greater than 32-bit addressing. 2663 2664 iommu.strict= [ARM64,X86,S390,EARLY] Configure TLB invalidation behaviour 2665 Format: { "0" | "1" } 2666 0 - Lazy mode. 2667 Request that DMA unmap operations use deferred 2668 invalidation of hardware TLBs, for increased 2669 throughput at the cost of reduced device isolation. 2670 Will fall back to strict mode if not supported by 2671 the relevant IOMMU driver. 2672 1 - Strict mode. 2673 DMA unmap operations invalidate IOMMU hardware TLBs 2674 synchronously. 2675 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}. 2676 Note: on x86, strict mode specified via one of the 2677 legacy driver-specific options takes precedence. 2678 2679 iommu.passthrough= 2680 [ARM64,X86,EARLY] Configure DMA to bypass the IOMMU by default. 2681 Format: { "0" | "1" } 2682 0 - Use IOMMU translation for DMA. 2683 1 - Bypass the IOMMU for DMA. 2684 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH. 2685 2686 iommu.debug_pagealloc= 2687 [KNL,EARLY] When CONFIG_IOMMU_DEBUG_PAGEALLOC is set, this 2688 parameter enables the feature at boot time. By default, it 2689 is disabled and the system behaves the same way as a kernel 2690 built without CONFIG_IOMMU_DEBUG_PAGEALLOC. 2691 Format: { "0" | "1" } 2692 0 - Sanitizer disabled. 2693 1 - Sanitizer enabled, expect runtime overhead. 2694 2695 io7= [HW] IO7 for Marvel-based Alpha systems 2696 See comment before marvel_specify_io7 in 2697 arch/alpha/kernel/core_marvel.c. 2698 2699 io_delay= [X86,EARLY] I/O delay method 2700 0x80 2701 Standard port 0x80 based delay 2702 0xed 2703 Alternate port 0xed based delay (needed on some systems) 2704 udelay 2705 Simple two microseconds delay 2706 none 2707 No delay 2708 2709 ip= [IP_PNP] 2710 See Documentation/admin-guide/nfs/nfsroot.rst. 2711 2712 ipcmni_extend [KNL,EARLY] Extend the maximum number of unique System V 2713 IPC identifiers from 32,768 to 16,777,216. 2714 2715 ipe.enforce= [IPE] 2716 Format: <bool> 2717 Determine whether IPE starts in permissive (0) or 2718 enforce (1) mode. The default is enforce. 2719 2720 ipe.success_audit= 2721 [IPE] 2722 Format: <bool> 2723 Start IPE with success auditing enabled, emitting 2724 an audit event when a binary is allowed. The default 2725 is 0. 2726 2727 irqaffinity= [SMP] Set the default irq affinity mask 2728 The argument is a cpu list, as described above. 2729 2730 irqchip.gicv2_force_probe= 2731 [ARM,ARM64,EARLY] 2732 Format: <bool> 2733 Force the kernel to look for the second 4kB page 2734 of a GICv2 controller even if the memory range 2735 exposed by the device tree is too small. 2736 2737 irqchip.gicv3_nolpi= 2738 [ARM,ARM64,EARLY] 2739 Force the kernel to ignore the availability of 2740 LPIs (and by consequence ITSs). Intended for system 2741 that use the kernel as a bootloader, and thus want 2742 to let secondary kernels in charge of setting up 2743 LPIs. 2744 2745 irqchip.gicv3_pseudo_nmi= [ARM64,EARLY] 2746 Enables support for pseudo-NMIs in the kernel. This 2747 requires the kernel to be built with 2748 CONFIG_ARM64_PSEUDO_NMI. 2749 2750 irqchip.riscv_imsic_noipi 2751 [RISC-V,EARLY] 2752 Force the kernel to not use IMSIC software injected MSIs 2753 as IPIs. Intended for system where IMSIC is trap-n-emulated, 2754 and thus want to reduce MMIO traps when triggering IPIs 2755 to multiple harts. 2756 2757 irqfixup [HW] 2758 When an interrupt is not handled search all handlers 2759 for it. Intended to get systems with badly broken 2760 firmware running. 2761 2762 irqhandler.duration_warn_us= [KNL] 2763 Warn if an IRQ handler exceeds the specified duration 2764 threshold in microseconds. Useful for identifying 2765 long-running IRQs in the system. 2766 2767 irqpoll [HW] 2768 When an interrupt is not handled search all handlers 2769 for it. Also check all handlers each timer 2770 interrupt. Intended to get systems with badly broken 2771 firmware running. 2772 2773 isapnp= [ISAPNP] 2774 Format: <RDP>,<reset>,<pci_scan>,<verbosity> 2775 2776 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance. 2777 [Deprecated - use cpusets instead] 2778 Format: [flag-list,]<cpu-list> 2779 2780 Specify one or more CPUs to isolate from disturbances 2781 specified in the flag list (default: domain): 2782 2783 nohz 2784 Disable the tick when a single task runs as well as 2785 disabling other kernel noises like having RCU callbacks 2786 offloaded. This is equivalent to the nohz_full parameter. 2787 2788 A residual 1Hz tick is offloaded to workqueues, which you 2789 need to affine to housekeeping through the global 2790 workqueue's affinity configured via the 2791 /sys/devices/virtual/workqueue/cpumask sysfs file, or 2792 by using the 'domain' flag described below. 2793 2794 NOTE: by default the global workqueue runs on all CPUs, 2795 so to protect individual CPUs the 'cpumask' file has to 2796 be configured manually after bootup. 2797 2798 domain 2799 Isolate from the general SMP balancing and scheduling 2800 algorithms. Note that performing domain isolation this way 2801 is irreversible: it's not possible to bring back a CPU to 2802 the domains once isolated through isolcpus. It's strongly 2803 advised to use cpusets instead to disable scheduler load 2804 balancing through the "cpuset.sched_load_balance" file. 2805 It offers a much more flexible interface where CPUs can 2806 move in and out of an isolated set anytime. 2807 2808 You can move a process onto or off an "isolated" CPU via 2809 the CPU affinity syscalls or cpuset. 2810 <cpu number> begins at 0 and the maximum value is 2811 "number of CPUs in system - 1". 2812 2813 managed_irq 2814 2815 Isolate from being targeted by managed interrupts 2816 which have an interrupt mask containing isolated 2817 CPUs. The affinity of managed interrupts is 2818 handled by the kernel and cannot be changed via 2819 the /proc/irq/* interfaces. 2820 2821 This isolation is best effort and only effective 2822 if the automatically assigned interrupt mask of a 2823 device queue contains isolated and housekeeping 2824 CPUs. If housekeeping CPUs are online then such 2825 interrupts are directed to the housekeeping CPU 2826 so that IO submitted on the housekeeping CPU 2827 cannot disturb the isolated CPU. 2828 2829 If a queue's affinity mask contains only isolated 2830 CPUs then this parameter has no effect on the 2831 interrupt routing decision, though interrupts are 2832 only delivered when tasks running on those 2833 isolated CPUs submit IO. IO submitted on 2834 housekeeping CPUs has no influence on those 2835 queues. 2836 2837 The format of <cpu-list> is described above. 2838 2839 iucv= [HW,NET] 2840 2841 ivrs_ioapic [HW,X86-64] 2842 Provide an override to the IOAPIC-ID<->DEVICE-ID 2843 mapping provided in the IVRS ACPI table. 2844 By default, PCI segment is 0, and can be omitted. 2845 2846 For example, to map IOAPIC-ID decimal 10 to 2847 PCI segment 0x1 and PCI device 00:14.0, 2848 write the parameter as: 2849 ivrs_ioapic=10@0001:00:14.0 2850 2851 Deprecated formats: 2852 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0 2853 write the parameter as: 2854 ivrs_ioapic[10]=00:14.0 2855 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and 2856 PCI device 00:14.0 write the parameter as: 2857 ivrs_ioapic[10]=0001:00:14.0 2858 2859 ivrs_hpet [HW,X86-64] 2860 Provide an override to the HPET-ID<->DEVICE-ID 2861 mapping provided in the IVRS ACPI table. 2862 By default, PCI segment is 0, and can be omitted. 2863 2864 For example, to map HPET-ID decimal 10 to 2865 PCI segment 0x1 and PCI device 00:14.0, 2866 write the parameter as: 2867 ivrs_hpet=10@0001:00:14.0 2868 2869 Deprecated formats: 2870 * To map HPET-ID decimal 0 to PCI device 00:14.0 2871 write the parameter as: 2872 ivrs_hpet[0]=00:14.0 2873 * To map HPET-ID decimal 10 to PCI segment 0x1 and 2874 PCI device 00:14.0 write the parameter as: 2875 ivrs_ioapic[10]=0001:00:14.0 2876 2877 ivrs_acpihid [HW,X86-64] 2878 Provide an override to the ACPI-HID:UID<->DEVICE-ID 2879 mapping provided in the IVRS ACPI table. 2880 By default, PCI segment is 0, and can be omitted. 2881 2882 For example, to map UART-HID:UID AMD0020:0 to 2883 PCI segment 0x1 and PCI device ID 00:14.5, 2884 write the parameter as: 2885 ivrs_acpihid=AMD0020:0@0001:00:14.5 2886 2887 Deprecated formats: 2888 * To map UART-HID:UID AMD0020:0 to PCI segment is 0, 2889 PCI device ID 00:14.5, write the parameter as: 2890 ivrs_acpihid[00:14.5]=AMD0020:0 2891 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and 2892 PCI device ID 00:14.5, write the parameter as: 2893 ivrs_acpihid[0001:00:14.5]=AMD0020:0 2894 2895 js= [HW,JOY] Analog joystick 2896 See Documentation/input/joydev/joystick.rst. 2897 2898 kasan_multi_shot 2899 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print 2900 report on every invalid memory access. Without this 2901 parameter KASAN will print report only for the first 2902 invalid access. 2903 2904 keep_bootcon [KNL,EARLY] 2905 Do not unregister boot console at start. This is only 2906 useful for debugging when something happens in the window 2907 between unregistering the boot console and initializing 2908 the real console. 2909 2910 keepinitrd [HW,ARM] See retain_initrd. 2911 2912 kernelcore= [KNL,X86,PPC,EARLY] 2913 Format: nn[KMGTPE] | nn% | "mirror" 2914 This parameter specifies the amount of memory usable by 2915 the kernel for non-movable allocations. The requested 2916 amount is spread evenly throughout all nodes in the 2917 system as ZONE_NORMAL. The remaining memory is used for 2918 movable memory in its own zone, ZONE_MOVABLE. In the 2919 event, a node is too small to have both ZONE_NORMAL and 2920 ZONE_MOVABLE, kernelcore memory will take priority and 2921 other nodes will have a larger ZONE_MOVABLE. 2922 2923 ZONE_MOVABLE is used for the allocation of pages that 2924 may be reclaimed or moved by the page migration 2925 subsystem. Note that allocations like PTEs-from-HighMem 2926 still use the HighMem zone if it exists, and the Normal 2927 zone if it does not. 2928 2929 It is possible to specify the exact amount of memory in 2930 the form of "nn[KMGTPE]", a percentage of total system 2931 memory in the form of "nn%", or "mirror". If "mirror" 2932 option is specified, mirrored (reliable) memory is used 2933 for non-movable allocations and remaining memory is used 2934 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror" 2935 are exclusive, so you cannot specify multiple forms. 2936 2937 kfence.burst= [MM,KFENCE] The number of additional successive 2938 allocations to be attempted through KFENCE for each 2939 sample interval. 2940 Format: <unsigned integer> 2941 Default: 0 2942 2943 kfence.check_on_panic= 2944 [MM,KFENCE] Whether to check all KFENCE-managed objects' 2945 canaries on panic. 2946 Format: <bool> 2947 Default: false 2948 2949 kfence.deferrable= 2950 [MM,KFENCE] Whether to use a deferrable timer to trigger 2951 allocations. This avoids forcing CPU wake-ups if the 2952 system is idle, at the risk of a less predictable 2953 sample interval. 2954 Format: <bool> 2955 Default: CONFIG_KFENCE_DEFERRABLE 2956 2957 kfence.sample_interval= 2958 [MM,KFENCE] KFENCE's sample interval in milliseconds. 2959 Format: <unsigned integer> 2960 0 - Disable KFENCE. 2961 >0 - Enabled KFENCE with given sample interval. 2962 Default: CONFIG_KFENCE_SAMPLE_INTERVAL 2963 2964 kfence.skip_covered_thresh= 2965 [MM,KFENCE] If pool utilization reaches this threshold 2966 (pool usage%), KFENCE limits currently covered 2967 allocations of the same source from further filling 2968 up the pool. 2969 Format: <unsigned integer> 2970 Default: 75 2971 2972 kgdbdbgp= [KGDB,HW,EARLY] kgdb over EHCI usb debug port. 2973 Format: <Controller#>[,poll interval] 2974 The controller # is the number of the ehci usb debug 2975 port as it is probed via PCI. The poll interval is 2976 optional and is the number seconds in between 2977 each poll cycle to the debug port in case you need 2978 the functionality for interrupting the kernel with 2979 gdb or control-c on the dbgp connection. When 2980 not using this parameter you use sysrq-g to break into 2981 the kernel debugger. 2982 2983 kgdboc= [KGDB,HW] kgdb over consoles. 2984 Requires a tty driver that supports console polling, 2985 or a supported polling keyboard driver (non-usb). 2986 Serial only format: <serial_device>[,baud] 2987 keyboard only format: kbd 2988 keyboard and serial format: kbd,<serial_device>[,baud] 2989 Optional Kernel mode setting: 2990 kms, kbd format: kms,kbd 2991 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud] 2992 2993 kgdboc_earlycon= [KGDB,HW,EARLY] 2994 If the boot console provides the ability to read 2995 characters and can work in polling mode, you can use 2996 this parameter to tell kgdb to use it as a backend 2997 until the normal console is registered. Intended to 2998 be used together with the kgdboc parameter which 2999 specifies the normal console to transition to. 3000 3001 The name of the early console should be specified 3002 as the value of this parameter. Note that the name of 3003 the early console might be different than the tty 3004 name passed to kgdboc. It's OK to leave the value 3005 blank and the first boot console that implements 3006 read() will be picked. 3007 3008 kgdbwait [KGDB,EARLY] Stop kernel execution and enter the 3009 kernel debugger at the earliest opportunity. 3010 3011 kho= [KEXEC,EARLY] 3012 Format: { "0" | "1" | "off" | "on" | "y" | "n" } 3013 Enables or disables Kexec HandOver. 3014 "0" | "off" | "n" - kexec handover is disabled 3015 "1" | "on" | "y" - kexec handover is enabled 3016 3017 kho_scratch= [KEXEC,EARLY] 3018 Format: ll[KMG],mm[KMG],nn[KMG] | nn% 3019 Defines the size of the KHO scratch region. The KHO 3020 scratch regions are physically contiguous memory 3021 ranges that can only be used for non-kernel 3022 allocations. That way, even when memory is heavily 3023 fragmented with handed over memory, the kexeced 3024 kernel will always have enough contiguous ranges to 3025 bootstrap itself. 3026 3027 It is possible to specify the exact amount of 3028 memory in the form of "ll[KMG],mm[KMG],nn[KMG]" 3029 where the first parameter defines the size of a low 3030 memory scratch area, the second parameter defines 3031 the size of a global scratch area and the third 3032 parameter defines the size of additional per-node 3033 scratch areas. The form "nn%" defines scale factor 3034 (in percents) of memory that was used during boot. 3035 3036 kmac= [MIPS] Korina ethernet MAC address. 3037 Configure the RouterBoard 532 series on-chip 3038 Ethernet adapter MAC address. 3039 3040 kmemleak= [KNL,EARLY] Boot-time kmemleak enable/disable 3041 Valid arguments: on, off 3042 Default: on 3043 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y, 3044 the default is off. 3045 3046 kprobe_event=[probe-list] 3047 [FTRACE] Add kprobe events and enable at boot time. 3048 The probe-list is a semicolon delimited list of probe 3049 definitions. Each definition is same as kprobe_events 3050 interface, but the parameters are comma delimited. 3051 For example, to add a kprobe event on vfs_read with 3052 arg1 and arg2, add to the command line; 3053 3054 kprobe_event=p,vfs_read,$arg1,$arg2 3055 3056 See also Documentation/trace/kprobetrace.rst "Kernel 3057 Boot Parameter" section. 3058 3059 kpti= [ARM64,EARLY] Control page table isolation of 3060 user and kernel address spaces. 3061 Default: enabled on cores which need mitigation. 3062 0: force disabled 3063 1: force enabled 3064 3065 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires 3066 CONFIG_KUNIT to be set to be fully enabled. The 3067 default value can be overridden via 3068 KUNIT_DEFAULT_ENABLED. 3069 Default is 1 (enabled) 3070 3071 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs. 3072 Default is 0 (don't ignore, but inject #GP) 3073 3074 kvm.eager_page_split= 3075 [KVM,X86] Controls whether or not KVM will try to 3076 proactively split all huge pages during dirty logging. 3077 Eager page splitting reduces interruptions to vCPU 3078 execution by eliminating the write-protection faults 3079 and MMU lock contention that would otherwise be 3080 required to split huge pages lazily. 3081 3082 VM workloads that rarely perform writes or that write 3083 only to a small region of VM memory may benefit from 3084 disabling eager page splitting to allow huge pages to 3085 still be used for reads. 3086 3087 The behavior of eager page splitting depends on whether 3088 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If 3089 disabled, all huge pages in a memslot will be eagerly 3090 split when dirty logging is enabled on that memslot. If 3091 enabled, eager page splitting will be performed during 3092 the KVM_CLEAR_DIRTY ioctl, and only for the pages being 3093 cleared. 3094 3095 Eager page splitting is only supported when kvm.tdp_mmu=Y. 3096 3097 Default is Y (on). 3098 3099 kvm.enable_pmu=[KVM,X86] 3100 If enabled, KVM will virtualize PMU functionality based 3101 on the virtual CPU model defined by userspace. This 3102 can be overridden on a per-VM basis via 3103 KVM_CAP_PMU_CAPABILITY. 3104 3105 If disabled, KVM will not virtualize PMU functionality, 3106 e.g. MSRs, PMCs, PMIs, etc., even if userspace defines 3107 a virtual CPU model that contains PMU assets. 3108 3109 Note, KVM's vPMU support implicitly requires running 3110 with an in-kernel local APIC, e.g. to deliver PMIs to 3111 the guest. Running without an in-kernel local APIC is 3112 not supported, though KVM will allow such a combination 3113 (with severely degraded functionality). 3114 3115 See also enable_mediated_pmu. 3116 3117 Default is Y (on). 3118 3119 kvm.enable_virt_at_load=[KVM,ARM64,LOONGARCH,MIPS,RISCV,X86] 3120 If enabled, KVM will enable virtualization in hardware 3121 when KVM is loaded, and disable virtualization when KVM 3122 is unloaded (if KVM is built as a module). 3123 3124 If disabled, KVM will dynamically enable and disable 3125 virtualization on-demand when creating and destroying 3126 VMs, i.e. on the 0=>1 and 1=>0 transitions of the 3127 number of VMs. 3128 3129 Enabling virtualization at module load avoids potential 3130 latency for creation of the 0=>1 VM, as KVM serializes 3131 virtualization enabling across all online CPUs. The 3132 "cost" of enabling virtualization when KVM is loaded, 3133 is that doing so may interfere with using out-of-tree 3134 hypervisors that want to "own" virtualization hardware. 3135 3136 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface. 3137 Default is false (don't support). 3138 3139 kvm.nx_huge_pages= 3140 [KVM] Controls the software workaround for the 3141 X86_BUG_ITLB_MULTIHIT bug. 3142 force : Always deploy workaround. 3143 off : Never deploy workaround. 3144 auto : Deploy workaround based on the presence of 3145 X86_BUG_ITLB_MULTIHIT. 3146 3147 Default is 'auto'. 3148 3149 If the software workaround is enabled for the host, 3150 guests do need not to enable it for nested guests. 3151 3152 kvm.nx_huge_pages_recovery_ratio= 3153 [KVM] Controls how many 4KiB pages are periodically zapped 3154 back to huge pages. 0 disables the recovery, otherwise if 3155 the value is N KVM will zap 1/Nth of the 4KiB pages every 3156 period (see below). The default is 60. 3157 3158 kvm.nx_huge_pages_recovery_period_ms= 3159 [KVM] Controls the time period at which KVM zaps 4KiB pages 3160 back to huge pages. If the value is a non-zero N, KVM will 3161 zap a portion (see ratio above) of the pages every N msecs. 3162 If the value is 0 (the default), KVM will pick a period based 3163 on the ratio, such that a page is zapped after 1 hour on average. 3164 3165 kvm-{amd,intel}.enable_mediated_pmu=[KVM,AMD,INTEL] 3166 If enabled, KVM will provide a mediated virtual PMU, 3167 instead of the default perf-based virtual PMU (if 3168 kvm.enable_pmu is true and PMU is enumerated via the 3169 virtual CPU model). 3170 3171 With a perf-based vPMU, KVM operates as a user of perf, 3172 i.e. emulates guest PMU counters using perf events. 3173 KVM-created perf events are managed by perf as regular 3174 (guest-only) events, e.g. are scheduled in/out, contend 3175 for hardware resources, etc. Using a perf-based vPMU 3176 allows guest and host usage of the PMU to co-exist, but 3177 incurs non-trivial overhead and can result in silently 3178 dropped guest events (due to resource contention). 3179 3180 With a mediated vPMU, hardware PMU state is context 3181 switched around the world switch to/from the guest. 3182 KVM mediates which events the guest can utilize, but 3183 gives the guest direct access to all other PMU assets 3184 when possible (KVM may intercept some accesses if the 3185 virtual CPU model provides a subset of hardware PMU 3186 functionality). Using a mediated vPMU significantly 3187 reduces PMU virtualization overhead and eliminates lost 3188 guest events, but is mutually exclusive with using perf 3189 to profile KVM guests and adds latency to most VM-Exits 3190 (to context switch PMU state). 3191 3192 Default is N (off). 3193 3194 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in 3195 KVM/SVM. Default is 1 (enabled). 3196 3197 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables, 3198 a.k.a. Two-Dimensional Page Tables. Default is 1 3199 (enabled). Disable by KVM if hardware lacks support 3200 for NPT. 3201 3202 kvm-amd.ciphertext_hiding_asids= 3203 [KVM,AMD] Ciphertext hiding prevents disallowed accesses 3204 to SNP private memory from reading ciphertext. Instead, 3205 reads will see constant default values (0xff). 3206 3207 If ciphertext hiding is enabled, the joint SEV-ES and 3208 SEV-SNP ASID space is partitioned into separate SEV-ES 3209 and SEV-SNP ASID ranges, with the SEV-SNP range being 3210 [1..max_snp_asid] and the SEV-ES range being 3211 (max_snp_asid..min_sev_asid), where min_sev_asid is 3212 enumerated by CPUID.0x.8000_001F[EDX]. 3213 3214 A non-zero value enables SEV-SNP ciphertext hiding and 3215 adjusts the ASID ranges for SEV-ES and SEV-SNP guests. 3216 KVM caps the number of SEV-SNP ASIDs at the maximum 3217 possible value, e.g. specifying -1u will assign all 3218 joint SEV-ES and SEV-SNP ASIDs to SEV-SNP. Note, 3219 assigning all joint ASIDs to SEV-SNP, i.e. configuring 3220 max_snp_asid == min_sev_asid-1, will effectively make 3221 SEV-ES unusable. 3222 3223 kvm-arm.mode= 3224 [KVM,ARM,EARLY] Select one of KVM/arm64's modes of 3225 operation. 3226 3227 none: Forcefully disable KVM. 3228 3229 nvhe: Standard nVHE-based mode, without support for 3230 protected guests. 3231 3232 protected: Mode with support for guests whose state is 3233 kept private from the host, using VHE or 3234 nVHE depending on HW support. 3235 3236 nested: VHE-based mode with support for nested 3237 virtualization. Requires at least ARMv8.4 3238 hardware (with FEAT_NV2). 3239 3240 Defaults to VHE/nVHE based on hardware support. Setting 3241 mode to "protected" will disable kexec and hibernation 3242 for the host. To force nVHE on VHE hardware, add 3243 "arm64_sw.hvhe=0 id_aa64mmfr1.vh=0" to the 3244 command-line. 3245 "nested" is experimental and should be used with 3246 extreme caution. 3247 3248 kvm-arm.vgic_v3_group0_trap= 3249 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-0 3250 system registers 3251 3252 kvm-arm.vgic_v3_group1_trap= 3253 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-1 3254 system registers 3255 3256 kvm-arm.vgic_v3_common_trap= 3257 [KVM,ARM,EARLY] Trap guest accesses to GICv3 common 3258 system registers 3259 3260 kvm-arm.vgic_v4_enable= 3261 [KVM,ARM,EARLY] Allow use of GICv4 for direct 3262 injection of LPIs. 3263 3264 kvm-arm.wfe_trap_policy= 3265 [KVM,ARM] Control when to set WFE instruction trap for 3266 KVM VMs. Traps are allowed but not guaranteed by the 3267 CPU architecture. 3268 3269 trap: set WFE instruction trap 3270 3271 notrap: clear WFE instruction trap 3272 3273 kvm-arm.wfi_trap_policy= 3274 [KVM,ARM] Control when to set WFI instruction trap for 3275 KVM VMs. Traps are allowed but not guaranteed by the 3276 CPU architecture. 3277 3278 trap: set WFI instruction trap 3279 3280 notrap: clear WFI instruction trap 3281 3282 kvm_cma_resv_ratio=n [PPC,EARLY] 3283 Reserves given percentage from system memory area for 3284 contiguous memory allocation for KVM hash pagetable 3285 allocation. 3286 By default it reserves 5% of total system memory. 3287 Format: <integer> 3288 Default: 5 3289 3290 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables, 3291 a.k.a. Two-Dimensional Page Tables. Default is 1 3292 (enabled). Disable by KVM if hardware lacks support 3293 for EPT. 3294 3295 kvm-intel.emulate_invalid_guest_state= 3296 [KVM,Intel] Control whether to emulate invalid guest 3297 state. Ignored if kvm-intel.enable_unrestricted_guest=1, 3298 as guest state is never invalid for unrestricted 3299 guests. This param doesn't apply to nested guests (L2), 3300 as KVM never emulates invalid L2 guest state. 3301 Default is 1 (enabled). 3302 3303 kvm-intel.flexpriority= 3304 [KVM,Intel] Control KVM's use of FlexPriority feature 3305 (TPR shadow). Default is 1 (enabled). Disable by KVM if 3306 hardware lacks support for it. 3307 3308 kvm-intel.nested= 3309 [KVM,Intel] Control nested virtualization feature in 3310 KVM/VMX. Default is 1 (enabled). 3311 3312 kvm-intel.unrestricted_guest= 3313 [KVM,Intel] Control KVM's use of unrestricted guest 3314 feature (virtualized real and unpaged mode). Default 3315 is 1 (enabled). Disable by KVM if EPT is disabled or 3316 hardware lacks support for it. 3317 3318 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault 3319 CVE-2018-3620. 3320 3321 Valid arguments: never, cond, always 3322 3323 always: L1D cache flush on every VMENTER. 3324 cond: Flush L1D on VMENTER only when the code between 3325 VMEXIT and VMENTER can leak host memory. 3326 never: Disables the mitigation 3327 3328 Default is cond (do L1 cache flush in specific instances) 3329 3330 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor 3331 Identification feature (tagged TLBs). Default is 1 3332 (enabled). Disable by KVM if hardware lacks support 3333 for it. 3334 3335 l1d_flush= [X86,INTEL,EARLY] 3336 Control mitigation for L1D based snooping vulnerability. 3337 3338 Certain CPUs are vulnerable to an exploit against CPU 3339 internal buffers which can forward information to a 3340 disclosure gadget under certain conditions. 3341 3342 In vulnerable processors, the speculatively 3343 forwarded data can be used in a cache side channel 3344 attack, to access data to which the attacker does 3345 not have direct access. 3346 3347 This parameter controls the mitigation. The 3348 options are: 3349 3350 on - enable the interface for the mitigation 3351 3352 l1tf= [X86,EARLY] Control mitigation of the L1TF vulnerability on 3353 affected CPUs 3354 3355 The kernel PTE inversion protection is unconditionally 3356 enabled and cannot be disabled. 3357 3358 full 3359 Provides all available mitigations for the 3360 L1TF vulnerability. Disables SMT and 3361 enables all mitigations in the 3362 hypervisors, i.e. unconditional L1D flush. 3363 3364 SMT control and L1D flush control via the 3365 sysfs interface is still possible after 3366 boot. Hypervisors will issue a warning 3367 when the first VM is started in a 3368 potentially insecure configuration, 3369 i.e. SMT enabled or L1D flush disabled. 3370 3371 full,force 3372 Same as 'full', but disables SMT and L1D 3373 flush runtime control. Implies the 3374 'nosmt=force' command line option. 3375 (i.e. sysfs control of SMT is disabled.) 3376 3377 flush 3378 Leaves SMT enabled and enables the default 3379 hypervisor mitigation, i.e. conditional 3380 L1D flush. 3381 3382 SMT control and L1D flush control via the 3383 sysfs interface is still possible after 3384 boot. Hypervisors will issue a warning 3385 when the first VM is started in a 3386 potentially insecure configuration, 3387 i.e. SMT enabled or L1D flush disabled. 3388 3389 flush,nosmt 3390 3391 Disables SMT and enables the default 3392 hypervisor mitigation. 3393 3394 SMT control and L1D flush control via the 3395 sysfs interface is still possible after 3396 boot. Hypervisors will issue a warning 3397 when the first VM is started in a 3398 potentially insecure configuration, 3399 i.e. SMT enabled or L1D flush disabled. 3400 3401 flush,nowarn 3402 Same as 'flush', but hypervisors will not 3403 warn when a VM is started in a potentially 3404 insecure configuration. 3405 3406 off 3407 Disables hypervisor mitigations and doesn't 3408 emit any warnings. 3409 It also drops the swap size and available 3410 RAM limit restriction on both hypervisor and 3411 bare metal. 3412 3413 Default is 'flush'. 3414 3415 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst 3416 3417 l2cr= [PPC] 3418 3419 l3cr= [PPC] 3420 3421 lapic [X86-32,APIC,EARLY] Enable the local APIC even if BIOS 3422 disabled it. 3423 3424 lapic= [X86,APIC] Do not use TSC deadline 3425 value for LAPIC timer one-shot implementation. Default 3426 back to the programmable timer unit in the LAPIC. 