1 acpi= [HW,ACPI,X86,ARM64,RISCV64] 2 Advanced Configuration and Power Interface 3 Format: { force | on | off | strict | noirq | rsdt | 4 copy_dsdt } 5 force -- enable ACPI if default was off 6 on -- enable ACPI but allow fallback to DT [arm64,riscv64] 7 off -- disable ACPI if default was on 8 noirq -- do not use ACPI for IRQ routing 9 strict -- Be less tolerant of platforms that are not 10 strictly ACPI specification compliant. 11 rsdt -- prefer RSDT over (default) XSDT 12 copy_dsdt -- copy DSDT to memory 13 For ARM64 and RISCV64, ONLY "acpi=off", "acpi=on" or 14 "acpi=force" are available 15 16 See also Documentation/power/runtime_pm.rst, pci=noacpi 17 18 acpi_apic_instance= [ACPI, IOAPIC] 19 Format: <int> 20 2: use 2nd APIC table, if available 21 1,0: use 1st APIC table 22 default: 0 23 24 acpi_backlight= [HW,ACPI] 25 { vendor | video | native | none } 26 If set to vendor, prefer vendor-specific driver 27 (e.g. thinkpad_acpi, sony_acpi, etc.) instead 28 of the ACPI video.ko driver. 29 If set to video, use the ACPI video.ko driver. 30 If set to native, use the device's native backlight mode. 31 If set to none, disable the ACPI backlight interface. 32 33 acpi_force_32bit_fadt_addr 34 force FADT to use 32 bit addresses rather than the 35 64 bit X_* addresses. Some firmware have broken 64 36 bit addresses for force ACPI ignore these and use 37 the older legacy 32 bit addresses. 38 39 acpica_no_return_repair [HW, ACPI] 40 Disable AML predefined validation mechanism 41 This mechanism can repair the evaluation result to make 42 the return objects more ACPI specification compliant. 43 This option is useful for developers to identify the 44 root cause of an AML interpreter issue when the issue 45 has something to do with the repair mechanism. 46 47 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG] 48 acpi.debug_level= [HW,ACPI,ACPI_DEBUG] 49 Format: <int> 50 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI 51 debug output. Bits in debug_layer correspond to a 52 _COMPONENT in an ACPI source file, e.g., 53 #define _COMPONENT ACPI_EVENTS 54 Bits in debug_level correspond to a level in 55 ACPI_DEBUG_PRINT statements, e.g., 56 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ... 57 The debug_level mask defaults to "info". See 58 Documentation/firmware-guide/acpi/debug.rst for more information about 59 debug layers and levels. 60 61 Enable processor driver info messages: 62 acpi.debug_layer=0x20000000 63 Enable AML "Debug" output, i.e., stores to the Debug 64 object while interpreting AML: 65 acpi.debug_layer=0xffffffff acpi.debug_level=0x2 66 Enable all messages related to ACPI hardware: 67 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff 68 69 Some values produce so much output that the system is 70 unusable. The "log_buf_len" parameter may be useful 71 if you need to capture more output. 72 73 acpi_enforce_resources= [ACPI] 74 { strict | lax | no } 75 Check for resource conflicts between native drivers 76 and ACPI OperationRegions (SystemIO and SystemMemory 77 only). IO ports and memory declared in ACPI might be 78 used by the ACPI subsystem in arbitrary AML code and 79 can interfere with legacy drivers. 80 strict (default): access to resources claimed by ACPI 81 is denied; legacy drivers trying to access reserved 82 resources will fail to bind to device using them. 83 lax: access to resources claimed by ACPI is allowed; 84 legacy drivers trying to access reserved resources 85 will bind successfully but a warning message is logged. 86 no: ACPI OperationRegions are not marked as reserved, 87 no further checks are performed. 88 89 acpi_force_table_verification [HW,ACPI] 90 Enable table checksum verification during early stage. 91 By default, this is disabled due to x86 early mapping 92 size limitation. 93 94 acpi_irq_balance [HW,ACPI] 95 ACPI will balance active IRQs 96 default in APIC mode 97 98 acpi_irq_nobalance [HW,ACPI] 99 ACPI will not move active IRQs (default) 100 default in PIC mode 101 102 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA 103 Format: <irq>,<irq>... 104 105 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for 106 use by PCI 107 Format: <irq>,<irq>... 108 109 acpi_mask_gpe= [HW,ACPI] 110 Due to the existence of _Lxx/_Exx, some GPEs triggered 111 by unsupported hardware/firmware features can result in 112 GPE floodings that cannot be automatically disabled by 113 the GPE dispatcher. 114 This facility can be used to prevent such uncontrolled 115 GPE floodings. 116 Format: <byte> or <bitmap-list> 117 118 acpi_no_auto_serialize [HW,ACPI] 119 Disable auto-serialization of AML methods 120 AML control methods that contain the opcodes to create 121 named objects will be marked as "Serialized" by the 122 auto-serialization feature. 123 This feature is enabled by default. 124 This option allows to turn off the feature. 125 126 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump 127 kernels. 128 129 acpi_no_static_ssdt [HW,ACPI] 130 Disable installation of static SSDTs at early boot time 131 By default, SSDTs contained in the RSDT/XSDT will be 132 installed automatically and they will appear under 133 /sys/firmware/acpi/tables. 134 This option turns off this feature. 135 Note that specifying this option does not affect 136 dynamic table installation which will install SSDT 137 tables to /sys/firmware/acpi/tables/dynamic. 138 139 acpi_no_watchdog [HW,ACPI,WDT] 140 Ignore the ACPI-based watchdog interface (WDAT) and let 141 a native driver control the watchdog device instead. 142 143 acpi_rsdp= [ACPI,EFI,KEXEC] 144 Pass the RSDP address to the kernel, mostly used 145 on machines running EFI runtime service to boot the 146 second kernel for kdump. 147 148 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS 149 Format: To spoof as Windows 98: ="Microsoft Windows" 150 151 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead 152 of 2 which is mandated by ACPI 6) as the supported ACPI 153 specification revision (when using this switch, it may 154 be necessary to carry out a cold reboot _twice_ in a 155 row to make it take effect on the platform firmware). 156 157 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings 158 acpi_osi="string1" # add string1 159 acpi_osi="!string2" # remove string2 160 acpi_osi=!* # remove all strings 161 acpi_osi=! # disable all built-in OS vendor 162 strings 163 acpi_osi=!! # enable all built-in OS vendor 164 strings 165 acpi_osi= # disable all strings 166 167 'acpi_osi=!' can be used in combination with single or 168 multiple 'acpi_osi="string1"' to support specific OS 169 vendor string(s). Note that such command can only 170 affect the default state of the OS vendor strings, thus 171 it cannot affect the default state of the feature group 172 strings and the current state of the OS vendor strings, 173 specifying it multiple times through kernel command line 174 is meaningless. This command is useful when one do not 175 care about the state of the feature group strings which 176 should be controlled by the OSPM. 177 Examples: 178 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent 179 to 'acpi_osi="Windows 2000" acpi_osi=!', they all 180 can make '_OSI("Windows 2000")' TRUE. 181 182 'acpi_osi=' cannot be used in combination with other 183 'acpi_osi=' command lines, the _OSI method will not 184 exist in the ACPI namespace. NOTE that such command can 185 only affect the _OSI support state, thus specifying it 186 multiple times through kernel command line is also 187 meaningless. 188 Examples: 189 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)' 190 FALSE. 191 192 'acpi_osi=!*' can be used in combination with single or 193 multiple 'acpi_osi="string1"' to support specific 194 string(s). Note that such command can affect the 195 current state of both the OS vendor strings and the 196 feature group strings, thus specifying it multiple times 197 through kernel command line is meaningful. But it may 198 still not able to affect the final state of a string if 199 there are quirks related to this string. This command 200 is useful when one want to control the state of the 201 feature group strings to debug BIOS issues related to 202 the OSPM features. 203 Examples: 204 1. 'acpi_osi="Module Device" acpi_osi=!*' can make 205 '_OSI("Module Device")' FALSE. 206 2. 'acpi_osi=!* acpi_osi="Module Device"' can make 207 '_OSI("Module Device")' TRUE. 208 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is 209 equivalent to 210 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"' 211 and 212 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!', 213 they all will make '_OSI("Windows 2000")' TRUE. 214 215 acpi_pm_good [X86] 216 Override the pmtimer bug detection: force the kernel 217 to assume that this machine's pmtimer latches its value 218 and always returns good values. 219 220 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode 221 Format: { level | edge | high | low } 222 223 acpi_skip_timer_override [HW,ACPI] 224 Recognize and ignore IRQ0/pin2 Interrupt Override. 225 For broken nForce2 BIOS resulting in XT-PIC timer. 226 227 acpi_sleep= [HW,ACPI] Sleep options 228 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig, 229 s4_nohwsig, old_ordering, nonvs, 230 sci_force_enable, nobl } 231 See Documentation/power/video.rst for information on 232 s3_bios and s3_mode. 233 s3_beep is for debugging; it makes the PC's speaker beep 234 as soon as the kernel's real-mode entry point is called. 235 s4_hwsig causes the kernel to check the ACPI hardware 236 signature during resume from hibernation, and gracefully 237 refuse to resume if it has changed. This complies with 238 the ACPI specification but not with reality, since 239 Windows does not do this and many laptops do change it 240 on docking. So the default behaviour is to allow resume 241 and simply warn when the signature changes, unless the 242 s4_hwsig option is enabled. 243 s4_nohwsig prevents ACPI hardware signature from being 244 used (or even warned about) during resume. 245 old_ordering causes the ACPI 1.0 ordering of the _PTS 246 control method, with respect to putting devices into 247 low power states, to be enforced (the ACPI 2.0 ordering 248 of _PTS is used by default). 249 nonvs prevents the kernel from saving/restoring the 250 ACPI NVS memory during suspend/hibernation and resume. 251 sci_force_enable causes the kernel to set SCI_EN directly 252 on resume from S1/S3 (which is against the ACPI spec, 253 but some broken systems don't work without it). 254 nobl causes the internal blacklist of systems known to 255 behave incorrectly in some ways with respect to system 256 suspend and resume to be ignored (use wisely). 257 258 acpi_use_timer_override [HW,ACPI] 259 Use timer override. For some broken Nvidia NF5 boards 260 that require a timer override, but don't have HPET 261 262 add_efi_memmap [EFI; X86] Include EFI memory map in 263 kernel's map of available physical RAM. 264 265 agp= [AGP] 266 { off | try_unsupported } 267 off: disable AGP support 268 try_unsupported: try to drive unsupported chipsets 269 (may crash computer or cause data corruption) 270 271 ALSA [HW,ALSA] 272 See Documentation/sound/alsa-configuration.rst 273 274 alignment= [KNL,ARM] 275 Allow the default userspace alignment fault handler 276 behaviour to be specified. Bit 0 enables warnings, 277 bit 1 enables fixups, and bit 2 sends a segfault. 278 279 align_va_addr= [X86-64] 280 Align virtual addresses by clearing slice [14:12] when 281 allocating a VMA at process creation time. This option 282 gives you up to 3% performance improvement on AMD F15h 283 machines (where it is enabled by default) for a 284 CPU-intensive style benchmark, and it can vary highly in 285 a microbenchmark depending on workload and compiler. 286 287 32: only for 32-bit processes 288 64: only for 64-bit processes 289 on: enable for both 32- and 64-bit processes 290 off: disable for both 32- and 64-bit processes 291 292 alloc_snapshot [FTRACE] 293 Allocate the ftrace snapshot buffer on boot up when the 294 main buffer is allocated. This is handy if debugging 295 and you need to use tracing_snapshot() on boot up, and 296 do not want to use tracing_snapshot_alloc() as it needs 297 to be done where GFP_KERNEL allocations are allowed. 298 299 allow_mismatched_32bit_el0 [ARM64] 300 Allow execve() of 32-bit applications and setting of the 301 PER_LINUX32 personality on systems where only a strict 302 subset of the CPUs support 32-bit EL0. When this 303 parameter is present, the set of CPUs supporting 32-bit 304 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0 305 and hot-unplug operations may be restricted. 306 307 See Documentation/arch/arm64/asymmetric-32bit.rst for more 308 information. 309 310 amd_iommu= [HW,X86-64] 311 Pass parameters to the AMD IOMMU driver in the system. 312 Possible values are: 313 fullflush - Deprecated, equivalent to iommu.strict=1 314 off - do not initialize any AMD IOMMU found in 315 the system 316 force_isolation - Force device isolation for all 317 devices. The IOMMU driver is not 318 allowed anymore to lift isolation 319 requirements as needed. This option 320 does not override iommu=pt 321 force_enable - Force enable the IOMMU on platforms known 322 to be buggy with IOMMU enabled. Use this 323 option with care. 324 pgtbl_v1 - Use v1 page table for DMA-API (Default). 325 pgtbl_v2 - Use v2 page table for DMA-API. 326 irtcachedis - Disable Interrupt Remapping Table (IRT) caching. 327 328 amd_iommu_dump= [HW,X86-64] 329 Enable AMD IOMMU driver option to dump the ACPI table 330 for AMD IOMMU. With this option enabled, AMD IOMMU 331 driver will print ACPI tables for AMD IOMMU during 332 IOMMU initialization. 333 334 amd_iommu_intr= [HW,X86-64] 335 Specifies one of the following AMD IOMMU interrupt 336 remapping modes: 337 legacy - Use legacy interrupt remapping mode. 338 vapic - Use virtual APIC mode, which allows IOMMU 339 to inject interrupts directly into guest. 340 This mode requires kvm-amd.avic=1. 341 (Default when IOMMU HW support is present.) 342 343 amd_pstate= [X86] 344 disable 345 Do not enable amd_pstate as the default 346 scaling driver for the supported processors 347 passive 348 Use amd_pstate with passive mode as a scaling driver. 349 In this mode autonomous selection is disabled. 350 Driver requests a desired performance level and platform 351 tries to match the same performance level if it is 352 satisfied by guaranteed performance level. 353 active 354 Use amd_pstate_epp driver instance as the scaling driver, 355 driver provides a hint to the hardware if software wants 356 to bias toward performance (0x0) or energy efficiency (0xff) 357 to the CPPC firmware. then CPPC power algorithm will 358 calculate the runtime workload and adjust the realtime cores 359 frequency. 360 guided 361 Activate guided autonomous mode. Driver requests minimum and 362 maximum performance level and the platform autonomously 363 selects a performance level in this range and appropriate 364 to the current workload. 365 366 amijoy.map= [HW,JOY] Amiga joystick support 367 Map of devices attached to JOY0DAT and JOY1DAT 368 Format: <a>,<b> 369 See also Documentation/input/joydev/joystick.rst 370 371 analog.map= [HW,JOY] Analog joystick and gamepad support 372 Specifies type or capabilities of an analog joystick 373 connected to one of 16 gameports 374 Format: <type1>,<type2>,..<type16> 375 376 apc= [HW,SPARC] 377 Power management functions (SPARCstation-4/5 + deriv.) 378 Format: noidle 379 Disable APC CPU standby support. SPARCstation-Fox does 380 not play well with APC CPU idle - disable it if you have 381 APC and your system crashes randomly. 382 383 apic= [APIC,X86] Advanced Programmable Interrupt Controller 384 Change the output verbosity while booting 385 Format: { quiet (default) | verbose | debug } 386 Change the amount of debugging information output 387 when initialising the APIC and IO-APIC components. 388 For X86-32, this can also be used to specify an APIC 389 driver name. 390 Format: apic=driver_name 391 Examples: apic=bigsmp 392 393 apic_extnmi= [APIC,X86] External NMI delivery setting 394 Format: { bsp (default) | all | none } 395 bsp: External NMI is delivered only to CPU 0 396 all: External NMIs are broadcast to all CPUs as a 397 backup of CPU 0 398 none: External NMI is masked for all CPUs. This is 399 useful so that a dump capture kernel won't be 400 shot down by NMI 401 402 autoconf= [IPV6] 403 See Documentation/networking/ipv6.rst. 404 405 apm= [APM] Advanced Power Management 406 See header of arch/x86/kernel/apm_32.c. 407 408 apparmor= [APPARMOR] Disable or enable AppArmor at boot time 409 Format: { "0" | "1" } 410 See security/apparmor/Kconfig help text 411 0 -- disable. 412 1 -- enable. 413 Default value is set via kernel config option. 414 415 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards 416 Format: <io>,<irq>,<nodeID> 417 418 arm64.nobti [ARM64] Unconditionally disable Branch Target 419 Identification support 420 421 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory 422 Set instructions support 423 424 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension 425 support 426 427 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication 428 support 429 430 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix 431 Extension support 432 433 arm64.nosve [ARM64] Unconditionally disable Scalable Vector 434 Extension support 435 436 ataflop= [HW,M68k] 437 438 atarimouse= [HW,MOUSE] Atari Mouse 439 440 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess, 441 EzKey and similar keyboards 442 443 atkbd.reset= [HW] Reset keyboard during initialization 444 445 atkbd.set= [HW] Select keyboard code set 446 Format: <int> (2 = AT (default), 3 = PS/2) 447 448 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar 449 keyboards 450 451 atkbd.softraw= [HW] Choose between synthetic and real raw mode 452 Format: <bool> (0 = real, 1 = synthetic (default)) 453 454 atkbd.softrepeat= [HW] 455 Use software keyboard repeat 456 457 audit= [KNL] Enable the audit sub-system 458 Format: { "0" | "1" | "off" | "on" } 459 0 | off - kernel audit is disabled and can not be 460 enabled until the next reboot 461 unset - kernel audit is initialized but disabled and 462 will be fully enabled by the userspace auditd. 463 1 | on - kernel audit is initialized and partially 464 enabled, storing at most audit_backlog_limit 465 messages in RAM until it is fully enabled by the 466 userspace auditd. 467 Default: unset 468 469 audit_backlog_limit= [KNL] Set the audit queue size limit. 470 Format: <int> (must be >=0) 471 Default: 64 472 473 bau= [X86_UV] Enable the BAU on SGI UV. The default 474 behavior is to disable the BAU (i.e. bau=0). 475 Format: { "0" | "1" } 476 0 - Disable the BAU. 477 1 - Enable the BAU. 478 unset - Disable the BAU. 479 480 baycom_epp= [HW,AX25] 481 Format: <io>,<mode> 482 483 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem 484 Format: <io>,<mode> 485 See header of drivers/net/hamradio/baycom_par.c. 486 487 baycom_ser_fdx= [HW,AX25] 488 BayCom Serial Port AX.25 Modem (Full Duplex Mode) 489 Format: <io>,<irq>,<mode>[,<baud>] 490 See header of drivers/net/hamradio/baycom_ser_fdx.c. 491 492 baycom_ser_hdx= [HW,AX25] 493 BayCom Serial Port AX.25 Modem (Half Duplex Mode) 494 Format: <io>,<irq>,<mode> 495 See header of drivers/net/hamradio/baycom_ser_hdx.c. 496 497 bert_disable [ACPI] 498 Disable BERT OS support on buggy BIOSes. 499 500 bgrt_disable [ACPI][X86] 501 Disable BGRT to avoid flickering OEM logo. 502 503 blkdevparts= Manual partition parsing of block device(s) for 504 embedded devices based on command line input. 505 See Documentation/block/cmdline-partition.rst 506 507 boot_delay= Milliseconds to delay each printk during boot. 508 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled, 509 and you may also have to specify "lpj=". Boot_delay 510 values larger than 10 seconds (10000) are assumed 511 erroneous and ignored. 512 Format: integer 513 514 bootconfig [KNL] 515 Extended command line options can be added to an initrd 516 and this will cause the kernel to look for it. 517 518 See Documentation/admin-guide/bootconfig.rst 519 520 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards) 521 bttv.radio= Most important insmod options are available as 522 kernel args too. 523 bttv.pll= See Documentation/admin-guide/media/bttv.rst 524 bttv.tuner= 525 526 bulk_remove=off [PPC] This parameter disables the use of the pSeries 527 firmware feature for flushing multiple hpte entries 528 at a time. 529 530 c101= [NET] Moxa C101 synchronous serial card 531 532 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection. 533 Sometimes CPU hardware bugs make them report the cache 534 size incorrectly. The kernel will attempt work arounds 535 to fix known problems, but for some CPUs it is not 536 possible to determine what the correct size should be. 537 This option provides an override for these situations. 538 539 carrier_timeout= 540 [NET] Specifies amount of time (in seconds) that 541 the kernel should wait for a network carrier. By default 542 it waits 120 seconds. 543 544 ca_keys= [KEYS] This parameter identifies a specific key(s) on 545 the system trusted keyring to be used for certificate 546 trust validation. 547 format: { id:<keyid> | builtin } 548 549 cca= [MIPS] Override the kernel pages' cache coherency 550 algorithm. Accepted values range from 0 to 7 551 inclusive. See arch/mips/include/asm/pgtable-bits.h 552 for platform specific values (SB1, Loongson3 and 553 others). 554 555 ccw_timeout_log [S390] 556 See Documentation/arch/s390/common_io.rst for details. 557 558 cgroup_disable= [KNL] Disable a particular controller or optional feature 559 Format: {name of the controller(s) or feature(s) to disable} 560 The effects of cgroup_disable=foo are: 561 - foo isn't auto-mounted if you mount all cgroups in 562 a single hierarchy 563 - foo isn't visible as an individually mountable 564 subsystem 565 - if foo is an optional feature then the feature is 566 disabled and corresponding cgroup files are not 567 created 568 {Currently only "memory" controller deal with this and 569 cut the overhead, others just disable the usage. So 570 only cgroup_disable=memory is actually worthy} 571 Specifying "pressure" disables per-cgroup pressure 572 stall information accounting feature 573 574 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1 575 Format: { { controller | "all" | "named" } 576 [,{ controller | "all" | "named" }...] } 577 Like cgroup_disable, but only applies to cgroup v1; 578 the blacklisted controllers remain available in cgroup2. 579 "all" blacklists all controllers and "named" disables 580 named mounts. Specifying both "all" and "named" disables 581 all v1 hierarchies. 582 583 cgroup_favordynmods= [KNL] Enable or Disable favordynmods. 584 Format: { "true" | "false" } 585 Defaults to the value of CONFIG_CGROUP_FAVOR_DYNMODS. 586 587 cgroup.memory= [KNL] Pass options to the cgroup memory controller. 588 Format: <string> 589 nosocket -- Disable socket memory accounting. 590 nokmem -- Disable kernel memory accounting. 591 nobpf -- Disable BPF memory accounting. 592 593 checkreqprot= [SELINUX] Set initial checkreqprot flag value. 594 Format: { "0" | "1" } 595 See security/selinux/Kconfig help text. 596 0 -- check protection applied by kernel (includes 597 any implied execute protection). 598 1 -- check protection requested by application. 599 Default value is set via a kernel config option. 600 Value can be changed at runtime via 601 /sys/fs/selinux/checkreqprot. 602 Setting checkreqprot to 1 is deprecated. 603 604 cio_ignore= [S390] 605 See Documentation/arch/s390/common_io.rst for details. 606 607 clearcpuid=X[,X...] [X86] 608 Disable CPUID feature X for the kernel. See 609 arch/x86/include/asm/cpufeatures.h for the valid bit 610 numbers X. Note the Linux-specific bits are not necessarily 611 stable over kernel options, but the vendor-specific 612 ones should be. 613 X can also be a string as appearing in the flags: line 614 in /proc/cpuinfo which does not have the above 615 instability issue. However, not all features have names 616 in /proc/cpuinfo. 617 Note that using this option will taint your kernel. 618 Also note that user programs calling CPUID directly 619 or using the feature without checking anything 620 will still see it. This just prevents it from 621 being used by the kernel or shown in /proc/cpuinfo. 622 Also note the kernel might malfunction if you disable 623 some critical bits. 624 625 clk_ignore_unused 626 [CLK] 627 Prevents the clock framework from automatically gating 628 clocks that have not been explicitly enabled by a Linux 629 device driver but are enabled in hardware at reset or 630 by the bootloader/firmware. Note that this does not 631 force such clocks to be always-on nor does it reserve 632 those clocks in any way. This parameter is useful for 633 debug and development, but should not be needed on a 634 platform with proper driver support. For more 635 information, see Documentation/driver-api/clk.rst. 636 637 clock= [BUGS=X86-32, HW] gettimeofday clocksource override. 638 [Deprecated] 639 Forces specified clocksource (if available) to be used 640 when calculating gettimeofday(). If specified 641 clocksource is not available, it defaults to PIT. 642 Format: { pit | tsc | cyclone | pmtmr } 643 644 clocksource= Override the default clocksource 645 Format: <string> 646 Override the default clocksource and use the clocksource 647 with the name specified. 648 Some clocksource names to choose from, depending on 649 the platform: 650 [all] jiffies (this is the base, fallback clocksource) 651 [ACPI] acpi_pm 652 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2, 653 pxa_timer,timer3,32k_counter,timer0_1 654 [X86-32] pit,hpet,tsc; 655 scx200_hrt on Geode; cyclone on IBM x440 656 [MIPS] MIPS 657 [PARISC] cr16 658 [S390] tod 659 [SH] SuperH 660 [SPARC64] tick 661 [X86-64] hpet,tsc 662 663 clocksource.arm_arch_timer.evtstrm= 664 [ARM,ARM64] 665 Format: <bool> 666 Enable/disable the eventstream feature of the ARM 667 architected timer so that code using WFE-based polling 668 loops can be debugged more effectively on production 669 systems. 670 671 clocksource.max_cswd_read_retries= [KNL] 672 Number of clocksource_watchdog() retries due to 673 external delays before the clock will be marked 674 unstable. Defaults to two retries, that is, 675 three attempts to read the clock under test. 676 677 clocksource.verify_n_cpus= [KNL] 678 Limit the number of CPUs checked for clocksources 679 marked with CLOCK_SOURCE_VERIFY_PERCPU that 680 are marked unstable due to excessive skew. 681 A negative value says to check all CPUs, while 682 zero says not to check any. Values larger than 683 nr_cpu_ids are silently truncated to nr_cpu_ids. 684 The actual CPUs are chosen randomly, with 685 no replacement if the same CPU is chosen twice. 686 687 clocksource-wdtest.holdoff= [KNL] 688 Set the time in seconds that the clocksource 689 watchdog test waits before commencing its tests. 690 Defaults to zero when built as a module and to 691 10 seconds when built into the kernel. 692 693 cma=nn[MG]@[start[MG][-end[MG]]] 694 [KNL,CMA] 695 Sets the size of kernel global memory area for 696 contiguous memory allocations and optionally the 697 placement constraint by the physical address range of 698 memory allocations. A value of 0 disables CMA 699 altogether. For more information, see 700 kernel/dma/contiguous.c 701 702 cma_pernuma=nn[MG] 703 [KNL,CMA] 704 Sets the size of kernel per-numa memory area for 705 contiguous memory allocations. A value of 0 disables 706 per-numa CMA altogether. And If this option is not 707 specified, the default value is 0. 708 With per-numa CMA enabled, DMA users on node nid will 709 first try to allocate buffer from the pernuma area 710 which is located in node nid, if the allocation fails, 711 they will fallback to the global default memory area. 712 713 numa_cma=<node>:nn[MG][,<node>:nn[MG]] 714 [KNL,CMA] 715 Sets the size of kernel numa memory area for 716 contiguous memory allocations. It will reserve CMA 717 area for the specified node. 718 719 With numa CMA enabled, DMA users on node nid will 720 first try to allocate buffer from the numa area 721 which is located in node nid, if the allocation fails, 722 they will fallback to the global default memory area. 723 724 cmo_free_hint= [PPC] Format: { yes | no } 725 Specify whether pages are marked as being inactive 726 when they are freed. This is used in CMO environments 727 to determine OS memory pressure for page stealing by 728 a hypervisor. 729 Default: yes 730 731 coherent_pool=nn[KMG] [ARM,KNL] 732 Sets the size of memory pool for coherent, atomic dma 733 allocations, by default set to 256K. 734 735 com20020= [HW,NET] ARCnet - COM20020 chipset 736 Format: 737 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]] 738 739 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers) 740 Format: <io>[,<irq>] 741 742 com90xx= [HW,NET] 743 ARCnet - COM90xx chipset (memory-mapped buffers) 744 Format: <io>[,<irq>[,<memstart>]] 745 746 condev= [HW,S390] console device 747 conmode= 748 749 con3215_drop= [S390] 3215 console drop mode. 750 Format: y|n|Y|N|1|0 751 When set to true, drop data on the 3215 console when 752 the console buffer is full. In this case the 753 operator using a 3270 terminal emulator (for example 754 x3270) does not have to enter the clear key for the 755 console output to advance and the kernel to continue. 756 This leads to a much faster boot time when a 3270 757 terminal emulator is active. If no 3270 terminal 758 emulator is used, this parameter has no effect. 759 760 console= [KNL] Output console device and options. 761 762 tty<n> Use the virtual console device <n>. 763 764 ttyS<n>[,options] 765 ttyUSB0[,options] 766 Use the specified serial port. The options are of 767 the form "bbbbpnf", where "bbbb" is the baud rate, 768 "p" is parity ("n", "o", or "e"), "n" is number of 769 bits, and "f" is flow control ("r" for RTS or 770 omit it). Default is "9600n8". 771 772 See Documentation/admin-guide/serial-console.rst for more 773 information. See 774 Documentation/networking/netconsole.rst for an 775 alternative. 776 777 uart[8250],io,<addr>[,options] 778 uart[8250],mmio,<addr>[,options] 779 uart[8250],mmio16,<addr>[,options] 780 uart[8250],mmio32,<addr>[,options] 781 uart[8250],0x<addr>[,options] 782 Start an early, polled-mode console on the 8250/16550 783 UART at the specified I/O port or MMIO address, 784 switching to the matching ttyS device later. 785 MMIO inter-register address stride is either 8-bit 786 (mmio), 16-bit (mmio16), or 32-bit (mmio32). 787 If none of [io|mmio|mmio16|mmio32], <addr> is assumed 788 to be equivalent to 'mmio'. 'options' are specified in 789 the same format described for ttyS above; if unspecified, 790 the h/w is not re-initialized. 791 792 hvc<n> Use the hypervisor console device <n>. This is for 793 both Xen and PowerPC hypervisors. 794 795 { null | "" } 796 Use to disable console output, i.e., to have kernel 797 console messages discarded. 798 This must be the only console= parameter used on the 799 kernel command line. 800 801 If the device connected to the port is not a TTY but a braille 802 device, prepend "brl," before the device type, for instance 803 console=brl,ttyS0 804 For now, only VisioBraille is supported. 805 806 console_msg_format= 807 [KNL] Change console messages format 808 default 809 By default we print messages on consoles in 810 "[time stamp] text\n" format (time stamp may not be 811 printed, depending on CONFIG_PRINTK_TIME or 812 `printk_time' param). 813 syslog 814 Switch to syslog format: "<%u>[time stamp] text\n" 815 IOW, each message will have a facility and loglevel 816 prefix. The format is similar to one used by syslog() 817 syscall, or to executing "dmesg -S --raw" or to reading 818 from /proc/kmsg. 819 820 consoleblank= [KNL] The console blank (screen saver) timeout in 821 seconds. A value of 0 disables the blank timer. 822 Defaults to 0. 823 824 coredump_filter= 825 [KNL] Change the default value for 826 /proc/<pid>/coredump_filter. 827 See also Documentation/filesystems/proc.rst. 828 829 coresight_cpu_debug.enable 830 [ARM,ARM64] 831 Format: <bool> 832 Enable/disable the CPU sampling based debugging. 833 0: default value, disable debugging 834 1: enable debugging at boot time 835 836 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver 837 Format: 838 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>] 839 840 cpuidle.off=1 [CPU_IDLE] 841 disable the cpuidle sub-system 842 843 cpuidle.governor= 844 [CPU_IDLE] Name of the cpuidle governor to use. 845 846 cpufreq.off=1 [CPU_FREQ] 847 disable the cpufreq sub-system 848 849 cpufreq.default_governor= 850 [CPU_FREQ] Name of the default cpufreq governor or 851 policy to use. This governor must be registered in the 852 kernel before the cpufreq driver probes. 853 854 cpu_init_udelay=N 855 [X86] Delay for N microsec between assert and de-assert 856 of APIC INIT to start processors. This delay occurs 857 on every CPU online, such as boot, and resume from suspend. 858 Default: 10000 859 860 cpuhp.parallel= 861 [SMP] Enable/disable parallel bringup of secondary CPUs 862 Format: <bool> 863 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise 864 the parameter has no effect. 865 866 crash_kexec_post_notifiers 867 Run kdump after running panic-notifiers and dumping 868 kmsg. This only for the users who doubt kdump always 869 succeeds in any situation. 870 Note that this also increases risks of kdump failure, 871 because some panic notifiers can make the crashed 872 kernel more unstable. 873 874 crashkernel=size[KMG][@offset[KMG]] 875 [KNL] Using kexec, Linux can switch to a 'crash kernel' 876 upon panic. This parameter reserves the physical 877 memory region [offset, offset + size] for that kernel 878 image. If '@offset' is omitted, then a suitable offset 879 is selected automatically. 880 [KNL, X86-64, ARM64, RISCV] Select a region under 4G first, and 881 fall back to reserve region above 4G when '@offset' 882 hasn't been specified. 883 See Documentation/admin-guide/kdump/kdump.rst for further details. 884 885 crashkernel=range1:size1[,range2:size2,...][@offset] 886 [KNL] Same as above, but depends on the memory 887 in the running system. The syntax of range is 888 start-[end] where start and end are both 889 a memory unit (amount[KMG]). See also 890 Documentation/admin-guide/kdump/kdump.rst for an example. 891 892 crashkernel=size[KMG],high 893 [KNL, X86-64, ARM64, RISCV] range could be above 4G. 894 Allow kernel to allocate physical memory region from top, 895 so could be above 4G if system have more than 4G ram 896 installed. Otherwise memory region will be allocated 897 below 4G, if available. 898 It will be ignored if crashkernel=X is specified. 899 crashkernel=size[KMG],low 900 [KNL, X86-64, ARM64, RISCV] range under 4G. When crashkernel=X,high 901 is passed, kernel could allocate physical memory region 902 above 4G, that cause second kernel crash on system 903 that require some amount of low memory, e.g. swiotlb 904 requires at least 64M+32K low memory, also enough extra 905 low memory is needed to make sure DMA buffers for 32-bit 906 devices won't run out. Kernel would try to allocate 907 default size of memory below 4G automatically. The default 908 size is platform dependent. 