Date: Mon, 3 Apr 2006 16:34:03 -0400 From: Kris Kennaway <kris@obsecurity.org> To: Peter Holm <peter@holm.cc> Cc: truckman@freebsd.org, current@freebsd.org Subject: Re: Livelock / softdep_flush "loop" Message-ID: <20060403203402.GA89233@xor.obsecurity.org> In-Reply-To: <20060403192808.GA91475@peter.osted.lan> References: <20060402094431.GA81954@peter.osted.lan> <20060402.180153.74658240.Tor.Egge@cvsup.no.freebsd.org> <20060402190910.GA12759@peter.osted.lan> <20060402.204851.78799557.Tor.Egge@cvsup.no.freebsd.org> <20060403192808.GA91475@peter.osted.lan>
next in thread | previous in thread | raw e-mail | index | archive | help
[-- Attachment #1 --] On Mon, Apr 03, 2006 at 09:28:08PM +0200, Peter Holm wrote: > On Sun, Apr 02, 2006 at 08:48:51PM +0000, Tor Egge wrote: > > > > The enclosed patch might help. > > > > > > > > > > I'm testing it right now. > > > > Similarly to the softdepflush process looping in softdep_flush(), the process > > creating a snapshot might loop in ffs_sync(). > > > > I ran the same test with your first patch for two hours and your > second patch for 7 hours, without seeing any livelocks. > > I added snapshots to the test and still did not get any livelocks, but > I'm not certain that this last test covers your change. Hmm, there is a PR reporting hangs when removing many files with snapshots enabled, on 6.0. I tried to reproduce it on 6.1 but failed with the test case that the PR said was a reliable cause for them, so I assumed it was fixed. It may be that I didn't try hard enough though, since I didn't see your problem either. Kris [-- Attachment #2 --] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.2 (FreeBSD) iD8DBQFEMYa6Wry0BWjoQKURAr6DAKD5iGz6aYO5LNPlgFj4RixFonaJfQCgt3S3 gut18DnWCkJsG6rmLzp6wx8= =hHdh -----END PGP SIGNATURE-----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060403203402.GA89233>