git format-patch series helper with changelog trail and checkpatch support
Project description
gfs – Git Format-patch Series helper
A CLI tool built on top of git format-patch that simplifies the Linux
kernel upstream patch workflow. Instead of manually juggling format-patch
flags, maintainer lists, and changelog trails across versions, gfs
handles it all in a single command:
- Versioned patch series — generate v1, v2, v3, … with one command.
- Automatic maintainer CCs — runs
get_maintainer.pland adds the right CCs for you. - Changelog trail injection — automatically carries forward the changelog history across versions, so reviewers see what changed.
- Checkpatch integration — run
checkpatch.plon your series before sending.
Requirements
- Python 3.10+
- Git
- All commands must be run from the root of a Linux kernel tree
(the tool needs
scripts/get_maintainer.plandscripts/checkpatch.pl).
Installation
Install from PyPI:
pip install gfs-tool
Or from source (from the gfs directory):
pip install .
Or in editable/development mode:
pip install -e .
After installation, the gfs command is available system-wide:
gfs --help
Quick start
# All gfs commands must be run from the root of a Linux kernel tree
cd /path/to/linux
# 1. Create v1 of a new series
gfs init -c <sha> -n 3 --prefix "PATCH" -t for-pm-upstream --to user@example.com --cc user@example.com
# OR: Initialize from existing patches (any version)
gfs sync my-topic/v2
# 2. Run checkpatch
gfs check -t for-pm-upstream
# 3. After review → generate v2
gfs -v 2 -c <sha> -n 3 --prefix "PATCH v2" -t for-pm-upstream
# 4. Fill in changelog notes manually, then checkpatch again
gfs check -t for-pm-upstream
Commands
gfs init — Initialise a new patch series (v1)
Creates the first version of a patch series and saves --to / --cc
in <topic>/.series.json so you don't have to repeat them.
gfs init -c <commit-sha> -n <num-patches> --prefix <prefix> -t <topic> [--to <email>] [--cc <email>] [--no-cc]
| Flag | Required | Description |
|---|---|---|
-c, --commit |
✅ | Base commit SHA for git format-patch |
-n, --num-patches |
✅ | Number of patches to generate |
-p, --prefix |
✅ | Subject prefix, e.g. "PATCH", "PATCH v2" |
-t, --topic |
✅ | Topic output directory, e.g. for-pm-upstream |
--to |
To: email address |
|
--cc |
Cc: email address |
|
--no-cc |
Skip get_maintainer.pl pass (single format-patch run) |
Example:
gfs init -c abc1234 -n 3 --prefix "PATCH" -t for-topic --to user@example.com --cc user@example.com
Output structure:
for-topic/
├── .series.json ← saved to/cc config
└── v1/
├── 0000-cover-letter.patch
├── 0001-subsys-Add-foo-support.patch
├── 0002-subsys-Add-bar-support.patch
└── 0003-subsys-Add-baz-support.patch
gfs -v N — Generate a new version (v2, v3, …)
Generates the next version of the patch series. Automatically injects changelog trail headers into every patch.
gfs -v <version> -c <commit-sha> -n <num-patches> --prefix <prefix> -t <topic> [--to <email>] [--cc <email>] [--no-cc]
| Flag | Required | Description |
|---|---|---|
-v, --version |
✅ | Version number (2, 3, 4, …) |
-c, --commit |
✅ | Base commit SHA |
-n, --num-patches |
✅ | Number of patches |
-p, --prefix |
✅ | Subject prefix, e.g. "PATCH v2" |
-t, --topic |
✅ | Topic directory |
--to |
Override saved To: address |
|
--cc |
Override saved Cc: address |
|
--no-cc |
Skip get_maintainer.pl pass (single format-patch run) |
Example:
gfs -v 2 -c abc1234 -n 3 --prefix "PATCH v2" -t for-topic
gfs check — Run checkpatch.pl
Runs ./scripts/checkpatch.pl --strict --codespell on all patches
in the given topic directory.
gfs check -t <topic> [-v <version>]
| Flag | Required | Description |
|---|---|---|
-t, --topic |
✅ | Topic directory |
-v, --version |
Version to check (default: latest) |
Examples:
# Check latest version
gfs check -t for-topic
# Check a specific version
gfs check -t for-topic -v 3
gfs sync — Initialize project from existing patches
Scans an existing patch directory and automatically creates the .series.json
configuration file by extracting metadata from the patches (To/Cc addresses,
prefix, patch count). Useful for taking over maintenance of an existing series
or working with patches received from others.
