Re: [PATCH] kernel-doc: fix struct_group_tagged() parsing

From: Alexander Lobakin
Date: Thu Apr 25 2024 - 05:13:38 EST


From: Jonathan Corbet <corbet@xxxxxxx>
Date: Wed, 24 Apr 2024 13:31:14 -0600

> Alexander Lobakin <aleksander.lobakin@xxxxxxxxx> writes:
>
>> From: Kees Cook <keescook@xxxxxxxxxxxx>
>>
>> kernel-doc emits a warning on struct_group_tagged() if you describe your
>> struct group member:
>>
>> include/net/libeth/rx.h:69: warning: Excess struct member 'fp' description in 'libeth_fq'
>>
>> The code:
>>
>> /**
>> * struct libeth_fq - structure representing a buffer queue
>> * @fp: hotpath part of the structure
>> * @pp: &page_pool for buffer management
>> [...]
>> */
>> struct libeth_fq {
>> struct_group_tagged(libeth_fq_fp, fp,
>> struct page_pool *pp;
>> [...]
>> );
>>
>> When a struct_group_tagged() is encountered, we need to build a
>> `struct TAG NAME;` from it, so that it will be treated as a valid
>> embedded struct.
>> Decouple the regex and do the replacement there. As far as I can see,
>> this doesn't produce any new warnings on the current mainline tree.
>>
>> Reported-by: Jakub Kicinski <kuba@xxxxxxxxxx>
>> Closes: https://lore.kernel.org/netdev/20240405212513.0d189968@xxxxxxxxxx
>> Fixes: 50d7bd38c3aa ("stddef: Introduce struct_group() helper macro")
>> Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>
>> Co-developed-by: Alexander Lobakin <aleksander.lobakin@xxxxxxxxx>
>> Signed-off-by: Alexander Lobakin <aleksander.lobakin@xxxxxxxxx>
>> ---
>> scripts/kernel-doc | 3 ++-
>> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> So, in docs-next, applying this *adds* two warnings:
>
>> ./include/net/page_pool/types.h:77: warning: Function parameter or struct member 'fast' not described in 'page_pool_params'
>> ./include/net/page_pool/types.h:77: warning: Function parameter or struct member 'slow' not described in 'page_pool_params'
>
> In truth, the warnings look correct. I guess I'll leave this applied,
> but perhaps a fix for the warnings should go into the net tree?

Sure, we'll fix this. Thanks!

>
> Thanks,
>
> jon

Olek