Re: [Xen-devel] [PATCH V2] arch: arm: vgic-v3: fix GICD_ISACTIVER range

[Julien Grall <julien@xxxxxxx>]

Hi,

On 28/11/2019 08:32, Jürgen Groß wrote:
> On 28.11.19 09:14, Julien Grall wrote:
> > > In short, I think the patch should go in now and there are no downsides
> > > to it. That's it, I rest my case. Julien, I hope you'll reconsider.
> > I don't really see the point to try to allow Linux 5.4 booting on Xen 
> > 4.13 without knowing whether we are not going to uncovered more BUG 
> > around I*ACTIVER.
> Sorry, but this is a rather weird statement.
>
> IIUC you are saying that a typo blocking boot of Linux 5.4 should not be
> fixed as you are not sure there are no other bugs?
The implementation of I*ACTIVER was incorrect but gone unnoticed because 
no-one used it until 5.4. It also happen that we didn't cover all the 
I*ACTIVER registers, so 5.4 crashes instead of using the wrong behavior.
This patch is basically replacing a guest crash by a behavior we don't 
fully understand.
> > If you really want this patch in Xen 4.13, then you should read my 
> > mail on the first version and trying to answer me why this 5.4 is 
> > going to be safe running on Xen 4.13.
> Or do you think that with the typo fixed and running Linux 5.4 guests
> will destabilize the host?
It is not going to destabilize the hosts. But this is not going to make 
5.4 running correctly as Xen guest.
Cheers,

--
Julien Grall

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel