Re: Backport request

[Krzysztof Kozlowski <krzk@kernel.org>]

On 23/04/2026 08:46, Greg KH wrote:
> On Thu, Apr 23, 2026 at 08:39:29AM +0200, Krzysztof Kozlowski wrote:
>> On 23/04/2026 06:55, Greg KH wrote:
>>> On Wed, Apr 22, 2026 at 06:04:50PM +0100, Josh Law wrote:
>>>> Hello, I would like backports for 
>>>>
>>>> Mainline hashes:
>>>>
>>>> https://github.com/torvalds/linux/commit/8cdf30813ea8ce881cecc08664144416dbdb3e16
>>>>
>>>> https://github.com/torvalds/linux/commit/9003ec6f7f394943880618737d797a9f257e6e1e
>>>
>>> None of those have showed up in an actual release yet, so why should
>>> they be included "early"?
>>
>> None of the code was tested as Josh Law lied more than once about tests
>> [1] or laughed at us when we asked for testing:
>>
>> "laugh my ass out and your test cases, absolutely ill add some test
>> cases" [2]
>>
>> and then Josh Law was pushing his patches to get merged:
>>
>> "This most definitely needs to be merged." [3]
>> "Yeah in my opinion I think this may need to be merged.. if you would
>> like I can add the NOWARN" [4]
>>
>> And now we see a push for these commits to stable!
>>
>> Nothing from Josh Law should be going to stable trees, because nothing
>> was ever tested.
> 
> Makes sense, is anyone going to send reverts for these?

Untested does not mean yet incorrect, so not sure if we need to act on
already accepted commits getting to stable. Especially that for a revert
I would need to provide some stronger arguments, IMO. Easier to drop
from the queue in review.

OTOH, if the actor is not trusted, a past correct patch is not a
positive indication for another patch. Untrusted, for whatever reason,
actor can write correct patches for some time...

When untrusted person pushes why things are not in stable, it feels to
me like a warning sign.

> 
> I'll drop them from my "to review" stable queues, that includes the
> following:
> 
> 	 mm/damon/sysfs: check contexts->nr in repeat_call_fn
> 	 mm/damon/sysfs: fix param_ctx leak on damon_sysfs_new_test_ctx() failure
> 	 mm/damon/sysfs: check contexts->nr before accessing contexts_arr[0]
> 	 lib/ts_bm: fix integer overflow in pattern length calculation
> 	 lib/ts_kmp: fix integer overflow in pattern length calculation


Best regards,
Krzysztof