[Bug 438263] New: b44: constantly "powering down PHY" during high network load after resume
https://bugzilla.novell.com/show_bug.cgi?id=438263
Summary: b44: constantly "powering down PHY" during high network load after resume Product: openSUSE 11.0 Version: Final Platform: 64bit OS/Version: openSUSE 11.0 Status: NEW Severity: Normal Priority: P5 - None Component: Network AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: andreas.pfaller@gmail.com QAContact: qa@suse.de Found By: Community User
On my Dell Inspiron 1720 Laptop the b44 network driver constantly reports "powering down PHY" during some high network load after a resume from "Suspend to RAM", e.g.
.. Oct 22 11:49:29 i1720 kernel: b44: eth0: powering down PHY Oct 22 11:49:30 i1720 kernel: b44: eth0: Link is down. Oct 22 11:49:30 i1720 nm-dispatcher.action: Script '/etc/NetworkManager/dispatcher.d/autofs' exited with error status 1. Oct 22 11:49:30 i1720 syslog-ng[2637]: io.c: do_write: write() failed (errno 22), Invalid argument Oct 22 11:49:30 i1720 syslog-ng[2637]: Connection broken to AF_INET(192.168.1.6:514), reopening in 60 seconds Oct 22 11:49:30 i1720 nm-dispatcher.action: Script '/etc/NetworkManager/dispatcher.d/smbfs' exited with error status 6. Oct 22 11:49:32 i1720 kernel: b44: eth0: Link is up at 100 Mbps, full duplex. Oct 22 11:49:32 i1720 kernel: b44: eth0: Flow control is off for TX and off for RX. Oct 22 11:49:33 i1720 nm-dispatcher.action: Script '/etc/NetworkManager/dispatcher.d/autofs' exited with error status 1. Oct 22 11:49:34 i1720 SuSEfirewall2: SuSEfirewall2 not active Oct 22 11:49:34 i1720 nm-dispatcher.action: Script '/etc/NetworkManager/dispatcher.d/netcontrol_global_hooks' exited with error status 1. Oct 22 11:49:34 i1720 nm-dispatcher.action: Script '/etc/NetworkManager/dispatcher.d/nfs' exited with error status 2. Oct 22 11:49:34 i1720 nm-dispatcher.action: Script '/etc/NetworkManager/dispatcher.d/smbfs' exited with error status 6. Oct 22 11:49:38 i1720 kernel: b44: eth0: powering down PHY Oct 22 11:49:39 i1720 kernel: b44: eth0: Link is down. Oct 22 11:49:39 i1720 nm-dispatcher.action: Script '/etc/NetworkManager/dispatcher.d/autofs' exited with error status 1. Oct 22 11:49:40 i1720 nm-dispatcher.action: Script '/etc/NetworkManager/dispatcher.d/smbfs' exited with error status 6. Oct 22 11:49:42 i1720 kernel: b44: eth0: Link is up at 100 Mbps, full duplex. Oct 22 11:49:42 i1720 kernel: b44: eth0: Flow control is off for TX and off for RX. Oct 22 11:49:43 i1720 nm-dispatcher.action: Script '/etc/NetworkManager/dispatcher.d/autofs' exited with error status 1. ..
A rebooted system works without problem. Light network usage after "Suspend to RAM" also does not trigger the problem. This problem is 100% reproducable e.g. with "bonnie -s 2048" on an mounted nfs share (the problem only starts at the start of bonnies "rewriting" test) or when the laptop takes part on a bigger "icecream" compile.
If the load decreases (e.g. ^C of bonnie) the interface recovers after some time and is usable with light network load (e.g. ssh).
https://bugzilla.novell.com/show_bug.cgi?id=438263
User andreas.pfaller@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=438263#c1
--- Comment #1 from Andreas Pfaller andreas.pfaller@gmail.com 2008-10-23 08:52:26 MDT --- Created an attachment (id=247527) --> (https://bugzilla.novell.com/attachment.cgi?id=247527) lspci -v output
https://bugzilla.novell.com/show_bug.cgi?id=438263
Michal Seben mseben@novell.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |mseben@novell.com AssignedTo|bnc-team-screening@forge.provo.novell.com |jbenc@novell.com
https://bugzilla.novell.com/show_bug.cgi?id=438263
User jbenc@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=438263#c2
Jiri Benc jbenc@novell.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO Info Provider| |andreas.pfaller@gmail.com
--- Comment #2 from Jiri Benc jbenc@novell.com 2008-10-30 13:09:08 MDT --- Can you please test the kernel from http://labs.suse.cz/jbenc/kernel/bug438263/? I believe it fixes the problem.
https://bugzilla.novell.com/show_bug.cgi?id=438263
User andreas.pfaller@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=438263#c3
Andreas Pfaller andreas.pfaller@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|andreas.pfaller@gmail.com |
--- Comment #3 from Andreas Pfaller andreas.pfaller@gmail.com 2008-10-30 17:00:28 MDT --- Sure. No change regarding the problem while using the provided 2.6.25.19-0.2-default kernel from comment #2.
