Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 15 Aug 1998 23:30:15 -0500
From:      Karl Denninger  <karl@mcs.net>
To:        Terry Lambert <tlambert@primenet.com>
Cc:        dg@root.com, current@FreeBSD.ORG
Subject:   Re: Better VM patches (was Tentative fix for VM bug)
Message-ID:  <19980815233015.14991@mcs.net>
In-Reply-To: <199808160113.SAA17303@usr07.primenet.com>; from Terry Lambert on Sun, Aug 16, 1998 at 01:13:33AM %2B0000
References:  <199808152349.QAA20449@implode.root.com> <199808160113.SAA17303@usr07.primenet.com>

next in thread | previous in thread | raw e-mail | index | archive | help
This has had an effect, in that it has changed behavior.

It has not completely made the problem go away, as I managed to provoke it
about an hour after I loaded these patches.

However, it has had a significantly positive effect.

I am NOT running with DIAGNOSTIC, so I won't get the panic (I don't want
it!)  

More on this tomorrow when its been up for a day or so.

--
-- 
Karl Denninger (karl@MCS.Net)| MCSNet - Serving Chicagoland and Wisconsin
http://www.mcs.net/          | T1's from $600 monthly / All Lines K56Flex/DOV
			     | NEW! Corporate ISDN Prices dropped by up to 50%!
Voice: [+1 312 803-MCS1 x219]| EXCLUSIVE NEW FEATURE ON ALL PERSONAL ACCOUNTS
Fax:   [+1 312 803-4929]     | *SPAMBLOCK* Technology now included at no cost

On Sun, Aug 16, 1998 at 01:13:33AM +0000, Terry Lambert wrote:
> > >yes, but is it a bug that he's reporting?
> > 
> >    I don't see anything wrong with vm_object_page_remove(). We'd have
> > far more serious problems if there was something wrong with it (the system
> > wouldn't last for more than a few seconds, if that), and further, the
> > problem wouldn't be specific to just NFS. My thinking at the moment is that
> > the problem is caused by the lack of NFSnode locking; we depend on VOP_LOCK
> > actually doing something in many parts of the code, and with nfs_lock()
> > being a noop, it just doesn't happen. There might be a way to fix this
> > without messing with nfs_lock(), but the exact failure scenario will need
> > to be looked at carefully before this can be determined.
> 
> Let me prefix this by saying that your NFS fixes are correct, and
> that my "backup patch" apparently only masks the EOF bug that
> results from bogus invalidations, and there's a better fix.

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980815233015.14991>