diff mbox series

[net-next,3/4] net/smc: fix one NULL pointer dereference in smc_ib_is_sg_need_sync()

Message ID 20240809083148.1989912-4-liujian56@huawei.com (mailing list archive)
State Not Applicable
Headers show
Series Make SMC-R can work with rxe devices | expand

Commit Message

Liu Jian Aug. 9, 2024, 8:31 a.m. UTC
BUG: kernel NULL pointer dereference, address: 0000000000000238
PGD 0 P4D 0
Oops: 0000 [#1] PREEMPT SMP PTI
CPU: 3 PID: 289 Comm: kworker/3:1 Kdump: loaded Tainted: G           OE
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.15.0-1 04/01/2014
Workqueue: smc_hs_wq smc_listen_work [smc]
RIP: 0010:dma_need_sync+0x5/0x60
...
Call Trace:
 <TASK>
 ? dma_need_sync+0x5/0x60
 ? smc_ib_is_sg_need_sync+0x61/0xf0 [smc]
 smcr_buf_map_link+0x24a/0x380 [smc]
 __smc_buf_create+0x483/0xb10 [smc]
 smc_buf_create+0x21/0xe0 [smc]
 smc_listen_work+0xf11/0x14f0 [smc]
 ? smc_tcp_listen_work+0x364/0x520 [smc]
 process_one_work+0x18d/0x3f0
 worker_thread+0x304/0x440
 kthread+0xe4/0x110
 ret_from_fork+0x47/0x70
 ret_from_fork_asm+0x1a/0x30
 </TASK>

If the software RoCE device is used, ibdev->dma_device is a null pointer.
As a result, the problem occurs. Null pointer detection is added to
prevent problems.

Signed-off-by: Liu Jian <liujian56@huawei.com>
---
 net/smc/smc_ib.c | 2 ++
 1 file changed, 2 insertions(+)

Comments

Wen Gu Aug. 9, 2024, 10:01 a.m. UTC | #1
On 2024/8/9 16:31, Liu Jian wrote:
> BUG: kernel NULL pointer dereference, address: 0000000000000238
> PGD 0 P4D 0
> Oops: 0000 [#1] PREEMPT SMP PTI
> CPU: 3 PID: 289 Comm: kworker/3:1 Kdump: loaded Tainted: G           OE
> Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.15.0-1 04/01/2014
> Workqueue: smc_hs_wq smc_listen_work [smc]
> RIP: 0010:dma_need_sync+0x5/0x60
> ...
> Call Trace:
>   <TASK>
>   ? dma_need_sync+0x5/0x60
>   ? smc_ib_is_sg_need_sync+0x61/0xf0 [smc]
>   smcr_buf_map_link+0x24a/0x380 [smc]
>   __smc_buf_create+0x483/0xb10 [smc]
>   smc_buf_create+0x21/0xe0 [smc]
>   smc_listen_work+0xf11/0x14f0 [smc]
>   ? smc_tcp_listen_work+0x364/0x520 [smc]
>   process_one_work+0x18d/0x3f0
>   worker_thread+0x304/0x440
>   kthread+0xe4/0x110
>   ret_from_fork+0x47/0x70
>   ret_from_fork_asm+0x1a/0x30
>   </TASK>
> 
> If the software RoCE device is used, ibdev->dma_device is a null pointer.
> As a result, the problem occurs. Null pointer detection is added to
> prevent problems.
> 
> Signed-off-by: Liu Jian <liujian56@huawei.com>
> ---
>   net/smc/smc_ib.c | 2 ++
>   1 file changed, 2 insertions(+)
> 
> diff --git a/net/smc/smc_ib.c b/net/smc/smc_ib.c
> index 382351ac9434..059822cc3fde 100644
> --- a/net/smc/smc_ib.c
> +++ b/net/smc/smc_ib.c
> @@ -748,6 +748,8 @@ bool smc_ib_is_sg_need_sync(struct smc_link *lnk,
>   		    buf_slot->sgt[lnk->link_idx].nents, i) {
>   		if (!sg_dma_len(sg))
>   			break;
> +		if (!lnk->smcibdev->ibdev->dma_device)
> +			break;

LGTM.

