This guide assumes that you have:
- Set up your environment to build and run OpenSwitch, and
- Made changes to source code that you want to contribute to OpenSwitch
For individual contributions, please sign the DCO For corporate engagements, please send an email to cla_admin@lists.openswitch.net
- Preparing to contribute changes
- Contributing changes
- Adding a new component
- Adding a new feature
- Documenting the code
- Commit messages
- OpenSwitch Coding Style
Before you can contribute your changes, you will need to prepare your development environment.
OpenSwitch runs Gerrit for online code reviews and Git as the distributed version control system. The OpenSwitch Gerrit site is the entry point for change, review, test, and inclusion in the OpenSwitch project.
Create SSH keys only if no keys have been created earlier for the designated development machine. To find out if keys exist for the designated development machine, check the contents of the ~/.ssh/
directory. The following two files exist if keys were already created:
id_rsa
id_rsa.pub
If the ~/.ssh
directory or the files do not exist, then generate the SSH keys with the following commands:
mkdir ~/.ssh
chmod 700 ~/.ssh
ssh-keygen -t rsa
You are prompted for both a location to save the keys, and for a pass-phrase for the keys. Accept the default location to save the keys and enterResubmitting a set of changes a pass-phrase. The pass-phrase is important as it keeps your keys secure while stored on your hard drive. Remember the pass-phrase as it will be needed to work with Gerrit. An example output follows:
Generating public/private rsa key pair.
Enter file in which to save the key (/home/b/.ssh/id_rsa):
Enter pass-phrase (empty for no pass-phrase):
Enter same pass-phrase again:
Your identification has been saved in /home/b/.ssh/id_rsa.
Your public key has been saved in /home/b/.ssh/id_rsa.pub.
The public key needed to work with Gerrit is saved at ~/.ssh/id_rsa.pub
. The contents of this file (this is a text file) is needed in the next step.
If you get the error message, Agent admitted failure to sign using the key.
as seen below, simply run the command ssh-add
after the message.
"Agent admitted failure to sign using the key."
# debug1: No more authentication methods to try.
# Permission denied (publickey).
- Log in to the OpenSwitch Review site by clicking on the top right
Sign in
link.- OpenSwitch uses a
GitHub
account to login to the Review Site (Gerrit). If you do not have an account, create one by clickingSign up
in the GitHub page after you try to login to the OpenSwitch Review. - Keep a record of the password and username as this is your Gerrit User information and is needed to work with OpenSwitch.
- OpenSwitch uses a
- Click
Settings
(in the menu under the user's name in the upper right-hand corner). - Select
SSH Public Keys
in the panel on the left. - Paste the SSH public key (
id_rsa.pub
) contents in the input area to use with this site. - Click
Add
.
- Log in to the OpenSwitch Review site by clicking on the top right
Sign in
link. - Click
Settings
in the menu under the user's name in the upper right-hand corner. - Click
Profile
on the left side. - Verify that a username exists, or define a username if one does not exist.
- Keep track of the username as this is the Gerrit User information and is used to work with OpenSwitch.
Add an entry similar to the following in your ~/.ssh/config
file (preserve the identification), which directs SSH to the proper SSH private key file to use for the OpenSwitch review website:
Host review.openswitch.net
User <gerrit-user>
IdentityFile ~/.ssh/id_rsa
The network port used by Gerrit (29418) on the review website may be blocked in some networks, for example on a corporate network. One workaround (besides asking your ISP or corporate IT team to open the port) is to tunnel in using your HTTP proxy. You can achieve that with the nc
or socat
command line utilities, or any other utilities of your preference. Follow the instructions below to use socat
or nc
to configure a proxy for the SSH traffic.
Take note of the proxy information.
- Proxy URL such as
web-proxy.location.domain.com
. The proxy URL is referred to as<proxy-url>
below. - Proxy port such as 8080. The proxy port is referred to as
<proxy-port>
below.
Using nc
- Install
nc
if not yet installed - Open (or create) the
~/.ssh/config
file and add the following lines (preserve the indentation):
Host review.openswitch.net
ProxyCommand nc -x <proxy-url>:<proxy-port> -X connect %h %p
Using socat
- Install
socat
if not yet installed - Open (or create) the
~/.ssh/config
file and add the following lines (preserve the indentation):
Host review.openswitch.net
ProxyCommand socat - PROXY:<proxy-url>:%h:%p,proxyport=<proxy-port>
For more details and examples for adding usernames or passwords, see this page.
