On 02/16/2014 01:07 PM, Patrick McHardy wrote: > On Sun, Feb 16, 2014 at 12:54:40PM +0100, Nikolay Aleksandrov wrote: >> Add a check if payload's offset is a power of 2 when selecting ops. >> The fast ops were meant for well aligned offsets, also this fixes a >> small bug when using unaligned offset and length of 3 which causes >> only 1 byte to be loaded because the fast ops are chosen. >> >> Signed-off-by: Nikolay Aleksandrov <nikolay@xxxxxxxxxx> >> --- >> I broke the line since it was >80 chars. >> This patch applies to Dave's -net tree. >> >> net/netfilter/nft_payload.c | 4 +++- >> 1 file changed, 3 insertions(+), 1 deletion(-) >> >> diff --git a/net/netfilter/nft_payload.c b/net/netfilter/nft_payload.c >> index a2aeb31..15a3dcb 100644 >> --- a/net/netfilter/nft_payload.c >> +++ b/net/netfilter/nft_payload.c >> @@ -135,7 +135,9 @@ nft_payload_select_ops(const struct nft_ctx *ctx, >> if (len == 0 || len > FIELD_SIZEOF(struct nft_data, data)) >> return ERR_PTR(-EINVAL); >> >> - if (len <= 4 && IS_ALIGNED(offset, len) && base != NFT_PAYLOAD_LL_HEADER) >> + if (len <= 4 && IS_ALIGNED(offset, len) && >> + base != NFT_PAYLOAD_LL_HEADER && >> + !(offset & (offset - 1))) > > Small cosmetic note: we have is_power_of_2(), which does exactly this. > I didn't choose is_power_of_2 because it also checks if x is != 0, so I thought we can allow for 0 offset with the fast ops. > Also I think the more logical order would be: > > len <= 4 && is_power_of_2 && IS_ALIGNED && base ... > Okay :-) >> return &nft_payload_fast_ops; >> else >> return &nft_payload_ops; >> -- >> 1.8.4.2 -- To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
