Date: Fri, 14 Apr 2017 19:32:33 -0700 From: Peter Wemm <peter@wemm.org> To: Maxim Sobolev <sobomax@freebsd.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers <src-committers@freebsd.org>, Hiren Panchasara <hiren@freebsd.org> Subject: Re: svn commit: r316874 - head/sys/kern Message-ID: <6546471.KH0i3bg3hj@overcee.wemm.org> In-Reply-To: <CAH7qZftToe23-E05JVV0CUmFV2bu2zFuE_O6rmFObYbTdb1h4A@mail.gmail.com> References: <201704141723.v3EHNS3B043902@repo.freebsd.org> <6942092.OsECkgAp13@overcee.wemm.org> <CAH7qZftToe23-E05JVV0CUmFV2bu2zFuE_O6rmFObYbTdb1h4A@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--nextPart2010007.ntouo2SN7c
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset="us-ascii"
On Friday, April 14, 2017 02:21:05 PM Maxim Sobolev wrote:
> Peter, It is actually the other way around. If you take syslogd code =
out of
> 11-stable and earlier that would work just fine with or without r3168=
74.
> But since r285910 syslogd in head had been refactored a lot and I thi=
nk
> that particular bug has introduced that has been masked by the shutdo=
wn()
> on datagram sockets becoming a NOP after r285910. Then r316874 restor=
ed our
> historical behavior for the shutdown(2) and bingo, bug in the new sys=
logd
> code is now causing it to spin when shutdown() !=3D NOP.
>=20
> -Max
Hmm, there's a new problem:
45104 auditd CALL socket(PF_LOCAL,0x10000002<SOCK_DGRAM|SOCK_CLOEXE=
C>,0)
45104 auditd RET socket 3
45104 auditd CALL connect(0x3,0x7fffffffdbd8,0x6a)
45104 auditd STRU struct sockaddr { AF_LOCAL, /var/run/logpriv }
45104 auditd NAMI "/var/run/logpriv"
45104 auditd RET connect 0
45104 auditd CALL sendto(0x3,0x7fffffffe130,0x2f,0,0,0)
45104 auditd RET sendto -1 errno 55 No buffer space available
45104 auditd CALL open(0x800da5c67,0x100005<O_WRONLY|O_NONBLOCK|
O_CLOEXEC>)
45104 auditd NAMI "/dev/console"
45104 auditd RET open 4
.. and it all goes to /dev/console instead.
On restarting syslogd:
Apr 15 02:17:43 repoman2 syslogd: exiting on signal 15
sonewconn: pcb 0xfffff80051e72680: Listen queue overflow: 16 already in=
queue=20
already
Umm.. did the patch forget to listen to incoming connections or somethi=
ng?
I haven't seen this before anywhere except when syslogd is wedged.
>=20
> On Fri, Apr 14, 2017 at 12:46 PM, Peter Wemm <peter@wemm.org> wrote:
> > On Friday, April 14, 2017 12:41:52 PM Maxim Sobolev wrote:
> > > Thanks, Peter. I will try to look into this asap.
> >=20
> > I don't understand what is going on yet. Presumably there must be o=
ther
> > changes in play that affect udp/select sometime between the origina=
l 2015
> > change and this. The syslogd -s code is Old(TM). I'm also wonderin=
g
> > whether
> > the -s code even works at all since the 2015 / r285910 change...
> >=20
> > > -Max
> > >=20
> > > On Apr 14, 2017 12:32 PM, "Peter Wemm" <peter@wemm.org> wrote:
> > > > On Friday, April 14, 2017 11:49:26 AM Peter Wemm wrote:
> > > > > On Friday, April 14, 2017 05:23:28 PM Maxim Sobolev wrote:
> > > > > > Author: sobomax
> > > > > > Date: Fri Apr 14 17:23:28 2017
> > > > > > New Revision: 316874
> > > > > > URL: https://svnweb.freebsd.org/changeset/base/316874
> > > > > >=20
> > > > > > Log:
> > > > > > Restore ability to shutdown DGRAM sockets, still forcing
> >=20
> > ENOTCONN to
> >=20
> > > > be
> > > >=20
> > > > > > returned by the shutdown(2) system call. This ability has b=
een
> >=20
> > lost as
> >=20
> > > > > > part
> > > > > > of the svn revision 285910.
