Re: [PATCH v2] kvm-xfstests, gce-xfstests: improve recommended kernel config

[Eric Biggers <ebiggers3@gmail.com>]

On Thu, May 25, 2017 at 01:24:11PM -0400, Theodore Ts'o wrote:
> On Tue, May 23, 2017 at 10:45:57AM -0700, Eric Biggers wrote:
> > From: Eric Biggers <ebiggers@google.com>
> > 
> > Trim more unneeded features from the recommended kernel config, then
> > enable some more relatively lightweight debugging options.
> > 
> > For now, just update the latest x86_64 config.  But if people are okay
> > with these changes I'll update some of the other configs too.
> > 
> > Tested with both kvm-xfstests and gce-xfstests.
> > 
> > Signed-off-by: Eric Biggers <ebiggers@google.com>
> 
> This looks good to me.  Would you mind sending a patch which updates
> the rest of the configs similarly, so we can update all of the configs
> at once?
> 
> As an aside, I *do* use the HW_RANDOM_VIRTIO from time to time (that's
> why it was enabled in the config), and in fact kvm-xfstests is wired
> up to supply the virtio random number generator by default (since it
> doesn't really cost much):
> 
> if test -z "$NO_VIRTIO_RNG" ; then
>     VIRTIO_RNG="-object rng-random,filename=/dev/urandom,id=rng0 \
> 	-device virtio-rng-pci,rng=rng0"
> fi
> 
> But that's for my kernel RNG testing, and it's fair to say that that
> most users of kvm-xfstests won't need it.
> 

Yes I'll do that.  I'll probably leave HW_RANDOM_VIRTIO on, given that the
virtio-rng-pci device is emulated by default, and the kernel config ought to
support the default emulated hardware.  (Or else we should turn the
virtio-rng-pci device off by default.)

Eric