[Bug 976250] New: Xen HVM guests fail to boot with OVMF
http://bugzilla.suse.com/show_bug.cgi?id=976250 Bug ID: 976250 Summary: Xen HVM guests fail to boot with OVMF Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: Other OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Virtualization:Tools Assignee: glin@suse.com Reporter: jfehlig@suse.com QA Contact: qa-bugs@suse.de CC: carnold@suse.com, mlatimer@suse.com, ohering@suse.com Found By: --- Blocker: --- The OBS Virtualization/xen package was recently changed to use the distro provided openSUSE:Factory:NonFree/ovmf package https://build.opensuse.org/package/rdiff/Virtualization/xen?linkrev=base&rev=416 Unfortunately, the latest ovmf package does not work with Xen. The upstream Xen community is aware of the problem since OSSTEST has been reporting xen.git+ovmf.git failures on a daily basis. Here's a link to today's failure report http://lists.xen.org/archives/html/xen-devel/2016-04/msg02428.html Rumor has it that Citrix is working on a fix for ovmf. Olaf is at the Xen hackathon and may be able to confirm the rumor. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
James Fehlig
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c1
James Fehlig
http://bugzilla.suse.com/show_bug.cgi?id=976250
Anja Stock
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c2
--- Comment #2 from sd fs
Rumor has it that Citrix is working on a fix for ovmf.
Checking in -- Has this been fixed? I see some ovmf recent additions to Xen pkgs, but dunno whether the (PV)HVM guest-booting is working yet. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c3
--- Comment #3 from Gary Ching-Pang Lin
Rumor has it that Citrix is working on a fix for ovmf.
Checking in -- Has this been fixed? I see some ovmf recent additions to Xen pkgs, but dunno whether the (PV)HVM guest-booting is working yet.
The edk2 upstream just fixed the PCI host bridge driver(*) for Xen and it works for me. I already pushed the patches into Virtualization and the patches will land factory soon. (*) http://thread.gmane.org/gmane.comp.bios.edk2.devel/11967 -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c4
--- Comment #4 from James Fehlig
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c5
Gary Ching-Pang Lin
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c6
sd fs
Then let's close this bug :)
I just tested this by switching a working setup using upstream Xen's bundled ovmf to the packages from the VVirtualization repo for Leap. Dom0 boots like always. But when I try to boot the same guest from console, it sticks at Trust openSUSE Certificate Do you agree to use the built-in openSUSE certificate to verify boot loaders and kernels? No Yes You can't boot the guest whether you choose Yes or No. I don't know what this is about. It was never needed before. But it sure doesn't boot. I guess I'm supposed to reopen this? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c7
Gary Ching-Pang Lin
I just tested this by switching a working setup using upstream Xen's bundled ovmf to the packages from the VVirtualization repo for Leap.
Dom0 boots like always.
But when I try to boot the same guest from console, it sticks at
Trust openSUSE Certificate Do you agree to use the built-in openSUSE certificate to verify boot loaders and kernels?
No Yes
You can't boot the guest whether you choose Yes or No.
I don't know what this is about. It was never needed before. But it sure doesn't boot.
I guess I'm supposed to reopen this?
Could you try to add this line to your xl config and attach debug.log? device_model_args = ["-debugcon", "file:debug.log", "-global", "isa-debugcon.iobase=0x402"] -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c8
sd fs
device_model_args = ["-debugcon", "file:debug.log", "-global", "isa-debugcon.iobase=0x402"]
debug.log after adding that line ... -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c9
--- Comment #9 from sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c10
Gary Ching-Pang Lin
Created attachment 676904 [details] debug.log with device_model_args = ...
device_model_args = ["-debugcon", "file:debug.log", "-global", "isa-debugcon.iobase=0x402"]
debug.log after adding that line ...
I didn't find anything wrong in the log :-( What if you add this line? bios_override='/usr/share/qemu/ovmf-x86_64.bin' -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c11
sd fs
I didn't find anything wrong in the log :-(
What if you add this line?
bios_override='/usr/share/qemu/ovmf-x86_64.bin'
I installed these latest packages kernel-default-4.5.4-2.1.gb14d2a1.x86_64 ovmf-2015+git1462940744.321151f-20.1.x86_64 qemu-ovmf-x86_64-2015+git1462940744.321151f-20.1.noarch xen-4.7.0_03-441.1.x86_64 xen-libs-4.7.0_03-441.1.x86_64 xen-tools-4.7.0_03-441.1.x86_64 and added to guest.cfg builder = 'hvm' xen_platform_pci = 1 device_model_version="qemu-xen" device_model_args = ["-debugcon", "file:debug.log", "-global","isa-debugcon.iobase=0x402"] bios = 'ovmf' + bios_override = '/usr/share/qemu/ovmf-x86_64.bin' ... Now on reboot I **can** boot the Guest and get to it like usual via SSH. I didn't check the Guest thoroughly yet, but its services seem to be running, at least. But when I check at the Dom0 console, I get "(null)" for Names, xl list Name ID Mem VCPUs State Time(s) (null) 0 4096 1 r----- 66.6 (null) 1 2049 1 -b---- 18.2 -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c12
--- Comment #12 from sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c13
Gary Ching-Pang Lin
I didn't find anything wrong in the log :-(
What if you add this line?
