Since commit 511631011d39706ac81ee5e4c9084d61e5b4fd34, entitled
"x86: fix pointer type warning in arch/x86/mm/init_64.c:early_memtest", the
following warning is produced for x86_64:
arch/x86/mm/init_64.c: In function ‘early_memtest’:
arch/x86/mm/init_64.c:520: warning: passing argument 2 of
‘find_e820_area_size’ from incompatible pointer type
Before this commit, the compilation of this routine on my system was clean. I
use gcc version 4.2.1. Was this the warning for i386 before the patch? I
tried simple changes in the definition of find_e820_area_size, but only
managed to push the warning on to the call to bad_addr_size. At that point I
decided to let the experts find a proper fix. Note that the warning does no
harm - I just get nervous about incompatible pointer warnings as it usually
means that I have screwed something up.
Larry
Larry Finger wrote:
> Since commit 511631011d39706ac81ee5e4c9084d61e5b4fd34, entitled
> "x86: fix pointer type warning in arch/x86/mm/init_64.c:early_memtest", the
> following warning is produced for x86_64:
>
> arch/x86/mm/init_64.c: In function ‘early_memtest’:
> arch/x86/mm/init_64.c:520: warning: passing argument 2 of
> ‘find_e820_area_size’ from incompatible pointer type
>
> Before this commit, the compilation of this routine on my system was clean. I
> use gcc version 4.2.1. Was this the warning for i386 before the patch? I
> tried simple changes in the definition of find_e820_area_size, but only
> managed to push the warning on to the call to bad_addr_size. At that point I
> decided to let the experts find a proper fix. Note that the warning does no
> harm - I just get nervous about incompatible pointer warnings as it usually
> means that I have screwed something up.
>
The patch was against linux-next, which seems to have changed
find_e820_area_size() - against mainline it introduced the warning
instead of fixing it.
--
Kevin Winchester