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: NChris Bainbridge <chris.bainbridge@gmail.com>
Signed-off-by: NBorislav Petkov <bp@suse.de>
Cc: hmh@hmh.eng.br
Link: http://lkml.kernel.org/r/1456834359-5132-1-git-send-email-chris.bainbridge@gmail.comSigned-off-by: NThomas Gleixner <tglx@linutronix.de>