3427 Format: notscdeadline 3428 3429 lapic_timer_c2_ok [X86,APIC,EARLY] trust the local apic timer 3430 in C2 power state. 3431 3432 libata.dma= [LIBATA] DMA control 3433 libata.dma=0 Disable all PATA and SATA DMA 3434 libata.dma=1 PATA and SATA Disk DMA only 3435 libata.dma=2 ATAPI (CDROM) DMA only 3436 libata.dma=4 Compact Flash DMA only 3437 Combinations also work, so libata.dma=3 enables DMA 3438 for disks and CDROMs, but not CFs. 3439 3440 libata.ignore_hpa= [LIBATA] Ignore HPA limit 3441 libata.ignore_hpa=0 keep BIOS limits (default) 3442 libata.ignore_hpa=1 ignore limits, using full disk 3443 3444 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume 3445 when set. 3446 Format: <int> 3447 3448 libata.force= [LIBATA] Force configurations. The format is a comma- 3449 separated list of "[ID:]VAL" where ID is PORT[.DEVICE]. 3450 PORT and DEVICE are decimal numbers matching port, link 3451 or device. Basically, it matches the ATA ID string 3452 printed on console by libata. If the whole ID part is 3453 omitted, the last PORT and DEVICE values are used. If 3454 ID hasn't been specified yet, the configuration applies 3455 to all ports, links and devices. 3456 3457 If only DEVICE is omitted, the parameter applies to 3458 the port and all links and devices behind it. DEVICE 3459 number of 0 either selects the first device or the 3460 first fan-out link behind PMP device. It does not 3461 select the host link. DEVICE number of 15 selects the 3462 host link and device attached to it. 3463 3464 The VAL specifies the configuration to force. As long 3465 as there is no ambiguity, shortcut notation is allowed. 3466 For example, both 1.5 and 1.5G would work for 1.5Gbps. 3467 The following configurations can be forced. 3468 3469 * Cable type: 40c, 80c, short40c, unk, ign or sata. 3470 Any ID with matching PORT is used. 3471 3472 * SATA link speed limit: 1.5Gbps or 3.0Gbps. 3473 3474 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7]. 3475 udma[/][16,25,33,44,66,100,133] notation is also 3476 allowed. 3477 3478 * nohrst, nosrst, norst: suppress hard, soft and both 3479 resets. 3480 3481 * rstonce: only attempt one reset during hot-unplug 3482 link recovery. 3483 3484 * [no]dbdelay: Enable or disable the extra 200ms delay 3485 before debouncing a link PHY and device presence 3486 detection. 3487 3488 * [no]ncq: Turn on or off NCQ. 3489 3490 * [no]ncqtrim: Enable or disable queued DSM TRIM. 3491 3492 * [no]ncqati: Enable or disable NCQ trim on ATI chipset. 3493 3494 * [no]trim: Enable or disable (unqueued) TRIM. 3495 3496 * trim_zero: Indicate that TRIM command zeroes data. 3497 3498 * max_trim_128m: Set 128M maximum trim size limit. 3499 3500 * [no]dma: Turn on or off DMA transfers. 3501 3502 * atapi_dmadir: Enable ATAPI DMADIR bridge support. 3503 3504 * atapi_mod16_dma: Enable the use of ATAPI DMA for 3505 commands that are not a multiple of 16 bytes. 3506 3507 * [no]dmalog: Enable or disable the use of the 3508 READ LOG DMA EXT command to access logs. 3509 3510 * [no]iddevlog: Enable or disable access to the 3511 identify device data log. 3512 3513 * [no]logdir: Enable or disable access to the general 3514 purpose log directory. 3515 3516 * max_sec=<sectors>: Set the transfer size limit, in 3517 number of 512-byte sectors, to the value specified in 3518 <sectors>. The value specified in <sectors> has to be 3519 a non-zero positive integer. 3520 3521 * max_sec_128: Set transfer size limit to 128 sectors. 3522 3523 * max_sec_1024: Set or clear transfer size limit to 3524 1024 sectors. 3525 3526 * max_sec_lba48: Set or clear transfer size limit to 3527 65535 sectors. 3528 3529 * external: Mark port as external (hotplug-capable). 3530 3531 * [no]lpm: Enable or disable link power management. 3532 3533 * [no]setxfer: Indicate if transfer speed mode setting 3534 should be skipped. 3535 3536 * [no]fua: Disable or enable FUA (Force Unit Access) 3537 support for devices supporting this feature. 3538 3539 * dump_id: Dump IDENTIFY data. 3540 3541 * disable: Disable this device. 3542 3543 If there are multiple matching configurations changing 3544 the same attribute, the last one is used. 3545 3546 liveupdate= [KNL,EARLY] 3547 Format: <bool> 3548 Enable Live Update Orchestrator (LUO). 3549 Default: off. 3550 3551 lockd.nlm_grace_period=P [NFS] Assign grace period. 3552 Format: <integer> 3553 3554 lockd.nlm_tcpport=N [NFS] Assign TCP port. 3555 Format: <integer> 3556 3557 lockd.nlm_timeout=T [NFS] Assign timeout value. 3558 Format: <integer> 3559 3560 lockd.nlm_udpport=M [NFS] Assign UDP port. 3561 Format: <integer> 3562 3563 lockdown= [SECURITY,EARLY] 3564 { integrity | confidentiality } 3565 Enable the kernel lockdown feature. If set to 3566 integrity, kernel features that allow userland to 3567 modify the running kernel are disabled. If set to 3568 confidentiality, kernel features that allow userland 3569 to extract confidential information from the kernel 3570 are also disabled. 3571 3572 locktorture.acq_writer_lim= [KNL] 3573 Set the time limit in jiffies for a lock 3574 acquisition. Acquisitions exceeding this limit 3575 will result in a splat once they do complete. 3576 3577 locktorture.bind_readers= [KNL] 3578 Specify the list of CPUs to which the readers are 3579 to be bound. 3580 3581 locktorture.bind_writers= [KNL] 3582 Specify the list of CPUs to which the writers are 3583 to be bound. 3584 3585 locktorture.call_rcu_chains= [KNL] 3586 Specify the number of self-propagating call_rcu() 3587 chains to set up. These are used to ensure that 3588 there is a high probability of an RCU grace period 3589 in progress at any given time. Defaults to 0, 3590 which disables these call_rcu() chains. 3591 3592 locktorture.long_hold= [KNL] 3593 Specify the duration in milliseconds for the 3594 occasional long-duration lock hold time. Defaults 3595 to 100 milliseconds. Select 0 to disable. 3596 3597 locktorture.nested_locks= [KNL] 3598 Specify the maximum lock nesting depth that 3599 locktorture is to exercise, up to a limit of 8 3600 (MAX_NESTED_LOCKS). Specify zero to disable. 3601 Note that this parameter is ineffective on types 3602 of locks that do not support nested acquisition. 3603 3604 locktorture.nreaders_stress= [KNL] 3605 Set the number of locking read-acquisition kthreads. 3606 Defaults to being automatically set based on the 3607 number of online CPUs. 3608 3609 locktorture.nwriters_stress= [KNL] 3610 Set the number of locking write-acquisition kthreads. 3611 3612 locktorture.onoff_holdoff= [KNL] 3613 Set time (s) after boot for CPU-hotplug testing. 3614 3615 locktorture.onoff_interval= [KNL] 3616 Set time (s) between CPU-hotplug operations, or 3617 zero to disable CPU-hotplug testing. 3618 3619 locktorture.rt_boost= [KNL] 3620 Do periodic testing of real-time lock priority 3621 boosting. Select 0 to disable, 1 to boost 3622 only rt_mutex, and 2 to boost unconditionally. 3623 Defaults to 2, which might seem to be an 3624 odd choice, but which should be harmless for 3625 non-real-time spinlocks, due to their disabling 3626 of preemption. Note that non-realtime mutexes 3627 disable boosting. 3628 3629 locktorture.rt_boost_factor= [KNL] 3630 Number that determines how often and for how 3631 long priority boosting is exercised. This is 3632 scaled down by the number of writers, so that the 3633 number of boosts per unit time remains roughly 3634 constant as the number of writers increases. 3635 On the other hand, the duration of each boost 3636 increases with the number of writers. 3637 3638 locktorture.shuffle_interval= [KNL] 3639 Set task-shuffle interval (jiffies). Shuffling 3640 tasks allows some CPUs to go into dyntick-idle 3641 mode during the locktorture test. 3642 3643 locktorture.shutdown_secs= [KNL] 3644 Set time (s) after boot system shutdown. This 3645 is useful for hands-off automated testing. 3646 3647 locktorture.stat_interval= [KNL] 3648 Time (s) between statistics printk()s. 3649 3650 locktorture.stutter= [KNL] 3651 Time (s) to stutter testing, for example, 3652 specifying five seconds causes the test to run for 3653 five seconds, wait for five seconds, and so on. 3654 This tests the locking primitive's ability to 3655 transition abruptly to and from idle. 3656 3657 locktorture.torture_type= [KNL] 3658 Specify the locking implementation to test. 3659 3660 locktorture.verbose= [KNL] 3661 Enable additional printk() statements. 3662 3663 locktorture.writer_fifo= [KNL] 3664 Run the write-side locktorture kthreads at 3665 sched_set_fifo() real-time priority. 3666 3667 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver 3668 Format: <irq> 3669 3670 loglevel= [KNL,EARLY] 3671 All Kernel Messages with a loglevel smaller than the 3672 console loglevel will be printed to the console. It can 3673 also be changed with klogd or other programs. The 3674 loglevels are defined as follows: 3675 3676 0 (KERN_EMERG) system is unusable 3677 1 (KERN_ALERT) action must be taken immediately 3678 2 (KERN_CRIT) critical conditions 3679 3 (KERN_ERR) error conditions 3680 4 (KERN_WARNING) warning conditions 3681 5 (KERN_NOTICE) normal but significant condition 3682 6 (KERN_INFO) informational 3683 7 (KERN_DEBUG) debug-level messages 3684 3685 log_buf_len=n[KMG] [KNL,EARLY] 3686 Sets the size of the printk ring buffer, in bytes. 3687 n must be a power of two and greater than the 3688 minimal size. The minimal size is defined by 3689 LOG_BUF_SHIFT kernel config parameter. There 3690 is also CONFIG_LOG_CPU_MAX_BUF_SHIFT config 3691 parameter that allows to increase the default size 3692 depending on the number of CPUs. See init/Kconfig 3693 for more details. 3694 3695 logo.nologo [FB] Disables display of the built-in Linux logo. 3696 This may be used to provide more screen space for 3697 kernel log messages and is useful when debugging 3698 kernel boot problems. 3699 3700 lp=0 [LP] Specify parallel ports to use, e.g, 3701 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses 3702 lp=reset first parallel port). 'lp=0' disables the 3703 lp=auto printer driver. 'lp=reset' (which can be 3704 specified in addition to the ports) causes 3705 attached printers to be reset. Using 3706 lp=port1,port2,... specifies the parallel ports 3707 to associate lp devices with, starting with 3708 lp0. A port specification may be 'none' to skip 3709 that lp device, or a parport name such as 3710 'parport0'. Specifying 'lp=auto' instead of a 3711 port specification list means that device IDs 3712 from each port should be examined, to see if 3713 an IEEE 1284-compliant printer is attached; if 3714 so, the driver will manage that printer. 3715 See also header of drivers/char/lp.c. 3716 3717 lpj=n [KNL] 3718 Sets loops_per_jiffy to given constant, thus avoiding 3719 time-consuming boot-time autodetection (up to 250 ms per 3720 CPU). 0 enables autodetection (default). To determine 3721 the correct value for your kernel, boot with normal 3722 autodetection and see what value is printed. Note that 3723 on SMP systems the preset will be applied to all CPUs, 3724 which is likely to cause problems if your CPUs need 3725 significantly divergent settings. An incorrect value 3726 will cause delays in the kernel to be wrong, leading to 3727 unpredictable I/O errors and other breakage. Although 3728 unlikely, in the extreme case this might damage your 3729 hardware. 3730 3731 lsm.debug [SECURITY] Enable LSM initialization debugging output. 3732 3733 lsm=lsm1,...,lsmN 3734 [SECURITY] Choose order of LSM initialization. This 3735 overrides CONFIG_LSM, and the "security=" parameter. 3736 3737 machtype= [Loongson] Share the same kernel image file between 3738 different yeeloong laptops. 3739 Example: machtype=lemote-yeeloong-2f-7inch 3740 3741 maxcpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 3742 will bring up during bootup. maxcpus=n : n >= 0 limits 3743 the kernel to bring up 'n' processors. Surely after 3744 bootup you can bring up the other plugged cpu by executing 3745 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus 3746 only takes effect during system bootup. 3747 While n=0 is a special case, it is equivalent to "nosmp", 3748 which also disables the IO APIC. 3749 3750 max_loop= [LOOP] The number of loop block devices that get 3751 (loop.max_loop) unconditionally pre-created at init time. The default 3752 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead 3753 of statically allocating a predefined number, loop 3754 devices can be requested on-demand with the 3755 /dev/loop-control interface. 3756 3757 mce= [X86-{32,64}] 3758 3759 Please see Documentation/arch/x86/x86_64/machinecheck.rst for sysfs runtime tunables. 3760 3761 off 3762 disable machine check 3763 3764 no_cmci 3765 disable CMCI(Corrected Machine Check Interrupt) that 3766 Intel processor supports. Usually this disablement is 3767 not recommended, but it might be handy if your 3768 hardware is misbehaving. 3769 3770 Note that you'll get more problems without CMCI than 3771 with due to the shared banks, i.e. you might get 3772 duplicated error logs. 3773 3774 dont_log_ce 3775 don't make logs for corrected errors. All events 3776 reported as corrected are silently cleared by OS. This 3777 option will be useful if you have no interest in any 3778 of corrected errors. 3779 3780 ignore_ce 3781 disable features for corrected errors, e.g. 3782 polling timer and CMCI. All events reported as 3783 corrected are not cleared by OS and remained in its 3784 error banks. 3785 3786 Usually this disablement is not recommended, however 3787 if there is an agent checking/clearing corrected 3788 errors (e.g. BIOS or hardware monitoring 3789 applications), conflicting with OS's error handling, 3790 and you cannot deactivate the agent, then this option 3791 will be a help. 3792 3793 no_lmce 3794 do not opt-in to Local MCE delivery. Use legacy method 3795 to broadcast MCEs. 3796 3797 bootlog 3798 enable logging of machine checks left over from 3799 booting. Disabled by default on AMD Fam10h and older 3800 because some BIOS leave bogus ones. 3801 3802 If your BIOS doesn't do that it's a good idea to 3803 enable though to make sure you log even machine check 3804 events that result in a reboot. On Intel systems it is 3805 enabled by default. 3806 3807 nobootlog 3808 disable boot machine check logging. 3809 3810 monarchtimeout (number) 3811 sets the time in us to wait for other CPUs on machine 3812 checks. 0 to disable. 3813 3814 bios_cmci_threshold 3815 don't overwrite the bios-set CMCI threshold. This boot 3816 option prevents Linux from overwriting the CMCI 3817 threshold set by the bios. Without this option, Linux 3818 always sets the CMCI threshold to 1. Enabling this may 3819 make memory predictive failure analysis less effective 3820 if the bios sets thresholds for memory errors since we 3821 will not see details for all errors. 3822 3823 recovery 3824 force-enable recoverable machine check code paths 3825 3826 Everything else is in sysfs now. 3827 3828 3829 md= [HW] RAID subsystems devices and level 3830 See Documentation/admin-guide/md.rst. 3831 3832 mdacon= [MDA] 3833 Format: <first>,<last> 3834 Specifies range of consoles to be captured by the MDA. 3835 3836 mds= [X86,INTEL,EARLY] 3837 Control mitigation for the Micro-architectural Data 3838 Sampling (MDS) vulnerability. 3839 3840 Certain CPUs are vulnerable to an exploit against CPU 3841 internal buffers which can forward information to a 3842 disclosure gadget under certain conditions. 3843 3844 In vulnerable processors, the speculatively 3845 forwarded data can be used in a cache side channel 3846 attack, to access data to which the attacker does 3847 not have direct access. 3848 3849 This parameter controls the MDS mitigation. The 3850 options are: 3851 3852 full - Enable MDS mitigation on vulnerable CPUs 3853 full,nosmt - Enable MDS mitigation and disable 3854 SMT on vulnerable CPUs 3855 off - Unconditionally disable MDS mitigation 3856 3857 On TAA-affected machines, mds=off can be prevented by 3858 an active TAA mitigation as both vulnerabilities are 3859 mitigated with the same mechanism so in order to disable 3860 this mitigation, you need to specify tsx_async_abort=off 3861 too. 3862 3863 Not specifying this option is equivalent to 3864 mds=full. 3865 3866 For details see: Documentation/admin-guide/hw-vuln/mds.rst 3867 3868 mem=nn[KMG] [HEXAGON,EARLY] Set the memory size. 3869 Must be specified, otherwise memory size will be 0. 3870 3871 mem=nn[KMG] [KNL,BOOT,EARLY] Force usage of a specific amount 3872 of memory Amount of memory to be used in cases 3873 as follows: 3874 3875 1 for test; 3876 2 when the kernel is not able to see the whole system memory; 3877 3 memory that lies after 'mem=' boundary is excluded from 3878 the hypervisor, then assigned to KVM guests. 3879 4 to limit the memory available for kdump kernel. 3880 3881 [ARC,MICROBLAZE] - the limit applies only to low memory, 3882 high memory is not affected. 3883 3884 [ARM64] - only limits memory covered by the linear 3885 mapping. The NOMAP regions are not affected. 3886 3887 [X86] Work as limiting max address. Use together 3888 with memmap= to avoid physical address space collisions. 3889 Without memmap= PCI devices could be placed at addresses 3890 belonging to unused RAM. 3891 3892 Note that this only takes effects during boot time since 3893 in above case 3, memory may need be hot added after boot 3894 if system memory of hypervisor is not sufficient. 3895 3896 mem=nn[KMG]@ss[KMG] 3897 [ARM,MIPS,EARLY] - override the memory layout 3898 reported by firmware. 3899 Define a memory region of size nn[KMG] starting at 3900 ss[KMG]. 3901 Multiple different regions can be specified with 3902 multiple mem= parameters on the command line. 3903 3904 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel 3905 memory. 3906 3907 memblock=debug [KNL,EARLY] Enable memblock debug messages. 3908 3909 memchunk=nn[KMG] 3910 [KNL,SH] Allow user to override the default size for 3911 per-device physically contiguous DMA buffers. 3912 3913 memhp_default_state=online/offline/online_kernel/online_movable 3914 [KNL] Set the initial state for the memory hotplug 3915 onlining policy. If not specified, the default value is 3916 set according to the 3917 CONFIG_MHP_DEFAULT_ONLINE_TYPE kernel config 3918 options. 3919 See Documentation/admin-guide/mm/memory-hotplug.rst. 3920 3921 memmap=exactmap [KNL,X86,EARLY] Enable setting of an exact 3922 E820 memory map, as specified by the user. 3923 Such memmap=exactmap lines can be constructed based on 3924 BIOS output or other requirements. See the memmap=nn@ss 3925 option description. 3926 3927 memmap=nn[KMG]@ss[KMG] 3928 [KNL, X86,MIPS,XTENSA,EARLY] Force usage of a specific region of memory. 3929 Region of memory to be used is from ss to ss+nn. 3930 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG], 3931 which limits max address to nn[KMG]. 3932 Multiple different regions can be specified, 3933 comma delimited. 3934 Example: 3935 memmap=100M@2G,100M#3G,1G!1024G 3936 3937 memmap=nn[KMG]#ss[KMG] 3938 [KNL,ACPI,EARLY] Mark specific memory as ACPI data. 3939 Region of memory to be marked is from ss to ss+nn. 3940 3941 memmap=nn[KMG]$ss[KMG] 3942 [KNL,ACPI,EARLY] Mark specific memory as reserved. 3943 Region of memory to be reserved is from ss to ss+nn. 3944 Example: Exclude memory from 0x18690000-0x1869ffff 3945 memmap=64K$0x18690000 3946 or 3947 memmap=0x10000$0x18690000 3948 Some bootloaders may need an escape character before '$', 3949 like Grub2, otherwise '$' and the following number 3950 will be eaten. 3951 3952 memmap=nn[KMG]!ss[KMG,EARLY] 3953 [KNL,X86] Mark specific memory as protected. 3954 Region of memory to be used, from ss to ss+nn. 3955 The memory region may be marked as e820 type 12 (0xc) 3956 and is NVDIMM or ADR memory. 3957 3958 memmap=<size>%<offset>-<oldtype>+<newtype> 3959 [KNL,ACPI,EARLY] Convert memory within the specified region 3960 from <oldtype> to <newtype>. If "-<oldtype>" is left 3961 out, the whole region will be marked as <newtype>, 3962 even if previously unavailable. If "+<newtype>" is left 3963 out, matching memory will be removed. Types are 3964 specified as e820 types, e.g., 1 = RAM, 2 = reserved, 3965 3 = ACPI, 12 = PRAM. 3966 3967 memory_corruption_check=0/1 [X86,EARLY] 3968 Some BIOSes seem to corrupt the first 64k of 3969 memory when doing things like suspend/resume. 3970 Setting this option will scan the memory 3971 looking for corruption. Enabling this will 3972 both detect corruption and prevent the kernel 3973 from using the memory being corrupted. 3974 However, it's intended as a diagnostic tool; if 3975 repeatable BIOS-originated corruption always 3976 affects the same memory, you can use memmap= 3977 to prevent the kernel from using that memory. 3978 3979 memory_corruption_check_size=size [X86,EARLY] 3980 By default it checks for corruption in the low 3981 64k, making this memory unavailable for normal 3982 use. Use this parameter to scan for 3983 corruption in more or less memory. 3984 3985 memory_corruption_check_period=seconds [X86,EARLY] 3986 By default it checks for corruption every 60 3987 seconds. Use this parameter to check at some 3988 other rate. 0 disables periodic checking. 3989 3990 memory_hotplug.memmap_on_memory 3991 [KNL,X86,ARM] Boolean flag to enable this feature. 3992 Format: {on | off (default)} 3993 When enabled, runtime hotplugged memory will 3994 allocate its internal metadata (struct pages, 3995 those vmemmap pages cannot be optimized even 3996 if hugetlb_free_vmemmap is enabled) from the 3997 hotadded memory which will allow to hotadd a 3998 lot of memory without requiring additional 3999 memory to do so. 4000 This feature is disabled by default because it 4001 has some implication on large (e.g. GB) 4002 allocations in some configurations (e.g. small 4003 memory blocks). 4004 The state of the flag can be read in 4005 /sys/module/memory_hotplug/parameters/memmap_on_memory. 4006 Note that even when enabled, there are a few cases where 4007 the feature is not effective. 4008 4009 memtest= [KNL,X86,ARM,M68K,PPC,RISCV,EARLY] Enable memtest 4010 Format: <integer> 4011 default : 0 <disable> 4012 Specifies the number of memtest passes to be 4013 performed. Each pass selects another test 4014 pattern from a given set of patterns. Memtest 4015 fills the memory with this pattern, validates 4016 memory contents and reserves bad memory 4017 regions that are detected. 4018 4019 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control 4020 Valid arguments: on, off 4021 Default: off 4022 mem_encrypt=on: Activate SME 4023 mem_encrypt=off: Do not activate SME 4024 4025 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst 4026 for details on when memory encryption can be activated. 4027 4028 mem_sleep_default= [SUSPEND] Default system suspend mode: 4029 s2idle - Suspend-To-Idle 4030 shallow - Power-On Suspend or equivalent (if supported) 4031 deep - Suspend-To-RAM or equivalent (if supported) 4032 See Documentation/admin-guide/pm/sleep-states.rst. 4033 4034 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when 4035 the BIOS has incorrectly applied a workaround. TinyBIOS 4036 version 0.98 is known to be affected, 0.99 fixes the 4037 problem by letting the user disable the workaround. 4038 4039 mga= [HW,DRM] 4040 4041 microcode= [X86] Control the behavior of the microcode loader. 4042 Available options, comma separated: 4043 4044 base_rev=X - with <X> with format: <u32> 4045 Set the base microcode revision of each thread when in 4046 debug mode. 4047 4048 dis_ucode_ldr: disable the microcode loader 4049 4050 force_minrev: 4051 Enable or disable the microcode minimal revision 4052 enforcement for the runtime microcode loader. 4053 4054 mini2440= [ARM,HW,KNL] 4055 Format:[0..2][b][c][t] 4056 Default: "0tb" 4057 MINI2440 configuration specification: 4058 0 - The attached screen is the 3.5" TFT 4059 1 - The attached screen is the 7" TFT 4060 2 - The VGA Shield is attached (1024x768) 4061 Leaving out the screen size parameter will not load 4062 the TFT driver, and the framebuffer will be left 4063 unconfigured. 4064 b - Enable backlight. The TFT backlight pin will be 4065 linked to the kernel VESA blanking code and a GPIO 4066 LED. This parameter is not necessary when using the 4067 VGA shield. 4068 c - Enable the s3c camera interface. 4069 t - Reserved for enabling touchscreen support. The 4070 touchscreen support is not enabled in the mainstream 4071 kernel as of 2.6.30, a preliminary port can be found 4072 in the "bleeding edge" mini2440 support kernel at 4073 https://repo.or.cz/w/linux-2.6/mini2440.git 4074 4075 mitigations= 4076 [X86,PPC,S390,ARM64,EARLY] Control optional mitigations for 4077 CPU vulnerabilities. This is a set of curated, 4078 arch-independent options, each of which is an 4079 aggregation of existing arch-specific options. 4080 4081 Note, "mitigations" is supported if and only if the 4082 kernel was built with CPU_MITIGATIONS=y. 4083 4084 off 4085 Disable all optional CPU mitigations. This 4086 improves system performance, but it may also 4087 expose users to several CPU vulnerabilities. 4088 Equivalent to: if nokaslr then kpti=0 [ARM64] 4089 gather_data_sampling=off [X86] 4090 indirect_target_selection=off [X86] 4091 kvm.nx_huge_pages=off [X86] 4092 l1tf=off [X86] 4093 mds=off [X86] 4094 mmio_stale_data=off [X86] 4095 no_entry_flush [PPC] 4096 no_uaccess_flush [PPC] 4097 nobp=0 [S390] 4098 nopti [X86,PPC] 4099 nospectre_bhb [ARM64] 4100 nospectre_v1 [X86,PPC] 4101 nospectre_v2 [X86,PPC,S390,ARM64] 4102 reg_file_data_sampling=off [X86] 4103 retbleed=off [X86] 4104 spec_rstack_overflow=off [X86] 4105 spec_store_bypass_disable=off [X86,PPC] 4106 spectre_bhi=off [X86] 4107 spectre_v2_user=off [X86] 4108 srbds=off [X86,INTEL] 4109 ssbd=force-off [ARM64] 4110 tsa=off [X86,AMD] 4111 tsx_async_abort=off [X86] 4112 vmscape=off [X86] 4113 4114 Exceptions: 4115 This does not have any effect on 4116 kvm.nx_huge_pages when 4117 kvm.nx_huge_pages=force. 4118 4119 auto (default) 4120 Mitigate all CPU vulnerabilities, but leave SMT 4121 enabled, even if it's vulnerable. This is for 4122 users who don't want to be surprised by SMT 4123 getting disabled across kernel upgrades, or who 4124 have other ways of avoiding SMT-based attacks. 4125 Equivalent to: (default behavior) 4126 4127 auto,nosmt 4128 Mitigate all CPU vulnerabilities, disabling SMT 4129 if needed. This is for users who always want to 4130 be fully mitigated, even if it means losing SMT. 4131 Equivalent to: l1tf=flush,nosmt [X86] 4132 mds=full,nosmt [X86] 4133 tsx_async_abort=full,nosmt [X86] 4134 mmio_stale_data=full,nosmt [X86] 4135 retbleed=auto,nosmt [X86] 4136 4137 [X86] After one of the above options, additionally 4138 supports attack-vector based controls as documented in 4139 Documentation/admin-guide/hw-vuln/attack_vector_controls.rst 4140 4141 mminit_loglevel= 4142 [KNL,EARLY] When CONFIG_DEBUG_MEMORY_INIT is set, this 4143 parameter allows control of the logging verbosity for 4144 the additional memory initialisation checks. A value 4145 of 0 disables mminit logging and a level of 4 will 4146 log everything. Information is printed at KERN_DEBUG 4147 so loglevel=8 may also need to be specified. 4148 4149 mmio_stale_data= 4150 [X86,INTEL,EARLY] Control mitigation for the Processor 4151 MMIO Stale Data vulnerabilities. 4152 4153 Processor MMIO Stale Data is a class of 4154 vulnerabilities that may expose data after an MMIO 4155 operation. Exposed data could originate or end in 4156 the same CPU buffers as affected by MDS and TAA. 4157 Therefore, similar to MDS and TAA, the mitigation 4158 is to clear the affected CPU buffers. 4159 4160 This parameter controls the mitigation. The 4161 options are: 4162 4163 full - Enable mitigation on vulnerable CPUs 4164 4165 full,nosmt - Enable mitigation and disable SMT on 4166 vulnerable CPUs. 4167 4168 off - Unconditionally disable mitigation 4169 4170 On MDS or TAA affected machines, 4171 mmio_stale_data=off can be prevented by an active 4172 MDS or TAA mitigation as these vulnerabilities are 4173 mitigated with the same mechanism so in order to 4174 disable this mitigation, you need to specify 4175 mds=off and tsx_async_abort=off too. 4176 4177 Not specifying this option is equivalent to 4178 mmio_stale_data=full. 4179 4180 For details see: 4181 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst 4182 4183 <module>.async_probe[=<bool>] [KNL] 4184 If no <bool> value is specified or if the value 4185 specified is not a valid <bool>, enable asynchronous 4186 probe on this module. Otherwise, enable/disable 4187 asynchronous probe on this module as indicated by the 4188 <bool> value. See also: module.async_probe 4189 4190 module.async_probe=<bool> 4191 [KNL] When set to true, modules will use async probing 4192 by default. To enable/disable async probing for a 4193 specific module, use the module specific control that 4194 is documented under <module>.async_probe. When both 4195 module.async_probe and <module>.async_probe are 4196 specified, <module>.async_probe takes precedence for 4197 the specific module. 