[REGRESSION] Cuttlefish boot issue on lts branches

[REGRESSION] Cuttlefish boot issue on lts branches

[Terry Tritton]

Hi all,
We've seen a regression on several lts branches(at least 5.10,5.15,6.1) causing
boot issues on cuttlefish/crossvm android emulation on arm64.

The offending patch is:
    PCI: Use preserve_config in place of pci_flags

It looks like this patch was added to stable by AUTOSEL but without the other
3 patches in the series:
    https://lore.kernel.org/all/20240508174138.3630283-1-vidyas@nvidia.com/

Applying the missing patches resolves the issue but they do not apply cleanly.

Can we revert this patch on the lts branches?

Re: [REGRESSION] Cuttlefish boot issue on lts branches

[Greg KH]

On Fri, Jan 17, 2025 at 11:01:17AM +0000, Terry Tritton wrote:
> Hi all,
> We've seen a regression on several lts branches(at least 5.10,5.15,6.1) causing
> boot issues on cuttlefish/crossvm android emulation on arm64.
> 
> The offending patch is:
>     PCI: Use preserve_config in place of pci_flags
> 
> It looks like this patch was added to stable by AUTOSEL but without the other
> 3 patches in the series:
>     https://lore.kernel.org/all/20240508174138.3630283-1-vidyas@nvidia.com/
> 
> Applying the missing patches resolves the issue but they do not apply cleanly.
> 
> Can we revert this patch on the lts branches?

Sure, can you please send the revert with all of this information so we
get it correct and you get proper credit for finding/fixing it?

thanks,

greg k-h

Re: [REGRESSION] Cuttlefish boot issue on lts branches

[Terry Tritton]

Thanks Greg,
whats the best way to do that?
Do you need a patch for each lts branch affected or just one based on stable?

Re: [REGRESSION] Cuttlefish boot issue on lts branches

[Greg KH]

On Fri, Jan 17, 2025 at 01:34:14PM +0000, Terry Tritton wrote:
> Thanks Greg,
> whats the best way to do that?
> Do you need a patch for each lts branch affected or just one based on stable?

One for each branch please, as the git ids for the commit is different
on each one, right?

Look at commit acddb8762014 ("Revert "rcu-tasks: Fix access non-existent
percpu rtpcp variable in rcu_tasks_need_gpcb()"") for an idea of how to
write the header.

thanks,

greg k-h

Re: [REGRESSION] Cuttlefish boot issue on lts branches

[Terry Tritton]

> One for each branch please, as the git ids for the commit is different
> on each one, right?

Thanks that's what I assumed but wasn't sure if there was some script
that did some magic.

Sent those to stable now, let me know if there are any issues.

Thanks
Terry

Re: [REGRESSION] Cuttlefish boot issue on lts branches

[Greg KH]

On Fri, Jan 17, 2025 at 03:18:01PM +0000, Terry Tritton wrote:
> > One for each branch please, as the git ids for the commit is different
> > on each one, right?
> 
> Thanks that's what I assumed but wasn't sure if there was some script
> that did some magic.
> 
> Sent those to stable now, let me know if there are any issues.

Looks good, I'll get to them next week...

greg k-h