virtio module backport

DISCUSSION Anthony Liguori 3 participants 4 messages Jan 2, 2008 – Jan 3, 2008 2 messages not in the archive KVM

Share & tools

Anthony Liguori

author

2008-01-02 21:04 UTC #0 lore raw

Hi,

I've posted a repo that contains my initial attempt at backporting the virtio modules. I've tested against a 2.6.22, 2.6.20, and 2.6.18 kernel (although I haven't tried the block device against 2.6.18). It requires Rusty's patch queue plus the three patches I just sent to virtualization@l-f.

The makefile supports a 'make sync LINUX=/path/to/linux' just like the kvm-userspace/kernel directory does.

This is still rough as I haven't went through and made sure all of the #if's are right.

http://hg.codemonkey.ws/virtio-ext-modules

Regards,

Anthony Liguori

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

A message that is not in the archive (sent off-list or not yet received): 477BFC6E.2040508-r/[email protected]

Anthony Liguori

2008-01-02 21:21 UTC #2 lore raw

Just to be clear, to use this repository, you need a kernel tree with Rusty's virtio patch queue applied.

For instance, if you had the latest KVM git tree, you would then need to
get Rusty's queue at http://ozlabs.org/~rusty/kernel/hg. I apply up
to virtio:virtio_net-performance-stats.patch.

Then you need to apply the three patches I just posted. Then you can do a "make sync LINUX=/path/to/kernel/tree" and then a make. You should end up with a virtio.ko, virtio_ring.ko, virtio_pci.ko, and a virtio_blk.ko.

Regards,

Anthony Liguori

Quoted 20 lines · Anthony Liguori

Hi,
I've posted a repo that contains my initial attempt at backporting the virtio modules. I've tested against a 2.6.22, 2.6.20, and 2.6.18 kernel (although I haven't tried the block device against 2.6.18). It requires Rusty's patch queue plus the three patches I just sent to virtualization@l-f.
The makefile supports a 'make sync LINUX=/path/to/linux' just like the kvm-userspace/kernel directory does.
This is still rough as I haven't went through and made sure all of the #if's are right.
http://hg.codemonkey.ws/virtio-ext-modules
Regards,
Anthony Liguori

Quoted 9 lines

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

Avi Kivity

2008-01-03 11:54 UTC #3 lore raw

In reply to a message not in the archive

Quoted 8 lines · Anthony Liguori

Hi,
I've posted a repo that contains my initial attempt at backporting the virtio modules. I've tested against a 2.6.22, 2.6.20, and 2.6.18 kernel (although I haven't tried the block device against 2.6.18). It requires Rusty's patch queue plus the three patches I just sent to virtualization@l-f.

Quoted 9 lines

The makefile supports a 'make sync LINUX=/path/to/linux' just like the kvm-userspace/kernel directory does.
This is still rough as I haven't went through and made sure all of the #if's are right.
http://hg.codemonkey.ws/virtio-ext-modules

I think that 'typedef _Bool bool' is better than unsigned int, as it forces conversions to 0/1.

Modern Linux uses _Bool as well.

Signature · 7 lines

error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

A message that is not in the archive (sent off-list or not yet received): 477CCCE2.5090307-atKUWr5tajBWk0Htik3J/[email protected]

Anthony Liguori

author

2008-01-03 15:16 UTC #5 lore raw

Quoted 15 lines · Avi Kivity

Quoted 14 lines · Anthony Liguori
Hi,
I've posted a repo that contains my initial attempt at backporting the virtio modules. I've tested against a 2.6.22, 2.6.20, and 2.6.18 kernel (although I haven't tried the block device against 2.6.18). It requires Rusty's patch queue plus the three patches I just sent to virtualization@l-f.
The makefile supports a 'make sync LINUX=/path/to/linux' just like the kvm-userspace/kernel directory does.
This is still rough as I haven't went through and made sure all of the #if's are right.

Quoted 9 lines

Quoted 3 lines
http://hg.codemonkey.ws/virtio-ext-modules
I think that 'typedef _Bool bool' is better than unsigned int, as it forces conversions to 0/1.
Modern Linux uses _Bool as well.

Fair enough. I've updated.

Regards,

Anthony Liguori