1 Oct 2006 02:40
Re: kern/103256: [em] em0: watchdog timeout -- resetting (6.1-STABLE)
Robert Edmonds <robert.edmonds <at> gmail.com>
2006-10-01 00:40:25 GMT
2006-10-01 00:40:25 GMT
The following reply was made to PR kern/103256; it has been noted by GNATS.
From: Robert Edmonds <robert.edmonds <at> gmail.com>
To: bug-followup <at> FreeBSD.org, imriz <at> co.zahav.net.il
Cc:
Subject: Re: kern/103256: [em] em0: watchdog timeout -- resetting (6.1-STABLE)
Date: Sat, 30 Sep 2006 20:38:27 -0400
Hello,
I can confirm this problem on an SMP amd64 machine (Core 2 Duo,
Supermicro PDSMI+ motherboard) with built-in Intel 82573V and 82573E
PCI-Express gigabit NICs and an Intel 82545GM PCI-X gigabit card.
Both 6.1 and 6.2-BETA1 exhibit the same behavior: starting a download
(e.g., portsnap fetch) is enough to trigger the timeout and a loss of
connectivity on the PCI-E NICs, but not the PCI-X one.
The 82545GM card works fine without debug.mpsafenet=0, but the 82573s
will only work with debug.mpsafenet=0. I have several of these servers
which are not in production, so I can test whatever patches may be
necessary.
Output of pciconf -lv and dmesg follow.
hostb0 <at> pci0:0:0: class=0x060000 card=0x798015d9 chip=0x27788086 rev=0xc0 hdr=0x00
vendor = 'Intel Corporation'
device = 'Server Memory Controller Hub'
class = bridge
subclass = HOST-PCI
(Continue reading)
The fix proposed in the PR should be replaced by simply MFCing
rev. 1.101 to login.c:
: date: 2006/03/28 15:30:42; author: cognet; state: Exp; lines: +5 -2
: Don't call audit_logout() if pwd is NULL, as audit_logout() attempts to
: dereference it.
: This will happen if we ^D at the Login: prompt without having provided a
: valid login before.
: Set pwd to NULL on bad login attempts to prevent audit_logout() from being
: called for a user which didn't actually log on.
:
: Reported by: Jerome Magnin jethro at docisland dot org
RSS Feed