You need to configure your local repository to use your Gerrit account. To configure your repository:
- Set
gitreview.username
with your Gerrit login.
$ git config --global gitreview.username <gerrit-user>
- Configure your email and your username in Git:
$ git config --global user.email <email>
$ git config --global user.name <name>
To contribute to OpenSwitch, every member must sign a DCO. Follow the simple steps below to sign the DCO agreement:
- Log in to the OpenSwitch Reveiw.
- At the top right, click "Settings" under your name.
- In the menu on the left, click Agreements and then click on
New Contributor Agreement
. - Select DCO and after carefully reviewing the content, sign at the bottom.
git-review
is used to contribute changes back to the OpenSwitch project.
- Install
git-review
as described here, if not already installed on the development machine. - Set up the git-review.
$ git-review -s
Now that you prepared your development environment to contribute changes, you can follow the following steps to actually make the changes.
Changes to the OpenSwitch code base go through a review process before being merged. This section describes how local code changes can be submitted for review and, if approved, made available upstream in the OpenSwitch project.
This section assumes that the changes are contained to adding new functionality to an existing feature, adding new capability to an existing component, or fixing a defect in existing code base. More elaborate changes that require adding a new component or a feature are addressed in subsequent sections.
Each change-set submitted to review should include, along with the changes to source code, corresponding additions or changes to test scripts and documentation. Without these, your change-set may not be approved by the reviewers.
Before attempting to commit changes, make sure that they are compliant with the OpenSwitch Coding Style. For non-OpenSwitch modules, follow the coding style that the module already uses.
In particular, the following file types are rejected:
- Files with trailing spaces
- Files with non-printable ASCII characters in their names.
Changes for review must be committed to a local branch with a commit message that:
- Begins with a single line of text which summarizes the contents of the change.
- Additionally provides a description following the summarization line.
If the change set has been composed by multiple commits to the local branch, consider squashing them into a single commit using git rebase
.
git-review
is used as an aid when submitting a change set for review.
The commit message for the set of changes must include a Change-Id
, as generated by the -i
option to git-review
. To send changes for review, commit the changes and then run git-review as follows:
$ git commit -s -m "Meaningful summary of this change set."
$ git-review -i
The git-review
command output includes an URL for the change review. In example below the URL generated was: https://review.openswitch.net/1148.
$ git review
remote: Resolving deltas: 100% (1/1)
remote: Processing changes: new: 1, refs: 1, done
remote: (W) 5421c73: commit subject >65 characters; use shorter first paragraph
remote:
remote: New Changes:
remote: https://review.openswitch.net/1148
Use the URL in the git-review
output to login to the Review Site and add reviewers as needed.
The reviewers with +2 permissions are automatically added for reviews on the master and release branches.
If you are submitting changes to multiple repositories and the changes are required to merge in a specific order, create dependencies among them. To create dependency among changes across different repositories use the "Depends-on" tag in the commit message.
While submitting the dependent patch for review, include "Depends-On: " in the footer of your commit message. The gerrit-change-id should be the entire text starting with the "I" that appears in Change-Id of the review this patch depends on.
Multiple cross-repository dependencies can be created but do not create a cycle.
Reviewers may request modifications to the set of changes previously submitted. Make the changes as requested by the reviewers and then submit a patch with the following commands. The -i
switch for git-review
is not needed as a Change-Id
was previously generated by the initial commit.
$ git commit --amend
$ git-review
To cancel the changes, click Abandon Change in the web interface from Gerrit.
- Log in to the change review URL using the link provided by the
git-review
command output. - Click on the
Review
button and give the change a+1 Approved
rate in theWorkflow
section. This initiates the process of merging your change with the main product.
Each component recipe file is pointing to a fixed SHA ID at which point the component was deemed stable. In each recipe file, this is indicated as follows:
SRCREV = "a358e75a97d9eb58ba89fe42473ac6ab8c481218"
This ensures that your local sandbox is guarded by other changes being committed since all the recipe files are fixed to a stable SHA across components. Once your review is tested, verified and merged to master, update the corresponding recipe file's SRCREV field with the recently merged code review commitID SHA. This is the “Commit-Id” in the “Commit Log” section of the code review.
