Documentation/x86/x86_64/boot-options.txt

   1 AMD64 specific boot options
   2
   3 There are many others (usually documented in driver documentation), but
   4 only the AMD64 specific ones are listed here.
   5
   6 Machine check
   7
   8    mce=off disable machine check
   9    mce=bootlog Enable logging of machine checks left over from booting.
  10                Disabled by default on AMD because some BIOS leave bogus ones.
  11                If your BIOS doesn't do that it's a good idea to enable though
  12                to make sure you log even machine check events that result
  13                in a reboot. On Intel systems it is enabled by default.
  14    mce=nobootlog
  15                 Disable boot machine check logging.
  16    mce=tolerancelevel (number)
  17                 0: always panic on uncorrected errors, log corrected errors
  18                 1: panic or SIGBUS on uncorrected errors, log corrected errors
  19                 2: SIGBUS or log uncorrected errors, log corrected errors
  20                 3: never panic or SIGBUS, log all errors (for testing only)
  21                 Default is 1
  22                 Can be also set using sysfs which is preferable.
  23
  24    nomce (for compatibility with i386): same as mce=off
  25
  26    Everything else is in sysfs now.
  27
  28 APICs
  29
  30    apic          Use IO-APIC. Default
  31
  32    noapic        Don't use the IO-APIC.
  33
  34    disableapic   Don't use the local APIC
  35
  36    nolapic       Don't use the local APIC (alias for i386 compatibility)
  37
  38    pirq=...      See Documentation/x86/i386/IO-APIC.txt
  39
  40    noapictimer   Don't set up the APIC timer
  41
  42    no_timer_check Don't check the IO-APIC timer. This can work around
  43                  problems with incorrect timer initialization on some boards.
  44
  45    apicmaintimer Run time keeping from the local APIC timer instead
  46                  of using the PIT/HPET interrupt for this. This is useful
  47                  when the PIT/HPET interrupts are unreliable.
  48
  49    noapicmaintimer  Don't do time keeping using the APIC timer.
  50                  Useful when this option was auto selected, but doesn't work.
  51
  52    apicpmtimer
  53                  Do APIC timer calibration using the pmtimer. Implies
  54                  apicmaintimer. Useful when your PIT timer is totally
  55                  broken.
  56
  57 Early Console
  58
  59    syntax: earlyprintk=vga
  60            earlyprintk=serial[,ttySn[,baudrate]]
  61
  62    The early console is useful when the kernel crashes before the
  63    normal console is initialized. It is not enabled by
  64    default because it has some cosmetic problems.
  65    Append ,keep to not disable it when the real console takes over.
  66    Only vga or serial at a time, not both.
  67    Currently only ttyS0 and ttyS1 are supported.
  68    Interaction with the standard serial driver is not very good.
  69    The VGA output is eventually overwritten by the real console.
  70
  71 Timing
  72
  73   notsc
  74   Don't use the CPU time stamp counter to read the wall time.
  75   This can be used to work around timing problems on multiprocessor systems
  76   with not properly synchronized CPUs.
  77
  78   report_lost_ticks
  79   Report when timer interrupts are lost because some code turned off
  80   interrupts for too long.
  81
  82   nmi_watchdog=NUMBER[,panic]
  83   NUMBER can be:
  84   0 don't use an NMI watchdog
  85   1 use the IO-APIC timer for the NMI watchdog
  86   2 use the local APIC for the NMI watchdog using a performance counter. Note
  87   This will use one performance counter and the local APIC's performance
  88   vector.
  89   When panic is specified panic when an NMI watchdog timeout occurs.
  90   This is useful when you use a panic=... timeout and need the box
  91   quickly up again.
  92
  93   nohpet
  94   Don't use the HPET timer.
  95
  96 Idle loop
  97
  98   idle=poll
  99   Don't do power saving in the idle loop using HLT, but poll for rescheduling
 100   event. This will make the CPUs eat a lot more power, but may be useful
 101   to get slightly better performance in multiprocessor benchmarks. It also
 102   makes some profiling using performance counters more accurate.
 103   Please note that on systems with MONITOR/MWAIT support (like Intel EM64T
 104   CPUs) this option has no performance advantage over the normal idle loop.
 105   It may also interact badly with hyperthreading.
 106
 107 Rebooting
 108
 109    reboot=b[ios] | t[riple] | k[bd] | a[cpi] | e[fi] [, [w]arm | [c]old]
 110    bios   Use the CPU reboot vector for warm reset
 111    warm   Don't set the cold reboot flag
 112    cold   Set the cold reboot flag
 113    triple Force a triple fault (init)
 114    kbd    Use the keyboard controller. cold reset (default)
 115    acpi   Use the ACPI RESET_REG in the FADT. If ACPI is not configured or the
 116           ACPI reset does not work, the reboot path attempts the reset using
 117           the keyboard controller.
