| 1 | acpi= [HW,ACPI,X86,ARM64] |
| 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] |
| 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, ONLY "acpi=off", "acpi=on" or "acpi=force" |
| 14 | 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/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 | |
| 325 | amd_iommu_dump= [HW,X86-64] |
| 326 | Enable AMD IOMMU driver option to dump the ACPI table |
| 327 | for AMD IOMMU. With this option enabled, AMD IOMMU |
| 328 | driver will print ACPI tables for AMD IOMMU during |
| 329 | IOMMU initialization. |
| 330 | |
| 331 | amd_iommu_intr= [HW,X86-64] |
| 332 | Specifies one of the following AMD IOMMU interrupt |
| 333 | remapping modes: |
| 334 | legacy - Use legacy interrupt remapping mode. |
| 335 | vapic - Use virtual APIC mode, which allows IOMMU |
| 336 | to inject interrupts directly into guest. |
| 337 | This mode requires kvm-amd.avic=1. |
| 338 | (Default when IOMMU HW support is present.) |
| 339 | |
| 340 | amijoy.map= [HW,JOY] Amiga joystick support |
| 341 | Map of devices attached to JOY0DAT and JOY1DAT |
| 342 | Format: <a>,<b> |
| 343 | See also Documentation/input/joydev/joystick.rst |
| 344 | |
| 345 | analog.map= [HW,JOY] Analog joystick and gamepad support |
| 346 | Specifies type or capabilities of an analog joystick |
| 347 | connected to one of 16 gameports |
| 348 | Format: <type1>,<type2>,..<type16> |
| 349 | |
| 350 | apc= [HW,SPARC] |
| 351 | Power management functions (SPARCstation-4/5 + deriv.) |
| 352 | Format: noidle |
| 353 | Disable APC CPU standby support. SPARCstation-Fox does |
| 354 | not play well with APC CPU idle - disable it if you have |
| 355 | APC and your system crashes randomly. |
| 356 | |
| 357 | apic= [APIC,X86] Advanced Programmable Interrupt Controller |
| 358 | Change the output verbosity while booting |
| 359 | Format: { quiet (default) | verbose | debug } |
| 360 | Change the amount of debugging information output |
| 361 | when initialising the APIC and IO-APIC components. |
| 362 | For X86-32, this can also be used to specify an APIC |
| 363 | driver name. |
| 364 | Format: apic=driver_name |
| 365 | Examples: apic=bigsmp |
| 366 | |
| 367 | apic_extnmi= [APIC,X86] External NMI delivery setting |
| 368 | Format: { bsp (default) | all | none } |
| 369 | bsp: External NMI is delivered only to CPU 0 |
| 370 | all: External NMIs are broadcast to all CPUs as a |
| 371 | backup of CPU 0 |
| 372 | none: External NMI is masked for all CPUs. This is |
| 373 | useful so that a dump capture kernel won't be |
| 374 | shot down by NMI |
| 375 | |
| 376 | autoconf= [IPV6] |
| 377 | See Documentation/networking/ipv6.rst. |
| 378 | |
| 379 | show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller |
| 380 | Limit apic dumping. The parameter defines the maximal |
| 381 | number of local apics being dumped. Also it is possible |
| 382 | to set it to "all" by meaning -- no limit here. |
| 383 | Format: { 1 (default) | 2 | ... | all }. |
| 384 | The parameter valid if only apic=debug or |
| 385 | apic=verbose is specified. |
| 386 | Example: apic=debug show_lapic=all |
| 387 | |
| 388 | apm= [APM] Advanced Power Management |
| 389 | See header of arch/x86/kernel/apm_32.c. |
| 390 | |
| 391 | arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards |
| 392 | Format: <io>,<irq>,<nodeID> |
| 393 | |
| 394 | arm64.nobti [ARM64] Unconditionally disable Branch Target |
| 395 | Identification support |
| 396 | |
| 397 | arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication |
| 398 | support |
| 399 | |
| 400 | arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension |
| 401 | support |
| 402 | |
| 403 | ataflop= [HW,M68k] |
| 404 | |
| 405 | atarimouse= [HW,MOUSE] Atari Mouse |
| 406 | |
| 407 | atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess, |
| 408 | EzKey and similar keyboards |
| 409 | |
| 410 | atkbd.reset= [HW] Reset keyboard during initialization |
| 411 | |
| 412 | atkbd.set= [HW] Select keyboard code set |
| 413 | Format: <int> (2 = AT (default), 3 = PS/2) |
| 414 | |
| 415 | atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar |
| 416 | keyboards |
| 417 | |
| 418 | atkbd.softraw= [HW] Choose between synthetic and real raw mode |
| 419 | Format: <bool> (0 = real, 1 = synthetic (default)) |
| 420 | |
| 421 | atkbd.softrepeat= [HW] |
| 422 | Use software keyboard repeat |
| 423 | |
| 424 | audit= [KNL] Enable the audit sub-system |
| 425 | Format: { "0" | "1" | "off" | "on" } |
| 426 | 0 | off - kernel audit is disabled and can not be |
| 427 | enabled until the next reboot |
| 428 | unset - kernel audit is initialized but disabled and |
| 429 | will be fully enabled by the userspace auditd. |
| 430 | 1 | on - kernel audit is initialized and partially |
| 431 | enabled, storing at most audit_backlog_limit |
| 432 | messages in RAM until it is fully enabled by the |
| 433 | userspace auditd. |
| 434 | Default: unset |
| 435 | |
| 436 | audit_backlog_limit= [KNL] Set the audit queue size limit. |
| 437 | Format: <int> (must be >=0) |
| 438 | Default: 64 |
| 439 | |
| 440 | bau= [X86_UV] Enable the BAU on SGI UV. The default |
| 441 | behavior is to disable the BAU (i.e. bau=0). |
| 442 | Format: { "0" | "1" } |
| 443 | 0 - Disable the BAU. |
| 444 | 1 - Enable the BAU. |
| 445 | unset - Disable the BAU. |
| 446 | |
| 447 | baycom_epp= [HW,AX25] |
| 448 | Format: <io>,<mode> |
| 449 | |
| 450 | baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem |
| 451 | Format: <io>,<mode> |
| 452 | See header of drivers/net/hamradio/baycom_par.c. |
| 453 | |
| 454 | baycom_ser_fdx= [HW,AX25] |
| 455 | BayCom Serial Port AX.25 Modem (Full Duplex Mode) |
| 456 | Format: <io>,<irq>,<mode>[,<baud>] |
| 457 | See header of drivers/net/hamradio/baycom_ser_fdx.c. |
| 458 | |
| 459 | baycom_ser_hdx= [HW,AX25] |
| 460 | BayCom Serial Port AX.25 Modem (Half Duplex Mode) |
| 461 | Format: <io>,<irq>,<mode> |
| 462 | See header of drivers/net/hamradio/baycom_ser_hdx.c. |
| 463 | |
| 464 | blkdevparts= Manual partition parsing of block device(s) for |
| 465 | embedded devices based on command line input. |
| 466 | See Documentation/block/cmdline-partition.rst |
| 467 | |
| 468 | boot_delay= Milliseconds to delay each printk during boot. |
| 469 | Values larger than 10 seconds (10000) are changed to |
| 470 | no delay (0). |
| 471 | Format: integer |
| 472 | |
| 473 | bootconfig [KNL] |
| 474 | Extended command line options can be added to an initrd |
| 475 | and this will cause the kernel to look for it. |
| 476 | |
| 477 | See Documentation/admin-guide/bootconfig.rst |
| 478 | |
| 479 | bert_disable [ACPI] |
| 480 | Disable BERT OS support on buggy BIOSes. |
| 481 | |
| 482 | bgrt_disable [ACPI][X86] |
| 483 | Disable BGRT to avoid flickering OEM logo. |
| 484 | |
| 485 | bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards) |
| 486 | bttv.radio= Most important insmod options are available as |
| 487 | kernel args too. |
| 488 | bttv.pll= See Documentation/admin-guide/media/bttv.rst |
| 489 | bttv.tuner= |
| 490 | |
| 491 | bulk_remove=off [PPC] This parameter disables the use of the pSeries |
| 492 | firmware feature for flushing multiple hpte entries |
| 493 | at a time. |
| 494 | |
| 495 | c101= [NET] Moxa C101 synchronous serial card |
| 496 | |
| 497 | cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection. |
| 498 | Sometimes CPU hardware bugs make them report the cache |
| 499 | size incorrectly. The kernel will attempt work arounds |
| 500 | to fix known problems, but for some CPUs it is not |
| 501 | possible to determine what the correct size should be. |
| 502 | This option provides an override for these situations. |
| 503 | |
| 504 | carrier_timeout= |
| 505 | [NET] Specifies amount of time (in seconds) that |
| 506 | the kernel should wait for a network carrier. By default |
| 507 | it waits 120 seconds. |
| 508 | |
| 509 | ca_keys= [KEYS] This parameter identifies a specific key(s) on |
| 510 | the system trusted keyring to be used for certificate |
| 511 | trust validation. |
| 512 | format: { id:<keyid> | builtin } |
| 513 | |
| 514 | cca= [MIPS] Override the kernel pages' cache coherency |
| 515 | algorithm. Accepted values range from 0 to 7 |
| 516 | inclusive. See arch/mips/include/asm/pgtable-bits.h |
| 517 | for platform specific values (SB1, Loongson3 and |
| 518 | others). |
| 519 | |
| 520 | ccw_timeout_log [S390] |
| 521 | See Documentation/s390/common_io.rst for details. |
| 522 | |
| 523 | cgroup_disable= [KNL] Disable a particular controller or optional feature |
| 524 | Format: {name of the controller(s) or feature(s) to disable} |
| 525 | The effects of cgroup_disable=foo are: |
| 526 | - foo isn't auto-mounted if you mount all cgroups in |
| 527 | a single hierarchy |
| 528 | - foo isn't visible as an individually mountable |
| 529 | subsystem |
| 530 | - if foo is an optional feature then the feature is |
| 531 | disabled and corresponding cgroup files are not |
| 532 | created |
| 533 | {Currently only "memory" controller deal with this and |
| 534 | cut the overhead, others just disable the usage. So |
| 535 | only cgroup_disable=memory is actually worthy} |
| 536 | Specifying "pressure" disables per-cgroup pressure |
| 537 | stall information accounting feature |
| 538 | |
| 539 | cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1 |
| 540 | Format: { { controller | "all" | "named" } |
| 541 | [,{ controller | "all" | "named" }...] } |
| 542 | Like cgroup_disable, but only applies to cgroup v1; |
| 543 | the blacklisted controllers remain available in cgroup2. |
| 544 | "all" blacklists all controllers and "named" disables |
| 545 | named mounts. Specifying both "all" and "named" disables |
| 546 | all v1 hierarchies. |
| 547 | |
| 548 | cgroup.memory= [KNL] Pass options to the cgroup memory controller. |
| 549 | Format: <string> |
| 550 | nosocket -- Disable socket memory accounting. |
| 551 | nokmem -- Disable kernel memory accounting. |
| 552 | |
| 553 | checkreqprot [SELINUX] Set initial checkreqprot flag value. |
| 554 | Format: { "0" | "1" } |
| 555 | See security/selinux/Kconfig help text. |
| 556 | 0 -- check protection applied by kernel (includes |
| 557 | any implied execute protection). |
| 558 | 1 -- check protection requested by application. |
| 559 | Default value is set via a kernel config option. |
| 560 | Value can be changed at runtime via |
| 561 | /sys/fs/selinux/checkreqprot. |
| 562 | Setting checkreqprot to 1 is deprecated. |
| 563 | |
| 564 | cio_ignore= [S390] |
| 565 | See Documentation/s390/common_io.rst for details. |
| 566 | clk_ignore_unused |
| 567 | [CLK] |
| 568 | Prevents the clock framework from automatically gating |
| 569 | clocks that have not been explicitly enabled by a Linux |
| 570 | device driver but are enabled in hardware at reset or |
| 571 | by the bootloader/firmware. Note that this does not |
| 572 | force such clocks to be always-on nor does it reserve |
| 573 | those clocks in any way. This parameter is useful for |
| 574 | debug and development, but should not be needed on a |
| 575 | platform with proper driver support. For more |
| 576 | information, see Documentation/driver-api/clk.rst. |
| 577 | |
| 578 | clock= [BUGS=X86-32, HW] gettimeofday clocksource override. |
| 579 | [Deprecated] |
| 580 | Forces specified clocksource (if available) to be used |
| 581 | when calculating gettimeofday(). If specified |
| 582 | clocksource is not available, it defaults to PIT. |
| 583 | Format: { pit | tsc | cyclone | pmtmr } |
| 584 | |
| 585 | clocksource= Override the default clocksource |
| 586 | Format: <string> |
| 587 | Override the default clocksource and use the clocksource |
| 588 | with the name specified. |
| 589 | Some clocksource names to choose from, depending on |
| 590 | the platform: |
| 591 | [all] jiffies (this is the base, fallback clocksource) |
| 592 | [ACPI] acpi_pm |
| 593 | [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2, |
| 594 | pxa_timer,timer3,32k_counter,timer0_1 |
| 595 | [X86-32] pit,hpet,tsc; |
| 596 | scx200_hrt on Geode; cyclone on IBM x440 |
| 597 | [MIPS] MIPS |
| 598 | [PARISC] cr16 |
| 599 | [S390] tod |
| 600 | [SH] SuperH |
| 601 | [SPARC64] tick |
| 602 | [X86-64] hpet,tsc |
| 603 | |
| 604 | clocksource.arm_arch_timer.evtstrm= |
| 605 | [ARM,ARM64] |
| 606 | Format: <bool> |
| 607 | Enable/disable the eventstream feature of the ARM |
| 608 | architected timer so that code using WFE-based polling |
| 609 | loops can be debugged more effectively on production |
| 610 | systems. |
| 611 | |
| 612 | clocksource.max_cswd_read_retries= [KNL] |
| 613 | Number of clocksource_watchdog() retries due to |
| 614 | external delays before the clock will be marked |
| 615 | unstable. Defaults to two retries, that is, |
| 616 | three attempts to read the clock under test. |
| 617 | |
| 618 | clocksource.verify_n_cpus= [KNL] |
| 619 | Limit the number of CPUs checked for clocksources |
| 620 | marked with CLOCK_SOURCE_VERIFY_PERCPU that |
| 621 | are marked unstable due to excessive skew. |
| 622 | A negative value says to check all CPUs, while |
| 623 | zero says not to check any. Values larger than |
| 624 | nr_cpu_ids are silently truncated to nr_cpu_ids. |
| 625 | The actual CPUs are chosen randomly, with |
| 626 | no replacement if the same CPU is chosen twice. |
| 627 | |
| 628 | clocksource-wdtest.holdoff= [KNL] |
| 629 | Set the time in seconds that the clocksource |
| 630 | watchdog test waits before commencing its tests. |
| 631 | Defaults to zero when built as a module and to |
| 632 | 10 seconds when built into the kernel. |
| 633 | |
| 634 | clearcpuid=BITNUM[,BITNUM...] [X86] |
| 635 | Disable CPUID feature X for the kernel. See |
| 636 | arch/x86/include/asm/cpufeatures.h for the valid bit |
| 637 | numbers. Note the Linux specific bits are not necessarily |
| 638 | stable over kernel options, but the vendor specific |
| 639 | ones should be. |
| 640 | Also note that user programs calling CPUID directly |
| 641 | or using the feature without checking anything |
| 642 | will still see it. This just prevents it from |
| 643 | being used by the kernel or shown in /proc/cpuinfo. |
| 644 | Also note the kernel might malfunction if you disable |
| 645 | some critical bits. |
| 646 | |
| 647 | cma=nn[MG]@[start[MG][-end[MG]]] |
| 648 | [KNL,CMA] |
| 649 | Sets the size of kernel global memory area for |
| 650 | contiguous memory allocations and optionally the |
| 651 | placement constraint by the physical address range of |
| 652 | memory allocations. A value of 0 disables CMA |
| 653 | altogether. For more information, see |
| 654 | kernel/dma/contiguous.c |
| 655 | |
| 656 | cma_pernuma=nn[MG] |
| 657 | [ARM64,KNL,CMA] |
| 658 | Sets the size of kernel per-numa memory area for |
| 659 | contiguous memory allocations. A value of 0 disables |
| 660 | per-numa CMA altogether. And If this option is not |
| 661 | specificed, the default value is 0. |
| 662 | With per-numa CMA enabled, DMA users on node nid will |
| 663 | first try to allocate buffer from the pernuma area |
| 664 | which is located in node nid, if the allocation fails, |
| 665 | they will fallback to the global default memory area. |
| 666 | |
| 667 | cmo_free_hint= [PPC] Format: { yes | no } |
| 668 | Specify whether pages are marked as being inactive |
| 669 | when they are freed. This is used in CMO environments |
| 670 | to determine OS memory pressure for page stealing by |
| 671 | a hypervisor. |
| 672 | Default: yes |
| 673 | |
| 674 | coherent_pool=nn[KMG] [ARM,KNL] |
| 675 | Sets the size of memory pool for coherent, atomic dma |
| 676 | allocations, by default set to 256K. |
| 677 | |
| 678 | com20020= [HW,NET] ARCnet - COM20020 chipset |
| 679 | Format: |
| 680 | <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]] |
| 681 | |
| 682 | com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers) |
| 683 | Format: <io>[,<irq>] |
| 684 | |
| 685 | com90xx= [HW,NET] |
| 686 | ARCnet - COM90xx chipset (memory-mapped buffers) |
| 687 | Format: <io>[,<irq>[,<memstart>]] |
| 688 | |
| 689 | condev= [HW,S390] console device |
| 690 | conmode= |
| 691 | |
| 692 | console= [KNL] Output console device and options. |
| 693 | |
| 694 | tty<n> Use the virtual console device <n>. |
| 695 | |
| 696 | ttyS<n>[,options] |
| 697 | ttyUSB0[,options] |
| 698 | Use the specified serial port. The options are of |
| 699 | the form "bbbbpnf", where "bbbb" is the baud rate, |
| 700 | "p" is parity ("n", "o", or "e"), "n" is number of |
| 701 | bits, and "f" is flow control ("r" for RTS or |
| 702 | omit it). Default is "9600n8". |
| 703 | |
| 704 | See Documentation/admin-guide/serial-console.rst for more |
| 705 | information. See |
| 706 | Documentation/networking/netconsole.rst for an |
| 707 | alternative. |
| 708 | |
| 709 | uart[8250],io,<addr>[,options] |
| 710 | uart[8250],mmio,<addr>[,options] |
| 711 | uart[8250],mmio16,<addr>[,options] |
| 712 | uart[8250],mmio32,<addr>[,options] |
| 713 | uart[8250],0x<addr>[,options] |
| 714 | Start an early, polled-mode console on the 8250/16550 |
| 715 | UART at the specified I/O port or MMIO address, |
| 716 | switching to the matching ttyS device later. |
| 717 | MMIO inter-register address stride is either 8-bit |
| 718 | (mmio), 16-bit (mmio16), or 32-bit (mmio32). |
| 719 | If none of [io|mmio|mmio16|mmio32], <addr> is assumed |
| 720 | to be equivalent to 'mmio'. 'options' are specified in |
| 721 | the same format described for ttyS above; if unspecified, |
| 722 | the h/w is not re-initialized. |
| 723 | |
| 724 | hvc<n> Use the hypervisor console device <n>. This is for |
| 725 | both Xen and PowerPC hypervisors. |
| 726 | |
| 727 | If the device connected to the port is not a TTY but a braille |
| 728 | device, prepend "brl," before the device type, for instance |
| 729 | console=brl,ttyS0 |
| 730 | For now, only VisioBraille is supported. |
| 731 | |
| 732 | console_msg_format= |
| 733 | [KNL] Change console messages format |
| 734 | default |
| 735 | By default we print messages on consoles in |
| 736 | "[time stamp] text\n" format (time stamp may not be |
| 737 | printed, depending on CONFIG_PRINTK_TIME or |
| 738 | `printk_time' param). |
| 739 | syslog |
| 740 | Switch to syslog format: "<%u>[time stamp] text\n" |
| 741 | IOW, each message will have a facility and loglevel |
| 742 | prefix. The format is similar to one used by syslog() |
| 743 | syscall, or to executing "dmesg -S --raw" or to reading |
| 744 | from /proc/kmsg. |
| 745 | |
| 746 | consoleblank= [KNL] The console blank (screen saver) timeout in |
| 747 | seconds. A value of 0 disables the blank timer. |
| 748 | Defaults to 0. |
| 749 | |
| 750 | coredump_filter= |
| 751 | [KNL] Change the default value for |
| 752 | /proc/<pid>/coredump_filter. |
| 753 | See also Documentation/filesystems/proc.rst. |
| 754 | |
| 755 | coresight_cpu_debug.enable |
| 756 | [ARM,ARM64] |
| 757 | Format: <bool> |
| 758 | Enable/disable the CPU sampling based debugging. |
| 759 | 0: default value, disable debugging |
| 760 | 1: enable debugging at boot time |
| 761 | |
| 762 | cpuidle.off=1 [CPU_IDLE] |
| 763 | disable the cpuidle sub-system |
| 764 | |
| 765 | cpuidle.governor= |
| 766 | [CPU_IDLE] Name of the cpuidle governor to use. |
| 767 | |
| 768 | cpufreq.off=1 [CPU_FREQ] |
| 769 | disable the cpufreq sub-system |
| 770 | |
| 771 | cpufreq.default_governor= |
| 772 | [CPU_FREQ] Name of the default cpufreq governor or |
| 773 | policy to use. This governor must be registered in the |
| 774 | kernel before the cpufreq driver probes. |
| 775 | |
| 776 | cpu_init_udelay=N |
| 777 | [X86] Delay for N microsec between assert and de-assert |
| 778 | of APIC INIT to start processors. This delay occurs |
| 779 | on every CPU online, such as boot, and resume from suspend. |
| 780 | Default: 10000 |
| 781 | |
| 782 | cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver |
| 783 | Format: |
| 784 | <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>] |
| 785 | |
| 786 | crashkernel=size[KMG][@offset[KMG]] |
| 787 | [KNL] Using kexec, Linux can switch to a 'crash kernel' |
| 788 | upon panic. This parameter reserves the physical |
| 789 | memory region [offset, offset + size] for that kernel |
| 790 | image. If '@offset' is omitted, then a suitable offset |
| 791 | is selected automatically. |
| 792 | [KNL, X86-64] Select a region under 4G first, and |
| 793 | fall back to reserve region above 4G when '@offset' |
| 794 | hasn't been specified. |
| 795 | See Documentation/admin-guide/kdump/kdump.rst for further details. |
| 796 | |
| 797 | crashkernel=range1:size1[,range2:size2,...][@offset] |
| 798 | [KNL] Same as above, but depends on the memory |
| 799 | in the running system. The syntax of range is |
| 800 | start-[end] where start and end are both |
| 801 | a memory unit (amount[KMG]). See also |
| 802 | Documentation/admin-guide/kdump/kdump.rst for an example. |
| 803 | |
| 804 | crashkernel=size[KMG],high |
| 805 | [KNL, X86-64] range could be above 4G. Allow kernel |
| 806 | to allocate physical memory region from top, so could |
| 807 | be above 4G if system have more than 4G ram installed. |
| 808 | Otherwise memory region will be allocated below 4G, if |
| 809 | available. |
| 810 | It will be ignored if crashkernel=X is specified. |
| 811 | crashkernel=size[KMG],low |
| 812 | [KNL, X86-64] range under 4G. When crashkernel=X,high |
| 813 | is passed, kernel could allocate physical memory region |
| 814 | above 4G, that cause second kernel crash on system |
| 815 | that require some amount of low memory, e.g. swiotlb |
| 816 | requires at least 64M+32K low memory, also enough extra |
| 817 | low memory is needed to make sure DMA buffers for 32-bit |
| 818 | devices won't run out. Kernel would try to allocate at |
| 819 | at least 256M below 4G automatically. |
| 820 | This one let user to specify own low range under 4G |
| 821 | for second kernel instead. |
| 822 | 0: to disable low allocation. |
| 823 | It will be ignored when crashkernel=X,high is not used |
| 824 | or memory reserved is below 4G. |
| 825 | |
| 826 | cryptomgr.notests |
| 827 | [KNL] Disable crypto self-tests |
| 828 | |
| 829 | cs89x0_dma= [HW,NET] |
| 830 | Format: <dma> |
| 831 | |
| 832 | cs89x0_media= [HW,NET] |
| 833 | Format: { rj45 | aui | bnc } |
| 834 | |
| 835 | csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call |
| 836 | handling. When switched on, additional debug data is |
| 837 | printed to the console in case a hanging CPU is |
| 838 | detected, and that CPU is pinged again in order to try |
| 839 | to resolve the hang situation. |
| 840 | 0: disable csdlock debugging (default) |
| 841 | 1: enable basic csdlock debugging (minor impact) |
| 842 | ext: enable extended csdlock debugging (more impact, |
| 843 | but more data) |
| 844 | |
| 845 | dasd= [HW,NET] |
| 846 | See header of drivers/s390/block/dasd_devmap.c. |
| 847 | |
| 848 | db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port |
| 849 | (one device per port) |
| 850 | Format: <port#>,<type> |
| 851 | See also Documentation/input/devices/joystick-parport.rst |
| 852 | |
| 853 | debug [KNL] Enable kernel debugging (events log level). |
| 854 | |
| 855 | debug_boot_weak_hash |
| 856 | [KNL] Enable printing [hashed] pointers early in the |
| 857 | boot sequence. If enabled, we use a weak hash instead |
| 858 | of siphash to hash pointers. Use this option if you are |
| 859 | seeing instances of '(___ptrval___)') and need to see a |
| 860 | value (hashed pointer) instead. Cryptographically |
| 861 | insecure, please do not use on production kernels. |
| 862 | |
| 863 | debug_locks_verbose= |
| 864 | [KNL] verbose locking self-tests |
| 865 | Format: <int> |
| 866 | Print debugging info while doing the locking API |
| 867 | self-tests. |
| 868 | Bitmask for the various LOCKTYPE_ tests. Defaults to 0 |
| 869 | (no extra messages), setting it to -1 (all bits set) |
| 870 | will print _a_lot_ more information - normally only |
| 871 | useful to lockdep developers. |
| 872 | |
| 873 | debug_objects [KNL] Enable object debugging |
| 874 | |
| 875 | no_debug_objects |
| 876 | [KNL] Disable object debugging |
| 877 | |
| 878 | debug_guardpage_minorder= |
| 879 | [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this |
| 880 | parameter allows control of the order of pages that will |
| 881 | be intentionally kept free (and hence protected) by the |
| 882 | buddy allocator. Bigger value increase the probability |
| 883 | of catching random memory corruption, but reduce the |
| 884 | amount of memory for normal system use. The maximum |
| 885 | possible value is MAX_ORDER/2. Setting this parameter |
| 886 | to 1 or 2 should be enough to identify most random |
| 887 | memory corruption problems caused by bugs in kernel or |
| 888 | driver code when a CPU writes to (or reads from) a |
| 889 | random memory location. Note that there exists a class |
| 890 | of memory corruptions problems caused by buggy H/W or |
| 891 | F/W or by drivers badly programing DMA (basically when |
| 892 | memory is written at bus level and the CPU MMU is |
| 893 | bypassed) which are not detectable by |
| 894 | CONFIG_DEBUG_PAGEALLOC, hence this option will not help |
| 895 | tracking down these problems. |
| 896 | |
| 897 | debug_pagealloc= |
| 898 | [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter |
| 899 | enables the feature at boot time. By default, it is |
| 900 | disabled and the system will work mostly the same as a |
| 901 | kernel built without CONFIG_DEBUG_PAGEALLOC. |
| 902 | Note: to get most of debug_pagealloc error reports, it's |
| 903 | useful to also enable the page_owner functionality. |
| 904 | on: enable the feature |
| 905 | |
| 906 | debugfs= [KNL] This parameter enables what is exposed to userspace |
| 907 | and debugfs internal clients. |
| 908 | Format: { on, no-mount, off } |
| 909 | on: All functions are enabled. |
| 910 | no-mount: |
| 911 | Filesystem is not registered but kernel clients can |
| 912 | access APIs and a crashkernel can be used to read |
| 913 | its content. There is nothing to mount. |
| 914 | off: Filesystem is not registered and clients |
| 915 | get a -EPERM as result when trying to register files |
| 916 | or directories within debugfs. |
| 917 | This is equivalent of the runtime functionality if |
| 918 | debugfs was not enabled in the kernel at all. |
| 919 | Default value is set in build-time with a kernel configuration. |
| 920 | |
| 921 | debugpat [X86] Enable PAT debugging |
| 922 | |
| 923 | decnet.addr= [HW,NET] |
| 924 | Format: <area>[,<node>] |
| 925 | See also Documentation/networking/decnet.rst. |
| 926 | |
| 927 | default_hugepagesz= |
| 928 | [HW] The size of the default HugeTLB page. This is |
| 929 | the size represented by the legacy /proc/ hugepages |
| 930 | APIs. In addition, this is the default hugetlb size |
| 931 | used for shmget(), mmap() and mounting hugetlbfs |
| 932 | filesystems. If not specified, defaults to the |
| 933 | architecture's default huge page size. Huge page |
| 934 | sizes are architecture dependent. See also |
| 935 | Documentation/admin-guide/mm/hugetlbpage.rst. |
| 936 | Format: size[KMG] |
| 937 | |
| 938 | deferred_probe_timeout= |
| 939 | [KNL] Debugging option to set a timeout in seconds for |
| 940 | deferred probe to give up waiting on dependencies to |
| 941 | probe. Only specific dependencies (subsystems or |
| 942 | drivers) that have opted in will be ignored. A timeout of 0 |
| 943 | will timeout at the end of initcalls. This option will also |
| 944 | dump out devices still on the deferred probe list after |
| 945 | retrying. |
| 946 | |
| 947 | dell_smm_hwmon.ignore_dmi= |
| 948 | [HW] Continue probing hardware even if DMI data |
| 949 | indicates that the driver is running on unsupported |
| 950 | hardware. |
| 951 | |
| 952 | dell_smm_hwmon.force= |
| 953 | [HW] Activate driver even if SMM BIOS signature does |
| 954 | not match list of supported models and enable otherwise |
| 955 | blacklisted features. |
| 956 | |
| 957 | dell_smm_hwmon.power_status= |
| 958 | [HW] Report power status in /proc/i8k |
| 959 | (disabled by default). |
| 960 | |
| 961 | dell_smm_hwmon.restricted= |
| 962 | [HW] Allow controlling fans only if SYS_ADMIN |
| 963 | capability is set. |
| 964 | |
| 965 | dfltcc= [HW,S390] |
| 966 | Format: { on | off | def_only | inf_only | always } |
| 967 | on: s390 zlib hardware support for compression on |
| 968 | level 1 and decompression (default) |
| 969 | off: No s390 zlib hardware support |
| 970 | def_only: s390 zlib hardware support for deflate |
| 971 | only (compression on level 1) |
| 972 | inf_only: s390 zlib hardware support for inflate |
| 973 | only (decompression) |
| 974 | always: Same as 'on' but ignores the selected compression |
| 975 | level always using hardware support (used for debugging) |
| 976 | |
| 977 | dhash_entries= [KNL] |
| 978 | Set number of hash buckets for dentry cache. |
| 979 | |
| 980 | disable_1tb_segments [PPC] |
| 981 | Disables the use of 1TB hash page table segments. This |
| 982 | causes the kernel to fall back to 256MB segments which |
| 983 | can be useful when debugging issues that require an SLB |
| 984 | miss to occur. |
| 985 | |
| 986 | stress_slb [PPC] |
| 987 | Limits the number of kernel SLB entries, and flushes |
| 988 | them frequently to increase the rate of SLB faults |
| 989 | on kernel addresses. |
| 990 | |
| 991 | disable= [IPV6] |
| 992 | See Documentation/networking/ipv6.rst. |
| 993 | |
| 994 | hardened_usercopy= |
| 995 | [KNL] Under CONFIG_HARDENED_USERCOPY, whether |
| 996 | hardening is enabled for this boot. Hardened |
| 997 | usercopy checking is used to protect the kernel |
| 998 | from reading or writing beyond known memory |
| 999 | allocation boundaries as a proactive defense |
| 1000 | against bounds-checking flaws in the kernel's |
| 1001 | copy_to_user()/copy_from_user() interface. |
| 1002 | on Perform hardened usercopy checks (default). |
| 1003 | off Disable hardened usercopy checks. |
| 1004 | |
| 1005 | disable_radix [PPC] |
| 1006 | Disable RADIX MMU mode on POWER9 |
| 1007 | |
| 1008 | radix_hcall_invalidate=on [PPC/PSERIES] |
| 1009 | Disable RADIX GTSE feature and use hcall for TLB |
| 1010 | invalidate. |
| 1011 | |
| 1012 | disable_tlbie [PPC] |
| 1013 | Disable TLBIE instruction. Currently does not work |
| 1014 | with KVM, with HASH MMU, or with coherent accelerators. |
| 1015 | |
| 1016 | disable_cpu_apicid= [X86,APIC,SMP] |
| 1017 | Format: <int> |
| 1018 | The number of initial APIC ID for the |
| 1019 | corresponding CPU to be disabled at boot, |
| 1020 | mostly used for the kdump 2nd kernel to |
| 1021 | disable BSP to wake up multiple CPUs without |
| 1022 | causing system reset or hang due to sending |
| 1023 | INIT from AP to BSP. |
| 1024 | |
| 1025 | disable_ddw [PPC/PSERIES] |
| 1026 | Disable Dynamic DMA Window support. Use this |
| 1027 | to workaround buggy firmware. |
| 1028 | |
| 1029 | disable_ipv6= [IPV6] |
| 1030 | See Documentation/networking/ipv6.rst. |
| 1031 | |
| 1032 | disable_mtrr_cleanup [X86] |
| 1033 | The kernel tries to adjust MTRR layout from continuous |
| 1034 | to discrete, to make X server driver able to add WB |
| 1035 | entry later. This parameter disables that. |
| 1036 | |
| 1037 | disable_mtrr_trim [X86, Intel and AMD only] |
| 1038 | By default the kernel will trim any uncacheable |
| 1039 | memory out of your available memory pool based on |
| 1040 | MTRR settings. This parameter disables that behavior, |
| 1041 | possibly causing your machine to run very slowly. |
| 1042 | |
| 1043 | disable_timer_pin_1 [X86] |
| 1044 | Disable PIN 1 of APIC timer |
| 1045 | Can be useful to work around chipset bugs. |
| 1046 | |
| 1047 | dis_ucode_ldr [X86] Disable the microcode loader. |
| 1048 | |
| 1049 | dma_debug=off If the kernel is compiled with DMA_API_DEBUG support, |
| 1050 | this option disables the debugging code at boot. |
| 1051 | |
| 1052 | dma_debug_entries=<number> |
| 1053 | This option allows to tune the number of preallocated |
| 1054 | entries for DMA-API debugging code. One entry is |
| 1055 | required per DMA-API allocation. Use this if the |
| 1056 | DMA-API debugging code disables itself because the |
| 1057 | architectural default is too low. |
| 1058 | |
| 1059 | dma_debug_driver=<driver_name> |
| 1060 | With this option the DMA-API debugging driver |
| 1061 | filter feature can be enabled at boot time. Just |
| 1062 | pass the driver to filter for as the parameter. |
| 1063 | The filter can be disabled or changed to another |
| 1064 | driver later using sysfs. |
| 1065 | |
| 1066 | driver_async_probe= [KNL] |
| 1067 | List of driver names to be probed asynchronously. |
| 1068 | Format: <driver_name1>,<driver_name2>... |
| 1069 | |
| 1070 | drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>] |
| 1071 | Broken monitors, graphic adapters, KVMs and EDIDless |
| 1072 | panels may send no or incorrect EDID data sets. |
| 1073 | This parameter allows to specify an EDID data sets |
| 1074 | in the /lib/firmware directory that are used instead. |
| 1075 | Generic built-in EDID data sets are used, if one of |
| 1076 | edid/1024x768.bin, edid/1280x1024.bin, |
| 1077 | edid/1680x1050.bin, or edid/1920x1080.bin is given |
| 1078 | and no file with the same name exists. Details and |
| 1079 | instructions how to build your own EDID data are |
| 1080 | available in Documentation/admin-guide/edid.rst. An EDID |
| 1081 | data set will only be used for a particular connector, |
| 1082 | if its name and a colon are prepended to the EDID |
| 1083 | name. Each connector may use a unique EDID data |
| 1084 | set by separating the files with a comma. An EDID |
| 1085 | data set with no connector name will be used for |
| 1086 | any connectors not explicitly specified. |
| 1087 | |
| 1088 | dscc4.setup= [NET] |
| 1089 | |
| 1090 | dt_cpu_ftrs= [PPC] |
| 1091 | Format: {"off" | "known"} |
| 1092 | Control how the dt_cpu_ftrs device-tree binding is |
| 1093 | used for CPU feature discovery and setup (if it |
| 1094 | exists). |
| 1095 | off: Do not use it, fall back to legacy cpu table. |
| 1096 | known: Do not pass through unknown features to guests |
| 1097 | or userspace, only those that the kernel is aware of. |
| 1098 | |
| 1099 | dump_apple_properties [X86] |
| 1100 | Dump name and content of EFI device properties on |
| 1101 | x86 Macs. Useful for driver authors to determine |
| 1102 | what data is available or for reverse-engineering. |
| 1103 | |
| 1104 | dyndbg[="val"] [KNL,DYNAMIC_DEBUG] |
| 1105 | <module>.dyndbg[="val"] |
| 1106 | Enable debug messages at boot time. See |
| 1107 | Documentation/admin-guide/dynamic-debug-howto.rst |
| 1108 | for details. |
| 1109 | |
| 1110 | nopku [X86] Disable Memory Protection Keys CPU feature found |
| 1111 | in some Intel CPUs. |
| 1112 | |
| 1113 | <module>.async_probe [KNL] |
| 1114 | Enable asynchronous probe on this module. |
| 1115 | |
| 1116 | early_ioremap_debug [KNL] |
| 1117 | Enable debug messages in early_ioremap support. This |
| 1118 | is useful for tracking down temporary early mappings |
| 1119 | which are not unmapped. |
| 1120 | |
| 1121 | earlycon= [KNL] Output early console device and options. |
| 1122 | |
| 1123 | When used with no options, the early console is |
| 1124 | determined by stdout-path property in device tree's |
| 1125 | chosen node or the ACPI SPCR table if supported by |
| 1126 | the platform. |
| 1127 | |
| 1128 | cdns,<addr>[,options] |
| 1129 | Start an early, polled-mode console on a Cadence |
| 1130 | (xuartps) serial port at the specified address. Only |
| 1131 | supported option is baud rate. If baud rate is not |
| 1132 | specified, the serial port must already be setup and |
| 1133 | configured. |
| 1134 | |
| 1135 | uart[8250],io,<addr>[,options] |
| 1136 | uart[8250],mmio,<addr>[,options] |
| 1137 | uart[8250],mmio32,<addr>[,options] |
| 1138 | uart[8250],mmio32be,<addr>[,options] |
| 1139 | uart[8250],0x<addr>[,options] |
| 1140 | Start an early, polled-mode console on the 8250/16550 |
| 1141 | UART at the specified I/O port or MMIO address. |
| 1142 | MMIO inter-register address stride is either 8-bit |
| 1143 | (mmio) or 32-bit (mmio32 or mmio32be). |
| 1144 | If none of [io|mmio|mmio32|mmio32be], <addr> is assumed |
| 1145 | to be equivalent to 'mmio'. 'options' are specified |
| 1146 | in the same format described for "console=ttyS<n>"; if |
| 1147 | unspecified, the h/w is not initialized. |
| 1148 | |
| 1149 | pl011,<addr> |
| 1150 | pl011,mmio32,<addr> |
| 1151 | Start an early, polled-mode console on a pl011 serial |
| 1152 | port at the specified address. The pl011 serial port |
| 1153 | must already be setup and configured. Options are not |
| 1154 | yet supported. If 'mmio32' is specified, then only |
| 1155 | the driver will use only 32-bit accessors to read/write |
| 1156 | the device registers. |
| 1157 | |
| 1158 | liteuart,<addr> |
| 1159 | Start an early console on a litex serial port at the |
| 1160 | specified address. The serial port must already be |
| 1161 | setup and configured. Options are not yet supported. |
| 1162 | |
| 1163 | meson,<addr> |
| 1164 | Start an early, polled-mode console on a meson serial |
| 1165 | port at the specified address. The serial port must |
| 1166 | already be setup and configured. Options are not yet |
| 1167 | supported. |
| 1168 | |
| 1169 | msm_serial,<addr> |
| 1170 | Start an early, polled-mode console on an msm serial |
| 1171 | port at the specified address. The serial port |
| 1172 | must already be setup and configured. Options are not |
| 1173 | yet supported. |
| 1174 | |
| 1175 | msm_serial_dm,<addr> |
| 1176 | Start an early, polled-mode console on an msm serial |
| 1177 | dm port at the specified address. The serial port |
| 1178 | must already be setup and configured. Options are not |
| 1179 | yet supported. |
| 1180 | |
| 1181 | owl,<addr> |
| 1182 | Start an early, polled-mode console on a serial port |
| 1183 | of an Actions Semi SoC, such as S500 or S900, at the |
| 1184 | specified address. The serial port must already be |
| 1185 | setup and configured. Options are not yet supported. |
| 1186 | |
| 1187 | rda,<addr> |
| 1188 | Start an early, polled-mode console on a serial port |
| 1189 | of an RDA Micro SoC, such as RDA8810PL, at the |
| 1190 | specified address. The serial port must already be |
| 1191 | setup and configured. Options are not yet supported. |
| 1192 | |
| 1193 | sbi |
| 1194 | Use RISC-V SBI (Supervisor Binary Interface) for early |
| 1195 | console. |
| 1196 | |
| 1197 | smh Use ARM semihosting calls for early console. |
| 1198 | |
| 1199 | s3c2410,<addr> |
| 1200 | s3c2412,<addr> |
| 1201 | s3c2440,<addr> |
| 1202 | s3c6400,<addr> |
| 1203 | s5pv210,<addr> |
| 1204 | exynos4210,<addr> |
| 1205 | Use early console provided by serial driver available |
| 1206 | on Samsung SoCs, requires selecting proper type and |
| 1207 | a correct base address of the selected UART port. The |
| 1208 | serial port must already be setup and configured. |
| 1209 | Options are not yet supported. |
| 1210 | |
| 1211 | lantiq,<addr> |
| 1212 | Start an early, polled-mode console on a lantiq serial |
| 1213 | (lqasc) port at the specified address. The serial port |
| 1214 | must already be setup and configured. Options are not |
| 1215 | yet supported. |
| 1216 | |
| 1217 | lpuart,<addr> |
| 1218 | lpuart32,<addr> |
| 1219 | Use early console provided by Freescale LP UART driver |
| 1220 | found on Freescale Vybrid and QorIQ LS1021A processors. |
| 1221 | A valid base address must be provided, and the serial |
| 1222 | port must already be setup and configured. |
| 1223 | |
| 1224 | ec_imx21,<addr> |
| 1225 | ec_imx6q,<addr> |
| 1226 | Start an early, polled-mode, output-only console on the |
| 1227 | Freescale i.MX UART at the specified address. The UART |
| 1228 | must already be setup and configured. |
| 1229 | |
| 1230 | ar3700_uart,<addr> |
| 1231 | Start an early, polled-mode console on the |
| 1232 | Armada 3700 serial port at the specified |
| 1233 | address. The serial port must already be setup |
| 1234 | and configured. Options are not yet supported. |
| 1235 | |
| 1236 | qcom_geni,<addr> |
| 1237 | Start an early, polled-mode console on a Qualcomm |
| 1238 | Generic Interface (GENI) based serial port at the |
| 1239 | specified address. The serial port must already be |
| 1240 | setup and configured. Options are not yet supported. |
| 1241 | |
| 1242 | efifb,[options] |
| 1243 | Start an early, unaccelerated console on the EFI |
| 1244 | memory mapped framebuffer (if available). On cache |
| 1245 | coherent non-x86 systems that use system memory for |
| 1246 | the framebuffer, pass the 'ram' option so that it is |
| 1247 | mapped with the correct attributes. |
| 1248 | |
| 1249 | linflex,<addr> |
| 1250 | Use early console provided by Freescale LINFlexD UART |
| 1251 | serial driver for NXP S32V234 SoCs. A valid base |
| 1252 | address must be provided, and the serial port must |
| 1253 | already be setup and configured. |
| 1254 | |
| 1255 | earlyprintk= [X86,SH,ARM,M68k,S390] |
| 1256 | earlyprintk=vga |
| 1257 | earlyprintk=sclp |
| 1258 | earlyprintk=xen |
| 1259 | earlyprintk=serial[,ttySn[,baudrate]] |
| 1260 | earlyprintk=serial[,0x...[,baudrate]] |
| 1261 | earlyprintk=ttySn[,baudrate] |
| 1262 | earlyprintk=dbgp[debugController#] |
| 1263 | earlyprintk=pciserial[,force],bus:device.function[,baudrate] |
| 1264 | earlyprintk=xdbc[xhciController#] |
| 1265 | |
| 1266 | earlyprintk is useful when the kernel crashes before |
| 1267 | the normal console is initialized. It is not enabled by |
| 1268 | default because it has some cosmetic problems. |
| 1269 | |
| 1270 | Append ",keep" to not disable it when the real console |
| 1271 | takes over. |
| 1272 | |
| 1273 | Only one of vga, efi, serial, or usb debug port can |
| 1274 | be used at a time. |
| 1275 | |
| 1276 | Currently only ttyS0 and ttyS1 may be specified by |
| 1277 | name. Other I/O ports may be explicitly specified |
| 1278 | on some architectures (x86 and arm at least) by |
| 1279 | replacing ttySn with an I/O port address, like this: |
| 1280 | earlyprintk=serial,0x1008,115200 |
| 1281 | You can find the port for a given device in |
| 1282 | /proc/tty/driver/serial: |
| 1283 | 2: uart:ST16650V2 port:00001008 irq:18 ... |
| 1284 | |
| 1285 | Interaction with the standard serial driver is not |
| 1286 | very good. |
| 1287 | |
| 1288 | The VGA and EFI output is eventually overwritten by |
| 1289 | the real console. |
| 1290 | |
| 1291 | The xen option can only be used in Xen domains. |
| 1292 | |
| 1293 | The sclp output can only be used on s390. |
| 1294 | |
| 1295 | The optional "force" to "pciserial" enables use of a |
| 1296 | PCI device even when its classcode is not of the |
| 1297 | UART class. |
| 1298 | |
| 1299 | edac_report= [HW,EDAC] Control how to report EDAC event |
| 1300 | Format: {"on" | "off" | "force"} |
| 1301 | on: enable EDAC to report H/W event. May be overridden |
| 1302 | by other higher priority error reporting module. |
| 1303 | off: disable H/W event reporting through EDAC. |
| 1304 | force: enforce the use of EDAC to report H/W event. |
| 1305 | default: on. |
| 1306 | |
| 1307 | ekgdboc= [X86,KGDB] Allow early kernel console debugging |
| 1308 | ekgdboc=kbd |
| 1309 | |
| 1310 | This is designed to be used in conjunction with |
| 1311 | the boot argument: earlyprintk=vga |
| 1312 | |
| 1313 | This parameter works in place of the kgdboc parameter |
| 1314 | but can only be used if the backing tty is available |
| 1315 | very early in the boot process. For early debugging |
| 1316 | via a serial port see kgdboc_earlycon instead. |
| 1317 | |
| 1318 | edd= [EDD] |
| 1319 | Format: {"off" | "on" | "skip[mbr]"} |
| 1320 | |
| 1321 | efi= [EFI] |
| 1322 | Format: { "debug", "disable_early_pci_dma", |
| 1323 | "nochunk", "noruntime", "nosoftreserve", |
| 1324 | "novamap", "no_disable_early_pci_dma" } |
| 1325 | debug: enable misc debug output. |
| 1326 | disable_early_pci_dma: disable the busmaster bit on all |
| 1327 | PCI bridges while in the EFI boot stub. |
| 1328 | nochunk: disable reading files in "chunks" in the EFI |
| 1329 | boot stub, as chunking can cause problems with some |
| 1330 | firmware implementations. |
| 1331 | noruntime : disable EFI runtime services support |
| 1332 | nosoftreserve: The EFI_MEMORY_SP (Specific Purpose) |
| 1333 | attribute may cause the kernel to reserve the |
| 1334 | memory range for a memory mapping driver to |
| 1335 | claim. Specify efi=nosoftreserve to disable this |
| 1336 | reservation and treat the memory by its base type |
| 1337 | (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM"). |
| 1338 | novamap: do not call SetVirtualAddressMap(). |
| 1339 | no_disable_early_pci_dma: Leave the busmaster bit set |
| 1340 | on all PCI bridges while in the EFI boot stub |
| 1341 | |
| 1342 | efi_no_storage_paranoia [EFI; X86] |
| 1343 | Using this parameter you can use more than 50% of |
| 1344 | your efi variable storage. Use this parameter only if |
| 1345 | you are really sure that your UEFI does sane gc and |
| 1346 | fulfills the spec otherwise your board may brick. |
| 1347 | |
| 1348 | efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86] |
| 1349 | Add arbitrary attribute to specific memory range by |
| 1350 | updating original EFI memory map. |
| 1351 | Region of memory which aa attribute is added to is |
| 1352 | from ss to ss+nn. |
| 1353 | |
| 1354 | If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000 |
| 1355 | is specified, EFI_MEMORY_MORE_RELIABLE(0x10000) |
| 1356 | attribute is added to range 0x100000000-0x180000000 and |
| 1357 | 0x10a0000000-0x1120000000. |
| 1358 | |
| 1359 | If efi_fake_mem=8G@9G:0x40000 is specified, the |
| 1360 | EFI_MEMORY_SP(0x40000) attribute is added to |
| 1361 | range 0x240000000-0x43fffffff. |
| 1362 | |
| 1363 | Using this parameter you can do debugging of EFI memmap |
| 1364 | related features. For example, you can do debugging of |
| 1365 | Address Range Mirroring feature even if your box |
| 1366 | doesn't support it, or mark specific memory as |
| 1367 | "soft reserved". |
| 1368 | |
| 1369 | efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT |
| 1370 | that is to be dynamically loaded by Linux. If there are |
| 1371 | multiple variables with the same name but with different |
| 1372 | vendor GUIDs, all of them will be loaded. See |
| 1373 | Documentation/admin-guide/acpi/ssdt-overlays.rst for details. |
| 1374 | |
| 1375 | |
| 1376 | eisa_irq_edge= [PARISC,HW] |
| 1377 | See header of drivers/parisc/eisa.c. |
| 1378 | |
| 1379 | elanfreq= [X86-32] |
| 1380 | See comment before function elanfreq_setup() in |
| 1381 | arch/x86/kernel/cpu/cpufreq/elanfreq.c. |
| 1382 | |
| 1383 | elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390] |
| 1384 | Specifies physical address of start of kernel core |
| 1385 | image elf header and optionally the size. Generally |
| 1386 | kexec loader will pass this option to capture kernel. |
| 1387 | See Documentation/admin-guide/kdump/kdump.rst for details. |
| 1388 | |
| 1389 | enable_mtrr_cleanup [X86] |
| 1390 | The kernel tries to adjust MTRR layout from continuous |
| 1391 | to discrete, to make X server driver able to add WB |
| 1392 | entry later. This parameter enables that. |
| 1393 | |
| 1394 | enable_timer_pin_1 [X86] |
| 1395 | Enable PIN 1 of APIC timer |
| 1396 | Can be useful to work around chipset bugs |
| 1397 | (in particular on some ATI chipsets). |
| 1398 | The kernel tries to set a reasonable default. |
| 1399 | |
| 1400 | enforcing [SELINUX] Set initial enforcing status. |
| 1401 | Format: {"0" | "1"} |
| 1402 | See security/selinux/Kconfig help text. |
| 1403 | 0 -- permissive (log only, no denials). |
| 1404 | 1 -- enforcing (deny and log). |
| 1405 | Default value is 0. |
| 1406 | Value can be changed at runtime via |
| 1407 | /sys/fs/selinux/enforce. |
| 1408 | |
| 1409 | erst_disable [ACPI] |
| 1410 | Disable Error Record Serialization Table (ERST) |
| 1411 | support. |
| 1412 | |
| 1413 | ether= [HW,NET] Ethernet cards parameters |
| 1414 | This option is obsoleted by the "netdev=" option, which |
| 1415 | has equivalent usage. See its documentation for details. |
| 1416 | |
| 1417 | evm= [EVM] |
| 1418 | Format: { "fix" } |
| 1419 | Permit 'security.evm' to be updated regardless of |
| 1420 | current integrity status. |
| 1421 | |
| 1422 | failslab= |
| 1423 | fail_usercopy= |
| 1424 | fail_page_alloc= |
| 1425 | fail_make_request=[KNL] |
| 1426 | General fault injection mechanism. |
| 1427 | Format: <interval>,<probability>,<space>,<times> |
| 1428 | See also Documentation/fault-injection/. |
| 1429 | |
| 1430 | fb_tunnels= [NET] |
| 1431 | Format: { initns | none } |
| 1432 | See Documentation/admin-guide/sysctl/net.rst for |
| 1433 | fb_tunnels_only_for_init_ns |
| 1434 | |
| 1435 | floppy= [HW] |
| 1436 | See Documentation/admin-guide/blockdev/floppy.rst. |
| 1437 | |
| 1438 | force_pal_cache_flush |
| 1439 | [IA-64] Avoid check_sal_cache_flush which may hang on |
| 1440 | buggy SAL_CACHE_FLUSH implementations. Using this |
| 1441 | parameter will force ia64_sal_cache_flush to call |
| 1442 | ia64_pal_cache_flush instead of SAL_CACHE_FLUSH. |
| 1443 | |
| 1444 | forcepae [X86-32] |
| 1445 | Forcefully enable Physical Address Extension (PAE). |
| 1446 | Many Pentium M systems disable PAE but may have a |
| 1447 | functionally usable PAE implementation. |
| 1448 | Warning: use of this parameter will taint the kernel |
| 1449 | and may cause unknown problems. |
| 1450 | |
| 1451 | ftrace=[tracer] |
| 1452 | [FTRACE] will set and start the specified tracer |
| 1453 | as early as possible in order to facilitate early |
| 1454 | boot debugging. |
| 1455 | |
| 1456 | ftrace_dump_on_oops[=orig_cpu] |
| 1457 | [FTRACE] will dump the trace buffers on oops. |
| 1458 | If no parameter is passed, ftrace will dump |
| 1459 | buffers of all CPUs, but if you pass orig_cpu, it will |
| 1460 | dump only the buffer of the CPU that triggered the |
| 1461 | oops. |
| 1462 | |
| 1463 | ftrace_filter=[function-list] |
| 1464 | [FTRACE] Limit the functions traced by the function |
| 1465 | tracer at boot up. function-list is a comma-separated |
| 1466 | list of functions. This list can be changed at run |
| 1467 | time by the set_ftrace_filter file in the debugfs |
| 1468 | tracing directory. |
| 1469 | |
| 1470 | ftrace_notrace=[function-list] |
| 1471 | [FTRACE] Do not trace the functions specified in |
| 1472 | function-list. This list can be changed at run time |
| 1473 | by the set_ftrace_notrace file in the debugfs |
| 1474 | tracing directory. |
| 1475 | |
| 1476 | ftrace_graph_filter=[function-list] |
| 1477 | [FTRACE] Limit the top level callers functions traced |
| 1478 | by the function graph tracer at boot up. |
| 1479 | function-list is a comma-separated list of functions |
| 1480 | that can be changed at run time by the |
| 1481 | set_graph_function file in the debugfs tracing directory. |
| 1482 | |
| 1483 | ftrace_graph_notrace=[function-list] |
| 1484 | [FTRACE] Do not trace from the functions specified in |
| 1485 | function-list. This list is a comma-separated list of |
| 1486 | functions that can be changed at run time by the |
| 1487 | set_graph_notrace file in the debugfs tracing directory. |
| 1488 | |
| 1489 | ftrace_graph_max_depth=<uint> |
| 1490 | [FTRACE] Used with the function graph tracer. This is |
| 1491 | the max depth it will trace into a function. This value |
| 1492 | can be changed at run time by the max_graph_depth file |
| 1493 | in the tracefs tracing directory. default: 0 (no limit) |
| 1494 | |
| 1495 | fw_devlink= [KNL] Create device links between consumer and supplier |
| 1496 | devices by scanning the firmware to infer the |
| 1497 | consumer/supplier relationships. This feature is |
| 1498 | especially useful when drivers are loaded as modules as |
| 1499 | it ensures proper ordering of tasks like device probing |
| 1500 | (suppliers first, then consumers), supplier boot state |
| 1501 | clean up (only after all consumers have probed), |
| 1502 | suspend/resume & runtime PM (consumers first, then |
| 1503 | suppliers). |
| 1504 | Format: { off | permissive | on | rpm } |
| 1505 | off -- Don't create device links from firmware info. |
| 1506 | permissive -- Create device links from firmware info |
| 1507 | but use it only for ordering boot state clean |
| 1508 | up (sync_state() calls). |
| 1509 | on -- Create device links from firmware info and use it |
| 1510 | to enforce probe and suspend/resume ordering. |
| 1511 | rpm -- Like "on", but also use to order runtime PM. |
| 1512 | |
| 1513 | fw_devlink.strict=<bool> |
| 1514 | [KNL] Treat all inferred dependencies as mandatory |
| 1515 | dependencies. This only applies for fw_devlink=on|rpm. |
| 1516 | Format: <bool> |
| 1517 | |
| 1518 | gamecon.map[2|3]= |
| 1519 | [HW,JOY] Multisystem joystick and NES/SNES/PSX pad |
| 1520 | support via parallel port (up to 5 devices per port) |
| 1521 | Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5> |
| 1522 | See also Documentation/input/devices/joystick-parport.rst |
| 1523 | |
| 1524 | gamma= [HW,DRM] |
| 1525 | |
| 1526 | gart_fix_e820= [X86-64] disable the fix e820 for K8 GART |
| 1527 | Format: off | on |
| 1528 | default: on |
| 1529 | |
| 1530 | gcov_persist= [GCOV] When non-zero (default), profiling data for |
| 1531 | kernel modules is saved and remains accessible via |
| 1532 | debugfs, even when the module is unloaded/reloaded. |
| 1533 | When zero, profiling data is discarded and associated |
| 1534 | debugfs files are removed at module unload time. |
| 1535 | |
| 1536 | goldfish [X86] Enable the goldfish android emulator platform. |
| 1537 | Don't use this when you are not running on the |
| 1538 | android emulator |
| 1539 | |
| 1540 | gpio-mockup.gpio_mockup_ranges |
| 1541 | [HW] Sets the ranges of gpiochip of for this device. |
| 1542 | Format: <start1>,<end1>,<start2>,<end2>... |
| 1543 | gpio-mockup.gpio_mockup_named_lines |
| 1544 | [HW] Let the driver know GPIO lines should be named. |
| 1545 | |
| 1546 | gpt [EFI] Forces disk with valid GPT signature but |
| 1547 | invalid Protective MBR to be treated as GPT. If the |
| 1548 | primary GPT is corrupted, it enables the backup/alternate |
| 1549 | GPT to be used instead. |
| 1550 | |
| 1551 | grcan.enable0= [HW] Configuration of physical interface 0. Determines |
| 1552 | the "Enable 0" bit of the configuration register. |
| 1553 | Format: 0 | 1 |
| 1554 | Default: 0 |
| 1555 | grcan.enable1= [HW] Configuration of physical interface 1. Determines |
| 1556 | the "Enable 0" bit of the configuration register. |
| 1557 | Format: 0 | 1 |
| 1558 | Default: 0 |
| 1559 | grcan.select= [HW] Select which physical interface to use. |
| 1560 | Format: 0 | 1 |
| 1561 | Default: 0 |
| 1562 | grcan.txsize= [HW] Sets the size of the tx buffer. |
| 1563 | Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0. |
| 1564 | Default: 1024 |
| 1565 | grcan.rxsize= [HW] Sets the size of the rx buffer. |
| 1566 | Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0. |
| 1567 | Default: 1024 |
| 1568 | |
| 1569 | hardlockup_all_cpu_backtrace= |
| 1570 | [KNL] Should the hard-lockup detector generate |
| 1571 | backtraces on all cpus. |
| 1572 | Format: 0 | 1 |
| 1573 | |
| 1574 | hashdist= [KNL,NUMA] Large hashes allocated during boot |
| 1575 | are distributed across NUMA nodes. Defaults on |
| 1576 | for 64-bit NUMA, off otherwise. |
| 1577 | Format: 0 | 1 (for off | on) |
| 1578 | |
| 1579 | hcl= [IA-64] SGI's Hardware Graph compatibility layer |
| 1580 | |
| 1581 | hd= [EIDE] (E)IDE hard drive subsystem geometry |
| 1582 | Format: <cyl>,<head>,<sect> |
| 1583 | |
| 1584 | hest_disable [ACPI] |
| 1585 | Disable Hardware Error Source Table (HEST) support; |
| 1586 | corresponding firmware-first mode error processing |
| 1587 | logic will be disabled. |
| 1588 | |
| 1589 | highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact |
| 1590 | size of <nn>. This works even on boxes that have no |
| 1591 | highmem otherwise. This also works to reduce highmem |
| 1592 | size on bigger boxes. |
| 1593 | |
| 1594 | highres= [KNL] Enable/disable high resolution timer mode. |
| 1595 | Valid parameters: "on", "off" |
| 1596 | Default: "on" |
| 1597 | |
| 1598 | hlt [BUGS=ARM,SH] |
| 1599 | |
| 1600 | hpet= [X86-32,HPET] option to control HPET usage |
| 1601 | Format: { enable (default) | disable | force | |
| 1602 | verbose } |
| 1603 | disable: disable HPET and use PIT instead |
| 1604 | force: allow force enabled of undocumented chips (ICH4, |
| 1605 | VIA, nVidia) |
| 1606 | verbose: show contents of HPET registers during setup |
| 1607 | |
| 1608 | hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET |
| 1609 | registers. Default set by CONFIG_HPET_MMAP_DEFAULT. |
| 1610 | |
| 1611 | hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation |
| 1612 | of gigantic hugepages. Or using node format, the size |
| 1613 | of a CMA area per node can be specified. |
| 1614 | Format: nn[KMGTPE] or (node format) |
| 1615 | <node>:nn[KMGTPE][,<node>:nn[KMGTPE]] |
| 1616 | |
| 1617 | Reserve a CMA area of given size and allocate gigantic |
| 1618 | hugepages using the CMA allocator. If enabled, the |
| 1619 | boot-time allocation of gigantic hugepages is skipped. |
| 1620 | |
| 1621 | hugepages= [HW] Number of HugeTLB pages to allocate at boot. |
| 1622 | If this follows hugepagesz (below), it specifies |
| 1623 | the number of pages of hugepagesz to be allocated. |
| 1624 | If this is the first HugeTLB parameter on the command |
| 1625 | line, it specifies the number of pages to allocate for |
| 1626 | the default huge page size. If using node format, the |
| 1627 | number of pages to allocate per-node can be specified. |
| 1628 | See also Documentation/admin-guide/mm/hugetlbpage.rst. |
| 1629 | Format: <integer> or (node format) |
| 1630 | <node>:<integer>[,<node>:<integer>] |
| 1631 | |
| 1632 | hugepagesz= |
| 1633 | [HW] The size of the HugeTLB pages. This is used in |
| 1634 | conjunction with hugepages (above) to allocate huge |
| 1635 | pages of a specific size at boot. The pair |
| 1636 | hugepagesz=X hugepages=Y can be specified once for |
| 1637 | each supported huge page size. Huge page sizes are |
| 1638 | architecture dependent. See also |
| 1639 | Documentation/admin-guide/mm/hugetlbpage.rst. |
| 1640 | Format: size[KMG] |
| 1641 | |
| 1642 | hugetlb_free_vmemmap= |
| 1643 | [KNL] Reguires CONFIG_HUGETLB_PAGE_FREE_VMEMMAP |
| 1644 | enabled. |
| 1645 | Allows heavy hugetlb users to free up some more |
| 1646 | memory (6 * PAGE_SIZE for each 2MB hugetlb page). |
| 1647 | Format: { on | off (default) } |
| 1648 | |
| 1649 | on: enable the feature |
| 1650 | off: disable the feature |
| 1651 | |
| 1652 | Built with CONFIG_HUGETLB_PAGE_FREE_VMEMMAP_DEFAULT_ON=y, |
| 1653 | the default is on. |
| 1654 | |
| 1655 | This is not compatible with memory_hotplug.memmap_on_memory. |
| 1656 | If both parameters are enabled, hugetlb_free_vmemmap takes |
| 1657 | precedence over memory_hotplug.memmap_on_memory. |
| 1658 | |
| 1659 | hung_task_panic= |
| 1660 | [KNL] Should the hung task detector generate panics. |
| 1661 | Format: 0 | 1 |
| 1662 | |
| 1663 | A value of 1 instructs the kernel to panic when a |
| 1664 | hung task is detected. The default value is controlled |
| 1665 | by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time |
| 1666 | option. The value selected by this boot parameter can |
| 1667 | be changed later by the kernel.hung_task_panic sysctl. |
| 1668 | |
| 1669 | hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC) |
| 1670 | terminal devices. Valid values: 0..8 |
| 1671 | hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs. |
| 1672 | If specified, z/VM IUCV HVC accepts connections |
| 1673 | from listed z/VM user IDs only. |
| 1674 | |
| 1675 | hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations |
| 1676 | which allow the hypervisor to 'idle' the |
| 1677 | guest on lock contention. |
| 1678 | |
| 1679 | keep_bootcon [KNL] |
| 1680 | Do not unregister boot console at start. This is only |
| 1681 | useful for debugging when something happens in the window |
| 1682 | between unregistering the boot console and initializing |
| 1683 | the real console. |
| 1684 | |
| 1685 | i2c_bus= [HW] Override the default board specific I2C bus speed |
| 1686 | or register an additional I2C bus that is not |
| 1687 | registered from board initialization code. |
| 1688 | Format: |
| 1689 | <bus_id>,<clkrate> |
| 1690 | |
| 1691 | i8042.debug [HW] Toggle i8042 debug mode |
| 1692 | i8042.unmask_kbd_data |
| 1693 | [HW] Enable printing of interrupt data from the KBD port |
| 1694 | (disabled by default, and as a pre-condition |
| 1695 | requires that i8042.debug=1 be enabled) |
| 1696 | i8042.direct [HW] Put keyboard port into non-translated mode |
| 1697 | i8042.dumbkbd [HW] Pretend that controller can only read data from |
| 1698 | keyboard and cannot control its state |
| 1699 | (Don't attempt to blink the leds) |
| 1700 | i8042.noaux [HW] Don't check for auxiliary (== mouse) port |
| 1701 | i8042.nokbd [HW] Don't check/create keyboard port |
| 1702 | i8042.noloop [HW] Disable the AUX Loopback command while probing |
| 1703 | for the AUX port |
| 1704 | i8042.nomux [HW] Don't check presence of an active multiplexing |
| 1705 | controller |
| 1706 | i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX |
| 1707 | controllers |
| 1708 | i8042.notimeout [HW] Ignore timeout condition signalled by controller |
| 1709 | i8042.reset [HW] Reset the controller during init, cleanup and |
| 1710 | suspend-to-ram transitions, only during s2r |
| 1711 | transitions, or never reset |
| 1712 | Format: { 1 | Y | y | 0 | N | n } |
| 1713 | 1, Y, y: always reset controller |
| 1714 | 0, N, n: don't ever reset controller |
| 1715 | Default: only on s2r transitions on x86; most other |
| 1716 | architectures force reset to be always executed |
| 1717 | i8042.unlock [HW] Unlock (ignore) the keylock |
| 1718 | i8042.kbdreset [HW] Reset device connected to KBD port |
| 1719 | i8042.probe_defer |
| 1720 | [HW] Allow deferred probing upon i8042 probe errors |
| 1721 | |
| 1722 | i810= [HW,DRM] |
| 1723 | |
| 1724 | i915.invert_brightness= |
| 1725 | [DRM] Invert the sense of the variable that is used to |
| 1726 | set the brightness of the panel backlight. Normally a |
| 1727 | brightness value of 0 indicates backlight switched off, |
| 1728 | and the maximum of the brightness value sets the backlight |
| 1729 | to maximum brightness. If this parameter is set to 0 |
| 1730 | (default) and the machine requires it, or this parameter |
| 1731 | is set to 1, a brightness value of 0 sets the backlight |
| 1732 | to maximum brightness, and the maximum of the brightness |
| 1733 | value switches the backlight off. |
| 1734 | -1 -- never invert brightness |
| 1735 | 0 -- machine default |
| 1736 | 1 -- force brightness inversion |
| 1737 | |
| 1738 | icn= [HW,ISDN] |
| 1739 | Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]] |
| 1740 | |
| 1741 | ide-core.nodma= [HW] (E)IDE subsystem |
| 1742 | Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc |
| 1743 | .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr |
| 1744 | .cdrom .chs .ignore_cable are additional options |
| 1745 | See Documentation/ide/ide.rst. |
| 1746 | |
| 1747 | ide-generic.probe-mask= [HW] (E)IDE subsystem |
| 1748 | Format: <int> |
| 1749 | Probe mask for legacy ISA IDE ports. Depending on |
| 1750 | platform up to 6 ports are supported, enabled by |
| 1751 | setting corresponding bits in the mask to 1. The |
| 1752 | default value is 0x0, which has a special meaning. |
| 1753 | On systems that have PCI, it triggers scanning the |
| 1754 | PCI bus for the first and the second port, which |
| 1755 | are then probed. On systems without PCI the value |
| 1756 | of 0x0 enables probing the two first ports as if it |
| 1757 | was 0x3. |
| 1758 | |
| 1759 | ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem |
| 1760 | Claim all unknown PCI IDE storage controllers. |
| 1761 | |
| 1762 | idle= [X86] |
| 1763 | Format: idle=poll, idle=halt, idle=nomwait |
| 1764 | Poll forces a polling idle loop that can slightly |
| 1765 | improve the performance of waking up a idle CPU, but |
| 1766 | will use a lot of power and make the system run hot. |
| 1767 | Not recommended. |
| 1768 | idle=halt: Halt is forced to be used for CPU idle. |
| 1769 | In such case C2/C3 won't be used again. |
| 1770 | idle=nomwait: Disable mwait for CPU C-states |
| 1771 | |
| 1772 | idxd.sva= [HW] |
| 1773 | Format: <bool> |
| 1774 | Allow force disabling of Shared Virtual Memory (SVA) |
| 1775 | support for the idxd driver. By default it is set to |
| 1776 | true (1). |
| 1777 | |
| 1778 | idxd.tc_override= [HW] |
| 1779 | Format: <bool> |
| 1780 | Allow override of default traffic class configuration |
| 1781 | for the device. By default it is set to false (0). |
| 1782 | |
| 1783 | ieee754= [MIPS] Select IEEE Std 754 conformance mode |
| 1784 | Format: { strict | legacy | 2008 | relaxed } |
| 1785 | Default: strict |
| 1786 | |
| 1787 | Choose which programs will be accepted for execution |
| 1788 | based on the IEEE 754 NaN encoding(s) supported by |
| 1789 | the FPU and the NaN encoding requested with the value |
| 1790 | of an ELF file header flag individually set by each |
| 1791 | binary. Hardware implementations are permitted to |
| 1792 | support either or both of the legacy and the 2008 NaN |
| 1793 | encoding mode. |
| 1794 | |
| 1795 | Available settings are as follows: |
| 1796 | strict accept binaries that request a NaN encoding |
| 1797 | supported by the FPU |
| 1798 | legacy only accept legacy-NaN binaries, if supported |
| 1799 | by the FPU |
| 1800 | 2008 only accept 2008-NaN binaries, if supported |
| 1801 | by the FPU |
| 1802 | relaxed accept any binaries regardless of whether |
| 1803 | supported by the FPU |
| 1804 | |
| 1805 | The FPU emulator is always able to support both NaN |
| 1806 | encodings, so if no FPU hardware is present or it has |
| 1807 | been disabled with 'nofpu', then the settings of |
| 1808 | 'legacy' and '2008' strap the emulator accordingly, |
| 1809 | 'relaxed' straps the emulator for both legacy-NaN and |
| 1810 | 2008-NaN, whereas 'strict' enables legacy-NaN only on |
| 1811 | legacy processors and both NaN encodings on MIPS32 or |
| 1812 | MIPS64 CPUs. |
| 1813 | |
| 1814 | The setting for ABS.fmt/NEG.fmt instruction execution |
| 1815 | mode generally follows that for the NaN encoding, |
| 1816 | except where unsupported by hardware. |
| 1817 | |
| 1818 | ignore_loglevel [KNL] |
| 1819 | Ignore loglevel setting - this will print /all/ |
| 1820 | kernel messages to the console. Useful for debugging. |
| 1821 | We also add it as printk module parameter, so users |
| 1822 | could change it dynamically, usually by |
| 1823 | /sys/module/printk/parameters/ignore_loglevel. |
| 1824 | |
| 1825 | ignore_rlimit_data |
| 1826 | Ignore RLIMIT_DATA setting for data mappings, |
| 1827 | print warning at first misuse. Can be changed via |
| 1828 | /sys/module/kernel/parameters/ignore_rlimit_data. |
| 1829 | |
| 1830 | ihash_entries= [KNL] |
| 1831 | Set number of hash buckets for inode cache. |
| 1832 | |
| 1833 | ima_appraise= [IMA] appraise integrity measurements |
| 1834 | Format: { "off" | "enforce" | "fix" | "log" } |
| 1835 | default: "enforce" |
| 1836 | |
| 1837 | ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead. |
| 1838 | The builtin appraise policy appraises all files |
| 1839 | owned by uid=0. |
| 1840 | |
| 1841 | ima_canonical_fmt [IMA] |
| 1842 | Use the canonical format for the binary runtime |
| 1843 | measurements, instead of host native format. |
| 1844 | |
| 1845 | ima_hash= [IMA] |
| 1846 | Format: { md5 | sha1 | rmd160 | sha256 | sha384 |
| 1847 | | sha512 | ... } |
| 1848 | default: "sha1" |
| 1849 | |
| 1850 | The list of supported hash algorithms is defined |
| 1851 | in crypto/hash_info.h. |
| 1852 | |
| 1853 | ima_policy= [IMA] |
| 1854 | The builtin policies to load during IMA setup. |
| 1855 | Format: "tcb | appraise_tcb | secure_boot | |
| 1856 | fail_securely | critical_data" |