Since I had no source for the provided kernel I had to do my bonnie test from runlevel 3 after "s2ram" (I normally use a self compiled nvidia driver and use the kde applet for suspend). The provided kernel also failed to enable the display after resume so bonnie was started from an ssh session. (I never tried a suspend from runlevel 3 before so I don't know if this problem exists with 2.6.25.18-0.2).
Some other observations:
bonnie throughput after fresh reboot (nearly identical with 2.6.25.18-0.2 and your 2.6.25.19-0.2):
Bonnie 1.4: File './Bonnie.4545', size: 2147483648, volumes: 1 Writing with putc()... done: 11379 kB/s 25.4 %CPU Rewriting... done: 5715 kB/s 2.0 %CPU Writing intelligently... done: 11439 kB/s 2.4 %CPU Reading with getc()... done: 11114 kB/s 33.2 %CPU Reading intelligently... done: 11319 kB/s 1.8 %CPU Seeker 1...Seeker 2...Seeker 3...start 'em...done...done...done... ---Sequential Output (nosync)--- ---Sequential Input-- --Rnd Seek- -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --04k (03)- Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU i1720 1*2048 11379 25.4 11439 2.4 5715 2.0 11114 33.2 11319 1.8 1430.1 5.0
bonnie throughput after resume (also nearly identical with both kernel versions). Note the greatly reduced throughput (there are no "powering down" messages during this part of the test):
Writing with putc()... done: 7874 kB/s 15.2 %CPU
As I said in my original comment the "powering down" problems start during the rewrite phase of bonnie so I can't provide throughput numbers for the later bonnie stages.
https://bugzilla.novell.com/show_bug.cgi?id=438263
User andreas.pfaller@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=438263#c4
--- Comment #4 from Andreas Pfaller andreas.pfaller@gmail.com 2008-10-30 17:03:29 MDT --- Created an attachment (id=249001) --> (https://bugzilla.novell.com/attachment.cgi?id=249001) dmesg output covering boot, suspend and resume with provided 2.6.25.19-0.2 kernel
https://bugzilla.novell.com/show_bug.cgi?id=438263
Jiri Benc jbenc@novell.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED Priority|P5 - None |P4 - Low
https://bugzilla.novell.com/show_bug.cgi?id=438263
User jbenc@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=438263#c5
Jiri Benc jbenc@novell.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |andreas.pfaller@gmail.com
--- Comment #5 from Jiri Benc jbenc@novell.com 2009-02-26 12:39:55 MST --- Sorry for not looking into the bug for such long time. Are you able to reproduce the bug with openSUSE 11.1?
https://bugzilla.novell.com/show_bug.cgi?id=438263
User andreas.pfaller@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=438263#c6
Andreas Pfaller andreas.pfaller@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED CC| |andreas.pfaller@gmail.com Info Provider|andreas.pfaller@gmail.com |
--- Comment #6 from Andreas Pfaller andreas.pfaller@gmail.com 2009-02-26 13:48:20 MST --- Sorry no, I stayed with 11.0 on all my systems.
https://bugzilla.novell.com/show_bug.cgi?id=438263
User jbenc@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=438263#c7
Jiri Benc jbenc@novell.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |andreas.pfaller@gmail.com
--- Comment #7 from Jiri Benc jbenc@novell.com 2009-02-27 07:32:02 MST --- Could you try the latest kernel of the day? (ftp://ftp.suse.com/pub/projects/kernel/kotd/SLE11_BRANCH/)
https://bugzilla.novell.com/show_bug.cgi?id=438263
User andreas.pfaller@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=438263#c8
Andreas Pfaller andreas.pfaller@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|andreas.pfaller@gmail.com |
--- Comment #8 from Andreas Pfaller andreas.pfaller@gmail.com 2009-02-27 10:28:12 MST --- No problems anymore after resume while using kernel-default-2.6.27.19-SLE11_BRANCH_20090226184338_a75dbd91.x86_64.rpm kernel-default-base-2.6.27.19-SLE11_BRANCH_20090226184338_a75dbd91.x86_64.rpm
kernel-default-extra-2.6.27.19-SLE11_BRANCH_20090226184338_a75dbd91.x86_64.rpm (on an up-to-date 11.0 system ignoring dependencies during install).
https://bugzilla.novell.com/show_bug.cgi?id=438263
User jbenc@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=438263#c9
Jiri Benc jbenc@novell.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED
--- Comment #9 from Jiri Benc jbenc@novell.com 2009-03-02 03:55:09 MST --- So this is apparently fixed in newer kernels. As a workaround for openSUSE 11.0, you can use the mentioned kernel from SLES11/openSUSE 11.1 (please note that you won't get automatic kernel security updates this way).
-
bugzilla_noreply＠novell.com