Date: Sun, 23 Sep 2012 16:19:13 -0700 (PDT) From: Pedro Giffuni <pfg@freebsd.org> To: David Wolfskill <david@catwhisker.org>, Konstantin Belousov <kostikbel@gmail.com> Cc: "stable@freebsd.org" <stable@freebsd.org> Subject: Re: rpc.lockd exiting just after startup @r240811 Message-ID: <1348442353.1084.YahooMailNeo@web113510.mail.gq1.yahoo.com> In-Reply-To: <20120923225021.GT30766@albert.catwhisker.org> References: <20120923130414.GQ30766@albert.catwhisker.org> <20120923154349.GM37286@deviant.kiev.zoral.com.ua> <20120923225021.GT30766@albert.catwhisker.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Hello David;=0A=0ASorry to have caused you trouble.=0A=0A----- Original Mes= sage -----=0A> From: David Wolfskill <david@catwhisker.org>=0A...=0A> =0A> = On Sun, Sep 23, 2012 at 06:43:49PM +0300, Konstantin Belousov wrote:=0A>> = ...=0A>> Try to revert the r240799. If this does not help, then some diggi= ng with=0A>> gdb would be needed to see why kernel dislikes the buffer. We= ll, the same=0A>> digging would be needed even if the revert helps.=0A> = =0A> Thanks for the suggestiion.=0A> =0A> As noted earlier, I first tried a= rebuild (not specifying -DNOCLEAN);=0A> that proved ineffective.=0A> =0A> = I then reverted r240799, and rpc.lockd runs; I am now able to issue=0A> svn= commands against an NFS-resident working copy.=A0 The procedure=0A> I used= (in case I screwed up -- or there's a less inelegant way to=0A> do it) was= :=0A> =0A> cd /usr/src && \=0A> svn diff -c 240799 >/tmp/d && \=0A> svn pat= ch --reverse-diff /tmp/d=0A> =0A> I then performed:=0A> =0A> make -DNOCLEAN= -j 4 buildworld && make -DNOCLEAN -j 4 kernel=0A> =0A> The machine is now = running:=0A> =0A> FreeBSD freebeast.catwhisker.org 9.1-PRERELEASE FreeBSD 9= .1-PRERELEASE #478 =0A> 240856M: Sun Sep 23 15:02:33 PDT 2012=A0 =A0 =0A> r= oot@freebeast.catwhisker.org:/usr/obj/usr/src/sys/GENERIC=A0 i386=0A> =0A> = =0A> I have subsequently seen similar symptoms on my work desktop, which=0A= > is running stable/9 as of r240849.=0A> =0A> =0A> Since rpc.lockd seems OK= on the systems running head, perhaps there=0A> was an issue with the MFC?= =0A>=A0=0A=0ABoth should have the same issue. It does depend on what kind o= f load=0Ayour -current box is running, of course.=0A=0A> I'll be happy to t= est further -- I just need a bit of guidance from=0A> time to time. :-}=0A>= =A0=0A=0A=0A=0AWhat I can do is send you individual patches to see if we=0A= can narrow down the culprit. If I can avoid reverting all the=0Achange it w= ill be better.=0A=0AIt will be a bit boring because you may have to build k= ernel=0Aand libc each time, so I would start with the most obvious=0Acandid= ates. For now just leave r240799 reverted.=0A=0APedro.=0A
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1348442353.1084.YahooMailNeo>