| 1857 | |
| 1858 | The "tcb" policy measures all programs exec'd, files |
| 1859 | mmap'd for exec, and all files opened with the read |
| 1860 | mode bit set by either the effective uid (euid=0) or |
| 1861 | uid=0. |
| 1862 | |
| 1863 | The "appraise_tcb" policy appraises the integrity of |
| 1864 | all files owned by root. |
| 1865 | |
| 1866 | The "secure_boot" policy appraises the integrity |
| 1867 | of files (eg. kexec kernel image, kernel modules, |
| 1868 | firmware, policy, etc) based on file signatures. |
| 1869 | |
| 1870 | The "fail_securely" policy forces file signature |
| 1871 | verification failure also on privileged mounted |
| 1872 | filesystems with the SB_I_UNVERIFIABLE_SIGNATURE |
| 1873 | flag. |
| 1874 | |
| 1875 | The "critical_data" policy measures kernel integrity |
| 1876 | critical data. |
| 1877 | |
| 1878 | ima_tcb [IMA] Deprecated. Use ima_policy= instead. |
| 1879 | Load a policy which meets the needs of the Trusted |
| 1880 | Computing Base. This means IMA will measure all |
| 1881 | programs exec'd, files mmap'd for exec, and all files |
| 1882 | opened for read by uid=0. |
| 1883 | |
| 1884 | ima_template= [IMA] |
| 1885 | Select one of defined IMA measurements template formats. |
| 1886 | Formats: { "ima" | "ima-ng" | "ima-sig" } |
| 1887 | Default: "ima-ng" |
| 1888 | |
| 1889 | ima_template_fmt= |
| 1890 | [IMA] Define a custom template format. |
| 1891 | Format: { "field1|...|fieldN" } |
| 1892 | |
| 1893 | ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage |
| 1894 | Format: <min_file_size> |
| 1895 | Set the minimal file size for using asynchronous hash. |
| 1896 | If left unspecified, ahash usage is disabled. |
| 1897 | |
| 1898 | ahash performance varies for different data sizes on |
| 1899 | different crypto accelerators. This option can be used |
| 1900 | to achieve the best performance for a particular HW. |
| 1901 | |
| 1902 | ima.ahash_bufsize= [IMA] Asynchronous hash buffer size |
| 1903 | Format: <bufsize> |
| 1904 | Set hashing buffer size. Default: 4k. |
| 1905 | |
| 1906 | ahash performance varies for different chunk sizes on |
| 1907 | different crypto accelerators. This option can be used |
| 1908 | to achieve best performance for particular HW. |
| 1909 | |
| 1910 | init= [KNL] |
| 1911 | Format: <full_path> |
| 1912 | Run specified binary instead of /sbin/init as init |
| 1913 | process. |
| 1914 | |
| 1915 | initcall_debug [KNL] Trace initcalls as they are executed. Useful |
| 1916 | for working out where the kernel is dying during |
| 1917 | startup. |
| 1918 | |
| 1919 | initcall_blacklist= [KNL] Do not execute a comma-separated list of |
| 1920 | initcall functions. Useful for debugging built-in |
| 1921 | modules and initcalls. |
| 1922 | |
| 1923 | initramfs_async= [KNL] |
| 1924 | Format: <bool> |
| 1925 | Default: 1 |
| 1926 | This parameter controls whether the initramfs |
| 1927 | image is unpacked asynchronously, concurrently |
| 1928 | with devices being probed and |
| 1929 | initialized. This should normally just work, |
| 1930 | but as a debugging aid, one can get the |
| 1931 | historical behaviour of the initramfs |
| 1932 | unpacking being completed before device_ and |
| 1933 | late_ initcalls. |
| 1934 | |
| 1935 | initrd= [BOOT] Specify the location of the initial ramdisk |
| 1936 | |
| 1937 | initrdmem= [KNL] Specify a physical address and size from which to |
| 1938 | load the initrd. If an initrd is compiled in or |
| 1939 | specified in the bootparams, it takes priority over this |
| 1940 | setting. |
| 1941 | Format: ss[KMG],nn[KMG] |
| 1942 | Default is 0, 0 |
| 1943 | |
| 1944 | init_on_alloc= [MM] Fill newly allocated pages and heap objects with |
| 1945 | zeroes. |
| 1946 | Format: 0 | 1 |
| 1947 | Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON. |
| 1948 | |
| 1949 | init_on_free= [MM] Fill freed pages and heap objects with zeroes. |
| 1950 | Format: 0 | 1 |
| 1951 | Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON. |
| 1952 | |
| 1953 | init_pkru= [X86] Specify the default memory protection keys rights |
| 1954 | register contents for all processes. 0x55555554 by |
| 1955 | default (disallow access to all but pkey 0). Can |
| 1956 | override in debugfs after boot. |
| 1957 | |
| 1958 | inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver |
| 1959 | Format: <irq> |
| 1960 | |
| 1961 | int_pln_enable [X86] Enable power limit notification interrupt |
| 1962 | |
| 1963 | integrity_audit=[IMA] |
| 1964 | Format: { "0" | "1" } |
| 1965 | 0 -- basic integrity auditing messages. (Default) |
| 1966 | 1 -- additional integrity auditing messages. |
| 1967 | |
| 1968 | intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option |
| 1969 | on |
| 1970 | Enable intel iommu driver. |
| 1971 | off |
| 1972 | Disable intel iommu driver. |
| 1973 | igfx_off [Default Off] |
| 1974 | By default, gfx is mapped as normal device. If a gfx |
| 1975 | device has a dedicated DMAR unit, the DMAR unit is |
| 1976 | bypassed by not enabling DMAR with this option. In |
| 1977 | this case, gfx device will use physical address for |
| 1978 | DMA. |
| 1979 | strict [Default Off] |
| 1980 | Deprecated, equivalent to iommu.strict=1. |
| 1981 | sp_off [Default Off] |
| 1982 | By default, super page will be supported if Intel IOMMU |
| 1983 | has the capability. With this option, super page will |
| 1984 | not be supported. |
| 1985 | sm_on |
| 1986 | Enable the Intel IOMMU scalable mode if the hardware |
| 1987 | advertises that it has support for the scalable mode |
| 1988 | translation. |
| 1989 | sm_off |
| 1990 | Disallow use of the Intel IOMMU scalable mode. |
| 1991 | tboot_noforce [Default Off] |
| 1992 | Do not force the Intel IOMMU enabled under tboot. |
| 1993 | By default, tboot will force Intel IOMMU on, which |
| 1994 | could harm performance of some high-throughput |
| 1995 | devices like 40GBit network cards, even if identity |
| 1996 | mapping is enabled. |
| 1997 | Note that using this option lowers the security |
| 1998 | provided by tboot because it makes the system |
| 1999 | vulnerable to DMA attacks. |
| 2000 | |
| 2001 | intel_idle.max_cstate= [KNL,HW,ACPI,X86] |
| 2002 | 0 disables intel_idle and fall back on acpi_idle. |
| 2003 | 1 to 9 specify maximum depth of C-state. |
| 2004 | |
| 2005 | intel_pstate= [X86] |
| 2006 | disable |
| 2007 | Do not enable intel_pstate as the default |
| 2008 | scaling driver for the supported processors |
| 2009 | passive |
| 2010 | Use intel_pstate as a scaling driver, but configure it |
| 2011 | to work with generic cpufreq governors (instead of |
| 2012 | enabling its internal governor). This mode cannot be |
| 2013 | used along with the hardware-managed P-states (HWP) |
| 2014 | feature. |
| 2015 | force |
| 2016 | Enable intel_pstate on systems that prohibit it by default |
| 2017 | in favor of acpi-cpufreq. Forcing the intel_pstate driver |
| 2018 | instead of acpi-cpufreq may disable platform features, such |
| 2019 | as thermal controls and power capping, that rely on ACPI |
| 2020 | P-States information being indicated to OSPM and therefore |
| 2021 | should be used with caution. This option does not work with |
| 2022 | processors that aren't supported by the intel_pstate driver |
| 2023 | or on platforms that use pcc-cpufreq instead of acpi-cpufreq. |
| 2024 | no_hwp |
| 2025 | Do not enable hardware P state control (HWP) |
| 2026 | if available. |
| 2027 | hwp_only |
| 2028 | Only load intel_pstate on systems which support |
| 2029 | hardware P state control (HWP) if available. |
| 2030 | support_acpi_ppc |
| 2031 | Enforce ACPI _PPC performance limits. If the Fixed ACPI |
| 2032 | Description Table, specifies preferred power management |
| 2033 | profile as "Enterprise Server" or "Performance Server", |
| 2034 | then this feature is turned on by default. |
| 2035 | per_cpu_perf_limits |
| 2036 | Allow per-logical-CPU P-State performance control limits using |
| 2037 | cpufreq sysfs interface |
| 2038 | |
| 2039 | intremap= [X86-64, Intel-IOMMU] |
| 2040 | on enable Interrupt Remapping (default) |
| 2041 | off disable Interrupt Remapping |
| 2042 | nosid disable Source ID checking |
| 2043 | no_x2apic_optout |
| 2044 | BIOS x2APIC opt-out request will be ignored |
| 2045 | nopost disable Interrupt Posting |
| 2046 | |
| 2047 | iomem= Disable strict checking of access to MMIO memory |
| 2048 | strict regions from userspace. |
| 2049 | relaxed |
| 2050 | |
| 2051 | iommu= [X86] |
| 2052 | off |
| 2053 | force |
| 2054 | noforce |
| 2055 | biomerge |
| 2056 | panic |
| 2057 | nopanic |
| 2058 | merge |
| 2059 | nomerge |
| 2060 | soft |
| 2061 | pt [X86] |
| 2062 | nopt [X86] |
| 2063 | nobypass [PPC/POWERNV] |
| 2064 | Disable IOMMU bypass, using IOMMU for PCI devices. |
| 2065 | |
| 2066 | iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices. |
| 2067 | Format: { "0" | "1" } |
| 2068 | 0 - Try to allocate a 32-bit DMA address first, before |
| 2069 | falling back to the full range if needed. |
| 2070 | 1 - Allocate directly from the full usable range, |
| 2071 | forcing Dual Address Cycle for PCI cards supporting |
| 2072 | greater than 32-bit addressing. |
| 2073 | |
| 2074 | iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour |
| 2075 | Format: { "0" | "1" } |
| 2076 | 0 - Lazy mode. |
| 2077 | Request that DMA unmap operations use deferred |
| 2078 | invalidation of hardware TLBs, for increased |
| 2079 | throughput at the cost of reduced device isolation. |
| 2080 | Will fall back to strict mode if not supported by |
| 2081 | the relevant IOMMU driver. |
| 2082 | 1 - Strict mode. |
| 2083 | DMA unmap operations invalidate IOMMU hardware TLBs |
| 2084 | synchronously. |
| 2085 | unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}. |
| 2086 | Note: on x86, strict mode specified via one of the |
| 2087 | legacy driver-specific options takes precedence. |
| 2088 | |
| 2089 | iommu.passthrough= |
| 2090 | [ARM64, X86] Configure DMA to bypass the IOMMU by default. |
| 2091 | Format: { "0" | "1" } |
| 2092 | 0 - Use IOMMU translation for DMA. |
| 2093 | 1 - Bypass the IOMMU for DMA. |
| 2094 | unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH. |
| 2095 | |
| 2096 | io7= [HW] IO7 for Marvel-based Alpha systems |
| 2097 | See comment before marvel_specify_io7 in |
| 2098 | arch/alpha/kernel/core_marvel.c. |
| 2099 | |
| 2100 | io_delay= [X86] I/O delay method |
| 2101 | 0x80 |
| 2102 | Standard port 0x80 based delay |
| 2103 | 0xed |
| 2104 | Alternate port 0xed based delay (needed on some systems) |
| 2105 | udelay |
| 2106 | Simple two microseconds delay |
| 2107 | none |
| 2108 | No delay |
| 2109 | |
| 2110 | ip= [IP_PNP] |
| 2111 | See Documentation/admin-guide/nfs/nfsroot.rst. |
| 2112 | |
| 2113 | ipcmni_extend [KNL] Extend the maximum number of unique System V |
| 2114 | IPC identifiers from 32,768 to 16,777,216. |
| 2115 | |
| 2116 | irqaffinity= [SMP] Set the default irq affinity mask |
| 2117 | The argument is a cpu list, as described above. |
| 2118 | |
| 2119 | irqchip.gicv2_force_probe= |
| 2120 | [ARM, ARM64] |
| 2121 | Format: <bool> |
| 2122 | Force the kernel to look for the second 4kB page |
| 2123 | of a GICv2 controller even if the memory range |
| 2124 | exposed by the device tree is too small. |
| 2125 | |
| 2126 | irqchip.gicv3_nolpi= |
| 2127 | [ARM, ARM64] |
| 2128 | Force the kernel to ignore the availability of |
| 2129 | LPIs (and by consequence ITSs). Intended for system |
| 2130 | that use the kernel as a bootloader, and thus want |
| 2131 | to let secondary kernels in charge of setting up |
| 2132 | LPIs. |
| 2133 | |
| 2134 | irqchip.gicv3_pseudo_nmi= [ARM64] |
| 2135 | Enables support for pseudo-NMIs in the kernel. This |
| 2136 | requires the kernel to be built with |
| 2137 | CONFIG_ARM64_PSEUDO_NMI. |
| 2138 | |
| 2139 | irqfixup [HW] |
| 2140 | When an interrupt is not handled search all handlers |
| 2141 | for it. Intended to get systems with badly broken |
| 2142 | firmware running. |
| 2143 | |
| 2144 | irqpoll [HW] |
| 2145 | When an interrupt is not handled search all handlers |
| 2146 | for it. Also check all handlers each timer |
| 2147 | interrupt. Intended to get systems with badly broken |
| 2148 | firmware running. |
| 2149 | |
| 2150 | isapnp= [ISAPNP] |
| 2151 | Format: <RDP>,<reset>,<pci_scan>,<verbosity> |
| 2152 | |
| 2153 | isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance. |
| 2154 | [Deprecated - use cpusets instead] |
| 2155 | Format: [flag-list,]<cpu-list> |
| 2156 | |
| 2157 | Specify one or more CPUs to isolate from disturbances |
| 2158 | specified in the flag list (default: domain): |
| 2159 | |
| 2160 | nohz |
| 2161 | Disable the tick when a single task runs. |
| 2162 | |
| 2163 | A residual 1Hz tick is offloaded to workqueues, which you |
| 2164 | need to affine to housekeeping through the global |
| 2165 | workqueue's affinity configured via the |
| 2166 | /sys/devices/virtual/workqueue/cpumask sysfs file, or |
| 2167 | by using the 'domain' flag described below. |
| 2168 | |
| 2169 | NOTE: by default the global workqueue runs on all CPUs, |
| 2170 | so to protect individual CPUs the 'cpumask' file has to |
| 2171 | be configured manually after bootup. |
| 2172 | |
| 2173 | domain |
| 2174 | Isolate from the general SMP balancing and scheduling |
| 2175 | algorithms. Note that performing domain isolation this way |
| 2176 | is irreversible: it's not possible to bring back a CPU to |
| 2177 | the domains once isolated through isolcpus. It's strongly |
| 2178 | advised to use cpusets instead to disable scheduler load |
| 2179 | balancing through the "cpuset.sched_load_balance" file. |
| 2180 | It offers a much more flexible interface where CPUs can |
| 2181 | move in and out of an isolated set anytime. |
| 2182 | |
| 2183 | You can move a process onto or off an "isolated" CPU via |
| 2184 | the CPU affinity syscalls or cpuset. |
| 2185 | <cpu number> begins at 0 and the maximum value is |
| 2186 | "number of CPUs in system - 1". |
| 2187 | |
| 2188 | managed_irq |
| 2189 | |
| 2190 | Isolate from being targeted by managed interrupts |
| 2191 | which have an interrupt mask containing isolated |
| 2192 | CPUs. The affinity of managed interrupts is |
| 2193 | handled by the kernel and cannot be changed via |
| 2194 | the /proc/irq/* interfaces. |
| 2195 | |
| 2196 | This isolation is best effort and only effective |
| 2197 | if the automatically assigned interrupt mask of a |
| 2198 | device queue contains isolated and housekeeping |
| 2199 | CPUs. If housekeeping CPUs are online then such |
| 2200 | interrupts are directed to the housekeeping CPU |
| 2201 | so that IO submitted on the housekeeping CPU |
| 2202 | cannot disturb the isolated CPU. |
| 2203 | |
| 2204 | If a queue's affinity mask contains only isolated |
| 2205 | CPUs then this parameter has no effect on the |
| 2206 | interrupt routing decision, though interrupts are |
| 2207 | only delivered when tasks running on those |
| 2208 | isolated CPUs submit IO. IO submitted on |
| 2209 | housekeeping CPUs has no influence on those |
| 2210 | queues. |
| 2211 | |
| 2212 | The format of <cpu-list> is described above. |
| 2213 | |
| 2214 | iucv= [HW,NET] |
| 2215 | |
| 2216 | ivrs_ioapic [HW,X86-64] |
| 2217 | Provide an override to the IOAPIC-ID<->DEVICE-ID |
| 2218 | mapping provided in the IVRS ACPI table. For |
| 2219 | example, to map IOAPIC-ID decimal 10 to |
| 2220 | PCI device 00:14.0 write the parameter as: |
| 2221 | ivrs_ioapic[10]=00:14.0 |
| 2222 | |
| 2223 | ivrs_hpet [HW,X86-64] |
| 2224 | Provide an override to the HPET-ID<->DEVICE-ID |
| 2225 | mapping provided in the IVRS ACPI table. For |
| 2226 | example, to map HPET-ID decimal 0 to |
| 2227 | PCI device 00:14.0 write the parameter as: |
| 2228 | ivrs_hpet[0]=00:14.0 |
| 2229 | |
| 2230 | ivrs_acpihid [HW,X86-64] |
| 2231 | Provide an override to the ACPI-HID:UID<->DEVICE-ID |
| 2232 | mapping provided in the IVRS ACPI table. For |
| 2233 | example, to map UART-HID:UID AMD0020:0 to |
| 2234 | PCI device 00:14.5 write the parameter as: |
| 2235 | ivrs_acpihid[00:14.5]=AMD0020:0 |
| 2236 | |
| 2237 | js= [HW,JOY] Analog joystick |
| 2238 | See Documentation/input/joydev/joystick.rst. |
| 2239 | |
| 2240 | nokaslr [KNL] |
| 2241 | When CONFIG_RANDOMIZE_BASE is set, this disables |
| 2242 | kernel and module base offset ASLR (Address Space |
| 2243 | Layout Randomization). |
| 2244 | |
| 2245 | kasan_multi_shot |
| 2246 | [KNL] Enforce KASAN (Kernel Address Sanitizer) to print |
| 2247 | report on every invalid memory access. Without this |
| 2248 | parameter KASAN will print report only for the first |
| 2249 | invalid access. |
| 2250 | |
| 2251 | keepinitrd [HW,ARM] |
| 2252 | |
| 2253 | kernelcore= [KNL,X86,IA-64,PPC] |
| 2254 | Format: nn[KMGTPE] | nn% | "mirror" |
| 2255 | This parameter specifies the amount of memory usable by |
| 2256 | the kernel for non-movable allocations. The requested |
| 2257 | amount is spread evenly throughout all nodes in the |
| 2258 | system as ZONE_NORMAL. The remaining memory is used for |
| 2259 | movable memory in its own zone, ZONE_MOVABLE. In the |
| 2260 | event, a node is too small to have both ZONE_NORMAL and |
| 2261 | ZONE_MOVABLE, kernelcore memory will take priority and |
| 2262 | other nodes will have a larger ZONE_MOVABLE. |
| 2263 | |
| 2264 | ZONE_MOVABLE is used for the allocation of pages that |
| 2265 | may be reclaimed or moved by the page migration |
| 2266 | subsystem. Note that allocations like PTEs-from-HighMem |
| 2267 | still use the HighMem zone if it exists, and the Normal |
| 2268 | zone if it does not. |
| 2269 | |
| 2270 | It is possible to specify the exact amount of memory in |
| 2271 | the form of "nn[KMGTPE]", a percentage of total system |
| 2272 | memory in the form of "nn%", or "mirror". If "mirror" |
| 2273 | option is specified, mirrored (reliable) memory is used |
| 2274 | for non-movable allocations and remaining memory is used |
| 2275 | for Movable pages. "nn[KMGTPE]", "nn%", and "mirror" |
| 2276 | are exclusive, so you cannot specify multiple forms. |
| 2277 | |
| 2278 | kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port. |
| 2279 | Format: <Controller#>[,poll interval] |
| 2280 | The controller # is the number of the ehci usb debug |
| 2281 | port as it is probed via PCI. The poll interval is |
| 2282 | optional and is the number seconds in between |
| 2283 | each poll cycle to the debug port in case you need |
| 2284 | the functionality for interrupting the kernel with |
| 2285 | gdb or control-c on the dbgp connection. When |
| 2286 | not using this parameter you use sysrq-g to break into |
| 2287 | the kernel debugger. |
| 2288 | |
| 2289 | kgdboc= [KGDB,HW] kgdb over consoles. |
| 2290 | Requires a tty driver that supports console polling, |
| 2291 | or a supported polling keyboard driver (non-usb). |
| 2292 | Serial only format: <serial_device>[,baud] |
| 2293 | keyboard only format: kbd |
| 2294 | keyboard and serial format: kbd,<serial_device>[,baud] |
| 2295 | Optional Kernel mode setting: |
| 2296 | kms, kbd format: kms,kbd |
| 2297 | kms, kbd and serial format: kms,kbd,<ser_dev>[,baud] |
| 2298 | |
| 2299 | kgdboc_earlycon= [KGDB,HW] |
| 2300 | If the boot console provides the ability to read |
| 2301 | characters and can work in polling mode, you can use |
| 2302 | this parameter to tell kgdb to use it as a backend |
| 2303 | until the normal console is registered. Intended to |
| 2304 | be used together with the kgdboc parameter which |
| 2305 | specifies the normal console to transition to. |
| 2306 | |
| 2307 | The name of the early console should be specified |
| 2308 | as the value of this parameter. Note that the name of |
| 2309 | the early console might be different than the tty |
| 2310 | name passed to kgdboc. It's OK to leave the value |
| 2311 | blank and the first boot console that implements |
| 2312 | read() will be picked. |
| 2313 | |
| 2314 | kgdbwait [KGDB] Stop kernel execution and enter the |
| 2315 | kernel debugger at the earliest opportunity. |
| 2316 | |
| 2317 | kmac= [MIPS] Korina ethernet MAC address. |
| 2318 | Configure the RouterBoard 532 series on-chip |
| 2319 | Ethernet adapter MAC address. |
| 2320 | |
| 2321 | kmemleak= [KNL] Boot-time kmemleak enable/disable |
| 2322 | Valid arguments: on, off |
| 2323 | Default: on |
| 2324 | Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y, |
| 2325 | the default is off. |
| 2326 | |
| 2327 | kprobe_event=[probe-list] |
| 2328 | [FTRACE] Add kprobe events and enable at boot time. |
| 2329 | The probe-list is a semicolon delimited list of probe |
| 2330 | definitions. Each definition is same as kprobe_events |
| 2331 | interface, but the parameters are comma delimited. |
| 2332 | For example, to add a kprobe event on vfs_read with |
| 2333 | arg1 and arg2, add to the command line; |
| 2334 | |
| 2335 | kprobe_event=p,vfs_read,$arg1,$arg2 |
| 2336 | |
| 2337 | See also Documentation/trace/kprobetrace.rst "Kernel |
| 2338 | Boot Parameter" section. |
| 2339 | |
| 2340 | kpti= [ARM64] Control page table isolation of user |
| 2341 | and kernel address spaces. |
| 2342 | Default: enabled on cores which need mitigation. |
| 2343 | 0: force disabled |
| 2344 | 1: force enabled |
| 2345 | |
| 2346 | kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs. |
| 2347 | Default is 0 (don't ignore, but inject #GP) |
| 2348 | |
| 2349 | kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface. |
| 2350 | Default is false (don't support). |
| 2351 | |
| 2352 | kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit |
| 2353 | KVM MMU at runtime. |
| 2354 | Default is 0 (off) |
| 2355 | |
| 2356 | kvm.nx_huge_pages= |
| 2357 | [KVM] Controls the software workaround for the |
| 2358 | X86_BUG_ITLB_MULTIHIT bug. |
| 2359 | force : Always deploy workaround. |
| 2360 | off : Never deploy workaround. |
| 2361 | auto : Deploy workaround based on the presence of |
| 2362 | X86_BUG_ITLB_MULTIHIT. |
| 2363 | |
| 2364 | Default is 'auto'. |
| 2365 | |
| 2366 | If the software workaround is enabled for the host, |
| 2367 | guests do need not to enable it for nested guests. |
| 2368 | |
| 2369 | kvm.nx_huge_pages_recovery_ratio= |
| 2370 | [KVM] Controls how many 4KiB pages are periodically zapped |
| 2371 | back to huge pages. 0 disables the recovery, otherwise if |
| 2372 | the value is N KVM will zap 1/Nth of the 4KiB pages every |
| 2373 | period (see below). The default is 60. |
| 2374 | |
| 2375 | kvm.nx_huge_pages_recovery_period_ms= |
| 2376 | [KVM] Controls the time period at which KVM zaps 4KiB pages |
| 2377 | back to huge pages. If the value is a non-zero N, KVM will |
| 2378 | zap a portion (see ratio above) of the pages every N msecs. |
| 2379 | If the value is 0 (the default), KVM will pick a period based |
| 2380 | on the ratio, such that a page is zapped after 1 hour on average. |
| 2381 | |
| 2382 | kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM. |
| 2383 | Default is 1 (enabled) |
| 2384 | |
| 2385 | kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU) |
| 2386 | for all guests. |
| 2387 | Default is 1 (enabled) if in 64-bit or 32-bit PAE mode. |
| 2388 | |
| 2389 | kvm-arm.mode= |
| 2390 | [KVM,ARM] Select one of KVM/arm64's modes of operation. |
| 2391 | |
| 2392 | none: Forcefully disable KVM. |
| 2393 | |
| 2394 | nvhe: Standard nVHE-based mode, without support for |
| 2395 | protected guests. |
| 2396 | |
| 2397 | protected: nVHE-based mode with support for guests whose |
| 2398 | state is kept private from the host. |
| 2399 | Not valid if the kernel is running in EL2. |
| 2400 | |
| 2401 | Defaults to VHE/nVHE based on hardware support. Setting |
| 2402 | mode to "protected" will disable kexec and hibernation |
| 2403 | for the host. |
| 2404 | |
| 2405 | kvm-arm.vgic_v3_group0_trap= |
| 2406 | [KVM,ARM] Trap guest accesses to GICv3 group-0 |
| 2407 | system registers |
| 2408 | |
| 2409 | kvm-arm.vgic_v3_group1_trap= |
| 2410 | [KVM,ARM] Trap guest accesses to GICv3 group-1 |
| 2411 | system registers |
| 2412 | |
| 2413 | kvm-arm.vgic_v3_common_trap= |
| 2414 | [KVM,ARM] Trap guest accesses to GICv3 common |
| 2415 | system registers |
| 2416 | |
| 2417 | kvm-arm.vgic_v4_enable= |
| 2418 | [KVM,ARM] Allow use of GICv4 for direct injection of |
| 2419 | LPIs. |
| 2420 | |
| 2421 | kvm_cma_resv_ratio=n [PPC] |
| 2422 | Reserves given percentage from system memory area for |
| 2423 | contiguous memory allocation for KVM hash pagetable |
| 2424 | allocation. |
| 2425 | By default it reserves 5% of total system memory. |
| 2426 | Format: <integer> |
| 2427 | Default: 5 |
| 2428 | |
| 2429 | kvm-intel.ept= [KVM,Intel] Disable extended page tables |
| 2430 | (virtualized MMU) support on capable Intel chips. |
| 2431 | Default is 1 (enabled) |
| 2432 | |
| 2433 | kvm-intel.emulate_invalid_guest_state= |
| 2434 | [KVM,Intel] Disable emulation of invalid guest state. |
| 2435 | Ignored if kvm-intel.enable_unrestricted_guest=1, as |
| 2436 | guest state is never invalid for unrestricted guests. |
| 2437 | This param doesn't apply to nested guests (L2), as KVM |
| 2438 | never emulates invalid L2 guest state. |
| 2439 | Default is 1 (enabled) |
| 2440 | |
| 2441 | kvm-intel.flexpriority= |
| 2442 | [KVM,Intel] Disable FlexPriority feature (TPR shadow). |
| 2443 | Default is 1 (enabled) |
| 2444 | |
| 2445 | kvm-intel.nested= |
| 2446 | [KVM,Intel] Enable VMX nesting (nVMX). |
| 2447 | Default is 0 (disabled) |
| 2448 | |
| 2449 | kvm-intel.unrestricted_guest= |
| 2450 | [KVM,Intel] Disable unrestricted guest feature |
| 2451 | (virtualized real and unpaged mode) on capable |
| 2452 | Intel chips. Default is 1 (enabled) |
| 2453 | |
| 2454 | kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault |
| 2455 | CVE-2018-3620. |
| 2456 | |
| 2457 | Valid arguments: never, cond, always |
| 2458 | |
| 2459 | always: L1D cache flush on every VMENTER. |
| 2460 | cond: Flush L1D on VMENTER only when the code between |
| 2461 | VMEXIT and VMENTER can leak host memory. |
| 2462 | never: Disables the mitigation |
| 2463 | |
| 2464 | Default is cond (do L1 cache flush in specific instances) |
| 2465 | |
| 2466 | kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification |
| 2467 | feature (tagged TLBs) on capable Intel chips. |
| 2468 | Default is 1 (enabled) |
| 2469 | |
| 2470 | l1d_flush= [X86,INTEL] |
| 2471 | Control mitigation for L1D based snooping vulnerability. |
| 2472 | |
| 2473 | Certain CPUs are vulnerable to an exploit against CPU |
| 2474 | internal buffers which can forward information to a |
| 2475 | disclosure gadget under certain conditions. |
| 2476 | |
| 2477 | In vulnerable processors, the speculatively |
| 2478 | forwarded data can be used in a cache side channel |
| 2479 | attack, to access data to which the attacker does |
| 2480 | not have direct access. |
| 2481 | |
| 2482 | This parameter controls the mitigation. The |
| 2483 | options are: |
| 2484 | |
| 2485 | on - enable the interface for the mitigation |
| 2486 | |
| 2487 | l1tf= [X86] Control mitigation of the L1TF vulnerability on |
| 2488 | affected CPUs |
| 2489 | |
| 2490 | The kernel PTE inversion protection is unconditionally |
| 2491 | enabled and cannot be disabled. |
| 2492 | |
| 2493 | full |
| 2494 | Provides all available mitigations for the |
| 2495 | L1TF vulnerability. Disables SMT and |
| 2496 | enables all mitigations in the |
| 2497 | hypervisors, i.e. unconditional L1D flush. |
| 2498 | |
| 2499 | SMT control and L1D flush control via the |
| 2500 | sysfs interface is still possible after |
| 2501 | boot. Hypervisors will issue a warning |
| 2502 | when the first VM is started in a |
| 2503 | potentially insecure configuration, |
| 2504 | i.e. SMT enabled or L1D flush disabled. |
| 2505 | |
| 2506 | full,force |
| 2507 | Same as 'full', but disables SMT and L1D |
| 2508 | flush runtime control. Implies the |
| 2509 | 'nosmt=force' command line option. |
| 2510 | (i.e. sysfs control of SMT is disabled.) |
| 2511 | |
| 2512 | flush |
| 2513 | Leaves SMT enabled and enables the default |
| 2514 | hypervisor mitigation, i.e. conditional |
| 2515 | L1D flush. |
| 2516 | |
| 2517 | SMT control and L1D flush control via the |
| 2518 | sysfs interface is still possible after |
| 2519 | boot. Hypervisors will issue a warning |
| 2520 | when the first VM is started in a |
| 2521 | potentially insecure configuration, |
| 2522 | i.e. SMT enabled or L1D flush disabled. |
| 2523 | |
| 2524 | flush,nosmt |
| 2525 | |
| 2526 | Disables SMT and enables the default |
| 2527 | hypervisor mitigation. |
| 2528 | |
| 2529 | SMT control and L1D flush control via the |
| 2530 | sysfs interface is still possible after |
| 2531 | boot. Hypervisors will issue a warning |
| 2532 | when the first VM is started in a |
| 2533 | potentially insecure configuration, |
| 2534 | i.e. SMT enabled or L1D flush disabled. |
| 2535 | |
| 2536 | flush,nowarn |
| 2537 | Same as 'flush', but hypervisors will not |
| 2538 | warn when a VM is started in a potentially |
| 2539 | insecure configuration. |
| 2540 | |
| 2541 | off |
| 2542 | Disables hypervisor mitigations and doesn't |
| 2543 | emit any warnings. |
| 2544 | It also drops the swap size and available |
| 2545 | RAM limit restriction on both hypervisor and |
| 2546 | bare metal. |
| 2547 | |
| 2548 | Default is 'flush'. |
| 2549 | |
| 2550 | For details see: Documentation/admin-guide/hw-vuln/l1tf.rst |
| 2551 | |
| 2552 | l2cr= [PPC] |
| 2553 | |
| 2554 | l3cr= [PPC] |
| 2555 | |
| 2556 | lapic [X86-32,APIC] Enable the local APIC even if BIOS |
| 2557 | disabled it. |
| 2558 | |
| 2559 | lapic= [X86,APIC] Do not use TSC deadline |
| 2560 | value for LAPIC timer one-shot implementation. Default |
| 2561 | back to the programmable timer unit in the LAPIC. |
| 2562 | Format: notscdeadline |
| 2563 | |
| 2564 | lapic_timer_c2_ok [X86,APIC] trust the local apic timer |
| 2565 | in C2 power state. |
| 2566 | |
| 2567 | libata.dma= [LIBATA] DMA control |
| 2568 | libata.dma=0 Disable all PATA and SATA DMA |
| 2569 | libata.dma=1 PATA and SATA Disk DMA only |
| 2570 | libata.dma=2 ATAPI (CDROM) DMA only |
| 2571 | libata.dma=4 Compact Flash DMA only |
| 2572 | Combinations also work, so libata.dma=3 enables DMA |
| 2573 | for disks and CDROMs, but not CFs. |
| 2574 | |
| 2575 | libata.ignore_hpa= [LIBATA] Ignore HPA limit |
| 2576 | libata.ignore_hpa=0 keep BIOS limits (default) |
| 2577 | libata.ignore_hpa=1 ignore limits, using full disk |
| 2578 | |
| 2579 | libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume |
| 2580 | when set. |
| 2581 | Format: <int> |
| 2582 | |
| 2583 | libata.force= [LIBATA] Force configurations. The format is comma- |
| 2584 | separated list of "[ID:]VAL" where ID is |
| 2585 | PORT[.DEVICE]. PORT and DEVICE are decimal numbers |
| 2586 | matching port, link or device. Basically, it matches |
| 2587 | the ATA ID string printed on console by libata. If |
| 2588 | the whole ID part is omitted, the last PORT and DEVICE |
| 2589 | values are used. If ID hasn't been specified yet, the |
| 2590 | configuration applies to all ports, links and devices. |
| 2591 | |
| 2592 | If only DEVICE is omitted, the parameter applies to |
| 2593 | the port and all links and devices behind it. DEVICE |
| 2594 | number of 0 either selects the first device or the |
| 2595 | first fan-out link behind PMP device. It does not |
| 2596 | select the host link. DEVICE number of 15 selects the |
| 2597 | host link and device attached to it. |
| 2598 | |
| 2599 | The VAL specifies the configuration to force. As long |
| 2600 | as there's no ambiguity shortcut notation is allowed. |
| 2601 | For example, both 1.5 and 1.5G would work for 1.5Gbps. |
| 2602 | The following configurations can be forced. |
| 2603 | |
| 2604 | * Cable type: 40c, 80c, short40c, unk, ign or sata. |
| 2605 | Any ID with matching PORT is used. |
| 2606 | |
| 2607 | * SATA link speed limit: 1.5Gbps or 3.0Gbps. |
| 2608 | |
| 2609 | * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7]. |
| 2610 | udma[/][16,25,33,44,66,100,133] notation is also |
| 2611 | allowed. |
| 2612 | |
| 2613 | * [no]ncq: Turn on or off NCQ. |
| 2614 | |
| 2615 | * [no]ncqtrim: Turn off queued DSM TRIM. |
| 2616 | |
| 2617 | * nohrst, nosrst, norst: suppress hard, soft |
| 2618 | and both resets. |
| 2619 | |
| 2620 | * rstonce: only attempt one reset during |
| 2621 | hot-unplug link recovery |
| 2622 | |
| 2623 | * dump_id: dump IDENTIFY data. |
| 2624 | |
| 2625 | * atapi_dmadir: Enable ATAPI DMADIR bridge support |
| 2626 | |
| 2627 | * disable: Disable this device. |
| 2628 | |
| 2629 | If there are multiple matching configurations changing |
| 2630 | the same attribute, the last one is used. |
| 2631 | |
| 2632 | memblock=debug [KNL] Enable memblock debug messages. |
| 2633 | |
| 2634 | load_ramdisk= [RAM] [Deprecated] |
| 2635 | |
| 2636 | lockd.nlm_grace_period=P [NFS] Assign grace period. |
| 2637 | Format: <integer> |
| 2638 | |
| 2639 | lockd.nlm_tcpport=N [NFS] Assign TCP port. |
| 2640 | Format: <integer> |
| 2641 | |
| 2642 | lockd.nlm_timeout=T [NFS] Assign timeout value. |
| 2643 | Format: <integer> |
| 2644 | |
| 2645 | lockd.nlm_udpport=M [NFS] Assign UDP port. |
| 2646 | Format: <integer> |
| 2647 | |
| 2648 | lockdown= [SECURITY] |
| 2649 | { integrity | confidentiality } |
| 2650 | Enable the kernel lockdown feature. If set to |
| 2651 | integrity, kernel features that allow userland to |
| 2652 | modify the running kernel are disabled. If set to |
| 2653 | confidentiality, kernel features that allow userland |
| 2654 | to extract confidential information from the kernel |
| 2655 | are also disabled. |
| 2656 | |
| 2657 | locktorture.nreaders_stress= [KNL] |
| 2658 | Set the number of locking read-acquisition kthreads. |
| 2659 | Defaults to being automatically set based on the |
| 2660 | number of online CPUs. |
| 2661 | |
| 2662 | locktorture.nwriters_stress= [KNL] |
| 2663 | Set the number of locking write-acquisition kthreads. |
| 2664 | |
| 2665 | locktorture.onoff_holdoff= [KNL] |
| 2666 | Set time (s) after boot for CPU-hotplug testing. |
| 2667 | |
