From owner-freebsd-stable Wed Mar 6 13:30:58 2002 Delivered-To: freebsd-stable@freebsd.org Received: from salmon.maths.tcd.ie (salmon.maths.tcd.ie [134.226.81.11]) by hub.freebsd.org (Postfix) with SMTP id 8B5BF37B416 for ; Wed, 6 Mar 2002 13:30:45 -0800 (PST) Received: from walton.maths.tcd.ie by salmon.maths.tcd.ie with SMTP id ; 6 Mar 2002 21:30:44 +0000 (GMT) To: "Jan L. Peterson" Cc: Matthew Dillon , freebsd-stable@FreeBSD.ORG Subject: Re: crashes on 4.5-RELEASE In-Reply-To: Your message of "Wed, 06 Mar 2002 12:11:09 MST." <20020306191109.5FF92422D8@mail.flipdog.com> Date: Wed, 06 Mar 2002 21:30:43 +0000 From: Ian Dowse Message-ID: <200203062130.aa57106@salmon.maths.tcd.ie> Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG In message <20020306191109.5FF92422D8@mail.flipdog.com>, "Jan L. Peterson" writ es: >I have not had a crash since applying this patch: > >- if (space < resid + clen && uio && >+ if (space < resid + clen && >even with INVARIANTS turned off. Nor have I seen an xl0 watchdog >timeout. I actually doubt that this patch could help with the panics you were seeing. The only problem this is known to fix is one where you have an NFS filesystem mounted, and you load an IPFW rule that blocks outgoing NFS packets. In that situation it causes the system to run out of mbufs which could result in a NULL-dereference in the NFS code. You should have seen errors like "m_retry failed, consider increase mbuf value" and much lower fault virtual addresses. The fact that turning on INVARIANTS made the crashes go away is much more interesting - is it possible that you might have been using some kernel module that didn't match the kernel, and that when recompiling the kernel with INVARIANTS you effectively fixed this? Are you using any kernel modules? Check the output of kldstat, as some ports/packages install modules, and these don't get updated automatically when you buildworld. A useful experiment would be to turn off INVARIANTS, undo this patch, and check that the panics still occur. It's quite possible that there is a hard-to-find bug that INVARIANTS or the sosend patch somehow disguise, but it's worth ruling out simpler explanations first! Ian To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message