Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.
For bugs related to Red Hat Enterprise Linux 5 product line. The current stable release is 5.10. For Red Hat Enterprise Linux 6 and above, please visit Red Hat JIRA https://issues.redhat.com/secure/CreateIssue!default.jspa?pid=12332745 to report new issues.

Bug 543979

Summary: gPXE fails to PXE boot on e1000 virtual NIC
Product: Red Hat Enterprise Linux 5 Reporter: Dan Yasny <dyasny>
Component: kvmAssignee: Eduardo Habkost <ehabkost>
Status: CLOSED ERRATA QA Contact: Virtualization Bugs <virt-bugs>
Severity: medium Docs Contact:
Priority: low    
Version: 5.5CC: acathrow, gcosta, lihuang, ovirt-maint, riek, virt-maint, ykaul
Target Milestone: rc   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: kvm-83-140.el5 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2010-03-30 07:55:57 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
screen shot none

Description Dan Yasny 2009-12-03 16:17:02 UTC 
Description of problem:
gPXE fails to PXE boot on e1000 virtual NIC
When tried to change the roms from etherboot to gpxe, e1000 would not boot with the following error, while rtl8139 booted fine

 QMon-2082::ERROR::2009-12-03 23:35:34,998::QemuMonitor::179::vds.vmlog.2131e1bd-2a4e-4a7a-b490-4308853a88ea::QemuMonitor
 Traceback (most recent call last):
   File "/usr/share/vdsm/QemuMonitor.py", line 177, in _work
   File "/usr/share/vdsm/QemuMonitor.py", line 105, in _work
   File "<string>", line 1, in sendall
 error: (32, 'Broken pipe')
 QMon-2082::INFO::2009-12-03 23:35:34,999::vm::582::vds.vmlog.2131e1bd-2a4e-4a7a-b490-4308853a88ea::Monitor stopped
 QMon-2082::DEBUG::2009-12-03 23:35:34,999::vm::1219::vds.vmlog.2131e1bd-2a4e-4a7a-b490-4308853a88ea::Changed state to Down: Lost connection with kvm process
 Thread-2078::WARNING::2009-12-03 23:35:35,006::vm::749::vds.vmlog.2131e1bd-2a4e-4a7a-b490-4308853a88ea::Traceback (most recent call last):
   File "/usr/share/vdsm/vm.py", line 747, in _run
   File "/usr/share/vdsm/vm.py", line 634, in _monitorDependentInit
   File "/usr/share/vdsm/vm.py", line 752, in _isKvmEnabled
   File "/usr/share/vdsm/vm.py", line 1004, in _sendMonitorCommand
   File "/usr/share/vdsm/QemuMonitor.py", line 73, in sendCommand
 error: (32, 'Broken pipe')


Version-Release number of selected component (if applicable):
sp215
 kmod-kvm-83-105.el5_4.9
 kvm-tools-83-105.el5_4.9
 etherboot-zroms-kvm-5.4.4-10.el5
 kvm-debuginfo-83-105.el5_4.9
 kvm-qemu-img-83-105.el5_4.9
 kvm-83-105.el5_4.9

gpxe-bootimgs-0.9.7-6.el5rhev.x86_64.rpm 
gpxe-roms-0.9.7-6.el5rhev.x86_64.rpm 
gpxe-roms-qemu-0.9.7-6.el5rhev.x86_64.rp

How reproducible:
Always

Steps to Reproduce:
1.install the gpxe packages
2.change the symlinks in /usr/share/kvm to
lrwxrwxrwx 1 root root     28 Dec  3 23:14 pxe-e1000.bin -> ../pxe-roms/e1000-0x100e.rom
lrwxrwxrwx 1 root root     18 Dec  3 23:15 pxe-ne2k_pci.bin -> ../pxe-roms/ne.rom
lrwxrwxrwx 1 root root     23 Dec  3 23:15 pxe-pcnet.bin -> ../pxe-roms/pcnet32.rom
lrwxrwxrwx 1 root root     23 Dec  3 23:15 pxe-rtl8139.bin -> ../pxe-roms/rtl8139.rom
lrwxrwxrwx 1 root root     26 Dec  3 23:15 pxe-virtio.bin -> ../pxe-roms/virtio-net.rom

3. boot a VM from PXE using e1000
  
Actual results:
as described in log above

Expected results:
VM to boot and load the gpxe

Additional info:
Comment 2 Glauber Costa 2009-12-09 15:34:30 UTC 
This is not a gpxe bug. It is a kvm bug.

Since RHEV does not have a kvm package, I'll change it to RHEL5.
Comment 3 Glauber Costa 2009-12-09 15:37:33 UTC 
Ok, now that the bug is targeted at the correct component, the problem here is that e1000 rom is too big to fit the current ROMs section.

The fix went upstream some months ago, and have a fairly easy backport.
Comment 8 lihuang 2009-12-23 07:44:18 UTC 
Created attachment 379983 [details]
screen shot

I have test with 

[root@t77 ~]# rpm -qa gpxe* kvm*
gpxe-roms-qemu-0.9.7-8.el5
kvm-83-140.el5
kvm-tools-83-140.el5
kvm-debuginfo-83-140.el5
kvm-qemu-img-83-140.el5

can not reproduce the " Too many option ROMS" error, when using e1000 nic

but the e1000 bin doesn't work.
when I use cli : 
  -net nic,model=e1000 -net tap -boot n
error in the VNC windows is 
--
 FATAL: No bootable device
--
when Press F12 on the boot menu.there is no _network_ in the boot menu,(refer to attachment)


also test -option-rom /usr/share/qemu-pxe-roms/e1000-82542.zrom result same 

using rtl8139 and virtio. boot from network is PASS.

[root@t77 kvm]# ls -l /usr/share/kvm/pxe-e1000.bin 
lrwxrwxrwx 1 root root 33 Dec 23 11:58 /usr/share/kvm/pxe-e1000.bin -> ../qemu-pxe-roms/e1000-82542.zrom
[root@t77 kvm]# rpm -qf ../qemu-pxe-roms/e1000-82542.zrom
gpxe-roms-qemu-0.9.7-8.el5
Comment 9 lihuang 2009-12-23 08:00:18 UTC 
Hi   Glauber
  the problem (Comment #8) is in kvm or gpxe ?
Comment 10 lihuang 2009-12-24 09:59:08 UTC 
I have created a new bug to track the new problem in comment#8.
--> Bug 550265

Thanks
Comment 11 Glauber Costa 2009-12-24 10:54:59 UTC 
Probably neither.

It sounds to me like a BIOS problem.

But since AFAIK, we ship the bios together with kvm, the component of the bug should be KVM.
Comment 12 lihuang 2009-12-24 15:29:30 UTC 
so. I am closing this bug since the original issue is fixed.

correct me if I am wrong.
Comment 15 errata-xmlrpc 2010-03-30 07:55:57 UTC 
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

http://rhn.redhat.com/errata/RHSA-2010-0271.html