bios_override='/usr/share/qemu/ovmf-x86_64.bin'
I installed these latest packages
kernel-default-4.5.4-2.1.gb14d2a1.x86_64 ovmf-2015+git1462940744.321151f-20.1.x86_64 qemu-ovmf-x86_64-2015+git1462940744.321151f-20.1.noarch xen-4.7.0_03-441.1.x86_64 xen-libs-4.7.0_03-441.1.x86_64 xen-tools-4.7.0_03-441.1.x86_64
My packages: kernel-default-4.6.0-1.1.gaf7ce24.x86_64 ovmf-2015+git1462940744.321151f-20.1.x86_64 qemu-ovmf-x86_64-2015+git1462940744.321151f-20.1.noarch qemu-x86-2.5.93-327.8.x86_64 xen-4.7.0_03-442.1.x86_64 xen-libs-4.7.0_03-442.1.x86_64 xen-tools-4.7.0_03-442.1.x86_64 I am using Leap 42.1 + Kernel:Stable + Virtualization
and added to guest.cfg
builder = 'hvm' xen_platform_pci = 1 device_model_version="qemu-xen" device_model_args = ["-debugcon", "file:debug.log", "-global","isa-debugcon.iobase=0x402"] bios = 'ovmf' + bios_override = '/usr/share/qemu/ovmf-x86_64.bin' ...
Here is my config: builder = 'hvm' name = 's0-efi' vcpus = 2 memory = 1024 disk = ['/home/linux/VM/s0-efi.qcow2,qcow2,xvda,w'] on_poweroff="destroy" on_reboot="destroy" on_crash="destroy" sdl=1 serial='pty' bios='ovmf' device_model_args = ["-debugcon", "file:debug.log", "-global", "isa-debugcon.iobase=0x402"]
Now on reboot I **can** boot the Guest and get to it like usual via SSH. I didn't check the Guest thoroughly yet, but its services seem to be running, at least.
But when I check at the Dom0 console, I get "(null)" for Names,
xl list Name ID Mem VCPUs State Time(s) (null) 0 4096 1 r----- 66.6 (null) 1 2049 1 -b---- 18.2
Mine: linux-hjgb:/home/linux # xl list Name ID Mem VCPUs State Time(s) Domain-0 0 2671 4 r----- 1107.0 s0-efi 3 1024 1 ------ 7.4 I don't know why (null) showed in your system but ovmf is unlikely to change the name of Domain0. This might be other problem. The problems I see in your case are: 1. The certificate prompt This is a prompt from shim in openSUSE when SecureBoot is enabled. Xen by default uses "/usr/share/qemu/ovmf-x86_64-ms.bin" which contains the UEFI KEK and CA keys. shim is supposed to ask the user once and saves the answer in a non-volatile variable. Unfortunately, Xen currently doesn't support pflash devices, so the variable couldn't be saved reliably. Maybe we should switch to ovmf-x86_64.bin for openSUSE. You mentioned that the guest didn't boot after choosing yes. Did you only check status with the serial console? or also connect to the guest with VNC? Shim should load grub2 immediately and I'm not aware of any bug in this part. Could you try to boot the guest again with SDL or VNC? 2. d1v0 Over-allocation for domain 1: 524545 > 524544 Did you find the message from the serial console of the host? My test machine doesn't have a serial port. Might need to find a way to get the message. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c14
--- Comment #14 from James Fehlig
But when I check at the Dom0 console, I get "(null)" for Names,
xl list Name ID Mem VCPUs State Time(s) (null) 0 4096 1 r----- 66.6 (null) 1 2049 1 -b---- 18.2
Is the xencommons service enabled and started? I've seen the '(null)' names in the past when I didn't have the xencommons service started. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c15
--- Comment #15 from Gary Ching-Pang Lin
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c16
sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c17
sd fs
I filed a bug for the non-volatile variables (bsc#982193). Once the bug is fixed, we'll be able to save the non-volatile variables. Although it's not as reliable as pflash, it's better than nothing.
When I try to access that bug I get "You are not authorized to access bug #982193." Can you turn on access? I'd like to maybe see if that's now causing the return of the guest boot issue. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c18
Gary Ching-Pang Lin
(In reply to Gary Ching-Pang Lin from comment #15)
I filed a bug for the non-volatile variables (bsc#982193). Once the bug is fixed, we'll be able to save the non-volatile variables. Although it's not as reliable as pflash, it's better than nothing.