| 2668 | locktorture.onoff_interval= [KNL] |
| 2669 | Set time (s) between CPU-hotplug operations, or |
| 2670 | zero to disable CPU-hotplug testing. |
| 2671 | |
| 2672 | locktorture.shuffle_interval= [KNL] |
| 2673 | Set task-shuffle interval (jiffies). Shuffling |
| 2674 | tasks allows some CPUs to go into dyntick-idle |
| 2675 | mode during the locktorture test. |
| 2676 | |
| 2677 | locktorture.shutdown_secs= [KNL] |
| 2678 | Set time (s) after boot system shutdown. This |
| 2679 | is useful for hands-off automated testing. |
| 2680 | |
| 2681 | locktorture.stat_interval= [KNL] |
| 2682 | Time (s) between statistics printk()s. |
| 2683 | |
| 2684 | locktorture.stutter= [KNL] |
| 2685 | Time (s) to stutter testing, for example, |
| 2686 | specifying five seconds causes the test to run for |
| 2687 | five seconds, wait for five seconds, and so on. |
| 2688 | This tests the locking primitive's ability to |
| 2689 | transition abruptly to and from idle. |
| 2690 | |
| 2691 | locktorture.torture_type= [KNL] |
| 2692 | Specify the locking implementation to test. |
| 2693 | |
| 2694 | locktorture.verbose= [KNL] |
| 2695 | Enable additional printk() statements. |
| 2696 | |
| 2697 | logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver |
| 2698 | Format: <irq> |
| 2699 | |
| 2700 | loglevel= All Kernel Messages with a loglevel smaller than the |
| 2701 | console loglevel will be printed to the console. It can |
| 2702 | also be changed with klogd or other programs. The |
| 2703 | loglevels are defined as follows: |
| 2704 | |
| 2705 | 0 (KERN_EMERG) system is unusable |
| 2706 | 1 (KERN_ALERT) action must be taken immediately |
| 2707 | 2 (KERN_CRIT) critical conditions |
| 2708 | 3 (KERN_ERR) error conditions |
| 2709 | 4 (KERN_WARNING) warning conditions |
| 2710 | 5 (KERN_NOTICE) normal but significant condition |
| 2711 | 6 (KERN_INFO) informational |
| 2712 | 7 (KERN_DEBUG) debug-level messages |
| 2713 | |
| 2714 | log_buf_len=n[KMG] Sets the size of the printk ring buffer, |
| 2715 | in bytes. n must be a power of two and greater |
| 2716 | than the minimal size. The minimal size is defined |
| 2717 | by LOG_BUF_SHIFT kernel config parameter. There is |
| 2718 | also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter |
| 2719 | that allows to increase the default size depending on |
| 2720 | the number of CPUs. See init/Kconfig for more details. |
| 2721 | |
| 2722 | logo.nologo [FB] Disables display of the built-in Linux logo. |
| 2723 | This may be used to provide more screen space for |
| 2724 | kernel log messages and is useful when debugging |
| 2725 | kernel boot problems. |
| 2726 | |
| 2727 | lp=0 [LP] Specify parallel ports to use, e.g, |
| 2728 | lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses |
| 2729 | lp=reset first parallel port). 'lp=0' disables the |
| 2730 | lp=auto printer driver. 'lp=reset' (which can be |
| 2731 | specified in addition to the ports) causes |
| 2732 | attached printers to be reset. Using |
| 2733 | lp=port1,port2,... specifies the parallel ports |
| 2734 | to associate lp devices with, starting with |
| 2735 | lp0. A port specification may be 'none' to skip |
| 2736 | that lp device, or a parport name such as |
| 2737 | 'parport0'. Specifying 'lp=auto' instead of a |
| 2738 | port specification list means that device IDs |
| 2739 | from each port should be examined, to see if |
| 2740 | an IEEE 1284-compliant printer is attached; if |
| 2741 | so, the driver will manage that printer. |
| 2742 | See also header of drivers/char/lp.c. |
| 2743 | |
| 2744 | lpj=n [KNL] |
| 2745 | Sets loops_per_jiffy to given constant, thus avoiding |
| 2746 | time-consuming boot-time autodetection (up to 250 ms per |
| 2747 | CPU). 0 enables autodetection (default). To determine |
| 2748 | the correct value for your kernel, boot with normal |
| 2749 | autodetection and see what value is printed. Note that |
| 2750 | on SMP systems the preset will be applied to all CPUs, |
| 2751 | which is likely to cause problems if your CPUs need |
| 2752 | significantly divergent settings. An incorrect value |
| 2753 | will cause delays in the kernel to be wrong, leading to |
| 2754 | unpredictable I/O errors and other breakage. Although |
| 2755 | unlikely, in the extreme case this might damage your |
| 2756 | hardware. |
| 2757 | |
| 2758 | ltpc= [NET] |
| 2759 | Format: <io>,<irq>,<dma> |
| 2760 | |
| 2761 | lsm.debug [SECURITY] Enable LSM initialization debugging output. |
| 2762 | |
| 2763 | lsm=lsm1,...,lsmN |
| 2764 | [SECURITY] Choose order of LSM initialization. This |
| 2765 | overrides CONFIG_LSM, and the "security=" parameter. |
| 2766 | |
| 2767 | machvec= [IA-64] Force the use of a particular machine-vector |
| 2768 | (machvec) in a generic kernel. |
| 2769 | Example: machvec=hpzx1 |
| 2770 | |
| 2771 | machtype= [Loongson] Share the same kernel image file between |
| 2772 | different yeeloong laptops. |
| 2773 | Example: machtype=lemote-yeeloong-2f-7inch |
| 2774 | |
| 2775 | max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater |
| 2776 | than or equal to this physical address is ignored. |
| 2777 | |
| 2778 | maxcpus= [SMP] Maximum number of processors that an SMP kernel |
| 2779 | will bring up during bootup. maxcpus=n : n >= 0 limits |
| 2780 | the kernel to bring up 'n' processors. Surely after |
| 2781 | bootup you can bring up the other plugged cpu by executing |
| 2782 | "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus |
| 2783 | only takes effect during system bootup. |
| 2784 | While n=0 is a special case, it is equivalent to "nosmp", |
| 2785 | which also disables the IO APIC. |
| 2786 | |
| 2787 | max_loop= [LOOP] The number of loop block devices that get |
| 2788 | (loop.max_loop) unconditionally pre-created at init time. The default |
| 2789 | number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead |
| 2790 | of statically allocating a predefined number, loop |
| 2791 | devices can be requested on-demand with the |
| 2792 | /dev/loop-control interface. |
| 2793 | |
| 2794 | mce [X86-32] Machine Check Exception |
| 2795 | |
| 2796 | mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst |
| 2797 | |
| 2798 | md= [HW] RAID subsystems devices and level |
| 2799 | See Documentation/admin-guide/md.rst. |
| 2800 | |
| 2801 | mdacon= [MDA] |
| 2802 | Format: <first>,<last> |
| 2803 | Specifies range of consoles to be captured by the MDA. |
| 2804 | |
| 2805 | mds= [X86,INTEL] |
| 2806 | Control mitigation for the Micro-architectural Data |
| 2807 | Sampling (MDS) vulnerability. |
| 2808 | |
| 2809 | Certain CPUs are vulnerable to an exploit against CPU |
| 2810 | internal buffers which can forward information to a |
| 2811 | disclosure gadget under certain conditions. |
| 2812 | |
| 2813 | In vulnerable processors, the speculatively |
| 2814 | forwarded data can be used in a cache side channel |
| 2815 | attack, to access data to which the attacker does |
| 2816 | not have direct access. |
| 2817 | |
| 2818 | This parameter controls the MDS mitigation. The |
| 2819 | options are: |
| 2820 | |
| 2821 | full - Enable MDS mitigation on vulnerable CPUs |
| 2822 | full,nosmt - Enable MDS mitigation and disable |
| 2823 | SMT on vulnerable CPUs |
| 2824 | off - Unconditionally disable MDS mitigation |
| 2825 | |
| 2826 | On TAA-affected machines, mds=off can be prevented by |
| 2827 | an active TAA mitigation as both vulnerabilities are |
| 2828 | mitigated with the same mechanism so in order to disable |
| 2829 | this mitigation, you need to specify tsx_async_abort=off |
| 2830 | too. |
| 2831 | |
| 2832 | Not specifying this option is equivalent to |
| 2833 | mds=full. |
| 2834 | |
| 2835 | For details see: Documentation/admin-guide/hw-vuln/mds.rst |
| 2836 | |
| 2837 | mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory |
| 2838 | Amount of memory to be used in cases as follows: |
| 2839 | |
| 2840 | 1 for test; |
| 2841 | 2 when the kernel is not able to see the whole system memory; |
| 2842 | 3 memory that lies after 'mem=' boundary is excluded from |
| 2843 | the hypervisor, then assigned to KVM guests. |
| 2844 | |
| 2845 | [X86] Work as limiting max address. Use together |
| 2846 | with memmap= to avoid physical address space collisions. |
| 2847 | Without memmap= PCI devices could be placed at addresses |
| 2848 | belonging to unused RAM. |
| 2849 | |
| 2850 | Note that this only takes effects during boot time since |
| 2851 | in above case 3, memory may need be hot added after boot |
| 2852 | if system memory of hypervisor is not sufficient. |
| 2853 | |
| 2854 | mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel |
| 2855 | memory. |
| 2856 | |
| 2857 | memchunk=nn[KMG] |
| 2858 | [KNL,SH] Allow user to override the default size for |
| 2859 | per-device physically contiguous DMA buffers. |
| 2860 | |
| 2861 | memhp_default_state=online/offline |
| 2862 | [KNL] Set the initial state for the memory hotplug |
| 2863 | onlining policy. If not specified, the default value is |
| 2864 | set according to the |
| 2865 | CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config |
| 2866 | option. |
| 2867 | See Documentation/admin-guide/mm/memory-hotplug.rst. |
| 2868 | |
| 2869 | memmap=exactmap [KNL,X86] Enable setting of an exact |
| 2870 | E820 memory map, as specified by the user. |
| 2871 | Such memmap=exactmap lines can be constructed based on |
| 2872 | BIOS output or other requirements. See the memmap=nn@ss |
| 2873 | option description. |
| 2874 | |
| 2875 | memmap=nn[KMG]@ss[KMG] |
| 2876 | [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory. |
| 2877 | Region of memory to be used is from ss to ss+nn. |
| 2878 | If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG], |
| 2879 | which limits max address to nn[KMG]. |
| 2880 | Multiple different regions can be specified, |
| 2881 | comma delimited. |
| 2882 | Example: |
| 2883 | memmap=100M@2G,100M#3G,1G!1024G |
| 2884 | |
| 2885 | memmap=nn[KMG]#ss[KMG] |
| 2886 | [KNL,ACPI] Mark specific memory as ACPI data. |
| 2887 | Region of memory to be marked is from ss to ss+nn. |
| 2888 | |
| 2889 | memmap=nn[KMG]$ss[KMG] |
| 2890 | [KNL,ACPI] Mark specific memory as reserved. |
| 2891 | Region of memory to be reserved is from ss to ss+nn. |
| 2892 | Example: Exclude memory from 0x18690000-0x1869ffff |
| 2893 | memmap=64K$0x18690000 |
| 2894 | or |
| 2895 | memmap=0x10000$0x18690000 |
| 2896 | Some bootloaders may need an escape character before '$', |
| 2897 | like Grub2, otherwise '$' and the following number |
| 2898 | will be eaten. |
| 2899 | |
| 2900 | memmap=nn[KMG]!ss[KMG] |
| 2901 | [KNL,X86] Mark specific memory as protected. |
| 2902 | Region of memory to be used, from ss to ss+nn. |
| 2903 | The memory region may be marked as e820 type 12 (0xc) |
| 2904 | and is NVDIMM or ADR memory. |
| 2905 | |
| 2906 | memmap=<size>%<offset>-<oldtype>+<newtype> |
| 2907 | [KNL,ACPI] Convert memory within the specified region |
| 2908 | from <oldtype> to <newtype>. If "-<oldtype>" is left |
| 2909 | out, the whole region will be marked as <newtype>, |
| 2910 | even if previously unavailable. If "+<newtype>" is left |
| 2911 | out, matching memory will be removed. Types are |
| 2912 | specified as e820 types, e.g., 1 = RAM, 2 = reserved, |
| 2913 | 3 = ACPI, 12 = PRAM. |
| 2914 | |
| 2915 | memory_corruption_check=0/1 [X86] |
| 2916 | Some BIOSes seem to corrupt the first 64k of |
| 2917 | memory when doing things like suspend/resume. |
| 2918 | Setting this option will scan the memory |
| 2919 | looking for corruption. Enabling this will |
| 2920 | both detect corruption and prevent the kernel |
| 2921 | from using the memory being corrupted. |
| 2922 | However, its intended as a diagnostic tool; if |
| 2923 | repeatable BIOS-originated corruption always |
| 2924 | affects the same memory, you can use memmap= |
| 2925 | to prevent the kernel from using that memory. |
| 2926 | |
| 2927 | memory_corruption_check_size=size [X86] |
| 2928 | By default it checks for corruption in the low |
| 2929 | 64k, making this memory unavailable for normal |
| 2930 | use. Use this parameter to scan for |
| 2931 | corruption in more or less memory. |
| 2932 | |
| 2933 | memory_corruption_check_period=seconds [X86] |
| 2934 | By default it checks for corruption every 60 |
| 2935 | seconds. Use this parameter to check at some |
| 2936 | other rate. 0 disables periodic checking. |
| 2937 | |
| 2938 | memory_hotplug.memmap_on_memory |
| 2939 | [KNL,X86,ARM] Boolean flag to enable this feature. |
| 2940 | Format: {on | off (default)} |
| 2941 | When enabled, runtime hotplugged memory will |
| 2942 | allocate its internal metadata (struct pages) |
| 2943 | from the hotadded memory which will allow to |
| 2944 | hotadd a lot of memory without requiring |
| 2945 | additional memory to do so. |
| 2946 | This feature is disabled by default because it |
| 2947 | has some implication on large (e.g. GB) |
| 2948 | allocations in some configurations (e.g. small |
| 2949 | memory blocks). |
| 2950 | The state of the flag can be read in |
| 2951 | /sys/module/memory_hotplug/parameters/memmap_on_memory. |
| 2952 | Note that even when enabled, there are a few cases where |
| 2953 | the feature is not effective. |
| 2954 | |
| 2955 | This is not compatible with hugetlb_free_vmemmap. If |
| 2956 | both parameters are enabled, hugetlb_free_vmemmap takes |
| 2957 | precedence over memory_hotplug.memmap_on_memory. |
| 2958 | |
| 2959 | memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest |
| 2960 | Format: <integer> |
| 2961 | default : 0 <disable> |
| 2962 | Specifies the number of memtest passes to be |
| 2963 | performed. Each pass selects another test |
| 2964 | pattern from a given set of patterns. Memtest |
| 2965 | fills the memory with this pattern, validates |
| 2966 | memory contents and reserves bad memory |
| 2967 | regions that are detected. |
| 2968 | |
| 2969 | mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control |
| 2970 | Valid arguments: on, off |
| 2971 | Default (depends on kernel configuration option): |
| 2972 | on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y) |
| 2973 | off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n) |
| 2974 | mem_encrypt=on: Activate SME |
| 2975 | mem_encrypt=off: Do not activate SME |
| 2976 | |
| 2977 | Refer to Documentation/virt/kvm/amd-memory-encryption.rst |
| 2978 | for details on when memory encryption can be activated. |
| 2979 | |
| 2980 | mem_sleep_default= [SUSPEND] Default system suspend mode: |
| 2981 | s2idle - Suspend-To-Idle |
| 2982 | shallow - Power-On Suspend or equivalent (if supported) |
| 2983 | deep - Suspend-To-RAM or equivalent (if supported) |
| 2984 | See Documentation/admin-guide/pm/sleep-states.rst. |
| 2985 | |
| 2986 | meye.*= [HW] Set MotionEye Camera parameters |
| 2987 | See Documentation/admin-guide/media/meye.rst. |
| 2988 | |
| 2989 | mfgpt_irq= [IA-32] Specify the IRQ to use for the |
| 2990 | Multi-Function General Purpose Timers on AMD Geode |
| 2991 | platforms. |
| 2992 | |
| 2993 | mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when |
| 2994 | the BIOS has incorrectly applied a workaround. TinyBIOS |
| 2995 | version 0.98 is known to be affected, 0.99 fixes the |
| 2996 | problem by letting the user disable the workaround. |
| 2997 | |
| 2998 | mga= [HW,DRM] |
| 2999 | |
| 3000 | min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this |
| 3001 | physical address is ignored. |
| 3002 | |
| 3003 | mini2440= [ARM,HW,KNL] |
| 3004 | Format:[0..2][b][c][t] |
| 3005 | Default: "0tb" |
| 3006 | MINI2440 configuration specification: |
| 3007 | 0 - The attached screen is the 3.5" TFT |
| 3008 | 1 - The attached screen is the 7" TFT |
| 3009 | 2 - The VGA Shield is attached (1024x768) |
| 3010 | Leaving out the screen size parameter will not load |
| 3011 | the TFT driver, and the framebuffer will be left |
| 3012 | unconfigured. |
| 3013 | b - Enable backlight. The TFT backlight pin will be |
| 3014 | linked to the kernel VESA blanking code and a GPIO |
| 3015 | LED. This parameter is not necessary when using the |
| 3016 | VGA shield. |
| 3017 | c - Enable the s3c camera interface. |
| 3018 | t - Reserved for enabling touchscreen support. The |
| 3019 | touchscreen support is not enabled in the mainstream |
| 3020 | kernel as of 2.6.30, a preliminary port can be found |
| 3021 | in the "bleeding edge" mini2440 support kernel at |
| 3022 | https://repo.or.cz/w/linux-2.6/mini2440.git |
| 3023 | |
| 3024 | mitigations= |
| 3025 | [X86,PPC,S390,ARM64] Control optional mitigations for |
| 3026 | CPU vulnerabilities. This is a set of curated, |
| 3027 | arch-independent options, each of which is an |
| 3028 | aggregation of existing arch-specific options. |
| 3029 | |
| 3030 | off |
| 3031 | Disable all optional CPU mitigations. This |
| 3032 | improves system performance, but it may also |
| 3033 | expose users to several CPU vulnerabilities. |
| 3034 | Equivalent to: nopti [X86,PPC] |
| 3035 | kpti=0 [ARM64] |
| 3036 | nospectre_v1 [X86,PPC] |
| 3037 | nobp=0 [S390] |
| 3038 | nospectre_v2 [X86,PPC,S390,ARM64] |
| 3039 | spectre_v2_user=off [X86] |
| 3040 | spec_store_bypass_disable=off [X86,PPC] |
| 3041 | ssbd=force-off [ARM64] |
| 3042 | l1tf=off [X86] |
| 3043 | mds=off [X86] |
| 3044 | tsx_async_abort=off [X86] |
| 3045 | kvm.nx_huge_pages=off [X86] |
| 3046 | no_entry_flush [PPC] |
| 3047 | no_uaccess_flush [PPC] |
| 3048 | |
| 3049 | Exceptions: |
| 3050 | This does not have any effect on |
| 3051 | kvm.nx_huge_pages when |
| 3052 | kvm.nx_huge_pages=force. |
| 3053 | |
| 3054 | auto (default) |
| 3055 | Mitigate all CPU vulnerabilities, but leave SMT |
| 3056 | enabled, even if it's vulnerable. This is for |
| 3057 | users who don't want to be surprised by SMT |
| 3058 | getting disabled across kernel upgrades, or who |
| 3059 | have other ways of avoiding SMT-based attacks. |
| 3060 | Equivalent to: (default behavior) |
| 3061 | |
| 3062 | auto,nosmt |
| 3063 | Mitigate all CPU vulnerabilities, disabling SMT |
| 3064 | if needed. This is for users who always want to |
| 3065 | be fully mitigated, even if it means losing SMT. |
| 3066 | Equivalent to: l1tf=flush,nosmt [X86] |
| 3067 | mds=full,nosmt [X86] |
| 3068 | tsx_async_abort=full,nosmt [X86] |
| 3069 | |
| 3070 | mminit_loglevel= |
| 3071 | [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this |
| 3072 | parameter allows control of the logging verbosity for |
| 3073 | the additional memory initialisation checks. A value |
| 3074 | of 0 disables mminit logging and a level of 4 will |
| 3075 | log everything. Information is printed at KERN_DEBUG |
| 3076 | so loglevel=8 may also need to be specified. |
| 3077 | |
| 3078 | module.sig_enforce |
| 3079 | [KNL] When CONFIG_MODULE_SIG is set, this means that |
| 3080 | modules without (valid) signatures will fail to load. |
| 3081 | Note that if CONFIG_MODULE_SIG_FORCE is set, that |
| 3082 | is always true, so this option does nothing. |
| 3083 | |
| 3084 | module_blacklist= [KNL] Do not load a comma-separated list of |
| 3085 | modules. Useful for debugging problem modules. |
| 3086 | |
| 3087 | mousedev.tap_time= |
| 3088 | [MOUSE] Maximum time between finger touching and |
| 3089 | leaving touchpad surface for touch to be considered |
| 3090 | a tap and be reported as a left button click (for |
| 3091 | touchpads working in absolute mode only). |
| 3092 | Format: <msecs> |
| 3093 | mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices |
| 3094 | reporting absolute coordinates, such as tablets |
| 3095 | mousedev.yres= [MOUSE] Vertical screen resolution, used for devices |
| 3096 | reporting absolute coordinates, such as tablets |
| 3097 | |
| 3098 | movablecore= [KNL,X86,IA-64,PPC] |
| 3099 | Format: nn[KMGTPE] | nn% |
| 3100 | This parameter is the complement to kernelcore=, it |
| 3101 | specifies the amount of memory used for migratable |
| 3102 | allocations. If both kernelcore and movablecore is |
| 3103 | specified, then kernelcore will be at *least* the |
| 3104 | specified value but may be more. If movablecore on its |
| 3105 | own is specified, the administrator must be careful |
| 3106 | that the amount of memory usable for all allocations |
| 3107 | is not too small. |
| 3108 | |
| 3109 | movable_node [KNL] Boot-time switch to make hotplugable memory |
| 3110 | NUMA nodes to be movable. This means that the memory |
| 3111 | of such nodes will be usable only for movable |
| 3112 | allocations which rules out almost all kernel |
| 3113 | allocations. Use with caution! |
| 3114 | |
| 3115 | MTD_Partition= [MTD] |
| 3116 | Format: <name>,<region-number>,<size>,<offset> |
| 3117 | |
| 3118 | MTD_Region= [MTD] Format: |
| 3119 | <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>] |
| 3120 | |
| 3121 | mtdparts= [MTD] |
| 3122 | See drivers/mtd/parsers/cmdlinepart.c |
| 3123 | |
| 3124 | multitce=off [PPC] This parameter disables the use of the pSeries |
| 3125 | firmware feature for updating multiple TCE entries |
| 3126 | at a time. |
| 3127 | |
| 3128 | onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration |
| 3129 | |
| 3130 | Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock] |
| 3131 | |
| 3132 | boundary - index of last SLC block on Flex-OneNAND. |
| 3133 | The remaining blocks are configured as MLC blocks. |
| 3134 | lock - Configure if Flex-OneNAND boundary should be locked. |
| 3135 | Once locked, the boundary cannot be changed. |
| 3136 | 1 indicates lock status, 0 indicates unlock status. |
| 3137 | |
| 3138 | mtdset= [ARM] |
| 3139 | ARM/S3C2412 JIVE boot control |
| 3140 | |
| 3141 | See arch/arm/mach-s3c/mach-jive.c |
| 3142 | |
| 3143 | mtouchusb.raw_coordinates= |
| 3144 | [HW] Make the MicroTouch USB driver use raw coordinates |
| 3145 | ('y', default) or cooked coordinates ('n') |
| 3146 | |
| 3147 | mtrr_chunk_size=nn[KMG] [X86] |
| 3148 | used for mtrr cleanup. It is largest continuous chunk |
| 3149 | that could hold holes aka. UC entries. |
| 3150 | |
| 3151 | mtrr_gran_size=nn[KMG] [X86] |
| 3152 | Used for mtrr cleanup. It is granularity of mtrr block. |
| 3153 | Default is 1. |
| 3154 | Large value could prevent small alignment from |
| 3155 | using up MTRRs. |
| 3156 | |
| 3157 | mtrr_spare_reg_nr=n [X86] |
| 3158 | Format: <integer> |
| 3159 | Range: 0,7 : spare reg number |
| 3160 | Default : 1 |
| 3161 | Used for mtrr cleanup. It is spare mtrr entries number. |
| 3162 | Set to 2 or more if your graphical card needs more. |
| 3163 | |
| 3164 | n2= [NET] SDL Inc. RISCom/N2 synchronous serial card |
| 3165 | |
| 3166 | netdev= [NET] Network devices parameters |
| 3167 | Format: <irq>,<io>,<mem_start>,<mem_end>,<name> |
| 3168 | Note that mem_start is often overloaded to mean |
| 3169 | something different and driver-specific. |
| 3170 | This usage is only documented in each driver source |
| 3171 | file if at all. |
| 3172 | |
| 3173 | nf_conntrack.acct= |
| 3174 | [NETFILTER] Enable connection tracking flow accounting |
| 3175 | 0 to disable accounting |
| 3176 | 1 to enable accounting |
| 3177 | Default value is 0. |
| 3178 | |
| 3179 | nfsaddrs= [NFS] Deprecated. Use ip= instead. |
| 3180 | See Documentation/admin-guide/nfs/nfsroot.rst. |
| 3181 | |
| 3182 | nfsroot= [NFS] nfs root filesystem for disk-less boxes. |
| 3183 | See Documentation/admin-guide/nfs/nfsroot.rst. |
| 3184 | |
| 3185 | nfsrootdebug [NFS] enable nfsroot debugging messages. |
| 3186 | See Documentation/admin-guide/nfs/nfsroot.rst. |
| 3187 | |
| 3188 | nfs.callback_nr_threads= |
| 3189 | [NFSv4] set the total number of threads that the |
| 3190 | NFS client will assign to service NFSv4 callback |
| 3191 | requests. |
| 3192 | |
| 3193 | nfs.callback_tcpport= |
| 3194 | [NFS] set the TCP port on which the NFSv4 callback |
| 3195 | channel should listen. |
| 3196 | |
| 3197 | nfs.cache_getent= |
| 3198 | [NFS] sets the pathname to the program which is used |
| 3199 | to update the NFS client cache entries. |
| 3200 | |
| 3201 | nfs.cache_getent_timeout= |
| 3202 | [NFS] sets the timeout after which an attempt to |
| 3203 | update a cache entry is deemed to have failed. |
| 3204 | |
| 3205 | nfs.idmap_cache_timeout= |
| 3206 | [NFS] set the maximum lifetime for idmapper cache |
| 3207 | entries. |
| 3208 | |
| 3209 | nfs.enable_ino64= |
| 3210 | [NFS] enable 64-bit inode numbers. |
| 3211 | If zero, the NFS client will fake up a 32-bit inode |
| 3212 | number for the readdir() and stat() syscalls instead |
| 3213 | of returning the full 64-bit number. |
| 3214 | The default is to return 64-bit inode numbers. |
| 3215 | |
| 3216 | nfs.max_session_cb_slots= |
| 3217 | [NFSv4.1] Sets the maximum number of session |
| 3218 | slots the client will assign to the callback |
| 3219 | channel. This determines the maximum number of |
| 3220 | callbacks the client will process in parallel for |
| 3221 | a particular server. |
| 3222 | |
| 3223 | nfs.max_session_slots= |
| 3224 | [NFSv4.1] Sets the maximum number of session slots |
| 3225 | the client will attempt to negotiate with the server. |
| 3226 | This limits the number of simultaneous RPC requests |
| 3227 | that the client can send to the NFSv4.1 server. |
| 3228 | Note that there is little point in setting this |
| 3229 | value higher than the max_tcp_slot_table_limit. |
| 3230 | |
| 3231 | nfs.nfs4_disable_idmapping= |
| 3232 | [NFSv4] When set to the default of '1', this option |
| 3233 | ensures that both the RPC level authentication |
| 3234 | scheme and the NFS level operations agree to use |
| 3235 | numeric uids/gids if the mount is using the |
| 3236 | 'sec=sys' security flavour. In effect it is |
| 3237 | disabling idmapping, which can make migration from |
| 3238 | legacy NFSv2/v3 systems to NFSv4 easier. |
| 3239 | Servers that do not support this mode of operation |
| 3240 | will be autodetected by the client, and it will fall |
| 3241 | back to using the idmapper. |
| 3242 | To turn off this behaviour, set the value to '0'. |
| 3243 | nfs.nfs4_unique_id= |
| 3244 | [NFS4] Specify an additional fixed unique ident- |
| 3245 | ification string that NFSv4 clients can insert into |
| 3246 | their nfs_client_id4 string. This is typically a |
| 3247 | UUID that is generated at system install time. |
| 3248 | |
| 3249 | nfs.send_implementation_id = |
| 3250 | [NFSv4.1] Send client implementation identification |
| 3251 | information in exchange_id requests. |
| 3252 | If zero, no implementation identification information |
| 3253 | will be sent. |
| 3254 | The default is to send the implementation identification |
| 3255 | information. |
| 3256 | |
| 3257 | nfs.recover_lost_locks = |
| 3258 | [NFSv4] Attempt to recover locks that were lost due |
| 3259 | to a lease timeout on the server. Please note that |
| 3260 | doing this risks data corruption, since there are |
| 3261 | no guarantees that the file will remain unchanged |
| 3262 | after the locks are lost. |
| 3263 | If you want to enable the kernel legacy behaviour of |
| 3264 | attempting to recover these locks, then set this |
| 3265 | parameter to '1'. |
| 3266 | The default parameter value of '0' causes the kernel |
| 3267 | not to attempt recovery of lost locks. |
| 3268 | |
| 3269 | nfs4.layoutstats_timer = |
| 3270 | [NFSv4.2] Change the rate at which the kernel sends |
| 3271 | layoutstats to the pNFS metadata server. |
| 3272 | |
| 3273 | Setting this to value to 0 causes the kernel to use |
| 3274 | whatever value is the default set by the layout |
| 3275 | driver. A non-zero value sets the minimum interval |
| 3276 | in seconds between layoutstats transmissions. |
| 3277 | |
| 3278 | nfsd.inter_copy_offload_enable = |
| 3279 | [NFSv4.2] When set to 1, the server will support |
| 3280 | server-to-server copies for which this server is |
| 3281 | the destination of the copy. |
| 3282 | |
| 3283 | nfsd.nfsd4_ssc_umount_timeout = |
| 3284 | [NFSv4.2] When used as the destination of a |
| 3285 | server-to-server copy, knfsd temporarily mounts |
| 3286 | the source server. It caches the mount in case |
| 3287 | it will be needed again, and discards it if not |
| 3288 | used for the number of milliseconds specified by |
| 3289 | this parameter. |
| 3290 | |
| 3291 | nfsd.nfs4_disable_idmapping= |
| 3292 | [NFSv4] When set to the default of '1', the NFSv4 |
| 3293 | server will return only numeric uids and gids to |
| 3294 | clients using auth_sys, and will accept numeric uids |
| 3295 | and gids from such clients. This is intended to ease |
| 3296 | migration from NFSv2/v3. |
| 3297 | |
| 3298 | |
| 3299 | nmi_backtrace.backtrace_idle [KNL] |
| 3300 | Dump stacks even of idle CPUs in response to an |
| 3301 | NMI stack-backtrace request. |
| 3302 | |
| 3303 | nmi_debug= [KNL,SH] Specify one or more actions to take |
| 3304 | when a NMI is triggered. |
| 3305 | Format: [state][,regs][,debounce][,die] |
| 3306 | |
| 3307 | nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels |
| 3308 | Format: [panic,][nopanic,][num] |
| 3309 | Valid num: 0 or 1 |
| 3310 | 0 - turn hardlockup detector in nmi_watchdog off |
| 3311 | 1 - turn hardlockup detector in nmi_watchdog on |
| 3312 | When panic is specified, panic when an NMI watchdog |
| 3313 | timeout occurs (or 'nopanic' to not panic on an NMI |
| 3314 | watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set) |
| 3315 | To disable both hard and soft lockup detectors, |
| 3316 | please see 'nowatchdog'. |
| 3317 | This is useful when you use a panic=... timeout and |
| 3318 | need the box quickly up again. |
| 3319 | |
| 3320 | These settings can be accessed at runtime via |
| 3321 | the nmi_watchdog and hardlockup_panic sysctls. |
| 3322 | |
| 3323 | netpoll.carrier_timeout= |
| 3324 | [NET] Specifies amount of time (in seconds) that |
| 3325 | netpoll should wait for a carrier. By default netpoll |
| 3326 | waits 4 seconds. |
| 3327 | |
| 3328 | no387 [BUGS=X86-32] Tells the kernel to use the 387 maths |
| 3329 | emulation library even if a 387 maths coprocessor |
| 3330 | is present. |
| 3331 | |
| 3332 | no5lvl [X86-64] Disable 5-level paging mode. Forces |
| 3333 | kernel to use 4-level paging instead. |
| 3334 | |
| 3335 | nofsgsbase [X86] Disables FSGSBASE instructions. |
| 3336 | |
| 3337 | no_console_suspend |
| 3338 | [HW] Never suspend the console |
| 3339 | Disable suspending of consoles during suspend and |
| 3340 | hibernate operations. Once disabled, debugging |
| 3341 | messages can reach various consoles while the rest |
| 3342 | of the system is being put to sleep (ie, while |
| 3343 | debugging driver suspend/resume hooks). This may |
| 3344 | not work reliably with all consoles, but is known |
| 3345 | to work with serial and VGA consoles. |
| 3346 | To facilitate more flexible debugging, we also add |
| 3347 | console_suspend, a printk module parameter to control |
| 3348 | it. Users could use console_suspend (usually |
| 3349 | /sys/module/printk/parameters/console_suspend) to |
| 3350 | turn on/off it dynamically. |
| 3351 | |
| 3352 | novmcoredd [KNL,KDUMP] |
| 3353 | Disable device dump. Device dump allows drivers to |
| 3354 | append dump data to vmcore so you can collect driver |
| 3355 | specified debug info. Drivers can append the data |
| 3356 | without any limit and this data is stored in memory, |
| 3357 | so this may cause significant memory stress. Disabling |
| 3358 | device dump can help save memory but the driver debug |
| 3359 | data will be no longer available. This parameter |
| 3360 | is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP |
| 3361 | is set. |
| 3362 | |
| 3363 | noaliencache [MM, NUMA, SLAB] Disables the allocation of alien |
| 3364 | caches in the slab allocator. Saves per-node memory, |
| 3365 | but will impact performance. |
| 3366 | |
| 3367 | noalign [KNL,ARM] |
| 3368 | |
| 3369 | noaltinstr [S390] Disables alternative instructions patching |
| 3370 | (CPU alternatives feature). |
| 3371 | |
| 3372 | noapic [SMP,APIC] Tells the kernel to not make use of any |
| 3373 | IOAPICs that may be present in the system. |
| 3374 | |
| 3375 | noautogroup Disable scheduler automatic task group creation. |
| 3376 | |
| 3377 | nobats [PPC] Do not use BATs for mapping kernel lowmem |
| 3378 | on "Classic" PPC cores. |
| 3379 | |
| 3380 | nocache [ARM] |
| 3381 | |
| 3382 | noclflush [BUGS=X86] Don't use the CLFLUSH instruction |
| 3383 | |
| 3384 | delayacct [KNL] Enable per-task delay accounting |
| 3385 | |
| 3386 | nodsp [SH] Disable hardware DSP at boot time. |
| 3387 | |
| 3388 | noefi Disable EFI runtime services support. |
| 3389 | |
| 3390 | no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel. |
| 3391 | |
| 3392 | noexec [IA-64] |
| 3393 | |
| 3394 | noexec [X86] |
| 3395 | On X86-32 available only on PAE configured kernels. |
| 3396 | noexec=on: enable non-executable mappings (default) |
| 3397 | noexec=off: disable non-executable mappings |
| 3398 | |
| 3399 | nosmap [X86,PPC] |
| 3400 | Disable SMAP (Supervisor Mode Access Prevention) |
| 3401 | even if it is supported by processor. |
| 3402 | |
| 3403 | nosmep [X86,PPC64s] |
| 3404 | Disable SMEP (Supervisor Mode Execution Prevention) |
| 3405 | even if it is supported by processor. |
| 3406 | |
| 3407 | noexec32 [X86-64] |
| 3408 | This affects only 32-bit executables. |
| 3409 | noexec32=on: enable non-executable mappings (default) |
| 3410 | read doesn't imply executable mappings |
| 3411 | noexec32=off: disable non-executable mappings |
| 3412 | read implies executable mappings |
| 3413 | |
| 3414 | nofpu [MIPS,SH] Disable hardware FPU at boot time. |
| 3415 | |
| 3416 | nofxsr [BUGS=X86-32] Disables x86 floating point extended |
| 3417 | register save and restore. The kernel will only save |
| 3418 | legacy floating-point registers on task switch. |
| 3419 | |
| 3420 | nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings. |
| 3421 | |
| 3422 | nohugevmalloc [PPC] Disable kernel huge vmalloc mappings. |
| 3423 | |
| 3424 | nosmt [KNL,S390] Disable symmetric multithreading (SMT). |
| 3425 | Equivalent to smt=1. |
| 3426 | |
| 3427 | [KNL,X86] Disable symmetric multithreading (SMT). |
| 3428 | nosmt=force: Force disable SMT, cannot be undone |
