[i2c] V2(S) Signoff #22108
Comments
Since i2c was versioned at 1.1.0 (in 0a8d1b2), we made multiple changes that change the HW and SW interfaces of this HW IP block (a `git diff 0a8d1b2..<this commit>^ -- hw/ip/i2c/data/i2c.hjson` gives a good overview) in a backward-incompatible way. Thus we need to give I2C a major version bump and reset its D and V stages. Issues lowRISC#20971 and lowRISC#22108 track signing i2c off at D2(S) and V2(S) again, respectively. Signed-off-by: Andreas Kurth <[email protected]>
|
I've set up "pretend meeting review notes" here, which is hopefully the leg-work needed for signing off the block as v1. I believe the main work item remaining is to tweak the testplan to reflect design changes that have come in since the ES tapeout. This task is tracked as #22168. For anyone who finds this and is confused: sorry! This note should have been added to the V1 signoff issue |
Since i2c was versioned at 1.1.0 (in 0a8d1b2), we made multiple changes that change the HW and SW interfaces of this HW IP block (a `git diff 0a8d1b2..<this commit>^ -- hw/ip/i2c/data/i2c.hjson` gives a good overview) in a backward-incompatible way. Thus we need to give I2C a major version bump and reset its D and V stages. Issues lowRISC#20971 and lowRISC#22108 track signing i2c off at D2(S) and V2(S) again, respectively. Signed-off-by: Andreas Kurth <[email protected]>
Since i2c was versioned at 1.1.0 (in 0a8d1b2), we made multiple changes that change the HW and SW interfaces of this HW IP block (a `git diff 0a8d1b2..<this commit>^ -- hw/ip/i2c/data/i2c.hjson` gives a good overview) in a backward-incompatible way. Thus we need to give I2C a major version bump and reset its D and V stages. Issues lowRISC#20971 and lowRISC#22108 track signing i2c off at D2(S) and V2(S) again, respectively. Signed-off-by: Andreas Kurth <[email protected]>
Since i2c was versioned at 1.1.0 (in 0a8d1b2), we made multiple changes that change the HW and SW interfaces of this HW IP block (a `git diff 0a8d1b2..<this commit>^ -- hw/ip/i2c/data/i2c.hjson` gives a good overview) in a backward-incompatible way. Thus we need to give I2C a major version bump and reset its D and V stages. Issues #20971 and #22108 track signing i2c off at D2(S) and V2(S) again, respectively. Signed-off-by: Andreas Kurth <[email protected]>
I2C V2/V2S Signoff
This is categorized as a focus area block. SummaryHighlight important RTL changes that have an impact on DV and the testplan. For each such change make sure the testplan is in sync with the change The I2C block has undergone significant changes since the ES.
Look over the V checklist items up to the stage we intend to sign off in order to double check that none of the checklist items are violated by the RTL changes. This issue tracks the primary outstanding items as per the M4 milestone : #23077
Commits since Earlgrey-ES tapeout$ git rev-parse --short HEAD PRs closed since the Earlgrey-ES tapeout
PRs open
Issues closed since the Earlgrey-ES tapeoutUse the following filter to search for relevant issues and capture them in the issue: is:issue is:closed closed:>2023-06-27 label:IP:i2c
Currently open issuesUse the following filter to search for relevant issues and capture them in the issue: is:issue is:open label:IP:i2c
For each issue that is still open, make a note why it does not have to be resolved for this signoff. If it can’t be deferred, the issue needs to be addressed before proceeding with the signoff. Regression Summary
Coverage Summary
|
|
Discussed moving as P1 to M5 due to RTL risk analysis performed during last issue triage. |
|
The above V2(S) review comment follows the lightweight signoff process for a V2(S) level signoff, but as agreed and discussed, this signoff does not meet the full signoff critera due to the addition of new RTL features at a late point in the overall development timeline. Importantly, the following qualifying critera apply:
Features out-of-scope for this signoff- Testpoints applicable to the following new RTL features
- Functional Coverage applicable to the new testpoints- Code Coverage of RTL directly contributory to the new features onlyConfidence-level based on this qualified signoffBy definition, this signoff cannot opine about the verification confidence in the new RTL features.
The following evidence supports the assertion that we are sufficiently confident to signoff to this qualified level.:
|
|
Thanks for the careful write-up. I'm convinced and happy that it's correct. I propose you send a PR updating the development stage in the hjson, with a note in the checklist about the limited scope ("we have only verified XYZ"). |
This moves the I2C HWIP block to V2S, as described in the signoff review issue #22108. > Note. there are a few in-flight PRs (#23987, #24010) that need to be merged before this one. When they are merged, the lightweight signoff criteria bullets in the signoff issue [comment here](#22108 (comment)) will be updated to reflect the latest status. Signed-off-by: Harry Callahan <[email protected]>
|
I've updated the signoff checklist comment above, based on the merger of #23987 and #24010. This includes updating the note about low FSM coverage based on a review of the local 10-seed All-Once regression I ran to support #23987, and the review discussion where it was agreed this was acceptable. As such, I think this is ready to signoff by merging #24011. Though if we wished to wait for a nightly regression to confirm the coverage metrics, that would also be acceptable. |
This moves the I2C HWIP block to V2S, as described in the signoff review issue lowRISC#22108. > Note. there are a few in-flight PRs (lowRISC#23987, lowRISC#24010) that need to be merged before this one. When they are merged, the lightweight signoff criteria bullets in the signoff issue [comment here](lowRISC#22108 (comment)) will be updated to reflect the latest status. Signed-off-by: Harry Callahan <[email protected]>
This moves the I2C HWIP block to V2S, as described in the signoff review issue #22108. > Note. there are a few in-flight PRs (#23987, #24010) that need to be merged before this one. When they are merged, the lightweight signoff criteria bullets in the signoff issue [comment here](#22108 (comment)) will be updated to reflect the latest status. Signed-off-by: Harry Callahan <[email protected]>
Ensure V2(S) signoff criteria are fulfilled.
This issue tracks V2(S) signoff for the M5-scoped components of the I2C hwip. As such, this is not a full signoff, but a signoff with qualifiers. The qualifying conditions can be seen in the following linked thread comment. (And are also linked in the
i2c.hjsonfile 'revisions' section under 'Notes'). The lightweight signoff template has also been completed in the following comment for the scoped features.Signoff Meeting Notes (OpentitanACL)
Depends on items captured in the following issues to close:
Effort tracked here is just for further signoff admin.
The text was updated successfully, but these errors were encountered: