Re: [PATCH V4 2/3] Revert "x86, efivars: firmware bug workaroundsshould be in platform code"

[Matthew Garrett]

On Thu, 2013-04-11 at 14:24 +0100, Matt Fleming wrote:
> On 10/04/13 18:46, Matthew Garrett wrote:
> > This reverts commit a6e4d5a03e9e3587e88aba687d8f225f4f04c792. Doing this
> > workaround properly requires us to work within the variable code.
> > 
> > Signed-off-by: Matthew Garrett <matthew.garrett@xxxxxxxxxx>
> > ---
> >  arch/x86/platform/efi/efi.c | 25 -------------------------
> >  drivers/firmware/efivars.c  | 18 +++++++++++++++---
> >  include/linux/efi.h         |  9 +--------
> >  3 files changed, 16 insertions(+), 36 deletions(-)
> 
> Does it really? Why can't you just hook into the get_next_variable() and
> set_variable() functions in arch/x86/platform/efi/efi.c?

struct efi_variable isn't exported from efivars.c, and chunks of the
workaround need to be called at efivars init anyway.

-- 
Matthew Garrett | mjg59@xxxxxxxxxxxxx
èº{.nÇ+‰·Ÿ®‰­†+%ŠËlzwm…ébëæìr¸›zX§»®w¥Š{ayºÊÚë,j­¢f£¢·hš‹àz¹®w¥¢¸¢·¦j:+v‰¨ŠwèjØm¶Ÿÿ¾«‘êçzZ+ƒùšŽŠÝj"ú!¶iO•æ¬z·švØ^¶m§ÿðÃnÆàþY&—