1 accept_memory= [MM] 2 Format: { eager | lazy } 3 default: lazy 4 By default, unaccepted memory is accepted lazily to 5 avoid prolonged boot times. The lazy option will add 6 some runtime overhead until all memory is eventually 7 accepted. In most cases the overhead is negligible. 8 For some workloads or for debugging purposes 9 accept_memory=eager can be used to accept all memory 10 at once during boot. 11 12 acpi= [HW,ACPI,X86,ARM64,RISCV64,EARLY] 13 Advanced Configuration and Power Interface 14 Format: { force | on | off | strict | noirq | rsdt | 15 copy_dsdt | nospcr } 16 force -- enable ACPI if default was off 17 on -- enable ACPI but allow fallback to DT [arm64,riscv64] 18 off -- disable ACPI if default was on 19 noirq -- do not use ACPI for IRQ routing 20 strict -- Be less tolerant of platforms that are not 21 strictly ACPI specification compliant. 22 rsdt -- prefer RSDT over (default) XSDT 23 copy_dsdt -- copy DSDT to memory 24 nocmcff -- Disable firmware first mode for corrected 25 errors. This disables parsing the HEST CMC error 26 source to check if firmware has set the FF flag. This 27 may result in duplicate corrected error reports. 28 nospcr -- disable console in ACPI SPCR table as 29 default _serial_ console on ARM64 30 For ARM64, ONLY "acpi=off", "acpi=on", "acpi=force" or 31 "acpi=nospcr" are available 32 For RISCV64, ONLY "acpi=off", "acpi=on" or "acpi=force" 33 are available 34 35 See also Documentation/power/runtime_pm.rst, pci=noacpi 36 37 acpi_apic_instance= [ACPI,IOAPIC,EARLY] 38 Format: <int> 39 2: use 2nd APIC table, if available 40 1,0: use 1st APIC table 41 default: 0 42 43 acpi_backlight= [HW,ACPI] 44 { vendor | video | native | none } 45 If set to vendor, prefer vendor-specific driver 46 (e.g. thinkpad_acpi, sony_acpi, etc.) instead 47 of the ACPI video.ko driver. 48 If set to video, use the ACPI video.ko driver. 49 If set to native, use the device's native backlight mode. 50 If set to none, disable the ACPI backlight interface. 51 52 acpi_force_32bit_fadt_addr [ACPI,EARLY] 53 force FADT to use 32 bit addresses rather than the 54 64 bit X_* addresses. Some firmware have broken 64 55 bit addresses for force ACPI ignore these and use 56 the older legacy 32 bit addresses. 57 58 acpica_no_return_repair [HW, ACPI] 59 Disable AML predefined validation mechanism 60 This mechanism can repair the evaluation result to make 61 the return objects more ACPI specification compliant. 62 This option is useful for developers to identify the 63 root cause of an AML interpreter issue when the issue 64 has something to do with the repair mechanism. 65 66 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG] 67 acpi.debug_level= [HW,ACPI,ACPI_DEBUG] 68 Format: <int> 69 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI 70 debug output. Bits in debug_layer correspond to a 71 _COMPONENT in an ACPI source file, e.g., 72 #define _COMPONENT ACPI_EVENTS 73 Bits in debug_level correspond to a level in 74 ACPI_DEBUG_PRINT statements, e.g., 75 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ... 76 The debug_level mask defaults to "info". See 77 Documentation/firmware-guide/acpi/debug.rst for more information about 78 debug layers and levels. 79 80 Enable processor driver info messages: 81 acpi.debug_layer=0x20000000 82 Enable AML "Debug" output, i.e., stores to the Debug 83 object while interpreting AML: 84 acpi.debug_layer=0xffffffff acpi.debug_level=0x2 85 Enable all messages related to ACPI hardware: 86 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff 87 88 Some values produce so much output that the system is 89 unusable. The "log_buf_len" parameter may be useful 90 if you need to capture more output. 91 92 acpi_enforce_resources= [ACPI] 93 { strict | lax | no } 94 Check for resource conflicts between native drivers 95 and ACPI OperationRegions (SystemIO and SystemMemory 96 only). IO ports and memory declared in ACPI might be 97 used by the ACPI subsystem in arbitrary AML code and 98 can interfere with legacy drivers. 99 strict (default): access to resources claimed by ACPI 100 is denied; legacy drivers trying to access reserved 101 resources will fail to bind to device using them. 102 lax: access to resources claimed by ACPI is allowed; 103 legacy drivers trying to access reserved resources 104 will bind successfully but a warning message is logged. 105 no: ACPI OperationRegions are not marked as reserved, 106 no further checks are performed. 107 108 acpi_force_table_verification [HW,ACPI,EARLY] 109 Enable table checksum verification during early stage. 110 By default, this is disabled due to x86 early mapping 111 size limitation. 112 113 acpi_irq_balance [HW,ACPI] 114 ACPI will balance active IRQs 115 default in APIC mode 116 117 acpi_irq_nobalance [HW,ACPI] 118 ACPI will not move active IRQs (default) 119 default in PIC mode 120 121 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA 122 Format: <irq>,<irq>... 123 124 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for 125 use by PCI 126 Format: <irq>,<irq>... 127 128 acpi_mask_gpe= [HW,ACPI] 129 Due to the existence of _Lxx/_Exx, some GPEs triggered 130 by unsupported hardware/firmware features can result in 131 GPE floodings that cannot be automatically disabled by 132 the GPE dispatcher. 133 This facility can be used to prevent such uncontrolled 134 GPE floodings. 135 Format: <byte> or <bitmap-list> 136 137 acpi_no_auto_serialize [HW,ACPI] 138 Disable auto-serialization of AML methods 139 AML control methods that contain the opcodes to create 140 named objects will be marked as "Serialized" by the 141 auto-serialization feature. 142 This feature is enabled by default. 143 This option allows to turn off the feature. 144 145 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump 146 kernels. 147 148 acpi_no_static_ssdt [HW,ACPI,EARLY] 149 Disable installation of static SSDTs at early boot time 150 By default, SSDTs contained in the RSDT/XSDT will be 151 installed automatically and they will appear under 152 /sys/firmware/acpi/tables. 153 This option turns off this feature. 154 Note that specifying this option does not affect 155 dynamic table installation which will install SSDT 156 tables to /sys/firmware/acpi/tables/dynamic. 157 158 acpi_no_watchdog [HW,ACPI,WDT] 159 Ignore the ACPI-based watchdog interface (WDAT) and let 160 a native driver control the watchdog device instead. 161 162 acpi_rsdp= [ACPI,EFI,KEXEC,EARLY] 163 Pass the RSDP address to the kernel, mostly used 164 on machines running EFI runtime service to boot the 165 second kernel for kdump. 166 167 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS 168 Format: To spoof as Windows 98: ="Microsoft Windows" 169 170 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead 171 of 2 which is mandated by ACPI 6) as the supported ACPI 172 specification revision (when using this switch, it may 173 be necessary to carry out a cold reboot _twice_ in a 174 row to make it take effect on the platform firmware). 175 176 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings 177 acpi_osi="string1" # add string1 178 acpi_osi="!string2" # remove string2 179 acpi_osi=!* # remove all strings 180 acpi_osi=! # disable all built-in OS vendor 181 strings 182 acpi_osi=!! # enable all built-in OS vendor 183 strings 184 acpi_osi= # disable all strings 185 186 'acpi_osi=!' can be used in combination with single or 187 multiple 'acpi_osi="string1"' to support specific OS 188 vendor string(s). Note that such command can only 189 affect the default state of the OS vendor strings, thus 190 it cannot affect the default state of the feature group 191 strings and the current state of the OS vendor strings, 192 specifying it multiple times through kernel command line 193 is meaningless. This command is useful when one do not 194 care about the state of the feature group strings which 195 should be controlled by the OSPM. 196 Examples: 197 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent 198 to 'acpi_osi="Windows 2000" acpi_osi=!', they all 199 can make '_OSI("Windows 2000")' TRUE. 200 201 'acpi_osi=' cannot be used in combination with other 202 'acpi_osi=' command lines, the _OSI method will not 203 exist in the ACPI namespace. NOTE that such command can 204 only affect the _OSI support state, thus specifying it 205 multiple times through kernel command line is also 206 meaningless. 207 Examples: 208 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)' 209 FALSE. 210 211 'acpi_osi=!*' can be used in combination with single or 212 multiple 'acpi_osi="string1"' to support specific 213 string(s). Note that such command can affect the 214 current state of both the OS vendor strings and the 215 feature group strings, thus specifying it multiple times 216 through kernel command line is meaningful. But it may 217 still not able to affect the final state of a string if 218 there are quirks related to this string. This command 219 is useful when one want to control the state of the 220 feature group strings to debug BIOS issues related to 221 the OSPM features. 222 Examples: 223 1. 'acpi_osi="Module Device" acpi_osi=!*' can make 224 '_OSI("Module Device")' FALSE. 225 2. 'acpi_osi=!* acpi_osi="Module Device"' can make 226 '_OSI("Module Device")' TRUE. 227 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is 228 equivalent to 229 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"' 230 and 231 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!', 232 they all will make '_OSI("Windows 2000")' TRUE. 233 234 acpi_pm_good [X86] 235 Override the pmtimer bug detection: force the kernel 236 to assume that this machine's pmtimer latches its value 237 and always returns good values. 238 239 acpi_sci= [HW,ACPI,EARLY] ACPI System Control Interrupt trigger mode 240 Format: { level | edge | high | low } 241 242 acpi_skip_timer_override [HW,ACPI,EARLY] 243 Recognize and ignore IRQ0/pin2 Interrupt Override. 244 For broken nForce2 BIOS resulting in XT-PIC timer. 245 246 acpi_sleep= [HW,ACPI] Sleep options 247 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig, 248 s4_nohwsig, old_ordering, nonvs, 249 sci_force_enable, nobl } 250 See Documentation/power/video.rst for information on 251 s3_bios and s3_mode. 252 s3_beep is for debugging; it makes the PC's speaker beep 253 as soon as the kernel's real-mode entry point is called. 254 s4_hwsig causes the kernel to check the ACPI hardware 255 signature during resume from hibernation, and gracefully 256 refuse to resume if it has changed. This complies with 257 the ACPI specification but not with reality, since 258 Windows does not do this and many laptops do change it 259 on docking. So the default behaviour is to allow resume 260 and simply warn when the signature changes, unless the 261 s4_hwsig option is enabled. 262 s4_nohwsig prevents ACPI hardware signature from being 263 used (or even warned about) during resume. 264 old_ordering causes the ACPI 1.0 ordering of the _PTS 265 control method, with respect to putting devices into 266 low power states, to be enforced (the ACPI 2.0 ordering 267 of _PTS is used by default). 268 nonvs prevents the kernel from saving/restoring the 269 ACPI NVS memory during suspend/hibernation and resume. 270 sci_force_enable causes the kernel to set SCI_EN directly 271 on resume from S1/S3 (which is against the ACPI spec, 272 but some broken systems don't work without it). 273 nobl causes the internal blacklist of systems known to 274 behave incorrectly in some ways with respect to system 275 suspend and resume to be ignored (use wisely). 276 277 acpi_use_timer_override [HW,ACPI,EARLY] 278 Use timer override. For some broken Nvidia NF5 boards 279 that require a timer override, but don't have HPET 280 281 add_efi_memmap [EFI,X86,EARLY] Include EFI memory map in 282 kernel's map of available physical RAM. 283 284 agp= [AGP] 285 { off | try_unsupported } 286 off: disable AGP support 287 try_unsupported: try to drive unsupported chipsets 288 (may crash computer or cause data corruption) 289 290 ALSA [HW,ALSA] 291 See Documentation/sound/alsa-configuration.rst 292 293 alignment= [KNL,ARM] 294 Allow the default userspace alignment fault handler 295 behaviour to be specified. Bit 0 enables warnings, 296 bit 1 enables fixups, and bit 2 sends a segfault. 297 298 align_va_addr= [X86-64] 299 Align virtual addresses by clearing slice [14:12] when 300 allocating a VMA at process creation time. This option 301 gives you up to 3% performance improvement on AMD F15h 302 machines (where it is enabled by default) for a 303 CPU-intensive style benchmark, and it can vary highly in 304 a microbenchmark depending on workload and compiler. 305 306 32: only for 32-bit processes 307 64: only for 64-bit processes 308 on: enable for both 32- and 64-bit processes 309 off: disable for both 32- and 64-bit processes 310 311 alloc_snapshot [FTRACE] 312 Allocate the ftrace snapshot buffer on boot up when the 313 main buffer is allocated. This is handy if debugging 314 and you need to use tracing_snapshot() on boot up, and 315 do not want to use tracing_snapshot_alloc() as it needs 316 to be done where GFP_KERNEL allocations are allowed. 317 318 allow_mismatched_32bit_el0 [ARM64,EARLY] 319 Allow execve() of 32-bit applications and setting of the 320 PER_LINUX32 personality on systems where only a strict 321 subset of the CPUs support 32-bit EL0. When this 322 parameter is present, the set of CPUs supporting 32-bit 323 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0 324 and hot-unplug operations may be restricted. 325 326 See Documentation/arch/arm64/asymmetric-32bit.rst for more 327 information. 328 329 amd_iommu= [HW,X86-64] 330 Pass parameters to the AMD IOMMU driver in the system. 331 Possible values are: 332 fullflush - Deprecated, equivalent to iommu.strict=1 333 off - do not initialize any AMD IOMMU found in 334 the system 335 force_isolation - Force device isolation for all 336 devices. The IOMMU driver is not 337 allowed anymore to lift isolation 338 requirements as needed. This option 339 does not override iommu=pt 340 force_enable - Force enable the IOMMU on platforms known 341 to be buggy with IOMMU enabled. Use this 342 option with care. 343 pgtbl_v1 - Use v1 page table for DMA-API (Default). 344 pgtbl_v2 - Use v2 page table for DMA-API. 345 irtcachedis - Disable Interrupt Remapping Table (IRT) caching. 346 nohugepages - Limit page-sizes used for v1 page-tables 347 to 4 KiB. 348 v2_pgsizes_only - Limit page-sizes used for v1 page-tables 349 to 4KiB/2Mib/1GiB. 350 351 352 amd_iommu_dump= [HW,X86-64] 353 Enable AMD IOMMU driver option to dump the ACPI table 354 for AMD IOMMU. With this option enabled, AMD IOMMU 355 driver will print ACPI tables for AMD IOMMU during 356 IOMMU initialization. 357 358 amd_iommu_intr= [HW,X86-64] 359 Specifies one of the following AMD IOMMU interrupt 360 remapping modes: 361 legacy - Use legacy interrupt remapping mode. 362 vapic - Use virtual APIC mode, which allows IOMMU 363 to inject interrupts directly into guest. 364 This mode requires kvm-amd.avic=1. 365 (Default when IOMMU HW support is present.) 366 367 amd_pstate= [X86,EARLY] 368 disable 369 Do not enable amd_pstate as the default 370 scaling driver for the supported processors 371 passive 372 Use amd_pstate with passive mode as a scaling driver. 373 In this mode autonomous selection is disabled. 374 Driver requests a desired performance level and platform 375 tries to match the same performance level if it is 376 satisfied by guaranteed performance level. 377 active 378 Use amd_pstate_epp driver instance as the scaling driver, 379 driver provides a hint to the hardware if software wants 380 to bias toward performance (0x0) or energy efficiency (0xff) 381 to the CPPC firmware. then CPPC power algorithm will 382 calculate the runtime workload and adjust the realtime cores 383 frequency. 384 guided 385 Activate guided autonomous mode. Driver requests minimum and 386 maximum performance level and the platform autonomously 387 selects a performance level in this range and appropriate 388 to the current workload. 389 390 amd_prefcore= 391 [X86] 392 disable 393 Disable amd-pstate preferred core. 394 395 amijoy.map= [HW,JOY] Amiga joystick support 396 Map of devices attached to JOY0DAT and JOY1DAT 397 Format: <a>,<b> 398 See also Documentation/input/joydev/joystick.rst 399 400 analog.map= [HW,JOY] Analog joystick and gamepad support 401 Specifies type or capabilities of an analog joystick 402 connected to one of 16 gameports 403 Format: <type1>,<type2>,..<type16> 404 405 apc= [HW,SPARC] 406 Power management functions (SPARCstation-4/5 + deriv.) 407 Format: noidle 408 Disable APC CPU standby support. SPARCstation-Fox does 409 not play well with APC CPU idle - disable it if you have 410 APC and your system crashes randomly. 411 412 apic [APIC,X86-64] Use IO-APIC. Default. 413 414 apic= [APIC,X86,EARLY] Advanced Programmable Interrupt Controller 415 Change the output verbosity while booting 416 Format: { quiet (default) | verbose | debug } 417 Change the amount of debugging information output 418 when initialising the APIC and IO-APIC components. 419 420 apic_extnmi= [APIC,X86,EARLY] External NMI delivery setting 421 Format: { bsp (default) | all | none } 422 bsp: External NMI is delivered only to CPU 0 423 all: External NMIs are broadcast to all CPUs as a 424 backup of CPU 0 425 none: External NMI is masked for all CPUs. This is 426 useful so that a dump capture kernel won't be 427 shot down by NMI 428 429 apicpmtimer Do APIC timer calibration using the pmtimer. Implies 430 apicmaintimer. Useful when your PIT timer is totally 431 broken. 432 433 autoconf= [IPV6] 434 See Documentation/networking/ipv6.rst. 435 436 apm= [APM] Advanced Power Management 437 See header of arch/x86/kernel/apm_32.c. 438 439 apparmor= [APPARMOR] Disable or enable AppArmor at boot time 440 Format: { "0" | "1" } 441 See security/apparmor/Kconfig help text 442 0 -- disable. 443 1 -- enable. 444 Default value is set via kernel config option. 445 446 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards 447 Format: <io>,<irq>,<nodeID> 448 449 arm64.no32bit_el0 [ARM64] Unconditionally disable the execution of 450 32 bit applications. 451 452 arm64.nobti [ARM64] Unconditionally disable Branch Target 453 Identification support 454 455 arm64.nogcs [ARM64] Unconditionally disable Guarded Control Stack 456 support 457 458 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory 459 Set instructions support 460 461 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension 462 support 463 464 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication 465 support 466 467 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix 468 Extension support 469 470 arm64.nosve [ARM64] Unconditionally disable Scalable Vector 471 Extension support 472 473 ataflop= [HW,M68k] 474 475 atarimouse= [HW,MOUSE] Atari Mouse 476 477 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess, 478 EzKey and similar keyboards 479 480 atkbd.reset= [HW] Reset keyboard during initialization 481 482 atkbd.set= [HW] Select keyboard code set 483 Format: <int> (2 = AT (default), 3 = PS/2) 484 485 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar 486 keyboards 487 488 atkbd.softraw= [HW] Choose between synthetic and real raw mode 489 Format: <bool> (0 = real, 1 = synthetic (default)) 490 491 atkbd.softrepeat= [HW] 492 Use software keyboard repeat 493 494 audit= [KNL] Enable the audit sub-system 495 Format: { "0" | "1" | "off" | "on" } 496 0 | off - kernel audit is disabled and can not be 497 enabled until the next reboot 498 unset - kernel audit is initialized but disabled and 499 will be fully enabled by the userspace auditd. 500 1 | on - kernel audit is initialized and partially 501 enabled, storing at most audit_backlog_limit 502 messages in RAM until it is fully enabled by the 503 userspace auditd. 504 Default: unset 505 506 audit_backlog_limit= [KNL] Set the audit queue size limit. 507 Format: <int> (must be >=0) 508 Default: 64 509 510 bau= [X86_UV] Enable the BAU on SGI UV. The default 511 behavior is to disable the BAU (i.e. bau=0). 512 Format: { "0" | "1" } 513 0 - Disable the BAU. 514 1 - Enable the BAU. 515 unset - Disable the BAU. 516 517 baycom_epp= [HW,AX25] 518 Format: <io>,<mode> 519 520 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem 521 Format: <io>,<mode> 522 See header of drivers/net/hamradio/baycom_par.c. 523 524 baycom_ser_fdx= [HW,AX25] 525 BayCom Serial Port AX.25 Modem (Full Duplex Mode) 526 Format: <io>,<irq>,<mode>[,<baud>] 527 See header of drivers/net/hamradio/baycom_ser_fdx.c. 528 529 baycom_ser_hdx= [HW,AX25] 530 BayCom Serial Port AX.25 Modem (Half Duplex Mode) 531 Format: <io>,<irq>,<mode> 532 See header of drivers/net/hamradio/baycom_ser_hdx.c. 533 534 bdev_allow_write_mounted= 535 Format: <bool> 536 Control the ability to open a mounted block device 537 for writing, i.e., allow / disallow writes that bypass 538 the FS. This was implemented as a means to prevent 539 fuzzers from crashing the kernel by overwriting the 540 metadata underneath a mounted FS without its awareness. 541 This also prevents destructive formatting of mounted 542 filesystems by naive storage tooling that don't use 543 O_EXCL. Default is Y and can be changed through the 544 Kconfig option CONFIG_BLK_DEV_WRITE_MOUNTED. 545 546 bert_disable [ACPI] 547 Disable BERT OS support on buggy BIOSes. 548 549 bgrt_disable [ACPI,X86,EARLY] 550 Disable BGRT to avoid flickering OEM logo. 551 552 blkdevparts= Manual partition parsing of block device(s) for 553 embedded devices based on command line input. 554 See Documentation/block/cmdline-partition.rst 555 556 boot_delay= [KNL,EARLY] 557 Milliseconds to delay each printk during boot. 558 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled, 559 and you may also have to specify "lpj=". Boot_delay 560 values larger than 10 seconds (10000) are assumed 561 erroneous and ignored. 562 Format: integer 563 564 bootconfig [KNL,EARLY] 565 Extended command line options can be added to an initrd 566 and this will cause the kernel to look for it. 567 568 See Documentation/admin-guide/bootconfig.rst 569 570 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards) 571 bttv.radio= Most important insmod options are available as 572 kernel args too. 573 bttv.pll= See Documentation/admin-guide/media/bttv.rst 574 bttv.tuner= 575 576 bulk_remove=off [PPC] This parameter disables the use of the pSeries 577 firmware feature for flushing multiple hpte entries 578 at a time. 579 580 c101= [NET] Moxa C101 synchronous serial card 581 582 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection. 583 Sometimes CPU hardware bugs make them report the cache 584 size incorrectly. The kernel will attempt work arounds 585 to fix known problems, but for some CPUs it is not 586 possible to determine what the correct size should be. 587 This option provides an override for these situations. 588 589 carrier_timeout= 590 [NET] Specifies amount of time (in seconds) that 591 the kernel should wait for a network carrier. By default 592 it waits 120 seconds. 593 594 ca_keys= [KEYS] This parameter identifies a specific key(s) on 595 the system trusted keyring to be used for certificate 596 trust validation. 597 format: { id:<keyid> | builtin } 598 599 cca= [MIPS,EARLY] Override the kernel pages' cache coherency 600 algorithm. Accepted values range from 0 to 7 601 inclusive. See arch/mips/include/asm/pgtable-bits.h 602 for platform specific values (SB1, Loongson3 and 603 others). 604 605 ccw_timeout_log [S390] 606 See Documentation/arch/s390/common_io.rst for details. 607 608 cgroup_disable= [KNL] Disable a particular controller or optional feature 609 Format: {name of the controller(s) or feature(s) to disable} 610 The effects of cgroup_disable=foo are: 611 - foo isn't auto-mounted if you mount all cgroups in 612 a single hierarchy 613 - foo isn't visible as an individually mountable 614 subsystem 615 - if foo is an optional feature then the feature is 616 disabled and corresponding cgroup files are not 617 created 618 {Currently only "memory" controller deal with this and 619 cut the overhead, others just disable the usage. So 620 only cgroup_disable=memory is actually worthy} 621 Specifying "pressure" disables per-cgroup pressure 622 stall information accounting feature 623 624 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1 625 Format: { { controller | "all" | "named" } 626 [,{ controller | "all" | "named" }...] } 627 Like cgroup_disable, but only applies to cgroup v1; 628 the blacklisted controllers remain available in cgroup2. 629 "all" blacklists all controllers and "named" disables 630 named mounts. Specifying both "all" and "named" disables 631 all v1 hierarchies. 632 633 cgroup_favordynmods= [KNL] Enable or Disable favordynmods. 634 Format: { "true" | "false" } 635 Defaults to the value of CONFIG_CGROUP_FAVOR_DYNMODS. 636 637 cgroup.memory= [KNL] Pass options to the cgroup memory controller. 638 Format: <string> 639 nosocket -- Disable socket memory accounting. 640 nokmem -- Disable kernel memory accounting. 641 nobpf -- Disable BPF memory accounting. 642 643 checkreqprot= [SELINUX] Set initial checkreqprot flag value. 644 Format: { "0" | "1" } 645 See security/selinux/Kconfig help text. 646 0 -- check protection applied by kernel (includes 647 any implied execute protection). 648 1 -- check protection requested by application. 649 Default value is set via a kernel config option. 650 Value can be changed at runtime via 651 /sys/fs/selinux/checkreqprot. 652 Setting checkreqprot to 1 is deprecated. 653 654 cio_ignore= [S390] 655 See Documentation/arch/s390/common_io.rst for details. 656 657 clearcpuid=X[,X...] [X86] 658 Disable CPUID feature X for the kernel. See 659 arch/x86/include/asm/cpufeatures.h for the valid bit 660 numbers X. Note the Linux-specific bits are not necessarily 661 stable over kernel options, but the vendor-specific 662 ones should be. 663 X can also be a string as appearing in the flags: line 664 in /proc/cpuinfo which does not have the above 665 instability issue. However, not all features have names 666 in /proc/cpuinfo. 667 Note that using this option will taint your kernel. 668 Also note that user programs calling CPUID directly 669 or using the feature without checking anything 670 will still see it. This just prevents it from 671 being used by the kernel or shown in /proc/cpuinfo. 672 Also note the kernel might malfunction if you disable 673 some critical bits. 674 675 clk_ignore_unused 676 [CLK] 677 Prevents the clock framework from automatically gating 678 clocks that have not been explicitly enabled by a Linux 679 device driver but are enabled in hardware at reset or 680 by the bootloader/firmware. Note that this does not 681 force such clocks to be always-on nor does it reserve 682 those clocks in any way. This parameter is useful for 683 debug and development, but should not be needed on a 684 platform with proper driver support. For more 685 information, see Documentation/driver-api/clk.rst. 686 687 clock= [BUGS=X86-32, HW] gettimeofday clocksource override. 688 [Deprecated] 689 Forces specified clocksource (if available) to be used 690 when calculating gettimeofday(). If specified 691 clocksource is not available, it defaults to PIT. 692 Format: { pit | tsc | cyclone | pmtmr } 693 694 clocksource= Override the default clocksource 695 Format: <string> 696 Override the default clocksource and use the clocksource 697 with the name specified. 698 Some clocksource names to choose from, depending on 699 the platform: 700 [all] jiffies (this is the base, fallback clocksource) 701 [ACPI] acpi_pm 702 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2, 703 pxa_timer,timer3,32k_counter,timer0_1 704 [X86-32] pit,hpet,tsc; 705 scx200_hrt on Geode; cyclone on IBM x440 706 [MIPS] MIPS 707 [PARISC] cr16 708 [S390] tod 709 [SH] SuperH 710 [SPARC64] tick 711 [X86-64] hpet,tsc 712 713 clocksource.arm_arch_timer.evtstrm= 714 [ARM,ARM64,EARLY] 715 Format: <bool> 716 Enable/disable the eventstream feature of the ARM 717 architected timer so that code using WFE-based polling 718 loops can be debugged more effectively on production 719 systems. 720 721 clocksource.verify_n_cpus= [KNL] 722 Limit the number of CPUs checked for clocksources 723 marked with CLOCK_SOURCE_VERIFY_PERCPU that 724 are marked unstable due to excessive skew. 725 A negative value says to check all CPUs, while 726 zero says not to check any. Values larger than 727 nr_cpu_ids are silently truncated to nr_cpu_ids. 728 The actual CPUs are chosen randomly, with 729 no replacement if the same CPU is chosen twice. 730 731 clocksource-wdtest.holdoff= [KNL] 732 Set the time in seconds that the clocksource 733 watchdog test waits before commencing its tests. 734 Defaults to zero when built as a module and to 735 10 seconds when built into the kernel. 736 737 cma=nn[MG]@[start[MG][-end[MG]]] 738 [KNL,CMA,EARLY] 739 Sets the size of kernel global memory area for 740 contiguous memory allocations and optionally the 741 placement constraint by the physical address range of 742 memory allocations. A value of 0 disables CMA 743 altogether. For more information, see 744 kernel/dma/contiguous.c 745 746 cma_pernuma=nn[MG] 747 [KNL,CMA,EARLY] 748 Sets the size of kernel per-numa memory area for 749 contiguous memory allocations. A value of 0 disables 750 per-numa CMA altogether. And If this option is not 751 specified, the default value is 0. 752 With per-numa CMA enabled, DMA users on node nid will 753 first try to allocate buffer from the pernuma area 754 which is located in node nid, if the allocation fails, 755 they will fallback to the global default memory area. 756 757 numa_cma=<node>:nn[MG][,<node>:nn[MG]] 758 [KNL,CMA,EARLY] 759 Sets the size of kernel numa memory area for 760 contiguous memory allocations. It will reserve CMA 761 area for the specified node. 762 763 With numa CMA enabled, DMA users on node nid will 764 first try to allocate buffer from the numa area 765 which is located in node nid, if the allocation fails, 766 they will fallback to the global default memory area. 767 768 cmo_free_hint= [PPC] Format: { yes | no } 769 Specify whether pages are marked as being inactive 770 when they are freed. This is used in CMO environments 771 to determine OS memory pressure for page stealing by 772 a hypervisor. 773 Default: yes 774 775 coherent_pool=nn[KMG] [ARM,KNL,EARLY] 776 Sets the size of memory pool for coherent, atomic dma 777 allocations, by default set to 256K. 778 779 com20020= [HW,NET] ARCnet - COM20020 chipset 780 Format: 781 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]] 782 783 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers) 784 Format: <io>[,<irq>] 785 786 com90xx= [HW,NET] 787 ARCnet - COM90xx chipset (memory-mapped buffers) 788 Format: <io>[,<irq>[,<memstart>]] 789 790 condev= [HW,S390] console device 791 conmode= 792 793 con3215_drop= [S390,EARLY] 3215 console drop mode. 794 Format: y|n|Y|N|1|0 795 When set to true, drop data on the 3215 console when 796 the console buffer is full. In this case the 797 operator using a 3270 terminal emulator (for example 798 x3270) does not have to enter the clear key for the 799 console output to advance and the kernel to continue. 800 This leads to a much faster boot time when a 3270 801 terminal emulator is active. If no 3270 terminal 802 emulator is used, this parameter has no effect. 803 804 console= [KNL] Output console device and options. 805 806 tty<n> Use the virtual console device <n>. 807 808 ttyS<n>[,options] 809 ttyUSB0[,options] 810 Use the specified serial port. The options are of 811 the form "bbbbpnf", where "bbbb" is the baud rate, 812 "p" is parity ("n", "o", or "e"), "n" is number of 813 bits, and "f" is flow control ("r" for RTS or 814 omit it). Default is "9600n8". 815 816 See Documentation/admin-guide/serial-console.rst for more 817 information. See 818 Documentation/networking/netconsole.rst for an 819 alternative. 820 821 <DEVNAME>:<n>.<n>[,options] 822 Use the specified serial port on the serial core bus. 823 The addressing uses DEVNAME of the physical serial port 824 device, followed by the serial core controller instance, 825 and the serial port instance. The options are the same 826 as documented for the ttyS addressing above. 827 828 The mapping of the serial ports to the tty instances 829 can be viewed with: 830 831 $ ls -d /sys/bus/serial-base/devices/*:*.*/tty/* 832 /sys/bus/serial-base/devices/00:04:0.0/tty/ttyS0 833 834 In the above example, the console can be addressed with 835 console=00:04:0.0. Note that a console addressed this 836 way will only get added when the related device driver 837 is ready. The use of an earlycon parameter in addition to 838 the console may be desired for console output early on. 839 840 uart[8250],io,<addr>[,options] 841 uart[8250],mmio,<addr>[,options] 842 uart[8250],mmio16,<addr>[,options] 843 uart[8250],mmio32,<addr>[,options] 844 uart[8250],0x<addr>[,options] 845 Start an early, polled-mode console on the 8250/16550 846 UART at the specified I/O port or MMIO address, 847 switching to the matching ttyS device later. 848 MMIO inter-register address stride is either 8-bit 849 (mmio), 16-bit (mmio16), or 32-bit (mmio32). 850 If none of [io|mmio|mmio16|mmio32], <addr> is assumed 851 to be equivalent to 'mmio'. 'options' are specified in 852 the same format described for ttyS above; if unspecified, 853 the h/w is not re-initialized. 854 855 hvc<n> Use the hypervisor console device <n>. This is for 856 both Xen and PowerPC hypervisors. 857 858 { null | "" } 859 Use to disable console output, i.e., to have kernel 860 console messages discarded. 861 This must be the only console= parameter used on the 862 kernel command line. 863 864 If the device connected to the port is not a TTY but a braille 865 device, prepend "brl," before the device type, for instance 866 console=brl,ttyS0 867 For now, only VisioBraille is supported. 868 869 console_msg_format= 870 [KNL] Change console messages format 871 default 872 By default we print messages on consoles in 873 "[time stamp] text\n" format (time stamp may not be 874 printed, depending on CONFIG_PRINTK_TIME or 875 `printk_time' param). 876 syslog 877 Switch to syslog format: "<%u>[time stamp] text\n" 878 IOW, each message will have a facility and loglevel 879 prefix. The format is similar to one used by syslog() 880 syscall, or to executing "dmesg -S --raw" or to reading 881 from /proc/kmsg. 882 883 consoleblank= [KNL] The console blank (screen saver) timeout in 884 seconds. A value of 0 disables the blank timer. 885 Defaults to 0. 886 887 coredump_filter= 888 [KNL] Change the default value for 889 /proc/<pid>/coredump_filter. 890 See also Documentation/filesystems/proc.rst. 891 892 coresight_cpu_debug.enable 893 [ARM,ARM64] 894 Format: <bool> 895 Enable/disable the CPU sampling based debugging. 896 0: default value, disable debugging 897 1: enable debugging at boot time 898 899 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver 900 Format: 901 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>] 902 903 cpuidle.off=1 [CPU_IDLE] 904 disable the cpuidle sub-system 905 906 cpuidle.governor= 907 [CPU_IDLE] Name of the cpuidle governor to use. 908 909 cpufreq.off=1 [CPU_FREQ] 910 disable the cpufreq sub-system 911 912 cpufreq.default_governor= 913 [CPU_FREQ] Name of the default cpufreq governor or 914 policy to use. This governor must be registered in the 915 kernel before the cpufreq driver probes. 916 917 cpu_init_udelay=N 918 [X86,EARLY] Delay for N microsec between assert and de-assert 919 of APIC INIT to start processors. This delay occurs 920 on every CPU online, such as boot, and resume from suspend. 921 Default: 10000 922 923 cpuhp.parallel= 924 [SMP] Enable/disable parallel bringup of secondary CPUs 925 Format: <bool> 926 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise 927 the parameter has no effect. 928 929 crash_kexec_post_notifiers 930 Only jump to kdump kernel after running the panic 931 notifiers and dumping kmsg. This option increases 932 the risks of a kdump failure, since some panic 933 notifiers can make the crashed kernel more unstable. 934 In configurations where kdump may not be reliable, 935 running the panic notifiers could allow collecting 936 more data on dmesg, like stack traces from other CPUS 937 or extra data dumped by panic_print. Note that some 938 configurations enable this option unconditionally, 939 like Hyper-V, PowerPC (fadump) and AMD SEV-SNP. 940 941 crashkernel=size[KMG][@offset[KMG]] 942 [KNL,EARLY] Using kexec, Linux can switch to a 'crash kernel' 943 upon panic. This parameter reserves the physical 944 memory region [offset, offset + size] for that kernel 945 image. If '@offset' is omitted, then a suitable offset 946 is selected automatically. 947 [KNL, X86-64, ARM64, RISCV, LoongArch] Select a region 948 under 4G first, and fall back to reserve region above 949 4G when '@offset' hasn't been specified. 950 See Documentation/admin-guide/kdump/kdump.rst for further details. 951 952 crashkernel=range1:size1[,range2:size2,...][@offset] 953 [KNL] Same as above, but depends on the memory 954 in the running system. The syntax of range is 955 start-[end] where start and end are both 956 a memory unit (amount[KMG]). See also 957 Documentation/admin-guide/kdump/kdump.rst for an example. 958 959 crashkernel=size[KMG],high 960 [KNL, X86-64, ARM64, RISCV, LoongArch] range could be 961 above 4G. 962 Allow kernel to allocate physical memory region from top, 963 so could be above 4G if system have more than 4G ram 964 installed. Otherwise memory region will be allocated 965 below 4G, if available. 966 It will be ignored if crashkernel=X is specified. 967 crashkernel=size[KMG],low 968 [KNL, X86-64, ARM64, RISCV, LoongArch] range under 4G. 969 When crashkernel=X,high is passed, kernel could allocate 970 physical memory region above 4G, that cause second kernel 971 crash on system that require some amount of low memory, 972 e.g. swiotlb requires at least 64M+32K low memory, also 973 enough extra low memory is needed to make sure DMA buffers 974 for 32-bit devices won't run out. Kernel would try to allocate 975 default size of memory below 4G automatically. The default 976 size is platform dependent. 977 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB) 978 --> arm64: 128MiB 979 --> riscv: 128MiB 980 --> loongarch: 128MiB 981 This one lets the user specify own low range under 4G 982 for second kernel instead. 983 0: to disable low allocation. 984 It will be ignored when crashkernel=X,high is not used 985 or memory reserved is below 4G. 986 987 cryptomgr.notests 988 [KNL] Disable crypto self-tests 989 990 cs89x0_dma= [HW,NET] 991 Format: <dma> 992 993 cs89x0_media= [HW,NET] 994 Format: { rj45 | aui | bnc } 995 996 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU 997 function call handling. When switched on, 998 additional debug data is printed to the console 999 in case a hanging CPU is detected, and that 1000 CPU is pinged again in order to try to resolve 1001 the hang situation. The default value of this 1002 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT 1003 Kconfig option. 1004 1005 dasd= [HW,NET] 1006 See header of drivers/s390/block/dasd_devmap.c. 1007 1008 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port 1009 (one device per port) 1010 Format: <port#>,<type> 1011 See also Documentation/input/devices/joystick-parport.rst 1012 1013 debug [KNL,EARLY] Enable kernel debugging (events log level). 1014 1015 debug_boot_weak_hash 1016 [KNL,EARLY] Enable printing [hashed] pointers early in the 1017 boot sequence. If enabled, we use a weak hash instead 1018 of siphash to hash pointers. Use this option if you are 1019 seeing instances of '(___ptrval___)') and need to see a 1020 value (hashed pointer) instead. Cryptographically 1021 insecure, please do not use on production kernels. 1022 1023 debug_locks_verbose= 1024 [KNL] verbose locking self-tests 1025 Format: <int> 1026 Print debugging info while doing the locking API 1027 self-tests. 1028 Bitmask for the various LOCKTYPE_ tests. Defaults to 0 1029 (no extra messages), setting it to -1 (all bits set) 1030 will print _a_lot_ more information - normally only 1031 useful to lockdep developers. 1032 1033 debug_objects [KNL,EARLY] Enable object debugging 1034 1035 debug_guardpage_minorder= 1036 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this 1037 parameter allows control of the order of pages that will 1038 be intentionally kept free (and hence protected) by the 1039 buddy allocator. Bigger value increase the probability 1040 of catching random memory corruption, but reduce the 1041 amount of memory for normal system use. The maximum 1042 possible value is MAX_PAGE_ORDER/2. Setting this 1043 parameter to 1 or 2 should be enough to identify most 1044 random memory corruption problems caused by bugs in 1045 kernel or driver code when a CPU writes to (or reads 1046 from) a random memory location. Note that there exists 1047 a class of memory corruptions problems caused by buggy 1048 H/W or F/W or by drivers badly programming DMA 1049 (basically when memory is written at bus level and the 1050 CPU MMU is bypassed) which are not detectable by 1051 CONFIG_DEBUG_PAGEALLOC, hence this option will not 1052 help tracking down these problems. 1053 1054 debug_pagealloc= 1055 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this parameter 1056 enables the feature at boot time. By default, it is 1057 disabled and the system will work mostly the same as a 1058 kernel built without CONFIG_DEBUG_PAGEALLOC. 1059 Note: to get most of debug_pagealloc error reports, it's 1060 useful to also enable the page_owner functionality. 1061 on: enable the feature 1062 1063 debugfs= [KNL,EARLY] This parameter enables what is exposed to 1064 userspace and debugfs internal clients. 1065 Format: { on, no-mount, off } 1066 on: All functions are enabled. 1067 no-mount: 1068 Filesystem is not registered but kernel clients can 1069 access APIs and a crashkernel can be used to read 1070 its content. There is nothing to mount. 1071 off: Filesystem is not registered and clients 1072 get a -EPERM as result when trying to register files 1073 or directories within debugfs. 1074 This is equivalent of the runtime functionality if 1075 debugfs was not enabled in the kernel at all. 1076 Default value is set in build-time with a kernel configuration. 1077 1078 debugpat [X86] Enable PAT debugging 1079 1080 default_hugepagesz= 1081 [HW] The size of the default HugeTLB page. This is 1082 the size represented by the legacy /proc/ hugepages 1083 APIs. In addition, this is the default hugetlb size 1084 used for shmget(), mmap() and mounting hugetlbfs 1085 filesystems. If not specified, defaults to the 1086 architecture's default huge page size. Huge page 1087 sizes are architecture dependent. See also 1088 Documentation/admin-guide/mm/hugetlbpage.rst. 1089 Format: size[KMG] 1090 1091 deferred_probe_timeout= 1092 [KNL] Debugging option to set a timeout in seconds for 1093 deferred probe to give up waiting on dependencies to 1094 probe. Only specific dependencies (subsystems or 1095 drivers) that have opted in will be ignored. A timeout 1096 of 0 will timeout at the end of initcalls. If the time 1097 out hasn't expired, it'll be restarted by each 1098 successful driver registration. This option will also 1099 dump out devices still on the deferred probe list after 1100 retrying. 1101 1102 delayacct [KNL] Enable per-task delay accounting 1103 1104 dell_smm_hwmon.ignore_dmi= 1105 [HW] Continue probing hardware even if DMI data 1106 indicates that the driver is running on unsupported 1107 hardware. 1108 1109 dell_smm_hwmon.force= 1110 [HW] Activate driver even if SMM BIOS signature does 1111 not match list of supported models and enable otherwise 1112 blacklisted features. 1113 1114 dell_smm_hwmon.power_status= 1115 [HW] Report power status in /proc/i8k 1116 (disabled by default). 1117 1118 dell_smm_hwmon.restricted= 1119 [HW] Allow controlling fans only if SYS_ADMIN 1120 capability is set. 1121 1122 dell_smm_hwmon.fan_mult= 1123 [HW] Factor to multiply fan speed with. 1124 1125 dell_smm_hwmon.fan_max= 1126 [HW] Maximum configurable fan speed. 1127 1128 dfltcc= [HW,S390] 1129 Format: { on | off | def_only | inf_only | always } 1130 on: s390 zlib hardware support for compression on 1131 level 1 and decompression (default) 1132 off: No s390 zlib hardware support 1133 def_only: s390 zlib hardware support for deflate 1134 only (compression on level 1) 1135 inf_only: s390 zlib hardware support for inflate 1136 only (decompression) 1137 always: Same as 'on' but ignores the selected compression 1138 level always using hardware support (used for debugging) 1139 1140 dhash_entries= [KNL] 1141 Set number of hash buckets for dentry cache. 1142 1143 disable_1tb_segments [PPC,EARLY] 1144 Disables the use of 1TB hash page table segments. This 1145 causes the kernel to fall back to 256MB segments which 1146 can be useful when debugging issues that require an SLB 1147 miss to occur. 1148 1149 disable= [IPV6] 1150 See Documentation/networking/ipv6.rst. 1151 1152 disable_radix [PPC,EARLY] 1153 Disable RADIX MMU mode on POWER9 1154 1155 disable_tlbie [PPC] 1156 Disable TLBIE instruction. Currently does not work 1157 with KVM, with HASH MMU, or with coherent accelerators. 1158 1159 disable_ddw [PPC/PSERIES,EARLY] 1160 Disable Dynamic DMA Window support. Use this 1161 to workaround buggy firmware. 1162 1163 disable_ipv6= [IPV6] 1164 See Documentation/networking/ipv6.rst. 1165 1166 disable_mtrr_cleanup [X86,EARLY] 1167 The kernel tries to adjust MTRR layout from continuous 1168 to discrete, to make X server driver able to add WB 1169 entry later. This parameter disables that. 1170 1171 disable_mtrr_trim [X86, Intel and AMD only,EARLY] 1172 By default the kernel will trim any uncacheable 1173 memory out of your available memory pool based on 1174 MTRR settings. This parameter disables that behavior, 1175 possibly causing your machine to run very slowly. 1176 1177 disable_timer_pin_1 [X86,EARLY] 1178 Disable PIN 1 of APIC timer 1179 Can be useful to work around chipset bugs. 1180 1181 dis_ucode_ldr [X86] Disable the microcode loader. 1182 1183 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support, 1184 this option disables the debugging code at boot. 1185 1186 dma_debug_entries=<number> 1187 This option allows to tune the number of preallocated 1188 entries for DMA-API debugging code. One entry is 1189 required per DMA-API allocation. Use this if the 1190 DMA-API debugging code disables itself because the 1191 architectural default is too low. 1192 1193 dma_debug_driver=<driver_name> 1194 With this option the DMA-API debugging driver 1195 filter feature can be enabled at boot time. Just 1196 pass the driver to filter for as the parameter. 1197 The filter can be disabled or changed to another 1198 driver later using sysfs. 1199 1200 reg_file_data_sampling= 1201 [X86] Controls mitigation for Register File Data 1202 Sampling (RFDS) vulnerability. RFDS is a CPU 1203 vulnerability which may allow userspace to infer 1204 kernel data values previously stored in floating point 1205 registers, vector registers, or integer registers. 1206 RFDS only affects Intel Atom processors. 1207 1208 on: Turns ON the mitigation. 1209 off: Turns OFF the mitigation. 1210 1211 This parameter overrides the compile time default set 1212 by CONFIG_MITIGATION_RFDS. Mitigation cannot be 1213 disabled when other VERW based mitigations (like MDS) 1214 are enabled. In order to disable RFDS mitigation all 1215 VERW based mitigations need to be disabled. 1216 1217 For details see: 1218 Documentation/admin-guide/hw-vuln/reg-file-data-sampling.rst 1219 1220 driver_async_probe= [KNL] 1221 List of driver names to be probed asynchronously. * 1222 matches with all driver names. If * is specified, the 1223 rest of the listed driver names are those that will NOT 1224 match the *. 1225 Format: <driver_name1>,<driver_name2>... 1226 1227 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>] 1228 Broken monitors, graphic adapters, KVMs and EDIDless 1229 panels may send no or incorrect EDID data sets. 1230 This parameter allows to specify an EDID data sets 1231 in the /lib/firmware directory that are used instead. 1232 An EDID data set will only be used for a particular 1233 connector, if its name and a colon are prepended to 1234 the EDID name. Each connector may use a unique EDID 1235 data set by separating the files with a comma. An EDID 1236 data set with no connector name will be used for 1237 any connectors not explicitly specified. 1238 1239 dscc4.setup= [NET] 1240 1241 dt_cpu_ftrs= [PPC,EARLY] 1242 Format: {"off" | "known"} 1243 Control how the dt_cpu_ftrs device-tree binding is 1244 used for CPU feature discovery and setup (if it 1245 exists). 1246 off: Do not use it, fall back to legacy cpu table. 1247 known: Do not pass through unknown features to guests 1248 or userspace, only those that the kernel is aware of. 1249 1250 dump_apple_properties [X86] 1251 Dump name and content of EFI device properties on 1252 x86 Macs. Useful for driver authors to determine 1253 what data is available or for reverse-engineering. 1254 1255 dyndbg[="val"] [KNL,DYNAMIC_DEBUG] 1256 <module>.dyndbg[="val"] 1257 Enable debug messages at boot time. See 1258 Documentation/admin-guide/dynamic-debug-howto.rst 1259 for details. 1260 1261 early_ioremap_debug [KNL,EARLY] 1262 Enable debug messages in early_ioremap support. This 1263 is useful for tracking down temporary early mappings 1264 which are not unmapped. 1265 1266 earlycon= [KNL,EARLY] Output early console device and options. 1267 1268 When used with no options, the early console is 1269 determined by stdout-path property in device tree's 1270 chosen node or the ACPI SPCR table if supported by 1271 the platform. 1272 1273 cdns,<addr>[,options] 1274 Start an early, polled-mode console on a Cadence 1275 (xuartps) serial port at the specified address. Only 1276 supported option is baud rate. If baud rate is not 1277 specified, the serial port must already be setup and 1278 configured. 1279 1280 uart[8250],io,<addr>[,options[,uartclk]] 1281 uart[8250],mmio,<addr>[,options[,uartclk]] 1282 uart[8250],mmio32,<addr>[,options[,uartclk]] 1283 uart[8250],mmio32be,<addr>[,options[,uartclk]] 1284 uart[8250],0x<addr>[,options] 1285 Start an early, polled-mode console on the 8250/16550 1286 UART at the specified I/O port or MMIO address. 1287 MMIO inter-register address stride is either 8-bit 1288 (mmio) or 32-bit (mmio32 or mmio32be). 1289 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed 1290 to be equivalent to 'mmio'. 'options' are specified 1291 in the same format described for "console=ttyS<n>"; if 1292 unspecified, the h/w is not initialized. 'uartclk' is 1293 the uart clock frequency; if unspecified, it is set 1294 to 'BASE_BAUD' * 16. 1295 1296 pl011,<addr> 1297 pl011,mmio32,<addr> 1298 Start an early, polled-mode console on a pl011 serial 1299 port at the specified address. The pl011 serial port 1300 must already be setup and configured. Options are not 1301 yet supported. If 'mmio32' is specified, then only 1302 the driver will use only 32-bit accessors to read/write 1303 the device registers. 1304 1305 liteuart,<addr> 1306 Start an early console on a litex serial port at the 1307 specified address. The serial port must already be 1308 setup and configured. Options are not yet supported. 1309 1310 meson,<addr> 1311 Start an early, polled-mode console on a meson serial 1312 port at the specified address. The serial port must 1313 already be setup and configured. Options are not yet 1314 supported. 1315 1316 msm_serial,<addr> 1317 Start an early, polled-mode console on an msm serial 1318 port at the specified address. The serial port 1319 must already be setup and configured. Options are not 1320 yet supported. 1321 1322 msm_serial_dm,<addr> 1323 Start an early, polled-mode console on an msm serial 1324 dm port at the specified address. The serial port 1325 must already be setup and configured. Options are not 1326 yet supported. 1327 1328 owl,<addr> 1329 Start an early, polled-mode console on a serial port 1330 of an Actions Semi SoC, such as S500 or S900, at the 1331 specified address. The serial port must already be 1332 setup and configured. Options are not yet supported. 1333 1334 rda,<addr> 1335 Start an early, polled-mode console on a serial port 1336 of an RDA Micro SoC, such as RDA8810PL, at the 1337 specified address. The serial port must already be 1338 setup and configured. Options are not yet supported. 1339 1340 sbi 1341 Use RISC-V SBI (Supervisor Binary Interface) for early 1342 console. 1343 1344 smh Use ARM semihosting calls for early console. 1345 1346 s3c2410,<addr> 1347 s3c2412,<addr> 1348 s3c2440,<addr> 1349 s3c6400,<addr> 1350 s5pv210,<addr> 1351 exynos4210,<addr> 1352 Use early console provided by serial driver available 1353 on Samsung SoCs, requires selecting proper type and 1354 a correct base address of the selected UART port. The 1355 serial port must already be setup and configured. 1356 Options are not yet supported. 1357 1358 lantiq,<addr> 1359 Start an early, polled-mode console on a lantiq serial 1360 (lqasc) port at the specified address. The serial port 1361 must already be setup and configured. Options are not 1362 yet supported. 1363 1364 lpuart,<addr> 1365 lpuart32,<addr> 1366 Use early console provided by Freescale LP UART driver 1367 found on Freescale Vybrid and QorIQ LS1021A processors. 1368 A valid base address must be provided, and the serial 1369 port must already be setup and configured. 1370 1371 ec_imx21,<addr> 1372 ec_imx6q,<addr> 1373 Start an early, polled-mode, output-only console on the 1374 Freescale i.MX UART at the specified address. The UART 1375 must already be setup and configured. 1376 1377 ar3700_uart,<addr> 1378 Start an early, polled-mode console on the 1379 Armada 3700 serial port at the specified 1380 address. The serial port must already be setup 1381 and configured. Options are not yet supported. 1382 1383 qcom_geni,<addr> 1384 Start an early, polled-mode console on a Qualcomm 1385 Generic Interface (GENI) based serial port at the 1386 specified address. The serial port must already be 1387 setup and configured. Options are not yet supported. 1388 1389 efifb,[options] 1390 Start an early, unaccelerated console on the EFI 1391 memory mapped framebuffer (if available). On cache 1392 coherent non-x86 systems that use system memory for 1393 the framebuffer, pass the 'ram' option so that it is 1394 mapped with the correct attributes. 1395 1396 linflex,<addr> 1397 Use early console provided by Freescale LINFlexD UART 1398 serial driver for NXP S32V234 SoCs. A valid base 1399 address must be provided, and the serial port must 1400 already be setup and configured. 1401 1402 earlyprintk= [X86,SH,ARM,M68k,S390,UM,EARLY] 1403 earlyprintk=vga 1404 earlyprintk=sclp 1405 earlyprintk=xen 1406 earlyprintk=serial[,ttySn[,baudrate]] 1407 earlyprintk=serial[,0x...[,baudrate]] 1408 earlyprintk=ttySn[,baudrate] 1409 earlyprintk=dbgp[debugController#] 1410 earlyprintk=pciserial[,force],bus:device.function[,baudrate] 1411 earlyprintk=xdbc[xhciController#] 1412 earlyprintk=bios 1413 1414 earlyprintk is useful when the kernel crashes before 1415 the normal console is initialized. It is not enabled by 1416 default because it has some cosmetic problems. 1417 1418 Append ",keep" to not disable it when the real console 1419 takes over. 1420 1421 Only one of vga, serial, or usb debug port can 1422 be used at a time. 1423 1424 Currently only ttyS0 and ttyS1 may be specified by 1425 name. Other I/O ports may be explicitly specified 1426 on some architectures (x86 and arm at least) by 1427 replacing ttySn with an I/O port address, like this: 1428 earlyprintk=serial,0x1008,115200 1429 You can find the port for a given device in 1430 /proc/tty/driver/serial: 1431 2: uart:ST16650V2 port:00001008 irq:18 ... 1432 1433 Interaction with the standard serial driver is not 1434 very good. 1435 1436 The VGA output is eventually overwritten by 1437 the real console. 1438 1439 The xen option can only be used in Xen domains. 1440 1441 The sclp output can only be used on s390. 1442 1443 The bios output can only be used on SuperH. 1444 1445 The optional "force" to "pciserial" enables use of a 1446 PCI device even when its classcode is not of the 1447 UART class. 1448 1449 edac_report= [HW,EDAC] Control how to report EDAC event 1450 Format: {"on" | "off" | "force"} 1451 on: enable EDAC to report H/W event. May be overridden 1452 by other higher priority error reporting module. 1453 off: disable H/W event reporting through EDAC. 1454 force: enforce the use of EDAC to report H/W event. 1455 default: on. 1456 1457 edd= [EDD] 1458 Format: {"off" | "on" | "skip[mbr]"} 1459 1460 efi= [EFI,EARLY] 1461 Format: { "debug", "disable_early_pci_dma", 1462 "nochunk", "noruntime", "nosoftreserve", 1463 "novamap", "no_disable_early_pci_dma" } 1464 debug: enable misc debug output. 1465 disable_early_pci_dma: disable the busmaster bit on all 1466 PCI bridges while in the EFI boot stub. 1467 nochunk: disable reading files in "chunks" in the EFI 1468 boot stub, as chunking can cause problems with some 1469 firmware implementations. 1470 noruntime : disable EFI runtime services support 1471 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose) 1472 attribute may cause the kernel to reserve the 1473 memory range for a memory mapping driver to 1474 claim. Specify efi=nosoftreserve to disable this 1475 reservation and treat the memory by its base type 1476 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM"). 1477 novamap: do not call SetVirtualAddressMap(). 1478 no_disable_early_pci_dma: Leave the busmaster bit set 1479 on all PCI bridges while in the EFI boot stub 1480 1481 efi_no_storage_paranoia [EFI,X86,EARLY] 1482 Using this parameter you can use more than 50% of 1483 your efi variable storage. Use this parameter only if 1484 you are really sure that your UEFI does sane gc and 1485 fulfills the spec otherwise your board may brick. 1486 1487 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT 1488 that is to be dynamically loaded by Linux. If there are 1489 multiple variables with the same name but with different 1490 vendor GUIDs, all of them will be loaded. See 1491 Documentation/admin-guide/acpi/ssdt-overlays.rst for details. 1492 1493 1494 eisa_irq_edge= [PARISC,HW] 1495 See header of drivers/parisc/eisa.c. 1496 1497 ekgdboc= [X86,KGDB,EARLY] Allow early kernel console debugging 1498 Format: ekgdboc=kbd 1499 1500 This is designed to be used in conjunction with 1501 the boot argument: earlyprintk=vga 1502 1503 This parameter works in place of the kgdboc parameter 1504 but can only be used if the backing tty is available 1505 very early in the boot process. For early debugging 1506 via a serial port see kgdboc_earlycon instead. 1507 1508 elanfreq= [X86-32] 1509 See comment before function elanfreq_setup() in 1510 arch/x86/kernel/cpu/cpufreq/elanfreq.c. 1511 1512 elfcorehdr=[size[KMG]@]offset[KMG] [PPC,SH,X86,S390,EARLY] 1513 Specifies physical address of start of kernel core 1514 image elf header and optionally the size. Generally 1515 kexec loader will pass this option to capture kernel. 1516 See Documentation/admin-guide/kdump/kdump.rst for details. 1517 1518 enable_mtrr_cleanup [X86,EARLY] 1519 The kernel tries to adjust MTRR layout from continuous 1520 to discrete, to make X server driver able to add WB 1521 entry later. This parameter enables that. 1522 1523 enable_timer_pin_1 [X86] 1524 Enable PIN 1 of APIC timer 1525 Can be useful to work around chipset bugs 1526 (in particular on some ATI chipsets). 1527 The kernel tries to set a reasonable default. 1528 1529 enforcing= [SELINUX] Set initial enforcing status. 1530 Format: {"0" | "1"} 1531 See security/selinux/Kconfig help text. 1532 0 -- permissive (log only, no denials). 1533 1 -- enforcing (deny and log). 1534 Default value is 0. 1535 Value can be changed at runtime via 1536 /sys/fs/selinux/enforce. 1537 1538 erst_disable [ACPI] 1539 Disable Error Record Serialization Table (ERST) 1540 support. 1541 1542 ether= [HW,NET] Ethernet cards parameters 1543 This option is obsoleted by the "netdev=" option, which 1544 has equivalent usage. See its documentation for details. 1545 1546 evm= [EVM] 1547 Format: { "fix" } 1548 Permit 'security.evm' to be updated regardless of 1549 current integrity status. 1550 1551 early_page_ext [KNL,EARLY] Enforces page_ext initialization to earlier 1552 stages so cover more early boot allocations. 1553 Please note that as side effect some optimizations 1554 might be disabled to achieve that (e.g. parallelized 1555 memory initialization is disabled) so the boot process 1556 might take longer, especially on systems with a lot of 1557 memory. Available with CONFIG_PAGE_EXTENSION=y. 1558 1559 failslab= 1560 fail_usercopy= 1561 fail_page_alloc= 1562 fail_skb_realloc= 1563 fail_make_request=[KNL] 1564 General fault injection mechanism. 1565 Format: <interval>,<probability>,<space>,<times> 1566 See also Documentation/fault-injection/. 1567 1568 fb_tunnels= [NET] 1569 Format: { initns | none } 1570 See Documentation/admin-guide/sysctl/net.rst for 1571 fb_tunnels_only_for_init_ns 1572 1573 floppy= [HW] 1574 See Documentation/admin-guide/blockdev/floppy.rst. 1575 1576 forcepae [X86-32] 1577 Forcefully enable Physical Address Extension (PAE). 1578 Many Pentium M systems disable PAE but may have a 1579 functionally usable PAE implementation. 1580 Warning: use of this parameter will taint the kernel 1581 and may cause unknown problems. 1582 1583 fred= [X86-64] 1584 Enable/disable Flexible Return and Event Delivery. 1585 Format: { on | off } 1586 on: enable FRED when it's present. 1587 off: disable FRED, the default setting. 1588 1589 ftrace=[tracer] 1590 [FTRACE] will set and start the specified tracer 1591 as early as possible in order to facilitate early 1592 boot debugging. 1593 1594 ftrace_boot_snapshot 1595 [FTRACE] On boot up, a snapshot will be taken of the 1596 ftrace ring buffer that can be read at: 1597 /sys/kernel/tracing/snapshot. 1598 This is useful if you need tracing information from kernel 1599 boot up that is likely to be overridden by user space 1600 start up functionality. 1601 1602 Optionally, the snapshot can also be defined for a tracing 1603 instance that was created by the trace_instance= command 1604 line parameter. 1605 1606 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo 1607 1608 The above will cause the "foo" tracing instance to trigger 1609 a snapshot at the end of boot up. 1610 1611 ftrace_dump_on_oops[=2(orig_cpu) | =<instance>][,<instance> | 1612 ,<instance>=2(orig_cpu)] 1613 [FTRACE] will dump the trace buffers on oops. 1614 If no parameter is passed, ftrace will dump global 1615 buffers of all CPUs, if you pass 2 or orig_cpu, it 1616 will dump only the buffer of the CPU that triggered 1617 the oops, or the specific instance will be dumped if 1618 its name is passed. Multiple instance dump is also 1619 supported, and instances are separated by commas. Each 1620 instance supports only dump on CPU that triggered the 1621 oops by passing 2 or orig_cpu to it. 1622 1623 ftrace_dump_on_oops=foo=orig_cpu 1624 1625 The above will dump only the buffer of "foo" instance 1626 on CPU that triggered the oops. 1627 1628 ftrace_dump_on_oops,foo,bar=orig_cpu 1629 1630 The above will dump global buffer on all CPUs, the 1631 buffer of "foo" instance on all CPUs and the buffer 1632 of "bar" instance on CPU that triggered the oops. 1633 1634 ftrace_filter=[function-list] 1635 [FTRACE] Limit the functions traced by the function 1636 tracer at boot up. function-list is a comma-separated 1637 list of functions. This list can be changed at run 1638 time by the set_ftrace_filter file in the debugfs 1639 tracing directory. 1640 1641 ftrace_notrace=[function-list] 1642 [FTRACE] Do not trace the functions specified in 1643 function-list. This list can be changed at run time 1644 by the set_ftrace_notrace file in the debugfs 1645 tracing directory. 1646 1647 ftrace_graph_filter=[function-list] 1648 [FTRACE] Limit the top level callers functions traced 1649 by the function graph tracer at boot up. 1650 function-list is a comma-separated list of functions 1651 that can be changed at run time by the 1652 set_graph_function file in the debugfs tracing directory. 1653 1654 ftrace_graph_notrace=[function-list] 1655 [FTRACE] Do not trace from the functions specified in 1656 function-list. This list is a comma-separated list of 1657 functions that can be changed at run time by the 1658 set_graph_notrace file in the debugfs tracing directory. 1659 1660 ftrace_graph_max_depth=<uint> 1661 [FTRACE] Used with the function graph tracer. This is 1662 the max depth it will trace into a function. This value 1663 can be changed at run time by the max_graph_depth file 1664 in the tracefs tracing directory. default: 0 (no limit) 1665 1666 fw_devlink= [KNL,EARLY] Create device links between consumer and supplier 1667 devices by scanning the firmware to infer the 1668 consumer/supplier relationships. This feature is 1669 especially useful when drivers are loaded as modules as 1670 it ensures proper ordering of tasks like device probing 1671 (suppliers first, then consumers), supplier boot state 1672 clean up (only after all consumers have probed), 1673 suspend/resume & runtime PM (consumers first, then 1674 suppliers). 1675 Format: { off | permissive | on | rpm } 1676 off -- Don't create device links from firmware info. 1677 permissive -- Create device links from firmware info 1678 but use it only for ordering boot state clean 1679 up (sync_state() calls). 1680 on -- Create device links from firmware info and use it 1681 to enforce probe and suspend/resume ordering. 1682 rpm -- Like "on", but also use to order runtime PM. 1683 1684 fw_devlink.strict=<bool> 1685 [KNL,EARLY] Treat all inferred dependencies as mandatory 1686 dependencies. This only applies for fw_devlink=on|rpm. 1687 Format: <bool> 1688 1689 fw_devlink.sync_state = 1690 [KNL,EARLY] When all devices that could probe have finished 1691 probing, this parameter controls what to do with 1692 devices that haven't yet received their sync_state() 1693 calls. 1694 Format: { strict | timeout } 1695 strict -- Default. Continue waiting on consumers to 1696 probe successfully. 1697 timeout -- Give up waiting on consumers and call 1698 sync_state() on any devices that haven't yet 1699 received their sync_state() calls after 1700 deferred_probe_timeout has expired or by 1701 late_initcall() if !CONFIG_MODULES. 1702 1703 gamecon.map[2|3]= 1704 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad 1705 support via parallel port (up to 5 devices per port) 1706 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5> 1707 See also Documentation/input/devices/joystick-parport.rst 1708 1709 gamma= [HW,DRM] 1710 1711 gart_fix_e820= [X86-64,EARLY] disable the fix e820 for K8 GART 1712 Format: off | on 1713 default: on 1714 1715 gather_data_sampling= 1716 [X86,INTEL,EARLY] Control the Gather Data Sampling (GDS) 1717 mitigation. 1718 1719 Gather Data Sampling is a hardware vulnerability which 1720 allows unprivileged speculative access to data which was 1721 previously stored in vector registers. 1722 1723 This issue is mitigated by default in updated microcode. 1724 The mitigation may have a performance impact but can be 1725 disabled. On systems without the microcode mitigation 1726 disabling AVX serves as a mitigation. 1727 1728 force: Disable AVX to mitigate systems without 1729 microcode mitigation. No effect if the microcode 1730 mitigation is present. Known to cause crashes in 1731 userspace with buggy AVX enumeration. 1732 1733 off: Disable GDS mitigation. 1734 1735 gbpages [X86] Use GB pages for kernel direct mappings. 1736 1737 gcov_persist= [GCOV] When non-zero (default), profiling data for 1738 kernel modules is saved and remains accessible via 1739 debugfs, even when the module is unloaded/reloaded. 1740 When zero, profiling data is discarded and associated 1741 debugfs files are removed at module unload time. 1742 1743 goldfish [X86] Enable the goldfish android emulator platform. 1744 Don't use this when you are not running on the 1745 android emulator 1746 1747 gpio-mockup.gpio_mockup_ranges 1748 [HW] Sets the ranges of gpiochip of for this device. 1749 Format: <start1>,<end1>,<start2>,<end2>... 1750 gpio-mockup.gpio_mockup_named_lines 1751 [HW] Let the driver know GPIO lines should be named. 1752 1753 gpt [EFI] Forces disk with valid GPT signature but 1754 invalid Protective MBR to be treated as GPT. If the 1755 primary GPT is corrupted, it enables the backup/alternate 1756 GPT to be used instead. 1757 1758 grcan.enable0= [HW] Configuration of physical interface 0. Determines 1759 the "Enable 0" bit of the configuration register. 1760 Format: 0 | 1 1761 Default: 0 1762 grcan.enable1= [HW] Configuration of physical interface 1. Determines 1763 the "Enable 0" bit of the configuration register. 1764 Format: 0 | 1 1765 Default: 0 1766 grcan.select= [HW] Select which physical interface to use. 1767 Format: 0 | 1 1768 Default: 0 1769 grcan.txsize= [HW] Sets the size of the tx buffer. 1770 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0. 1771 Default: 1024 1772 grcan.rxsize= [HW] Sets the size of the rx buffer. 1773 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0. 1774 Default: 1024 1775 1776 hardened_usercopy= 1777 [KNL] Under CONFIG_HARDENED_USERCOPY, whether 1778 hardening is enabled for this boot. Hardened 1779 usercopy checking is used to protect the kernel 1780 from reading or writing beyond known memory 1781 allocation boundaries as a proactive defense 1782 against bounds-checking flaws in the kernel's 1783 copy_to_user()/copy_from_user() interface. 1784 The default is determined by 1785 CONFIG_HARDENED_USERCOPY_DEFAULT_ON. 1786 on Perform hardened usercopy checks. 1787 off Disable hardened usercopy checks. 1788 1789 hardlockup_all_cpu_backtrace= 1790 [KNL] Should the hard-lockup detector generate 1791 backtraces on all cpus. 1792 Format: 0 | 1 1793 1794 hashdist= [KNL,NUMA] Large hashes allocated during boot 1795 are distributed across NUMA nodes. Defaults on 1796 for 64-bit NUMA, off otherwise. 1797 Format: 0 | 1 (for off | on) 1798 1799 hd= [EIDE] (E)IDE hard drive subsystem geometry 1800 Format: <cyl>,<head>,<sect> 1801 1802 hest_disable [ACPI] 1803 Disable Hardware Error Source Table (HEST) support; 1804 corresponding firmware-first mode error processing 1805 logic will be disabled. 1806 1807 hibernate= [HIBERNATION] 1808 noresume Don't check if there's a hibernation image 1809 present during boot. 1810 nocompress Don't compress/decompress hibernation images. 1811 no Disable hibernation and resume. 1812 protect_image Turn on image protection during restoration 1813 (that will set all pages holding image data 1814 during restoration read-only). 1815 1816 hibernate.compressor= [HIBERNATION] Compression algorithm to be 1817 used with hibernation. 1818 Format: { lzo | lz4 } 1819 Default: lzo 1820 1821 lzo: Select LZO compression algorithm to 1822 compress/decompress hibernation image. 1823 1824 lz4: Select LZ4 compression algorithm to 1825 compress/decompress hibernation image. 1826 1827 highmem=nn[KMG] [KNL,BOOT,EARLY] forces the highmem zone to have an exact 1828 size of <nn>. This works even on boxes that have no 1829 highmem otherwise. This also works to reduce highmem 1830 size on bigger boxes. 1831 1832 highres= [KNL] Enable/disable high resolution timer mode. 1833 Valid parameters: "on", "off" 1834 Default: "on" 1835 1836 hlt [BUGS=ARM,SH] 1837 1838 hostname= [KNL,EARLY] Set the hostname (aka UTS nodename). 1839 Format: <string> 1840 This allows setting the system's hostname during early 1841 startup. This sets the name returned by gethostname. 1842 Using this parameter to set the hostname makes it 1843 possible to ensure the hostname is correctly set before 1844 any userspace processes run, avoiding the possibility 1845 that a process may call gethostname before the hostname 1846 has been explicitly set, resulting in the calling 1847 process getting an incorrect result. The string must 1848 not exceed the maximum allowed hostname length (usually 1849 64 characters) and will be truncated otherwise. 1850 1851 hpet= [X86-32,HPET] option to control HPET usage 1852 Format: { enable (default) | disable | force | 1853 verbose } 1854 disable: disable HPET and use PIT instead 1855 force: allow force enabled of undocumented chips (ICH4, 1856 VIA, nVidia) 1857 verbose: show contents of HPET registers during setup 1858 1859 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET 1860 registers. Default set by CONFIG_HPET_MMAP_DEFAULT. 1861 1862 hugepages= [HW] Number of HugeTLB pages to allocate at boot. 1863 If this follows hugepagesz (below), it specifies 1864 the number of pages of hugepagesz to be allocated. 1865 If this is the first HugeTLB parameter on the command 1866 line, it specifies the number of pages to allocate for 1867 the default huge page size. If using node format, the 1868 number of pages to allocate per-node can be specified. 1869 See also Documentation/admin-guide/mm/hugetlbpage.rst. 1870 Format: <integer> or (node format) 1871 <node>:<integer>[,<node>:<integer>] 1872 1873 hugepagesz= 1874 [HW] The size of the HugeTLB pages. This is used in 1875 conjunction with hugepages (above) to allocate huge 1876 pages of a specific size at boot. The pair 1877 hugepagesz=X hugepages=Y can be specified once for 1878 each supported huge page size. Huge page sizes are 1879 architecture dependent. See also 1880 Documentation/admin-guide/mm/hugetlbpage.rst. 1881 Format: size[KMG] 1882 1883 hugetlb_cma= [HW,CMA,EARLY] The size of a CMA area used for allocation 1884 of gigantic hugepages. Or using node format, the size 1885 of a CMA area per node can be specified. 1886 Format: nn[KMGTPE] or (node format) 1887 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]] 1888 1889 Reserve a CMA area of given size and allocate gigantic 1890 hugepages using the CMA allocator. If enabled, the 1891 boot-time allocation of gigantic hugepages is skipped. 1892 1893 hugetlb_free_vmemmap= 1894 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP 1895 enabled. 1896 Control if HugeTLB Vmemmap Optimization (HVO) is enabled. 1897 Allows heavy hugetlb users to free up some more 1898 memory (7 * PAGE_SIZE for each 2MB hugetlb page). 1899 Format: { on | off (default) } 1900 1901 on: enable HVO 1902 off: disable HVO 1903 1904 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y, 1905 the default is on. 1906 1907 Note that the vmemmap pages may be allocated from the added 1908 memory block itself when memory_hotplug.memmap_on_memory is 1909 enabled, those vmemmap pages cannot be optimized even if this 1910 feature is enabled. Other vmemmap pages not allocated from 1911 the added memory block itself do not be affected. 1912 1913 hung_task_panic= 1914 [KNL] Should the hung task detector generate panics. 1915 Format: 0 | 1 1916 1917 A value of 1 instructs the kernel to panic when a 1918 hung task is detected. The default value is controlled 1919 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time 1920 option. The value selected by this boot parameter can 1921 be changed later by the kernel.hung_task_panic sysctl. 1922 1923 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC) 1924 terminal devices. Valid values: 0..8 1925 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs. 1926 If specified, z/VM IUCV HVC accepts connections 1927 from listed z/VM user IDs only. 1928 1929 hv_nopvspin [X86,HYPER_V,EARLY] 1930 Disables the paravirt spinlock optimizations 1931 which allow the hypervisor to 'idle' the guest 1932 on lock contention. 1933 1934 i2c_bus= [HW] Override the default board specific I2C bus speed 1935 or register an additional I2C bus that is not 1936 registered from board initialization code. 1937 Format: 1938 <bus_id>,<clkrate> 1939 1940 i2c_touchscreen_props= [HW,ACPI,X86] 1941 Set device-properties for ACPI-enumerated I2C-attached 1942 touchscreen, to e.g. fix coordinates of upside-down 1943 mounted touchscreens. If you need this option please 1944 submit a drivers/platform/x86/touchscreen_dmi.c patch 1945 adding a DMI quirk for this. 1946 1947 Format: 1948 <ACPI_HW_ID>:<prop_name>=<val>[:prop_name=val][:...] 1949 Where <val> is one of: 1950 Omit "=<val>" entirely Set a boolean device-property 1951 Unsigned number Set a u32 device-property 1952 Anything else Set a string device-property 1953 1954 Examples (split over multiple lines): 1955 i2c_touchscreen_props=GDIX1001:touchscreen-inverted-x: 1956 touchscreen-inverted-y 1957 1958 i2c_touchscreen_props=MSSL1680:touchscreen-size-x=1920: 1959 touchscreen-size-y=1080:touchscreen-inverted-y: 1960 firmware-name=gsl1680-vendor-model.fw:silead,home-button 1961 1962 i8042.debug [HW] Toggle i8042 debug mode 1963 i8042.unmask_kbd_data 1964 [HW] Enable printing of interrupt data from the KBD port 1965 (disabled by default, and as a pre-condition 1966 requires that i8042.debug=1 be enabled) 1967 i8042.direct [HW] Put keyboard port into non-translated mode 1968 i8042.dumbkbd [HW] Pretend that controller can only read data from 1969 keyboard and cannot control its state 1970 (Don't attempt to blink the leds) 1971 i8042.noaux [HW] Don't check for auxiliary (== mouse) port 1972 i8042.nokbd [HW] Don't check/create keyboard port 1973 i8042.noloop [HW] Disable the AUX Loopback command while probing 1974 for the AUX port 1975 i8042.nomux [HW] Don't check presence of an active multiplexing 1976 controller 1977 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX 1978 controllers 1979 i8042.notimeout [HW] Ignore timeout condition signalled by controller 1980 i8042.reset [HW] Reset the controller during init, cleanup and 1981 suspend-to-ram transitions, only during s2r 1982 transitions, or never reset 1983 Format: { 1 | Y | y | 0 | N | n } 1984 1, Y, y: always reset controller 1985 0, N, n: don't ever reset controller 1986 Default: only on s2r transitions on x86; most other 1987 architectures force reset to be always executed 1988 i8042.unlock [HW] Unlock (ignore) the keylock 1989 i8042.kbdreset [HW] Reset device connected to KBD port 1990 i8042.probe_defer 1991 [HW] Allow deferred probing upon i8042 probe errors 1992 1993 i810= [HW,DRM] 1994 1995 i915.invert_brightness= 1996 [DRM] Invert the sense of the variable that is used to 1997 set the brightness of the panel backlight. Normally a 1998 brightness value of 0 indicates backlight switched off, 1999 and the maximum of the brightness value sets the backlight 2000 to maximum brightness. If this parameter is set to 0 2001 (default) and the machine requires it, or this parameter 2002 is set to 1, a brightness value of 0 sets the backlight 2003 to maximum brightness, and the maximum of the brightness 2004 value switches the backlight off. 2005 -1 -- never invert brightness 2006 0 -- machine default 2007 1 -- force brightness inversion 2008 2009 ia32_emulation= [X86-64] 2010 Format: <bool> 2011 When true, allows loading 32-bit programs and executing 32-bit 2012 syscalls, essentially overriding IA32_EMULATION_DEFAULT_DISABLED at 2013 boot time. When false, unconditionally disables IA32 emulation. 