4198 4199 module.enable_dups_trace 4200 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set, 4201 this means that duplicate request_module() calls will 4202 trigger a WARN_ON() instead of a pr_warn(). Note that 4203 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s 4204 will always be issued and this option does nothing. 4205 module.sig_enforce 4206 [KNL] When CONFIG_MODULE_SIG is set, this means that 4207 modules without (valid) signatures will fail to load. 4208 Note that if CONFIG_MODULE_SIG_FORCE is set, that 4209 is always true, so this option does nothing. 4210 4211 module_blacklist= [KNL] Do not load a comma-separated list of 4212 modules. Useful for debugging problem modules. 4213 4214 mousedev.tap_time= 4215 [MOUSE] Maximum time between finger touching and 4216 leaving touchpad surface for touch to be considered 4217 a tap and be reported as a left button click (for 4218 touchpads working in absolute mode only). 4219 Format: <msecs> 4220 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices 4221 reporting absolute coordinates, such as tablets 4222 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices 4223 reporting absolute coordinates, such as tablets 4224 4225 movablecore= [KNL,X86,PPC,EARLY] 4226 Format: nn[KMGTPE] | nn% 4227 This parameter is the complement to kernelcore=, it 4228 specifies the amount of memory used for migratable 4229 allocations. If both kernelcore and movablecore is 4230 specified, then kernelcore will be at *least* the 4231 specified value but may be more. If movablecore on its 4232 own is specified, the administrator must be careful 4233 that the amount of memory usable for all allocations 4234 is not too small. 4235 4236 movable_node [KNL,EARLY] Boot-time switch to make hotplugable memory 4237 NUMA nodes to be movable. This means that the memory 4238 of such nodes will be usable only for movable 4239 allocations which rules out almost all kernel 4240 allocations. Use with caution! 4241 4242 MTD_Partition= [MTD] 4243 Format: <name>,<region-number>,<size>,<offset> 4244 4245 MTD_Region= [MTD] Format: 4246 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>] 4247 4248 mtdparts= [MTD] 4249 See drivers/mtd/parsers/cmdlinepart.c 4250 4251 mtouchusb.raw_coordinates= 4252 [HW] Make the MicroTouch USB driver use raw coordinates 4253 ('y', default) or cooked coordinates ('n') 4254 4255 mtrr=debug [X86,EARLY] 4256 Enable printing debug information related to MTRR 4257 registers at boot time. 4258 4259 mtrr_chunk_size=nn[KMG,X86,EARLY] 4260 used for mtrr cleanup. It is largest continuous chunk 4261 that could hold holes aka. UC entries. 4262 4263 mtrr_gran_size=nn[KMG,X86,EARLY] 4264 Used for mtrr cleanup. It is granularity of mtrr block. 4265 Default is 1. 4266 Large value could prevent small alignment from 4267 using up MTRRs. 4268 4269 mtrr_spare_reg_nr=n [X86,EARLY] 4270 Format: <integer> 4271 Range: 0,7 : spare reg number 4272 Default : 1 4273 Used for mtrr cleanup. It is spare mtrr entries number. 4274 Set to 2 or more if your graphical card needs more. 4275 4276 multitce=off [PPC] This parameter disables the use of the pSeries 4277 firmware feature for updating multiple TCE entries 4278 at a time. 4279 4280 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card 4281 4282 netdev= [NET] Network devices parameters 4283 Format: <irq>,<io>,<mem_start>,<mem_end>,<name> 4284 Note that mem_start is often overloaded to mean 4285 something different and driver-specific. 4286 This usage is only documented in each driver source 4287 file if at all. 4288 4289 netpoll.carrier_timeout= 4290 [NET] Specifies amount of time (in seconds) that 4291 netpoll should wait for a carrier. By default netpoll 4292 waits 4 seconds. 4293 4294 nf_conntrack.acct= 4295 [NETFILTER] Enable connection tracking flow accounting 4296 0 to disable accounting 4297 1 to enable accounting 4298 Default value is 0. 4299 4300 nfs.cache_getent= 4301 [NFS] sets the pathname to the program which is used 4302 to update the NFS client cache entries. 4303 4304 nfs.cache_getent_timeout= 4305 [NFS] sets the timeout after which an attempt to 4306 update a cache entry is deemed to have failed. 4307 4308 nfs.callback_nr_threads= 4309 [NFSv4] set the total number of threads that the 4310 NFS client will assign to service NFSv4 callback 4311 requests. 4312 4313 nfs.callback_tcpport= 4314 [NFS] set the TCP port on which the NFSv4 callback 4315 channel should listen. 4316 4317 nfs.delay_retrans= 4318 [NFS] specifies the number of times the NFSv4 client 4319 retries the request before returning an EAGAIN error, 4320 after a reply of NFS4ERR_DELAY from the server. 4321 Only applies if the softerr mount option is enabled, 4322 and the specified value is >= 0. 4323 4324 nfs.enable_ino64= 4325 [NFS] enable 64-bit inode numbers. 4326 If zero, the NFS client will fake up a 32-bit inode 4327 number for the readdir() and stat() syscalls instead 4328 of returning the full 64-bit number. 4329 The default is to return 64-bit inode numbers. 4330 4331 nfs.idmap_cache_timeout= 4332 [NFS] set the maximum lifetime for idmapper cache 4333 entries. 4334 4335 nfs.max_session_cb_slots= 4336 [NFSv4.1] Sets the maximum number of session 4337 slots the client will assign to the callback 4338 channel. This determines the maximum number of 4339 callbacks the client will process in parallel for 4340 a particular server. 4341 4342 nfs.max_session_slots= 4343 [NFSv4.1] Sets the maximum number of session slots 4344 the client will attempt to negotiate with the server. 4345 This limits the number of simultaneous RPC requests 4346 that the client can send to the NFSv4.1 server. 4347 Note that there is little point in setting this 4348 value higher than the max_tcp_slot_table_limit. 4349 4350 nfs.nfs4_disable_idmapping= 4351 [NFSv4] When set to the default of '1', this option 4352 ensures that both the RPC level authentication 4353 scheme and the NFS level operations agree to use 4354 numeric uids/gids if the mount is using the 4355 'sec=sys' security flavour. In effect it is 4356 disabling idmapping, which can make migration from 4357 legacy NFSv2/v3 systems to NFSv4 easier. 4358 Servers that do not support this mode of operation 4359 will be autodetected by the client, and it will fall 4360 back to using the idmapper. 4361 To turn off this behaviour, set the value to '0'. 4362 4363 nfs.nfs4_unique_id= 4364 [NFS4] Specify an additional fixed unique ident- 4365 ification string that NFSv4 clients can insert into 4366 their nfs_client_id4 string. This is typically a 4367 UUID that is generated at system install time. 4368 4369 nfs.recover_lost_locks= 4370 [NFSv4] Attempt to recover locks that were lost due 4371 to a lease timeout on the server. Please note that 4372 doing this risks data corruption, since there are 4373 no guarantees that the file will remain unchanged 4374 after the locks are lost. 4375 If you want to enable the kernel legacy behaviour of 4376 attempting to recover these locks, then set this 4377 parameter to '1'. 4378 The default parameter value of '0' causes the kernel 4379 not to attempt recovery of lost locks. 4380 4381 nfs.send_implementation_id= 4382 [NFSv4.1] Send client implementation identification 4383 information in exchange_id requests. 4384 If zero, no implementation identification information 4385 will be sent. 4386 The default is to send the implementation identification 4387 information. 4388 4389 nfs4.layoutstats_timer= 4390 [NFSv4.2] Change the rate at which the kernel sends 4391 layoutstats to the pNFS metadata server. 4392 4393 Setting this to value to 0 causes the kernel to use 4394 whatever value is the default set by the layout 4395 driver. A non-zero value sets the minimum interval 4396 in seconds between layoutstats transmissions. 4397 4398 nfsd.inter_copy_offload_enable= 4399 [NFSv4.2] When set to 1, the server will support 4400 server-to-server copies for which this server is 4401 the destination of the copy. 4402 4403 nfsd.nfs4_disable_idmapping= 4404 [NFSv4] When set to the default of '1', the NFSv4 4405 server will return only numeric uids and gids to 4406 clients using auth_sys, and will accept numeric uids 4407 and gids from such clients. This is intended to ease 4408 migration from NFSv2/v3. 4409 4410 nfsd.nfsd4_ssc_umount_timeout= 4411 [NFSv4.2] When used as the destination of a 4412 server-to-server copy, knfsd temporarily mounts 4413 the source server. It caches the mount in case 4414 it will be needed again, and discards it if not 4415 used for the number of milliseconds specified by 4416 this parameter. 4417 4418 nfsaddrs= [NFS] Deprecated. Use ip= instead. 4419 See Documentation/admin-guide/nfs/nfsroot.rst. 4420 4421 nfsroot= [NFS] nfs root filesystem for disk-less boxes. 4422 See Documentation/admin-guide/nfs/nfsroot.rst. 4423 4424 nfsrootdebug [NFS] enable nfsroot debugging messages. 4425 See Documentation/admin-guide/nfs/nfsroot.rst. 4426 4427 nmi_backtrace.backtrace_idle [KNL] 4428 Dump stacks even of idle CPUs in response to an 4429 NMI stack-backtrace request. 4430 4431 nmi_debug= [KNL,SH] Specify one or more actions to take 4432 when a NMI is triggered. 4433 Format: [state][,regs][,debounce][,die] 4434 4435 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels 4436 Format: [panic,][nopanic,][rNNN,][num] 4437 Valid num: 0 or 1 4438 0 - turn hardlockup detector in nmi_watchdog off 4439 1 - turn hardlockup detector in nmi_watchdog on 4440 rNNN - configure the watchdog with raw perf event 0xNNN 4441 4442 When panic is specified, panic when an NMI watchdog 4443 timeout occurs (or 'nopanic' to not panic on an NMI 4444 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set) 4445 To disable both hard and soft lockup detectors, 4446 please see 'nowatchdog'. 4447 This is useful when you use a panic=... timeout and 4448 need the box quickly up again. 4449 4450 These settings can be accessed at runtime via 4451 the nmi_watchdog and hardlockup_panic sysctls. 4452 4453 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths 4454 emulation library even if a 387 maths coprocessor 4455 is present. 4456 4457 no4lvl [RISCV,EARLY] Disable 4-level and 5-level paging modes. 4458 Forces kernel to use 3-level paging instead. 4459 4460 no5lvl [X86-64,RISCV,EARLY] Disable 5-level paging mode. Forces 4461 kernel to use 4-level paging instead. 4462 4463 noalign [KNL,ARM] 4464 4465 noapic [SMP,APIC,EARLY] Tells the kernel to not make use of any 4466 IOAPICs that may be present in the system. 4467 4468 noapictimer [APIC,X86] Don't set up the APIC timer 4469 4470 noautogroup Disable scheduler automatic task group creation. 4471 4472 nocache [ARM,EARLY] 4473 4474 no_console_suspend 4475 [HW] Never suspend the console 4476 Disable suspending of consoles during suspend and 4477 hibernate operations. Once disabled, debugging 4478 messages can reach various consoles while the rest 4479 of the system is being put to sleep (ie, while 4480 debugging driver suspend/resume hooks). This may 4481 not work reliably with all consoles, but is known 4482 to work with serial and VGA consoles. 4483 To facilitate more flexible debugging, we also add 4484 console_suspend, a printk module parameter to control 4485 it. Users could use console_suspend (usually 4486 /sys/module/printk/parameters/console_suspend) to 4487 turn on/off it dynamically. 4488 4489 no_debug_objects 4490 [KNL,EARLY] Disable object debugging 4491 4492 nodsp [SH] Disable hardware DSP at boot time. 4493 4494 noefi [EFI,EARLY] Disable EFI runtime services support. 4495 4496 no_entry_flush [PPC,EARLY] Don't flush the L1-D cache when entering the kernel. 4497 4498 noexec32 [X86-64] 4499 This affects only 32-bit executables. 4500 noexec32=on: enable non-executable mappings (default) 4501 read doesn't imply executable mappings 4502 noexec32=off: disable non-executable mappings 4503 read implies executable mappings 4504 4505 no_file_caps Tells the kernel not to honor file capabilities. The 4506 only way then for a file to be executed with privilege 4507 is to be setuid root or executed by root. 4508 4509 nofpu [MIPS,SH] Disable hardware FPU at boot time. 4510 4511 nofsgsbase [X86] Disables FSGSBASE instructions. 4512 4513 nofxsr [BUGS=X86-32] Disables x86 floating point extended 4514 register save and restore. The kernel will only save 4515 legacy floating-point registers on task switch. 4516 4517 nogbpages [X86] Do not use GB pages for kernel direct mappings. 4518 4519 no_hash_pointers 4520 [KNL,EARLY] 4521 Alias for "hash_pointers=never". 4522 4523 nohibernate [HIBERNATION] Disable hibernation and resume. 4524 4525 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,RISCV,SH] Forces the kernel to 4526 busy wait in do_idle() and not use the arch_cpu_idle() 4527 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP 4528 to be effective. This is useful on platforms where the 4529 sleep(SH) or wfi(ARM,ARM64) instructions do not work 4530 correctly or when doing power measurements to evaluate 4531 the impact of the sleep instructions. This is also 4532 useful when using JTAG debugger. 4533 4534 nohpet [X86] Don't use the HPET timer. 4535 4536 nohugeiomap [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge I/O mappings. 4537 4538 nohugevmalloc [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge vmalloc mappings. 4539 4540 nohz= [KNL] Boottime enable/disable dynamic ticks 4541 Valid arguments: on, off 4542 Default: on 4543 4544 nohz_full= [KNL,BOOT,SMP,ISOL] 4545 The argument is a cpu list, as described above. 4546 In kernels built with CONFIG_NO_HZ_FULL=y, set 4547 the specified list of CPUs whose tick will be stopped 4548 whenever possible. The boot CPU will be forced outside 4549 the range to maintain the timekeeping. Any CPUs 4550 in this list will have their RCU callbacks offloaded, 4551 just as if they had also been called out in the 4552 rcu_nocbs= boot parameter. 4553 4554 Note that this argument takes precedence over 4555 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 4556 4557 noinitrd [Deprecated,RAM] Tells the kernel not to load any configured 4558 initial RAM disk. Currently this parameter applies to 4559 initrd only, not to initramfs. But it applies to both 4560 in EFI mode. 4561 4562 nointremap [X86-64,Intel-IOMMU,EARLY] Do not enable interrupt 4563 remapping. 4564 [Deprecated - use intremap=off] 4565 4566 noinvpcid [X86,EARLY] Disable the INVPCID cpu feature. 4567 4568 noiotrap [SH] Disables trapped I/O port accesses. 4569 4570 noirqdebug [X86-32] Disables the code which attempts to detect and 4571 disable unhandled interrupt sources. 4572 4573 noisapnp [ISAPNP] Disables ISA PnP code. 4574 4575 nokaslr [KNL,EARLY] 4576 When CONFIG_RANDOMIZE_BASE is set, this disables 4577 kernel and module base offset ASLR (Address Space 4578 Layout Randomization). 4579 4580 no-kvmapf [X86,KVM,EARLY] Disable paravirtualized asynchronous page 4581 fault handling. 4582 4583 no-kvmclock [X86,KVM,EARLY] Disable paravirtualized KVM clock driver 4584 4585 nolapic [X86-32,APIC,EARLY] Do not enable or use the local APIC. 4586 4587 nolapic_timer [X86-32,APIC,EARLY] Do not use the local APIC timer. 4588 4589 nomce [X86-32] Disable Machine Check Exception 4590 4591 nomfgpt [X86-32] Disable Multi-Function General Purpose 4592 Timer usage (for AMD Geode machines). 4593 4594 nomodeset Disable kernel modesetting. Most systems' firmware 4595 sets up a display mode and provides framebuffer memory 4596 for output. With nomodeset, DRM and fbdev drivers will 4597 not load if they could possibly displace the pre- 4598 initialized output. Only the system framebuffer will 4599 be available for use. The respective drivers will not 4600 perform display-mode changes or accelerated rendering. 4601 4602 Useful as error fallback, or for testing and debugging. 4603 4604 nomodule Disable module load 4605 4606 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to 4607 shutdown the other cpus. Instead use the REBOOT_VECTOR 4608 irq. 4609 4610 nopat [X86,EARLY] Disable PAT (page attribute table extension of 4611 pagetables) support. 4612 4613 nopcid [X86-64,EARLY] Disable the PCID cpu feature. 4614 4615 nopku [X86] Disable Memory Protection Keys CPU feature found 4616 in some Intel CPUs. 4617 4618 nopti [X86-64,EARLY] 4619 Equivalent to pti=off 4620 4621 nopv= [X86,XEN,KVM,HYPER_V,VMWARE,EARLY] 4622 Disables the PV optimizations forcing the guest to run 4623 as generic guest with no PV drivers. Currently support 4624 XEN HVM, KVM, HYPER_V and VMWARE guest. 4625 4626 nopvspin [X86,XEN,KVM,EARLY] 4627 Disables the qspinlock slow path using PV optimizations 4628 which allow the hypervisor to 'idle' the guest on lock 4629 contention. 4630 4631 norandmaps Don't use address space randomization. Equivalent to 4632 echo 0 > /proc/sys/kernel/randomize_va_space 4633 4634 noreplace-smp [X86-32,SMP] Don't replace SMP instructions 4635 with UP alternatives 4636 4637 noresume [SWSUSP] Disables resume and restores original swap 4638 space. 4639 4640 no-scroll [VGA] Disables scrollback. 4641 This is required for the Braillex ib80-piezo Braille 4642 reader made by F.H. Papenmeier (Germany). 4643 4644 nosgx [X86-64,SGX,EARLY] Disables Intel SGX kernel support. 4645 4646 nosmap [PPC,EARLY] 4647 Disable SMAP (Supervisor Mode Access Prevention) 4648 even if it is supported by processor. 4649 4650 nosmep [PPC64s,EARLY] 4651 Disable SMEP (Supervisor Mode Execution Prevention) 4652 even if it is supported by processor. 4653 4654 nosmp [SMP,EARLY] Tells an SMP kernel to act as a UP kernel, 4655 and disable the IO APIC. legacy for "maxcpus=0". 4656 4657 nosmt [KNL,MIPS,PPC,EARLY] Disable symmetric multithreading (SMT). 4658 Equivalent to smt=1. 4659 4660 [KNL,LOONGARCH,X86,PPC,S390] Disable symmetric multithreading (SMT). 4661 nosmt=force: Force disable SMT, cannot be undone 4662 via the sysfs control file. 4663 4664 nosoftlockup [KNL] Disable the soft-lockup detector. 4665 4666 nospec_store_bypass_disable 4667 [HW,EARLY] Disable all mitigations for the Speculative 4668 Store Bypass vulnerability 4669 4670 nospectre_bhb [ARM64,EARLY] Disable all mitigations for Spectre-BHB (branch 4671 history injection) vulnerability. System may allow data leaks 4672 with this option. 4673 4674 nospectre_v1 [X86,PPC,EARLY] Disable mitigations for Spectre Variant 1 4675 (bounds check bypass). With this option data leaks are 4676 possible in the system. 4677 4678 nospectre_v2 [X86,PPC_E500,ARM64,EARLY] Disable all mitigations 4679 for the Spectre variant 2 (indirect branch 4680 prediction) vulnerability. System may allow data 4681 leaks with this option. 4682 4683 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES,RISCV,LOONGARCH,EARLY] 4684 Disable paravirtualized steal time accounting. steal time 4685 is computed, but won't influence scheduler behaviour 4686 4687 nosync [HW,M68K] Disables sync negotiation for all devices. 4688 4689 no_timer_check [X86,APIC] Disables the code which tests for broken 4690 timer IRQ sources, i.e., the IO-APIC timer. This can 4691 work around problems with incorrect timer 4692 initialization on some boards. 4693 4694 no_uaccess_flush 4695 [PPC,EARLY] Don't flush the L1-D cache after accessing user data. 4696 4697 novmcoredd [KNL,KDUMP] 4698 Disable device dump. Device dump allows drivers to 4699 append dump data to vmcore so you can collect driver 4700 specified debug info. Drivers can append the data 4701 without any limit and this data is stored in memory, 4702 so this may cause significant memory stress. Disabling 4703 device dump can help save memory but the driver debug 4704 data will be no longer available. This parameter 4705 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP 4706 is set. 4707 4708 no-vmw-sched-clock 4709 [X86,PV_OPS,EARLY] Disable paravirtualized VMware 4710 scheduler clock and use the default one. 4711 4712 nowatchdog [KNL] Disable both lockup detectors, i.e. 4713 soft-lockup and NMI watchdog (hard-lockup). 4714 4715 nowb [ARM,EARLY] 4716 4717 nox2apic [X86-64,APIC,EARLY] Do not enable x2APIC mode. 4718 4719 NOTE: this parameter will be ignored on systems with the 4720 LEGACY_XAPIC_DISABLED bit set in the 4721 IA32_XAPIC_DISABLE_STATUS MSR. 4722 4723 noxsave [BUGS=X86] Disables x86 extended register state save 4724 and restore using xsave. The kernel will fallback to 4725 enabling legacy floating-point and sse state. 4726 4727 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended 4728 register states. The kernel will fall back to use 4729 xsave to save the states. By using this parameter, 4730 performance of saving the states is degraded because 4731 xsave doesn't support modified optimization while 4732 xsaveopt supports it on xsaveopt enabled systems. 4733 4734 noxsaves [X86] Disables xsaves and xrstors used in saving and 4735 restoring x86 extended register state in compacted 4736 form of xsave area. The kernel will fall back to use 4737 xsaveopt and xrstor to save and restore the states 4738 in standard form of xsave area. By using this 4739 parameter, xsave area per process might occupy more 4740 memory on xsaves enabled systems. 4741 4742 nr_cpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 4743 could support. nr_cpus=n : n >= 1 limits the kernel to 4744 support 'n' processors. It could be larger than the 4745 number of already plugged CPU during bootup, later in 4746 runtime you can physically add extra cpu until it reaches 4747 n. So during boot up some boot time memory for per-cpu 4748 variables need be pre-allocated for later physical cpu 4749 hot plugging. 4750 4751 nr_uarts= [SERIAL] maximum number of UARTs to be registered. 4752 4753 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86, EARLY] 4754 Disable NUMA, Only set up a single NUMA node 4755 spanning all memory. 4756 4757 numa=fake=<size>[MG] 4758 [KNL, ARM64, RISCV, X86, EARLY] 4759 If given as a memory unit, fills all system RAM with 4760 nodes of size interleaved over physical nodes. 4761 4762 numa=fake=<N> 4763 [KNL, ARM64, RISCV, X86, EARLY] 4764 If given as an integer, fills all system RAM with N 4765 fake nodes interleaved over physical nodes. 4766 4767 numa=fake=<N>U 4768 [KNL, ARM64, RISCV, X86, EARLY] 4769 If given as an integer followed by 'U', it will 4770 divide each physical node into N emulated nodes. 4771 4772 numa=noacpi [X86] Don't parse the SRAT table for NUMA setup 4773 4774 numa=nohmat [X86] Don't parse the HMAT table for NUMA setup, or 4775 soft-reserved memory partitioning. 4776 4777 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic 4778 NUMA balancing. 4779 Allowed values are enable and disable 4780 4781 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA. 4782 'node', 'default' can be specified 4783 This can be set from sysctl after boot. 4784 See Documentation/admin-guide/sysctl/vm.rst for details. 4785 4786 ohci1394_dma=early [HW,EARLY] enable debugging via the ohci1394 driver. 4787 See Documentation/core-api/debugging-via-ohci1394.rst for more 4788 info. 4789 4790 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands 4791 Rather than timing out after 20 ms if an EC 4792 command is not properly ACKed, override the length 4793 of the timeout. We have interrupts disabled while 4794 waiting for the ACK, so if this is set too high 4795 interrupts *may* be lost! 4796 4797 omap_mux= [OMAP] Override bootloader pin multiplexing. 4798 Format: <mux_mode0.mode_name=value>... 4799 For example, to override I2C bus2: 4800 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100 4801 4802 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration 4803 4804 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock] 4805 4806 boundary - index of last SLC block on Flex-OneNAND. 4807 The remaining blocks are configured as MLC blocks. 4808 lock - Configure if Flex-OneNAND boundary should be locked. 4809 Once locked, the boundary cannot be changed. 4810 1 indicates lock status, 0 indicates unlock status. 4811 4812 oops=panic [KNL,EARLY] 4813 Always panic on oopses. Default is to just kill the 4814 process, but there is a small probability of 4815 deadlocking the machine. 4816 This will also cause panics on machine check exceptions. 4817 Useful together with panic=30 to trigger a reboot. 4818 4819 page_alloc.shuffle= 4820 [KNL] Boolean flag to control whether the page allocator 4821 should randomize its free lists. This parameter can be 4822 used to enable/disable page randomization. The state of 4823 the flag can be read from sysfs at: 4824 /sys/module/page_alloc/parameters/shuffle. 4825 This parameter is only available if CONFIG_SHUFFLE_PAGE_ALLOCATOR=y. 4826 4827 page_owner= [KNL,EARLY] Boot-time page_owner enabling option. 4828 Storage of the information about who allocated 4829 each page is disabled in default. With this switch, 4830 we can turn it on. 4831 on: enable the feature 4832 4833 page_poison= [KNL,EARLY] Boot-time parameter changing the state of 4834 poisoning on the buddy allocator, available with 4835 CONFIG_PAGE_POISONING=y. 4836 off: turn off poisoning (default) 4837 on: turn on poisoning 4838 4839 page_reporting.page_reporting_order= 4840 [KNL] Minimal page reporting order 4841 Format: <integer> 4842 Adjust the minimal page reporting order. The page 4843 reporting is disabled when it exceeds MAX_PAGE_ORDER. 4844 4845 panic= [KNL] Kernel behaviour on panic: delay <timeout> 4846 timeout > 0: seconds before rebooting 4847 timeout = 0: wait forever 4848 timeout < 0: reboot immediately 4849 Format: <timeout> 4850 4851 panic_on_taint= [KNL,EARLY] 4852 Bitmask for conditionally calling panic() in add_taint() 4853 Format: <hex>[,nousertaint] 4854 Hexadecimal bitmask representing the set of TAINT flags 4855 that will cause the kernel to panic when add_taint() is 4856 called with any of the flags in this set. 4857 The optional switch "nousertaint" can be utilized to 4858 prevent userspace forced crashes by writing to sysctl 4859 /proc/sys/kernel/tainted any flagset matching with the 4860 bitmask set on panic_on_taint. 4861 See Documentation/admin-guide/tainted-kernels.rst for 4862 extra details on the taint flags that users can pick 4863 to compose the bitmask to assign to panic_on_taint. 4864 4865 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump 4866 on a WARN(). 4867 4868 panic_force_cpu= 4869 [KNL,SMP] Force panic handling to execute on a specific CPU. 4870 Format: <cpu number> 4871 Some platforms require panic handling to occur on a 4872 specific CPU for the crash kernel to function correctly. 4873 This can be due to firmware limitations, interrupt routing 4874 constraints, or platform-specific requirements where only 4875 a particular CPU can safely enter the crash kernel. 4876 When set, panic() will redirect execution to the specified 4877 CPU before proceeding with the normal panic and kexec flow. 4878 If the target CPU is offline or unavailable, panic proceeds 4879 on the current CPU. 4880 This option should only be used for systems with the above 4881 constraints as it might cause the panic operation to be less reliable. 4882 4883 panic_print= Bitmask for printing system info when panic happens. 4884 User can chose combination of the following bits: 4885 bit 0: print all tasks info 4886 bit 1: print system memory info 4887 bit 2: print timer info 4888 bit 3: print locks info if CONFIG_LOCKDEP is on 4889 bit 4: print ftrace buffer 4890 bit 5: replay all kernel messages on consoles at the end of panic 4891 bit 6: print all CPUs backtrace (if available in the arch) 4892 bit 7: print only tasks in uninterruptible (blocked) state 4893 *Be aware* that this option may print a _lot_ of lines, 4894 so there are risks of losing older messages in the log. 4895 Use this option carefully, maybe worth to setup a 4896 bigger log buffer with "log_buf_len" along with this. 4897 4898 panic_sys_info= A comma separated list of extra information to be dumped 4899 on panic. 4900 Format: val[,val...] 4901 Where @val can be any of the following: 4902 4903 tasks: print all tasks info 4904 mem: print system memory info 4905 timers: print timers info 4906 locks: print locks info if CONFIG_LOCKDEP is on 4907 ftrace: print ftrace buffer 4908 all_bt: print all CPUs backtrace (if available in the arch) 4909 blocked_tasks: print only tasks in uninterruptible (blocked) state 4910 4911 This is a human readable alternative to the 'panic_print' option. 4912 4913 panic_console_replay 4914 When panic happens, replay all kernel messages on 4915 consoles at the end of panic. 4916 4917 parkbd.port= [HW] Parallel port number the keyboard adapter is 4918 connected to, default is 0. 4919 Format: <parport#> 4920 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation, 4921 0 for XT, 1 for AT (default is AT). 4922 Format: <mode> 4923 4924 parport= [HW,PPT] Specify parallel ports. 0 disables. 4925 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] } 4926 Use 'auto' to force the driver to use any 4927 IRQ/DMA settings detected (the default is to 4928 ignore detected IRQ/DMA settings because of 4929 possible conflicts). You can specify the base 4930 address, IRQ, and DMA settings; IRQ and DMA 4931 should be numbers, or 'auto' (for using detected 4932 settings on that particular port), or 'nofifo' 4933 (to avoid using a FIFO even if it is detected). 