| 3429 | via the sysfs control file. |
| 3430 | |
| 3431 | nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1 |
| 3432 | (bounds check bypass). With this option data leaks are |
| 3433 | possible in the system. |
| 3434 | |
| 3435 | nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for |
| 3436 | the Spectre variant 2 (indirect branch prediction) |
| 3437 | vulnerability. System may allow data leaks with this |
| 3438 | option. |
| 3439 | |
| 3440 | nospec_store_bypass_disable |
| 3441 | [HW] Disable all mitigations for the Speculative Store Bypass vulnerability |
| 3442 | |
| 3443 | no_uaccess_flush |
| 3444 | [PPC] Don't flush the L1-D cache after accessing user data. |
| 3445 | |
| 3446 | noxsave [BUGS=X86] Disables x86 extended register state save |
| 3447 | and restore using xsave. The kernel will fallback to |
| 3448 | enabling legacy floating-point and sse state. |
| 3449 | |
| 3450 | noxsaveopt [X86] Disables xsaveopt used in saving x86 extended |
| 3451 | register states. The kernel will fall back to use |
| 3452 | xsave to save the states. By using this parameter, |
| 3453 | performance of saving the states is degraded because |
| 3454 | xsave doesn't support modified optimization while |
| 3455 | xsaveopt supports it on xsaveopt enabled systems. |
| 3456 | |
| 3457 | noxsaves [X86] Disables xsaves and xrstors used in saving and |
| 3458 | restoring x86 extended register state in compacted |
| 3459 | form of xsave area. The kernel will fall back to use |
| 3460 | xsaveopt and xrstor to save and restore the states |
| 3461 | in standard form of xsave area. By using this |
| 3462 | parameter, xsave area per process might occupy more |
| 3463 | memory on xsaves enabled systems. |
| 3464 | |
| 3465 | nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait |
| 3466 | in do_idle() and not use the arch_cpu_idle() |
| 3467 | implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP |
| 3468 | to be effective. This is useful on platforms where the |
| 3469 | sleep(SH) or wfi(ARM,ARM64) instructions do not work |
| 3470 | correctly or when doing power measurements to evalute |
| 3471 | the impact of the sleep instructions. This is also |
| 3472 | useful when using JTAG debugger. |
| 3473 | |
| 3474 | no_file_caps Tells the kernel not to honor file capabilities. The |
| 3475 | only way then for a file to be executed with privilege |
| 3476 | is to be setuid root or executed by root. |
| 3477 | |
| 3478 | nohalt [IA-64] Tells the kernel not to use the power saving |
| 3479 | function PAL_HALT_LIGHT when idle. This increases |
| 3480 | power-consumption. On the positive side, it reduces |
| 3481 | interrupt wake-up latency, which may improve performance |
| 3482 | in certain environments such as networked servers or |
| 3483 | real-time systems. |
| 3484 | |
| 3485 | no_hash_pointers |
| 3486 | Force pointers printed to the console or buffers to be |
| 3487 | unhashed. By default, when a pointer is printed via %p |
| 3488 | format string, that pointer is "hashed", i.e. obscured |
| 3489 | by hashing the pointer value. This is a security feature |
| 3490 | that hides actual kernel addresses from unprivileged |
| 3491 | users, but it also makes debugging the kernel more |
| 3492 | difficult since unequal pointers can no longer be |
| 3493 | compared. However, if this command-line option is |
| 3494 | specified, then all normal pointers will have their true |
| 3495 | value printed. Pointers printed via %pK may still be |
| 3496 | hashed. This option should only be specified when |
| 3497 | debugging the kernel. Please do not use on production |
| 3498 | kernels. |
| 3499 | |
| 3500 | nohibernate [HIBERNATION] Disable hibernation and resume. |
| 3501 | |
| 3502 | nohz= [KNL] Boottime enable/disable dynamic ticks |
| 3503 | Valid arguments: on, off |
| 3504 | Default: on |
| 3505 | |
| 3506 | nohz_full= [KNL,BOOT,SMP,ISOL] |
| 3507 | The argument is a cpu list, as described above. |
| 3508 | In kernels built with CONFIG_NO_HZ_FULL=y, set |
| 3509 | the specified list of CPUs whose tick will be stopped |
| 3510 | whenever possible. The boot CPU will be forced outside |
| 3511 | the range to maintain the timekeeping. Any CPUs |
| 3512 | in this list will have their RCU callbacks offloaded, |
| 3513 | just as if they had also been called out in the |
| 3514 | rcu_nocbs= boot parameter. |
| 3515 | |
| 3516 | noiotrap [SH] Disables trapped I/O port accesses. |
| 3517 | |
| 3518 | noirqdebug [X86-32] Disables the code which attempts to detect and |
| 3519 | disable unhandled interrupt sources. |
| 3520 | |
| 3521 | no_timer_check [X86,APIC] Disables the code which tests for |
| 3522 | broken timer IRQ sources. |
| 3523 | |
| 3524 | noisapnp [ISAPNP] Disables ISA PnP code. |
| 3525 | |
| 3526 | noinitrd [RAM] Tells the kernel not to load any configured |
| 3527 | initial RAM disk. |
| 3528 | |
| 3529 | nointremap [X86-64, Intel-IOMMU] Do not enable interrupt |
| 3530 | remapping. |
| 3531 | [Deprecated - use intremap=off] |
| 3532 | |
| 3533 | nointroute [IA-64] |
| 3534 | |
| 3535 | noinvpcid [X86] Disable the INVPCID cpu feature. |
| 3536 | |
| 3537 | nojitter [IA-64] Disables jitter checking for ITC timers. |
| 3538 | |
| 3539 | no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver |
| 3540 | |
| 3541 | no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page |
| 3542 | fault handling. |
| 3543 | |
| 3544 | no-vmw-sched-clock |
| 3545 | [X86,PV_OPS] Disable paravirtualized VMware scheduler |
| 3546 | clock and use the default one. |
| 3547 | |
| 3548 | no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time |
| 3549 | accounting. steal time is computed, but won't |
| 3550 | influence scheduler behaviour |
| 3551 | |
| 3552 | nolapic [X86-32,APIC] Do not enable or use the local APIC. |
| 3553 | |
| 3554 | nolapic_timer [X86-32,APIC] Do not use the local APIC timer. |
| 3555 | |
| 3556 | noltlbs [PPC] Do not use large page/tlb entries for kernel |
| 3557 | lowmem mapping on PPC40x and PPC8xx |
| 3558 | |
| 3559 | nomca [IA-64] Disable machine check abort handling |
| 3560 | |
| 3561 | nomce [X86-32] Disable Machine Check Exception |
| 3562 | |
| 3563 | nomfgpt [X86-32] Disable Multi-Function General Purpose |
| 3564 | Timer usage (for AMD Geode machines). |
| 3565 | |
| 3566 | nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to |
| 3567 | shutdown the other cpus. Instead use the REBOOT_VECTOR |
| 3568 | irq. |
| 3569 | |
| 3570 | nomodeset Disable kernel modesetting. DRM drivers will not perform |
| 3571 | display-mode changes or accelerated rendering. Only the |
| 3572 | system framebuffer will be available for use if this was |
| 3573 | set-up by the firmware or boot loader. |
| 3574 | |
| 3575 | Useful as fallback, or for testing and debugging. |
| 3576 | |
| 3577 | nomodule Disable module load |
| 3578 | |
| 3579 | nopat [X86] Disable PAT (page attribute table extension of |
| 3580 | pagetables) support. |
| 3581 | |
| 3582 | nopcid [X86-64] Disable the PCID cpu feature. |
| 3583 | |
| 3584 | norandmaps Don't use address space randomization. Equivalent to |
| 3585 | echo 0 > /proc/sys/kernel/randomize_va_space |
| 3586 | |
| 3587 | noreplace-smp [X86-32,SMP] Don't replace SMP instructions |
| 3588 | with UP alternatives |
| 3589 | |
| 3590 | nordrand [X86] Disable kernel use of the RDRAND and |
| 3591 | RDSEED instructions even if they are supported |
| 3592 | by the processor. RDRAND and RDSEED are still |
| 3593 | available to user space applications. |
| 3594 | |
| 3595 | noresume [SWSUSP] Disables resume and restores original swap |
| 3596 | space. |
| 3597 | |
| 3598 | no-scroll [VGA] Disables scrollback. |
| 3599 | This is required for the Braillex ib80-piezo Braille |
| 3600 | reader made by F.H. Papenmeier (Germany). |
| 3601 | |
| 3602 | nosbagart [IA-64] |
| 3603 | |
| 3604 | nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support. |
| 3605 | |
| 3606 | nosgx [X86-64,SGX] Disables Intel SGX kernel support. |
| 3607 | |
| 3608 | nosmp [SMP] Tells an SMP kernel to act as a UP kernel, |
| 3609 | and disable the IO APIC. legacy for "maxcpus=0". |
| 3610 | |
| 3611 | nosoftlockup [KNL] Disable the soft-lockup detector. |
| 3612 | |
| 3613 | nosync [HW,M68K] Disables sync negotiation for all devices. |
| 3614 | |
| 3615 | nowatchdog [KNL] Disable both lockup detectors, i.e. |
| 3616 | soft-lockup and NMI watchdog (hard-lockup). |
| 3617 | |
| 3618 | nowb [ARM] |
| 3619 | |
| 3620 | nox2apic [X86-64,APIC] Do not enable x2APIC mode. |
| 3621 | |
| 3622 | cpu0_hotplug [X86] Turn on CPU0 hotplug feature when |
| 3623 | CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off. |
| 3624 | Some features depend on CPU0. Known dependencies are: |
| 3625 | 1. Resume from suspend/hibernate depends on CPU0. |
| 3626 | Suspend/hibernate will fail if CPU0 is offline and you |
| 3627 | need to online CPU0 before suspend/hibernate. |
| 3628 | 2. PIC interrupts also depend on CPU0. CPU0 can't be |
| 3629 | removed if a PIC interrupt is detected. |
| 3630 | It's said poweroff/reboot may depend on CPU0 on some |
| 3631 | machines although I haven't seen such issues so far |
| 3632 | after CPU0 is offline on a few tested machines. |
| 3633 | If the dependencies are under your control, you can |
| 3634 | turn on cpu0_hotplug. |
| 3635 | |
| 3636 | nps_mtm_hs_ctr= [KNL,ARC] |
| 3637 | This parameter sets the maximum duration, in |
| 3638 | cycles, each HW thread of the CTOP can run |
| 3639 | without interruptions, before HW switches it. |
| 3640 | The actual maximum duration is 16 times this |
| 3641 | parameter's value. |
| 3642 | Format: integer between 1 and 255 |
| 3643 | Default: 255 |
| 3644 | |
| 3645 | nptcg= [IA-64] Override max number of concurrent global TLB |
| 3646 | purges which is reported from either PAL_VM_SUMMARY or |
| 3647 | SAL PALO. |
| 3648 | |
| 3649 | nr_cpus= [SMP] Maximum number of processors that an SMP kernel |
| 3650 | could support. nr_cpus=n : n >= 1 limits the kernel to |
| 3651 | support 'n' processors. It could be larger than the |
| 3652 | number of already plugged CPU during bootup, later in |
| 3653 | runtime you can physically add extra cpu until it reaches |
| 3654 | n. So during boot up some boot time memory for per-cpu |
| 3655 | variables need be pre-allocated for later physical cpu |
| 3656 | hot plugging. |
| 3657 | |
| 3658 | nr_uarts= [SERIAL] maximum number of UARTs to be registered. |
| 3659 | |
| 3660 | numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only |
| 3661 | set up a single NUMA node spanning all memory. |
| 3662 | |
| 3663 | numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic |
| 3664 | NUMA balancing. |
| 3665 | Allowed values are enable and disable |
| 3666 | |
| 3667 | numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA. |
| 3668 | 'node', 'default' can be specified |
| 3669 | This can be set from sysctl after boot. |
| 3670 | See Documentation/admin-guide/sysctl/vm.rst for details. |
| 3671 | |
| 3672 | ohci1394_dma=early [HW] enable debugging via the ohci1394 driver. |
| 3673 | See Documentation/core-api/debugging-via-ohci1394.rst for more |
| 3674 | info. |
| 3675 | |
| 3676 | olpc_ec_timeout= [OLPC] ms delay when issuing EC commands |
| 3677 | Rather than timing out after 20 ms if an EC |
| 3678 | command is not properly ACKed, override the length |
| 3679 | of the timeout. We have interrupts disabled while |
| 3680 | waiting for the ACK, so if this is set too high |
| 3681 | interrupts *may* be lost! |
| 3682 | |
| 3683 | omap_mux= [OMAP] Override bootloader pin multiplexing. |
| 3684 | Format: <mux_mode0.mode_name=value>... |
| 3685 | For example, to override I2C bus2: |
| 3686 | omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100 |
| 3687 | |
| 3688 | oops=panic Always panic on oopses. Default is to just kill the |
| 3689 | process, but there is a small probability of |
| 3690 | deadlocking the machine. |
| 3691 | This will also cause panics on machine check exceptions. |
| 3692 | Useful together with panic=30 to trigger a reboot. |
| 3693 | |
| 3694 | page_alloc.shuffle= |
| 3695 | [KNL] Boolean flag to control whether the page allocator |
| 3696 | should randomize its free lists. The randomization may |
| 3697 | be automatically enabled if the kernel detects it is |
| 3698 | running on a platform with a direct-mapped memory-side |
| 3699 | cache, and this parameter can be used to |
| 3700 | override/disable that behavior. The state of the flag |
| 3701 | can be read from sysfs at: |
| 3702 | /sys/module/page_alloc/parameters/shuffle. |
| 3703 | |
| 3704 | page_owner= [KNL] Boot-time page_owner enabling option. |
| 3705 | Storage of the information about who allocated |
| 3706 | each page is disabled in default. With this switch, |
| 3707 | we can turn it on. |
| 3708 | on: enable the feature |
| 3709 | |
| 3710 | page_poison= [KNL] Boot-time parameter changing the state of |
| 3711 | poisoning on the buddy allocator, available with |
| 3712 | CONFIG_PAGE_POISONING=y. |
| 3713 | off: turn off poisoning (default) |
| 3714 | on: turn on poisoning |
| 3715 | |
| 3716 | page_reporting.page_reporting_order= |
| 3717 | [KNL] Minimal page reporting order |
| 3718 | Format: <integer> |
| 3719 | Adjust the minimal page reporting order. The page |
| 3720 | reporting is disabled when it exceeds (MAX_ORDER-1). |
| 3721 | |
| 3722 | panic= [KNL] Kernel behaviour on panic: delay <timeout> |
| 3723 | timeout > 0: seconds before rebooting |
| 3724 | timeout = 0: wait forever |
| 3725 | timeout < 0: reboot immediately |
| 3726 | Format: <timeout> |
| 3727 | |
| 3728 | panic_print= Bitmask for printing system info when panic happens. |
| 3729 | User can chose combination of the following bits: |
| 3730 | bit 0: print all tasks info |
| 3731 | bit 1: print system memory info |
| 3732 | bit 2: print timer info |
| 3733 | bit 3: print locks info if CONFIG_LOCKDEP is on |
| 3734 | bit 4: print ftrace buffer |
| 3735 | bit 5: print all printk messages in buffer |
| 3736 | |
| 3737 | panic_on_taint= Bitmask for conditionally calling panic() in add_taint() |
| 3738 | Format: <hex>[,nousertaint] |
| 3739 | Hexadecimal bitmask representing the set of TAINT flags |
| 3740 | that will cause the kernel to panic when add_taint() is |
| 3741 | called with any of the flags in this set. |
| 3742 | The optional switch "nousertaint" can be utilized to |
| 3743 | prevent userspace forced crashes by writing to sysctl |
| 3744 | /proc/sys/kernel/tainted any flagset matching with the |
| 3745 | bitmask set on panic_on_taint. |
| 3746 | See Documentation/admin-guide/tainted-kernels.rst for |
| 3747 | extra details on the taint flags that users can pick |
| 3748 | to compose the bitmask to assign to panic_on_taint. |
| 3749 | |
| 3750 | panic_on_warn panic() instead of WARN(). Useful to cause kdump |
| 3751 | on a WARN(). |
| 3752 | |
| 3753 | crash_kexec_post_notifiers |
| 3754 | Run kdump after running panic-notifiers and dumping |
| 3755 | kmsg. This only for the users who doubt kdump always |
| 3756 | succeeds in any situation. |
| 3757 | Note that this also increases risks of kdump failure, |
| 3758 | because some panic notifiers can make the crashed |
| 3759 | kernel more unstable. |
| 3760 | |
| 3761 | parkbd.port= [HW] Parallel port number the keyboard adapter is |
| 3762 | connected to, default is 0. |
| 3763 | Format: <parport#> |
| 3764 | parkbd.mode= [HW] Parallel port keyboard adapter mode of operation, |
| 3765 | 0 for XT, 1 for AT (default is AT). |
| 3766 | Format: <mode> |
| 3767 | |
| 3768 | parport= [HW,PPT] Specify parallel ports. 0 disables. |
| 3769 | Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] } |
| 3770 | Use 'auto' to force the driver to use any |
| 3771 | IRQ/DMA settings detected (the default is to |
| 3772 | ignore detected IRQ/DMA settings because of |
| 3773 | possible conflicts). You can specify the base |
| 3774 | address, IRQ, and DMA settings; IRQ and DMA |
| 3775 | should be numbers, or 'auto' (for using detected |
| 3776 | settings on that particular port), or 'nofifo' |
| 3777 | (to avoid using a FIFO even if it is detected). |
| 3778 | Parallel ports are assigned in the order they |
| 3779 | are specified on the command line, starting |
| 3780 | with parport0. |
| 3781 | |
| 3782 | parport_init_mode= [HW,PPT] |
| 3783 | Configure VIA parallel port to operate in |
| 3784 | a specific mode. This is necessary on Pegasos |
| 3785 | computer where firmware has no options for setting |
| 3786 | up parallel port mode and sets it to spp. |
| 3787 | Currently this function knows 686a and 8231 chips. |
| 3788 | Format: [spp|ps2|epp|ecp|ecpepp] |
| 3789 | |
| 3790 | pata_legacy.all= [HW,LIBATA] |
| 3791 | Format: <int> |
| 3792 | Set to non-zero to probe primary and secondary ISA |
| 3793 | port ranges on PCI systems where no PCI PATA device |
| 3794 | has been found at either range. Disabled by default. |
| 3795 | |
| 3796 | pata_legacy.autospeed= [HW,LIBATA] |
| 3797 | Format: <int> |
| 3798 | Set to non-zero if a chip is present that snoops speed |
| 3799 | changes. Disabled by default. |
| 3800 | |
| 3801 | pata_legacy.ht6560a= [HW,LIBATA] |
| 3802 | Format: <int> |
| 3803 | Set to 1, 2, or 3 for HT 6560A on the primary channel, |
| 3804 | the secondary channel, or both channels respectively. |
| 3805 | Disabled by default. |
| 3806 | |
| 3807 | pata_legacy.ht6560b= [HW,LIBATA] |
| 3808 | Format: <int> |
| 3809 | Set to 1, 2, or 3 for HT 6560B on the primary channel, |
| 3810 | the secondary channel, or both channels respectively. |
| 3811 | Disabled by default. |
| 3812 | |
| 3813 | pata_legacy.iordy_mask= [HW,LIBATA] |
| 3814 | Format: <int> |
| 3815 | IORDY enable mask. Set individual bits to allow IORDY |
| 3816 | for the respective channel. Bit 0 is for the first |
| 3817 | legacy channel handled by this driver, bit 1 is for |
| 3818 | the second channel, and so on. The sequence will often |
| 3819 | correspond to the primary legacy channel, the secondary |
| 3820 | legacy channel, and so on, but the handling of a PCI |
| 3821 | bus and the use of other driver options may interfere |
| 3822 | with the sequence. By default IORDY is allowed across |
| 3823 | all channels. |
| 3824 | |
| 3825 | pata_legacy.opti82c46x= [HW,LIBATA] |
| 3826 | Format: <int> |
| 3827 | Set to 1, 2, or 3 for Opti 82c611A on the primary |
| 3828 | channel, the secondary channel, or both channels |
| 3829 | respectively. Disabled by default. |
| 3830 | |
| 3831 | pata_legacy.opti82c611a= [HW,LIBATA] |
| 3832 | Format: <int> |
| 3833 | Set to 1, 2, or 3 for Opti 82c465MV on the primary |
| 3834 | channel, the secondary channel, or both channels |
| 3835 | respectively. Disabled by default. |
| 3836 | |
| 3837 | pata_legacy.pio_mask= [HW,LIBATA] |
| 3838 | Format: <int> |
| 3839 | PIO mode mask for autospeed devices. Set individual |
| 3840 | bits to allow the use of the respective PIO modes. |
| 3841 | Bit 0 is for mode 0, bit 1 is for mode 1, and so on. |
| 3842 | All modes allowed by default. |
| 3843 | |
| 3844 | pata_legacy.probe_all= [HW,LIBATA] |
| 3845 | Format: <int> |
| 3846 | Set to non-zero to probe tertiary and further ISA |
| 3847 | port ranges on PCI systems. Disabled by default. |
| 3848 | |
| 3849 | pata_legacy.probe_mask= [HW,LIBATA] |
| 3850 | Format: <int> |
| 3851 | Probe mask for legacy ISA PATA ports. Depending on |
| 3852 | platform configuration and the use of other driver |
| 3853 | options up to 6 legacy ports are supported: 0x1f0, |
| 3854 | 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing |
| 3855 | of individual ports can be disabled by setting the |
| 3856 | corresponding bits in the mask to 1. Bit 0 is for |
| 3857 | the first port in the list above (0x1f0), and so on. |
| 3858 | By default all supported ports are probed. |
| 3859 | |
| 3860 | pata_legacy.qdi= [HW,LIBATA] |
| 3861 | Format: <int> |
| 3862 | Set to non-zero to probe QDI controllers. By default |
| 3863 | set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise. |
| 3864 | |
| 3865 | pata_legacy.winbond= [HW,LIBATA] |
| 3866 | Format: <int> |
| 3867 | Set to non-zero to probe Winbond controllers. Use |
| 3868 | the standard I/O port (0x130) if 1, otherwise the |
| 3869 | value given is the I/O port to use (typically 0x1b0). |
| 3870 | By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE, |
| 3871 | 0 otherwise. |
| 3872 | |
| 3873 | pata_platform.pio_mask= [HW,LIBATA] |
| 3874 | Format: <int> |
| 3875 | Supported PIO mode mask. Set individual bits to allow |
| 3876 | the use of the respective PIO modes. Bit 0 is for |
| 3877 | mode 0, bit 1 is for mode 1, and so on. Mode 0 only |
| 3878 | allowed by default. |
| 3879 | |
| 3880 | pause_on_oops= |
| 3881 | Halt all CPUs after the first oops has been printed for |
| 3882 | the specified number of seconds. This is to be used if |
| 3883 | your oopses keep scrolling off the screen. |
| 3884 | |
| 3885 | pcbit= [HW,ISDN] |
| 3886 | |
| 3887 | pcd. [PARIDE] |
| 3888 | See header of drivers/block/paride/pcd.c. |
| 3889 | See also Documentation/admin-guide/blockdev/paride.rst. |
| 3890 | |
| 3891 | pci=option[,option...] [PCI] various PCI subsystem options. |
| 3892 | |
| 3893 | Some options herein operate on a specific device |
| 3894 | or a set of devices (<pci_dev>). These are |
| 3895 | specified in one of the following formats: |
| 3896 | |
| 3897 | [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]* |
| 3898 | pci:<vendor>:<device>[:<subvendor>:<subdevice>] |
| 3899 | |
| 3900 | Note: the first format specifies a PCI |
| 3901 | bus/device/function address which may change |
| 3902 | if new hardware is inserted, if motherboard |
| 3903 | firmware changes, or due to changes caused |
| 3904 | by other kernel parameters. If the |
| 3905 | domain is left unspecified, it is |
| 3906 | taken to be zero. Optionally, a path |
| 3907 | to a device through multiple device/function |
| 3908 | addresses can be specified after the base |
| 3909 | address (this is more robust against |
| 3910 | renumbering issues). The second format |
| 3911 | selects devices using IDs from the |
| 3912 | configuration space which may match multiple |
| 3913 | devices in the system. |
| 3914 | |
| 3915 | earlydump dump PCI config space before the kernel |
| 3916 | changes anything |
| 3917 | off [X86] don't probe for the PCI bus |
| 3918 | bios [X86-32] force use of PCI BIOS, don't access |
| 3919 | the hardware directly. Use this if your machine |
| 3920 | has a non-standard PCI host bridge. |
| 3921 | nobios [X86-32] disallow use of PCI BIOS, only direct |
| 3922 | hardware access methods are allowed. Use this |
| 3923 | if you experience crashes upon bootup and you |
| 3924 | suspect they are caused by the BIOS. |
| 3925 | conf1 [X86] Force use of PCI Configuration Access |
| 3926 | Mechanism 1 (config address in IO port 0xCF8, |
| 3927 | data in IO port 0xCFC, both 32-bit). |
| 3928 | conf2 [X86] Force use of PCI Configuration Access |
| 3929 | Mechanism 2 (IO port 0xCF8 is an 8-bit port for |
| 3930 | the function, IO port 0xCFA, also 8-bit, sets |
| 3931 | bus number. The config space is then accessed |
| 3932 | through ports 0xC000-0xCFFF). |
| 3933 | See http://wiki.osdev.org/PCI for more info |
| 3934 | on the configuration access mechanisms. |
| 3935 | noaer [PCIE] If the PCIEAER kernel config parameter is |
| 3936 | enabled, this kernel boot option can be used to |
| 3937 | disable the use of PCIE advanced error reporting. |
| 3938 | nodomains [PCI] Disable support for multiple PCI |
| 3939 | root domains (aka PCI segments, in ACPI-speak). |
| 3940 | nommconf [X86] Disable use of MMCONFIG for PCI |
| 3941 | Configuration |
| 3942 | check_enable_amd_mmconf [X86] check for and enable |
| 3943 | properly configured MMIO access to PCI |
| 3944 | config space on AMD family 10h CPU |
| 3945 | nomsi [MSI] If the PCI_MSI kernel config parameter is |
| 3946 | enabled, this kernel boot option can be used to |
| 3947 | disable the use of MSI interrupts system-wide. |
| 3948 | noioapicquirk [APIC] Disable all boot interrupt quirks. |
| 3949 | Safety option to keep boot IRQs enabled. This |
| 3950 | should never be necessary. |
| 3951 | ioapicreroute [APIC] Enable rerouting of boot IRQs to the |
| 3952 | primary IO-APIC for bridges that cannot disable |
| 3953 | boot IRQs. This fixes a source of spurious IRQs |
| 3954 | when the system masks IRQs. |
| 3955 | noioapicreroute [APIC] Disable workaround that uses the |
| 3956 | boot IRQ equivalent of an IRQ that connects to |
| 3957 | a chipset where boot IRQs cannot be disabled. |
| 3958 | The opposite of ioapicreroute. |
| 3959 | biosirq [X86-32] Use PCI BIOS calls to get the interrupt |
| 3960 | routing table. These calls are known to be buggy |
| 3961 | on several machines and they hang the machine |
| 3962 | when used, but on other computers it's the only |
| 3963 | way to get the interrupt routing table. Try |
| 3964 | this option if the kernel is unable to allocate |
| 3965 | IRQs or discover secondary PCI buses on your |
| 3966 | motherboard. |
| 3967 | rom [X86] Assign address space to expansion ROMs. |
| 3968 | Use with caution as certain devices share |
| 3969 | address decoders between ROMs and other |
| 3970 | resources. |
| 3971 | norom [X86] Do not assign address space to |
| 3972 | expansion ROMs that do not already have |
| 3973 | BIOS assigned address ranges. |
| 3974 | nobar [X86] Do not assign address space to the |
| 3975 | BARs that weren't assigned by the BIOS. |
| 3976 | irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be |
| 3977 | assigned automatically to PCI devices. You can |
| 3978 | make the kernel exclude IRQs of your ISA cards |
| 3979 | this way. |
| 3980 | pirqaddr=0xAAAAA [X86] Specify the physical address |
| 3981 | of the PIRQ table (normally generated |
| 3982 | by the BIOS) if it is outside the |
| 3983 | F0000h-100000h range. |
| 3984 | lastbus=N [X86] Scan all buses thru bus #N. Can be |
| 3985 | useful if the kernel is unable to find your |
| 3986 | secondary buses and you want to tell it |
| 3987 | explicitly which ones they are. |
| 3988 | assign-busses [X86] Always assign all PCI bus |
| 3989 | numbers ourselves, overriding |
| 3990 | whatever the firmware may have done. |
| 3991 | usepirqmask [X86] Honor the possible IRQ mask stored |
| 3992 | in the BIOS $PIR table. This is needed on |
| 3993 | some systems with broken BIOSes, notably |
| 3994 | some HP Pavilion N5400 and Omnibook XE3 |
| 3995 | notebooks. This will have no effect if ACPI |
| 3996 | IRQ routing is enabled. |
| 3997 | noacpi [X86] Do not use ACPI for IRQ routing |
| 3998 | or for PCI scanning. |
| 3999 | use_crs [X86] Use PCI host bridge window information |
| 4000 | from ACPI. On BIOSes from 2008 or later, this |
| 4001 | is enabled by default. If you need to use this, |
| 4002 | please report a bug. |
| 4003 | nocrs [X86] Ignore PCI host bridge windows from ACPI. |
| 4004 | If you need to use this, please report a bug. |
| 4005 | routeirq Do IRQ routing for all PCI devices. |
| 4006 | This is normally done in pci_enable_device(), |
| 4007 | so this option is a temporary workaround |
| 4008 | for broken drivers that don't call it. |
| 4009 | skip_isa_align [X86] do not align io start addr, so can |
| 4010 | handle more pci cards |
| 4011 | noearly [X86] Don't do any early type 1 scanning. |
| 4012 | This might help on some broken boards which |
| 4013 | machine check when some devices' config space |
| 4014 | is read. But various workarounds are disabled |
| 4015 | and some IOMMU drivers will not work. |
| 4016 | bfsort Sort PCI devices into breadth-first order. |
| 4017 | This sorting is done to get a device |
| 4018 | order compatible with older (<= 2.4) kernels. |
| 4019 | nobfsort Don't sort PCI devices into breadth-first order. |
| 4020 | pcie_bus_tune_off Disable PCIe MPS (Max Payload Size) |
| 4021 | tuning and use the BIOS-configured MPS defaults. |
| 4022 | pcie_bus_safe Set every device's MPS to the largest value |
| 4023 | supported by all devices below the root complex. |
| 4024 | pcie_bus_perf Set device MPS to the largest allowable MPS |
| 4025 | based on its parent bus. Also set MRRS (Max |
| 4026 | Read Request Size) to the largest supported |
| 4027 | value (no larger than the MPS that the device |
| 4028 | or bus can support) for best performance. |
| 4029 | pcie_bus_peer2peer Set every device's MPS to 128B, which |
| 4030 | every device is guaranteed to support. This |
| 4031 | configuration allows peer-to-peer DMA between |
| 4032 | any pair of devices, possibly at the cost of |
| 4033 | reduced performance. This also guarantees |
| 4034 | that hot-added devices will work. |
| 4035 | cbiosize=nn[KMG] The fixed amount of bus space which is |
| 4036 | reserved for the CardBus bridge's IO window. |
| 4037 | The default value is 256 bytes. |
| 4038 | cbmemsize=nn[KMG] The fixed amount of bus space which is |
| 4039 | reserved for the CardBus bridge's memory |
| 4040 | window. The default value is 64 megabytes. |
| 4041 | resource_alignment= |
| 4042 | Format: |
| 4043 | [<order of align>@]<pci_dev>[; ...] |
| 4044 | Specifies alignment and device to reassign |
| 4045 | aligned memory resources. How to |
| 4046 | specify the device is described above. |
| 4047 | If <order of align> is not specified, |
| 4048 | PAGE_SIZE is used as alignment. |
| 4049 | A PCI-PCI bridge can be specified if resource |
| 4050 | windows need to be expanded. |
| 4051 | To specify the alignment for several |
| 4052 | instances of a device, the PCI vendor, |
| 4053 | device, subvendor, and subdevice may be |
| 4054 | specified, e.g., 12@pci:8086:9c22:103c:198f |
| 4055 | for 4096-byte alignment. |
| 4056 | ecrc= Enable/disable PCIe ECRC (transaction layer |
| 4057 | end-to-end CRC checking). |
| 4058 | bios: Use BIOS/firmware settings. This is the |
| 4059 | the default. |
| 4060 | off: Turn ECRC off |
| 4061 | on: Turn ECRC on. |
| 4062 | hpiosize=nn[KMG] The fixed amount of bus space which is |
| 4063 | reserved for hotplug bridge's IO window. |
| 4064 | Default size is 256 bytes. |
| 4065 | hpmmiosize=nn[KMG] The fixed amount of bus space which is |
| 4066 | reserved for hotplug bridge's MMIO window. |
| 4067 | Default size is 2 megabytes. |
| 4068 | hpmmioprefsize=nn[KMG] The fixed amount of bus space which is |
| 4069 | reserved for hotplug bridge's MMIO_PREF window. |
| 4070 | Default size is 2 megabytes. |
| 4071 | hpmemsize=nn[KMG] The fixed amount of bus space which is |
| 4072 | reserved for hotplug bridge's MMIO and |
| 4073 | MMIO_PREF window. |
| 4074 | Default size is 2 megabytes. |
| 4075 | hpbussize=nn The minimum amount of additional bus numbers |
| 4076 | reserved for buses below a hotplug bridge. |
| 4077 | Default is 1. |
| 4078 | realloc= Enable/disable reallocating PCI bridge resources |
| 4079 | if allocations done by BIOS are too small to |
| 4080 | accommodate resources required by all child |
| 4081 | devices. |
| 4082 | off: Turn realloc off |
| 4083 | on: Turn realloc on |
| 4084 | realloc same as realloc=on |
| 4085 | noari do not use PCIe ARI. |
| 4086 | noats [PCIE, Intel-IOMMU, AMD-IOMMU] |
| 4087 | do not use PCIe ATS (and IOMMU device IOTLB). |
| 4088 | pcie_scan_all Scan all possible PCIe devices. Otherwise we |
| 4089 | only look for one device below a PCIe downstream |
| 4090 | port. |
| 4091 | big_root_window Try to add a big 64bit memory window to the PCIe |
| 4092 | root complex on AMD CPUs. Some GFX hardware |
| 4093 | can resize a BAR to allow access to all VRAM. |
| 4094 | Adding the window is slightly risky (it may |
| 4095 | conflict with unreported devices), so this |
| 4096 | taints the kernel. |
| 4097 | disable_acs_redir=<pci_dev>[; ...] |
| 4098 | Specify one or more PCI devices (in the format |
| 4099 | specified above) separated by semicolons. |
| 4100 | Each device specified will have the PCI ACS |
| 4101 | redirect capabilities forced off which will |
| 4102 | allow P2P traffic between devices through |
| 4103 | bridges without forcing it upstream. Note: |
| 4104 | this removes isolation between devices and |
| 4105 | may put more devices in an IOMMU group. |
| 4106 | force_floating [S390] Force usage of floating interrupts. |
| 4107 | nomio [S390] Do not use MIO instructions. |
| 4108 | norid [S390] ignore the RID field and force use of |
| 4109 | one PCI domain per PCI function |
| 4110 | |
| 4111 | pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power |
| 4112 | Management. |
| 4113 | off Disable ASPM. |
| 4114 | force Enable ASPM even on devices that claim not to support it. |
| 4115 | WARNING: Forcing ASPM on may cause system lockups. |
| 4116 | |
| 4117 | pcie_ports= [PCIE] PCIe port services handling: |
| 4118 | native Use native PCIe services (PME, AER, DPC, PCIe hotplug) |
| 4119 | even if the platform doesn't give the OS permission to |
| 4120 | use them. This may cause conflicts if the platform |
| 4121 | also tries to use these services. |
| 4122 | dpc-native Use native PCIe service for DPC only. May |
| 4123 | cause conflicts if firmware uses AER or DPC. |
| 4124 | compat Disable native PCIe services (PME, AER, DPC, PCIe |
| 4125 | hotplug). |
| 4126 | |
| 4127 | pcie_port_pm= [PCIE] PCIe port power management handling: |
| 4128 | off Disable power management of all PCIe ports |
| 4129 | force Forcibly enable power management of all PCIe ports |
| 4130 | |
| 4131 | pcie_pme= [PCIE,PM] Native PCIe PME signaling options: |
| 4132 | nomsi Do not use MSI for native PCIe PME signaling (this makes |
| 4133 | all PCIe root ports use INTx for all services). |
| 4134 | |
| 4135 | pcmv= [HW,PCMCIA] BadgePAD 4 |
| 4136 | |
| 4137 | pd_ignore_unused |
| 4138 | [PM] |
| 4139 | Keep all power-domains already enabled by bootloader on, |
| 4140 | even if no driver has claimed them. This is useful |
| 4141 | for debug and development, but should not be |
| 4142 | needed on a platform with proper driver support. |
| 4143 | |
| 4144 | pd. [PARIDE] |
| 4145 | See Documentation/admin-guide/blockdev/paride.rst. |
| 4146 | |
| 4147 | pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at |
| 4148 | boot time. |
| 4149 | Format: { 0 | 1 } |
| 4150 | See arch/parisc/kernel/pdc_chassis.c |
