| Message ID | 483d9eec159b22172fe04dacd58d7f88dfc2f301.1602150362.git.lucien.xin@gmail.com |
|---|---|
| State | New |
| Headers | show |
| Series | sctp: Implement RFC6951: UDP Encapsulation of SCTP | expand |
CCing Tom Herbert On Thu, Oct 8, 2020 at 5:48 PM Xin Long <lucien.xin@gmail.com> wrote: > > This patch fixes two things: > > When skb->ip_summed == CHECKSUM_PARTIAL, skb_checksum_help() should be > called do the checksum, instead of gso_make_checksum(), which is used > to do the checksum for current proto after calling skb_segment(), not > after the inner proto's gso_segment(). > > When offload_csum is disabled, the hardware will not do the checksum > for the current proto, udp. So instead of calling gso_make_checksum(), > it should calculate checksum for udp itself. > > Cc: Tom Herbert <tom@herbertland.com> > Signed-off-by: Xin Long <lucien.xin@gmail.com> > --- > net/ipv4/udp_offload.c | 11 ++++++----- > 1 file changed, 6 insertions(+), 5 deletions(-) > > diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c > index e67a66f..c0b010b 100644 > --- a/net/ipv4/udp_offload.c > +++ b/net/ipv4/udp_offload.c > @@ -131,14 +131,15 @@ static struct sk_buff *__skb_udp_tunnel_segment(struct sk_buff *skb, > uh->check = ~csum_fold(csum_add(partial, > (__force __wsum)htonl(len))); > > - if (skb->encapsulation || !offload_csum) { > - uh->check = gso_make_checksum(skb, ~uh->check); > - if (uh->check == 0) > - uh->check = CSUM_MANGLED_0; > - } else { > + if (skb->encapsulation) > + skb_checksum_help(skb); > + > + if (offload_csum) { > skb->ip_summed = CHECKSUM_PARTIAL; > skb->csum_start = skb_transport_header(skb) - skb->head; > skb->csum_offset = offsetof(struct udphdr, check); > + } else { > + uh->check = csum_fold(skb_checksum(skb, udp_offset, len, 0)); > } > } while ((skb = skb->next)); > out: > -- > 2.1.0 >
diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c index e67a66f..c0b010b 100644 --- a/net/ipv4/udp_offload.c +++ b/net/ipv4/udp_offload.c @@ -131,14 +131,15 @@ static struct sk_buff *__skb_udp_tunnel_segment(struct sk_buff *skb, uh->check = ~csum_fold(csum_add(partial, (__force __wsum)htonl(len))); - if (skb->encapsulation || !offload_csum) { - uh->check = gso_make_checksum(skb, ~uh->check); - if (uh->check == 0) - uh->check = CSUM_MANGLED_0; - } else { + if (skb->encapsulation) + skb_checksum_help(skb); + + if (offload_csum) { skb->ip_summed = CHECKSUM_PARTIAL; skb->csum_start = skb_transport_header(skb) - skb->head; skb->csum_offset = offsetof(struct udphdr, check); + } else { + uh->check = csum_fold(skb_checksum(skb, udp_offset, len, 0)); } } while ((skb = skb->next)); out:
This patch fixes two things: When skb->ip_summed == CHECKSUM_PARTIAL, skb_checksum_help() should be called do the checksum, instead of gso_make_checksum(), which is used to do the checksum for current proto after calling skb_segment(), not after the inner proto's gso_segment(). When offload_csum is disabled, the hardware will not do the checksum for the current proto, udp. So instead of calling gso_make_checksum(), it should calculate checksum for udp itself. Cc: Tom Herbert <tom@herbertland.com> Signed-off-by: Xin Long <lucien.xin@gmail.com> --- net/ipv4/udp_offload.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-)