4934 Parallel ports are assigned in the order they 4935 are specified on the command line, starting 4936 with parport0. 4937 4938 parport_init_mode= [HW,PPT] 4939 Configure VIA parallel port to operate in 4940 a specific mode. This is necessary on Pegasos 4941 computer where firmware has no options for setting 4942 up parallel port mode and sets it to spp. 4943 Currently this function knows 686a and 8231 chips. 4944 Format: [spp|ps2|epp|ecp|ecpepp] 4945 4946 pata_legacy.all= [HW,LIBATA] 4947 Format: <int> 4948 Set to non-zero to probe primary and secondary ISA 4949 port ranges on PCI systems where no PCI PATA device 4950 has been found at either range. Disabled by default. 4951 4952 pata_legacy.autospeed= [HW,LIBATA] 4953 Format: <int> 4954 Set to non-zero if a chip is present that snoops speed 4955 changes. Disabled by default. 4956 4957 pata_legacy.ht6560a= [HW,LIBATA] 4958 Format: <int> 4959 Set to 1, 2, or 3 for HT 6560A on the primary channel, 4960 the secondary channel, or both channels respectively. 4961 Disabled by default. 4962 4963 pata_legacy.ht6560b= [HW,LIBATA] 4964 Format: <int> 4965 Set to 1, 2, or 3 for HT 6560B on the primary channel, 4966 the secondary channel, or both channels respectively. 4967 Disabled by default. 4968 4969 pata_legacy.iordy_mask= [HW,LIBATA] 4970 Format: <int> 4971 IORDY enable mask. Set individual bits to allow IORDY 4972 for the respective channel. Bit 0 is for the first 4973 legacy channel handled by this driver, bit 1 is for 4974 the second channel, and so on. The sequence will often 4975 correspond to the primary legacy channel, the secondary 4976 legacy channel, and so on, but the handling of a PCI 4977 bus and the use of other driver options may interfere 4978 with the sequence. By default IORDY is allowed across 4979 all channels. 4980 4981 pata_legacy.opti82c46x= [HW,LIBATA] 4982 Format: <int> 4983 Set to 1, 2, or 3 for Opti 82c611A on the primary 4984 channel, the secondary channel, or both channels 4985 respectively. Disabled by default. 4986 4987 pata_legacy.opti82c611a= [HW,LIBATA] 4988 Format: <int> 4989 Set to 1, 2, or 3 for Opti 82c465MV on the primary 4990 channel, the secondary channel, or both channels 4991 respectively. Disabled by default. 4992 4993 pata_legacy.pio_mask= [HW,LIBATA] 4994 Format: <int> 4995 PIO mode mask for autospeed devices. Set individual 4996 bits to allow the use of the respective PIO modes. 4997 Bit 0 is for mode 0, bit 1 is for mode 1, and so on. 4998 All modes allowed by default. 4999 5000 pata_legacy.probe_all= [HW,LIBATA] 5001 Format: <int> 5002 Set to non-zero to probe tertiary and further ISA 5003 port ranges on PCI systems. Disabled by default. 5004 5005 pata_legacy.probe_mask= [HW,LIBATA] 5006 Format: <int> 5007 Probe mask for legacy ISA PATA ports. Depending on 5008 platform configuration and the use of other driver 5009 options up to 6 legacy ports are supported: 0x1f0, 5010 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing 5011 of individual ports can be disabled by setting the 5012 corresponding bits in the mask to 1. Bit 0 is for 5013 the first port in the list above (0x1f0), and so on. 5014 By default all supported ports are probed. 5015 5016 pata_legacy.qdi= [HW,LIBATA] 5017 Format: <int> 5018 Set to non-zero to probe QDI controllers. By default 5019 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise. 5020 5021 pata_legacy.winbond= [HW,LIBATA] 5022 Format: <int> 5023 Set to non-zero to probe Winbond controllers. Use 5024 the standard I/O port (0x130) if 1, otherwise the 5025 value given is the I/O port to use (typically 0x1b0). 5026 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE, 5027 0 otherwise. 5028 5029 pata_platform.pio_mask= [HW,LIBATA] 5030 Format: <int> 5031 Supported PIO mode mask. Set individual bits to allow 5032 the use of the respective PIO modes. Bit 0 is for 5033 mode 0, bit 1 is for mode 1, and so on. Mode 0 only 5034 allowed by default. 5035 5036 pause_on_oops=<int> 5037 Halt all CPUs after the first oops has been printed for 5038 the specified number of seconds. This is to be used if 5039 your oopses keep scrolling off the screen. 5040 5041 pcbit= [HW,ISDN] 5042 5043 pci=option[,option...] [PCI,EARLY] various PCI subsystem options. 5044 5045 Some options herein operate on a specific device 5046 or a set of devices (<pci_dev>). These are 5047 specified in one of the following formats: 5048 5049 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]* 5050 pci:<vendor>:<device>[:<subvendor>:<subdevice>] 5051 5052 Note: the first format specifies a PCI 5053 bus/device/function address which may change 5054 if new hardware is inserted, if motherboard 5055 firmware changes, or due to changes caused 5056 by other kernel parameters. If the 5057 domain is left unspecified, it is 5058 taken to be zero. Optionally, a path 5059 to a device through multiple device/function 5060 addresses can be specified after the base 5061 address (this is more robust against 5062 renumbering issues). The second format 5063 selects devices using IDs from the 5064 configuration space which may match multiple 5065 devices in the system. 5066 5067 earlydump dump PCI config space before the kernel 5068 changes anything 5069 off [X86] don't probe for the PCI bus 5070 bios [X86-32] force use of PCI BIOS, don't access 5071 the hardware directly. Use this if your machine 5072 has a non-standard PCI host bridge. 5073 nobios [X86-32] disallow use of PCI BIOS, only direct 5074 hardware access methods are allowed. Use this 5075 if you experience crashes upon bootup and you 5076 suspect they are caused by the BIOS. 5077 conf1 [X86] Force use of PCI Configuration Access 5078 Mechanism 1 (config address in IO port 0xCF8, 5079 data in IO port 0xCFC, both 32-bit). 5080 conf2 [X86] Force use of PCI Configuration Access 5081 Mechanism 2 (IO port 0xCF8 is an 8-bit port for 5082 the function, IO port 0xCFA, also 8-bit, sets 5083 bus number. The config space is then accessed 5084 through ports 0xC000-0xCFFF). 5085 See http://wiki.osdev.org/PCI for more info 5086 on the configuration access mechanisms. 5087 noaer [PCIE] If the PCIEAER kernel config parameter is 5088 enabled, this kernel boot option can be used to 5089 disable the use of PCIE advanced error reporting. 5090 nodomains [PCI] Disable support for multiple PCI 5091 root domains (aka PCI segments, in ACPI-speak). 5092 nommconf [X86] Disable use of MMCONFIG for PCI 5093 Configuration 5094 check_enable_amd_mmconf [X86] check for and enable 5095 properly configured MMIO access to PCI 5096 config space on AMD family 10h CPU 5097 nomsi [MSI] If the PCI_MSI kernel config parameter is 5098 enabled, this kernel boot option can be used to 5099 disable the use of MSI interrupts system-wide. 5100 noioapicquirk [APIC] Disable all boot interrupt quirks. 5101 Safety option to keep boot IRQs enabled. This 5102 should never be necessary. 5103 ioapicreroute [APIC] Enable rerouting of boot IRQs to the 5104 primary IO-APIC for bridges that cannot disable 5105 boot IRQs. This fixes a source of spurious IRQs 5106 when the system masks IRQs. 5107 noioapicreroute [APIC] Disable workaround that uses the 5108 boot IRQ equivalent of an IRQ that connects to 5109 a chipset where boot IRQs cannot be disabled. 5110 The opposite of ioapicreroute. 5111 biosirq [X86-32] Use PCI BIOS calls to get the interrupt 5112 routing table. These calls are known to be buggy 5113 on several machines and they hang the machine 5114 when used, but on other computers it's the only 5115 way to get the interrupt routing table. Try 5116 this option if the kernel is unable to allocate 5117 IRQs or discover secondary PCI buses on your 5118 motherboard. 5119 rom [X86] Assign address space to expansion ROMs. 5120 Use with caution as certain devices share 5121 address decoders between ROMs and other 5122 resources. 5123 norom [X86] Do not assign address space to 5124 expansion ROMs that do not already have 5125 BIOS assigned address ranges. 5126 nobar [X86] Do not assign address space to the 5127 BARs that weren't assigned by the BIOS. 5128 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be 5129 assigned automatically to PCI devices. You can 5130 make the kernel exclude IRQs of your ISA cards 5131 this way. 5132 pirqaddr=0xAAAAA [X86] Specify the physical address 5133 of the PIRQ table (normally generated 5134 by the BIOS) if it is outside the 5135 F0000h-100000h range. 5136 lastbus=N [X86] Scan all buses thru bus #N. Can be 5137 useful if the kernel is unable to find your 5138 secondary buses and you want to tell it 5139 explicitly which ones they are. 5140 assign-busses [X86] Always assign all PCI bus 5141 numbers ourselves, overriding 5142 whatever the firmware may have done. 5143 usepirqmask [X86] Honor the possible IRQ mask stored 5144 in the BIOS $PIR table. This is needed on 5145 some systems with broken BIOSes, notably 5146 some HP Pavilion N5400 and Omnibook XE3 5147 notebooks. This will have no effect if ACPI 5148 IRQ routing is enabled. 5149 noacpi [X86] Do not use ACPI for IRQ routing 5150 or for PCI scanning. 5151 use_crs [X86] Use PCI host bridge window information 5152 from ACPI. On BIOSes from 2008 or later, this 5153 is enabled by default. If you need to use this, 5154 please report a bug. 5155 nocrs [X86] Ignore PCI host bridge windows from ACPI. 5156 If you need to use this, please report a bug. 5157 use_e820 [X86] Use E820 reservations to exclude parts of 5158 PCI host bridge windows. This is a workaround 5159 for BIOS defects in host bridge _CRS methods. 5160 If you need to use this, please report a bug to 5161 <linux-pci@vger.kernel.org>. 5162 no_e820 [X86] Ignore E820 reservations for PCI host 5163 bridge windows. This is the default on modern 5164 hardware. If you need to use this, please report 5165 a bug to <linux-pci@vger.kernel.org>. 5166 routeirq Do IRQ routing for all PCI devices. 5167 This is normally done in pci_enable_device(), 5168 so this option is a temporary workaround 5169 for broken drivers that don't call it. 5170 skip_isa_align [X86] do not align io start addr, so can 5171 handle more pci cards 5172 noearly [X86] Don't do any early type 1 scanning. 5173 This might help on some broken boards which 5174 machine check when some devices' config space 5175 is read. But various workarounds are disabled 5176 and some IOMMU drivers will not work. 5177 bfsort Sort PCI devices into breadth-first order. 5178 This sorting is done to get a device 5179 order compatible with older (<= 2.4) kernels. 5180 nobfsort Don't sort PCI devices into breadth-first order. 5181 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size) 5182 tuning and use the BIOS-configured MPS defaults. 5183 pcie_bus_safe Set every device's MPS to the largest value 5184 supported by all devices below the root complex. 5185 pcie_bus_perf Set device MPS to the largest allowable MPS 5186 based on its parent bus. Also set MRRS (Max 5187 Read Request Size) to the largest supported 5188 value (no larger than the MPS that the device 5189 or bus can support) for best performance. 5190 pcie_bus_peer2peer Set every device's MPS to 128B, which 5191 every device is guaranteed to support. This 5192 configuration allows peer-to-peer DMA between 5193 any pair of devices, possibly at the cost of 5194 reduced performance. This also guarantees 5195 that hot-added devices will work. 5196 cbiosize=nn[KMG] The fixed amount of bus space which is 5197 reserved for the CardBus bridge's IO window. 5198 The default value is 256 bytes. 5199 cbmemsize=nn[KMG] The fixed amount of bus space which is 5200 reserved for the CardBus bridge's memory 5201 window. The default value is 64 megabytes. 5202 resource_alignment= 5203 Format: 5204 [<order of align>@]<pci_dev>[; ...] 5205 Specifies alignment and device to reassign 5206 aligned memory resources. How to 5207 specify the device is described above. 5208 If <order of align> is not specified, 5209 PAGE_SIZE is used as alignment. 5210 A PCI-PCI bridge can be specified if resource 5211 windows need to be expanded. 5212 To specify the alignment for several 5213 instances of a device, the PCI vendor, 5214 device, subvendor, and subdevice may be 5215 specified, e.g., 12@pci:8086:9c22:103c:198f 5216 for 4096-byte alignment. 5217 ecrc= Enable/disable PCIe ECRC (transaction layer 5218 end-to-end CRC checking). Only effective if 5219 OS has native AER control (either granted by 5220 ACPI _OSC or forced via "pcie_ports=native") 5221 bios: Use BIOS/firmware settings. This is the 5222 the default. 5223 off: Turn ECRC off 5224 on: Turn ECRC on. 5225 hpiosize=nn[KMG] The fixed amount of bus space which is 5226 reserved for hotplug bridge's IO window. 5227 Default size is 256 bytes. 5228 hpmmiosize=nn[KMG] The fixed amount of bus space which is 5229 reserved for hotplug bridge's MMIO window. 5230 Default size is 2 megabytes. 5231 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is 5232 reserved for hotplug bridge's MMIO_PREF window. 5233 Default size is 2 megabytes. 5234 hpmemsize=nn[KMG] The fixed amount of bus space which is 5235 reserved for hotplug bridge's MMIO and 5236 MMIO_PREF window. 5237 Default size is 2 megabytes. 5238 hpbussize=nn The minimum amount of additional bus numbers 5239 reserved for buses below a hotplug bridge. 5240 Default is 1. 5241 realloc= Enable/disable reallocating PCI bridge resources 5242 if allocations done by BIOS are too small to 5243 accommodate resources required by all child 5244 devices. 5245 off: Turn realloc off 5246 on: Turn realloc on 5247 realloc same as realloc=on 5248 noari do not use PCIe ARI. 5249 noats [PCIE, Intel-IOMMU, AMD-IOMMU] 5250 do not use PCIe ATS (and IOMMU device IOTLB). 5251 pcie_scan_all Scan all possible PCIe devices. Otherwise we 5252 only look for one device below a PCIe downstream 5253 port. 5254 big_root_window Try to add a big 64bit memory window to the PCIe 5255 root complex on AMD CPUs. Some GFX hardware 5256 can resize a BAR to allow access to all VRAM. 5257 Adding the window is slightly risky (it may 5258 conflict with unreported devices), so this 5259 taints the kernel. 5260 disable_acs_redir=<pci_dev>[; ...] 5261 Specify one or more PCI devices (in the format 5262 specified above) separated by semicolons. 5263 Each device specified will have the PCI ACS 5264 redirect capabilities forced off which will 5265 allow P2P traffic between devices through 5266 bridges without forcing it upstream. Note: 5267 this removes isolation between devices and 5268 may put more devices in an IOMMU group. 5269 config_acs= 5270 Format: 5271 <ACS flags>@<pci_dev>[; ...] 5272 Specify one or more PCI devices (in the format 5273 specified above) optionally prepended with flags 5274 and separated by semicolons. The respective 5275 capabilities will be enabled, disabled or 5276 unchanged based on what is specified in 5277 flags. 5278 5279 ACS Flags is defined as follows: 5280 bit-0 : ACS Source Validation 5281 bit-1 : ACS Translation Blocking 5282 bit-2 : ACS P2P Request Redirect 5283 bit-3 : ACS P2P Completion Redirect 5284 bit-4 : ACS Upstream Forwarding 5285 bit-5 : ACS P2P Egress Control 5286 bit-6 : ACS Direct Translated P2P 5287 Each bit can be marked as: 5288 '0' – force disabled 5289 '1' – force enabled 5290 'x' – unchanged 5291 For example, 5292 pci=config_acs=10x@pci:0:0 5293 would configure all devices that support 5294 ACS to enable P2P Request Redirect, disable 5295 Translation Blocking, and leave Source 5296 Validation unchanged from whatever power-up 5297 or firmware set it to. 5298 5299 Note: this may remove isolation between devices 5300 and may put more devices in an IOMMU group. 5301 force_floating [S390] Force usage of floating interrupts. 5302 nomio [S390] Do not use MIO instructions. 5303 norid [S390] ignore the RID field and force use of 5304 one PCI domain per PCI function 5305 notph [PCIE] If the PCIE_TPH kernel config parameter 5306 is enabled, this kernel boot option can be used 5307 to disable PCIe TLP Processing Hints support 5308 system-wide. 5309 5310 pcie_aspm= [PCIE] Forcibly enable or ignore PCIe Active State Power 5311 Management. 5312 off Don't touch ASPM configuration at all. Leave any 5313 configuration done by firmware unchanged. 5314 force Enable ASPM even on devices that claim not to support it. 5315 WARNING: Forcing ASPM on may cause system lockups. 5316 5317 pcie_ports= [PCIE] PCIe port services handling: 5318 native Use native PCIe services (PME, AER, DPC, PCIe hotplug) 5319 even if the platform doesn't give the OS permission to 5320 use them. This may cause conflicts if the platform 5321 also tries to use these services. 5322 dpc-native Use native PCIe service for DPC only. May 5323 cause conflicts if firmware uses AER or DPC. 5324 compat Disable native PCIe services (PME, AER, DPC, PCIe 5325 hotplug). 5326 5327 pcie_port_pm= [PCIE] PCIe port power management handling: 5328 off Disable power management of all PCIe ports 5329 force Forcibly enable power management of all PCIe ports 5330 5331 pcie_pme= [PCIE,PM] Native PCIe PME signaling options: 5332 nomsi Do not use MSI for native PCIe PME signaling (this makes 5333 all PCIe root ports use INTx for all services). 5334 5335 pcmv= [HW,PCMCIA] BadgePAD 4 5336 5337 pd_ignore_unused 5338 [PM] 5339 Keep all power-domains already enabled by bootloader on, 5340 even if no driver has claimed them. This is useful 5341 for debug and development, but should not be 5342 needed on a platform with proper driver support. 5343 5344 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at 5345 boot time. 5346 Format: { 0 | 1 } 5347 See arch/parisc/kernel/pdc_chassis.c 5348 5349 percpu_alloc= [MM,EARLY] 5350 Select which percpu first chunk allocator to use. 5351 Currently supported values are "embed" and "page". 5352 Archs may support subset or none of the selections. 5353 See comments in mm/percpu.c for details on each 5354 allocator. This parameter is primarily for debugging 5355 and performance comparison. 5356 5357 pirq= [SMP,APIC] Manual mp-table setup 5358 See Documentation/arch/x86/i386/IO-APIC.rst. 5359 5360 plip= [PPT,NET] Parallel port network link 5361 Format: { parport<nr> | timid | 0 } 5362 See also Documentation/admin-guide/parport.rst. 5363 5364 pmtmr= [X86] Manual setup of pmtmr I/O Port. 5365 Override pmtimer IOPort with a hex value. 5366 e.g. pmtmr=0x508 5367 5368 pmu_override= [PPC] Override the PMU. 5369 This option takes over the PMU facility, so it is no 5370 longer usable by perf. Setting this option starts the 5371 PMU counters by setting MMCR0 to 0 (the FC bit is 5372 cleared). If a number is given, then MMCR1 is set to 5373 that number, otherwise (e.g., 'pmu_override=on'), MMCR1 5374 remains 0. 5375 5376 pm_async= [PM] 5377 Format: off 5378 This parameter sets the initial value of the 5379 /sys/power/pm_async sysfs knob at boot time. 5380 If set to "off", disables asynchronous suspend and 5381 resume of devices during system-wide power transitions. 5382 This can be useful on platforms where device 5383 dependencies are not well-defined, or for debugging 5384 power management issues. Asynchronous operations are 5385 enabled by default. 5386 5387 5388 pm_debug_messages [SUSPEND,KNL] 5389 Enable suspend/resume debug messages during boot up. 5390 5391 pnp.debug=1 [PNP] 5392 Enable PNP debug messages (depends on the 5393 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time 5394 via /sys/module/pnp/parameters/debug. We always show 5395 current resource usage; turning this on also shows 5396 possible settings and some assignment information. 5397 5398 pnpacpi= [ACPI] 5399 { off } 5400 5401 pnpbios= [ISAPNP] 5402 { on | off | curr | res | no-curr | no-res } 5403 5404 pnp_reserve_irq= 5405 [ISAPNP] Exclude IRQs for the autoconfiguration 5406 5407 pnp_reserve_dma= 5408 [ISAPNP] Exclude DMAs for the autoconfiguration 5409 5410 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration 5411 Ranges are in pairs (I/O port base and size). 5412 5413 pnp_reserve_mem= 5414 [ISAPNP] Exclude memory regions for the 5415 autoconfiguration. 5416 Ranges are in pairs (memory base and size). 5417 5418 ports= [IP_VS_FTP] IPVS ftp helper module 5419 Default is 21. 5420 Up to 8 (IP_VS_APP_MAX_PORTS) ports 5421 may be specified. 5422 Format: <port>,<port>.... 5423 5424 possible_cpus= [SMP,S390,X86] 5425 Format: <unsigned int> 5426 Set the number of possible CPUs, overriding the 5427 regular discovery mechanisms (such as ACPI/FW, etc). 5428 5429 powersave=off [PPC] This option disables power saving features. 5430 It specifically disables cpuidle and sets the 5431 platform machine description specific power_save 5432 function to NULL. On Idle the CPU just reduces 5433 execution priority. 5434 5435 ppc_strict_facility_enable 5436 [PPC,ENABLE] This option catches any kernel floating point, 5437 Altivec, VSX and SPE outside of regions specifically 5438 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()). 5439 There is some performance impact when enabling this. 5440 5441 ppc_tm= [PPC,EARLY] 5442 Format: {"off"} 5443 Disable Hardware Transactional Memory 5444 5445 preempt= [KNL] 5446 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC 5447 none - Limited to cond_resched() calls 5448 voluntary - Limited to cond_resched() and might_sleep() calls 5449 full - Any section that isn't explicitly preempt disabled 5450 can be preempted anytime. Tasks will also yield 5451 contended spinlocks (if the critical section isn't 5452 explicitly preempt disabled beyond the lock itself). 5453 lazy - Scheduler controlled. Similar to full but instead 5454 of preempting the task immediately, the task gets 5455 one HZ tick time to yield itself before the 5456 preemption will be forced. One preemption is when the 5457 task returns to user space. 5458 5459 print-fatal-signals= 5460 [KNL] debug: print fatal signals 5461 5462 If enabled, warn about various signal handling 5463 related application anomalies: too many signals, 5464 too many POSIX.1 timers, fatal signals causing a 5465 coredump - etc. 5466 5467 If you hit the warning due to signal overflow, 5468 you might want to try "ulimit -i unlimited". 5469 5470 default: off. 5471 5472 printk.always_kmsg_dump= 5473 Trigger kmsg_dump for cases other than kernel oops or 5474 panics 5475 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5476 default: disabled 5477 5478 printk.console_no_auto_verbose= 5479 Disable console loglevel raise on oops, panic 5480 or lockdep-detected issues (only if lock debug is on). 5481 With an exception to setups with low baudrate on 5482 serial console, keeping this 0 is a good choice 5483 in order to provide more debug information. 5484 Format: <bool> 5485 default: 0 (auto_verbose is enabled) 5486 5487 printk.debug_non_panic_cpus= 5488 Allows storing messages from non-panic CPUs into 5489 the printk log buffer during panic(). They are 5490 flushed to consoles by the panic-CPU on 5491 a best-effort basis. 5492 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5493 Default: disabled 5494 5495 printk.devkmsg={on,off,ratelimit} 5496 Control writing to /dev/kmsg. 5497 on - unlimited logging to /dev/kmsg from userspace 5498 off - logging to /dev/kmsg disabled 5499 ratelimit - ratelimit the logging 5500 Default: ratelimit 5501 5502 printk.time= Show timing data prefixed to each printk message line 5503 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5504 5505 proc_mem.force_override= [KNL] 5506 Format: {always | ptrace | never} 5507 Traditionally /proc/pid/mem allows memory permissions to be 5508 overridden without restrictions. This option may be set to 5509 restrict that. Can be one of: 5510 - 'always': traditional behavior always allows mem overrides. 5511 - 'ptrace': only allow mem overrides for active ptracers. 5512 - 'never': never allow mem overrides. 5513 If not specified, default is the CONFIG_PROC_MEM_* choice. 5514 5515 processor.max_cstate= [HW,ACPI] 5516 Limit processor to maximum C-state 5517 max_cstate=9 overrides any DMI blacklist limit. 5518 5519 processor.nocst [HW,ACPI] 5520 Ignore the _CST method to determine C-states, 5521 instead using the legacy FADT method 5522 5523 profile= [KNL] Enable kernel profiling via /proc/profile 5524 Format: [<profiletype>,]<number> 5525 Param: <profiletype>: "schedule" or "kvm" 5526 [defaults to kernel profiling] 5527 Param: "schedule" - profile schedule points. 5528 Param: "kvm" - profile VM exits. 5529 Param: <number> - step/bucket size as a power of 2 for 5530 statistical time based profiling. 5531 5532 prot_virt= [S390] enable hosting protected virtual machines 5533 isolated from the hypervisor (if hardware supports 5534 that). If enabled, the default kernel base address 5535 might be overridden even when Kernel Address Space 5536 Layout Randomization is disabled. 5537 Format: <bool> 5538 5539 psi= [KNL] Enable or disable pressure stall information 5540 tracking. 5541 Format: <bool> 5542 5543 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to 5544 probe for; one of (bare|imps|exps|lifebook|any). 5545 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports 5546 per second. 5547 psmouse.resetafter= [HW,MOUSE] 5548 Try to reset the device after so many bad packets 5549 (0 = never). 5550 psmouse.resolution= 5551 [HW,MOUSE] Set desired mouse resolution, in dpi. 5552 psmouse.smartscroll= 5553 [HW,MOUSE] Controls Logitech smartscroll autorepeat. 5554 0 = disabled, 1 = enabled (default). 5555 5556 pstore.backend= Specify the name of the pstore backend to use 5557 5558 pti= [X86-64] Control Page Table Isolation of user and 5559 kernel address spaces. Disabling this feature 5560 removes hardening, but improves performance of 5561 system calls and interrupts. 5562 5563 on - unconditionally enable 5564 off - unconditionally disable 5565 auto - kernel detects whether your CPU model is 5566 vulnerable to issues that PTI mitigates 5567 5568 Not specifying this option is equivalent to pti=auto. 5569 5570 pty.legacy_count= 5571 [KNL] Number of legacy pty's. Overwrites compiled-in 5572 default number. 5573 5574 quiet [KNL,EARLY] Disable most log messages 5575 5576 r128= [HW,DRM] 5577 5578 radix_hcall_invalidate=on [PPC/PSERIES] 5579 Disable RADIX GTSE feature and use hcall for TLB 5580 invalidate. 5581 5582 raid= [HW,RAID] 5583 See Documentation/admin-guide/md.rst. 5584 5585 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes 5586 See Documentation/admin-guide/blockdev/ramdisk.rst. 5587 5588 ramdisk_start= [Deprecated,RAM] RAM disk image start address 5589 5590 random.trust_cpu=off 5591 [KNL,EARLY] Disable trusting the use of the CPU's 5592 random number generator (if available) to 5593 initialize the kernel's RNG. 5594 5595 random.trust_bootloader=off 5596 [KNL,EARLY] Disable trusting the use of the a seed 5597 passed by the bootloader (if available) to 5598 initialize the kernel's RNG. 5599 5600 randomize_kstack_offset= 5601 [KNL,EARLY] Enable or disable kernel stack offset 5602 randomization, which provides roughly 5 bits of 5603 entropy, frustrating memory corruption attacks 5604 that depend on stack address determinism or 5605 cross-syscall address exposures. This is only 5606 available on architectures that have defined 5607 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET. 5608 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5609 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT. 5610 5611 ras=option[,option,...] [KNL] RAS-specific options 5612 5613 cec_disable [X86] 5614 Disable the Correctable Errors Collector, 5615 see CONFIG_RAS_CEC help text. 5616 5617 rcu_nocbs[=cpu-list] 5618 [KNL] The optional argument is a cpu list, 5619 as described above. 5620 5621 In kernels built with CONFIG_RCU_NOCB_CPU=y, 5622 enable the no-callback CPU mode, which prevents 5623 such CPUs' callbacks from being invoked in 5624 softirq context. Invocation of such CPUs' RCU 5625 callbacks will instead be offloaded to "rcuox/N" 5626 kthreads created for that purpose, where "x" is 5627 "p" for RCU-preempt, "s" for RCU-sched, and "g" 5628 for the kthreads that mediate grace periods; and 5629 "N" is the CPU number. This reduces OS jitter on 5630 the offloaded CPUs, which can be useful for HPC 5631 and real-time workloads. It can also improve 5632 energy efficiency for asymmetric multiprocessors. 5633 5634 If a cpulist is passed as an argument, the specified 5635 list of CPUs is set to no-callback mode from boot. 5636 5637 Otherwise, if the '=' sign and the cpulist 5638 arguments are omitted, no CPU will be set to 5639 no-callback mode from boot but the mode may be 5640 toggled at runtime via cpusets. 5641 5642 Note that this argument takes precedence over 5643 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 5644 5645 rcu_nocb_poll [KNL] 5646 Rather than requiring that offloaded CPUs 5647 (specified by rcu_nocbs= above) explicitly 5648 awaken the corresponding "rcuoN" kthreads, 5649 make these kthreads poll for callbacks. 5650 This improves the real-time response for the 5651 offloaded CPUs by relieving them of the need to 5652 wake up the corresponding kthread, but degrades 5653 energy efficiency by requiring that the kthreads 5654 periodically wake up to do the polling. 5655 5656 rcutree.blimit= [KNL] 5657 Set maximum number of finished RCU callbacks to 5658 process in one batch. 5659 5660 rcutree.csd_lock_suppress_rcu_stall= [KNL] 5661 Do only a one-line RCU CPU stall warning when 5662 there is an ongoing too-long CSD-lock wait. 5663 5664 rcutree.do_rcu_barrier= [KNL] 5665 Request a call to rcu_barrier(). This is 5666 throttled so that userspace tests can safely 5667 hammer on the sysfs variable if they so choose. 5668 If triggered before the RCU grace-period machinery 5669 is fully active, this will error out with EAGAIN. 5670 5671 rcutree.dump_tree= [KNL] 5672 Dump the structure of the rcu_node combining tree 5673 out at early boot. This is used for diagnostic 5674 purposes, to verify correct tree setup. 5675 5676 rcutree.gp_cleanup_delay= [KNL] 5677 Set the number of jiffies to delay each step of 5678 RCU grace-period cleanup. 5679 5680 rcutree.gp_init_delay= [KNL] 5681 Set the number of jiffies to delay each step of 5682 RCU grace-period initialization. 5683 5684 rcutree.gp_preinit_delay= [KNL] 5685 Set the number of jiffies to delay each step of 5686 RCU grace-period pre-initialization, that is, 5687 the propagation of recent CPU-hotplug changes up 5688 the rcu_node combining tree. 5689 5690 rcutree.jiffies_till_first_fqs= [KNL] 5691 Set delay from grace-period initialization to 5692 first attempt to force quiescent states. 5693 Units are jiffies, minimum value is zero, 5694 and maximum value is HZ. 5695 5696 rcutree.jiffies_till_next_fqs= [KNL] 5697 Set delay between subsequent attempts to force 5698 quiescent states. Units are jiffies, minimum 5699 value is one, and maximum value is HZ. 5700 5701 rcutree.jiffies_till_sched_qs= [KNL] 5702 Set required age in jiffies for a 5703 given grace period before RCU starts 5704 soliciting quiescent-state help from 5705 rcu_note_context_switch() and cond_resched(). 5706 If not specified, the kernel will calculate 5707 a value based on the most recent settings 5708 of rcutree.jiffies_till_first_fqs 5709 and rcutree.jiffies_till_next_fqs. 5710 This calculated value may be viewed in 5711 rcutree.jiffies_to_sched_qs. Any attempt to set 5712 rcutree.jiffies_to_sched_qs will be cheerfully 5713 overwritten. 5714 5715 rcutree.kthread_prio= [KNL,BOOT] 5716 Set the SCHED_FIFO priority of the RCU per-CPU 5717 kthreads (rcuc/N). This value is also used for 5718 the priority of the RCU boost threads (rcub/N) 5719 and for the RCU grace-period kthreads (rcu_bh, 5720 rcu_preempt, and rcu_sched). If RCU_BOOST is 5721 set, valid values are 1-99 and the default is 1 5722 (the least-favored priority). Otherwise, when 5723 RCU_BOOST is not set, valid values are 0-99 and 5724 the default is zero (non-realtime operation). 5725 When RCU_NOCB_CPU is set, also adjust the 5726 priority of NOCB callback kthreads. 5727 5728 rcutree.nocb_nobypass_lim_per_jiffy= [KNL] 5729 On callback-offloaded (rcu_nocbs) CPUs, 5730 RCU reduces the lock contention that would 5731 otherwise be caused by callback floods through 5732 use of the ->nocb_bypass list. However, in the 5733 common non-flooded case, RCU queues directly to 5734 the main ->cblist in order to avoid the extra 5735 overhead of the ->nocb_bypass list and its lock. 5736 But if there are too many callbacks queued during 5737 a single jiffy, RCU pre-queues the callbacks into 5738 the ->nocb_bypass queue. The definition of "too 5739 many" is supplied by this kernel boot parameter. 5740 5741 rcutree.nohz_full_patience_delay= [KNL] 5742 On callback-offloaded (rcu_nocbs) CPUs, avoid 5743 disturbing RCU unless the grace period has 5744 reached the specified age in milliseconds. 5745 Defaults to zero. Large values will be capped 5746 at five seconds. All values will be rounded down 5747 to the nearest value representable by jiffies. 5748 5749 rcutree.qhimark= [KNL] 5750 Set threshold of queued RCU callbacks beyond which 5751 batch limiting is disabled. 5752 5753 rcutree.qlowmark= [KNL] 5754 Set threshold of queued RCU callbacks below which 5755 batch limiting is re-enabled. 5756 5757 rcutree.qovld= [KNL] 5758 Set threshold of queued RCU callbacks beyond which 5759 RCU's force-quiescent-state scan will aggressively 5760 enlist help from cond_resched() and sched IPIs to 5761 help CPUs more quickly reach quiescent states. 5762 Set to less than zero to make this be set based 5763 on rcutree.qhimark at boot time and to zero to 5764 disable more aggressive help enlistment. 5765 5766 rcutree.rcu_delay_page_cache_fill_msec= [KNL] 5767 Set the page-cache refill delay (in milliseconds) 5768 in response to low-memory conditions. The range 5769 of permitted values is in the range 0:100000. 5770 5771 rcutree.rcu_divisor= [KNL] 5772 Set the shift-right count to use to compute 5773 the callback-invocation batch limit bl from 5774 the number of callbacks queued on this CPU. 5775 The result will be bounded below by the value of 5776 the rcutree.blimit kernel parameter. Every bl 5777 callbacks, the softirq handler will exit in 5778 order to allow the CPU to do other work. 5779 5780 Please note that this callback-invocation batch 5781 limit applies only to non-offloaded callback 5782 invocation. Offloaded callbacks are instead 5783 invoked in the context of an rcuoc kthread, which 5784 scheduler will preempt as it does any other task. 5785 5786 rcutree.rcu_fanout_exact= [KNL] 5787 Disable autobalancing of the rcu_node combining 5788 tree. This is used by rcutorture, and might 5789 possibly be useful for architectures having high 5790 cache-to-cache transfer latencies. 5791 5792 rcutree.rcu_fanout_leaf= [KNL] 5793 Change the number of CPUs assigned to each 5794 leaf rcu_node structure. Useful for very 5795 large systems, which will choose the value 64, 5796 and for NUMA systems with large remote-access 5797 latencies, which will choose a value aligned 5798 with the appropriate hardware boundaries. 5799 5800 rcutree.rcu_min_cached_objs= [KNL] 5801 Minimum number of objects which are cached and 5802 maintained per one CPU. Object size is equal 5803 to PAGE_SIZE. The cache allows to reduce the 5804 pressure to page allocator, also it makes the 5805 whole algorithm to behave better in low memory 5806 condition. 5807 5808 rcutree.rcu_nocb_gp_stride= [KNL] 5809 Set the number of NOCB callback kthreads in 5810 each group, which defaults to the square root 5811 of the number of CPUs. Larger numbers reduce 5812 the wakeup overhead on the global grace-period 5813 kthread, but increases that same overhead on 5814 each group's NOCB grace-period kthread. 5815 5816 rcutree.rcu_kick_kthreads= [KNL] 5817 Cause the grace-period kthread to get an extra 5818 wake_up() if it sleeps three times longer than 5819 it should at force-quiescent-state time. 5820 This wake_up() will be accompanied by a 5821 WARN_ONCE() splat and an ftrace_dump(). 5822 5823 rcutree.rcu_resched_ns= [KNL] 5824 Limit the time spend invoking a batch of RCU 5825 callbacks to the specified number of nanoseconds. 5826 By default, this limit is checked only once 5827 every 32 callbacks in order to limit the pain 5828 inflicted by local_clock() overhead. 5829 5830 rcutree.rcu_unlock_delay= [KNL] 5831 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels, 5832 this specifies an rcu_read_unlock()-time delay 5833 in microseconds. This defaults to zero. 5834 Larger delays increase the probability of 5835 catching RCU pointer leaks, that is, buggy use 5836 of RCU-protected pointers after the relevant 5837 rcu_read_unlock() has completed. 5838 5839 rcutree.sysrq_rcu= [KNL] 5840 Commandeer a sysrq key to dump out Tree RCU's 5841 rcu_node tree with an eye towards determining 5842 why a new grace period has not yet started. 5843 5844 rcutree.use_softirq= [KNL] 5845 If set to zero, move all RCU_SOFTIRQ processing to 5846 per-CPU rcuc kthreads. Defaults to a non-zero 5847 value, meaning that RCU_SOFTIRQ is used by default. 5848 Specify rcutree.use_softirq=0 to use rcuc kthreads. 5849 5850 But note that CONFIG_PREEMPT_RT=y kernels disable 5851 this kernel boot parameter, forcibly setting it 5852 to zero. 5853 5854 rcutree.enable_rcu_lazy= [KNL] 5855 To save power, batch RCU callbacks and flush after 5856 delay, memory pressure or callback list growing too 5857 big. 5858 5859 rcutree.rcu_normal_wake_from_gp= [KNL] 5860 Reduces a latency of synchronize_rcu() call. This approach 5861 maintains its own track of synchronize_rcu() callers, so it 5862 does not interact with regular callbacks because it does not 5863 use a call_rcu[_hurry]() path. Please note, this is for a 5864 normal grace period. 5865 5866 How to enable it: 5867 5868 echo 1 > /sys/module/rcutree/parameters/rcu_normal_wake_from_gp 5869 or pass a boot parameter "rcutree.rcu_normal_wake_from_gp=1" 5870 5871 Default is 1 if num_possible_cpus() <= 16 and it is not explicitly 5872 disabled by the boot parameter passing 0. 5873 5874 rcuscale.gp_async= [KNL] 5875 Measure performance of asynchronous 5876 grace-period primitives such as call_rcu(). 5877 5878 rcuscale.gp_async_max= [KNL] 5879 Specify the maximum number of outstanding 5880 callbacks per writer thread. When a writer 5881 thread exceeds this limit, it invokes the 5882 corresponding flavor of rcu_barrier() to allow 5883 previously posted callbacks to drain. 5884 5885 rcuscale.gp_exp= [KNL] 5886 Measure performance of expedited synchronous 5887 grace-period primitives. 5888 5889 rcuscale.holdoff= [KNL] 5890 Set test-start holdoff period. The purpose of 5891 this parameter is to delay the start of the 5892 test until boot completes in order to avoid 5893 interference. 5894 5895 rcuscale.kfree_by_call_rcu= [KNL] 5896 In kernels built with CONFIG_RCU_LAZY=y, test 5897 call_rcu() instead of kfree_rcu(). 5898 5899 rcuscale.kfree_mult= [KNL] 5900 Instead of allocating an object of size kfree_obj, 5901 allocate one of kfree_mult * sizeof(kfree_obj). 5902 Defaults to 1. 5903 5904 rcuscale.kfree_rcu_test= [KNL] 5905 Set to measure performance of kfree_rcu() flooding. 5906 5907 rcuscale.kfree_rcu_test_double= [KNL] 5908 Test the double-argument variant of kfree_rcu(). 5909 If this parameter has the same value as 5910 rcuscale.kfree_rcu_test_single, both the single- 5911 and double-argument variants are tested. 5912 5913 rcuscale.kfree_rcu_test_single= [KNL] 5914 Test the single-argument variant of kfree_rcu(). 5915 If this parameter has the same value as 5916 rcuscale.kfree_rcu_test_double, both the single- 5917 and double-argument variants are tested. 5918 5919 rcuscale.kfree_nthreads= [KNL] 5920 The number of threads running loops of kfree_rcu(). 5921 5922 rcuscale.kfree_alloc_num= [KNL] 5923 Number of allocations and frees done in an iteration. 5924 5925 rcuscale.kfree_loops= [KNL] 5926 Number of loops doing rcuscale.kfree_alloc_num number 5927 of allocations and frees. 5928 5929 rcuscale.minruntime= [KNL] 5930 Set the minimum test run time in seconds. This 5931 does not affect the data-collection interval, 5932 but instead allows better measurement of things 5933 like CPU consumption. 5934 5935 rcuscale.nreaders= [KNL] 5936 Set number of RCU readers. The value -1 selects 5937 N, where N is the number of CPUs. A value 5938 "n" less than -1 selects N-n+1, where N is again 5939 the number of CPUs. For example, -2 selects N 5940 (the number of CPUs), -3 selects N+1, and so on. 5941 A value of "n" less than or equal to -N selects 5942 a single reader. 5943 5944 rcuscale.nwriters= [KNL] 5945 Set number of RCU writers. The values operate 5946 the same as for rcuscale.nreaders. 5947 N, where N is the number of CPUs 5948 5949 rcuscale.scale_type= [KNL] 5950 Specify the RCU implementation to test. 5951 5952 rcuscale.shutdown= [KNL] 5953 Shut the system down after performance tests 5954 complete. This is useful for hands-off automated 5955 testing. 5956 5957 rcuscale.verbose= [KNL] 5958 Enable additional printk() statements. 5959 5960 rcuscale.writer_holdoff= [KNL] 5961 Write-side holdoff between grace periods, 5962 in microseconds. The default of zero says 5963 no holdoff. 5964 5965 rcuscale.writer_holdoff_jiffies= [KNL] 5966 Additional write-side holdoff between grace 5967 periods, but in jiffies. The default of zero 5968 says no holdoff. 5969 5970 rcutorture.fqs_duration= [KNL] 5971 Set duration of force_quiescent_state bursts 5972 in microseconds. 5973 5974 rcutorture.fqs_holdoff= [KNL] 5975 Set holdoff time within force_quiescent_state bursts 5976 in microseconds. 5977 5978 rcutorture.fqs_stutter= [KNL] 5979 Set wait time between force_quiescent_state bursts 5980 in seconds. 5981 5982 rcutorture.fwd_progress= [KNL] 5983 Specifies the number of kthreads to be used 5984 for RCU grace-period forward-progress testing 5985 for the types of RCU supporting this notion. 5986 Defaults to 1 kthread, values less than zero or 5987 greater than the number of CPUs cause the number 5988 of CPUs to be used. 5989 5990 rcutorture.fwd_progress_div= [KNL] 5991 Specify the fraction of a CPU-stall-warning 5992 period to do tight-loop forward-progress testing. 5993 5994 rcutorture.fwd_progress_holdoff= [KNL] 5995 Number of seconds to wait between successive 5996 forward-progress tests. 5997 5998 rcutorture.fwd_progress_need_resched= [KNL] 5999 Enclose cond_resched() calls within checks for 6000 need_resched() during tight-loop forward-progress 6001 testing. 6002 6003 rcutorture.gp_cond= [KNL] 6004 Use conditional/asynchronous update-side 6005 normal-grace-period primitives, if available. 6006 6007 rcutorture.gp_cond_exp= [KNL] 6008 Use conditional/asynchronous update-side 6009 expedited-grace-period primitives, if available. 6010 6011 rcutorture.gp_cond_full= [KNL] 6012 Use conditional/asynchronous update-side 6013 normal-grace-period primitives that also take 6014 concurrent expedited grace periods into account, 6015 if available. 6016 6017 rcutorture.gp_cond_exp_full= [KNL] 6018 Use conditional/asynchronous update-side 6019 expedited-grace-period primitives that also take 6020 concurrent normal grace periods into account, 6021 if available. 6022 6023 rcutorture.gp_cond_wi= [KNL] 6024 Nominal wait interval for normal conditional 6025 grace periods (specified by rcutorture's 6026 gp_cond and gp_cond_full module parameters), 6027 in microseconds. The actual wait interval will 6028 be randomly selected to nanosecond granularity up 6029 to this wait interval. Defaults to 16 jiffies, 6030 for example, 16,000 microseconds on a system 6031 with HZ=1000. 6032 6033 rcutorture.gp_cond_wi_exp= [KNL] 6034 Nominal wait interval for expedited conditional 6035 grace periods (specified by rcutorture's 6036 gp_cond_exp and gp_cond_exp_full module 6037 parameters), in microseconds. The actual wait 6038 interval will be randomly selected to nanosecond 6039 granularity up to this wait interval. Defaults to 6040 128 microseconds. 6041 6042 rcutorture.gp_exp= [KNL] 6043 Use expedited update-side primitives, if available. 6044 6045 rcutorture.gp_normal= [KNL] 6046 Use normal (non-expedited) asynchronous 6047 update-side primitives, if available. 6048 6049 rcutorture.gp_poll= [KNL] 6050 Use polled update-side normal-grace-period 6051 primitives, if available. 6052 6053 rcutorture.gp_poll_exp= [KNL] 6054 Use polled update-side expedited-grace-period 6055 primitives, if available. 6056 6057 rcutorture.gp_poll_full= [KNL] 6058 Use polled update-side normal-grace-period 6059 primitives that also take concurrent expedited 6060 grace periods into account, if available. 6061 6062 rcutorture.gp_poll_exp_full= [KNL] 6063 Use polled update-side expedited-grace-period 6064 primitives that also take concurrent normal 6065 grace periods into account, if available. 6066 6067 rcutorture.gp_poll_wi= [KNL] 6068 Nominal wait interval for normal conditional 6069 grace periods (specified by rcutorture's 6070 gp_poll and gp_poll_full module parameters), 6071 in microseconds. The actual wait interval will 6072 be randomly selected to nanosecond granularity up 6073 to this wait interval. Defaults to 16 jiffies, 6074 for example, 16,000 microseconds on a system 6075 with HZ=1000. 6076 6077 rcutorture.gp_poll_wi_exp= [KNL] 6078 Nominal wait interval for expedited conditional 6079 grace periods (specified by rcutorture's 6080 gp_poll_exp and gp_poll_exp_full module 6081 parameters), in microseconds. The actual wait 6082 interval will be randomly selected to nanosecond 6083 granularity up to this wait interval. Defaults to 6084 128 microseconds. 6085 6086 rcutorture.gp_sync= [KNL] 6087 Use normal (non-expedited) synchronous 6088 update-side primitives, if available. If all 6089 of rcutorture.gp_cond=, rcutorture.gp_exp=, 6090 rcutorture.gp_normal=, and rcutorture.gp_sync= 6091 are zero, rcutorture acts as if is interpreted 6092 they are all non-zero. 6093 6094 rcutorture.gpwrap_lag= [KNL] 6095 Enable grace-period wrap lag testing. Setting 6096 to false prevents the gpwrap lag test from 6097 running. Default is true. 6098 6099 rcutorture.gpwrap_lag_gps= [KNL] 6100 Set the value for grace-period wrap lag during 6101 active lag testing periods. This controls how many 6102 grace periods differences we tolerate between 6103 rdp and rnp's gp_seq before setting overflow flag. 6104 The default is always set to 8. 6105 6106 rcutorture.gpwrap_lag_cycle_mins= [KNL] 6107 Set the total cycle duration for gpwrap lag 6108 testing in minutes. This is the total time for 6109 one complete cycle of active and inactive 6110 testing periods. Default is 30 minutes. 6111 6112 rcutorture.gpwrap_lag_active_mins= [KNL] 6113 Set the duration for which gpwrap lag is active 6114 within each cycle, in minutes. During this time, 6115 the grace-period wrap lag will be set to the 6116 value specified by gpwrap_lag_gps. Default is 6117 5 minutes. 6118 6119 rcutorture.irqreader= [KNL] 6120 Run RCU readers from irq handlers, or, more 6121 accurately, from a timer handler. Not all RCU 6122 flavors take kindly to this sort of thing. 6123 6124 rcutorture.leakpointer= [KNL] 6125 Leak an RCU-protected pointer out of the reader. 6126 This can of course result in splats, and is 6127 intended to test the ability of things like 6128 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect 6129 such leaks. 6130 6131 rcutorture.n_barrier_cbs= [KNL] 6132 Set callbacks/threads for rcu_barrier() testing. 6133 6134 rcutorture.nfakewriters= [KNL] 6135 Set number of concurrent RCU writers. These just 6136 stress RCU, they don't participate in the actual 6137 test, hence the "fake". 6138 6139 rcutorture.nocbs_nthreads= [KNL] 6140 Set number of RCU callback-offload togglers. 6141 Zero (the default) disables toggling. 6142 6143 rcutorture.nocbs_toggle= [KNL] 6144 Set the delay in milliseconds between successive 6145 callback-offload toggling attempts. 6146 6147 rcutorture.nreaders= [KNL] 6148 Set number of RCU readers. The value -1 selects 6149 N-1, where N is the number of CPUs. A value 6150 "n" less than -1 selects N-n-2, where N is again 6151 the number of CPUs. For example, -2 selects N 6152 (the number of CPUs), -3 selects N+1, and so on. 6153 6154 rcutorture.object_debug= [KNL] 6155 Enable debug-object double-call_rcu() testing. 6156 6157 rcutorture.onoff_holdoff= [KNL] 6158 Set time (s) after boot for CPU-hotplug testing. 6159 6160 rcutorture.onoff_interval= [KNL] 6161 Set time (jiffies) between CPU-hotplug operations, 6162 or zero to disable CPU-hotplug testing. 6163 6164 rcutorture.preempt_duration= [KNL] 6165 Set duration (in milliseconds) of preemptions 6166 by a high-priority FIFO real-time task. Set to 6167 zero (the default) to disable. The CPUs to 6168 preempt are selected randomly from the set that 6169 are online at a given point in time. Races with 6170 CPUs going offline are ignored, with that attempt 6171 at preemption skipped. 6172 6173 rcutorture.preempt_interval= [KNL] 6174 Set interval (in milliseconds, defaulting to one 6175 second) between preemptions by a high-priority 6176 FIFO real-time task. This delay is mediated 6177 by an hrtimer and is further fuzzed to avoid 6178 inadvertent synchronizations. 6179 6180 rcutorture.read_exit_burst= [KNL] 6181 The number of times in a given read-then-exit 6182 episode that a set of read-then-exit kthreads 6183 is spawned. 6184 6185 rcutorture.read_exit_delay= [KNL] 6186 The delay, in seconds, between successive 6187 read-then-exit testing episodes. 6188 6189 rcutorture.reader_flavor= [KNL] 6190 A bit mask indicating which readers to use. 6191 If there is more than one bit set, the readers 6192 are entered from low-order bit up, and are 6193 exited in the opposite order. For SRCU, the 6194 0x1 bit is normal readers, 0x2 NMI-safe readers, 6195 and 0x4 light-weight readers. 6196 6197 rcutorture.shuffle_interval= [KNL] 6198 Set task-shuffle interval (s). Shuffling tasks 6199 allows some CPUs to go into dyntick-idle mode 6200 during the rcutorture test. 6201 6202 rcutorture.shutdown_secs= [KNL] 6203 Set time (s) after boot system shutdown. This 6204 is useful for hands-off automated testing. 6205 6206 rcutorture.stall_cpu= [KNL] 6207 Duration of CPU stall (s) to test RCU CPU stall 6208 warnings, zero to disable. 6209 6210 rcutorture.stall_cpu_block= [KNL] 6211 Sleep while stalling if set. This will result 6212 in warnings from preemptible RCU in addition to 6213 any other stall-related activity. Note that 6214 in kernels built with CONFIG_PREEMPTION=n and 6215 CONFIG_PREEMPT_COUNT=y, this parameter will 6216 cause the CPU to pass through a quiescent state. 6217 Given CONFIG_PREEMPTION=n, this will suppress 6218 RCU CPU stall warnings, but will instead result 6219 in scheduling-while-atomic splats. 6220 6221 Use of this module parameter results in splats. 6222 6223 6224 rcutorture.stall_cpu_holdoff= [KNL] 6225 Time to wait (s) after boot before inducing stall. 6226 6227 rcutorture.stall_cpu_irqsoff= [KNL] 6228 Disable interrupts while stalling if set, but only 6229 on the first stall in the set. 6230 6231 rcutorture.stall_cpu_repeat= [KNL] 6232 Number of times to repeat the stall sequence, 6233 so that rcutorture.stall_cpu_repeat=3 will result 6234 in four stall sequences. 6235 6236 rcutorture.stall_gp_kthread= [KNL] 6237 Duration (s) of forced sleep within RCU 6238 grace-period kthread to test RCU CPU stall 6239 warnings, zero to disable. If both stall_cpu 6240 and stall_gp_kthread are specified, the 6241 kthread is starved first, then the CPU. 6242 6243 rcutorture.stat_interval= [KNL] 6244 Time (s) between statistics printk()s. 6245 6246 rcutorture.stutter= [KNL] 6247 Time (s) to stutter testing, for example, specifying 6248 five seconds causes the test to run for five seconds, 6249 wait for five seconds, and so on. This tests RCU's 6250 ability to transition abruptly to and from idle. 6251 6252 rcutorture.test_boost= [KNL] 6253 Test RCU priority boosting? 0=no, 1=maybe, 2=yes. 6254 "Maybe" means test if the RCU implementation 6255 under test support RCU priority boosting. 6256 6257 rcutorture.test_boost_duration= [KNL] 6258 Duration (s) of each individual boost test. 6259 6260 rcutorture.test_boost_holdoff= [KNL] 6261 Holdoff time (s) from start of test to the start 6262 of RCU priority-boost testing. Defaults to zero, 6263 that is, no holdoff. 6264 6265 rcutorture.test_boost_interval= [KNL] 6266 Interval (s) between each boost test. 6267 6268 rcutorture.test_no_idle_hz= [KNL] 6269 Test RCU's dyntick-idle handling. See also the 6270 rcutorture.shuffle_interval parameter. 6271 6272 rcutorture.torture_type= [KNL] 6273 Specify the RCU implementation to test. 6274 6275 rcutorture.verbose= [KNL] 6276 Enable additional printk() statements. 6277 6278 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL] 6279 Dump ftrace buffer after reporting RCU CPU 6280 stall warning. 6281 6282 rcupdate.rcu_cpu_stall_notifiers= [KNL] 6283 Provide RCU CPU stall notifiers, but see the 6284 warnings in the RCU_CPU_STALL_NOTIFIER Kconfig 6285 option's help text. TL;DR: You almost certainly 6286 do not want rcupdate.rcu_cpu_stall_notifiers. 6287 6288 rcupdate.rcu_cpu_stall_suppress= [KNL] 6289 Suppress RCU CPU stall warning messages. 6290 6291 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL] 6292 Suppress RCU CPU stall warning messages and 6293 rcutorture writer stall warnings that occur 6294 during early boot, that is, during the time 6295 before the init task is spawned. 6296 6297 rcupdate.rcu_cpu_stall_timeout= [KNL] 6298 Set timeout for RCU CPU stall warning messages. 6299 The value is in seconds and the maximum allowed 6300 value is 300 seconds. 6301 6302 rcupdate.rcu_exp_cpu_stall_timeout= [KNL] 6303 Set timeout for expedited RCU CPU stall warning 6304 messages. The value is in milliseconds 6305 and the maximum allowed value is 21000 6306 milliseconds. Please note that this value is 6307 adjusted to an arch timer tick resolution. 6308 Setting this to zero causes the value from 6309 rcupdate.rcu_cpu_stall_timeout to be used (after 6310 conversion from seconds to milliseconds). 6311 6312 rcupdate.rcu_cpu_stall_cputime= [KNL] 6313 Provide statistics on the cputime and count of 6314 interrupts and tasks during the sampling period. For 6315 multiple continuous RCU stalls, all sampling periods 6316 begin at half of the first RCU stall timeout. 6317 6318 rcupdate.rcu_exp_stall_task_details= [KNL] 6319 Print stack dumps of any tasks blocking the 6320 current expedited RCU grace period during an 6321 expedited RCU CPU stall warning. 6322 6323 rcupdate.rcu_expedited= [KNL] 6324 Use expedited grace-period primitives, for 6325 example, synchronize_rcu_expedited() instead 6326 of synchronize_rcu(). This reduces latency, 6327 but can increase CPU utilization, degrade 6328 real-time latency, and degrade energy efficiency. 6329 No effect on CONFIG_TINY_RCU kernels. 6330 6331 rcupdate.rcu_normal= [KNL] 6332 Use only normal grace-period primitives, 6333 for example, synchronize_rcu() instead of 6334 synchronize_rcu_expedited(). This improves 6335 real-time latency, CPU utilization, and 6336 energy efficiency, but can expose users to 6337 increased grace-period latency. This parameter 6338 overrides rcupdate.rcu_expedited. No effect on 6339 CONFIG_TINY_RCU kernels. 6340 6341 rcupdate.rcu_normal_after_boot= [KNL] 6342 Once boot has completed (that is, after 6343 rcu_end_inkernel_boot() has been invoked), use 6344 only normal grace-period primitives. No effect 6345 on CONFIG_TINY_RCU kernels. 6346 6347 But note that CONFIG_PREEMPT_RT=y kernels enables 6348 this kernel boot parameter, forcibly setting 6349 it to the value one, that is, converting any 6350 post-boot attempt at an expedited RCU grace 6351 period to instead use normal non-expedited 6352 grace-period processing. 6353 6354 rcupdate.rcu_task_collapse_lim= [KNL] 6355 Set the maximum number of callbacks present 6356 at the beginning of a grace period that allows 6357 the RCU Tasks flavors to collapse back to using 6358 a single callback queue. This switching only 6359 occurs when rcupdate.rcu_task_enqueue_lim is 6360 set to the default value of -1. 6361 6362 rcupdate.rcu_task_contend_lim= [KNL] 6363 Set the minimum number of callback-queuing-time 6364 lock-contention events per jiffy required to 6365 cause the RCU Tasks flavors to switch to per-CPU 6366 callback queuing. This switching only occurs 6367 when rcupdate.rcu_task_enqueue_lim is set to 6368 the default value of -1. 6369 6370 rcupdate.rcu_task_enqueue_lim= [KNL] 6371 Set the number of callback queues to use for the 6372 RCU Tasks family of RCU flavors. The default 6373 of -1 allows this to be automatically (and 6374 dynamically) adjusted. This parameter is intended 6375 for use in testing. 6376 6377 rcupdate.rcu_task_lazy_lim= [KNL] 6378 Number of callbacks on a given CPU that will 6379 cancel laziness on that CPU. Use -1 to disable 6380 cancellation of laziness, but be advised that 6381 doing so increases the danger of OOM due to 6382 callback flooding. 6383 6384 rcupdate.rcu_task_stall_info= [KNL] 6385 Set initial timeout in jiffies for RCU task stall 6386 informational messages, which give some indication 6387 of the problem for those not patient enough to 6388 wait for ten minutes. Informational messages are 6389 only printed prior to the stall-warning message 6390 for a given grace period. Disable with a value 6391 less than or equal to zero. Defaults to ten 6392 seconds. A change in value does not take effect 6393 until the beginning of the next grace period. 6394 6395 rcupdate.rcu_task_stall_info_mult= [KNL] 6396 Multiplier for time interval between successive 6397 RCU task stall informational messages for a given 6398 RCU tasks grace period. This value is clamped 6399 to one through ten, inclusive. It defaults to 6400 the value three, so that the first informational 6401 message is printed 10 seconds into the grace 6402 period, the second at 40 seconds, the third at 6403 160 seconds, and then the stall warning at 600 6404 seconds would prevent a fourth at 640 seconds. 