2021-05-13 20:51:52

by Maximilian Luz

Subject: [PATCH] platform/surface: dtx: Fix poll function

The poll function should not return -ERESTARTSYS.

Furthermore, locking in this function is completely unnecessary. The
ddev->lock protects access to the main device and controller (ddev->dev
and ddev->ctrl), ensuring that both are and remain valid while being
accessed by clients. Both are, however, never accessed in the poll
function. The shutdown test (via atomic bit flags) be safely done
without locking, so drop locking here entirely.

Reported-by: kernel test robot <[email protected]>
Fixes: 1d609992832e ("platform/surface: Add DTX driver)
Signed-off-by: Maximilian Luz <[email protected]>
---
drivers/platform/surface/surface_dtx.c | 8 +-------
1 file changed, 1 insertion(+), 7 deletions(-)

diff --git a/drivers/platform/surface/surface_dtx.c b/drivers/platform/surface/surface_dtx.c
index 63ce587e79e3..5d9b758a99bb 100644
--- a/drivers/platform/surface/surface_dtx.c
+++ b/drivers/platform/surface/surface_dtx.c
@@ -527,20 +527,14 @@ static __poll_t surface_dtx_poll(struct file *file, struct poll_table_struct *pt
struct sdtx_client *client = file->private_data;
__poll_t events = 0;

- if (down_read_killable(&client->ddev->lock))
- return -ERESTARTSYS;
-
- if (test_bit(SDTX_DEVICE_SHUTDOWN_BIT, &client->ddev->flags)) {
- up_read(&client->ddev->lock);
+ if (test_bit(SDTX_DEVICE_SHUTDOWN_BIT, &client->ddev->flags))
return EPOLLHUP | EPOLLERR;
- }

poll_wait(file, &client->ddev->waitq, pt);

if (!kfifo_is_empty(&client->buffer))
events |= EPOLLIN | EPOLLRDNORM;

- up_read(&client->ddev->lock);
return events;
}

--
2.31.1

2021-05-19 21:13:27

by Hans de Goede

Subject: Re: [PATCH] platform/surface: dtx: Fix poll function

Hi,

On 5/13/21 3:44 PM, Maximilian Luz wrote:
> The poll function should not return -ERESTARTSYS.
>
> Furthermore, locking in this function is completely unnecessary. The
> ddev->lock protects access to the main device and controller (ddev->dev
> and ddev->ctrl), ensuring that both are and remain valid while being
> accessed by clients. Both are, however, never accessed in the poll
> function. The shutdown test (via atomic bit flags) be safely done
> without locking, so drop locking here entirely.
>
> Reported-by: kernel test robot <[email protected]>
> Fixes: 1d609992832e ("platform/surface: Add DTX driver> Signed-off-by: Maximilian Luz <[email protected]>

Thank you for your patch, I've applied this patch to my review-hans
branch:
https://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86.git/log/?h=review-hans

I will also include this in the next pdx86-fixes pull-req for 5.13.

Note it will show up in my review-hans branch once I've pushed my
local branch there, which might take a while.

Once I've run some tests on this branch the patches there will be
added to the platform-drivers-x86/for-next branch and eventually
will be included in the pdx86 pull-request to Linus for the next
merge-window.

Regards,

Hans

> ---
> drivers/platform/surface/surface_dtx.c | 8 +-------
> 1 file changed, 1 insertion(+), 7 deletions(-)
>
> diff --git a/drivers/platform/surface/surface_dtx.c b/drivers/platform/surface/surface_dtx.c
> index 63ce587e79e3..5d9b758a99bb 100644
> --- a/drivers/platform/surface/surface_dtx.c
> +++ b/drivers/platform/surface/surface_dtx.c
> @@ -527,20 +527,14 @@ static __poll_t surface_dtx_poll(struct file *file, struct poll_table_struct *pt
> struct sdtx_client *client = file->private_data;
> __poll_t events = 0;
>
> - if (down_read_killable(&client->ddev->lock))
> - return -ERESTARTSYS;
> -
> - if (test_bit(SDTX_DEVICE_SHUTDOWN_BIT, &client->ddev->flags)) {
> - up_read(&client->ddev->lock);
> + if (test_bit(SDTX_DEVICE_SHUTDOWN_BIT, &client->ddev->flags))
> return EPOLLHUP | EPOLLERR;
> - }
>
> poll_wait(file, &client->ddev->waitq, pt);
>
> if (!kfifo_is_empty(&client->buffer))
> events |= EPOLLIN | EPOLLRDNORM;
>
> - up_read(&client->ddev->lock);
> return events;
> }
>
>