[libvirt] [PATCH] Read PCI class from sysfs class file instead of config space.

[Thadeu Lima de Souza Cascardo]

On Tue, Jan 07, 2014 at 09:53:51AM -0700, Eric Blake wrote:
> On 01/07/2014 09:35 AM, Michal Privoznik wrote:
> 
> >> diff --git a/tests/virpcitestdata/0001:00:00.0.config b/tests/virpcitestdata/0001:00:00.0.config
> >> new file mode 100644
> >> index 0000000000000000000000000000000000000000..808d48993cfc0f41223fcb5f49deffc594f136b7
> >> GIT binary patch
> >> literal 4096
> >> zcmeH at I}XAy5Jca`5~RZg60MJb!~ys;a0<jx5^hCDNy!mt=t)n(8Yfir2x&(4YG(bB
> >> z{ig90wibyn0^=hytn<`78f&|D=zEvd5U8+Sxa1hw*nI*%I6fEDtkd58IUH=(- at Ei&
> >> z`ONZ@#r(MD|GagrnWu5_32tAW*RPg6sv;l)A|L`HAOa#F0wN#+A|L`H at J9q*PZkdc
> >>
> >> literal 0
> >> HcmV?d00001
> >>
> > 
> > Some binary garbage ...
> 
> I know this was copied from a working sample file, as opposed to
> something you generated (if it were generated, I'd request that we
> instead version-control the human-readable source that got converted
> into the binary blob).  As long as it doesn't trip up 'make
> syntax-check' or git server side hooks, I think we can live with this.
> Does the resulting file contain any newlines or carriage returns that
> might get botched on Windows platforms, if we don't add a .gitattributes
> listing that mentions it is explicitly a binary file?
> 

This was copied from a system where it was failing before. I also copied
the config files from a working device + bridge.

make syntax-check seems to work, but some of the files have a newline
character. I haven't worked with git on Windows. Is that going to be a
problem?

Regards.
Cascardo.

> 
> -- 
> Eric Blake   eblake redhat com    +1-919-301-3266
> Libvirt virtualization library http://libvirt.org
>