6405 6406 rcupdate.rcu_task_stall_timeout= [KNL] 6407 Set timeout in jiffies for RCU task stall 6408 warning messages. Disable with a value less 6409 than or equal to zero. Defaults to ten minutes. 6410 A change in value does not take effect until 6411 the beginning of the next grace period. 6412 6413 rcupdate.rcu_tasks_lazy_ms= [KNL] 6414 Set timeout in milliseconds RCU Tasks asynchronous 6415 callback batching for call_rcu_tasks(). 6416 A negative value will take the default. A value 6417 of zero will disable batching. Batching is 6418 always disabled for synchronize_rcu_tasks(). 6419 6420 rcupdate.rcu_self_test= [KNL] 6421 Run the RCU early boot self tests 6422 6423 rdinit= [KNL] 6424 Format: <full_path> 6425 Run specified binary instead of /init from the ramdisk, 6426 used for early userspace startup. See initrd. 6427 6428 rdrand= [X86,EARLY] 6429 force - Override the decision by the kernel to hide the 6430 advertisement of RDRAND support (this affects 6431 certain AMD processors because of buggy BIOS 6432 support, specifically around the suspend/resume 6433 path). 6434 6435 rdt= [HW,X86,RDT] 6436 Turn on/off individual RDT features. List is: 6437 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp, 6438 mba, smba, bmec, abmc, sdciae, energy[:guid], 6439 perf[:guid]. 6440 E.g. to turn on cmt and turn off mba use: 6441 rdt=cmt,!mba 6442 To turn off all energy telemetry monitoring and ensure that 6443 perf telemetry monitoring associated with guid 0x12345 6444 is enabled use: 6445 rdt=!energy,perf:0x12345 6446 6447 reboot= [KNL] 6448 Format (x86 or x86_64): 6449 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \ 6450 [[,]s[mp]#### \ 6451 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \ 6452 [[,]f[orce] 6453 Where reboot_mode is one of warm (soft) or cold (hard) or gpio 6454 (prefix with 'panic_' to set mode for panic 6455 reboot only), 6456 reboot_type is one of bios, acpi, kbd, triple, efi, or pci, 6457 reboot_force is either force or not specified, 6458 reboot_cpu is s[mp]#### with #### being the processor 6459 to be used for rebooting. 6460 6461 acpi 6462 Use the ACPI RESET_REG in the FADT. If ACPI is not 6463 configured or the ACPI reset does not work, the reboot 6464 path attempts the reset using the keyboard controller. 6465 6466 bios 6467 Use the CPU reboot vector for warm reset 6468 6469 cold 6470 Set the cold reboot flag 6471 6472 default 6473 There are some built-in platform specific "quirks" 6474 - you may see: "reboot: <name> series board detected. 6475 Selecting <type> for reboots." In the case where you 6476 think the quirk is in error (e.g. you have newer BIOS, 6477 or newer board) using this option will ignore the 6478 built-in quirk table, and use the generic default 6479 reboot actions. 6480 6481 efi 6482 Use efi reset_system runtime service. If EFI is not 6483 configured or the EFI reset does not work, the reboot 6484 path attempts the reset using the keyboard controller. 6485 6486 force 6487 Don't stop other CPUs on reboot. This can make reboot 6488 more reliable in some cases. 6489 6490 kbd 6491 Use the keyboard controller. cold reset (default) 6492 6493 pci 6494 Use a write to the PCI config space register 0xcf9 to 6495 trigger reboot. 6496 6497 triple 6498 Force a triple fault (init) 6499 6500 warm 6501 Don't set the cold reboot flag 6502 6503 Using warm reset will be much faster especially on big 6504 memory systems because the BIOS will not go through 6505 the memory check. Disadvantage is that not all 6506 hardware will be completely reinitialized on reboot so 6507 there may be boot problems on some systems. 6508 6509 6510 refscale.holdoff= [KNL] 6511 Set test-start holdoff period. The purpose of 6512 this parameter is to delay the start of the 6513 test until boot completes in order to avoid 6514 interference. 6515 6516 refscale.lookup_instances= [KNL] 6517 Number of data elements to use for the forms of 6518 SLAB_TYPESAFE_BY_RCU testing. A negative number 6519 is negated and multiplied by nr_cpu_ids, while 6520 zero specifies nr_cpu_ids. 6521 6522 refscale.loops= [KNL] 6523 Set the number of loops over the synchronization 6524 primitive under test. Increasing this number 6525 reduces noise due to loop start/end overhead, 6526 but the default has already reduced the per-pass 6527 noise to a handful of picoseconds on ca. 2020 6528 x86 laptops. 6529 6530 refscale.nreaders= [KNL] 6531 Set number of readers. The default value of -1 6532 selects N, where N is roughly 75% of the number 6533 of CPUs. A value of zero is an interesting choice. 6534 6535 refscale.nruns= [KNL] 6536 Set number of runs, each of which is dumped onto 6537 the console log. 6538 6539 refscale.readdelay= [KNL] 6540 Set the read-side critical-section duration, 6541 measured in microseconds. 6542 6543 refscale.scale_type= [KNL] 6544 Specify the read-protection implementation to test. 6545 6546 refscale.shutdown= [KNL] 6547 Shut down the system at the end of the performance 6548 test. This defaults to 1 (shut it down) when 6549 refscale is built into the kernel and to 0 (leave 6550 it running) when refscale is built as a module. 6551 6552 refscale.verbose= [KNL] 6553 Enable additional printk() statements. 6554 6555 refscale.verbose_batched= [KNL] 6556 Batch the additional printk() statements. If zero 6557 (the default) or negative, print everything. Otherwise, 6558 print every Nth verbose statement, where N is the value 6559 specified. 6560 6561 regulator_ignore_unused 6562 [REGULATOR] 6563 Prevents regulator framework from disabling regulators 6564 that are unused, due no driver claiming them. This may 6565 be useful for debug and development, but should not be 6566 needed on a platform with proper driver support. 6567 6568 relax_domain_level= 6569 [KNL, SMP] Set scheduler's default relax_domain_level. 6570 See Documentation/admin-guide/cgroup-v1/cpusets.rst. 6571 6572 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory 6573 Format: <base1>,<size1>[,<base2>,<size2>,...] 6574 Reserve I/O ports or memory so the kernel won't use 6575 them. If <base> is less than 0x10000, the region 6576 is assumed to be I/O ports; otherwise it is memory. 6577 6578 reserve_mem= [RAM] 6579 Format: nn[KMG]:<align>:<label> 6580 Reserve physical memory and label it with a name that 6581 other subsystems can use to access it. This is typically 6582 used for systems that do not wipe the RAM, and this command 6583 line will try to reserve the same physical memory on 6584 soft reboots. Note, it is not guaranteed to be the same 6585 location. For example, if anything about the system changes 6586 or if booting a different kernel. It can also fail if KASLR 6587 places the kernel at the location of where the RAM reservation 6588 was from a previous boot, the new reservation will be at a 6589 different location. 6590 Any subsystem using this feature must add a way to verify 6591 that the contents of the physical memory is from a previous 6592 boot, as there may be cases where the memory will not be 6593 located at the same location. 6594 6595 The format is size:align:label for example, to request 6596 12 megabytes of 4096 alignment for ramoops: 6597 6598 reserve_mem=12M:4096:oops ramoops.mem_name=oops 6599 6600 reservetop= [X86-32,EARLY] 6601 Format: nn[KMG] 6602 Reserves a hole at the top of the kernel virtual 6603 address space. 6604 6605 reset_devices [KNL] Force drivers to reset the underlying device 6606 during initialization. 6607 6608 resume= [SWSUSP] 6609 Specify the partition device for software suspend 6610 Format: 6611 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>} 6612 6613 resume_offset= [SWSUSP] 6614 Specify the offset from the beginning of the partition 6615 given by "resume=" at which the swap header is located, 6616 in <PAGE_SIZE> units (needed only for swap files). 6617 See Documentation/power/swsusp-and-swap-files.rst 6618 6619 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to 6620 read the resume files 6621 6622 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up. 6623 Useful for devices that are detected asynchronously 6624 (e.g. USB and MMC devices). 6625 6626 retain_initrd [RAM] Keep initrd memory after extraction. After boot, it will 6627 be accessible via /sys/firmware/initrd. 6628 6629 retbleed= [X86] Control mitigation of RETBleed (Arbitrary 6630 Speculative Code Execution with Return Instructions) 6631 vulnerability. 6632 6633 AMD-based UNRET and IBPB mitigations alone do not stop 6634 sibling threads from influencing the predictions of other 6635 sibling threads. For that reason, STIBP is used on pro- 6636 cessors that support it, and mitigate SMT on processors 6637 that don't. 6638 6639 off - no mitigation 6640 auto - automatically select a mitigation 6641 auto,nosmt - automatically select a mitigation, 6642 disabling SMT if necessary for 6643 the full mitigation (only on Zen1 6644 and older without STIBP). 6645 ibpb - On AMD, mitigate short speculation 6646 windows on basic block boundaries too. 6647 Safe, highest perf impact. It also 6648 enables STIBP if present. Not suitable 6649 on Intel. 6650 ibpb,nosmt - Like "ibpb" above but will disable SMT 6651 when STIBP is not available. This is 6652 the alternative for systems which do not 6653 have STIBP. 6654 unret - Force enable untrained return thunks, 6655 only effective on AMD f15h-f17h based 6656 systems. 6657 unret,nosmt - Like unret, but will disable SMT when STIBP 6658 is not available. This is the alternative for 6659 systems which do not have STIBP. 6660 6661 Selecting 'auto' will choose a mitigation method at run 6662 time according to the CPU. 6663 6664 Not specifying this option is equivalent to retbleed=auto. 6665 6666 rfkill.default_state= 6667 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm, 6668 etc. communication is blocked by default. 6669 1 Unblocked. 6670 6671 rfkill.master_switch_mode= 6672 0 The "airplane mode" button does nothing. 6673 1 The "airplane mode" button toggles between everything 6674 blocked and the previous configuration. 6675 2 The "airplane mode" button toggles between everything 6676 blocked and everything unblocked. 6677 6678 ring3mwait=disable 6679 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported 6680 CPUs. 6681 6682 riscv_isa_fallback [RISCV,EARLY] 6683 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit 6684 falling back to detecting extension support by parsing 6685 "riscv,isa" property on devicetree systems when the 6686 replacement properties are not found. See the Kconfig 6687 entry for RISCV_ISA_FALLBACK. 6688 6689 riscv_nousercfi= 6690 all Disable user CFI ABI to userspace even if cpu extension 6691 are available. 6692 bcfi Disable user backward CFI ABI to userspace even if 6693 the shadow stack extension is available. 6694 fcfi Disable user forward CFI ABI to userspace even if the 6695 landing pad extension is available. 6696 6697 ro [KNL] Mount root device read-only on boot 6698 6699 rodata= [KNL,EARLY] 6700 on Mark read-only kernel memory as read-only (default). 6701 off Leave read-only kernel memory writable for debugging. 6702 noalias Mark read-only kernel memory as read-only but retain 6703 writable aliases in the direct map for regions outside 6704 of the kernel image. [arm64] 6705 6706 rockchip.usb_uart 6707 [EARLY] 6708 Enable the uart passthrough on the designated usb port 6709 on Rockchip SoCs. When active, the signals of the 6710 debug-uart get routed to the D+ and D- pins of the usb 6711 port and the regular usb controller gets disabled. 6712 6713 root= [KNL] Root filesystem 6714 Usually this is a block device specifier of some kind, 6715 see the early_lookup_bdev comment in 6716 block/early-lookup.c for details. 6717 Alternatively this can be "ram" for the legacy initial 6718 ramdisk, "nfs" and "cifs" for root on a network file 6719 system, or "mtd" and "ubi" for mounting from raw flash. 6720 6721 rootdelay= [KNL] Delay (in seconds) to pause before attempting to 6722 mount the root filesystem 6723 6724 rootflags= [KNL] Set root filesystem mount option string 6725 6726 rseq_slice_ext= [KNL] RSEQ based time slice extension 6727 Format: boolean 6728 Control enablement of RSEQ based time slice extension. 6729 Default is 'on'. 6730 6731 initramfs_options= [KNL] 6732 Specify mount options for the initramfs mount. 6733 6734 rootfstype= [KNL] Set root filesystem type 6735 6736 rootwait [KNL] Wait (indefinitely) for root device to show up. 6737 Useful for devices that are detected asynchronously 6738 (e.g. USB and MMC devices). 6739 6740 rootwait= [KNL] Maximum time (in seconds) to wait for root device 6741 to show up before attempting to mount the root 6742 filesystem. 6743 6744 rproc_mem=nn[KMG][@address] 6745 [KNL,ARM,CMA] Remoteproc physical memory block. 6746 Memory area to be used by remote processor image, 6747 managed by CMA. 6748 6749 rseq_debug= [KNL] Enable or disable restartable sequence 6750 debug mode. Defaults to CONFIG_RSEQ_DEBUG_DEFAULT_ENABLE. 6751 Format: <bool> 6752 6753 rt_group_sched= [KNL] Enable or disable SCHED_RR/FIFO group scheduling 6754 when CONFIG_RT_GROUP_SCHED=y. Defaults to 6755 !CONFIG_RT_GROUP_SCHED_DEFAULT_DISABLED. 6756 Format: <bool> 6757 6758 rw [KNL] Mount root device read-write on boot 6759 6760 S [KNL] Run init in single mode 6761 6762 s390_iommu= [HW,S390] 6763 Set s390 IOTLB flushing mode 6764 strict 6765 With strict flushing every unmap operation will result 6766 in an IOTLB flush. Default is lazy flushing before 6767 reuse, which is faster. Deprecated, equivalent to 6768 iommu.strict=1. 6769 6770 s390_iommu_aperture= [KNL,S390] 6771 Specifies the size of the per device DMA address space 6772 accessible through the DMA and IOMMU APIs as a decimal 6773 factor of the size of main memory. 6774 The default is 1 meaning that one can concurrently use 6775 as many DMA addresses as physical memory is installed, 6776 if supported by hardware, and thus map all of memory 6777 once. With a value of 2 one can map all of memory twice 6778 and so on. As a special case a factor of 0 imposes no 6779 restrictions other than those given by hardware at the 6780 cost of significant additional memory use for tables. 6781 6782 sa1100ir [NET] 6783 See drivers/net/irda/sa1100_ir.c. 6784 6785 sched_proxy_exec= [KNL] 6786 Enables or disables "proxy execution" style 6787 solution to mutex-based priority inversion. 6788 Format: <bool> 6789 6790 sched_verbose [KNL,EARLY] Enables verbose scheduler debug messages. 6791 6792 schedstats= [KNL,X86] Enable or disable scheduled statistics. 6793 Allowed values are enable and disable. This feature 6794 incurs a small amount of overhead in the scheduler 6795 but is useful for debugging and performance tuning. 6796 6797 sched_thermal_decay_shift= 6798 [Deprecated] 6799 [KNL, SMP] Set a decay shift for scheduler thermal 6800 pressure signal. Thermal pressure signal follows the 6801 default decay period of other scheduler pelt 6802 signals(usually 32 ms but configurable). Setting 6803 sched_thermal_decay_shift will left shift the decay 6804 period for the thermal pressure signal by the shift 6805 value. 6806 i.e. with the default pelt decay period of 32 ms 6807 sched_thermal_decay_shift thermal pressure decay pr 6808 1 64 ms 6809 2 128 ms 6810 and so on. 6811 Format: integer between 0 and 10 6812 Default is 0. 6813 6814 scftorture.holdoff= [KNL] 6815 Number of seconds to hold off before starting 6816 test. Defaults to zero for module insertion and 6817 to 10 seconds for built-in smp_call_function() 6818 tests. 6819 6820 scftorture.longwait= [KNL] 6821 Request ridiculously long waits randomly selected 6822 up to the chosen limit in seconds. Zero (the 6823 default) disables this feature. Please note 6824 that requesting even small non-zero numbers of 6825 seconds can result in RCU CPU stall warnings, 6826 softlockup complaints, and so on. 6827 6828 scftorture.nthreads= [KNL] 6829 Number of kthreads to spawn to invoke the 6830 smp_call_function() family of functions. 6831 The default of -1 specifies a number of kthreads 6832 equal to the number of CPUs. 6833 6834 scftorture.onoff_holdoff= [KNL] 6835 Number seconds to wait after the start of the 6836 test before initiating CPU-hotplug operations. 6837 6838 scftorture.onoff_interval= [KNL] 6839 Number seconds to wait between successive 6840 CPU-hotplug operations. Specifying zero (which 6841 is the default) disables CPU-hotplug operations. 6842 6843 scftorture.shutdown_secs= [KNL] 6844 The number of seconds following the start of the 6845 test after which to shut down the system. The 6846 default of zero avoids shutting down the system. 6847 Non-zero values are useful for automated tests. 6848 6849 scftorture.stat_interval= [KNL] 6850 The number of seconds between outputting the 6851 current test statistics to the console. A value 6852 of zero disables statistics output. 6853 6854 scftorture.stutter_cpus= [KNL] 6855 The number of jiffies to wait between each change 6856 to the set of CPUs under test. 6857 6858 scftorture.use_cpus_read_lock= [KNL] 6859 Use use_cpus_read_lock() instead of the default 6860 preempt_disable() to disable CPU hotplug 6861 while invoking one of the smp_call_function*() 6862 functions. 6863 6864 scftorture.verbose= [KNL] 6865 Enable additional printk() statements. 6866 6867 scftorture.weight_single= [KNL] 6868 The probability weighting to use for the 6869 smp_call_function_single() function with a zero 6870 "wait" parameter. A value of -1 selects the 6871 default if all other weights are -1. However, 6872 if at least one weight has some other value, a 6873 value of -1 will instead select a weight of zero. 6874 6875 scftorture.weight_single_wait= [KNL] 6876 The probability weighting to use for the 6877 smp_call_function_single() function with a 6878 non-zero "wait" parameter. See weight_single. 6879 6880 scftorture.weight_many= [KNL] 6881 The probability weighting to use for the 6882 smp_call_function_many() function with a zero 6883 "wait" parameter. See weight_single. 6884 Note well that setting a high probability for 6885 this weighting can place serious IPI load 6886 on the system. 6887 6888 scftorture.weight_many_wait= [KNL] 6889 The probability weighting to use for the 6890 smp_call_function_many() function with a 6891 non-zero "wait" parameter. See weight_single 6892 and weight_many. 6893 6894 scftorture.weight_all= [KNL] 6895 The probability weighting to use for the 6896 smp_call_function_all() function with a zero 6897 "wait" parameter. See weight_single and 6898 weight_many. 6899 6900 scftorture.weight_all_wait= [KNL] 6901 The probability weighting to use for the 6902 smp_call_function_all() function with a 6903 non-zero "wait" parameter. See weight_single 6904 and weight_many. 6905 6906 sdw_mclk_divider=[SDW] 6907 Specify the MCLK divider for Intel SoundWire buses in 6908 case the BIOS does not provide the clock rate properly. 6909 6910 skew_tick= [KNL,EARLY] Offset the periodic timer tick per cpu to mitigate 6911 xtime_lock contention on larger systems, and/or RCU lock 6912 contention on all systems with CONFIG_MAXSMP set. 6913 Format: { "0" | "1" } 6914 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1" 6915 1 -- enable. 6916 Note: increases power consumption, thus should only be 6917 enabled if running jitter sensitive (HPC/RT) workloads. 6918 6919 security= [SECURITY] Choose a legacy "major" security module to 6920 enable at boot. This has been deprecated by the 6921 "lsm=" parameter. 6922 6923 selinux= [SELINUX] Disable or enable SELinux at boot time. 6924 Format: { "0" | "1" } 6925 See security/selinux/Kconfig help text. 6926 0 -- disable. 6927 1 -- enable. 6928 Default value is 1. 6929 6930 serialnumber [BUGS=X86-32] 6931 6932 sev=option[,option...] [X86-64] 6933 6934 debug 6935 Enable debug messages. 6936 6937 nosnp 6938 Do not enable SEV-SNP (applies to host/hypervisor 6939 only). Setting 'nosnp' avoids the RMP check overhead 6940 in memory accesses when users do not want to run 6941 SEV-SNP guests. 6942 6943 shapers= [NET] 6944 Maximal number of shapers. 6945 6946 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller 6947 Limit apic dumping. The parameter defines the maximal 6948 number of local apics being dumped. Also it is possible 6949 to set it to "all" by meaning -- no limit here. 6950 Format: { 1 (default) | 2 | ... | all }. 6951 The parameter valid if only apic=debug or 6952 apic=verbose is specified. 6953 Example: apic=debug show_lapic=all 6954 6955 slab_debug[=options[,slabs][;[options[,slabs]]...] [MM] 6956 Enabling slab_debug allows one to determine the 6957 culprit if slab objects become corrupted. Enabling 6958 slab_debug can create guard zones around objects and 6959 may poison objects when not in use. Also tracks the 6960 last alloc / free. For more information see 6961 Documentation/admin-guide/mm/slab.rst. 6962 (slub_debug legacy name also accepted for now) 6963 6964 Using this option implies the "no_hash_pointers" 6965 option which can be undone by adding the 6966 "hash_pointers=always" option. 6967 6968 slab_max_order= [MM] 6969 Determines the maximum allowed order for slabs. 6970 A high setting may cause OOMs due to memory 6971 fragmentation. For more information see 6972 Documentation/admin-guide/mm/slab.rst. 6973 (slub_max_order legacy name also accepted for now) 6974 6975 slab_merge [MM] 6976 Enable merging of slabs with similar size when the 6977 kernel is built without CONFIG_SLAB_MERGE_DEFAULT. 6978 (slub_merge legacy name also accepted for now) 6979 6980 slab_min_objects= [MM] 6981 The minimum number of objects per slab. SLUB will 6982 increase the slab order up to slab_max_order to 6983 generate a sufficiently large slab able to contain 6984 the number of objects indicated. The higher the number 6985 of objects the smaller the overhead of tracking slabs 6986 and the less frequently locks need to be acquired. 6987 For more information see 6988 Documentation/admin-guide/mm/slab.rst. 6989 (slub_min_objects legacy name also accepted for now) 6990 6991 slab_min_order= [MM] 6992 Determines the minimum page order for slabs. Must be 6993 lower or equal to slab_max_order. For more information see 6994 Documentation/admin-guide/mm/slab.rst. 6995 (slub_min_order legacy name also accepted for now) 6996 6997 slab_nomerge [MM] 6998 Disable merging of slabs with similar size. May be 6999 necessary if there is some reason to distinguish 7000 allocs to different slabs, especially in hardened 7001 environments where the risk of heap overflows and 7002 layout control by attackers can usually be 7003 frustrated by disabling merging. This will reduce 7004 most of the exposure of a heap attack to a single 7005 cache (risks via metadata attacks are mostly 7006 unchanged). Debug options disable merging on their 7007 own. 7008 For more information see 7009 Documentation/admin-guide/mm/slab.rst. 7010 (slub_nomerge legacy name also accepted for now) 7011 7012 slab_strict_numa [MM] 7013 Support memory policies on a per object level 7014 in the slab allocator. The default is for memory 7015 policies to be applied at the folio level when 7016 a new folio is needed or a partial folio is 7017 retrieved from the lists. Increases overhead 7018 in the slab fastpaths but gains more accurate 7019 NUMA kernel object placement which helps with slow 7020 interconnects in NUMA systems. 7021 7022 slram= [HW,MTD] 7023 7024 smart2= [HW] 7025 Format: <io1>[,<io2>[,...,<io8>]] 7026 7027 smp.csd_lock_timeout= [KNL] 7028 Specify the period of time in milliseconds 7029 that smp_call_function() and friends will wait 7030 for a CPU to release the CSD lock. This is 7031 useful when diagnosing bugs involving CPUs 7032 disabling interrupts for extended periods 7033 of time. Defaults to 5,000 milliseconds, and 7034 setting a value of zero disables this feature. 7035 This feature may be more efficiently disabled 7036 using the csdlock_debug- kernel parameter. 7037 7038 smp.panic_on_ipistall= [KNL] 7039 If a csd_lock_timeout extends for more than 7040 the specified number of milliseconds, panic the 7041 system. By default, let CSD-lock acquisition 7042 take as long as they take. Specifying 300,000 7043 for this value provides a 5-minute timeout. 7044 7045 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices 7046 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port 7047 smsc-ircc2.ircc_sir= [HW] SIR base I/O port 7048 smsc-ircc2.ircc_fir= [HW] FIR base I/O port 7049 smsc-ircc2.ircc_irq= [HW] IRQ line 7050 smsc-ircc2.ircc_dma= [HW] DMA channel 7051 smsc-ircc2.ircc_transceiver= [HW] Transceiver type: 7052 0: Toshiba Satellite 1800 (GP data pin select) 7053 1: Fast pin select (default) 7054 2: ATC IRMode 7055 7056 smt= [KNL,MIPS,S390,EARLY] Set the maximum number of threads 7057 (logical CPUs) to use per physical CPU on systems 7058 capable of symmetric multithreading (SMT). Will 7059 be capped to the actual hardware limit. 7060 Format: <integer> 7061 Default: -1 (no limit) 7062 7063 softlockup_panic= 7064 [KNL] Should the soft-lockup detector generate panics. 7065 Format: <int> 7066 7067 A value of non-zero instructs the soft-lockup detector 7068 to panic the machine when a soft-lockup duration exceeds 7069 N thresholds. It is also controlled by the kernel.softlockup_panic 7070 sysctl and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the 7071 respective build-time switch to that functionality. 7072 7073 softlockup_all_cpu_backtrace= 7074 [KNL] Should the soft-lockup detector generate 7075 backtraces on all cpus. 7076 Format: 0 | 1 7077 7078 sonypi.*= [HW] Sony Programmable I/O Control Device driver 7079 See Documentation/admin-guide/laptops/sonypi.rst 7080 7081 spectre_bhi= [X86] Control mitigation of Branch History Injection 7082 (BHI) vulnerability. This setting affects the 7083 deployment of the HW BHI control and the SW BHB 7084 clearing sequence. 7085 7086 on - (default) Enable the HW or SW mitigation as 7087 needed. This protects the kernel from 7088 both syscalls and VMs. 7089 vmexit - On systems which don't have the HW mitigation 7090 available, enable the SW mitigation on vmexit 7091 ONLY. On such systems, the host kernel is 7092 protected from VM-originated BHI attacks, but 7093 may still be vulnerable to syscall attacks. 7094 off - Disable the mitigation. 7095 7096 spectre_v2= [X86,EARLY] Control mitigation of Spectre variant 2 7097 (indirect branch speculation) vulnerability. 7098 The default operation protects the kernel from 7099 user space attacks. 7100 7101 on - unconditionally enable, implies 7102 spectre_v2_user=on 7103 off - unconditionally disable, implies 7104 spectre_v2_user=off 7105 auto - kernel detects whether your CPU model is 7106 vulnerable 7107 7108 Selecting 'on' will, and 'auto' may, choose a 7109 mitigation method at run time according to the 7110 CPU, the available microcode, the setting of the 7111 CONFIG_MITIGATION_RETPOLINE configuration option, 7112 and the compiler with which the kernel was built. 7113 7114 Selecting 'on' will also enable the mitigation 7115 against user space to user space task attacks. 7116 Selecting specific mitigation does not force enable 7117 user mitigations. 7118 7119 Selecting 'off' will disable both the kernel and 7120 the user space protections. 7121 7122 Specific mitigations can also be selected manually: 7123 7124 retpoline - replace indirect branches 7125 retpoline,generic - Retpolines 7126 retpoline,lfence - LFENCE; indirect branch 7127 retpoline,amd - alias for retpoline,lfence 7128 eibrs - Enhanced/Auto IBRS 7129 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines 7130 eibrs,lfence - Enhanced/Auto IBRS + LFENCE 7131 ibrs - use IBRS to protect kernel 7132 7133 Not specifying this option is equivalent to 7134 spectre_v2=auto. 7135 7136 spectre_v2_user= 7137 [X86] Control mitigation of Spectre variant 2 7138 (indirect branch speculation) vulnerability between 7139 user space tasks 7140 7141 on - Unconditionally enable mitigations. Is 7142 enforced by spectre_v2=on 7143 7144 off - Unconditionally disable mitigations. Is 7145 enforced by spectre_v2=off 7146 7147 prctl - Indirect branch speculation is enabled, 7148 but mitigation can be enabled via prctl 7149 per thread. The mitigation control state 7150 is inherited on fork. 7151 7152 prctl,ibpb 7153 - Like "prctl" above, but only STIBP is 7154 controlled per thread. IBPB is issued 7155 always when switching between different user 7156 space processes. 7157 7158 seccomp 7159 - Same as "prctl" above, but all seccomp 7160 threads will enable the mitigation unless 7161 they explicitly opt out. 7162 7163 seccomp,ibpb 7164 - Like "seccomp" above, but only STIBP is 7165 controlled per thread. IBPB is issued 7166 always when switching between different 7167 user space processes. 7168 7169 auto - Kernel selects the mitigation depending on 7170 the available CPU features and vulnerability. 7171 7172 Default mitigation: "prctl" 7173 7174 Not specifying this option is equivalent to 7175 spectre_v2_user=auto. 7176 7177 spec_rstack_overflow= 7178 [X86,EARLY] Control RAS overflow mitigation on AMD Zen CPUs 7179 7180 off - Disable mitigation 7181 microcode - Enable microcode mitigation only 7182 safe-ret - Enable sw-only safe RET mitigation (default) 7183 ibpb - Enable mitigation by issuing IBPB on 7184 kernel entry 7185 ibpb-vmexit - Issue IBPB only on VMEXIT 7186 (cloud-specific mitigation) 7187 7188 spec_store_bypass_disable= 7189 [HW,EARLY] Control Speculative Store Bypass (SSB) Disable mitigation 7190 (Speculative Store Bypass vulnerability) 7191 7192 Certain CPUs are vulnerable to an exploit against a 7193 a common industry wide performance optimization known 7194 as "Speculative Store Bypass" in which recent stores 7195 to the same memory location may not be observed by 7196 later loads during speculative execution. The idea 7197 is that such stores are unlikely and that they can 7198 be detected prior to instruction retirement at the 7199 end of a particular speculation execution window. 7200 7201 In vulnerable processors, the speculatively forwarded 7202 store can be used in a cache side channel attack, for 7203 example to read memory to which the attacker does not 7204 directly have access (e.g. inside sandboxed code). 7205 7206 This parameter controls whether the Speculative Store 7207 Bypass optimization is used. 7208 7209 On x86 the options are: 7210 7211 on - Unconditionally disable Speculative Store Bypass 7212 off - Unconditionally enable Speculative Store Bypass 7213 auto - Kernel detects whether the CPU model contains an 7214 implementation of Speculative Store Bypass and 7215 picks the most appropriate mitigation. If the 7216 CPU is not vulnerable, "off" is selected. If the 7217 CPU is vulnerable the default mitigation is 7218 architecture and Kconfig dependent. See below. 7219 prctl - Control Speculative Store Bypass per thread 7220 via prctl. Speculative Store Bypass is enabled 7221 for a process by default. The state of the control 7222 is inherited on fork. 7223 seccomp - Same as "prctl" above, but all seccomp threads 7224 will disable SSB unless they explicitly opt out. 7225 7226 Default mitigations: 7227 X86: "prctl" 7228 7229 On powerpc the options are: 7230 7231 on,auto - On Power8 and Power9 insert a store-forwarding 7232 barrier on kernel entry and exit. On Power7 7233 perform a software flush on kernel entry and 7234 exit. 7235 off - No action. 7236 7237 Not specifying this option is equivalent to 7238 spec_store_bypass_disable=auto. 7239 7240 split_lock_detect= 7241 [X86] Enable split lock detection or bus lock detection 7242 7243 When enabled (and if hardware support is present), atomic 7244 instructions that access data across cache line 7245 boundaries will result in an alignment check exception 7246 for split lock detection or a debug exception for 7247 bus lock detection. 7248 7249 off - not enabled 7250 7251 warn - the kernel will emit rate-limited warnings 7252 about applications triggering the #AC 7253 exception or the #DB exception. This mode is 7254 the default on CPUs that support split lock 7255 detection or bus lock detection. Default 7256 behavior is by #AC if both features are 7257 enabled in hardware. 7258 7259 fatal - the kernel will send SIGBUS to applications 7260 that trigger the #AC exception or the #DB 7261 exception. Default behavior is by #AC if 7262 both features are enabled in hardware. 7263 7264 ratelimit:N - 7265 Set system wide rate limit to N bus locks 7266 per second for bus lock detection. 7267 0 < N <= 1000. 7268 7269 N/A for split lock detection. 7270 7271 7272 If an #AC exception is hit in the kernel or in 7273 firmware (i.e. not while executing in user mode) 7274 the kernel will oops in either "warn" or "fatal" 7275 mode. 7276 7277 #DB exception for bus lock is triggered only when 7278 CPL > 0. 7279 7280 srbds= [X86,INTEL,EARLY] 7281 Control the Special Register Buffer Data Sampling 7282 (SRBDS) mitigation. 7283 7284 Certain CPUs are vulnerable to an MDS-like 7285 exploit which can leak bits from the random 7286 number generator. 7287 7288 By default, this issue is mitigated by 7289 microcode. However, the microcode fix can cause 7290 the RDRAND and RDSEED instructions to become 7291 much slower. Among other effects, this will 7292 result in reduced throughput from /dev/urandom. 7293 7294 The microcode mitigation can be disabled with 7295 the following option: 7296 7297 off: Disable mitigation and remove 7298 performance impact to RDRAND and RDSEED 7299 7300 srcutree.big_cpu_lim [KNL] 7301 Specifies the number of CPUs constituting a 7302 large system, such that srcu_struct structures 7303 should immediately allocate an srcu_node array. 7304 This kernel-boot parameter defaults to 128, 7305 but takes effect only when the low-order four 7306 bits of srcutree.convert_to_big is equal to 3 7307 (decide at boot). 7308 7309 srcutree.convert_to_big [KNL] 7310 Specifies under what conditions an SRCU tree 7311 srcu_struct structure will be converted to big 7312 form, that is, with an rcu_node tree: 7313 7314 0: Never. 7315 1: At init_srcu_struct() time. 7316 2: When rcutorture decides to. 7317 3: Decide at boot time (default). 7318 0x1X: Above plus if high contention. 7319 7320 Either way, the srcu_node tree will be sized based 7321 on the actual runtime number of CPUs (nr_cpu_ids) 7322 instead of the compile-time CONFIG_NR_CPUS. 7323 7324 srcutree.counter_wrap_check [KNL] 7325 Specifies how frequently to check for 7326 grace-period sequence counter wrap for the 7327 srcu_data structure's ->srcu_gp_seq_needed field. 7328 The greater the number of bits set in this kernel 7329 parameter, the less frequently counter wrap will 7330 be checked for. Note that the bottom two bits 7331 are ignored. 7332 7333 srcutree.exp_holdoff [KNL] 7334 Specifies how many nanoseconds must elapse 7335 since the end of the last SRCU grace period for 7336 a given srcu_struct until the next normal SRCU 7337 grace period will be considered for automatic 7338 expediting. Set to zero to disable automatic 7339 expediting. 7340 7341 srcutree.srcu_max_nodelay [KNL] 7342 Specifies the number of no-delay instances 7343 per jiffy for which the SRCU grace period 7344 worker thread will be rescheduled with zero 7345 delay. Beyond this limit, worker thread will 7346 be rescheduled with a sleep delay of one jiffy. 7347 7348 srcutree.srcu_max_nodelay_phase [KNL] 7349 Specifies the per-grace-period phase, number of 7350 non-sleeping polls of readers. Beyond this limit, 7351 grace period worker thread will be rescheduled 7352 with a sleep delay of one jiffy, between each 7353 rescan of the readers, for a grace period phase. 7354 7355 srcutree.srcu_retry_check_delay [KNL] 7356 Specifies number of microseconds of non-sleeping 7357 delay between each non-sleeping poll of readers. 7358 7359 srcutree.small_contention_lim [KNL] 7360 Specifies the number of update-side contention 7361 events per jiffy will be tolerated before 7362 initiating a conversion of an srcu_struct 7363 structure to big form. Note that the value of 7364 srcutree.convert_to_big must have the 0x10 bit 7365 set for contention-based conversions to occur. 7366 7367 ssbd= [ARM64,HW,EARLY] 7368 Speculative Store Bypass Disable control 7369 7370 On CPUs that are vulnerable to the Speculative 7371 Store Bypass vulnerability and offer a 7372 firmware based mitigation, this parameter 7373 indicates how the mitigation should be used: 7374 7375 force-on: Unconditionally enable mitigation for 7376 for both kernel and userspace 7377 force-off: Unconditionally disable mitigation for 7378 for both kernel and userspace 7379 kernel: Always enable mitigation in the 7380 kernel, and offer a prctl interface 7381 to allow userspace to register its 7382 interest in being mitigated too. 7383 7384 stack_guard_gap= [MM] 7385 override the default stack gap protection. The value 7386 is in page units and it defines how many pages prior 7387 to (for stacks growing down) resp. after (for stacks 7388 growing up) the main stack are reserved for no other 7389 mapping. Default value is 256 pages. 7390 7391 stack_depot_disable= [KNL,EARLY] 7392 Setting this to true through kernel command line will 7393 disable the stack depot thereby saving the static memory 7394 consumed by the stack hash table. By default this is set 7395 to false. 7396 7397 stack_depot_max_pools= [KNL,EARLY] 7398 Specify the maximum number of pools to use for storing 7399 stack traces. Pools are allocated on-demand up to this 7400 limit. Default value is 8191 pools. 7401 7402 stacktrace [FTRACE] 7403 Enable the stack tracer on boot up. 7404 7405 stacktrace_filter=[function-list] 7406 [FTRACE] Limit the functions that the stack tracer 7407 will trace at boot up. function-list is a comma-separated 7408 list of functions. This list can be changed at run 7409 time by the stack_trace_filter file in the debugfs 7410 tracing directory. Note, this enables stack tracing 7411 and the stacktrace above is not needed. 7412 7413 sti= [PARISC,HW] 7414 Format: <num> 7415 Set the STI (builtin display/keyboard on the HP-PARISC 7416 machines) console (graphic card) which should be used 7417 as the initial boot-console. 7418 See also comment in drivers/video/console/sticore.c. 7419 7420 sti_font= [HW] 7421 See comment in drivers/video/console/sticore.c. 7422 7423 stifb= [HW] 7424 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]] 7425 7426 strict_sas_size= 7427 [X86] 7428 Format: <bool> 7429 Enable or disable strict sigaltstack size checks 7430 against the required signal frame size which 7431 depends on the supported FPU features. This can 7432 be used to filter out binaries which have 7433 not yet been made aware of AT_MINSIGSTKSZ. 7434 7435 stress_hpt [PPC,EARLY] 7436 Limits the number of kernel HPT entries in the hash 7437 page table to increase the rate of hash page table 7438 faults on kernel addresses. 7439 7440 stress_slb [PPC,EARLY] 7441 Limits the number of kernel SLB entries, and flushes 7442 them frequently to increase the rate of SLB faults 7443 on kernel addresses. 7444 7445 no_slb_preload [PPC,EARLY] 7446 Disables slb preloading for userspace. 7447 7448 sunrpc.min_resvport= 7449 sunrpc.max_resvport= 7450 [NFS,SUNRPC] 7451 SunRPC servers often require that client requests 7452 originate from a privileged port (i.e. a port in the 7453 range 0 < portnr < 1024). 7454 An administrator who wishes to reserve some of these 7455 ports for other uses may adjust the range that the 7456 kernel's sunrpc client considers to be privileged 7457 using these two parameters to set the minimum and 7458 maximum port values. 7459 7460 sunrpc.svc_rpc_per_connection_limit= 7461 [NFS,SUNRPC] 7462 Limit the number of requests that the server will 7463 process in parallel from a single connection. 7464 The default value is 0 (no limit). 7465 7466 sunrpc.pool_mode= 7467 [NFS] 7468 Control how the NFS server code allocates CPUs to 7469 service thread pools. Depending on how many NICs 7470 you have and where their interrupts are bound, this 7471 option will affect which CPUs will do NFS serving. 7472 Note: this parameter cannot be changed while the 7473 NFS server is running. 7474 7475 auto the server chooses an appropriate mode 7476 automatically using heuristics 7477 global a single global pool contains all CPUs 7478 percpu one pool for each CPU 7479 pernode one pool for each NUMA node (equivalent 7480 to global on non-NUMA machines) 7481 7482 sunrpc.tcp_slot_table_entries= 7483 sunrpc.udp_slot_table_entries= 7484 [NFS,SUNRPC] 7485 Sets the upper limit on the number of simultaneous 7486 RPC calls that can be sent from the client to a 7487 server. Increasing these values may allow you to 7488 improve throughput, but will also increase the 7489 amount of memory reserved for use by the client. 7490 7491 suspend.pm_test_delay= 7492 [SUSPEND] 7493 Sets the number of seconds to remain in a suspend test 7494 mode before resuming the system (see 7495 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 7496 is set. Default value is 5. 7497 7498 svm= [PPC] 7499 Format: { on | off | y | n | 1 | 0 } 7500 This parameter controls use of the Protected 7501 Execution Facility on pSeries. 7502 7503 swiotlb= [ARM,PPC,MIPS,X86,S390,EARLY] 7504 Format: { <int> [,<int>] | force | noforce } 7505 <int> -- Number of I/O TLB slabs 7506 <int> -- Second integer after comma. Number of swiotlb 7507 areas with their own lock. Will be rounded up 7508 to a power of 2. 7509 force -- force using of bounce buffers even if they 7510 wouldn't be automatically used by the kernel 7511 noforce -- Never use bounce buffers (for debugging) 7512 7513 switches= [HW,M68k,EARLY] 7514 7515 sysctl.*= [KNL] 7516 Set a sysctl parameter, right before loading the init 7517 process, as if the value was written to the respective 7518 /proc/sys/... file. Both '.' and '/' are recognized as 7519 separators. Unrecognized parameters and invalid values 7520 are reported in the kernel log. Sysctls registered 7521 later by a loaded module cannot be set this way. 7522 Example: sysctl.vm.swappiness=40 7523 7524 sysrq_always_enabled 7525 [KNL] 7526 Ignore sysrq setting - this boot parameter will 7527 neutralize any effect of /proc/sys/kernel/sysrq. 7528 Useful for debugging. 7529 7530 tcpmhash_entries= [KNL,NET] 7531 Set the number of tcp_metrics_hash slots. 7532 Default value is 8192 or 16384 depending on total 7533 ram pages. This is used to specify the TCP metrics 7534 cache size. See Documentation/networking/ip-sysctl.rst 7535 "tcp_no_metrics_save" section for more details. 7536 7537 tdfx= [HW,DRM] 7538 7539 test_suspend= [SUSPEND] 7540 Format: { "mem" | "standby" | "freeze" }[,N] 7541 Specify "mem" (for Suspend-to-RAM) or "standby" (for 7542 standby suspend) or "freeze" (for suspend type freeze) 7543 as the system sleep state during system startup with 7544 the optional capability to repeat N number of times. 7545 The system is woken from this state using a 7546 wakeup-capable RTC alarm. 7547 7548 thash_entries= [KNL,NET] 7549 Set number of hash buckets for TCP connection 7550 7551 thermal.act= [HW,ACPI] 7552 -1: disable all active trip points in all thermal zones 7553 <degrees C>: override all lowest active trip points 7554 7555 thermal.crt= [HW,ACPI] 7556 -1: disable all critical trip points in all thermal zones 7557 <degrees C>: override all critical trip points 7558 7559 thermal.off= [HW,ACPI] 7560 1: disable ACPI thermal control 7561 7562 thermal.psv= [HW,ACPI] 7563 -1: disable all passive trip points 7564 <degrees C>: override all passive trip points to this 7565 value 7566 7567 thermal.tzp= [HW,ACPI] 7568 Specify global default ACPI thermal zone polling rate 7569 <deci-seconds>: poll all this frequency 7570 0: no polling (default) 7571 7572 thp_anon= [KNL] 7573 Format: <size>[KMG],<size>[KMG]:<state>;<size>[KMG]-<size>[KMG]:<state> 7574 state is one of "always", "madvise", "never" or "inherit". 7575 Control the default behavior of the system with respect 7576 to anonymous transparent hugepages. 7577 Can be used multiple times for multiple anon THP sizes. 7578 See Documentation/admin-guide/mm/transhuge.rst for more 7579 details. 7580 7581 threadirqs [KNL,EARLY] 7582 Force threading of all interrupt handlers except those 7583 marked explicitly IRQF_NO_THREAD. 7584 7585 thp_shmem= [KNL] 7586 Format: <size>[KMG],<size>[KMG]:<policy>;<size>[KMG]-<size>[KMG]:<policy> 7587 Control the default policy of each hugepage size for the 7588 internal shmem mount. <policy> is one of policies available 7589 for the shmem mount ("always", "inherit", "never", "within_size", 7590 and "advise"). 7591 It can be used multiple times for multiple shmem THP sizes. 7592 See Documentation/admin-guide/mm/transhuge.rst for more 7593 details. 7594 7595 topology= [S390,EARLY] 7596 Format: {off | on} 7597 Specify if the kernel should make use of the cpu 7598 topology information if the hardware supports this. 7599 The scheduler will make use of this information and 7600 e.g. base its process migration decisions on it. 7601 Default is on. 7602 7603 torture.disable_onoff_at_boot= [KNL] 7604 Prevent the CPU-hotplug component of torturing 7605 until after init has spawned. 7606 7607 torture.ftrace_dump_at_shutdown= [KNL] 7608 Dump the ftrace buffer at torture-test shutdown, 7609 even if there were no errors. This can be a 7610 very costly operation when many torture tests 7611 are running concurrently, especially on systems 7612 with rotating-rust storage. 7613 7614 torture.verbose_sleep_frequency= [KNL] 7615 Specifies how many verbose printk()s should be 7616 emitted between each sleep. The default of zero 7617 disables verbose-printk() sleeping. 7618 7619 torture.verbose_sleep_duration= [KNL] 7620 Duration of each verbose-printk() sleep in jiffies. 7621 7622 tpm.disable_pcr_integrity= [HW,TPM] 7623 Do not protect PCR registers from unintended physical 7624 access, or interposers in the bus by the means of 7625 having an integrity protected session wrapped around 7626 TPM2_PCR_Extend command. Consider this in a situation 7627 where TPM is heavily utilized by IMA, thus protection 7628 causing a major performance hit, and the space where 7629 machines are deployed is by other means guarded. 7630 7631 tpm_crb_ffa.busy_timeout_ms= [ARM64,TPM] 7632 Maximum time in milliseconds to retry sending a message 7633 to the TPM service before giving up. This parameter controls 7634 how long the system will continue retrying when the TPM 7635 service is busy. 7636 Format: <unsigned int> 7637 Default: 2000 (2 seconds) 7638 7639 tpm_suspend_pcr=[HW,TPM] 7640 Format: integer pcr id 7641 Specify that at suspend time, the tpm driver 7642 should extend the specified pcr with zeros, 7643 as a workaround for some chips which fail to 7644 flush the last written pcr on TPM_SaveState. 7645 This will guarantee that all the other pcrs 7646 are saved. 7647 7648 tpm_tis.interrupts= [HW,TPM] 7649 Enable interrupts for the MMIO based physical layer 7650 for the FIFO interface. By default it is set to false 7651 (0). For more information about TPM hardware interfaces 7652 defined by Trusted Computing Group (TCG) see 7653 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/ 7654 7655 tp_printk [FTRACE] 7656 Have the tracepoints sent to printk as well as the 7657 tracing ring buffer. This is useful for early boot up 7658 where the system hangs or reboots and does not give the 7659 option for reading the tracing buffer or performing a 7660 ftrace_dump_on_oops. 7661 7662 To turn off having tracepoints sent to printk, 7663 echo 0 > /proc/sys/kernel/tracepoint_printk 7664 Note, echoing 1 into this file without the 7665 tp_printk kernel cmdline option has no effect. 7666 7667 The tp_printk_stop_on_boot (see below) can also be used 7668 to stop the printing of events to console at 7669 late_initcall_sync. 7670 7671 ** CAUTION ** 7672 7673 Having tracepoints sent to printk() and activating high 7674 frequency tracepoints such as irq or sched, can cause 7675 the system to live lock. 7676 7677 tp_printk_stop_on_boot [FTRACE] 7678 When tp_printk (above) is set, it can cause a lot of noise 7679 on the console. It may be useful to only include the 7680 printing of events during boot up, as user space may 7681 make the system inoperable. 7682 7683 This command line option will stop the printing of events 7684 to console at the late_initcall_sync() time frame. 7685 7686 trace_buf_size=nn[KMG] 7687 [FTRACE] will set tracing buffer size on each cpu. 7688 7689 trace_clock= [FTRACE] Set the clock used for tracing events 7690 at boot up. 7691 local - Use the per CPU time stamp counter 7692 (converted into nanoseconds). Fast, but 7693 depending on the architecture, may not be 7694 in sync between CPUs. 7695 global - Event time stamps are synchronized across 7696 CPUs. May be slower than the local clock, 7697 but better for some race conditions. 7698 counter - Simple counting of events (1, 2, ..) 7699 note, some counts may be skipped due to the 7700 infrastructure grabbing the clock more than 7701 once per event. 7702 uptime - Use jiffies as the time stamp. 7703 perf - Use the same clock that perf uses. 7704 mono - Use ktime_get_mono_fast_ns() for time stamps. 7705 mono_raw - Use ktime_get_raw_fast_ns() for time 7706 stamps. 7707 boot - Use ktime_get_boot_fast_ns() for time stamps. 7708 Architectures may add more clocks. See 7709 Documentation/trace/ftrace.rst for more details. 7710 7711 trace_event=[event-list] 7712 [FTRACE] Set and start specified trace events in order 7713 to facilitate early boot debugging. The event-list is a 7714 comma-separated list of trace events to enable. See 7715 also Documentation/trace/events.rst 7716 7717 To enable modules, use :mod: keyword: 7718 7719 trace_event=:mod:<module> 7720 7721 The value before :mod: will only enable specific events 7722 that are part of the module. See the above mentioned 7723 document for more information. 7724 7725 trace_instance=[instance-info] 7726 [FTRACE] Create a ring buffer instance early in boot up. 7727 This will be listed in: 7728 7729 /sys/kernel/tracing/instances 7730 7731 Events can be enabled at the time the instance is created 7732 via: 7733 7734 trace_instance=<name>,<system1>:<event1>,<system2>:<event2> 7735 7736 Note, the "<system*>:" portion is optional if the event is 7737 unique. 7738 7739 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall 7740 7741 will enable the "sched_switch" event (note, the "sched:" is optional, and 7742 the same thing would happen if it was left off). The irq_handler_entry 7743 event, and all events under the "initcall" system. 7744 7745 Flags can be added to the instance to modify its behavior when it is 7746 created. The flags are separated by '^'. 7747 7748 The available flags are: 7749 7750 traceoff - Have the tracing instance tracing disabled after it is created. 7751 traceprintk - Have trace_printk() write into this trace instance 7752 (note, "printk" and "trace_printk" can also be used) 7753 7754 trace_instance=foo^traceoff^traceprintk,sched,irq 7755 7756 The flags must come before the defined events. 7757 7758 If memory has been reserved (see memmap for x86), the instance 7759 can use that memory: 7760 7761 memmap=12M$0x284500000 trace_instance=boot_map@0x284500000:12M 7762 7763 The above will create a "boot_map" instance that uses the physical 7764 memory at 0x284500000 that is 12Megs. The per CPU buffers of that 7765 instance will be split up accordingly. 7766 7767 Alternatively, the memory can be reserved by the reserve_mem option: 7768 7769 reserve_mem=12M:4096:trace trace_instance=boot_map@trace 7770 7771 This will reserve 12 megabytes at boot up with a 4096 byte alignment 7772 and place the ring buffer in this memory. Note that due to KASLR, the 7773 memory may not be the same location each time, which will not preserve 7774 the buffer content. 7775 7776 Also note that the layout of the ring buffer data may change between 7777 kernel versions where the validator will fail and reset the ring buffer 7778 if the layout is not the same as the previous kernel. 7779 7780 If the ring buffer is used for persistent bootups and has events enabled, 7781 it is recommend to disable tracing so that events from a previous boot do not 7782 mix with events of the current boot (unless you are debugging a random crash 7783 at boot up). 7784 7785 reserve_mem=12M:4096:trace trace_instance=boot_map^traceoff^traceprintk@trace,sched,irq 7786 7787 Note, saving the trace buffer across reboots does require that the system 7788 is set up to not wipe memory. For instance, CONFIG_RESET_ATTACK_MITIGATION 7789 can force a memory reset on boot which will clear any trace that was stored. 7790 This is just one of many ways that can clear memory. Make sure your system 7791 keeps the content of memory across reboots before relying on this option. 7792 7793 NB: Both the mapped address and size must be page aligned for the architecture. 7794 7795 See also Documentation/trace/debugging.rst 7796 7797 7798 trace_options=[option-list] 7799 [FTRACE] Enable or disable tracer options at boot. 7800 The option-list is a comma delimited list of options 7801 that can be enabled or disabled just as if you were 7802 to echo the option name into 7803 7804 /sys/kernel/tracing/trace_options 7805 7806 For example, to enable stacktrace option (to dump the 7807 stack trace of each event), add to the command line: 7808 7809 trace_options=stacktrace 7810 7811 See also Documentation/trace/ftrace.rst "trace options" 7812 section. 7813 7814 trace_trigger=[trigger-list] 7815 [FTRACE] Add an event trigger on specific events. 7816 Set a trigger on top of a specific event, with an optional 7817 filter. 7818 7819 The format is "trace_trigger=<event>.<trigger>[ if <filter>],..." 7820 Where more than one trigger may be specified that are comma delimited. 7821 7822 For example: 7823 7824 trace_trigger="sched_switch.stacktrace if prev_state == 2" 7825 7826 The above will enable the "stacktrace" trigger on the "sched_switch" 7827 event but only trigger it if the "prev_state" of the "sched_switch" 7828 event is "2" (TASK_UNINTERRUPTIBLE). 7829 7830 See also "Event triggers" in Documentation/trace/events.rst 7831 7832 7833 traceoff_after_boot 7834 [FTRACE] Sometimes tracing is used to debug issues 7835 during the boot process. Since the trace buffer has a 7836 limited amount of storage, it may be prudent to 7837 disable tracing after the boot is finished, otherwise 7838 the critical information may be overwritten. With this 7839 option, the main tracing buffer will be turned off at 7840 the end of the boot process. 7841 7842 traceoff_on_warning 7843 [FTRACE] enable this option to disable tracing when a 7844 warning is hit. This turns off "tracing_on". Tracing can 7845 be enabled again by echoing '1' into the "tracing_on" 7846 file located in /sys/kernel/tracing/ 7847 7848 This option is useful, as it disables the trace before 7849 the WARNING dump is called, which prevents the trace to 7850 be filled with content caused by the warning output. 7851 7852 This option can also be set at run time via the sysctl 7853 option: kernel/traceoff_on_warning 7854 7855 transparent_hugepage= 7856 [KNL] 7857 Format: [always|madvise|never] 7858 Can be used to control the default behavior of the system 7859 with respect to transparent hugepages. 7860 See Documentation/admin-guide/mm/transhuge.rst 7861 for more details. 7862 7863 transparent_hugepage_shmem= [KNL] 7864 Format: [always|within_size|advise|never|deny|force] 7865 Can be used to control the hugepage allocation policy for 7866 the internal shmem mount. 7867 See Documentation/admin-guide/mm/transhuge.rst 7868 for more details. 7869 7870 transparent_hugepage_tmpfs= [KNL] 7871 Format: [always|within_size|advise|never] 7872 Can be used to control the default hugepage allocation policy 7873 for the tmpfs mount. 7874 See Documentation/admin-guide/mm/transhuge.rst 7875 for more details. 7876 7877 trusted.source= [KEYS] 7878 Format: <string> 7879 This parameter identifies the trust source as a backend 7880 for trusted keys implementation. Supported trust 7881 sources: 7882 - "tpm" 7883 - "tee" 7884 - "caam" 7885 - "dcp" 7886 - "pkwm" 7887 If not specified then it defaults to iterating through 7888 the trust source list starting with TPM and assigns the 7889 first trust source as a backend which is initialized 7890 successfully during iteration. 7891 7892 trusted.rng= [KEYS] 7893 Format: <string> 7894 The RNG used to generate key material for trusted keys. 7895 Can be one of: 7896 - "kernel" 7897 - the same value as trusted.source: "tpm" or "tee" 7898 - "default" 7899 If not specified, "default" is used. In this case, 7900 the RNG's choice is left to each individual trust source. 7901 7902 trusted.dcp_use_otp_key 7903 This is intended to be used in combination with 7904 trusted.source=dcp and will select the DCP OTP key 7905 instead of the DCP UNIQUE key blob encryption. 7906 7907 trusted.dcp_skip_zk_test 7908 This is intended to be used in combination with 7909 trusted.source=dcp and will disable the check if the 7910 blob key is all zeros. This is helpful for situations where 7911 having this key zero'ed is acceptable. E.g. in testing 7912 scenarios. 7913 7914 tsa= [X86] Control mitigation for Transient Scheduler 7915 Attacks on AMD CPUs. Search the following in your 7916 favourite search engine for more details: 7917 7918 "Technical guidance for mitigating transient scheduler 7919 attacks". 7920 7921 off - disable the mitigation 7922 on - enable the mitigation (default) 7923 user - mitigate only user/kernel transitions 7924 vm - mitigate only guest/host transitions 7925 7926 7927 tsc= Disable clocksource stability checks for TSC. 7928 Format: <string> 7929 [x86] reliable: mark tsc clocksource as reliable, this 7930 disables clocksource verification at runtime, as well 7931 as the stability checks done at bootup. Used to enable 7932 high-resolution timer mode on older hardware, and in 7933 virtualized environment. 