| 4151 | |
| 4152 | percpu_alloc= Select which percpu first chunk allocator to use. |
| 4153 | Currently supported values are "embed" and "page". |
| 4154 | Archs may support subset or none of the selections. |
| 4155 | See comments in mm/percpu.c for details on each |
| 4156 | allocator. This parameter is primarily for debugging |
| 4157 | and performance comparison. |
| 4158 | |
| 4159 | pf. [PARIDE] |
| 4160 | See Documentation/admin-guide/blockdev/paride.rst. |
| 4161 | |
| 4162 | pg. [PARIDE] |
| 4163 | See Documentation/admin-guide/blockdev/paride.rst. |
| 4164 | |
| 4165 | pirq= [SMP,APIC] Manual mp-table setup |
| 4166 | See Documentation/x86/i386/IO-APIC.rst. |
| 4167 | |
| 4168 | plip= [PPT,NET] Parallel port network link |
| 4169 | Format: { parport<nr> | timid | 0 } |
| 4170 | See also Documentation/admin-guide/parport.rst. |
| 4171 | |
| 4172 | pmtmr= [X86] Manual setup of pmtmr I/O Port. |
| 4173 | Override pmtimer IOPort with a hex value. |
| 4174 | e.g. pmtmr=0x508 |
| 4175 | |
| 4176 | pmu_override= [PPC] Override the PMU. |
| 4177 | This option takes over the PMU facility, so it is no |
| 4178 | longer usable by perf. Setting this option starts the |
| 4179 | PMU counters by setting MMCR0 to 0 (the FC bit is |
| 4180 | cleared). If a number is given, then MMCR1 is set to |
| 4181 | that number, otherwise (e.g., 'pmu_override=on'), MMCR1 |
| 4182 | remains 0. |
| 4183 | |
| 4184 | pm_debug_messages [SUSPEND,KNL] |
| 4185 | Enable suspend/resume debug messages during boot up. |
| 4186 | |
| 4187 | pnp.debug=1 [PNP] |
| 4188 | Enable PNP debug messages (depends on the |
| 4189 | CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time |
| 4190 | via /sys/module/pnp/parameters/debug. We always show |
| 4191 | current resource usage; turning this on also shows |
| 4192 | possible settings and some assignment information. |
| 4193 | |
| 4194 | pnpacpi= [ACPI] |
| 4195 | { off } |
| 4196 | |
| 4197 | pnpbios= [ISAPNP] |
| 4198 | { on | off | curr | res | no-curr | no-res } |
| 4199 | |
| 4200 | pnp_reserve_irq= |
| 4201 | [ISAPNP] Exclude IRQs for the autoconfiguration |
| 4202 | |
| 4203 | pnp_reserve_dma= |
| 4204 | [ISAPNP] Exclude DMAs for the autoconfiguration |
| 4205 | |
| 4206 | pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration |
| 4207 | Ranges are in pairs (I/O port base and size). |
| 4208 | |
| 4209 | pnp_reserve_mem= |
| 4210 | [ISAPNP] Exclude memory regions for the |
| 4211 | autoconfiguration. |
| 4212 | Ranges are in pairs (memory base and size). |
| 4213 | |
| 4214 | ports= [IP_VS_FTP] IPVS ftp helper module |
| 4215 | Default is 21. |
| 4216 | Up to 8 (IP_VS_APP_MAX_PORTS) ports |
| 4217 | may be specified. |
| 4218 | Format: <port>,<port>.... |
| 4219 | |
| 4220 | powersave=off [PPC] This option disables power saving features. |
| 4221 | It specifically disables cpuidle and sets the |
| 4222 | platform machine description specific power_save |
| 4223 | function to NULL. On Idle the CPU just reduces |
| 4224 | execution priority. |
| 4225 | |
| 4226 | ppc_strict_facility_enable |
| 4227 | [PPC] This option catches any kernel floating point, |
| 4228 | Altivec, VSX and SPE outside of regions specifically |
| 4229 | allowed (eg kernel_enable_fpu()/kernel_disable_fpu()). |
| 4230 | There is some performance impact when enabling this. |
| 4231 | |
| 4232 | ppc_tm= [PPC] |
| 4233 | Format: {"off"} |
| 4234 | Disable Hardware Transactional Memory |
| 4235 | |
| 4236 | preempt= [KNL] |
| 4237 | Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC |
| 4238 | none - Limited to cond_resched() calls |
| 4239 | voluntary - Limited to cond_resched() and might_sleep() calls |
| 4240 | full - Any section that isn't explicitly preempt disabled |
| 4241 | can be preempted anytime. |
| 4242 | |
| 4243 | print-fatal-signals= |
| 4244 | [KNL] debug: print fatal signals |
| 4245 | |
| 4246 | If enabled, warn about various signal handling |
| 4247 | related application anomalies: too many signals, |
| 4248 | too many POSIX.1 timers, fatal signals causing a |
| 4249 | coredump - etc. |
| 4250 | |
| 4251 | If you hit the warning due to signal overflow, |
| 4252 | you might want to try "ulimit -i unlimited". |
| 4253 | |
| 4254 | default: off. |
| 4255 | |
| 4256 | printk.always_kmsg_dump= |
| 4257 | Trigger kmsg_dump for cases other than kernel oops or |
| 4258 | panics |
| 4259 | Format: <bool> (1/Y/y=enable, 0/N/n=disable) |
| 4260 | default: disabled |
| 4261 | |
| 4262 | printk.console_no_auto_verbose= |
| 4263 | Disable console loglevel raise on oops, panic |
| 4264 | or lockdep-detected issues (only if lock debug is on). |
| 4265 | With an exception to setups with low baudrate on |
| 4266 | serial console, keeping this 0 is a good choice |
| 4267 | in order to provide more debug information. |
| 4268 | Format: <bool> |
| 4269 | default: 0 (auto_verbose is enabled) |
| 4270 | |
| 4271 | printk.devkmsg={on,off,ratelimit} |
| 4272 | Control writing to /dev/kmsg. |
| 4273 | on - unlimited logging to /dev/kmsg from userspace |
| 4274 | off - logging to /dev/kmsg disabled |
| 4275 | ratelimit - ratelimit the logging |
| 4276 | Default: ratelimit |
| 4277 | |
| 4278 | printk.time= Show timing data prefixed to each printk message line |
| 4279 | Format: <bool> (1/Y/y=enable, 0/N/n=disable) |
| 4280 | |
| 4281 | processor.max_cstate= [HW,ACPI] |
| 4282 | Limit processor to maximum C-state |
| 4283 | max_cstate=9 overrides any DMI blacklist limit. |
| 4284 | |
| 4285 | processor.nocst [HW,ACPI] |
| 4286 | Ignore the _CST method to determine C-states, |
| 4287 | instead using the legacy FADT method |
| 4288 | |
| 4289 | profile= [KNL] Enable kernel profiling via /proc/profile |
| 4290 | Format: [<profiletype>,]<number> |
| 4291 | Param: <profiletype>: "schedule", "sleep", or "kvm" |
| 4292 | [defaults to kernel profiling] |
| 4293 | Param: "schedule" - profile schedule points. |
| 4294 | Param: "sleep" - profile D-state sleeping (millisecs). |
| 4295 | Requires CONFIG_SCHEDSTATS |
| 4296 | Param: "kvm" - profile VM exits. |
| 4297 | Param: <number> - step/bucket size as a power of 2 for |
| 4298 | statistical time based profiling. |
| 4299 | |
| 4300 | prompt_ramdisk= [RAM] [Deprecated] |
| 4301 | |
| 4302 | prot_virt= [S390] enable hosting protected virtual machines |
| 4303 | isolated from the hypervisor (if hardware supports |
| 4304 | that). |
| 4305 | Format: <bool> |
| 4306 | |
| 4307 | psi= [KNL] Enable or disable pressure stall information |
| 4308 | tracking. |
| 4309 | Format: <bool> |
| 4310 | |
| 4311 | psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to |
| 4312 | probe for; one of (bare|imps|exps|lifebook|any). |
| 4313 | psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports |
| 4314 | per second. |
| 4315 | psmouse.resetafter= [HW,MOUSE] |
| 4316 | Try to reset the device after so many bad packets |
| 4317 | (0 = never). |
| 4318 | psmouse.resolution= |
| 4319 | [HW,MOUSE] Set desired mouse resolution, in dpi. |
| 4320 | psmouse.smartscroll= |
| 4321 | [HW,MOUSE] Controls Logitech smartscroll autorepeat. |
| 4322 | 0 = disabled, 1 = enabled (default). |
| 4323 | |
| 4324 | pstore.backend= Specify the name of the pstore backend to use |
| 4325 | |
| 4326 | pt. [PARIDE] |
| 4327 | See Documentation/admin-guide/blockdev/paride.rst. |
| 4328 | |
| 4329 | pti= [X86-64] Control Page Table Isolation of user and |
| 4330 | kernel address spaces. Disabling this feature |
| 4331 | removes hardening, but improves performance of |
| 4332 | system calls and interrupts. |
| 4333 | |
| 4334 | on - unconditionally enable |
| 4335 | off - unconditionally disable |
| 4336 | auto - kernel detects whether your CPU model is |
| 4337 | vulnerable to issues that PTI mitigates |
| 4338 | |
| 4339 | Not specifying this option is equivalent to pti=auto. |
| 4340 | |
| 4341 | nopti [X86-64] |
| 4342 | Equivalent to pti=off |
| 4343 | |
| 4344 | pty.legacy_count= |
| 4345 | [KNL] Number of legacy pty's. Overwrites compiled-in |
| 4346 | default number. |
| 4347 | |
| 4348 | quiet [KNL] Disable most log messages |
| 4349 | |
| 4350 | r128= [HW,DRM] |
| 4351 | |
| 4352 | raid= [HW,RAID] |
| 4353 | See Documentation/admin-guide/md.rst. |
| 4354 | |
| 4355 | ramdisk_size= [RAM] Sizes of RAM disks in kilobytes |
| 4356 | See Documentation/admin-guide/blockdev/ramdisk.rst. |
| 4357 | |
| 4358 | ramdisk_start= [RAM] RAM disk image start address |
| 4359 | |
| 4360 | random.trust_cpu={on,off} |
| 4361 | [KNL] Enable or disable trusting the use of the |
| 4362 | CPU's random number generator (if available) to |
| 4363 | fully seed the kernel's CRNG. Default is controlled |
| 4364 | by CONFIG_RANDOM_TRUST_CPU. |
| 4365 | |
| 4366 | randomize_kstack_offset= |
| 4367 | [KNL] Enable or disable kernel stack offset |
| 4368 | randomization, which provides roughly 5 bits of |
| 4369 | entropy, frustrating memory corruption attacks |
| 4370 | that depend on stack address determinism or |
| 4371 | cross-syscall address exposures. This is only |
| 4372 | available on architectures that have defined |
| 4373 | CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET. |
| 4374 | Format: <bool> (1/Y/y=enable, 0/N/n=disable) |
| 4375 | Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT. |
| 4376 | |
| 4377 | ras=option[,option,...] [KNL] RAS-specific options |
| 4378 | |
| 4379 | cec_disable [X86] |
| 4380 | Disable the Correctable Errors Collector, |
| 4381 | see CONFIG_RAS_CEC help text. |
| 4382 | |
| 4383 | rcu_nocbs[=cpu-list] |
| 4384 | [KNL] The optional argument is a cpu list, |
| 4385 | as described above. |
| 4386 | |
| 4387 | In kernels built with CONFIG_RCU_NOCB_CPU=y, |
| 4388 | enable the no-callback CPU mode, which prevents |
| 4389 | such CPUs' callbacks from being invoked in |
| 4390 | softirq context. Invocation of such CPUs' RCU |
| 4391 | callbacks will instead be offloaded to "rcuox/N" |
| 4392 | kthreads created for that purpose, where "x" is |
| 4393 | "p" for RCU-preempt, "s" for RCU-sched, and "g" |
| 4394 | for the kthreads that mediate grace periods; and |
| 4395 | "N" is the CPU number. This reduces OS jitter on |
| 4396 | the offloaded CPUs, which can be useful for HPC |
| 4397 | and real-time workloads. It can also improve |
| 4398 | energy efficiency for asymmetric multiprocessors. |
| 4399 | |
| 4400 | If a cpulist is passed as an argument, the specified |
| 4401 | list of CPUs is set to no-callback mode from boot. |
| 4402 | |
| 4403 | Otherwise, if the '=' sign and the cpulist |
| 4404 | arguments are omitted, no CPU will be set to |
| 4405 | no-callback mode from boot but the mode may be |
| 4406 | toggled at runtime via cpusets. |
| 4407 | |
| 4408 | rcu_nocb_poll [KNL] |
| 4409 | Rather than requiring that offloaded CPUs |
| 4410 | (specified by rcu_nocbs= above) explicitly |
| 4411 | awaken the corresponding "rcuoN" kthreads, |
| 4412 | make these kthreads poll for callbacks. |
| 4413 | This improves the real-time response for the |
| 4414 | offloaded CPUs by relieving them of the need to |
| 4415 | wake up the corresponding kthread, but degrades |
| 4416 | energy efficiency by requiring that the kthreads |
| 4417 | periodically wake up to do the polling. |
| 4418 | |
| 4419 | rcutree.blimit= [KNL] |
| 4420 | Set maximum number of finished RCU callbacks to |
| 4421 | process in one batch. |
| 4422 | |
| 4423 | rcutree.dump_tree= [KNL] |
| 4424 | Dump the structure of the rcu_node combining tree |
| 4425 | out at early boot. This is used for diagnostic |
| 4426 | purposes, to verify correct tree setup. |
| 4427 | |
| 4428 | rcutree.gp_cleanup_delay= [KNL] |
| 4429 | Set the number of jiffies to delay each step of |
| 4430 | RCU grace-period cleanup. |
| 4431 | |
| 4432 | rcutree.gp_init_delay= [KNL] |
| 4433 | Set the number of jiffies to delay each step of |
| 4434 | RCU grace-period initialization. |
| 4435 | |
| 4436 | rcutree.gp_preinit_delay= [KNL] |
| 4437 | Set the number of jiffies to delay each step of |
| 4438 | RCU grace-period pre-initialization, that is, |
| 4439 | the propagation of recent CPU-hotplug changes up |
| 4440 | the rcu_node combining tree. |
| 4441 | |
| 4442 | rcutree.use_softirq= [KNL] |
| 4443 | If set to zero, move all RCU_SOFTIRQ processing to |
| 4444 | per-CPU rcuc kthreads. Defaults to a non-zero |
| 4445 | value, meaning that RCU_SOFTIRQ is used by default. |
| 4446 | Specify rcutree.use_softirq=0 to use rcuc kthreads. |
| 4447 | |
| 4448 | But note that CONFIG_PREEMPT_RT=y kernels disable |
| 4449 | this kernel boot parameter, forcibly setting it |
| 4450 | to zero. |
| 4451 | |
| 4452 | rcutree.rcu_fanout_exact= [KNL] |
| 4453 | Disable autobalancing of the rcu_node combining |
| 4454 | tree. This is used by rcutorture, and might |
| 4455 | possibly be useful for architectures having high |
| 4456 | cache-to-cache transfer latencies. |
| 4457 | |
| 4458 | rcutree.rcu_fanout_leaf= [KNL] |
| 4459 | Change the number of CPUs assigned to each |
| 4460 | leaf rcu_node structure. Useful for very |
| 4461 | large systems, which will choose the value 64, |
| 4462 | and for NUMA systems with large remote-access |
| 4463 | latencies, which will choose a value aligned |
| 4464 | with the appropriate hardware boundaries. |
| 4465 | |
| 4466 | rcutree.rcu_min_cached_objs= [KNL] |
| 4467 | Minimum number of objects which are cached and |
| 4468 | maintained per one CPU. Object size is equal |
| 4469 | to PAGE_SIZE. The cache allows to reduce the |
| 4470 | pressure to page allocator, also it makes the |
| 4471 | whole algorithm to behave better in low memory |
| 4472 | condition. |
| 4473 | |
| 4474 | rcutree.rcu_delay_page_cache_fill_msec= [KNL] |
| 4475 | Set the page-cache refill delay (in milliseconds) |
| 4476 | in response to low-memory conditions. The range |
| 4477 | of permitted values is in the range 0:100000. |
| 4478 | |
| 4479 | rcutree.jiffies_till_first_fqs= [KNL] |
| 4480 | Set delay from grace-period initialization to |
| 4481 | first attempt to force quiescent states. |
| 4482 | Units are jiffies, minimum value is zero, |
| 4483 | and maximum value is HZ. |
| 4484 | |
| 4485 | rcutree.jiffies_till_next_fqs= [KNL] |
| 4486 | Set delay between subsequent attempts to force |
| 4487 | quiescent states. Units are jiffies, minimum |
| 4488 | value is one, and maximum value is HZ. |
| 4489 | |
| 4490 | rcutree.jiffies_till_sched_qs= [KNL] |
| 4491 | Set required age in jiffies for a |
| 4492 | given grace period before RCU starts |
| 4493 | soliciting quiescent-state help from |
| 4494 | rcu_note_context_switch() and cond_resched(). |
| 4495 | If not specified, the kernel will calculate |
| 4496 | a value based on the most recent settings |
| 4497 | of rcutree.jiffies_till_first_fqs |
| 4498 | and rcutree.jiffies_till_next_fqs. |
| 4499 | This calculated value may be viewed in |
| 4500 | rcutree.jiffies_to_sched_qs. Any attempt to set |
| 4501 | rcutree.jiffies_to_sched_qs will be cheerfully |
| 4502 | overwritten. |
| 4503 | |
| 4504 | rcutree.kthread_prio= [KNL,BOOT] |
| 4505 | Set the SCHED_FIFO priority of the RCU per-CPU |
| 4506 | kthreads (rcuc/N). This value is also used for |
| 4507 | the priority of the RCU boost threads (rcub/N) |
| 4508 | and for the RCU grace-period kthreads (rcu_bh, |
| 4509 | rcu_preempt, and rcu_sched). If RCU_BOOST is |
| 4510 | set, valid values are 1-99 and the default is 1 |
| 4511 | (the least-favored priority). Otherwise, when |
| 4512 | RCU_BOOST is not set, valid values are 0-99 and |
| 4513 | the default is zero (non-realtime operation). |
| 4514 | |
| 4515 | rcutree.rcu_nocb_gp_stride= [KNL] |
| 4516 | Set the number of NOCB callback kthreads in |
| 4517 | each group, which defaults to the square root |
| 4518 | of the number of CPUs. Larger numbers reduce |
| 4519 | the wakeup overhead on the global grace-period |
| 4520 | kthread, but increases that same overhead on |
| 4521 | each group's NOCB grace-period kthread. |
| 4522 | |
| 4523 | rcutree.qhimark= [KNL] |
| 4524 | Set threshold of queued RCU callbacks beyond which |
| 4525 | batch limiting is disabled. |
| 4526 | |
| 4527 | rcutree.qlowmark= [KNL] |
| 4528 | Set threshold of queued RCU callbacks below which |
| 4529 | batch limiting is re-enabled. |
| 4530 | |
| 4531 | rcutree.qovld= [KNL] |
| 4532 | Set threshold of queued RCU callbacks beyond which |
| 4533 | RCU's force-quiescent-state scan will aggressively |
| 4534 | enlist help from cond_resched() and sched IPIs to |
| 4535 | help CPUs more quickly reach quiescent states. |
| 4536 | Set to less than zero to make this be set based |
| 4537 | on rcutree.qhimark at boot time and to zero to |
| 4538 | disable more aggressive help enlistment. |
| 4539 | |
| 4540 | rcutree.rcu_kick_kthreads= [KNL] |
| 4541 | Cause the grace-period kthread to get an extra |
| 4542 | wake_up() if it sleeps three times longer than |
| 4543 | it should at force-quiescent-state time. |
| 4544 | This wake_up() will be accompanied by a |
| 4545 | WARN_ONCE() splat and an ftrace_dump(). |
| 4546 | |
| 4547 | rcutree.rcu_unlock_delay= [KNL] |
| 4548 | In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels, |
| 4549 | this specifies an rcu_read_unlock()-time delay |
| 4550 | in microseconds. This defaults to zero. |
| 4551 | Larger delays increase the probability of |
| 4552 | catching RCU pointer leaks, that is, buggy use |
| 4553 | of RCU-protected pointers after the relevant |
| 4554 | rcu_read_unlock() has completed. |
| 4555 | |
| 4556 | rcutree.sysrq_rcu= [KNL] |
| 4557 | Commandeer a sysrq key to dump out Tree RCU's |
| 4558 | rcu_node tree with an eye towards determining |
| 4559 | why a new grace period has not yet started. |
| 4560 | |
| 4561 | rcuscale.gp_async= [KNL] |
| 4562 | Measure performance of asynchronous |
| 4563 | grace-period primitives such as call_rcu(). |
| 4564 | |
| 4565 | rcuscale.gp_async_max= [KNL] |
| 4566 | Specify the maximum number of outstanding |
| 4567 | callbacks per writer thread. When a writer |
| 4568 | thread exceeds this limit, it invokes the |
| 4569 | corresponding flavor of rcu_barrier() to allow |
| 4570 | previously posted callbacks to drain. |
| 4571 | |
| 4572 | rcuscale.gp_exp= [KNL] |
| 4573 | Measure performance of expedited synchronous |
| 4574 | grace-period primitives. |
| 4575 | |
| 4576 | rcuscale.holdoff= [KNL] |
| 4577 | Set test-start holdoff period. The purpose of |
| 4578 | this parameter is to delay the start of the |
| 4579 | test until boot completes in order to avoid |
| 4580 | interference. |
| 4581 | |
| 4582 | rcuscale.kfree_rcu_test= [KNL] |
| 4583 | Set to measure performance of kfree_rcu() flooding. |
| 4584 | |
| 4585 | rcuscale.kfree_rcu_test_double= [KNL] |
| 4586 | Test the double-argument variant of kfree_rcu(). |
| 4587 | If this parameter has the same value as |
| 4588 | rcuscale.kfree_rcu_test_single, both the single- |
| 4589 | and double-argument variants are tested. |
| 4590 | |
| 4591 | rcuscale.kfree_rcu_test_single= [KNL] |
| 4592 | Test the single-argument variant of kfree_rcu(). |
| 4593 | If this parameter has the same value as |
| 4594 | rcuscale.kfree_rcu_test_double, both the single- |
| 4595 | and double-argument variants are tested. |
| 4596 | |
| 4597 | rcuscale.kfree_nthreads= [KNL] |
| 4598 | The number of threads running loops of kfree_rcu(). |
| 4599 | |
| 4600 | rcuscale.kfree_alloc_num= [KNL] |
| 4601 | Number of allocations and frees done in an iteration. |
| 4602 | |
| 4603 | rcuscale.kfree_loops= [KNL] |
| 4604 | Number of loops doing rcuscale.kfree_alloc_num number |
| 4605 | of allocations and frees. |
| 4606 | |
| 4607 | rcuscale.nreaders= [KNL] |
| 4608 | Set number of RCU readers. The value -1 selects |
| 4609 | N, where N is the number of CPUs. A value |
| 4610 | "n" less than -1 selects N-n+1, where N is again |
| 4611 | the number of CPUs. For example, -2 selects N |
| 4612 | (the number of CPUs), -3 selects N+1, and so on. |
| 4613 | A value of "n" less than or equal to -N selects |
| 4614 | a single reader. |
| 4615 | |
| 4616 | rcuscale.nwriters= [KNL] |
| 4617 | Set number of RCU writers. The values operate |
| 4618 | the same as for rcuscale.nreaders. |
| 4619 | N, where N is the number of CPUs |
| 4620 | |
| 4621 | rcuscale.perf_type= [KNL] |
| 4622 | Specify the RCU implementation to test. |
| 4623 | |
| 4624 | rcuscale.shutdown= [KNL] |
| 4625 | Shut the system down after performance tests |
| 4626 | complete. This is useful for hands-off automated |
| 4627 | testing. |
| 4628 | |
| 4629 | rcuscale.verbose= [KNL] |
| 4630 | Enable additional printk() statements. |
| 4631 | |
| 4632 | rcuscale.writer_holdoff= [KNL] |
| 4633 | Write-side holdoff between grace periods, |
| 4634 | in microseconds. The default of zero says |
| 4635 | no holdoff. |
| 4636 | |
| 4637 | rcutorture.fqs_duration= [KNL] |
| 4638 | Set duration of force_quiescent_state bursts |
| 4639 | in microseconds. |
| 4640 | |
| 4641 | rcutorture.fqs_holdoff= [KNL] |
| 4642 | Set holdoff time within force_quiescent_state bursts |
| 4643 | in microseconds. |
| 4644 | |
| 4645 | rcutorture.fqs_stutter= [KNL] |
| 4646 | Set wait time between force_quiescent_state bursts |
| 4647 | in seconds. |
| 4648 | |
| 4649 | rcutorture.fwd_progress= [KNL] |
| 4650 | Specifies the number of kthreads to be used |
| 4651 | for RCU grace-period forward-progress testing |
| 4652 | for the types of RCU supporting this notion. |
| 4653 | Defaults to 1 kthread, values less than zero or |
| 4654 | greater than the number of CPUs cause the number |
| 4655 | of CPUs to be used. |
| 4656 | |
| 4657 | rcutorture.fwd_progress_div= [KNL] |
| 4658 | Specify the fraction of a CPU-stall-warning |
| 4659 | period to do tight-loop forward-progress testing. |
| 4660 | |
| 4661 | rcutorture.fwd_progress_holdoff= [KNL] |
| 4662 | Number of seconds to wait between successive |
| 4663 | forward-progress tests. |
| 4664 | |
| 4665 | rcutorture.fwd_progress_need_resched= [KNL] |
| 4666 | Enclose cond_resched() calls within checks for |
| 4667 | need_resched() during tight-loop forward-progress |
| 4668 | testing. |
| 4669 | |
| 4670 | rcutorture.gp_cond= [KNL] |
| 4671 | Use conditional/asynchronous update-side |
| 4672 | primitives, if available. |
| 4673 | |
| 4674 | rcutorture.gp_exp= [KNL] |
| 4675 | Use expedited update-side primitives, if available. |
| 4676 | |
| 4677 | rcutorture.gp_normal= [KNL] |
| 4678 | Use normal (non-expedited) asynchronous |
| 4679 | update-side primitives, if available. |
| 4680 | |
| 4681 | rcutorture.gp_sync= [KNL] |
| 4682 | Use normal (non-expedited) synchronous |
| 4683 | update-side primitives, if available. If all |
| 4684 | of rcutorture.gp_cond=, rcutorture.gp_exp=, |
| 4685 | rcutorture.gp_normal=, and rcutorture.gp_sync= |
| 4686 | are zero, rcutorture acts as if is interpreted |
| 4687 | they are all non-zero. |
| 4688 | |
| 4689 | rcutorture.irqreader= [KNL] |
| 4690 | Run RCU readers from irq handlers, or, more |
| 4691 | accurately, from a timer handler. Not all RCU |
| 4692 | flavors take kindly to this sort of thing. |
| 4693 | |
| 4694 | rcutorture.leakpointer= [KNL] |
| 4695 | Leak an RCU-protected pointer out of the reader. |
| 4696 | This can of course result in splats, and is |
| 4697 | intended to test the ability of things like |
| 4698 | CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect |
| 4699 | such leaks. |
| 4700 | |
| 4701 | rcutorture.n_barrier_cbs= [KNL] |
| 4702 | Set callbacks/threads for rcu_barrier() testing. |
| 4703 | |
| 4704 | rcutorture.nfakewriters= [KNL] |
| 4705 | Set number of concurrent RCU writers. These just |
| 4706 | stress RCU, they don't participate in the actual |
| 4707 | test, hence the "fake". |
| 4708 | |
| 4709 | rcutorture.nocbs_nthreads= [KNL] |
| 4710 | Set number of RCU callback-offload togglers. |
| 4711 | Zero (the default) disables toggling. |
| 4712 | |
| 4713 | rcutorture.nocbs_toggle= [KNL] |
| 4714 | Set the delay in milliseconds between successive |
| 4715 | callback-offload toggling attempts. |
| 4716 | |
| 4717 | rcutorture.nreaders= [KNL] |
| 4718 | Set number of RCU readers. The value -1 selects |
| 4719 | N-1, where N is the number of CPUs. A value |
| 4720 | "n" less than -1 selects N-n-2, where N is again |
| 4721 | the number of CPUs. For example, -2 selects N |
| 4722 | (the number of CPUs), -3 selects N+1, and so on. |
| 4723 | |
| 4724 | rcutorture.object_debug= [KNL] |
| 4725 | Enable debug-object double-call_rcu() testing. |
| 4726 | |
| 4727 | rcutorture.onoff_holdoff= [KNL] |
| 4728 | Set time (s) after boot for CPU-hotplug testing. |
| 4729 | |
| 4730 | rcutorture.onoff_interval= [KNL] |
| 4731 | Set time (jiffies) between CPU-hotplug operations, |
| 4732 | or zero to disable CPU-hotplug testing. |
| 4733 | |
| 4734 | rcutorture.read_exit= [KNL] |
| 4735 | Set the number of read-then-exit kthreads used |
| 4736 | to test the interaction of RCU updaters and |
| 4737 | task-exit processing. |
| 4738 | |
| 4739 | rcutorture.read_exit_burst= [KNL] |
| 4740 | The number of times in a given read-then-exit |
| 4741 | episode that a set of read-then-exit kthreads |
| 4742 | is spawned. |
| 4743 | |
| 4744 | rcutorture.read_exit_delay= [KNL] |
| 4745 | The delay, in seconds, between successive |
| 4746 | read-then-exit testing episodes. |
| 4747 | |
| 4748 | rcutorture.shuffle_interval= [KNL] |
| 4749 | Set task-shuffle interval (s). Shuffling tasks |
| 4750 | allows some CPUs to go into dyntick-idle mode |
| 4751 | during the rcutorture test. |
| 4752 | |
| 4753 | rcutorture.shutdown_secs= [KNL] |
| 4754 | Set time (s) after boot system shutdown. This |
| 4755 | is useful for hands-off automated testing. |
| 4756 | |
| 4757 | rcutorture.stall_cpu= [KNL] |
| 4758 | Duration of CPU stall (s) to test RCU CPU stall |
| 4759 | warnings, zero to disable. |
| 4760 | |
| 4761 | rcutorture.stall_cpu_block= [KNL] |
| 4762 | Sleep while stalling if set. This will result |
| 4763 | in warnings from preemptible RCU in addition |
| 4764 | to any other stall-related activity. |
| 4765 | |
| 4766 | rcutorture.stall_cpu_holdoff= [KNL] |
| 4767 | Time to wait (s) after boot before inducing stall. |
| 4768 | |
| 4769 | rcutorture.stall_cpu_irqsoff= [KNL] |
| 4770 | Disable interrupts while stalling if set. |
| 4771 | |
| 4772 | rcutorture.stall_gp_kthread= [KNL] |
| 4773 | Duration (s) of forced sleep within RCU |
| 4774 | grace-period kthread to test RCU CPU stall |
| 4775 | warnings, zero to disable. If both stall_cpu |
| 4776 | and stall_gp_kthread are specified, the |
| 4777 | kthread is starved first, then the CPU. |
| 4778 | |
| 4779 | rcutorture.stat_interval= [KNL] |
| 4780 | Time (s) between statistics printk()s. |
| 4781 | |
| 4782 | rcutorture.stutter= [KNL] |
| 4783 | Time (s) to stutter testing, for example, specifying |
| 4784 | five seconds causes the test to run for five seconds, |
| 4785 | wait for five seconds, and so on. This tests RCU's |
| 4786 | ability to transition abruptly to and from idle. |
| 4787 | |
| 4788 | rcutorture.test_boost= [KNL] |
| 4789 | Test RCU priority boosting? 0=no, 1=maybe, 2=yes. |
| 4790 | "Maybe" means test if the RCU implementation |
| 4791 | under test support RCU priority boosting. |
| 4792 | |
| 4793 | rcutorture.test_boost_duration= [KNL] |
| 4794 | Duration (s) of each individual boost test. |
| 4795 | |
| 4796 | rcutorture.test_boost_interval= [KNL] |
| 4797 | Interval (s) between each boost test. |
| 4798 | |
| 4799 | rcutorture.test_no_idle_hz= [KNL] |
| 4800 | Test RCU's dyntick-idle handling. See also the |
| 4801 | rcutorture.shuffle_interval parameter. |
| 4802 | |
| 4803 | rcutorture.torture_type= [KNL] |
| 4804 | Specify the RCU implementation to test. |
| 4805 | |
| 4806 | rcutorture.verbose= [KNL] |
| 4807 | Enable additional printk() statements. |
| 4808 | |
| 4809 | rcupdate.rcu_cpu_stall_ftrace_dump= [KNL] |
| 4810 | Dump ftrace buffer after reporting RCU CPU |
| 4811 | stall warning. |
| 4812 | |
| 4813 | rcupdate.rcu_cpu_stall_suppress= [KNL] |
| 4814 | Suppress RCU CPU stall warning messages. |
| 4815 | |
| 4816 | rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL] |
| 4817 | Suppress RCU CPU stall warning messages and |
| 4818 | rcutorture writer stall warnings that occur |
| 4819 | during early boot, that is, during the time |
| 4820 | before the init task is spawned. |
| 4821 | |
| 4822 | rcupdate.rcu_cpu_stall_timeout= [KNL] |
| 4823 | Set timeout for RCU CPU stall warning messages. |
| 4824 | |
| 4825 | rcupdate.rcu_expedited= [KNL] |
| 4826 | Use expedited grace-period primitives, for |
| 4827 | example, synchronize_rcu_expedited() instead |
| 4828 | of synchronize_rcu(). This reduces latency, |
| 4829 | but can increase CPU utilization, degrade |
| 4830 | real-time latency, and degrade energy efficiency. |
| 4831 | No effect on CONFIG_TINY_RCU kernels. |
| 4832 | |
| 4833 | rcupdate.rcu_normal= [KNL] |
| 4834 | Use only normal grace-period primitives, |
| 4835 | for example, synchronize_rcu() instead of |
| 4836 | synchronize_rcu_expedited(). This improves |
| 4837 | real-time latency, CPU utilization, and |
| 4838 | energy efficiency, but can expose users to |
| 4839 | increased grace-period latency. This parameter |
| 4840 | overrides rcupdate.rcu_expedited. No effect on |
| 4841 | CONFIG_TINY_RCU kernels. |
| 4842 | |
| 4843 | rcupdate.rcu_normal_after_boot= [KNL] |
| 4844 | Once boot has completed (that is, after |
| 4845 | rcu_end_inkernel_boot() has been invoked), use |
| 4846 | only normal grace-period primitives. No effect |
| 4847 | on CONFIG_TINY_RCU kernels. |
| 4848 | |
| 4849 | But note that CONFIG_PREEMPT_RT=y kernels enables |
| 4850 | this kernel boot parameter, forcibly setting |
| 4851 | it to the value one, that is, converting any |
| 4852 | post-boot attempt at an expedited RCU grace |
| 4853 | period to instead use normal non-expedited |
| 4854 | grace-period processing. |
| 4855 | |
| 4856 | rcupdate.rcu_task_collapse_lim= [KNL] |
| 4857 | Set the maximum number of callbacks present |
| 4858 | at the beginning of a grace period that allows |
| 4859 | the RCU Tasks flavors to collapse back to using |
| 4860 | a single callback queue. This switching only |
| 4861 | occurs when rcupdate.rcu_task_enqueue_lim is |
| 4862 | set to the default value of -1. |
| 4863 | |
| 4864 | rcupdate.rcu_task_contend_lim= [KNL] |
| 4865 | Set the minimum number of callback-queuing-time |
| 4866 | lock-contention events per jiffy required to |
| 4867 | cause the RCU Tasks flavors to switch to per-CPU |
| 4868 | callback queuing. This switching only occurs |
| 4869 | when rcupdate.rcu_task_enqueue_lim is set to |
| 4870 | the default value of -1. |
| 4871 | |
| 4872 | rcupdate.rcu_task_enqueue_lim= [KNL] |
| 4873 | Set the number of callback queues to use for the |
| 4874 | RCU Tasks family of RCU flavors. The default |
| 4875 | of -1 allows this to be automatically (and |
| 4876 | dynamically) adjusted. This parameter is intended |
| 4877 | for use in testing. |
| 4878 | |
| 4879 | rcupdate.rcu_task_ipi_delay= [KNL] |
| 4880 | Set time in jiffies during which RCU tasks will |
| 4881 | avoid sending IPIs, starting with the beginning |
| 4882 | of a given grace period. Setting a large |
| 4883 | number avoids disturbing real-time workloads, |
| 4884 | but lengthens grace periods. |
| 4885 | |
| 4886 | rcupdate.rcu_task_stall_timeout= [KNL] |
| 4887 | Set timeout in jiffies for RCU task stall warning |
| 4888 | messages. Disable with a value less than or equal |
| 4889 | to zero. |
| 4890 | |
| 4891 | rcupdate.rcu_self_test= [KNL] |
| 4892 | Run the RCU early boot self tests |
| 4893 | |
| 4894 | rdinit= [KNL] |
| 4895 | Format: <full_path> |
| 4896 | Run specified binary instead of /init from the ramdisk, |
| 4897 | used for early userspace startup. See initrd. |
| 4898 | |
| 4899 | rdrand= [X86] |
| 4900 | force - Override the decision by the kernel to hide the |
| 4901 | advertisement of RDRAND support (this affects |
| 4902 | certain AMD processors because of buggy BIOS |
| 4903 | support, specifically around the suspend/resume |
| 4904 | path). |
| 4905 | |
| 4906 | rdt= [HW,X86,RDT] |
| 4907 | Turn on/off individual RDT features. List is: |
| 4908 | cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp, |
| 4909 | mba. |
| 4910 | E.g. to turn on cmt and turn off mba use: |
| 4911 | rdt=cmt,!mba |
| 4912 | |
| 4913 | reboot= [KNL] |
| 4914 | Format (x86 or x86_64): |
| 4915 | [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \ |
| 4916 | [[,]s[mp]#### \ |
| 4917 | [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \ |
| 4918 | [[,]f[orce] |
| 4919 | Where reboot_mode is one of warm (soft) or cold (hard) or gpio |
| 4920 | (prefix with 'panic_' to set mode for panic |
| 4921 | reboot only), |
| 4922 | reboot_type is one of bios, acpi, kbd, triple, efi, or pci, |
| 4923 | reboot_force is either force or not specified, |
| 4924 | reboot_cpu is s[mp]#### with #### being the processor |
| 4925 | to be used for rebooting. |
| 4926 | |
| 4927 | refscale.holdoff= [KNL] |
| 4928 | Set test-start holdoff period. The purpose of |
| 4929 | this parameter is to delay the start of the |
| 4930 | test until boot completes in order to avoid |
| 4931 | interference. |
| 4932 | |
| 4933 | refscale.loops= [KNL] |
| 4934 | Set the number of loops over the synchronization |
| 4935 | primitive under test. Increasing this number |
| 4936 | reduces noise due to loop start/end overhead, |
| 4937 | but the default has already reduced the per-pass |
| 4938 | noise to a handful of picoseconds on ca. 2020 |
| 4939 | x86 laptops. |
| 4940 | |
| 4941 | refscale.nreaders= [KNL] |
| 4942 | Set number of readers. The default value of -1 |
| 4943 | selects N, where N is roughly 75% of the number |
| 4944 | of CPUs. A value of zero is an interesting choice. |
| 4945 | |
| 4946 | refscale.nruns= [KNL] |
| 4947 | Set number of runs, each of which is dumped onto |
| 4948 | the console log. |
| 4949 | |
| 4950 | refscale.readdelay= [KNL] |
| 4951 | Set the read-side critical-section duration, |
| 4952 | measured in microseconds. |
| 4953 | |
| 4954 | refscale.scale_type= [KNL] |
| 4955 | Specify the read-protection implementation to test. |
| 4956 | |
| 4957 | refscale.shutdown= [KNL] |
