Re: [tip:x86/x32] x32: Fix alignment fail in struct compat_siginfo

[H.J. Lu]

On Wed, Mar 14, 2012 at 2:43 PM, tip-bot for H. Peter Anvin
<hpa@xxxxxxxxx> wrote:
> Commit-ID:  31796ac4e8f0e88f5c10f1ad6dab8f19bebe44a4
> Gitweb:     http://git.kernel.org/tip/31796ac4e8f0e88f5c10f1ad6dab8f19bebe44a4
> Author:     H. Peter Anvin <hpa@xxxxxxxxx>
> AuthorDate: Wed, 14 Mar 2012 14:27:52 -0700
> Committer:  H. Peter Anvin <hpa@xxxxxxxxx>
> CommitDate: Wed, 14 Mar 2012 14:27:52 -0700
>
> x32: Fix alignment fail in struct compat_siginfo
>
> Adding struct _sigchld_x32 caused a misalignment cascade in struct
> siginfo, because union _sifields is located on an 4-byte boundary
> (8-byte misaligned.)
>
> Adding new fields that are 8-byte aligned caused the intermediate
> structures to also be aligned to 8 bytes, thereby adding padding in
> unexpected places.
>
> Thus, change s64 to compat_s64 here, which makes it "misaligned on
> paper".  In reality these fields *are* actually aligned (there are 3
> preceeding ints outside the union and 3 inside struct _sigchld_x32),
> but because of the intervening union and struct it is not possible for
> gcc to avoid padding without breaking the ABI.
>

Just for the record, although the offsets of _utime and _stime are
multiple of 8 bytes, struct siginfo is only aligned at 4 bytes.  So
the addresses of  _utime and _stime may not be 8byte aligned.
But misaligned int64 load/store are OK for x86-64.

-- 
H.J.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/