2017-12-13 09:09:00

by Jia-Ju Bai

Subject: [PATCH] qedi: Fix a possible sleep-in-atomic bug in qedi_process_tmf_resp

The driver may sleep under a spinlock.
The function call path is:
qedi_cpu_offline (acquire the spinlock)
qedi_fp_process_cqes
qedi_mtask_completion
qedi_process_tmf_resp
kzalloc(GFP_KERNEL) --> may sleep

To fix it, GFP_KERNEL is replaced with GFP_ATOMIC.

This bug is found by my static analysis tool(DSAC) and checked by my code review.

Signed-off-by: Jia-Ju Bai <[email protected]>
---
drivers/scsi/qedi/qedi_fw.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/scsi/qedi/qedi_fw.c b/drivers/scsi/qedi/qedi_fw.c
index bd302d3..20a9259 100644
--- a/drivers/scsi/qedi/qedi_fw.c
+++ b/drivers/scsi/qedi/qedi_fw.c
@@ -198,7 +198,7 @@ static void qedi_process_tmf_resp(struct qedi_ctx *qedi,
cqe_tmp_response = &cqe->cqe_common.iscsi_hdr.tmf_response;

qedi_cmd = task->dd_data;
- qedi_cmd->tmf_resp_buf = kzalloc(sizeof(*resp_hdr_ptr), GFP_KERNEL);
+ qedi_cmd->tmf_resp_buf = kzalloc(sizeof(*resp_hdr_ptr), GFP_ATOMIC);
if (!qedi_cmd->tmf_resp_buf) {
QEDI_ERR(&qedi->dbg_ctx,
"Failed to allocate resp buf, cid=0x%x\n",
--
1.7.9.5

2017-12-13 09:12:32

by Rangankar, Manish

Subject: Re: [PATCH] qedi: Fix a possible sleep-in-atomic bug in qedi_process_tmf_resp

On 13/12/17 2:41 PM, "Jia-Ju Bai" <[email protected]> wrote:

>The driver may sleep under a spinlock.
>The function call path is:
>qedi_cpu_offline (acquire the spinlock)
> qedi_fp_process_cqes
> qedi_mtask_completion
> qedi_process_tmf_resp
> kzalloc(GFP_KERNEL) --> may sleep
>
>To fix it, GFP_KERNEL is replaced with GFP_ATOMIC.
>
>This bug is found by my static analysis tool(DSAC) and checked by my code
>review.
>
>Signed-off-by: Jia-Ju Bai <[email protected]>
>---
> drivers/scsi/qedi/qedi_fw.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
>diff --git a/drivers/scsi/qedi/qedi_fw.c b/drivers/scsi/qedi/qedi_fw.c
>index bd302d3..20a9259 100644
>--- a/drivers/scsi/qedi/qedi_fw.c
>+++ b/drivers/scsi/qedi/qedi_fw.c
>@@ -198,7 +198,7 @@ static void qedi_process_tmf_resp(struct qedi_ctx
>*qedi,
> cqe_tmp_response = &cqe->cqe_common.iscsi_hdr.tmf_response;
>
> qedi_cmd = task->dd_data;
>- qedi_cmd->tmf_resp_buf = kzalloc(sizeof(*resp_hdr_ptr), GFP_KERNEL);
>+ qedi_cmd->tmf_resp_buf = kzalloc(sizeof(*resp_hdr_ptr), GFP_ATOMIC);
> if (!qedi_cmd->tmf_resp_buf) {
> QEDI_ERR(&qedi->dbg_ctx,
> "Failed to allocate resp buf, cid=0x%x\n",
>--
>1.7.9.5

Thanks,

Acked-by: Manish Rangankar <[email protected]>

>

2017-12-19 03:33:43

by Martin K. Petersen

Subject: Re: [PATCH] qedi: Fix a possible sleep-in-atomic bug in qedi_process_tmf_resp

Jia-Ju,

> To fix it, GFP_KERNEL is replaced with GFP_ATOMIC.
>
> This bug is found by my static analysis tool(DSAC) and checked by my
> code review.

Applied to 4.16/scsi-queue. Thank you!

--
Martin K. Petersen Oracle Linux Engineering