Skip to contentSkip to conversation

doc: update CI and pre-merge details in committer guidelines

PATCH In review Jani Nikula 3 participants 10 messages Jun 27, 2018 – Jul 6, 2018

lore
Share & tools
lore
#0 lore raw

PATCH 1/2 doc: update CI and pre-merge details in committer guidelines

Lots has happened in the CI front since the first version was added.

---
 drm-intel.rst | 45 ++++++++++++++++++++++++++++-----------------
 1 file changed, 28 insertions(+), 17 deletions(-)
drm-intel.rst +28 -17
index c68949a41c95..baf48f459dd9 100644
@@ -172,6 +172,8 @@ An inexhaustive list of details to check:
   `details on testing requirements
   <http://blog.ffwll.ch/2013/11/testing-requirements-for-drmi915.html>`_.
 
+* The patch series has passed CI pre-merge testing. See CI details below.
+
 * An open source userspace, reviewed and ready for merging by the upstream
   project, must be available for new kernel ABI. Please see `details on
   upstreaming requirements
@@ -186,11 +188,6 @@ An inexhaustive list of details to check:
   (or the author) stand a chance to fairly quickly understand what goes wrong if
   the commit is reported to cause a regression?
 
-* `checkpatch.pl` does not complain. (Some of the more subjective warnings may
-  be ignored at the committer's discretion.)
-
-* The patch does not introduce new `sparse` warnings.
-
 * When pushing someone else's patch you must add your own signed-off per
   http://developercertificate.org/. dim apply-branch should do this
   automatically for you.
@@ -244,8 +241,6 @@ On Confidence, Complexity, and Transparency
   you have involved enough people to feel comfortable if the justification for
   the commit is questioned afterwards.
 
-* Make sure pre-merge testing is completed successfully.
-
 On Rough Consensus
 ------------------
 
@@ -290,18 +285,34 @@ discussions happen in public forums, and make sure there's a searchable
 permanent record of any discussions for later reference. This means that for
 most things internal meetings are not the most suitable venue.
 
-Pre-Merge Testing
------------------
+Continuous Integration and Pre-Merge Testing
+--------------------------------------------
+
+The requirements for CI_ pre-merge testing are:
+
+* ``checkpatch.pl`` does not complain. (Some of the more subjective warnings may
+  be ignored at the committer's discretion.)
+
+* The patch does not introduce new ``sparse`` warnings.
+
+* Patch series must pass IGT Basic Acceptance Tests (BAT) on all the CI machines
+  without causing regressions.
+
+* Patch series must pass full IGT tests on CI shard machines without causing
+  regressions.
+
+The CI bots will send results to the patch author and intel-gfx for any patches
+tracked by patchwork. The results are also available on patchwork_ and the CI_
+site.
+
+Check CI failures and make sure any sporadic failures are a) pre-existing,
+and b) tracked in bugzilla. If there's anything dubious that you can't track
+down to pre-existing and tracked issues please don't push, but instead figure
+out what's going on.
 
-Our CI infrastructure is being built up and currently requirements for pre-merge
-testing are fairly simple:
+.. _CI: https://intel-gfx-ci.01.org/
 
-* All patches must past IGT Basic Acceptance Tests (BAT) on all the CI machines
-  without causing regressions.  The CI bots will send results to intel-gfx for
-  any patches tracked by patchwork. Check CI failures and make sure any sporadic
-  failures are a) pre-existing b) tracked in bugzilla. If there's anything
-  dubious that you can't track down to pre-existing&tracked issues please don't
-  push, but instead figure out what's going on.
+.. _patchwork: https://patchwork.freedesktop.org/project/intel-gfx/series/
 
 Tooling
 =======
Signature · 6 lines
2.11.0

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw

PATCH 2/2 doc: clarify what type of changes are acceptable at commit time

As a rule of thumb, don't change patches while committing.

---
 drm-intel.rst | 7 +++++++
 1 file changed, 7 insertions(+)
drm-intel.rst +7 -0
index baf48f459dd9..ad8ff9739336 100644
@@ -196,6 +196,13 @@ An inexhaustive list of details to check:
   coordinate with maintainers to avoid unnecessary pain with conflicts. Usually
   some explicit merges are needed to avoid git getting lost.
 