Note: If SRCREV=${AUTOREV}
for a component's recipe file, above step is not required. It will always pick up the latest changes.
In OpenSwitch, a module is known as a component. Each component resides in a seperate repository. Each feature will have several components associated, depending in the size of the feature. This section shows you how you can add a new component to an existing feature or create your own component.
Adding a new component to OpenSwitch comes with responsibility to invest in ongoing maintenance of the component. To enable this, per the Governance document, identify individuals who are willing to take on the roles of Maintainer, Reviewer, and Bug Czar for the component.
Start by sending an email to the infra@lists.openswitch.net mailing list, with the Github IDs of the individuals identified for the above roles, and the name of the component/repository you would like to add. You will receive back a confirmation email including a new group, of the form ops-<repo-name>-maintainers
, created for the above individuals that you can use in the steps below.
Once these individuals have been identified and the maintainers group created, the website needs to be update with this information.
git clone https://git.openswitch.net/openswitch/ops-docs
Update the code-repositories.md file and send out for review.
Now you are ready to follow the following steps in creating a new component.
To add a new repository to the OpenSwitch project (referred to as <repo-name>
):
- Clone the project with the following command:
git clone https://review.openswitch.net/infra/project-config
-
Create a file
gerrit/acls/openswitch/ops-<repo-name>.config
with the following contents:Note: The code review and abandon group name should be
ops-<repo-name>-maintainers
that you received n the earlier step, as shown in this example below.[access "refs/heads/*"] abandon = group Change Owner abandon = group ops-<repo-name>-maintainers label-Code-Review = -2..+2 group ops-<repo-name>-maintainers label-Workflow = -1..+1 group Change Owner [submit] mergeContent = true action = rebase if necessary
-
Modify
gerrit/projects.yaml
to add the repository.- project: openswitch/<repo-name> description: <Repo Description>
-
Commit the changes with the following command:
git commit --signoff
-
Review the changes with the following command:
git review
-
The following people have code review +2 permission to approve the new repo creation:
- Contact Infra team infra@lists.openswitch.net to create the maintainers group
ops-<repo-name>-maintainers
in gerrit. In the email, specify the full name and email address for the users inGitHub
that you want to add as maintainers of this repository. - To update the website with information about the new repository, update the code-repositories.md file in the ops-docs repository:
git clone https://review.openswitch.net/openswitch/ops-docs
- Modify
code-repositories.md
to add the new maintainers group.
[openswitch/<repo-name>](http://git.openswitch.net/cgit/openswitch/<repo-name>/tree/) | <Repository Description> | <Module Maintainer> | [ops-<repo-name>-maintainers](https://review.openswitch.net/#/admin/groups/<##>,members) | <Bug czar> |
- Commit the changes with the following command:
git commit --signoff
- Review the changes with the following command:
git review
-
The following people have code review +2 permission to approve the request for updating code-repositories.md:
-
The exsiting members in the -maintainers group have permission to add more people in the same group.
