xfs
[Top] [All Lists]

Re: Work Items for XFS port (UPDATED)

To: "Andi Kleen" <ak@xxxxxxx>
Subject: Re: Work Items for XFS port (UPDATED)
From: Steve Lord <lord@xxxxxxx>
Date: Thu, 27 Apr 2000 17:28:16 -0500
Cc: Jim Mostek <mostek@xxxxxxx>, linux-xfs@xxxxxxxxxxx
In-reply-to: Message from "Andi Kleen" <ak@suse.de> of "Fri, 28 Apr 2000 00:21:15 +0200." <20000428002115.A8685@gruyere.muc.suse.de>
Sender: owner-linux-xfs@xxxxxxxxxxx
> On Thu, Apr 27, 2000 at 04:55:16PM -0500, Jim Mostek wrote:
> > If you see any problems or want more items added, let me know.
> 
> XFS/pagebuf never checks for Linux signals ATM when it sleeps. I guess 
> Irix signal handling is a bit different. It should be probably fixed though.
> Currently e.g. find /xfs is very hard to Ctrl-C. 
> I started with adding a few signal checks, but it seems to require
> a lot more work.
> 
> -Andi

How is this different from other filesystems which will sit in wait_on_buffer
using TASK_UNINTERRUPTIBLE? My finds interrupt very quickly.

I did try out a couple of things after you last mentioned this, but I never
managed to find anything which was slow to interrupt. This rather implies
that you have a very long running system call stuck in xfs, I suspect the
problem is not so much that we do not pay attention to interrupts, but that
something is taking much longer than it should. It is possible you found
another case of the missing run_task_queue(&tq_disk). Do things behave
better if you have activity in other filesystems on going at the same time?

Steve








<Prev in Thread] Current Thread [Next in Thread>