909 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB) 910 --> arm64: 128MiB 911 --> riscv: 128MiB 912 This one lets the user specify own low range under 4G 913 for second kernel instead. 914 0: to disable low allocation. 915 It will be ignored when crashkernel=X,high is not used 916 or memory reserved is below 4G. 917 918 cryptomgr.notests 919 [KNL] Disable crypto self-tests 920 921 cs89x0_dma= [HW,NET] 922 Format: <dma> 923 924 cs89x0_media= [HW,NET] 925 Format: { rj45 | aui | bnc } 926 927 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU 928 function call handling. When switched on, 929 additional debug data is printed to the console 930 in case a hanging CPU is detected, and that 931 CPU is pinged again in order to try to resolve 932 the hang situation. The default value of this 933 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT 934 Kconfig option. 935 936 dasd= [HW,NET] 937 See header of drivers/s390/block/dasd_devmap.c. 938 939 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port 940 (one device per port) 941 Format: <port#>,<type> 942 See also Documentation/input/devices/joystick-parport.rst 943 944 debug [KNL] Enable kernel debugging (events log level). 945 946 debug_boot_weak_hash 947 [KNL] Enable printing [hashed] pointers early in the 948 boot sequence. If enabled, we use a weak hash instead 949 of siphash to hash pointers. Use this option if you are 950 seeing instances of '(___ptrval___)') and need to see a 951 value (hashed pointer) instead. Cryptographically 952 insecure, please do not use on production kernels. 953 954 debug_locks_verbose= 955 [KNL] verbose locking self-tests 956 Format: <int> 957 Print debugging info while doing the locking API 958 self-tests. 959 Bitmask for the various LOCKTYPE_ tests. Defaults to 0 960 (no extra messages), setting it to -1 (all bits set) 961 will print _a_lot_ more information - normally only 962 useful to lockdep developers. 963 964 debug_objects [KNL] Enable object debugging 965 966 debug_guardpage_minorder= 967 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this 968 parameter allows control of the order of pages that will 969 be intentionally kept free (and hence protected) by the 970 buddy allocator. Bigger value increase the probability 971 of catching random memory corruption, but reduce the 972 amount of memory for normal system use. The maximum 973 possible value is MAX_ORDER/2. Setting this parameter 974 to 1 or 2 should be enough to identify most random 975 memory corruption problems caused by bugs in kernel or 976 driver code when a CPU writes to (or reads from) a 977 random memory location. Note that there exists a class 978 of memory corruptions problems caused by buggy H/W or 979 F/W or by drivers badly programming DMA (basically when 980 memory is written at bus level and the CPU MMU is 981 bypassed) which are not detectable by 982 CONFIG_DEBUG_PAGEALLOC, hence this option will not help 983 tracking down these problems. 984 985 debug_pagealloc= 986 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter 987 enables the feature at boot time. By default, it is 988 disabled and the system will work mostly the same as a 989 kernel built without CONFIG_DEBUG_PAGEALLOC. 990 Note: to get most of debug_pagealloc error reports, it's 991 useful to also enable the page_owner functionality. 992 on: enable the feature 993 994 debugfs= [KNL] This parameter enables what is exposed to userspace 995 and debugfs internal clients. 996 Format: { on, no-mount, off } 997 on: All functions are enabled. 998 no-mount: 999 Filesystem is not registered but kernel clients can 1000 access APIs and a crashkernel can be used to read 1001 its content. There is nothing to mount. 1002 off: Filesystem is not registered and clients 1003 get a -EPERM as result when trying to register files 1004 or directories within debugfs. 1005 This is equivalent of the runtime functionality if 1006 debugfs was not enabled in the kernel at all. 1007 Default value is set in build-time with a kernel configuration. 1008 1009 debugpat [X86] Enable PAT debugging 1010 1011 default_hugepagesz= 1012 [HW] The size of the default HugeTLB page. This is 1013 the size represented by the legacy /proc/ hugepages 1014 APIs. In addition, this is the default hugetlb size 1015 used for shmget(), mmap() and mounting hugetlbfs 1016 filesystems. If not specified, defaults to the 1017 architecture's default huge page size. Huge page 1018 sizes are architecture dependent. See also 1019 Documentation/admin-guide/mm/hugetlbpage.rst. 1020 Format: size[KMG] 1021 1022 deferred_probe_timeout= 1023 [KNL] Debugging option to set a timeout in seconds for 1024 deferred probe to give up waiting on dependencies to 1025 probe. Only specific dependencies (subsystems or 1026 drivers) that have opted in will be ignored. A timeout 1027 of 0 will timeout at the end of initcalls. If the time 1028 out hasn't expired, it'll be restarted by each 1029 successful driver registration. This option will also 1030 dump out devices still on the deferred probe list after 1031 retrying. 1032 1033 delayacct [KNL] Enable per-task delay accounting 1034 1035 dell_smm_hwmon.ignore_dmi= 1036 [HW] Continue probing hardware even if DMI data 1037 indicates that the driver is running on unsupported 1038 hardware. 1039 1040 dell_smm_hwmon.force= 1041 [HW] Activate driver even if SMM BIOS signature does 1042 not match list of supported models and enable otherwise 1043 blacklisted features. 1044 1045 dell_smm_hwmon.power_status= 1046 [HW] Report power status in /proc/i8k 1047 (disabled by default). 1048 1049 dell_smm_hwmon.restricted= 1050 [HW] Allow controlling fans only if SYS_ADMIN 1051 capability is set. 1052 1053 dell_smm_hwmon.fan_mult= 1054 [HW] Factor to multiply fan speed with. 1055 1056 dell_smm_hwmon.fan_max= 1057 [HW] Maximum configurable fan speed. 1058 1059 dfltcc= [HW,S390] 1060 Format: { on | off | def_only | inf_only | always } 1061 on: s390 zlib hardware support for compression on 1062 level 1 and decompression (default) 1063 off: No s390 zlib hardware support 1064 def_only: s390 zlib hardware support for deflate 1065 only (compression on level 1) 1066 inf_only: s390 zlib hardware support for inflate 1067 only (decompression) 1068 always: Same as 'on' but ignores the selected compression 1069 level always using hardware support (used for debugging) 1070 1071 dhash_entries= [KNL] 1072 Set number of hash buckets for dentry cache. 1073 1074 disable_1tb_segments [PPC] 1075 Disables the use of 1TB hash page table segments. This 1076 causes the kernel to fall back to 256MB segments which 1077 can be useful when debugging issues that require an SLB 1078 miss to occur. 1079 1080 disable= [IPV6] 1081 See Documentation/networking/ipv6.rst. 1082 1083 disable_radix [PPC] 1084 Disable RADIX MMU mode on POWER9 1085 1086 disable_tlbie [PPC] 1087 Disable TLBIE instruction. Currently does not work 1088 with KVM, with HASH MMU, or with coherent accelerators. 1089 1090 disable_cpu_apicid= [X86,APIC,SMP] 1091 Format: <int> 1092 The number of initial APIC ID for the 1093 corresponding CPU to be disabled at boot, 1094 mostly used for the kdump 2nd kernel to 1095 disable BSP to wake up multiple CPUs without 1096 causing system reset or hang due to sending 1097 INIT from AP to BSP. 1098 1099 disable_ddw [PPC/PSERIES] 1100 Disable Dynamic DMA Window support. Use this 1101 to workaround buggy firmware. 1102 1103 disable_ipv6= [IPV6] 1104 See Documentation/networking/ipv6.rst. 1105 1106 disable_mtrr_cleanup [X86] 1107 The kernel tries to adjust MTRR layout from continuous 1108 to discrete, to make X server driver able to add WB 1109 entry later. This parameter disables that. 1110 1111 disable_mtrr_trim [X86, Intel and AMD only] 1112 By default the kernel will trim any uncacheable 1113 memory out of your available memory pool based on 1114 MTRR settings. This parameter disables that behavior, 1115 possibly causing your machine to run very slowly. 1116 1117 disable_timer_pin_1 [X86] 1118 Disable PIN 1 of APIC timer 1119 Can be useful to work around chipset bugs. 1120 1121 dis_ucode_ldr [X86] Disable the microcode loader. 1122 1123 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support, 1124 this option disables the debugging code at boot. 1125 1126 dma_debug_entries=<number> 1127 This option allows to tune the number of preallocated 1128 entries for DMA-API debugging code. One entry is 1129 required per DMA-API allocation. Use this if the 1130 DMA-API debugging code disables itself because the 1131 architectural default is too low. 1132 1133 dma_debug_driver=<driver_name> 1134 With this option the DMA-API debugging driver 1135 filter feature can be enabled at boot time. Just 1136 pass the driver to filter for as the parameter. 1137 The filter can be disabled or changed to another 1138 driver later using sysfs. 1139 1140 driver_async_probe= [KNL] 1141 List of driver names to be probed asynchronously. * 1142 matches with all driver names. If * is specified, the 1143 rest of the listed driver names are those that will NOT 1144 match the *. 1145 Format: <driver_name1>,<driver_name2>... 1146 1147 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>] 1148 Broken monitors, graphic adapters, KVMs and EDIDless 1149 panels may send no or incorrect EDID data sets. 1150 This parameter allows to specify an EDID data sets 1151 in the /lib/firmware directory that are used instead. 1152 Generic built-in EDID data sets are used, if one of 1153 edid/1024x768.bin, edid/1280x1024.bin, 1154 edid/1680x1050.bin, or edid/1920x1080.bin is given 1155 and no file with the same name exists. Details and 1156 instructions how to build your own EDID data are 1157 available in Documentation/admin-guide/edid.rst. An EDID 1158 data set will only be used for a particular connector, 1159 if its name and a colon are prepended to the EDID 1160 name. Each connector may use a unique EDID data 1161 set by separating the files with a comma. An EDID 1162 data set with no connector name will be used for 1163 any connectors not explicitly specified. 1164 1165 dscc4.setup= [NET] 1166 1167 dt_cpu_ftrs= [PPC] 1168 Format: {"off" | "known"} 1169 Control how the dt_cpu_ftrs device-tree binding is 1170 used for CPU feature discovery and setup (if it 1171 exists). 1172 off: Do not use it, fall back to legacy cpu table. 1173 known: Do not pass through unknown features to guests 1174 or userspace, only those that the kernel is aware of. 1175 1176 dump_apple_properties [X86] 1177 Dump name and content of EFI device properties on 1178 x86 Macs. Useful for driver authors to determine 1179 what data is available or for reverse-engineering. 1180 1181 dyndbg[="val"] [KNL,DYNAMIC_DEBUG] 1182 <module>.dyndbg[="val"] 1183 Enable debug messages at boot time. See 1184 Documentation/admin-guide/dynamic-debug-howto.rst 1185 for details. 1186 1187 early_ioremap_debug [KNL] 1188 Enable debug messages in early_ioremap support. This 1189 is useful for tracking down temporary early mappings 1190 which are not unmapped. 1191 1192 earlycon= [KNL] Output early console device and options. 1193 1194 When used with no options, the early console is 1195 determined by stdout-path property in device tree's 1196 chosen node or the ACPI SPCR table if supported by 1197 the platform. 1198 1199 cdns,<addr>[,options] 1200 Start an early, polled-mode console on a Cadence 1201 (xuartps) serial port at the specified address. Only 1202 supported option is baud rate. If baud rate is not 1203 specified, the serial port must already be setup and 1204 configured. 1205 1206 uart[8250],io,<addr>[,options[,uartclk]] 1207 uart[8250],mmio,<addr>[,options[,uartclk]] 1208 uart[8250],mmio32,<addr>[,options[,uartclk]] 1209 uart[8250],mmio32be,<addr>[,options[,uartclk]] 1210 uart[8250],0x<addr>[,options] 1211 Start an early, polled-mode console on the 8250/16550 1212 UART at the specified I/O port or MMIO address. 1213 MMIO inter-register address stride is either 8-bit 1214 (mmio) or 32-bit (mmio32 or mmio32be). 1215 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed 1216 to be equivalent to 'mmio'. 'options' are specified 1217 in the same format described for "console=ttyS<n>"; if 1218 unspecified, the h/w is not initialized. 'uartclk' is 1219 the uart clock frequency; if unspecified, it is set 1220 to 'BASE_BAUD' * 16. 1221 1222 pl011,<addr> 1223 pl011,mmio32,<addr> 1224 Start an early, polled-mode console on a pl011 serial 1225 port at the specified address. The pl011 serial port 1226 must already be setup and configured. Options are not 1227 yet supported. If 'mmio32' is specified, then only 1228 the driver will use only 32-bit accessors to read/write 1229 the device registers. 1230 1231 liteuart,<addr> 1232 Start an early console on a litex serial port at the 1233 specified address. The serial port must already be 1234 setup and configured. Options are not yet supported. 1235 1236 meson,<addr> 1237 Start an early, polled-mode console on a meson serial 1238 port at the specified address. The serial port must 1239 already be setup and configured. Options are not yet 1240 supported. 1241 1242 msm_serial,<addr> 1243 Start an early, polled-mode console on an msm serial 1244 port at the specified address. The serial port 1245 must already be setup and configured. Options are not 1246 yet supported. 1247 1248 msm_serial_dm,<addr> 1249 Start an early, polled-mode console on an msm serial 1250 dm port at the specified address. The serial port 1251 must already be setup and configured. Options are not 1252 yet supported. 1253 1254 owl,<addr> 1255 Start an early, polled-mode console on a serial port 1256 of an Actions Semi SoC, such as S500 or S900, at the 1257 specified address. The serial port must already be 1258 setup and configured. Options are not yet supported. 1259 1260 rda,<addr> 1261 Start an early, polled-mode console on a serial port 1262 of an RDA Micro SoC, such as RDA8810PL, at the 1263 specified address. The serial port must already be 1264 setup and configured. Options are not yet supported. 1265 1266 sbi 1267 Use RISC-V SBI (Supervisor Binary Interface) for early 1268 console. 1269 1270 smh Use ARM semihosting calls for early console. 1271 1272 s3c2410,<addr> 1273 s3c2412,<addr> 1274 s3c2440,<addr> 1275 s3c6400,<addr> 1276 s5pv210,<addr> 1277 exynos4210,<addr> 1278 Use early console provided by serial driver available 1279 on Samsung SoCs, requires selecting proper type and 1280 a correct base address of the selected UART port. The 1281 serial port must already be setup and configured. 1282 Options are not yet supported. 1283 1284 lantiq,<addr> 1285 Start an early, polled-mode console on a lantiq serial 1286 (lqasc) port at the specified address. The serial port 1287 must already be setup and configured. Options are not 1288 yet supported. 1289 1290 lpuart,<addr> 1291 lpuart32,<addr> 1292 Use early console provided by Freescale LP UART driver 1293 found on Freescale Vybrid and QorIQ LS1021A processors. 1294 A valid base address must be provided, and the serial 1295 port must already be setup and configured. 1296 1297 ec_imx21,<addr> 1298 ec_imx6q,<addr> 1299 Start an early, polled-mode, output-only console on the 1300 Freescale i.MX UART at the specified address. The UART 1301 must already be setup and configured. 1302 1303 ar3700_uart,<addr> 1304 Start an early, polled-mode console on the 1305 Armada 3700 serial port at the specified 1306 address. The serial port must already be setup 1307 and configured. Options are not yet supported. 1308 1309 qcom_geni,<addr> 1310 Start an early, polled-mode console on a Qualcomm 1311 Generic Interface (GENI) based serial port at the 1312 specified address. The serial port must already be 1313 setup and configured. Options are not yet supported. 1314 1315 efifb,[options] 1316 Start an early, unaccelerated console on the EFI 1317 memory mapped framebuffer (if available). On cache 1318 coherent non-x86 systems that use system memory for 1319 the framebuffer, pass the 'ram' option so that it is 1320 mapped with the correct attributes. 1321 1322 linflex,<addr> 1323 Use early console provided by Freescale LINFlexD UART 1324 serial driver for NXP S32V234 SoCs. A valid base 1325 address must be provided, and the serial port must 1326 already be setup and configured. 1327 1328 earlyprintk= [X86,SH,ARM,M68k,S390] 1329 earlyprintk=vga 1330 earlyprintk=sclp 1331 earlyprintk=xen 1332 earlyprintk=serial[,ttySn[,baudrate]] 1333 earlyprintk=serial[,0x...[,baudrate]] 1334 earlyprintk=ttySn[,baudrate] 1335 earlyprintk=dbgp[debugController#] 1336 earlyprintk=pciserial[,force],bus:device.function[,baudrate] 1337 earlyprintk=xdbc[xhciController#] 1338 1339 earlyprintk is useful when the kernel crashes before 1340 the normal console is initialized. It is not enabled by 1341 default because it has some cosmetic problems. 1342 1343 Append ",keep" to not disable it when the real console 1344 takes over. 1345 1346 Only one of vga, serial, or usb debug port can 1347 be used at a time. 1348 1349 Currently only ttyS0 and ttyS1 may be specified by 1350 name. Other I/O ports may be explicitly specified 1351 on some architectures (x86 and arm at least) by 1352 replacing ttySn with an I/O port address, like this: 1353 earlyprintk=serial,0x1008,115200 1354 You can find the port for a given device in 1355 /proc/tty/driver/serial: 1356 2: uart:ST16650V2 port:00001008 irq:18 ... 1357 1358 Interaction with the standard serial driver is not 1359 very good. 1360 1361 The VGA output is eventually overwritten by 1362 the real console. 1363 1364 The xen option can only be used in Xen domains. 1365 1366 The sclp output can only be used on s390. 1367 1368 The optional "force" to "pciserial" enables use of a 1369 PCI device even when its classcode is not of the 1370 UART class. 1371 1372 edac_report= [HW,EDAC] Control how to report EDAC event 1373 Format: {"on" | "off" | "force"} 1374 on: enable EDAC to report H/W event. May be overridden 1375 by other higher priority error reporting module. 1376 off: disable H/W event reporting through EDAC. 1377 force: enforce the use of EDAC to report H/W event. 1378 default: on. 1379 1380 edd= [EDD] 1381 Format: {"off" | "on" | "skip[mbr]"} 1382 1383 efi= [EFI] 1384 Format: { "debug", "disable_early_pci_dma", 1385 "nochunk", "noruntime", "nosoftreserve", 1386 "novamap", "no_disable_early_pci_dma" } 1387 debug: enable misc debug output. 1388 disable_early_pci_dma: disable the busmaster bit on all 1389 PCI bridges while in the EFI boot stub. 1390 nochunk: disable reading files in "chunks" in the EFI 1391 boot stub, as chunking can cause problems with some 1392 firmware implementations. 1393 noruntime : disable EFI runtime services support 1394 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose) 1395 attribute may cause the kernel to reserve the 1396 memory range for a memory mapping driver to 1397 claim. Specify efi=nosoftreserve to disable this 1398 reservation and treat the memory by its base type 1399 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM"). 1400 novamap: do not call SetVirtualAddressMap(). 1401 no_disable_early_pci_dma: Leave the busmaster bit set 1402 on all PCI bridges while in the EFI boot stub 1403 1404 efi_no_storage_paranoia [EFI; X86] 1405 Using this parameter you can use more than 50% of 1406 your efi variable storage. Use this parameter only if 1407 you are really sure that your UEFI does sane gc and 1408 fulfills the spec otherwise your board may brick. 1409 1410 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86] 1411 Add arbitrary attribute to specific memory range by 1412 updating original EFI memory map. 1413 Region of memory which aa attribute is added to is 1414 from ss to ss+nn. 1415 1416 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000 1417 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000) 1418 attribute is added to range 0x100000000-0x180000000 and 1419 0x10a0000000-0x1120000000. 1420 1421 If efi_fake_mem=8G@9G:0x40000 is specified, the 1422 EFI_MEMORY_SP(0x40000) attribute is added to 1423 range 0x240000000-0x43fffffff. 1424 1425 Using this parameter you can do debugging of EFI memmap 1426 related features. For example, you can do debugging of 1427 Address Range Mirroring feature even if your box 1428 doesn't support it, or mark specific memory as 1429 "soft reserved". 1430 1431 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT 1432 that is to be dynamically loaded by Linux. If there are 1433 multiple variables with the same name but with different 1434 vendor GUIDs, all of them will be loaded. See 1435 Documentation/admin-guide/acpi/ssdt-overlays.rst for details. 1436 1437 1438 eisa_irq_edge= [PARISC,HW] 1439 See header of drivers/parisc/eisa.c. 1440 1441 ekgdboc= [X86,KGDB] Allow early kernel console debugging 1442 Format: ekgdboc=kbd 1443 1444 This is designed to be used in conjunction with 1445 the boot argument: earlyprintk=vga 1446 1447 This parameter works in place of the kgdboc parameter 1448 but can only be used if the backing tty is available 1449 very early in the boot process. For early debugging 1450 via a serial port see kgdboc_earlycon instead. 1451 1452 elanfreq= [X86-32] 1453 See comment before function elanfreq_setup() in 1454 arch/x86/kernel/cpu/cpufreq/elanfreq.c. 1455 1456 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390] 1457 Specifies physical address of start of kernel core 1458 image elf header and optionally the size. Generally 1459 kexec loader will pass this option to capture kernel. 1460 See Documentation/admin-guide/kdump/kdump.rst for details. 1461 1462 enable_mtrr_cleanup [X86] 1463 The kernel tries to adjust MTRR layout from continuous 1464 to discrete, to make X server driver able to add WB 1465 entry later. This parameter enables that. 1466 1467 enable_timer_pin_1 [X86] 1468 Enable PIN 1 of APIC timer 1469 Can be useful to work around chipset bugs 1470 (in particular on some ATI chipsets). 1471 The kernel tries to set a reasonable default. 1472 1473 enforcing= [SELINUX] Set initial enforcing status. 1474 Format: {"0" | "1"} 1475 See security/selinux/Kconfig help text. 1476 0 -- permissive (log only, no denials). 1477 1 -- enforcing (deny and log). 1478 Default value is 0. 1479 Value can be changed at runtime via 1480 /sys/fs/selinux/enforce. 1481 1482 erst_disable [ACPI] 1483 Disable Error Record Serialization Table (ERST) 1484 support. 1485 1486 ether= [HW,NET] Ethernet cards parameters 1487 This option is obsoleted by the "netdev=" option, which 1488 has equivalent usage. See its documentation for details. 1489 1490 evm= [EVM] 1491 Format: { "fix" } 1492 Permit 'security.evm' to be updated regardless of 1493 current integrity status. 1494 1495 early_page_ext [KNL] Enforces page_ext initialization to earlier 1496 stages so cover more early boot allocations. 1497 Please note that as side effect some optimizations 1498 might be disabled to achieve that (e.g. parallelized 1499 memory initialization is disabled) so the boot process 1500 might take longer, especially on systems with a lot of 1501 memory. Available with CONFIG_PAGE_EXTENSION=y. 1502 1503 failslab= 1504 fail_usercopy= 1505 fail_page_alloc= 1506 fail_make_request=[KNL] 1507 General fault injection mechanism. 1508 Format: <interval>,<probability>,<space>,<times> 1509 See also Documentation/fault-injection/. 1510 1511 fb_tunnels= [NET] 1512 Format: { initns | none } 1513 See Documentation/admin-guide/sysctl/net.rst for 1514 fb_tunnels_only_for_init_ns 1515 1516 floppy= [HW] 1517 See Documentation/admin-guide/blockdev/floppy.rst. 1518 1519 force_pal_cache_flush 1520 [IA-64] Avoid check_sal_cache_flush which may hang on 1521 buggy SAL_CACHE_FLUSH implementations. Using this 1522 parameter will force ia64_sal_cache_flush to call 1523 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH. 1524 1525 forcepae [X86-32] 1526 Forcefully enable Physical Address Extension (PAE). 1527 Many Pentium M systems disable PAE but may have a 1528 functionally usable PAE implementation. 1529 Warning: use of this parameter will taint the kernel 1530 and may cause unknown problems. 1531 1532 ftrace=[tracer] 1533 [FTRACE] will set and start the specified tracer 1534 as early as possible in order to facilitate early 1535 boot debugging. 1536 1537 ftrace_boot_snapshot 1538 [FTRACE] On boot up, a snapshot will be taken of the 1539 ftrace ring buffer that can be read at: 1540 /sys/kernel/tracing/snapshot. 1541 This is useful if you need tracing information from kernel 1542 boot up that is likely to be overridden by user space 1543 start up functionality. 1544 1545 Optionally, the snapshot can also be defined for a tracing 1546 instance that was created by the trace_instance= command 1547 line parameter. 1548 1549 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo 1550 1551 The above will cause the "foo" tracing instance to trigger 1552 a snapshot at the end of boot up. 1553 1554 ftrace_dump_on_oops[=orig_cpu] 1555 [FTRACE] will dump the trace buffers on oops. 1556 If no parameter is passed, ftrace will dump 1557 buffers of all CPUs, but if you pass orig_cpu, it will 1558 dump only the buffer of the CPU that triggered the 1559 oops. 1560 1561 ftrace_filter=[function-list] 1562 [FTRACE] Limit the functions traced by the function 1563 tracer at boot up. function-list is a comma-separated 1564 list of functions. This list can be changed at run 1565 time by the set_ftrace_filter file in the debugfs 1566 tracing directory. 1567 1568 ftrace_notrace=[function-list] 1569 [FTRACE] Do not trace the functions specified in 1570 function-list. This list can be changed at run time 1571 by the set_ftrace_notrace file in the debugfs 1572 tracing directory. 1573 1574 ftrace_graph_filter=[function-list] 1575 [FTRACE] Limit the top level callers functions traced 1576 by the function graph tracer at boot up. 1577 function-list is a comma-separated list of functions 1578 that can be changed at run time by the 1579 set_graph_function file in the debugfs tracing directory. 1580 1581 ftrace_graph_notrace=[function-list] 1582 [FTRACE] Do not trace from the functions specified in 1583 function-list. This list is a comma-separated list of 1584 functions that can be changed at run time by the 1585 set_graph_notrace file in the debugfs tracing directory. 1586 1587 ftrace_graph_max_depth=<uint> 1588 [FTRACE] Used with the function graph tracer. This is 1589 the max depth it will trace into a function. This value 1590 can be changed at run time by the max_graph_depth file 1591 in the tracefs tracing directory. default: 0 (no limit) 1592 1593 fw_devlink= [KNL] Create device links between consumer and supplier 1594 devices by scanning the firmware to infer the 1595 consumer/supplier relationships. This feature is 1596 especially useful when drivers are loaded as modules as 1597 it ensures proper ordering of tasks like device probing 1598 (suppliers first, then consumers), supplier boot state 1599 clean up (only after all consumers have probed), 1600 suspend/resume & runtime PM (consumers first, then 1601 suppliers). 1602 Format: { off | permissive | on | rpm } 1603 off -- Don't create device links from firmware info. 1604 permissive -- Create device links from firmware info 1605 but use it only for ordering boot state clean 1606 up (sync_state() calls). 1607 on -- Create device links from firmware info and use it 1608 to enforce probe and suspend/resume ordering. 1609 rpm -- Like "on", but also use to order runtime PM. 1610 1611 fw_devlink.strict=<bool> 1612 [KNL] Treat all inferred dependencies as mandatory 1613 dependencies. This only applies for fw_devlink=on|rpm. 1614 Format: <bool> 1615 1616 fw_devlink.sync_state = 1617 [KNL] When all devices that could probe have finished 1618 probing, this parameter controls what to do with 1619 devices that haven't yet received their sync_state() 1620 calls. 1621 Format: { strict | timeout } 1622 strict -- Default. Continue waiting on consumers to 1623 probe successfully. 1624 timeout -- Give up waiting on consumers and call 1625 sync_state() on any devices that haven't yet 1626 received their sync_state() calls after 1627 deferred_probe_timeout has expired or by 1628 late_initcall() if !CONFIG_MODULES. 1629 1630 gamecon.map[2|3]= 1631 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad 1632 support via parallel port (up to 5 devices per port) 1633 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5> 1634 See also Documentation/input/devices/joystick-parport.rst 1635 1636 gamma= [HW,DRM] 1637 1638 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART 1639 Format: off | on 1640 default: on 1641 1642 gather_data_sampling= 1643 [X86,INTEL] Control the Gather Data Sampling (GDS) 1644 mitigation. 1645 1646 Gather Data Sampling is a hardware vulnerability which 1647 allows unprivileged speculative access to data which was 1648 previously stored in vector registers. 1649 1650 This issue is mitigated by default in updated microcode. 1651 The mitigation may have a performance impact but can be 1652 disabled. On systems without the microcode mitigation 1653 disabling AVX serves as a mitigation. 1654 1655 force: Disable AVX to mitigate systems without 1656 microcode mitigation. No effect if the microcode 1657 mitigation is present. Known to cause crashes in 1658 userspace with buggy AVX enumeration. 1659 1660 off: Disable GDS mitigation. 1661 1662 gcov_persist= [GCOV] When non-zero (default), profiling data for 1663 kernel modules is saved and remains accessible via 1664 debugfs, even when the module is unloaded/reloaded. 1665 When zero, profiling data is discarded and associated 1666 debugfs files are removed at module unload time. 1667 1668 goldfish [X86] Enable the goldfish android emulator platform. 1669 Don't use this when you are not running on the 1670 android emulator 1671 1672 gpio-mockup.gpio_mockup_ranges 1673 [HW] Sets the ranges of gpiochip of for this device. 1674 Format: <start1>,<end1>,<start2>,<end2>... 1675 gpio-mockup.gpio_mockup_named_lines 1676 [HW] Let the driver know GPIO lines should be named. 1677 1678 gpt [EFI] Forces disk with valid GPT signature but 1679 invalid Protective MBR to be treated as GPT. If the 1680 primary GPT is corrupted, it enables the backup/alternate 1681 GPT to be used instead. 1682 1683 grcan.enable0= [HW] Configuration of physical interface 0. Determines 1684 the "Enable 0" bit of the configuration register. 1685 Format: 0 | 1 1686 Default: 0 1687 grcan.enable1= [HW] Configuration of physical interface 1. Determines 1688 the "Enable 0" bit of the configuration register. 1689 Format: 0 | 1 1690 Default: 0 1691 grcan.select= [HW] Select which physical interface to use. 1692 Format: 0 | 1 1693 Default: 0 1694 grcan.txsize= [HW] Sets the size of the tx buffer. 1695 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0. 1696 Default: 1024 1697 grcan.rxsize= [HW] Sets the size of the rx buffer. 1698 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0. 1699 Default: 1024 1700 1701 hardened_usercopy= 1702 [KNL] Under CONFIG_HARDENED_USERCOPY, whether 1703 hardening is enabled for this boot. Hardened 1704 usercopy checking is used to protect the kernel 1705 from reading or writing beyond known memory 1706 allocation boundaries as a proactive defense 1707 against bounds-checking flaws in the kernel's 1708 copy_to_user()/copy_from_user() interface. 1709 on Perform hardened usercopy checks (default). 1710 off Disable hardened usercopy checks. 1711 1712 hardlockup_all_cpu_backtrace= 1713 [KNL] Should the hard-lockup detector generate 1714 backtraces on all cpus. 1715 Format: 0 | 1 1716 1717 hashdist= [KNL,NUMA] Large hashes allocated during boot 1718 are distributed across NUMA nodes. Defaults on 1719 for 64-bit NUMA, off otherwise. 1720 Format: 0 | 1 (for off | on) 1721 1722 hcl= [IA-64] SGI's Hardware Graph compatibility layer 1723 1724 hd= [EIDE] (E)IDE hard drive subsystem geometry 1725 Format: <cyl>,<head>,<sect> 1726 1727 hest_disable [ACPI] 1728 Disable Hardware Error Source Table (HEST) support; 1729 corresponding firmware-first mode error processing 1730 logic will be disabled. 1731 1732 hibernate= [HIBERNATION] 1733 noresume Don't check if there's a hibernation image 1734 present during boot. 1735 nocompress Don't compress/decompress hibernation images. 1736 no Disable hibernation and resume. 1737 protect_image Turn on image protection during restoration 1738 (that will set all pages holding image data 1739 during restoration read-only). 1740 1741 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact 1742 size of <nn>. This works even on boxes that have no 1743 highmem otherwise. This also works to reduce highmem 1744 size on bigger boxes. 1745 1746 highres= [KNL] Enable/disable high resolution timer mode. 1747 Valid parameters: "on", "off" 1748 Default: "on" 1749 1750 hlt [BUGS=ARM,SH] 1751 1752 hostname= [KNL] Set the hostname (aka UTS nodename). 1753 Format: <string> 1754 This allows setting the system's hostname during early 1755 startup. This sets the name returned by gethostname. 1756 Using this parameter to set the hostname makes it 1757 possible to ensure the hostname is correctly set before 1758 any userspace processes run, avoiding the possibility 1759 that a process may call gethostname before the hostname 1760 has been explicitly set, resulting in the calling 1761 process getting an incorrect result. The string must 1762 not exceed the maximum allowed hostname length (usually 1763 64 characters) and will be truncated otherwise. 1764 1765 hpet= [X86-32,HPET] option to control HPET usage 1766 Format: { enable (default) | disable | force | 1767 verbose } 1768 disable: disable HPET and use PIT instead 1769 force: allow force enabled of undocumented chips (ICH4, 1770 VIA, nVidia) 1771 verbose: show contents of HPET registers during setup 1772 1773 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET 1774 registers. Default set by CONFIG_HPET_MMAP_DEFAULT. 1775 1776 hugepages= [HW] Number of HugeTLB pages to allocate at boot. 1777 If this follows hugepagesz (below), it specifies 1778 the number of pages of hugepagesz to be allocated. 1779 If this is the first HugeTLB parameter on the command 1780 line, it specifies the number of pages to allocate for 1781 the default huge page size. If using node format, the 1782 number of pages to allocate per-node can be specified. 1783 See also Documentation/admin-guide/mm/hugetlbpage.rst. 1784 Format: <integer> or (node format) 1785 <node>:<integer>[,<node>:<integer>] 1786 1787 hugepagesz= 1788 [HW] The size of the HugeTLB pages. This is used in 1789 conjunction with hugepages (above) to allocate huge 1790 pages of a specific size at boot. The pair 1791 hugepagesz=X hugepages=Y can be specified once for 1792 each supported huge page size. Huge page sizes are 1793 architecture dependent. See also 1794 Documentation/admin-guide/mm/hugetlbpage.rst. 1795 Format: size[KMG] 1796 1797 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation 1798 of gigantic hugepages. Or using node format, the size 1799 of a CMA area per node can be specified. 1800 Format: nn[KMGTPE] or (node format) 1801 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]] 1802 1803 Reserve a CMA area of given size and allocate gigantic 1804 hugepages using the CMA allocator. If enabled, the 1805 boot-time allocation of gigantic hugepages is skipped. 