On Tue, Mar 25, 2025 at 02:04:07PM +0000, Emmanuel Nyarko wrote: > So then I guess we will need to init rv to 0.
The ixl_other_intr() handler seems to hard code the assumption that the interrupt is to be handled by it, unconditionally clears all enabled interrupts, calls ixl_flush() etc. This needs more fixing than just initializing rv = 0. Maybe the handler is even correct and the interrupt can not come from any other device, so always return 1 would be correct - but then it needs a comment explaining why this is the case. I wouldn't touch it and find someone who can actually test changes, maybe file a PR instead and/or ping Yamaguchi-san. Martin