+* As a general rule, do not modify the patches while applying, apart from the
+  commit message. If the patch conflicts, or needs to be changed due to review,
+  have the author rebase, update and resend. Any change at this stage is a
+  potential issue bypassing CI.
+
+  At most, minor comment and whitespace tweaks are acceptable.
+
 On Confidence, Complexity, and Transparency
 -------------------------------------------
 
Signature · 6 lines
2.11.0

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw

Re: [DIM DOCS PATCH 2/2] doc: clarify what type of changes are acceptable at commit time

Quoted 13 lines · Jani Nikula

As a rule of thumb, don't change patches while committing.

---
 drm-intel.rst | 7 +++++++
 1 file changed, 7 insertions(+)
drm-intel.rst +0 -0
index baf48f459dd9..ad8ff9739336 100644

drm-intel.rst An inexhaustive list of details to check: · line 203

Quoted 8 lines
drm-intel.rst +4 -0
@@ -196,6 +196,13 @@ An inexhaustive list of details to check:
   coordinate with maintainers to avoid unnecessary pain with conflicts. Usually
   some explicit merges are needed to avoid git getting lost.
 
+* As a general rule, do not modify the patches while applying, apart from the
+  commit message. If the patch conflicts, or needs to be changed due to review,
+  have the author rebase, update and resend. Any change at this stage is a
+  potential issue bypassing CI.
 

Should we also mention that merge conflicts need to be told to maintainers, so that they can do a backmerge? Just because this blew up recently for drm-misc ...
-Daniel

Quoted 12 lines
drm-intel.rst +2 -1
+  At most, minor comment and whitespace tweaks are acceptable.
+
 On Confidence, Complexity, and Transparency
 -------------------------------------------
 
--

2.11.0

_______________________________________________
dim-tools mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dim-tools

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#3 lore raw

Re: [DIM DOCS PATCH 2/2] doc: clarify what type of changes are acceptable at commit time

Quoted 17 lines · Daniel Vetter
Quoted 16 lines · Jani Nikula

As a rule of thumb, don't change patches while committing.

---
 drm-intel.rst | 7 +++++++
 1 file changed, 7 insertions(+)
drm-intel.rst +0 -0
index baf48f459dd9..ad8ff9739336 100644
@@ -196,6 +196,13 @@ An inexhaustive list of details to check:
   coordinate with maintainers to avoid unnecessary pain with conflicts. Usually
   some explicit merges are needed to avoid git getting lost.
 
Quoted 8 lines

drm-intel.rst An inexhaustive list of details to check: · line 208

Quoted 5 lines
drm-intel.rst +4 -0
+* As a general rule, do not modify the patches while applying, apart from the
+  commit message. If the patch conflicts, or needs to be changed due to review,
+  have the author rebase, update and resend. Any change at this stage is a
+  potential issue bypassing CI.
 

Should we also mention that merge conflicts need to be told to maintainers, so that they can do a backmerge? Just because this blew up recently for drm-misc ...

Added bullet

* Please notify maintainers (IRC is fine) of new merge conflicts during drm-tip rebuild, so that they can do backmerges as needed.

and pushed to get this done. Please amend if that's not what you wanted.

Thanks,
Jani.

Quoted 14 lines

-Daniel

Quoted 12 lines
drm-intel.rst +2 -1
+  At most, minor comment and whitespace tweaks are acceptable.
+
 On Confidence, Complexity, and Transparency
 -------------------------------------------
 
--

2.11.0

_______________________________________________
dim-tools mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dim-tools

Signature · 5 lines
Jani Nikula, Intel Open Source Graphics Center
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#4 lore raw

Re: [DIM DOCS PATCH 2/2] doc: clarify what type of changes are acceptable at commit time

Quoted 27 lines · Jani Nikula
Quoted 17 lines · Daniel Vetter
Quoted 16 lines · Jani Nikula

As a rule of thumb, don't change patches while committing.

---
 drm-intel.rst | 7 +++++++
 1 file changed, 7 insertions(+)
drm-intel.rst +0 -0
index baf48f459dd9..ad8ff9739336 100644
@@ -196,6 +196,13 @@ An inexhaustive list of details to check:
   coordinate with maintainers to avoid unnecessary pain with conflicts. Usually
   some explicit merges are needed to avoid git getting lost.
 
