Re: [PATCH 0/8] hw/cxl: CXL emulation cleanups and minor fixes for upstream

[Lukas Wunner <lukas@wunner.de>]

On Fri, Jan 13, 2023 at 02:45:11PM +0000, Jonathan Cameron wrote:
> On Fri, 13 Jan 2023 14:40:26 +0000 Jonathan Cameron <Jonathan.Cameron@Huawei.com> wrote:
> > On Fri, 13 Jan 2023 09:19:59 -0500 Gregory Price <gregory.price@memverge.com> wrote:
> > > On Fri, Jan 13, 2023 at 09:12:13AM +0000, Jonathan Cameron wrote:  
> > > > Just to check, are these different from the on stack problem you reported
> > > > previously?  Doesn't look like the fix for that has made it upstream yet.
> > > > 
> > > > What kernel are you running?
> > > 
> > > The prior issue I saw was related to the CXL Fixed Memory Window having
> > > an e820 region registered during machine initialization.  That fix is
> > > upstream.
> > > 
> > > On 2023-1-11 branch it is commit 2486dd045794d65598fbca9fd1224c27b9732dce
> > > 
> > > This one appears when registering any kind of type-3 device, during
> > > boot.  
> > 
> > I meant this one
> > 
> > https://lore.kernel.org/all/20221118000524.1477383-1-ira.weiny@intel.com/
> > 
> > Sorry, should have dug out a link in earlier reply and save a round trip.
> 
> Ah I'd forgotten we were going round the houses somewhat on the right fix...
> https://lore.kernel.org/all/20221128040338.1936529-3-ira.weiny@intel.com/
> was another proposal and Lukas had yet another.
> https://lore.kernel.org/all/cover.1669608950.git.lukas@wunner.de/

I'll respin those fixes shortly, together with the other DOE patches
I've accumulated on my development branch for DOE+SPDM:

https://github.com/l1k/linux/commits/doe

Ira kindly tested them this week and reports that they don't regress
CDAT retrieval for him.

Thanks,

Lukas