OS Status Meeting Minutes
October 15, 2007
- Important Trinity 2 dates to be aware of:
- M1 test results review is today
- Build for M2 is this Wednesday (Oct 17) for OS testing on the weekend
- M2 testing by other groups starts next Wed (Oct 24)
- Peter is a special guest for this meeting.
- Introduced Peter to the team. Peter has been recently hired as a manager in the support organization and will be working closely with the OS team.
- Working on test case development for Trinity 2. Have two test cases ready for code review; one passes and one fails. For the failure case will consult with the developer to resolve the issue.
- Last week, updated a number of PRs for Trinity 2.
- Resolved issues reported by a customer that is using a Trinity 2 source drop.
- Spent some time investigating SMP scheduling interrupt thread latency jitter. Will continue looking at this. (PR 40658)
- Spent time on Memory Partitioning code reviews for Trinity 2 which will continue this week.
- Will start working on wired tlb fix for 1Mb size limitation for Trinity 2.
- Have been working on test case development for Trinity 2. Will have 6 or 7 ready for review tomorrow.
- Working on 2 high priority customer issues.
- One is for a crash that happened on 6.2.1 that was triggered by a msgReplyv on a stale rcvid. Have a test case and a theory as to what is happening that still needs to be verified. (PR 51782)
- The other is with SP3 and is a possible mutual exclusion hole. Now have a test procnto to catch events to try and prove the theory. (PR 51538)
- Spent time working on point fixes for a Trinity 2 source drop.
- Working with a customer to see if a Trinity 2 patch that was provided to fix a stat() EINVAL failure has resolved the issue. (PR 50959)
- Spent some time investigating SMP scheduling interrupt thread latency jitter.(PR 40658)
- Will be working on some more fixes for Trinity 2.
- Spent time on some manual testing and clean up of test cases for Memory Partitioning for Trinity 2.
- Ran lint on Trinity 2 which reported some new errors. Will give to Mosa to make changes and do the fixes and submit for review.
- Time spent getting report card ready for M1 test results review today. No info yet on benchmarks.
- Updated Trinity 2 Master Test Plan. Working with the install tester, Gianluca, to capture updates to test cases due resulting from changes in the installer requirements.
- Updated and wrote some new test cases for Trinity 2. Added a number of d-unit tests that hadn’t been added yet to the regression.
- Had an nfs problem in the lab on the weekend which affected the testing. Restarted the test and will monitor.
- One of the terminal servers in the lab went down. Shahram will be upgrading the software on the third terminal server and we’ll swap that one in to correct the issue.
- Most of week spent on code reviews for Memory Partitioning.
- Working on completion of Trinity 2 user documentation.
- Initial focus is applying editorial comments to the ARM tech notes.
- Also applying changes to Memory Partitioning doc based on feedback from Attilla/Mike.
- Completed Trinity 2 file system testing for M1 and created PRs.
- Retested proc fix for Trinity 2.
- Plan to rerun Trinity 2 TDK compatibility testing.
- Last week spent time on customer reported stat() EINVAL failure on Trinity 2 that Colin eventually fixed. (PR 50959)
- Have a couple of critical PRs to work on for Trinity 2. One is thread pools issue which is likely resolved as part of a fix that was done a number of weeks ago so will confirm if it’s reproducible in the current load. (PR 19408)
- Another PR is a test stopper for Trinity 2. Colin has suggested a possible fix so will check if it works. (PR 47739)
- Working on a couple of crashes seen on a customer modified kernel.