Reviewed-by: Wen Gu <guwen@linux.alibaba.com>

>   		if (dma_need_sync(lnk->smcibdev->ibdev->dma_device,
>   				  sg_dma_address(sg))) {
>   			ret = true;
Dust Li Aug. 9, 2024, 2:59 p.m. UTC | #2
On 2024-08-09 16:31:47, Liu Jian wrote:
>BUG: kernel NULL pointer dereference, address: 0000000000000238
>PGD 0 P4D 0
>Oops: 0000 [#1] PREEMPT SMP PTI
>CPU: 3 PID: 289 Comm: kworker/3:1 Kdump: loaded Tainted: G           OE
>Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.15.0-1 04/01/2014
>Workqueue: smc_hs_wq smc_listen_work [smc]
>RIP: 0010:dma_need_sync+0x5/0x60
>...
>Call Trace:
> <TASK>
> ? dma_need_sync+0x5/0x60
> ? smc_ib_is_sg_need_sync+0x61/0xf0 [smc]
> smcr_buf_map_link+0x24a/0x380 [smc]
> __smc_buf_create+0x483/0xb10 [smc]
> smc_buf_create+0x21/0xe0 [smc]
> smc_listen_work+0xf11/0x14f0 [smc]
> ? smc_tcp_listen_work+0x364/0x520 [smc]
> process_one_work+0x18d/0x3f0
> worker_thread+0x304/0x440
> kthread+0xe4/0x110
> ret_from_fork+0x47/0x70
> ret_from_fork_asm+0x1a/0x30
> </TASK>
>
>If the software RoCE device is used, ibdev->dma_device is a null pointer.
>As a result, the problem occurs. Null pointer detection is added to
>prevent problems.
>
>Signed-off-by: Liu Jian <liujian56@huawei.com>

Reviewed-by: Dust Li <dust.li@linux.alibaba.com>

Best regard,
Dust

>---
> net/smc/smc_ib.c | 2 ++
> 1 file changed, 2 insertions(+)
>
>diff --git a/net/smc/smc_ib.c b/net/smc/smc_ib.c
>index 382351ac9434..059822cc3fde 100644
>--- a/net/smc/smc_ib.c
>+++ b/net/smc/smc_ib.c
>@@ -748,6 +748,8 @@ bool smc_ib_is_sg_need_sync(struct smc_link *lnk,
> 		    buf_slot->sgt[lnk->link_idx].nents, i) {
> 		if (!sg_dma_len(sg))
> 			break;
>+		if (!lnk->smcibdev->ibdev->dma_device)
>+			break;
> 		if (dma_need_sync(lnk->smcibdev->ibdev->dma_device,
> 				  sg_dma_address(sg))) {
> 			ret = true;
>-- 
>2.34.1
>
D. Wythe Aug. 12, 2024, 3:30 a.m. UTC | #3
On 8/9/24 4:31 PM, Liu Jian wrote:
> BUG: kernel NULL pointer dereference, address: 0000000000000238
> PGD 0 P4D 0
> Oops: 0000 [#1] PREEMPT SMP PTI
> CPU: 3 PID: 289 Comm: kworker/3:1 Kdump: loaded Tainted: G           OE
> Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.15.0-1 04/01/2014
> Workqueue: smc_hs_wq smc_listen_work [smc]
> RIP: 0010:dma_need_sync+0x5/0x60
> ...
> Call Trace:
>   <TASK>
>   ? dma_need_sync+0x5/0x60
>   ? smc_ib_is_sg_need_sync+0x61/0xf0 [smc]
>   smcr_buf_map_link+0x24a/0x380 [smc]
>   __smc_buf_create+0x483/0xb10 [smc]
>   smc_buf_create+0x21/0xe0 [smc]
>   smc_listen_work+0xf11/0x14f0 [smc]
>   ? smc_tcp_listen_work+0x364/0x520 [smc]
>   process_one_work+0x18d/0x3f0
>   worker_thread+0x304/0x440
>   kthread+0xe4/0x110
>   ret_from_fork+0x47/0x70
>   ret_from_fork_asm+0x1a/0x30
>   </TASK>
>
> If the software RoCE device is used, ibdev->dma_device is a null pointer.
> As a result, the problem occurs. Null pointer detection is added to
> prevent problems.
>
> Signed-off-by: Liu Jian <liujian56@huawei.com>
> ---
>   net/smc/smc_ib.c | 2 ++
>   1 file changed, 2 insertions(+)
>
> diff --git a/net/smc/smc_ib.c b/net/smc/smc_ib.c
> index 382351ac9434..059822cc3fde 100644
> --- a/net/smc/smc_ib.c
> +++ b/net/smc/smc_ib.c
> @@ -748,6 +748,8 @@ bool smc_ib_is_sg_need_sync(struct smc_link *lnk,
>   		    buf_slot->sgt[lnk->link_idx].nents, i) {
>   		if (!sg_dma_len(sg))
>   			break;
> +		if (!lnk->smcibdev->ibdev->dma_device)
> +			break;
>   		if (dma_need_sync(lnk->smcibdev->ibdev->dma_device,
>   				  sg_dma_address(sg))) {
>   			ret = true;

Maybe you need add a fix tag ?
diff mbox series

Patch

diff --git a/net/smc/smc_ib.c b/net/smc/smc_ib.c
index 382351ac9434..059822cc3fde 100644
--- a/net/smc/smc_ib.c
+++ b/net/smc/smc_ib.c
@@ -748,6 +748,8 @@  bool smc_ib_is_sg_need_sync(struct smc_link *lnk,
 		    buf_slot->sgt[lnk->link_idx].nents, i) {
 		if (!sg_dma_len(sg))
 			break;
+		if (!lnk->smcibdev->ibdev->dma_device)
+			break;
 		if (dma_need_sync(lnk->smcibdev->ibdev->dma_device,
 				  sg_dma_address(sg))) {
 			ret = true;