| 4958 | Shut down the system at the end of the performance |
| 4959 | test. This defaults to 1 (shut it down) when |
| 4960 | refscale is built into the kernel and to 0 (leave |
| 4961 | it running) when refscale is built as a module. |
| 4962 | |
| 4963 | refscale.verbose= [KNL] |
| 4964 | Enable additional printk() statements. |
| 4965 | |
| 4966 | refscale.verbose_batched= [KNL] |
| 4967 | Batch the additional printk() statements. If zero |
| 4968 | (the default) or negative, print everything. Otherwise, |
| 4969 | print every Nth verbose statement, where N is the value |
| 4970 | specified. |
| 4971 | |
| 4972 | relax_domain_level= |
| 4973 | [KNL, SMP] Set scheduler's default relax_domain_level. |
| 4974 | See Documentation/admin-guide/cgroup-v1/cpusets.rst. |
| 4975 | |
| 4976 | reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory |
| 4977 | Format: <base1>,<size1>[,<base2>,<size2>,...] |
| 4978 | Reserve I/O ports or memory so the kernel won't use |
| 4979 | them. If <base> is less than 0x10000, the region |
| 4980 | is assumed to be I/O ports; otherwise it is memory. |
| 4981 | |
| 4982 | reservetop= [X86-32] |
| 4983 | Format: nn[KMG] |
| 4984 | Reserves a hole at the top of the kernel virtual |
| 4985 | address space. |
| 4986 | |
| 4987 | reset_devices [KNL] Force drivers to reset the underlying device |
| 4988 | during initialization. |
| 4989 | |
| 4990 | resume= [SWSUSP] |
| 4991 | Specify the partition device for software suspend |
| 4992 | Format: |
| 4993 | {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>} |
| 4994 | |
| 4995 | resume_offset= [SWSUSP] |
| 4996 | Specify the offset from the beginning of the partition |
| 4997 | given by "resume=" at which the swap header is located, |
| 4998 | in <PAGE_SIZE> units (needed only for swap files). |
| 4999 | See Documentation/power/swsusp-and-swap-files.rst |
| 5000 | |
| 5001 | resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to |
| 5002 | read the resume files |
| 5003 | |
| 5004 | resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up. |
| 5005 | Useful for devices that are detected asynchronously |
| 5006 | (e.g. USB and MMC devices). |
| 5007 | |
| 5008 | hibernate= [HIBERNATION] |
| 5009 | noresume Don't check if there's a hibernation image |
| 5010 | present during boot. |
| 5011 | nocompress Don't compress/decompress hibernation images. |
| 5012 | no Disable hibernation and resume. |
| 5013 | protect_image Turn on image protection during restoration |
| 5014 | (that will set all pages holding image data |
| 5015 | during restoration read-only). |
| 5016 | |
| 5017 | retain_initrd [RAM] Keep initrd memory after extraction |
| 5018 | |
| 5019 | rfkill.default_state= |
| 5020 | 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm, |
| 5021 | etc. communication is blocked by default. |
| 5022 | 1 Unblocked. |
| 5023 | |
| 5024 | rfkill.master_switch_mode= |
| 5025 | 0 The "airplane mode" button does nothing. |
| 5026 | 1 The "airplane mode" button toggles between everything |
| 5027 | blocked and the previous configuration. |
| 5028 | 2 The "airplane mode" button toggles between everything |
| 5029 | blocked and everything unblocked. |
| 5030 | |
| 5031 | rhash_entries= [KNL,NET] |
| 5032 | Set number of hash buckets for route cache |
| 5033 | |
| 5034 | ring3mwait=disable |
| 5035 | [KNL] Disable ring 3 MONITOR/MWAIT feature on supported |
| 5036 | CPUs. |
| 5037 | |
| 5038 | ro [KNL] Mount root device read-only on boot |
| 5039 | |
| 5040 | rodata= [KNL] |
| 5041 | on Mark read-only kernel memory as read-only (default). |
| 5042 | off Leave read-only kernel memory writable for debugging. |
| 5043 | |
| 5044 | rockchip.usb_uart |
| 5045 | Enable the uart passthrough on the designated usb port |
| 5046 | on Rockchip SoCs. When active, the signals of the |
| 5047 | debug-uart get routed to the D+ and D- pins of the usb |
| 5048 | port and the regular usb controller gets disabled. |
| 5049 | |
| 5050 | root= [KNL] Root filesystem |
| 5051 | See name_to_dev_t comment in init/do_mounts.c. |
| 5052 | |
| 5053 | rootdelay= [KNL] Delay (in seconds) to pause before attempting to |
| 5054 | mount the root filesystem |
| 5055 | |
| 5056 | rootflags= [KNL] Set root filesystem mount option string |
| 5057 | |
| 5058 | rootfstype= [KNL] Set root filesystem type |
| 5059 | |
| 5060 | rootwait [KNL] Wait (indefinitely) for root device to show up. |
| 5061 | Useful for devices that are detected asynchronously |
| 5062 | (e.g. USB and MMC devices). |
| 5063 | |
| 5064 | rproc_mem=nn[KMG][@address] |
| 5065 | [KNL,ARM,CMA] Remoteproc physical memory block. |
| 5066 | Memory area to be used by remote processor image, |
| 5067 | managed by CMA. |
| 5068 | |
| 5069 | rw [KNL] Mount root device read-write on boot |
| 5070 | |
| 5071 | S [KNL] Run init in single mode |
| 5072 | |
| 5073 | s390_iommu= [HW,S390] |
| 5074 | Set s390 IOTLB flushing mode |
| 5075 | strict |
| 5076 | With strict flushing every unmap operation will result in |
| 5077 | an IOTLB flush. Default is lazy flushing before reuse, |
| 5078 | which is faster. |
| 5079 | |
| 5080 | s390_iommu_aperture= [KNL,S390] |
| 5081 | Specifies the size of the per device DMA address space |
| 5082 | accessible through the DMA and IOMMU APIs as a decimal |
| 5083 | factor of the size of main memory. |
| 5084 | The default is 1 meaning that one can concurrently use |
| 5085 | as many DMA addresses as physical memory is installed, |
| 5086 | if supported by hardware, and thus map all of memory |
| 5087 | once. With a value of 2 one can map all of memory twice |
| 5088 | and so on. As a special case a factor of 0 imposes no |
| 5089 | restrictions other than those given by hardware at the |
| 5090 | cost of significant additional memory use for tables. |
| 5091 | |
| 5092 | sa1100ir [NET] |
| 5093 | See drivers/net/irda/sa1100_ir.c. |
| 5094 | |
| 5095 | sched_verbose [KNL] Enables verbose scheduler debug messages. |
| 5096 | |
| 5097 | schedstats= [KNL,X86] Enable or disable scheduled statistics. |
| 5098 | Allowed values are enable and disable. This feature |
| 5099 | incurs a small amount of overhead in the scheduler |
| 5100 | but is useful for debugging and performance tuning. |
| 5101 | |
| 5102 | sched_thermal_decay_shift= |
| 5103 | [KNL, SMP] Set a decay shift for scheduler thermal |
| 5104 | pressure signal. Thermal pressure signal follows the |
| 5105 | default decay period of other scheduler pelt |
| 5106 | signals(usually 32 ms but configurable). Setting |
| 5107 | sched_thermal_decay_shift will left shift the decay |
| 5108 | period for the thermal pressure signal by the shift |
| 5109 | value. |
| 5110 | i.e. with the default pelt decay period of 32 ms |
| 5111 | sched_thermal_decay_shift thermal pressure decay pr |
| 5112 | 1 64 ms |
| 5113 | 2 128 ms |
| 5114 | and so on. |
| 5115 | Format: integer between 0 and 10 |
| 5116 | Default is 0. |
| 5117 | |
| 5118 | scftorture.holdoff= [KNL] |
| 5119 | Number of seconds to hold off before starting |
| 5120 | test. Defaults to zero for module insertion and |
| 5121 | to 10 seconds for built-in smp_call_function() |
| 5122 | tests. |
| 5123 | |
| 5124 | scftorture.longwait= [KNL] |
| 5125 | Request ridiculously long waits randomly selected |
| 5126 | up to the chosen limit in seconds. Zero (the |
| 5127 | default) disables this feature. Please note |
| 5128 | that requesting even small non-zero numbers of |
| 5129 | seconds can result in RCU CPU stall warnings, |
| 5130 | softlockup complaints, and so on. |
| 5131 | |
| 5132 | scftorture.nthreads= [KNL] |
| 5133 | Number of kthreads to spawn to invoke the |
| 5134 | smp_call_function() family of functions. |
| 5135 | The default of -1 specifies a number of kthreads |
| 5136 | equal to the number of CPUs. |
| 5137 | |
| 5138 | scftorture.onoff_holdoff= [KNL] |
| 5139 | Number seconds to wait after the start of the |
| 5140 | test before initiating CPU-hotplug operations. |
| 5141 | |
| 5142 | scftorture.onoff_interval= [KNL] |
| 5143 | Number seconds to wait between successive |
| 5144 | CPU-hotplug operations. Specifying zero (which |
| 5145 | is the default) disables CPU-hotplug operations. |
| 5146 | |
| 5147 | scftorture.shutdown_secs= [KNL] |
| 5148 | The number of seconds following the start of the |
| 5149 | test after which to shut down the system. The |
| 5150 | default of zero avoids shutting down the system. |
| 5151 | Non-zero values are useful for automated tests. |
| 5152 | |
| 5153 | scftorture.stat_interval= [KNL] |
| 5154 | The number of seconds between outputting the |
| 5155 | current test statistics to the console. A value |
| 5156 | of zero disables statistics output. |
| 5157 | |
| 5158 | scftorture.stutter_cpus= [KNL] |
| 5159 | The number of jiffies to wait between each change |
| 5160 | to the set of CPUs under test. |
| 5161 | |
| 5162 | scftorture.use_cpus_read_lock= [KNL] |
| 5163 | Use use_cpus_read_lock() instead of the default |
| 5164 | preempt_disable() to disable CPU hotplug |
| 5165 | while invoking one of the smp_call_function*() |
| 5166 | functions. |
| 5167 | |
| 5168 | scftorture.verbose= [KNL] |
| 5169 | Enable additional printk() statements. |
| 5170 | |
| 5171 | scftorture.weight_single= [KNL] |
| 5172 | The probability weighting to use for the |
| 5173 | smp_call_function_single() function with a zero |
| 5174 | "wait" parameter. A value of -1 selects the |
| 5175 | default if all other weights are -1. However, |
| 5176 | if at least one weight has some other value, a |
| 5177 | value of -1 will instead select a weight of zero. |
| 5178 | |
| 5179 | scftorture.weight_single_wait= [KNL] |
| 5180 | The probability weighting to use for the |
| 5181 | smp_call_function_single() function with a |
| 5182 | non-zero "wait" parameter. See weight_single. |
| 5183 | |
| 5184 | scftorture.weight_many= [KNL] |
| 5185 | The probability weighting to use for the |
| 5186 | smp_call_function_many() function with a zero |
| 5187 | "wait" parameter. See weight_single. |
| 5188 | Note well that setting a high probability for |
| 5189 | this weighting can place serious IPI load |
| 5190 | on the system. |
| 5191 | |
| 5192 | scftorture.weight_many_wait= [KNL] |
| 5193 | The probability weighting to use for the |
| 5194 | smp_call_function_many() function with a |
| 5195 | non-zero "wait" parameter. See weight_single |
| 5196 | and weight_many. |
| 5197 | |
| 5198 | scftorture.weight_all= [KNL] |
| 5199 | The probability weighting to use for the |
| 5200 | smp_call_function_all() function with a zero |
| 5201 | "wait" parameter. See weight_single and |
| 5202 | weight_many. |
| 5203 | |
| 5204 | scftorture.weight_all_wait= [KNL] |
| 5205 | The probability weighting to use for the |
| 5206 | smp_call_function_all() function with a |
| 5207 | non-zero "wait" parameter. See weight_single |
| 5208 | and weight_many. |
| 5209 | |
| 5210 | skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate |
| 5211 | xtime_lock contention on larger systems, and/or RCU lock |
| 5212 | contention on all systems with CONFIG_MAXSMP set. |
| 5213 | Format: { "0" | "1" } |
| 5214 | 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1" |
| 5215 | 1 -- enable. |
| 5216 | Note: increases power consumption, thus should only be |
| 5217 | enabled if running jitter sensitive (HPC/RT) workloads. |
| 5218 | |
| 5219 | security= [SECURITY] Choose a legacy "major" security module to |
| 5220 | enable at boot. This has been deprecated by the |
| 5221 | "lsm=" parameter. |
| 5222 | |
| 5223 | selinux= [SELINUX] Disable or enable SELinux at boot time. |
| 5224 | Format: { "0" | "1" } |
| 5225 | See security/selinux/Kconfig help text. |
| 5226 | 0 -- disable. |
| 5227 | 1 -- enable. |
| 5228 | Default value is 1. |
| 5229 | |
| 5230 | apparmor= [APPARMOR] Disable or enable AppArmor at boot time |
| 5231 | Format: { "0" | "1" } |
| 5232 | See security/apparmor/Kconfig help text |
| 5233 | 0 -- disable. |
| 5234 | 1 -- enable. |
| 5235 | Default value is set via kernel config option. |
| 5236 | |
| 5237 | serialnumber [BUGS=X86-32] |
| 5238 | |
| 5239 | shapers= [NET] |
| 5240 | Maximal number of shapers. |
| 5241 | |
| 5242 | simeth= [IA-64] |
| 5243 | simscsi= |
| 5244 | |
| 5245 | slram= [HW,MTD] |
| 5246 | |
| 5247 | slab_merge [MM] |
| 5248 | Enable merging of slabs with similar size when the |
| 5249 | kernel is built without CONFIG_SLAB_MERGE_DEFAULT. |
| 5250 | |
| 5251 | slab_nomerge [MM] |
| 5252 | Disable merging of slabs with similar size. May be |
| 5253 | necessary if there is some reason to distinguish |
| 5254 | allocs to different slabs, especially in hardened |
| 5255 | environments where the risk of heap overflows and |
| 5256 | layout control by attackers can usually be |
| 5257 | frustrated by disabling merging. This will reduce |
| 5258 | most of the exposure of a heap attack to a single |
| 5259 | cache (risks via metadata attacks are mostly |
| 5260 | unchanged). Debug options disable merging on their |
| 5261 | own. |
| 5262 | For more information see Documentation/vm/slub.rst. |
| 5263 | |
| 5264 | slab_max_order= [MM, SLAB] |
| 5265 | Determines the maximum allowed order for slabs. |
| 5266 | A high setting may cause OOMs due to memory |
| 5267 | fragmentation. Defaults to 1 for systems with |
| 5268 | more than 32MB of RAM, 0 otherwise. |
| 5269 | |
| 5270 | slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB] |
| 5271 | Enabling slub_debug allows one to determine the |
| 5272 | culprit if slab objects become corrupted. Enabling |
| 5273 | slub_debug can create guard zones around objects and |
| 5274 | may poison objects when not in use. Also tracks the |
| 5275 | last alloc / free. For more information see |
| 5276 | Documentation/vm/slub.rst. |
| 5277 | |
| 5278 | slub_max_order= [MM, SLUB] |
| 5279 | Determines the maximum allowed order for slabs. |
| 5280 | A high setting may cause OOMs due to memory |
| 5281 | fragmentation. For more information see |
| 5282 | Documentation/vm/slub.rst. |
| 5283 | |
| 5284 | slub_min_objects= [MM, SLUB] |
| 5285 | The minimum number of objects per slab. SLUB will |
| 5286 | increase the slab order up to slub_max_order to |
| 5287 | generate a sufficiently large slab able to contain |
| 5288 | the number of objects indicated. The higher the number |
| 5289 | of objects the smaller the overhead of tracking slabs |
| 5290 | and the less frequently locks need to be acquired. |
| 5291 | For more information see Documentation/vm/slub.rst. |
| 5292 | |
| 5293 | slub_min_order= [MM, SLUB] |
| 5294 | Determines the minimum page order for slabs. Must be |
| 5295 | lower than slub_max_order. |
| 5296 | For more information see Documentation/vm/slub.rst. |
| 5297 | |
| 5298 | slub_merge [MM, SLUB] |
| 5299 | Same with slab_merge. |
| 5300 | |
| 5301 | slub_nomerge [MM, SLUB] |
| 5302 | Same with slab_nomerge. This is supported for legacy. |
| 5303 | See slab_nomerge for more information. |
| 5304 | |
| 5305 | smart2= [HW] |
| 5306 | Format: <io1>[,<io2>[,...,<io8>]] |
| 5307 | |
| 5308 | smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices |
| 5309 | smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port |
| 5310 | smsc-ircc2.ircc_sir= [HW] SIR base I/O port |
| 5311 | smsc-ircc2.ircc_fir= [HW] FIR base I/O port |
| 5312 | smsc-ircc2.ircc_irq= [HW] IRQ line |
| 5313 | smsc-ircc2.ircc_dma= [HW] DMA channel |
| 5314 | smsc-ircc2.ircc_transceiver= [HW] Transceiver type: |
| 5315 | 0: Toshiba Satellite 1800 (GP data pin select) |
| 5316 | 1: Fast pin select (default) |
| 5317 | 2: ATC IRMode |
| 5318 | |
| 5319 | smt [KNL,S390] Set the maximum number of threads (logical |
| 5320 | CPUs) to use per physical CPU on systems capable of |
| 5321 | symmetric multithreading (SMT). Will be capped to the |
| 5322 | actual hardware limit. |
| 5323 | Format: <integer> |
| 5324 | Default: -1 (no limit) |
| 5325 | |
| 5326 | softlockup_panic= |
| 5327 | [KNL] Should the soft-lockup detector generate panics. |
| 5328 | Format: 0 | 1 |
| 5329 | |
| 5330 | A value of 1 instructs the soft-lockup detector |
| 5331 | to panic the machine when a soft-lockup occurs. It is |
| 5332 | also controlled by the kernel.softlockup_panic sysctl |
| 5333 | and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the |
| 5334 | respective build-time switch to that functionality. |
| 5335 | |
| 5336 | softlockup_all_cpu_backtrace= |
| 5337 | [KNL] Should the soft-lockup detector generate |
| 5338 | backtraces on all cpus. |
| 5339 | Format: 0 | 1 |
| 5340 | |
| 5341 | sonypi.*= [HW] Sony Programmable I/O Control Device driver |
| 5342 | See Documentation/admin-guide/laptops/sonypi.rst |
| 5343 | |
| 5344 | spectre_v2= [X86] Control mitigation of Spectre variant 2 |
| 5345 | (indirect branch speculation) vulnerability. |
| 5346 | The default operation protects the kernel from |
| 5347 | user space attacks. |
| 5348 | |
| 5349 | on - unconditionally enable, implies |
| 5350 | spectre_v2_user=on |
| 5351 | off - unconditionally disable, implies |
| 5352 | spectre_v2_user=off |
| 5353 | auto - kernel detects whether your CPU model is |
| 5354 | vulnerable |
| 5355 | |
| 5356 | Selecting 'on' will, and 'auto' may, choose a |
| 5357 | mitigation method at run time according to the |
| 5358 | CPU, the available microcode, the setting of the |
| 5359 | CONFIG_RETPOLINE configuration option, and the |
| 5360 | compiler with which the kernel was built. |
| 5361 | |
| 5362 | Selecting 'on' will also enable the mitigation |
| 5363 | against user space to user space task attacks. |
| 5364 | |
| 5365 | Selecting 'off' will disable both the kernel and |
| 5366 | the user space protections. |
| 5367 | |
| 5368 | Specific mitigations can also be selected manually: |
| 5369 | |
| 5370 | retpoline - replace indirect branches |
| 5371 | retpoline,generic - google's original retpoline |
| 5372 | retpoline,amd - AMD-specific minimal thunk |
| 5373 | |
| 5374 | Not specifying this option is equivalent to |
| 5375 | spectre_v2=auto. |
| 5376 | |
| 5377 | spectre_v2_user= |
| 5378 | [X86] Control mitigation of Spectre variant 2 |
| 5379 | (indirect branch speculation) vulnerability between |
| 5380 | user space tasks |
| 5381 | |
| 5382 | on - Unconditionally enable mitigations. Is |
| 5383 | enforced by spectre_v2=on |
| 5384 | |
| 5385 | off - Unconditionally disable mitigations. Is |
| 5386 | enforced by spectre_v2=off |
| 5387 | |
| 5388 | prctl - Indirect branch speculation is enabled, |
| 5389 | but mitigation can be enabled via prctl |
| 5390 | per thread. The mitigation control state |
| 5391 | is inherited on fork. |
| 5392 | |
| 5393 | prctl,ibpb |
| 5394 | - Like "prctl" above, but only STIBP is |
| 5395 | controlled per thread. IBPB is issued |
| 5396 | always when switching between different user |
| 5397 | space processes. |
| 5398 | |
| 5399 | seccomp |
| 5400 | - Same as "prctl" above, but all seccomp |
| 5401 | threads will enable the mitigation unless |
| 5402 | they explicitly opt out. |
| 5403 | |
| 5404 | seccomp,ibpb |
| 5405 | - Like "seccomp" above, but only STIBP is |
| 5406 | controlled per thread. IBPB is issued |
| 5407 | always when switching between different |
| 5408 | user space processes. |
| 5409 | |
| 5410 | auto - Kernel selects the mitigation depending on |
| 5411 | the available CPU features and vulnerability. |
| 5412 | |
| 5413 | Default mitigation: "prctl" |
| 5414 | |
| 5415 | Not specifying this option is equivalent to |
| 5416 | spectre_v2_user=auto. |
| 5417 | |
| 5418 | spec_store_bypass_disable= |
| 5419 | [HW] Control Speculative Store Bypass (SSB) Disable mitigation |
| 5420 | (Speculative Store Bypass vulnerability) |
| 5421 | |
| 5422 | Certain CPUs are vulnerable to an exploit against a |
| 5423 | a common industry wide performance optimization known |
| 5424 | as "Speculative Store Bypass" in which recent stores |
| 5425 | to the same memory location may not be observed by |
| 5426 | later loads during speculative execution. The idea |
| 5427 | is that such stores are unlikely and that they can |
| 5428 | be detected prior to instruction retirement at the |
| 5429 | end of a particular speculation execution window. |
| 5430 | |
| 5431 | In vulnerable processors, the speculatively forwarded |
| 5432 | store can be used in a cache side channel attack, for |
| 5433 | example to read memory to which the attacker does not |
| 5434 | directly have access (e.g. inside sandboxed code). |
| 5435 | |
| 5436 | This parameter controls whether the Speculative Store |
| 5437 | Bypass optimization is used. |
| 5438 | |
| 5439 | On x86 the options are: |
| 5440 | |
| 5441 | on - Unconditionally disable Speculative Store Bypass |
| 5442 | off - Unconditionally enable Speculative Store Bypass |
| 5443 | auto - Kernel detects whether the CPU model contains an |
| 5444 | implementation of Speculative Store Bypass and |
| 5445 | picks the most appropriate mitigation. If the |
| 5446 | CPU is not vulnerable, "off" is selected. If the |
| 5447 | CPU is vulnerable the default mitigation is |
| 5448 | architecture and Kconfig dependent. See below. |
| 5449 | prctl - Control Speculative Store Bypass per thread |
| 5450 | via prctl. Speculative Store Bypass is enabled |
| 5451 | for a process by default. The state of the control |
| 5452 | is inherited on fork. |
| 5453 | seccomp - Same as "prctl" above, but all seccomp threads |
| 5454 | will disable SSB unless they explicitly opt out. |
| 5455 | |
| 5456 | Default mitigations: |
| 5457 | X86: "prctl" |
| 5458 | |
| 5459 | On powerpc the options are: |
| 5460 | |
| 5461 | on,auto - On Power8 and Power9 insert a store-forwarding |
| 5462 | barrier on kernel entry and exit. On Power7 |
| 5463 | perform a software flush on kernel entry and |
| 5464 | exit. |
| 5465 | off - No action. |
| 5466 | |
| 5467 | Not specifying this option is equivalent to |
| 5468 | spec_store_bypass_disable=auto. |
| 5469 | |
| 5470 | spia_io_base= [HW,MTD] |
| 5471 | spia_fio_base= |
| 5472 | spia_pedr= |
| 5473 | spia_peddr= |
| 5474 | |
| 5475 | split_lock_detect= |
| 5476 | [X86] Enable split lock detection or bus lock detection |
| 5477 | |
| 5478 | When enabled (and if hardware support is present), atomic |
| 5479 | instructions that access data across cache line |
| 5480 | boundaries will result in an alignment check exception |
| 5481 | for split lock detection or a debug exception for |
| 5482 | bus lock detection. |
| 5483 | |
| 5484 | off - not enabled |
| 5485 | |
| 5486 | warn - the kernel will emit rate-limited warnings |
| 5487 | about applications triggering the #AC |
| 5488 | exception or the #DB exception. This mode is |
| 5489 | the default on CPUs that support split lock |
| 5490 | detection or bus lock detection. Default |
| 5491 | behavior is by #AC if both features are |
| 5492 | enabled in hardware. |
| 5493 | |
| 5494 | fatal - the kernel will send SIGBUS to applications |
| 5495 | that trigger the #AC exception or the #DB |
| 5496 | exception. Default behavior is by #AC if |
| 5497 | both features are enabled in hardware. |
| 5498 | |
| 5499 | ratelimit:N - |
| 5500 | Set system wide rate limit to N bus locks |
| 5501 | per second for bus lock detection. |
| 5502 | 0 < N <= 1000. |
| 5503 | |
| 5504 | N/A for split lock detection. |
| 5505 | |
| 5506 | |
| 5507 | If an #AC exception is hit in the kernel or in |
| 5508 | firmware (i.e. not while executing in user mode) |
| 5509 | the kernel will oops in either "warn" or "fatal" |
| 5510 | mode. |
| 5511 | |
| 5512 | #DB exception for bus lock is triggered only when |
| 5513 | CPL > 0. |
| 5514 | |
| 5515 | srbds= [X86,INTEL] |
| 5516 | Control the Special Register Buffer Data Sampling |
| 5517 | (SRBDS) mitigation. |
| 5518 | |
| 5519 | Certain CPUs are vulnerable to an MDS-like |
| 5520 | exploit which can leak bits from the random |
| 5521 | number generator. |
| 5522 | |
| 5523 | By default, this issue is mitigated by |
| 5524 | microcode. However, the microcode fix can cause |
| 5525 | the RDRAND and RDSEED instructions to become |
| 5526 | much slower. Among other effects, this will |
| 5527 | result in reduced throughput from /dev/urandom. |
| 5528 | |
| 5529 | The microcode mitigation can be disabled with |
| 5530 | the following option: |
| 5531 | |
| 5532 | off: Disable mitigation and remove |
| 5533 | performance impact to RDRAND and RDSEED |
| 5534 | |
| 5535 | srcutree.counter_wrap_check [KNL] |
| 5536 | Specifies how frequently to check for |
| 5537 | grace-period sequence counter wrap for the |
| 5538 | srcu_data structure's ->srcu_gp_seq_needed field. |
| 5539 | The greater the number of bits set in this kernel |
| 5540 | parameter, the less frequently counter wrap will |
| 5541 | be checked for. Note that the bottom two bits |
| 5542 | are ignored. |
| 5543 | |
| 5544 | srcutree.exp_holdoff [KNL] |
| 5545 | Specifies how many nanoseconds must elapse |
| 5546 | since the end of the last SRCU grace period for |
| 5547 | a given srcu_struct until the next normal SRCU |
| 5548 | grace period will be considered for automatic |
| 5549 | expediting. Set to zero to disable automatic |
| 5550 | expediting. |
| 5551 | |
| 5552 | ssbd= [ARM64,HW] |
| 5553 | Speculative Store Bypass Disable control |
| 5554 | |
| 5555 | On CPUs that are vulnerable to the Speculative |
| 5556 | Store Bypass vulnerability and offer a |
| 5557 | firmware based mitigation, this parameter |
| 5558 | indicates how the mitigation should be used: |
| 5559 | |
| 5560 | force-on: Unconditionally enable mitigation for |
| 5561 | for both kernel and userspace |
| 5562 | force-off: Unconditionally disable mitigation for |
| 5563 | for both kernel and userspace |
| 5564 | kernel: Always enable mitigation in the |
| 5565 | kernel, and offer a prctl interface |
| 5566 | to allow userspace to register its |
| 5567 | interest in being mitigated too. |
| 5568 | |
| 5569 | stack_guard_gap= [MM] |
| 5570 | override the default stack gap protection. The value |
| 5571 | is in page units and it defines how many pages prior |
| 5572 | to (for stacks growing down) resp. after (for stacks |
| 5573 | growing up) the main stack are reserved for no other |
| 5574 | mapping. Default value is 256 pages. |
| 5575 | |
| 5576 | stack_depot_disable= [KNL] |
| 5577 | Setting this to true through kernel command line will |
| 5578 | disable the stack depot thereby saving the static memory |
| 5579 | consumed by the stack hash table. By default this is set |
| 5580 | to false. |
| 5581 | |
| 5582 | stacktrace [FTRACE] |
| 5583 | Enabled the stack tracer on boot up. |
| 5584 | |
| 5585 | stacktrace_filter=[function-list] |
| 5586 | [FTRACE] Limit the functions that the stack tracer |
| 5587 | will trace at boot up. function-list is a comma-separated |
| 5588 | list of functions. This list can be changed at run |
| 5589 | time by the stack_trace_filter file in the debugfs |
| 5590 | tracing directory. Note, this enables stack tracing |
| 5591 | and the stacktrace above is not needed. |
| 5592 | |
| 5593 | sti= [PARISC,HW] |
| 5594 | Format: <num> |
| 5595 | Set the STI (builtin display/keyboard on the HP-PARISC |
| 5596 | machines) console (graphic card) which should be used |
| 5597 | as the initial boot-console. |
| 5598 | See also comment in drivers/video/console/sticore.c. |
| 5599 | |
| 5600 | sti_font= [HW] |
| 5601 | See comment in drivers/video/console/sticore.c. |
| 5602 | |
| 5603 | stifb= [HW] |
| 5604 | Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]] |
| 5605 | |
| 5606 | strict_sas_size= |
| 5607 | [X86] |
| 5608 | Format: <bool> |
| 5609 | Enable or disable strict sigaltstack size checks |
| 5610 | against the required signal frame size which |
| 5611 | depends on the supported FPU features. This can |
| 5612 | be used to filter out binaries which have |
| 5613 | not yet been made aware of AT_MINSIGSTKSZ. |
| 5614 | |
| 5615 | sunrpc.min_resvport= |
| 5616 | sunrpc.max_resvport= |
| 5617 | [NFS,SUNRPC] |
| 5618 | SunRPC servers often require that client requests |
| 5619 | originate from a privileged port (i.e. a port in the |
| 5620 | range 0 < portnr < 1024). |
| 5621 | An administrator who wishes to reserve some of these |
| 5622 | ports for other uses may adjust the range that the |
| 5623 | kernel's sunrpc client considers to be privileged |
| 5624 | using these two parameters to set the minimum and |
| 5625 | maximum port values. |
| 5626 | |
| 5627 | sunrpc.svc_rpc_per_connection_limit= |
| 5628 | [NFS,SUNRPC] |
| 5629 | Limit the number of requests that the server will |
| 5630 | process in parallel from a single connection. |
| 5631 | The default value is 0 (no limit). |
| 5632 | |
| 5633 | sunrpc.pool_mode= |
| 5634 | [NFS] |
| 5635 | Control how the NFS server code allocates CPUs to |
| 5636 | service thread pools. Depending on how many NICs |
| 5637 | you have and where their interrupts are bound, this |
| 5638 | option will affect which CPUs will do NFS serving. |
| 5639 | Note: this parameter cannot be changed while the |
| 5640 | NFS server is running. |
| 5641 | |
| 5642 | auto the server chooses an appropriate mode |
| 5643 | automatically using heuristics |
| 5644 | global a single global pool contains all CPUs |
| 5645 | percpu one pool for each CPU |
| 5646 | pernode one pool for each NUMA node (equivalent |
| 5647 | to global on non-NUMA machines) |
| 5648 | |
| 5649 | sunrpc.tcp_slot_table_entries= |
| 5650 | sunrpc.udp_slot_table_entries= |
| 5651 | [NFS,SUNRPC] |
| 5652 | Sets the upper limit on the number of simultaneous |
| 5653 | RPC calls that can be sent from the client to a |
| 5654 | server. Increasing these values may allow you to |
| 5655 | improve throughput, but will also increase the |
| 5656 | amount of memory reserved for use by the client. |
| 5657 | |
| 5658 | suspend.pm_test_delay= |
| 5659 | [SUSPEND] |
| 5660 | Sets the number of seconds to remain in a suspend test |
| 5661 | mode before resuming the system (see |
| 5662 | /sys/power/pm_test). Only available when CONFIG_PM_DEBUG |
| 5663 | is set. Default value is 5. |
| 5664 | |
| 5665 | svm= [PPC] |
| 5666 | Format: { on | off | y | n | 1 | 0 } |
| 5667 | This parameter controls use of the Protected |
| 5668 | Execution Facility on pSeries. |
| 5669 | |
| 5670 | swapaccount=[0|1] |
| 5671 | [KNL] Enable accounting of swap in memory resource |
| 5672 | controller if no parameter or 1 is given or disable |
| 5673 | it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst) |
| 5674 | |
| 5675 | swiotlb= [ARM,IA-64,PPC,MIPS,X86] |
| 5676 | Format: { <int> | force | noforce } |
| 5677 | <int> -- Number of I/O TLB slabs |
| 5678 | force -- force using of bounce buffers even if they |
| 5679 | wouldn't be automatically used by the kernel |
| 5680 | noforce -- Never use bounce buffers (for debugging) |
| 5681 | |
| 5682 | switches= [HW,M68k] |
| 5683 | |
| 5684 | sysctl.*= [KNL] |
| 5685 | Set a sysctl parameter, right before loading the init |
| 5686 | process, as if the value was written to the respective |
| 5687 | /proc/sys/... file. Both '.' and '/' are recognized as |
| 5688 | separators. Unrecognized parameters and invalid values |
| 5689 | are reported in the kernel log. Sysctls registered |
| 5690 | later by a loaded module cannot be set this way. |
| 5691 | Example: sysctl.vm.swappiness=40 |
| 5692 | |
| 5693 | sysfs.deprecated=0|1 [KNL] |
| 5694 | Enable/disable old style sysfs layout for old udev |
| 5695 | on older distributions. When this option is enabled |
| 5696 | very new udev will not work anymore. When this option |
| 5697 | is disabled (or CONFIG_SYSFS_DEPRECATED not compiled) |
| 5698 | in older udev will not work anymore. |
| 5699 | Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in |
| 5700 | the kernel configuration. |
| 5701 | |
| 5702 | sysrq_always_enabled |
| 5703 | [KNL] |
| 5704 | Ignore sysrq setting - this boot parameter will |
| 5705 | neutralize any effect of /proc/sys/kernel/sysrq. |
| 5706 | Useful for debugging. |
| 5707 | |
| 5708 | tcpmhash_entries= [KNL,NET] |
| 5709 | Set the number of tcp_metrics_hash slots. |
| 5710 | Default value is 8192 or 16384 depending on total |
| 5711 | ram pages. This is used to specify the TCP metrics |
| 5712 | cache size. See Documentation/networking/ip-sysctl.rst |
| 5713 | "tcp_no_metrics_save" section for more details. |
| 5714 | |
| 5715 | tdfx= [HW,DRM] |
| 5716 | |
| 5717 | test_suspend= [SUSPEND][,N] |
| 5718 | Specify "mem" (for Suspend-to-RAM) or "standby" (for |
| 5719 | standby suspend) or "freeze" (for suspend type freeze) |
| 5720 | as the system sleep state during system startup with |
| 5721 | the optional capability to repeat N number of times. |
| 5722 | The system is woken from this state using a |
| 5723 | wakeup-capable RTC alarm. |
| 5724 | |
| 5725 | thash_entries= [KNL,NET] |
| 5726 | Set number of hash buckets for TCP connection |
| 5727 | |
| 5728 | thermal.act= [HW,ACPI] |
| 5729 | -1: disable all active trip points in all thermal zones |
| 5730 | <degrees C>: override all lowest active trip points |
| 5731 | |
| 5732 | thermal.crt= [HW,ACPI] |
| 5733 | -1: disable all critical trip points in all thermal zones |
| 5734 | <degrees C>: override all critical trip points |
| 5735 | |
| 5736 | thermal.nocrt= [HW,ACPI] |
| 5737 | Set to disable actions on ACPI thermal zone |
| 5738 | critical and hot trip points. |
| 5739 | |
| 5740 | thermal.off= [HW,ACPI] |
| 5741 | 1: disable ACPI thermal control |
| 5742 | |
| 5743 | thermal.psv= [HW,ACPI] |
| 5744 | -1: disable all passive trip points |
| 5745 | <degrees C>: override all passive trip points to this |
| 5746 | value |
| 5747 | |
| 5748 | thermal.tzp= [HW,ACPI] |
| 5749 | Specify global default ACPI thermal zone polling rate |
| 5750 | <deci-seconds>: poll all this frequency |
| 5751 | 0: no polling (default) |
| 5752 | |
| 5753 | threadirqs [KNL] |
| 5754 | Force threading of all interrupt handlers except those |
| 5755 | marked explicitly IRQF_NO_THREAD. |
| 5756 | |
| 5757 | topology= [S390] |
| 5758 | Format: {off | on} |
| 5759 | Specify if the kernel should make use of the cpu |
| 5760 | topology information if the hardware supports this. |
| 5761 | The scheduler will make use of this information and |