Quoted 8 lines

drm-intel.rst An inexhaustive list of details to check: · line 208

Quoted 5 lines
drm-intel.rst +4 -0
+* As a general rule, do not modify the patches while applying, apart from the
+  commit message. If the patch conflicts, or needs to be changed due to review,
+  have the author rebase, update and resend. Any change at this stage is a
+  potential issue bypassing CI.
 

Should we also mention that merge conflicts need to be told to maintainers, so that they can do a backmerge? Just because this blew up recently for drm-misc ...

Quoted 6 lines

Added bullet

* Please notify maintainers (IRC is fine) of new merge conflicts during drm-tip rebuild, so that they can do backmerges as needed.

and pushed to get this done. Please amend if that's not what you wanted.

Personally I preferred committers told me before they wrestled a patch into submission because it missed a required patch, then wrestled drm-tip into submission with a nasty conflict on top. Doing a backmerge first, pushing 2nd is much simpler, and fits into the bullet about not modifying patches.

I'd have added something like "Also don't modify patches due to conflicts. Instead ask maintainers to backmerge your required baseline patches first, since that avoids tons of headaches later on with conflicts in integration trees and pull requests." Or something like that.
-Daniel

Quoted 22 lines

Thanks,
Jani.

Quoted 14 lines

-Daniel

Quoted 12 lines
drm-intel.rst +2 -1
+  At most, minor comment and whitespace tweaks are acceptable.
+
 On Confidence, Complexity, and Transparency
 -------------------------------------------
 
--

2.11.0

_______________________________________________
dim-tools mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dim-tools

--
Jani Nikula, Intel Open Source Graphics Center

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#5 lore raw

Re: [DIM DOCS PATCH 2/2] doc: clarify what type of changes are acceptable at commit time

Quoted 41 lines · Daniel Vetter
Quoted 27 lines · Jani Nikula
Quoted 17 lines · Daniel Vetter
Quoted 16 lines · Jani Nikula

As a rule of thumb, don't change patches while committing.

---
 drm-intel.rst | 7 +++++++
 1 file changed, 7 insertions(+)
drm-intel.rst +0 -0
index baf48f459dd9..ad8ff9739336 100644
@@ -196,6 +196,13 @@ An inexhaustive list of details to check:
   coordinate with maintainers to avoid unnecessary pain with conflicts. Usually
   some explicit merges are needed to avoid git getting lost.
 
Quoted 8 lines

drm-intel.rst An inexhaustive list of details to check: · line 208

Quoted 5 lines
drm-intel.rst +4 -0
+* As a general rule, do not modify the patches while applying, apart from the
+  commit message. If the patch conflicts, or needs to be changed due to review,
+  have the author rebase, update and resend. Any change at this stage is a
+  potential issue bypassing CI.
 

Should we also mention that merge conflicts need to be told to maintainers, so that they can do a backmerge? Just because this blew up recently for drm-misc ...

Quoted 6 lines

Added bullet

* Please notify maintainers (IRC is fine) of new merge conflicts during drm-tip rebuild, so that they can do backmerges as needed.

and pushed to get this done. Please amend if that's not what you wanted.

Personally I preferred committers told me before they wrestled a patch into submission because it missed a required patch, then wrestled drm-tip into submission with a nasty conflict on top. Doing a backmerge first, pushing 2nd is much simpler, and fits into the bullet about not modifying patches.

Quoted 5 lines

I'd have added something like "Also don't modify patches due to conflicts. Instead ask maintainers to backmerge your required baseline patches first, since that avoids tons of headaches later on with conflicts in integration trees and pull requests." Or something like that.

Oh, I misunderstood you, in part because one of my original bullets already says don't modify patches due to conflicts. I thought with "merge conflicts" you specifically meant drm-tip conflicts.

BR,
Jani.

Quoted 24 lines

-Daniel

Quoted 22 lines

Thanks,
Jani.

Quoted 14 lines

-Daniel

Quoted 12 lines
drm-intel.rst +2 -1
+  At most, minor comment and whitespace tweaks are acceptable.
+
 On Confidence, Complexity, and Transparency
 -------------------------------------------
 
--

2.11.0

_______________________________________________
dim-tools mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dim-tools

--
Jani Nikula, Intel Open Source Graphics Center

Signature · 5 lines
Jani Nikula, Intel Open Source Graphics Center
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Acked-by on v1 2/2
#6 lore raw

Re: [DIM DOCS PATCH 2/2] doc: clarify what type of changes are acceptable at commit time

In reply to Jani Nikula
Quoted 4 lines · Jani Nikula

As a rule of thumb, don't change patches while committing.

