Re: [PATCH v2 for-4.15 3/7] CHANGELOG.md: Add dom0/domU zstd compression support

[Julien Grall <julien@xxxxxxx>]

On 30/03/2021 11:56, Jan Beulich wrote:
> On 30.03.2021 12:14, Julien Grall wrote:
> > On 30/03/2021 11:08, George Dunlap wrote:
> > > > On Mar 29, 2021, at 7:54 PM, Julien Grall <julien.grall.oss@xxxxxxxxx> wrote:
> > > >
> > > > Hi George,
> > > >
> > > > On Mon, 29 Mar 2021 at 17:15, George Dunlap <george.dunlap@xxxxxxxxxx> wrote:
> > > > > diff --git a/CHANGELOG.md b/CHANGELOG.md
> > > > > index 8c89212f14..538eae611c 100644
> > > > > --- a/CHANGELOG.md
> > > > > +++ b/CHANGELOG.md
> > > > > @@ -12,6 +12,7 @@ The format is based on [Keep a 
> > > > > Changelog](https://keepachangelog.com/en/1.0.0/)
> > > > >    - ARM SMMUv3 (Tech Preview)
> > > > >    - Intel Processor Trace support (Tech Preview)
> > > > >    - Named PCI devices for xl/libxl
> > > > > + - Support for zstd-compressed dom0 and domU kernels
> > > > Looking at the log, the feature looks x86 only (at least the dom0 part).
> > > Oh, interesting.  So what about the following?
> > >
> > > - Support for zstd-compressed dom0 (x86) and domU kernels
> > Sounds fine to me. Note that I haven't tried zstd-compressed for domu
> > Kernel on Arm.
> >
> > Maybe the author (Jan) can confirm whether this was plumbed in common
> > libxc code?
> Well, I've followed the model used for other compression methods. I have
> no idea what this means or does not mean for Arm; in fact so far I was
> under the impression that like in the hypervisor decompression support
> was an x86-only feature, but from what you say I conclude I have been
> wrong with this.
We support gzip for dom0. For guests, I am not entirely sure what the 
state as I tend to uncompressed kernel every time.
Cheers,

--
Julien Grall