2014 2015 icn= [HW,ISDN] 2016 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]] 2017 2018 2019 idle= [X86,EARLY] 2020 Format: idle=poll, idle=halt, idle=nomwait 2021 2022 idle=poll: Don't do power saving in the idle loop 2023 using HLT, but poll for rescheduling event. This will 2024 make the CPUs eat a lot more power, but may be useful 2025 to get slightly better performance in multiprocessor 2026 benchmarks. It also makes some profiling using 2027 performance counters more accurate. Please note that 2028 on systems with MONITOR/MWAIT support (like Intel 2029 EM64T CPUs) this option has no performance advantage 2030 over the normal idle loop. It may also interact badly 2031 with hyperthreading. 2032 2033 idle=halt: Halt is forced to be used for CPU idle. 2034 In such case C2/C3 won't be used again. 2035 2036 idle=nomwait: Disable mwait for CPU C-states 2037 2038 idxd.sva= [HW] 2039 Format: <bool> 2040 Allow force disabling of Shared Virtual Memory (SVA) 2041 support for the idxd driver. By default it is set to 2042 true (1). 2043 2044 idxd.tc_override= [HW] 2045 Format: <bool> 2046 Allow override of default traffic class configuration 2047 for the device. By default it is set to false (0). 2048 2049 ieee754= [MIPS] Select IEEE Std 754 conformance mode 2050 Format: { strict | legacy | 2008 | relaxed | emulated } 2051 Default: strict 2052 2053 Choose which programs will be accepted for execution 2054 based on the IEEE 754 NaN encoding(s) supported by 2055 the FPU and the NaN encoding requested with the value 2056 of an ELF file header flag individually set by each 2057 binary. Hardware implementations are permitted to 2058 support either or both of the legacy and the 2008 NaN 2059 encoding mode. 2060 2061 Available settings are as follows: 2062 strict accept binaries that request a NaN encoding 2063 supported by the FPU 2064 legacy only accept legacy-NaN binaries, if supported 2065 by the FPU 2066 2008 only accept 2008-NaN binaries, if supported 2067 by the FPU 2068 relaxed accept any binaries regardless of whether 2069 supported by the FPU 2070 emulated accept any binaries but enable FPU emulator 2071 if binary mode is unsupported by the FPU. 2072 2073 The FPU emulator is always able to support both NaN 2074 encodings, so if no FPU hardware is present or it has 2075 been disabled with 'nofpu', then the settings of 2076 'legacy' and '2008' strap the emulator accordingly, 2077 'relaxed' straps the emulator for both legacy-NaN and 2078 2008-NaN, whereas 'strict' enables legacy-NaN only on 2079 legacy processors and both NaN encodings on MIPS32 or 2080 MIPS64 CPUs. 2081 2082 The setting for ABS.fmt/NEG.fmt instruction execution 2083 mode generally follows that for the NaN encoding, 2084 except where unsupported by hardware. 2085 2086 ignore_loglevel [KNL,EARLY] 2087 Ignore loglevel setting - this will print /all/ 2088 kernel messages to the console. Useful for debugging. 2089 We also add it as printk module parameter, so users 2090 could change it dynamically, usually by 2091 /sys/module/printk/parameters/ignore_loglevel. 2092 2093 ignore_rlimit_data 2094 Ignore RLIMIT_DATA setting for data mappings, 2095 print warning at first misuse. Can be changed via 2096 /sys/module/kernel/parameters/ignore_rlimit_data. 2097 2098 ihash_entries= [KNL] 2099 Set number of hash buckets for inode cache. 2100 2101 ima_appraise= [IMA] appraise integrity measurements 2102 Format: { "off" | "enforce" | "fix" | "log" } 2103 default: "enforce" 2104 2105 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead. 2106 The builtin appraise policy appraises all files 2107 owned by uid=0. 2108 2109 ima_canonical_fmt [IMA] 2110 Use the canonical format for the binary runtime 2111 measurements, instead of host native format. 2112 2113 ima_hash= [IMA] 2114 Format: { md5 | sha1 | rmd160 | sha256 | sha384 2115 | sha512 | ... } 2116 default: "sha1" 2117 2118 The list of supported hash algorithms is defined 2119 in crypto/hash_info.h. 2120 2121 ima_policy= [IMA] 2122 The builtin policies to load during IMA setup. 2123 Format: "tcb | appraise_tcb | secure_boot | 2124 fail_securely | critical_data" 2125 2126 The "tcb" policy measures all programs exec'd, files 2127 mmap'd for exec, and all files opened with the read 2128 mode bit set by either the effective uid (euid=0) or 2129 uid=0. 2130 2131 The "appraise_tcb" policy appraises the integrity of 2132 all files owned by root. 2133 2134 The "secure_boot" policy appraises the integrity 2135 of files (eg. kexec kernel image, kernel modules, 2136 firmware, policy, etc) based on file signatures. 2137 2138 The "fail_securely" policy forces file signature 2139 verification failure also on privileged mounted 2140 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE 2141 flag. 2142 2143 The "critical_data" policy measures kernel integrity 2144 critical data. 2145 2146 ima_tcb [IMA] Deprecated. Use ima_policy= instead. 2147 Load a policy which meets the needs of the Trusted 2148 Computing Base. This means IMA will measure all 2149 programs exec'd, files mmap'd for exec, and all files 2150 opened for read by uid=0. 2151 2152 ima_template= [IMA] 2153 Select one of defined IMA measurements template formats. 2154 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" | 2155 "ima-sigv2" } 2156 Default: "ima-ng" 2157 2158 ima_template_fmt= 2159 [IMA] Define a custom template format. 2160 Format: { "field1|...|fieldN" } 2161 2162 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage 2163 Format: <min_file_size> 2164 Set the minimal file size for using asynchronous hash. 2165 If left unspecified, ahash usage is disabled. 2166 2167 ahash performance varies for different data sizes on 2168 different crypto accelerators. This option can be used 2169 to achieve the best performance for a particular HW. 2170 2171 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size 2172 Format: <bufsize> 2173 Set hashing buffer size. Default: 4k. 2174 2175 ahash performance varies for different chunk sizes on 2176 different crypto accelerators. This option can be used 2177 to achieve best performance for particular HW. 2178 2179 init= [KNL] 2180 Format: <full_path> 2181 Run specified binary instead of /sbin/init as init 2182 process. 2183 2184 initcall_debug [KNL] Trace initcalls as they are executed. Useful 2185 for working out where the kernel is dying during 2186 startup. 2187 2188 initcall_blacklist= [KNL] Do not execute a comma-separated list of 2189 initcall functions. Useful for debugging built-in 2190 modules and initcalls. 2191 2192 initramfs_async= [KNL] 2193 Format: <bool> 2194 Default: 1 2195 This parameter controls whether the initramfs 2196 image is unpacked asynchronously, concurrently 2197 with devices being probed and 2198 initialized. This should normally just work, 2199 but as a debugging aid, one can get the 2200 historical behaviour of the initramfs 2201 unpacking being completed before device_ and 2202 late_ initcalls. 2203 2204 initrd= [BOOT,EARLY] Specify the location of the initial ramdisk 2205 2206 initrdmem= [KNL,EARLY] Specify a physical address and size from which to 2207 load the initrd. If an initrd is compiled in or 2208 specified in the bootparams, it takes priority over this 2209 setting. 2210 Format: ss[KMG],nn[KMG] 2211 Default is 0, 0 2212 2213 init_on_alloc= [MM,EARLY] Fill newly allocated pages and heap objects with 2214 zeroes. 2215 Format: 0 | 1 2216 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON. 2217 2218 init_on_free= [MM,EARLY] Fill freed pages and heap objects with zeroes. 2219 Format: 0 | 1 2220 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON. 2221 2222 init_pkru= [X86] Specify the default memory protection keys rights 2223 register contents for all processes. 0x55555554 by 2224 default (disallow access to all but pkey 0). Can 2225 override in debugfs after boot. 2226 2227 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver 2228 Format: <irq> 2229 2230 int_pln_enable [X86] Enable power limit notification interrupt 2231 2232 integrity_audit=[IMA] 2233 Format: { "0" | "1" } 2234 0 -- basic integrity auditing messages. (Default) 2235 1 -- additional integrity auditing messages. 2236 2237 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option 2238 on 2239 Enable intel iommu driver. 2240 off 2241 Disable intel iommu driver. 2242 igfx_off [Default Off] 2243 By default, gfx is mapped as normal device. If a gfx 2244 device has a dedicated DMAR unit, the DMAR unit is 2245 bypassed by not enabling DMAR with this option. In 2246 this case, gfx device will use physical address for 2247 DMA. 2248 strict [Default Off] 2249 Deprecated, equivalent to iommu.strict=1. 2250 sp_off [Default Off] 2251 By default, super page will be supported if Intel IOMMU 2252 has the capability. With this option, super page will 2253 not be supported. 2254 sm_on 2255 Enable the Intel IOMMU scalable mode if the hardware 2256 advertises that it has support for the scalable mode 2257 translation. 2258 sm_off 2259 Disallow use of the Intel IOMMU scalable mode. 2260 tboot_noforce [Default Off] 2261 Do not force the Intel IOMMU enabled under tboot. 2262 By default, tboot will force Intel IOMMU on, which 2263 could harm performance of some high-throughput 2264 devices like 40GBit network cards, even if identity 2265 mapping is enabled. 2266 Note that using this option lowers the security 2267 provided by tboot because it makes the system 2268 vulnerable to DMA attacks. 2269 2270 intel_idle.max_cstate= [KNL,HW,ACPI,X86] 2271 0 disables intel_idle and fall back on acpi_idle. 2272 1 to 9 specify maximum depth of C-state. 2273 2274 intel_pstate= [X86,EARLY] 2275 disable 2276 Do not enable intel_pstate as the default 2277 scaling driver for the supported processors 2278 active 2279 Use intel_pstate driver to bypass the scaling 2280 governors layer of cpufreq and provides it own 2281 algorithms for p-state selection. There are two 2282 P-state selection algorithms provided by 2283 intel_pstate in the active mode: powersave and 2284 performance. The way they both operate depends 2285 on whether or not the hardware managed P-states 2286 (HWP) feature has been enabled in the processor 2287 and possibly on the processor model. 2288 passive 2289 Use intel_pstate as a scaling driver, but configure it 2290 to work with generic cpufreq governors (instead of 2291 enabling its internal governor). This mode cannot be 2292 used along with the hardware-managed P-states (HWP) 2293 feature. 2294 force 2295 Enable intel_pstate on systems that prohibit it by default 2296 in favor of acpi-cpufreq. Forcing the intel_pstate driver 2297 instead of acpi-cpufreq may disable platform features, such 2298 as thermal controls and power capping, that rely on ACPI 2299 P-States information being indicated to OSPM and therefore 2300 should be used with caution. This option does not work with 2301 processors that aren't supported by the intel_pstate driver 2302 or on platforms that use pcc-cpufreq instead of acpi-cpufreq. 2303 no_hwp 2304 Do not enable hardware P state control (HWP) 2305 if available. 2306 hwp_only 2307 Only load intel_pstate on systems which support 2308 hardware P state control (HWP) if available. 2309 support_acpi_ppc 2310 Enforce ACPI _PPC performance limits. If the Fixed ACPI 2311 Description Table, specifies preferred power management 2312 profile as "Enterprise Server" or "Performance Server", 2313 then this feature is turned on by default. 2314 per_cpu_perf_limits 2315 Allow per-logical-CPU P-State performance control limits using 2316 cpufreq sysfs interface 2317 no_cas 2318 Do not enable capacity-aware scheduling (CAS) on 2319 hybrid systems 2320 2321 intremap= [X86-64,Intel-IOMMU,EARLY] 2322 on enable Interrupt Remapping (default) 2323 off disable Interrupt Remapping 2324 nosid disable Source ID checking 2325 no_x2apic_optout 2326 BIOS x2APIC opt-out request will be ignored 2327 nopost disable Interrupt Posting 2328 posted_msi 2329 enable MSIs delivered as posted interrupts 2330 2331 iomem= Disable strict checking of access to MMIO memory 2332 strict regions from userspace. 2333 relaxed 2334 2335 iommu= [X86,EARLY] 2336 2337 off 2338 Don't initialize and use any kind of IOMMU. 2339 2340 force 2341 Force the use of the hardware IOMMU even when 2342 it is not actually needed (e.g. because < 3 GB 2343 memory). 2344 2345 noforce 2346 Don't force hardware IOMMU usage when it is not 2347 needed. (default). 2348 2349 biomerge 2350 panic 2351 nopanic 2352 merge 2353 nomerge 2354 2355 soft 2356 Use software bounce buffering (SWIOTLB) (default for 2357 Intel machines). This can be used to prevent the usage 2358 of an available hardware IOMMU. 2359 2360 [X86] 2361 pt 2362 [X86] 2363 nopt 2364 [PPC/POWERNV] 2365 nobypass 2366 Disable IOMMU bypass, using IOMMU for PCI devices. 2367 2368 [X86] 2369 AMD Gart HW IOMMU-specific options: 2370 2371 <size> 2372 Set the size of the remapping area in bytes. 2373 2374 allowed 2375 Overwrite iommu off workarounds for specific chipsets 2376 2377 fullflush 2378 Flush IOMMU on each allocation (default). 2379 2380 nofullflush 2381 Don't use IOMMU fullflush. 2382 2383 memaper[=<order>] 2384 Allocate an own aperture over RAM with size 2385 32MB<<order. (default: order=1, i.e. 64MB) 2386 2387 merge 2388 Do scatter-gather (SG) merging. Implies "force" 2389 (experimental). 2390 2391 nomerge 2392 Don't do scatter-gather (SG) merging. 2393 2394 noaperture 2395 Ask the IOMMU not to touch the aperture for AGP. 2396 2397 noagp 2398 Don't initialize the AGP driver and use full aperture. 2399 2400 panic 2401 Always panic when IOMMU overflows. 2402 2403 iommu.forcedac= [ARM64,X86,EARLY] Control IOVA allocation for PCI devices. 2404 Format: { "0" | "1" } 2405 0 - Try to allocate a 32-bit DMA address first, before 2406 falling back to the full range if needed. 2407 1 - Allocate directly from the full usable range, 2408 forcing Dual Address Cycle for PCI cards supporting 2409 greater than 32-bit addressing. 2410 2411 iommu.strict= [ARM64,X86,S390,EARLY] Configure TLB invalidation behaviour 2412 Format: { "0" | "1" } 2413 0 - Lazy mode. 2414 Request that DMA unmap operations use deferred 2415 invalidation of hardware TLBs, for increased 2416 throughput at the cost of reduced device isolation. 2417 Will fall back to strict mode if not supported by 2418 the relevant IOMMU driver. 2419 1 - Strict mode. 2420 DMA unmap operations invalidate IOMMU hardware TLBs 2421 synchronously. 2422 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}. 2423 Note: on x86, strict mode specified via one of the 2424 legacy driver-specific options takes precedence. 2425 2426 iommu.passthrough= 2427 [ARM64,X86,EARLY] Configure DMA to bypass the IOMMU by default. 2428 Format: { "0" | "1" } 2429 0 - Use IOMMU translation for DMA. 2430 1 - Bypass the IOMMU for DMA. 2431 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH. 2432 2433 io7= [HW] IO7 for Marvel-based Alpha systems 2434 See comment before marvel_specify_io7 in 2435 arch/alpha/kernel/core_marvel.c. 2436 2437 io_delay= [X86,EARLY] I/O delay method 2438 0x80 2439 Standard port 0x80 based delay 2440 0xed 2441 Alternate port 0xed based delay (needed on some systems) 2442 udelay 2443 Simple two microseconds delay 2444 none 2445 No delay 2446 2447 ip= [IP_PNP] 2448 See Documentation/admin-guide/nfs/nfsroot.rst. 2449 2450 ipcmni_extend [KNL,EARLY] Extend the maximum number of unique System V 2451 IPC identifiers from 32,768 to 16,777,216. 2452 2453 ipe.enforce= [IPE] 2454 Format: <bool> 2455 Determine whether IPE starts in permissive (0) or 2456 enforce (1) mode. The default is enforce. 2457 2458 ipe.success_audit= 2459 [IPE] 2460 Format: <bool> 2461 Start IPE with success auditing enabled, emitting 2462 an audit event when a binary is allowed. The default 2463 is 0. 2464 2465 irqaffinity= [SMP] Set the default irq affinity mask 2466 The argument is a cpu list, as described above. 2467 2468 irqchip.gicv2_force_probe= 2469 [ARM,ARM64,EARLY] 2470 Format: <bool> 2471 Force the kernel to look for the second 4kB page 2472 of a GICv2 controller even if the memory range 2473 exposed by the device tree is too small. 2474 2475 irqchip.gicv3_nolpi= 2476 [ARM,ARM64,EARLY] 2477 Force the kernel to ignore the availability of 2478 LPIs (and by consequence ITSs). Intended for system 2479 that use the kernel as a bootloader, and thus want 2480 to let secondary kernels in charge of setting up 2481 LPIs. 2482 2483 irqchip.gicv3_pseudo_nmi= [ARM64,EARLY] 2484 Enables support for pseudo-NMIs in the kernel. This 2485 requires the kernel to be built with 2486 CONFIG_ARM64_PSEUDO_NMI. 2487 2488 irqfixup [HW] 2489 When an interrupt is not handled search all handlers 2490 for it. Intended to get systems with badly broken 2491 firmware running. 2492 2493 irqpoll [HW] 2494 When an interrupt is not handled search all handlers 2495 for it. Also check all handlers each timer 2496 interrupt. Intended to get systems with badly broken 2497 firmware running. 2498 2499 isapnp= [ISAPNP] 2500 Format: <RDP>,<reset>,<pci_scan>,<verbosity> 2501 2502 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance. 2503 [Deprecated - use cpusets instead] 2504 Format: [flag-list,]<cpu-list> 2505 2506 Specify one or more CPUs to isolate from disturbances 2507 specified in the flag list (default: domain): 2508 2509 nohz 2510 Disable the tick when a single task runs as well as 2511 disabling other kernel noises like having RCU callbacks 2512 offloaded. This is equivalent to the nohz_full parameter. 2513 2514 A residual 1Hz tick is offloaded to workqueues, which you 2515 need to affine to housekeeping through the global 2516 workqueue's affinity configured via the 2517 /sys/devices/virtual/workqueue/cpumask sysfs file, or 2518 by using the 'domain' flag described below. 2519 2520 NOTE: by default the global workqueue runs on all CPUs, 2521 so to protect individual CPUs the 'cpumask' file has to 2522 be configured manually after bootup. 2523 2524 domain 2525 Isolate from the general SMP balancing and scheduling 2526 algorithms. Note that performing domain isolation this way 2527 is irreversible: it's not possible to bring back a CPU to 2528 the domains once isolated through isolcpus. It's strongly 2529 advised to use cpusets instead to disable scheduler load 2530 balancing through the "cpuset.sched_load_balance" file. 2531 It offers a much more flexible interface where CPUs can 2532 move in and out of an isolated set anytime. 2533 2534 You can move a process onto or off an "isolated" CPU via 2535 the CPU affinity syscalls or cpuset. 2536 <cpu number> begins at 0 and the maximum value is 2537 "number of CPUs in system - 1". 2538 2539 managed_irq 2540 2541 Isolate from being targeted by managed interrupts 2542 which have an interrupt mask containing isolated 2543 CPUs. The affinity of managed interrupts is 2544 handled by the kernel and cannot be changed via 2545 the /proc/irq/* interfaces. 2546 2547 This isolation is best effort and only effective 2548 if the automatically assigned interrupt mask of a 2549 device queue contains isolated and housekeeping 2550 CPUs. If housekeeping CPUs are online then such 2551 interrupts are directed to the housekeeping CPU 2552 so that IO submitted on the housekeeping CPU 2553 cannot disturb the isolated CPU. 2554 2555 If a queue's affinity mask contains only isolated 2556 CPUs then this parameter has no effect on the 2557 interrupt routing decision, though interrupts are 2558 only delivered when tasks running on those 2559 isolated CPUs submit IO. IO submitted on 2560 housekeeping CPUs has no influence on those 2561 queues. 2562 2563 The format of <cpu-list> is described above. 2564 2565 iucv= [HW,NET] 2566 2567 ivrs_ioapic [HW,X86-64] 2568 Provide an override to the IOAPIC-ID<->DEVICE-ID 2569 mapping provided in the IVRS ACPI table. 2570 By default, PCI segment is 0, and can be omitted. 2571 2572 For example, to map IOAPIC-ID decimal 10 to 2573 PCI segment 0x1 and PCI device 00:14.0, 2574 write the parameter as: 2575 ivrs_ioapic=10@0001:00:14.0 2576 2577 Deprecated formats: 2578 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0 2579 write the parameter as: 2580 ivrs_ioapic[10]=00:14.0 2581 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and 2582 PCI device 00:14.0 write the parameter as: 2583 ivrs_ioapic[10]=0001:00:14.0 2584 2585 ivrs_hpet [HW,X86-64] 2586 Provide an override to the HPET-ID<->DEVICE-ID 2587 mapping provided in the IVRS ACPI table. 2588 By default, PCI segment is 0, and can be omitted. 2589 2590 For example, to map HPET-ID decimal 10 to 2591 PCI segment 0x1 and PCI device 00:14.0, 2592 write the parameter as: 2593 ivrs_hpet=10@0001:00:14.0 2594 2595 Deprecated formats: 2596 * To map HPET-ID decimal 0 to PCI device 00:14.0 2597 write the parameter as: 2598 ivrs_hpet[0]=00:14.0 2599 * To map HPET-ID decimal 10 to PCI segment 0x1 and 2600 PCI device 00:14.0 write the parameter as: 2601 ivrs_ioapic[10]=0001:00:14.0 2602 2603 ivrs_acpihid [HW,X86-64] 2604 Provide an override to the ACPI-HID:UID<->DEVICE-ID 2605 mapping provided in the IVRS ACPI table. 2606 By default, PCI segment is 0, and can be omitted. 2607 2608 For example, to map UART-HID:UID AMD0020:0 to 2609 PCI segment 0x1 and PCI device ID 00:14.5, 2610 write the parameter as: 2611 ivrs_acpihid=AMD0020:0@0001:00:14.5 2612 2613 Deprecated formats: 2614 * To map UART-HID:UID AMD0020:0 to PCI segment is 0, 2615 PCI device ID 00:14.5, write the parameter as: 2616 ivrs_acpihid[00:14.5]=AMD0020:0 2617 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and 2618 PCI device ID 00:14.5, write the parameter as: 2619 ivrs_acpihid[0001:00:14.5]=AMD0020:0 2620 2621 js= [HW,JOY] Analog joystick 2622 See Documentation/input/joydev/joystick.rst. 2623 2624 kasan_multi_shot 2625 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print 2626 report on every invalid memory access. Without this 2627 parameter KASAN will print report only for the first 2628 invalid access. 2629 2630 keep_bootcon [KNL,EARLY] 2631 Do not unregister boot console at start. This is only 2632 useful for debugging when something happens in the window 2633 between unregistering the boot console and initializing 2634 the real console. 2635 2636 keepinitrd [HW,ARM] See retain_initrd. 2637 2638 kernelcore= [KNL,X86,PPC,EARLY] 2639 Format: nn[KMGTPE] | nn% | "mirror" 2640 This parameter specifies the amount of memory usable by 2641 the kernel for non-movable allocations. The requested 2642 amount is spread evenly throughout all nodes in the 2643 system as ZONE_NORMAL. The remaining memory is used for 2644 movable memory in its own zone, ZONE_MOVABLE. In the 2645 event, a node is too small to have both ZONE_NORMAL and 2646 ZONE_MOVABLE, kernelcore memory will take priority and 2647 other nodes will have a larger ZONE_MOVABLE. 2648 2649 ZONE_MOVABLE is used for the allocation of pages that 2650 may be reclaimed or moved by the page migration 2651 subsystem. Note that allocations like PTEs-from-HighMem 2652 still use the HighMem zone if it exists, and the Normal 2653 zone if it does not. 2654 2655 It is possible to specify the exact amount of memory in 2656 the form of "nn[KMGTPE]", a percentage of total system 2657 memory in the form of "nn%", or "mirror". If "mirror" 2658 option is specified, mirrored (reliable) memory is used 2659 for non-movable allocations and remaining memory is used 2660 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror" 2661 are exclusive, so you cannot specify multiple forms. 2662 2663 kgdbdbgp= [KGDB,HW,EARLY] kgdb over EHCI usb debug port. 2664 Format: <Controller#>[,poll interval] 2665 The controller # is the number of the ehci usb debug 2666 port as it is probed via PCI. The poll interval is 2667 optional and is the number seconds in between 2668 each poll cycle to the debug port in case you need 2669 the functionality for interrupting the kernel with 2670 gdb or control-c on the dbgp connection. When 2671 not using this parameter you use sysrq-g to break into 2672 the kernel debugger. 2673 2674 kgdboc= [KGDB,HW] kgdb over consoles. 2675 Requires a tty driver that supports console polling, 2676 or a supported polling keyboard driver (non-usb). 2677 Serial only format: <serial_device>[,baud] 2678 keyboard only format: kbd 2679 keyboard and serial format: kbd,<serial_device>[,baud] 2680 Optional Kernel mode setting: 2681 kms, kbd format: kms,kbd 2682 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud] 2683 2684 kgdboc_earlycon= [KGDB,HW,EARLY] 2685 If the boot console provides the ability to read 2686 characters and can work in polling mode, you can use 2687 this parameter to tell kgdb to use it as a backend 2688 until the normal console is registered. Intended to 2689 be used together with the kgdboc parameter which 2690 specifies the normal console to transition to. 2691 2692 The name of the early console should be specified 2693 as the value of this parameter. Note that the name of 2694 the early console might be different than the tty 2695 name passed to kgdboc. It's OK to leave the value 2696 blank and the first boot console that implements 2697 read() will be picked. 2698 2699 kgdbwait [KGDB,EARLY] Stop kernel execution and enter the 2700 kernel debugger at the earliest opportunity. 2701 2702 kmac= [MIPS] Korina ethernet MAC address. 2703 Configure the RouterBoard 532 series on-chip 2704 Ethernet adapter MAC address. 2705 2706 kmemleak= [KNL,EARLY] Boot-time kmemleak enable/disable 2707 Valid arguments: on, off 2708 Default: on 2709 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y, 2710 the default is off. 2711 2712 kprobe_event=[probe-list] 2713 [FTRACE] Add kprobe events and enable at boot time. 2714 The probe-list is a semicolon delimited list of probe 2715 definitions. Each definition is same as kprobe_events 2716 interface, but the parameters are comma delimited. 2717 For example, to add a kprobe event on vfs_read with 2718 arg1 and arg2, add to the command line; 2719 2720 kprobe_event=p,vfs_read,$arg1,$arg2 2721 2722 See also Documentation/trace/kprobetrace.rst "Kernel 2723 Boot Parameter" section. 2724 2725 kpti= [ARM64,EARLY] Control page table isolation of 2726 user and kernel address spaces. 2727 Default: enabled on cores which need mitigation. 2728 0: force disabled 2729 1: force enabled 2730 2731 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires 2732 CONFIG_KUNIT to be set to be fully enabled. The 2733 default value can be overridden via 2734 KUNIT_DEFAULT_ENABLED. 2735 Default is 1 (enabled) 2736 2737 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs. 2738 Default is 0 (don't ignore, but inject #GP) 2739 2740 kvm.eager_page_split= 2741 [KVM,X86] Controls whether or not KVM will try to 2742 proactively split all huge pages during dirty logging. 2743 Eager page splitting reduces interruptions to vCPU 2744 execution by eliminating the write-protection faults 2745 and MMU lock contention that would otherwise be 2746 required to split huge pages lazily. 2747 2748 VM workloads that rarely perform writes or that write 2749 only to a small region of VM memory may benefit from 2750 disabling eager page splitting to allow huge pages to 2751 still be used for reads. 2752 2753 The behavior of eager page splitting depends on whether 2754 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If 2755 disabled, all huge pages in a memslot will be eagerly 2756 split when dirty logging is enabled on that memslot. If 2757 enabled, eager page splitting will be performed during 2758 the KVM_CLEAR_DIRTY ioctl, and only for the pages being 2759 cleared. 2760 2761 Eager page splitting is only supported when kvm.tdp_mmu=Y. 2762 2763 Default is Y (on). 2764 2765 kvm.enable_virt_at_load=[KVM,ARM64,LOONGARCH,MIPS,RISCV,X86] 2766 If enabled, KVM will enable virtualization in hardware 2767 when KVM is loaded, and disable virtualization when KVM 2768 is unloaded (if KVM is built as a module). 2769 2770 If disabled, KVM will dynamically enable and disable 2771 virtualization on-demand when creating and destroying 2772 VMs, i.e. on the 0=>1 and 1=>0 transitions of the 2773 number of VMs. 2774 2775 Enabling virtualization at module load avoids potential 2776 latency for creation of the 0=>1 VM, as KVM serializes 2777 virtualization enabling across all online CPUs. The 2778 "cost" of enabling virtualization when KVM is loaded, 2779 is that doing so may interfere with using out-of-tree 2780 hypervisors that want to "own" virtualization hardware. 2781 2782 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface. 2783 Default is false (don't support). 2784 2785 kvm.nx_huge_pages= 2786 [KVM] Controls the software workaround for the 2787 X86_BUG_ITLB_MULTIHIT bug. 2788 force : Always deploy workaround. 2789 off : Never deploy workaround. 2790 auto : Deploy workaround based on the presence of 2791 X86_BUG_ITLB_MULTIHIT. 2792 2793 Default is 'auto'. 2794 2795 If the software workaround is enabled for the host, 2796 guests do need not to enable it for nested guests. 2797 2798 kvm.nx_huge_pages_recovery_ratio= 2799 [KVM] Controls how many 4KiB pages are periodically zapped 2800 back to huge pages. 0 disables the recovery, otherwise if 2801 the value is N KVM will zap 1/Nth of the 4KiB pages every 2802 period (see below). The default is 60. 2803 2804 kvm.nx_huge_pages_recovery_period_ms= 2805 [KVM] Controls the time period at which KVM zaps 4KiB pages 2806 back to huge pages. If the value is a non-zero N, KVM will 2807 zap a portion (see ratio above) of the pages every N msecs. 2808 If the value is 0 (the default), KVM will pick a period based 2809 on the ratio, such that a page is zapped after 1 hour on average. 2810 2811 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in 2812 KVM/SVM. Default is 1 (enabled). 2813 2814 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables, 2815 a.k.a. Two-Dimensional Page Tables. Default is 1 2816 (enabled). Disable by KVM if hardware lacks support 2817 for NPT. 2818 2819 kvm-arm.mode= 2820 [KVM,ARM,EARLY] Select one of KVM/arm64's modes of 2821 operation. 2822 2823 none: Forcefully disable KVM. 2824 2825 nvhe: Standard nVHE-based mode, without support for 2826 protected guests. 2827 2828 protected: Mode with support for guests whose state is 2829 kept private from the host, using VHE or 2830 nVHE depending on HW support. 2831 2832 nested: VHE-based mode with support for nested 2833 virtualization. Requires at least ARMv8.4 2834 hardware (with FEAT_NV2). 2835 2836 Defaults to VHE/nVHE based on hardware support. Setting 2837 mode to "protected" will disable kexec and hibernation 2838 for the host. To force nVHE on VHE hardware, add 2839 "arm64_sw.hvhe=0 id_aa64mmfr1.vh=0" to the 2840 command-line. 2841 "nested" is experimental and should be used with 2842 extreme caution. 2843 2844 kvm-arm.vgic_v3_group0_trap= 2845 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-0 2846 system registers 2847 2848 kvm-arm.vgic_v3_group1_trap= 2849 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-1 2850 system registers 2851 2852 kvm-arm.vgic_v3_common_trap= 2853 [KVM,ARM,EARLY] Trap guest accesses to GICv3 common 2854 system registers 2855 2856 kvm-arm.vgic_v4_enable= 2857 [KVM,ARM,EARLY] Allow use of GICv4 for direct 2858 injection of LPIs. 2859 2860 kvm-arm.wfe_trap_policy= 2861 [KVM,ARM] Control when to set WFE instruction trap for 2862 KVM VMs. Traps are allowed but not guaranteed by the 2863 CPU architecture. 2864 2865 trap: set WFE instruction trap 2866 2867 notrap: clear WFE instruction trap 2868 2869 kvm-arm.wfi_trap_policy= 2870 [KVM,ARM] Control when to set WFI instruction trap for 2871 KVM VMs. Traps are allowed but not guaranteed by the 2872 CPU architecture. 2873 2874 trap: set WFI instruction trap 2875 2876 notrap: clear WFI instruction trap 2877 2878 kvm_cma_resv_ratio=n [PPC,EARLY] 2879 Reserves given percentage from system memory area for 2880 contiguous memory allocation for KVM hash pagetable 2881 allocation. 2882 By default it reserves 5% of total system memory. 2883 Format: <integer> 2884 Default: 5 2885 2886 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables, 2887 a.k.a. Two-Dimensional Page Tables. Default is 1 2888 (enabled). Disable by KVM if hardware lacks support 2889 for EPT. 2890 2891 kvm-intel.emulate_invalid_guest_state= 2892 [KVM,Intel] Control whether to emulate invalid guest 2893 state. Ignored if kvm-intel.enable_unrestricted_guest=1, 2894 as guest state is never invalid for unrestricted 2895 guests. This param doesn't apply to nested guests (L2), 2896 as KVM never emulates invalid L2 guest state. 2897 Default is 1 (enabled). 2898 2899 kvm-intel.flexpriority= 2900 [KVM,Intel] Control KVM's use of FlexPriority feature 2901 (TPR shadow). Default is 1 (enabled). Disable by KVM if 2902 hardware lacks support for it. 2903 2904 kvm-intel.nested= 2905 [KVM,Intel] Control nested virtualization feature in 2906 KVM/VMX. Default is 1 (enabled). 2907 2908 kvm-intel.unrestricted_guest= 2909 [KVM,Intel] Control KVM's use of unrestricted guest 2910 feature (virtualized real and unpaged mode). Default 2911 is 1 (enabled). Disable by KVM if EPT is disabled or 2912 hardware lacks support for it. 2913 2914 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault 2915 CVE-2018-3620. 2916 2917 Valid arguments: never, cond, always 2918 2919 always: L1D cache flush on every VMENTER. 2920 cond: Flush L1D on VMENTER only when the code between 2921 VMEXIT and VMENTER can leak host memory. 2922 never: Disables the mitigation 2923 2924 Default is cond (do L1 cache flush in specific instances) 2925 2926 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor 2927 Identification feature (tagged TLBs). Default is 1 2928 (enabled). Disable by KVM if hardware lacks support 2929 for it. 2930 2931 l1d_flush= [X86,INTEL,EARLY] 2932 Control mitigation for L1D based snooping vulnerability. 2933 2934 Certain CPUs are vulnerable to an exploit against CPU 2935 internal buffers which can forward information to a 2936 disclosure gadget under certain conditions. 2937 2938 In vulnerable processors, the speculatively 2939 forwarded data can be used in a cache side channel 2940 attack, to access data to which the attacker does 2941 not have direct access. 2942 2943 This parameter controls the mitigation. The 2944 options are: 2945 2946 on - enable the interface for the mitigation 2947 2948 l1tf= [X86,EARLY] Control mitigation of the L1TF vulnerability on 2949 affected CPUs 2950 2951 The kernel PTE inversion protection is unconditionally 2952 enabled and cannot be disabled. 2953 2954 full 2955 Provides all available mitigations for the 2956 L1TF vulnerability. Disables SMT and 2957 enables all mitigations in the 2958 hypervisors, i.e. unconditional L1D flush. 2959 2960 SMT control and L1D flush control via the 2961 sysfs interface is still possible after 2962 boot. Hypervisors will issue a warning 2963 when the first VM is started in a 2964 potentially insecure configuration, 2965 i.e. SMT enabled or L1D flush disabled. 2966 2967 full,force 2968 Same as 'full', but disables SMT and L1D 2969 flush runtime control. Implies the 2970 'nosmt=force' command line option. 2971 (i.e. sysfs control of SMT is disabled.) 2972 2973 flush 2974 Leaves SMT enabled and enables the default 2975 hypervisor mitigation, i.e. conditional 2976 L1D flush. 2977 2978 SMT control and L1D flush control via the 2979 sysfs interface is still possible after 2980 boot. Hypervisors will issue a warning 2981 when the first VM is started in a 2982 potentially insecure configuration, 2983 i.e. SMT enabled or L1D flush disabled. 2984 2985 flush,nosmt 2986 2987 Disables SMT and enables the default 2988 hypervisor mitigation. 2989 2990 SMT control and L1D flush control via the 2991 sysfs interface is still possible after 2992 boot. Hypervisors will issue a warning 2993 when the first VM is started in a 2994 potentially insecure configuration, 2995 i.e. SMT enabled or L1D flush disabled. 2996 2997 flush,nowarn 2998 Same as 'flush', but hypervisors will not 2999 warn when a VM is started in a potentially 3000 insecure configuration. 3001 3002 off 3003 Disables hypervisor mitigations and doesn't 3004 emit any warnings. 3005 It also drops the swap size and available 3006 RAM limit restriction on both hypervisor and 3007 bare metal. 3008 3009 Default is 'flush'. 3010 3011 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst 3012 3013 l2cr= [PPC] 3014 3015 l3cr= [PPC] 3016 3017 lapic [X86-32,APIC,EARLY] Enable the local APIC even if BIOS 3018 disabled it. 3019 3020 lapic= [X86,APIC] Do not use TSC deadline 3021 value for LAPIC timer one-shot implementation. Default 3022 back to the programmable timer unit in the LAPIC. 3023 Format: notscdeadline 3024 3025 lapic_timer_c2_ok [X86,APIC,EARLY] trust the local apic timer 3026 in C2 power state. 3027 3028 libata.dma= [LIBATA] DMA control 3029 libata.dma=0 Disable all PATA and SATA DMA 3030 libata.dma=1 PATA and SATA Disk DMA only 3031 libata.dma=2 ATAPI (CDROM) DMA only 3032 libata.dma=4 Compact Flash DMA only 3033 Combinations also work, so libata.dma=3 enables DMA 3034 for disks and CDROMs, but not CFs. 3035 3036 libata.ignore_hpa= [LIBATA] Ignore HPA limit 3037 libata.ignore_hpa=0 keep BIOS limits (default) 3038 libata.ignore_hpa=1 ignore limits, using full disk 3039 3040 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume 3041 when set. 3042 Format: <int> 3043 3044 libata.force= [LIBATA] Force configurations. The format is a comma- 3045 separated list of "[ID:]VAL" where ID is PORT[.DEVICE]. 