Every repository is gated by at least two jenkins(CI) jobs. To create a set of basic jenkins(CI) jobs using yaml files
- Git clone project infra: git clone https://review.openswitch.net/infra/project-config
- Create a new file jenkins/jobs/ops-myrepo-jobs.yaml with following contents, change the content according to your needs (Note: replace ops-myrepo with your repo name):
- job-template:
name: 'ops-myrepo-check-{platform}'
node: openswitch
wrappers:
- build-timeout:
timeout: 120
- timestamps
- ansicolor
builders:
- revoke-sudo
- module-build-branch:
module: 'ops-myrepo'
platform: '{platform}'
- job-template:
name: 'ops-myrepo-gate-{platform}'
node: openswitch
wrappers:
- build-timeout:
timeout: 120
- timestamps
- ansicolor
builders:
- revoke-sudo
- module-build-branch:
module: 'ops-myrepo'
platform: '{platform}'
- job-group:
name: 'ops-myrepo-jobs'
jobs:
- 'ops-myrepo-check-{platform}'
- 'ops-myrepo-gate-{platform}'
- modify jenkins/jobs/projects.yaml to add the CI to Jenkins (Note: replace ops-myrepo with your repo name):
- project:
name: openswitch/ops-myrepo
node: 'openswitch'
platform:
- as5712
- genericx86-64
jobs:
- 'ops-myrepo-jobs'
- modify zuul/layout.yaml to add the CI to zuul (Note: replace ops-myrepo with your repo name):
- name: openswitch/ops-myrepo
template:
- name: merge-check
- name: module-build
module: ops-myrepo
platform: genericx86-64
- name: module-build
module: ops-myrepo
platform: as5712
- check in for review (Note: replace ops-myrepo with your repo name):
$ git add jenkins/jobs/ops-myrepo-jobs.yaml
$ git commit jenkins/jobs/ops-myrepo-jobs.yaml jenkins/jobs/projects.yaml zuul/layout.yaml -m "initial CI for new repo"
$ git review
- To later modify your repo to be gated with feature test cases. You will modify jenkins/jobs/ops-myrepo-jobs.yaml similar to http://git.openswitch.net/cgit/infra/project-config/tree/jenkins/jobs/ops-arpmgrd-jobs.yaml along with jenkins/jobs/projects.yaml and zuul/layout.yaml
For your component work properly, you must have a recipe. This recipe should be placed in the ops-build
repo, inside yocto/openswitch/meta-distro-openswitch
.
There are different directories for the recipes:
- recipes-asic
- recipes-core
- recipes-kernel
- recipes-networking
- recipes-onie
- recipes-ops
Choose the directory that applies and add the .bb
file to it.
The following is an example of a recipe:
SUMMARY = "Management Interface Configuration Daemon"
LICENSE = "Apache-2.0"
LIC_FILES_CHKSUM = "file://${COMMON_LICENSE_DIR}/Apache-2.0;md5=89aea4e17d99a7cacdbeed46a0096b10"
DEPENDS = "ops-utils ops-ovsdb"
RDEPENDS_${PN} = "python-argparse python-json python-ops-ovsdb python-distribute"
SRC_URI = "git://git.openswitch.net/openswitch/ops-mgmt-intf;protocol=http \
file://mgmt-intf.service \
"
SRCREV="${AUTOREV}"
# When using AUTOREV, we need to force the package version to the revision of git
# in order to avoid stale shared states.
PV = "git${SRCPV}"
S = "${WORKDIR}/git"
do_install_prepend() {
install -d ${D}${systemd_unitdir}/system
install -m 0644 ${WORKDIR}/mgmt-intf.service ${D}${systemd_unitdir}/system/
}
SYSTEMD_PACKAGES = "${PN}"
SYSTEMD_SERVICE_${PN} = "mgmt-intf.service"
inherit openswitch setuptools systemd
The recipes have some important parts:
- SUMMARY--Summary of your recipe.
- LICENSE--License for your recipe.
- LIC_FILES_CHKSUM--Path of the license file with the md5.
- DEPENDS--Label we have to include all the dependencies that our recipe has in order to compile, in this case, depending on
ops-utils
andops-ovsdb
- SRC_URI--List of source files, it can be local or remote, Yocto supports several protocols. You can find the complete list in the Yocto Ref Manual
- S--Location in the build directory where the recipe source code resides. It is the work directory.
- inherit--Variable used to inherit the functionality of a class, in our case we need
cmake
in order to build this recipe.
If you need more details of how the BitBake works, see BitBake documentation.
The new repo needs a CMake file to be able to index with the rest of the OpenSwitch project. The easiest way to create a CMake file is to take a similar repo, copy the CMake and adapt it for the new repo.
All executable that OpenSwitch repositories generate must have ops- prefix (except openvswitch). Similarly, all Python scripts/daemons names also must have ops- prefix.
The following is an example code taken from ops-vland.