When I try to access that bug I get
"You are not authorized to access bug #982193."
Can you turn on access? I'd like to maybe see if that's now causing the return of the guest boot issue.
Sorry, I filed the bug for SLE-12-SP2 and currently this type of bug is only open for SUSE Enterprise Partners, and I couldn't change it. Anyway, there are two patches related to the bug and I already pushed them into upstream and openSUSE. https://github.com/tianocore/edk2/commit/db827286e2839102c5b0a45f88a99b8ef94... https://github.com/tianocore/edk2/commit/da2369d21d2e57a0de8fa7ae954812122c8... -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c19
--- Comment #19 from sd fs
Anyway, there are two patches related to the bug and I already pushed them into upstream and openSUSE.
Ok thanks. IIUC then I should look for those updates in 'ovmf' package on Opensuse? Are they likely the fix for the 'new' issue in comment #16? Or is that still something else? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c20
sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c21
--- Comment #21 from James Fehlig
cat /etc/xen/auto/pvhvm.cfg name = 's0-efi' builder = 'hvm' xen_platform_pci = 1 device_model_version="qemu-xen" device_model_args = ["-debugcon", "file:debug.log", "-global","isa-debugcon.iobase=0x402"] boot = 'cd' acpi = 1 apic = 1 ... hap = 1 shadow_memory = 16 ... bios = 'ovmf' bios_override = '/usr/share/qemu/ovmf-x86_64.bin'
Just out of curiosity, does it work if you remove bios_override and use the default firmware path (/usr/share/qemu/ovmf-x86_64-ms-code.bin) configured in Xen? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c22
sd fs
Just out of curiosity, does it work if you remove bios_override and use the default firmware path (/usr/share/qemu/ovmf-x86_64-ms-code.bin) configured in Xen?
Doesn't look like it :-/ changed in cfg - bios_override = '/usr/share/qemu/ovmf-x86_64.bin' + #bios_override = '/usr/share/qemu/ovmf-x86_64.bin' verified ls -al /usr/share/qemu/ovmf-x86_64-ms-code.bin -rw-r--r-- 1 root root 1.9M Jun 12 19:22 /usr/share/qemu/ovmf-x86_64-ms-code.bin (by 'use the default firmware path', I assumed that I do NOT need to configure it ...) xl create test.cfg ... ~ 10 secs ... xl list Name ID Mem VCPUs State Time(s) Domain-0 0 4096 1 r----- 86.2 test 4 2048 1 r----- 7.0 ... ~ 30 secs ... xl list Name ID Mem VCPUs State Time(s) Domain-0 0 4096 1 r----- 86.8 logs ==> http://pastebin.com/raw/d5EgGd9u Note, I see an "ConvertPages: Incompatible memory types" message in there, which I'm just starting to find in some old posts. Don't know if that's related. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c23
--- Comment #23 from James Fehlig
(by 'use the default firmware path', I assumed that I do NOT need to configure it ...)
Right. The default path is set by the '--with-system-ovmf=' configure option when building xen.
Note, I see an "ConvertPages: Incompatible memory types" message in there, which I'm just starting to find in some old posts. Don't know if that's related.
Heh, maybe switching firmwares after the fact is not such a good idea :-). But Gary is our OVMF expert, I only work on the tools that use it. However, I'm sure debug output from it would be useful. You can add something like device_model_args=[ "-debugcon", "file:/tmp/debug.log", "-global", "isa-debugcon.iobase=0x402"] to the domU config and attach debug.log -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c24
--- Comment #24 from sd fs
device_model_args=[ "-debugcon", "file:/tmp/debug.log", "-global", "isa-debugcon.iobase=0x402"]
to the domU config and attach debug.log
Already did that, couple of times ^^^ Are you looking for something else, or just haven't yet looked? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c25
--- Comment #25 from James Fehlig
Already did that, couple of times
Yes you did. Sorry for not paying closer attention to your comments. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c26
Gary Ching-Pang Lin
xen logs @ DomU
cd /var/log/xen cat *log
===> paste @ : http://pastebin.com/raw/Yifg2a6Q
cat /debug.log
===> paste @: http://pastebin.com/raw/Rzv62KuT
Has req'd config changed?