3046 PORT and DEVICE are decimal numbers matching port, link 3047 or device. Basically, it matches the ATA ID string 3048 printed on console by libata. If the whole ID part is 3049 omitted, the last PORT and DEVICE values are used. If 3050 ID hasn't been specified yet, the configuration applies 3051 to all ports, links and devices. 3052 3053 If only DEVICE is omitted, the parameter applies to 3054 the port and all links and devices behind it. DEVICE 3055 number of 0 either selects the first device or the 3056 first fan-out link behind PMP device. It does not 3057 select the host link. DEVICE number of 15 selects the 3058 host link and device attached to it. 3059 3060 The VAL specifies the configuration to force. As long 3061 as there is no ambiguity, shortcut notation is allowed. 3062 For example, both 1.5 and 1.5G would work for 1.5Gbps. 3063 The following configurations can be forced. 3064 3065 * Cable type: 40c, 80c, short40c, unk, ign or sata. 3066 Any ID with matching PORT is used. 3067 3068 * SATA link speed limit: 1.5Gbps or 3.0Gbps. 3069 3070 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7]. 3071 udma[/][16,25,33,44,66,100,133] notation is also 3072 allowed. 3073 3074 * nohrst, nosrst, norst: suppress hard, soft and both 3075 resets. 3076 3077 * rstonce: only attempt one reset during hot-unplug 3078 link recovery. 3079 3080 * [no]dbdelay: Enable or disable the extra 200ms delay 3081 before debouncing a link PHY and device presence 3082 detection. 3083 3084 * [no]ncq: Turn on or off NCQ. 3085 3086 * [no]ncqtrim: Enable or disable queued DSM TRIM. 3087 3088 * [no]ncqati: Enable or disable NCQ trim on ATI chipset. 3089 3090 * [no]trim: Enable or disable (unqueued) TRIM. 3091 3092 * trim_zero: Indicate that TRIM command zeroes data. 3093 3094 * max_trim_128m: Set 128M maximum trim size limit. 3095 3096 * [no]dma: Turn on or off DMA transfers. 3097 3098 * atapi_dmadir: Enable ATAPI DMADIR bridge support. 3099 3100 * atapi_mod16_dma: Enable the use of ATAPI DMA for 3101 commands that are not a multiple of 16 bytes. 3102 3103 * [no]dmalog: Enable or disable the use of the 3104 READ LOG DMA EXT command to access logs. 3105 3106 * [no]iddevlog: Enable or disable access to the 3107 identify device data log. 3108 3109 * [no]logdir: Enable or disable access to the general 3110 purpose log directory. 3111 3112 * max_sec_128: Set transfer size limit to 128 sectors. 3113 3114 * max_sec_1024: Set or clear transfer size limit to 3115 1024 sectors. 3116 3117 * max_sec_lba48: Set or clear transfer size limit to 3118 65535 sectors. 3119 3120 * external: Mark port as external (hotplug-capable). 3121 3122 * [no]lpm: Enable or disable link power management. 3123 3124 * [no]setxfer: Indicate if transfer speed mode setting 3125 should be skipped. 3126 3127 * [no]fua: Disable or enable FUA (Force Unit Access) 3128 support for devices supporting this feature. 3129 3130 * dump_id: Dump IDENTIFY data. 3131 3132 * disable: Disable this device. 3133 3134 If there are multiple matching configurations changing 3135 the same attribute, the last one is used. 3136 3137 load_ramdisk= [RAM] [Deprecated] 3138 3139 lockd.nlm_grace_period=P [NFS] Assign grace period. 3140 Format: <integer> 3141 3142 lockd.nlm_tcpport=N [NFS] Assign TCP port. 3143 Format: <integer> 3144 3145 lockd.nlm_timeout=T [NFS] Assign timeout value. 3146 Format: <integer> 3147 3148 lockd.nlm_udpport=M [NFS] Assign UDP port. 3149 Format: <integer> 3150 3151 lockdown= [SECURITY,EARLY] 3152 { integrity | confidentiality } 3153 Enable the kernel lockdown feature. If set to 3154 integrity, kernel features that allow userland to 3155 modify the running kernel are disabled. If set to 3156 confidentiality, kernel features that allow userland 3157 to extract confidential information from the kernel 3158 are also disabled. 3159 3160 locktorture.acq_writer_lim= [KNL] 3161 Set the time limit in jiffies for a lock 3162 acquisition. Acquisitions exceeding this limit 3163 will result in a splat once they do complete. 3164 3165 locktorture.bind_readers= [KNL] 3166 Specify the list of CPUs to which the readers are 3167 to be bound. 3168 3169 locktorture.bind_writers= [KNL] 3170 Specify the list of CPUs to which the writers are 3171 to be bound. 3172 3173 locktorture.call_rcu_chains= [KNL] 3174 Specify the number of self-propagating call_rcu() 3175 chains to set up. These are used to ensure that 3176 there is a high probability of an RCU grace period 3177 in progress at any given time. Defaults to 0, 3178 which disables these call_rcu() chains. 3179 3180 locktorture.long_hold= [KNL] 3181 Specify the duration in milliseconds for the 3182 occasional long-duration lock hold time. Defaults 3183 to 100 milliseconds. Select 0 to disable. 3184 3185 locktorture.nested_locks= [KNL] 3186 Specify the maximum lock nesting depth that 3187 locktorture is to exercise, up to a limit of 8 3188 (MAX_NESTED_LOCKS). Specify zero to disable. 3189 Note that this parameter is ineffective on types 3190 of locks that do not support nested acquisition. 3191 3192 locktorture.nreaders_stress= [KNL] 3193 Set the number of locking read-acquisition kthreads. 3194 Defaults to being automatically set based on the 3195 number of online CPUs. 3196 3197 locktorture.nwriters_stress= [KNL] 3198 Set the number of locking write-acquisition kthreads. 3199 3200 locktorture.onoff_holdoff= [KNL] 3201 Set time (s) after boot for CPU-hotplug testing. 3202 3203 locktorture.onoff_interval= [KNL] 3204 Set time (s) between CPU-hotplug operations, or 3205 zero to disable CPU-hotplug testing. 3206 3207 locktorture.rt_boost= [KNL] 3208 Do periodic testing of real-time lock priority 3209 boosting. Select 0 to disable, 1 to boost 3210 only rt_mutex, and 2 to boost unconditionally. 3211 Defaults to 2, which might seem to be an 3212 odd choice, but which should be harmless for 3213 non-real-time spinlocks, due to their disabling 3214 of preemption. Note that non-realtime mutexes 3215 disable boosting. 3216 3217 locktorture.rt_boost_factor= [KNL] 3218 Number that determines how often and for how 3219 long priority boosting is exercised. This is 3220 scaled down by the number of writers, so that the 3221 number of boosts per unit time remains roughly 3222 constant as the number of writers increases. 3223 On the other hand, the duration of each boost 3224 increases with the number of writers. 3225 3226 locktorture.shuffle_interval= [KNL] 3227 Set task-shuffle interval (jiffies). Shuffling 3228 tasks allows some CPUs to go into dyntick-idle 3229 mode during the locktorture test. 3230 3231 locktorture.shutdown_secs= [KNL] 3232 Set time (s) after boot system shutdown. This 3233 is useful for hands-off automated testing. 3234 3235 locktorture.stat_interval= [KNL] 3236 Time (s) between statistics printk()s. 3237 3238 locktorture.stutter= [KNL] 3239 Time (s) to stutter testing, for example, 3240 specifying five seconds causes the test to run for 3241 five seconds, wait for five seconds, and so on. 3242 This tests the locking primitive's ability to 3243 transition abruptly to and from idle. 3244 3245 locktorture.torture_type= [KNL] 3246 Specify the locking implementation to test. 3247 3248 locktorture.verbose= [KNL] 3249 Enable additional printk() statements. 3250 3251 locktorture.writer_fifo= [KNL] 3252 Run the write-side locktorture kthreads at 3253 sched_set_fifo() real-time priority. 3254 3255 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver 3256 Format: <irq> 3257 3258 loglevel= [KNL,EARLY] 3259 All Kernel Messages with a loglevel smaller than the 3260 console loglevel will be printed to the console. It can 3261 also be changed with klogd or other programs. The 3262 loglevels are defined as follows: 3263 3264 0 (KERN_EMERG) system is unusable 3265 1 (KERN_ALERT) action must be taken immediately 3266 2 (KERN_CRIT) critical conditions 3267 3 (KERN_ERR) error conditions 3268 4 (KERN_WARNING) warning conditions 3269 5 (KERN_NOTICE) normal but significant condition 3270 6 (KERN_INFO) informational 3271 7 (KERN_DEBUG) debug-level messages 3272 3273 log_buf_len=n[KMG] [KNL,EARLY] 3274 Sets the size of the printk ring buffer, in bytes. 3275 n must be a power of two and greater than the 3276 minimal size. The minimal size is defined by 3277 LOG_BUF_SHIFT kernel config parameter. There 3278 is also CONFIG_LOG_CPU_MAX_BUF_SHIFT config 3279 parameter that allows to increase the default size 3280 depending on the number of CPUs. See init/Kconfig 3281 for more details. 3282 3283 logo.nologo [FB] Disables display of the built-in Linux logo. 3284 This may be used to provide more screen space for 3285 kernel log messages and is useful when debugging 3286 kernel boot problems. 3287 3288 lp=0 [LP] Specify parallel ports to use, e.g, 3289 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses 3290 lp=reset first parallel port). 'lp=0' disables the 3291 lp=auto printer driver. 'lp=reset' (which can be 3292 specified in addition to the ports) causes 3293 attached printers to be reset. Using 3294 lp=port1,port2,... specifies the parallel ports 3295 to associate lp devices with, starting with 3296 lp0. A port specification may be 'none' to skip 3297 that lp device, or a parport name such as 3298 'parport0'. Specifying 'lp=auto' instead of a 3299 port specification list means that device IDs 3300 from each port should be examined, to see if 3301 an IEEE 1284-compliant printer is attached; if 3302 so, the driver will manage that printer. 3303 See also header of drivers/char/lp.c. 3304 3305 lpj=n [KNL] 3306 Sets loops_per_jiffy to given constant, thus avoiding 3307 time-consuming boot-time autodetection (up to 250 ms per 3308 CPU). 0 enables autodetection (default). To determine 3309 the correct value for your kernel, boot with normal 3310 autodetection and see what value is printed. Note that 3311 on SMP systems the preset will be applied to all CPUs, 3312 which is likely to cause problems if your CPUs need 3313 significantly divergent settings. An incorrect value 3314 will cause delays in the kernel to be wrong, leading to 3315 unpredictable I/O errors and other breakage. Although 3316 unlikely, in the extreme case this might damage your 3317 hardware. 3318 3319 lsm.debug [SECURITY] Enable LSM initialization debugging output. 3320 3321 lsm=lsm1,...,lsmN 3322 [SECURITY] Choose order of LSM initialization. This 3323 overrides CONFIG_LSM, and the "security=" parameter. 3324 3325 machtype= [Loongson] Share the same kernel image file between 3326 different yeeloong laptops. 3327 Example: machtype=lemote-yeeloong-2f-7inch 3328 3329 maxcpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 3330 will bring up during bootup. maxcpus=n : n >= 0 limits 3331 the kernel to bring up 'n' processors. Surely after 3332 bootup you can bring up the other plugged cpu by executing 3333 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus 3334 only takes effect during system bootup. 3335 While n=0 is a special case, it is equivalent to "nosmp", 3336 which also disables the IO APIC. 3337 3338 max_loop= [LOOP] The number of loop block devices that get 3339 (loop.max_loop) unconditionally pre-created at init time. The default 3340 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead 3341 of statically allocating a predefined number, loop 3342 devices can be requested on-demand with the 3343 /dev/loop-control interface. 3344 3345 mce= [X86-{32,64}] 3346 3347 Please see Documentation/arch/x86/x86_64/machinecheck.rst for sysfs runtime tunables. 3348 3349 off 3350 disable machine check 3351 3352 no_cmci 3353 disable CMCI(Corrected Machine Check Interrupt) that 3354 Intel processor supports. Usually this disablement is 3355 not recommended, but it might be handy if your 3356 hardware is misbehaving. 3357 3358 Note that you'll get more problems without CMCI than 3359 with due to the shared banks, i.e. you might get 3360 duplicated error logs. 3361 3362 dont_log_ce 3363 don't make logs for corrected errors. All events 3364 reported as corrected are silently cleared by OS. This 3365 option will be useful if you have no interest in any 3366 of corrected errors. 3367 3368 ignore_ce 3369 disable features for corrected errors, e.g. 3370 polling timer and CMCI. All events reported as 3371 corrected are not cleared by OS and remained in its 3372 error banks. 3373 3374 Usually this disablement is not recommended, however 3375 if there is an agent checking/clearing corrected 3376 errors (e.g. BIOS or hardware monitoring 3377 applications), conflicting with OS's error handling, 3378 and you cannot deactivate the agent, then this option 3379 will be a help. 3380 3381 no_lmce 3382 do not opt-in to Local MCE delivery. Use legacy method 3383 to broadcast MCEs. 3384 3385 bootlog 3386 enable logging of machine checks left over from 3387 booting. Disabled by default on AMD Fam10h and older 3388 because some BIOS leave bogus ones. 3389 3390 If your BIOS doesn't do that it's a good idea to 3391 enable though to make sure you log even machine check 3392 events that result in a reboot. On Intel systems it is 3393 enabled by default. 3394 3395 nobootlog 3396 disable boot machine check logging. 3397 3398 monarchtimeout (number) 3399 sets the time in us to wait for other CPUs on machine 3400 checks. 0 to disable. 3401 3402 bios_cmci_threshold 3403 don't overwrite the bios-set CMCI threshold. This boot 3404 option prevents Linux from overwriting the CMCI 3405 threshold set by the bios. Without this option, Linux 3406 always sets the CMCI threshold to 1. Enabling this may 3407 make memory predictive failure analysis less effective 3408 if the bios sets thresholds for memory errors since we 3409 will not see details for all errors. 3410 3411 recovery 3412 force-enable recoverable machine check code paths 3413 3414 Everything else is in sysfs now. 3415 3416 3417 md= [HW] RAID subsystems devices and level 3418 See Documentation/admin-guide/md.rst. 3419 3420 mdacon= [MDA] 3421 Format: <first>,<last> 3422 Specifies range of consoles to be captured by the MDA. 3423 3424 mds= [X86,INTEL,EARLY] 3425 Control mitigation for the Micro-architectural Data 3426 Sampling (MDS) vulnerability. 3427 3428 Certain CPUs are vulnerable to an exploit against CPU 3429 internal buffers which can forward information to a 3430 disclosure gadget under certain conditions. 3431 3432 In vulnerable processors, the speculatively 3433 forwarded data can be used in a cache side channel 3434 attack, to access data to which the attacker does 3435 not have direct access. 3436 3437 This parameter controls the MDS mitigation. The 3438 options are: 3439 3440 full - Enable MDS mitigation on vulnerable CPUs 3441 full,nosmt - Enable MDS mitigation and disable 3442 SMT on vulnerable CPUs 3443 off - Unconditionally disable MDS mitigation 3444 3445 On TAA-affected machines, mds=off can be prevented by 3446 an active TAA mitigation as both vulnerabilities are 3447 mitigated with the same mechanism so in order to disable 3448 this mitigation, you need to specify tsx_async_abort=off 3449 too. 3450 3451 Not specifying this option is equivalent to 3452 mds=full. 3453 3454 For details see: Documentation/admin-guide/hw-vuln/mds.rst 3455 3456 mem=nn[KMG] [HEXAGON,EARLY] Set the memory size. 3457 Must be specified, otherwise memory size will be 0. 3458 3459 mem=nn[KMG] [KNL,BOOT,EARLY] Force usage of a specific amount 3460 of memory Amount of memory to be used in cases 3461 as follows: 3462 3463 1 for test; 3464 2 when the kernel is not able to see the whole system memory; 3465 3 memory that lies after 'mem=' boundary is excluded from 3466 the hypervisor, then assigned to KVM guests. 3467 4 to limit the memory available for kdump kernel. 3468 3469 [ARC,MICROBLAZE] - the limit applies only to low memory, 3470 high memory is not affected. 3471 3472 [ARM64] - only limits memory covered by the linear 3473 mapping. The NOMAP regions are not affected. 3474 3475 [X86] Work as limiting max address. Use together 3476 with memmap= to avoid physical address space collisions. 3477 Without memmap= PCI devices could be placed at addresses 3478 belonging to unused RAM. 3479 3480 Note that this only takes effects during boot time since 3481 in above case 3, memory may need be hot added after boot 3482 if system memory of hypervisor is not sufficient. 3483 3484 mem=nn[KMG]@ss[KMG] 3485 [ARM,MIPS,EARLY] - override the memory layout 3486 reported by firmware. 3487 Define a memory region of size nn[KMG] starting at 3488 ss[KMG]. 3489 Multiple different regions can be specified with 3490 multiple mem= parameters on the command line. 3491 3492 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel 3493 memory. 3494 3495 memblock=debug [KNL,EARLY] Enable memblock debug messages. 3496 3497 memchunk=nn[KMG] 3498 [KNL,SH] Allow user to override the default size for 3499 per-device physically contiguous DMA buffers. 3500 3501 memhp_default_state=online/offline/online_kernel/online_movable 3502 [KNL] Set the initial state for the memory hotplug 3503 onlining policy. If not specified, the default value is 3504 set according to the 3505 CONFIG_MHP_DEFAULT_ONLINE_TYPE kernel config 3506 options. 3507 See Documentation/admin-guide/mm/memory-hotplug.rst. 3508 3509 memmap=exactmap [KNL,X86,EARLY] Enable setting of an exact 3510 E820 memory map, as specified by the user. 3511 Such memmap=exactmap lines can be constructed based on 3512 BIOS output or other requirements. See the memmap=nn@ss 3513 option description. 3514 3515 memmap=nn[KMG]@ss[KMG] 3516 [KNL, X86,MIPS,XTENSA,EARLY] Force usage of a specific region of memory. 3517 Region of memory to be used is from ss to ss+nn. 3518 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG], 3519 which limits max address to nn[KMG]. 3520 Multiple different regions can be specified, 3521 comma delimited. 3522 Example: 3523 memmap=100M@2G,100M#3G,1G!1024G 3524 3525 memmap=nn[KMG]#ss[KMG] 3526 [KNL,ACPI,EARLY] Mark specific memory as ACPI data. 3527 Region of memory to be marked is from ss to ss+nn. 3528 3529 memmap=nn[KMG]$ss[KMG] 3530 [KNL,ACPI,EARLY] Mark specific memory as reserved. 3531 Region of memory to be reserved is from ss to ss+nn. 3532 Example: Exclude memory from 0x18690000-0x1869ffff 3533 memmap=64K$0x18690000 3534 or 3535 memmap=0x10000$0x18690000 3536 Some bootloaders may need an escape character before '$', 3537 like Grub2, otherwise '$' and the following number 3538 will be eaten. 3539 3540 memmap=nn[KMG]!ss[KMG,EARLY] 3541 [KNL,X86] Mark specific memory as protected. 3542 Region of memory to be used, from ss to ss+nn. 3543 The memory region may be marked as e820 type 12 (0xc) 3544 and is NVDIMM or ADR memory. 3545 3546 memmap=<size>%<offset>-<oldtype>+<newtype> 3547 [KNL,ACPI,EARLY] Convert memory within the specified region 3548 from <oldtype> to <newtype>. If "-<oldtype>" is left 3549 out, the whole region will be marked as <newtype>, 3550 even if previously unavailable. If "+<newtype>" is left 3551 out, matching memory will be removed. Types are 3552 specified as e820 types, e.g., 1 = RAM, 2 = reserved, 3553 3 = ACPI, 12 = PRAM. 3554 3555 memory_corruption_check=0/1 [X86,EARLY] 3556 Some BIOSes seem to corrupt the first 64k of 3557 memory when doing things like suspend/resume. 3558 Setting this option will scan the memory 3559 looking for corruption. Enabling this will 3560 both detect corruption and prevent the kernel 3561 from using the memory being corrupted. 3562 However, its intended as a diagnostic tool; if 3563 repeatable BIOS-originated corruption always 3564 affects the same memory, you can use memmap= 3565 to prevent the kernel from using that memory. 3566 3567 memory_corruption_check_size=size [X86,EARLY] 3568 By default it checks for corruption in the low 3569 64k, making this memory unavailable for normal 3570 use. Use this parameter to scan for 3571 corruption in more or less memory. 3572 3573 memory_corruption_check_period=seconds [X86,EARLY] 3574 By default it checks for corruption every 60 3575 seconds. Use this parameter to check at some 3576 other rate. 0 disables periodic checking. 3577 3578 memory_hotplug.memmap_on_memory 3579 [KNL,X86,ARM] Boolean flag to enable this feature. 3580 Format: {on | off (default)} 3581 When enabled, runtime hotplugged memory will 3582 allocate its internal metadata (struct pages, 3583 those vmemmap pages cannot be optimized even 3584 if hugetlb_free_vmemmap is enabled) from the 3585 hotadded memory which will allow to hotadd a 3586 lot of memory without requiring additional 3587 memory to do so. 3588 This feature is disabled by default because it 3589 has some implication on large (e.g. GB) 3590 allocations in some configurations (e.g. small 3591 memory blocks). 3592 The state of the flag can be read in 3593 /sys/module/memory_hotplug/parameters/memmap_on_memory. 3594 Note that even when enabled, there are a few cases where 3595 the feature is not effective. 3596 3597 memtest= [KNL,X86,ARM,M68K,PPC,RISCV,EARLY] Enable memtest 3598 Format: <integer> 3599 default : 0 <disable> 3600 Specifies the number of memtest passes to be 3601 performed. Each pass selects another test 3602 pattern from a given set of patterns. Memtest 3603 fills the memory with this pattern, validates 3604 memory contents and reserves bad memory 3605 regions that are detected. 3606 3607 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control 3608 Valid arguments: on, off 3609 Default: off 3610 mem_encrypt=on: Activate SME 3611 mem_encrypt=off: Do not activate SME 3612 3613 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst 3614 for details on when memory encryption can be activated. 3615 3616 mem_sleep_default= [SUSPEND] Default system suspend mode: 3617 s2idle - Suspend-To-Idle 3618 shallow - Power-On Suspend or equivalent (if supported) 3619 deep - Suspend-To-RAM or equivalent (if supported) 3620 See Documentation/admin-guide/pm/sleep-states.rst. 3621 3622 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when 3623 the BIOS has incorrectly applied a workaround. TinyBIOS 3624 version 0.98 is known to be affected, 0.99 fixes the 3625 problem by letting the user disable the workaround. 3626 3627 mga= [HW,DRM] 3628 3629 microcode.force_minrev= [X86] 3630 Format: <bool> 3631 Enable or disable the microcode minimal revision 3632 enforcement for the runtime microcode loader. 3633 3634 mini2440= [ARM,HW,KNL] 3635 Format:[0..2][b][c][t] 3636 Default: "0tb" 3637 MINI2440 configuration specification: 3638 0 - The attached screen is the 3.5" TFT 3639 1 - The attached screen is the 7" TFT 3640 2 - The VGA Shield is attached (1024x768) 3641 Leaving out the screen size parameter will not load 3642 the TFT driver, and the framebuffer will be left 3643 unconfigured. 3644 b - Enable backlight. The TFT backlight pin will be 3645 linked to the kernel VESA blanking code and a GPIO 3646 LED. This parameter is not necessary when using the 3647 VGA shield. 3648 c - Enable the s3c camera interface. 3649 t - Reserved for enabling touchscreen support. The 3650 touchscreen support is not enabled in the mainstream 3651 kernel as of 2.6.30, a preliminary port can be found 3652 in the "bleeding edge" mini2440 support kernel at 3653 https://repo.or.cz/w/linux-2.6/mini2440.git 3654 3655 mitigations= 3656 [X86,PPC,S390,ARM64,EARLY] Control optional mitigations for 3657 CPU vulnerabilities. This is a set of curated, 3658 arch-independent options, each of which is an 3659 aggregation of existing arch-specific options. 3660 3661 Note, "mitigations" is supported if and only if the 3662 kernel was built with CPU_MITIGATIONS=y. 3663 3664 off 3665 Disable all optional CPU mitigations. This 3666 improves system performance, but it may also 3667 expose users to several CPU vulnerabilities. 3668 Equivalent to: if nokaslr then kpti=0 [ARM64] 3669 gather_data_sampling=off [X86] 3670 kvm.nx_huge_pages=off [X86] 3671 l1tf=off [X86] 3672 mds=off [X86] 3673 mmio_stale_data=off [X86] 3674 no_entry_flush [PPC] 3675 no_uaccess_flush [PPC] 3676 nobp=0 [S390] 3677 nopti [X86,PPC] 3678 nospectre_bhb [ARM64] 3679 nospectre_v1 [X86,PPC] 3680 nospectre_v2 [X86,PPC,S390,ARM64] 3681 reg_file_data_sampling=off [X86] 3682 retbleed=off [X86] 3683 spec_rstack_overflow=off [X86] 3684 spec_store_bypass_disable=off [X86,PPC] 3685 spectre_bhi=off [X86] 3686 spectre_v2_user=off [X86] 3687 srbds=off [X86,INTEL] 3688 ssbd=force-off [ARM64] 3689 tsx_async_abort=off [X86] 3690 3691 Exceptions: 3692 This does not have any effect on 3693 kvm.nx_huge_pages when 3694 kvm.nx_huge_pages=force. 3695 3696 auto (default) 3697 Mitigate all CPU vulnerabilities, but leave SMT 3698 enabled, even if it's vulnerable. This is for 3699 users who don't want to be surprised by SMT 3700 getting disabled across kernel upgrades, or who 3701 have other ways of avoiding SMT-based attacks. 3702 Equivalent to: (default behavior) 3703 3704 auto,nosmt 3705 Mitigate all CPU vulnerabilities, disabling SMT 3706 if needed. This is for users who always want to 3707 be fully mitigated, even if it means losing SMT. 3708 Equivalent to: l1tf=flush,nosmt [X86] 3709 mds=full,nosmt [X86] 3710 tsx_async_abort=full,nosmt [X86] 3711 mmio_stale_data=full,nosmt [X86] 3712 retbleed=auto,nosmt [X86] 3713 3714 mminit_loglevel= 3715 [KNL,EARLY] When CONFIG_DEBUG_MEMORY_INIT is set, this 3716 parameter allows control of the logging verbosity for 3717 the additional memory initialisation checks. A value 3718 of 0 disables mminit logging and a level of 4 will 3719 log everything. Information is printed at KERN_DEBUG 3720 so loglevel=8 may also need to be specified. 3721 3722 mmio_stale_data= 3723 [X86,INTEL,EARLY] Control mitigation for the Processor 3724 MMIO Stale Data vulnerabilities. 3725 3726 Processor MMIO Stale Data is a class of 3727 vulnerabilities that may expose data after an MMIO 3728 operation. Exposed data could originate or end in 3729 the same CPU buffers as affected by MDS and TAA. 3730 Therefore, similar to MDS and TAA, the mitigation 3731 is to clear the affected CPU buffers. 3732 3733 This parameter controls the mitigation. The 3734 options are: 3735 3736 full - Enable mitigation on vulnerable CPUs 3737 3738 full,nosmt - Enable mitigation and disable SMT on 3739 vulnerable CPUs. 3740 3741 off - Unconditionally disable mitigation 3742 3743 On MDS or TAA affected machines, 3744 mmio_stale_data=off can be prevented by an active 3745 MDS or TAA mitigation as these vulnerabilities are 3746 mitigated with the same mechanism so in order to 3747 disable this mitigation, you need to specify 3748 mds=off and tsx_async_abort=off too. 3749 3750 Not specifying this option is equivalent to 3751 mmio_stale_data=full. 3752 3753 For details see: 3754 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst 3755 3756 <module>.async_probe[=<bool>] [KNL] 3757 If no <bool> value is specified or if the value 3758 specified is not a valid <bool>, enable asynchronous 3759 probe on this module. Otherwise, enable/disable 3760 asynchronous probe on this module as indicated by the 3761 <bool> value. See also: module.async_probe 3762 3763 module.async_probe=<bool> 3764 [KNL] When set to true, modules will use async probing 3765 by default. To enable/disable async probing for a 3766 specific module, use the module specific control that 3767 is documented under <module>.async_probe. When both 3768 module.async_probe and <module>.async_probe are 3769 specified, <module>.async_probe takes precedence for 3770 the specific module. 3771 3772 module.enable_dups_trace 3773 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set, 3774 this means that duplicate request_module() calls will 3775 trigger a WARN_ON() instead of a pr_warn(). Note that 3776 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s 3777 will always be issued and this option does nothing. 3778 module.sig_enforce 3779 [KNL] When CONFIG_MODULE_SIG is set, this means that 3780 modules without (valid) signatures will fail to load. 3781 Note that if CONFIG_MODULE_SIG_FORCE is set, that 3782 is always true, so this option does nothing. 3783 3784 module_blacklist= [KNL] Do not load a comma-separated list of 3785 modules. Useful for debugging problem modules. 3786 3787 mousedev.tap_time= 3788 [MOUSE] Maximum time between finger touching and 3789 leaving touchpad surface for touch to be considered 3790 a tap and be reported as a left button click (for 3791 touchpads working in absolute mode only). 3792 Format: <msecs> 3793 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices 3794 reporting absolute coordinates, such as tablets 3795 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices 3796 reporting absolute coordinates, such as tablets 3797 3798 movablecore= [KNL,X86,PPC,EARLY] 3799 Format: nn[KMGTPE] | nn% 3800 This parameter is the complement to kernelcore=, it 3801 specifies the amount of memory used for migratable 3802 allocations. If both kernelcore and movablecore is 3803 specified, then kernelcore will be at *least* the 3804 specified value but may be more. If movablecore on its 3805 own is specified, the administrator must be careful 3806 that the amount of memory usable for all allocations 3807 is not too small. 3808 3809 movable_node [KNL,EARLY] Boot-time switch to make hotplugable memory 3810 NUMA nodes to be movable. This means that the memory 3811 of such nodes will be usable only for movable 3812 allocations which rules out almost all kernel 3813 allocations. Use with caution! 3814 3815 MTD_Partition= [MTD] 3816 Format: <name>,<region-number>,<size>,<offset> 3817 3818 MTD_Region= [MTD] Format: 3819 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>] 3820 3821 mtdparts= [MTD] 3822 See drivers/mtd/parsers/cmdlinepart.c 3823 3824 mtouchusb.raw_coordinates= 3825 [HW] Make the MicroTouch USB driver use raw coordinates 3826 ('y', default) or cooked coordinates ('n') 3827 3828 mtrr=debug [X86,EARLY] 3829 Enable printing debug information related to MTRR 3830 registers at boot time. 3831 3832 mtrr_chunk_size=nn[KMG,X86,EARLY] 3833 used for mtrr cleanup. It is largest continuous chunk 3834 that could hold holes aka. UC entries. 3835 3836 mtrr_gran_size=nn[KMG,X86,EARLY] 3837 Used for mtrr cleanup. It is granularity of mtrr block. 3838 Default is 1. 3839 Large value could prevent small alignment from 3840 using up MTRRs. 3841 3842 mtrr_spare_reg_nr=n [X86,EARLY] 3843 Format: <integer> 3844 Range: 0,7 : spare reg number 3845 Default : 1 3846 Used for mtrr cleanup. It is spare mtrr entries number. 3847 Set to 2 or more if your graphical card needs more. 3848 3849 multitce=off [PPC] This parameter disables the use of the pSeries 3850 firmware feature for updating multiple TCE entries 3851 at a time. 3852 3853 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card 3854 3855 netdev= [NET] Network devices parameters 3856 Format: <irq>,<io>,<mem_start>,<mem_end>,<name> 3857 Note that mem_start is often overloaded to mean 3858 something different and driver-specific. 3859 This usage is only documented in each driver source 3860 file if at all. 3861 3862 netpoll.carrier_timeout= 3863 [NET] Specifies amount of time (in seconds) that 3864 netpoll should wait for a carrier. By default netpoll 3865 waits 4 seconds. 3866 3867 nf_conntrack.acct= 3868 [NETFILTER] Enable connection tracking flow accounting 3869 0 to disable accounting 3870 1 to enable accounting 3871 Default value is 0. 3872 3873 nfs.cache_getent= 3874 [NFS] sets the pathname to the program which is used 3875 to update the NFS client cache entries. 3876 3877 nfs.cache_getent_timeout= 3878 [NFS] sets the timeout after which an attempt to 3879 update a cache entry is deemed to have failed. 3880 3881 nfs.callback_nr_threads= 3882 [NFSv4] set the total number of threads that the 3883 NFS client will assign to service NFSv4 callback 3884 requests. 3885 3886 nfs.callback_tcpport= 3887 [NFS] set the TCP port on which the NFSv4 callback 3888 channel should listen. 3889 3890 nfs.delay_retrans= 3891 [NFS] specifies the number of times the NFSv4 client 3892 retries the request before returning an EAGAIN error, 3893 after a reply of NFS4ERR_DELAY from the server. 3894 Only applies if the softerr mount option is enabled, 3895 and the specified value is >= 0. 3896 3897 nfs.enable_ino64= 3898 [NFS] enable 64-bit inode numbers. 3899 If zero, the NFS client will fake up a 32-bit inode 3900 number for the readdir() and stat() syscalls instead 3901 of returning the full 64-bit number. 3902 The default is to return 64-bit inode numbers. 3903 3904 nfs.idmap_cache_timeout= 3905 [NFS] set the maximum lifetime for idmapper cache 3906 entries. 3907 3908 nfs.max_session_cb_slots= 3909 [NFSv4.1] Sets the maximum number of session 3910 slots the client will assign to the callback 3911 channel. This determines the maximum number of 3912 callbacks the client will process in parallel for 3913 a particular server. 3914 3915 nfs.max_session_slots= 3916 [NFSv4.1] Sets the maximum number of session slots 3917 the client will attempt to negotiate with the server. 3918 This limits the number of simultaneous RPC requests 3919 that the client can send to the NFSv4.1 server. 3920 Note that there is little point in setting this 3921 value higher than the max_tcp_slot_table_limit. 3922 3923 nfs.nfs4_disable_idmapping= 3924 [NFSv4] When set to the default of '1', this option 3925 ensures that both the RPC level authentication 3926 scheme and the NFS level operations agree to use 3927 numeric uids/gids if the mount is using the 3928 'sec=sys' security flavour. In effect it is 3929 disabling idmapping, which can make migration from 3930 legacy NFSv2/v3 systems to NFSv4 easier. 3931 Servers that do not support this mode of operation 3932 will be autodetected by the client, and it will fall 3933 back to using the idmapper. 3934 To turn off this behaviour, set the value to '0'. 3935 3936 nfs.nfs4_unique_id= 3937 [NFS4] Specify an additional fixed unique ident- 3938 ification string that NFSv4 clients can insert into 3939 their nfs_client_id4 string. This is typically a 3940 UUID that is generated at system install time. 3941 3942 nfs.recover_lost_locks= 3943 [NFSv4] Attempt to recover locks that were lost due 3944 to a lease timeout on the server. Please note that 3945 doing this risks data corruption, since there are 3946 no guarantees that the file will remain unchanged 3947 after the locks are lost. 3948 If you want to enable the kernel legacy behaviour of 3949 attempting to recover these locks, then set this 3950 parameter to '1'. 3951 The default parameter value of '0' causes the kernel 3952 not to attempt recovery of lost locks. 3953 3954 nfs.send_implementation_id= 3955 [NFSv4.1] Send client implementation identification 3956 information in exchange_id requests. 3957 If zero, no implementation identification information 3958 will be sent. 3959 The default is to send the implementation identification 3960 information. 3961 3962 nfs4.layoutstats_timer= 3963 [NFSv4.2] Change the rate at which the kernel sends 3964 layoutstats to the pNFS metadata server. 3965 3966 Setting this to value to 0 causes the kernel to use 3967 whatever value is the default set by the layout 3968 driver. A non-zero value sets the minimum interval 3969 in seconds between layoutstats transmissions. 3970 3971 nfsd.inter_copy_offload_enable= 3972 [NFSv4.2] When set to 1, the server will support 3973 server-to-server copies for which this server is 3974 the destination of the copy. 3975 3976 nfsd.nfs4_disable_idmapping= 3977 [NFSv4] When set to the default of '1', the NFSv4 3978 server will return only numeric uids and gids to 3979 clients using auth_sys, and will accept numeric uids 3980 and gids from such clients. This is intended to ease 3981 migration from NFSv2/v3. 3982 3983 nfsd.nfsd4_ssc_umount_timeout= 3984 [NFSv4.2] When used as the destination of a 3985 server-to-server copy, knfsd temporarily mounts 3986 the source server. It caches the mount in case 3987 it will be needed again, and discards it if not 3988 used for the number of milliseconds specified by 3989 this parameter. 3990 3991 nfsaddrs= [NFS] Deprecated. Use ip= instead. 3992 See Documentation/admin-guide/nfs/nfsroot.rst. 3993 3994 nfsroot= [NFS] nfs root filesystem for disk-less boxes. 3995 See Documentation/admin-guide/nfs/nfsroot.rst. 3996 3997 nfsrootdebug [NFS] enable nfsroot debugging messages. 3998 See Documentation/admin-guide/nfs/nfsroot.rst. 3999 4000 nmi_backtrace.backtrace_idle [KNL] 4001 Dump stacks even of idle CPUs in response to an 4002 NMI stack-backtrace request. 4003 4004 nmi_debug= [KNL,SH] Specify one or more actions to take 4005 when a NMI is triggered. 4006 Format: [state][,regs][,debounce][,die] 4007 4008 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels 4009 Format: [panic,][nopanic,][rNNN,][num] 4010 Valid num: 0 or 1 4011 0 - turn hardlockup detector in nmi_watchdog off 4012 1 - turn hardlockup detector in nmi_watchdog on 4013 rNNN - configure the watchdog with raw perf event 0xNNN 4014 4015 When panic is specified, panic when an NMI watchdog 4016 timeout occurs (or 'nopanic' to not panic on an NMI 4017 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set) 4018 To disable both hard and soft lockup detectors, 4019 please see 'nowatchdog'. 4020 This is useful when you use a panic=... timeout and 4021 need the box quickly up again. 4022 4023 These settings can be accessed at runtime via 4024 the nmi_watchdog and hardlockup_panic sysctls. 4025 4026 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths 4027 emulation library even if a 387 maths coprocessor 4028 is present. 4029 4030 no4lvl [RISCV,EARLY] Disable 4-level and 5-level paging modes. 4031 Forces kernel to use 3-level paging instead. 4032 4033 no5lvl [X86-64,RISCV,EARLY] Disable 5-level paging mode. Forces 4034 kernel to use 4-level paging instead. 4035 4036 noalign [KNL,ARM] 4037 4038 noapic [SMP,APIC,EARLY] Tells the kernel to not make use of any 4039 IOAPICs that may be present in the system. 4040 4041 noapictimer [APIC,X86] Don't set up the APIC timer 4042 4043 noautogroup Disable scheduler automatic task group creation. 4044 4045 nocache [ARM,EARLY] 4046 4047 no_console_suspend 4048 [HW] Never suspend the console 4049 Disable suspending of consoles during suspend and 4050 hibernate operations. Once disabled, debugging 4051 messages can reach various consoles while the rest 4052 of the system is being put to sleep (ie, while 4053 debugging driver suspend/resume hooks). This may 4054 not work reliably with all consoles, but is known 4055 to work with serial and VGA consoles. 4056 To facilitate more flexible debugging, we also add 4057 console_suspend, a printk module parameter to control 4058 it. Users could use console_suspend (usually 4059 /sys/module/printk/parameters/console_suspend) to 4060 turn on/off it dynamically. 4061 4062 no_debug_objects 4063 [KNL,EARLY] Disable object debugging 4064 4065 nodsp [SH] Disable hardware DSP at boot time. 4066 4067 noefi [EFI,EARLY] Disable EFI runtime services support. 4068 4069 no_entry_flush [PPC,EARLY] Don't flush the L1-D cache when entering the kernel. 4070 4071 noexec32 [X86-64] 4072 This affects only 32-bit executables. 4073 noexec32=on: enable non-executable mappings (default) 4074 read doesn't imply executable mappings 4075 noexec32=off: disable non-executable mappings 4076 read implies executable mappings 4077 4078 no_file_caps Tells the kernel not to honor file capabilities. The 4079 only way then for a file to be executed with privilege 4080 is to be setuid root or executed by root. 4081 4082 nofpu [MIPS,SH] Disable hardware FPU at boot time. 4083 4084 nofsgsbase [X86] Disables FSGSBASE instructions. 4085 4086 nofxsr [BUGS=X86-32] Disables x86 floating point extended 4087 register save and restore. The kernel will only save 4088 legacy floating-point registers on task switch. 4089 4090 nogbpages [X86] Do not use GB pages for kernel direct mappings. 4091 4092 no_hash_pointers 4093 [KNL,EARLY] 4094 Force pointers printed to the console or buffers to be 4095 unhashed. By default, when a pointer is printed via %p 4096 format string, that pointer is "hashed", i.e. obscured 4097 by hashing the pointer value. This is a security feature 4098 that hides actual kernel addresses from unprivileged 4099 users, but it also makes debugging the kernel more 4100 difficult since unequal pointers can no longer be 4101 compared. However, if this command-line option is 4102 specified, then all normal pointers will have their true 4103 value printed. This option should only be specified when 4104 debugging the kernel. Please do not use on production 4105 kernels. 4106 4107 nohibernate [HIBERNATION] Disable hibernation and resume. 4108 4109 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,RISCV,SH] Forces the kernel to 4110 busy wait in do_idle() and not use the arch_cpu_idle() 4111 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP 4112 to be effective. This is useful on platforms where the 4113 sleep(SH) or wfi(ARM,ARM64) instructions do not work 4114 correctly or when doing power measurements to evaluate 4115 the impact of the sleep instructions. This is also 4116 useful when using JTAG debugger. 4117 4118 nohpet [X86] Don't use the HPET timer. 4119 4120 nohugeiomap [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge I/O mappings. 4121 4122 nohugevmalloc [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge vmalloc mappings. 4123 4124 nohz= [KNL] Boottime enable/disable dynamic ticks 4125 Valid arguments: on, off 4126 Default: on 4127 4128 nohz_full= [KNL,BOOT,SMP,ISOL] 4129 The argument is a cpu list, as described above. 4130 In kernels built with CONFIG_NO_HZ_FULL=y, set 4131 the specified list of CPUs whose tick will be stopped 4132 whenever possible. The boot CPU will be forced outside 4133 the range to maintain the timekeeping. Any CPUs 4134 in this list will have their RCU callbacks offloaded, 4135 just as if they had also been called out in the 4136 rcu_nocbs= boot parameter. 4137 4138 Note that this argument takes precedence over 4139 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 4140 4141 noinitrd [RAM] Tells the kernel not to load any configured 4142 initial RAM disk. 4143 4144 nointremap [X86-64,Intel-IOMMU,EARLY] Do not enable interrupt 4145 remapping. 4146 [Deprecated - use intremap=off] 4147 4148 noinvpcid [X86,EARLY] Disable the INVPCID cpu feature. 4149 4150 noiotrap [SH] Disables trapped I/O port accesses. 4151 4152 noirqdebug [X86-32] Disables the code which attempts to detect and 4153 disable unhandled interrupt sources. 4154 4155 noisapnp [ISAPNP] Disables ISA PnP code. 4156 4157 nokaslr [KNL,EARLY] 4158 When CONFIG_RANDOMIZE_BASE is set, this disables 4159 kernel and module base offset ASLR (Address Space 4160 Layout Randomization). 4161 4162 no-kvmapf [X86,KVM,EARLY] Disable paravirtualized asynchronous page 4163 fault handling. 4164 4165 no-kvmclock [X86,KVM,EARLY] Disable paravirtualized KVM clock driver 4166 4167 nolapic [X86-32,APIC,EARLY] Do not enable or use the local APIC. 4168 4169 nolapic_timer [X86-32,APIC,EARLY] Do not use the local APIC timer. 4170 4171 nomce [X86-32] Disable Machine Check Exception 4172 4173 nomfgpt [X86-32] Disable Multi-Function General Purpose 4174 Timer usage (for AMD Geode machines). 4175 4176 nomodeset Disable kernel modesetting. Most systems' firmware 4177 sets up a display mode and provides framebuffer memory 4178 for output. With nomodeset, DRM and fbdev drivers will 4179 not load if they could possibly displace the pre- 4180 initialized output. Only the system framebuffer will 4181 be available for use. The respective drivers will not 4182 perform display-mode changes or accelerated rendering. 4183 4184 Useful as error fallback, or for testing and debugging. 4185 4186 nomodule Disable module load 4187 4188 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to 4189 shutdown the other cpus. Instead use the REBOOT_VECTOR 4190 irq. 4191 4192 nopat [X86,EARLY] Disable PAT (page attribute table extension of 4193 pagetables) support. 4194 4195 nopcid [X86-64,EARLY] Disable the PCID cpu feature. 4196 4197 nopku [X86] Disable Memory Protection Keys CPU feature found 4198 in some Intel CPUs. 4199 4200 nopti [X86-64,EARLY] 4201 Equivalent to pti=off 4202 4203 nopv= [X86,XEN,KVM,HYPER_V,VMWARE,EARLY] 4204 Disables the PV optimizations forcing the guest to run 4205 as generic guest with no PV drivers. Currently support 4206 XEN HVM, KVM, HYPER_V and VMWARE guest. 4207 4208 nopvspin [X86,XEN,KVM,EARLY] 4209 Disables the qspinlock slow path using PV optimizations 4210 which allow the hypervisor to 'idle' the guest on lock 4211 contention. 4212 4213 norandmaps Don't use address space randomization. Equivalent to 4214 echo 0 > /proc/sys/kernel/randomize_va_space 4215 4216 noreplace-smp [X86-32,SMP] Don't replace SMP instructions 4217 with UP alternatives 4218 4219 noresume [SWSUSP] Disables resume and restores original swap 4220 space. 4221 4222 no-scroll [VGA] Disables scrollback. 4223 This is required for the Braillex ib80-piezo Braille 4224 reader made by F.H. Papenmeier (Germany). 4225 4226 nosgx [X86-64,SGX,EARLY] Disables Intel SGX kernel support. 4227 4228 nosmap [PPC,EARLY] 4229 Disable SMAP (Supervisor Mode Access Prevention) 4230 even if it is supported by processor. 4231 4232 nosmep [PPC64s,EARLY] 4233 Disable SMEP (Supervisor Mode Execution Prevention) 4234 even if it is supported by processor. 4235 4236 nosmp [SMP,EARLY] Tells an SMP kernel to act as a UP kernel, 4237 and disable the IO APIC. legacy for "maxcpus=0". 4238 4239 nosmt [KNL,MIPS,PPC,S390,EARLY] Disable symmetric multithreading (SMT). 4240 Equivalent to smt=1. 4241 4242 [KNL,X86,PPC] Disable symmetric multithreading (SMT). 4243 nosmt=force: Force disable SMT, cannot be undone 4244 via the sysfs control file. 4245 4246 nosoftlockup [KNL] Disable the soft-lockup detector. 4247 4248 nospec_store_bypass_disable 4249 [HW,EARLY] Disable all mitigations for the Speculative 4250 Store Bypass vulnerability 4251 4252 nospectre_bhb [ARM64,EARLY] Disable all mitigations for Spectre-BHB (branch 4253 history injection) vulnerability. System may allow data leaks 4254 with this option. 4255 4256 nospectre_v1 [X86,PPC,EARLY] Disable mitigations for Spectre Variant 1 4257 (bounds check bypass). With this option data leaks are 4258 possible in the system. 4259 4260 nospectre_v2 [X86,PPC_E500,ARM64,EARLY] Disable all mitigations 4261 for the Spectre variant 2 (indirect branch 4262 prediction) vulnerability. System may allow data 4263 leaks with this option. 4264 4265 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES,RISCV,LOONGARCH,EARLY] 4266 Disable paravirtualized steal time accounting. steal time 4267 is computed, but won't influence scheduler behaviour 4268 4269 nosync [HW,M68K] Disables sync negotiation for all devices. 4270 4271 no_timer_check [X86,APIC] Disables the code which tests for broken 4272 timer IRQ sources, i.e., the IO-APIC timer. This can 4273 work around problems with incorrect timer 4274 initialization on some boards. 4275 4276 no_uaccess_flush 4277 [PPC,EARLY] Don't flush the L1-D cache after accessing user data. 4278 4279 novmcoredd [KNL,KDUMP] 4280 Disable device dump. Device dump allows drivers to 4281 append dump data to vmcore so you can collect driver 4282 specified debug info. Drivers can append the data 4283 without any limit and this data is stored in memory, 4284 so this may cause significant memory stress. Disabling 4285 device dump can help save memory but the driver debug 4286 data will be no longer available. This parameter 4287 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP 4288 is set. 4289 4290 no-vmw-sched-clock 4291 [X86,PV_OPS,EARLY] Disable paravirtualized VMware 4292 scheduler clock and use the default one. 4293 4294 nowatchdog [KNL] Disable both lockup detectors, i.e. 4295 soft-lockup and NMI watchdog (hard-lockup). 4296 4297 nowb [ARM,EARLY] 4298 4299 nox2apic [X86-64,APIC,EARLY] Do not enable x2APIC mode. 4300 4301 NOTE: this parameter will be ignored on systems with the 4302 LEGACY_XAPIC_DISABLED bit set in the 4303 IA32_XAPIC_DISABLE_STATUS MSR. 4304 4305 noxsave [BUGS=X86] Disables x86 extended register state save 4306 and restore using xsave. The kernel will fallback to 4307 enabling legacy floating-point and sse state. 4308 4309 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended 4310 register states. The kernel will fall back to use 4311 xsave to save the states. By using this parameter, 4312 performance of saving the states is degraded because 4313 xsave doesn't support modified optimization while 4314 xsaveopt supports it on xsaveopt enabled systems. 4315 4316 noxsaves [X86] Disables xsaves and xrstors used in saving and 4317 restoring x86 extended register state in compacted 4318 form of xsave area. The kernel will fall back to use 4319 xsaveopt and xrstor to save and restore the states 4320 in standard form of xsave area. By using this 4321 parameter, xsave area per process might occupy more 4322 memory on xsaves enabled systems. 4323 4324 nr_cpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 4325 could support. nr_cpus=n : n >= 1 limits the kernel to 4326 support 'n' processors. It could be larger than the 4327 number of already plugged CPU during bootup, later in 4328 runtime you can physically add extra cpu until it reaches 4329 n. So during boot up some boot time memory for per-cpu 4330 variables need be pre-allocated for later physical cpu 4331 hot plugging. 4332 4333 nr_uarts= [SERIAL] maximum number of UARTs to be registered. 4334 4335 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86, EARLY] 4336 Disable NUMA, Only set up a single NUMA node 4337 spanning all memory. 4338 4339 numa=fake=<size>[MG] 4340 [KNL, ARM64, RISCV, X86, EARLY] 4341 If given as a memory unit, fills all system RAM with 4342 nodes of size interleaved over physical nodes. 4343 4344 numa=fake=<N> 4345 [KNL, ARM64, RISCV, X86, EARLY] 4346 If given as an integer, fills all system RAM with N 4347 fake nodes interleaved over physical nodes. 4348 4349 numa=fake=<N>U 4350 [KNL, ARM64, RISCV, X86, EARLY] 4351 If given as an integer followed by 'U', it will 4352 divide each physical node into N emulated nodes. 4353 4354 numa=noacpi [X86] Don't parse the SRAT table for NUMA setup 4355 4356 numa=nohmat [X86] Don't parse the HMAT table for NUMA setup, or 4357 soft-reserved memory partitioning. 4358 4359 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic 4360 NUMA balancing. 4361 Allowed values are enable and disable 4362 4363 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA. 4364 'node', 'default' can be specified 4365 This can be set from sysctl after boot. 4366 See Documentation/admin-guide/sysctl/vm.rst for details. 4367 4368 ohci1394_dma=early [HW,EARLY] enable debugging via the ohci1394 driver. 4369 See Documentation/core-api/debugging-via-ohci1394.rst for more 4370 info. 4371 4372 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands 4373 Rather than timing out after 20 ms if an EC 4374 command is not properly ACKed, override the length 4375 of the timeout. We have interrupts disabled while 4376 waiting for the ACK, so if this is set too high 4377 interrupts *may* be lost! 4378 4379 omap_mux= [OMAP] Override bootloader pin multiplexing. 4380 Format: <mux_mode0.mode_name=value>... 4381 For example, to override I2C bus2: 4382 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100 4383 4384 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration 4385 4386 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock] 4387 4388 boundary - index of last SLC block on Flex-OneNAND. 4389 The remaining blocks are configured as MLC blocks. 4390 lock - Configure if Flex-OneNAND boundary should be locked. 4391 Once locked, the boundary cannot be changed. 4392 1 indicates lock status, 0 indicates unlock status. 4393 4394 oops=panic [KNL,EARLY] 4395 Always panic on oopses. Default is to just kill the 4396 process, but there is a small probability of 4397 deadlocking the machine. 4398 This will also cause panics on machine check exceptions. 4399 Useful together with panic=30 to trigger a reboot. 4400 4401 page_alloc.shuffle= 4402 [KNL] Boolean flag to control whether the page allocator 4403 should randomize its free lists. This parameter can be 4404 used to enable/disable page randomization. The state of 4405 the flag can be read from sysfs at: 4406 /sys/module/page_alloc/parameters/shuffle. 4407 This parameter is only available if CONFIG_SHUFFLE_PAGE_ALLOCATOR=y. 4408 4409 page_owner= [KNL,EARLY] Boot-time page_owner enabling option. 4410 Storage of the information about who allocated 4411 each page is disabled in default. With this switch, 4412 we can turn it on. 4413 on: enable the feature 4414 4415 page_poison= [KNL,EARLY] Boot-time parameter changing the state of 4416 poisoning on the buddy allocator, available with 4417 CONFIG_PAGE_POISONING=y. 4418 off: turn off poisoning (default) 4419 on: turn on poisoning 4420 4421 page_reporting.page_reporting_order= 4422 [KNL] Minimal page reporting order 4423 Format: <integer> 4424 Adjust the minimal page reporting order. The page 4425 reporting is disabled when it exceeds MAX_PAGE_ORDER. 4426 4427 panic= [KNL] Kernel behaviour on panic: delay <timeout> 4428 timeout > 0: seconds before rebooting 4429 timeout = 0: wait forever 4430 timeout < 0: reboot immediately 4431 Format: <timeout> 4432 4433 panic_on_taint= [KNL,EARLY] 4434 Bitmask for conditionally calling panic() in add_taint() 4435 Format: <hex>[,nousertaint] 4436 Hexadecimal bitmask representing the set of TAINT flags 4437 that will cause the kernel to panic when add_taint() is 4438 called with any of the flags in this set. 4439 The optional switch "nousertaint" can be utilized to 4440 prevent userspace forced crashes by writing to sysctl 4441 /proc/sys/kernel/tainted any flagset matching with the 4442 bitmask set on panic_on_taint. 4443 See Documentation/admin-guide/tainted-kernels.rst for 4444 extra details on the taint flags that users can pick 4445 to compose the bitmask to assign to panic_on_taint. 4446 4447 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump 4448 on a WARN(). 4449 4450 panic_print= Bitmask for printing system info when panic happens. 4451 User can chose combination of the following bits: 4452 bit 0: print all tasks info 4453 bit 1: print system memory info 4454 bit 2: print timer info 4455 bit 3: print locks info if CONFIG_LOCKDEP is on 4456 bit 4: print ftrace buffer 4457 bit 5: print all printk messages in buffer 4458 bit 6: print all CPUs backtrace (if available in the arch) 4459 bit 7: print only tasks in uninterruptible (blocked) state 4460 *Be aware* that this option may print a _lot_ of lines, 4461 so there are risks of losing older messages in the log. 4462 Use this option carefully, maybe worth to setup a 4463 bigger log buffer with "log_buf_len" along with this. 4464 4465 parkbd.port= [HW] Parallel port number the keyboard adapter is 4466 connected to, default is 0. 4467 Format: <parport#> 4468 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation, 4469 0 for XT, 1 for AT (default is AT). 4470 Format: <mode> 4471 4472 parport= [HW,PPT] Specify parallel ports. 0 disables. 4473 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] } 4474 Use 'auto' to force the driver to use any 4475 IRQ/DMA settings detected (the default is to 4476 ignore detected IRQ/DMA settings because of 4477 possible conflicts). You can specify the base 4478 address, IRQ, and DMA settings; IRQ and DMA 4479 should be numbers, or 'auto' (for using detected 4480 settings on that particular port), or 'nofifo' 4481 (to avoid using a FIFO even if it is detected). 4482 Parallel ports are assigned in the order they 4483 are specified on the command line, starting 4484 with parport0. 4485 4486 parport_init_mode= [HW,PPT] 4487 Configure VIA parallel port to operate in 4488 a specific mode. This is necessary on Pegasos 4489 computer where firmware has no options for setting 4490 up parallel port mode and sets it to spp. 4491 Currently this function knows 686a and 8231 chips. 4492 Format: [spp|ps2|epp|ecp|ecpepp] 4493 4494 pata_legacy.all= [HW,LIBATA] 4495 Format: <int> 4496 Set to non-zero to probe primary and secondary ISA 4497 port ranges on PCI systems where no PCI PATA device 4498 has been found at either range. Disabled by default. 4499 4500 pata_legacy.autospeed= [HW,LIBATA] 4501 Format: <int> 4502 Set to non-zero if a chip is present that snoops speed 4503 changes. Disabled by default. 4504 4505 pata_legacy.ht6560a= [HW,LIBATA] 4506 Format: <int> 4507 Set to 1, 2, or 3 for HT 6560A on the primary channel, 4508 the secondary channel, or both channels respectively. 4509 Disabled by default. 4510 4511 pata_legacy.ht6560b= [HW,LIBATA] 4512 Format: <int> 4513 Set to 1, 2, or 3 for HT 6560B on the primary channel, 4514 the secondary channel, or both channels respectively. 4515 Disabled by default. 4516 4517 pata_legacy.iordy_mask= [HW,LIBATA] 4518 Format: <int> 4519 IORDY enable mask. Set individual bits to allow IORDY 4520 for the respective channel. Bit 0 is for the first 4521 legacy channel handled by this driver, bit 1 is for 4522 the second channel, and so on. The sequence will often 4523 correspond to the primary legacy channel, the secondary 4524 legacy channel, and so on, but the handling of a PCI 4525 bus and the use of other driver options may interfere 4526 with the sequence. By default IORDY is allowed across 4527 all channels. 4528 4529 pata_legacy.opti82c46x= [HW,LIBATA] 4530 Format: <int> 4531 Set to 1, 2, or 3 for Opti 82c611A on the primary 4532 channel, the secondary channel, or both channels 4533 respectively. Disabled by default. 4534 4535 pata_legacy.opti82c611a= [HW,LIBATA] 4536 Format: <int> 4537 Set to 1, 2, or 3 for Opti 82c465MV on the primary 4538 channel, the secondary channel, or both channels 4539 respectively. Disabled by default. 4540 4541 pata_legacy.pio_mask= [HW,LIBATA] 4542 Format: <int> 4543 PIO mode mask for autospeed devices. Set individual 4544 bits to allow the use of the respective PIO modes. 4545 Bit 0 is for mode 0, bit 1 is for mode 1, and so on. 4546 All modes allowed by default. 4547 4548 pata_legacy.probe_all= [HW,LIBATA] 4549 Format: <int> 4550 Set to non-zero to probe tertiary and further ISA 4551 port ranges on PCI systems. Disabled by default. 4552 4553 pata_legacy.probe_mask= [HW,LIBATA] 4554 Format: <int> 4555 Probe mask for legacy ISA PATA ports. Depending on 4556 platform configuration and the use of other driver 4557 options up to 6 legacy ports are supported: 0x1f0, 4558 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing 4559 of individual ports can be disabled by setting the 4560 corresponding bits in the mask to 1. Bit 0 is for 4561 the first port in the list above (0x1f0), and so on. 4562 By default all supported ports are probed. 4563 4564 pata_legacy.qdi= [HW,LIBATA] 4565 Format: <int> 4566 Set to non-zero to probe QDI controllers. By default 4567 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise. 4568 4569 pata_legacy.winbond= [HW,LIBATA] 4570 Format: <int> 4571 Set to non-zero to probe Winbond controllers. Use 4572 the standard I/O port (0x130) if 1, otherwise the 4573 value given is the I/O port to use (typically 0x1b0). 4574 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE, 4575 0 otherwise. 4576 4577 pata_platform.pio_mask= [HW,LIBATA] 4578 Format: <int> 4579 Supported PIO mode mask. Set individual bits to allow 4580 the use of the respective PIO modes. Bit 0 is for 4581 mode 0, bit 1 is for mode 1, and so on. Mode 0 only 4582 allowed by default. 4583 4584 pause_on_oops=<int> 4585 Halt all CPUs after the first oops has been printed for 4586 the specified number of seconds. This is to be used if 4587 your oopses keep scrolling off the screen. 4588 4589 pcbit= [HW,ISDN] 4590 4591 pci=option[,option...] [PCI,EARLY] various PCI subsystem options. 4592 4593 Some options herein operate on a specific device 4594 or a set of devices (<pci_dev>). These are 4595 specified in one of the following formats: 4596 4597 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]* 4598 pci:<vendor>:<device>[:<subvendor>:<subdevice>] 4599 4600 Note: the first format specifies a PCI 4601 bus/device/function address which may change 4602 if new hardware is inserted, if motherboard 4603 firmware changes, or due to changes caused 4604 by other kernel parameters. If the 4605 domain is left unspecified, it is 4606 taken to be zero. Optionally, a path 4607 to a device through multiple device/function 4608 addresses can be specified after the base 4609 address (this is more robust against 4610 renumbering issues). The second format 4611 selects devices using IDs from the 4612 configuration space which may match multiple 4613 devices in the system. 4614 4615 earlydump dump PCI config space before the kernel 4616 changes anything 4617 off [X86] don't probe for the PCI bus 4618 bios [X86-32] force use of PCI BIOS, don't access 4619 the hardware directly. Use this if your machine 4620 has a non-standard PCI host bridge. 4621 nobios [X86-32] disallow use of PCI BIOS, only direct 4622 hardware access methods are allowed. Use this 4623 if you experience crashes upon bootup and you 4624 suspect they are caused by the BIOS. 4625 conf1 [X86] Force use of PCI Configuration Access 4626 Mechanism 1 (config address in IO port 0xCF8, 4627 data in IO port 0xCFC, both 32-bit). 4628 conf2 [X86] Force use of PCI Configuration Access 4629 Mechanism 2 (IO port 0xCF8 is an 8-bit port for 4630 the function, IO port 0xCFA, also 8-bit, sets 4631 bus number. The config space is then accessed 4632 through ports 0xC000-0xCFFF). 4633 See http://wiki.osdev.org/PCI for more info 4634 on the configuration access mechanisms. 4635 noaer [PCIE] If the PCIEAER kernel config parameter is 4636 enabled, this kernel boot option can be used to 4637 disable the use of PCIE advanced error reporting. 4638 nodomains [PCI] Disable support for multiple PCI 4639 root domains (aka PCI segments, in ACPI-speak). 4640 nommconf [X86] Disable use of MMCONFIG for PCI 4641 Configuration 4642 check_enable_amd_mmconf [X86] check for and enable 4643 properly configured MMIO access to PCI 4644 config space on AMD family 10h CPU 4645 nomsi [MSI] If the PCI_MSI kernel config parameter is 4646 enabled, this kernel boot option can be used to 4647 disable the use of MSI interrupts system-wide. 4648 noioapicquirk [APIC] Disable all boot interrupt quirks. 4649 Safety option to keep boot IRQs enabled. This 4650 should never be necessary. 4651 ioapicreroute [APIC] Enable rerouting of boot IRQs to the 4652 primary IO-APIC for bridges that cannot disable 4653 boot IRQs. This fixes a source of spurious IRQs 4654 when the system masks IRQs. 4655 noioapicreroute [APIC] Disable workaround that uses the 4656 boot IRQ equivalent of an IRQ that connects to 4657 a chipset where boot IRQs cannot be disabled. 4658 The opposite of ioapicreroute. 4659 biosirq [X86-32] Use PCI BIOS calls to get the interrupt 4660 routing table. These calls are known to be buggy 4661 on several machines and they hang the machine 4662 when used, but on other computers it's the only 4663 way to get the interrupt routing table. Try 4664 this option if the kernel is unable to allocate 4665 IRQs or discover secondary PCI buses on your 4666 motherboard. 4667 rom [X86] Assign address space to expansion ROMs. 4668 Use with caution as certain devices share 4669 address decoders between ROMs and other 4670 resources. 4671 norom [X86] Do not assign address space to 4672 expansion ROMs that do not already have 4673 BIOS assigned address ranges. 4674 nobar [X86] Do not assign address space to the 4675 BARs that weren't assigned by the BIOS. 4676 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be 4677 assigned automatically to PCI devices. You can 4678 make the kernel exclude IRQs of your ISA cards 4679 this way. 4680 pirqaddr=0xAAAAA [X86] Specify the physical address 4681 of the PIRQ table (normally generated 4682 by the BIOS) if it is outside the 4683 F0000h-100000h range. 4684 lastbus=N [X86] Scan all buses thru bus #N. Can be 4685 useful if the kernel is unable to find your 4686 secondary buses and you want to tell it 4687 explicitly which ones they are. 4688 assign-busses [X86] Always assign all PCI bus 4689 numbers ourselves, overriding 4690 whatever the firmware may have done. 4691 usepirqmask [X86] Honor the possible IRQ mask stored 4692 in the BIOS $PIR table. This is needed on 4693 some systems with broken BIOSes, notably 4694 some HP Pavilion N5400 and Omnibook XE3 4695 notebooks. This will have no effect if ACPI 4696 IRQ routing is enabled. 4697 noacpi [X86] Do not use ACPI for IRQ routing 4698 or for PCI scanning. 4699 use_crs [X86] Use PCI host bridge window information 4700 from ACPI. On BIOSes from 2008 or later, this 4701 is enabled by default. If you need to use this, 4702 please report a bug. 4703 nocrs [X86] Ignore PCI host bridge windows from ACPI. 4704 If you need to use this, please report a bug. 4705 use_e820 [X86] Use E820 reservations to exclude parts of 4706 PCI host bridge windows. This is a workaround 4707 for BIOS defects in host bridge _CRS methods. 4708 If you need to use this, please report a bug to 4709 <linux-pci@vger.kernel.org>. 4710 no_e820 [X86] Ignore E820 reservations for PCI host 4711 bridge windows. This is the default on modern 4712 hardware. If you need to use this, please report 4713 a bug to <linux-pci@vger.kernel.org>. 4714 routeirq Do IRQ routing for all PCI devices. 4715 This is normally done in pci_enable_device(), 4716 so this option is a temporary workaround 4717 for broken drivers that don't call it. 4718 skip_isa_align [X86] do not align io start addr, so can 4719 handle more pci cards 4720 noearly [X86] Don't do any early type 1 scanning. 4721 This might help on some broken boards which 4722 machine check when some devices' config space 4723 is read. But various workarounds are disabled 4724 and some IOMMU drivers will not work. 4725 bfsort Sort PCI devices into breadth-first order. 4726 This sorting is done to get a device 4727 order compatible with older (<= 2.4) kernels. 4728 nobfsort Don't sort PCI devices into breadth-first order. 4729 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size) 4730 tuning and use the BIOS-configured MPS defaults. 4731 pcie_bus_safe Set every device's MPS to the largest value 4732 supported by all devices below the root complex. 4733 pcie_bus_perf Set device MPS to the largest allowable MPS 4734 based on its parent bus. Also set MRRS (Max 4735 Read Request Size) to the largest supported 4736 value (no larger than the MPS that the device 4737 or bus can support) for best performance. 4738 pcie_bus_peer2peer Set every device's MPS to 128B, which 4739 every device is guaranteed to support. This 4740 configuration allows peer-to-peer DMA between 4741 any pair of devices, possibly at the cost of 4742 reduced performance. This also guarantees 4743 that hot-added devices will work. 4744 cbiosize=nn[KMG] The fixed amount of bus space which is 4745 reserved for the CardBus bridge's IO window. 4746 The default value is 256 bytes. 4747 cbmemsize=nn[KMG] The fixed amount of bus space which is 4748 reserved for the CardBus bridge's memory 4749 window. The default value is 64 megabytes. 4750 resource_alignment= 4751 Format: 4752 [<order of align>@]<pci_dev>[; ...] 4753 Specifies alignment and device to reassign 4754 aligned memory resources. How to 4755 specify the device is described above. 4756 If <order of align> is not specified, 4757 PAGE_SIZE is used as alignment. 4758 A PCI-PCI bridge can be specified if resource 4759 windows need to be expanded. 4760 To specify the alignment for several 4761 instances of a device, the PCI vendor, 4762 device, subvendor, and subdevice may be 4763 specified, e.g., 12@pci:8086:9c22:103c:198f 4764 for 4096-byte alignment. 4765 ecrc= Enable/disable PCIe ECRC (transaction layer 4766 end-to-end CRC checking). Only effective if 4767 OS has native AER control (either granted by 4768 ACPI _OSC or forced via "pcie_ports=native") 4769 bios: Use BIOS/firmware settings. This is the 4770 the default. 4771 off: Turn ECRC off 4772 on: Turn ECRC on. 4773 hpiosize=nn[KMG] The fixed amount of bus space which is 4774 reserved for hotplug bridge's IO window. 4775 Default size is 256 bytes. 4776 hpmmiosize=nn[KMG] The fixed amount of bus space which is 4777 reserved for hotplug bridge's MMIO window. 4778 Default size is 2 megabytes. 4779 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is 4780 reserved for hotplug bridge's MMIO_PREF window. 4781 Default size is 2 megabytes. 4782 hpmemsize=nn[KMG] The fixed amount of bus space which is 4783 reserved for hotplug bridge's MMIO and 4784 MMIO_PREF window. 4785 Default size is 2 megabytes. 4786 hpbussize=nn The minimum amount of additional bus numbers 4787 reserved for buses below a hotplug bridge. 4788 Default is 1. 4789 realloc= Enable/disable reallocating PCI bridge resources 4790 if allocations done by BIOS are too small to 4791 accommodate resources required by all child 4792 devices. 4793 off: Turn realloc off 4794 on: Turn realloc on 4795 realloc same as realloc=on 4796 noari do not use PCIe ARI. 4797 noats [PCIE, Intel-IOMMU, AMD-IOMMU] 4798 do not use PCIe ATS (and IOMMU device IOTLB). 4799 pcie_scan_all Scan all possible PCIe devices. Otherwise we 4800 only look for one device below a PCIe downstream 4801 port. 4802 big_root_window Try to add a big 64bit memory window to the PCIe 4803 root complex on AMD CPUs. Some GFX hardware 4804 can resize a BAR to allow access to all VRAM. 4805 Adding the window is slightly risky (it may 4806 conflict with unreported devices), so this 4807 taints the kernel. 4808 disable_acs_redir=<pci_dev>[; ...] 4809 Specify one or more PCI devices (in the format 4810 specified above) separated by semicolons. 4811 Each device specified will have the PCI ACS 4812 redirect capabilities forced off which will 4813 allow P2P traffic between devices through 4814 bridges without forcing it upstream. Note: 4815 this removes isolation between devices and 4816 may put more devices in an IOMMU group. 4817 config_acs= 4818 Format: 4819 <ACS flags>@<pci_dev>[; ...] 4820 Specify one or more PCI devices (in the format 4821 specified above) optionally prepended with flags 4822 and separated by semicolons. The respective 4823 capabilities will be enabled, disabled or 4824 unchanged based on what is specified in 4825 flags. 4826 4827 ACS Flags is defined as follows: 4828 bit-0 : ACS Source Validation 4829 bit-1 : ACS Translation Blocking 4830 bit-2 : ACS P2P Request Redirect 4831 bit-3 : ACS P2P Completion Redirect 4832 bit-4 : ACS Upstream Forwarding 4833 bit-5 : ACS P2P Egress Control 4834 bit-6 : ACS Direct Translated P2P 4835 Each bit can be marked as: 4836 '0' – force disabled 4837 '1' – force enabled 4838 'x' – unchanged 4839 For example, 4840 pci=config_acs=10x@pci:0:0 4841 would configure all devices that support 4842 ACS to enable P2P Request Redirect, disable 4843 Translation Blocking, and leave Source 4844 Validation unchanged from whatever power-up 4845 or firmware set it to. 4846 4847 Note: this may remove isolation between devices 4848 and may put more devices in an IOMMU group. 4849 force_floating [S390] Force usage of floating interrupts. 4850 nomio [S390] Do not use MIO instructions. 4851 norid [S390] ignore the RID field and force use of 4852 one PCI domain per PCI function 4853 notph [PCIE] If the PCIE_TPH kernel config parameter 4854 is enabled, this kernel boot option can be used 4855 to disable PCIe TLP Processing Hints support 4856 system-wide. 4857 4858 pcie_aspm= [PCIE] Forcibly enable or ignore PCIe Active State Power 4859 Management. 4860 off Don't touch ASPM configuration at all. Leave any 4861 configuration done by firmware unchanged. 4862 force Enable ASPM even on devices that claim not to support it. 4863 WARNING: Forcing ASPM on may cause system lockups. 4864 4865 pcie_ports= [PCIE] PCIe port services handling: 4866 native Use native PCIe services (PME, AER, DPC, PCIe hotplug) 4867 even if the platform doesn't give the OS permission to 4868 use them. This may cause conflicts if the platform 4869 also tries to use these services. 4870 dpc-native Use native PCIe service for DPC only. May 4871 cause conflicts if firmware uses AER or DPC. 4872 compat Disable native PCIe services (PME, AER, DPC, PCIe 4873 hotplug). 4874 4875 pcie_port_pm= [PCIE] PCIe port power management handling: 4876 off Disable power management of all PCIe ports 4877 force Forcibly enable power management of all PCIe ports 4878 4879 pcie_pme= [PCIE,PM] Native PCIe PME signaling options: 4880 nomsi Do not use MSI for native PCIe PME signaling (this makes 4881 all PCIe root ports use INTx for all services). 4882 4883 pcmv= [HW,PCMCIA] BadgePAD 4 4884 4885 pd_ignore_unused 4886 [PM] 4887 Keep all power-domains already enabled by bootloader on, 4888 even if no driver has claimed them. This is useful 4889 for debug and development, but should not be 4890 needed on a platform with proper driver support. 4891 4892 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at 4893 boot time. 4894 Format: { 0 | 1 } 4895 See arch/parisc/kernel/pdc_chassis.c 4896 4897 percpu_alloc= [MM,EARLY] 4898 Select which percpu first chunk allocator to use. 4899 Currently supported values are "embed" and "page". 4900 Archs may support subset or none of the selections. 4901 See comments in mm/percpu.c for details on each 4902 allocator. This parameter is primarily for debugging 4903 and performance comparison. 4904 4905 pirq= [SMP,APIC] Manual mp-table setup 4906 See Documentation/arch/x86/i386/IO-APIC.rst. 4907 4908 plip= [PPT,NET] Parallel port network link 4909 Format: { parport<nr> | timid | 0 } 4910 See also Documentation/admin-guide/parport.rst. 4911 4912 pmtmr= [X86] Manual setup of pmtmr I/O Port. 4913 Override pmtimer IOPort with a hex value. 4914 e.g. pmtmr=0x508 4915 4916 pmu_override= [PPC] Override the PMU. 4917 This option takes over the PMU facility, so it is no 4918 longer usable by perf. Setting this option starts the 4919 PMU counters by setting MMCR0 to 0 (the FC bit is 4920 cleared). If a number is given, then MMCR1 is set to 4921 that number, otherwise (e.g., 'pmu_override=on'), MMCR1 4922 remains 0. 4923 4924 pm_debug_messages [SUSPEND,KNL] 4925 Enable suspend/resume debug messages during boot up. 4926 4927 pnp.debug=1 [PNP] 4928 Enable PNP debug messages (depends on the 4929 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time 4930 via /sys/module/pnp/parameters/debug. We always show 4931 current resource usage; turning this on also shows 4932 possible settings and some assignment information. 4933 4934 pnpacpi= [ACPI] 4935 { off } 4936 4937 pnpbios= [ISAPNP] 4938 { on | off | curr | res | no-curr | no-res } 4939 4940 pnp_reserve_irq= 4941 [ISAPNP] Exclude IRQs for the autoconfiguration 4942 4943 pnp_reserve_dma= 4944 [ISAPNP] Exclude DMAs for the autoconfiguration 4945 4946 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration 4947 Ranges are in pairs (I/O port base and size). 4948 4949 pnp_reserve_mem= 4950 [ISAPNP] Exclude memory regions for the 4951 autoconfiguration. 4952 Ranges are in pairs (memory base and size). 4953 4954 ports= [IP_VS_FTP] IPVS ftp helper module 4955 Default is 21. 4956 Up to 8 (IP_VS_APP_MAX_PORTS) ports 4957 may be specified. 4958 Format: <port>,<port>.... 4959 4960 possible_cpus= [SMP,S390,X86] 4961 Format: <unsigned int> 4962 Set the number of possible CPUs, overriding the 4963 regular discovery mechanisms (such as ACPI/FW, etc). 4964 4965 powersave=off [PPC] This option disables power saving features. 4966 It specifically disables cpuidle and sets the 4967 platform machine description specific power_save 4968 function to NULL. On Idle the CPU just reduces 4969 execution priority. 4970 4971 ppc_strict_facility_enable 4972 [PPC,ENABLE] This option catches any kernel floating point, 4973 Altivec, VSX and SPE outside of regions specifically 4974 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()). 4975 There is some performance impact when enabling this. 4976 4977 ppc_tm= [PPC,EARLY] 4978 Format: {"off"} 4979 Disable Hardware Transactional Memory 4980 4981 preempt= [KNL] 4982 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC 4983 none - Limited to cond_resched() calls 4984 voluntary - Limited to cond_resched() and might_sleep() calls 4985 full - Any section that isn't explicitly preempt disabled 4986 can be preempted anytime. Tasks will also yield 4987 contended spinlocks (if the critical section isn't 4988 explicitly preempt disabled beyond the lock itself). 4989 lazy - Scheduler controlled. Similar to full but instead 4990 of preempting the task immediately, the task gets 4991 one HZ tick time to yield itself before the 4992 preemption will be forced. One preemption is when the 4993 task returns to user space. 4994 4995 print-fatal-signals= 4996 [KNL] debug: print fatal signals 4997 4998 If enabled, warn about various signal handling 4999 related application anomalies: too many signals, 5000 too many POSIX.1 timers, fatal signals causing a 5001 coredump - etc. 5002 5003 If you hit the warning due to signal overflow, 5004 you might want to try "ulimit -i unlimited". 5005 5006 default: off. 5007 5008 printk.always_kmsg_dump= 5009 Trigger kmsg_dump for cases other than kernel oops or 5010 panics 5011 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5012 default: disabled 5013 5014 printk.console_no_auto_verbose= 5015 Disable console loglevel raise on oops, panic 5016 or lockdep-detected issues (only if lock debug is on). 5017 With an exception to setups with low baudrate on 5018 serial console, keeping this 0 is a good choice 5019 in order to provide more debug information. 5020 Format: <bool> 5021 default: 0 (auto_verbose is enabled) 5022 5023 printk.debug_non_panic_cpus= 5024 Allows storing messages from non-panic CPUs into 5025 the printk log buffer during panic(). They are 5026 flushed to consoles by the panic-CPU on 5027 a best-effort basis. 