cxgb4: prevent potential integer overflow on 32bit

PATCH Applied Dan Carpenter 2 participants 3 messages Nov 30, 2024 – Dec 10, 2024

Subsystems (3)

Kernel janitors InfiniBand CXGB4 iWARP RNIC driver (IW_CXGB4)

lore

Share & tools

lore

Dan Carpenter

author

2024-11-30 10:01 UTC #0 lore raw

PATCH 1/1 cxgb4: prevent potential integer overflow on 32bit

The "gl->tot_len" variable is controlled by the user. It comes from process_responses(). On 32bit systems, the "gl->tot_len + sizeof(struct cpl_pass_accept_req) + sizeof(struct rss_header)" addition could have an integer wrapping bug. Use size_add() to prevent this.

Fixes a08943947873 ("crypto: chtls - Register chtls with net tls")

Signed-off-by Dan Carpenter <[email protected]>

---
This is from static analysis. I've spent some time reviewing this code but I might be wrong.

 drivers/infiniband/hw/cxgb4/device.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

drivers/infiniband/hw/cxgb4/device.c +4 -2

index 80970a1738f8..034b85c42255 100644
@@ -1114,8 +1114,10 @@ static inline struct sk_buff *copy_gl_to_skb_pkt(const struct pkt_gl *gl,
 	 * The math here assumes sizeof cpl_pass_accept_req >= sizeof
 	 * cpl_rx_pkt.
 	 */
-	skb = alloc_skb(gl->tot_len + sizeof(struct cpl_pass_accept_req) +
-			sizeof(struct rss_header) - pktshift, GFP_ATOMIC);
+	skb = alloc_skb(size_add(gl->tot_len,
+				 sizeof(struct cpl_pass_accept_req) +
+				 sizeof(struct rss_header)) - pktshift,
+			GFP_ATOMIC);
 	if (unlikely(!skb))
 		return NULL;

Signature · 1 line

2.45.2

Jason Gunthorpe

maintainer Applied

2024-12-09 18:56 UTC #1 lore raw

Quoted 10 lines · Dan Carpenter

The "gl->tot_len" variable is controlled by the user. It comes from process_responses(). On 32bit systems, the "gl->tot_len + sizeof(struct cpl_pass_accept_req) + sizeof(struct rss_header)" addition could have an integer wrapping bug. Use size_add() to prevent this.
Fixes a08943947873 ("crypto: chtls - Register chtls with net tls")
Signed-off-by Dan Carpenter <[email protected]>
---
This is from static analysis. I've spent some time reviewing this code but I might be wrong.

Applied to for-next

I fixed the Fixes line:

Fixes 1cab775c3e75 ("RDMA/cxgb4: Fix LE hash collision bug for passive open connection")

Thanks,
Jason

Dan Carpenter

author

2024-12-10 09:18 UTC #2 lore raw

Quoted 12 lines · Jason Gunthorpe

Quoted 10 lines · Dan Carpenter
The "gl->tot_len" variable is controlled by the user. It comes from process_responses(). On 32bit systems, the "gl->tot_len + sizeof(struct cpl_pass_accept_req) + sizeof(struct rss_header)" addition could have an integer wrapping bug. Use size_add() to prevent this.
Fixes a08943947873 ("crypto: chtls - Register chtls with net tls")
Signed-off-by Dan Carpenter <[email protected]>
---
This is from static analysis. I've spent some time reviewing this code but I might be wrong.

Quoted 5 lines

Applied to for-next
I fixed the Fixes line:
Fixes 1cab775c3e75 ("RDMA/cxgb4: Fix LE hash collision bug for passive open connection")

Aw crud. There are two implementations of copy_gl_to_skb_pkt() and I only patched one. It's pretty weird how I mixed up the Fixes tags. Anyway, I'll patch drivers/net/ethernet/chelsio/inline_crypto/chtls/chtls_main.c as well.

regards,
dan carpenter