Re: scatterlist.c: bug in sg_next()?

[Boaz Harrosh]

Leon Woestenberg wrote:
> Hello,
> 
> I was code-inspecting 2.6.27-r7 through git web, when I came across this:
> 
> In sg_next(), after following a chain_ptr, a few more checks should be
> performed.
> The rare case exists that the first entry in the chained list is a
> last marker, in case NULL must be returned.
> 
> Can someone confirm and cook a patch?
> 
> struct scatterlist *sg_next(struct scatterlist *sg)
> {
>   if (sg_is_last(sg))
>     return NULL;
>   sg++;
>   if (unlikely(sg_is_chain(sg))) {
>      sg = sg_chain_ptr(sg);
> +#ifdef CONFIG_DEBUG_SG
> + BUG_ON(sg->sg_magic != SG_MAGIC);
> +#endif
> +   if (sg_is_last(sg))
> +     return NULL;
>   }
>   return sg;
> }
> 
> Signed-off-by: Leon Woestenberg <leon@xxxxxxxxxxxxxx>
> 
> Regards,

No! the last marker is set on a valid sg entry. Only it's next
is no longer valid. So the check at the top is for the Next-sg
not the passed-in-sg. What you thought of is a NULL terminating
sg-list. The end marker is so to save the extra NULL entry.

Boaz

--
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/