Hi, Sagi
I notice that this series "nvmet-tcp: Don't map pages which can't come from HIGHMEM"
has no reply for two weeks. Is it possible to fix this critical issue firstly?
Cc: Fabio M. De Francesco <[email protected]>
v3 -> v4:
- free CMD buffers only nvmet_tcp_need_data_in().
v2 -> v3:
- Keep original function name as nvmet_tcp_uninit_data_in_cmds()
- Add new function nvmet_tcp_free_cmd_data_in_buffers() to free CMD buffers
- Rework against http://git.infradead.org/nvme.git/shortlog/refs/heads/nvme-6.0
- Move nvmet_tcp_free_cmd_data_in_buffers() after cancel_work_sync(&queue->io_work).
Hi @Sagi, if you still prefer to clear the commands that are still waiting for
data from the wire, please let me know.
v1 -> v2:
- Separate nvmet_tcp_uninit_data_in_cmds() into nvmet_tcp_uninit_req()
& nvmet_tcp_uninit_data_in_cmds()(free buffers only).
v1:
- Move nvmet_tcp_uninit_data_in_cmds() after nvmet_sq_destroy()
zhenwei pi (1):
nvmet-tcp: Fix NULL pointer dereference during release
drivers/nvme/target/tcp.c | 20 ++++++++++++++++++--
1 file changed, 18 insertions(+), 2 deletions(-)
--
2.20.1
nvmet-tcp frees CMD buffers in nvmet_tcp_uninit_data_in_cmds(),
and waits the inflight IO requests in nvmet_sq_destroy(). During wait
the inflight IO requests, the callback nvmet_tcp_queue_response()
is called from backend after IO complete, this leads a typical
Use-After-Free issue like this:
BUG: kernel NULL pointer dereference, address: 0000000000000008
#PF: supervisor read access in kernel mode
#PF: error_code(0x0000) - not-present page
PGD 0 P4D 0
Oops: 0000 [#1] PREEMPT SMP NOPTI
CPU: 14 PID: 207 Comm: kworker/14:1H Kdump: loaded Tainted: G E 6.0.0-rc2.bm.1-amd64 #12
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
Workqueue: nvmet_tcp_wq nvmet_tcp_io_work [nvmet_tcp]
RIP: 0010:shash_ahash_digest+0x2b/0x110
Code: 1f 44 00 00 41 57 41 56 41 55 41 54 55 48 89 fd 53 48 89 f3 48 83 ec 08 44 8b 67 30 45 85 e4 74 1c 48 8b 57 38 b8 00 10 00 00 <44> 8b 7a 08 44 29 f8 39 42 0c 0f 46 42 0c 41 39 c4 76 43 48 8b 03
RSP: 0018:ffffc900006e3dd8 EFLAGS: 00010206
RAX: 0000000000001000 RBX: ffff888104ac1650 RCX: 0000000000000000
RDX: 0000000000000000 RSI: ffff888104ac1650 RDI: ffff888104ac1600
RBP: ffff888104ac1600 R08: ffff8881073980c8 R09: ffff8881057798b8
R10: 8080808080808080 R11: 0000000000000000 R12: 0000000000001000
R13: 0000000000000000 R14: ffff88810601a1cc R15: ffff888107398000
FS: 0000000000000000(0000) GS:ffff88823fd80000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000008 CR3: 000000010a8e8000 CR4: 0000000000350ee0
Call Trace:
<TASK>
nvmet_tcp_io_work+0xa1c/0xb1c [nvmet_tcp]
? __switch_to+0x106/0x420
process_one_work+0x1ae/0x380
? process_one_work+0x380/0x380
worker_thread+0x30/0x360
? process_one_work+0x380/0x380
kthread+0xe6/0x110
? kthread_complete_and_exit+0x20/0x20
ret_from_fork+0x1f/0x30
Suggested by Sagi, separate nvmet_tcp_uninit_data_in_cmds() into two
steps:
uninit data in cmds <- new step 1
nvmet_sq_destroy();
cancel_work_sync(&queue->io_work);
free CMD buffers <- new step 2
Signed-off-by: zhenwei pi <[email protected]>
---
drivers/nvme/target/tcp.c | 20 ++++++++++++++++++--
1 file changed, 18 insertions(+), 2 deletions(-)
diff --git a/drivers/nvme/target/tcp.c b/drivers/nvme/target/tcp.c
index a3694a32f6d5..3ac9750b41ad 100644
--- a/drivers/nvme/target/tcp.c
+++ b/drivers/nvme/target/tcp.c
@@ -1423,14 +1423,29 @@ static void nvmet_tcp_uninit_data_in_cmds(struct nvmet_tcp_queue *queue)
for (i = 0; i < queue->nr_cmds; i++, cmd++) {
if (nvmet_tcp_need_data_in(cmd))
nvmet_req_uninit(&cmd->req);
+ }
+
+ if (!queue->nr_cmds && nvmet_tcp_need_data_in(&queue->connect)) {
+ /* failed in connect */
+ nvmet_req_uninit(&queue->connect.req);
+ }
+}
+static void nvmet_tcp_free_cmd_data_in_buffers(struct nvmet_tcp_queue *queue)
+{
+ struct nvmet_tcp_cmd *cmd = queue->cmds;
+ int i;
+
+ for (i = 0; i < queue->nr_cmds; i++, cmd++) {
+ if (!nvmet_tcp_need_data_in(cmd))
+ continue;
nvmet_tcp_unmap_pdu_iovec(cmd);
nvmet_tcp_free_cmd_buffers(cmd);
}
if (!queue->nr_cmds && nvmet_tcp_need_data_in(&queue->connect)) {
- /* failed in connect */