The debug.log showed the firmware booted "opensuse-secureboot" successfully. In theory, the system was already booted. Could you add "sdl=1" to show the monitor? Maybe it could gives us some hint... -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c27
sd fs
Could you add "sdl=1" to show the monitor? Maybe it could gives us some hint...
changed - sdl=0 + sdl=1 now, on guest launch, (1) the sdl console wimdow opens (2) i see the tianocore logo (3) i see the grub2 menu (4) the grub2 boot timer counts down, then the guest simply quits and the sdl window closes -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c28
Gary Ching-Pang Lin
Could you add "sdl=1" to show the monitor? Maybe it could gives us some hint...
changed
- sdl=0 + sdl=1
now, on guest launch,
(1) the sdl console wimdow opens (2) i see the tianocore logo (3) i see the grub2 menu (4) the grub2 boot timer counts down, then the guest simply quits and the sdl window closes
That's weird... So grub2 didn't show anything like "loading kernel" or "loading initrd"? What's the status of you guest? Is it a freshly installed leap or updated? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c29
--- Comment #29 from sd fs
So grub2 didn't show anything like "loading kernel" or "loading initrd"?
No, not that I could see.
What's the status of you guest? Is it a freshly installed leap or updated?
It's an upgraded guest. One of a bunch that were working until a week or so ago. Now none of them do, all getting the same messages and fails. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c30
--- Comment #30 from sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c31
--- Comment #31 from sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c32
--- Comment #32 from Gary Ching-Pang Lin
I mounted & chroot'd into the Guest's partition
There I did a `zypper up`, `mkinitrd` and `grub2-mkconfig -o /boot/grub/grub2.cfg`.
Then exited and rebooted the guest.
It failed to boot as before.
Here's the latest "debug.log" from that attempt,
Could you attach /boot/grub2/grub.cfg of both good and bad guests? Hope we can find some clues. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c33
--- Comment #33 from sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c34
--- Comment #34 from sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c35
sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c36
Gary Ching-Pang Lin
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c37
Gary Ching-Pang Lin
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c38
sd fs
The other two patches are for bsc#982193 to enable the file-based NvVar. I build an OVMF without those two patches in my branch(*).
Upgraded Dom0 to rpm -qa | grep -i ovmf ovmf-2015+git1462940744.321151f-31.1.x86_64 ovmf-tools-2015+git1462940744.321151f-31.1.x86_64 qemu-ovmf-x86_64-2015+git1462940744.321151f-31.1.noarch re-exec'd mkinitrd grub2-mkconfig -o /boot/grub2/grub.cfg rebooted Dom0 Guest launch still fails. Here's the debug.log http://pastebin.com/raw/7kYSMm7a Notice still the references to file-based NvVar, e.g. "Saved NV Variables to NvVars file" ? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c39
sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c40
--- Comment #40 from sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c41
--- Comment #41 from Gary Ching-Pang Lin
The other two patches are for bsc#982193 to enable the file-based NvVar. I build an OVMF without those two patches in my branch(*).
Upgraded Dom0 to
rpm -qa | grep -i ovmf ovmf-2015+git1462940744.321151f-31.1.x86_64 ovmf-tools-2015+git1462940744.321151f-31.1.x86_64 qemu-ovmf-x86_64-2015+git1462940744.321151f-31.1.noarch
re-exec'd
mkinitrd grub2-mkconfig -o /boot/grub2/grub.cfg
rebooted Dom0
Guest launch still fails.
That's weird... The only difference between the OVMF of comment#11 and the one in my branch is a HTTP/TCP patch which your system will never execute.
Here's the debug.log
http://pastebin.com/raw/7kYSMm7a
Notice still the references to file-based NvVar, e.g.
"Saved NV Variables to NvVars file"
? It's because you were using the partition in a real disk, so OVMF can access it without connecting the XenBus. This error still showed:
Variable Check ReadOnly variable fail Write Protected - 04B37FE8-F6AE-480B-BDD5-37D98C5E89AA:VarErrorFlag so your system needs ovmf-bsc982193-dont-restore-readonly-var.patch to make file-based NvVars really work. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c42
--- Comment #42 from sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c43
--- Comment #43 from Gary Ching-Pang Lin
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c44
--- Comment #44 from sd fs
http://bugzilla.suse.com/show_bug.cgi?id=976250
Joey Lee
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c45
--- Comment #45 from Gary Ching-Pang Lin
Okay, then. I guess I'll switch back to the upstream Xen builds with internal ovmf since I know those worked.
Have to figure out how to build those myself it seems.
It seems xen upstream is using this git for OVMF. http://xenbits.xen.org/gitweb/?p=ovmf.git;a=summary I cloned the git repo and it's actually a snapshot of edk2 upstream. It's quite easy to build ovmf with the script in OvmfPkg. If you don't want to enable secure boot, IPv6 support, or http boot, just run the script OvmfPkg/build.sh and the resultant firmware file will be Build/OvmfX64/DEBUG_GCC49/FV/OVMF.fd Hope you could get a usable ovmf so that I can try to figure out what we are still missing. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=976250
http://bugzilla.suse.com/show_bug.cgi?id=976250#c46
--- Comment #46 from sd fs
participants (1)
-
bugzilla_noreply@novell.com