Hello, On Thu, May 12, 2022 at 08:18:24PM +0900, Byungchul Park wrote: > > 1. wait_for_completion_killable_timeout() doesn't need someone to wake it up > > to make forward progress because it will unstick itself after timeout > > expires. > > I have a question about this one. Yes, it would never been stuck thanks > to timeout. However, IIUC, timeouts are not supposed to expire in normal > cases. So I thought a timeout expiration means not a normal case so need > to inform it in terms of dependency so as to prevent further expiraton. > That's why I have been trying to track even timeout'ed APIs. > > Do you think DEPT shouldn't track timeout APIs? If I was wrong, I > shouldn't track the timeout APIs any more. Without actually surveying the use cases, I can't say for sure but my experience has been that we often get pretty creative with timeouts and it's something people actively think about and monitor (and it's usually not subtle). Given that, I'm skeptical about how much value it'd add for a dependency checker to warn about timeouts. It might be net negative than the other way around. > > 2. complete_all() from __fw_load_abort() isn't the only source of wakeup. > > The fw loader can be, and mainly should be, woken up by firmware loading > > actually completing instead of being aborted. > > This is the point I'd like to ask. In normal cases, fw_load_done() might > happen, of course, if the loading gets completed. However, I was > wondering if the kernel ensures either fw_load_done() or fw_load_abort() > to be called by *another* context while kernel_halt(). We'll have to walk through the code to tell that. On a cursory look tho, up until that point (just before shutting down usermode helper), I don't see anything which would actively block firmware loading. Thanks. -- tejun