- nvmet_tcp_finish_cmd(&queue->connect);
+ nvmet_tcp_unmap_pdu_iovec(&queue->connect);
+ nvmet_tcp_free_cmd_buffers(&queue->connect);
}
}
@@ -1452,6 +1467,7 @@ static void nvmet_tcp_release_queue_work(struct work_struct *w)
nvmet_tcp_uninit_data_in_cmds(queue);
nvmet_sq_destroy(&queue->nvme_sq);
cancel_work_sync(&queue->io_work);
+ nvmet_tcp_free_cmd_data_in_buffers(queue);
sock_release(queue->sock);
nvmet_tcp_free_cmds(queue);
if (queue->hdr_digest || queue->data_digest)
--
2.20.1
On 9/13/22 04:42, zhenwei pi wrote:
> nvmet-tcp frees CMD buffers in nvmet_tcp_uninit_data_in_cmds(),
> and waits the inflight IO requests in nvmet_sq_destroy(). During wait
> the inflight IO requests, the callback nvmet_tcp_queue_response()
> is called from backend after IO complete, this leads a typical
> Use-After-Free issue like this:
>
> BUG: kernel NULL pointer dereference, address: 0000000000000008
> #PF: supervisor read access in kernel mode
> #PF: error_code(0x0000) - not-present page
> PGD 0 P4D 0
> Oops: 0000 [#1] PREEMPT SMP NOPTI
> CPU: 14 PID: 207 Comm: kworker/14:1H Kdump: loaded Tainted: G E 6.0.0-rc2.bm.1-amd64 #12
> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
> Workqueue: nvmet_tcp_wq nvmet_tcp_io_work [nvmet_tcp]
> RIP: 0010:shash_ahash_digest+0x2b/0x110
> Code: 1f 44 00 00 41 57 41 56 41 55 41 54 55 48 89 fd 53 48 89 f3 48 83 ec 08 44 8b 67 30 45 85 e4 74 1c 48 8b 57 38 b8 00 10 00 00 <44> 8b 7a 08 44 29 f8 39 42 0c 0f 46 42 0c 41 39 c4 76 43 48 8b 03
> RSP: 0018:ffffc900006e3dd8 EFLAGS: 00010206
> RAX: 0000000000001000 RBX: ffff888104ac1650 RCX: 0000000000000000
> RDX: 0000000000000000 RSI: ffff888104ac1650 RDI: ffff888104ac1600
> RBP: ffff888104ac1600 R08: ffff8881073980c8 R09: ffff8881057798b8
> R10: 8080808080808080 R11: 0000000000000000 R12: 0000000000001000
> R13: 0000000000000000 R14: ffff88810601a1cc R15: ffff888107398000
> FS: 0000000000000000(0000) GS:ffff88823fd80000(0000) knlGS:0000000000000000
> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 0000000000000008 CR3: 000000010a8e8000 CR4: 0000000000350ee0
> Call Trace:
> <TASK>
> nvmet_tcp_io_work+0xa1c/0xb1c [nvmet_tcp]
> ? __switch_to+0x106/0x420
> process_one_work+0x1ae/0x380
> ? process_one_work+0x380/0x380
> worker_thread+0x30/0x360
> ? process_one_work+0x380/0x380
> kthread+0xe6/0x110
> ? kthread_complete_and_exit+0x20/0x20
> ret_from_fork+0x1f/0x30
>
> Suggested by Sagi, separate nvmet_tcp_uninit_data_in_cmds() into two
> steps:
> uninit data in cmds <- new step 1
> nvmet_sq_destroy();
> cancel_work_sync(&queue->io_work);
> free CMD buffers <- new step 2
>
> Signed-off-by: zhenwei pi <[email protected]>
> ---
> drivers/nvme/target/tcp.c | 20 ++++++++++++++++++--
> 1 file changed, 18 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/nvme/target/tcp.c b/drivers/nvme/target/tcp.c
> index a3694a32f6d5..3ac9750b41ad 100644
> --- a/drivers/nvme/target/tcp.c
> +++ b/drivers/nvme/target/tcp.c
> @@ -1423,14 +1423,29 @@ static void nvmet_tcp_uninit_data_in_cmds(struct nvmet_tcp_queue *queue)
> for (i = 0; i < queue->nr_cmds; i++, cmd++) {
> if (nvmet_tcp_need_data_in(cmd))
> nvmet_req_uninit(&cmd->req);
> + }
> +
> + if (!queue->nr_cmds && nvmet_tcp_need_data_in(&queue->connect)) {
> + /* failed in connect */
> + nvmet_req_uninit(&queue->connect.req);
> + }
> +}
>
> +static void nvmet_tcp_free_cmd_data_in_buffers(struct nvmet_tcp_queue *queue)
> +{
> + struct nvmet_tcp_cmd *cmd = queue->cmds;
> + int i;
> +
> + for (i = 0; i < queue->nr_cmds; i++, cmd++) {
> + if (!nvmet_tcp_need_data_in(cmd))
> + continue;
> nvmet_tcp_unmap_pdu_iovec(cmd);
This will cause a merge conflict, because nvme-6.1
nvmet_tcp_unmap_pdu_iovec does no longer exist.
Other than that, looks fine to me,
Reviewed-by: Sagi Grimberg <[email protected]>
> nvmet-tcp frees CMD buffers in nvmet_tcp_uninit_data_in_cmds(),
> and waits the inflight IO requests in nvmet_sq_destroy(). During wait
> the inflight IO requests, the callback nvmet_tcp_queue_response()
> is called from backend after IO complete, this leads a typical
> Use-After-Free issue like this:
Would it be possible to resend this patch rebased on top of nvme-6.1?