gfs sync [path]
| Argument | Required | Description |
|---|---|---|
path |
Path to topic directory or version subdirectory (default: current directory) |
How it works:
- Scans the directory structure to find topic name and version number
- Counts patches in the version directory (excluding cover letter)
- Extracts metadata from patches:
- Subject prefix from the
[PREFIX N/M]pattern To:addresses from all patchesCc:addresses from all patches
- Subject prefix from the
- Creates
.series.jsonwith the extracted configuration - Displays suggested next command to generate the next version
Examples:
# Sync from a specific version directory
gfs sync my-topic/v2
# Sync from topic directory (uses latest version)
gfs sync my-topic
# Sync from current directory
gfs sync .
cd my-topic && gfs sync
Expected directory structure:
my-topic/
├── v1/
│ ├── 0000-cover-letter.patch
│ ├── 0001-first-patch.patch
│ └── 0002-second-patch.patch
└── v2/
├── 0000-cover-letter.patch
├── 0001-first-patch.patch
└── 0002-second-patch.patch
Output:
After running gfs sync my-topic/v2, you'll see:
── gfs sync ──
📂 Topic: my-topic
📌 Version: v2
📝 Patches: 2
🏷️ Prefix: PATCH v2
📧 To: maintainer@example.com
📧 Cc: reviewer@example.com, list@vger.kernel.org
✅ Project initialized from existing patches.
Next steps:
• Edit my-topic/.series.json if needed
• Run: gfs -v 3 -c <sha> -n 2 -p 'PATCH v2 v3' -t my-topic
When to use:
- Taking over maintenance of a patch series from someone else
- Recovering from lost configuration files
- Working with patch series created by other tools
- Initializing
gfsworkflow for pre-existing patch directories
How it works
Two-pass git format-patch
Every init and -v N invocation runs git format-patch twice:
-
Pass 1 — generates the patch files:
git format-patch <sha> -N --to=<to> --cc=<cc> --subject-prefix="PATCH vN" --thread --cover-letter -o <topic>/vN/ -
Pass 2 — re-generates adding maintainers in CC:
git format-patch <sha> -N --to=<to> --cc=<cc> --subject-prefix="PATCH vN" --thread --cover-letter -o <topic>/vN/ --cc="$(scripts/get_maintainer.pl --no-rolestats --separator=, <topic>/vN/000*)"The second pass needs the files from the first pass to exist so that
get_maintainer.plcan parse them and determine the correct maintainers/reviewers.
Changelog trail injection
For version ≥ 2, gfs automatically injects changelog headers between
--- and the diffstat in every patch file. It reads the trail from the
previous version and prepends a new empty header.
You then fill in the details manually.
v2 patches will contain:
Signed-off-by: You <you@example.com>
---
v1->v2:
-
drivers/subsys/foo.c | 131 +++...
v3 patches will contain (after you filled v2):
Signed-off-by: You <you@example.com>
---
v2->v3:
-
v1->v2:
- Refactored initialization logic.
drivers/subsys/foo.c | 118 +++...
v4 patches will contain:
Signed-off-by: You <you@example.com>
---
v3->v4:
-
v2->v3:
- Simplified error handling.
v1->v2:
- Refactored initialization logic.
drivers/subsys/foo.c | 117 +++...
Per-topic configuration
Each topic directory stores a .series.json with saved to and cc
addresses, so you don't need to repeat them every time:
{
"to": "user@example.com",
"cc": "user@example.com"
}
This also means you can work on multiple series in parallel, each with its own independent config.
Typical workflow
# All gfs commands must be run from the root of a Linux kernel tree
cd /path/to/linux
# ── Start a new series ───────────────────────────────────────────
gfs init -c abc123 -n 3 --prefix "PATCH" -t for-pm-upstream --to user@example.com --cc user@example.com
# ── Check patches ────────────────────────────────────────────────
gfs check -t for-pm-upstream
# ── Send v1 with git send-email ──────────────────────────────────
git send-email for-pm-upstream/v1/*.patch
# ── After review: amend commits, then generate v2 ────────────────
gfs -v 2 -c abc123 -n 3 --prefix "PATCH v2" -t for-pm-upstream
# ── Fill in changelog notes in each patch file ───────────────────
vim for-pm-upstream/v2/0001-*.patch # add notes under "v1->v2:"
# ── Check & send v2 ─────────────────────────────────────────────
gfs check -t for-pm-upstream
git send-email for-pm-upstream/v2/*.patch
# ── v3, v4, … repeat ────────────────────────────────────────────
gfs -v 3 -c abc123 -n 3 --prefix "PATCH v3" -t for-pm-upstream
Directory layout
After several iterations your topic directory will look like:
for-topic/
├── .series.json
├── v1/
│ ├── 0000-cover-letter.patch
│ ├── 0001-subsys-Add-foo-support.patch
│ ├── 0002-subsys-Add-bar-support.patch
│ └── 0003-subsys-Add-baz-support.patch
├── v2/
│ ├── 0000-cover-letter.patch
│ ├── ...