5028 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5029 Default: disabled 5030 5031 printk.devkmsg={on,off,ratelimit} 5032 Control writing to /dev/kmsg. 5033 on - unlimited logging to /dev/kmsg from userspace 5034 off - logging to /dev/kmsg disabled 5035 ratelimit - ratelimit the logging 5036 Default: ratelimit 5037 5038 printk.time= Show timing data prefixed to each printk message line 5039 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5040 5041 proc_mem.force_override= [KNL] 5042 Format: {always | ptrace | never} 5043 Traditionally /proc/pid/mem allows memory permissions to be 5044 overridden without restrictions. This option may be set to 5045 restrict that. Can be one of: 5046 - 'always': traditional behavior always allows mem overrides. 5047 - 'ptrace': only allow mem overrides for active ptracers. 5048 - 'never': never allow mem overrides. 5049 If not specified, default is the CONFIG_PROC_MEM_* choice. 5050 5051 processor.max_cstate= [HW,ACPI] 5052 Limit processor to maximum C-state 5053 max_cstate=9 overrides any DMI blacklist limit. 5054 5055 processor.nocst [HW,ACPI] 5056 Ignore the _CST method to determine C-states, 5057 instead using the legacy FADT method 5058 5059 profile= [KNL] Enable kernel profiling via /proc/profile 5060 Format: [<profiletype>,]<number> 5061 Param: <profiletype>: "schedule" or "kvm" 5062 [defaults to kernel profiling] 5063 Param: "schedule" - profile schedule points. 5064 Param: "kvm" - profile VM exits. 5065 Param: <number> - step/bucket size as a power of 2 for 5066 statistical time based profiling. 5067 5068 prompt_ramdisk= [RAM] [Deprecated] 5069 5070 prot_virt= [S390] enable hosting protected virtual machines 5071 isolated from the hypervisor (if hardware supports 5072 that). If enabled, the default kernel base address 5073 might be overridden even when Kernel Address Space 5074 Layout Randomization is disabled. 5075 Format: <bool> 5076 5077 psi= [KNL] Enable or disable pressure stall information 5078 tracking. 5079 Format: <bool> 5080 5081 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to 5082 probe for; one of (bare|imps|exps|lifebook|any). 5083 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports 5084 per second. 5085 psmouse.resetafter= [HW,MOUSE] 5086 Try to reset the device after so many bad packets 5087 (0 = never). 5088 psmouse.resolution= 5089 [HW,MOUSE] Set desired mouse resolution, in dpi. 5090 psmouse.smartscroll= 5091 [HW,MOUSE] Controls Logitech smartscroll autorepeat. 5092 0 = disabled, 1 = enabled (default). 5093 5094 pstore.backend= Specify the name of the pstore backend to use 5095 5096 pti= [X86-64] Control Page Table Isolation of user and 5097 kernel address spaces. Disabling this feature 5098 removes hardening, but improves performance of 5099 system calls and interrupts. 5100 5101 on - unconditionally enable 5102 off - unconditionally disable 5103 auto - kernel detects whether your CPU model is 5104 vulnerable to issues that PTI mitigates 5105 5106 Not specifying this option is equivalent to pti=auto. 5107 5108 pty.legacy_count= 5109 [KNL] Number of legacy pty's. Overwrites compiled-in 5110 default number. 5111 5112 quiet [KNL,EARLY] Disable most log messages 5113 5114 r128= [HW,DRM] 5115 5116 radix_hcall_invalidate=on [PPC/PSERIES] 5117 Disable RADIX GTSE feature and use hcall for TLB 5118 invalidate. 5119 5120 raid= [HW,RAID] 5121 See Documentation/admin-guide/md.rst. 5122 5123 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes 5124 See Documentation/admin-guide/blockdev/ramdisk.rst. 5125 5126 ramdisk_start= [RAM] RAM disk image start address 5127 5128 random.trust_cpu=off 5129 [KNL,EARLY] Disable trusting the use of the CPU's 5130 random number generator (if available) to 5131 initialize the kernel's RNG. 5132 5133 random.trust_bootloader=off 5134 [KNL,EARLY] Disable trusting the use of the a seed 5135 passed by the bootloader (if available) to 5136 initialize the kernel's RNG. 5137 5138 randomize_kstack_offset= 5139 [KNL,EARLY] Enable or disable kernel stack offset 5140 randomization, which provides roughly 5 bits of 5141 entropy, frustrating memory corruption attacks 5142 that depend on stack address determinism or 5143 cross-syscall address exposures. This is only 5144 available on architectures that have defined 5145 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET. 5146 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5147 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT. 5148 5149 ras=option[,option,...] [KNL] RAS-specific options 5150 5151 cec_disable [X86] 5152 Disable the Correctable Errors Collector, 5153 see CONFIG_RAS_CEC help text. 5154 5155 rcu_nocbs[=cpu-list] 5156 [KNL] The optional argument is a cpu list, 5157 as described above. 5158 5159 In kernels built with CONFIG_RCU_NOCB_CPU=y, 5160 enable the no-callback CPU mode, which prevents 5161 such CPUs' callbacks from being invoked in 5162 softirq context. Invocation of such CPUs' RCU 5163 callbacks will instead be offloaded to "rcuox/N" 5164 kthreads created for that purpose, where "x" is 5165 "p" for RCU-preempt, "s" for RCU-sched, and "g" 5166 for the kthreads that mediate grace periods; and 5167 "N" is the CPU number. This reduces OS jitter on 5168 the offloaded CPUs, which can be useful for HPC 5169 and real-time workloads. It can also improve 5170 energy efficiency for asymmetric multiprocessors. 5171 5172 If a cpulist is passed as an argument, the specified 5173 list of CPUs is set to no-callback mode from boot. 5174 5175 Otherwise, if the '=' sign and the cpulist 5176 arguments are omitted, no CPU will be set to 5177 no-callback mode from boot but the mode may be 5178 toggled at runtime via cpusets. 5179 5180 Note that this argument takes precedence over 5181 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 5182 5183 rcu_nocb_poll [KNL] 5184 Rather than requiring that offloaded CPUs 5185 (specified by rcu_nocbs= above) explicitly 5186 awaken the corresponding "rcuoN" kthreads, 5187 make these kthreads poll for callbacks. 5188 This improves the real-time response for the 5189 offloaded CPUs by relieving them of the need to 5190 wake up the corresponding kthread, but degrades 5191 energy efficiency by requiring that the kthreads 5192 periodically wake up to do the polling. 5193 5194 rcutree.blimit= [KNL] 5195 Set maximum number of finished RCU callbacks to 5196 process in one batch. 5197 5198 rcutree.csd_lock_suppress_rcu_stall= [KNL] 5199 Do only a one-line RCU CPU stall warning when 5200 there is an ongoing too-long CSD-lock wait. 5201 5202 rcutree.do_rcu_barrier= [KNL] 5203 Request a call to rcu_barrier(). This is 5204 throttled so that userspace tests can safely 5205 hammer on the sysfs variable if they so choose. 5206 If triggered before the RCU grace-period machinery 5207 is fully active, this will error out with EAGAIN. 5208 5209 rcutree.dump_tree= [KNL] 5210 Dump the structure of the rcu_node combining tree 5211 out at early boot. This is used for diagnostic 5212 purposes, to verify correct tree setup. 5213 5214 rcutree.gp_cleanup_delay= [KNL] 5215 Set the number of jiffies to delay each step of 5216 RCU grace-period cleanup. 5217 5218 rcutree.gp_init_delay= [KNL] 5219 Set the number of jiffies to delay each step of 5220 RCU grace-period initialization. 5221 5222 rcutree.gp_preinit_delay= [KNL] 5223 Set the number of jiffies to delay each step of 5224 RCU grace-period pre-initialization, that is, 5225 the propagation of recent CPU-hotplug changes up 5226 the rcu_node combining tree. 5227 5228 rcutree.jiffies_till_first_fqs= [KNL] 5229 Set delay from grace-period initialization to 5230 first attempt to force quiescent states. 5231 Units are jiffies, minimum value is zero, 5232 and maximum value is HZ. 5233 5234 rcutree.jiffies_till_next_fqs= [KNL] 5235 Set delay between subsequent attempts to force 5236 quiescent states. Units are jiffies, minimum 5237 value is one, and maximum value is HZ. 5238 5239 rcutree.jiffies_till_sched_qs= [KNL] 5240 Set required age in jiffies for a 5241 given grace period before RCU starts 5242 soliciting quiescent-state help from 5243 rcu_note_context_switch() and cond_resched(). 5244 If not specified, the kernel will calculate 5245 a value based on the most recent settings 5246 of rcutree.jiffies_till_first_fqs 5247 and rcutree.jiffies_till_next_fqs. 5248 This calculated value may be viewed in 5249 rcutree.jiffies_to_sched_qs. Any attempt to set 5250 rcutree.jiffies_to_sched_qs will be cheerfully 5251 overwritten. 5252 5253 rcutree.kthread_prio= [KNL,BOOT] 5254 Set the SCHED_FIFO priority of the RCU per-CPU 5255 kthreads (rcuc/N). This value is also used for 5256 the priority of the RCU boost threads (rcub/N) 5257 and for the RCU grace-period kthreads (rcu_bh, 5258 rcu_preempt, and rcu_sched). If RCU_BOOST is 5259 set, valid values are 1-99 and the default is 1 5260 (the least-favored priority). Otherwise, when 5261 RCU_BOOST is not set, valid values are 0-99 and 5262 the default is zero (non-realtime operation). 5263 When RCU_NOCB_CPU is set, also adjust the 5264 priority of NOCB callback kthreads. 5265 5266 rcutree.nocb_nobypass_lim_per_jiffy= [KNL] 5267 On callback-offloaded (rcu_nocbs) CPUs, 5268 RCU reduces the lock contention that would 5269 otherwise be caused by callback floods through 5270 use of the ->nocb_bypass list. However, in the 5271 common non-flooded case, RCU queues directly to 5272 the main ->cblist in order to avoid the extra 5273 overhead of the ->nocb_bypass list and its lock. 5274 But if there are too many callbacks queued during 5275 a single jiffy, RCU pre-queues the callbacks into 5276 the ->nocb_bypass queue. The definition of "too 5277 many" is supplied by this kernel boot parameter. 5278 5279 rcutree.nohz_full_patience_delay= [KNL] 5280 On callback-offloaded (rcu_nocbs) CPUs, avoid 5281 disturbing RCU unless the grace period has 5282 reached the specified age in milliseconds. 5283 Defaults to zero. Large values will be capped 5284 at five seconds. All values will be rounded down 5285 to the nearest value representable by jiffies. 5286 5287 rcutree.qhimark= [KNL] 5288 Set threshold of queued RCU callbacks beyond which 5289 batch limiting is disabled. 5290 5291 rcutree.qlowmark= [KNL] 5292 Set threshold of queued RCU callbacks below which 5293 batch limiting is re-enabled. 5294 5295 rcutree.qovld= [KNL] 5296 Set threshold of queued RCU callbacks beyond which 5297 RCU's force-quiescent-state scan will aggressively 5298 enlist help from cond_resched() and sched IPIs to 5299 help CPUs more quickly reach quiescent states. 5300 Set to less than zero to make this be set based 5301 on rcutree.qhimark at boot time and to zero to 5302 disable more aggressive help enlistment. 5303 5304 rcutree.rcu_delay_page_cache_fill_msec= [KNL] 5305 Set the page-cache refill delay (in milliseconds) 5306 in response to low-memory conditions. The range 5307 of permitted values is in the range 0:100000. 5308 5309 rcutree.rcu_divisor= [KNL] 5310 Set the shift-right count to use to compute 5311 the callback-invocation batch limit bl from 5312 the number of callbacks queued on this CPU. 5313 The result will be bounded below by the value of 5314 the rcutree.blimit kernel parameter. Every bl 5315 callbacks, the softirq handler will exit in 5316 order to allow the CPU to do other work. 5317 5318 Please note that this callback-invocation batch 5319 limit applies only to non-offloaded callback 5320 invocation. Offloaded callbacks are instead 5321 invoked in the context of an rcuoc kthread, which 5322 scheduler will preempt as it does any other task. 5323 5324 rcutree.rcu_fanout_exact= [KNL] 5325 Disable autobalancing of the rcu_node combining 5326 tree. This is used by rcutorture, and might 5327 possibly be useful for architectures having high 5328 cache-to-cache transfer latencies. 5329 5330 rcutree.rcu_fanout_leaf= [KNL] 5331 Change the number of CPUs assigned to each 5332 leaf rcu_node structure. Useful for very 5333 large systems, which will choose the value 64, 5334 and for NUMA systems with large remote-access 5335 latencies, which will choose a value aligned 5336 with the appropriate hardware boundaries. 5337 5338 rcutree.rcu_min_cached_objs= [KNL] 5339 Minimum number of objects which are cached and 5340 maintained per one CPU. Object size is equal 5341 to PAGE_SIZE. The cache allows to reduce the 5342 pressure to page allocator, also it makes the 5343 whole algorithm to behave better in low memory 5344 condition. 5345 5346 rcutree.rcu_nocb_gp_stride= [KNL] 5347 Set the number of NOCB callback kthreads in 5348 each group, which defaults to the square root 5349 of the number of CPUs. Larger numbers reduce 5350 the wakeup overhead on the global grace-period 5351 kthread, but increases that same overhead on 5352 each group's NOCB grace-period kthread. 5353 5354 rcutree.rcu_kick_kthreads= [KNL] 5355 Cause the grace-period kthread to get an extra 5356 wake_up() if it sleeps three times longer than 5357 it should at force-quiescent-state time. 5358 This wake_up() will be accompanied by a 5359 WARN_ONCE() splat and an ftrace_dump(). 5360 5361 rcutree.rcu_resched_ns= [KNL] 5362 Limit the time spend invoking a batch of RCU 5363 callbacks to the specified number of nanoseconds. 5364 By default, this limit is checked only once 5365 every 32 callbacks in order to limit the pain 5366 inflicted by local_clock() overhead. 5367 5368 rcutree.rcu_unlock_delay= [KNL] 5369 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels, 5370 this specifies an rcu_read_unlock()-time delay 5371 in microseconds. This defaults to zero. 5372 Larger delays increase the probability of 5373 catching RCU pointer leaks, that is, buggy use 5374 of RCU-protected pointers after the relevant 5375 rcu_read_unlock() has completed. 5376 5377 rcutree.sysrq_rcu= [KNL] 5378 Commandeer a sysrq key to dump out Tree RCU's 5379 rcu_node tree with an eye towards determining 5380 why a new grace period has not yet started. 5381 5382 rcutree.use_softirq= [KNL] 5383 If set to zero, move all RCU_SOFTIRQ processing to 5384 per-CPU rcuc kthreads. Defaults to a non-zero 5385 value, meaning that RCU_SOFTIRQ is used by default. 5386 Specify rcutree.use_softirq=0 to use rcuc kthreads. 5387 5388 But note that CONFIG_PREEMPT_RT=y kernels disable 5389 this kernel boot parameter, forcibly setting it 5390 to zero. 5391 5392 rcutree.enable_rcu_lazy= [KNL] 5393 To save power, batch RCU callbacks and flush after 5394 delay, memory pressure or callback list growing too 5395 big. 5396 5397 rcutree.rcu_normal_wake_from_gp= [KNL] 5398 Reduces a latency of synchronize_rcu() call. This approach 5399 maintains its own track of synchronize_rcu() callers, so it 5400 does not interact with regular callbacks because it does not 5401 use a call_rcu[_hurry]() path. Please note, this is for a 5402 normal grace period. 5403 5404 How to enable it: 5405 5406 echo 1 > /sys/module/rcutree/parameters/rcu_normal_wake_from_gp 5407 or pass a boot parameter "rcutree.rcu_normal_wake_from_gp=1" 5408 5409 Default is 0. 5410 5411 rcuscale.gp_async= [KNL] 5412 Measure performance of asynchronous 5413 grace-period primitives such as call_rcu(). 5414 5415 rcuscale.gp_async_max= [KNL] 5416 Specify the maximum number of outstanding 5417 callbacks per writer thread. When a writer 5418 thread exceeds this limit, it invokes the 5419 corresponding flavor of rcu_barrier() to allow 5420 previously posted callbacks to drain. 5421 5422 rcuscale.gp_exp= [KNL] 5423 Measure performance of expedited synchronous 5424 grace-period primitives. 5425 5426 rcuscale.holdoff= [KNL] 5427 Set test-start holdoff period. The purpose of 5428 this parameter is to delay the start of the 5429 test until boot completes in order to avoid 5430 interference. 5431 5432 rcuscale.kfree_by_call_rcu= [KNL] 5433 In kernels built with CONFIG_RCU_LAZY=y, test 5434 call_rcu() instead of kfree_rcu(). 5435 5436 rcuscale.kfree_mult= [KNL] 5437 Instead of allocating an object of size kfree_obj, 5438 allocate one of kfree_mult * sizeof(kfree_obj). 5439 Defaults to 1. 5440 5441 rcuscale.kfree_rcu_test= [KNL] 5442 Set to measure performance of kfree_rcu() flooding. 5443 5444 rcuscale.kfree_rcu_test_double= [KNL] 5445 Test the double-argument variant of kfree_rcu(). 5446 If this parameter has the same value as 5447 rcuscale.kfree_rcu_test_single, both the single- 5448 and double-argument variants are tested. 5449 5450 rcuscale.kfree_rcu_test_single= [KNL] 5451 Test the single-argument variant of kfree_rcu(). 5452 If this parameter has the same value as 5453 rcuscale.kfree_rcu_test_double, both the single- 5454 and double-argument variants are tested. 5455 5456 rcuscale.kfree_nthreads= [KNL] 5457 The number of threads running loops of kfree_rcu(). 5458 5459 rcuscale.kfree_alloc_num= [KNL] 5460 Number of allocations and frees done in an iteration. 5461 5462 rcuscale.kfree_loops= [KNL] 5463 Number of loops doing rcuscale.kfree_alloc_num number 5464 of allocations and frees. 5465 5466 rcuscale.minruntime= [KNL] 5467 Set the minimum test run time in seconds. This 5468 does not affect the data-collection interval, 5469 but instead allows better measurement of things 5470 like CPU consumption. 5471 5472 rcuscale.nreaders= [KNL] 5473 Set number of RCU readers. The value -1 selects 5474 N, where N is the number of CPUs. A value 5475 "n" less than -1 selects N-n+1, where N is again 5476 the number of CPUs. For example, -2 selects N 5477 (the number of CPUs), -3 selects N+1, and so on. 5478 A value of "n" less than or equal to -N selects 5479 a single reader. 5480 5481 rcuscale.nwriters= [KNL] 5482 Set number of RCU writers. The values operate 5483 the same as for rcuscale.nreaders. 5484 N, where N is the number of CPUs 5485 5486 rcuscale.scale_type= [KNL] 5487 Specify the RCU implementation to test. 5488 5489 rcuscale.shutdown= [KNL] 5490 Shut the system down after performance tests 5491 complete. This is useful for hands-off automated 5492 testing. 5493 5494 rcuscale.verbose= [KNL] 5495 Enable additional printk() statements. 5496 5497 rcuscale.writer_holdoff= [KNL] 5498 Write-side holdoff between grace periods, 5499 in microseconds. The default of zero says 5500 no holdoff. 5501 5502 rcuscale.writer_holdoff_jiffies= [KNL] 5503 Additional write-side holdoff between grace 5504 periods, but in jiffies. The default of zero 5505 says no holdoff. 5506 5507 rcutorture.fqs_duration= [KNL] 5508 Set duration of force_quiescent_state bursts 5509 in microseconds. 5510 5511 rcutorture.fqs_holdoff= [KNL] 5512 Set holdoff time within force_quiescent_state bursts 5513 in microseconds. 5514 5515 rcutorture.fqs_stutter= [KNL] 5516 Set wait time between force_quiescent_state bursts 5517 in seconds. 5518 5519 rcutorture.fwd_progress= [KNL] 5520 Specifies the number of kthreads to be used 5521 for RCU grace-period forward-progress testing 5522 for the types of RCU supporting this notion. 5523 Defaults to 1 kthread, values less than zero or 5524 greater than the number of CPUs cause the number 5525 of CPUs to be used. 5526 5527 rcutorture.fwd_progress_div= [KNL] 5528 Specify the fraction of a CPU-stall-warning 5529 period to do tight-loop forward-progress testing. 5530 5531 rcutorture.fwd_progress_holdoff= [KNL] 5532 Number of seconds to wait between successive 5533 forward-progress tests. 5534 5535 rcutorture.fwd_progress_need_resched= [KNL] 5536 Enclose cond_resched() calls within checks for 5537 need_resched() during tight-loop forward-progress 5538 testing. 5539 5540 rcutorture.gp_cond= [KNL] 5541 Use conditional/asynchronous update-side 5542 normal-grace-period primitives, if available. 5543 5544 rcutorture.gp_cond_exp= [KNL] 5545 Use conditional/asynchronous update-side 5546 expedited-grace-period primitives, if available. 5547 5548 rcutorture.gp_cond_full= [KNL] 5549 Use conditional/asynchronous update-side 5550 normal-grace-period primitives that also take 5551 concurrent expedited grace periods into account, 5552 if available. 5553 5554 rcutorture.gp_cond_exp_full= [KNL] 5555 Use conditional/asynchronous update-side 5556 expedited-grace-period primitives that also take 5557 concurrent normal grace periods into account, 5558 if available. 5559 5560 rcutorture.gp_cond_wi= [KNL] 5561 Nominal wait interval for normal conditional 5562 grace periods (specified by rcutorture's 5563 gp_cond and gp_cond_full module parameters), 5564 in microseconds. The actual wait interval will 5565 be randomly selected to nanosecond granularity up 5566 to this wait interval. Defaults to 16 jiffies, 5567 for example, 16,000 microseconds on a system 5568 with HZ=1000. 5569 5570 rcutorture.gp_cond_wi_exp= [KNL] 5571 Nominal wait interval for expedited conditional 5572 grace periods (specified by rcutorture's 5573 gp_cond_exp and gp_cond_exp_full module 5574 parameters), in microseconds. The actual wait 5575 interval will be randomly selected to nanosecond 5576 granularity up to this wait interval. Defaults to 5577 128 microseconds. 5578 5579 rcutorture.gp_exp= [KNL] 5580 Use expedited update-side primitives, if available. 5581 5582 rcutorture.gp_normal= [KNL] 5583 Use normal (non-expedited) asynchronous 5584 update-side primitives, if available. 5585 5586 rcutorture.gp_poll= [KNL] 5587 Use polled update-side normal-grace-period 5588 primitives, if available. 5589 5590 rcutorture.gp_poll_exp= [KNL] 5591 Use polled update-side expedited-grace-period 5592 primitives, if available. 5593 5594 rcutorture.gp_poll_full= [KNL] 5595 Use polled update-side normal-grace-period 5596 primitives that also take concurrent expedited 5597 grace periods into account, if available. 5598 5599 rcutorture.gp_poll_exp_full= [KNL] 5600 Use polled update-side expedited-grace-period 5601 primitives that also take concurrent normal 5602 grace periods into account, if available. 5603 5604 rcutorture.gp_poll_wi= [KNL] 5605 Nominal wait interval for normal conditional 5606 grace periods (specified by rcutorture's 5607 gp_poll and gp_poll_full module parameters), 5608 in microseconds. The actual wait interval will 5609 be randomly selected to nanosecond granularity up 5610 to this wait interval. Defaults to 16 jiffies, 5611 for example, 16,000 microseconds on a system 5612 with HZ=1000. 5613 5614 rcutorture.gp_poll_wi_exp= [KNL] 5615 Nominal wait interval for expedited conditional 5616 grace periods (specified by rcutorture's 5617 gp_poll_exp and gp_poll_exp_full module 5618 parameters), in microseconds. The actual wait 5619 interval will be randomly selected to nanosecond 5620 granularity up to this wait interval. Defaults to 5621 128 microseconds. 5622 5623 rcutorture.gp_sync= [KNL] 5624 Use normal (non-expedited) synchronous 5625 update-side primitives, if available. If all 5626 of rcutorture.gp_cond=, rcutorture.gp_exp=, 5627 rcutorture.gp_normal=, and rcutorture.gp_sync= 5628 are zero, rcutorture acts as if is interpreted 5629 they are all non-zero. 5630 5631 rcutorture.irqreader= [KNL] 5632 Run RCU readers from irq handlers, or, more 5633 accurately, from a timer handler. Not all RCU 5634 flavors take kindly to this sort of thing. 5635 5636 rcutorture.leakpointer= [KNL] 5637 Leak an RCU-protected pointer out of the reader. 5638 This can of course result in splats, and is 5639 intended to test the ability of things like 5640 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect 5641 such leaks. 5642 5643 rcutorture.n_barrier_cbs= [KNL] 5644 Set callbacks/threads for rcu_barrier() testing. 5645 5646 rcutorture.nfakewriters= [KNL] 5647 Set number of concurrent RCU writers. These just 5648 stress RCU, they don't participate in the actual 5649 test, hence the "fake". 5650 5651 rcutorture.nocbs_nthreads= [KNL] 5652 Set number of RCU callback-offload togglers. 5653 Zero (the default) disables toggling. 5654 5655 rcutorture.nocbs_toggle= [KNL] 5656 Set the delay in milliseconds between successive 5657 callback-offload toggling attempts. 5658 5659 rcutorture.nreaders= [KNL] 5660 Set number of RCU readers. The value -1 selects 5661 N-1, where N is the number of CPUs. A value 5662 "n" less than -1 selects N-n-2, where N is again 5663 the number of CPUs. For example, -2 selects N 5664 (the number of CPUs), -3 selects N+1, and so on. 5665 5666 rcutorture.object_debug= [KNL] 5667 Enable debug-object double-call_rcu() testing. 5668 5669 rcutorture.onoff_holdoff= [KNL] 5670 Set time (s) after boot for CPU-hotplug testing. 5671 5672 rcutorture.onoff_interval= [KNL] 5673 Set time (jiffies) between CPU-hotplug operations, 5674 or zero to disable CPU-hotplug testing. 5675 5676 rcutorture.preempt_duration= [KNL] 5677 Set duration (in milliseconds) of preemptions 5678 by a high-priority FIFO real-time task. Set to 5679 zero (the default) to disable. The CPUs to 5680 preempt are selected randomly from the set that 5681 are online at a given point in time. Races with 5682 CPUs going offline are ignored, with that attempt 5683 at preemption skipped. 5684 5685 rcutorture.preempt_interval= [KNL] 5686 Set interval (in milliseconds, defaulting to one 5687 second) between preemptions by a high-priority 5688 FIFO real-time task. This delay is mediated 5689 by an hrtimer and is further fuzzed to avoid 5690 inadvertent synchronizations. 5691 5692 rcutorture.read_exit_burst= [KNL] 5693 The number of times in a given read-then-exit 5694 episode that a set of read-then-exit kthreads 5695 is spawned. 5696 5697 rcutorture.read_exit_delay= [KNL] 5698 The delay, in seconds, between successive 5699 read-then-exit testing episodes. 5700 5701 rcutorture.reader_flavor= [KNL] 5702 A bit mask indicating which readers to use. 5703 If there is more than one bit set, the readers 5704 are entered from low-order bit up, and are 5705 exited in the opposite order. For SRCU, the 5706 0x1 bit is normal readers, 0x2 NMI-safe readers, 5707 and 0x4 light-weight readers. 5708 5709 rcutorture.shuffle_interval= [KNL] 5710 Set task-shuffle interval (s). Shuffling tasks 5711 allows some CPUs to go into dyntick-idle mode 5712 during the rcutorture test. 5713 5714 rcutorture.shutdown_secs= [KNL] 5715 Set time (s) after boot system shutdown. This 5716 is useful for hands-off automated testing. 5717 5718 rcutorture.stall_cpu= [KNL] 5719 Duration of CPU stall (s) to test RCU CPU stall 5720 warnings, zero to disable. 5721 5722 rcutorture.stall_cpu_block= [KNL] 5723 Sleep while stalling if set. This will result 5724 in warnings from preemptible RCU in addition to 5725 any other stall-related activity. Note that 5726 in kernels built with CONFIG_PREEMPTION=n and 5727 CONFIG_PREEMPT_COUNT=y, this parameter will 5728 cause the CPU to pass through a quiescent state. 5729 Given CONFIG_PREEMPTION=n, this will suppress 5730 RCU CPU stall warnings, but will instead result 5731 in scheduling-while-atomic splats. 5732 5733 Use of this module parameter results in splats. 5734 5735 5736 rcutorture.stall_cpu_holdoff= [KNL] 5737 Time to wait (s) after boot before inducing stall. 5738 5739 rcutorture.stall_cpu_irqsoff= [KNL] 5740 Disable interrupts while stalling if set, but only 5741 on the first stall in the set. 5742 5743 rcutorture.stall_cpu_repeat= [KNL] 5744 Number of times to repeat the stall sequence, 5745 so that rcutorture.stall_cpu_repeat=3 will result 5746 in four stall sequences. 5747 5748 rcutorture.stall_gp_kthread= [KNL] 5749 Duration (s) of forced sleep within RCU 5750 grace-period kthread to test RCU CPU stall 5751 warnings, zero to disable. If both stall_cpu 5752 and stall_gp_kthread are specified, the 5753 kthread is starved first, then the CPU. 5754 5755 rcutorture.stat_interval= [KNL] 5756 Time (s) between statistics printk()s. 5757 5758 rcutorture.stutter= [KNL] 5759 Time (s) to stutter testing, for example, specifying 5760 five seconds causes the test to run for five seconds, 5761 wait for five seconds, and so on. This tests RCU's 5762 ability to transition abruptly to and from idle. 5763 5764 rcutorture.test_boost= [KNL] 5765 Test RCU priority boosting? 0=no, 1=maybe, 2=yes. 5766 "Maybe" means test if the RCU implementation 5767 under test support RCU priority boosting. 5768 5769 rcutorture.test_boost_duration= [KNL] 5770 Duration (s) of each individual boost test. 5771 5772 rcutorture.test_boost_holdoff= [KNL] 5773 Holdoff time (s) from start of test to the start 5774 of RCU priority-boost testing. Defaults to zero, 5775 that is, no holdoff. 5776 5777 rcutorture.test_boost_interval= [KNL] 5778 Interval (s) between each boost test. 5779 5780 rcutorture.test_no_idle_hz= [KNL] 5781 Test RCU's dyntick-idle handling. See also the 5782 rcutorture.shuffle_interval parameter. 5783 5784 rcutorture.torture_type= [KNL] 5785 Specify the RCU implementation to test. 5786 5787 rcutorture.verbose= [KNL] 5788 Enable additional printk() statements. 5789 5790 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL] 5791 Dump ftrace buffer after reporting RCU CPU 5792 stall warning. 5793 5794 rcupdate.rcu_cpu_stall_notifiers= [KNL] 5795 Provide RCU CPU stall notifiers, but see the 5796 warnings in the RCU_CPU_STALL_NOTIFIER Kconfig 5797 option's help text. TL;DR: You almost certainly 5798 do not want rcupdate.rcu_cpu_stall_notifiers. 5799 5800 rcupdate.rcu_cpu_stall_suppress= [KNL] 5801 Suppress RCU CPU stall warning messages. 5802 5803 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL] 5804 Suppress RCU CPU stall warning messages and 5805 rcutorture writer stall warnings that occur 5806 during early boot, that is, during the time 5807 before the init task is spawned. 5808 5809 rcupdate.rcu_cpu_stall_timeout= [KNL] 5810 Set timeout for RCU CPU stall warning messages. 5811 The value is in seconds and the maximum allowed 5812 value is 300 seconds. 5813 5814 rcupdate.rcu_exp_cpu_stall_timeout= [KNL] 5815 Set timeout for expedited RCU CPU stall warning 5816 messages. The value is in milliseconds 5817 and the maximum allowed value is 21000 5818 milliseconds. Please note that this value is 5819 adjusted to an arch timer tick resolution. 5820 Setting this to zero causes the value from 5821 rcupdate.rcu_cpu_stall_timeout to be used (after 5822 conversion from seconds to milliseconds). 5823 5824 rcupdate.rcu_cpu_stall_cputime= [KNL] 5825 Provide statistics on the cputime and count of 5826 interrupts and tasks during the sampling period. For 5827 multiple continuous RCU stalls, all sampling periods 5828 begin at half of the first RCU stall timeout. 5829 5830 rcupdate.rcu_exp_stall_task_details= [KNL] 5831 Print stack dumps of any tasks blocking the 5832 current expedited RCU grace period during an 5833 expedited RCU CPU stall warning. 5834 5835 rcupdate.rcu_expedited= [KNL] 5836 Use expedited grace-period primitives, for 5837 example, synchronize_rcu_expedited() instead 5838 of synchronize_rcu(). This reduces latency, 5839 but can increase CPU utilization, degrade 5840 real-time latency, and degrade energy efficiency. 5841 No effect on CONFIG_TINY_RCU kernels. 5842 5843 rcupdate.rcu_normal= [KNL] 5844 Use only normal grace-period primitives, 5845 for example, synchronize_rcu() instead of 5846 synchronize_rcu_expedited(). This improves 5847 real-time latency, CPU utilization, and 5848 energy efficiency, but can expose users to 5849 increased grace-period latency. This parameter 5850 overrides rcupdate.rcu_expedited. No effect on 5851 CONFIG_TINY_RCU kernels. 5852 5853 rcupdate.rcu_normal_after_boot= [KNL] 5854 Once boot has completed (that is, after 5855 rcu_end_inkernel_boot() has been invoked), use 5856 only normal grace-period primitives. No effect 5857 on CONFIG_TINY_RCU kernels. 5858 5859 But note that CONFIG_PREEMPT_RT=y kernels enables 5860 this kernel boot parameter, forcibly setting 5861 it to the value one, that is, converting any 5862 post-boot attempt at an expedited RCU grace 5863 period to instead use normal non-expedited 5864 grace-period processing. 5865 5866 rcupdate.rcu_task_collapse_lim= [KNL] 5867 Set the maximum number of callbacks present 5868 at the beginning of a grace period that allows 5869 the RCU Tasks flavors to collapse back to using 5870 a single callback queue. This switching only 5871 occurs when rcupdate.rcu_task_enqueue_lim is 5872 set to the default value of -1. 5873 5874 rcupdate.rcu_task_contend_lim= [KNL] 5875 Set the minimum number of callback-queuing-time 5876 lock-contention events per jiffy required to 5877 cause the RCU Tasks flavors to switch to per-CPU 5878 callback queuing. This switching only occurs 5879 when rcupdate.rcu_task_enqueue_lim is set to 5880 the default value of -1. 5881 5882 rcupdate.rcu_task_enqueue_lim= [KNL] 5883 Set the number of callback queues to use for the 5884 RCU Tasks family of RCU flavors. The default 5885 of -1 allows this to be automatically (and 5886 dynamically) adjusted. This parameter is intended 5887 for use in testing. 5888 5889 rcupdate.rcu_task_ipi_delay= [KNL] 5890 Set time in jiffies during which RCU tasks will 5891 avoid sending IPIs, starting with the beginning 5892 of a given grace period. Setting a large 5893 number avoids disturbing real-time workloads, 5894 but lengthens grace periods. 5895 5896 rcupdate.rcu_task_lazy_lim= [KNL] 5897 Number of callbacks on a given CPU that will 5898 cancel laziness on that CPU. Use -1 to disable 5899 cancellation of laziness, but be advised that 5900 doing so increases the danger of OOM due to 5901 callback flooding. 5902 5903 rcupdate.rcu_task_stall_info= [KNL] 5904 Set initial timeout in jiffies for RCU task stall 5905 informational messages, which give some indication 5906 of the problem for those not patient enough to 5907 wait for ten minutes. Informational messages are 5908 only printed prior to the stall-warning message 5909 for a given grace period. Disable with a value 5910 less than or equal to zero. Defaults to ten 5911 seconds. A change in value does not take effect 5912 until the beginning of the next grace period. 5913 5914 rcupdate.rcu_task_stall_info_mult= [KNL] 5915 Multiplier for time interval between successive 5916 RCU task stall informational messages for a given 5917 RCU tasks grace period. This value is clamped 5918 to one through ten, inclusive. It defaults to 5919 the value three, so that the first informational 5920 message is printed 10 seconds into the grace 5921 period, the second at 40 seconds, the third at 5922 160 seconds, and then the stall warning at 600 5923 seconds would prevent a fourth at 640 seconds. 5924 5925 rcupdate.rcu_task_stall_timeout= [KNL] 5926 Set timeout in jiffies for RCU task stall 5927 warning messages. Disable with a value less 5928 than or equal to zero. Defaults to ten minutes. 5929 A change in value does not take effect until 5930 the beginning of the next grace period. 5931 5932 rcupdate.rcu_tasks_lazy_ms= [KNL] 5933 Set timeout in milliseconds RCU Tasks asynchronous 5934 callback batching for call_rcu_tasks(). 5935 A negative value will take the default. A value 5936 of zero will disable batching. Batching is 5937 always disabled for synchronize_rcu_tasks(). 5938 5939 rcupdate.rcu_tasks_trace_lazy_ms= [KNL] 5940 Set timeout in milliseconds RCU Tasks 5941 Trace asynchronous callback batching for 5942 call_rcu_tasks_trace(). A negative value 5943 will take the default. A value of zero will 5944 disable batching. Batching is always disabled 5945 for synchronize_rcu_tasks_trace(). 5946 5947 rcupdate.rcu_self_test= [KNL] 5948 Run the RCU early boot self tests 5949 5950 rdinit= [KNL] 5951 Format: <full_path> 5952 Run specified binary instead of /init from the ramdisk, 5953 used for early userspace startup. See initrd. 5954 5955 rdrand= [X86,EARLY] 5956 force - Override the decision by the kernel to hide the 5957 advertisement of RDRAND support (this affects 5958 certain AMD processors because of buggy BIOS 5959 support, specifically around the suspend/resume 5960 path). 5961 5962 rdt= [HW,X86,RDT] 5963 Turn on/off individual RDT features. List is: 5964 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp, 5965 mba, smba, bmec. 5966 E.g. to turn on cmt and turn off mba use: 5967 rdt=cmt,!mba 5968 5969 reboot= [KNL] 5970 Format (x86 or x86_64): 5971 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \ 5972 [[,]s[mp]#### \ 5973 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \ 5974 [[,]f[orce] 5975 Where reboot_mode is one of warm (soft) or cold (hard) or gpio 5976 (prefix with 'panic_' to set mode for panic 5977 reboot only), 5978 reboot_type is one of bios, acpi, kbd, triple, efi, or pci, 5979 reboot_force is either force or not specified, 5980 reboot_cpu is s[mp]#### with #### being the processor 5981 to be used for rebooting. 5982 5983 acpi 5984 Use the ACPI RESET_REG in the FADT. If ACPI is not 5985 configured or the ACPI reset does not work, the reboot 5986 path attempts the reset using the keyboard controller. 5987 5988 bios 5989 Use the CPU reboot vector for warm reset 5990 5991 cold 5992 Set the cold reboot flag 5993 5994 default 5995 There are some built-in platform specific "quirks" 5996 - you may see: "reboot: <name> series board detected. 5997 Selecting <type> for reboots." In the case where you 5998 think the quirk is in error (e.g. you have newer BIOS, 5999 or newer board) using this option will ignore the 6000 built-in quirk table, and use the generic default 6001 reboot actions. 6002 6003 efi 6004 Use efi reset_system runtime service. If EFI is not 6005 configured or the EFI reset does not work, the reboot 6006 path attempts the reset using the keyboard controller. 6007 6008 force 6009 Don't stop other CPUs on reboot. This can make reboot 6010 more reliable in some cases. 6011 6012 kbd 6013 Use the keyboard controller. cold reset (default) 6014 6015 pci 6016 Use a write to the PCI config space register 0xcf9 to 6017 trigger reboot. 6018 6019 triple 6020 Force a triple fault (init) 6021 6022 warm 6023 Don't set the cold reboot flag 6024 6025 Using warm reset will be much faster especially on big 6026 memory systems because the BIOS will not go through 6027 the memory check. Disadvantage is that not all 6028 hardware will be completely reinitialized on reboot so 6029 there may be boot problems on some systems. 6030 6031 6032 refscale.holdoff= [KNL] 6033 Set test-start holdoff period. The purpose of 6034 this parameter is to delay the start of the 6035 test until boot completes in order to avoid 6036 interference. 6037 6038 refscale.lookup_instances= [KNL] 6039 Number of data elements to use for the forms of 6040 SLAB_TYPESAFE_BY_RCU testing. A negative number 6041 is negated and multiplied by nr_cpu_ids, while 6042 zero specifies nr_cpu_ids. 6043 6044 refscale.loops= [KNL] 6045 Set the number of loops over the synchronization 6046 primitive under test. Increasing this number 6047 reduces noise due to loop start/end overhead, 6048 but the default has already reduced the per-pass 6049 noise to a handful of picoseconds on ca. 2020 6050 x86 laptops. 6051 6052 refscale.nreaders= [KNL] 6053 Set number of readers. The default value of -1 6054 selects N, where N is roughly 75% of the number 6055 of CPUs. A value of zero is an interesting choice. 6056 6057 refscale.nruns= [KNL] 6058 Set number of runs, each of which is dumped onto 6059 the console log. 6060 6061 refscale.readdelay= [KNL] 6062 Set the read-side critical-section duration, 6063 measured in microseconds. 