1. 20 Oct, 2010 1 commit
  2. 25 Aug, 2010 1 commit
  3. 18 Aug, 2010 1 commit
    • Joerg Roedel's avatar
      x86-32: Separate 1:1 pagetables from swapper_pg_dir · fd89a137
      Joerg Roedel authored
      
      
      This patch fixes machine crashes which occur when heavily exercising the
      CPU hotplug codepaths on a 32-bit kernel. These crashes are caused by
      AMD Erratum 383 and result in a fatal machine check exception. Here's
      the scenario:
      
      1. On 32-bit, the swapper_pg_dir page table is used as the initial page
      table for booting a secondary CPU.
      
      2. To make this work, swapper_pg_dir needs a direct mapping of physical
      memory in it (the low mappings). By adding those low, large page (2M)
      mappings (PAE kernel), we create the necessary conditions for Erratum
      383 to occur.
      
      3. Other CPUs which do not participate in the off- and onlining game may
      use swapper_pg_dir while the low mappings are present (when leave_mm is
      called). For all steps below, the CPU referred to is a CPU that is using
      swapper_pg_dir, and not the CPU which is being onlined.
      
      4. The presence of the low mappings in swapper_pg_dir can result
      in TLB entries for addresses below __PAGE_OFFSET to be established
      speculatively. These TLB entries are marked global and large.
      
      5. When the CPU with such TLB entry switches to another page table, this
      TLB entry remains because it is global.
      
      6. The process then generates an access to an address covered by the
      above TLB entry but there is a permission mismatch - the TLB entry
      covers a large global page not accessible to userspace.
      
      7. Due to this permission mismatch a new 4kb, user TLB entry gets
      established. Further, Erratum 383 provides for a small window of time
      where both TLB entries are present. This results in an uncorrectable
      machine check exception signalling a TLB multimatch which panics the
      machine.
      
      There are two ways to fix this issue:
      
              1. Always do a global TLB flush when a new cr3 is loaded and the
              old page table was swapper_pg_dir. I consider this a hack hard
              to understand and with performance implications
      
              2. Do not use swapper_pg_dir to boot secondary CPUs like 64-bit
              does.
      
      This patch implements solution 2. It introduces a trampoline_pg_dir
      which has the same layout as swapper_pg_dir with low_mappings. This page
      table is used as the initial page table of the booting CPU. Later in the
      bringup process, it switches to swapper_pg_dir and does a global TLB
      flush. This fixes the crashes in our test cases.
      
      -v2: switch to swapper_pg_dir right after entering start_secondary() so
      that we are able to access percpu data which might not be mapped in the
      trampoline page table.
      Signed-off-by: default avatarJoerg Roedel <joerg.roedel@amd.com>
      LKML-Reference: <20100816123833.GB28147@aftab>
      Signed-off-by: default avatarBorislav Petkov <borislav.petkov@amd.com>
      Signed-off-by: default avatarH. Peter Anvin <hpa@zytor.com>
      fd89a137
  4. 11 Dec, 2009 1 commit
    • Yinghai Lu's avatar
      x86: Use find_e820() instead of hard coded trampoline address · 893f38d1
      Yinghai Lu authored
      Jens found the following crash/regression:
      
      [    0.000000] found SMP MP-table at [ffff8800000fdd80] fdd80
      [    0.000000] Kernel panic - not syncing: Overlapping early reservations 12-f011 MP-table mpc to 0-fff BIOS data page
      
      and
      
      [    0.000000] Kernel panic - not syncing: Overlapping early reservations 12-f011 MP-table mpc to 6000-7fff TRAMPOLINE
      
      and bisected it to b24c2a92
      
       ("x86: Move find_smp_config()
      earlier and avoid bootmem usage").
      
      It turns out the BIOS is using the first 64k for mptable,
      without reserving it.
      
      So try to find good range for the real-mode trampoline instead of
      hard coding it, in case some bios tries to use that range for sth.
      Reported-by: default avatarJens Axboe <jens.axboe@oracle.com>
      Signed-off-by: default avatarYinghai Lu <yinghai@kernel.org>
      Tested-by: default avatarJens Axboe <jens.axboe@oracle.com>
      Cc: Randy Dunlap <randy.dunlap@oracle.com>
      LKML-Reference: <4B21630A.6000308@kernel.org>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      893f38d1
  5. 12 Oct, 2009 1 commit
  6. 20 Sep, 2009 1 commit
  7. 08 Dec, 2008 1 commit
    • Rafael J. Wysocki's avatar
      x86: change static allocation of trampoline area · 3e1e9002
      Rafael J. Wysocki authored
      
      
      Impact: fix trampoline sizing bug, save space
      
      While debugging a suspend-to-RAM related issue it occured to me that
      if the trampoline code had grown past 4 KB, we would have been
      allocating too little memory for it, since the 4 KB size of the
      trampoline is hardcoded into arch/x86/kernel/e820.c .  Change that
      by making the kernel compute the trampoline size and allocate as much
      memory as necessary.
      Signed-off-by: default avatarRafael J. Wysocki <rjw@sisk.pl>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      3e1e9002
  8. 25 May, 2008 1 commit
    • Yinghai Lu's avatar
      x86: extend e820 ealy_res support 32bit · a4c81cf6
      Yinghai Lu authored
      
      
      move early_res related from e820_64.c to e820.c
      make edba detection to be done in head32.c
      remove smp_alloc_memory, because we have fixed trampoline address now.
      Signed-off-by: default avatarYinghai Lu <yhlu.kernel@gmail.com>
      
       arch/x86/kernel/e820.c              |  214 ++++++++++++++++++++++++++++++++++++
       arch/x86/kernel/e820_64.c           |  196 --------------------------------
       arch/x86/kernel/head32.c            |   76 ++++++++++++
       arch/x86/kernel/setup_32.c          |  109 +++---------------
       arch/x86/kernel/smpboot.c           |   17 --
       arch/x86/kernel/trampoline.c        |    2
       arch/x86/mach-voyager/voyager_smp.c |    9 -
       include/asm-x86/e820.h              |    6 +
       include/asm-x86/e820_64.h           |    9 -
       include/asm-x86/smp.h               |    1
       arch/x86/kernel/e820.c              |  214 ++++++++++++++++++++++++++++++++++++
       arch/x86/kernel/e820_64.c           |  196 --------------------------------
       arch/x86/kernel/head32.c            |   76 ++++++++++++
       arch/x86/kernel/setup_32.c          |  109 +++---------------
       arch/x86/kernel/smpboot.c           |   17 --
       arch/x86/kernel/trampoline.c        |    2
       arch/x86/mach-voyager/voyager_smp.c |    9 -
       include/asm-x86/e820.h              |    6 +
       include/asm-x86/e820_64.h           |    9 -
       include/asm-x86/smp.h               |    1
       arch/x86/kernel/e820.c              |  214 ++++++++++++++++++++++++++++++++++++
       arch/x86/kernel/e820_64.c           |  196 --------------------------------
       arch/x86/kernel/head32.c            |   76 ++++++++++++
       arch/x86/kernel/setup_32.c          |  109 +++---------------
       arch/x86/kernel/smpboot.c           |   17 --
       arch/x86/kernel/trampoline.c        |    2
       arch/x86/mach-voyager/voyager_smp.c |    9 -
       include/asm-x86/e820.h              |    6 +
       include/asm-x86/e820_64.h           |    9 -
       include/asm-x86/smp.h               |    1
       10 files changed, 320 insertions(+), 319 deletions(-)
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      a4c81cf6
  9. 17 Apr, 2008 1 commit