Re: [PATCH v5 0/8] mm/slab: introduce kfree_rcu_nolock() and improve slub_kunit coverage

[Vlastimil Babka (SUSE)]

On 7/30/26 05:08, Harry Yoo wrote:
> 
> 
> On 7/29/26 11:14 PM, Vlastimil Babka (SUSE) wrote:
>> On 7/29/26 10:20, Harry Yoo (Oracle) wrote:
>> 
>> ...
>> 
>>> 8:  8e69c83a1e52 ! 8:  aeaf2ab60046 slub_kunit: extend the test for kfree_rcu_nolock()
>>>     @@ lib/tests/slub_kunit.c: struct test_nolock_context {
>>>      -			gfp_t gfp = (i % 2) ? GFP_KERNEL : GFP_KERNEL_ACCOUNT;
>>>      +			gfp_t gfp = (i & 1) ? GFP_KERNEL : GFP_KERNEL_ACCOUNT;
>>>       
>>>     - 			objects[j] = kmalloc_obj(*objects[j], gfp);
>>>     +-			objects[j] = kmalloc(64, gfp);
>>>     ++			objects[j] = kmalloc_obj(*objects[j], gfp);
>>>       			if (!objects[j]) {
>>>     + 				j--;
>>>     + 				while (j >= 0)
>>>      @@ lib/tests/slub_kunit.c: static void test_kmalloc_kfree(void)
>>>       			}
>>>       		}
>>>
>>> Link: https://lore.kernel.org/linux-mm/20260427-nolock-api-fix-v2-0-a6b83a92d9a4@xxxxxxxxxx [1]
>>> Link: https://lore.kernel.org/linux-mm/6edebc2b-5f5a-4b9c-9a4c-564310acee1b@xxxxxxxxxx  [2]
>>> Acked-by: Vlastimil Babka (SUSE) <vbabka@xxxxxxxxxx>
>>> Signed-off-by: Harry Yoo (Oracle) <harry@xxxxxxxxxx>
>>> Link: https://lore.kernel.org/linux-mm/872bd673-3d45-4111-8a41-31185db3ece5@xxxxxxxxxx
>>> Signed-off-by: Harry Yoo (Oracle) <harry@xxxxxxxxxx>
>>> Link: https://lore.kernel.org/linux-mm/6811cc17-8ee4-48c8-8cbf-6bf4d9f98162@xxxxxxxxxx [1]
>>> Link: https://lore.kernel.org/linux-mm/40591888-3a87-433e-b3d2-cda1cab543be@xxxxxxxxxx [2]
>>> Suggested-by: Vlastimil Babka (SUSE) <vbabka@xxxxxxxxxx>
>>> Signed-off-by: Harry Yoo (Oracle) <harry@xxxxxxxxxx>
>>> Signed-off-by: Harry Yoo (Oracle) <harry@xxxxxxxxxx>
>>> Signed-off-by: Harry Yoo (Oracle) <harry@xxxxxxxxxx>
>> 
>> Wonder what produced this weird block of assorted trailers? "b4 prep
>> --compare-to vX" never did that for me?
> 
> Interesting. I checked it again with `b4 prep --compare-to v4`, but the
> command didn't produce this. Looking at the cover letter
> `b4 prep --edit-cover` It's not even there.
> 
> Probably it was produced during b4 send --no-sign? hmm...

Hmm let's see if it ever happens again.

>>> ---
>>> Harry Yoo (Oracle) (8):
>>>       mm/slab, slub_kunit: register kprobe to trigger _nolock APIs
>>>       mm/slab: handle the !allow_spin case in kfree_rcu_sheaf()
>>>       mm/slab: use call_rcu() in unknown context if irqs are enabled
>>>       mm/slab: extend deferred free mechanism to handle rcu sheaves
>>>       mm/slab: allow kfree_rcu_sheaf() on PREEMPT_RT
>>>       mm/slab: introduce struct kvfree_rcu_head for kvfree_rcu batching
>>>       mm/slab: introduce kfree_rcu_nolock()
>>>       slub_kunit: extend the test for kfree_rcu_nolock()
>>>
>>>  include/linux/rcupdate.h   |  32 ++++++-
>>>  include/linux/slab.h       |  16 +---
>>>  include/linux/types.h      |   8 ++
>>>  include/trace/events/rcu.h |   2 +-
>>>  lib/tests/slub_kunit.c     | 190 ++++++++++++++++++++++++++++-----------
>>>  mm/slab.h                  |  52 ++++++++++-
>>>  mm/slab_common.c           |  81 +++++++++++++----
>>>  mm/slub.c                  | 220 ++++++++++++++++++++++++++++-----------------
>>>  8 files changed, 433 insertions(+), 168 deletions(-)
>> 
>> Applied to slab/for-7.3/kfree_rcu_nolock and merged go slab/for-next, thanks!
> 
> Thanks a lot!
> 
>> Used slab/for-next-fixes as a base for a better topic branch indepdencence.
> 
> Ah, I see. thanks.
> 
> So slab/for-next is not a good base unless it depends too much on
> existing commits on the branch?
> 
> And usually it's better to use the base (slab/for-next-fixes,
> in this case) of slab/for-next as a base?

Yeah. Hopefully we'll figure out some documented guidelines for this in the
upcoming mm-next meetings.

>>> ---
>>> base-commit: ffc64805cebac330b2a60b2f1f144fcc35b8ab46
>>> change-id: 20260615-kfree_rcu_nolock-e5502555992f
>>>
>>> Best regards,
>