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