KernelUpdates

Differences between revisions 10 and 11
Revision 10 as of 2009-08-03 11:44:03
Size: 3985
Editor: 89
Comment:
Revision 11 as of 2009-11-26 18:51:59
Size: 4502
Editor: p5B2E6F05
Comment:
Deletions are marked like this. Additions are marked like this.
Line 8: Line 8:
There are several categories of updates, in addition to normal security updates: Additionally to the generic requirements acceptable bugs fall into one of the following criteria:
Line 10: Line 10:
 * Critical bug fixes. These are categorized as non-security issues relating to bugs that affect a large range of users. These are bugs that keep users from reliably using their systems, or prevent booting at all. These patches must pass through rigorous testing by Canonical/Ubuntu and the community at large.
 * Supported vendor patches. These are patches generally related to hardware support. If they are very specific to a piece of hardware, the nature of the patch makes regression on other hardware unlikely or impossible, and we have the hardware available for thorough testing, then it can be part of an SRU. Otherwise they need to be maintained in a separate repository/pocket/PPA specific for that vendor or, if appropriate, in `linux-backports-modules`, and thus need to be maintained separately in the future. This also depends on whether a change relates to a LTS or a non-LTS release. While small changes to add drivers can be considered throughout the lifetime of an LTS, this will be limited to the point in time the next kernel hits beta status.

Other changes are generally avoided on stable kernels, since the regression potential is so exceptionally high.

All non-security changes need to follow the standard SRU procedure in terms of having a bug associated with them, which is fixed in the development release and signed off by `ubuntu-sru` or `canonical-qa`, and the changelog needs to include the bug number.
 1. It fixes a critical issue (data-loss, OOPs, crashes) or is security related (security related issues might be covered by security releases which are special in handling and publication).
 1. Simple, obvious and short fixes or hardware enablement patches as long as the next release has not reached beta status. If there is a related upstream stable tree open (see below) this class of patches is required to come through the upstream process. Patches sent upstream for that reason must include their ''!BugLink'' reference.
 1. For 3 months after a non-LTS release or the lifetime of a LTS release we will be pulling upstream stable updates from the corresponding series. There will be one tracking bug report for each stable update but additional references to existing bugs will be added to the contained patches (on best can do base). Uploads containing a upstream stable patchset increase the retention period to two weeks, while it is one week otherwise.
 1. Fixes to drivers which are not upstream are accepted directly if they fall into the first two categories.
 1. ''$DEITY'' intervention. Might happen, but very very rarely and will not be explainable.
Line 26: Line 25:
 * If the fix for a problem applies to the requirements for a SRU and has also been tested to successfully solved the bug, then the next step is to send the proposed patch for SRU review to the kernel-team mailing list, where it needs to receive at least two ACKs from core kernel developers. The review should include a generic header which describes the impact, and the fix, and how the issue can be tested. The same information should go into the description of the launchpad bug, so the stable release team can quickly find it.  * The patch or patches '''must''' contain the link to the Launchpad bug and contain a Signed-off-by line from the submitter.
 * The beginning of the description area of the bug needs to have a SRU justification which should look like this example: {{{
    SRU Justification:

    Impact: <a short description about the symptoms and the impact of the bug>
    Fix: <how was this fixed, where did the fix come from>
    Testcase: <how can the fix be tested>
}}}
 * If the fix for a problem applies to the requirements for a SRU and has also been tested to successfully solved the bug, then the next step depends on whether the fix is serious enough to be directly applied or whether it should go in via upstream stable (as long as that is appropriate).
 * For serious problem fixes, the patch must be sent the the kernel-team mailing list. There is requires ACKs from two senior kernel-team members and then will be applied to the kernel tree. Even though going into the tree on the faster path, the next step should be done. {{{
    To: kernel-team@lists.ubuntu.com
    Subject: [<series>] SRU: <bug/patch title>

