Bug 794579

Summary: RHEL6.2 xen HVM guest fails to boot up when there are 16 VBDs and ioemu vif attached
Product: Red Hat Enterprise Linux 6 Reporter: Shengnan Wang <shwang>
Component: kernelAssignee: Igor Mammedov <imammedo>
Status: CLOSED DUPLICATE QA Contact: Virtualization Bugs <virt-bugs>
Severity: medium Docs Contact:
Priority: medium    
Version: 6.2CC: drjones, imammedo, leiwang, moli, qguan, qwan, yuzhou
Target Milestone: rc   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard: xen
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2012-05-14 15:15:56 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:
Bug Depends On:    
Bug Blocks: 653816    
Attachments:
Description Flags
config file info
none
boot log
none
boot log none

Description Shengnan Wang 2012-02-17 06:21:39 UTC 
RHEL6.2 HVM guest fails to boot up when there are more than 15 physical lvs (phy:/dev/VolGroup/$lvname) or blktap (tap:aio:) xvdx disks and one ioemu vif (rtl8139 or e1000) attached. It hang while booting at the step of "set up logical volume management". It may print call trace after a while. Guest can boot up successfully when there is no ioemu vif attached or only netfront vif attached.

Call trace info:
Setting up Logical Volume Management: BUG: soft lockup - CPU#0 stuck for 63s! [swapper:0]
Process swapper (pid: 0, ti=c0a06000 task=c0a2e560 task.ti=c0a06000)
Stack:
Call Trace:
Code: c0 7c e3 89 d8 5b 5e 5f c3 90 8d 74 26 00 31 c0 5b 5e 5f c3 66 90 55 57 56 53 89 d3 83 ec 18 89 44 24 0c f6 42 04 20 75 06 fb 90 <8d> 74 26 00 31 ed 31 ff eb 0f 90 83 fe 02 74 63 8b 5b 10 09 f7 
BUG: soft lockup - CPU#0 stuck for 61s! [swapper:0]


Version-Release number of selected component (if applicable):
host: 2.6.18-308.el5xen
guest: 2.6.32-220.el5xen
xen: xen-3.0.3-135.el5

How reproducible:
hang 100%
call trace 40%

Steps to Reproduce:
1.Create a 6.2 hvm guest with more than 15 phy or blktap xvdx disks and the ioemu nic. Details in the cfgfile.txt in the file attached.
2.Follow the booting process via vnc or 'xm create -c guest.cfg'.

Actual results:
At step 2, the guest hangs when set up logical volume management.

Expected results:
The guest boot up successfully or the appropriate error message is dispalyed in the host.

Additional info:
1. Test it on 2.6.18-308.el5(5.8-20120202.0) with xen-135. The problem occurs in both i386 and x86_64 6.2 hvm guest (2.6.32-220.el6).
2. The same problem occurs on 274(kernel)+132(xen) host.
3. It works well for the 5.8 guest.
Comment 1 Shengnan Wang 2012-02-17 06:22:55 UTC 
Created attachment 563836 [details]
config file info
Comment 2 Shengnan Wang 2012-02-17 06:23:38 UTC 
Created attachment 563838 [details]
boot log
Comment 4 Shengnan Wang 2012-02-17 06:49:02 UTC 
Created attachment 563843 [details]
boot log
Comment 5 Andrew Jones 2012-04-30 11:39:41 UTC 
Too late for 6.3, assuming it's not a regression. Setting flag for 6.4.
Comment 6 Igor Mammedov 2012-05-14 15:15:56 UTC 
If we boot with ignore_loglevel, then it is pretty strait-forward:
--- 
blkfront: xvdd: barriers disabled
  alloc irq_desc for 31 on node 0
  alloc kstat_irqs on node 0
 xvdd:
  alloc irq_desc for 32 on node 0
  alloc kstat_irqs on node 0
  alloc irq_desc for 33 on node 0
  alloc kstat_irqs on node 0
...
8139cp: 10/100 PCI Ethernet driver v1.3 (Mar 22, 2004)
8139cp 0000:00:04.0: PCI INT A -> GSI 32 (level, low) -> IRQ 32
eth0: RTL-8139C+ at 0xffffc90000d16000, 06:16:36:b6:e3:c8, IRQ 32
---

IRQ 32 collision.

Applying patches from bug 756307 to 6.2GA kernel fixes issue, tested locally.
To test one should use "xen_emul_unplug=ide-disks" kernel option since patches disable pv drivers if "xen_emul_unplug=never".

Fixed in RHEL6.3 Beta.

*** This bug has been marked as a duplicate of bug 756307 ***