Jeff King wrote:
> This _might_ still suffer from the issue fixed in 5f6a112 (block-sha1:
> avoid pointer conversion that violates alignment constraints,
> 2012-07-22), as we are taking the pointer of a uint32 in a struct.
No conversion, so no issue there.
Line 1484 looks more problematic:
disk_ce = (struct ondisk_cache_entry *)((char *)mmap +
In v4 indexes, src_offset doesn't have any particular alignment so
this conversion has undefined behavior.
Do you know if any tests exercise this code with paths that don't
have convenient length?
> I'm inclined to leave it for now, as we haven't made anything worse, and
> nobody has reported a problem.
Probably the simplest fix would be to take a char *, memcpy into a
new (aligned) buffer and then byteswap in place, but that's
orthogonal to this series.
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html