# Copyright (C) 2015 Hewlett-Packard Development Company, L.P.
# All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
cmake_minimum_required (VERSION 2.8)
set (VLAND ops-vland)
project (${VLAND})
set (SRC_DIR src)
set (INCL_DIR include)
# Rules to locate needed libraries
include(FindPkgConfig)
pkg_check_modules(OVSCOMMON REQUIRED libovscommon)
pkg_check_modules(OVSDB REQUIRED libovsdb)
include_directories (${PROJECT_BINARY_DIR} ${PROJECT_SOURCE_DIR}/${INCL_DIR}
${OVSCOMMON_INCLUDE_DIRS}
)
# Define compile flags
set (CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -std=gnu99 -Wall -Werror")
# Source files to build ops-vland
set (SOURCES ${SRC_DIR}/vland.c ${SRC_DIR}/vland_ovsdb_if.c)
# Rules to build ops-vland
add_executable (${VLAND} ${SOURCES})
target_link_libraries (${VLAND} ${OVSCOMMON_LIBRARIES} ${OVSDB_LIBRARIES}
-lpthread -lrt)
# Rules to install ops-vland binary in rootfs
install(TARGETS ${VLAND}
RUNTIME DESTINATION bin)
For more details of how CMake works, please go to CMake documentation.
All components should follow the following structure for top-level files and directories:
Directory or File | Name | Purpose |
---|---|---|
File | AUTHORS | List of everyone who has contributed to this repository |
File | COPYING | Explanation of licenses applicable to this repository |
File | README.md | Overall explanation of the component residing in this repository |
File | DESIGN.md | Explanation of the design of internals of this component |
File | FAQ.md | (Optional) Collection of frequently asked questions regarding this component |
File | NEWS | (Optional) Should specify what changed from one release to another |
File | NOTICE | (Optional) Includes references for all the projects that were used for this repository |
Directory | /src | Source code |
Directory | /include | Header files |
Directory | /tests | Test scripts and test case documentation. Each [component]*.py test script should be accompanied by corresponding [component]*.md file that document the tests cases in that test script |
Directory | /docs | (Optional) Any other documentation that you prefer to add |
Now that you have created a new repository, and set up the recipe, CMake and top-level files and directories, follow the following last steps to complete the process of adding a new component.
- Add the code for your component by following the steps under Contributing changes.
- If this is a new daemon that runs when the switch boots up, add a service file in the same location as the recipe file. Set up makefiles/recipes, such that any executable that we generate has ops- prefix (except openvswitch, see below). Similarly, all Python scripts/daemons names must have ops- prefix.
- Add the necessary component tests and corresponding documentation as per Adding top-level files and directories. CIT will pick up these test cases and run them for every commit to this module.
yocto/openswitch/meta-distro-openswitch/recipes-core/packagegroups$ vim packagegroup-openswitch.bb
. - Commit your changes.
To add a new feature that does not belong to any of the existing code:
- Create the necessary module as mentioned under Adding a new component.
- Fetch any existing modules that will be modified as part of this feature.
- Additionally, add and integrate sufficient Feature test cases to CIT infra so that future commits are validated against the feature.
- Update the documentation on feature usage and design details, follow the Documenting the code section.
- Commit your changes.
The following table lists the type of documents, the target locations, and the expected file names. For a description of each type of document, see the following table:
| Doc Type | Repo | Directory | File Name | Purpose | |------------------------- |-------------------------- |----------- |------------------------- | | Functionality Guide | openswitch/ops | /docs | [functionality]_guide.md | Details on the functionality and its corresponding CLIs | | Feature Designs | openswitch/ops | /docs | [feature]_design.md | Details of how the various components come together to deliver feature functionality to end user | | Feature Test Plans | openswitch/ops | /tests | [feature]_test.md | Document each test case from corresponding [feature]_test.py test script |
A staging website is compiled every hour to include new and modified content across all the repositories. All content changes can be verified on the staging area before it makes it to http://openswitch.net. For any new document, since it is not already linked on the website, the webpage can be verified by manually entering the URL. For the above set of documents, the URL is in the following format:
http://egats.openswitch.net/documents/user/<document-name>
(NOTE: Document name is without .md) Once the content is verified on the staging area, content contributors must send an email to mailto:infra@lists.openswitch.net. In the email:
- Explain the content changes.
- For new documents requiring links, provide:
- The section to where you need to add the link.
- The text of the link.
