If get_ep_from_tid() fails to lookup non-NULL value for ep, ep is
dereferenced later regardless of whether it is empty.
This patch adds a simple sanity check to fix the issue.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: 944661dd97f4 ("RDMA/iw_cxgb4: atomically lookup ep and get a reference")
Signed-off-by: Nikita Zhandarovich <n.zhandarovich@fintech.ru>
---
v2: do not use pr_warn() when get_ep_from_tid() returns NULL as
Jason Gunthorpe <jgg@ziepe.ca> suggested.

 drivers/infiniband/hw/cxgb4/cm.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index 499a425a3379..f5f4579f037c 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -2676,6 +2676,8 @@ static int pass_establish(struct c4iw_dev *dev, struct sk_buff *skb)
 	u16 tcp_opt = ntohs(req->tcp_opt);
 
 	ep = get_ep_from_tid(dev, tid);
+	if (!ep)
+		return 0;
 	pr_debug("ep %p tid %u\n", ep, ep->hwtid);
 	ep->snd_seq = be32_to_cpu(req->snd_isn);
 	ep->rcv_seq = be32_to_cpu(req->rcv_isn);
-- 
2.25.1

On Thu, 02 Feb 2023 10:48:50 -0800, Nikita Zhandarovich wrote:
> If get_ep_from_tid() fails to lookup non-NULL value for ep, ep is
> dereferenced later regardless of whether it is empty.
> This patch adds a simple sanity check to fix the issue.
> 
> Found by Linux Verification Center (linuxtesting.org) with SVACE.
> 
> 
> [...]

Applied, thanks!

[1/1] RDMA/cxgb4: Fix potential null-ptr-deref in pass_establish()
      https://git.kernel.org/rdma/rdma/c/283861a4c52c1e

Best regards,
-- 
Leon Romanovsky <leon@kernel.org>

On Thu, Feb 02, 2023 at 10:48:50AM -0800, Nikita Zhandarovich wrote:
> If get_ep_from_tid() fails to lookup non-NULL value for ep, ep is
> dereferenced later regardless of whether it is empty.
> This patch adds a simple sanity check to fix the issue.
> 
> Found by Linux Verification Center (linuxtesting.org) with SVACE.
> 
> Fixes: 944661dd97f4 ("RDMA/iw_cxgb4: atomically lookup ep and get a reference")
> Signed-off-by: Nikita Zhandarovich <n.zhandarovich@fintech.ru>
> ---
> v2: do not use pr_warn() when get_ep_from_tid() returns NULL as
> Jason Gunthorpe <jgg@ziepe.ca> suggested.
> 
>  drivers/infiniband/hw/cxgb4/cm.c | 2 ++
>  1 file changed, 2 insertions(+)

I applied, but please next time.
1. Don't send patches as reply-to. It messes whole patch flow.
2. Use target in subject line, e.g. [PATCH rdma-next] or [PATCH rdma-rc]

Thanks


> 
> diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
> index 499a425a3379..f5f4579f037c 100644
> --- a/drivers/infiniband/hw/cxgb4/cm.c
> +++ b/drivers/infiniband/hw/cxgb4/cm.c
> @@ -2676,6 +2676,8 @@ static int pass_establish(struct c4iw_dev *dev, struct sk_buff *skb)
>  	u16 tcp_opt = ntohs(req->tcp_opt);
>  
>  	ep = get_ep_from_tid(dev, tid);
> +	if (!ep)
> +		return 0;
>  	pr_debug("ep %p tid %u\n", ep, ep->hwtid);
>  	ep->snd_seq = be32_to_cpu(req->snd_isn);
>  	ep->rcv_seq = be32_to_cpu(req->rcv_isn);
> -- 
> 2.25.1
>