net: fix stale TX skb pointers on DMA map failure
Subsystems (7)
cover net: fix stale TX skb pointers on DMA map failure
From: Xuanqiang Luo <[email protected]>
While I was backporting commit 1a303baa715e6 ("ice: fix double-free of tx_buf skb"), an AI-assisted scan identified several suspected TX error paths. I reviewed the results and found this issue in the three drivers fixed here.
The drivers differ, but the bug is the same. On a DMA mapping failure, the TX path frees an skb while its ring entry still points to it. A later transmission normally overwrites the entry. If the interface is stopped first, teardown can instead access or free the skb again.
I do not have these adapters, so I have not tested the drivers on hardware. I checked the error and teardown paths by inspection. Still, these small fixes seem worth posting for review. They are independent, but are sent as one series because they address the same issue.
Xuanqiang Luo (3):
bna: fix use-after-free on DMA mapping failure
hinic3: fix use-after-free on DMA mapping failure
net: hibmcge: fix double-free of tx skb on DMA mapping failure
drivers/net/ethernet/brocade/bna/bnad.c | 2 +- drivers/net/ethernet/hisilicon/hibmcge/hbg_txrx.c | 1 + drivers/net/ethernet/huawei/hinic3/hinic3_tx.c | 5 +++-- 3 files changed, 5 insertions(+), 3 deletions(-)
Signature · 1 line
2.43.0
PATCH 1/3 bna: fix use-after-free on DMA mapping failure
From: Xuanqiang Luo <[email protected]>
If dma_map_single() fails in bnad_start_xmit(), the skb is freed, but head_unmap->skb was set before the mapping attempt and is not cleared. The producer index is not advanced, so later transmissions normally overwrite the entry.
However, if the interface is brought down first, bnad_txq_cleanup() scans the entire unmap queue, finds the stale pointer, and calls bnad_tx_buff_unmap() on it. That function dereferences the freed skb in skb_headlen(). Its zero nvecs count is decremented to -1, causing its while (nvecs) loop to repeatedly unmap entries around the TX ring and potentially hang cleanup.
Set head_unmap->skb after the first DMA mapping succeeds. This prevents the stale entry from reaching bnad_tx_buff_unmap().
--- drivers/net/ethernet/brocade/bna/bnad.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/brocade/bna/bnad.c +1 -1
@@ -3006,7 +3006,6 @@ bnad_start_xmit(struct sk_buff *skb, struct net_device *netdev) txqent->hdr.wi.reserved =0 ; txqent->hdr.wi.num_vectors = vectors; - head_unmap->skb = skb; head_unmap->nvecs =0 ;/* Program the vectors */
@@ -3018,6 +3017,7 @@ bnad_start_xmit(struct sk_buff *skb, struct net_device *netdev)BNAD_UPDATE_CTR (bnad, tx_skb_map_failed);return NETDEV_TX_OK ; } + head_unmap->skb = skb;BNA_SET_DMA_ADDR (dma_addr, &txqent->vector[0 ].host_addr); txqent->vector[0 ].length =htons (len);dma_unmap_addr_set (&unmap->vectors[0 ], dma_addr, dma_addr);
Signature · 1 line
2.43.0
PATCH 2/3 hinic3: fix use-after-free on DMA mapping failure
In reply to [email protected]From: Xuanqiang Luo <[email protected]>
If hinic3_tx_map_skb() fails in hinic3_send_one_skb(), the skb is freed, but tx_info->skb was set before the mapping attempt and is not cleared. The SQ producer index is rolled back, so later transmissions normally overwrite the entry.
If the interface is brought down first, hinic3_free_txqs_res() calls free_all_tx_skbs(). It scans the entire tx_info array and finds the stale pointer. hinic3_tx_unmap_skb() then dereferences the freed skb in skb_shinfo(), before it is freed again.
Set tx_info->skb and its WQEBB count only after DMA mapping succeeds, preventing the stale pointer from reaching free_all_tx_skbs().
--- drivers/net/ethernet/huawei/hinic3/hinic3_tx.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/net/ethernet/huawei/hinic3/hinic3_tx.c +3 -2
@@ -578,8 +578,6 @@ static netdev_tx_t hinic3_send_one_skb(struct sk_buff *skb, *wqe_combo.task = task; tx_info = &txq->tx_info[pi]; - tx_info->skb = skb; - tx_info->wqebb_cnt = wqebb_cnt; err =hinic3_tx_map_skb (netdev, skb, txq, tx_info, &wqe_combo);if (err) {
@@ -589,6 +587,9 @@ static netdev_tx_t hinic3_send_one_skb(struct sk_buff *skb, goto err_drop_pkt; } + tx_info->skb = skb; + tx_info->wqebb_cnt = wqebb_cnt; +netif_subqueue_sent (netdev, txq->sq->q_id, skb->len); netif_subqueue_maybe_stop(netdev, txq->sq->q_id, hinic3_wq_free_wqebbs(&txq->sq->wq),
Signature · 1 line
2.43.0
Re: [PATCH net v1 2/3] hinic3: fix use-after-free on DMA mapping failure
Quoted 17 lines
drivers/net/ethernet/huawei/hinic3/hinic3_tx.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)drivers/net/ethernet/huawei/hinic3/hinic3_tx.c +0
-2@@ -578,8 +578,6 @@ static netdev_tx_t hinic3_send_one_skb(struct sk_buff *skb, *wqe_combo.task = task; tx_info = &txq->tx_info[pi]; - tx_info->skb = skb; - tx_info->wqebb_cnt = wqebb_cnt; err =hinic3_tx_map_skb (netdev, skb, txq, tx_info, &wqe_combo);if (err) {
drivers/net/ethernet/huawei/hinic3/hinic3_tx.c static netdev_tx_t hinic3_send_one_skb(struct sk_buff *skb, · line 595
Quoted 9 lines
drivers/net/ethernet/huawei/hinic3/hinic3_tx.c +3
-0@@ -589,6 +587,9 @@ static netdev_tx_t hinic3_send_one_skb(struct sk_buff *skb, goto err_drop_pkt; } + tx_info->skb = skb; + tx_info->wqebb_cnt = wqebb_cnt; +netif_subqueue_sent (netdev, txq->sq->q_id, skb->len); netif_subqueue_maybe_stop(netdev, txq->sq->q_id, hinic3_wq_free_wqebbs(&txq->sq->wq),
Thanks for your fixes, the proposed modification of the hinic3 driver has been tested on hardware, and it works fine.
PATCH 3/3 net: hibmcge: fix double-free of tx skb on DMA mapping failure
In reply to [email protected]From: Xuanqiang Luo <[email protected]>
If hbg_dma_map() fails, hbg_net_start_xmit() frees the skb, but buffer->skb is left pointing to it. ring->ntu is not advanced, so the buffer is not visible to the TX cleanup path.
A subsequent transmit normally overwrites the buffer. However, if the interface is brought down first, hbg_ring_uninit() calls hbg_buffer_free(). It sees the stale pointer, attempts to unmap the failed mapping, and frees the skb again.
Clear buffer->skb before freeing the skb in the error path, preventing hbg_buffer_free() from treating it as an outstanding TX buffer.
--- drivers/net/ethernet/hisilicon/hibmcge/hbg_txrx.c | 1 + 1 file changed, 1 insertion(+)
drivers/net/ethernet/hisilicon/hibmcge/hbg_txrx.c +1 -0
@@ -155,6 +155,7 @@ netdev_tx_t hbg_net_start_xmit(struct sk_buff *skb, struct net_device *netdev) buffer->skb = skb; buffer->skb_len = skb->len;if (unlikely (hbg_dma_map (buffer))) { + buffer->skb =NULL ;dev_kfree_skb_any (skb);return NETDEV_TX_OK ; }
Signature · 1 line
2.43.0
Re: [PATCH net v1 3/3] net: hibmcge: fix double-free of tx skb on DMA mapping failure
Quoted 11 lines · on 2026/7/10 17:05, [email protected] wrote:
From: Xuanqiang Luo <[email protected]>
If
hbg_dma_map()fails,hbg_net_start_xmit()frees the skb, butbuffer->skbis left pointing to it.ring->ntuis not advanced, so the buffer is not visible to the TX cleanup path.A subsequent transmit normally overwrites the buffer. However, if the interface is brought down first,
hbg_ring_uninit()callshbg_buffer_free(). It sees the stale pointer, attempts to unmap the failed mapping, and frees the skb again.
Quoted 7 lines
Clear
buffer->skbbefore freeing the skb in the error path, preventinghbg_buffer_free()from treating it as an outstanding TX buffer.Fixes 40735e7543f9 ("net: hibmcge: Implement .ndo_start_xmit function")Assisted-by Opencode:deepseek-v4-pro[1m]Signed-off-by Xuanqiang Luo <[email protected]>
Thanks for fixing this. The patch looks good to me.
Quoted 16 lines
--- drivers/net/ethernet/hisilicon/hibmcge/hbg_txrx.c | 1 + 1 file changed, 1 insertion(+)drivers/net/ethernet/hisilicon/hibmcge/hbg_txrx.c +1
-0@@ -155,6 +155,7 @@ netdev_tx_t hbg_net_start_xmit(struct sk_buff *skb, struct net_device *netdev) buffer->skb = skb; buffer->skb_len = skb->len;if (unlikely (hbg_dma_map (buffer))) { + buffer->skb =NULL ;dev_kfree_skb_any (skb);return NETDEV_TX_OK ; }
Quoted 17 lines · On Fri, Jul 10, 2026 at 05:05:21PM +0800, [email protected] wrote:
From: Xuanqiang Luo <[email protected]>
While I was backporting commit
1a303baa715e6("ice: fix double-free oftx_bufskb"), an AI-assisted scan identified several suspected TX error paths. I reviewed the results and found this issue in the three drivers fixed here.The drivers differ, but the bug is the same. On a DMA mapping failure, the TX path frees an skb while its ring entry still points to it. A later transmission normally overwrites the entry. If the interface is stopped first, teardown can instead access or free the skb again.
I do not have these adapters, so I have not tested the drivers on hardware. I checked the error and teardown paths by inspection. Still, these small fixes seem worth posting for review. They are independent, but are sent as one series because they address the same issue.
Quoted 4 lines
Xuanqiang Luo (3):
bna: fix use-after-free on DMA mapping failure
hinic3: fix use-after-free on DMA mapping failure
net: hibmcge: fix double-free of tx skb on DMA mapping failure
For the series:
This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:
Hello:
This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:
Quoted 8 lines · On Fri, 10 Jul 2026 17:05:21 +0800 you wrote:
From: Xuanqiang Luo <[email protected]>
While I was backporting commit
1a303baa715e6("ice: fix double-free oftx_bufskb"), an AI-assisted scan identified several suspected TX error paths. I reviewed the results and found this issue in the three drivers fixed here.[...]
Here is the summary with links:
- [net,v1,1/3] bna: fix use-after-free on DMA mapping failure
https://git.kernel.org/netdev/net-next/c/be9381d5776c - [net,v1,2/3] hinic3: fix use-after-free on DMA mapping failure https://git.kernel.org/netdev/net-next/c/dd5de39af5ef - [net,v1,3/3] net: hibmcge: fix double-free of tx skb on DMA mapping failure https://git.kernel.org/netdev/net-next/c/569595dc92e0
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html