Re: [Bug Report] Linux v6.15-rc7 boot failure on Xen-4.17

[Xin Li <xin@zytor.com>]

On 5/24/2025 11:58 PM, Jürgen Groß wrote:
>> Just noticed vanilla v6.15-rc7 fails to boot on Xen 4.17 (Ubuntu Xen
>> package) in a KVM guest.  So I did a bisect, and it says:
>>
>> 872df34d7c51a79523820ea6a14860398c639b87 is the first bad commit
>> commit 872df34d7c51a79523820ea6a14860398c639b87
>> Author: Peter Zijlstra <peterz@infradead.org>
>> Date:   Mon Oct 14 10:05:48 2024 -0700
>>
>>      x86/its: Use dynamic thunks for indirect branches
>>
>>
>> Attached is the serial console log and my kernel config.
> 
> So you are saying that the host kernel is to blame, given that above
> commit isn't in the dom0 kernel.

No, it's the KVM guest.  I didn't change host kernel during the
bisecting.

I didn't dig further, but it looks related to changes in:
     arch/x86/kernel/alternative.c
> Can you start e.g. a normal Linux guest on that host?
The same kernel binary boots fine as native.

> Any host messages to be seen when that happens?

[469584.422141] virbr0: port 1(vnet12) entered blocking state
[469584.422149] virbr0: port 1(vnet12) entered disabled state
[469584.422164] vnet12: entered allmulticast mode
[469584.422297] vnet12: entered promiscuous mode
[469584.422716] virbr0: port 1(vnet12) entered blocking state
[469584.422721] virbr0: port 1(vnet12) entered listening state
[469586.477553] virbr0: port 1(vnet12) entered learning state
[469588.526074] virbr0: port 1(vnet12) entered forwarding state
[469588.526082] virbr0: topology change detected, propagating
[469624.321914] virbr0: port 1(vnet12) entered disabled state
[469624.332737] vnet12 (unregistering): left allmulticast mode
[469624.332743] vnet12 (unregistering): left promiscuous mode
[469624.332765] virbr0: port 1(vnet12) entered disabled state

Thanks!
     Xin