7934 [x86] noirqtime: Do not use TSC to do irq accounting. 7935 Used to run time disable IRQ_TIME_ACCOUNTING on any 7936 platforms where RDTSC is slow and this accounting 7937 can add overhead. 7938 [x86] unstable: mark the TSC clocksource as unstable, this 7939 marks the TSC unconditionally unstable at bootup and 7940 avoids any further wobbles once the TSC watchdog notices. 7941 [x86] nowatchdog: disable clocksource watchdog. Used 7942 in situations with strict latency requirements (where 7943 interruptions from clocksource watchdog are not 7944 acceptable). 7945 [x86] recalibrate: force recalibration against a HW timer 7946 (HPET or PM timer) on systems whose TSC frequency was 7947 obtained from HW or FW using either an MSR or CPUID(0x15). 7948 Warn if the difference is more than 500 ppm. 7949 [x86] watchdog: Use TSC as the watchdog clocksource with 7950 which to check other HW timers (HPET or PM timer), but 7951 only on systems where TSC has been deemed trustworthy. 7952 This will be suppressed by an earlier tsc=nowatchdog and 7953 can be overridden by a later tsc=nowatchdog. A console 7954 message will flag any such suppression or overriding. 7955 7956 tsc_early_khz= [X86,EARLY] Skip early TSC calibration and use the given 7957 value instead. Useful when the early TSC frequency discovery 7958 procedure is not reliable, such as on overclocked systems 7959 with CPUID.16h support and partial CPUID.15h support. 7960 Format: <unsigned int> 7961 7962 tsx= [X86] Control Transactional Synchronization 7963 Extensions (TSX) feature in Intel processors that 7964 support TSX control. 7965 7966 This parameter controls the TSX feature. The options are: 7967 7968 on - Enable TSX on the system. Although there are 7969 mitigations for all known security vulnerabilities, 7970 TSX has been known to be an accelerator for 7971 several previous speculation-related CVEs, and 7972 so there may be unknown security risks associated 7973 with leaving it enabled. 7974 7975 off - Disable TSX on the system. (Note that this 7976 option takes effect only on newer CPUs which are 7977 not vulnerable to MDS, i.e., have 7978 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get 7979 the new IA32_TSX_CTRL MSR through a microcode 7980 update. This new MSR allows for the reliable 7981 deactivation of the TSX functionality.) 7982 7983 auto - Disable TSX if X86_BUG_TAA is present, 7984 otherwise enable TSX on the system. 7985 7986 Not specifying this option is equivalent to tsx=off. 7987 7988 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 7989 for more details. 7990 7991 tsx_async_abort= [X86,INTEL,EARLY] Control mitigation for the TSX Async 7992 Abort (TAA) vulnerability. 7993 7994 Similar to Micro-architectural Data Sampling (MDS) 7995 certain CPUs that support Transactional 7996 Synchronization Extensions (TSX) are vulnerable to an 7997 exploit against CPU internal buffers which can forward 7998 information to a disclosure gadget under certain 7999 conditions. 8000 8001 In vulnerable processors, the speculatively forwarded 8002 data can be used in a cache side channel attack, to 8003 access data to which the attacker does not have direct 8004 access. 8005 8006 This parameter controls the TAA mitigation. The 8007 options are: 8008 8009 full - Enable TAA mitigation on vulnerable CPUs 8010 if TSX is enabled. 8011 8012 full,nosmt - Enable TAA mitigation and disable SMT on 8013 vulnerable CPUs. If TSX is disabled, SMT 8014 is not disabled because CPU is not 8015 vulnerable to cross-thread TAA attacks. 8016 off - Unconditionally disable TAA mitigation 8017 8018 On MDS-affected machines, tsx_async_abort=off can be 8019 prevented by an active MDS mitigation as both vulnerabilities 8020 are mitigated with the same mechanism so in order to disable 8021 this mitigation, you need to specify mds=off too. 8022 8023 Not specifying this option is equivalent to 8024 tsx_async_abort=full. On CPUs which are MDS affected 8025 and deploy MDS mitigation, TAA mitigation is not 8026 required and doesn't provide any additional 8027 mitigation. 8028 8029 For details see: 8030 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 8031 8032 turbografx.map[2|3]= [HW,JOY] 8033 TurboGraFX parallel port interface 8034 Format: 8035 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7> 8036 See also Documentation/input/devices/joystick-parport.rst 8037 8038 udbg-immortal [PPC] When debugging early kernel crashes that 8039 happen after console_init() and before a proper 8040 console driver takes over, this boot options might 8041 help "seeing" what's going on. 8042 8043 uhash_entries= [KNL,NET] 8044 Set number of hash buckets for UDP/UDP-Lite connections 8045 8046 uhci-hcd.ignore_oc= 8047 [USB] Ignore overcurrent events (default N). 8048 Some badly-designed motherboards generate lots of 8049 bogus events, for ports that aren't wired to 8050 anything. Set this parameter to avoid log spamming. 8051 Note that genuine overcurrent events won't be 8052 reported either. 8053 8054 unaligned_scalar_speed= 8055 [RISCV] 8056 Format: {slow | fast | unsupported} 8057 Allow skipping scalar unaligned access speed tests. This 8058 is useful for testing alternative code paths and to skip 8059 the tests in environments where they run too slowly. All 8060 CPUs must have the same scalar unaligned access speed. 8061 8062 unaligned_vector_speed= 8063 [RISCV] 8064 Format: {slow | fast | unsupported} 8065 Allow skipping vector unaligned access speed tests. This 8066 is useful for testing alternative code paths and to skip 8067 the tests in environments where they run too slowly. All 8068 CPUs must have the same vector unaligned access speed. 8069 8070 unknown_nmi_panic 8071 [X86] Cause panic on unknown NMI. 8072 8073 unwind_debug [X86-64,EARLY] 8074 Enable unwinder debug output. This can be 8075 useful for debugging certain unwinder error 8076 conditions, including corrupt stacks and 8077 bad/missing unwinder metadata. 8078 8079 usbcore.authorized_default= 8080 [USB] Default USB device authorization: 8081 (default -1 = authorized (same as 1), 8082 0 = not authorized, 1 = authorized, 2 = authorized 8083 if device connected to internal port) 8084 8085 usbcore.autosuspend= 8086 [USB] The autosuspend time delay (in seconds) used 8087 for newly-detected USB devices (default 2). This 8088 is the time required before an idle device will be 8089 autosuspended. Devices for which the delay is set 8090 to a negative value won't be autosuspended at all. 8091 8092 usbcore.usbfs_snoop= 8093 [USB] Set to log all usbfs traffic (default 0 = off). 8094 8095 usbcore.usbfs_snoop_max= 8096 [USB] Maximum number of bytes to snoop in each URB 8097 (default = 65536). 8098 8099 usbcore.blinkenlights= 8100 [USB] Set to cycle leds on hubs (default 0 = off). 8101 8102 usbcore.old_scheme_first= 8103 [USB] Start with the old device initialization 8104 scheme (default 0 = off). 8105 8106 usbcore.usbfs_memory_mb= 8107 [USB] Memory limit (in MB) for buffers allocated by 8108 usbfs (default = 16, 0 = max = 2047). 8109 8110 usbcore.use_both_schemes= 8111 [USB] Try the other device initialization scheme 8112 if the first one fails (default 1 = enabled). 8113 8114 usbcore.initial_descriptor_timeout= 8115 [USB] Specifies timeout for the initial 64-byte 8116 USB_REQ_GET_DESCRIPTOR request in milliseconds 8117 (default 5000 = 5.0 seconds). 8118 8119 usbcore.nousb [USB] Disable the USB subsystem 8120 8121 usbcore.quirks= 8122 [USB] A list of quirk entries to augment the built-in 8123 usb core quirk list. List entries are separated by 8124 commas. Each entry has the form 8125 VendorID:ProductID:Flags. The IDs are 4-digit hex 8126 numbers and Flags is a set of letters. Each letter 8127 will change the built-in quirk; setting it if it is 8128 clear and clearing it if it is set. The letters have 8129 the following meanings: 8130 a = USB_QUIRK_STRING_FETCH_255 (string 8131 descriptors must not be fetched using 8132 a 255-byte read); 8133 b = USB_QUIRK_RESET_RESUME (device can't resume 8134 correctly so reset it instead); 8135 c = USB_QUIRK_NO_SET_INTF (device can't handle 8136 Set-Interface requests); 8137 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't 8138 handle its Configuration or Interface 8139 strings); 8140 e = USB_QUIRK_RESET (device can't be reset 8141 (e.g morph devices), don't use reset); 8142 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has 8143 more interface descriptions than the 8144 bNumInterfaces count, and can't handle 8145 talking to these interfaces); 8146 g = USB_QUIRK_DELAY_INIT (device needs a pause 8147 during initialization, after we read 8148 the device descriptor); 8149 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For 8150 high speed and super speed interrupt 8151 endpoints, the USB 2.0 and USB 3.0 spec 8152 require the interval in microframes (1 8153 microframe = 125 microseconds) to be 8154 calculated as interval = 2 ^ 8155 (bInterval-1). 8156 Devices with this quirk report their 8157 bInterval as the result of this 8158 calculation instead of the exponent 8159 variable used in the calculation); 8160 i = USB_QUIRK_DEVICE_QUALIFIER (device can't 8161 handle device_qualifier descriptor 8162 requests); 8163 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device 8164 generates spurious wakeup, ignore 8165 remote wakeup capability); 8166 k = USB_QUIRK_NO_LPM (device can't handle Link 8167 Power Management); 8168 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL 8169 (Device reports its bInterval as linear 8170 frames instead of the USB 2.0 8171 calculation); 8172 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs 8173 to be disconnected before suspend to 8174 prevent spurious wakeup); 8175 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a 8176 pause after every control message); 8177 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra 8178 delay after resetting its port); 8179 p = USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT 8180 (Reduce timeout of the SET_ADDRESS 8181 request from 5000 ms to 500 ms); 8182 Example: quirks=0781:5580:bk,0a5c:5834:gij 8183 8184 usbhid.mousepoll= 8185 [USBHID] The interval which mice are to be polled at. 8186 8187 usbhid.jspoll= 8188 [USBHID] The interval which joysticks are to be polled at. 8189 8190 usbhid.kbpoll= 8191 [USBHID] The interval which keyboards are to be polled at. 8192 8193 usb-storage.delay_use= 8194 [UMS] The delay in seconds before a new device is 8195 scanned for Logical Units (default 1). 8196 Optionally the delay in milliseconds if the value has 8197 suffix with "ms". 8198 Example: delay_use=2567ms 8199 8200 usb-storage.quirks= 8201 [UMS] A list of quirks entries to supplement or 8202 override the built-in unusual_devs list. List 8203 entries are separated by commas. Each entry has 8204 the form VID:PID:Flags where VID and PID are Vendor 8205 and Product ID values (4-digit hex numbers) and 8206 Flags is a set of characters, each corresponding 8207 to a common usb-storage quirk flag as follows: 8208 a = SANE_SENSE (collect more than 18 bytes 8209 of sense data, not on uas); 8210 b = BAD_SENSE (don't collect more than 18 8211 bytes of sense data, not on uas); 8212 c = FIX_CAPACITY (decrease the reported 8213 device capacity by one sector); 8214 d = NO_READ_DISC_INFO (don't use 8215 READ_DISC_INFO command, not on uas); 8216 e = NO_READ_CAPACITY_16 (don't use 8217 READ_CAPACITY_16 command); 8218 f = NO_REPORT_OPCODES (don't use report opcodes 8219 command, uas only); 8220 g = MAX_SECTORS_240 (don't transfer more than 8221 240 sectors at a time, uas only); 8222 h = CAPACITY_HEURISTICS (decrease the 8223 reported device capacity by one 8224 sector if the number is odd); 8225 i = IGNORE_DEVICE (don't bind to this 8226 device); 8227 j = NO_REPORT_LUNS (don't use report luns 8228 command, uas only); 8229 k = NO_SAME (do not use WRITE_SAME, uas only) 8230 l = NOT_LOCKABLE (don't try to lock and 8231 unlock ejectable media, not on uas); 8232 m = MAX_SECTORS_64 (don't transfer more 8233 than 64 sectors = 32 KB at a time, 8234 not on uas); 8235 n = INITIAL_READ10 (force a retry of the 8236 initial READ(10) command, not on uas); 8237 o = CAPACITY_OK (accept the capacity 8238 reported by the device, not on uas); 8239 p = WRITE_CACHE (the device cache is ON 8240 by default, not on uas); 8241 r = IGNORE_RESIDUE (the device reports 8242 bogus residue values, not on uas); 8243 s = SINGLE_LUN (the device has only one 8244 Logical Unit); 8245 t = NO_ATA_1X (don't allow ATA(12) and ATA(16) 8246 commands, uas only); 8247 u = IGNORE_UAS (don't bind to the uas driver); 8248 w = NO_WP_DETECT (don't test whether the 8249 medium is write-protected). 8250 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE 8251 even if the device claims no cache, 8252 not on uas) 8253 Example: quirks=0419:aaf5:rl,0421:0433:rc 8254 8255 user_debug= [KNL,ARM] 8256 Format: <int> 8257 See arch/arm/Kconfig.debug help text. 8258 1 - undefined instruction events 8259 2 - system calls 8260 4 - invalid data aborts 8261 8 - SIGSEGV faults 8262 16 - SIGBUS faults 8263 Example: user_debug=31 8264 8265 vdso= [X86,SH,SPARC] 8266 On X86_32, this is an alias for vdso32=. Otherwise: 8267 8268 vdso=1: enable VDSO (the default) 8269 vdso=0: disable VDSO mapping 8270 8271 vdso32= [X86] Control the 32-bit vDSO 8272 vdso32=1: enable 32-bit VDSO 8273 vdso32=0 or vdso32=2: disable 32-bit VDSO 8274 8275 See the help text for CONFIG_COMPAT_VDSO for more 8276 details. If CONFIG_COMPAT_VDSO is set, the default is 8277 vdso32=0; otherwise, the default is vdso32=1. 8278 8279 For compatibility with older kernels, vdso32=2 is an 8280 alias for vdso32=0. 8281 8282 Try vdso32=0 if you encounter an error that says: 8283 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed! 8284 8285 video= [FB,EARLY] Frame buffer configuration 8286 See Documentation/fb/modedb.rst. 8287 8288 video.brightness_switch_enabled= [ACPI] 8289 Format: [0|1] 8290 If set to 1, on receiving an ACPI notify event 8291 generated by hotkey, video driver will adjust brightness 8292 level and then send out the event to user space through 8293 the allocated input device. If set to 0, video driver 8294 will only send out the event without touching backlight 8295 brightness level. 8296 default: 1 8297 8298 virtio_mmio.device= 8299 [VMMIO] Memory mapped virtio (platform) device. 8300 8301 <size>@<baseaddr>:<irq>[:<id>] 8302 where: 8303 <size> := size (can use standard suffixes 8304 like K, M and G) 8305 <baseaddr> := physical base address 8306 <irq> := interrupt number (as passed to 8307 request_irq()) 8308 <id> := (optional) platform device id 8309 example: 8310 virtio_mmio.device=1K@0x100b0000:48:7 8311 8312 Can be used multiple times for multiple devices. 8313 8314 vga= [BOOT,X86-32] Select a particular video mode 8315 See Documentation/arch/x86/boot.rst and 8316 Documentation/admin-guide/svga.rst. 8317 Use vga=ask for menu. 8318 This is actually a boot loader parameter; the value is 8319 passed to the kernel using a special protocol. 8320 8321 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y. 8322 May slow down system boot speed, especially when 8323 enabled on systems with a large amount of memory. 8324 All options are enabled by default, and this 8325 interface is meant to allow for selectively 8326 enabling or disabling specific virtual memory 8327 debugging features. 8328 8329 Available options are: 8330 P Enable page structure init time poisoning 8331 - Disable all of the above options 8332 8333 vmalloc=nn[KMG] [KNL,BOOT,EARLY] Forces the vmalloc area to have an 8334 exact size of <nn>. This can be used to increase 8335 the minimum size (128MB on x86, arm32 platforms). 8336 It can also be used to decrease the size and leave more room 8337 for directly mapped kernel RAM. Note that this parameter does 8338 not exist on many other platforms (including arm64, alpha, 8339 loongarch, arc, csky, hexagon, microblaze, mips, nios2, openrisc, 8340 parisc, m64k, powerpc, riscv, sh, um, xtensa, s390, sparc). 8341 8342 vmcp_cma=nn[MG] [KNL,S390,EARLY] 8343 Sets the memory size reserved for contiguous memory 8344 allocations for the vmcp device driver. 8345 8346 vmhalt= [KNL,S390] Perform z/VM CP command after system halt. 8347 Format: <command> 8348 8349 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic. 8350 Format: <command> 8351 8352 vmpoff= [KNL,S390] Perform z/VM CP command after power off. 8353 Format: <command> 8354 8355 vmscape= [X86] Controls mitigation for VMscape attacks. 8356 VMscape attacks can leak information from a userspace 8357 hypervisor to a guest via speculative side-channels. 8358 8359 off - disable the mitigation 8360 ibpb - use Indirect Branch Prediction Barrier 8361 (IBPB) mitigation (default) 8362 force - force vulnerability detection even on 8363 unaffected processors 8364 8365 vsyscall= [X86-64,EARLY] 8366 Controls the behavior of vsyscalls (i.e. calls to 8367 fixed addresses of 0xffffffffff600x00 from legacy 8368 code). Most statically-linked binaries and older 8369 versions of glibc use these calls. Because these 8370 functions are at fixed addresses, they make nice 8371 targets for exploits that can control RIP. 8372 8373 emulate Vsyscalls turn into traps and are emulated 8374 reasonably safely. The vsyscall page is 8375 readable. 8376 8377 xonly [default] Vsyscalls turn into traps and are 8378 emulated reasonably safely. The vsyscall 8379 page is not readable. 8380 8381 none Vsyscalls don't work at all. This makes 8382 them quite hard to use for exploits but 8383 might break your system. 8384 8385 vt.color= [VT] Default text color. 8386 Format: 0xYX, X = foreground, Y = background. 8387 Default: 0x07 = light gray on black. 8388 8389 vt.cur_default= [VT] Default cursor shape. 8390 Format: 0xCCBBAA, where AA, BB, and CC are the same as 8391 the parameters of the <Esc>[?A;B;Cc escape sequence; 8392 see vga-softcursor.rst. Default: 2 = underline. 8393 8394 vt.default_blu= [VT] 8395 Format: <blue0>,<blue1>,<blue2>,...,<blue15> 8396 Change the default blue palette of the console. 8397 This is a 16-member array composed of values 8398 ranging from 0-255. 8399 8400 vt.default_grn= [VT] 8401 Format: <green0>,<green1>,<green2>,...,<green15> 8402 Change the default green palette of the console. 8403 This is a 16-member array composed of values 8404 ranging from 0-255. 8405 8406 vt.default_red= [VT] 8407 Format: <red0>,<red1>,<red2>,...,<red15> 8408 Change the default red palette of the console. 8409 This is a 16-member array composed of values 8410 ranging from 0-255. 8411 8412 vt.default_utf8= 8413 [VT] 8414 Format=<0|1> 8415 Set system-wide default UTF-8 mode for all tty's. 8416 Default is 1, i.e. UTF-8 mode is enabled for all 8417 newly opened terminals. 8418 8419 vt.global_cursor_default= 8420 [VT] 8421 Format=<-1|0|1> 8422 Set system-wide default for whether a cursor 8423 is shown on new VTs. Default is -1, 8424 i.e. cursors will be created by default unless 8425 overridden by individual drivers. 0 will hide 8426 cursors, 1 will display them. 8427 8428 vt.italic= [VT] Default color for italic text; 0-15. 8429 Default: 2 = green. 8430 8431 vt.underline= [VT] Default color for underlined text; 0-15. 8432 Default: 3 = cyan. 8433 8434 watchdog timers [HW,WDT] For information on watchdog timers, 8435 see Documentation/watchdog/watchdog-parameters.rst 8436 or other driver-specific files in the 8437 Documentation/watchdog/ directory. 8438 8439 watchdog_thresh= 8440 [KNL] 8441 Set the hard lockup detector stall duration 8442 threshold in seconds. The soft lockup detector 8443 threshold is set to twice the value. A value of 0 8444 disables both lockup detectors. Default is 10 8445 seconds. 8446 8447 workqueue.unbound_cpus= 8448 [KNL,SMP] Specify to constrain one or some CPUs 8449 to use in unbound workqueues. 8450 Format: <cpu-list> 8451 By default, all online CPUs are available for 8452 unbound workqueues. 8453 8454 workqueue.watchdog_thresh= 8455 If CONFIG_WQ_WATCHDOG is configured, workqueue can 8456 warn stall conditions and dump internal state to 8457 help debugging. 0 disables workqueue stall 8458 detection; otherwise, it's the stall threshold 8459 duration in seconds. The default value is 30 and 8460 it can be updated at runtime by writing to the 8461 corresponding sysfs file. 8462 8463 workqueue.panic_on_stall=<uint> 8464 Panic when workqueue stall is detected by 8465 CONFIG_WQ_WATCHDOG. It sets the number times of the 8466 stall to trigger panic. 8467 8468 The default is set by CONFIG_BOOTPARAM_WQ_STALL_PANIC, 8469 which is 0 (disabled) if not configured. 8470 8471 workqueue.panic_on_stall_time=<uint> 8472 Panic when a workqueue stall has been continuous for 8473 the specified number of seconds. Unlike panic_on_stall 8474 which counts accumulated stall events, this triggers 8475 based on the duration of a single continuous stall. 8476 8477 The default is 0, which disables the time-based panic. 8478 8479 workqueue.cpu_intensive_thresh_us= 8480 Per-cpu work items which run for longer than this 8481 threshold are automatically considered CPU intensive 8482 and excluded from concurrency management to prevent 8483 them from noticeably delaying other per-cpu work 8484 items. Default is 10000 (10ms). 8485 8486 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 8487 will report the work functions which violate this 8488 threshold repeatedly. They are likely good 8489 candidates for using WQ_UNBOUND workqueues instead. 8490 8491 workqueue.cpu_intensive_warning_thresh=<uint> 8492 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 8493 will report the work functions which violate the 8494 intensive_threshold_us repeatedly. In order to prevent 8495 spurious warnings, start printing only after a work 8496 function has violated this threshold number of times. 8497 8498 The default is 4 times. 0 disables the warning. 8499 8500 workqueue.power_efficient 8501 Per-cpu workqueues are generally preferred because 8502 they show better performance thanks to cache 8503 locality; unfortunately, per-cpu workqueues tend to 8504 be more power hungry than unbound workqueues. 8505 8506 Enabling this makes the per-cpu workqueues which 8507 were observed to contribute significantly to power 8508 consumption unbound, leading to measurably lower 8509 power usage at the cost of small performance 8510 overhead. 8511 8512 The default value of this parameter is determined by 8513 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT. 8514 8515 workqueue.default_affinity_scope= 8516 Select the default affinity scope to use for unbound 8517 workqueues. Can be one of "cpu", "smt", "cache", 8518 "numa" and "system". Default is "cache". For more 8519 information, see the Affinity Scopes section in 8520 Documentation/core-api/workqueue.rst. 8521 8522 This can be changed after boot by writing to the 8523 matching /sys/module/workqueue/parameters file. All 8524 workqueues with the "default" affinity scope will be 8525 updated accordingly. 8526 8527 workqueue.debug_force_rr_cpu 8528 Workqueue used to implicitly guarantee that work 8529 items queued without explicit CPU specified are put 8530 on the local CPU. This guarantee is no longer true 8531 and while local CPU is still preferred work items 8532 may be put on foreign CPUs. This debug option 8533 forces round-robin CPU selection to flush out 8534 usages which depend on the now broken guarantee. 8535 When enabled, memory and cache locality will be 8536 impacted. 8537 8538 writecombine= [LOONGARCH,EARLY] Control the MAT (Memory Access 8539 Type) of ioremap_wc(). 8540 8541 on - Enable writecombine, use WUC for ioremap_wc() 8542 off - Disable writecombine, use SUC for ioremap_wc() 8543 8544 x2apic_phys [X86-64,APIC,EARLY] Use x2apic physical mode instead of 8545 default x2apic cluster mode on platforms 8546 supporting x2apic. 8547 8548 xen_512gb_limit [KNL,X86-64,XEN] 8549 Restricts the kernel running paravirtualized under Xen 8550 to use only up to 512 GB of RAM. The reason to do so is 8551 crash analysis tools and Xen tools for doing domain 8552 save/restore/migration must be enabled to handle larger 8553 domains. 8554 8555 xen_console_io [XEN,EARLY] 8556 Boolean option to enable/disable the usage of the Xen 8557 console_io hypercalls to read and write to the console. 8558 Mostly useful for debugging and development. 8559 8560 xen_emul_unplug= [HW,X86,XEN,EARLY] 8561 Unplug Xen emulated devices 8562 Format: [unplug0,][unplug1] 8563 ide-disks -- unplug primary master IDE devices 8564 aux-ide-disks -- unplug non-primary-master IDE devices 8565 nics -- unplug network devices 8566 all -- unplug all emulated devices (NICs and IDE disks) 8567 unnecessary -- unplugging emulated devices is 8568 unnecessary even if the host did not respond to 8569 the unplug protocol 8570 never -- do not unplug even if version check succeeds 8571 8572 xen_legacy_crash [X86,XEN,EARLY] 8573 Crash from Xen panic notifier, without executing late 8574 panic() code such as dumping handler. 8575 8576 xen_mc_debug [X86,XEN,EARLY] 8577 Enable multicall debugging when running as a Xen PV guest. 8578 Enabling this feature will reduce performance a little 8579 bit, so it should only be enabled for obtaining extended 8580 debug data in case of multicall errors. 8581 8582 xen_msr_safe= [X86,XEN,EARLY] 8583 Format: <bool> 8584 Select whether to always use non-faulting (safe) MSR 8585 access functions when running as Xen PV guest. The 8586 default value is controlled by CONFIG_XEN_PV_MSR_SAFE. 8587 8588 xen_nopv [X86] 8589 Disables the PV optimizations forcing the HVM guest to 8590 run as generic HVM guest with no PV drivers. 8591 This option is obsoleted by the "nopv" option, which 8592 has equivalent effect for XEN platform. 8593 8594 xen_no_vector_callback 8595 [KNL,X86,XEN,EARLY] Disable the vector callback for Xen 8596 event channel interrupts. 8597 8598 xen_scrub_pages= [XEN] 8599 Boolean option to control scrubbing pages before giving them back 8600 to Xen, for use by other domains. Can be also changed at runtime 8601 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages. 8602 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT. 8603 8604 xen_timer_slop= [X86-64,XEN,EARLY] 8605 Set the timer slop (in nanoseconds) for the virtual Xen 8606 timers (default is 100000). This adjusts the minimum 8607 delta of virtualized Xen timers, where lower values 8608 improve timer resolution at the expense of processing 8609 more timer interrupts. 8610 8611 xen.balloon_boot_timeout= [XEN] 8612 The time (in seconds) to wait before giving up to boot 8613 in case initial ballooning fails to free enough memory. 8614 Applies only when running as HVM or PVH guest and 8615 started with less memory configured than allowed at 8616 max. Default is 180. 8617 8618 xen.event_eoi_delay= [XEN] 8619 How long to delay EOI handling in case of event 8620 storms (jiffies). Default is 10. 8621 8622 xen.event_loop_timeout= [XEN] 8623 After which time (jiffies) the event handling loop 8624 should start to delay EOI handling. Default is 2. 8625 8626 xen.fifo_events= [XEN] 8627 Boolean parameter to disable using fifo event handling 8628 even if available. Normally fifo event handling is 8629 preferred over the 2-level event handling, as it is 8630 fairer and the number of possible event channels is 8631 much higher. Default is on (use fifo events). 8632 8633 xirc2ps_cs= [NET,PCMCIA] 8634 Format: 8635 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]] 8636 8637 xive= [PPC] 8638 By default on POWER9 and above, the kernel will 8639 natively use the XIVE interrupt controller. This option 8640 allows the fallback firmware mode to be used: 8641 8642 off Fallback to firmware control of XIVE interrupt 8643 controller on both pseries and powernv 8644 platforms. Only useful on POWER9 and above. 8645 8646 xive.store-eoi=off [PPC] 8647 By default on POWER10 and above, the kernel will use 8648 stores for EOI handling when the XIVE interrupt mode 8649 is active. This option allows the XIVE driver to use 8650 loads instead, as on POWER9. 8651 8652 xhci-hcd.quirks [USB,KNL] 8653 A hex value specifying bitmask with supplemental xhci 8654 host controller quirks. Meaning of each bit can be 8655 consulted in header drivers/usb/host/xhci.h. 8656 8657 xmon [PPC,EARLY] 8658 Format: { early | on | rw | ro | off } 8659 Controls if xmon debugger is enabled. Default is off. 8660 Passing only "xmon" is equivalent to "xmon=early". 8661 early Call xmon as early as possible on boot; xmon 8662 debugger is called from setup_arch(). 8663 on xmon debugger hooks will be installed so xmon 8664 is only called on a kernel crash. Default mode, 8665 i.e. either "ro" or "rw" mode, is controlled 8666 with CONFIG_XMON_DEFAULT_RO_MODE. 8667 rw xmon debugger hooks will be installed so xmon 8668 is called only on a kernel crash, mode is write, 8669 meaning SPR registers, memory and, other data 8670 can be written using xmon commands. 8671 ro same as "rw" option above but SPR registers, 8672 memory, and other data can't be written using 8673 xmon commands. 8674 off xmon is disabled. 8675