Quoted 29 lines
---
 drm-intel.rst | 7 +++++++
 1 file changed, 7 insertions(+)
drm-intel.rst +7 -0
index baf48f459dd9..ad8ff9739336 100644
@@ -196,6 +196,13 @@ An inexhaustive list of details to check:
   coordinate with maintainers to avoid unnecessary pain with conflicts. Usually
   some explicit merges are needed to avoid git getting lost.
 
+* As a general rule, do not modify the patches while applying, apart from the
+  commit message. If the patch conflicts, or needs to be changed due to review,
+  have the author rebase, update and resend. Any change at this stage is a
+  potential issue bypassing CI.
+
+  At most, minor comment and whitespace tweaks are acceptable.
+
 On Confidence, Complexity, and Transparency
 -------------------------------------------
 
Signature · 6 lines
2.11.0

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#7 lore raw
In reply to Jani Nikula
Quoted 52 lines · Jani Nikula

Lots has happened in the CI front since the first version was added.

---
 drm-intel.rst | 45 ++++++++++++++++++++++++++++-----------------
 1 file changed, 28 insertions(+), 17 deletions(-)
drm-intel.rst +7 -9
index c68949a41c95..baf48f459dd9 100644
@@ -172,6 +172,8 @@ An inexhaustive list of details to check:
   `details on testing requirements
   <http://blog.ffwll.ch/2013/11/testing-requirements-for-drmi915.html>`_.
 
+* The patch series has passed CI pre-merge testing. See CI details below.
+
 * An open source userspace, reviewed and ready for merging by the upstream
   project, must be available for new kernel ABI. Please see `details on
   upstreaming requirements
@@ -186,11 +188,6 @@ An inexhaustive list of details to check:
   (or the author) stand a chance to fairly quickly understand what goes wrong if
   the commit is reported to cause a regression?
 
-* `checkpatch.pl` does not complain. (Some of the more subjective warnings may
-  be ignored at the committer's discretion.)
-
-* The patch does not introduce new `sparse` warnings.
-
 * When pushing someone else's patch you must add your own signed-off per
   http://developercertificate.org/. dim apply-branch should do this
   automatically for you.
@@ -244,8 +241,6 @@ On Confidence, Complexity, and Transparency
   you have involved enough people to feel comfortable if the justification for
   the commit is questioned afterwards.
 
-* Make sure pre-merge testing is completed successfully.
-
 On Rough Consensus
 ------------------
 
@@ -290,18 +285,34 @@ discussions happen in public forums, and make sure there's a searchable
 permanent record of any discussions for later reference. This means that for
 most things internal meetings are not the most suitable venue.
 
-Pre-Merge Testing
------------------
+Continuous Integration and Pre-Merge Testing
+--------------------------------------------
+
+The requirements for CI_ pre-merge testing are:
+

drm-intel.rst discussions happen in public forums, and make sure there's a searchable · line 302

Quoted 10 lines
drm-intel.rst +10 -0
+* ``checkpatch.pl`` does not complain. (Some of the more subjective warnings may
+  be ignored at the committer's discretion.)
+
+* The patch does not introduce new ``sparse`` warnings.
+
+* Patch series must pass IGT Basic Acceptance Tests (BAT) on all the CI machines
+  without causing regressions.
+
+* Patch series must pass full IGT tests on CI shard machines without causing
+  regressions.

* Patch series must pass gpu piglit tests on all CI machines without causing regressions.

Very recent addition, and thus far hasn't really resulted in breakage/regression reports, but it's there&in production.

Otherwise lgtm, ack.
-Daniel

Quoted 31 lines
drm-intel.rst +11 -9
+
+The CI bots will send results to the patch author and intel-gfx for any patches
+tracked by patchwork. The results are also available on patchwork_ and the CI_
+site.
+
+Check CI failures and make sure any sporadic failures are a) pre-existing,
+and b) tracked in bugzilla. If there's anything dubious that you can't track
+down to pre-existing and tracked issues please don't push, but instead figure
+out what's going on.
 