 118    efi    Use efi reset_system runtime service. If EFI is not configured or the
 119           EFI reset does not work, the reboot path attempts the reset using
 120           the keyboard controller.
 121
 122    Using warm reset will be much faster especially on big memory
 123    systems because the BIOS will not go through the memory check.
 124    Disadvantage is that not all hardware will be completely reinitialized
 125    on reboot so there may be boot problems on some systems.
 126
 127    reboot=force
 128
 129    Don't stop other CPUs on reboot. This can make reboot more reliable
 130    in some cases.
 131
 132 Non Executable Mappings
 133
 134   noexec=on|off
 135
 136   on      Enable(default)
 137   off     Disable
 138
 139 SMP
 140
 141   additional_cpus=NUM Allow NUM more CPUs for hotplug
 142                  (defaults are specified by the BIOS, see Documentation/x86/x86_64/cpu-hotplug-spec)
 143
 144 NUMA
 145
 146   numa=off      Only set up a single NUMA node spanning all memory.
 147
 148   numa=noacpi   Don't parse the SRAT table for NUMA setup
 149
 150   numa=fake=CMDLINE
 151                 If a number, fakes CMDLINE nodes and ignores NUMA setup of the
 152                 actual machine.  Otherwise, system memory is configured
 153                 depending on the sizes and coefficients listed.  For example:
 154                         numa=fake=2*512,1024,4*256,*128
 155                 gives two 512M nodes, a 1024M node, four 256M nodes, and the
 156                 rest split into 128M chunks.  If the last character of CMDLINE
 157                 is a *, the remaining memory is divided up equally among its
 158                 coefficient:
 159                         numa=fake=2*512,2*
 160                 gives two 512M nodes and the rest split into two nodes.
 161                 Otherwise, the remaining system RAM is allocated to an
 162                 additional node.
 163
 164   numa=hotadd=percent
 165                 Only allow hotadd memory to preallocate page structures upto
 166                 percent of already available memory.
 167                 numa=hotadd=0 will disable hotadd memory.
 168
 169 ACPI
 170
 171   acpi=off      Don't enable ACPI
 172   acpi=ht       Use ACPI boot table parsing, but don't enable ACPI
 173                 interpreter
 174   acpi=force    Force ACPI on (currently not needed)
 175
 176   acpi=strict   Disable out of spec ACPI workarounds.
 177
 178   acpi_sci={edge,level,high,low}  Set up ACPI SCI interrupt.
 179
 180   acpi=noirq    Don't route interrupts
 181
 182 PCI
 183
 184   pci=off       Don't use PCI
 185   pci=conf1     Use conf1 access.
 186   pci=conf2     Use conf2 access.
 187   pci=rom       Assign ROMs.
 188   pci=assign-busses    Assign busses
 189   pci=irqmask=MASK             Set PCI interrupt mask to MASK
 190   pci=lastbus=NUMBER           Scan upto NUMBER busses, no matter what the mptable says.
 191   pci=noacpi            Don't use ACPI to set up PCI interrupt routing.
 192
 193 IOMMU (input/output memory management unit)
 194
 195  Currently four x86-64 PCI-DMA mapping implementations exist:
 196
 197    1. <arch/x86_64/kernel/pci-nommu.c>: use no hardware/software IOMMU at all
 198       (e.g. because you have < 3 GB memory).
 199       Kernel boot message: "PCI-DMA: Disabling IOMMU"
 200
 201    2. <arch/x86_64/kernel/pci-gart.c>: AMD GART based hardware IOMMU.
 202       Kernel boot message: "PCI-DMA: using GART IOMMU"
 203
 204    3. <arch/x86_64/kernel/pci-swiotlb.c> : Software IOMMU implementation. Used
 205       e.g. if there is no hardware IOMMU in the system and it is need because
 206       you have >3GB memory or told the kernel to us it (iommu=soft))
 207       Kernel boot message: "PCI-DMA: Using software bounce buffering
 208       for IO (SWIOTLB)"
 209
 210    4. <arch/x86_64/pci-calgary.c> : IBM Calgary hardware IOMMU. Used in IBM
 211       pSeries and xSeries servers. This hardware IOMMU supports DMA address
 212       mapping with memory protection, etc.
