Bug 190620

Summary: kernel-2.6.16-1.2107_FC5, NFS server stop working
Product: [Fedora] Fedora Reporter: Kazutoshi Morioka <morioka>
Component: kernelAssignee: Dave Jones <davej>
Status: CLOSED ERRATA QA Contact: Brian Brock <bbrock>
Severity: medium Docs Contact:
Priority: medium    
Version: 5CC: cks-rhbugzilla, gdr, pfrields, wtogami
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2006-05-07 00:10:20 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
dmesg with "pci=nomsi"
none
dmesg with noapic
none
dmesg of kernel-2.6.16-1.2108_FC5.i686
none
this is dmesg of 2096, which shows no problem none

Description Kazutoshi Morioka 2006-05-04 00:06:57 UTC 
Description of problem:
After upgrading to kernel-2.6.16-1.2107_FC5,
NFS server do not respond to mount request.

Version-Release number of selected component (if applicable):
kernel-2.6.16-1.2107_FC5
nfs-utils-1.0.8.rc2-5.FC5
portmap-4.0-65.2.2
util-linux-2.13-0.20.3

How reproducible:
depended on hardware. I have tested this on four PCs.
a. Athlon XP 2200, 8139too : occur
b. Sempron 2600+, e1000 : occur
c. Celeron 950MHz, r8169 : not occur
d. K6-2 450MHz, tulip : not occur

I switched HDD between a and c, then tested again:
a. Athlon XP 2200, 8139too: occur
c. Celeron 950MHz, r8169: not occur
It seems CPU related?

Steps to Reproduce:
1. upgrade to kernel-2.6.16-1.2107_FC5
2. reboot
3. service nfs start
4. exportfs '*:/home'
5. mount foobar:/home /mnt # from other host
  
Actual results:
NFS server not responding, and cause timeout.
mount: mount to NFS server 'foobar' failed: timed out (retrying).

Expected results:
foobar:/home is mounted on /mnt

Additional info:
Comment 1 Dave Jones 2006-05-04 03:37:54 UTC 
does booting with pci=nomsi make this go away?
Comment 2 Kazutoshi Morioka 2006-05-04 03:47:37 UTC 
I tried "pci=nomsi" and got a message "PCI: Unknown option `nomsi'".
All problems are stay there.
Comment 3 Kazutoshi Morioka 2006-05-04 03:50:51 UTC 
Created attachment 128574 [details]
dmesg with "pci=nomsi"
Comment 4 Herbert Xu 2006-05-04 04:14:02 UTC 
Sigh, this option was added post 2.6.16.  Please try noapic instead.  Thanks.
Comment 5 Dave Jones 2006-05-04 04:16:53 UTC 
alternatively, 2108 from http://people.redhat.com/davej/kernels/Fedora/FC5 has
this disabled by default.
Comment 6 Kazutoshi Morioka 2006-05-04 04:22:09 UTC 
Created attachment 128578 [details]
dmesg with noapic

I already tried "noapic" and "pci=irqroute", and got same problems.

# lspci
00:00.0 Host bridge: VIA Technologies, Inc. VT8377 [KT400/KT600 AGP] Host
Bridge00:01.0 PCI bridge: VIA Technologies, Inc. VT8235 PCI Bridge
00:10.0 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1
Controller (rev 80)
00:10.1 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1
Controller (rev 80)
00:10.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1
Controller (rev 80)
00:10.3 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 82)
00:11.0 ISA bridge: VIA Technologies, Inc. VT8235 ISA Bridge
00:11.1 IDE interface: VIA Technologies, Inc.
VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06)
00:11.5 Multimedia audio controller: VIA Technologies, Inc. VT8233/A/8235/8237
AC97 Audio Controller (rev 50)
00:13.0 Ethernet controller: Realtek Semiconductor Co., Ltd.
RTL-8139/8139C/8139C+ (rev 10)
01:00.0 VGA compatible controller: ATI Technologies Inc Radeon RV100 QY [Radeon
7000/VE]
Comment 7 Kazutoshi Morioka 2006-05-04 04:29:53 UTC 
I tried kernel-2.6.16-1.2108_FC5.i686.rpm, and, got same problems again.
Comment 8 Kazutoshi Morioka 2006-05-04 04:31:23 UTC 
Created attachment 128579 [details]
dmesg of kernel-2.6.16-1.2108_FC5.i686
Comment 9 Kazutoshi Morioka 2006-05-04 04:35:33 UTC 
Created attachment 128580 [details]
this is dmesg of 2096, which shows no problem
Comment 10 josip 2006-05-04 12:37:16 UTC 
After upgrading to kernel-2.6.16-1.2107_FC5, my network is dead.  Old
kernel-2.6.16-1.2096_FC5 works just fine, though.

Configuration:  Athlon 64 X2 on an ASUS A8N-SLI Premium motherboard

alias eth1 forcedeth
alias eth0 skge

where eth0 is not connected to any network, and eth1 is the main interface on a
GigE network. 

Results: Boot hangs at NFS mount.  Booting  kernel-2.6.16-1.2107_FC5 in single
user mode then starting network shows normal routing etc, but pings to machines
on the same subnet fail.  Old kernel-2.6.16-1.2096_FC5 works just fine, though.
Comment 11 Chris Siebenmann 2006-05-04 20:28:33 UTC 
For what it is worth, I have more or less the same problem as comment #10.
I have a Tyan Tomkat K8E (same nForce Ultra 4 chipset as the ASUS A8N-SLI
I believe) with an Athlon X2 4400+ and the x86_64 version of FC5. Under
.2107 the forcedeth driver doesn't work; under .2096 it does.

More data:
- a stock kernel.org 2.6.16.13 works fine.
- 'noapic' and 'pci=irqroute' make no difference.
- on .2107 the driver emits no complaints about life or anything, as far as
  I can see; it just doesn't emit or receive packets.
Comment 12 Devin Reade 2006-05-04 21:00:07 UTC 
It looks like the original submission was wrt the linux box running an NFS
server.  For the record, I'm seeing what appears to be the same problem where
the linux box is the NFS _client_ (the server is Solaris 9).  Rebooting into
the previous 2.6.16-1.2096_FC5smp allows me to work.
Comment 13 Kazutoshi Morioka 2006-05-05 02:47:09 UTC 
I tested 2.6.16-1.2109_FC5, and all problems are gone.
NFS server works again, graphical boot works again.
Thank you Dave!
Comment 14 Chris Siebenmann 2006-05-06 00:56:58 UTC 
I can confirm that 2.6.16-1.2111_FC5 (which is what I found as current at the
URL in comment #5) fixes the problem for me.
Comment 15 josip 2006-05-07 16:49:50 UTC 
New kernel-2.6.16-1.2111_FC5 fixes the problem for me, too.