Re: [Qemu-devel] [PULL 0/8] 9p patches for 2.8 20160916

[Greg Kurz]

On Fri, 16 Sep 2016 15:25:42 +0100
Peter Maydell <address@hidden> wrote:

> On 16 September 2016 at 14:09, Greg Kurz <address@hidden> wrote:
> > The following changes since commit 5f473241ac595452ae0638dc63e7af2a2294f5ec:
> >
> >   Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into 
> > staging (2016-09-15 18:12:40 +0100)
> >
> > are available in the git repository at:
> >
> >   https://github.com/gkurz/qemu.git tags/for-upstream
> >
> > for you to fetch changes up to 5c574f1bf92cfaf4aae6cbb66b79066c654920e1:
> >
> >   9pfs: fix potential segfault during walk (2016-09-16 12:16:29 +0200)
> >
> > ----------------------------------------------------------------
> > This pull request contains:
> > - a fix for a regression introduced in 2.7
> > - basic functional testing for virtio-9p
> > - some code cleanups for 9pfs
> >
> > ----------------------------------------------------------------  
> 
> Fails test on ppc64be:
> 
>   /i386/virtio/9p/pci/basic/transaction:                               **
> ERROR:/home/pm215/qemu/tests/virtio-9p-test.c:167:pci_basic_transaction:
> assertion failed (hdr.size
> < (uint32_t) P9_MAX_SIZE): (603979776 < 8192)
> FAIL
> GTester: last random seed: R02S0ce79f32586824bc8cbd5461f009c62a
> (pid=33294)
> 
> Looks like an endianness bug.
> 

Ok, I'll look into it... but the important part in this pull request is
the "9pfs: fix potential segfault during walk" patch. It fixes a regression
introduced in 2.7 by the 9P security fixes. And IIUC, Michael Roth is about
to release 2.6.1.1 with these fixes and the regression...

Should I send a new pull request without the qtest patches ? Or with the
regression fix only ?

Please advise.

Thanks.

--
Greg

> thanks
> -- PMM