 213       Kernel boot message: "PCI-DMA: Using Calgary IOMMU"
 214
 215  iommu=[<size>][,noagp][,off][,force][,noforce][,leak[=<nr_of_leak_pages>]
 216         [,memaper[=<order>]][,merge][,forcesac][,fullflush][,nomerge]
 217         [,noaperture][,calgary]
 218
 219   General iommu options:
 220     off                Don't initialize and use any kind of IOMMU.
 221     noforce            Don't force hardware IOMMU usage when it is not needed.
 222                        (default).
 223     force              Force the use of the hardware IOMMU even when it is
 224                        not actually needed (e.g. because < 3 GB memory).
 225     soft               Use software bounce buffering (SWIOTLB) (default for
 226                        Intel machines). This can be used to prevent the usage
 227                        of an available hardware IOMMU.
 228
 229   iommu options only relevant to the AMD GART hardware IOMMU:
 230     <size>             Set the size of the remapping area in bytes.
 231     allowed            Overwrite iommu off workarounds for specific chipsets.
 232     fullflush          Flush IOMMU on each allocation (default).
 233     nofullflush        Don't use IOMMU fullflush.
 234     leak               Turn on simple iommu leak tracing (only when
 235                        CONFIG_IOMMU_LEAK is on). Default number of leak pages
 236                        is 20.
 237     memaper[=<order>]  Allocate an own aperture over RAM with size 32MB<<order.
 238                        (default: order=1, i.e. 64MB)
 239     merge              Do scatter-gather (SG) merging. Implies "force"
 240                        (experimental).
 241     nomerge            Don't do scatter-gather (SG) merging.
 242     noaperture         Ask the IOMMU not to touch the aperture for AGP.
 243     forcesac           Force single-address cycle (SAC) mode for masks <40bits
 244                        (experimental).
 245     noagp              Don't initialize the AGP driver and use full aperture.
 246     allowdac           Allow double-address cycle (DAC) mode, i.e. DMA >4GB.
 247                        DAC is used with 32-bit PCI to push a 64-bit address in
 248                        two cycles. When off all DMA over >4GB is forced through
 249                        an IOMMU or software bounce buffering.
 250     nodac              Forbid DAC mode, i.e. DMA >4GB.
 251     panic              Always panic when IOMMU overflows.
 252     calgary            Use the Calgary IOMMU if it is available
 253
 254   iommu options only relevant to the software bounce buffering (SWIOTLB) IOMMU
 255   implementation:
 256     swiotlb=<pages>[,force]
 257     <pages>            Prereserve that many 128K pages for the software IO
 258                        bounce buffering.
 259     force              Force all IO through the software TLB.
 260
 261   Settings for the IBM Calgary hardware IOMMU currently found in IBM
 262   pSeries and xSeries machines:
 263
 264     calgary=[64k,128k,256k,512k,1M,2M,4M,8M]
 265     calgary=[translate_empty_slots]
 266     calgary=[disable=<PCI bus number>]
 267     panic              Always panic when IOMMU overflows
 268
 269     64k,...,8M - Set the size of each PCI slot's translation table
 270     when using the Calgary IOMMU. This is the size of the translation
 271     table itself in main memory. The smallest table, 64k, covers an IO
 272     space of 32MB; the largest, 8MB table, can cover an IO space of
 273     4GB. Normally the kernel will make the right choice by itself.
 274
 275     translate_empty_slots - Enable translation even on slots that have
 276     no devices attached to them, in case a device will be hotplugged
 277     in the future.
 278
 279     disable=<PCI bus number> - Disable translation on a given PHB. For
 280     example, the built-in graphics adapter resides on the first bridge
 281     (PCI bus number 0); if translation (isolation) is enabled on this
 282     bridge, X servers that access the hardware directly from user
 283     space might stop working. Use this option if you have devices that
 284     are accessed from userspace directly on some PCI host bridge.
 285
 286 Debugging
 287
 288   oops=panic    Always panic on oopses. Default is to just kill the process,
 289                 but there is a small probability of deadlocking the machine.
 290                 This will also cause panics on machine check exceptions.
 291                 Useful together with panic=30 to trigger a reboot.
 292
 293   kstack=N      Print N words from the kernel stack in oops dumps.
 294
 295   pagefaulttrace  Dump all page faults. Only useful for extreme debugging
 296                 and will create a lot of output.
 297
 298   call_trace=[old|both|newfallback|new]
 299                 old: use old inexact backtracer
 300                 new: use new exact dwarf2 unwinder
 301                 both: print entries from both
 302                 newfallback: use new unwinder but fall back to old if it gets
 303                         stuck (default)
 304
 305 Miscellaneous
 306
 307         nogbpages
 308                 Do not use GB pages for kernel direct mappings.
 309         gbpages
 310                 Use GB pages for kernel direct mappings.