- The location of the file to be linked. For example, "In Feature Design, add text 'BGP' and link it to http://git.openswitch.net/cgit/openswitch/ops/tree/docs/BGP_feature_design.md". Weekly the content is pushed from the staging area to the production website when the infrastructure team receives email in that week. If there are no changes, the weekly push from staging site to the production site is skipped.
The OpenSwitch project uses the Markdown markup language to generate its documentation. All documentation that is contributed to the project needs to be written in this format so that it can be built into the project. You can leverage the following tools to generate the markdown documentation.
- WYSIWYG-like markdown editor stackedit or install Atom
- Build ASCII-based block and flow diagrams to embed in markdown documents using asciiflow
Commit messages for change sets are recorded in the git repository when the changes are merged. Keeping with a simple, consistent format for these messages makes it easier for other users (and tools) to understand the changes.
The following excerpt is from the Git Pro, 2nd Edition book, chapter 5. We gratefully acknowledge the book's CC BY-NC-SA 3.0 license which permits replication without providing endorsement.
Getting in the habit of creating quality commit messages makes using and collaborating with Git a lot easier. As a general rule, your messages should start with a single line that’s no more than about 50 characters and that describes the changeset concisely, followed by a blank line, followed by a more detailed explanation. The Git project requires that the more detailed explanation include your motivation for the change and contrast its implementation with previous behavior – this is a good guideline to follow. It’s also a good idea to use the imperative present tense in these messages. In other words, use commands. Instead of “I added tests for” or “Adding tests for,” use “Add tests for.” Here is a template originally written by Tim Pope:
Short (50 chars or less) summary of changes
More detailed explanatory text, if necessary. Wrap it to
about 72 characters or so. In some contexts, the first
line is treated as the subject of an email and the rest of
the text as the body. The blank line separating the
summary from the body is critical (unless you omit the body
entirely); tools like rebase can get confused if you run
the two together.
Further paragraphs come after blank lines.
- Bullet points are okay, too
- Typically a hyphen or asterisk is used for the bullet,
preceded by a single space, with blank lines in
between, but conventions vary here
If all your commit messages look like this, things will be a lot easier for you and the developers you work with.
Note: there is no period ('.') at the end of the summary (first) line.
The commit messages are of great importance on a development process. They help the developers and final users to understand why changes were made. Therefore it is important to follow these guidelines. The first thing to take into account is that ALL our commit messages must have a subject and a body, which is explained in detail in the next sections.
The commit messages should be written impersonally. Use verbs like Add, Fix, Change, Update or Implement. Don't use Added or Adding, for example. Another important point is that each commit must change/add/fix only one thing. Keep in mind that each published commit, will live forever and sometimes when you go back to the project history for some issue, it could save you time due to a good hint in an informative commit message.
The first line of the commit message is known as the subject. It describes the change briefly and helps reviewers to see at a glance what the commit is about. It should not be longer than 50 characters.
Use the body of the commit message to describe your commit in detail and follow these guidelines:
- Separate the body from the subject with an empty line.
- Give an overview of why you're committing this change.
- Don't use bullet points.
- What the commit changes.
- Any new design choices made.
- Areas to focus on for recommendations or to verify correct implementation.
- Any research you might have done.
- Wrap the body of the message between 70 and 100 characters.
In order to enable automated tools to produce a meaningful changelog for OpenSwitch releases, the body of the commit message should include a line with tags associated with the change. The format of the line is like this:
Tags: tag, tag, ...
where "Tags" should be written verbatim and signals the start of the tags for the commit, followed by a colon (:) and a comma-separated list of tags.
The recognized tags fall into several categories described in the following sections.
An action classifies the type of changes made in the current commit and could be:
new
is used to say that new features were added to the codefix
is used to say that there was a bug fixedchg
is used to say that changes (refactor, small improvement, cosmetic changes) have been made to the code.
These tags describe the intended audience of the commit message could be:
dev
the commit message is intended to developersusr
the commit message is intended to the final user
These tags describe the type of change:
minor
is a minor change that you don't want to show in any changelogcosmetic
is cosmetic change like delete white spaces, indent the code or edit the comments; and you don't want this commit to appear in the any changelogrefactor
is refactor you don't want to appear in the any changelogwip
is work in progress you don't want to appear in any changelog
All the commit messages with the audience usr are parsed to generate the release notes. Be extra careful with those messages!.
