Re: task btrfs-cleaner:770 blocked for more than 120 seconds.

[Liu Bo <bo.li.liu@oracle.com>]

Hi Mike,
On Sun, Feb 14, 2016 at 12:23:16AM +0500, Михаил Гаврилов wrote:
> Sorry, I have not yet had time to apply your patch.
> 
> And get hang again when launch web browser.
> 
> Here new logs: http://btrfs.sy24.ru/kernel-sysrqw-btrfscleaner770blocked-3.txt

The logs show,
every hung process is waiting for fs tree root's WRITE_LOCK, but someone who's
holding the WRITE_LOCK is not reflected here.

Only one suspicious process is 
"Chrome_FileThre D ffff8807fe1d7a18 10360 22486   2826 0x00000000"
it's also waiting on btrfs_tree_lock(), but I can't tell whether it
waits for the fs tree root node's lock or the children node's lock.

Hmm, sorry for not being helpful, but I was wondering if this also
occurs on the latest btrfs(4.5) or is it possible to give it a shot?

The best result is that it's been fixed by one commit in 4.5.

> 
> Addition info: when btrfs partition hangs also hang GUI it is normal?
> For grab logs I switch to tty3 (Ctrl - Alt - F3)

I'm afraid the answer is yes because GUI needs to read files (like
config files) from the underlying partition.

Thanks,

-liubo

> 
> 
> 
> --
> Best Regards,
> Mike Gavrilov.
> 
> 
> 2016-02-13 1:34 GMT+05:00 Liu Bo <bo.li.liu@oracle.com>:
> > On Sat, Feb 13, 2016 at 12:15:08AM +0500, Михаил Гаврилов wrote:
> >> 2016-02-12 8:22 GMT+05:00 Liu Bo <bo.li.liu@oracle.com>:
> >> > You can try it on your 4.2.3 kernel or the latest 4.5, but I guess it
> >> > doesn't not fix the real deadlock you're hitting...
> >>
> >> It means it is not final patch? you continue investigate problem? Can
> >> I help you?
> >
> > Yeah, it's not the final patch, you can apply it and see if the deadlock
> > will happen again.
> >
> > Unless we can get a way to reproduce it, I'm afraid there's little things we can  do here.
> >
> > Thanks,
> >
> > -liubo