linux-toradex.git/arch/x86/kernel/cpu/microcode, branch v4.6

Merge branch 'x86-microcode-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2016-03-15T17:39:22+00:00

Pull x86 microcode updates from Ingo Molnar:
 "The biggest change in this cycle was the separation of the microcode
  loading mechanism from the initrd code plus the support of built-in
  microcode images.

  There were also lots cleanups and general restructuring (by Borislav
  Petkov)"

* 'x86-microcode-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (24 commits)
  x86/microcode/intel: Drop orig_sum from ext signature checksum
  x86/microcode/intel: Improve microcode sanity-checking error messages
  x86/microcode/intel: Merge two consecutive if-statements
  x86/microcode/intel: Get rid of DWSIZE
  x86/microcode/intel: Change checksum variables to u32
  x86/microcode: Use kmemdup() rather than duplicating its implementation
  x86/microcode: Remove unnecessary paravirt_enabled check
  x86/microcode: Document builtin microcode loading method
  x86/microcode/AMD: Issue microcode updated message later
  x86/microcode/intel: Cleanup get_matching_model_microcode()
  x86/microcode/intel: Remove unused arg of get_matching_model_microcode()
  x86/microcode/intel: Rename mc_saved_in_initrd
  x86/microcode/intel: Use *wrmsrl variants
  x86/microcode/intel: Cleanup apply_microcode_intel()
  x86/microcode/intel: Move the BUG_ON up and turn it into WARN_ON
  x86/microcode/intel: Rename mc_intel variable to mc
  x86/microcode/intel: Rename mc_saved_count to num_saved
  x86/microcode/intel: Rename local variables of type struct mc_saved_data
  x86/microcode/AMD: Drop redundant printk prefix
  x86/microcode: Issue update message only once
  ...

x86/microcode/intel: Drop orig_sum from ext signature checksum

2016-03-08T08:08:45+00:00

It is 0 because for !0 values we would have exited already.

Signed-off-by: Borislav Petkov 
Link: http://lkml.kernel.org/r/1457345404-28884-6-git-send-email-bp@alien8.de
Signed-off-by: Thomas Gleixner

x86/microcode/intel: Improve microcode sanity-checking error messages

2016-03-08T08:08:45+00:00

Turn them into proper sentences. Add comments to microcode_sanity_check() to
explain what it does.

Signed-off-by: Borislav Petkov 
Link: http://lkml.kernel.org/r/1457345404-28884-5-git-send-email-bp@alien8.de
Signed-off-by: Thomas Gleixner

x86/microcode/intel: Merge two consecutive if-statements

2016-03-08T08:08:45+00:00

Merge the two consecutive "if (ext_table_size)". No functional change.

Signed-off-by: Borislav Petkov 
Link: http://lkml.kernel.org/r/1457345404-28884-4-git-send-email-bp@alien8.de
Signed-off-by: Thomas Gleixner

x86/microcode/intel: Get rid of DWSIZE

2016-03-08T08:08:44+00:00

sizeof(u32) is perfectly clear as it is.

Signed-off-by: Borislav Petkov 
Link: http://lkml.kernel.org/r/1457345404-28884-3-git-send-email-bp@alien8.de
Signed-off-by: Thomas Gleixner

x86/microcode/intel: Change checksum variables to u32

2016-03-08T08:08:44+00:00

Microcode checksum verification should be done using unsigned 32-bit
values otherwise the calculation overflow results in undefined
behaviour.

This is also nicely documented in the SDM, section "Microcode Update
Checksum":

  "To check for a corrupt microcode update, software must perform a
  unsigned DWORD (32-bit) checksum of the microcode update. Even though
  some fields are signed, the checksum procedure treats all DWORDs as
  unsigned. Microcode updates with a header version equal to 00000001H
  must sum all DWORDs that comprise the microcode update. A valid
  checksum check will yield a value of 00000000H."

but for some reason the code has been using ints from the very
beginning.

