On 7/22/19 6:06 PM, Ming Lei wrote:
On Mon, Jul 22, 2019 at 08:25:07AM -0700, Bart Van Assche wrote:
On 7/21/19 10:39 PM, Ming Lei wrote:
blk-mq may schedule to call queue's complete function on remote CPU via
IPI, but doesn't provide any way to synchronize the request's complete
fn.
In some driver's EH(such as NVMe), hardware queue's resource may be freed &
re-allocated. If the completed request's complete fn is run finally after the
hardware queue's resource is released, kernel crash will be triggered.
Prepare for fixing this kind of issue by introducing
blk_mq_tagset_wait_completed_request().
An explanation is missing of why the block layer is modified to fix this
instead of the NVMe driver.
The above commit log has explained that there isn't sync mechanism in
blk-mq wrt. request completion, and there might be similar issue in other
future drivers.
That is not sufficient as a motivation to modify the block layer because
there is already a way to wait until request completions have finished,
namely the request queue freeze mechanism. Have you considered to use
that mechanism instead of introducing
blk_mq_tagset_wait_completed_request()?
Thanks,
Bart.