    <SRU justification copied from the bug>

    <patch attached or pull request>
}}}
 * For all other patches (as long as the upstream stable is appropriate) the fix has to be sent upstream (when the problem is there as well and the patch is not a backport) and to stable@kernel.org (if it has not been sent there before). As soon as that is accepted there, it will come back its way when we pull stable updates.
Line 30: Line 46:
 * Urgent security updates will be uploaded directly into -security without other changes. This just requires a temporary GIT fork which will be immediately merged back into the main branch for that stable release.
 * Less urgent security updates and non-security patches will be uploaded to -proposed and then just follow the normal SRU QA procedure (testing, confirming bugs, etc). After verification, these kernels are copied verbatim to -security, and the USN is issued. This avoids maintaining two GIT trees for stable releases while still keeping the testing period.
 * Non-security updates which change the ABI should be either avoided at all, or be combined with security updates which require ABI change anyway.
 * Security updates will be uploaded directly into -security without other changes. This just requires a temporary GIT fork which will be immediately merged back into the main branch for that stable release.
 * Normal updates will be provided as pre-releases through the kernel-ppa users PPA. At certain points those get made into proposed releases which are uploaded to the proposed pocket. Then again they have to get verified to fix the problems and not to cause regressions.

Kernel security and update policy for post-release trees

This document describes the process and criteria for post-release kernel updates. The kernel is a very complex source package, and it is fundamentally different than other packages in the archive. The described process and criteria are built on the normal StableReleaseUpdates document, and where these documents conflict, this document takes precedence.

What sort of updates are allowed for post-release kernels?

Additionally to the generic requirements acceptable bugs fall into one of the following criteria:

  1. It fixes a critical issue (data-loss, OOPs, crashes) or is security related (security related issues might be covered by security releases which are special in handling and publication).
  2. Simple, obvious and short fixes or hardware enablement patches as long as the next release has not reached beta status. If there is a related upstream stable tree open (see below) this class of patches is required to come through the upstream process. Patches sent upstream for that reason must include their BugLink reference.

  3. For 3 months after a non-LTS release or the lifetime of a LTS release we will be pulling upstream stable updates from the corresponding series. There will be one tracking bug report for each stable update but additional references to existing bugs will be added to the contained patches (on best can do base). Uploads containing a upstream stable patchset increase the retention period to two weeks, while it is one week otherwise.
  4. Fixes to drivers which are not upstream are accepted directly if they fall into the first two categories.
  5. $DEITY intervention. Might happen, but very very rarely and will not be explainable.

How long will updates be allowed for a release

This answer is directed at non-LTS releases. For normal 18-month releases, we will only accept updates to the kernel for 3-4 months after release. At this point we consider the in-development release to be stable enough for testing, and the primary target for fixing bugs. Plus, 3-4 months after release, most major bugs are either reported and fixed in the stable release, or deemed unfixable.

There may be a few exceptions to this, but don't count on them.

How does the process work

  • First step for every SRU is to have a bug associated.
  • The patch or patches must contain the link to the Launchpad bug and contain a Signed-off-by line from the submitter.

  • The beginning of the description area of the bug needs to have a SRU justification which should look like this example:

        SRU Justification:
    
        Impact: <a short description about the symptoms and the impact of the bug>
        Fix: <how was this fixed, where did the fix come from>
        Testcase: <how can the fix be tested>
  • If the fix for a problem applies to the requirements for a SRU and has also been tested to successfully solved the bug, then the next step depends on whether the fix is serious enough to be directly applied or whether it should go in via upstream stable (as long as that is appropriate).
  • For serious problem fixes, the patch must be sent the the kernel-team mailing list. There is requires ACKs from two senior kernel-team members and then will be applied to the kernel tree. Even though going into the tree on the faster path, the next step should be done.

        To: kernel-team@lists.ubuntu.com
        Subject: [<series>] SRU: <bug/patch title>
    
        <SRU justification copied from the bug>
    
        <patch attached or pull request>
  • For all other patches (as long as the upstream stable is appropriate) the fix has to be sent upstream (when the problem is there as well and the patch is not a backport) and to stable@kernel.org (if it has not been sent there before). As soon as that is accepted there, it will come back its way when we pull stable updates.

How will updates be provided in the archive

  • Security updates will be uploaded directly into -security without other changes. This just requires a temporary GIT fork which will be immediately merged back into the main branch for that stable release.
  • Normal updates will be provided as pre-releases through the kernel-ppa users PPA. At certain points those get made into proposed releases which are uploaded to the proposed pocket. Then again they have to get verified to fix the problems and not to cause regressions.

KernelTeam/KernelUpdates (last edited 2020-10-22 12:03:42 by anthonywong)