Re: [syzbot] [netfs?] INFO: task hung in netfs_unbuffered_write_iter

[Oleg Nesterov]

David,

I know that I will regret my email tomorrow, but

On 03/28, David Howells wrote:
>
> #syz test: git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git netfs-fixes
>
>   There's inconsistency with some wakeups between using rreq->waitq (a private
>   waitqueue) and using clear_and_wake_up_bit() (a shared global waitqueue).

Too late for me, I didn't even try to read the "netfs-fixes" changes.
And if I even tried, I probably wouldn't understand them ;)

But, afaics (I can be easily wrong), the curent logic in net/9p/ doesn't look
right regardless of any other fixes, no?

Oleg.