Skip to contentSkip to conversation

Config.mk: update OVMF changeset

PATCH In review Wei Liu 3 participants 6 messages Dec 8, 2013 – Dec 9, 2013

lore
Share & tools
lore

Wei Liu

author
#0 lore raw

PATCH 1/1 Config.mk: update OVMF changeset

Cc Ian Campbell <[email protected]>
Cc Ian Jackson <[email protected]>
Cc George Dunlap <[email protected]>

---
To IanC,

OVMF tree can be found at:
http://xenbits.xen.org/git-http/people/liuw/ovmf.git for-xenbits
Please push that branch to Xenbits.

I presume that
tools: enable OVMF build for Linux by default
is in your queue so I don't bother resending that. This patch should probably go before that one.

To George,

Good: users can get their hands dirty with OVMF.
Bad : if somebody spots a problem in interface we have to live with it.

But, if nobody gets a chance to use it then nobody can spot a problem. We've merged the interface in tree. I think the advantage of having OVMF for 4.4 outweights the disadvantage. I suggest we let OVMF go in and mark it as experimental.

---
 Config.mk |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Config.mk +1 -1
index 7eeed31..9b5d1bd 100644
@@ -233,7 +233,7 @@ OVMF_UPSTREAM_URL ?= git://xenbits.xen.org/ovmf.git
 QEMU_UPSTREAM_URL ?= git://xenbits.xen.org/qemu-upstream-unstable.git
 SEABIOS_UPSTREAM_URL ?= git://xenbits.xen.org/seabios.git
 endif
-OVMF_UPSTREAM_REVISION ?= a93b0e3f6895a074b99c8817181dfa6dbc7a4807
+OVMF_UPSTREAM_REVISION ?= 447d264115c476142f884af0be287622cd244423
 QEMU_UPSTREAM_REVISION ?= 1c514a7734b7f98625a0d18d5e8ee7581f26e50c
 SEABIOS_UPSTREAM_TAG ?= rel-1.7.3.1
 # Fri Aug 2 14:12:09 2013 -0400
Signature · 1 line
1.7.10.4
author
#2 lore raw
Quoted 10 lines · George Dunlap
Quoted 4 lines · Wei Liu
Cc Ian Campbell <[email protected]>
Cc Ian Jackson <[email protected]>
Cc George Dunlap <[email protected]>

This lacks a description of what the effect will be of updating to this changeset. I'm guessing that the OVMF side of your recent fixes have gotten checked in?

Yes, necessary patches to integrate OVMF with Xen is upstreamed now.

The original changeset in Config.mk:
1. doesn't work with guest > 4G ram
2. messes up PCI resource mapping

The version of OVMF fixes those bugs and works well for me.

Wei.

Quoted 1 line

-George

#3 lore raw
Quoted 10 lines · Wei Liu
Quoted 9 lines · George Dunlap
Quoted 4 lines · Wei Liu
Cc Ian Campbell <[email protected]>
Cc Ian Jackson <[email protected]>
Cc George Dunlap <[email protected]>

This lacks a description of what the effect will be of updating to this changeset. I'm guessing that the OVMF side of your recent fixes have gotten checked in?

Quoted 7 lines

Yes, necessary patches to integrate OVMF with Xen is upstreamed now.

The original changeset in Config.mk:
1. doesn't work with guest > 4G ram
2. messes up PCI resource mapping

The version of OVMF fixes those bugs and works well for me.

re the interface, we generally plan on building OVMF ourselves, right? I don't think the interface between Xen and OVMF necessarily needs to be stable at this point.

In any case, the interface changes are already in Xen; this doesn't make the interface available, it just makes it easier for people to build OVMF.

#4 lore raw
Quoted 17 lines · George Dunlap
Quoted 16 lines · Wei Liu
Quoted 9 lines · George Dunlap
Quoted 4 lines · Wei Liu
Cc Ian Campbell <[email protected]>
Cc Ian Jackson <[email protected]>
Cc George Dunlap <[email protected]>

This lacks a description of what the effect will be of updating to this changeset. I'm guessing that the OVMF side of your recent fixes have gotten checked in?

Yes, necessary patches to integrate OVMF with Xen is upstreamed now.

The original changeset in Config.mk:
1. doesn't work with guest > 4G ram
2. messes up PCI resource mapping

Quoted 5 lines
Quoted 1 line

The version of OVMF fixes those bugs and works well for me.

re the interface, we generally plan on building OVMF ourselves, right? I don't think the interface between Xen and OVMF necessarily needs to be stable at this point.

It is an ABI between Xen and an external 3rd party codebase. Although we might build it ourselves by default we cannot enforce this, and our users might use e.g. their distro binary. This is similar to the situation with upstream qemu.

This ABI absolutely has to be stable from now on (or make use of the inbuilt capability to extend itself).

I don't think there is going to be any reason for this interface to change so this is fine.

Quoted 4 lines

In any case, the interface changes are already in Xen; this doesn't make the interface available, it just makes it easier for people to build OVMF.

#5 lore raw
In reply to George Dunlap
Quoted 19 lines · George Dunlap
Quoted 10 lines · Wei Liu
Quoted 9 lines · George Dunlap
Quoted 4 lines · Wei Liu
Cc Ian Campbell <[email protected]>
Cc Ian Jackson <[email protected]>
Cc George Dunlap <[email protected]>

This lacks a description of what the effect will be of updating to this changeset. I'm guessing that the OVMF side of your recent fixes have gotten checked in?

Quoted 7 lines

Yes, necessary patches to integrate OVMF with Xen is upstreamed now.

The original changeset in Config.mk:
1. doesn't work with guest > 4G ram
2. messes up PCI resource mapping

The version of OVMF fixes those bugs and works well for me.

Quoted 8 lines

re the interface, we generally plan on building OVMF ourselves, right? I don't think the interface between Xen and OVMF necessarily needs to be stable at this point.

In any case, the interface changes are already in Xen; this doesn't make the interface available, it just makes it easier for people to build OVMF.

acked + applied. Also pushed the requested update to ovmf.git.

For now I have held off on applying the patch which enables OVMF by default ("tools: enable OVMF build for Linux by default"). I think that deserves its own Release ack/nack.

Ian.