Skip to content

linux: remove two patch files that are not in the series - #328

Merged
daniel-rossier merged 1 commit into
mainfrom
cleanup/dead-linux-patches
Sep 8, 2026
Merged

linux: remove two patch files that are not in the series#328
daniel-rossier merged 1 commit into
mainfrom
cleanup/dead-linux-patches

Conversation

@daniel-rossier

Copy link
Copy Markdown
Contributor

build/meta-linux/recipes-linux/linux/files/0001-linux-6.12-r0/ holds eight .patch files, but 0001-linux-6.12-r0-patches.inc references only six. The two extras are never applied by bitbake — they just sit there looking like part of the boot path, which is exactly how they cost me a debugging detour: reading the guest DT, the obvious file to trust is the higher-numbered one.

Neither carries anything the live series does not:

file verdict
0022-virt64.dts.patch body byte-identical to the live 0004. A pure duplicate.
0023-virt64_guest.dts.patch identical to the live 0002 except the CMA size0x6000000 (96 MB) where 0002 has 0xa000000 (160 MB). A stale revision from before that bump.

Both carry --- a/<path> headers where their live counterparts create the file from /dev/null, which points at an updiff run that regenerated the set without replacing it. (0004's own header still names an absolute path inside an infrabase build tree, from the same era — not fixed here.)

No functional change

The .inc is untouched, so exactly the same six patches are applied before and after. Verified that nothing else in the tree references either filename, and that the rpi4 series is clean (17 files, 17 entries).

Companion to #327, which fixes the content of the live guest DT; this one only removes the misleading dead copies.

🤖 Generated with Claude Code

https://claude.ai/code/session_01LQQDPHFAGVU67bEXrdboRS

build/meta-linux/recipes-linux/linux/files/0001-linux-6.12-r0/ holds
eight .patch files, but 0001-linux-6.12-r0-patches.inc references only
six. The two extras are never applied by bitbake — they just sit there
looking like part of the boot path, which is how they cost me a debugging
detour: reading the guest DT, the obvious file to trust is the
higher-numbered one.

Neither carries anything the live series does not:

  0022-virt64.dts.patch        body byte-identical to the live 0004.
                               A pure duplicate.
  0023-virt64_guest.dts.patch  identical to the live 0002 except for the
                               CMA size — 0x6000000 (96 MB) where 0002
                               has 0xa000000 (160 MB). A stale revision
                               from before that bump.

Both carry `--- a/<path>` headers where their live counterparts create
the file from /dev/null, which points at an updiff run that regenerated
the set without replacing it. (0004's own header still names an absolute
path inside an infrabase build tree, from the same era.)

No functional change: the .inc is untouched, so exactly the same six
patches are applied before and after. Verified that nothing else in the
tree references either filename, and that the rpi4 series is clean
(17 files, 17 entries).
@daniel-rossier
daniel-rossier merged commit e696580 into main Sep 8, 2026
6 checks passed
@daniel-rossier
daniel-rossier deleted the cleanup/dead-linux-patches branch September 8, 2026 15:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant