http://bugzilla.novell.com/show_bug.cgi?id=133718
User jack@novell.com added comment
http://bugzilla.novell.com/show_bug.cgi?id=133718#c28
--- Comment #28 from Jan Kara 2009-07-13 11:29:07 MDT ---
I've forgotten about the bug as well :).
Your patch looks good. I'm only afraid it might unveil some bugs in filesystems
because with your patch cleaning of filesystem inodes can happen from several
threads in parallel. But those can be probably fixed.
Also your patch does something a bit different from what I had in mind - my
idea was that if some filesystem gets congested, then we just skip cleansing
inodes there and hopefully make progress somewhere else (so something like you
wrote in Comment #25). That would IMHO also make some sence. As far as UDF is
concerned, I'll probably move it from doing any IO from udf_clear_inode anyway
because that turned out to be a bad idea for several reasons but there are
other filesystems which do non-trivial stuff in their clear_inode methods (e.g.
cluster filesystems have to do some network communication before they can
release the inode) and they are going to have similar problems as UDF if the
network gets slow...
--
Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.