1806 1807 hugetlb_free_vmemmap= 1808 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP 1809 enabled. 1810 Control if HugeTLB Vmemmap Optimization (HVO) is enabled. 1811 Allows heavy hugetlb users to free up some more 1812 memory (7 * PAGE_SIZE for each 2MB hugetlb page). 1813 Format: { on | off (default) } 1814 1815 on: enable HVO 1816 off: disable HVO 1817 1818 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y, 1819 the default is on. 1820 1821 Note that the vmemmap pages may be allocated from the added 1822 memory block itself when memory_hotplug.memmap_on_memory is 1823 enabled, those vmemmap pages cannot be optimized even if this 1824 feature is enabled. Other vmemmap pages not allocated from 1825 the added memory block itself do not be affected. 1826 1827 hung_task_panic= 1828 [KNL] Should the hung task detector generate panics. 1829 Format: 0 | 1 1830 1831 A value of 1 instructs the kernel to panic when a 1832 hung task is detected. The default value is controlled 1833 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time 1834 option. The value selected by this boot parameter can 1835 be changed later by the kernel.hung_task_panic sysctl. 1836 1837 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC) 1838 terminal devices. Valid values: 0..8 1839 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs. 1840 If specified, z/VM IUCV HVC accepts connections 1841 from listed z/VM user IDs only. 1842 1843 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations 1844 which allow the hypervisor to 'idle' the 1845 guest on lock contention. 1846 1847 i2c_bus= [HW] Override the default board specific I2C bus speed 1848 or register an additional I2C bus that is not 1849 registered from board initialization code. 1850 Format: 1851 <bus_id>,<clkrate> 1852 1853 i8042.debug [HW] Toggle i8042 debug mode 1854 i8042.unmask_kbd_data 1855 [HW] Enable printing of interrupt data from the KBD port 1856 (disabled by default, and as a pre-condition 1857 requires that i8042.debug=1 be enabled) 1858 i8042.direct [HW] Put keyboard port into non-translated mode 1859 i8042.dumbkbd [HW] Pretend that controller can only read data from 1860 keyboard and cannot control its state 1861 (Don't attempt to blink the leds) 1862 i8042.noaux [HW] Don't check for auxiliary (== mouse) port 1863 i8042.nokbd [HW] Don't check/create keyboard port 1864 i8042.noloop [HW] Disable the AUX Loopback command while probing 1865 for the AUX port 1866 i8042.nomux [HW] Don't check presence of an active multiplexing 1867 controller 1868 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX 1869 controllers 1870 i8042.notimeout [HW] Ignore timeout condition signalled by controller 1871 i8042.reset [HW] Reset the controller during init, cleanup and 1872 suspend-to-ram transitions, only during s2r 1873 transitions, or never reset 1874 Format: { 1 | Y | y | 0 | N | n } 1875 1, Y, y: always reset controller 1876 0, N, n: don't ever reset controller 1877 Default: only on s2r transitions on x86; most other 1878 architectures force reset to be always executed 1879 i8042.unlock [HW] Unlock (ignore) the keylock 1880 i8042.kbdreset [HW] Reset device connected to KBD port 1881 i8042.probe_defer 1882 [HW] Allow deferred probing upon i8042 probe errors 1883 1884 i810= [HW,DRM] 1885 1886 i915.invert_brightness= 1887 [DRM] Invert the sense of the variable that is used to 1888 set the brightness of the panel backlight. Normally a 1889 brightness value of 0 indicates backlight switched off, 1890 and the maximum of the brightness value sets the backlight 1891 to maximum brightness. If this parameter is set to 0 1892 (default) and the machine requires it, or this parameter 1893 is set to 1, a brightness value of 0 sets the backlight 1894 to maximum brightness, and the maximum of the brightness 1895 value switches the backlight off. 1896 -1 -- never invert brightness 1897 0 -- machine default 1898 1 -- force brightness inversion 1899 1900 ia32_emulation= [X86-64] 1901 Format: <bool> 1902 When true, allows loading 32-bit programs and executing 32-bit 1903 syscalls, essentially overriding IA32_EMULATION_DEFAULT_DISABLED at 1904 boot time. When false, unconditionally disables IA32 emulation. 1905 1906 icn= [HW,ISDN] 1907 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]] 1908 1909 1910 idle= [X86] 1911 Format: idle=poll, idle=halt, idle=nomwait 1912 Poll forces a polling idle loop that can slightly 1913 improve the performance of waking up a idle CPU, but 1914 will use a lot of power and make the system run hot. 1915 Not recommended. 1916 idle=halt: Halt is forced to be used for CPU idle. 1917 In such case C2/C3 won't be used again. 1918 idle=nomwait: Disable mwait for CPU C-states 1919 1920 idxd.sva= [HW] 1921 Format: <bool> 1922 Allow force disabling of Shared Virtual Memory (SVA) 1923 support for the idxd driver. By default it is set to 1924 true (1). 1925 1926 idxd.tc_override= [HW] 1927 Format: <bool> 1928 Allow override of default traffic class configuration 1929 for the device. By default it is set to false (0). 1930 1931 ieee754= [MIPS] Select IEEE Std 754 conformance mode 1932 Format: { strict | legacy | 2008 | relaxed } 1933 Default: strict 1934 1935 Choose which programs will be accepted for execution 1936 based on the IEEE 754 NaN encoding(s) supported by 1937 the FPU and the NaN encoding requested with the value 1938 of an ELF file header flag individually set by each 1939 binary. Hardware implementations are permitted to 1940 support either or both of the legacy and the 2008 NaN 1941 encoding mode. 1942 1943 Available settings are as follows: 1944 strict accept binaries that request a NaN encoding 1945 supported by the FPU 1946 legacy only accept legacy-NaN binaries, if supported 1947 by the FPU 1948 2008 only accept 2008-NaN binaries, if supported 1949 by the FPU 1950 relaxed accept any binaries regardless of whether 1951 supported by the FPU 1952 1953 The FPU emulator is always able to support both NaN 1954 encodings, so if no FPU hardware is present or it has 1955 been disabled with 'nofpu', then the settings of 1956 'legacy' and '2008' strap the emulator accordingly, 1957 'relaxed' straps the emulator for both legacy-NaN and 1958 2008-NaN, whereas 'strict' enables legacy-NaN only on 1959 legacy processors and both NaN encodings on MIPS32 or 1960 MIPS64 CPUs. 1961 1962 The setting for ABS.fmt/NEG.fmt instruction execution 1963 mode generally follows that for the NaN encoding, 1964 except where unsupported by hardware. 1965 1966 ignore_loglevel [KNL] 1967 Ignore loglevel setting - this will print /all/ 1968 kernel messages to the console. Useful for debugging. 1969 We also add it as printk module parameter, so users 1970 could change it dynamically, usually by 1971 /sys/module/printk/parameters/ignore_loglevel. 1972 1973 ignore_rlimit_data 1974 Ignore RLIMIT_DATA setting for data mappings, 1975 print warning at first misuse. Can be changed via 1976 /sys/module/kernel/parameters/ignore_rlimit_data. 1977 1978 ihash_entries= [KNL] 1979 Set number of hash buckets for inode cache. 1980 1981 ima_appraise= [IMA] appraise integrity measurements 1982 Format: { "off" | "enforce" | "fix" | "log" } 1983 default: "enforce" 1984 1985 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead. 1986 The builtin appraise policy appraises all files 1987 owned by uid=0. 1988 1989 ima_canonical_fmt [IMA] 1990 Use the canonical format for the binary runtime 1991 measurements, instead of host native format. 1992 1993 ima_hash= [IMA] 1994 Format: { md5 | sha1 | rmd160 | sha256 | sha384 1995 | sha512 | ... } 1996 default: "sha1" 1997 1998 The list of supported hash algorithms is defined 1999 in crypto/hash_info.h. 2000 2001 ima_policy= [IMA] 2002 The builtin policies to load during IMA setup. 2003 Format: "tcb | appraise_tcb | secure_boot | 2004 fail_securely | critical_data" 2005 2006 The "tcb" policy measures all programs exec'd, files 2007 mmap'd for exec, and all files opened with the read 2008 mode bit set by either the effective uid (euid=0) or 2009 uid=0. 2010 2011 The "appraise_tcb" policy appraises the integrity of 2012 all files owned by root. 2013 2014 The "secure_boot" policy appraises the integrity 2015 of files (eg. kexec kernel image, kernel modules, 2016 firmware, policy, etc) based on file signatures. 2017 2018 The "fail_securely" policy forces file signature 2019 verification failure also on privileged mounted 2020 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE 2021 flag. 2022 2023 The "critical_data" policy measures kernel integrity 2024 critical data. 2025 2026 ima_tcb [IMA] Deprecated. Use ima_policy= instead. 2027 Load a policy which meets the needs of the Trusted 2028 Computing Base. This means IMA will measure all 2029 programs exec'd, files mmap'd for exec, and all files 2030 opened for read by uid=0. 2031 2032 ima_template= [IMA] 2033 Select one of defined IMA measurements template formats. 2034 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" | 2035 "ima-sigv2" } 2036 Default: "ima-ng" 2037 2038 ima_template_fmt= 2039 [IMA] Define a custom template format. 2040 Format: { "field1|...|fieldN" } 2041 2042 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage 2043 Format: <min_file_size> 2044 Set the minimal file size for using asynchronous hash. 2045 If left unspecified, ahash usage is disabled. 2046 2047 ahash performance varies for different data sizes on 2048 different crypto accelerators. This option can be used 2049 to achieve the best performance for a particular HW. 2050 2051 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size 2052 Format: <bufsize> 2053 Set hashing buffer size. Default: 4k. 2054 2055 ahash performance varies for different chunk sizes on 2056 different crypto accelerators. This option can be used 2057 to achieve best performance for particular HW. 2058 2059 init= [KNL] 2060 Format: <full_path> 2061 Run specified binary instead of /sbin/init as init 2062 process. 2063 2064 initcall_debug [KNL] Trace initcalls as they are executed. Useful 2065 for working out where the kernel is dying during 2066 startup. 2067 2068 initcall_blacklist= [KNL] Do not execute a comma-separated list of 2069 initcall functions. Useful for debugging built-in 2070 modules and initcalls. 2071 2072 initramfs_async= [KNL] 2073 Format: <bool> 2074 Default: 1 2075 This parameter controls whether the initramfs 2076 image is unpacked asynchronously, concurrently 2077 with devices being probed and 2078 initialized. This should normally just work, 2079 but as a debugging aid, one can get the 2080 historical behaviour of the initramfs 2081 unpacking being completed before device_ and 2082 late_ initcalls. 2083 2084 initrd= [BOOT] Specify the location of the initial ramdisk 2085 2086 initrdmem= [KNL] Specify a physical address and size from which to 2087 load the initrd. If an initrd is compiled in or 2088 specified in the bootparams, it takes priority over this 2089 setting. 2090 Format: ss[KMG],nn[KMG] 2091 Default is 0, 0 2092 2093 init_on_alloc= [MM] Fill newly allocated pages and heap objects with 2094 zeroes. 2095 Format: 0 | 1 2096 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON. 2097 2098 init_on_free= [MM] Fill freed pages and heap objects with zeroes. 2099 Format: 0 | 1 2100 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON. 2101 2102 init_pkru= [X86] Specify the default memory protection keys rights 2103 register contents for all processes. 0x55555554 by 2104 default (disallow access to all but pkey 0). Can 2105 override in debugfs after boot. 2106 2107 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver 2108 Format: <irq> 2109 2110 int_pln_enable [X86] Enable power limit notification interrupt 2111 2112 integrity_audit=[IMA] 2113 Format: { "0" | "1" } 2114 0 -- basic integrity auditing messages. (Default) 2115 1 -- additional integrity auditing messages. 2116 2117 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option 2118 on 2119 Enable intel iommu driver. 2120 off 2121 Disable intel iommu driver. 2122 igfx_off [Default Off] 2123 By default, gfx is mapped as normal device. If a gfx 2124 device has a dedicated DMAR unit, the DMAR unit is 2125 bypassed by not enabling DMAR with this option. In 2126 this case, gfx device will use physical address for 2127 DMA. 2128 strict [Default Off] 2129 Deprecated, equivalent to iommu.strict=1. 2130 sp_off [Default Off] 2131 By default, super page will be supported if Intel IOMMU 2132 has the capability. With this option, super page will 2133 not be supported. 2134 sm_on 2135 Enable the Intel IOMMU scalable mode if the hardware 2136 advertises that it has support for the scalable mode 2137 translation. 2138 sm_off 2139 Disallow use of the Intel IOMMU scalable mode. 2140 tboot_noforce [Default Off] 2141 Do not force the Intel IOMMU enabled under tboot. 2142 By default, tboot will force Intel IOMMU on, which 2143 could harm performance of some high-throughput 2144 devices like 40GBit network cards, even if identity 2145 mapping is enabled. 2146 Note that using this option lowers the security 2147 provided by tboot because it makes the system 2148 vulnerable to DMA attacks. 2149 2150 intel_idle.max_cstate= [KNL,HW,ACPI,X86] 2151 0 disables intel_idle and fall back on acpi_idle. 2152 1 to 9 specify maximum depth of C-state. 2153 2154 intel_pstate= [X86] 2155 disable 2156 Do not enable intel_pstate as the default 2157 scaling driver for the supported processors 2158 active 2159 Use intel_pstate driver to bypass the scaling 2160 governors layer of cpufreq and provides it own 2161 algorithms for p-state selection. There are two 2162 P-state selection algorithms provided by 2163 intel_pstate in the active mode: powersave and 2164 performance. The way they both operate depends 2165 on whether or not the hardware managed P-states 2166 (HWP) feature has been enabled in the processor 2167 and possibly on the processor model. 2168 passive 2169 Use intel_pstate as a scaling driver, but configure it 2170 to work with generic cpufreq governors (instead of 2171 enabling its internal governor). This mode cannot be 2172 used along with the hardware-managed P-states (HWP) 2173 feature. 2174 force 2175 Enable intel_pstate on systems that prohibit it by default 2176 in favor of acpi-cpufreq. Forcing the intel_pstate driver 2177 instead of acpi-cpufreq may disable platform features, such 2178 as thermal controls and power capping, that rely on ACPI 2179 P-States information being indicated to OSPM and therefore 2180 should be used with caution. This option does not work with 2181 processors that aren't supported by the intel_pstate driver 2182 or on platforms that use pcc-cpufreq instead of acpi-cpufreq. 2183 no_hwp 2184 Do not enable hardware P state control (HWP) 2185 if available. 2186 hwp_only 2187 Only load intel_pstate on systems which support 2188 hardware P state control (HWP) if available. 2189 support_acpi_ppc 2190 Enforce ACPI _PPC performance limits. If the Fixed ACPI 2191 Description Table, specifies preferred power management 2192 profile as "Enterprise Server" or "Performance Server", 2193 then this feature is turned on by default. 2194 per_cpu_perf_limits 2195 Allow per-logical-CPU P-State performance control limits using 2196 cpufreq sysfs interface 2197 2198 intremap= [X86-64, Intel-IOMMU] 2199 on enable Interrupt Remapping (default) 2200 off disable Interrupt Remapping 2201 nosid disable Source ID checking 2202 no_x2apic_optout 2203 BIOS x2APIC opt-out request will be ignored 2204 nopost disable Interrupt Posting 2205 2206 iomem= Disable strict checking of access to MMIO memory 2207 strict regions from userspace. 2208 relaxed 2209 2210 iommu= [X86] 2211 off 2212 force 2213 noforce 2214 biomerge 2215 panic 2216 nopanic 2217 merge 2218 nomerge 2219 soft 2220 pt [X86] 2221 nopt [X86] 2222 nobypass [PPC/POWERNV] 2223 Disable IOMMU bypass, using IOMMU for PCI devices. 2224 2225 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices. 2226 Format: { "0" | "1" } 2227 0 - Try to allocate a 32-bit DMA address first, before 2228 falling back to the full range if needed. 2229 1 - Allocate directly from the full usable range, 2230 forcing Dual Address Cycle for PCI cards supporting 2231 greater than 32-bit addressing. 2232 2233 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour 2234 Format: { "0" | "1" } 2235 0 - Lazy mode. 2236 Request that DMA unmap operations use deferred 2237 invalidation of hardware TLBs, for increased 2238 throughput at the cost of reduced device isolation. 2239 Will fall back to strict mode if not supported by 2240 the relevant IOMMU driver. 2241 1 - Strict mode. 2242 DMA unmap operations invalidate IOMMU hardware TLBs 2243 synchronously. 2244 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}. 2245 Note: on x86, strict mode specified via one of the 2246 legacy driver-specific options takes precedence. 2247 2248 iommu.passthrough= 2249 [ARM64, X86] Configure DMA to bypass the IOMMU by default. 2250 Format: { "0" | "1" } 2251 0 - Use IOMMU translation for DMA. 2252 1 - Bypass the IOMMU for DMA. 2253 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH. 2254 2255 io7= [HW] IO7 for Marvel-based Alpha systems 2256 See comment before marvel_specify_io7 in 2257 arch/alpha/kernel/core_marvel.c. 2258 2259 io_delay= [X86] I/O delay method 2260 0x80 2261 Standard port 0x80 based delay 2262 0xed 2263 Alternate port 0xed based delay (needed on some systems) 2264 udelay 2265 Simple two microseconds delay 2266 none 2267 No delay 2268 2269 ip= [IP_PNP] 2270 See Documentation/admin-guide/nfs/nfsroot.rst. 2271 2272 ipcmni_extend [KNL] Extend the maximum number of unique System V 2273 IPC identifiers from 32,768 to 16,777,216. 2274 2275 irqaffinity= [SMP] Set the default irq affinity mask 2276 The argument is a cpu list, as described above. 2277 2278 irqchip.gicv2_force_probe= 2279 [ARM, ARM64] 2280 Format: <bool> 2281 Force the kernel to look for the second 4kB page 2282 of a GICv2 controller even if the memory range 2283 exposed by the device tree is too small. 2284 2285 irqchip.gicv3_nolpi= 2286 [ARM, ARM64] 2287 Force the kernel to ignore the availability of 2288 LPIs (and by consequence ITSs). Intended for system 2289 that use the kernel as a bootloader, and thus want 2290 to let secondary kernels in charge of setting up 2291 LPIs. 2292 2293 irqchip.gicv3_pseudo_nmi= [ARM64] 2294 Enables support for pseudo-NMIs in the kernel. This 2295 requires the kernel to be built with 2296 CONFIG_ARM64_PSEUDO_NMI. 2297 2298 irqfixup [HW] 2299 When an interrupt is not handled search all handlers 2300 for it. Intended to get systems with badly broken 2301 firmware running. 2302 2303 irqpoll [HW] 2304 When an interrupt is not handled search all handlers 2305 for it. Also check all handlers each timer 2306 interrupt. Intended to get systems with badly broken 2307 firmware running. 2308 2309 isapnp= [ISAPNP] 2310 Format: <RDP>,<reset>,<pci_scan>,<verbosity> 2311 2312 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance. 2313 [Deprecated - use cpusets instead] 2314 Format: [flag-list,]<cpu-list> 2315 2316 Specify one or more CPUs to isolate from disturbances 2317 specified in the flag list (default: domain): 2318 2319 nohz 2320 Disable the tick when a single task runs. 2321 2322 A residual 1Hz tick is offloaded to workqueues, which you 2323 need to affine to housekeeping through the global 2324 workqueue's affinity configured via the 2325 /sys/devices/virtual/workqueue/cpumask sysfs file, or 2326 by using the 'domain' flag described below. 2327 2328 NOTE: by default the global workqueue runs on all CPUs, 2329 so to protect individual CPUs the 'cpumask' file has to 2330 be configured manually after bootup. 2331 2332 domain 2333 Isolate from the general SMP balancing and scheduling 2334 algorithms. Note that performing domain isolation this way 2335 is irreversible: it's not possible to bring back a CPU to 2336 the domains once isolated through isolcpus. It's strongly 2337 advised to use cpusets instead to disable scheduler load 2338 balancing through the "cpuset.sched_load_balance" file. 2339 It offers a much more flexible interface where CPUs can 2340 move in and out of an isolated set anytime. 2341 2342 You can move a process onto or off an "isolated" CPU via 2343 the CPU affinity syscalls or cpuset. 2344 <cpu number> begins at 0 and the maximum value is 2345 "number of CPUs in system - 1". 2346 2347 managed_irq 2348 2349 Isolate from being targeted by managed interrupts 2350 which have an interrupt mask containing isolated 2351 CPUs. The affinity of managed interrupts is 2352 handled by the kernel and cannot be changed via 2353 the /proc/irq/* interfaces. 2354 2355 This isolation is best effort and only effective 2356 if the automatically assigned interrupt mask of a 2357 device queue contains isolated and housekeeping 2358 CPUs. If housekeeping CPUs are online then such 2359 interrupts are directed to the housekeeping CPU 2360 so that IO submitted on the housekeeping CPU 2361 cannot disturb the isolated CPU. 2362 2363 If a queue's affinity mask contains only isolated 2364 CPUs then this parameter has no effect on the 2365 interrupt routing decision, though interrupts are 2366 only delivered when tasks running on those 2367 isolated CPUs submit IO. IO submitted on 2368 housekeeping CPUs has no influence on those 2369 queues. 2370 2371 The format of <cpu-list> is described above. 2372 2373 iucv= [HW,NET] 2374 2375 ivrs_ioapic [HW,X86-64] 2376 Provide an override to the IOAPIC-ID<->DEVICE-ID 2377 mapping provided in the IVRS ACPI table. 2378 By default, PCI segment is 0, and can be omitted. 2379 2380 For example, to map IOAPIC-ID decimal 10 to 2381 PCI segment 0x1 and PCI device 00:14.0, 2382 write the parameter as: 2383 ivrs_ioapic=10@0001:00:14.0 2384 2385 Deprecated formats: 2386 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0 2387 write the parameter as: 2388 ivrs_ioapic[10]=00:14.0 2389 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and 2390 PCI device 00:14.0 write the parameter as: 2391 ivrs_ioapic[10]=0001:00:14.0 2392 2393 ivrs_hpet [HW,X86-64] 2394 Provide an override to the HPET-ID<->DEVICE-ID 2395 mapping provided in the IVRS ACPI table. 2396 By default, PCI segment is 0, and can be omitted. 2397 2398 For example, to map HPET-ID decimal 10 to 2399 PCI segment 0x1 and PCI device 00:14.0, 2400 write the parameter as: 2401 ivrs_hpet=10@0001:00:14.0 2402 2403 Deprecated formats: 2404 * To map HPET-ID decimal 0 to PCI device 00:14.0 2405 write the parameter as: 2406 ivrs_hpet[0]=00:14.0 2407 * To map HPET-ID decimal 10 to PCI segment 0x1 and 2408 PCI device 00:14.0 write the parameter as: 2409 ivrs_ioapic[10]=0001:00:14.0 2410 2411 ivrs_acpihid [HW,X86-64] 2412 Provide an override to the ACPI-HID:UID<->DEVICE-ID 2413 mapping provided in the IVRS ACPI table. 2414 By default, PCI segment is 0, and can be omitted. 2415 2416 For example, to map UART-HID:UID AMD0020:0 to 2417 PCI segment 0x1 and PCI device ID 00:14.5, 2418 write the parameter as: 2419 ivrs_acpihid=AMD0020:0@0001:00:14.5 2420 2421 Deprecated formats: 2422 * To map UART-HID:UID AMD0020:0 to PCI segment is 0, 2423 PCI device ID 00:14.5, write the parameter as: 2424 ivrs_acpihid[00:14.5]=AMD0020:0 2425 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and 2426 PCI device ID 00:14.5, write the parameter as: 2427 ivrs_acpihid[0001:00:14.5]=AMD0020:0 2428 2429 js= [HW,JOY] Analog joystick 2430 See Documentation/input/joydev/joystick.rst. 2431 2432 kasan_multi_shot 2433 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print 2434 report on every invalid memory access. Without this 2435 parameter KASAN will print report only for the first 2436 invalid access. 2437 2438 keep_bootcon [KNL] 2439 Do not unregister boot console at start. This is only 2440 useful for debugging when something happens in the window 2441 between unregistering the boot console and initializing 2442 the real console. 2443 2444 keepinitrd [HW,ARM] 2445 2446 kernelcore= [KNL,X86,IA-64,PPC] 2447 Format: nn[KMGTPE] | nn% | "mirror" 2448 This parameter specifies the amount of memory usable by 2449 the kernel for non-movable allocations. The requested 2450 amount is spread evenly throughout all nodes in the 2451 system as ZONE_NORMAL. The remaining memory is used for 2452 movable memory in its own zone, ZONE_MOVABLE. In the 2453 event, a node is too small to have both ZONE_NORMAL and 2454 ZONE_MOVABLE, kernelcore memory will take priority and 2455 other nodes will have a larger ZONE_MOVABLE. 2456 2457 ZONE_MOVABLE is used for the allocation of pages that 2458 may be reclaimed or moved by the page migration 2459 subsystem. Note that allocations like PTEs-from-HighMem 2460 still use the HighMem zone if it exists, and the Normal 2461 zone if it does not. 2462 2463 It is possible to specify the exact amount of memory in 2464 the form of "nn[KMGTPE]", a percentage of total system 2465 memory in the form of "nn%", or "mirror". If "mirror" 2466 option is specified, mirrored (reliable) memory is used 2467 for non-movable allocations and remaining memory is used 2468 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror" 2469 are exclusive, so you cannot specify multiple forms. 2470 2471 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port. 2472 Format: <Controller#>[,poll interval] 2473 The controller # is the number of the ehci usb debug 2474 port as it is probed via PCI. The poll interval is 2475 optional and is the number seconds in between 2476 each poll cycle to the debug port in case you need 2477 the functionality for interrupting the kernel with 2478 gdb or control-c on the dbgp connection. When 2479 not using this parameter you use sysrq-g to break into 2480 the kernel debugger. 2481 2482 kgdboc= [KGDB,HW] kgdb over consoles. 2483 Requires a tty driver that supports console polling, 2484 or a supported polling keyboard driver (non-usb). 2485 Serial only format: <serial_device>[,baud] 2486 keyboard only format: kbd 2487 keyboard and serial format: kbd,<serial_device>[,baud] 2488 Optional Kernel mode setting: 2489 kms, kbd format: kms,kbd 2490 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud] 2491 2492 kgdboc_earlycon= [KGDB,HW] 2493 If the boot console provides the ability to read 2494 characters and can work in polling mode, you can use 2495 this parameter to tell kgdb to use it as a backend 2496 until the normal console is registered. Intended to 2497 be used together with the kgdboc parameter which 2498 specifies the normal console to transition to. 2499 2500 The name of the early console should be specified 2501 as the value of this parameter. Note that the name of 2502 the early console might be different than the tty 2503 name passed to kgdboc. It's OK to leave the value 2504 blank and the first boot console that implements 2505 read() will be picked. 2506 2507 kgdbwait [KGDB] Stop kernel execution and enter the 2508 kernel debugger at the earliest opportunity. 2509 2510 kmac= [MIPS] Korina ethernet MAC address. 2511 Configure the RouterBoard 532 series on-chip 2512 Ethernet adapter MAC address. 2513 2514 kmemleak= [KNL] Boot-time kmemleak enable/disable 2515 Valid arguments: on, off 2516 Default: on 2517 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y, 2518 the default is off. 2519 2520 kprobe_event=[probe-list] 2521 [FTRACE] Add kprobe events and enable at boot time. 2522 The probe-list is a semicolon delimited list of probe 2523 definitions. Each definition is same as kprobe_events 2524 interface, but the parameters are comma delimited. 2525 For example, to add a kprobe event on vfs_read with 2526 arg1 and arg2, add to the command line; 2527 2528 kprobe_event=p,vfs_read,$arg1,$arg2 2529 2530 See also Documentation/trace/kprobetrace.rst "Kernel 2531 Boot Parameter" section. 2532 2533 kpti= [ARM64] Control page table isolation of user 2534 and kernel address spaces. 2535 Default: enabled on cores which need mitigation. 2536 0: force disabled 2537 1: force enabled 2538 2539 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires 2540 CONFIG_KUNIT to be set to be fully enabled. The 2541 default value can be overridden via 2542 KUNIT_DEFAULT_ENABLED. 2543 Default is 1 (enabled) 2544 2545 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs. 2546 Default is 0 (don't ignore, but inject #GP) 2547 2548 kvm.eager_page_split= 2549 [KVM,X86] Controls whether or not KVM will try to 2550 proactively split all huge pages during dirty logging. 2551 Eager page splitting reduces interruptions to vCPU 2552 execution by eliminating the write-protection faults 2553 and MMU lock contention that would otherwise be 2554 required to split huge pages lazily. 2555 2556 VM workloads that rarely perform writes or that write 2557 only to a small region of VM memory may benefit from 2558 disabling eager page splitting to allow huge pages to 2559 still be used for reads. 2560 2561 The behavior of eager page splitting depends on whether 2562 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If 2563 disabled, all huge pages in a memslot will be eagerly 2564 split when dirty logging is enabled on that memslot. If 2565 enabled, eager page splitting will be performed during 2566 the KVM_CLEAR_DIRTY ioctl, and only for the pages being 2567 cleared. 2568 2569 Eager page splitting is only supported when kvm.tdp_mmu=Y. 2570 2571 Default is Y (on). 2572 2573 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface. 2574 Default is false (don't support). 2575 2576 kvm.nx_huge_pages= 2577 [KVM] Controls the software workaround for the 2578 X86_BUG_ITLB_MULTIHIT bug. 2579 force : Always deploy workaround. 2580 off : Never deploy workaround. 2581 auto : Deploy workaround based on the presence of 2582 X86_BUG_ITLB_MULTIHIT. 2583 2584 Default is 'auto'. 2585 2586 If the software workaround is enabled for the host, 2587 guests do need not to enable it for nested guests. 2588 2589 kvm.nx_huge_pages_recovery_ratio= 2590 [KVM] Controls how many 4KiB pages are periodically zapped 2591 back to huge pages. 0 disables the recovery, otherwise if 2592 the value is N KVM will zap 1/Nth of the 4KiB pages every 2593 period (see below). The default is 60. 2594 2595 kvm.nx_huge_pages_recovery_period_ms= 2596 [KVM] Controls the time period at which KVM zaps 4KiB pages 2597 back to huge pages. If the value is a non-zero N, KVM will 2598 zap a portion (see ratio above) of the pages every N msecs. 2599 If the value is 0 (the default), KVM will pick a period based 2600 on the ratio, such that a page is zapped after 1 hour on average. 2601 2602 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in 2603 KVM/SVM. Default is 1 (enabled). 2604 2605 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables, 2606 a.k.a. Two-Dimensional Page Tables. Default is 1 2607 (enabled). Disable by KVM if hardware lacks support 2608 for NPT. 2609 2610 kvm-arm.mode= 2611 [KVM,ARM] Select one of KVM/arm64's modes of operation. 2612 2613 none: Forcefully disable KVM. 2614 2615 nvhe: Standard nVHE-based mode, without support for 2616 protected guests. 2617 2618 protected: nVHE-based mode with support for guests whose 2619 state is kept private from the host. 2620 2621 nested: VHE-based mode with support for nested 2622 virtualization. Requires at least ARMv8.3 2623 hardware. 2624 2625 Defaults to VHE/nVHE based on hardware support. Setting 2626 mode to "protected" will disable kexec and hibernation 2627 for the host. "nested" is experimental and should be 2628 used with extreme caution. 2629 2630 kvm-arm.vgic_v3_group0_trap= 2631 [KVM,ARM] Trap guest accesses to GICv3 group-0 2632 system registers 2633 2634 kvm-arm.vgic_v3_group1_trap= 2635 [KVM,ARM] Trap guest accesses to GICv3 group-1 2636 system registers 2637 2638 kvm-arm.vgic_v3_common_trap= 2639 [KVM,ARM] Trap guest accesses to GICv3 common 2640 system registers 2641 2642 kvm-arm.vgic_v4_enable= 2643 [KVM,ARM] Allow use of GICv4 for direct injection of 2644 LPIs. 2645 2646 kvm_cma_resv_ratio=n [PPC] 2647 Reserves given percentage from system memory area for 2648 contiguous memory allocation for KVM hash pagetable 2649 allocation. 2650 By default it reserves 5% of total system memory. 2651 Format: <integer> 2652 Default: 5 2653 2654 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables, 2655 a.k.a. Two-Dimensional Page Tables. Default is 1 2656 (enabled). Disable by KVM if hardware lacks support 2657 for EPT. 2658 2659 kvm-intel.emulate_invalid_guest_state= 2660 [KVM,Intel] Control whether to emulate invalid guest 2661 state. Ignored if kvm-intel.enable_unrestricted_guest=1, 2662 as guest state is never invalid for unrestricted 2663 guests. This param doesn't apply to nested guests (L2), 2664 as KVM never emulates invalid L2 guest state. 2665 Default is 1 (enabled). 2666 2667 kvm-intel.flexpriority= 2668 [KVM,Intel] Control KVM's use of FlexPriority feature 2669 (TPR shadow). Default is 1 (enabled). Disable by KVM if 2670 hardware lacks support for it. 2671 2672 kvm-intel.nested= 2673 [KVM,Intel] Control nested virtualization feature in 2674 KVM/VMX. Default is 1 (enabled). 2675 2676 kvm-intel.unrestricted_guest= 2677 [KVM,Intel] Control KVM's use of unrestricted guest 2678 feature (virtualized real and unpaged mode). Default 2679 is 1 (enabled). Disable by KVM if EPT is disabled or 2680 hardware lacks support for it. 2681 2682 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault 2683 CVE-2018-3620. 2684 2685 Valid arguments: never, cond, always 2686 2687 always: L1D cache flush on every VMENTER. 2688 cond: Flush L1D on VMENTER only when the code between 2689 VMEXIT and VMENTER can leak host memory. 2690 never: Disables the mitigation 2691 2692 Default is cond (do L1 cache flush in specific instances) 2693 2694 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor 2695 Identification feature (tagged TLBs). Default is 1 2696 (enabled). Disable by KVM if hardware lacks support 2697 for it. 2698 2699 l1d_flush= [X86,INTEL] 2700 Control mitigation for L1D based snooping vulnerability. 2701 2702 Certain CPUs are vulnerable to an exploit against CPU 2703 internal buffers which can forward information to a 2704 disclosure gadget under certain conditions. 2705 2706 In vulnerable processors, the speculatively 2707 forwarded data can be used in a cache side channel 2708 attack, to access data to which the attacker does 2709 not have direct access. 2710 2711 This parameter controls the mitigation. The 2712 options are: 2713 2714 on - enable the interface for the mitigation 2715 2716 l1tf= [X86] Control mitigation of the L1TF vulnerability on 2717 affected CPUs 2718 2719 The kernel PTE inversion protection is unconditionally 2720 enabled and cannot be disabled. 2721 2722 full 2723 Provides all available mitigations for the 2724 L1TF vulnerability. Disables SMT and 2725 enables all mitigations in the 2726 hypervisors, i.e. unconditional L1D flush. 2727 2728 SMT control and L1D flush control via the 2729 sysfs interface is still possible after 2730 boot. Hypervisors will issue a warning 2731 when the first VM is started in a 2732 potentially insecure configuration, 2733 i.e. SMT enabled or L1D flush disabled. 2734 2735 full,force 2736 Same as 'full', but disables SMT and L1D 2737 flush runtime control. Implies the 2738 'nosmt=force' command line option. 2739 (i.e. sysfs control of SMT is disabled.) 2740 2741 flush 2742 Leaves SMT enabled and enables the default 2743 hypervisor mitigation, i.e. conditional 2744 L1D flush. 2745 2746 SMT control and L1D flush control via the 2747 sysfs interface is still possible after 2748 boot. Hypervisors will issue a warning 2749 when the first VM is started in a 2750 potentially insecure configuration, 2751 i.e. SMT enabled or L1D flush disabled. 2752 2753 flush,nosmt 2754 2755 Disables SMT and enables the default 2756 hypervisor mitigation. 2757 2758 SMT control and L1D flush control via the 2759 sysfs interface is still possible after 2760 boot. Hypervisors will issue a warning 2761 when the first VM is started in a 2762 potentially insecure configuration, 2763 i.e. SMT enabled or L1D flush disabled. 2764 2765 flush,nowarn 2766 Same as 'flush', but hypervisors will not 2767 warn when a VM is started in a potentially 2768 insecure configuration. 2769 2770 off 2771 Disables hypervisor mitigations and doesn't 2772 emit any warnings. 2773 It also drops the swap size and available 2774 RAM limit restriction on both hypervisor and 2775 bare metal. 2776 2777 Default is 'flush'. 2778 2779 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst 2780 2781 l2cr= [PPC] 2782 2783 l3cr= [PPC] 2784 2785 lapic [X86-32,APIC] Enable the local APIC even if BIOS 2786 disabled it. 2787 2788 lapic= [X86,APIC] Do not use TSC deadline 2789 value for LAPIC timer one-shot implementation. Default 2790 back to the programmable timer unit in the LAPIC. 2791 Format: notscdeadline 2792 2793 lapic_timer_c2_ok [X86,APIC] trust the local apic timer 2794 in C2 power state. 2795 2796 libata.dma= [LIBATA] DMA control 2797 libata.dma=0 Disable all PATA and SATA DMA 2798 libata.dma=1 PATA and SATA Disk DMA only 2799 libata.dma=2 ATAPI (CDROM) DMA only 2800 libata.dma=4 Compact Flash DMA only 2801 Combinations also work, so libata.dma=3 enables DMA 2802 for disks and CDROMs, but not CFs. 2803 2804 libata.ignore_hpa= [LIBATA] Ignore HPA limit 2805 libata.ignore_hpa=0 keep BIOS limits (default) 2806 libata.ignore_hpa=1 ignore limits, using full disk 2807 2808 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume 2809 when set. 2810 Format: <int> 2811 2812 libata.force= [LIBATA] Force configurations. The format is a comma- 2813 separated list of "[ID:]VAL" where ID is PORT[.DEVICE]. 2814 PORT and DEVICE are decimal numbers matching port, link 2815 or device. Basically, it matches the ATA ID string 2816 printed on console by libata. If the whole ID part is 2817 omitted, the last PORT and DEVICE values are used. If 2818 ID hasn't been specified yet, the configuration applies 2819 to all ports, links and devices. 2820 2821 If only DEVICE is omitted, the parameter applies to 2822 the port and all links and devices behind it. DEVICE 2823 number of 0 either selects the first device or the 2824 first fan-out link behind PMP device. It does not 2825 select the host link. DEVICE number of 15 selects the 2826 host link and device attached to it. 2827 2828 The VAL specifies the configuration to force. As long 2829 as there is no ambiguity, shortcut notation is allowed. 2830 For example, both 1.5 and 1.5G would work for 1.5Gbps. 2831 The following configurations can be forced. 2832 2833 * Cable type: 40c, 80c, short40c, unk, ign or sata. 2834 Any ID with matching PORT is used. 2835 2836 * SATA link speed limit: 1.5Gbps or 3.0Gbps. 2837 2838 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7]. 2839 udma[/][16,25,33,44,66,100,133] notation is also 2840 allowed. 2841 2842 * nohrst, nosrst, norst: suppress hard, soft and both 2843 resets. 2844 2845 * rstonce: only attempt one reset during hot-unplug 2846 link recovery. 2847 2848 * [no]dbdelay: Enable or disable the extra 200ms delay 2849 before debouncing a link PHY and device presence 2850 detection. 2851 2852 * [no]ncq: Turn on or off NCQ. 2853 2854 * [no]ncqtrim: Enable or disable queued DSM TRIM. 2855 2856 * [no]ncqati: Enable or disable NCQ trim on ATI chipset. 2857 2858 * [no]trim: Enable or disable (unqueued) TRIM. 2859 2860 * trim_zero: Indicate that TRIM command zeroes data. 2861 2862 * max_trim_128m: Set 128M maximum trim size limit. 2863 2864 * [no]dma: Turn on or off DMA transfers. 2865 2866 * atapi_dmadir: Enable ATAPI DMADIR bridge support. 2867 2868 * atapi_mod16_dma: Enable the use of ATAPI DMA for 2869 commands that are not a multiple of 16 bytes. 2870 2871 * [no]dmalog: Enable or disable the use of the 2872 READ LOG DMA EXT command to access logs. 2873 2874 * [no]iddevlog: Enable or disable access to the 2875 identify device data log. 2876 2877 * [no]logdir: Enable or disable access to the general 2878 purpose log directory. 2879 2880 * max_sec_128: Set transfer size limit to 128 sectors. 2881 2882 * max_sec_1024: Set or clear transfer size limit to 2883 1024 sectors. 2884 2885 * max_sec_lba48: Set or clear transfer size limit to 2886 65535 sectors. 2887 2888 * [no]lpm: Enable or disable link power management. 2889 2890 * [no]setxfer: Indicate if transfer speed mode setting 2891 should be skipped. 2892 2893 * [no]fua: Disable or enable FUA (Force Unit Access) 2894 support for devices supporting this feature. 2895 2896 * dump_id: Dump IDENTIFY data. 2897 2898 * disable: Disable this device. 2899 2900 If there are multiple matching configurations changing 2901 the same attribute, the last one is used. 2902 2903 load_ramdisk= [RAM] [Deprecated] 2904 2905 lockd.nlm_grace_period=P [NFS] Assign grace period. 2906 Format: <integer> 2907 2908 lockd.nlm_tcpport=N [NFS] Assign TCP port. 2909 Format: <integer> 2910 2911 lockd.nlm_timeout=T [NFS] Assign timeout value. 2912 Format: <integer> 2913 2914 lockd.nlm_udpport=M [NFS] Assign UDP port. 2915 Format: <integer> 2916 2917 lockdown= [SECURITY] 2918 { integrity | confidentiality } 2919 Enable the kernel lockdown feature. If set to 2920 integrity, kernel features that allow userland to 2921 modify the running kernel are disabled. If set to 2922 confidentiality, kernel features that allow userland 2923 to extract confidential information from the kernel 2924 are also disabled. 2925 2926 locktorture.acq_writer_lim= [KNL] 2927 Set the time limit in jiffies for a lock 2928 acquisition. Acquisitions exceeding this limit 2929 will result in a splat once they do complete. 2930 2931 locktorture.bind_readers= [KNL] 2932 Specify the list of CPUs to which the readers are 2933 to be bound. 2934 2935 locktorture.bind_writers= [KNL] 2936 Specify the list of CPUs to which the writers are 2937 to be bound. 2938 2939 locktorture.call_rcu_chains= [KNL] 2940 Specify the number of self-propagating call_rcu() 2941 chains to set up. These are used to ensure that 2942 there is a high probability of an RCU grace period 2943 in progress at any given time. Defaults to 0, 2944 which disables these call_rcu() chains. 2945 2946 locktorture.long_hold= [KNL] 2947 Specify the duration in milliseconds for the 2948 occasional long-duration lock hold time. Defaults 2949 to 100 milliseconds. Select 0 to disable. 2950 2951 locktorture.nested_locks= [KNL] 2952 Specify the maximum lock nesting depth that 2953 locktorture is to exercise, up to a limit of 8 2954 (MAX_NESTED_LOCKS). Specify zero to disable. 2955 Note that this parameter is ineffective on types 2956 of locks that do not support nested acquisition. 2957 2958 locktorture.nreaders_stress= [KNL] 2959 Set the number of locking read-acquisition kthreads. 2960 Defaults to being automatically set based on the 2961 number of online CPUs. 2962 2963 locktorture.nwriters_stress= [KNL] 2964 Set the number of locking write-acquisition kthreads. 2965 2966 locktorture.onoff_holdoff= [KNL] 2967 Set time (s) after boot for CPU-hotplug testing. 2968 2969 locktorture.onoff_interval= [KNL] 2970 Set time (s) between CPU-hotplug operations, or 2971 zero to disable CPU-hotplug testing. 2972 2973 locktorture.rt_boost= [KNL] 2974 Do periodic testing of real-time lock priority 2975 boosting. Select 0 to disable, 1 to boost 2976 only rt_mutex, and 2 to boost unconditionally. 2977 Defaults to 2, which might seem to be an 2978 odd choice, but which should be harmless for 2979 non-real-time spinlocks, due to their disabling 2980 of preemption. Note that non-realtime mutexes 2981 disable boosting. 2982 2983 locktorture.rt_boost_factor= [KNL] 2984 Number that determines how often and for how 2985 long priority boosting is exercised. This is 2986 scaled down by the number of writers, so that the 2987 number of boosts per unit time remains roughly 2988 constant as the number of writers increases. 2989 On the other hand, the duration of each boost 2990 increases with the number of writers. 2991 2992 locktorture.shuffle_interval= [KNL] 2993 Set task-shuffle interval (jiffies). Shuffling 2994 tasks allows some CPUs to go into dyntick-idle 2995 mode during the locktorture test. 2996 2997 locktorture.shutdown_secs= [KNL] 2998 Set time (s) after boot system shutdown. This 2999 is useful for hands-off automated testing. 3000 3001 locktorture.stat_interval= [KNL] 3002 Time (s) between statistics printk()s. 3003 3004 locktorture.stutter= [KNL] 3005 Time (s) to stutter testing, for example, 3006 specifying five seconds causes the test to run for 3007 five seconds, wait for five seconds, and so on. 3008 This tests the locking primitive's ability to 3009 transition abruptly to and from idle. 3010 3011 locktorture.torture_type= [KNL] 3012 Specify the locking implementation to test. 3013 3014 locktorture.verbose= [KNL] 3015 Enable additional printk() statements. 3016 3017 locktorture.writer_fifo= [KNL] 3018 Run the write-side locktorture kthreads at 3019 sched_set_fifo() real-time priority. 3020 3021 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver 3022 Format: <irq> 3023 3024 loglevel= All Kernel Messages with a loglevel smaller than the 3025 console loglevel will be printed to the console. It can 3026 also be changed with klogd or other programs. The 3027 loglevels are defined as follows: 3028 3029 0 (KERN_EMERG) system is unusable 3030 1 (KERN_ALERT) action must be taken immediately 3031 2 (KERN_CRIT) critical conditions 3032 3 (KERN_ERR) error conditions 3033 4 (KERN_WARNING) warning conditions 3034 5 (KERN_NOTICE) normal but significant condition 3035 6 (KERN_INFO) informational 3036 7 (KERN_DEBUG) debug-level messages 3037 3038 log_buf_len=n[KMG] Sets the size of the printk ring buffer, 3039 in bytes. n must be a power of two and greater 3040 than the minimal size. The minimal size is defined 3041 by LOG_BUF_SHIFT kernel config parameter. There is 3042 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter 3043 that allows to increase the default size depending on 3044 the number of CPUs. See init/Kconfig for more details. 3045 3046 logo.nologo [FB] Disables display of the built-in Linux logo. 3047 This may be used to provide more screen space for 3048 kernel log messages and is useful when debugging 3049 kernel boot problems. 3050 3051 lp=0 [LP] Specify parallel ports to use, e.g, 3052 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses 3053 lp=reset first parallel port). 'lp=0' disables the 3054 lp=auto printer driver. 'lp=reset' (which can be 3055 specified in addition to the ports) causes 3056 attached printers to be reset. Using 3057 lp=port1,port2,... specifies the parallel ports 3058 to associate lp devices with, starting with 3059 lp0. A port specification may be 'none' to skip 3060 that lp device, or a parport name such as 3061 'parport0'. Specifying 'lp=auto' instead of a 3062 port specification list means that device IDs 3063 from each port should be examined, to see if 3064 an IEEE 1284-compliant printer is attached; if 3065 so, the driver will manage that printer. 3066 See also header of drivers/char/lp.c. 3067 3068 lpj=n [KNL] 3069 Sets loops_per_jiffy to given constant, thus avoiding 3070 time-consuming boot-time autodetection (up to 250 ms per 3071 CPU). 0 enables autodetection (default). To determine 3072 the correct value for your kernel, boot with normal 3073 autodetection and see what value is printed. Note that 3074 on SMP systems the preset will be applied to all CPUs, 3075 which is likely to cause problems if your CPUs need 3076 significantly divergent settings. An incorrect value 3077 will cause delays in the kernel to be wrong, leading to 3078 unpredictable I/O errors and other breakage. Although 3079 unlikely, in the extreme case this might damage your 3080 hardware. 3081 3082 ltpc= [NET] 3083 Format: <io>,<irq>,<dma> 3084 3085 lsm.debug [SECURITY] Enable LSM initialization debugging output. 3086 3087 lsm=lsm1,...,lsmN 3088 [SECURITY] Choose order of LSM initialization. This 3089 overrides CONFIG_LSM, and the "security=" parameter. 3090 3091 machvec= [IA-64] Force the use of a particular machine-vector 3092 (machvec) in a generic kernel. 3093 Example: machvec=hpzx1 3094 3095 machtype= [Loongson] Share the same kernel image file between 3096 different yeeloong laptops. 3097 Example: machtype=lemote-yeeloong-2f-7inch 3098 3099 max_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory greater 3100 than or equal to this physical address is ignored. 3101 3102 maxcpus= [SMP] Maximum number of processors that an SMP kernel 3103 will bring up during bootup. maxcpus=n : n >= 0 limits 3104 the kernel to bring up 'n' processors. Surely after 3105 bootup you can bring up the other plugged cpu by executing 3106 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus 3107 only takes effect during system bootup. 3108 While n=0 is a special case, it is equivalent to "nosmp", 3109 which also disables the IO APIC. 3110 3111 max_loop= [LOOP] The number of loop block devices that get 3112 (loop.max_loop) unconditionally pre-created at init time. The default 3113 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead 3114 of statically allocating a predefined number, loop 3115 devices can be requested on-demand with the 3116 /dev/loop-control interface. 3117 3118 mce [X86-32] Machine Check Exception 3119 3120 mce=option [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst 3121 3122 md= [HW] RAID subsystems devices and level 3123 See Documentation/admin-guide/md.rst. 3124 3125 mdacon= [MDA] 3126 Format: <first>,<last> 3127 Specifies range of consoles to be captured by the MDA. 3128 3129 mds= [X86,INTEL] 3130 Control mitigation for the Micro-architectural Data 3131 Sampling (MDS) vulnerability. 3132 3133 Certain CPUs are vulnerable to an exploit against CPU 3134 internal buffers which can forward information to a 3135 disclosure gadget under certain conditions. 3136 3137 In vulnerable processors, the speculatively 3138 forwarded data can be used in a cache side channel 3139 attack, to access data to which the attacker does 3140 not have direct access. 3141 3142 This parameter controls the MDS mitigation. The 3143 options are: 3144 3145 full - Enable MDS mitigation on vulnerable CPUs 3146 full,nosmt - Enable MDS mitigation and disable 3147 SMT on vulnerable CPUs 3148 off - Unconditionally disable MDS mitigation 3149 3150 On TAA-affected machines, mds=off can be prevented by 3151 an active TAA mitigation as both vulnerabilities are 3152 mitigated with the same mechanism so in order to disable 3153 this mitigation, you need to specify tsx_async_abort=off 3154 too. 3155 3156 Not specifying this option is equivalent to 3157 mds=full. 3158 3159 For details see: Documentation/admin-guide/hw-vuln/mds.rst 3160 3161 mem=nn[KMG] [HEXAGON] Set the memory size. 3162 Must be specified, otherwise memory size will be 0. 3163 3164 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory 3165 Amount of memory to be used in cases as follows: 3166 3167 1 for test; 3168 2 when the kernel is not able to see the whole system memory; 3169 3 memory that lies after 'mem=' boundary is excluded from 3170 the hypervisor, then assigned to KVM guests. 3171 4 to limit the memory available for kdump kernel. 3172 3173 [ARC,MICROBLAZE] - the limit applies only to low memory, 3174 high memory is not affected. 3175 3176 [ARM64] - only limits memory covered by the linear 3177 mapping. The NOMAP regions are not affected. 3178 3179 [X86] Work as limiting max address. Use together 3180 with memmap= to avoid physical address space collisions. 3181 Without memmap= PCI devices could be placed at addresses 3182 belonging to unused RAM. 3183 3184 Note that this only takes effects during boot time since 3185 in above case 3, memory may need be hot added after boot 3186 if system memory of hypervisor is not sufficient. 3187 3188 mem=nn[KMG]@ss[KMG] 3189 [ARM,MIPS] - override the memory layout reported by 3190 firmware. 3191 Define a memory region of size nn[KMG] starting at 3192 ss[KMG]. 3193 Multiple different regions can be specified with 3194 multiple mem= parameters on the command line. 3195 3196 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel 3197 memory. 3198 3199 memblock=debug [KNL] Enable memblock debug messages. 3200 3201 memchunk=nn[KMG] 3202 [KNL,SH] Allow user to override the default size for 3203 per-device physically contiguous DMA buffers. 3204 3205 memhp_default_state=online/offline/online_kernel/online_movable 3206 [KNL] Set the initial state for the memory hotplug 3207 onlining policy. If not specified, the default value is 3208 set according to the 3209 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config 3210 option. 3211 See Documentation/admin-guide/mm/memory-hotplug.rst. 3212 3213 memmap=exactmap [KNL,X86] Enable setting of an exact 3214 E820 memory map, as specified by the user. 3215 Such memmap=exactmap lines can be constructed based on 3216 BIOS output or other requirements. See the memmap=nn@ss 3217 option description. 3218 3219 memmap=nn[KMG]@ss[KMG] 3220 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory. 3221 Region of memory to be used is from ss to ss+nn. 3222 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG], 3223 which limits max address to nn[KMG]. 3224 Multiple different regions can be specified, 3225 comma delimited. 3226 Example: 3227 memmap=100M@2G,100M#3G,1G!1024G 3228 3229 memmap=nn[KMG]#ss[KMG] 3230 [KNL,ACPI] Mark specific memory as ACPI data. 3231 Region of memory to be marked is from ss to ss+nn. 3232 3233 memmap=nn[KMG]$ss[KMG] 3234 [KNL,ACPI] Mark specific memory as reserved. 3235 Region of memory to be reserved is from ss to ss+nn. 3236 Example: Exclude memory from 0x18690000-0x1869ffff 3237 memmap=64K$0x18690000 3238 or 3239 memmap=0x10000$0x18690000 3240 Some bootloaders may need an escape character before '$', 3241 like Grub2, otherwise '$' and the following number 3242 will be eaten. 3243 3244 memmap=nn[KMG]!ss[KMG] 3245 [KNL,X86] Mark specific memory as protected. 3246 Region of memory to be used, from ss to ss+nn. 3247 The memory region may be marked as e820 type 12 (0xc) 3248 and is NVDIMM or ADR memory. 3249 3250 memmap=<size>%<offset>-<oldtype>+<newtype> 3251 [KNL,ACPI] Convert memory within the specified region 3252 from <oldtype> to <newtype>. If "-<oldtype>" is left 3253 out, the whole region will be marked as <newtype>, 3254 even if previously unavailable. If "+<newtype>" is left 3255 out, matching memory will be removed. Types are 3256 specified as e820 types, e.g., 1 = RAM, 2 = reserved, 3257 3 = ACPI, 12 = PRAM. 3258 3259 memory_corruption_check=0/1 [X86] 3260 Some BIOSes seem to corrupt the first 64k of 3261 memory when doing things like suspend/resume. 3262 Setting this option will scan the memory 3263 looking for corruption. Enabling this will 3264 both detect corruption and prevent the kernel 3265 from using the memory being corrupted. 3266 However, its intended as a diagnostic tool; if 3267 repeatable BIOS-originated corruption always 3268 affects the same memory, you can use memmap= 3269 to prevent the kernel from using that memory. 3270 3271 memory_corruption_check_size=size [X86] 3272 By default it checks for corruption in the low 3273 64k, making this memory unavailable for normal 3274 use. Use this parameter to scan for 3275 corruption in more or less memory. 3276 3277 memory_corruption_check_period=seconds [X86] 3278 By default it checks for corruption every 60 3279 seconds. Use this parameter to check at some 3280 other rate. 0 disables periodic checking. 3281 3282 memory_hotplug.memmap_on_memory 3283 [KNL,X86,ARM] Boolean flag to enable this feature. 3284 Format: {on | off (default)} 3285 When enabled, runtime hotplugged memory will 3286 allocate its internal metadata (struct pages, 3287 those vmemmap pages cannot be optimized even 3288 if hugetlb_free_vmemmap is enabled) from the 3289 hotadded memory which will allow to hotadd a 3290 lot of memory without requiring additional 3291 memory to do so. 3292 This feature is disabled by default because it 3293 has some implication on large (e.g. GB) 3294 allocations in some configurations (e.g. small 3295 memory blocks). 3296 The state of the flag can be read in 3297 /sys/module/memory_hotplug/parameters/memmap_on_memory. 3298 Note that even when enabled, there are a few cases where 3299 the feature is not effective. 3300 3301 memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest 3302 Format: <integer> 3303 default : 0 <disable> 3304 Specifies the number of memtest passes to be 3305 performed. Each pass selects another test 3306 pattern from a given set of patterns. Memtest 3307 fills the memory with this pattern, validates 3308 memory contents and reserves bad memory 3309 regions that are detected. 3310 3311 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control 3312 Valid arguments: on, off 3313 Default (depends on kernel configuration option): 3314 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y) 3315 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n) 3316 mem_encrypt=on: Activate SME 3317 mem_encrypt=off: Do not activate SME 3318 3319 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst 3320 for details on when memory encryption can be activated. 3321 3322 mem_sleep_default= [SUSPEND] Default system suspend mode: 3323 s2idle - Suspend-To-Idle 3324 shallow - Power-On Suspend or equivalent (if supported) 3325 deep - Suspend-To-RAM or equivalent (if supported) 3326 See Documentation/admin-guide/pm/sleep-states.rst. 3327 3328 mfgpt_irq= [IA-32] Specify the IRQ to use for the 3329 Multi-Function General Purpose Timers on AMD Geode 3330 platforms. 3331 3332 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when 3333 the BIOS has incorrectly applied a workaround. TinyBIOS 3334 version 0.98 is known to be affected, 0.99 fixes the 3335 problem by letting the user disable the workaround. 3336 3337 mga= [HW,DRM] 3338 3339 min_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory below this 3340 physical address is ignored. 3341 3342 mini2440= [ARM,HW,KNL] 3343 Format:[0..2][b][c][t] 3344 Default: "0tb" 3345 MINI2440 configuration specification: 3346 0 - The attached screen is the 3.5" TFT 3347 1 - The attached screen is the 7" TFT 3348 2 - The VGA Shield is attached (1024x768) 3349 Leaving out the screen size parameter will not load 3350 the TFT driver, and the framebuffer will be left 3351 unconfigured. 3352 b - Enable backlight. The TFT backlight pin will be 3353 linked to the kernel VESA blanking code and a GPIO 3354 LED. This parameter is not necessary when using the 3355 VGA shield. 3356 c - Enable the s3c camera interface. 3357 t - Reserved for enabling touchscreen support. The 3358 touchscreen support is not enabled in the mainstream 3359 kernel as of 2.6.30, a preliminary port can be found 3360 in the "bleeding edge" mini2440 support kernel at 3361 https://repo.or.cz/w/linux-2.6/mini2440.git 3362 3363 mitigations= 3364 [X86,PPC,S390,ARM64] Control optional mitigations for 3365 CPU vulnerabilities. This is a set of curated, 3366 arch-independent options, each of which is an 3367 aggregation of existing arch-specific options. 3368 3369 off 3370 Disable all optional CPU mitigations. This 3371 improves system performance, but it may also 3372 expose users to several CPU vulnerabilities. 3373 Equivalent to: if nokaslr then kpti=0 [ARM64] 3374 gather_data_sampling=off [X86] 3375 kvm.nx_huge_pages=off [X86] 3376 l1tf=off [X86] 3377 mds=off [X86] 3378 mmio_stale_data=off [X86] 3379 no_entry_flush [PPC] 3380 no_uaccess_flush [PPC] 3381 nobp=0 [S390] 3382 nopti [X86,PPC] 3383 nospectre_bhb [ARM64] 3384 nospectre_v1 [X86,PPC] 3385 nospectre_v2 [X86,PPC,S390,ARM64] 3386 retbleed=off [X86] 3387 spec_store_bypass_disable=off [X86,PPC] 3388 spectre_v2_user=off [X86] 3389 srbds=off [X86,INTEL] 3390 ssbd=force-off [ARM64] 3391 tsx_async_abort=off [X86] 3392 3393 Exceptions: 3394 This does not have any effect on 3395 kvm.nx_huge_pages when 3396 kvm.nx_huge_pages=force. 3397 3398 auto (default) 3399 Mitigate all CPU vulnerabilities, but leave SMT 3400 enabled, even if it's vulnerable. This is for 3401 users who don't want to be surprised by SMT 3402 getting disabled across kernel upgrades, or who 3403 have other ways of avoiding SMT-based attacks. 3404 Equivalent to: (default behavior) 3405 3406 auto,nosmt 3407 Mitigate all CPU vulnerabilities, disabling SMT 3408 if needed. This is for users who always want to 3409 be fully mitigated, even if it means losing SMT. 3410 Equivalent to: l1tf=flush,nosmt [X86] 3411 mds=full,nosmt [X86] 3412 tsx_async_abort=full,nosmt [X86] 3413 mmio_stale_data=full,nosmt [X86] 3414 retbleed=auto,nosmt [X86] 3415 3416 mminit_loglevel= 3417 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this 3418 parameter allows control of the logging verbosity for 3419 the additional memory initialisation checks. A value 3420 of 0 disables mminit logging and a level of 4 will 3421 log everything. Information is printed at KERN_DEBUG 3422 so loglevel=8 may also need to be specified. 3423 3424 mmio_stale_data= 3425 [X86,INTEL] Control mitigation for the Processor 3426 MMIO Stale Data vulnerabilities. 3427 3428 Processor MMIO Stale Data is a class of 3429 vulnerabilities that may expose data after an MMIO 3430 operation. Exposed data could originate or end in 3431 the same CPU buffers as affected by MDS and TAA. 3432 Therefore, similar to MDS and TAA, the mitigation 3433 is to clear the affected CPU buffers. 3434 3435 This parameter controls the mitigation. The 3436 options are: 3437 3438 full - Enable mitigation on vulnerable CPUs 3439 3440 full,nosmt - Enable mitigation and disable SMT on 3441 vulnerable CPUs. 3442 3443 off - Unconditionally disable mitigation 3444 3445 On MDS or TAA affected machines, 3446 mmio_stale_data=off can be prevented by an active 3447 MDS or TAA mitigation as these vulnerabilities are 3448 mitigated with the same mechanism so in order to 3449 disable this mitigation, you need to specify 3450 mds=off and tsx_async_abort=off too. 3451 3452 Not specifying this option is equivalent to 3453 mmio_stale_data=full. 3454 3455 For details see: 3456 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst 3457 3458 <module>.async_probe[=<bool>] [KNL] 3459 If no <bool> value is specified or if the value 3460 specified is not a valid <bool>, enable asynchronous 3461 probe on this module. Otherwise, enable/disable 3462 asynchronous probe on this module as indicated by the 3463 <bool> value. See also: module.async_probe 3464 3465 module.async_probe=<bool> 3466 [KNL] When set to true, modules will use async probing 3467 by default. To enable/disable async probing for a 3468 specific module, use the module specific control that 3469 is documented under <module>.async_probe. When both 3470 module.async_probe and <module>.async_probe are 3471 specified, <module>.async_probe takes precedence for 3472 the specific module. 3473 3474 module.enable_dups_trace 3475 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set, 3476 this means that duplicate request_module() calls will 3477 trigger a WARN_ON() instead of a pr_warn(). Note that 3478 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s 3479 will always be issued and this option does nothing. 3480 module.sig_enforce 3481 [KNL] When CONFIG_MODULE_SIG is set, this means that 3482 modules without (valid) signatures will fail to load. 3483 Note that if CONFIG_MODULE_SIG_FORCE is set, that 3484 is always true, so this option does nothing. 3485 3486 module_blacklist= [KNL] Do not load a comma-separated list of 3487 modules. Useful for debugging problem modules. 3488 3489 mousedev.tap_time= 3490 [MOUSE] Maximum time between finger touching and 3491 leaving touchpad surface for touch to be considered 3492 a tap and be reported as a left button click (for 3493 touchpads working in absolute mode only). 3494 Format: <msecs> 3495 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices 3496 reporting absolute coordinates, such as tablets 3497 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices 3498 reporting absolute coordinates, such as tablets 3499 3500 movablecore= [KNL,X86,IA-64,PPC] 3501 Format: nn[KMGTPE] | nn% 3502 This parameter is the complement to kernelcore=, it 3503 specifies the amount of memory used for migratable 3504 allocations. If both kernelcore and movablecore is 3505 specified, then kernelcore will be at *least* the 3506 specified value but may be more. If movablecore on its 3507 own is specified, the administrator must be careful 3508 that the amount of memory usable for all allocations 3509 is not too small. 3510 3511 movable_node [KNL] Boot-time switch to make hotplugable memory 3512 NUMA nodes to be movable. This means that the memory 3513 of such nodes will be usable only for movable 3514 allocations which rules out almost all kernel 3515 allocations. Use with caution! 3516 3517 MTD_Partition= [MTD] 3518 Format: <name>,<region-number>,<size>,<offset> 3519 3520 MTD_Region= [MTD] Format: 3521 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>] 3522 3523 mtdparts= [MTD] 3524 See drivers/mtd/parsers/cmdlinepart.c 3525 3526 mtdset= [ARM] 3527 ARM/S3C2412 JIVE boot control 3528 3529 See arch/arm/mach-s3c/mach-jive.c 3530 3531 mtouchusb.raw_coordinates= 3532 [HW] Make the MicroTouch USB driver use raw coordinates 3533 ('y', default) or cooked coordinates ('n') 3534 3535 mtrr=debug [X86] 3536 Enable printing debug information related to MTRR 3537 registers at boot time. 3538 3539 mtrr_chunk_size=nn[KMG] [X86] 3540 used for mtrr cleanup. It is largest continuous chunk 3541 that could hold holes aka. UC entries. 3542 3543 mtrr_gran_size=nn[KMG] [X86] 3544 Used for mtrr cleanup. It is granularity of mtrr block. 3545 Default is 1. 3546 Large value could prevent small alignment from 3547 using up MTRRs. 3548 3549 mtrr_spare_reg_nr=n [X86] 3550 Format: <integer> 3551 Range: 0,7 : spare reg number 3552 Default : 1 3553 Used for mtrr cleanup. It is spare mtrr entries number. 3554 Set to 2 or more if your graphical card needs more. 3555 3556 multitce=off [PPC] This parameter disables the use of the pSeries 3557 firmware feature for updating multiple TCE entries 3558 at a time. 3559 3560 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card 3561 3562 netdev= [NET] Network devices parameters 3563 Format: <irq>,<io>,<mem_start>,<mem_end>,<name> 3564 Note that mem_start is often overloaded to mean 3565 something different and driver-specific. 3566 This usage is only documented in each driver source 3567 file if at all. 3568 3569 netpoll.carrier_timeout= 3570 [NET] Specifies amount of time (in seconds) that 3571 netpoll should wait for a carrier. By default netpoll 3572 waits 4 seconds. 3573 3574 nf_conntrack.acct= 3575 [NETFILTER] Enable connection tracking flow accounting 3576 0 to disable accounting 3577 1 to enable accounting 3578 Default value is 0. 3579 3580 nfs.cache_getent= 3581 [NFS] sets the pathname to the program which is used 3582 to update the NFS client cache entries. 3583 3584 nfs.cache_getent_timeout= 3585 [NFS] sets the timeout after which an attempt to 3586 update a cache entry is deemed to have failed. 3587 3588 nfs.callback_nr_threads= 3589 [NFSv4] set the total number of threads that the 3590 NFS client will assign to service NFSv4 callback 3591 requests. 3592 3593 nfs.callback_tcpport= 3594 [NFS] set the TCP port on which the NFSv4 callback 3595 channel should listen. 3596 3597 nfs.enable_ino64= 3598 [NFS] enable 64-bit inode numbers. 3599 If zero, the NFS client will fake up a 32-bit inode 3600 number for the readdir() and stat() syscalls instead 3601 of returning the full 64-bit number. 3602 The default is to return 64-bit inode numbers. 3603 3604 nfs.idmap_cache_timeout= 3605 [NFS] set the maximum lifetime for idmapper cache 3606 entries. 3607 3608 nfs.max_session_cb_slots= 3609 [NFSv4.1] Sets the maximum number of session 3610 slots the client will assign to the callback 3611 channel. This determines the maximum number of 3612 callbacks the client will process in parallel for 3613 a particular server. 3614 3615 nfs.max_session_slots= 3616 [NFSv4.1] Sets the maximum number of session slots 3617 the client will attempt to negotiate with the server. 3618 This limits the number of simultaneous RPC requests 3619 that the client can send to the NFSv4.1 server. 3620 Note that there is little point in setting this 3621 value higher than the max_tcp_slot_table_limit. 3622 3623 nfs.nfs4_disable_idmapping= 3624 [NFSv4] When set to the default of '1', this option 3625 ensures that both the RPC level authentication 3626 scheme and the NFS level operations agree to use 3627 numeric uids/gids if the mount is using the 3628 'sec=sys' security flavour. In effect it is 3629 disabling idmapping, which can make migration from 3630 legacy NFSv2/v3 systems to NFSv4 easier. 3631 Servers that do not support this mode of operation 3632 will be autodetected by the client, and it will fall 3633 back to using the idmapper. 3634 To turn off this behaviour, set the value to '0'. 3635 3636 nfs.nfs4_unique_id= 3637 [NFS4] Specify an additional fixed unique ident- 3638 ification string that NFSv4 clients can insert into 3639 their nfs_client_id4 string. This is typically a 3640 UUID that is generated at system install time. 3641 3642 nfs.recover_lost_locks= 3643 [NFSv4] Attempt to recover locks that were lost due 3644 to a lease timeout on the server. Please note that 3645 doing this risks data corruption, since there are 3646 no guarantees that the file will remain unchanged 3647 after the locks are lost. 3648 If you want to enable the kernel legacy behaviour of 3649 attempting to recover these locks, then set this 3650 parameter to '1'. 3651 The default parameter value of '0' causes the kernel 3652 not to attempt recovery of lost locks. 3653 3654 nfs.send_implementation_id= 3655 [NFSv4.1] Send client implementation identification 3656 information in exchange_id requests. 3657 If zero, no implementation identification information 3658 will be sent. 3659 The default is to send the implementation identification 3660 information. 3661 3662 nfs4.layoutstats_timer= 3663 [NFSv4.2] Change the rate at which the kernel sends 3664 layoutstats to the pNFS metadata server. 3665 3666 Setting this to value to 0 causes the kernel to use 3667 whatever value is the default set by the layout 3668 driver. A non-zero value sets the minimum interval 3669 in seconds between layoutstats transmissions. 