├── v3/
│ ├── ...
└── v4/
└── ...
Example session
A complete walk-through from v1 to v3 of a 3-patch series.
1. Initialise the series (v1)
$ cd /path/to/linux
$ gfs init -c abc1234 -n 3 --prefix "PATCH" -t for-topic --to user@example.com --cc user@example.com
✓ config saved → for-topic/.series.json
── Pass 1: generate patches ──
✓ for-topic/v1/0000-cover-letter.patch
✓ for-topic/v1/0001-subsys-Add-foo-support.patch
✓ for-topic/v1/0002-subsys-Add-bar-support.patch
✓ for-topic/v1/0003-subsys-Add-baz-support.patch
── Pass 2: adding get_maintainer.pl cc ──
✓ for-topic/v1/0000-cover-letter.patch
✓ for-topic/v1/0001-subsys-Add-foo-support.patch
✓ for-topic/v1/0002-subsys-Add-bar-support.patch
✓ for-topic/v1/0003-subsys-Add-baz-support.patch
2. Check with checkpatch and send v1
$ gfs check -t for-topic
── checkpatch.pl on for-topic/v1 ──
▸ ./scripts/checkpatch.pl --strict --codespell for-topic/v1/*.patch
$ git send-email for-topic/v1/*.patch
3. After review — generate v2
Amend commits based on review feedback, then:
$ gfs -v 2 -c abc1234 -n 3 --prefix "PATCH v2" -t for-topic
── Pass 1: generate patches ──
...
── Pass 2: adding get_maintainer.pl cc ──
...
📝 Adding changelog trail (v1->v2):
✓ trail injected → 0000-cover-letter.patch
✓ trail injected → 0001-subsys-Add-foo-support.patch
✓ trail injected → 0002-subsys-Add-bar-support.patch
✓ trail injected → 0003-subsys-Add-baz-support.patch
Each patch and the cover letter now contain:
v1->v2:
-
4. Fill in the changelog notes
Edit each patch and write what changed:
$ vim for-topic/v2/0001-*.patch
Replace the empty - with actual notes:
v1->v2:
- Refactored initialization logic.
5. Check and send v2
$ gfs check -t for-topic
$ git send-email for-topic/v2/*.patch
6. After another review — generate v3
$ gfs -v 3 -c abc1234 -n 3 --prefix "PATCH v3" -t for-topic
The changelog trail now accumulates automatically. Each patch contains:
Signed-off-by: You <you@example.com>
---
v2->v3:
-
v1->v2:
- Refactored initialization logic.
drivers/subsys/foo.c | 118 +++...
And the cover letter contains the same trail before the shortlog:
v2->v3:
-
v1->v2:
- Refactored initialization logic.
User Name (3):
subsys: Add foo support
...
Fill in v2->v3:, run checkpatch, send — and repeat for v4, v5, …
Sending the series
Once your patches are ready, send them with git send-email:
git send-email --cc linux-kernel@vger.kernel.org ./for-topic/v1/*.patch
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file gfs_tool-0.0.4.tar.gz.
File metadata
- Download URL: gfs_tool-0.0.4.tar.gz
- Upload date:
- Size: 17.2 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.12.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
a5d76699bc8aae7e4bccef979e9413ab9d3fd02bd015f4972e1cd2a611dd4f74
|
|
| MD5 |
f764f842d77d8182b270a73fa2376510
|
|
| BLAKE2b-256 |
396b6c32c040dbc3b6705e63ed98e5fb489144bcfe755f9320706203a80f2e25
|
File details
Details for the file gfs_tool-0.0.4-py3-none-any.whl.
File metadata
- Download URL: gfs_tool-0.0.4-py3-none-any.whl
- Upload date:
- Size: 14.2 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.12.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
2c00e758563bcab98c86748a427af73e2fa0ec629424949d8e01ae95dc1ed210
|
|
| MD5 |
28eaa32ae7013c5a5b3bbf78bbe06ac8
|
|
| BLAKE2b-256 |
baea7d3ce88f42800f4ad7e49e0c35cc410d49c56ba2e56d49a28e3caff0c496
|