[REGRESSION] Unexpected section title false positive warnings on DOC: directive

[Bagas Sanjaya]

Hi,

Building htmldocs on docs-next currenly produces about 50 new warnings; which
all of them are unexpected section title on DOC: kernel-doc directive, like:

/home/bagas/repo/linux-kernel/Documentation/driver-api/target:25: ./drivers/target/target_core_user.c:35: CRITICAL: Unexpected section title.

Userspace I/O
------------- [docutils]
WARNING: kernel-doc 'scripts/kernel-doc.py -rst -enable-lineno -function 'Userspace I/O' ./drivers/target/target_core_user.c' processing failed with: SystemMessage('/home/bagas/repo/linux-kernel/Documentation/driver-api/target:25: ./drivers/target/target_core_user.c:35: (SEVERE/4) Unexpected section title.\n\nUserspace I/O\n-------------')
/home/bagas/repo/linux-kernel/Documentation/driver-api/target:28: ./include/uapi/linux/target_core_user.h:14: CRITICAL: Unexpected section title.

Ring Design
----------- [docutils]
WARNING: kernel-doc 'scripts/kernel-doc.py -rst -enable-lineno -function 'Ring Design' ./include/uapi/linux/target_core_user.h' processing failed with: SystemMessage('/home/bagas/repo/linux-kernel/Documentation/driver-api/target:28: ./include/uapi/linux/target_core_user.h:14: (SEVERE/4) Unexpected section title.\n\nRing Design\n-----------')

These turns out to be false-positive as touching the source file in question
(e.g. drivers/target/target_core_user.c) and making htmldocs again makes the
regression go away for the corresponding docs.

Bisection (with git-bisect(1)) points to bea467aa5da1f5 ("docs: media:
v4l2-ioctl.h: document two global variables") as the first bad commit, even
though that I suspect that bdd1cf87847ff6 ("kernel-doc: add support to handle
DEFINE_ variables") may be the actual culprit (regressor).

The full bisection log:

git bisect start
# status: waiting for both good and bad commits
# bad: [7f3c3a0a9103dc92c823f27db3284ac2914e7558] MAINTAINERS: Add doc files on real-time support to Real-time Linux
git bisect bad 7f3c3a0a9103dc92c823f27db3284ac2914e7558
# status: waiting for good commit(s), bad commit known
# good: [9448598b22c50c8a5bb77a9103e2d49f134c9578] Linux 6.19-rc2
git bisect good 9448598b22c50c8a5bb77a9103e2d49f134c9578
# bad: [5188f6bd408f937d81c0c37eb59ddc1035cd912c] docs: admin: devices: /dev/sr<N> for SCSI CD-ROM
git bisect bad 5188f6bd408f937d81c0c37eb59ddc1035cd912c
# bad: [bea467aa5da1f51834501da3ac3c40204027a221] docs: media: v4l2-ioctl.h: document two global variables
git bisect bad bea467aa5da1f51834501da3ac3c40204027a221
# good: [82e87387f6e2af9f69a7528733e953fd22e815aa] Documentation: kernel-hacking: Remove comma
git bisect good 82e87387f6e2af9f69a7528733e953fd22e815aa
# good: [bdd1cf87847ff6aaadd53a185209d2bb2db72165] kernel-doc: add support to handle DEFINE_ variables
git bisect good bdd1cf87847ff6aaadd53a185209d2bb2db72165
# first bad commit: [bea467aa5da1f51834501da3ac3c40204027a221] docs: media: v4l2-ioctl.h: document two global variables

Mauro: Since you're the author of regressor, can you please take a look on it?

Thanks.

#regzbot introduced: bea467aa5da1f5

-- 
An old man doll... just what I always wanted! - Clara

Attachment: signature.asc
Description: PGP signature