[REGRESSION] v6.6.120: i3c crash caused by commit 82a09b9965ed

[REGRESSION] v6.6.120: i3c crash caused by commit 82a09b9965ed

[Kunihiko Hayashi]

Dear stable maintainers,

After updating from v6.6.119 to v6.6.120, I noticed a kernel crash
when I3C was enabled.

This regression is caused by:

     commit 82a09b9965ed ("i3c: fix refcount inconsistency in i3c_master_register")

The issue is resolved when the following upstream fix commit is applied:

     commit 3502cea99c7c ("i3c: Move device name assignment after i3c_bus_init")

I also confirmed that commit 82a09b9965ed is applied to other stable trees,
including at least linux-6.1.y, so the same regression may affect them
as well.

Applying the fix commit or reverting 82a09b9965ed resolves the issue.

Thank you,

---
Best Regards
Kunihiko Hayashi

Re: [REGRESSION] v6.6.120: i3c crash caused by commit 82a09b9965ed

[Greg Kroah-Hartman]

On Wed, Jan 21, 2026 at 08:04:03PM +0900, Kunihiko Hayashi wrote:
> Dear stable maintainers,
> 
> After updating from v6.6.119 to v6.6.120, I noticed a kernel crash
> when I3C was enabled.
> 
> This regression is caused by:
> 
>     commit 82a09b9965ed ("i3c: fix refcount inconsistency in i3c_master_register")
> 
> The issue is resolved when the following upstream fix commit is applied:
> 
>     commit 3502cea99c7c ("i3c: Move device name assignment after i3c_bus_init")

That does not seem to be a valid git id, where did it come from?

thanks,

greg k-h

Re: [REGRESSION] v6.6.120: i3c crash caused by commit 82a09b9965ed

[Alexandre Belloni]

On 21/01/2026 15:56:01+0100, Greg Kroah-Hartman wrote:
> On Wed, Jan 21, 2026 at 08:04:03PM +0900, Kunihiko Hayashi wrote:
> > Dear stable maintainers,
> > 
> > After updating from v6.6.119 to v6.6.120, I noticed a kernel crash
> > when I3C was enabled.
> > 
> > This regression is caused by:
> > 
> >     commit 82a09b9965ed ("i3c: fix refcount inconsistency in i3c_master_register")
> > 
> > The issue is resolved when the following upstream fix commit is applied:
> > 
> >     commit 3502cea99c7c ("i3c: Move device name assignment after i3c_bus_init")
> 
> That does not seem to be a valid git id, where did it come from?
> 

This has not yet been sent to Linus and my plan was to wait for the
merge window as the fix didn't make it clear this was actually happening
in the field.


-- 
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

Re: [REGRESSION] v6.6.120: i3c crash caused by commit 82a09b9965ed

[Greg Kroah-Hartman]

On Wed, Jan 21, 2026 at 05:57:38PM +0100, Alexandre Belloni wrote:
> On 21/01/2026 15:56:01+0100, Greg Kroah-Hartman wrote:
> > On Wed, Jan 21, 2026 at 08:04:03PM +0900, Kunihiko Hayashi wrote:
> > > Dear stable maintainers,
> > > 
> > > After updating from v6.6.119 to v6.6.120, I noticed a kernel crash
> > > when I3C was enabled.
> > > 
> > > This regression is caused by:
> > > 
> > >     commit 82a09b9965ed ("i3c: fix refcount inconsistency in i3c_master_register")
> > > 
> > > The issue is resolved when the following upstream fix commit is applied:
> > > 
> > >     commit 3502cea99c7c ("i3c: Move device name assignment after i3c_bus_init")
> > 
> > That does not seem to be a valid git id, where did it come from?
> > 
> 
> This has not yet been sent to Linus and my plan was to wait for the
> merge window as the fix didn't make it clear this was actually happening
> in the field.

So we are bug-compatible with Linus's tree right now?  Great, all should
be fine :)

thanks,

greg k-h

Re: [REGRESSION] v6.6.120: i3c crash caused by commit 82a09b9965ed

[Kunihiko Hayashi]

Hi Greg, Alexandre,

On 2026/01/22 2:20, Greg Kroah-Hartman wrote:
> On Wed, Jan 21, 2026 at 05:57:38PM +0100, Alexandre Belloni wrote:
>> On 21/01/2026 15:56:01+0100, Greg Kroah-Hartman wrote:
>>> On Wed, Jan 21, 2026 at 08:04:03PM +0900, Kunihiko Hayashi wrote:
>>>> Dear stable maintainers,
>>>>
>>>> After updating from v6.6.119 to v6.6.120, I noticed a kernel crash
>>>> when I3C was enabled.
>>>>
>>>> This regression is caused by:
>>>>
>>>>      commit 82a09b9965ed ("i3c: fix refcount inconsistency in
> i3c_master_register")
>>>>
>>>> The issue is resolved when the following upstream fix commit is
> applied:
>>>>
>>>>      commit 3502cea99c7c ("i3c: Move device name assignment after
> i3c_bus_init")
>>>
>>> That does not seem to be a valid git id, where did it come from?
>>>
>>
>> This has not yet been sent to Linus and my plan was to wait for the
>> merge window as the fix didn't make it clear this was actually happening
>> in the field.
> 
> So we are bug-compatible with Linus's tree right now?  Great, all should
> be fine :)

Thanks for the clarification.

I saw the commit in linux-next tree.
Understood that the fix hasn't yet landed in Linus's tree and
therefore cannot be picked up by stable at this point.

Sorry for the early report, and thank you for taking a look.

Thank you,

---
Best Regards
Kunihiko Hayashi

Re: [REGRESSION] v6.6.120: i3c crash caused by commit 82a09b9965ed

[Kunihiko Hayashi]

Hi,

On 2026/01/22 14:58, Kunihiko Hayashi wrote:
> Hi Greg, Alexandre,
> 
> On 2026/01/22 2:20, Greg Kroah-Hartman wrote:
>> On Wed, Jan 21, 2026 at 05:57:38PM +0100, Alexandre Belloni wrote:
>>> On 21/01/2026 15:56:01+0100, Greg Kroah-Hartman wrote:
>>>> On Wed, Jan 21, 2026 at 08:04:03PM +0900, Kunihiko Hayashi wrote:
>>>>> Dear stable maintainers,
>>>>>
>>>>> After updating from v6.6.119 to v6.6.120, I noticed a kernel crash
>>>>> when I3C was enabled.
>>>>>
>>>>> This regression is caused by:
>>>>>
>>>>>      commit 82a09b9965ed ("i3c: fix refcount inconsistency in
>> i3c_master_register")
>>>>>
>>>>> The issue is resolved when the following upstream fix commit is
>> applied:
>>>>>
>>>>>      commit 3502cea99c7c ("i3c: Move device name assignment after
>> i3c_bus_init")
>>>>
>>>> That does not seem to be a valid git id, where did it come from?
>>>>
>>>
>>> This has not yet been sent to Linus and my plan was to wait for the
>>> merge window as the fix didn't make it clear this was actually happening
>>> in the field.
>>
>> So we are bug-compatible with Linus's tree right now?  Great, all should
>> be fine :)
> 
> Thanks for the clarification.
> 
> I saw the commit in linux-next tree.
> Understood that the fix hasn't yet landed in Linus's tree and
> therefore cannot be picked up by stable at this point.

The upstream commit
   3502cea99c7c ("i3c: Move device name assignment after i3c_bus_init")
has now been merged into Linus's tree.

As far as I've checked, v6.12 and older versions are still affected.
Could you please consider this for stable?

Thank you,

---
Best Regards
Kunihiko Hayashi