Hi all,
Today's linux-next merge of the rdma tree got a conflict in:
drivers/infiniband/sw/rxe/rxe_net.c
between commit:
f1b0a8ea9f12 ("Revert "RDMA/rxe: Remove VLAN code leftovers from RXE"")
from Linus' tree and commit:
899aba891cab ("RDMA/rxe: Fix FIXME in rxe_udp_encap_recv()")
from the rdma tree.
I fixed it up (see below) and can carry the fix as necessary. This
is now fixed as far as linux-next is concerned, but any non trivial
conflicts should be mentioned to your upstream maintainer when your tree
is submitted for merging. You may also want to consider cooperating
with the maintainer of the conflicting tree to minimise any particularly
complex conflicts.
--
Cheers,
Stephen Rothwell
diff --cc drivers/infiniband/sw/rxe/rxe_net.c
index 943914c2a50c,36d56163afac..000000000000
--- a/drivers/infiniband/sw/rxe/rxe_net.c
+++ b/drivers/infiniband/sw/rxe/rxe_net.c
@@@ -153,15 -152,14 +153,19 @@@ static struct dst_entry *rxe_find_route
static int rxe_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
{
struct udphdr *udph;
+ struct rxe_dev *rxe;
struct net_device *ndev = skb->dev;
+ struct net_device *rdev = ndev;
- struct rxe_dev *rxe = rxe_get_dev_from_net(ndev);
struct rxe_pkt_info *pkt = SKB_TO_PKT(skb);
+ /* takes a reference on rxe->ib_dev
+ * drop when skb is freed
+ */
+ rxe = rxe_get_dev_from_net(ndev);
+ if (!rxe && is_vlan_dev(rdev)) {
+ rdev = vlan_dev_real_dev(ndev);
+ rxe = rxe_get_dev_from_net(rdev);
+ }
if (!rxe)
goto drop;
It looks like f1b0a8ea9f12 ("Revert "RDMA/rxe: Remove VLAN code leftovers from RXE"") has not been applied to rdma for-next which is where I do my work. Not sure how it got upstream.
bob
-----Original Message-----
From: Stephen Rothwell <[email protected]>
Sent: Tuesday, February 9, 2021 8:16 PM
To: Doug Ledford <[email protected]>; Jason Gunthorpe <[email protected]>
Cc: Pearson, Robert B <[email protected]>; Bob Pearson <[email protected]>; Jason Gunthorpe <[email protected]>; Jason Gunthorpe <[email protected]>; Linux Kernel Mailing List <[email protected]>; Linux Next Mailing List <[email protected]>; Martin Wilck <[email protected]>
Subject: linux-next: manual merge of the rdma tree with Linus' tree
Hi all,
Today's linux-next merge of the rdma tree got a conflict in:
drivers/infiniband/sw/rxe/rxe_net.c
between commit:
f1b0a8ea9f12 ("Revert "RDMA/rxe: Remove VLAN code leftovers from RXE"")
from Linus' tree and commit:
899aba891cab ("RDMA/rxe: Fix FIXME in rxe_udp_encap_recv()")
from the rdma tree.
I fixed it up (see below) and can carry the fix as necessary. This is now fixed as far as linux-next is concerned, but any non trivial conflicts should be mentioned to your upstream maintainer when your tree is submitted for merging. You may also want to consider cooperating with the maintainer of the conflicting tree to minimise any particularly complex conflicts.
--
Cheers,
Stephen Rothwell
diff --cc drivers/infiniband/sw/rxe/rxe_net.c
index 943914c2a50c,36d56163afac..000000000000
--- a/drivers/infiniband/sw/rxe/rxe_net.c
+++ b/drivers/infiniband/sw/rxe/rxe_net.c
@@@ -153,15 -152,14 +153,19 @@@ static struct dst_entry *rxe_find_route
static int rxe_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
{
struct udphdr *udph;
+ struct rxe_dev *rxe;
struct net_device *ndev = skb->dev;
+ struct net_device *rdev = ndev;
- struct rxe_dev *rxe = rxe_get_dev_from_net(ndev);
struct rxe_pkt_info *pkt = SKB_TO_PKT(skb);
+ /* takes a reference on rxe->ib_dev
+ * drop when skb is freed
+ */
+ rxe = rxe_get_dev_from_net(ndev);
+ if (!rxe && is_vlan_dev(rdev)) {
+ rdev = vlan_dev_real_dev(ndev);
+ rxe = rxe_get_dev_from_net(rdev);
+ }
if (!rxe)
goto drop;
On Wed, Feb 10, 2021 at 06:57:24PM +0000, Pearson, Robert B wrote:
> It looks like f1b0a8ea9f12 ("Revert "RDMA/rxe: Remove VLAN code
> leftovers from RXE"") has not been applied to rdma for-next which is
> where I do my work. Not sure how it got upstream.
It was sent to the rc tree, don't worry this is stuff I fix up.
Stephen's note is to help ensure it is fixed properly, does the change
look OK? It looks like this when applied (I changed Stephen's slightly)
static int rxe_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
{
struct udphdr *udph;
struct rxe_dev *rxe;
struct net_device *ndev = skb->dev;
struct rxe_pkt_info *pkt = SKB_TO_PKT(skb);
/* takes a reference on rxe->ib_dev
* drop when skb is freed
*/
rxe = rxe_get_dev_from_net(ndev);
if (!rxe && is_vlan_dev(rdev))
rxe = rxe_get_dev_from_net(vlan_dev_real_dev(ndev));
if (!rxe)
goto drop;
Jason
Looks perfect. Thanks.
-----Original Message-----
From: Jason Gunthorpe <[email protected]>
Sent: Wednesday, February 10, 2021 2:37 PM
To: Pearson, Robert B <[email protected]>
Cc: Stephen Rothwell <[email protected]>; Doug Ledford <[email protected]>; Bob Pearson <[email protected]>; Linux Kernel Mailing List <[email protected]>; Linux Next Mailing List <[email protected]>; Martin Wilck <[email protected]>
Subject: Re: linux-next: manual merge of the rdma tree with Linus' tree
On Wed, Feb 10, 2021 at 06:57:24PM +0000, Pearson, Robert B wrote:
> It looks like f1b0a8ea9f12 ("Revert "RDMA/rxe: Remove VLAN code
> leftovers from RXE"") has not been applied to rdma for-next which is
> where I do my work. Not sure how it got upstream.
It was sent to the rc tree, don't worry this is stuff I fix up.
Stephen's note is to help ensure it is fixed properly, does the change look OK? It looks like this when applied (I changed Stephen's slightly)
static int rxe_udp_encap_recv(struct sock *sk, struct sk_buff *skb) {
struct udphdr *udph;
struct rxe_dev *rxe;
struct net_device *ndev = skb->dev;
struct rxe_pkt_info *pkt = SKB_TO_PKT(skb);
/* takes a reference on rxe->ib_dev
* drop when skb is freed
*/
rxe = rxe_get_dev_from_net(ndev);
if (!rxe && is_vlan_dev(rdev))
rxe = rxe_get_dev_from_net(vlan_dev_real_dev(ndev));
if (!rxe)
goto drop;
Jason
On Wed, 2021-02-10 at 22:08 +0000, Pearson, Robert B wrote:
> Looks perfect. Thanks.
+1
Thanks everyone.