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