-Our CI infrastructure is being built up and currently requirements for pre-merge
-testing are fairly simple:
+.. _CI: https://intel-gfx-ci.01.org/
 
-* All patches must past IGT Basic Acceptance Tests (BAT) on all the CI machines
-  without causing regressions.  The CI bots will send results to intel-gfx for
-  any patches tracked by patchwork. Check CI failures and make sure any sporadic
-  failures are a) pre-existing b) tracked in bugzilla. If there's anything
-  dubious that you can't track down to pre-existing&tracked issues please don't
-  push, but instead figure out what's going on.
+.. _patchwork: https://patchwork.freedesktop.org/project/intel-gfx/series/
 
 Tooling
 =======
--

2.11.0

_______________________________________________
dim-tools mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dim-tools

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#8 lore raw
Quoted 67 lines · Daniel Vetter
Quoted 52 lines · Jani Nikula

Lots has happened in the CI front since the first version was added.

---
 drm-intel.rst | 45 ++++++++++++++++++++++++++++-----------------
 1 file changed, 28 insertions(+), 17 deletions(-)
drm-intel.rst +7 -9
index c68949a41c95..baf48f459dd9 100644
@@ -172,6 +172,8 @@ An inexhaustive list of details to check:
   `details on testing requirements
   <http://blog.ffwll.ch/2013/11/testing-requirements-for-drmi915.html>`_.
 
+* The patch series has passed CI pre-merge testing. See CI details below.
+
 * An open source userspace, reviewed and ready for merging by the upstream
   project, must be available for new kernel ABI. Please see `details on
   upstreaming requirements
@@ -186,11 +188,6 @@ An inexhaustive list of details to check:
   (or the author) stand a chance to fairly quickly understand what goes wrong if
   the commit is reported to cause a regression?
 
-* `checkpatch.pl` does not complain. (Some of the more subjective warnings may
-  be ignored at the committer's discretion.)
-
-* The patch does not introduce new `sparse` warnings.
-
 * When pushing someone else's patch you must add your own signed-off per
   http://developercertificate.org/. dim apply-branch should do this
   automatically for you.
@@ -244,8 +241,6 @@ On Confidence, Complexity, and Transparency
   you have involved enough people to feel comfortable if the justification for
   the commit is questioned afterwards.
 
-* Make sure pre-merge testing is completed successfully.
-
 On Rough Consensus
 ------------------
 
@@ -290,18 +285,34 @@ discussions happen in public forums, and make sure there's a searchable
 permanent record of any discussions for later reference. This means that for
 most things internal meetings are not the most suitable venue.
 
-Pre-Merge Testing
------------------
+Continuous Integration and Pre-Merge Testing
+--------------------------------------------
+
+The requirements for CI_ pre-merge testing are:
+

drm-intel.rst discussions happen in public forums, and make sure there's a searchable · line 302

Quoted 10 lines
drm-intel.rst +10 -0
+* ``checkpatch.pl`` does not complain. (Some of the more subjective warnings may
+  be ignored at the committer's discretion.)
+
+* The patch does not introduce new ``sparse`` warnings.
+
+* Patch series must pass IGT Basic Acceptance Tests (BAT) on all the CI machines
+  without causing regressions.
+
+* Patch series must pass full IGT tests on CI shard machines without causing
+  regressions.

* Patch series must pass gpu piglit tests on all CI machines without causing regressions.

Quoted 2 lines

Very recent addition, and thus far hasn't really resulted in breakage/regression reports, but it's there&in production.

Copy-pasted that and pushed, thanks.

BR,
Jani.

Quoted 35 lines

Otherwise lgtm, ack.
-Daniel

Quoted 31 lines
drm-intel.rst +11 -9
+
+The CI bots will send results to the patch author and intel-gfx for any patches
+tracked by patchwork. The results are also available on patchwork_ and the CI_
+site.
+
+Check CI failures and make sure any sporadic failures are a) pre-existing,
+and b) tracked in bugzilla. If there's anything dubious that you can't track
+down to pre-existing and tracked issues please don't push, but instead figure
+out what's going on.
 
