[PATCH 0 of 12] backport critical fixes from qemu upstream

[PATCH 0 of 12] backport critical fixes from qemu upstream

[Stefano Stabellini]

Hi all,
the current qemu block layer contains few critical problems, one of them
was fixed by "fix qemu memory leak in block interface" but unfortunately
others remain.
In particular every time the guest cancels a dma request qemu
segfaults.
In order to completely fix any remaining bug I suggest to revert the
previous fix "fix qemu memory leak in block interface" and backport
instead the following changesets from qemu upstream:

6bbff9a0b495918309074ac60375be5f9dc868b3 # Refactor aio callback allocation to use an aiocb pool (Avi Kivity)
c07a9008ac6985cd5a15909c2b9977d982defc12 # Convert vectored aio emulation to use a dedicated pool (Avi Kivity)
6512a2a7106480c19183d6466a6845bc9bdf6ec0 # Implement cancellation method for dma async I/O (Avi Kivity)
3fb94d56c6adc96ffc4a81b58c752cc4ccfae39c # Use vectored aiocb storage to store vector translation state (Avi Kivity)
37b7842c2fb405c270efdce714425c17af3c78cd # Move block dma helpers aiocb to store dma state (Avi Kivity)
c240b9af599d20e06a58090366be682684bd8555 # Fix vectored aio bounce handling immediate errors (Avi Kivity)
7403b14eeb4670d54497284b110ca3e3be4a99a4 # Fix DMA API when handling an immediate error from block layer (Avi Kivity)
081501daceca74d2091e798d5dcd342c14d66d1a # check for bs->drv in bdrv_flush (Christoph Hellwig)
fbb7b4e0804d2168f24142eebf7552adde1968dc # Improve block range checks
6a7ad2998ca47711eddd191caa1a95f14459793f # Call qemu_bh_delete at bdrv_aio_bh_cb.
36afc451599175eb8f03f1962de74e7f7a8af4db # block: Clean up after deleting BHs

The same should be done to qemu-xen-3.4-testing as well.
Cheers,

Stefano

Re: [PATCH 0 of 12] backport critical fixes from qemu upstream

[Mark Johnson]

Ian,

On Mon, Oct 5, 2009 at 9:40 AM, Stefano Stabellini
<stefano.stabellini@eu.citrix.com> wrote:
> Hi all,
> the current qemu block layer contains few critical problems, one of them
> was fixed by "fix qemu memory leak in block interface" but unfortunately
> others remain.
> In particular every time the guest cancels a dma request qemu
> segfaults.
> In order to completely fix any remaining bug I suggest to revert the
> previous fix "fix qemu memory leak in block interface" and backport
> instead the following changesets from qemu upstream:


> The same should be done to qemu-xen-3.4-testing as well.
> Cheers,

Are you planning to push these to qemu-xen-3.4-testing  soon too?



Thanks,

MRJ

Re: [PATCH 0 of 12] backport critical fixes from qemu upstream

[Ian Jackson]

Mark Johnson writes ("Re: [Xen-devel] [PATCH 0 of 12] backport critical fixes from qemu upstream"):
> On Mon, Oct 5, 2009 at 9:40 AM, Stefano Stabellini
> <stefano.stabellini@eu.citrix.com> wrote:
> > The same should be done to qemu-xen-3.4-testing as well.
> > Cheers,
> 
> Are you planning to push these to qemu-xen-3.4-testing  soon too?

Yes, but I'm giving it a few days to settle down in qemu-xen-unstable
first, as that seemed sensible ...

Ian.

Re: [PATCH 0 of 12] backport critical fixes from qemu upstream

[Florian Manschwetus]

[-- Attachment #1.1: Type: text/plain, Size: 916 bytes --]

Am 09.10.2009 15:58, schrieb Ian Jackson:
> Mark Johnson writes ("Re: [Xen-devel] [PATCH 0 of 12] backport critical fixes from qemu upstream"):
>> On Mon, Oct 5, 2009 at 9:40 AM, Stefano Stabellini
>> <stefano.stabellini@eu.citrix.com> wrote:
>>> The same should be done to qemu-xen-3.4-testing as well.
>>> Cheers,
>>
>> Are you planning to push these to qemu-xen-3.4-testing  soon too?
> 
> Yes, but I'm giving it a few days to settle down in qemu-xen-unstable
> first, as that seemed sensible ...
> 
> Ian.
Ok, but how much are a few days?

This,
2106 xvm         3  59    0 5700M 3743M sleep  758:48  0.28% qemu-dm

is quite a real showstopper and a big pain in my...

So, sorry but this fix is really a critical one.

Florian
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
> 



[-- Attachment #1.2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 3686 bytes --]

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: [PATCH 0 of 12] backport critical fixes from qemu upstream

[Mark Johnson]

On Fri, Oct 9, 2009 at 9:58 AM, Ian Jackson <Ian.Jackson@eu.citrix.com> wrote:
> Mark Johnson writes ("Re: [Xen-devel] [PATCH 0 of 12] backport critical fixes from qemu upstream"):
>> On Mon, Oct 5, 2009 at 9:40 AM, Stefano Stabellini
>> <stefano.stabellini@eu.citrix.com> wrote:
>> > The same should be done to qemu-xen-3.4-testing as well.
>> > Cheers,
>>
>> Are you planning to push these to qemu-xen-3.4-testing  soon too?
>
> Yes, but I'm giving it a few days to settle down in qemu-xen-unstable
> first, as that seemed sensible ...

FYI, I've pull these changes into 3.4 and qemu-dm is still leaking
a ton of memory for HVM guests which do not have PV drivers..

qemu-dm does not appear to leak memory for a HVM guest with
PV net & disk drivers.



MRJ