kvm-62 release

Fixes the longstanding RHEL 5 i386 smp boot problem as well as the recent Intel Core regression. Large page support improves performance for dedicated virtualization hosts, while vmware vga and more std-vga resolutions should improve the home user experience.

Changes from kvm-61:
- fix oops running external module on 2.6.25 due to mmdrop (Andrea Arcangeli)
- backward compatibility for 2.6.25 desc_struct changes
- fix memory alias mixup (Uri Lublin)
- add file:// migration protocol (Uri Lublin)
- fix reading from character devices with flow control (Jan Kiszka)
- large host page support (Marcelo Tosatti)
- fix screen corruption with high resolutions (Andreas Winkelbauer)
- implement vmware vga for kvm (Soren Hansen, Anthony Liguori)
- even more resolutions for -std-vga (Andreas Winkelbauer)
- code cleanups (Carlo Marcelo Arenas Belon, Harvey Harrison)
- allow up to 16 vcpus (Anthony Liguori, me)
- allow up to 16 memory slots
- improve vcpu_run error reporting
- fix host crash on hosts without EFER (Intel Core not 2s)
- drop kvm private segment descriptor in favor of Linux' desc_struct
- stat counter for hypercalls (Amit Shah)
- fake values for MSR_PERF_STATUS (Alexander Graf)
- needed for Darwin/OS X
- better root pagetable unpinning (Marcelo Tosatti)
- paravirt clock fixes (Glauber Costa, Marcelo Tosatti)
- avoid host hang when guest sets APIC timer to zero
- also fixes live migration of Windows 2000 guests
- fix RHEL 5 i386 smp boot due to timer interrupts being routed to all vcpus

Notes:
If you use the modules bundled with kvm-62, you can use any version
of Linux from 2.6.17 upwards.
If you use the modules bundled with Linux 2.6.20, you need to use
kvm-12.
If you use the modules bundled with Linux 2.6.21, you need to use
kvm-17.
Modules from Linux 2.6.22 and up will work with any kvm version from
kvm-22. Some features may only be available in newer releases.
For best performance, use Linux 2.6.23-rc2 or later as the host.

http://kvm.qumranet.com

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

- add file:// migration protocol (Uri Lublin)

Hi,

Quoted 2 lines

- add file:// migration protocol (Uri Lublin)

commit 70d2a9dc3594446f3ad66b32abc94c08b74118d4

Author: Uri Lublin <[email protected]>
Date:   Thu Feb 21 15:00:50 2008 +0200

kvm: qemu: migration: added file:// migration protocol

Instead of writing into a file with fork+exec, just open the file and write the data. This solves a bug that sometimes not all the data was kept in the state-file.

I haven't tested it yet, but this sounds like the problems I've always had with migration. It seems this is just a new way to work around it, rather than an actual fix for the bug though -- won't exec: still have the same problem even when not writing to a file? Was this ever tracked down any further? Or is there a known reason why "exec:dd" is expected to be broken? I never had any luck finding it..

Quoted 4 lines

- add file:// migration protocol (Uri Lublin)

commit 70d2a9dc3594446f3ad66b32abc94c08b74118d4

Quoted 9 lines

Author: Uri Lublin <[email protected]>
Date:   Thu Feb 21 15:00:50 2008 +0200

kvm: qemu: migration: added file:// migration protocol

Instead of writing into a file with fork+exec, just open the file and write the data. This solves a bug that sometimes not all the data was kept in the state-file.

I haven't tested it yet, but this sounds like the problems I've always had with migration. It seems this is just a new way to work around it, rather than an actual fix for the bug though -- won't exec: still have the same problem even when not writing to a file? Was this ever tracked down any further? Or is there a known reason why "exec:dd" is expected to be broken? I never had any luck finding it..

-jim

Opening a file, and writing into it is simpler anyway. I too do not know yet why "exec:dd" sometimes fails.