Bug 632982

Summary:	PXE Booting via a pcnet device pauses the VM
Product:	Red Hat Enterprise Linux 6	Reporter:	Gordan Bobic <gordan>
Component:	qemu-kvm	Assignee:	Virtualization Maintenance <virt-maint>
Status:	CLOSED WONTFIX	QA Contact:	Virtualization Bugs <virt-bugs>
Severity:	medium	Docs Contact:
Priority:	low
Version:	6.0	CC:	juzhang, mkenneth, tburke, virt-maint
Target Milestone:	rc
Target Release:	6.1
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2011-01-03 11:12:14 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:	580953

Description Gordan Bobic 2010-09-12 10:46:18 UTC

Description of problem:
When PXE booting a VM using a pcnet emulated interface, the VM enters a paused state and cannot be unpaused.

Version-Release number of selected component (if applicable):
qemu-kvm-0.12.1.2-2.90.el6.x86_64

How reproducible:
Every time.

Steps to Reproduce:
1. Create a VM with a pcnet NIC.
2. PXE boot it.
  
Actual results:
VM pauses at:

gPXE initialising devices...
pcnet32.c: Found pcnet32, Vendor=0x1022 Device=0x2000
10

Expected results:
Link should be initiated and DHCP invoked.

Additional info:
Changing the emulated NIC type fixes the problem.

Comment 2 Dor Laor 2010-11-22 10:59:26 UTC

Have you tried other models like virtio or e1000/rtl8139?

Comment 3 Gordan Bobic 2010-11-22 11:07:07 UTC

Does the part in the "Additional info" section of the original post answer your question?

Comment 4 Dor Laor 2010-11-22 11:52:00 UTC

(In reply to comment #3)
> Does the part in the "Additional info" section of the original post answer your
> question?

yes, sorry for missing it. 

Is there a reason for you to pick pcnet specifically?

Comment 5 Gordan Bobic 2010-11-22 12:02:45 UTC

Yes, I was trying different emulations to try to work around the following bugs:
https://bugzilla.redhat.com/show_bug.cgi?id=632712
https://bugzilla.redhat.com/show_bug.cgi?id=632716

Comment 6 Dor Laor 2011-01-03 11:12:14 UTC

We'll fix the above two and close this one as won't fix since pcnet isn't prime time device

Comment 7 Gordan Bobic 2011-01-03 12:55:32 UTC

It seems very shoddy and unprofessional to have things that are broken in an enterprise distribution. The pcnet emulation should either be fixed or removed completely. Having it broken yet existing as an option is the worst possible thing to do.

Comment 8 Dor Laor 2011-01-03 13:40:27 UTC

Bug 620657 - Unsupported device uses need to be documented in release notes

The above bug depicts the supported devices. Pcnet is not one of them.
At the time it was too intrusive to remove pcnet from the code so we used release notes approach for it. Sorry for the noise.

Comment 9 Gordan Bobic 2011-01-03 15:43:15 UTC

IMO, treating "documented as broken" as a "fix" is perhaps something that might be good enough as a bodge on Fedora, but it really doesn't do much good for the perception of quality in what is supposed to be an enterprise grade distribution that has been in stabilization and polish process for over a year.

Comment 10 Dor Laor 2011-01-03 15:57:29 UTC

It's documented as not supported.
I admit that we should have compiled it out but there are other restrictions to consider - we're 100% open source and we like to stay close as possible to upstream. There are times where we manage to change upstream on time so it will be easy to configure features out. This was not the case so this was the mechanism we used this time. 

We never intended to support it since there are other better alternatives.
As you can see the alternatives should be fixed in respect to gPxe so that's why we like to focus on fewer/better devices than what's supported upstream.

On a perfect world, this should have been compiled out. You're right here.