Messages in this thread | | | From | YE Chengfeng <> | Subject | Re: [PATCH] gpio: gpiolib-cdev: Fix potential &lr->wait.lock deadlock issue | Date | Mon, 26 Jun 2023 16:51:30 +0000 |
| |
> Yeah, I'm not sure this is correct. edge_irq_thread() runs in process > > context, so the whole premise of the patch seems to be flawed. What > tool reported this? Can this be a false positive? Have you seen this > > happen in real life?
It seems like it’s my mistake, I misidentify thread_irq as runing inside irq context. Then it’s truth that the patch is unnecessary.
The static analysis tool is built by me, I noticed lockdep occasionally reports such similar deadlocks in other place thus intent to build a static tool to locate such bugs. It has detected other true bugs these days but it’s a pity this one is really a false positive.
Indeed thanks so much for your time and sorry for bothering you with the false report.
Thanks again, Chengfeng | |