The general format of a commit message should look like the following:
Short description of the work done in this commit
Explanation of the work done in this commit. This explanation
must explain the change in detail . Don't use bullet points.
Tags: tag, tag, ...
If the issue #1043 in Taiga has been fixed and you want it to be in the release notes, then the commit message should look like the following:
Now Action_set could increase size.
Pass write_to_gpt algorithm to write first on high GPT half and write to
hardware when low GPT half is written. gpt_location_idx_get is giving a
relative position to the action_set label in UCO.
Tags: fix, usr
TG-1043 #closed
Using "TG-1043" causes Taiga to associate the commit with the issue.
If the same bug has been fixed but you don't want it in the release notes then you'll have to change usr to dev.
Now Action_set could increase size.
Pass write_to_gpt algorithm to write first on high GPT half and write to
hardware when low GPT half is written. gpt_location_idx_get is giving a
relative position to the action_set label in UCO.
Tags: fix, dev
TG-1043 #closed
If more than one bug has been fixed with the same change, then the commit message should look like the following:
Now Action_set could increase size.
Pass write_to_gpt algorithm to write first on high GPT half and write to
hardware when low GPT half is written. gpt_location_idx_get is giving a
relative position to the action_set label in UCO.
Tags: fix, usr
TG-1043 #closed
TG-1045 #closed
TG-1056 #closed
If each bug requires a separate change then don't put all them together in the same commit. Write a separate commit for each bug.
You have made a small improvement on the pv_pm_neo_indirectLookUpEntry and you want to share it with you coworkers, but not with the final user (release notes). Your commit message should look like the following:
Change pv_pm_neo_indirectLookUpEntry SPs handling (no ucx support)
Move ind_lut entries for extension mode to entries 125 & 126
(now INDLUT_MODE_ENTRY on pm.h can be used only to change it)
Tags: chg, dev
After you did this commit you realize that the name of a non-important variable should be change. The message of the commit changing the name of the variable should look like the following:
Change pv_pm_neo_indirectLookUpEntry variable name
Change the name of the variable the variable old_name that stored the
a previous state to new_name.
Tags: chg, dev, minor
Perhaps you need to edit some comments in the code. Your commit message should look like the following:
Edit comments of pv_pm_neo_indirectLookUpEntry
Edit the comments of the pv_pm_neo_indirectLookUpEntry. Explain the usage
of some variables inside the function function_name.
Tags: chg, dev, cosmetic
You need to refactor the code and you don't want that commit to appear in any changelog. Your commit message should look like the following:
Refactor pv_pm_neo_indirectLookUpEntry
Change the way the loop works to save clock cycles.
Tags: chg, dev, refactor
All the commits with a tag like minor, cosmetic and refactor won't appear in any changelog.
Don't include many different changes into one commit. Each change must have a separate commit.
For Python source code, the coding style used for OpenSwitch is PEP8.
For C source code, the coding style used for OpenSwitch is an extension of the Open vSwitch Coding Style, which is an extension of the One True Brace Style for indenting.
Where the repository is based on an external open source project, all code must align the coding style of that project.
"ops" is the acronym adopted by OpenSwitch project. If any new file name or new function name prefix is required, "ops_" should be used. (NOTE: Existing external open source project file and function names should not be changed for easier upstreaming.)
Code must be heavily commented assuming someone completely new to the codebase.
Avoid adding personal references in the code or comments such as, "I", "me", "he", "we" etc.
Add TODO/FIXME comments as needed.
The additions below are clarifications and do not conflict with the Open vSwitch style.
Write comparisons in a form that reads naturally. Examples:
bool foo_b = some_bool_function(x);
int foo_i = some_int_function(x);
if (true == foo_b) /* No, unnecessary comparison and doesn't read well */
if (foo_b) /* Yes */
if (11 == foo_i) /* No, doesn't read well */
if (foo_i == 11) /* Yes */
if (foo_i == 0) /* Yes */
if (!foo_i) /* Yes */
Do not do a gratuitous initialization of variables. Doing this prevents the compiler from detecting accidental use before initialization.
struct Goofus *goofus = NULL; /* Wrong */
struct Gallant *gallant; /* Right */