Re: showstopper race condition in sync() ??? [2.1.119|120]

Bill Hawes (whawes@star.net)
Fri, 18 Sep 1998 07:33:47 -0400


Cyrille Chepelov wrote:

> Bill: Great intuition !!
>
> Stephen: it stalled in wait_on_dquota(). Among other oddities :
> * top reports the machine is 98% idle, but the loadavg rises close
> to 1.0.
> * My only quota-enabled partitions are /home and /var/spool/mail.
> / is not.
>
> I'll try to write a fix monday, but I'd not count on my own code for
> kernel matters <grin />

Well, I cheated -- I looked at the dquot code after your second posting
mentioning quota. The dquot code is full of race-conditions in its list
management, as it's basically lifted from the early 2.0.xx inode code,
and that had many problems.

It really needs to be rewritten, though it may be possible to give it a
band-aid for the short term.

Regards,
Bill

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/