| 5762 | e.g. base its process migration decisions on it. |
| 5763 | Default is on. |
| 5764 | |
| 5765 | topology_updates= [KNL, PPC, NUMA] |
| 5766 | Format: {off} |
| 5767 | Specify if the kernel should ignore (off) |
| 5768 | topology updates sent by the hypervisor to this |
| 5769 | LPAR. |
| 5770 | |
| 5771 | torture.disable_onoff_at_boot= [KNL] |
| 5772 | Prevent the CPU-hotplug component of torturing |
| 5773 | until after init has spawned. |
| 5774 | |
| 5775 | torture.ftrace_dump_at_shutdown= [KNL] |
| 5776 | Dump the ftrace buffer at torture-test shutdown, |
| 5777 | even if there were no errors. This can be a |
| 5778 | very costly operation when many torture tests |
| 5779 | are running concurrently, especially on systems |
| 5780 | with rotating-rust storage. |
| 5781 | |
| 5782 | torture.verbose_sleep_frequency= [KNL] |
| 5783 | Specifies how many verbose printk()s should be |
| 5784 | emitted between each sleep. The default of zero |
| 5785 | disables verbose-printk() sleeping. |
| 5786 | |
| 5787 | torture.verbose_sleep_duration= [KNL] |
| 5788 | Duration of each verbose-printk() sleep in jiffies. |
| 5789 | |
| 5790 | tp720= [HW,PS2] |
| 5791 | |
| 5792 | tpm_suspend_pcr=[HW,TPM] |
| 5793 | Format: integer pcr id |
| 5794 | Specify that at suspend time, the tpm driver |
| 5795 | should extend the specified pcr with zeros, |
| 5796 | as a workaround for some chips which fail to |
| 5797 | flush the last written pcr on TPM_SaveState. |
| 5798 | This will guarantee that all the other pcrs |
| 5799 | are saved. |
| 5800 | |
| 5801 | trace_buf_size=nn[KMG] |
| 5802 | [FTRACE] will set tracing buffer size on each cpu. |
| 5803 | |
| 5804 | trace_event=[event-list] |
| 5805 | [FTRACE] Set and start specified trace events in order |
| 5806 | to facilitate early boot debugging. The event-list is a |
| 5807 | comma-separated list of trace events to enable. See |
| 5808 | also Documentation/trace/events.rst |
| 5809 | |
| 5810 | trace_options=[option-list] |
| 5811 | [FTRACE] Enable or disable tracer options at boot. |
| 5812 | The option-list is a comma delimited list of options |
| 5813 | that can be enabled or disabled just as if you were |
| 5814 | to echo the option name into |
| 5815 | |
| 5816 | /sys/kernel/debug/tracing/trace_options |
| 5817 | |
| 5818 | For example, to enable stacktrace option (to dump the |
| 5819 | stack trace of each event), add to the command line: |
| 5820 | |
| 5821 | trace_options=stacktrace |
| 5822 | |
| 5823 | See also Documentation/trace/ftrace.rst "trace options" |
| 5824 | section. |
| 5825 | |
| 5826 | tp_printk[FTRACE] |
| 5827 | Have the tracepoints sent to printk as well as the |
| 5828 | tracing ring buffer. This is useful for early boot up |
| 5829 | where the system hangs or reboots and does not give the |
| 5830 | option for reading the tracing buffer or performing a |
| 5831 | ftrace_dump_on_oops. |
| 5832 | |
| 5833 | To turn off having tracepoints sent to printk, |
| 5834 | echo 0 > /proc/sys/kernel/tracepoint_printk |
| 5835 | Note, echoing 1 into this file without the |
| 5836 | tracepoint_printk kernel cmdline option has no effect. |
| 5837 | |
| 5838 | The tp_printk_stop_on_boot (see below) can also be used |
| 5839 | to stop the printing of events to console at |
| 5840 | late_initcall_sync. |
| 5841 | |
| 5842 | ** CAUTION ** |
| 5843 | |
| 5844 | Having tracepoints sent to printk() and activating high |
| 5845 | frequency tracepoints such as irq or sched, can cause |
| 5846 | the system to live lock. |
| 5847 | |
| 5848 | tp_printk_stop_on_boot[FTRACE] |
| 5849 | When tp_printk (above) is set, it can cause a lot of noise |
| 5850 | on the console. It may be useful to only include the |
| 5851 | printing of events during boot up, as user space may |
| 5852 | make the system inoperable. |
| 5853 | |
| 5854 | This command line option will stop the printing of events |
| 5855 | to console at the late_initcall_sync() time frame. |
| 5856 | |
| 5857 | traceoff_on_warning |
| 5858 | [FTRACE] enable this option to disable tracing when a |
| 5859 | warning is hit. This turns off "tracing_on". Tracing can |
| 5860 | be enabled again by echoing '1' into the "tracing_on" |
| 5861 | file located in /sys/kernel/debug/tracing/ |
| 5862 | |
| 5863 | This option is useful, as it disables the trace before |
| 5864 | the WARNING dump is called, which prevents the trace to |
| 5865 | be filled with content caused by the warning output. |
| 5866 | |
| 5867 | This option can also be set at run time via the sysctl |
| 5868 | option: kernel/traceoff_on_warning |
| 5869 | |
| 5870 | transparent_hugepage= |
| 5871 | [KNL] |
| 5872 | Format: [always|madvise|never] |
| 5873 | Can be used to control the default behavior of the system |
| 5874 | with respect to transparent hugepages. |
| 5875 | See Documentation/admin-guide/mm/transhuge.rst |
| 5876 | for more details. |
| 5877 | |
| 5878 | trusted.source= [KEYS] |
| 5879 | Format: <string> |
| 5880 | This parameter identifies the trust source as a backend |
| 5881 | for trusted keys implementation. Supported trust |
| 5882 | sources: |
| 5883 | - "tpm" |
| 5884 | - "tee" |
| 5885 | If not specified then it defaults to iterating through |
| 5886 | the trust source list starting with TPM and assigns the |
| 5887 | first trust source as a backend which is initialized |
| 5888 | successfully during iteration. |
| 5889 | |
| 5890 | tsc= Disable clocksource stability checks for TSC. |
| 5891 | Format: <string> |
| 5892 | [x86] reliable: mark tsc clocksource as reliable, this |
| 5893 | disables clocksource verification at runtime, as well |
| 5894 | as the stability checks done at bootup. Used to enable |
| 5895 | high-resolution timer mode on older hardware, and in |
| 5896 | virtualized environment. |
| 5897 | [x86] noirqtime: Do not use TSC to do irq accounting. |
| 5898 | Used to run time disable IRQ_TIME_ACCOUNTING on any |
| 5899 | platforms where RDTSC is slow and this accounting |
| 5900 | can add overhead. |
| 5901 | [x86] unstable: mark the TSC clocksource as unstable, this |
| 5902 | marks the TSC unconditionally unstable at bootup and |
| 5903 | avoids any further wobbles once the TSC watchdog notices. |
| 5904 | [x86] nowatchdog: disable clocksource watchdog. Used |
| 5905 | in situations with strict latency requirements (where |
| 5906 | interruptions from clocksource watchdog are not |
| 5907 | acceptable). |
| 5908 | |
| 5909 | tsc_early_khz= [X86] Skip early TSC calibration and use the given |
| 5910 | value instead. Useful when the early TSC frequency discovery |
| 5911 | procedure is not reliable, such as on overclocked systems |
| 5912 | with CPUID.16h support and partial CPUID.15h support. |
| 5913 | Format: <unsigned int> |
| 5914 | |
| 5915 | tsx= [X86] Control Transactional Synchronization |
| 5916 | Extensions (TSX) feature in Intel processors that |
| 5917 | support TSX control. |
| 5918 | |
| 5919 | This parameter controls the TSX feature. The options are: |
| 5920 | |
| 5921 | on - Enable TSX on the system. Although there are |
| 5922 | mitigations for all known security vulnerabilities, |
| 5923 | TSX has been known to be an accelerator for |
| 5924 | several previous speculation-related CVEs, and |
| 5925 | so there may be unknown security risks associated |
| 5926 | with leaving it enabled. |
| 5927 | |
| 5928 | off - Disable TSX on the system. (Note that this |
| 5929 | option takes effect only on newer CPUs which are |
| 5930 | not vulnerable to MDS, i.e., have |
| 5931 | MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get |
| 5932 | the new IA32_TSX_CTRL MSR through a microcode |
| 5933 | update. This new MSR allows for the reliable |
| 5934 | deactivation of the TSX functionality.) |
| 5935 | |
| 5936 | auto - Disable TSX if X86_BUG_TAA is present, |
| 5937 | otherwise enable TSX on the system. |
| 5938 | |
| 5939 | Not specifying this option is equivalent to tsx=off. |
| 5940 | |
| 5941 | See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst |
| 5942 | for more details. |
| 5943 | |
| 5944 | tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async |
| 5945 | Abort (TAA) vulnerability. |
| 5946 | |
| 5947 | Similar to Micro-architectural Data Sampling (MDS) |
| 5948 | certain CPUs that support Transactional |
| 5949 | Synchronization Extensions (TSX) are vulnerable to an |
| 5950 | exploit against CPU internal buffers which can forward |
| 5951 | information to a disclosure gadget under certain |
| 5952 | conditions. |
| 5953 | |
| 5954 | In vulnerable processors, the speculatively forwarded |
| 5955 | data can be used in a cache side channel attack, to |
| 5956 | access data to which the attacker does not have direct |
| 5957 | access. |
| 5958 | |
| 5959 | This parameter controls the TAA mitigation. The |
| 5960 | options are: |
| 5961 | |
| 5962 | full - Enable TAA mitigation on vulnerable CPUs |
| 5963 | if TSX is enabled. |
| 5964 | |
| 5965 | full,nosmt - Enable TAA mitigation and disable SMT on |
| 5966 | vulnerable CPUs. If TSX is disabled, SMT |
| 5967 | is not disabled because CPU is not |
| 5968 | vulnerable to cross-thread TAA attacks. |
| 5969 | off - Unconditionally disable TAA mitigation |
| 5970 | |
| 5971 | On MDS-affected machines, tsx_async_abort=off can be |
| 5972 | prevented by an active MDS mitigation as both vulnerabilities |
| 5973 | are mitigated with the same mechanism so in order to disable |
| 5974 | this mitigation, you need to specify mds=off too. |
| 5975 | |
| 5976 | Not specifying this option is equivalent to |
| 5977 | tsx_async_abort=full. On CPUs which are MDS affected |
| 5978 | and deploy MDS mitigation, TAA mitigation is not |
| 5979 | required and doesn't provide any additional |
| 5980 | mitigation. |
| 5981 | |
| 5982 | For details see: |
| 5983 | Documentation/admin-guide/hw-vuln/tsx_async_abort.rst |
| 5984 | |
| 5985 | turbografx.map[2|3]= [HW,JOY] |
| 5986 | TurboGraFX parallel port interface |
| 5987 | Format: |
| 5988 | <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7> |
| 5989 | See also Documentation/input/devices/joystick-parport.rst |
| 5990 | |
| 5991 | udbg-immortal [PPC] When debugging early kernel crashes that |
| 5992 | happen after console_init() and before a proper |
| 5993 | console driver takes over, this boot options might |
| 5994 | help "seeing" what's going on. |
| 5995 | |
| 5996 | uhash_entries= [KNL,NET] |
| 5997 | Set number of hash buckets for UDP/UDP-Lite connections |
| 5998 | |
| 5999 | uhci-hcd.ignore_oc= |
| 6000 | [USB] Ignore overcurrent events (default N). |
| 6001 | Some badly-designed motherboards generate lots of |
| 6002 | bogus events, for ports that aren't wired to |
| 6003 | anything. Set this parameter to avoid log spamming. |
| 6004 | Note that genuine overcurrent events won't be |
| 6005 | reported either. |
| 6006 | |
| 6007 | unknown_nmi_panic |
| 6008 | [X86] Cause panic on unknown NMI. |
| 6009 | |
| 6010 | usbcore.authorized_default= |
| 6011 | [USB] Default USB device authorization: |
| 6012 | (default -1 = authorized except for wireless USB, |
| 6013 | 0 = not authorized, 1 = authorized, 2 = authorized |
| 6014 | if device connected to internal port) |
| 6015 | |
| 6016 | usbcore.autosuspend= |
| 6017 | [USB] The autosuspend time delay (in seconds) used |
| 6018 | for newly-detected USB devices (default 2). This |
| 6019 | is the time required before an idle device will be |
| 6020 | autosuspended. Devices for which the delay is set |
| 6021 | to a negative value won't be autosuspended at all. |
| 6022 | |
| 6023 | usbcore.usbfs_snoop= |
| 6024 | [USB] Set to log all usbfs traffic (default 0 = off). |
| 6025 | |
| 6026 | usbcore.usbfs_snoop_max= |
| 6027 | [USB] Maximum number of bytes to snoop in each URB |
| 6028 | (default = 65536). |
| 6029 | |
| 6030 | usbcore.blinkenlights= |
| 6031 | [USB] Set to cycle leds on hubs (default 0 = off). |
| 6032 | |
| 6033 | usbcore.old_scheme_first= |
| 6034 | [USB] Start with the old device initialization |
| 6035 | scheme (default 0 = off). |
| 6036 | |
| 6037 | usbcore.usbfs_memory_mb= |
| 6038 | [USB] Memory limit (in MB) for buffers allocated by |
| 6039 | usbfs (default = 16, 0 = max = 2047). |
| 6040 | |
| 6041 | usbcore.use_both_schemes= |
| 6042 | [USB] Try the other device initialization scheme |
| 6043 | if the first one fails (default 1 = enabled). |
| 6044 | |
| 6045 | usbcore.initial_descriptor_timeout= |
| 6046 | [USB] Specifies timeout for the initial 64-byte |
| 6047 | USB_REQ_GET_DESCRIPTOR request in milliseconds |
| 6048 | (default 5000 = 5.0 seconds). |
| 6049 | |
| 6050 | usbcore.nousb [USB] Disable the USB subsystem |
| 6051 | |
| 6052 | usbcore.quirks= |
| 6053 | [USB] A list of quirk entries to augment the built-in |
| 6054 | usb core quirk list. List entries are separated by |
| 6055 | commas. Each entry has the form |
| 6056 | VendorID:ProductID:Flags. The IDs are 4-digit hex |
| 6057 | numbers and Flags is a set of letters. Each letter |
| 6058 | will change the built-in quirk; setting it if it is |
| 6059 | clear and clearing it if it is set. The letters have |
| 6060 | the following meanings: |
| 6061 | a = USB_QUIRK_STRING_FETCH_255 (string |
| 6062 | descriptors must not be fetched using |
| 6063 | a 255-byte read); |
| 6064 | b = USB_QUIRK_RESET_RESUME (device can't resume |
| 6065 | correctly so reset it instead); |
| 6066 | c = USB_QUIRK_NO_SET_INTF (device can't handle |
| 6067 | Set-Interface requests); |
| 6068 | d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't |
| 6069 | handle its Configuration or Interface |
| 6070 | strings); |
| 6071 | e = USB_QUIRK_RESET (device can't be reset |
| 6072 | (e.g morph devices), don't use reset); |
| 6073 | f = USB_QUIRK_HONOR_BNUMINTERFACES (device has |
| 6074 | more interface descriptions than the |
| 6075 | bNumInterfaces count, and can't handle |
| 6076 | talking to these interfaces); |
| 6077 | g = USB_QUIRK_DELAY_INIT (device needs a pause |
| 6078 | during initialization, after we read |
| 6079 | the device descriptor); |
| 6080 | h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For |
| 6081 | high speed and super speed interrupt |
| 6082 | endpoints, the USB 2.0 and USB 3.0 spec |
| 6083 | require the interval in microframes (1 |
| 6084 | microframe = 125 microseconds) to be |
| 6085 | calculated as interval = 2 ^ |
| 6086 | (bInterval-1). |
| 6087 | Devices with this quirk report their |
| 6088 | bInterval as the result of this |
| 6089 | calculation instead of the exponent |
| 6090 | variable used in the calculation); |
| 6091 | i = USB_QUIRK_DEVICE_QUALIFIER (device can't |
| 6092 | handle device_qualifier descriptor |
| 6093 | requests); |
| 6094 | j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device |
| 6095 | generates spurious wakeup, ignore |
| 6096 | remote wakeup capability); |
| 6097 | k = USB_QUIRK_NO_LPM (device can't handle Link |
| 6098 | Power Management); |
| 6099 | l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL |
| 6100 | (Device reports its bInterval as linear |
| 6101 | frames instead of the USB 2.0 |
| 6102 | calculation); |
| 6103 | m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs |
| 6104 | to be disconnected before suspend to |
| 6105 | prevent spurious wakeup); |
| 6106 | n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a |
| 6107 | pause after every control message); |
| 6108 | o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra |
| 6109 | delay after resetting its port); |
| 6110 | Example: quirks=0781:5580:bk,0a5c:5834:gij |
| 6111 | |
| 6112 | usbhid.mousepoll= |
| 6113 | [USBHID] The interval which mice are to be polled at. |
| 6114 | |
| 6115 | usbhid.jspoll= |
| 6116 | [USBHID] The interval which joysticks are to be polled at. |
| 6117 | |
| 6118 | usbhid.kbpoll= |
| 6119 | [USBHID] The interval which keyboards are to be polled at. |
| 6120 | |
| 6121 | usb-storage.delay_use= |
| 6122 | [UMS] The delay in seconds before a new device is |
| 6123 | scanned for Logical Units (default 1). |
| 6124 | |
| 6125 | usb-storage.quirks= |
| 6126 | [UMS] A list of quirks entries to supplement or |
| 6127 | override the built-in unusual_devs list. List |
| 6128 | entries are separated by commas. Each entry has |
| 6129 | the form VID:PID:Flags where VID and PID are Vendor |
| 6130 | and Product ID values (4-digit hex numbers) and |
| 6131 | Flags is a set of characters, each corresponding |
| 6132 | to a common usb-storage quirk flag as follows: |
| 6133 | a = SANE_SENSE (collect more than 18 bytes |
| 6134 | of sense data, not on uas); |
| 6135 | b = BAD_SENSE (don't collect more than 18 |
| 6136 | bytes of sense data, not on uas); |
| 6137 | c = FIX_CAPACITY (decrease the reported |
| 6138 | device capacity by one sector); |
| 6139 | d = NO_READ_DISC_INFO (don't use |
| 6140 | READ_DISC_INFO command, not on uas); |
| 6141 | e = NO_READ_CAPACITY_16 (don't use |
| 6142 | READ_CAPACITY_16 command); |
| 6143 | f = NO_REPORT_OPCODES (don't use report opcodes |
| 6144 | command, uas only); |
| 6145 | g = MAX_SECTORS_240 (don't transfer more than |
| 6146 | 240 sectors at a time, uas only); |
| 6147 | h = CAPACITY_HEURISTICS (decrease the |
| 6148 | reported device capacity by one |
| 6149 | sector if the number is odd); |
| 6150 | i = IGNORE_DEVICE (don't bind to this |
| 6151 | device); |
| 6152 | j = NO_REPORT_LUNS (don't use report luns |
| 6153 | command, uas only); |
| 6154 | k = NO_SAME (do not use WRITE_SAME, uas only) |
| 6155 | l = NOT_LOCKABLE (don't try to lock and |
| 6156 | unlock ejectable media, not on uas); |
| 6157 | m = MAX_SECTORS_64 (don't transfer more |
| 6158 | than 64 sectors = 32 KB at a time, |
| 6159 | not on uas); |
| 6160 | n = INITIAL_READ10 (force a retry of the |
| 6161 | initial READ(10) command, not on uas); |
| 6162 | o = CAPACITY_OK (accept the capacity |
| 6163 | reported by the device, not on uas); |
| 6164 | p = WRITE_CACHE (the device cache is ON |
| 6165 | by default, not on uas); |
| 6166 | r = IGNORE_RESIDUE (the device reports |
| 6167 | bogus residue values, not on uas); |
| 6168 | s = SINGLE_LUN (the device has only one |
| 6169 | Logical Unit); |
| 6170 | t = NO_ATA_1X (don't allow ATA(12) and ATA(16) |
| 6171 | commands, uas only); |
| 6172 | u = IGNORE_UAS (don't bind to the uas driver); |
| 6173 | w = NO_WP_DETECT (don't test whether the |
| 6174 | medium is write-protected). |
| 6175 | y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE |
| 6176 | even if the device claims no cache, |
| 6177 | not on uas) |
| 6178 | Example: quirks=0419:aaf5:rl,0421:0433:rc |
| 6179 | |
| 6180 | user_debug= [KNL,ARM] |
| 6181 | Format: <int> |
| 6182 | See arch/arm/Kconfig.debug help text. |
| 6183 | 1 - undefined instruction events |
| 6184 | 2 - system calls |
| 6185 | 4 - invalid data aborts |
| 6186 | 8 - SIGSEGV faults |
| 6187 | 16 - SIGBUS faults |
| 6188 | Example: user_debug=31 |
| 6189 | |
| 6190 | userpte= |
| 6191 | [X86] Flags controlling user PTE allocations. |
| 6192 | |
| 6193 | nohigh = do not allocate PTE pages in |
| 6194 | HIGHMEM regardless of setting |
| 6195 | of CONFIG_HIGHPTE. |
| 6196 | |
| 6197 | vdso= [X86,SH] |
| 6198 | On X86_32, this is an alias for vdso32=. Otherwise: |
| 6199 | |
| 6200 | vdso=1: enable VDSO (the default) |
| 6201 | vdso=0: disable VDSO mapping |
| 6202 | |
| 6203 | vdso32= [X86] Control the 32-bit vDSO |
| 6204 | vdso32=1: enable 32-bit VDSO |
| 6205 | vdso32=0 or vdso32=2: disable 32-bit VDSO |
| 6206 | |
| 6207 | See the help text for CONFIG_COMPAT_VDSO for more |
| 6208 | details. If CONFIG_COMPAT_VDSO is set, the default is |
| 6209 | vdso32=0; otherwise, the default is vdso32=1. |
| 6210 | |
| 6211 | For compatibility with older kernels, vdso32=2 is an |
| 6212 | alias for vdso32=0. |
| 6213 | |
| 6214 | Try vdso32=0 if you encounter an error that says: |
| 6215 | dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed! |
| 6216 | |
| 6217 | vector= [IA-64,SMP] |
| 6218 | vector=percpu: enable percpu vector domain |
| 6219 | |
| 6220 | video= [FB] Frame buffer configuration |
| 6221 | See Documentation/fb/modedb.rst. |
| 6222 | |
| 6223 | video.brightness_switch_enabled= [0,1] |
| 6224 | If set to 1, on receiving an ACPI notify event |
| 6225 | generated by hotkey, video driver will adjust brightness |
| 6226 | level and then send out the event to user space through |
| 6227 | the allocated input device; If set to 0, video driver |
| 6228 | will only send out the event without touching backlight |
| 6229 | brightness level. |
| 6230 | default: 1 |
| 6231 | |
| 6232 | virtio_mmio.device= |
| 6233 | [VMMIO] Memory mapped virtio (platform) device. |
| 6234 | |
| 6235 | <size>@<baseaddr>:<irq>[:<id>] |
| 6236 | where: |
| 6237 | <size> := size (can use standard suffixes |
| 6238 | like K, M and G) |
| 6239 | <baseaddr> := physical base address |
| 6240 | <irq> := interrupt number (as passed to |
| 6241 | request_irq()) |
| 6242 | <id> := (optional) platform device id |
| 6243 | example: |
| 6244 | virtio_mmio.device=1K@0x100b0000:48:7 |
| 6245 | |
| 6246 | Can be used multiple times for multiple devices. |
| 6247 | |
| 6248 | vga= [BOOT,X86-32] Select a particular video mode |
| 6249 | See Documentation/x86/boot.rst and |
| 6250 | Documentation/admin-guide/svga.rst. |
| 6251 | Use vga=ask for menu. |
| 6252 | This is actually a boot loader parameter; the value is |
| 6253 | passed to the kernel using a special protocol. |
| 6254 | |
| 6255 | vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y. |
| 6256 | May slow down system boot speed, especially when |
| 6257 | enabled on systems with a large amount of memory. |
| 6258 | All options are enabled by default, and this |
| 6259 | interface is meant to allow for selectively |
| 6260 | enabling or disabling specific virtual memory |
| 6261 | debugging features. |
| 6262 | |
| 6263 | Available options are: |
| 6264 | P Enable page structure init time poisoning |
| 6265 | - Disable all of the above options |
| 6266 | |
| 6267 | vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact |
| 6268 | size of <nn>. This can be used to increase the |
| 6269 | minimum size (128MB on x86). It can also be used to |
| 6270 | decrease the size and leave more room for directly |
| 6271 | mapped kernel RAM. |
| 6272 | |
| 6273 | vmcp_cma=nn[MG] [KNL,S390] |
| 6274 | Sets the memory size reserved for contiguous memory |
| 6275 | allocations for the vmcp device driver. |
| 6276 | |
| 6277 | vmhalt= [KNL,S390] Perform z/VM CP command after system halt. |
| 6278 | Format: <command> |
| 6279 | |
| 6280 | vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic. |
| 6281 | Format: <command> |
| 6282 | |
| 6283 | vmpoff= [KNL,S390] Perform z/VM CP command after power off. |
| 6284 | Format: <command> |
| 6285 | |
| 6286 | vsyscall= [X86-64] |
| 6287 | Controls the behavior of vsyscalls (i.e. calls to |
| 6288 | fixed addresses of 0xffffffffff600x00 from legacy |
| 6289 | code). Most statically-linked binaries and older |
| 6290 | versions of glibc use these calls. Because these |
| 6291 | functions are at fixed addresses, they make nice |
| 6292 | targets for exploits that can control RIP. |
| 6293 | |
| 6294 | emulate [default] Vsyscalls turn into traps and are |
| 6295 | emulated reasonably safely. The vsyscall |
| 6296 | page is readable. |
| 6297 | |
| 6298 | xonly Vsyscalls turn into traps and are |
| 6299 | emulated reasonably safely. The vsyscall |
| 6300 | page is not readable. |
| 6301 | |
| 6302 | none Vsyscalls don't work at all. This makes |
| 6303 | them quite hard to use for exploits but |
| 6304 | might break your system. |
| 6305 | |
| 6306 | vt.color= [VT] Default text color. |
| 6307 | Format: 0xYX, X = foreground, Y = background. |
| 6308 | Default: 0x07 = light gray on black. |
| 6309 | |
| 6310 | vt.cur_default= [VT] Default cursor shape. |
| 6311 | Format: 0xCCBBAA, where AA, BB, and CC are the same as |
| 6312 | the parameters of the <Esc>[?A;B;Cc escape sequence; |
| 6313 | see VGA-softcursor.txt. Default: 2 = underline. |
| 6314 | |
| 6315 | vt.default_blu= [VT] |
| 6316 | Format: <blue0>,<blue1>,<blue2>,...,<blue15> |
| 6317 | Change the default blue palette of the console. |
| 6318 | This is a 16-member array composed of values |
| 6319 | ranging from 0-255. |
| 6320 | |
| 6321 | vt.default_grn= [VT] |
| 6322 | Format: <green0>,<green1>,<green2>,...,<green15> |
| 6323 | Change the default green palette of the console. |
| 6324 | This is a 16-member array composed of values |
| 6325 | ranging from 0-255. |
| 6326 | |
| 6327 | vt.default_red= [VT] |
| 6328 | Format: <red0>,<red1>,<red2>,...,<red15> |
| 6329 | Change the default red palette of the console. |
| 6330 | This is a 16-member array composed of values |
| 6331 | ranging from 0-255. |
| 6332 | |
| 6333 | vt.default_utf8= |
| 6334 | [VT] |
| 6335 | Format=<0|1> |
| 6336 | Set system-wide default UTF-8 mode for all tty's. |
| 6337 | Default is 1, i.e. UTF-8 mode is enabled for all |
| 6338 | newly opened terminals. |
| 6339 | |
| 6340 | vt.global_cursor_default= |
| 6341 | [VT] |
| 6342 | Format=<-1|0|1> |
| 6343 | Set system-wide default for whether a cursor |
| 6344 | is shown on new VTs. Default is -1, |
| 6345 | i.e. cursors will be created by default unless |
| 6346 | overridden by individual drivers. 0 will hide |
| 6347 | cursors, 1 will display them. |
| 6348 | |
| 6349 | vt.italic= [VT] Default color for italic text; 0-15. |
| 6350 | Default: 2 = green. |
| 6351 | |
| 6352 | vt.underline= [VT] Default color for underlined text; 0-15. |
| 6353 | Default: 3 = cyan. |
| 6354 | |
| 6355 | watchdog timers [HW,WDT] For information on watchdog timers, |
| 6356 | see Documentation/watchdog/watchdog-parameters.rst |
| 6357 | or other driver-specific files in the |
| 6358 | Documentation/watchdog/ directory. |
| 6359 | |
| 6360 | watchdog_thresh= |
| 6361 | [KNL] |
| 6362 | Set the hard lockup detector stall duration |
| 6363 | threshold in seconds. The soft lockup detector |
| 6364 | threshold is set to twice the value. A value of 0 |
| 6365 | disables both lockup detectors. Default is 10 |
| 6366 | seconds. |
| 6367 | |
| 6368 | workqueue.watchdog_thresh= |
| 6369 | If CONFIG_WQ_WATCHDOG is configured, workqueue can |
| 6370 | warn stall conditions and dump internal state to |
| 6371 | help debugging. 0 disables workqueue stall |
| 6372 | detection; otherwise, it's the stall threshold |
| 6373 | duration in seconds. The default value is 30 and |
| 6374 | it can be updated at runtime by writing to the |
| 6375 | corresponding sysfs file. |
| 6376 | |
| 6377 | workqueue.disable_numa |
| 6378 | By default, all work items queued to unbound |
| 6379 | workqueues are affine to the NUMA nodes they're |
| 6380 | issued on, which results in better behavior in |
| 6381 | general. If NUMA affinity needs to be disabled for |
| 6382 | whatever reason, this option can be used. Note |
| 6383 | that this also can be controlled per-workqueue for |
| 6384 | workqueues visible under /sys/bus/workqueue/. |
| 6385 | |
| 6386 | workqueue.power_efficient |
| 6387 | Per-cpu workqueues are generally preferred because |
| 6388 | they show better performance thanks to cache |
| 6389 | locality; unfortunately, per-cpu workqueues tend to |
| 6390 | be more power hungry than unbound workqueues. |
| 6391 | |
| 6392 | Enabling this makes the per-cpu workqueues which |
| 6393 | were observed to contribute significantly to power |
| 6394 | consumption unbound, leading to measurably lower |
| 6395 | power usage at the cost of small performance |
| 6396 | overhead. |
| 6397 | |
| 6398 | The default value of this parameter is determined by |
| 6399 | the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT. |
| 6400 | |
| 6401 | workqueue.debug_force_rr_cpu |
| 6402 | Workqueue used to implicitly guarantee that work |
| 6403 | items queued without explicit CPU specified are put |
| 6404 | on the local CPU. This guarantee is no longer true |
| 6405 | and while local CPU is still preferred work items |
| 6406 | may be put on foreign CPUs. This debug option |
| 6407 | forces round-robin CPU selection to flush out |
| 6408 | usages which depend on the now broken guarantee. |
| 6409 | When enabled, memory and cache locality will be |
| 6410 | impacted. |
| 6411 | |
| 6412 | x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of |
| 6413 | default x2apic cluster mode on platforms |
| 6414 | supporting x2apic. |
| 6415 | |
| 6416 | xen_512gb_limit [KNL,X86-64,XEN] |
| 6417 | Restricts the kernel running paravirtualized under Xen |
| 6418 | to use only up to 512 GB of RAM. The reason to do so is |
| 6419 | crash analysis tools and Xen tools for doing domain |
| 6420 | save/restore/migration must be enabled to handle larger |
| 6421 | domains. |
| 6422 | |
| 6423 | xen_emul_unplug= [HW,X86,XEN] |
| 6424 | Unplug Xen emulated devices |
| 6425 | Format: [unplug0,][unplug1] |
| 6426 | ide-disks -- unplug primary master IDE devices |
| 6427 | aux-ide-disks -- unplug non-primary-master IDE devices |
| 6428 | nics -- unplug network devices |
| 6429 | all -- unplug all emulated devices (NICs and IDE disks) |
| 6430 | unnecessary -- unplugging emulated devices is |
| 6431 | unnecessary even if the host did not respond to |
| 6432 | the unplug protocol |
| 6433 | never -- do not unplug even if version check succeeds |
| 6434 | |
| 6435 | xen_legacy_crash [X86,XEN] |
| 6436 | Crash from Xen panic notifier, without executing late |
| 6437 | panic() code such as dumping handler. |
| 6438 | |
| 6439 | xen_nopvspin [X86,XEN] |
| 6440 | Disables the qspinlock slowpath using Xen PV optimizations. |
| 6441 | This parameter is obsoleted by "nopvspin" parameter, which |
| 6442 | has equivalent effect for XEN platform. |
| 6443 | |
| 6444 | xen_nopv [X86] |
| 6445 | Disables the PV optimizations forcing the HVM guest to |
| 6446 | run as generic HVM guest with no PV drivers. |
| 6447 | This option is obsoleted by the "nopv" option, which |
| 6448 | has equivalent effect for XEN platform. |
| 6449 | |
| 6450 | xen_no_vector_callback |
| 6451 | [KNL,X86,XEN] Disable the vector callback for Xen |
| 6452 | event channel interrupts. |
| 6453 | |
| 6454 | xen_scrub_pages= [XEN] |
| 6455 | Boolean option to control scrubbing pages before giving them back |
| 6456 | to Xen, for use by other domains. Can be also changed at runtime |
| 6457 | with /sys/devices/system/xen_memory/xen_memory0/scrub_pages. |
| 6458 | Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT. |
| 6459 | |
| 6460 | xen_timer_slop= [X86-64,XEN] |
| 6461 | Set the timer slop (in nanoseconds) for the virtual Xen |
| 6462 | timers (default is 100000). This adjusts the minimum |
| 6463 | delta of virtualized Xen timers, where lower values |
| 6464 | improve timer resolution at the expense of processing |
| 6465 | more timer interrupts. |
| 6466 | |
| 6467 | xen.balloon_boot_timeout= [XEN] |
| 6468 | The time (in seconds) to wait before giving up to boot |
| 6469 | in case initial ballooning fails to free enough memory. |
| 6470 | Applies only when running as HVM or PVH guest and |
| 6471 | started with less memory configured than allowed at |
| 6472 | max. Default is 180. |
| 6473 | |
| 6474 | xen.event_eoi_delay= [XEN] |
| 6475 | How long to delay EOI handling in case of event |
| 6476 | storms (jiffies). Default is 10. |
| 6477 | |
| 6478 | xen.event_loop_timeout= [XEN] |
| 6479 | After which time (jiffies) the event handling loop |
| 6480 | should start to delay EOI handling. Default is 2. |
| 6481 | |
| 6482 | xen.fifo_events= [XEN] |
| 6483 | Boolean parameter to disable using fifo event handling |
| 6484 | even if available. Normally fifo event handling is |
| 6485 | preferred over the 2-level event handling, as it is |
| 6486 | fairer and the number of possible event channels is |
| 6487 | much higher. Default is on (use fifo events). |
| 6488 | |
| 6489 | nopv= [X86,XEN,KVM,HYPER_V,VMWARE] |
| 6490 | Disables the PV optimizations forcing the guest to run |
| 6491 | as generic guest with no PV drivers. Currently support |
| 6492 | XEN HVM, KVM, HYPER_V and VMWARE guest. |
| 6493 | |
| 6494 | nopvspin [X86,XEN,KVM] |
| 6495 | Disables the qspinlock slow path using PV optimizations |
| 6496 | which allow the hypervisor to 'idle' the guest on lock |
| 6497 | contention. |
| 6498 | |
| 6499 | xirc2ps_cs= [NET,PCMCIA] |
| 6500 | Format: |
| 6501 | <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]] |
| 6502 | |
| 6503 | xive= [PPC] |
| 6504 | By default on POWER9 and above, the kernel will |
| 6505 | natively use the XIVE interrupt controller. This option |
| 6506 | allows the fallback firmware mode to be used: |
| 6507 | |
| 6508 | off Fallback to firmware control of XIVE interrupt |
| 6509 | controller on both pseries and powernv |
| 6510 | platforms. Only useful on POWER9 and above. |
| 6511 | |
| 6512 | xive.store-eoi=off [PPC] |
| 6513 | By default on POWER10 and above, the kernel will use |
| 6514 | stores for EOI handling when the XIVE interrupt mode |
| 6515 | is active. This option allows the XIVE driver to use |
| 6516 | loads instead, as on POWER9. |
| 6517 | |
| 6518 | xhci-hcd.quirks [USB,KNL] |
| 6519 | A hex value specifying bitmask with supplemental xhci |
| 6520 | host controller quirks. Meaning of each bit can be |
| 6521 | consulted in header drivers/usb/host/xhci.h. |
| 6522 | |
| 6523 | xmon [PPC] |
| 6524 | Format: { early | on | rw | ro | off } |
| 6525 | Controls if xmon debugger is enabled. Default is off. |
| 6526 | Passing only "xmon" is equivalent to "xmon=early". |
| 6527 | early Call xmon as early as possible on boot; xmon |
| 6528 | debugger is called from setup_arch(). |
| 6529 | on xmon debugger hooks will be installed so xmon |
| 6530 | is only called on a kernel crash. Default mode, |
| 6531 | i.e. either "ro" or "rw" mode, is controlled |
| 6532 | with CONFIG_XMON_DEFAULT_RO_MODE. |
| 6533 | rw xmon debugger hooks will be installed so xmon |
| 6534 | is called only on a kernel crash, mode is write, |
| 6535 | meaning SPR registers, memory and, other data |
| 6536 | can be written using xmon commands. |
| 6537 | ro same as "rw" option above but SPR registers, |
| 6538 | memory, and other data can't be written using |
| 6539 | xmon commands. |
| 6540 | off xmon is disabled. |