In practice, this bug possibly manifested itself only when doing the
microcode data checksum - apparently, currently shipped Intel microcode
doesn't have an extended signature table for which we do checksum
verification too.

  UBSAN: Undefined behaviour in arch/x86/kernel/cpu/microcode/intel_lib.c:105:12
  signed integer overflow:
  -1500151068 + -2125470173 cannot be represented in type 'int'
  CPU: 0 PID: 0 Comm: swapper Not tainted 4.5.0-rc5+ #495
  ...
  Call Trace:
   dump_stack
   ? inotify_ioctl
   ubsan_epilogue
   handle_overflow
   __ubsan_handle_add_overflow
   microcode_sanity_check
   get_matching_model_microcode.isra.2.constprop.8
   ? early_idt_handler_common
   ? strlcpy
   ? find_cpio_data
   load_ucode_intel_bsp
   load_ucode_bsp
   ? load_ucode_bsp
   x86_64_start_kernel

[ Expand and massage commit message. ]
Signed-off-by: Chris Bainbridge 
Signed-off-by: Borislav Petkov 
Cc: hmh@hmh.eng.br
Link: http://lkml.kernel.org/r/1456834359-5132-1-git-send-email-chris.bainbridge@gmail.com
Signed-off-by: Thomas Gleixner

x86/microcode: Use kmemdup() rather than duplicating its implementation

2016-02-17T07:46:08+00:00

The patch was generated using fixed coccinelle semantic patch
scripts/coccinelle/api/memdup.cocci.

Signed-off-by: Andrzej Hajda 
Signed-off-by: Borislav Petkov 
Cc: Bartlomiej Zolnierkiewicz 
Cc: Linus Torvalds 
Cc: Marek Szyprowski 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: http://lkml.kernel.org/r/1455612202-14414-3-git-send-email-bp@alien8.de
Signed-off-by: Ingo Molnar

x86/microcode: Remove unnecessary paravirt_enabled check

2016-02-17T07:46:07+00:00

Commit:

  a18a0f6850d4 ("x86, microcode: Don't initialize microcode code on paravirt")

added a paravirt test in microcode_init(), primarily to avoid making
mc_bp_resume()->load_ucode_ap()->check_loader_disabled_ap() calls
because on 32-bit kernels this callchain ends up using __pa_nodebug()
macro which is invalid for Xen PV guests.

A subsequent commit:

  fbae4ba8c4a3 ("x86, microcode: Reload microcode on resume")

eliminated this callchain thus making a18a0f6850d4 unnecessary.

Signed-off-by: Boris Ostrovsky 
Signed-off-by: Borislav Petkov 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: david.vrabel@citrix.com
Cc: konrad.wilk@oracle.com
Cc: xen-devel@lists.xenproject.org
Link: http://lkml.kernel.org/r/1455612202-14414-2-git-send-email-bp@alien8.de
Signed-off-by: Ingo Molnar

x86/microcode/AMD: Issue microcode updated message later

2016-02-09T10:41:18+00:00

Before this, we issued this message from save_microcode_in_initrd()
which is called from free_initrd_mem(), i.e., only when we have an
initrd enabled. However, we can update from builtin microcode too but
then we don't issue the update message.

Fix it by issuing that message on the generic driver init path.

Tested-by: Thomas Voegtle 
Signed-off-by: Borislav Petkov 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: http://lkml.kernel.org/r/1454499225-21544-17-git-send-email-bp@alien8.de
Signed-off-by: Ingo Molnar

x86/microcode/intel: Cleanup get_matching_model_microcode()

2016-02-09T10:41:18+00:00

Reflow arguments, sort local variables in reverse christmas tree, kill
"out" label.

No functionality change.

Tested-by: Thomas Voegtle 
Signed-off-by: Borislav Petkov 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: http://lkml.kernel.org/r/1454499225-21544-16-git-send-email-bp@alien8.de
Signed-off-by: Ingo Molnar