3670 3671 nfsd.inter_copy_offload_enable= 3672 [NFSv4.2] When set to 1, the server will support 3673 server-to-server copies for which this server is 3674 the destination of the copy. 3675 3676 nfsd.nfs4_disable_idmapping= 3677 [NFSv4] When set to the default of '1', the NFSv4 3678 server will return only numeric uids and gids to 3679 clients using auth_sys, and will accept numeric uids 3680 and gids from such clients. This is intended to ease 3681 migration from NFSv2/v3. 3682 3683 nfsd.nfsd4_ssc_umount_timeout= 3684 [NFSv4.2] When used as the destination of a 3685 server-to-server copy, knfsd temporarily mounts 3686 the source server. It caches the mount in case 3687 it will be needed again, and discards it if not 3688 used for the number of milliseconds specified by 3689 this parameter. 3690 3691 nfsaddrs= [NFS] Deprecated. Use ip= instead. 3692 See Documentation/admin-guide/nfs/nfsroot.rst. 3693 3694 nfsroot= [NFS] nfs root filesystem for disk-less boxes. 3695 See Documentation/admin-guide/nfs/nfsroot.rst. 3696 3697 nfsrootdebug [NFS] enable nfsroot debugging messages. 3698 See Documentation/admin-guide/nfs/nfsroot.rst. 3699 3700 nmi_backtrace.backtrace_idle [KNL] 3701 Dump stacks even of idle CPUs in response to an 3702 NMI stack-backtrace request. 3703 3704 nmi_debug= [KNL,SH] Specify one or more actions to take 3705 when a NMI is triggered. 3706 Format: [state][,regs][,debounce][,die] 3707 3708 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels 3709 Format: [panic,][nopanic,][num] 3710 Valid num: 0 or 1 3711 0 - turn hardlockup detector in nmi_watchdog off 3712 1 - turn hardlockup detector in nmi_watchdog on 3713 When panic is specified, panic when an NMI watchdog 3714 timeout occurs (or 'nopanic' to not panic on an NMI 3715 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set) 3716 To disable both hard and soft lockup detectors, 3717 please see 'nowatchdog'. 3718 This is useful when you use a panic=... timeout and 3719 need the box quickly up again. 3720 3721 These settings can be accessed at runtime via 3722 the nmi_watchdog and hardlockup_panic sysctls. 3723 3724 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths 3725 emulation library even if a 387 maths coprocessor 3726 is present. 3727 3728 no4lvl [RISCV] Disable 4-level and 5-level paging modes. Forces 3729 kernel to use 3-level paging instead. 3730 3731 no5lvl [X86-64,RISCV] Disable 5-level paging mode. Forces 3732 kernel to use 4-level paging instead. 3733 3734 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien 3735 caches in the slab allocator. Saves per-node memory, 3736 but will impact performance. 3737 3738 noalign [KNL,ARM] 3739 3740 noaltinstr [S390] Disables alternative instructions patching 3741 (CPU alternatives feature). 3742 3743 noapic [SMP,APIC] Tells the kernel to not make use of any 3744 IOAPICs that may be present in the system. 3745 3746 noautogroup Disable scheduler automatic task group creation. 3747 3748 nocache [ARM] 3749 3750 no_console_suspend 3751 [HW] Never suspend the console 3752 Disable suspending of consoles during suspend and 3753 hibernate operations. Once disabled, debugging 3754 messages can reach various consoles while the rest 3755 of the system is being put to sleep (ie, while 3756 debugging driver suspend/resume hooks). This may 3757 not work reliably with all consoles, but is known 3758 to work with serial and VGA consoles. 3759 To facilitate more flexible debugging, we also add 3760 console_suspend, a printk module parameter to control 3761 it. Users could use console_suspend (usually 3762 /sys/module/printk/parameters/console_suspend) to 3763 turn on/off it dynamically. 3764 3765 no_debug_objects 3766 [KNL] Disable object debugging 3767 3768 nodsp [SH] Disable hardware DSP at boot time. 3769 3770 noefi Disable EFI runtime services support. 3771 3772 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel. 3773 3774 noexec [IA-64] 3775 3776 noexec32 [X86-64] 3777 This affects only 32-bit executables. 3778 noexec32=on: enable non-executable mappings (default) 3779 read doesn't imply executable mappings 3780 noexec32=off: disable non-executable mappings 3781 read implies executable mappings 3782 3783 no_file_caps Tells the kernel not to honor file capabilities. The 3784 only way then for a file to be executed with privilege 3785 is to be setuid root or executed by root. 3786 3787 nofpu [MIPS,SH] Disable hardware FPU at boot time. 3788 3789 nofsgsbase [X86] Disables FSGSBASE instructions. 3790 3791 nofxsr [BUGS=X86-32] Disables x86 floating point extended 3792 register save and restore. The kernel will only save 3793 legacy floating-point registers on task switch. 3794 3795 nohalt [IA-64] Tells the kernel not to use the power saving 3796 function PAL_HALT_LIGHT when idle. This increases 3797 power-consumption. On the positive side, it reduces 3798 interrupt wake-up latency, which may improve performance 3799 in certain environments such as networked servers or 3800 real-time systems. 3801 3802 no_hash_pointers 3803 Force pointers printed to the console or buffers to be 3804 unhashed. By default, when a pointer is printed via %p 3805 format string, that pointer is "hashed", i.e. obscured 3806 by hashing the pointer value. This is a security feature 3807 that hides actual kernel addresses from unprivileged 3808 users, but it also makes debugging the kernel more 3809 difficult since unequal pointers can no longer be 3810 compared. However, if this command-line option is 3811 specified, then all normal pointers will have their true 3812 value printed. This option should only be specified when 3813 debugging the kernel. Please do not use on production 3814 kernels. 3815 3816 nohibernate [HIBERNATION] Disable hibernation and resume. 3817 3818 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,SH] Forces the kernel to 3819 busy wait in do_idle() and not use the arch_cpu_idle() 3820 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP 3821 to be effective. This is useful on platforms where the 3822 sleep(SH) or wfi(ARM,ARM64) instructions do not work 3823 correctly or when doing power measurements to evaluate 3824 the impact of the sleep instructions. This is also 3825 useful when using JTAG debugger. 3826 3827 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings. 3828 3829 nohugevmalloc [KNL,X86,PPC,ARM64] Disable kernel huge vmalloc mappings. 3830 3831 nohz= [KNL] Boottime enable/disable dynamic ticks 3832 Valid arguments: on, off 3833 Default: on 3834 3835 nohz_full= [KNL,BOOT,SMP,ISOL] 3836 The argument is a cpu list, as described above. 3837 In kernels built with CONFIG_NO_HZ_FULL=y, set 3838 the specified list of CPUs whose tick will be stopped 3839 whenever possible. The boot CPU will be forced outside 3840 the range to maintain the timekeeping. Any CPUs 3841 in this list will have their RCU callbacks offloaded, 3842 just as if they had also been called out in the 3843 rcu_nocbs= boot parameter. 3844 3845 Note that this argument takes precedence over 3846 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 3847 3848 noinitrd [RAM] Tells the kernel not to load any configured 3849 initial RAM disk. 3850 3851 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt 3852 remapping. 3853 [Deprecated - use intremap=off] 3854 3855 nointroute [IA-64] 3856 3857 noinvpcid [X86] Disable the INVPCID cpu feature. 3858 3859 noiotrap [SH] Disables trapped I/O port accesses. 3860 3861 noirqdebug [X86-32] Disables the code which attempts to detect and 3862 disable unhandled interrupt sources. 3863 3864 noisapnp [ISAPNP] Disables ISA PnP code. 3865 3866 nojitter [IA-64] Disables jitter checking for ITC timers. 3867 3868 nokaslr [KNL] 3869 When CONFIG_RANDOMIZE_BASE is set, this disables 3870 kernel and module base offset ASLR (Address Space 3871 Layout Randomization). 3872 3873 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page 3874 fault handling. 3875 3876 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver 3877 3878 nolapic [X86-32,APIC] Do not enable or use the local APIC. 3879 3880 nolapic_timer [X86-32,APIC] Do not use the local APIC timer. 3881 3882 nomca [IA-64] Disable machine check abort handling 3883 3884 nomce [X86-32] Disable Machine Check Exception 3885 3886 nomfgpt [X86-32] Disable Multi-Function General Purpose 3887 Timer usage (for AMD Geode machines). 3888 3889 nomodeset Disable kernel modesetting. Most systems' firmware 3890 sets up a display mode and provides framebuffer memory 3891 for output. With nomodeset, DRM and fbdev drivers will 3892 not load if they could possibly displace the pre- 3893 initialized output. Only the system framebuffer will 3894 be available for use. The respective drivers will not 3895 perform display-mode changes or accelerated rendering. 3896 3897 Useful as error fallback, or for testing and debugging. 3898 3899 nomodule Disable module load 3900 3901 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to 3902 shutdown the other cpus. Instead use the REBOOT_VECTOR 3903 irq. 3904 3905 nopat [X86] Disable PAT (page attribute table extension of 3906 pagetables) support. 3907 3908 nopcid [X86-64] Disable the PCID cpu feature. 3909 3910 nopku [X86] Disable Memory Protection Keys CPU feature found 3911 in some Intel CPUs. 3912 3913 nopti [X86-64] 3914 Equivalent to pti=off 3915 3916 nopv= [X86,XEN,KVM,HYPER_V,VMWARE] 3917 Disables the PV optimizations forcing the guest to run 3918 as generic guest with no PV drivers. Currently support 3919 XEN HVM, KVM, HYPER_V and VMWARE guest. 3920 3921 nopvspin [X86,XEN,KVM] 3922 Disables the qspinlock slow path using PV optimizations 3923 which allow the hypervisor to 'idle' the guest on lock 3924 contention. 3925 3926 norandmaps Don't use address space randomization. Equivalent to 3927 echo 0 > /proc/sys/kernel/randomize_va_space 3928 3929 noreplace-smp [X86-32,SMP] Don't replace SMP instructions 3930 with UP alternatives 3931 3932 noresume [SWSUSP] Disables resume and restores original swap 3933 space. 3934 3935 nosbagart [IA-64] 3936 3937 no-scroll [VGA] Disables scrollback. 3938 This is required for the Braillex ib80-piezo Braille 3939 reader made by F.H. Papenmeier (Germany). 3940 3941 nosgx [X86-64,SGX] Disables Intel SGX kernel support. 3942 3943 nosmap [PPC] 3944 Disable SMAP (Supervisor Mode Access Prevention) 3945 even if it is supported by processor. 3946 3947 nosmep [PPC64s] 3948 Disable SMEP (Supervisor Mode Execution Prevention) 3949 even if it is supported by processor. 3950 3951 nosmp [SMP] Tells an SMP kernel to act as a UP kernel, 3952 and disable the IO APIC. legacy for "maxcpus=0". 3953 3954 nosmt [KNL,MIPS,PPC,S390] Disable symmetric multithreading (SMT). 3955 Equivalent to smt=1. 3956 3957 [KNL,X86,PPC] Disable symmetric multithreading (SMT). 3958 nosmt=force: Force disable SMT, cannot be undone 3959 via the sysfs control file. 3960 3961 nosoftlockup [KNL] Disable the soft-lockup detector. 3962 3963 nospec_store_bypass_disable 3964 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability 3965 3966 nospectre_bhb [ARM64] Disable all mitigations for Spectre-BHB (branch 3967 history injection) vulnerability. System may allow data leaks 3968 with this option. 3969 3970 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1 3971 (bounds check bypass). With this option data leaks are 3972 possible in the system. 3973 3974 nospectre_v2 [X86,PPC_E500,ARM64] Disable all mitigations for 3975 the Spectre variant 2 (indirect branch prediction) 3976 vulnerability. System may allow data leaks with this 3977 option. 3978 3979 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES] Disable paravirtualized 3980 steal time accounting. steal time is computed, but 3981 won't influence scheduler behaviour 3982 3983 nosync [HW,M68K] Disables sync negotiation for all devices. 3984 3985 no_timer_check [X86,APIC] Disables the code which tests for 3986 broken timer IRQ sources. 3987 3988 no_uaccess_flush 3989 [PPC] Don't flush the L1-D cache after accessing user data. 3990 3991 novmcoredd [KNL,KDUMP] 3992 Disable device dump. Device dump allows drivers to 3993 append dump data to vmcore so you can collect driver 3994 specified debug info. Drivers can append the data 3995 without any limit and this data is stored in memory, 3996 so this may cause significant memory stress. Disabling 3997 device dump can help save memory but the driver debug 3998 data will be no longer available. This parameter 3999 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP 4000 is set. 4001 4002 no-vmw-sched-clock 4003 [X86,PV_OPS] Disable paravirtualized VMware scheduler 4004 clock and use the default one. 4005 4006 nowatchdog [KNL] Disable both lockup detectors, i.e. 4007 soft-lockup and NMI watchdog (hard-lockup). 4008 4009 nowb [ARM] 4010 4011 nox2apic [X86-64,APIC] Do not enable x2APIC mode. 4012 4013 NOTE: this parameter will be ignored on systems with the 4014 LEGACY_XAPIC_DISABLED bit set in the 4015 IA32_XAPIC_DISABLE_STATUS MSR. 4016 4017 noxsave [BUGS=X86] Disables x86 extended register state save 4018 and restore using xsave. The kernel will fallback to 4019 enabling legacy floating-point and sse state. 4020 4021 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended 4022 register states. The kernel will fall back to use 4023 xsave to save the states. By using this parameter, 4024 performance of saving the states is degraded because 4025 xsave doesn't support modified optimization while 4026 xsaveopt supports it on xsaveopt enabled systems. 4027 4028 noxsaves [X86] Disables xsaves and xrstors used in saving and 4029 restoring x86 extended register state in compacted 4030 form of xsave area. The kernel will fall back to use 4031 xsaveopt and xrstor to save and restore the states 4032 in standard form of xsave area. By using this 4033 parameter, xsave area per process might occupy more 4034 memory on xsaves enabled systems. 4035 4036 nps_mtm_hs_ctr= [KNL,ARC] 4037 This parameter sets the maximum duration, in 4038 cycles, each HW thread of the CTOP can run 4039 without interruptions, before HW switches it. 4040 The actual maximum duration is 16 times this 4041 parameter's value. 4042 Format: integer between 1 and 255 4043 Default: 255 4044 4045 nptcg= [IA-64] Override max number of concurrent global TLB 4046 purges which is reported from either PAL_VM_SUMMARY or 4047 SAL PALO. 4048 4049 nr_cpus= [SMP] Maximum number of processors that an SMP kernel 4050 could support. nr_cpus=n : n >= 1 limits the kernel to 4051 support 'n' processors. It could be larger than the 4052 number of already plugged CPU during bootup, later in 4053 runtime you can physically add extra cpu until it reaches 4054 n. So during boot up some boot time memory for per-cpu 4055 variables need be pre-allocated for later physical cpu 4056 hot plugging. 4057 4058 nr_uarts= [SERIAL] maximum number of UARTs to be registered. 4059 4060 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only 4061 set up a single NUMA node spanning all memory. 4062 4063 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic 4064 NUMA balancing. 4065 Allowed values are enable and disable 4066 4067 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA. 4068 'node', 'default' can be specified 4069 This can be set from sysctl after boot. 4070 See Documentation/admin-guide/sysctl/vm.rst for details. 4071 4072 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver. 4073 See Documentation/core-api/debugging-via-ohci1394.rst for more 4074 info. 4075 4076 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands 4077 Rather than timing out after 20 ms if an EC 4078 command is not properly ACKed, override the length 4079 of the timeout. We have interrupts disabled while 4080 waiting for the ACK, so if this is set too high 4081 interrupts *may* be lost! 4082 4083 omap_mux= [OMAP] Override bootloader pin multiplexing. 4084 Format: <mux_mode0.mode_name=value>... 4085 For example, to override I2C bus2: 4086 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100 4087 4088 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration 4089 4090 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock] 4091 4092 boundary - index of last SLC block on Flex-OneNAND. 4093 The remaining blocks are configured as MLC blocks. 4094 lock - Configure if Flex-OneNAND boundary should be locked. 4095 Once locked, the boundary cannot be changed. 4096 1 indicates lock status, 0 indicates unlock status. 4097 4098 oops=panic Always panic on oopses. Default is to just kill the 4099 process, but there is a small probability of 4100 deadlocking the machine. 4101 This will also cause panics on machine check exceptions. 4102 Useful together with panic=30 to trigger a reboot. 4103 4104 page_alloc.shuffle= 4105 [KNL] Boolean flag to control whether the page allocator 4106 should randomize its free lists. The randomization may 4107 be automatically enabled if the kernel detects it is 4108 running on a platform with a direct-mapped memory-side 4109 cache, and this parameter can be used to 4110 override/disable that behavior. The state of the flag 4111 can be read from sysfs at: 4112 /sys/module/page_alloc/parameters/shuffle. 4113 4114 page_owner= [KNL] Boot-time page_owner enabling option. 4115 Storage of the information about who allocated 4116 each page is disabled in default. With this switch, 4117 we can turn it on. 4118 on: enable the feature 4119 4120 page_poison= [KNL] Boot-time parameter changing the state of 4121 poisoning on the buddy allocator, available with 4122 CONFIG_PAGE_POISONING=y. 4123 off: turn off poisoning (default) 4124 on: turn on poisoning 4125 4126 page_reporting.page_reporting_order= 4127 [KNL] Minimal page reporting order 4128 Format: <integer> 4129 Adjust the minimal page reporting order. The page 4130 reporting is disabled when it exceeds MAX_ORDER. 4131 4132 panic= [KNL] Kernel behaviour on panic: delay <timeout> 4133 timeout > 0: seconds before rebooting 4134 timeout = 0: wait forever 4135 timeout < 0: reboot immediately 4136 Format: <timeout> 4137 4138 panic_on_taint= Bitmask for conditionally calling panic() in add_taint() 4139 Format: <hex>[,nousertaint] 4140 Hexadecimal bitmask representing the set of TAINT flags 4141 that will cause the kernel to panic when add_taint() is 4142 called with any of the flags in this set. 4143 The optional switch "nousertaint" can be utilized to 4144 prevent userspace forced crashes by writing to sysctl 4145 /proc/sys/kernel/tainted any flagset matching with the 4146 bitmask set on panic_on_taint. 4147 See Documentation/admin-guide/tainted-kernels.rst for 4148 extra details on the taint flags that users can pick 4149 to compose the bitmask to assign to panic_on_taint. 4150 4151 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump 4152 on a WARN(). 4153 4154 panic_print= Bitmask for printing system info when panic happens. 4155 User can chose combination of the following bits: 4156 bit 0: print all tasks info 4157 bit 1: print system memory info 4158 bit 2: print timer info 4159 bit 3: print locks info if CONFIG_LOCKDEP is on 4160 bit 4: print ftrace buffer 4161 bit 5: print all printk messages in buffer 4162 bit 6: print all CPUs backtrace (if available in the arch) 4163 *Be aware* that this option may print a _lot_ of lines, 4164 so there are risks of losing older messages in the log. 4165 Use this option carefully, maybe worth to setup a 4166 bigger log buffer with "log_buf_len" along with this. 4167 4168 parkbd.port= [HW] Parallel port number the keyboard adapter is 4169 connected to, default is 0. 4170 Format: <parport#> 4171 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation, 4172 0 for XT, 1 for AT (default is AT). 4173 Format: <mode> 4174 4175 parport= [HW,PPT] Specify parallel ports. 0 disables. 4176 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] } 4177 Use 'auto' to force the driver to use any 4178 IRQ/DMA settings detected (the default is to 4179 ignore detected IRQ/DMA settings because of 4180 possible conflicts). You can specify the base 4181 address, IRQ, and DMA settings; IRQ and DMA 4182 should be numbers, or 'auto' (for using detected 4183 settings on that particular port), or 'nofifo' 4184 (to avoid using a FIFO even if it is detected). 4185 Parallel ports are assigned in the order they 4186 are specified on the command line, starting 4187 with parport0. 4188 4189 parport_init_mode= [HW,PPT] 4190 Configure VIA parallel port to operate in 4191 a specific mode. This is necessary on Pegasos 4192 computer where firmware has no options for setting 4193 up parallel port mode and sets it to spp. 4194 Currently this function knows 686a and 8231 chips. 4195 Format: [spp|ps2|epp|ecp|ecpepp] 4196 4197 pata_legacy.all= [HW,LIBATA] 4198 Format: <int> 4199 Set to non-zero to probe primary and secondary ISA 4200 port ranges on PCI systems where no PCI PATA device 4201 has been found at either range. Disabled by default. 4202 4203 pata_legacy.autospeed= [HW,LIBATA] 4204 Format: <int> 4205 Set to non-zero if a chip is present that snoops speed 4206 changes. Disabled by default. 4207 4208 pata_legacy.ht6560a= [HW,LIBATA] 4209 Format: <int> 4210 Set to 1, 2, or 3 for HT 6560A on the primary channel, 4211 the secondary channel, or both channels respectively. 4212 Disabled by default. 4213 4214 pata_legacy.ht6560b= [HW,LIBATA] 4215 Format: <int> 4216 Set to 1, 2, or 3 for HT 6560B on the primary channel, 4217 the secondary channel, or both channels respectively. 4218 Disabled by default. 4219 4220 pata_legacy.iordy_mask= [HW,LIBATA] 4221 Format: <int> 4222 IORDY enable mask. Set individual bits to allow IORDY 4223 for the respective channel. Bit 0 is for the first 4224 legacy channel handled by this driver, bit 1 is for 4225 the second channel, and so on. The sequence will often 4226 correspond to the primary legacy channel, the secondary 4227 legacy channel, and so on, but the handling of a PCI 4228 bus and the use of other driver options may interfere 4229 with the sequence. By default IORDY is allowed across 4230 all channels. 4231 4232 pata_legacy.opti82c46x= [HW,LIBATA] 4233 Format: <int> 4234 Set to 1, 2, or 3 for Opti 82c611A on the primary 4235 channel, the secondary channel, or both channels 4236 respectively. Disabled by default. 4237 4238 pata_legacy.opti82c611a= [HW,LIBATA] 4239 Format: <int> 4240 Set to 1, 2, or 3 for Opti 82c465MV on the primary 4241 channel, the secondary channel, or both channels 4242 respectively. Disabled by default. 4243 4244 pata_legacy.pio_mask= [HW,LIBATA] 4245 Format: <int> 4246 PIO mode mask for autospeed devices. Set individual 4247 bits to allow the use of the respective PIO modes. 4248 Bit 0 is for mode 0, bit 1 is for mode 1, and so on. 4249 All modes allowed by default. 4250 4251 pata_legacy.probe_all= [HW,LIBATA] 4252 Format: <int> 4253 Set to non-zero to probe tertiary and further ISA 4254 port ranges on PCI systems. Disabled by default. 4255 4256 pata_legacy.probe_mask= [HW,LIBATA] 4257 Format: <int> 4258 Probe mask for legacy ISA PATA ports. Depending on 4259 platform configuration and the use of other driver 4260 options up to 6 legacy ports are supported: 0x1f0, 4261 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing 4262 of individual ports can be disabled by setting the 4263 corresponding bits in the mask to 1. Bit 0 is for 4264 the first port in the list above (0x1f0), and so on. 4265 By default all supported ports are probed. 4266 4267 pata_legacy.qdi= [HW,LIBATA] 4268 Format: <int> 4269 Set to non-zero to probe QDI controllers. By default 4270 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise. 4271 4272 pata_legacy.winbond= [HW,LIBATA] 4273 Format: <int> 4274 Set to non-zero to probe Winbond controllers. Use 4275 the standard I/O port (0x130) if 1, otherwise the 4276 value given is the I/O port to use (typically 0x1b0). 4277 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE, 4278 0 otherwise. 4279 4280 pata_platform.pio_mask= [HW,LIBATA] 4281 Format: <int> 4282 Supported PIO mode mask. Set individual bits to allow 4283 the use of the respective PIO modes. Bit 0 is for 4284 mode 0, bit 1 is for mode 1, and so on. Mode 0 only 4285 allowed by default. 4286 4287 pause_on_oops=<int> 4288 Halt all CPUs after the first oops has been printed for 4289 the specified number of seconds. This is to be used if 4290 your oopses keep scrolling off the screen. 4291 4292 pcbit= [HW,ISDN] 4293 4294 pci=option[,option...] [PCI] various PCI subsystem options. 4295 4296 Some options herein operate on a specific device 4297 or a set of devices (<pci_dev>). These are 4298 specified in one of the following formats: 4299 4300 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]* 4301 pci:<vendor>:<device>[:<subvendor>:<subdevice>] 4302 4303 Note: the first format specifies a PCI 4304 bus/device/function address which may change 4305 if new hardware is inserted, if motherboard 4306 firmware changes, or due to changes caused 4307 by other kernel parameters. If the 4308 domain is left unspecified, it is 4309 taken to be zero. Optionally, a path 4310 to a device through multiple device/function 4311 addresses can be specified after the base 4312 address (this is more robust against 4313 renumbering issues). The second format 4314 selects devices using IDs from the 4315 configuration space which may match multiple 4316 devices in the system. 4317 4318 earlydump dump PCI config space before the kernel 4319 changes anything 4320 off [X86] don't probe for the PCI bus 4321 bios [X86-32] force use of PCI BIOS, don't access 4322 the hardware directly. Use this if your machine 4323 has a non-standard PCI host bridge. 4324 nobios [X86-32] disallow use of PCI BIOS, only direct 4325 hardware access methods are allowed. Use this 4326 if you experience crashes upon bootup and you 4327 suspect they are caused by the BIOS. 4328 conf1 [X86] Force use of PCI Configuration Access 4329 Mechanism 1 (config address in IO port 0xCF8, 4330 data in IO port 0xCFC, both 32-bit). 4331 conf2 [X86] Force use of PCI Configuration Access 4332 Mechanism 2 (IO port 0xCF8 is an 8-bit port for 4333 the function, IO port 0xCFA, also 8-bit, sets 4334 bus number. The config space is then accessed 4335 through ports 0xC000-0xCFFF). 4336 See http://wiki.osdev.org/PCI for more info 4337 on the configuration access mechanisms. 4338 noaer [PCIE] If the PCIEAER kernel config parameter is 4339 enabled, this kernel boot option can be used to 4340 disable the use of PCIE advanced error reporting. 4341 nodomains [PCI] Disable support for multiple PCI 4342 root domains (aka PCI segments, in ACPI-speak). 4343 nommconf [X86] Disable use of MMCONFIG for PCI 4344 Configuration 4345 check_enable_amd_mmconf [X86] check for and enable 4346 properly configured MMIO access to PCI 4347 config space on AMD family 10h CPU 4348 nomsi [MSI] If the PCI_MSI kernel config parameter is 4349 enabled, this kernel boot option can be used to 4350 disable the use of MSI interrupts system-wide. 4351 noioapicquirk [APIC] Disable all boot interrupt quirks. 4352 Safety option to keep boot IRQs enabled. This 4353 should never be necessary. 4354 ioapicreroute [APIC] Enable rerouting of boot IRQs to the 4355 primary IO-APIC for bridges that cannot disable 4356 boot IRQs. This fixes a source of spurious IRQs 4357 when the system masks IRQs. 4358 noioapicreroute [APIC] Disable workaround that uses the 4359 boot IRQ equivalent of an IRQ that connects to 4360 a chipset where boot IRQs cannot be disabled. 4361 The opposite of ioapicreroute. 4362 biosirq [X86-32] Use PCI BIOS calls to get the interrupt 4363 routing table. These calls are known to be buggy 4364 on several machines and they hang the machine 4365 when used, but on other computers it's the only 4366 way to get the interrupt routing table. Try 4367 this option if the kernel is unable to allocate 4368 IRQs or discover secondary PCI buses on your 4369 motherboard. 4370 rom [X86] Assign address space to expansion ROMs. 4371 Use with caution as certain devices share 4372 address decoders between ROMs and other 4373 resources. 4374 norom [X86] Do not assign address space to 4375 expansion ROMs that do not already have 4376 BIOS assigned address ranges. 4377 nobar [X86] Do not assign address space to the 4378 BARs that weren't assigned by the BIOS. 4379 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be 4380 assigned automatically to PCI devices. You can 4381 make the kernel exclude IRQs of your ISA cards 4382 this way. 4383 pirqaddr=0xAAAAA [X86] Specify the physical address 4384 of the PIRQ table (normally generated 4385 by the BIOS) if it is outside the 4386 F0000h-100000h range. 4387 lastbus=N [X86] Scan all buses thru bus #N. Can be 4388 useful if the kernel is unable to find your 4389 secondary buses and you want to tell it 4390 explicitly which ones they are. 4391 assign-busses [X86] Always assign all PCI bus 4392 numbers ourselves, overriding 4393 whatever the firmware may have done. 4394 usepirqmask [X86] Honor the possible IRQ mask stored 4395 in the BIOS $PIR table. This is needed on 4396 some systems with broken BIOSes, notably 4397 some HP Pavilion N5400 and Omnibook XE3 4398 notebooks. This will have no effect if ACPI 4399 IRQ routing is enabled. 4400 noacpi [X86] Do not use ACPI for IRQ routing 4401 or for PCI scanning. 4402 use_crs [X86] Use PCI host bridge window information 4403 from ACPI. On BIOSes from 2008 or later, this 4404 is enabled by default. If you need to use this, 4405 please report a bug. 4406 nocrs [X86] Ignore PCI host bridge windows from ACPI. 4407 If you need to use this, please report a bug. 4408 use_e820 [X86] Use E820 reservations to exclude parts of 4409 PCI host bridge windows. This is a workaround 4410 for BIOS defects in host bridge _CRS methods. 4411 If you need to use this, please report a bug to 4412 <linux-pci@vger.kernel.org>. 4413 no_e820 [X86] Ignore E820 reservations for PCI host 4414 bridge windows. This is the default on modern 4415 hardware. If you need to use this, please report 4416 a bug to <linux-pci@vger.kernel.org>. 4417 routeirq Do IRQ routing for all PCI devices. 4418 This is normally done in pci_enable_device(), 4419 so this option is a temporary workaround 4420 for broken drivers that don't call it. 4421 skip_isa_align [X86] do not align io start addr, so can 4422 handle more pci cards 4423 noearly [X86] Don't do any early type 1 scanning. 4424 This might help on some broken boards which 4425 machine check when some devices' config space 4426 is read. But various workarounds are disabled 4427 and some IOMMU drivers will not work. 4428 bfsort Sort PCI devices into breadth-first order. 4429 This sorting is done to get a device 4430 order compatible with older (<= 2.4) kernels. 4431 nobfsort Don't sort PCI devices into breadth-first order. 4432 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size) 4433 tuning and use the BIOS-configured MPS defaults. 4434 pcie_bus_safe Set every device's MPS to the largest value 4435 supported by all devices below the root complex. 4436 pcie_bus_perf Set device MPS to the largest allowable MPS 4437 based on its parent bus. Also set MRRS (Max 4438 Read Request Size) to the largest supported 4439 value (no larger than the MPS that the device 4440 or bus can support) for best performance. 4441 pcie_bus_peer2peer Set every device's MPS to 128B, which 4442 every device is guaranteed to support. This 4443 configuration allows peer-to-peer DMA between 4444 any pair of devices, possibly at the cost of 4445 reduced performance. This also guarantees 4446 that hot-added devices will work. 4447 cbiosize=nn[KMG] The fixed amount of bus space which is 4448 reserved for the CardBus bridge's IO window. 4449 The default value is 256 bytes. 4450 cbmemsize=nn[KMG] The fixed amount of bus space which is 4451 reserved for the CardBus bridge's memory 4452 window. The default value is 64 megabytes. 4453 resource_alignment= 4454 Format: 4455 [<order of align>@]<pci_dev>[; ...] 4456 Specifies alignment and device to reassign 4457 aligned memory resources. How to 4458 specify the device is described above. 4459 If <order of align> is not specified, 4460 PAGE_SIZE is used as alignment. 4461 A PCI-PCI bridge can be specified if resource 4462 windows need to be expanded. 4463 To specify the alignment for several 4464 instances of a device, the PCI vendor, 4465 device, subvendor, and subdevice may be 4466 specified, e.g., 12@pci:8086:9c22:103c:198f 4467 for 4096-byte alignment. 4468 ecrc= Enable/disable PCIe ECRC (transaction layer 4469 end-to-end CRC checking). Only effective if 4470 OS has native AER control (either granted by 4471 ACPI _OSC or forced via "pcie_ports=native") 4472 bios: Use BIOS/firmware settings. This is the 4473 the default. 4474 off: Turn ECRC off 4475 on: Turn ECRC on. 4476 hpiosize=nn[KMG] The fixed amount of bus space which is 4477 reserved for hotplug bridge's IO window. 4478 Default size is 256 bytes. 4479 hpmmiosize=nn[KMG] The fixed amount of bus space which is 4480 reserved for hotplug bridge's MMIO window. 4481 Default size is 2 megabytes. 4482 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is 4483 reserved for hotplug bridge's MMIO_PREF window. 4484 Default size is 2 megabytes. 4485 hpmemsize=nn[KMG] The fixed amount of bus space which is 4486 reserved for hotplug bridge's MMIO and 4487 MMIO_PREF window. 4488 Default size is 2 megabytes. 4489 hpbussize=nn The minimum amount of additional bus numbers 4490 reserved for buses below a hotplug bridge. 4491 Default is 1. 4492 realloc= Enable/disable reallocating PCI bridge resources 4493 if allocations done by BIOS are too small to 4494 accommodate resources required by all child 4495 devices. 4496 off: Turn realloc off 4497 on: Turn realloc on 4498 realloc same as realloc=on 4499 noari do not use PCIe ARI. 4500 noats [PCIE, Intel-IOMMU, AMD-IOMMU] 4501 do not use PCIe ATS (and IOMMU device IOTLB). 4502 pcie_scan_all Scan all possible PCIe devices. Otherwise we 4503 only look for one device below a PCIe downstream 4504 port. 4505 big_root_window Try to add a big 64bit memory window to the PCIe 4506 root complex on AMD CPUs. Some GFX hardware 4507 can resize a BAR to allow access to all VRAM. 4508 Adding the window is slightly risky (it may 4509 conflict with unreported devices), so this 4510 taints the kernel. 4511 disable_acs_redir=<pci_dev>[; ...] 4512 Specify one or more PCI devices (in the format 4513 specified above) separated by semicolons. 4514 Each device specified will have the PCI ACS 4515 redirect capabilities forced off which will 4516 allow P2P traffic between devices through 4517 bridges without forcing it upstream. Note: 4518 this removes isolation between devices and 4519 may put more devices in an IOMMU group. 4520 force_floating [S390] Force usage of floating interrupts. 4521 nomio [S390] Do not use MIO instructions. 4522 norid [S390] ignore the RID field and force use of 4523 one PCI domain per PCI function 4524 4525 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power 4526 Management. 4527 off Disable ASPM. 4528 force Enable ASPM even on devices that claim not to support it. 4529 WARNING: Forcing ASPM on may cause system lockups. 4530 4531 pcie_ports= [PCIE] PCIe port services handling: 4532 native Use native PCIe services (PME, AER, DPC, PCIe hotplug) 4533 even if the platform doesn't give the OS permission to 4534 use them. This may cause conflicts if the platform 4535 also tries to use these services. 4536 dpc-native Use native PCIe service for DPC only. May 4537 cause conflicts if firmware uses AER or DPC. 4538 compat Disable native PCIe services (PME, AER, DPC, PCIe 4539 hotplug). 4540 4541 pcie_port_pm= [PCIE] PCIe port power management handling: 4542 off Disable power management of all PCIe ports 4543 force Forcibly enable power management of all PCIe ports 4544 4545 pcie_pme= [PCIE,PM] Native PCIe PME signaling options: 4546 nomsi Do not use MSI for native PCIe PME signaling (this makes 4547 all PCIe root ports use INTx for all services). 4548 4549 pcmv= [HW,PCMCIA] BadgePAD 4 4550 4551 pd_ignore_unused 4552 [PM] 4553 Keep all power-domains already enabled by bootloader on, 4554 even if no driver has claimed them. This is useful 4555 for debug and development, but should not be 4556 needed on a platform with proper driver support. 4557 4558 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at 4559 boot time. 4560 Format: { 0 | 1 } 4561 See arch/parisc/kernel/pdc_chassis.c 4562 4563 percpu_alloc= Select which percpu first chunk allocator to use. 4564 Currently supported values are "embed" and "page". 4565 Archs may support subset or none of the selections. 4566 See comments in mm/percpu.c for details on each 4567 allocator. This parameter is primarily for debugging 4568 and performance comparison. 4569 4570 pirq= [SMP,APIC] Manual mp-table setup 4571 See Documentation/arch/x86/i386/IO-APIC.rst. 4572 4573 plip= [PPT,NET] Parallel port network link 4574 Format: { parport<nr> | timid | 0 } 4575 See also Documentation/admin-guide/parport.rst. 4576 4577 pmtmr= [X86] Manual setup of pmtmr I/O Port. 4578 Override pmtimer IOPort with a hex value. 4579 e.g. pmtmr=0x508 4580 4581 pmu_override= [PPC] Override the PMU. 4582 This option takes over the PMU facility, so it is no 4583 longer usable by perf. Setting this option starts the 4584 PMU counters by setting MMCR0 to 0 (the FC bit is 4585 cleared). If a number is given, then MMCR1 is set to 4586 that number, otherwise (e.g., 'pmu_override=on'), MMCR1 4587 remains 0. 4588 4589 pm_debug_messages [SUSPEND,KNL] 4590 Enable suspend/resume debug messages during boot up. 4591 4592 pnp.debug=1 [PNP] 4593 Enable PNP debug messages (depends on the 4594 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time 4595 via /sys/module/pnp/parameters/debug. We always show 4596 current resource usage; turning this on also shows 4597 possible settings and some assignment information. 4598 4599 pnpacpi= [ACPI] 4600 { off } 4601 4602 pnpbios= [ISAPNP] 4603 { on | off | curr | res | no-curr | no-res } 4604 4605 pnp_reserve_irq= 4606 [ISAPNP] Exclude IRQs for the autoconfiguration 4607 4608 pnp_reserve_dma= 4609 [ISAPNP] Exclude DMAs for the autoconfiguration 4610 4611 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration 4612 Ranges are in pairs (I/O port base and size). 4613 4614 pnp_reserve_mem= 4615 [ISAPNP] Exclude memory regions for the 4616 autoconfiguration. 4617 Ranges are in pairs (memory base and size). 4618 4619 ports= [IP_VS_FTP] IPVS ftp helper module 4620 Default is 21. 4621 Up to 8 (IP_VS_APP_MAX_PORTS) ports 4622 may be specified. 4623 Format: <port>,<port>.... 4624 4625 powersave=off [PPC] This option disables power saving features. 4626 It specifically disables cpuidle and sets the 4627 platform machine description specific power_save 4628 function to NULL. On Idle the CPU just reduces 4629 execution priority. 4630 4631 ppc_strict_facility_enable 4632 [PPC] This option catches any kernel floating point, 4633 Altivec, VSX and SPE outside of regions specifically 4634 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()). 4635 There is some performance impact when enabling this. 4636 4637 ppc_tm= [PPC] 4638 Format: {"off"} 4639 Disable Hardware Transactional Memory 4640 4641 preempt= [KNL] 4642 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC 4643 none - Limited to cond_resched() calls 4644 voluntary - Limited to cond_resched() and might_sleep() calls 4645 full - Any section that isn't explicitly preempt disabled 4646 can be preempted anytime. 4647 4648 print-fatal-signals= 4649 [KNL] debug: print fatal signals 4650 4651 If enabled, warn about various signal handling 4652 related application anomalies: too many signals, 4653 too many POSIX.1 timers, fatal signals causing a 4654 coredump - etc. 4655 4656 If you hit the warning due to signal overflow, 4657 you might want to try "ulimit -i unlimited". 4658 4659 default: off. 4660 4661 printk.always_kmsg_dump= 4662 Trigger kmsg_dump for cases other than kernel oops or 4663 panics 4664 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 4665 default: disabled 4666 4667 printk.console_no_auto_verbose= 4668 Disable console loglevel raise on oops, panic 4669 or lockdep-detected issues (only if lock debug is on). 4670 With an exception to setups with low baudrate on 4671 serial console, keeping this 0 is a good choice 4672 in order to provide more debug information. 4673 Format: <bool> 4674 default: 0 (auto_verbose is enabled) 4675 4676 printk.devkmsg={on,off,ratelimit} 4677 Control writing to /dev/kmsg. 4678 on - unlimited logging to /dev/kmsg from userspace 4679 off - logging to /dev/kmsg disabled 4680 ratelimit - ratelimit the logging 4681 Default: ratelimit 4682 4683 printk.time= Show timing data prefixed to each printk message line 4684 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 4685 4686 processor.max_cstate= [HW,ACPI] 4687 Limit processor to maximum C-state 4688 max_cstate=9 overrides any DMI blacklist limit. 4689 4690 processor.nocst [HW,ACPI] 4691 Ignore the _CST method to determine C-states, 4692 instead using the legacy FADT method 4693 4694 profile= [KNL] Enable kernel profiling via /proc/profile 4695 Format: [<profiletype>,]<number> 4696 Param: <profiletype>: "schedule", "sleep", or "kvm" 4697 [defaults to kernel profiling] 4698 Param: "schedule" - profile schedule points. 4699 Param: "sleep" - profile D-state sleeping (millisecs). 4700 Requires CONFIG_SCHEDSTATS 4701 Param: "kvm" - profile VM exits. 4702 Param: <number> - step/bucket size as a power of 2 for 4703 statistical time based profiling. 4704 4705 prompt_ramdisk= [RAM] [Deprecated] 4706 4707 prot_virt= [S390] enable hosting protected virtual machines 4708 isolated from the hypervisor (if hardware supports 4709 that). 4710 Format: <bool> 4711 4712 psi= [KNL] Enable or disable pressure stall information 4713 tracking. 4714 Format: <bool> 4715 4716 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to 4717 probe for; one of (bare|imps|exps|lifebook|any). 4718 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports 4719 per second. 4720 psmouse.resetafter= [HW,MOUSE] 4721 Try to reset the device after so many bad packets 4722 (0 = never). 4723 psmouse.resolution= 4724 [HW,MOUSE] Set desired mouse resolution, in dpi. 4725 psmouse.smartscroll= 4726 [HW,MOUSE] Controls Logitech smartscroll autorepeat. 4727 0 = disabled, 1 = enabled (default). 4728 4729 pstore.backend= Specify the name of the pstore backend to use 4730 4731 pti= [X86-64] Control Page Table Isolation of user and 4732 kernel address spaces. Disabling this feature 4733 removes hardening, but improves performance of 4734 system calls and interrupts. 4735 4736 on - unconditionally enable 4737 off - unconditionally disable 4738 auto - kernel detects whether your CPU model is 4739 vulnerable to issues that PTI mitigates 4740 4741 Not specifying this option is equivalent to pti=auto. 4742 4743 pty.legacy_count= 4744 [KNL] Number of legacy pty's. Overwrites compiled-in 4745 default number. 4746 4747 quiet [KNL] Disable most log messages 4748 4749 r128= [HW,DRM] 4750 4751 radix_hcall_invalidate=on [PPC/PSERIES] 4752 Disable RADIX GTSE feature and use hcall for TLB 4753 invalidate. 4754 4755 raid= [HW,RAID] 4756 See Documentation/admin-guide/md.rst. 4757 4758 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes 4759 See Documentation/admin-guide/blockdev/ramdisk.rst. 4760 4761 ramdisk_start= [RAM] RAM disk image start address 4762 4763 random.trust_cpu=off 4764 [KNL] Disable trusting the use of the CPU's 4765 random number generator (if available) to 4766 initialize the kernel's RNG. 4767 4768 random.trust_bootloader=off 4769 [KNL] Disable trusting the use of the a seed 4770 passed by the bootloader (if available) to 4771 initialize the kernel's RNG. 4772 4773 randomize_kstack_offset= 4774 [KNL] Enable or disable kernel stack offset 4775 randomization, which provides roughly 5 bits of 4776 entropy, frustrating memory corruption attacks 4777 that depend on stack address determinism or 4778 cross-syscall address exposures. This is only 4779 available on architectures that have defined 4780 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET. 4781 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 4782 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT. 4783 4784 ras=option[,option,...] [KNL] RAS-specific options 4785 4786 cec_disable [X86] 4787 Disable the Correctable Errors Collector, 4788 see CONFIG_RAS_CEC help text. 4789 4790 rcu_nocbs[=cpu-list] 4791 [KNL] The optional argument is a cpu list, 4792 as described above. 4793 4794 In kernels built with CONFIG_RCU_NOCB_CPU=y, 4795 enable the no-callback CPU mode, which prevents 4796 such CPUs' callbacks from being invoked in 4797 softirq context. Invocation of such CPUs' RCU 4798 callbacks will instead be offloaded to "rcuox/N" 4799 kthreads created for that purpose, where "x" is 4800 "p" for RCU-preempt, "s" for RCU-sched, and "g" 4801 for the kthreads that mediate grace periods; and 4802 "N" is the CPU number. This reduces OS jitter on 4803 the offloaded CPUs, which can be useful for HPC 4804 and real-time workloads. It can also improve 4805 energy efficiency for asymmetric multiprocessors. 4806 4807 If a cpulist is passed as an argument, the specified 4808 list of CPUs is set to no-callback mode from boot. 4809 4810 Otherwise, if the '=' sign and the cpulist 4811 arguments are omitted, no CPU will be set to 4812 no-callback mode from boot but the mode may be 4813 toggled at runtime via cpusets. 4814 4815 Note that this argument takes precedence over 4816 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 4817 4818 rcu_nocb_poll [KNL] 4819 Rather than requiring that offloaded CPUs 4820 (specified by rcu_nocbs= above) explicitly 4821 awaken the corresponding "rcuoN" kthreads, 4822 make these kthreads poll for callbacks. 4823 This improves the real-time response for the 4824 offloaded CPUs by relieving them of the need to 4825 wake up the corresponding kthread, but degrades 4826 energy efficiency by requiring that the kthreads 4827 periodically wake up to do the polling. 4828 4829 rcutree.blimit= [KNL] 4830 Set maximum number of finished RCU callbacks to 4831 process in one batch. 4832 4833 rcutree.do_rcu_barrier= [KNL] 4834 Request a call to rcu_barrier(). This is 4835 throttled so that userspace tests can safely 4836 hammer on the sysfs variable if they so choose. 4837 If triggered before the RCU grace-period machinery 4838 is fully active, this will error out with EAGAIN. 4839 4840 rcutree.dump_tree= [KNL] 4841 Dump the structure of the rcu_node combining tree 4842 out at early boot. This is used for diagnostic 4843 purposes, to verify correct tree setup. 4844 4845 rcutree.gp_cleanup_delay= [KNL] 4846 Set the number of jiffies to delay each step of 4847 RCU grace-period cleanup. 4848 4849 rcutree.gp_init_delay= [KNL] 4850 Set the number of jiffies to delay each step of 4851 RCU grace-period initialization. 4852 4853 rcutree.gp_preinit_delay= [KNL] 4854 Set the number of jiffies to delay each step of 4855 RCU grace-period pre-initialization, that is, 4856 the propagation of recent CPU-hotplug changes up 4857 the rcu_node combining tree. 4858 4859 rcutree.jiffies_till_first_fqs= [KNL] 4860 Set delay from grace-period initialization to 4861 first attempt to force quiescent states. 4862 Units are jiffies, minimum value is zero, 4863 and maximum value is HZ. 4864 4865 rcutree.jiffies_till_next_fqs= [KNL] 4866 Set delay between subsequent attempts to force 4867 quiescent states. Units are jiffies, minimum 4868 value is one, and maximum value is HZ. 4869 4870 rcutree.jiffies_till_sched_qs= [KNL] 4871 Set required age in jiffies for a 4872 given grace period before RCU starts 4873 soliciting quiescent-state help from 4874 rcu_note_context_switch() and cond_resched(). 4875 If not specified, the kernel will calculate 4876 a value based on the most recent settings 4877 of rcutree.jiffies_till_first_fqs 4878 and rcutree.jiffies_till_next_fqs. 4879 This calculated value may be viewed in 4880 rcutree.jiffies_to_sched_qs. Any attempt to set 4881 rcutree.jiffies_to_sched_qs will be cheerfully 4882 overwritten. 4883 4884 rcutree.kthread_prio= [KNL,BOOT] 4885 Set the SCHED_FIFO priority of the RCU per-CPU 4886 kthreads (rcuc/N). This value is also used for 4887 the priority of the RCU boost threads (rcub/N) 4888 and for the RCU grace-period kthreads (rcu_bh, 4889 rcu_preempt, and rcu_sched). If RCU_BOOST is 4890 set, valid values are 1-99 and the default is 1 4891 (the least-favored priority). Otherwise, when 4892 RCU_BOOST is not set, valid values are 0-99 and 4893 the default is zero (non-realtime operation). 4894 When RCU_NOCB_CPU is set, also adjust the 4895 priority of NOCB callback kthreads. 4896 4897 rcutree.nocb_nobypass_lim_per_jiffy= [KNL] 4898 On callback-offloaded (rcu_nocbs) CPUs, 4899 RCU reduces the lock contention that would 4900 otherwise be caused by callback floods through 4901 use of the ->nocb_bypass list. However, in the 4902 common non-flooded case, RCU queues directly to 4903 the main ->cblist in order to avoid the extra 4904 overhead of the ->nocb_bypass list and its lock. 4905 But if there are too many callbacks queued during 4906 a single jiffy, RCU pre-queues the callbacks into 4907 the ->nocb_bypass queue. The definition of "too 4908 many" is supplied by this kernel boot parameter. 4909 4910 rcutree.qhimark= [KNL] 4911 Set threshold of queued RCU callbacks beyond which 4912 batch limiting is disabled. 4913 4914 rcutree.qlowmark= [KNL] 4915 Set threshold of queued RCU callbacks below which 4916 batch limiting is re-enabled. 4917 4918 rcutree.qovld= [KNL] 4919 Set threshold of queued RCU callbacks beyond which 4920 RCU's force-quiescent-state scan will aggressively 4921 enlist help from cond_resched() and sched IPIs to 4922 help CPUs more quickly reach quiescent states. 4923 Set to less than zero to make this be set based 4924 on rcutree.qhimark at boot time and to zero to 4925 disable more aggressive help enlistment. 4926 4927 rcutree.rcu_delay_page_cache_fill_msec= [KNL] 4928 Set the page-cache refill delay (in milliseconds) 4929 in response to low-memory conditions. The range 4930 of permitted values is in the range 0:100000. 4931 4932 rcutree.rcu_divisor= [KNL] 4933 Set the shift-right count to use to compute 4934 the callback-invocation batch limit bl from 4935 the number of callbacks queued on this CPU. 4936 The result will be bounded below by the value of 4937 the rcutree.blimit kernel parameter. Every bl 4938 callbacks, the softirq handler will exit in 4939 order to allow the CPU to do other work. 4940 4941 Please note that this callback-invocation batch 4942 limit applies only to non-offloaded callback 4943 invocation. Offloaded callbacks are instead 4944 invoked in the context of an rcuoc kthread, which 4945 scheduler will preempt as it does any other task. 4946 4947 rcutree.rcu_fanout_exact= [KNL] 4948 Disable autobalancing of the rcu_node combining 4949 tree. This is used by rcutorture, and might 4950 possibly be useful for architectures having high 4951 cache-to-cache transfer latencies. 4952 4953 rcutree.rcu_fanout_leaf= [KNL] 4954 Change the number of CPUs assigned to each 4955 leaf rcu_node structure. Useful for very 4956 large systems, which will choose the value 64, 4957 and for NUMA systems with large remote-access 4958 latencies, which will choose a value aligned 4959 with the appropriate hardware boundaries. 4960 4961 rcutree.rcu_min_cached_objs= [KNL] 4962 Minimum number of objects which are cached and 4963 maintained per one CPU. Object size is equal 4964 to PAGE_SIZE. The cache allows to reduce the 4965 pressure to page allocator, also it makes the 4966 whole algorithm to behave better in low memory 4967 condition. 4968 4969 rcutree.rcu_nocb_gp_stride= [KNL] 4970 Set the number of NOCB callback kthreads in 4971 each group, which defaults to the square root 4972 of the number of CPUs. Larger numbers reduce 4973 the wakeup overhead on the global grace-period 4974 kthread, but increases that same overhead on 4975 each group's NOCB grace-period kthread. 4976 4977 rcutree.rcu_kick_kthreads= [KNL] 4978 Cause the grace-period kthread to get an extra 4979 wake_up() if it sleeps three times longer than 4980 it should at force-quiescent-state time. 4981 This wake_up() will be accompanied by a 4982 WARN_ONCE() splat and an ftrace_dump(). 4983 4984 rcutree.rcu_resched_ns= [KNL] 4985 Limit the time spend invoking a batch of RCU 4986 callbacks to the specified number of nanoseconds. 4987 By default, this limit is checked only once 4988 every 32 callbacks in order to limit the pain 4989 inflicted by local_clock() overhead. 4990 4991 rcutree.rcu_unlock_delay= [KNL] 4992 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels, 4993 this specifies an rcu_read_unlock()-time delay 4994 in microseconds. This defaults to zero. 4995 Larger delays increase the probability of 4996 catching RCU pointer leaks, that is, buggy use 4997 of RCU-protected pointers after the relevant 4998 rcu_read_unlock() has completed. 4999 5000 rcutree.sysrq_rcu= [KNL] 5001 Commandeer a sysrq key to dump out Tree RCU's 5002 rcu_node tree with an eye towards determining 5003 why a new grace period has not yet started. 5004 5005 rcutree.use_softirq= [KNL] 5006 If set to zero, move all RCU_SOFTIRQ processing to 5007 per-CPU rcuc kthreads. Defaults to a non-zero 5008 value, meaning that RCU_SOFTIRQ is used by default. 5009 Specify rcutree.use_softirq=0 to use rcuc kthreads. 5010 5011 But note that CONFIG_PREEMPT_RT=y kernels disable 5012 this kernel boot parameter, forcibly setting it 5013 to zero. 5014 5015 rcuscale.gp_async= [KNL] 5016 Measure performance of asynchronous 5017 grace-period primitives such as call_rcu(). 5018 5019 rcuscale.gp_async_max= [KNL] 5020 Specify the maximum number of outstanding 5021 callbacks per writer thread. When a writer 5022 thread exceeds this limit, it invokes the 5023 corresponding flavor of rcu_barrier() to allow 5024 previously posted callbacks to drain. 5025 5026 rcuscale.gp_exp= [KNL] 5027 Measure performance of expedited synchronous 5028 grace-period primitives. 5029 5030 rcuscale.holdoff= [KNL] 5031 Set test-start holdoff period. The purpose of 5032 this parameter is to delay the start of the 5033 test until boot completes in order to avoid 5034 interference. 5035 5036 rcuscale.kfree_by_call_rcu= [KNL] 5037 In kernels built with CONFIG_RCU_LAZY=y, test 5038 call_rcu() instead of kfree_rcu(). 5039 5040 rcuscale.kfree_mult= [KNL] 5041 Instead of allocating an object of size kfree_obj, 5042 allocate one of kfree_mult * sizeof(kfree_obj). 5043 Defaults to 1. 5044 5045 rcuscale.kfree_rcu_test= [KNL] 5046 Set to measure performance of kfree_rcu() flooding. 5047 5048 rcuscale.kfree_rcu_test_double= [KNL] 5049 Test the double-argument variant of kfree_rcu(). 5050 If this parameter has the same value as 5051 rcuscale.kfree_rcu_test_single, both the single- 5052 and double-argument variants are tested. 5053 5054 rcuscale.kfree_rcu_test_single= [KNL] 5055 Test the single-argument variant of kfree_rcu(). 5056 If this parameter has the same value as 5057 rcuscale.kfree_rcu_test_double, both the single- 5058 and double-argument variants are tested. 5059 5060 rcuscale.kfree_nthreads= [KNL] 5061 The number of threads running loops of kfree_rcu(). 5062 5063 rcuscale.kfree_alloc_num= [KNL] 5064 Number of allocations and frees done in an iteration. 5065 5066 rcuscale.kfree_loops= [KNL] 5067 Number of loops doing rcuscale.kfree_alloc_num number 5068 of allocations and frees. 5069 5070 rcuscale.minruntime= [KNL] 5071 Set the minimum test run time in seconds. This 5072 does not affect the data-collection interval, 5073 but instead allows better measurement of things 5074 like CPU consumption. 5075 5076 rcuscale.nreaders= [KNL] 5077 Set number of RCU readers. The value -1 selects 5078 N, where N is the number of CPUs. A value 5079 "n" less than -1 selects N-n+1, where N is again 5080 the number of CPUs. For example, -2 selects N 5081 (the number of CPUs), -3 selects N+1, and so on. 5082 A value of "n" less than or equal to -N selects 5083 a single reader. 5084 5085 rcuscale.nwriters= [KNL] 5086 Set number of RCU writers. The values operate 5087 the same as for rcuscale.nreaders. 5088 N, where N is the number of CPUs 5089 5090 rcuscale.scale_type= [KNL] 5091 Specify the RCU implementation to test. 5092 5093 rcuscale.shutdown= [KNL] 5094 Shut the system down after performance tests 5095 complete. This is useful for hands-off automated 5096 testing. 5097 5098 rcuscale.verbose= [KNL] 5099 Enable additional printk() statements. 5100 5101 rcuscale.writer_holdoff= [KNL] 5102 Write-side holdoff between grace periods, 5103 in microseconds. The default of zero says 5104 no holdoff. 5105 5106 rcuscale.writer_holdoff_jiffies= [KNL] 5107 Additional write-side holdoff between grace 5108 periods, but in jiffies. The default of zero 5109 says no holdoff. 5110 5111 rcutorture.fqs_duration= [KNL] 5112 Set duration of force_quiescent_state bursts 5113 in microseconds. 5114 5115 rcutorture.fqs_holdoff= [KNL] 5116 Set holdoff time within force_quiescent_state bursts 5117 in microseconds. 5118 5119 rcutorture.fqs_stutter= [KNL] 5120 Set wait time between force_quiescent_state bursts 5121 in seconds. 5122 5123 rcutorture.fwd_progress= [KNL] 5124 Specifies the number of kthreads to be used 5125 for RCU grace-period forward-progress testing 5126 for the types of RCU supporting this notion. 5127 Defaults to 1 kthread, values less than zero or 5128 greater than the number of CPUs cause the number 5129 of CPUs to be used. 5130 5131 rcutorture.fwd_progress_div= [KNL] 5132 Specify the fraction of a CPU-stall-warning 5133 period to do tight-loop forward-progress testing. 5134 5135 rcutorture.fwd_progress_holdoff= [KNL] 5136 Number of seconds to wait between successive 5137 forward-progress tests. 5138 5139 rcutorture.fwd_progress_need_resched= [KNL] 5140 Enclose cond_resched() calls within checks for 5141 need_resched() during tight-loop forward-progress 5142 testing. 5143 5144 rcutorture.gp_cond= [KNL] 5145 Use conditional/asynchronous update-side 5146 primitives, if available. 5147 5148 rcutorture.gp_exp= [KNL] 5149 Use expedited update-side primitives, if available. 5150 5151 rcutorture.gp_normal= [KNL] 5152 Use normal (non-expedited) asynchronous 5153 update-side primitives, if available. 5154 5155 rcutorture.gp_sync= [KNL] 5156 Use normal (non-expedited) synchronous 5157 update-side primitives, if available. If all 5158 of rcutorture.gp_cond=, rcutorture.gp_exp=, 5159 rcutorture.gp_normal=, and rcutorture.gp_sync= 5160 are zero, rcutorture acts as if is interpreted 5161 they are all non-zero. 5162 5163 rcutorture.irqreader= [KNL] 5164 Run RCU readers from irq handlers, or, more 5165 accurately, from a timer handler. Not all RCU 5166 flavors take kindly to this sort of thing. 5167 5168 rcutorture.leakpointer= [KNL] 5169 Leak an RCU-protected pointer out of the reader. 5170 This can of course result in splats, and is 5171 intended to test the ability of things like 5172 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect 5173 such leaks. 5174 5175 rcutorture.n_barrier_cbs= [KNL] 5176 Set callbacks/threads for rcu_barrier() testing. 5177 5178 rcutorture.nfakewriters= [KNL] 5179 Set number of concurrent RCU writers. These just 5180 stress RCU, they don't participate in the actual 5181 test, hence the "fake". 5182 5183 rcutorture.nocbs_nthreads= [KNL] 5184 Set number of RCU callback-offload togglers. 5185 Zero (the default) disables toggling. 5186 5187 rcutorture.nocbs_toggle= [KNL] 5188 Set the delay in milliseconds between successive 5189 callback-offload toggling attempts. 5190 5191 rcutorture.nreaders= [KNL] 5192 Set number of RCU readers. The value -1 selects 5193 N-1, where N is the number of CPUs. A value 5194 "n" less than -1 selects N-n-2, where N is again 5195 the number of CPUs. For example, -2 selects N 5196 (the number of CPUs), -3 selects N+1, and so on. 5197 5198 rcutorture.object_debug= [KNL] 5199 Enable debug-object double-call_rcu() testing. 5200 5201 rcutorture.onoff_holdoff= [KNL] 5202 Set time (s) after boot for CPU-hotplug testing. 5203 5204 rcutorture.onoff_interval= [KNL] 5205 Set time (jiffies) between CPU-hotplug operations, 5206 or zero to disable CPU-hotplug testing. 5207 5208 rcutorture.read_exit= [KNL] 5209 Set the number of read-then-exit kthreads used 5210 to test the interaction of RCU updaters and 5211 task-exit processing. 5212 5213 rcutorture.read_exit_burst= [KNL] 5214 The number of times in a given read-then-exit 5215 episode that a set of read-then-exit kthreads 5216 is spawned. 5217 5218 rcutorture.read_exit_delay= [KNL] 5219 The delay, in seconds, between successive 5220 read-then-exit testing episodes. 5221 5222 rcutorture.shuffle_interval= [KNL] 5223 Set task-shuffle interval (s). Shuffling tasks 5224 allows some CPUs to go into dyntick-idle mode 5225 during the rcutorture test. 5226 5227 rcutorture.shutdown_secs= [KNL] 5228 Set time (s) after boot system shutdown. This 5229 is useful for hands-off automated testing. 5230 5231 rcutorture.stall_cpu= [KNL] 5232 Duration of CPU stall (s) to test RCU CPU stall 5233 warnings, zero to disable. 5234 5235 rcutorture.stall_cpu_block= [KNL] 5236 Sleep while stalling if set. This will result 5237 in warnings from preemptible RCU in addition to 5238 any other stall-related activity. Note that 5239 in kernels built with CONFIG_PREEMPTION=n and 5240 CONFIG_PREEMPT_COUNT=y, this parameter will 5241 cause the CPU to pass through a quiescent state. 5242 Given CONFIG_PREEMPTION=n, this will suppress 5243 RCU CPU stall warnings, but will instead result 5244 in scheduling-while-atomic splats. 5245 5246 Use of this module parameter results in splats. 5247 5248 5249 rcutorture.stall_cpu_holdoff= [KNL] 5250 Time to wait (s) after boot before inducing stall. 5251 5252 rcutorture.stall_cpu_irqsoff= [KNL] 5253 Disable interrupts while stalling if set. 5254 5255 rcutorture.stall_gp_kthread= [KNL] 5256 Duration (s) of forced sleep within RCU 5257 grace-period kthread to test RCU CPU stall 5258 warnings, zero to disable. If both stall_cpu 5259 and stall_gp_kthread are specified, the 5260 kthread is starved first, then the CPU. 5261 5262 rcutorture.stat_interval= [KNL] 5263 Time (s) between statistics printk()s. 5264 5265 rcutorture.stutter= [KNL] 5266 Time (s) to stutter testing, for example, specifying 5267 five seconds causes the test to run for five seconds, 5268 wait for five seconds, and so on. This tests RCU's 5269 ability to transition abruptly to and from idle. 5270 5271 rcutorture.test_boost= [KNL] 5272 Test RCU priority boosting? 0=no, 1=maybe, 2=yes. 5273 "Maybe" means test if the RCU implementation 5274 under test support RCU priority boosting. 5275 5276 rcutorture.test_boost_duration= [KNL] 5277 Duration (s) of each individual boost test. 5278 5279 rcutorture.test_boost_interval= [KNL] 5280 Interval (s) between each boost test. 5281 5282 rcutorture.test_no_idle_hz= [KNL] 5283 Test RCU's dyntick-idle handling. See also the 5284 rcutorture.shuffle_interval parameter. 5285 5286 rcutorture.torture_type= [KNL] 5287 Specify the RCU implementation to test. 5288 5289 rcutorture.verbose= [KNL] 5290 Enable additional printk() statements. 5291 5292 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL] 5293 Dump ftrace buffer after reporting RCU CPU 5294 stall warning. 5295 5296 rcupdate.rcu_cpu_stall_suppress= [KNL] 5297 Suppress RCU CPU stall warning messages. 5298 5299 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL] 5300 Suppress RCU CPU stall warning messages and 5301 rcutorture writer stall warnings that occur 5302 during early boot, that is, during the time 5303 before the init task is spawned. 5304 5305 rcupdate.rcu_cpu_stall_timeout= [KNL] 5306 Set timeout for RCU CPU stall warning messages. 5307 The value is in seconds and the maximum allowed 5308 value is 300 seconds. 5309 5310 rcupdate.rcu_exp_cpu_stall_timeout= [KNL] 5311 Set timeout for expedited RCU CPU stall warning 5312 messages. The value is in milliseconds 5313 and the maximum allowed value is 21000 5314 milliseconds. Please note that this value is 5315 adjusted to an arch timer tick resolution. 5316 Setting this to zero causes the value from 5317 rcupdate.rcu_cpu_stall_timeout to be used (after 5318 conversion from seconds to milliseconds). 5319 5320 rcupdate.rcu_cpu_stall_cputime= [KNL] 5321 Provide statistics on the cputime and count of 5322 interrupts and tasks during the sampling period. For 5323 multiple continuous RCU stalls, all sampling periods 5324 begin at half of the first RCU stall timeout. 5325 5326 rcupdate.rcu_exp_stall_task_details= [KNL] 5327 Print stack dumps of any tasks blocking the 5328 current expedited RCU grace period during an 5329 expedited RCU CPU stall warning. 5330 5331 rcupdate.rcu_expedited= [KNL] 5332 Use expedited grace-period primitives, for 5333 example, synchronize_rcu_expedited() instead 5334 of synchronize_rcu(). This reduces latency, 5335 but can increase CPU utilization, degrade 5336 real-time latency, and degrade energy efficiency. 5337 No effect on CONFIG_TINY_RCU kernels. 5338 5339 rcupdate.rcu_normal= [KNL] 5340 Use only normal grace-period primitives, 5341 for example, synchronize_rcu() instead of 5342 synchronize_rcu_expedited(). This improves 5343 real-time latency, CPU utilization, and 5344 energy efficiency, but can expose users to 5345 increased grace-period latency. This parameter 5346 overrides rcupdate.rcu_expedited. No effect on 5347 CONFIG_TINY_RCU kernels. 5348 5349 rcupdate.rcu_normal_after_boot= [KNL] 5350 Once boot has completed (that is, after 5351 rcu_end_inkernel_boot() has been invoked), use 5352 only normal grace-period primitives. No effect 5353 on CONFIG_TINY_RCU kernels. 5354 5355 But note that CONFIG_PREEMPT_RT=y kernels enables 5356 this kernel boot parameter, forcibly setting 5357 it to the value one, that is, converting any 5358 post-boot attempt at an expedited RCU grace 5359 period to instead use normal non-expedited 5360 grace-period processing. 5361 5362 rcupdate.rcu_task_collapse_lim= [KNL] 5363 Set the maximum number of callbacks present 5364 at the beginning of a grace period that allows 5365 the RCU Tasks flavors to collapse back to using 5366 a single callback queue. This switching only 5367 occurs when rcupdate.rcu_task_enqueue_lim is 5368 set to the default value of -1. 5369 5370 rcupdate.rcu_task_contend_lim= [KNL] 5371 Set the minimum number of callback-queuing-time 5372 lock-contention events per jiffy required to 5373 cause the RCU Tasks flavors to switch to per-CPU 5374 callback queuing. This switching only occurs 5375 when rcupdate.rcu_task_enqueue_lim is set to 5376 the default value of -1. 5377 5378 rcupdate.rcu_task_enqueue_lim= [KNL] 5379 Set the number of callback queues to use for the 5380 RCU Tasks family of RCU flavors. The default 5381 of -1 allows this to be automatically (and 5382 dynamically) adjusted. This parameter is intended 5383 for use in testing. 5384 5385 rcupdate.rcu_task_ipi_delay= [KNL] 5386 Set time in jiffies during which RCU tasks will 5387 avoid sending IPIs, starting with the beginning 5388 of a given grace period. Setting a large 5389 number avoids disturbing real-time workloads, 5390 but lengthens grace periods. 5391 5392 rcupdate.rcu_task_lazy_lim= [KNL] 5393 Number of callbacks on a given CPU that will 5394 cancel laziness on that CPU. Use -1 to disable 5395 cancellation of laziness, but be advised that 5396 doing so increases the danger of OOM due to 5397 callback flooding. 5398 5399 rcupdate.rcu_task_stall_info= [KNL] 5400 Set initial timeout in jiffies for RCU task stall 5401 informational messages, which give some indication 5402 of the problem for those not patient enough to 5403 wait for ten minutes. Informational messages are 5404 only printed prior to the stall-warning message 5405 for a given grace period. Disable with a value 5406 less than or equal to zero. Defaults to ten 5407 seconds. A change in value does not take effect 5408 until the beginning of the next grace period. 5409 5410 rcupdate.rcu_task_stall_info_mult= [KNL] 5411 Multiplier for time interval between successive 5412 RCU task stall informational messages for a given 5413 RCU tasks grace period. This value is clamped 5414 to one through ten, inclusive. It defaults to 5415 the value three, so that the first informational 5416 message is printed 10 seconds into the grace 5417 period, the second at 40 seconds, the third at 5418 160 seconds, and then the stall warning at 600 5419 seconds would prevent a fourth at 640 seconds. 5420 5421 rcupdate.rcu_task_stall_timeout= [KNL] 5422 Set timeout in jiffies for RCU task stall 5423 warning messages. Disable with a value less 5424 than or equal to zero. Defaults to ten minutes. 5425 A change in value does not take effect until 5426 the beginning of the next grace period. 5427 5428 rcupdate.rcu_tasks_lazy_ms= [KNL] 5429 Set timeout in milliseconds RCU Tasks asynchronous 5430 callback batching for call_rcu_tasks(). 5431 A negative value will take the default. A value 5432 of zero will disable batching. Batching is 5433 always disabled for synchronize_rcu_tasks(). 5434 5435 rcupdate.rcu_tasks_rude_lazy_ms= [KNL] 5436 Set timeout in milliseconds RCU Tasks 5437 Rude asynchronous callback batching for 5438 call_rcu_tasks_rude(). A negative value 5439 will take the default. A value of zero will 5440 disable batching. Batching is always disabled 5441 for synchronize_rcu_tasks_rude(). 5442 5443 rcupdate.rcu_tasks_trace_lazy_ms= [KNL] 5444 Set timeout in milliseconds RCU Tasks 5445 Trace asynchronous callback batching for 5446 call_rcu_tasks_trace(). A negative value 5447 will take the default. A value of zero will 5448 disable batching. Batching is always disabled 5449 for synchronize_rcu_tasks_trace(). 5450 5451 rcupdate.rcu_self_test= [KNL] 5452 Run the RCU early boot self tests 5453 5454 rdinit= [KNL] 5455 Format: <full_path> 5456 Run specified binary instead of /init from the ramdisk, 5457 used for early userspace startup. See initrd. 5458 5459 rdrand= [X86] 5460 force - Override the decision by the kernel to hide the 5461 advertisement of RDRAND support (this affects 5462 certain AMD processors because of buggy BIOS 5463 support, specifically around the suspend/resume 5464 path). 5465 5466 rdt= [HW,X86,RDT] 5467 Turn on/off individual RDT features. List is: 5468 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp, 5469 mba, smba, bmec. 5470 E.g. to turn on cmt and turn off mba use: 5471 rdt=cmt,!mba 5472 5473 reboot= [KNL] 5474 Format (x86 or x86_64): 5475 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \ 5476 [[,]s[mp]#### \ 5477 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \ 5478 [[,]f[orce] 5479 Where reboot_mode is one of warm (soft) or cold (hard) or gpio 5480 (prefix with 'panic_' to set mode for panic 5481 reboot only), 5482 reboot_type is one of bios, acpi, kbd, triple, efi, or pci, 5483 reboot_force is either force or not specified, 5484 reboot_cpu is s[mp]#### with #### being the processor 5485 to be used for rebooting. 5486 5487 refscale.holdoff= [KNL] 5488 Set test-start holdoff period. The purpose of 5489 this parameter is to delay the start of the 5490 test until boot completes in order to avoid 5491 interference. 5492 5493 refscale.lookup_instances= [KNL] 5494 Number of data elements to use for the forms of 5495 SLAB_TYPESAFE_BY_RCU testing. A negative number 5496 is negated and multiplied by nr_cpu_ids, while 5497 zero specifies nr_cpu_ids. 5498 5499 refscale.loops= [KNL] 5500 Set the number of loops over the synchronization 5501 primitive under test. Increasing this number 5502 reduces noise due to loop start/end overhead, 5503 but the default has already reduced the per-pass 5504 noise to a handful of picoseconds on ca. 2020 5505 x86 laptops. 5506 5507 refscale.nreaders= [KNL] 5508 Set number of readers. The default value of -1 5509 selects N, where N is roughly 75% of the number 5510 of CPUs. A value of zero is an interesting choice. 5511 5512 refscale.nruns= [KNL] 5513 Set number of runs, each of which is dumped onto 5514 the console log. 5515 5516 refscale.readdelay= [KNL] 5517 Set the read-side critical-section duration, 5518 measured in microseconds. 5519 5520 refscale.scale_type= [KNL] 5521 Specify the read-protection implementation to test. 5522 5523 refscale.shutdown= [KNL] 5524 Shut down the system at the end of the performance 5525 test. This defaults to 1 (shut it down) when 5526 refscale is built into the kernel and to 0 (leave 5527 it running) when refscale is built as a module. 5528 5529 refscale.verbose= [KNL] 5530 Enable additional printk() statements. 5531 5532 refscale.verbose_batched= [KNL] 5533 Batch the additional printk() statements. If zero 5534 (the default) or negative, print everything. Otherwise, 5535 print every Nth verbose statement, where N is the value 5536 specified. 5537 5538 relax_domain_level= 5539 [KNL, SMP] Set scheduler's default relax_domain_level. 5540 See Documentation/admin-guide/cgroup-v1/cpusets.rst. 5541 5542 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory 5543 Format: <base1>,<size1>[,<base2>,<size2>,...] 5544 Reserve I/O ports or memory so the kernel won't use 5545 them. If <base> is less than 0x10000, the region 5546 is assumed to be I/O ports; otherwise it is memory. 5547 5548 reservetop= [X86-32] 5549 Format: nn[KMG] 5550 Reserves a hole at the top of the kernel virtual 5551 address space. 5552 5553 reset_devices [KNL] Force drivers to reset the underlying device 5554 during initialization. 5555 5556 resume= [SWSUSP] 5557 Specify the partition device for software suspend 5558 Format: 5559 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>} 5560 5561 resume_offset= [SWSUSP] 5562 Specify the offset from the beginning of the partition 5563 given by "resume=" at which the swap header is located, 5564 in <PAGE_SIZE> units (needed only for swap files). 5565 See Documentation/power/swsusp-and-swap-files.rst 5566 5567 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to 5568 read the resume files 5569 5570 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up. 5571 Useful for devices that are detected asynchronously 5572 (e.g. USB and MMC devices). 5573 5574 retain_initrd [RAM] Keep initrd memory after extraction 5575 5576 retbleed= [X86] Control mitigation of RETBleed (Arbitrary 5577 Speculative Code Execution with Return Instructions) 5578 vulnerability. 5579 5580 AMD-based UNRET and IBPB mitigations alone do not stop 5581 sibling threads from influencing the predictions of other 5582 sibling threads. For that reason, STIBP is used on pro- 5583 cessors that support it, and mitigate SMT on processors 5584 that don't. 5585 5586 off - no mitigation 5587 auto - automatically select a migitation 5588 auto,nosmt - automatically select a mitigation, 5589 disabling SMT if necessary for 5590 the full mitigation (only on Zen1 5591 and older without STIBP). 5592 ibpb - On AMD, mitigate short speculation 5593 windows on basic block boundaries too. 5594 Safe, highest perf impact. It also 5595 enables STIBP if present. Not suitable 5596 on Intel. 5597 ibpb,nosmt - Like "ibpb" above but will disable SMT 5598 when STIBP is not available. This is 5599 the alternative for systems which do not 5600 have STIBP. 5601 unret - Force enable untrained return thunks, 5602 only effective on AMD f15h-f17h based 5603 systems. 5604 unret,nosmt - Like unret, but will disable SMT when STIBP 5605 is not available. This is the alternative for 5606 systems which do not have STIBP. 5607 5608 Selecting 'auto' will choose a mitigation method at run 5609 time according to the CPU. 5610 5611 Not specifying this option is equivalent to retbleed=auto. 5612 5613 rfkill.default_state= 5614 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm, 5615 etc. communication is blocked by default. 5616 1 Unblocked. 5617 5618 rfkill.master_switch_mode= 5619 0 The "airplane mode" button does nothing. 5620 1 The "airplane mode" button toggles between everything 5621 blocked and the previous configuration. 5622 2 The "airplane mode" button toggles between everything 5623 blocked and everything unblocked. 5624 5625 rhash_entries= [KNL,NET] 5626 Set number of hash buckets for route cache 5627 5628 ring3mwait=disable 5629 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported 5630 CPUs. 5631 5632 riscv_isa_fallback [RISCV] 5633 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit 5634 falling back to detecting extension support by parsing 5635 "riscv,isa" property on devicetree systems when the 5636 replacement properties are not found. See the Kconfig 5637 entry for RISCV_ISA_FALLBACK. 5638 5639 ro [KNL] Mount root device read-only on boot 5640 5641 rodata= [KNL] 5642 on Mark read-only kernel memory as read-only (default). 5643 off Leave read-only kernel memory writable for debugging. 5644 full Mark read-only kernel memory and aliases as read-only 5645 [arm64] 5646 5647 rockchip.usb_uart 5648 Enable the uart passthrough on the designated usb port 5649 on Rockchip SoCs. When active, the signals of the 5650 debug-uart get routed to the D+ and D- pins of the usb 5651 port and the regular usb controller gets disabled. 5652 5653 root= [KNL] Root filesystem 5654 Usually this a a block device specifier of some kind, 5655 see the early_lookup_bdev comment in 5656 block/early-lookup.c for details. 5657 Alternatively this can be "ram" for the legacy initial 5658 ramdisk, "nfs" and "cifs" for root on a network file 5659 system, or "mtd" and "ubi" for mounting from raw flash. 5660 5661 rootdelay= [KNL] Delay (in seconds) to pause before attempting to 5662 mount the root filesystem 5663 5664 rootflags= [KNL] Set root filesystem mount option string 5665 5666 rootfstype= [KNL] Set root filesystem type 5667 5668 rootwait [KNL] Wait (indefinitely) for root device to show up. 5669 Useful for devices that are detected asynchronously 5670 (e.g. USB and MMC devices). 5671 5672 rootwait= [KNL] Maximum time (in seconds) to wait for root device 5673 to show up before attempting to mount the root 5674 filesystem. 5675 5676 rproc_mem=nn[KMG][@address] 5677 [KNL,ARM,CMA] Remoteproc physical memory block. 5678 Memory area to be used by remote processor image, 5679 managed by CMA. 5680 5681 rw [KNL] Mount root device read-write on boot 5682 5683 S [KNL] Run init in single mode 5684 5685 s390_iommu= [HW,S390] 5686 Set s390 IOTLB flushing mode 5687 strict 5688 With strict flushing every unmap operation will result in 5689 an IOTLB flush. Default is lazy flushing before reuse, 5690 which is faster. 5691 5692 s390_iommu_aperture= [KNL,S390] 5693 Specifies the size of the per device DMA address space 5694 accessible through the DMA and IOMMU APIs as a decimal 5695 factor of the size of main memory. 5696 The default is 1 meaning that one can concurrently use 5697 as many DMA addresses as physical memory is installed, 5698 if supported by hardware, and thus map all of memory 5699 once. With a value of 2 one can map all of memory twice 5700 and so on. As a special case a factor of 0 imposes no 5701 restrictions other than those given by hardware at the 5702 cost of significant additional memory use for tables. 5703 5704 sa1100ir [NET] 5705 See drivers/net/irda/sa1100_ir.c. 5706 5707 sched_verbose [KNL] Enables verbose scheduler debug messages. 5708 5709 schedstats= [KNL,X86] Enable or disable scheduled statistics. 5710 Allowed values are enable and disable. This feature 5711 incurs a small amount of overhead in the scheduler 5712 but is useful for debugging and performance tuning. 5713 5714 sched_thermal_decay_shift= 5715 [KNL, SMP] Set a decay shift for scheduler thermal 5716 pressure signal. Thermal pressure signal follows the 5717 default decay period of other scheduler pelt 5718 signals(usually 32 ms but configurable). Setting 5719 sched_thermal_decay_shift will left shift the decay 5720 period for the thermal pressure signal by the shift 5721 value. 5722 i.e. with the default pelt decay period of 32 ms 5723 sched_thermal_decay_shift thermal pressure decay pr 5724 1 64 ms 5725 2 128 ms 5726 and so on. 5727 Format: integer between 0 and 10 5728 Default is 0. 5729 5730 scftorture.holdoff= [KNL] 5731 Number of seconds to hold off before starting 5732 test. Defaults to zero for module insertion and 5733 to 10 seconds for built-in smp_call_function() 5734 tests. 5735 5736 scftorture.longwait= [KNL] 5737 Request ridiculously long waits randomly selected 5738 up to the chosen limit in seconds. Zero (the 5739 default) disables this feature. Please note 5740 that requesting even small non-zero numbers of 5741 seconds can result in RCU CPU stall warnings, 5742 softlockup complaints, and so on. 5743 5744 scftorture.nthreads= [KNL] 5745 Number of kthreads to spawn to invoke the 5746 smp_call_function() family of functions. 5747 The default of -1 specifies a number of kthreads 5748 equal to the number of CPUs. 5749 5750 scftorture.onoff_holdoff= [KNL] 5751 Number seconds to wait after the start of the 5752 test before initiating CPU-hotplug operations. 5753 5754 scftorture.onoff_interval= [KNL] 5755 Number seconds to wait between successive 5756 CPU-hotplug operations. Specifying zero (which 5757 is the default) disables CPU-hotplug operations. 5758 5759 scftorture.shutdown_secs= [KNL] 5760 The number of seconds following the start of the 5761 test after which to shut down the system. The 5762 default of zero avoids shutting down the system. 5763 Non-zero values are useful for automated tests. 5764 5765 scftorture.stat_interval= [KNL] 5766 The number of seconds between outputting the 5767 current test statistics to the console. A value 5768 of zero disables statistics output. 5769 5770 scftorture.stutter_cpus= [KNL] 5771 The number of jiffies to wait between each change 5772 to the set of CPUs under test. 5773 5774 scftorture.use_cpus_read_lock= [KNL] 5775 Use use_cpus_read_lock() instead of the default 5776 preempt_disable() to disable CPU hotplug 5777 while invoking one of the smp_call_function*() 5778 functions. 5779 5780 scftorture.verbose= [KNL] 5781 Enable additional printk() statements. 5782 5783 scftorture.weight_single= [KNL] 5784 The probability weighting to use for the 5785 smp_call_function_single() function with a zero 5786 "wait" parameter. A value of -1 selects the 5787 default if all other weights are -1. However, 5788 if at least one weight has some other value, a 5789 value of -1 will instead select a weight of zero. 5790 5791 scftorture.weight_single_wait= [KNL] 5792 The probability weighting to use for the 5793 smp_call_function_single() function with a 5794 non-zero "wait" parameter. See weight_single. 5795 5796 scftorture.weight_many= [KNL] 5797 The probability weighting to use for the 5798 smp_call_function_many() function with a zero 5799 "wait" parameter. See weight_single. 5800 Note well that setting a high probability for 5801 this weighting can place serious IPI load 5802 on the system. 5803 5804 scftorture.weight_many_wait= [KNL] 5805 The probability weighting to use for the 5806 smp_call_function_many() function with a 5807 non-zero "wait" parameter. See weight_single 5808 and weight_many. 5809 5810 scftorture.weight_all= [KNL] 5811 The probability weighting to use for the 5812 smp_call_function_all() function with a zero 5813 "wait" parameter. See weight_single and 5814 weight_many. 5815 5816 scftorture.weight_all_wait= [KNL] 5817 The probability weighting to use for the 5818 smp_call_function_all() function with a 5819 non-zero "wait" parameter. See weight_single 5820 and weight_many. 5821 5822 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate 5823 xtime_lock contention on larger systems, and/or RCU lock 5824 contention on all systems with CONFIG_MAXSMP set. 5825 Format: { "0" | "1" } 5826 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1" 5827 1 -- enable. 5828 Note: increases power consumption, thus should only be 5829 enabled if running jitter sensitive (HPC/RT) workloads. 5830 5831 security= [SECURITY] Choose a legacy "major" security module to 5832 enable at boot. This has been deprecated by the 5833 "lsm=" parameter. 5834 5835 selinux= [SELINUX] Disable or enable SELinux at boot time. 5836 Format: { "0" | "1" } 5837 See security/selinux/Kconfig help text. 5838 0 -- disable. 5839 1 -- enable. 5840 Default value is 1. 5841 5842 serialnumber [BUGS=X86-32] 5843 5844 sev=option[,option...] [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst 5845 5846 shapers= [NET] 5847 Maximal number of shapers. 5848 5849 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller 5850 Limit apic dumping. The parameter defines the maximal 5851 number of local apics being dumped. Also it is possible 5852 to set it to "all" by meaning -- no limit here. 5853 Format: { 1 (default) | 2 | ... | all }. 5854 The parameter valid if only apic=debug or 5855 apic=verbose is specified. 5856 Example: apic=debug show_lapic=all 5857 5858 simeth= [IA-64] 5859 simscsi= 5860 5861 slram= [HW,MTD] 5862 5863 slab_merge [MM] 5864 Enable merging of slabs with similar size when the 5865 kernel is built without CONFIG_SLAB_MERGE_DEFAULT. 5866 5867 slab_nomerge [MM] 5868 Disable merging of slabs with similar size. May be 5869 necessary if there is some reason to distinguish 5870 allocs to different slabs, especially in hardened 5871 environments where the risk of heap overflows and 5872 layout control by attackers can usually be 5873 frustrated by disabling merging. This will reduce 5874 most of the exposure of a heap attack to a single 5875 cache (risks via metadata attacks are mostly 5876 unchanged). Debug options disable merging on their 5877 own. 5878 For more information see Documentation/mm/slub.rst. 5879 5880 slab_max_order= [MM, SLAB] 5881 Determines the maximum allowed order for slabs. 5882 A high setting may cause OOMs due to memory 5883 fragmentation. Defaults to 1 for systems with 5884 more than 32MB of RAM, 0 otherwise. 5885 5886 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB] 5887 Enabling slub_debug allows one to determine the 5888 culprit if slab objects become corrupted. Enabling 5889 slub_debug can create guard zones around objects and 5890 may poison objects when not in use. Also tracks the 5891 last alloc / free. For more information see 5892 Documentation/mm/slub.rst. 5893 5894 slub_max_order= [MM, SLUB] 5895 Determines the maximum allowed order for slabs. 5896 A high setting may cause OOMs due to memory 5897 fragmentation. For more information see 5898 Documentation/mm/slub.rst. 5899 5900 slub_min_objects= [MM, SLUB] 5901 The minimum number of objects per slab. SLUB will 5902 increase the slab order up to slub_max_order to 5903 generate a sufficiently large slab able to contain 5904 the number of objects indicated. The higher the number 5905 of objects the smaller the overhead of tracking slabs 5906 and the less frequently locks need to be acquired. 5907 For more information see Documentation/mm/slub.rst. 5908 5909 slub_min_order= [MM, SLUB] 5910 Determines the minimum page order for slabs. Must be 5911 lower than slub_max_order. 5912 For more information see Documentation/mm/slub.rst. 5913 5914 slub_merge [MM, SLUB] 5915 Same with slab_merge. 5916 5917 slub_nomerge [MM, SLUB] 5918 Same with slab_nomerge. This is supported for legacy. 5919 See slab_nomerge for more information. 5920 5921 smart2= [HW] 5922 Format: <io1>[,<io2>[,...,<io8>]] 5923 5924 smp.csd_lock_timeout= [KNL] 5925 Specify the period of time in milliseconds 5926 that smp_call_function() and friends will wait 5927 for a CPU to release the CSD lock. This is 5928 useful when diagnosing bugs involving CPUs 5929 disabling interrupts for extended periods 5930 of time. Defaults to 5,000 milliseconds, and 5931 setting a value of zero disables this feature. 5932 This feature may be more efficiently disabled 5933 using the csdlock_debug- kernel parameter. 5934 5935 smp.panic_on_ipistall= [KNL] 5936 If a csd_lock_timeout extends for more than 5937 the specified number of milliseconds, panic the 5938 system. By default, let CSD-lock acquisition 5939 take as long as they take. Specifying 300,000 5940 for this value provides a 5-minute timeout. 5941 5942 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices 5943 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port 5944 smsc-ircc2.ircc_sir= [HW] SIR base I/O port 5945 smsc-ircc2.ircc_fir= [HW] FIR base I/O port 5946 smsc-ircc2.ircc_irq= [HW] IRQ line 5947 smsc-ircc2.ircc_dma= [HW] DMA channel 5948 smsc-ircc2.ircc_transceiver= [HW] Transceiver type: 5949 0: Toshiba Satellite 1800 (GP data pin select) 5950 1: Fast pin select (default) 5951 2: ATC IRMode 5952 5953 smt= [KNL,MIPS,S390] Set the maximum number of threads (logical 5954 CPUs) to use per physical CPU on systems capable of 5955 symmetric multithreading (SMT). Will be capped to the 5956 actual hardware limit. 5957 Format: <integer> 5958 Default: -1 (no limit) 5959 5960 softlockup_panic= 5961 [KNL] Should the soft-lockup detector generate panics. 5962 Format: 0 | 1 5963 5964 A value of 1 instructs the soft-lockup detector 5965 to panic the machine when a soft-lockup occurs. It is 5966 also controlled by the kernel.softlockup_panic sysctl 5967 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the 5968 respective build-time switch to that functionality. 5969 5970 softlockup_all_cpu_backtrace= 5971 [KNL] Should the soft-lockup detector generate 5972 backtraces on all cpus. 5973 Format: 0 | 1 5974 5975 sonypi.*= [HW] Sony Programmable I/O Control Device driver 5976 See Documentation/admin-guide/laptops/sonypi.rst 5977 5978 spectre_v2= [X86] Control mitigation of Spectre variant 2 5979 (indirect branch speculation) vulnerability. 5980 The default operation protects the kernel from 5981 user space attacks. 5982 5983 on - unconditionally enable, implies 5984 spectre_v2_user=on 5985 off - unconditionally disable, implies 5986 spectre_v2_user=off 5987 auto - kernel detects whether your CPU model is 5988 vulnerable 5989 5990 Selecting 'on' will, and 'auto' may, choose a 5991 mitigation method at run time according to the 5992 CPU, the available microcode, the setting of the 5993 CONFIG_RETPOLINE configuration option, and the 5994 compiler with which the kernel was built. 5995 5996 Selecting 'on' will also enable the mitigation 5997 against user space to user space task attacks. 5998 5999 Selecting 'off' will disable both the kernel and 6000 the user space protections. 6001 6002 Specific mitigations can also be selected manually: 6003 6004 retpoline - replace indirect branches 6005 retpoline,generic - Retpolines 6006 retpoline,lfence - LFENCE; indirect branch 6007 retpoline,amd - alias for retpoline,lfence 6008 eibrs - Enhanced/Auto IBRS 6009 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines 6010 eibrs,lfence - Enhanced/Auto IBRS + LFENCE 6011 ibrs - use IBRS to protect kernel 6012 6013 Not specifying this option is equivalent to 6014 spectre_v2=auto. 6015 6016 spectre_v2_user= 6017 [X86] Control mitigation of Spectre variant 2 6018 (indirect branch speculation) vulnerability between 6019 user space tasks 6020 6021 on - Unconditionally enable mitigations. Is 6022 enforced by spectre_v2=on 6023 6024 off - Unconditionally disable mitigations. Is 6025 enforced by spectre_v2=off 6026 6027 prctl - Indirect branch speculation is enabled, 6028 but mitigation can be enabled via prctl 6029 per thread. The mitigation control state 6030 is inherited on fork. 6031 6032 prctl,ibpb 6033 - Like "prctl" above, but only STIBP is 6034 controlled per thread. IBPB is issued 6035 always when switching between different user 6036 space processes. 6037 6038 seccomp 6039 - Same as "prctl" above, but all seccomp 6040 threads will enable the mitigation unless 6041 they explicitly opt out. 6042 6043 seccomp,ibpb 6044 - Like "seccomp" above, but only STIBP is 6045 controlled per thread. IBPB is issued 6046 always when switching between different 6047 user space processes. 6048 6049 auto - Kernel selects the mitigation depending on 6050 the available CPU features and vulnerability. 6051 6052 Default mitigation: "prctl" 6053 6054 Not specifying this option is equivalent to 6055 spectre_v2_user=auto. 6056 6057 spec_rstack_overflow= 6058 [X86] Control RAS overflow mitigation on AMD Zen CPUs 6059 6060 off - Disable mitigation 6061 microcode - Enable microcode mitigation only 6062 safe-ret - Enable sw-only safe RET mitigation (default) 6063 ibpb - Enable mitigation by issuing IBPB on 6064 kernel entry 6065 ibpb-vmexit - Issue IBPB only on VMEXIT 6066 (cloud-specific mitigation) 6067 6068 spec_store_bypass_disable= 6069 [HW] Control Speculative Store Bypass (SSB) Disable mitigation 6070 (Speculative Store Bypass vulnerability) 6071 6072 Certain CPUs are vulnerable to an exploit against a 6073 a common industry wide performance optimization known 6074 as "Speculative Store Bypass" in which recent stores 6075 to the same memory location may not be observed by 6076 later loads during speculative execution. The idea 6077 is that such stores are unlikely and that they can 6078 be detected prior to instruction retirement at the 6079 end of a particular speculation execution window. 6080 6081 In vulnerable processors, the speculatively forwarded 6082 store can be used in a cache side channel attack, for 6083 example to read memory to which the attacker does not 6084 directly have access (e.g. inside sandboxed code). 6085 6086 This parameter controls whether the Speculative Store 6087 Bypass optimization is used. 6088 6089 On x86 the options are: 6090 6091 on - Unconditionally disable Speculative Store Bypass 6092 off - Unconditionally enable Speculative Store Bypass 6093 auto - Kernel detects whether the CPU model contains an 6094 implementation of Speculative Store Bypass and 6095 picks the most appropriate mitigation. If the 6096 CPU is not vulnerable, "off" is selected. If the 6097 CPU is vulnerable the default mitigation is 6098 architecture and Kconfig dependent. See below. 6099 prctl - Control Speculative Store Bypass per thread 6100 via prctl. Speculative Store Bypass is enabled 6101 for a process by default. The state of the control 6102 is inherited on fork. 6103 seccomp - Same as "prctl" above, but all seccomp threads 6104 will disable SSB unless they explicitly opt out. 6105 6106 Default mitigations: 6107 X86: "prctl" 6108 6109 On powerpc the options are: 6110 6111 on,auto - On Power8 and Power9 insert a store-forwarding 6112 barrier on kernel entry and exit. On Power7 6113 perform a software flush on kernel entry and 6114 exit. 6115 off - No action. 6116 6117 Not specifying this option is equivalent to 6118 spec_store_bypass_disable=auto. 6119 6120 spia_io_base= [HW,MTD] 6121 spia_fio_base= 6122 spia_pedr= 6123 spia_peddr= 6124 6125 split_lock_detect= 6126 [X86] Enable split lock detection or bus lock detection 6127 6128 When enabled (and if hardware support is present), atomic 6129 instructions that access data across cache line 6130 boundaries will result in an alignment check exception 6131 for split lock detection or a debug exception for 6132 bus lock detection. 6133 6134 off - not enabled 6135 6136 warn - the kernel will emit rate-limited warnings 6137 about applications triggering the #AC 6138 exception or the #DB exception. This mode is 6139 the default on CPUs that support split lock 6140 detection or bus lock detection. Default 6141 behavior is by #AC if both features are 6142 enabled in hardware. 6143 6144 fatal - the kernel will send SIGBUS to applications 6145 that trigger the #AC exception or the #DB 6146 exception. Default behavior is by #AC if 6147 both features are enabled in hardware. 6148 6149 ratelimit:N - 6150 Set system wide rate limit to N bus locks 6151 per second for bus lock detection. 6152 0 < N <= 1000. 6153 6154 N/A for split lock detection. 6155 6156 6157 If an #AC exception is hit in the kernel or in 6158 firmware (i.e. not while executing in user mode) 6159 the kernel will oops in either "warn" or "fatal" 6160 mode. 6161 6162 #DB exception for bus lock is triggered only when 6163 CPL > 0. 6164 6165 srbds= [X86,INTEL] 6166 Control the Special Register Buffer Data Sampling 6167 (SRBDS) mitigation. 6168 6169 Certain CPUs are vulnerable to an MDS-like 6170 exploit which can leak bits from the random 6171 number generator. 6172 6173 By default, this issue is mitigated by 6174 microcode. However, the microcode fix can cause 6175 the RDRAND and RDSEED instructions to become 6176 much slower. Among other effects, this will 6177 result in reduced throughput from /dev/urandom. 6178 6179 The microcode mitigation can be disabled with 6180 the following option: 6181 6182 off: Disable mitigation and remove 6183 performance impact to RDRAND and RDSEED 6184 6185 srcutree.big_cpu_lim [KNL] 6186 Specifies the number of CPUs constituting a 6187 large system, such that srcu_struct structures 6188 should immediately allocate an srcu_node array. 6189 This kernel-boot parameter defaults to 128, 6190 but takes effect only when the low-order four 6191 bits of srcutree.convert_to_big is equal to 3 6192 (decide at boot). 6193 6194 srcutree.convert_to_big [KNL] 6195 Specifies under what conditions an SRCU tree 6196 srcu_struct structure will be converted to big 6197 form, that is, with an rcu_node tree: 6198 6199 0: Never. 6200 1: At init_srcu_struct() time. 6201 2: When rcutorture decides to. 6202 3: Decide at boot time (default). 6203 0x1X: Above plus if high contention. 6204 6205 Either way, the srcu_node tree will be sized based 6206 on the actual runtime number of CPUs (nr_cpu_ids) 6207 instead of the compile-time CONFIG_NR_CPUS. 6208 6209 srcutree.counter_wrap_check [KNL] 6210 Specifies how frequently to check for 6211 grace-period sequence counter wrap for the 6212 srcu_data structure's ->srcu_gp_seq_needed field. 6213 The greater the number of bits set in this kernel 6214 parameter, the less frequently counter wrap will 6215 be checked for. Note that the bottom two bits 6216 are ignored. 6217 6218 srcutree.exp_holdoff [KNL] 6219 Specifies how many nanoseconds must elapse 6220 since the end of the last SRCU grace period for 6221 a given srcu_struct until the next normal SRCU 6222 grace period will be considered for automatic 6223 expediting. Set to zero to disable automatic 6224 expediting. 6225 6226 srcutree.srcu_max_nodelay [KNL] 6227 Specifies the number of no-delay instances 6228 per jiffy for which the SRCU grace period 6229 worker thread will be rescheduled with zero 6230 delay. Beyond this limit, worker thread will 6231 be rescheduled with a sleep delay of one jiffy. 6232 6233 srcutree.srcu_max_nodelay_phase [KNL] 6234 Specifies the per-grace-period phase, number of 6235 non-sleeping polls of readers. Beyond this limit, 6236 grace period worker thread will be rescheduled 6237 with a sleep delay of one jiffy, between each 6238 rescan of the readers, for a grace period phase. 6239 6240 srcutree.srcu_retry_check_delay [KNL] 6241 Specifies number of microseconds of non-sleeping 6242 delay between each non-sleeping poll of readers. 6243 6244 srcutree.small_contention_lim [KNL] 6245 Specifies the number of update-side contention 6246 events per jiffy will be tolerated before 6247 initiating a conversion of an srcu_struct 6248 structure to big form. Note that the value of 6249 srcutree.convert_to_big must have the 0x10 bit 6250 set for contention-based conversions to occur. 6251 6252 ssbd= [ARM64,HW] 6253 Speculative Store Bypass Disable control 6254 6255 On CPUs that are vulnerable to the Speculative 6256 Store Bypass vulnerability and offer a 6257 firmware based mitigation, this parameter 6258 indicates how the mitigation should be used: 6259 6260 force-on: Unconditionally enable mitigation for 6261 for both kernel and userspace 6262 force-off: Unconditionally disable mitigation for 6263 for both kernel and userspace 6264 kernel: Always enable mitigation in the 6265 kernel, and offer a prctl interface 6266 to allow userspace to register its 6267 interest in being mitigated too. 6268 6269 stack_guard_gap= [MM] 6270 override the default stack gap protection. The value 6271 is in page units and it defines how many pages prior 6272 to (for stacks growing down) resp. after (for stacks 6273 growing up) the main stack are reserved for no other 6274 mapping. Default value is 256 pages. 6275 6276 stack_depot_disable= [KNL] 6277 Setting this to true through kernel command line will 6278 disable the stack depot thereby saving the static memory 6279 consumed by the stack hash table. By default this is set 6280 to false. 6281 6282 stacktrace [FTRACE] 6283 Enabled the stack tracer on boot up. 6284 6285 stacktrace_filter=[function-list] 6286 [FTRACE] Limit the functions that the stack tracer 6287 will trace at boot up. function-list is a comma-separated 6288 list of functions. This list can be changed at run 6289 time by the stack_trace_filter file in the debugfs 6290 tracing directory. Note, this enables stack tracing 6291 and the stacktrace above is not needed. 6292 6293 sti= [PARISC,HW] 6294 Format: <num> 6295 Set the STI (builtin display/keyboard on the HP-PARISC 6296 machines) console (graphic card) which should be used 6297 as the initial boot-console. 6298 See also comment in drivers/video/console/sticore.c. 6299 6300 sti_font= [HW] 6301 See comment in drivers/video/console/sticore.c. 6302 6303 stifb= [HW] 6304 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]] 6305 6306 strict_sas_size= 6307 [X86] 6308 Format: <bool> 6309 Enable or disable strict sigaltstack size checks 6310 against the required signal frame size which 6311 depends on the supported FPU features. This can 6312 be used to filter out binaries which have 6313 not yet been made aware of AT_MINSIGSTKSZ. 6314 6315 stress_hpt [PPC] 6316 Limits the number of kernel HPT entries in the hash 6317 page table to increase the rate of hash page table 6318 faults on kernel addresses. 6319 6320 stress_slb [PPC] 6321 Limits the number of kernel SLB entries, and flushes 6322 them frequently to increase the rate of SLB faults 6323 on kernel addresses. 6324 6325 sunrpc.min_resvport= 6326 sunrpc.max_resvport= 6327 [NFS,SUNRPC] 6328 SunRPC servers often require that client requests 6329 originate from a privileged port (i.e. a port in the 6330 range 0 < portnr < 1024). 6331 An administrator who wishes to reserve some of these 6332 ports for other uses may adjust the range that the 6333 kernel's sunrpc client considers to be privileged 6334 using these two parameters to set the minimum and 6335 maximum port values. 6336 6337 sunrpc.svc_rpc_per_connection_limit= 6338 [NFS,SUNRPC] 6339 Limit the number of requests that the server will 6340 process in parallel from a single connection. 6341 The default value is 0 (no limit). 6342 6343 sunrpc.pool_mode= 6344 [NFS] 6345 Control how the NFS server code allocates CPUs to 6346 service thread pools. Depending on how many NICs 6347 you have and where their interrupts are bound, this 6348 option will affect which CPUs will do NFS serving. 6349 Note: this parameter cannot be changed while the 6350 NFS server is running. 6351 6352 auto the server chooses an appropriate mode 6353 automatically using heuristics 6354 global a single global pool contains all CPUs 6355 percpu one pool for each CPU 6356 pernode one pool for each NUMA node (equivalent 6357 to global on non-NUMA machines) 6358 6359 sunrpc.tcp_slot_table_entries= 6360 sunrpc.udp_slot_table_entries= 6361 [NFS,SUNRPC] 6362 Sets the upper limit on the number of simultaneous 6363 RPC calls that can be sent from the client to a 6364 server. Increasing these values may allow you to 6365 improve throughput, but will also increase the 6366 amount of memory reserved for use by the client. 6367 6368 suspend.pm_test_delay= 6369 [SUSPEND] 6370 Sets the number of seconds to remain in a suspend test 6371 mode before resuming the system (see 6372 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 6373 is set. Default value is 5. 6374 6375 svm= [PPC] 6376 Format: { on | off | y | n | 1 | 0 } 6377 This parameter controls use of the Protected 6378 Execution Facility on pSeries. 6379 6380 swiotlb= [ARM,IA-64,PPC,MIPS,X86] 6381 Format: { <int> [,<int>] | force | noforce } 6382 <int> -- Number of I/O TLB slabs 6383 <int> -- Second integer after comma. Number of swiotlb 6384 areas with their own lock. Will be rounded up 6385 to a power of 2. 6386 force -- force using of bounce buffers even if they 6387 wouldn't be automatically used by the kernel 6388 noforce -- Never use bounce buffers (for debugging) 6389 6390 switches= [HW,M68k] 6391 6392 sysctl.*= [KNL] 6393 Set a sysctl parameter, right before loading the init 6394 process, as if the value was written to the respective 6395 /proc/sys/... file. Both '.' and '/' are recognized as 6396 separators. Unrecognized parameters and invalid values 6397 are reported in the kernel log. Sysctls registered 6398 later by a loaded module cannot be set this way. 6399 Example: sysctl.vm.swappiness=40 6400 6401 sysrq_always_enabled 6402 [KNL] 6403 Ignore sysrq setting - this boot parameter will 6404 neutralize any effect of /proc/sys/kernel/sysrq. 6405 Useful for debugging. 6406 6407 tcpmhash_entries= [KNL,NET] 6408 Set the number of tcp_metrics_hash slots. 6409 Default value is 8192 or 16384 depending on total 6410 ram pages. This is used to specify the TCP metrics 6411 cache size. See Documentation/networking/ip-sysctl.rst 6412 "tcp_no_metrics_save" section for more details. 6413 6414 tdfx= [HW,DRM] 6415 6416 test_suspend= [SUSPEND] 6417 Format: { "mem" | "standby" | "freeze" }[,N] 6418 Specify "mem" (for Suspend-to-RAM) or "standby" (for 6419 standby suspend) or "freeze" (for suspend type freeze) 6420 as the system sleep state during system startup with 6421 the optional capability to repeat N number of times. 6422 The system is woken from this state using a 6423 wakeup-capable RTC alarm. 6424 6425 thash_entries= [KNL,NET] 6426 Set number of hash buckets for TCP connection 6427 6428 thermal.act= [HW,ACPI] 6429 -1: disable all active trip points in all thermal zones 6430 <degrees C>: override all lowest active trip points 6431 6432 thermal.crt= [HW,ACPI] 6433 -1: disable all critical trip points in all thermal zones 6434 <degrees C>: override all critical trip points 6435 6436 thermal.off= [HW,ACPI] 6437 1: disable ACPI thermal control 6438 6439 thermal.psv= [HW,ACPI] 6440 -1: disable all passive trip points 6441 <degrees C>: override all passive trip points to this 6442 value 6443 6444 thermal.tzp= [HW,ACPI] 6445 Specify global default ACPI thermal zone polling rate 6446 <deci-seconds>: poll all this frequency 6447 0: no polling (default) 6448 6449 threadirqs [KNL] 6450 Force threading of all interrupt handlers except those 6451 marked explicitly IRQF_NO_THREAD. 6452 6453 topology= [S390] 6454 Format: {off | on} 6455 Specify if the kernel should make use of the cpu 6456 topology information if the hardware supports this. 6457 The scheduler will make use of this information and 6458 e.g. base its process migration decisions on it. 6459 Default is on. 6460 6461 topology_updates= [KNL, PPC, NUMA] 6462 Format: {off} 6463 Specify if the kernel should ignore (off) 6464 topology updates sent by the hypervisor to this 6465 LPAR. 6466 6467 torture.disable_onoff_at_boot= [KNL] 6468 Prevent the CPU-hotplug component of torturing 6469 until after init has spawned. 6470 6471 torture.ftrace_dump_at_shutdown= [KNL] 6472 Dump the ftrace buffer at torture-test shutdown, 6473 even if there were no errors. This can be a 6474 very costly operation when many torture tests 6475 are running concurrently, especially on systems 6476 with rotating-rust storage. 6477 6478 torture.verbose_sleep_frequency= [KNL] 6479 Specifies how many verbose printk()s should be 6480 emitted between each sleep. The default of zero 6481 disables verbose-printk() sleeping. 6482 6483 torture.verbose_sleep_duration= [KNL] 6484 Duration of each verbose-printk() sleep in jiffies. 6485 6486 tp720= [HW,PS2] 6487 6488 tpm_suspend_pcr=[HW,TPM] 6489 Format: integer pcr id 6490 Specify that at suspend time, the tpm driver 6491 should extend the specified pcr with zeros, 6492 as a workaround for some chips which fail to 6493 flush the last written pcr on TPM_SaveState. 6494 This will guarantee that all the other pcrs 6495 are saved. 6496 6497 tpm_tis.interrupts= [HW,TPM] 6498 Enable interrupts for the MMIO based physical layer 6499 for the FIFO interface. By default it is set to false 6500 (0). For more information about TPM hardware interfaces 6501 defined by Trusted Computing Group (TCG) see 6502 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/ 6503 6504 tp_printk [FTRACE] 6505 Have the tracepoints sent to printk as well as the 6506 tracing ring buffer. This is useful for early boot up 6507 where the system hangs or reboots and does not give the 6508 option for reading the tracing buffer or performing a 6509 ftrace_dump_on_oops. 6510 6511 To turn off having tracepoints sent to printk, 6512 echo 0 > /proc/sys/kernel/tracepoint_printk 6513 Note, echoing 1 into this file without the 6514 tracepoint_printk kernel cmdline option has no effect. 6515 6516 The tp_printk_stop_on_boot (see below) can also be used 6517 to stop the printing of events to console at 6518 late_initcall_sync. 6519 6520 ** CAUTION ** 6521 6522 Having tracepoints sent to printk() and activating high 6523 frequency tracepoints such as irq or sched, can cause 6524 the system to live lock. 6525 6526 tp_printk_stop_on_boot [FTRACE] 6527 When tp_printk (above) is set, it can cause a lot of noise 6528 on the console. It may be useful to only include the 6529 printing of events during boot up, as user space may 6530 make the system inoperable. 6531 6532 This command line option will stop the printing of events 6533 to console at the late_initcall_sync() time frame. 6534 6535 trace_buf_size=nn[KMG] 6536 [FTRACE] will set tracing buffer size on each cpu. 6537 6538 trace_clock= [FTRACE] Set the clock used for tracing events 6539 at boot up. 6540 local - Use the per CPU time stamp counter 6541 (converted into nanoseconds). Fast, but 6542 depending on the architecture, may not be 6543 in sync between CPUs. 6544 global - Event time stamps are synchronize across 6545 CPUs. May be slower than the local clock, 6546 but better for some race conditions. 6547 counter - Simple counting of events (1, 2, ..) 6548 note, some counts may be skipped due to the 6549 infrastructure grabbing the clock more than 6550 once per event. 6551 uptime - Use jiffies as the time stamp. 6552 perf - Use the same clock that perf uses. 6553 mono - Use ktime_get_mono_fast_ns() for time stamps. 6554 mono_raw - Use ktime_get_raw_fast_ns() for time 6555 stamps. 6556 boot - Use ktime_get_boot_fast_ns() for time stamps. 6557 Architectures may add more clocks. See 6558 Documentation/trace/ftrace.rst for more details. 6559 6560 trace_event=[event-list] 6561 [FTRACE] Set and start specified trace events in order 6562 to facilitate early boot debugging. The event-list is a 6563 comma-separated list of trace events to enable. See 6564 also Documentation/trace/events.rst 6565 6566 trace_instance=[instance-info] 6567 [FTRACE] Create a ring buffer instance early in boot up. 6568 This will be listed in: 6569 6570 /sys/kernel/tracing/instances 6571 6572 Events can be enabled at the time the instance is created 6573 via: 6574 6575 trace_instance=<name>,<system1>:<event1>,<system2>:<event2> 6576 6577 Note, the "<system*>:" portion is optional if the event is 6578 unique. 6579 6580 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall 6581 6582 will enable the "sched_switch" event (note, the "sched:" is optional, and 6583 the same thing would happen if it was left off). The irq_handler_entry 6584 event, and all events under the "initcall" system. 6585 6586 trace_options=[option-list] 6587 [FTRACE] Enable or disable tracer options at boot. 6588 The option-list is a comma delimited list of options 6589 that can be enabled or disabled just as if you were 6590 to echo the option name into 6591 6592 /sys/kernel/tracing/trace_options 6593 6594 For example, to enable stacktrace option (to dump the 6595 stack trace of each event), add to the command line: 6596 6597 trace_options=stacktrace 6598 6599 See also Documentation/trace/ftrace.rst "trace options" 6600 section. 6601 6602 trace_trigger=[trigger-list] 6603 [FTRACE] Add a event trigger on specific events. 6604 Set a trigger on top of a specific event, with an optional 6605 filter. 6606 6607 The format is is "trace_trigger=<event>.<trigger>[ if <filter>],..." 6608 Where more than one trigger may be specified that are comma deliminated. 6609 6610 For example: 6611 6612 trace_trigger="sched_switch.stacktrace if prev_state == 2" 6613 6614 The above will enable the "stacktrace" trigger on the "sched_switch" 6615 event but only trigger it if the "prev_state" of the "sched_switch" 6616 event is "2" (TASK_UNINTERUPTIBLE). 6617 6618 See also "Event triggers" in Documentation/trace/events.rst 6619 6620 6621 traceoff_on_warning 6622 [FTRACE] enable this option to disable tracing when a 6623 warning is hit. This turns off "tracing_on". Tracing can 6624 be enabled again by echoing '1' into the "tracing_on" 6625 file located in /sys/kernel/tracing/ 6626 6627 This option is useful, as it disables the trace before 6628 the WARNING dump is called, which prevents the trace to 6629 be filled with content caused by the warning output. 6630 6631 This option can also be set at run time via the sysctl 6632 option: kernel/traceoff_on_warning 6633 6634 transparent_hugepage= 6635 [KNL] 6636 Format: [always|madvise|never] 6637 Can be used to control the default behavior of the system 6638 with respect to transparent hugepages. 6639 See Documentation/admin-guide/mm/transhuge.rst 6640 for more details. 6641 6642 trusted.source= [KEYS] 6643 Format: <string> 6644 This parameter identifies the trust source as a backend 6645 for trusted keys implementation. Supported trust 6646 sources: 6647 - "tpm" 6648 - "tee" 6649 - "caam" 6650 If not specified then it defaults to iterating through 6651 the trust source list starting with TPM and assigns the 6652 first trust source as a backend which is initialized 6653 successfully during iteration. 6654 6655 trusted.rng= [KEYS] 6656 Format: <string> 6657 The RNG used to generate key material for trusted keys. 6658 Can be one of: 6659 - "kernel" 6660 - the same value as trusted.source: "tpm" or "tee" 6661 - "default" 6662 If not specified, "default" is used. In this case, 6663 the RNG's choice is left to each individual trust source. 6664 6665 tsc= Disable clocksource stability checks for TSC. 6666 Format: <string> 6667 [x86] reliable: mark tsc clocksource as reliable, this 6668 disables clocksource verification at runtime, as well 6669 as the stability checks done at bootup. Used to enable 6670 high-resolution timer mode on older hardware, and in 6671 virtualized environment. 6672 [x86] noirqtime: Do not use TSC to do irq accounting. 6673 Used to run time disable IRQ_TIME_ACCOUNTING on any 6674 platforms where RDTSC is slow and this accounting 6675 can add overhead. 6676 [x86] unstable: mark the TSC clocksource as unstable, this 6677 marks the TSC unconditionally unstable at bootup and 6678 avoids any further wobbles once the TSC watchdog notices. 6679 [x86] nowatchdog: disable clocksource watchdog. Used 6680 in situations with strict latency requirements (where 6681 interruptions from clocksource watchdog are not 6682 acceptable). 6683 [x86] recalibrate: force recalibration against a HW timer 6684 (HPET or PM timer) on systems whose TSC frequency was 6685 obtained from HW or FW using either an MSR or CPUID(0x15). 6686 Warn if the difference is more than 500 ppm. 6687 [x86] watchdog: Use TSC as the watchdog clocksource with 6688 which to check other HW timers (HPET or PM timer), but 6689 only on systems where TSC has been deemed trustworthy. 6690 This will be suppressed by an earlier tsc=nowatchdog and 6691 can be overridden by a later tsc=nowatchdog. A console 6692 message will flag any such suppression or overriding. 6693 6694 tsc_early_khz= [X86] Skip early TSC calibration and use the given 6695 value instead. Useful when the early TSC frequency discovery 6696 procedure is not reliable, such as on overclocked systems 6697 with CPUID.16h support and partial CPUID.15h support. 6698 Format: <unsigned int> 6699 6700 tsx= [X86] Control Transactional Synchronization 6701 Extensions (TSX) feature in Intel processors that 6702 support TSX control. 6703 6704 This parameter controls the TSX feature. The options are: 6705 6706 on - Enable TSX on the system. Although there are 6707 mitigations for all known security vulnerabilities, 6708 TSX has been known to be an accelerator for 6709 several previous speculation-related CVEs, and 6710 so there may be unknown security risks associated 6711 with leaving it enabled. 6712 6713 off - Disable TSX on the system. (Note that this 6714 option takes effect only on newer CPUs which are 6715 not vulnerable to MDS, i.e., have 6716 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get 6717 the new IA32_TSX_CTRL MSR through a microcode 6718 update. This new MSR allows for the reliable 6719 deactivation of the TSX functionality.) 6720 6721 auto - Disable TSX if X86_BUG_TAA is present, 6722 otherwise enable TSX on the system. 6723 6724 Not specifying this option is equivalent to tsx=off. 6725 6726 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 6727 for more details. 6728 6729 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async 6730 Abort (TAA) vulnerability. 6731 6732 Similar to Micro-architectural Data Sampling (MDS) 6733 certain CPUs that support Transactional 6734 Synchronization Extensions (TSX) are vulnerable to an 6735 exploit against CPU internal buffers which can forward 6736 information to a disclosure gadget under certain 6737 conditions. 6738 6739 In vulnerable processors, the speculatively forwarded 6740 data can be used in a cache side channel attack, to 6741 access data to which the attacker does not have direct 6742 access. 6743 6744 This parameter controls the TAA mitigation. The 6745 options are: 6746 6747 full - Enable TAA mitigation on vulnerable CPUs 6748 if TSX is enabled. 6749 6750 full,nosmt - Enable TAA mitigation and disable SMT on 6751 vulnerable CPUs. If TSX is disabled, SMT 6752 is not disabled because CPU is not 6753 vulnerable to cross-thread TAA attacks. 6754 off - Unconditionally disable TAA mitigation 6755 6756 On MDS-affected machines, tsx_async_abort=off can be 6757 prevented by an active MDS mitigation as both vulnerabilities 6758 are mitigated with the same mechanism so in order to disable 6759 this mitigation, you need to specify mds=off too. 6760 6761 Not specifying this option is equivalent to 6762 tsx_async_abort=full. On CPUs which are MDS affected 6763 and deploy MDS mitigation, TAA mitigation is not 6764 required and doesn't provide any additional 6765 mitigation. 6766 6767 For details see: 6768 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 6769 6770 turbografx.map[2|3]= [HW,JOY] 6771 TurboGraFX parallel port interface 6772 Format: 6773 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7> 6774 See also Documentation/input/devices/joystick-parport.rst 6775 6776 udbg-immortal [PPC] When debugging early kernel crashes that 6777 happen after console_init() and before a proper 6778 console driver takes over, this boot options might 6779 help "seeing" what's going on. 6780 6781 uhash_entries= [KNL,NET] 6782 Set number of hash buckets for UDP/UDP-Lite connections 6783 6784 uhci-hcd.ignore_oc= 6785 [USB] Ignore overcurrent events (default N). 6786 Some badly-designed motherboards generate lots of 6787 bogus events, for ports that aren't wired to 6788 anything. Set this parameter to avoid log spamming. 6789 Note that genuine overcurrent events won't be 6790 reported either. 6791 6792 unknown_nmi_panic 6793 [X86] Cause panic on unknown NMI. 6794 6795 unwind_debug [X86-64] 6796 Enable unwinder debug output. This can be 6797 useful for debugging certain unwinder error 6798 conditions, including corrupt stacks and 6799 bad/missing unwinder metadata. 6800 6801 usbcore.authorized_default= 6802 [USB] Default USB device authorization: 6803 (default -1 = authorized (same as 1), 6804 0 = not authorized, 1 = authorized, 2 = authorized 6805 if device connected to internal port) 6806 6807 usbcore.autosuspend= 6808 [USB] The autosuspend time delay (in seconds) used 6809 for newly-detected USB devices (default 2). This 6810 is the time required before an idle device will be 6811 autosuspended. Devices for which the delay is set 6812 to a negative value won't be autosuspended at all. 6813 6814 usbcore.usbfs_snoop= 6815 [USB] Set to log all usbfs traffic (default 0 = off). 6816 6817 usbcore.usbfs_snoop_max= 6818 [USB] Maximum number of bytes to snoop in each URB 6819 (default = 65536). 6820 6821 usbcore.blinkenlights= 6822 [USB] Set to cycle leds on hubs (default 0 = off). 6823 6824 usbcore.old_scheme_first= 6825 [USB] Start with the old device initialization 6826 scheme (default 0 = off). 6827 6828 usbcore.usbfs_memory_mb= 6829 [USB] Memory limit (in MB) for buffers allocated by 6830 usbfs (default = 16, 0 = max = 2047). 6831 6832 usbcore.use_both_schemes= 6833 [USB] Try the other device initialization scheme 6834 if the first one fails (default 1 = enabled). 6835 6836 usbcore.initial_descriptor_timeout= 6837 [USB] Specifies timeout for the initial 64-byte 6838 USB_REQ_GET_DESCRIPTOR request in milliseconds 6839 (default 5000 = 5.0 seconds). 6840 6841 usbcore.nousb [USB] Disable the USB subsystem 6842 6843 usbcore.quirks= 6844 [USB] A list of quirk entries to augment the built-in 6845 usb core quirk list. List entries are separated by 6846 commas. Each entry has the form 6847 VendorID:ProductID:Flags. The IDs are 4-digit hex 6848 numbers and Flags is a set of letters. Each letter 6849 will change the built-in quirk; setting it if it is 6850 clear and clearing it if it is set. The letters have 6851 the following meanings: 6852 a = USB_QUIRK_STRING_FETCH_255 (string 6853 descriptors must not be fetched using 6854 a 255-byte read); 6855 b = USB_QUIRK_RESET_RESUME (device can't resume 6856 correctly so reset it instead); 6857 c = USB_QUIRK_NO_SET_INTF (device can't handle 6858 Set-Interface requests); 6859 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't 6860 handle its Configuration or Interface 6861 strings); 6862 e = USB_QUIRK_RESET (device can't be reset 6863 (e.g morph devices), don't use reset); 6864 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has 6865 more interface descriptions than the 6866 bNumInterfaces count, and can't handle 6867 talking to these interfaces); 6868 g = USB_QUIRK_DELAY_INIT (device needs a pause 6869 during initialization, after we read 6870 the device descriptor); 6871 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For 6872 high speed and super speed interrupt 6873 endpoints, the USB 2.0 and USB 3.0 spec 6874 require the interval in microframes (1 6875 microframe = 125 microseconds) to be 6876 calculated as interval = 2 ^ 6877 (bInterval-1). 6878 Devices with this quirk report their 6879 bInterval as the result of this 6880 calculation instead of the exponent 6881 variable used in the calculation); 6882 i = USB_QUIRK_DEVICE_QUALIFIER (device can't 6883 handle device_qualifier descriptor 6884 requests); 6885 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device 6886 generates spurious wakeup, ignore 6887 remote wakeup capability); 6888 k = USB_QUIRK_NO_LPM (device can't handle Link 6889 Power Management); 6890 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL 6891 (Device reports its bInterval as linear 6892 frames instead of the USB 2.0 6893 calculation); 6894 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs 6895 to be disconnected before suspend to 6896 prevent spurious wakeup); 6897 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a 6898 pause after every control message); 6899 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra 6900 delay after resetting its port); 6901 Example: quirks=0781:5580:bk,0a5c:5834:gij 6902 6903 usbhid.mousepoll= 6904 [USBHID] The interval which mice are to be polled at. 6905 6906 usbhid.jspoll= 6907 [USBHID] The interval which joysticks are to be polled at. 6908 6909 usbhid.kbpoll= 6910 [USBHID] The interval which keyboards are to be polled at. 6911 6912 usb-storage.delay_use= 6913 [UMS] The delay in seconds before a new device is 6914 scanned for Logical Units (default 1). 6915 6916 usb-storage.quirks= 6917 [UMS] A list of quirks entries to supplement or 6918 override the built-in unusual_devs list. List 6919 entries are separated by commas. Each entry has 6920 the form VID:PID:Flags where VID and PID are Vendor 6921 and Product ID values (4-digit hex numbers) and 6922 Flags is a set of characters, each corresponding 6923 to a common usb-storage quirk flag as follows: 6924 a = SANE_SENSE (collect more than 18 bytes 6925 of sense data, not on uas); 6926 b = BAD_SENSE (don't collect more than 18 6927 bytes of sense data, not on uas); 6928 c = FIX_CAPACITY (decrease the reported 6929 device capacity by one sector); 6930 d = NO_READ_DISC_INFO (don't use 6931 READ_DISC_INFO command, not on uas); 6932 e = NO_READ_CAPACITY_16 (don't use 6933 READ_CAPACITY_16 command); 6934 f = NO_REPORT_OPCODES (don't use report opcodes 6935 command, uas only); 6936 g = MAX_SECTORS_240 (don't transfer more than 6937 240 sectors at a time, uas only); 6938 h = CAPACITY_HEURISTICS (decrease the 6939 reported device capacity by one 6940 sector if the number is odd); 6941 i = IGNORE_DEVICE (don't bind to this 6942 device); 6943 j = NO_REPORT_LUNS (don't use report luns 6944 command, uas only); 6945 k = NO_SAME (do not use WRITE_SAME, uas only) 6946 l = NOT_LOCKABLE (don't try to lock and 6947 unlock ejectable media, not on uas); 6948 m = MAX_SECTORS_64 (don't transfer more 6949 than 64 sectors = 32 KB at a time, 6950 not on uas); 6951 n = INITIAL_READ10 (force a retry of the 6952 initial READ(10) command, not on uas); 6953 o = CAPACITY_OK (accept the capacity 6954 reported by the device, not on uas); 6955 p = WRITE_CACHE (the device cache is ON 6956 by default, not on uas); 6957 r = IGNORE_RESIDUE (the device reports 6958 bogus residue values, not on uas); 6959 s = SINGLE_LUN (the device has only one 6960 Logical Unit); 6961 t = NO_ATA_1X (don't allow ATA(12) and ATA(16) 6962 commands, uas only); 6963 u = IGNORE_UAS (don't bind to the uas driver); 6964 w = NO_WP_DETECT (don't test whether the 6965 medium is write-protected). 6966 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE 6967 even if the device claims no cache, 6968 not on uas) 6969 Example: quirks=0419:aaf5:rl,0421:0433:rc 6970 6971 user_debug= [KNL,ARM] 6972 Format: <int> 6973 See arch/arm/Kconfig.debug help text. 6974 1 - undefined instruction events 6975 2 - system calls 6976 4 - invalid data aborts 6977 8 - SIGSEGV faults 6978 16 - SIGBUS faults 6979 Example: user_debug=31 6980 6981 userpte= 6982 [X86] Flags controlling user PTE allocations. 6983 6984 nohigh = do not allocate PTE pages in 6985 HIGHMEM regardless of setting 6986 of CONFIG_HIGHPTE. 6987 6988 vdso= [X86,SH,SPARC] 6989 On X86_32, this is an alias for vdso32=. Otherwise: 6990 6991 vdso=1: enable VDSO (the default) 6992 vdso=0: disable VDSO mapping 6993 6994 vdso32= [X86] Control the 32-bit vDSO 6995 vdso32=1: enable 32-bit VDSO 6996 vdso32=0 or vdso32=2: disable 32-bit VDSO 6997 6998 See the help text for CONFIG_COMPAT_VDSO for more 6999 details. If CONFIG_COMPAT_VDSO is set, the default is 7000 vdso32=0; otherwise, the default is vdso32=1. 7001 7002 For compatibility with older kernels, vdso32=2 is an 7003 alias for vdso32=0. 7004 7005 Try vdso32=0 if you encounter an error that says: 7006 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed! 7007 7008 vector= [IA-64,SMP] 7009 vector=percpu: enable percpu vector domain 7010 7011 video= [FB] Frame buffer configuration 7012 See Documentation/fb/modedb.rst. 7013 7014 video.brightness_switch_enabled= [ACPI] 7015 Format: [0|1] 7016 If set to 1, on receiving an ACPI notify event 7017 generated by hotkey, video driver will adjust brightness 7018 level and then send out the event to user space through 7019 the allocated input device. If set to 0, video driver 7020 will only send out the event without touching backlight 7021 brightness level. 7022 default: 1 7023 7024 virtio_mmio.device= 7025 [VMMIO] Memory mapped virtio (platform) device. 7026 7027 <size>@<baseaddr>:<irq>[:<id>] 7028 where: 7029 <size> := size (can use standard suffixes 7030 like K, M and G) 7031 <baseaddr> := physical base address 7032 <irq> := interrupt number (as passed to 7033 request_irq()) 7034 <id> := (optional) platform device id 7035 example: 7036 virtio_mmio.device=1K@0x100b0000:48:7 7037 7038 Can be used multiple times for multiple devices. 7039 7040 vga= [BOOT,X86-32] Select a particular video mode 7041 See Documentation/arch/x86/boot.rst and 7042 Documentation/admin-guide/svga.rst. 7043 Use vga=ask for menu. 7044 This is actually a boot loader parameter; the value is 7045 passed to the kernel using a special protocol. 7046 7047 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y. 7048 May slow down system boot speed, especially when 7049 enabled on systems with a large amount of memory. 7050 All options are enabled by default, and this 7051 interface is meant to allow for selectively 7052 enabling or disabling specific virtual memory 7053 debugging features. 7054 7055 Available options are: 7056 P Enable page structure init time poisoning 7057 - Disable all of the above options 7058 7059 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact 7060 size of <nn>. This can be used to increase the 7061 minimum size (128MB on x86). It can also be used to 7062 decrease the size and leave more room for directly 7063 mapped kernel RAM. 7064 7065 vmcp_cma=nn[MG] [KNL,S390] 7066 Sets the memory size reserved for contiguous memory 7067 allocations for the vmcp device driver. 7068 7069 vmhalt= [KNL,S390] Perform z/VM CP command after system halt. 7070 Format: <command> 7071 7072 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic. 7073 Format: <command> 7074 7075 vmpoff= [KNL,S390] Perform z/VM CP command after power off. 7076 Format: <command> 7077 7078 vsyscall= [X86-64] 7079 Controls the behavior of vsyscalls (i.e. calls to 7080 fixed addresses of 0xffffffffff600x00 from legacy 7081 code). Most statically-linked binaries and older 7082 versions of glibc use these calls. Because these 7083 functions are at fixed addresses, they make nice 7084 targets for exploits that can control RIP. 7085 7086 emulate Vsyscalls turn into traps and are emulated 7087 reasonably safely. The vsyscall page is 7088 readable. 7089 7090 xonly [default] Vsyscalls turn into traps and are 7091 emulated reasonably safely. The vsyscall 7092 page is not readable. 7093 7094 none Vsyscalls don't work at all. This makes 7095 them quite hard to use for exploits but 7096 might break your system. 7097 7098 vt.color= [VT] Default text color. 7099 Format: 0xYX, X = foreground, Y = background. 7100 Default: 0x07 = light gray on black. 7101 7102 vt.cur_default= [VT] Default cursor shape. 7103 Format: 0xCCBBAA, where AA, BB, and CC are the same as 7104 the parameters of the <Esc>[?A;B;Cc escape sequence; 7105 see VGA-softcursor.txt. Default: 2 = underline. 7106 7107 vt.default_blu= [VT] 7108 Format: <blue0>,<blue1>,<blue2>,...,<blue15> 7109 Change the default blue palette of the console. 7110 This is a 16-member array composed of values 7111 ranging from 0-255. 7112 7113 vt.default_grn= [VT] 7114 Format: <green0>,<green1>,<green2>,...,<green15> 7115 Change the default green palette of the console. 7116 This is a 16-member array composed of values 7117 ranging from 0-255. 7118 7119 vt.default_red= [VT] 7120 Format: <red0>,<red1>,<red2>,...,<red15> 7121 Change the default red palette of the console. 7122 This is a 16-member array composed of values 7123 ranging from 0-255. 7124 7125 vt.default_utf8= 7126 [VT] 7127 Format=<0|1> 7128 Set system-wide default UTF-8 mode for all tty's. 7129 Default is 1, i.e. UTF-8 mode is enabled for all 7130 newly opened terminals. 7131 7132 vt.global_cursor_default= 7133 [VT] 7134 Format=<-1|0|1> 7135 Set system-wide default for whether a cursor 7136 is shown on new VTs. Default is -1, 7137 i.e. cursors will be created by default unless 7138 overridden by individual drivers. 0 will hide 7139 cursors, 1 will display them. 7140 7141 vt.italic= [VT] Default color for italic text; 0-15. 7142 Default: 2 = green. 7143 7144 vt.underline= [VT] Default color for underlined text; 0-15. 7145 Default: 3 = cyan. 7146 7147 watchdog timers [HW,WDT] For information on watchdog timers, 7148 see Documentation/watchdog/watchdog-parameters.rst 7149 or other driver-specific files in the 7150 Documentation/watchdog/ directory. 7151 7152 watchdog_thresh= 7153 [KNL] 7154 Set the hard lockup detector stall duration 7155 threshold in seconds. The soft lockup detector 7156 threshold is set to twice the value. A value of 0 7157 disables both lockup detectors. Default is 10 7158 seconds. 7159 7160 workqueue.unbound_cpus= 7161 [KNL,SMP] Specify to constrain one or some CPUs 7162 to use in unbound workqueues. 7163 Format: <cpu-list> 7164 By default, all online CPUs are available for 7165 unbound workqueues. 7166 7167 workqueue.watchdog_thresh= 7168 If CONFIG_WQ_WATCHDOG is configured, workqueue can 7169 warn stall conditions and dump internal state to 7170 help debugging. 0 disables workqueue stall 7171 detection; otherwise, it's the stall threshold 7172 duration in seconds. The default value is 30 and 7173 it can be updated at runtime by writing to the 7174 corresponding sysfs file. 7175 7176 workqueue.cpu_intensive_thresh_us= 7177 Per-cpu work items which run for longer than this 7178 threshold are automatically considered CPU intensive 7179 and excluded from concurrency management to prevent 7180 them from noticeably delaying other per-cpu work 7181 items. Default is 10000 (10ms). 7182 7183 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 7184 will report the work functions which violate this 7185 threshold repeatedly. They are likely good 7186 candidates for using WQ_UNBOUND workqueues instead. 7187 7188 workqueue.power_efficient 7189 Per-cpu workqueues are generally preferred because 7190 they show better performance thanks to cache 7191 locality; unfortunately, per-cpu workqueues tend to 7192 be more power hungry than unbound workqueues. 7193 7194 Enabling this makes the per-cpu workqueues which 7195 were observed to contribute significantly to power 7196 consumption unbound, leading to measurably lower 7197 power usage at the cost of small performance 7198 overhead. 7199 7200 The default value of this parameter is determined by 7201 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT. 7202 7203 workqueue.default_affinity_scope= 7204 Select the default affinity scope to use for unbound 7205 workqueues. Can be one of "cpu", "smt", "cache", 7206 "numa" and "system". Default is "cache". For more 7207 information, see the Affinity Scopes section in 7208 Documentation/core-api/workqueue.rst. 7209 7210 This can be changed after boot by writing to the 7211 matching /sys/module/workqueue/parameters file. All 7212 workqueues with the "default" affinity scope will be 7213 updated accordignly. 7214 7215 workqueue.debug_force_rr_cpu 7216 Workqueue used to implicitly guarantee that work 7217 items queued without explicit CPU specified are put 7218 on the local CPU. This guarantee is no longer true 7219 and while local CPU is still preferred work items 7220 may be put on foreign CPUs. This debug option 7221 forces round-robin CPU selection to flush out 7222 usages which depend on the now broken guarantee. 7223 When enabled, memory and cache locality will be 7224 impacted. 7225 7226 writecombine= [LOONGARCH] Control the MAT (Memory Access Type) of 7227 ioremap_wc(). 7228 7229 on - Enable writecombine, use WUC for ioremap_wc() 7230 off - Disable writecombine, use SUC for ioremap_wc() 7231 7232 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of 7233 default x2apic cluster mode on platforms 7234 supporting x2apic. 7235 7236 xen_512gb_limit [KNL,X86-64,XEN] 7237 Restricts the kernel running paravirtualized under Xen 7238 to use only up to 512 GB of RAM. The reason to do so is 7239 crash analysis tools and Xen tools for doing domain 7240 save/restore/migration must be enabled to handle larger 7241 domains. 7242 7243 xen_emul_unplug= [HW,X86,XEN] 7244 Unplug Xen emulated devices 7245 Format: [unplug0,][unplug1] 7246 ide-disks -- unplug primary master IDE devices 7247 aux-ide-disks -- unplug non-primary-master IDE devices 7248 nics -- unplug network devices 7249 all -- unplug all emulated devices (NICs and IDE disks) 7250 unnecessary -- unplugging emulated devices is 7251 unnecessary even if the host did not respond to 7252 the unplug protocol 7253 never -- do not unplug even if version check succeeds 7254 7255 xen_legacy_crash [X86,XEN] 7256 Crash from Xen panic notifier, without executing late 7257 panic() code such as dumping handler. 7258 7259 xen_msr_safe= [X86,XEN] 7260 Format: <bool> 7261 Select whether to always use non-faulting (safe) MSR 7262 access functions when running as Xen PV guest. The 7263 default value is controlled by CONFIG_XEN_PV_MSR_SAFE. 7264 7265 xen_nopvspin [X86,XEN] 7266 Disables the qspinlock slowpath using Xen PV optimizations. 7267 This parameter is obsoleted by "nopvspin" parameter, which 7268 has equivalent effect for XEN platform. 7269 7270 xen_nopv [X86] 7271 Disables the PV optimizations forcing the HVM guest to 7272 run as generic HVM guest with no PV drivers. 7273 This option is obsoleted by the "nopv" option, which 7274 has equivalent effect for XEN platform. 7275 7276 xen_no_vector_callback 7277 [KNL,X86,XEN] Disable the vector callback for Xen 7278 event channel interrupts. 7279 7280 xen_scrub_pages= [XEN] 7281 Boolean option to control scrubbing pages before giving them back 7282 to Xen, for use by other domains. Can be also changed at runtime 7283 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages. 7284 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT. 7285 7286 xen_timer_slop= [X86-64,XEN] 7287 Set the timer slop (in nanoseconds) for the virtual Xen 7288 timers (default is 100000). This adjusts the minimum 7289 delta of virtualized Xen timers, where lower values 7290 improve timer resolution at the expense of processing 7291 more timer interrupts. 7292 7293 xen.balloon_boot_timeout= [XEN] 7294 The time (in seconds) to wait before giving up to boot 7295 in case initial ballooning fails to free enough memory. 7296 Applies only when running as HVM or PVH guest and 7297 started with less memory configured than allowed at 7298 max. Default is 180. 7299 7300 xen.event_eoi_delay= [XEN] 7301 How long to delay EOI handling in case of event 7302 storms (jiffies). Default is 10. 7303 7304 xen.event_loop_timeout= [XEN] 7305 After which time (jiffies) the event handling loop 7306 should start to delay EOI handling. Default is 2. 7307 7308 xen.fifo_events= [XEN] 7309 Boolean parameter to disable using fifo event handling 7310 even if available. Normally fifo event handling is 7311 preferred over the 2-level event handling, as it is 7312 fairer and the number of possible event channels is 7313 much higher. Default is on (use fifo events). 7314 7315 xirc2ps_cs= [NET,PCMCIA] 7316 Format: 7317 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]] 7318 7319 xive= [PPC] 7320 By default on POWER9 and above, the kernel will 7321 natively use the XIVE interrupt controller. This option 7322 allows the fallback firmware mode to be used: 7323 7324 off Fallback to firmware control of XIVE interrupt 7325 controller on both pseries and powernv 7326 platforms. Only useful on POWER9 and above. 7327 7328 xive.store-eoi=off [PPC] 7329 By default on POWER10 and above, the kernel will use 7330 stores for EOI handling when the XIVE interrupt mode 7331 is active. This option allows the XIVE driver to use 7332 loads instead, as on POWER9. 7333 7334 xhci-hcd.quirks [USB,KNL] 7335 A hex value specifying bitmask with supplemental xhci 7336 host controller quirks. Meaning of each bit can be 7337 consulted in header drivers/usb/host/xhci.h. 7338 7339 xmon [PPC] 7340 Format: { early | on | rw | ro | off } 7341 Controls if xmon debugger is enabled. Default is off. 7342 Passing only "xmon" is equivalent to "xmon=early". 7343 early Call xmon as early as possible on boot; xmon 7344 debugger is called from setup_arch(). 7345 on xmon debugger hooks will be installed so xmon 7346 is only called on a kernel crash. Default mode, 7347 i.e. either "ro" or "rw" mode, is controlled 7348 with CONFIG_XMON_DEFAULT_RO_MODE. 7349 rw xmon debugger hooks will be installed so xmon 7350 is called only on a kernel crash, mode is write, 7351 meaning SPR registers, memory and, other data 7352 can be written using xmon commands. 7353 ro same as "rw" option above but SPR registers, 7354 memory, and other data can't be written using 7355 xmon commands. 7356 off xmon is disabled. 7357 7358