-Our CI infrastructure is being built up and currently requirements for pre-merge
-testing are fairly simple:
+.. _CI: https://intel-gfx-ci.01.org/
 
-* All patches must past IGT Basic Acceptance Tests (BAT) on all the CI machines
-  without causing regressions.  The CI bots will send results to intel-gfx for
-  any patches tracked by patchwork. Check CI failures and make sure any sporadic
-  failures are a) pre-existing b) tracked in bugzilla. If there's anything
-  dubious that you can't track down to pre-existing&tracked issues please don't
-  push, but instead figure out what's going on.
+.. _patchwork: https://patchwork.freedesktop.org/project/intel-gfx/series/
 
 Tooling
 =======
--

2.11.0

_______________________________________________
dim-tools mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dim-tools

Signature · 5 lines
Jani Nikula, Intel Open Source Graphics Center
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Acked-by on v1 1/2
#9 lore raw
In reply to Jani Nikula
Quoted 3 lines · Jani Nikula

Lots has happened in the CI front since the first version was added.

Quoted 90 lines
---
 drm-intel.rst | 45 ++++++++++++++++++++++++++++-----------------
 1 file changed, 28 insertions(+), 17 deletions(-)
drm-intel.rst +28 -17
index c68949a41c95..baf48f459dd9 100644
@@ -172,6 +172,8 @@ An inexhaustive list of details to check:
   `details on testing requirements
   <http://blog.ffwll.ch/2013/11/testing-requirements-for-drmi915.html>`_.
 
+* The patch series has passed CI pre-merge testing. See CI details below.
+
 * An open source userspace, reviewed and ready for merging by the upstream
   project, must be available for new kernel ABI. Please see `details on
   upstreaming requirements
@@ -186,11 +188,6 @@ An inexhaustive list of details to check:
   (or the author) stand a chance to fairly quickly understand what goes wrong if
   the commit is reported to cause a regression?
 
-* `checkpatch.pl` does not complain. (Some of the more subjective warnings may
-  be ignored at the committer's discretion.)
-
-* The patch does not introduce new `sparse` warnings.
-
 * When pushing someone else's patch you must add your own signed-off per
   http://developercertificate.org/. dim apply-branch should do this
   automatically for you.
@@ -244,8 +241,6 @@ On Confidence, Complexity, and Transparency
   you have involved enough people to feel comfortable if the justification for
   the commit is questioned afterwards.
 
-* Make sure pre-merge testing is completed successfully.
-
 On Rough Consensus
 ------------------
 
@@ -290,18 +285,34 @@ discussions happen in public forums, and make sure there's a searchable
 permanent record of any discussions for later reference. This means that for
 most things internal meetings are not the most suitable venue.
 
-Pre-Merge Testing
------------------
+Continuous Integration and Pre-Merge Testing
+--------------------------------------------
+
+The requirements for CI_ pre-merge testing are:
+
+* ``checkpatch.pl`` does not complain. (Some of the more subjective warnings may
+  be ignored at the committer's discretion.)
+
+* The patch does not introduce new ``sparse`` warnings.
+
+* Patch series must pass IGT Basic Acceptance Tests (BAT) on all the CI machines
+  without causing regressions.
+
+* Patch series must pass full IGT tests on CI shard machines without causing
+  regressions.
+
+The CI bots will send results to the patch author and intel-gfx for any patches
+tracked by patchwork. The results are also available on patchwork_ and the CI_
+site.
+
+Check CI failures and make sure any sporadic failures are a) pre-existing,
+and b) tracked in bugzilla. If there's anything dubious that you can't track
+down to pre-existing and tracked issues please don't push, but instead figure
+out what's going on.
 
-Our CI infrastructure is being built up and currently requirements for pre-merge
-testing are fairly simple:
+.. _CI: https://intel-gfx-ci.01.org/
 
-* All patches must past IGT Basic Acceptance Tests (BAT) on all the CI machines
-  without causing regressions.  The CI bots will send results to intel-gfx for
-  any patches tracked by patchwork. Check CI failures and make sure any sporadic
-  failures are a) pre-existing b) tracked in bugzilla. If there's anything
-  dubious that you can't track down to pre-existing&tracked issues please don't
-  push, but instead figure out what's going on.
+.. _patchwork: https://patchwork.freedesktop.org/project/intel-gfx/series/
 
 Tooling
 =======
Signature · 6 lines
2.11.0

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx