On Mon, 3 Feb 2025 at 16:23, Arnd Bergmann arnd@arndb.de wrote:
On Mon, Feb 3, 2025, at 10:31, Jens Wiklander wrote:
On Mon, Feb 3, 2025 at 9:00 AM Sumit Garg sumit.garg@linaro.org wrote:
Why not mutex_lock()? If we fail to acquire the mutex here, we will quite likely free the req list item below at the end of this function while it remains in the list.
Right, I had mentioned mutex_lock_killable in an earlier reply, as I didn't know exactly where it hang. If we know that the wait_event_interruptible() was causing the hang, then the normal mutex_lock should be fine.
Yeah for my current test scenario mutex_lock() works fine too but I added mutex_lock_killable() just in another theoretical corner case being stuck acquiring mutex while someone is trying to kill the process. But let's avoid over complexity for the time being with a simpler mutex_lock() approach.
-Sumit