> > > > > >=20
> > > > > > Reviewed by: ed, rwatson, glebius, hiren
> > > > > > MFC after: 2 weeks
> > > > > > Differential Revision: https://reviews.freebsd.org/D10=
351
> > > > >=20
> > > > > This appears to have broken syslogd and had a major change in=
> >=20
> > behavior
> >=20
> > > > with
> > > >=20
> > > > > regards to select(2).
> > > > >=20
> > > > > If you run syslogd with the -s flag (which is default), it no=
w spins
> >=20
> > at
> >=20
> > > > 100%
> > > >=20
> > > > > cpu as all the shutdown sockets now return readable from sele=
ct.
> > > > >=20
> > > > > Old releases / jails also manifest this behavor. If it wasn'=
t for
> > > > > losing
> > > > > the ability to run old branch binaries I'd suggest changing s=
yslogd
> > > > > instead, but we depend on this in the cluster and I expect ot=
hers do
> > > > > too.
> > > > >=20
> > > > > I'm not 100% certain that this change is the culprit but a he=
ads-up
> > > > > can't
> > > > > hurt. I'll try reverting it on the freebsd cluster next, afte=
r
> > > > > fixing
> > > > > the
> > > > > broken auditing changes.
> > > > >=20
> > > > > -Peter
> > > >=20
> > > > I can confirm that reverting r316874 fixes syslogd and backward=
s
> > > > compatability
> > > > with old branches.
> > > >=20
> > > > --
> > > > Peter Wemm - peter@wemm.org; peter@FreeBSD.org; peter@yahoo-inc=
.com;
> > > > KI6FJV
> > > > UTF-8: for when a ' or ... just won\342\200\231t do\342\200\246=
> >=20
> > --
> > Peter Wemm - peter@wemm.org; peter@FreeBSD.org; peter@yahoo-inc.com=
;
> > KI6FJV
> > UTF-8: for when a ' or ... just won\342\200\231t do\342\200\246
=2D-=20
Peter Wemm - peter@wemm.org; peter@FreeBSD.org; peter@yahoo-inc.com; KI=
6FJV
UTF-8: for when a ' or ... just won\342\200\231t do\342\200\246
--nextPart2010007.ntouo2SN7c
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: This is a digitally signed message part.
Content-Transfer-Encoding: 7Bit
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEEBgrA0Vr/vfNVuPoUNdaXCeyAngQFAljxhkEACgkQNdaXCeyA
ngSYuwgAiYaKQFKr1C2yjuahP9uc782BCIvuvn7w+Eb+WzfU966Jh87jQlmzswxT
ukf1hVk96LPALs55NqX6Z1ZfZ2oa0YN92dZ1px0+AryrAx+CPpgYYe+2Yrf3vYHu
i84Z6ZgZjPb0Uf5Nmk7l8ypWHNxvCHZoyJDvFzwFbMECVrur1r6gYY76M3YCR2UK
xQsKIrypebjc8j0Qe+GPLCOpaibYXRSA0Vl63pyp05aEf9KVN5PWjwmlnRHFNSh3
DH6O5vucb5IugU88FVNoxBJX8LmoRy6JI/ewyuDphKgquu45gRxFffaTAsO9wG/Y
fNr5yzkas9EbmwLaTjOiOsh6/u4Jyg==
=KM3L
-----END PGP SIGNATURE-----
--nextPart2010007.ntouo2SN7c--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?6546471.KH0i3bg3hj>