diff mbox series

[1/2] io_uring/cmd: let cmds to know about dying task

Message ID 20241031163257.3616106-1-maharmstone@fb.com (mailing list archive)
State New
Headers show
Series [1/2] io_uring/cmd: let cmds to know about dying task | expand

Commit Message

Mark Harmstone Oct. 31, 2024, 4:32 p.m. UTC
From: Pavel Begunkov <asml.silence@gmail.com>

When the taks that submitted a request is dying, a task work for that
request might get run by a kernel thread or even worse by a half
dismantled task. We can't just cancel the task work without running the
callback as the cmd might need to do some clean up, so pass a flag
instead. If set, it's not safe to access any task resources and the
callback is expected to cancel the cmd ASAP.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
---
 include/linux/io_uring_types.h | 1 +
 io_uring/uring_cmd.c           | 6 +++++-
 2 files changed, 6 insertions(+), 1 deletion(-)

Comments

Ming Lei Nov. 1, 2024, 3:14 a.m. UTC | #1
On Fri, Nov 1, 2024 at 12:33 AM Mark Harmstone <maharmstone@fb.com> wrote:
>
> From: Pavel Begunkov <asml.silence@gmail.com>
>
> When the taks that submitted a request is dying, a task work for that
> request might get run by a kernel thread or even worse by a half
> dismantled task. We can't just cancel the task work without running the
> callback as the cmd might need to do some clean up, so pass a flag
> instead. If set, it's not safe to access any task resources and the
> callback is expected to cancel the cmd ASAP.
>
> Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
> ---
>  include/linux/io_uring_types.h | 1 +
>  io_uring/uring_cmd.c           | 6 +++++-
>  2 files changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h
> index 4b9ba523978d..2ee5dc105b58 100644
> --- a/include/linux/io_uring_types.h
> +++ b/include/linux/io_uring_types.h
> @@ -37,6 +37,7 @@ enum io_uring_cmd_flags {
>         /* set when uring wants to cancel a previously issued command */
>         IO_URING_F_CANCEL               = (1 << 11),
>         IO_URING_F_COMPAT               = (1 << 12),
> +       IO_URING_F_TASK_DEAD            = (1 << 13),
>  };
>
>  struct io_wq_work_node {
> diff --git a/io_uring/uring_cmd.c b/io_uring/uring_cmd.c
> index 39c3c816ec78..78a8ba5d39ae 100644
> --- a/io_uring/uring_cmd.c
> +++ b/io_uring/uring_cmd.c
> @@ -119,9 +119,13 @@ EXPORT_SYMBOL_GPL(io_uring_cmd_mark_cancelable);
>  static void io_uring_cmd_work(struct io_kiocb *req, struct io_tw_state *ts)
>  {
>         struct io_uring_cmd *ioucmd = io_kiocb_to_cmd(req, struct io_uring_cmd);
> +       unsigned int flags = IO_URING_F_COMPLETE_DEFER;
> +
> +       if (req->task != current)
> +               flags |= IO_URING_F_TASK_DEAD;

Looks fine,

Reviewed-by: Ming Lei <ming.lei@redhat.com>

BTW,  uring_cmd can get notified when the ring/task is dying if
io_uring_cmd_mark_cancelable() is called on the command.


Thanks,
Ming
Pavel Begunkov Nov. 3, 2024, 10:01 p.m. UTC | #2
On 10/31/24 16:32, Mark Harmstone wrote:
> From: Pavel Begunkov <asml.silence@gmail.com>
> 
> When the taks that submitted a request is dying, a task work for that
> request might get run by a kernel thread or even worse by a half
> dismantled task. We can't just cancel the task work without running the
> callback as the cmd might need to do some clean up, so pass a flag
> instead. If set, it's not safe to access any task resources and the
> callback is expected to cancel the cmd ASAP.

I was just going to write that you didn't CC io_uring for the
rest of the series, but I can't find it in the btrfs list, did
did something go wrong?

Regardless, I think it should be fine to merge it through
the btrfs tree
Pavel Begunkov Nov. 4, 2024, 12:13 a.m. UTC | #3
On 10/31/24 16:32, Mark Harmstone wrote:
...
>   struct io_wq_work_node {
> diff --git a/io_uring/uring_cmd.c b/io_uring/uring_cmd.c
> index 39c3c816ec78..78a8ba5d39ae 100644
> --- a/io_uring/uring_cmd.c
> +++ b/io_uring/uring_cmd.c
> @@ -119,9 +119,13 @@ EXPORT_SYMBOL_GPL(io_uring_cmd_mark_cancelable);
>   static void io_uring_cmd_work(struct io_kiocb *req, struct io_tw_state *ts)
>   {
>   	struct io_uring_cmd *ioucmd = io_kiocb_to_cmd(req, struct io_uring_cmd);
> +	unsigned int flags = IO_URING_F_COMPLETE_DEFER;
> +
> +	if (req->task != current)
> +		flags |= IO_URING_F_TASK_DEAD;

Actually, the patch is broken. I have no clue why Bernd changed
it and left my name, here was the original:

https://lore.kernel.org/io-uring/d2528a1c-3d7c-4124-953c-02e8e415529e@gmail.com/

also now it'll also conflict with patches Jens is queuing, I'll send
a replacement btrfs can take.
diff mbox series

Patch

diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h
index 4b9ba523978d..2ee5dc105b58 100644
--- a/include/linux/io_uring_types.h
+++ b/include/linux/io_uring_types.h
@@ -37,6 +37,7 @@  enum io_uring_cmd_flags {
 	/* set when uring wants to cancel a previously issued command */
 	IO_URING_F_CANCEL		= (1 << 11),
 	IO_URING_F_COMPAT		= (1 << 12),
+	IO_URING_F_TASK_DEAD		= (1 << 13),
 };
 
 struct io_wq_work_node {
diff --git a/io_uring/uring_cmd.c b/io_uring/uring_cmd.c
index 39c3c816ec78..78a8ba5d39ae 100644
--- a/io_uring/uring_cmd.c
+++ b/io_uring/uring_cmd.c
@@ -119,9 +119,13 @@  EXPORT_SYMBOL_GPL(io_uring_cmd_mark_cancelable);
 static void io_uring_cmd_work(struct io_kiocb *req, struct io_tw_state *ts)
 {
 	struct io_uring_cmd *ioucmd = io_kiocb_to_cmd(req, struct io_uring_cmd);
+	unsigned int flags = IO_URING_F_COMPLETE_DEFER;
+
+	if (req->task != current)
+		flags |= IO_URING_F_TASK_DEAD;
 
 	/* task_work executor checks the deffered list completion */
-	ioucmd->task_work_cb(ioucmd, IO_URING_F_COMPLETE_DEFER);
+	ioucmd->task_work_cb(ioucmd, flags);
 }
 
 void __io_uring_cmd_do_in_task(struct io_uring_cmd *ioucmd,