On Thu, Jun 23, 2022 at 01:59:12PM +0800, Duoming Zhou wrote:
> There are sleep in atomic context bugs when dm_fsync_timer_callback is
> executing. The root cause is that the memory allocation functions with
> GFP_KERNEL or GFP_NOIO parameters are called in dm_fsync_timer_callback
> which is a timer handler. The call paths that could trigger bugs are
> shown below:
>
> (interrupt context)
> dm_fsync_timer_callback
> write_nic_byte
> kzalloc(sizeof(data), GFP_KERNEL); //may sleep
> usb_control_msg
> kmalloc(.., GFP_NOIO); //may sleep
> write_nic_dword
> kzalloc(sizeof(data), GFP_KERNEL); //may sleep
> usb_control_msg
> kmalloc(.., GFP_NOIO); //may sleep
>
> This patch uses delayed work to replace timer and moves the operations
> that may sleep into the delayed work in order to mitigate bugs.
>
> Fixes: 8fc8598e61f6 ("Staging: Added Realtek rtl8192u driver to staging")
> Signed-off-by: Duoming Zhou <duoming@xxxxxxxxxx>
> ---
> Changes in v2:
> - Use delayed work to replace timer.
Did you test this with real hardware to verify it still works?
thanks,
greg k-h