6064 6065 refscale.scale_type= [KNL] 6066 Specify the read-protection implementation to test. 6067 6068 refscale.shutdown= [KNL] 6069 Shut down the system at the end of the performance 6070 test. This defaults to 1 (shut it down) when 6071 refscale is built into the kernel and to 0 (leave 6072 it running) when refscale is built as a module. 6073 6074 refscale.verbose= [KNL] 6075 Enable additional printk() statements. 6076 6077 refscale.verbose_batched= [KNL] 6078 Batch the additional printk() statements. If zero 6079 (the default) or negative, print everything. Otherwise, 6080 print every Nth verbose statement, where N is the value 6081 specified. 6082 6083 regulator_ignore_unused 6084 [REGULATOR] 6085 Prevents regulator framework from disabling regulators 6086 that are unused, due no driver claiming them. This may 6087 be useful for debug and development, but should not be 6088 needed on a platform with proper driver support. 6089 6090 relax_domain_level= 6091 [KNL, SMP] Set scheduler's default relax_domain_level. 6092 See Documentation/admin-guide/cgroup-v1/cpusets.rst. 6093 6094 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory 6095 Format: <base1>,<size1>[,<base2>,<size2>,...] 6096 Reserve I/O ports or memory so the kernel won't use 6097 them. If <base> is less than 0x10000, the region 6098 is assumed to be I/O ports; otherwise it is memory. 6099 6100 reserve_mem= [RAM] 6101 Format: nn[KMG]:<align>:<label> 6102 Reserve physical memory and label it with a name that 6103 other subsystems can use to access it. This is typically 6104 used for systems that do not wipe the RAM, and this command 6105 line will try to reserve the same physical memory on 6106 soft reboots. Note, it is not guaranteed to be the same 6107 location. For example, if anything about the system changes 6108 or if booting a different kernel. It can also fail if KASLR 6109 places the kernel at the location of where the RAM reservation 6110 was from a previous boot, the new reservation will be at a 6111 different location. 6112 Any subsystem using this feature must add a way to verify 6113 that the contents of the physical memory is from a previous 6114 boot, as there may be cases where the memory will not be 6115 located at the same location. 6116 6117 The format is size:align:label for example, to request 6118 12 megabytes of 4096 alignment for ramoops: 6119 6120 reserve_mem=12M:4096:oops ramoops.mem_name=oops 6121 6122 reservetop= [X86-32,EARLY] 6123 Format: nn[KMG] 6124 Reserves a hole at the top of the kernel virtual 6125 address space. 6126 6127 reset_devices [KNL] Force drivers to reset the underlying device 6128 during initialization. 6129 6130 resume= [SWSUSP] 6131 Specify the partition device for software suspend 6132 Format: 6133 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>} 6134 6135 resume_offset= [SWSUSP] 6136 Specify the offset from the beginning of the partition 6137 given by "resume=" at which the swap header is located, 6138 in <PAGE_SIZE> units (needed only for swap files). 6139 See Documentation/power/swsusp-and-swap-files.rst 6140 6141 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to 6142 read the resume files 6143 6144 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up. 6145 Useful for devices that are detected asynchronously 6146 (e.g. USB and MMC devices). 6147 6148 retain_initrd [RAM] Keep initrd memory after extraction. After boot, it will 6149 be accessible via /sys/firmware/initrd. 6150 6151 retbleed= [X86] Control mitigation of RETBleed (Arbitrary 6152 Speculative Code Execution with Return Instructions) 6153 vulnerability. 6154 6155 AMD-based UNRET and IBPB mitigations alone do not stop 6156 sibling threads from influencing the predictions of other 6157 sibling threads. For that reason, STIBP is used on pro- 6158 cessors that support it, and mitigate SMT on processors 6159 that don't. 6160 6161 off - no mitigation 6162 auto - automatically select a migitation 6163 auto,nosmt - automatically select a mitigation, 6164 disabling SMT if necessary for 6165 the full mitigation (only on Zen1 6166 and older without STIBP). 6167 ibpb - On AMD, mitigate short speculation 6168 windows on basic block boundaries too. 6169 Safe, highest perf impact. It also 6170 enables STIBP if present. Not suitable 6171 on Intel. 6172 ibpb,nosmt - Like "ibpb" above but will disable SMT 6173 when STIBP is not available. This is 6174 the alternative for systems which do not 6175 have STIBP. 6176 unret - Force enable untrained return thunks, 6177 only effective on AMD f15h-f17h based 6178 systems. 6179 unret,nosmt - Like unret, but will disable SMT when STIBP 6180 is not available. This is the alternative for 6181 systems which do not have STIBP. 6182 6183 Selecting 'auto' will choose a mitigation method at run 6184 time according to the CPU. 6185 6186 Not specifying this option is equivalent to retbleed=auto. 6187 6188 rfkill.default_state= 6189 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm, 6190 etc. communication is blocked by default. 6191 1 Unblocked. 6192 6193 rfkill.master_switch_mode= 6194 0 The "airplane mode" button does nothing. 6195 1 The "airplane mode" button toggles between everything 6196 blocked and the previous configuration. 6197 2 The "airplane mode" button toggles between everything 6198 blocked and everything unblocked. 6199 6200 ring3mwait=disable 6201 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported 6202 CPUs. 6203 6204 riscv_isa_fallback [RISCV,EARLY] 6205 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit 6206 falling back to detecting extension support by parsing 6207 "riscv,isa" property on devicetree systems when the 6208 replacement properties are not found. See the Kconfig 6209 entry for RISCV_ISA_FALLBACK. 6210 6211 ro [KNL] Mount root device read-only on boot 6212 6213 rodata= [KNL,EARLY] 6214 on Mark read-only kernel memory as read-only (default). 6215 off Leave read-only kernel memory writable for debugging. 6216 full Mark read-only kernel memory and aliases as read-only 6217 [arm64] 6218 6219 rockchip.usb_uart 6220 [EARLY] 6221 Enable the uart passthrough on the designated usb port 6222 on Rockchip SoCs. When active, the signals of the 6223 debug-uart get routed to the D+ and D- pins of the usb 6224 port and the regular usb controller gets disabled. 6225 6226 root= [KNL] Root filesystem 6227 Usually this a a block device specifier of some kind, 6228 see the early_lookup_bdev comment in 6229 block/early-lookup.c for details. 6230 Alternatively this can be "ram" for the legacy initial 6231 ramdisk, "nfs" and "cifs" for root on a network file 6232 system, or "mtd" and "ubi" for mounting from raw flash. 6233 6234 rootdelay= [KNL] Delay (in seconds) to pause before attempting to 6235 mount the root filesystem 6236 6237 rootflags= [KNL] Set root filesystem mount option string 6238 6239 rootfstype= [KNL] Set root filesystem type 6240 6241 rootwait [KNL] Wait (indefinitely) for root device to show up. 6242 Useful for devices that are detected asynchronously 6243 (e.g. USB and MMC devices). 6244 6245 rootwait= [KNL] Maximum time (in seconds) to wait for root device 6246 to show up before attempting to mount the root 6247 filesystem. 6248 6249 rproc_mem=nn[KMG][@address] 6250 [KNL,ARM,CMA] Remoteproc physical memory block. 6251 Memory area to be used by remote processor image, 6252 managed by CMA. 6253 6254 rw [KNL] Mount root device read-write on boot 6255 6256 S [KNL] Run init in single mode 6257 6258 s390_iommu= [HW,S390] 6259 Set s390 IOTLB flushing mode 6260 strict 6261 With strict flushing every unmap operation will result 6262 in an IOTLB flush. Default is lazy flushing before 6263 reuse, which is faster. Deprecated, equivalent to 6264 iommu.strict=1. 6265 6266 s390_iommu_aperture= [KNL,S390] 6267 Specifies the size of the per device DMA address space 6268 accessible through the DMA and IOMMU APIs as a decimal 6269 factor of the size of main memory. 6270 The default is 1 meaning that one can concurrently use 6271 as many DMA addresses as physical memory is installed, 6272 if supported by hardware, and thus map all of memory 6273 once. With a value of 2 one can map all of memory twice 6274 and so on. As a special case a factor of 0 imposes no 6275 restrictions other than those given by hardware at the 6276 cost of significant additional memory use for tables. 6277 6278 sa1100ir [NET] 6279 See drivers/net/irda/sa1100_ir.c. 6280 6281 sched_verbose [KNL,EARLY] Enables verbose scheduler debug messages. 6282 6283 schedstats= [KNL,X86] Enable or disable scheduled statistics. 6284 Allowed values are enable and disable. This feature 6285 incurs a small amount of overhead in the scheduler 6286 but is useful for debugging and performance tuning. 6287 6288 sched_thermal_decay_shift= 6289 [Deprecated] 6290 [KNL, SMP] Set a decay shift for scheduler thermal 6291 pressure signal. Thermal pressure signal follows the 6292 default decay period of other scheduler pelt 6293 signals(usually 32 ms but configurable). Setting 6294 sched_thermal_decay_shift will left shift the decay 6295 period for the thermal pressure signal by the shift 6296 value. 6297 i.e. with the default pelt decay period of 32 ms 6298 sched_thermal_decay_shift thermal pressure decay pr 6299 1 64 ms 6300 2 128 ms 6301 and so on. 6302 Format: integer between 0 and 10 6303 Default is 0. 6304 6305 scftorture.holdoff= [KNL] 6306 Number of seconds to hold off before starting 6307 test. Defaults to zero for module insertion and 6308 to 10 seconds for built-in smp_call_function() 6309 tests. 6310 6311 scftorture.longwait= [KNL] 6312 Request ridiculously long waits randomly selected 6313 up to the chosen limit in seconds. Zero (the 6314 default) disables this feature. Please note 6315 that requesting even small non-zero numbers of 6316 seconds can result in RCU CPU stall warnings, 6317 softlockup complaints, and so on. 6318 6319 scftorture.nthreads= [KNL] 6320 Number of kthreads to spawn to invoke the 6321 smp_call_function() family of functions. 6322 The default of -1 specifies a number of kthreads 6323 equal to the number of CPUs. 6324 6325 scftorture.onoff_holdoff= [KNL] 6326 Number seconds to wait after the start of the 6327 test before initiating CPU-hotplug operations. 6328 6329 scftorture.onoff_interval= [KNL] 6330 Number seconds to wait between successive 6331 CPU-hotplug operations. Specifying zero (which 6332 is the default) disables CPU-hotplug operations. 6333 6334 scftorture.shutdown_secs= [KNL] 6335 The number of seconds following the start of the 6336 test after which to shut down the system. The 6337 default of zero avoids shutting down the system. 6338 Non-zero values are useful for automated tests. 6339 6340 scftorture.stat_interval= [KNL] 6341 The number of seconds between outputting the 6342 current test statistics to the console. A value 6343 of zero disables statistics output. 6344 6345 scftorture.stutter_cpus= [KNL] 6346 The number of jiffies to wait between each change 6347 to the set of CPUs under test. 6348 6349 scftorture.use_cpus_read_lock= [KNL] 6350 Use use_cpus_read_lock() instead of the default 6351 preempt_disable() to disable CPU hotplug 6352 while invoking one of the smp_call_function*() 6353 functions. 6354 6355 scftorture.verbose= [KNL] 6356 Enable additional printk() statements. 6357 6358 scftorture.weight_single= [KNL] 6359 The probability weighting to use for the 6360 smp_call_function_single() function with a zero 6361 "wait" parameter. A value of -1 selects the 6362 default if all other weights are -1. However, 6363 if at least one weight has some other value, a 6364 value of -1 will instead select a weight of zero. 6365 6366 scftorture.weight_single_wait= [KNL] 6367 The probability weighting to use for the 6368 smp_call_function_single() function with a 6369 non-zero "wait" parameter. See weight_single. 6370 6371 scftorture.weight_many= [KNL] 6372 The probability weighting to use for the 6373 smp_call_function_many() function with a zero 6374 "wait" parameter. See weight_single. 6375 Note well that setting a high probability for 6376 this weighting can place serious IPI load 6377 on the system. 6378 6379 scftorture.weight_many_wait= [KNL] 6380 The probability weighting to use for the 6381 smp_call_function_many() function with a 6382 non-zero "wait" parameter. See weight_single 6383 and weight_many. 6384 6385 scftorture.weight_all= [KNL] 6386 The probability weighting to use for the 6387 smp_call_function_all() function with a zero 6388 "wait" parameter. See weight_single and 6389 weight_many. 6390 6391 scftorture.weight_all_wait= [KNL] 6392 The probability weighting to use for the 6393 smp_call_function_all() function with a 6394 non-zero "wait" parameter. See weight_single 6395 and weight_many. 6396 6397 sdw_mclk_divider=[SDW] 6398 Specify the MCLK divider for Intel SoundWire buses in 6399 case the BIOS does not provide the clock rate properly. 6400 6401 skew_tick= [KNL,EARLY] Offset the periodic timer tick per cpu to mitigate 6402 xtime_lock contention on larger systems, and/or RCU lock 6403 contention on all systems with CONFIG_MAXSMP set. 6404 Format: { "0" | "1" } 6405 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1" 6406 1 -- enable. 6407 Note: increases power consumption, thus should only be 6408 enabled if running jitter sensitive (HPC/RT) workloads. 6409 6410 security= [SECURITY] Choose a legacy "major" security module to 6411 enable at boot. This has been deprecated by the 6412 "lsm=" parameter. 6413 6414 selinux= [SELINUX] Disable or enable SELinux at boot time. 6415 Format: { "0" | "1" } 6416 See security/selinux/Kconfig help text. 6417 0 -- disable. 6418 1 -- enable. 6419 Default value is 1. 6420 6421 serialnumber [BUGS=X86-32] 6422 6423 sev=option[,option...] [X86-64] 6424 6425 debug 6426 Enable debug messages. 6427 6428 nosnp 6429 Do not enable SEV-SNP (applies to host/hypervisor 6430 only). Setting 'nosnp' avoids the RMP check overhead 6431 in memory accesses when users do not want to run 6432 SEV-SNP guests. 6433 6434 shapers= [NET] 6435 Maximal number of shapers. 6436 6437 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller 6438 Limit apic dumping. The parameter defines the maximal 6439 number of local apics being dumped. Also it is possible 6440 to set it to "all" by meaning -- no limit here. 6441 Format: { 1 (default) | 2 | ... | all }. 6442 The parameter valid if only apic=debug or 6443 apic=verbose is specified. 6444 Example: apic=debug show_lapic=all 6445 6446 slab_debug[=options[,slabs][;[options[,slabs]]...] [MM] 6447 Enabling slab_debug allows one to determine the 6448 culprit if slab objects become corrupted. Enabling 6449 slab_debug can create guard zones around objects and 6450 may poison objects when not in use. Also tracks the 6451 last alloc / free. For more information see 6452 Documentation/mm/slub.rst. 6453 (slub_debug legacy name also accepted for now) 6454 6455 slab_max_order= [MM] 6456 Determines the maximum allowed order for slabs. 6457 A high setting may cause OOMs due to memory 6458 fragmentation. For more information see 6459 Documentation/mm/slub.rst. 6460 (slub_max_order legacy name also accepted for now) 6461 6462 slab_merge [MM] 6463 Enable merging of slabs with similar size when the 6464 kernel is built without CONFIG_SLAB_MERGE_DEFAULT. 6465 (slub_merge legacy name also accepted for now) 6466 6467 slab_min_objects= [MM] 6468 The minimum number of objects per slab. SLUB will 6469 increase the slab order up to slab_max_order to 6470 generate a sufficiently large slab able to contain 6471 the number of objects indicated. The higher the number 6472 of objects the smaller the overhead of tracking slabs 6473 and the less frequently locks need to be acquired. 6474 For more information see Documentation/mm/slub.rst. 6475 (slub_min_objects legacy name also accepted for now) 6476 6477 slab_min_order= [MM] 6478 Determines the minimum page order for slabs. Must be 6479 lower or equal to slab_max_order. For more information see 6480 Documentation/mm/slub.rst. 6481 (slub_min_order legacy name also accepted for now) 6482 6483 slab_nomerge [MM] 6484 Disable merging of slabs with similar size. May be 6485 necessary if there is some reason to distinguish 6486 allocs to different slabs, especially in hardened 6487 environments where the risk of heap overflows and 6488 layout control by attackers can usually be 6489 frustrated by disabling merging. This will reduce 6490 most of the exposure of a heap attack to a single 6491 cache (risks via metadata attacks are mostly 6492 unchanged). Debug options disable merging on their 6493 own. 6494 For more information see Documentation/mm/slub.rst. 6495 (slub_nomerge legacy name also accepted for now) 6496 6497 slab_strict_numa [MM] 6498 Support memory policies on a per object level 6499 in the slab allocator. The default is for memory 6500 policies to be applied at the folio level when 6501 a new folio is needed or a partial folio is 6502 retrieved from the lists. Increases overhead 6503 in the slab fastpaths but gains more accurate 6504 NUMA kernel object placement which helps with slow 6505 interconnects in NUMA systems. 6506 6507 slram= [HW,MTD] 6508 6509 smart2= [HW] 6510 Format: <io1>[,<io2>[,...,<io8>]] 6511 6512 smp.csd_lock_timeout= [KNL] 6513 Specify the period of time in milliseconds 6514 that smp_call_function() and friends will wait 6515 for a CPU to release the CSD lock. This is 6516 useful when diagnosing bugs involving CPUs 6517 disabling interrupts for extended periods 6518 of time. Defaults to 5,000 milliseconds, and 6519 setting a value of zero disables this feature. 6520 This feature may be more efficiently disabled 6521 using the csdlock_debug- kernel parameter. 6522 6523 smp.panic_on_ipistall= [KNL] 6524 If a csd_lock_timeout extends for more than 6525 the specified number of milliseconds, panic the 6526 system. By default, let CSD-lock acquisition 6527 take as long as they take. Specifying 300,000 6528 for this value provides a 5-minute timeout. 6529 6530 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices 6531 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port 6532 smsc-ircc2.ircc_sir= [HW] SIR base I/O port 6533 smsc-ircc2.ircc_fir= [HW] FIR base I/O port 6534 smsc-ircc2.ircc_irq= [HW] IRQ line 6535 smsc-ircc2.ircc_dma= [HW] DMA channel 6536 smsc-ircc2.ircc_transceiver= [HW] Transceiver type: 6537 0: Toshiba Satellite 1800 (GP data pin select) 6538 1: Fast pin select (default) 6539 2: ATC IRMode 6540 6541 smt= [KNL,MIPS,S390,EARLY] Set the maximum number of threads 6542 (logical CPUs) to use per physical CPU on systems 6543 capable of symmetric multithreading (SMT). Will 6544 be capped to the actual hardware limit. 6545 Format: <integer> 6546 Default: -1 (no limit) 6547 6548 softlockup_panic= 6549 [KNL] Should the soft-lockup detector generate panics. 6550 Format: 0 | 1 6551 6552 A value of 1 instructs the soft-lockup detector 6553 to panic the machine when a soft-lockup occurs. It is 6554 also controlled by the kernel.softlockup_panic sysctl 6555 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the 6556 respective build-time switch to that functionality. 6557 6558 softlockup_all_cpu_backtrace= 6559 [KNL] Should the soft-lockup detector generate 6560 backtraces on all cpus. 6561 Format: 0 | 1 6562 6563 sonypi.*= [HW] Sony Programmable I/O Control Device driver 6564 See Documentation/admin-guide/laptops/sonypi.rst 6565 6566 spectre_bhi= [X86] Control mitigation of Branch History Injection 6567 (BHI) vulnerability. This setting affects the 6568 deployment of the HW BHI control and the SW BHB 6569 clearing sequence. 6570 6571 on - (default) Enable the HW or SW mitigation as 6572 needed. This protects the kernel from 6573 both syscalls and VMs. 6574 vmexit - On systems which don't have the HW mitigation 6575 available, enable the SW mitigation on vmexit 6576 ONLY. On such systems, the host kernel is 6577 protected from VM-originated BHI attacks, but 6578 may still be vulnerable to syscall attacks. 6579 off - Disable the mitigation. 6580 6581 spectre_v2= [X86,EARLY] Control mitigation of Spectre variant 2 6582 (indirect branch speculation) vulnerability. 6583 The default operation protects the kernel from 6584 user space attacks. 6585 6586 on - unconditionally enable, implies 6587 spectre_v2_user=on 6588 off - unconditionally disable, implies 6589 spectre_v2_user=off 6590 auto - kernel detects whether your CPU model is 6591 vulnerable 6592 6593 Selecting 'on' will, and 'auto' may, choose a 6594 mitigation method at run time according to the 6595 CPU, the available microcode, the setting of the 6596 CONFIG_MITIGATION_RETPOLINE configuration option, 6597 and the compiler with which the kernel was built. 6598 6599 Selecting 'on' will also enable the mitigation 6600 against user space to user space task attacks. 6601 Selecting specific mitigation does not force enable 6602 user mitigations. 6603 6604 Selecting 'off' will disable both the kernel and 6605 the user space protections. 6606 6607 Specific mitigations can also be selected manually: 6608 6609 retpoline - replace indirect branches 6610 retpoline,generic - Retpolines 6611 retpoline,lfence - LFENCE; indirect branch 6612 retpoline,amd - alias for retpoline,lfence 6613 eibrs - Enhanced/Auto IBRS 6614 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines 6615 eibrs,lfence - Enhanced/Auto IBRS + LFENCE 6616 ibrs - use IBRS to protect kernel 6617 6618 Not specifying this option is equivalent to 6619 spectre_v2=auto. 6620 6621 spectre_v2_user= 6622 [X86] Control mitigation of Spectre variant 2 6623 (indirect branch speculation) vulnerability between 6624 user space tasks 6625 6626 on - Unconditionally enable mitigations. Is 6627 enforced by spectre_v2=on 6628 6629 off - Unconditionally disable mitigations. Is 6630 enforced by spectre_v2=off 6631 6632 prctl - Indirect branch speculation is enabled, 6633 but mitigation can be enabled via prctl 6634 per thread. The mitigation control state 6635 is inherited on fork. 6636 6637 prctl,ibpb 6638 - Like "prctl" above, but only STIBP is 6639 controlled per thread. IBPB is issued 6640 always when switching between different user 6641 space processes. 6642 6643 seccomp 6644 - Same as "prctl" above, but all seccomp 6645 threads will enable the mitigation unless 6646 they explicitly opt out. 6647 6648 seccomp,ibpb 6649 - Like "seccomp" above, but only STIBP is 6650 controlled per thread. IBPB is issued 6651 always when switching between different 6652 user space processes. 6653 6654 auto - Kernel selects the mitigation depending on 6655 the available CPU features and vulnerability. 6656 6657 Default mitigation: "prctl" 6658 6659 Not specifying this option is equivalent to 6660 spectre_v2_user=auto. 6661 6662 spec_rstack_overflow= 6663 [X86,EARLY] Control RAS overflow mitigation on AMD Zen CPUs 6664 6665 off - Disable mitigation 6666 microcode - Enable microcode mitigation only 6667 safe-ret - Enable sw-only safe RET mitigation (default) 6668 ibpb - Enable mitigation by issuing IBPB on 6669 kernel entry 6670 ibpb-vmexit - Issue IBPB only on VMEXIT 6671 (cloud-specific mitigation) 6672 6673 spec_store_bypass_disable= 6674 [HW,EARLY] Control Speculative Store Bypass (SSB) Disable mitigation 6675 (Speculative Store Bypass vulnerability) 6676 6677 Certain CPUs are vulnerable to an exploit against a 6678 a common industry wide performance optimization known 6679 as "Speculative Store Bypass" in which recent stores 6680 to the same memory location may not be observed by 6681 later loads during speculative execution. The idea 6682 is that such stores are unlikely and that they can 6683 be detected prior to instruction retirement at the 6684 end of a particular speculation execution window. 6685 6686 In vulnerable processors, the speculatively forwarded 6687 store can be used in a cache side channel attack, for 6688 example to read memory to which the attacker does not 6689 directly have access (e.g. inside sandboxed code). 6690 6691 This parameter controls whether the Speculative Store 6692 Bypass optimization is used. 6693 6694 On x86 the options are: 6695 6696 on - Unconditionally disable Speculative Store Bypass 6697 off - Unconditionally enable Speculative Store Bypass 6698 auto - Kernel detects whether the CPU model contains an 6699 implementation of Speculative Store Bypass and 6700 picks the most appropriate mitigation. If the 6701 CPU is not vulnerable, "off" is selected. If the 6702 CPU is vulnerable the default mitigation is 6703 architecture and Kconfig dependent. See below. 6704 prctl - Control Speculative Store Bypass per thread 6705 via prctl. Speculative Store Bypass is enabled 6706 for a process by default. The state of the control 6707 is inherited on fork. 6708 seccomp - Same as "prctl" above, but all seccomp threads 6709 will disable SSB unless they explicitly opt out. 6710 6711 Default mitigations: 6712 X86: "prctl" 6713 6714 On powerpc the options are: 6715 6716 on,auto - On Power8 and Power9 insert a store-forwarding 6717 barrier on kernel entry and exit. On Power7 6718 perform a software flush on kernel entry and 6719 exit. 6720 off - No action. 6721 6722 Not specifying this option is equivalent to 6723 spec_store_bypass_disable=auto. 6724 6725 split_lock_detect= 6726 [X86] Enable split lock detection or bus lock detection 6727 6728 When enabled (and if hardware support is present), atomic 6729 instructions that access data across cache line 6730 boundaries will result in an alignment check exception 6731 for split lock detection or a debug exception for 6732 bus lock detection. 6733 6734 off - not enabled 6735 6736 warn - the kernel will emit rate-limited warnings 6737 about applications triggering the #AC 6738 exception or the #DB exception. This mode is 6739 the default on CPUs that support split lock 6740 detection or bus lock detection. Default 6741 behavior is by #AC if both features are 6742 enabled in hardware. 6743 6744 fatal - the kernel will send SIGBUS to applications 6745 that trigger the #AC exception or the #DB 6746 exception. Default behavior is by #AC if 6747 both features are enabled in hardware. 6748 6749 ratelimit:N - 6750 Set system wide rate limit to N bus locks 6751 per second for bus lock detection. 6752 0 < N <= 1000. 6753 6754 N/A for split lock detection. 6755 6756 6757 If an #AC exception is hit in the kernel or in 6758 firmware (i.e. not while executing in user mode) 6759 the kernel will oops in either "warn" or "fatal" 6760 mode. 6761 6762 #DB exception for bus lock is triggered only when 6763 CPL > 0. 6764 6765 srbds= [X86,INTEL,EARLY] 6766 Control the Special Register Buffer Data Sampling 6767 (SRBDS) mitigation. 6768 6769 Certain CPUs are vulnerable to an MDS-like 6770 exploit which can leak bits from the random 6771 number generator. 6772 6773 By default, this issue is mitigated by 6774 microcode. However, the microcode fix can cause 6775 the RDRAND and RDSEED instructions to become 6776 much slower. Among other effects, this will 6777 result in reduced throughput from /dev/urandom. 6778 6779 The microcode mitigation can be disabled with 6780 the following option: 6781 6782 off: Disable mitigation and remove 6783 performance impact to RDRAND and RDSEED 6784 6785 srcutree.big_cpu_lim [KNL] 6786 Specifies the number of CPUs constituting a 6787 large system, such that srcu_struct structures 6788 should immediately allocate an srcu_node array. 6789 This kernel-boot parameter defaults to 128, 6790 but takes effect only when the low-order four 6791 bits of srcutree.convert_to_big is equal to 3 6792 (decide at boot). 6793 6794 srcutree.convert_to_big [KNL] 6795 Specifies under what conditions an SRCU tree 6796 srcu_struct structure will be converted to big 6797 form, that is, with an rcu_node tree: 6798 6799 0: Never. 6800 1: At init_srcu_struct() time. 6801 2: When rcutorture decides to. 6802 3: Decide at boot time (default). 6803 0x1X: Above plus if high contention. 6804 6805 Either way, the srcu_node tree will be sized based 6806 on the actual runtime number of CPUs (nr_cpu_ids) 6807 instead of the compile-time CONFIG_NR_CPUS. 6808 6809 srcutree.counter_wrap_check [KNL] 6810 Specifies how frequently to check for 6811 grace-period sequence counter wrap for the 6812 srcu_data structure's ->srcu_gp_seq_needed field. 6813 The greater the number of bits set in this kernel 6814 parameter, the less frequently counter wrap will 6815 be checked for. Note that the bottom two bits 6816 are ignored. 6817 6818 srcutree.exp_holdoff [KNL] 6819 Specifies how many nanoseconds must elapse 6820 since the end of the last SRCU grace period for 6821 a given srcu_struct until the next normal SRCU 6822 grace period will be considered for automatic 6823 expediting. Set to zero to disable automatic 6824 expediting. 6825 6826 srcutree.srcu_max_nodelay [KNL] 6827 Specifies the number of no-delay instances 6828 per jiffy for which the SRCU grace period 6829 worker thread will be rescheduled with zero 6830 delay. Beyond this limit, worker thread will 6831 be rescheduled with a sleep delay of one jiffy. 6832 6833 srcutree.srcu_max_nodelay_phase [KNL] 6834 Specifies the per-grace-period phase, number of 6835 non-sleeping polls of readers. Beyond this limit, 6836 grace period worker thread will be rescheduled 6837 with a sleep delay of one jiffy, between each 6838 rescan of the readers, for a grace period phase. 6839 6840 srcutree.srcu_retry_check_delay [KNL] 6841 Specifies number of microseconds of non-sleeping 6842 delay between each non-sleeping poll of readers. 6843 6844 srcutree.small_contention_lim [KNL] 6845 Specifies the number of update-side contention 6846 events per jiffy will be tolerated before 6847 initiating a conversion of an srcu_struct 6848 structure to big form. Note that the value of 6849 srcutree.convert_to_big must have the 0x10 bit 6850 set for contention-based conversions to occur. 6851 6852 ssbd= [ARM64,HW,EARLY] 6853 Speculative Store Bypass Disable control 6854 6855 On CPUs that are vulnerable to the Speculative 6856 Store Bypass vulnerability and offer a 6857 firmware based mitigation, this parameter 6858 indicates how the mitigation should be used: 6859 6860 force-on: Unconditionally enable mitigation for 6861 for both kernel and userspace 6862 force-off: Unconditionally disable mitigation for 6863 for both kernel and userspace 6864 kernel: Always enable mitigation in the 6865 kernel, and offer a prctl interface 6866 to allow userspace to register its 6867 interest in being mitigated too. 6868 6869 stack_guard_gap= [MM] 6870 override the default stack gap protection. The value 6871 is in page units and it defines how many pages prior 6872 to (for stacks growing down) resp. after (for stacks 6873 growing up) the main stack are reserved for no other 6874 mapping. Default value is 256 pages. 6875 6876 stack_depot_disable= [KNL,EARLY] 6877 Setting this to true through kernel command line will 6878 disable the stack depot thereby saving the static memory 6879 consumed by the stack hash table. By default this is set 6880 to false. 6881 6882 stacktrace [FTRACE] 6883 Enabled the stack tracer on boot up. 6884 6885 stacktrace_filter=[function-list] 6886 [FTRACE] Limit the functions that the stack tracer 6887 will trace at boot up. function-list is a comma-separated 6888 list of functions. This list can be changed at run 6889 time by the stack_trace_filter file in the debugfs 6890 tracing directory. Note, this enables stack tracing 6891 and the stacktrace above is not needed. 6892 6893 sti= [PARISC,HW] 6894 Format: <num> 6895 Set the STI (builtin display/keyboard on the HP-PARISC 6896 machines) console (graphic card) which should be used 6897 as the initial boot-console. 6898 See also comment in drivers/video/console/sticore.c. 6899 6900 sti_font= [HW] 6901 See comment in drivers/video/console/sticore.c. 6902 6903 stifb= [HW] 6904 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]] 6905 6906 strict_sas_size= 6907 [X86] 6908 Format: <bool> 6909 Enable or disable strict sigaltstack size checks 6910 against the required signal frame size which 6911 depends on the supported FPU features. This can 6912 be used to filter out binaries which have 6913 not yet been made aware of AT_MINSIGSTKSZ. 6914 6915 stress_hpt [PPC,EARLY] 6916 Limits the number of kernel HPT entries in the hash 6917 page table to increase the rate of hash page table 6918 faults on kernel addresses. 6919 6920 stress_slb [PPC,EARLY] 6921 Limits the number of kernel SLB entries, and flushes 6922 them frequently to increase the rate of SLB faults 6923 on kernel addresses. 6924 6925 sunrpc.min_resvport= 6926 sunrpc.max_resvport= 6927 [NFS,SUNRPC] 6928 SunRPC servers often require that client requests 6929 originate from a privileged port (i.e. a port in the 6930 range 0 < portnr < 1024). 6931 An administrator who wishes to reserve some of these 6932 ports for other uses may adjust the range that the 6933 kernel's sunrpc client considers to be privileged 6934 using these two parameters to set the minimum and 6935 maximum port values. 6936 6937 sunrpc.svc_rpc_per_connection_limit= 6938 [NFS,SUNRPC] 6939 Limit the number of requests that the server will 6940 process in parallel from a single connection. 6941 The default value is 0 (no limit). 6942 6943 sunrpc.pool_mode= 6944 [NFS] 6945 Control how the NFS server code allocates CPUs to 6946 service thread pools. Depending on how many NICs 6947 you have and where their interrupts are bound, this 6948 option will affect which CPUs will do NFS serving. 6949 Note: this parameter cannot be changed while the 6950 NFS server is running. 6951 6952 auto the server chooses an appropriate mode 6953 automatically using heuristics 6954 global a single global pool contains all CPUs 6955 percpu one pool for each CPU 6956 pernode one pool for each NUMA node (equivalent 6957 to global on non-NUMA machines) 6958 6959 sunrpc.tcp_slot_table_entries= 6960 sunrpc.udp_slot_table_entries= 6961 [NFS,SUNRPC] 6962 Sets the upper limit on the number of simultaneous 6963 RPC calls that can be sent from the client to a 6964 server. Increasing these values may allow you to 6965 improve throughput, but will also increase the 6966 amount of memory reserved for use by the client. 6967 6968 suspend.pm_test_delay= 6969 [SUSPEND] 6970 Sets the number of seconds to remain in a suspend test 6971 mode before resuming the system (see 6972 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 6973 is set. Default value is 5. 6974 6975 svm= [PPC] 6976 Format: { on | off | y | n | 1 | 0 } 6977 This parameter controls use of the Protected 6978 Execution Facility on pSeries. 6979 6980 swiotlb= [ARM,PPC,MIPS,X86,S390,EARLY] 6981 Format: { <int> [,<int>] | force | noforce } 6982 <int> -- Number of I/O TLB slabs 6983 <int> -- Second integer after comma. Number of swiotlb 6984 areas with their own lock. Will be rounded up 6985 to a power of 2. 6986 force -- force using of bounce buffers even if they 6987 wouldn't be automatically used by the kernel 6988 noforce -- Never use bounce buffers (for debugging) 6989 6990 switches= [HW,M68k,EARLY] 6991 6992 sysctl.*= [KNL] 6993 Set a sysctl parameter, right before loading the init 6994 process, as if the value was written to the respective 6995 /proc/sys/... file. Both '.' and '/' are recognized as 6996 separators. Unrecognized parameters and invalid values 6997 are reported in the kernel log. Sysctls registered 6998 later by a loaded module cannot be set this way. 6999 Example: sysctl.vm.swappiness=40 7000 7001 sysrq_always_enabled 7002 [KNL] 7003 Ignore sysrq setting - this boot parameter will 7004 neutralize any effect of /proc/sys/kernel/sysrq. 7005 Useful for debugging. 7006 7007 tcpmhash_entries= [KNL,NET] 7008 Set the number of tcp_metrics_hash slots. 7009 Default value is 8192 or 16384 depending on total 7010 ram pages. This is used to specify the TCP metrics 7011 cache size. See Documentation/networking/ip-sysctl.rst 7012 "tcp_no_metrics_save" section for more details. 7013 7014 tdfx= [HW,DRM] 7015 7016 test_suspend= [SUSPEND] 7017 Format: { "mem" | "standby" | "freeze" }[,N] 7018 Specify "mem" (for Suspend-to-RAM) or "standby" (for 7019 standby suspend) or "freeze" (for suspend type freeze) 7020 as the system sleep state during system startup with 7021 the optional capability to repeat N number of times. 7022 The system is woken from this state using a 7023 wakeup-capable RTC alarm. 7024 7025 thash_entries= [KNL,NET] 7026 Set number of hash buckets for TCP connection 7027 7028 thermal.act= [HW,ACPI] 7029 -1: disable all active trip points in all thermal zones 7030 <degrees C>: override all lowest active trip points 7031 7032 thermal.crt= [HW,ACPI] 7033 -1: disable all critical trip points in all thermal zones 7034 <degrees C>: override all critical trip points 7035 7036 thermal.off= [HW,ACPI] 7037 1: disable ACPI thermal control 7038 7039 thermal.psv= [HW,ACPI] 7040 -1: disable all passive trip points 7041 <degrees C>: override all passive trip points to this 7042 value 7043 7044 thermal.tzp= [HW,ACPI] 7045 Specify global default ACPI thermal zone polling rate 7046 <deci-seconds>: poll all this frequency 7047 0: no polling (default) 7048 7049 thp_anon= [KNL] 7050 Format: <size>[KMG],<size>[KMG]:<state>;<size>[KMG]-<size>[KMG]:<state> 7051 state is one of "always", "madvise", "never" or "inherit". 7052 Control the default behavior of the system with respect 7053 to anonymous transparent hugepages. 7054 Can be used multiple times for multiple anon THP sizes. 7055 See Documentation/admin-guide/mm/transhuge.rst for more 7056 details. 7057 7058 threadirqs [KNL,EARLY] 7059 Force threading of all interrupt handlers except those 7060 marked explicitly IRQF_NO_THREAD. 7061 7062 thp_shmem= [KNL] 7063 Format: <size>[KMG],<size>[KMG]:<policy>;<size>[KMG]-<size>[KMG]:<policy> 7064 Control the default policy of each hugepage size for the 7065 internal shmem mount. <policy> is one of policies available 7066 for the shmem mount ("always", "inherit", "never", "within_size", 7067 and "advise"). 7068 It can be used multiple times for multiple shmem THP sizes. 7069 See Documentation/admin-guide/mm/transhuge.rst for more 7070 details. 7071 7072 topology= [S390,EARLY] 7073 Format: {off | on} 7074 Specify if the kernel should make use of the cpu 7075 topology information if the hardware supports this. 7076 The scheduler will make use of this information and 7077 e.g. base its process migration decisions on it. 7078 Default is on. 7079 7080 torture.disable_onoff_at_boot= [KNL] 7081 Prevent the CPU-hotplug component of torturing 7082 until after init has spawned. 7083 7084 torture.ftrace_dump_at_shutdown= [KNL] 7085 Dump the ftrace buffer at torture-test shutdown, 7086 even if there were no errors. This can be a 7087 very costly operation when many torture tests 7088 are running concurrently, especially on systems 7089 with rotating-rust storage. 7090 7091 torture.verbose_sleep_frequency= [KNL] 7092 Specifies how many verbose printk()s should be 7093 emitted between each sleep. The default of zero 7094 disables verbose-printk() sleeping. 7095 7096 torture.verbose_sleep_duration= [KNL] 7097 Duration of each verbose-printk() sleep in jiffies. 7098 7099 tpm.disable_pcr_integrity= [HW,TPM] 7100 Do not protect PCR registers from unintended physical 7101 access, or interposers in the bus by the means of 7102 having an integrity protected session wrapped around 7103 TPM2_PCR_Extend command. Consider this in a situation 7104 where TPM is heavily utilized by IMA, thus protection 7105 causing a major performance hit, and the space where 7106 machines are deployed is by other means guarded. 7107 7108 tpm_suspend_pcr=[HW,TPM] 7109 Format: integer pcr id 7110 Specify that at suspend time, the tpm driver 7111 should extend the specified pcr with zeros, 7112 as a workaround for some chips which fail to 7113 flush the last written pcr on TPM_SaveState. 7114 This will guarantee that all the other pcrs 7115 are saved. 7116 7117 tpm_tis.interrupts= [HW,TPM] 7118 Enable interrupts for the MMIO based physical layer 7119 for the FIFO interface. By default it is set to false 7120 (0). For more information about TPM hardware interfaces 7121 defined by Trusted Computing Group (TCG) see 7122 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/ 7123 7124 tp_printk [FTRACE] 7125 Have the tracepoints sent to printk as well as the 7126 tracing ring buffer. This is useful for early boot up 7127 where the system hangs or reboots and does not give the 7128 option for reading the tracing buffer or performing a 7129 ftrace_dump_on_oops. 7130 7131 To turn off having tracepoints sent to printk, 7132 echo 0 > /proc/sys/kernel/tracepoint_printk 7133 Note, echoing 1 into this file without the 7134 tp_printk kernel cmdline option has no effect. 7135 7136 The tp_printk_stop_on_boot (see below) can also be used 7137 to stop the printing of events to console at 7138 late_initcall_sync. 7139 7140 ** CAUTION ** 7141 7142 Having tracepoints sent to printk() and activating high 7143 frequency tracepoints such as irq or sched, can cause 7144 the system to live lock. 7145 7146 tp_printk_stop_on_boot [FTRACE] 7147 When tp_printk (above) is set, it can cause a lot of noise 7148 on the console. It may be useful to only include the 7149 printing of events during boot up, as user space may 7150 make the system inoperable. 7151 7152 This command line option will stop the printing of events 7153 to console at the late_initcall_sync() time frame. 7154 7155 trace_buf_size=nn[KMG] 7156 [FTRACE] will set tracing buffer size on each cpu. 7157 7158 trace_clock= [FTRACE] Set the clock used for tracing events 7159 at boot up. 7160 local - Use the per CPU time stamp counter 7161 (converted into nanoseconds). Fast, but 7162 depending on the architecture, may not be 7163 in sync between CPUs. 7164 global - Event time stamps are synchronize across 7165 CPUs. May be slower than the local clock, 7166 but better for some race conditions. 7167 counter - Simple counting of events (1, 2, ..) 7168 note, some counts may be skipped due to the 7169 infrastructure grabbing the clock more than 7170 once per event. 7171 uptime - Use jiffies as the time stamp. 7172 perf - Use the same clock that perf uses. 7173 mono - Use ktime_get_mono_fast_ns() for time stamps. 7174 mono_raw - Use ktime_get_raw_fast_ns() for time 7175 stamps. 7176 boot - Use ktime_get_boot_fast_ns() for time stamps. 7177 Architectures may add more clocks. See 7178 Documentation/trace/ftrace.rst for more details. 7179 7180 trace_event=[event-list] 7181 [FTRACE] Set and start specified trace events in order 7182 to facilitate early boot debugging. The event-list is a 7183 comma-separated list of trace events to enable. See 7184 also Documentation/trace/events.rst 7185 7186 To enable modules, use :mod: keyword: 7187 7188 trace_event=:mod:<module> 7189 7190 The value before :mod: will only enable specific events 7191 that are part of the module. See the above mentioned 7192 document for more information. 7193 7194 trace_instance=[instance-info] 7195 [FTRACE] Create a ring buffer instance early in boot up. 7196 This will be listed in: 7197 7198 /sys/kernel/tracing/instances 7199 7200 Events can be enabled at the time the instance is created 7201 via: 7202 7203 trace_instance=<name>,<system1>:<event1>,<system2>:<event2> 7204 7205 Note, the "<system*>:" portion is optional if the event is 7206 unique. 7207 7208 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall 7209 7210 will enable the "sched_switch" event (note, the "sched:" is optional, and 7211 the same thing would happen if it was left off). The irq_handler_entry 7212 event, and all events under the "initcall" system. 7213 7214 Flags can be added to the instance to modify its behavior when it is 7215 created. The flags are separated by '^'. 7216 7217 The available flags are: 7218 7219 traceoff - Have the tracing instance tracing disabled after it is created. 7220 traceprintk - Have trace_printk() write into this trace instance 7221 (note, "printk" and "trace_printk" can also be used) 7222 7223 trace_instance=foo^traceoff^traceprintk,sched,irq 7224 7225 The flags must come before the defined events. 7226 7227 If memory has been reserved (see memmap for x86), the instance 7228 can use that memory: 7229 7230 memmap=12M$0x284500000 trace_instance=boot_map@0x284500000:12M 7231 7232 The above will create a "boot_map" instance that uses the physical 7233 memory at 0x284500000 that is 12Megs. The per CPU buffers of that 7234 instance will be split up accordingly. 7235 7236 Alternatively, the memory can be reserved by the reserve_mem option: 7237 7238 reserve_mem=12M:4096:trace trace_instance=boot_map@trace 7239 7240 This will reserve 12 megabytes at boot up with a 4096 byte alignment 7241 and place the ring buffer in this memory. Note that due to KASLR, the 7242 memory may not be the same location each time, which will not preserve 7243 the buffer content. 7244 7245 Also note that the layout of the ring buffer data may change between 7246 kernel versions where the validator will fail and reset the ring buffer 7247 if the layout is not the same as the previous kernel. 7248 7249 If the ring buffer is used for persistent bootups and has events enabled, 7250 it is recommend to disable tracing so that events from a previous boot do not 7251 mix with events of the current boot (unless you are debugging a random crash 7252 at boot up). 7253 7254 reserve_mem=12M:4096:trace trace_instance=boot_map^traceoff^traceprintk@trace,sched,irq 7255 7256 Note, saving the trace buffer across reboots does require that the system 7257 is set up to not wipe memory. For instance, CONFIG_RESET_ATTACK_MITIGATION 7258 can force a memory reset on boot which will clear any trace that was stored. 7259 This is just one of many ways that can clear memory. Make sure your system 7260 keeps the content of memory across reboots before relying on this option. 7261 7262 See also Documentation/trace/debugging.rst 7263 7264 7265 trace_options=[option-list] 7266 [FTRACE] Enable or disable tracer options at boot. 7267 The option-list is a comma delimited list of options 7268 that can be enabled or disabled just as if you were 7269 to echo the option name into 7270 7271 /sys/kernel/tracing/trace_options 7272 7273 For example, to enable stacktrace option (to dump the 7274 stack trace of each event), add to the command line: 7275 7276 trace_options=stacktrace 7277 7278 See also Documentation/trace/ftrace.rst "trace options" 7279 section. 7280 7281 trace_trigger=[trigger-list] 7282 [FTRACE] Add a event trigger on specific events. 7283 Set a trigger on top of a specific event, with an optional 7284 filter. 7285 7286 The format is is "trace_trigger=<event>.<trigger>[ if <filter>],..." 7287 Where more than one trigger may be specified that are comma deliminated. 7288 7289 For example: 7290 7291 trace_trigger="sched_switch.stacktrace if prev_state == 2" 7292 7293 The above will enable the "stacktrace" trigger on the "sched_switch" 7294 event but only trigger it if the "prev_state" of the "sched_switch" 7295 event is "2" (TASK_UNINTERUPTIBLE). 7296 7297 See also "Event triggers" in Documentation/trace/events.rst 7298 7299 7300 traceoff_after_boot 7301 [FTRACE] Sometimes tracing is used to debug issues 7302 during the boot process. Since the trace buffer has a 7303 limited amount of storage, it may be prudent to 7304 disable tracing after the boot is finished, otherwise 7305 the critical information may be overwritten. With this 7306 option, the main tracing buffer will be turned off at 7307 the end of the boot process. 7308 7309 traceoff_on_warning 7310 [FTRACE] enable this option to disable tracing when a 7311 warning is hit. This turns off "tracing_on". Tracing can 7312 be enabled again by echoing '1' into the "tracing_on" 7313 file located in /sys/kernel/tracing/ 7314 7315 This option is useful, as it disables the trace before 7316 the WARNING dump is called, which prevents the trace to 7317 be filled with content caused by the warning output. 7318 7319 This option can also be set at run time via the sysctl 7320 option: kernel/traceoff_on_warning 7321 7322 transparent_hugepage= 7323 [KNL] 7324 Format: [always|madvise|never] 7325 Can be used to control the default behavior of the system 7326 with respect to transparent hugepages. 7327 See Documentation/admin-guide/mm/transhuge.rst 7328 for more details. 7329 7330 transparent_hugepage_shmem= [KNL] 7331 Format: [always|within_size|advise|never|deny|force] 7332 Can be used to control the hugepage allocation policy for 7333 the internal shmem mount. 7334 See Documentation/admin-guide/mm/transhuge.rst 7335 for more details. 7336 7337 transparent_hugepage_tmpfs= [KNL] 7338 Format: [always|within_size|advise|never] 7339 Can be used to control the default hugepage allocation policy 7340 for the tmpfs mount. 7341 See Documentation/admin-guide/mm/transhuge.rst 7342 for more details. 7343 7344 trusted.source= [KEYS] 7345 Format: <string> 7346 This parameter identifies the trust source as a backend 7347 for trusted keys implementation. Supported trust 7348 sources: 7349 - "tpm" 7350 - "tee" 7351 - "caam" 7352 - "dcp" 7353 If not specified then it defaults to iterating through 7354 the trust source list starting with TPM and assigns the 7355 first trust source as a backend which is initialized 7356 successfully during iteration. 7357 7358 trusted.rng= [KEYS] 7359 Format: <string> 7360 The RNG used to generate key material for trusted keys. 7361 Can be one of: 7362 - "kernel" 7363 - the same value as trusted.source: "tpm" or "tee" 7364 - "default" 7365 If not specified, "default" is used. In this case, 7366 the RNG's choice is left to each individual trust source. 7367 7368 trusted.dcp_use_otp_key 7369 This is intended to be used in combination with 7370 trusted.source=dcp and will select the DCP OTP key 7371 instead of the DCP UNIQUE key blob encryption. 7372 7373 trusted.dcp_skip_zk_test 7374 This is intended to be used in combination with 7375 trusted.source=dcp and will disable the check if the 7376 blob key is all zeros. This is helpful for situations where 7377 having this key zero'ed is acceptable. E.g. in testing 7378 scenarios. 7379 7380 tsc= Disable clocksource stability checks for TSC. 7381 Format: <string> 7382 [x86] reliable: mark tsc clocksource as reliable, this 7383 disables clocksource verification at runtime, as well 7384 as the stability checks done at bootup. Used to enable 7385 high-resolution timer mode on older hardware, and in 7386 virtualized environment. 7387 [x86] noirqtime: Do not use TSC to do irq accounting. 7388 Used to run time disable IRQ_TIME_ACCOUNTING on any 7389 platforms where RDTSC is slow and this accounting 7390 can add overhead. 7391 [x86] unstable: mark the TSC clocksource as unstable, this 7392 marks the TSC unconditionally unstable at bootup and 7393 avoids any further wobbles once the TSC watchdog notices. 7394 [x86] nowatchdog: disable clocksource watchdog. Used 7395 in situations with strict latency requirements (where 7396 interruptions from clocksource watchdog are not 7397 acceptable). 7398 [x86] recalibrate: force recalibration against a HW timer 7399 (HPET or PM timer) on systems whose TSC frequency was 7400 obtained from HW or FW using either an MSR or CPUID(0x15). 7401 Warn if the difference is more than 500 ppm. 7402 [x86] watchdog: Use TSC as the watchdog clocksource with 7403 which to check other HW timers (HPET or PM timer), but 7404 only on systems where TSC has been deemed trustworthy. 7405 This will be suppressed by an earlier tsc=nowatchdog and 7406 can be overridden by a later tsc=nowatchdog. A console 7407 message will flag any such suppression or overriding. 7408 7409 tsc_early_khz= [X86,EARLY] Skip early TSC calibration and use the given 7410 value instead. Useful when the early TSC frequency discovery 7411 procedure is not reliable, such as on overclocked systems 7412 with CPUID.16h support and partial CPUID.15h support. 7413 Format: <unsigned int> 7414 7415 tsx= [X86] Control Transactional Synchronization 7416 Extensions (TSX) feature in Intel processors that 7417 support TSX control. 7418 7419 This parameter controls the TSX feature. The options are: 7420 7421 on - Enable TSX on the system. Although there are 7422 mitigations for all known security vulnerabilities, 7423 TSX has been known to be an accelerator for 7424 several previous speculation-related CVEs, and 7425 so there may be unknown security risks associated 7426 with leaving it enabled. 7427 7428 off - Disable TSX on the system. (Note that this 7429 option takes effect only on newer CPUs which are 7430 not vulnerable to MDS, i.e., have 7431 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get 7432 the new IA32_TSX_CTRL MSR through a microcode 7433 update. This new MSR allows for the reliable 7434 deactivation of the TSX functionality.) 7435 7436 auto - Disable TSX if X86_BUG_TAA is present, 7437 otherwise enable TSX on the system. 7438 7439 Not specifying this option is equivalent to tsx=off. 7440 7441 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 7442 for more details. 7443 7444 tsx_async_abort= [X86,INTEL,EARLY] Control mitigation for the TSX Async 7445 Abort (TAA) vulnerability. 7446 7447 Similar to Micro-architectural Data Sampling (MDS) 7448 certain CPUs that support Transactional 7449 Synchronization Extensions (TSX) are vulnerable to an 7450 exploit against CPU internal buffers which can forward 7451 information to a disclosure gadget under certain 7452 conditions. 7453 7454 In vulnerable processors, the speculatively forwarded 7455 data can be used in a cache side channel attack, to 7456 access data to which the attacker does not have direct 7457 access. 7458 7459 This parameter controls the TAA mitigation. The 7460 options are: 7461 7462 full - Enable TAA mitigation on vulnerable CPUs 7463 if TSX is enabled. 7464 7465 full,nosmt - Enable TAA mitigation and disable SMT on 7466 vulnerable CPUs. If TSX is disabled, SMT 7467 is not disabled because CPU is not 7468 vulnerable to cross-thread TAA attacks. 7469 off - Unconditionally disable TAA mitigation 7470 7471 On MDS-affected machines, tsx_async_abort=off can be 7472 prevented by an active MDS mitigation as both vulnerabilities 7473 are mitigated with the same mechanism so in order to disable 7474 this mitigation, you need to specify mds=off too. 7475 7476 Not specifying this option is equivalent to 7477 tsx_async_abort=full. On CPUs which are MDS affected 7478 and deploy MDS mitigation, TAA mitigation is not 7479 required and doesn't provide any additional 7480 mitigation. 7481 7482 For details see: 7483 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 7484 7485 turbografx.map[2|3]= [HW,JOY] 7486 TurboGraFX parallel port interface 7487 Format: 7488 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7> 7489 See also Documentation/input/devices/joystick-parport.rst 7490 7491 udbg-immortal [PPC] When debugging early kernel crashes that 7492 happen after console_init() and before a proper 7493 console driver takes over, this boot options might 7494 help "seeing" what's going on. 7495 7496 uhash_entries= [KNL,NET] 7497 Set number of hash buckets for UDP/UDP-Lite connections 7498 7499 uhci-hcd.ignore_oc= 7500 [USB] Ignore overcurrent events (default N). 7501 Some badly-designed motherboards generate lots of 7502 bogus events, for ports that aren't wired to 7503 anything. Set this parameter to avoid log spamming. 7504 Note that genuine overcurrent events won't be 7505 reported either. 7506 7507 unknown_nmi_panic 7508 [X86] Cause panic on unknown NMI. 7509 7510 unwind_debug [X86-64,EARLY] 7511 Enable unwinder debug output. This can be 7512 useful for debugging certain unwinder error 7513 conditions, including corrupt stacks and 7514 bad/missing unwinder metadata. 7515 7516 usbcore.authorized_default= 7517 [USB] Default USB device authorization: 7518 (default -1 = authorized (same as 1), 7519 0 = not authorized, 1 = authorized, 2 = authorized 7520 if device connected to internal port) 7521 7522 usbcore.autosuspend= 7523 [USB] The autosuspend time delay (in seconds) used 7524 for newly-detected USB devices (default 2). This 7525 is the time required before an idle device will be 7526 autosuspended. Devices for which the delay is set 7527 to a negative value won't be autosuspended at all. 7528 7529 usbcore.usbfs_snoop= 7530 [USB] Set to log all usbfs traffic (default 0 = off). 7531 7532 usbcore.usbfs_snoop_max= 7533 [USB] Maximum number of bytes to snoop in each URB 7534 (default = 65536). 7535 7536 usbcore.blinkenlights= 7537 [USB] Set to cycle leds on hubs (default 0 = off). 7538 7539 usbcore.old_scheme_first= 7540 [USB] Start with the old device initialization 7541 scheme (default 0 = off). 7542 7543 usbcore.usbfs_memory_mb= 7544 [USB] Memory limit (in MB) for buffers allocated by 7545 usbfs (default = 16, 0 = max = 2047). 7546 7547 usbcore.use_both_schemes= 7548 [USB] Try the other device initialization scheme 7549 if the first one fails (default 1 = enabled). 7550 7551 usbcore.initial_descriptor_timeout= 7552 [USB] Specifies timeout for the initial 64-byte 7553 USB_REQ_GET_DESCRIPTOR request in milliseconds 7554 (default 5000 = 5.0 seconds). 7555 7556 usbcore.nousb [USB] Disable the USB subsystem 7557 7558 usbcore.quirks= 7559 [USB] A list of quirk entries to augment the built-in 7560 usb core quirk list. List entries are separated by 7561 commas. Each entry has the form 7562 VendorID:ProductID:Flags. The IDs are 4-digit hex 7563 numbers and Flags is a set of letters. Each letter 7564 will change the built-in quirk; setting it if it is 7565 clear and clearing it if it is set. The letters have 7566 the following meanings: 7567 a = USB_QUIRK_STRING_FETCH_255 (string 7568 descriptors must not be fetched using 7569 a 255-byte read); 7570 b = USB_QUIRK_RESET_RESUME (device can't resume 7571 correctly so reset it instead); 7572 c = USB_QUIRK_NO_SET_INTF (device can't handle 7573 Set-Interface requests); 7574 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't 7575 handle its Configuration or Interface 7576 strings); 7577 e = USB_QUIRK_RESET (device can't be reset 7578 (e.g morph devices), don't use reset); 7579 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has 7580 more interface descriptions than the 7581 bNumInterfaces count, and can't handle 7582 talking to these interfaces); 7583 g = USB_QUIRK_DELAY_INIT (device needs a pause 7584 during initialization, after we read 7585 the device descriptor); 7586 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For 7587 high speed and super speed interrupt 7588 endpoints, the USB 2.0 and USB 3.0 spec 7589 require the interval in microframes (1 7590 microframe = 125 microseconds) to be 7591 calculated as interval = 2 ^ 7592 (bInterval-1). 7593 Devices with this quirk report their 7594 bInterval as the result of this 7595 calculation instead of the exponent 7596 variable used in the calculation); 7597 i = USB_QUIRK_DEVICE_QUALIFIER (device can't 7598 handle device_qualifier descriptor 7599 requests); 7600 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device 7601 generates spurious wakeup, ignore 7602 remote wakeup capability); 7603 k = USB_QUIRK_NO_LPM (device can't handle Link 7604 Power Management); 7605 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL 7606 (Device reports its bInterval as linear 7607 frames instead of the USB 2.0 7608 calculation); 7609 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs 7610 to be disconnected before suspend to 7611 prevent spurious wakeup); 7612 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a 7613 pause after every control message); 7614 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra 7615 delay after resetting its port); 7616 p = USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT 7617 (Reduce timeout of the SET_ADDRESS 7618 request from 5000 ms to 500 ms); 7619 Example: quirks=0781:5580:bk,0a5c:5834:gij 7620 7621 usbhid.mousepoll= 7622 [USBHID] The interval which mice are to be polled at. 7623 7624 usbhid.jspoll= 7625 [USBHID] The interval which joysticks are to be polled at. 7626 7627 usbhid.kbpoll= 7628 [USBHID] The interval which keyboards are to be polled at. 7629 7630 usb-storage.delay_use= 7631 [UMS] The delay in seconds before a new device is 7632 scanned for Logical Units (default 1). 7633 Optionally the delay in milliseconds if the value has 7634 suffix with "ms". 7635 Example: delay_use=2567ms 7636 7637 usb-storage.quirks= 7638 [UMS] A list of quirks entries to supplement or 7639 override the built-in unusual_devs list. List 7640 entries are separated by commas. Each entry has 7641 the form VID:PID:Flags where VID and PID are Vendor 7642 and Product ID values (4-digit hex numbers) and 7643 Flags is a set of characters, each corresponding 7644 to a common usb-storage quirk flag as follows: 7645 a = SANE_SENSE (collect more than 18 bytes 7646 of sense data, not on uas); 7647 b = BAD_SENSE (don't collect more than 18 7648 bytes of sense data, not on uas); 7649 c = FIX_CAPACITY (decrease the reported 7650 device capacity by one sector); 7651 d = NO_READ_DISC_INFO (don't use 7652 READ_DISC_INFO command, not on uas); 7653 e = NO_READ_CAPACITY_16 (don't use 7654 READ_CAPACITY_16 command); 7655 f = NO_REPORT_OPCODES (don't use report opcodes 7656 command, uas only); 7657 g = MAX_SECTORS_240 (don't transfer more than 7658 240 sectors at a time, uas only); 7659 h = CAPACITY_HEURISTICS (decrease the 7660 reported device capacity by one 7661 sector if the number is odd); 7662 i = IGNORE_DEVICE (don't bind to this 7663 device); 7664 j = NO_REPORT_LUNS (don't use report luns 7665 command, uas only); 7666 k = NO_SAME (do not use WRITE_SAME, uas only) 7667 l = NOT_LOCKABLE (don't try to lock and 7668 unlock ejectable media, not on uas); 7669 m = MAX_SECTORS_64 (don't transfer more 7670 than 64 sectors = 32 KB at a time, 7671 not on uas); 7672 n = INITIAL_READ10 (force a retry of the 7673 initial READ(10) command, not on uas); 7674 o = CAPACITY_OK (accept the capacity 7675 reported by the device, not on uas); 7676 p = WRITE_CACHE (the device cache is ON 7677 by default, not on uas); 7678 r = IGNORE_RESIDUE (the device reports 7679 bogus residue values, not on uas); 7680 s = SINGLE_LUN (the device has only one 7681 Logical Unit); 7682 t = NO_ATA_1X (don't allow ATA(12) and ATA(16) 7683 commands, uas only); 7684 u = IGNORE_UAS (don't bind to the uas driver); 7685 w = NO_WP_DETECT (don't test whether the 7686 medium is write-protected). 7687 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE 7688 even if the device claims no cache, 7689 not on uas) 7690 Example: quirks=0419:aaf5:rl,0421:0433:rc 7691 7692 user_debug= [KNL,ARM] 7693 Format: <int> 7694 See arch/arm/Kconfig.debug help text. 7695 1 - undefined instruction events 7696 2 - system calls 7697 4 - invalid data aborts 7698 8 - SIGSEGV faults 7699 16 - SIGBUS faults 7700 Example: user_debug=31 7701 7702 vdso= [X86,SH,SPARC] 7703 On X86_32, this is an alias for vdso32=. Otherwise: 7704 7705 vdso=1: enable VDSO (the default) 7706 vdso=0: disable VDSO mapping 7707 7708 vdso32= [X86] Control the 32-bit vDSO 7709 vdso32=1: enable 32-bit VDSO 7710 vdso32=0 or vdso32=2: disable 32-bit VDSO 7711 7712 See the help text for CONFIG_COMPAT_VDSO for more 7713 details. If CONFIG_COMPAT_VDSO is set, the default is 7714 vdso32=0; otherwise, the default is vdso32=1. 7715 7716 For compatibility with older kernels, vdso32=2 is an 7717 alias for vdso32=0. 7718 7719 Try vdso32=0 if you encounter an error that says: 7720 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed! 7721 7722 video= [FB,EARLY] Frame buffer configuration 7723 See Documentation/fb/modedb.rst. 7724 7725 video.brightness_switch_enabled= [ACPI] 7726 Format: [0|1] 7727 If set to 1, on receiving an ACPI notify event 7728 generated by hotkey, video driver will adjust brightness 7729 level and then send out the event to user space through 7730 the allocated input device. If set to 0, video driver 7731 will only send out the event without touching backlight 7732 brightness level. 7733 default: 1 7734 7735 virtio_mmio.device= 7736 [VMMIO] Memory mapped virtio (platform) device. 7737 7738 <size>@<baseaddr>:<irq>[:<id>] 7739 where: 7740 <size> := size (can use standard suffixes 7741 like K, M and G) 7742 <baseaddr> := physical base address 7743 <irq> := interrupt number (as passed to 7744 request_irq()) 7745 <id> := (optional) platform device id 7746 example: 7747 virtio_mmio.device=1K@0x100b0000:48:7 7748 7749 Can be used multiple times for multiple devices. 7750 7751 vga= [BOOT,X86-32] Select a particular video mode 7752 See Documentation/arch/x86/boot.rst and 7753 Documentation/admin-guide/svga.rst. 7754 Use vga=ask for menu. 7755 This is actually a boot loader parameter; the value is 7756 passed to the kernel using a special protocol. 7757 7758 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y. 7759 May slow down system boot speed, especially when 7760 enabled on systems with a large amount of memory. 7761 All options are enabled by default, and this 7762 interface is meant to allow for selectively 7763 enabling or disabling specific virtual memory 7764 debugging features. 7765 7766 Available options are: 7767 P Enable page structure init time poisoning 7768 - Disable all of the above options 7769 7770 vmalloc=nn[KMG] [KNL,BOOT,EARLY] Forces the vmalloc area to have an 7771 exact size of <nn>. This can be used to increase 7772 the minimum size (128MB on x86, arm32 platforms). 7773 It can also be used to decrease the size and leave more room 7774 for directly mapped kernel RAM. Note that this parameter does 7775 not exist on many other platforms (including arm64, alpha, 7776 loongarch, arc, csky, hexagon, microblaze, mips, nios2, openrisc, 7777 parisc, m64k, powerpc, riscv, sh, um, xtensa, s390, sparc). 7778 7779 vmcp_cma=nn[MG] [KNL,S390,EARLY] 7780 Sets the memory size reserved for contiguous memory 7781 allocations for the vmcp device driver. 7782 7783 vmhalt= [KNL,S390] Perform z/VM CP command after system halt. 7784 Format: <command> 7785 7786 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic. 7787 Format: <command> 7788 7789 vmpoff= [KNL,S390] Perform z/VM CP command after power off. 7790 Format: <command> 7791 7792 vsyscall= [X86-64,EARLY] 7793 Controls the behavior of vsyscalls (i.e. calls to 7794 fixed addresses of 0xffffffffff600x00 from legacy 7795 code). Most statically-linked binaries and older 7796 versions of glibc use these calls. Because these 7797 functions are at fixed addresses, they make nice 7798 targets for exploits that can control RIP. 7799 7800 emulate Vsyscalls turn into traps and are emulated 7801 reasonably safely. The vsyscall page is 7802 readable. 7803 7804 xonly [default] Vsyscalls turn into traps and are 7805 emulated reasonably safely. The vsyscall 7806 page is not readable. 7807 7808 none Vsyscalls don't work at all. This makes 7809 them quite hard to use for exploits but 7810 might break your system. 7811 7812 vt.color= [VT] Default text color. 7813 Format: 0xYX, X = foreground, Y = background. 7814 Default: 0x07 = light gray on black. 7815 7816 vt.cur_default= [VT] Default cursor shape. 7817 Format: 0xCCBBAA, where AA, BB, and CC are the same as 7818 the parameters of the <Esc>[?A;B;Cc escape sequence; 7819 see vga-softcursor.rst. Default: 2 = underline. 7820 7821 vt.default_blu= [VT] 7822 Format: <blue0>,<blue1>,<blue2>,...,<blue15> 7823 Change the default blue palette of the console. 7824 This is a 16-member array composed of values 7825 ranging from 0-255. 7826 7827 vt.default_grn= [VT] 7828 Format: <green0>,<green1>,<green2>,...,<green15> 7829 Change the default green palette of the console. 7830 This is a 16-member array composed of values 7831 ranging from 0-255. 7832 7833 vt.default_red= [VT] 7834 Format: <red0>,<red1>,<red2>,...,<red15> 7835 Change the default red palette of the console. 7836 This is a 16-member array composed of values 7837 ranging from 0-255. 7838 7839 vt.default_utf8= 7840 [VT] 7841 Format=<0|1> 7842 Set system-wide default UTF-8 mode for all tty's. 7843 Default is 1, i.e. UTF-8 mode is enabled for all 7844 newly opened terminals. 7845 7846 vt.global_cursor_default= 7847 [VT] 7848 Format=<-1|0|1> 7849 Set system-wide default for whether a cursor 7850 is shown on new VTs. Default is -1, 7851 i.e. cursors will be created by default unless 7852 overridden by individual drivers. 0 will hide 7853 cursors, 1 will display them. 7854 7855 vt.italic= [VT] Default color for italic text; 0-15. 7856 Default: 2 = green. 7857 7858 vt.underline= [VT] Default color for underlined text; 0-15. 7859 Default: 3 = cyan. 7860 7861 watchdog timers [HW,WDT] For information on watchdog timers, 7862 see Documentation/watchdog/watchdog-parameters.rst 7863 or other driver-specific files in the 7864 Documentation/watchdog/ directory. 7865 7866 watchdog_thresh= 7867 [KNL] 7868 Set the hard lockup detector stall duration 7869 threshold in seconds. The soft lockup detector 7870 threshold is set to twice the value. A value of 0 7871 disables both lockup detectors. Default is 10 7872 seconds. 7873 7874 workqueue.unbound_cpus= 7875 [KNL,SMP] Specify to constrain one or some CPUs 7876 to use in unbound workqueues. 7877 Format: <cpu-list> 7878 By default, all online CPUs are available for 7879 unbound workqueues. 7880 7881 workqueue.watchdog_thresh= 7882 If CONFIG_WQ_WATCHDOG is configured, workqueue can 7883 warn stall conditions and dump internal state to 7884 help debugging. 0 disables workqueue stall 7885 detection; otherwise, it's the stall threshold 7886 duration in seconds. The default value is 30 and 7887 it can be updated at runtime by writing to the 7888 corresponding sysfs file. 7889 7890 workqueue.panic_on_stall=<uint> 7891 Panic when workqueue stall is detected by 7892 CONFIG_WQ_WATCHDOG. It sets the number times of the 7893 stall to trigger panic. 7894 7895 The default is 0, which disables the panic on stall. 7896 7897 workqueue.cpu_intensive_thresh_us= 7898 Per-cpu work items which run for longer than this 7899 threshold are automatically considered CPU intensive 7900 and excluded from concurrency management to prevent 7901 them from noticeably delaying other per-cpu work 7902 items. Default is 10000 (10ms). 7903 7904 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 7905 will report the work functions which violate this 7906 threshold repeatedly. They are likely good 7907 candidates for using WQ_UNBOUND workqueues instead. 7908 7909 workqueue.cpu_intensive_warning_thresh=<uint> 7910 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 7911 will report the work functions which violate the 7912 intensive_threshold_us repeatedly. In order to prevent 7913 spurious warnings, start printing only after a work 7914 function has violated this threshold number of times. 7915 7916 The default is 4 times. 0 disables the warning. 7917 7918 workqueue.power_efficient 7919 Per-cpu workqueues are generally preferred because 7920 they show better performance thanks to cache 7921 locality; unfortunately, per-cpu workqueues tend to 7922 be more power hungry than unbound workqueues. 7923 7924 Enabling this makes the per-cpu workqueues which 7925 were observed to contribute significantly to power 7926 consumption unbound, leading to measurably lower 7927 power usage at the cost of small performance 7928 overhead. 7929 7930 The default value of this parameter is determined by 7931 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT. 7932 7933 workqueue.default_affinity_scope= 7934 Select the default affinity scope to use for unbound 7935 workqueues. Can be one of "cpu", "smt", "cache", 7936 "numa" and "system". Default is "cache". For more 7937 information, see the Affinity Scopes section in 7938 Documentation/core-api/workqueue.rst. 7939 7940 This can be changed after boot by writing to the 7941 matching /sys/module/workqueue/parameters file. All 7942 workqueues with the "default" affinity scope will be 7943 updated accordingly. 7944 7945 workqueue.debug_force_rr_cpu 7946 Workqueue used to implicitly guarantee that work 7947 items queued without explicit CPU specified are put 7948 on the local CPU. This guarantee is no longer true 7949 and while local CPU is still preferred work items 7950 may be put on foreign CPUs. This debug option 7951 forces round-robin CPU selection to flush out 7952 usages which depend on the now broken guarantee. 7953 When enabled, memory and cache locality will be 7954 impacted. 7955 7956 writecombine= [LOONGARCH,EARLY] Control the MAT (Memory Access 7957 Type) of ioremap_wc(). 7958 7959 on - Enable writecombine, use WUC for ioremap_wc() 7960 off - Disable writecombine, use SUC for ioremap_wc() 7961 7962 x2apic_phys [X86-64,APIC,EARLY] Use x2apic physical mode instead of 7963 default x2apic cluster mode on platforms 7964 supporting x2apic. 7965 7966 xen_512gb_limit [KNL,X86-64,XEN] 7967 Restricts the kernel running paravirtualized under Xen 7968 to use only up to 512 GB of RAM. The reason to do so is 7969 crash analysis tools and Xen tools for doing domain 7970 save/restore/migration must be enabled to handle larger 7971 domains. 7972 7973 xen_emul_unplug= [HW,X86,XEN,EARLY] 7974 Unplug Xen emulated devices 7975 Format: [unplug0,][unplug1] 7976 ide-disks -- unplug primary master IDE devices 7977 aux-ide-disks -- unplug non-primary-master IDE devices 7978 nics -- unplug network devices 7979 all -- unplug all emulated devices (NICs and IDE disks) 7980 unnecessary -- unplugging emulated devices is 7981 unnecessary even if the host did not respond to 7982 the unplug protocol 7983 never -- do not unplug even if version check succeeds 7984 7985 xen_legacy_crash [X86,XEN,EARLY] 7986 Crash from Xen panic notifier, without executing late 7987 panic() code such as dumping handler. 7988 7989 xen_mc_debug [X86,XEN,EARLY] 7990 Enable multicall debugging when running as a Xen PV guest. 7991 Enabling this feature will reduce performance a little 7992 bit, so it should only be enabled for obtaining extended 7993 debug data in case of multicall errors. 7994 7995 xen_msr_safe= [X86,XEN,EARLY] 7996 Format: <bool> 7997 Select whether to always use non-faulting (safe) MSR 7998 access functions when running as Xen PV guest. The 7999 default value is controlled by CONFIG_XEN_PV_MSR_SAFE. 8000 8001 xen_nopv [X86] 8002 Disables the PV optimizations forcing the HVM guest to 8003 run as generic HVM guest with no PV drivers. 8004 This option is obsoleted by the "nopv" option, which 8005 has equivalent effect for XEN platform. 8006 8007 xen_no_vector_callback 8008 [KNL,X86,XEN,EARLY] Disable the vector callback for Xen 8009 event channel interrupts. 8010 8011 xen_scrub_pages= [XEN] 8012 Boolean option to control scrubbing pages before giving them back 8013 to Xen, for use by other domains. Can be also changed at runtime 8014 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages. 8015 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT. 8016 8017 xen_timer_slop= [X86-64,XEN,EARLY] 8018 Set the timer slop (in nanoseconds) for the virtual Xen 8019 timers (default is 100000). This adjusts the minimum 8020 delta of virtualized Xen timers, where lower values 8021 improve timer resolution at the expense of processing 8022 more timer interrupts. 8023 8024 xen.balloon_boot_timeout= [XEN] 8025 The time (in seconds) to wait before giving up to boot 8026 in case initial ballooning fails to free enough memory. 8027 Applies only when running as HVM or PVH guest and 8028 started with less memory configured than allowed at 8029 max. Default is 180. 8030 8031 xen.event_eoi_delay= [XEN] 8032 How long to delay EOI handling in case of event 8033 storms (jiffies). Default is 10. 8034 8035 xen.event_loop_timeout= [XEN] 8036 After which time (jiffies) the event handling loop 8037 should start to delay EOI handling. Default is 2. 8038 8039 xen.fifo_events= [XEN] 8040 Boolean parameter to disable using fifo event handling 8041 even if available. Normally fifo event handling is 8042 preferred over the 2-level event handling, as it is 8043 fairer and the number of possible event channels is 8044 much higher. Default is on (use fifo events). 8045 8046 xirc2ps_cs= [NET,PCMCIA] 8047 Format: 8048 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]] 8049 8050 xive= [PPC] 8051 By default on POWER9 and above, the kernel will 8052 natively use the XIVE interrupt controller. This option 8053 allows the fallback firmware mode to be used: 8054 8055 off Fallback to firmware control of XIVE interrupt 8056 controller on both pseries and powernv 8057 platforms. Only useful on POWER9 and above. 8058 8059 xive.store-eoi=off [PPC] 8060 By default on POWER10 and above, the kernel will use 8061 stores for EOI handling when the XIVE interrupt mode 8062 is active. This option allows the XIVE driver to use 8063 loads instead, as on POWER9. 8064 8065 xhci-hcd.quirks [USB,KNL] 8066 A hex value specifying bitmask with supplemental xhci 8067 host controller quirks. Meaning of each bit can be 8068 consulted in header drivers/usb/host/xhci.h. 8069 8070 xmon [PPC,EARLY] 8071 Format: { early | on | rw | ro | off } 8072 Controls if xmon debugger is enabled. Default is off. 8073 Passing only "xmon" is equivalent to "xmon=early". 8074 early Call xmon as early as possible on boot; xmon 8075 debugger is called from setup_arch(). 8076 on xmon debugger hooks will be installed so xmon 8077 is only called on a kernel crash. Default mode, 8078 i.e. either "ro" or "rw" mode, is controlled 8079 with CONFIG_XMON_DEFAULT_RO_MODE. 8080 rw xmon debugger hooks will be installed so xmon 8081 is called only on a kernel crash, mode is write, 8082 meaning SPR registers, memory and, other data 8083 can be written using xmon commands. 8084 ro same as "rw" option above but SPR registers, 8085 memory, and other data can't be written using 8086 xmon commands. 8087 off xmon is disabled. 8088