OS Status Meeting Minutes

November 5, 2007

Steve Bergwerff

• Shiv is here in Ottawa for 1 more week
• Colin is in Germany this week.
• Attilla is at a process meeting.
• Important Trinity 2 dates to be aware of:
  • M2 full testing started October 24
  • M2 test results review is scheduled for November 9

Doug Bailey

• Worked on a number of PRs last week
  • Implemented a fix for a Trinity 2 test stopper PR raised by the GUI team on M1. Fix has been implemented, reviewed, regressed, and checked in. (PR 51940)
  • Fixed an issue for a 6.2.0 based customer. The fix has been sent to the customer for risk evaluation of including the change. Waiting on a response/decision. (PR 51782)
  • PR 51538 was reopened in order to apply the fix to another release branch for the customer.
  • Started looking at Trinity 2 PR 21182 remote spawn issue.

Peter Mitsis

• Working to get Trinity 2 benchmarks for file system read/write performance based on a customer report of a possible slowdown.
• Sent out a code review for a fix to fdisk.
• Investigating implementation of CDROM in preparation for looking at some PRs in this area.

Chris Travis

• Last week was mostly spent analyzing the M2 regression test logs and raising PRs for any issues.
• Have requested an official build of a Trinity 2 procnto patch that includes a few fixes that are currently test stoppers. It includes a fix to freemem reporting (PR 52390). Plan to run the regression tonight on selected targets.
• Trinity 2 Dinkum tests are now running, had an issue with not seeing stop message properly but that is resolved. Re-ran the tests and currently have two tests not working. This has been PR’d and assigned to Doug.
• Worked with tools group on getting qconn working with memory partitioning for Trinity 2 so they can test with it. Don’t currently have the binary needed but Mike Kisel is working on this. It is contingent on resolving some design decisions for external interfaces. The proposal is currently submitted for code review so waiting on responses.
• Will be running the Memory Partitioning tests as part of M2. Marion is running them as well.
• Posted the M2 build on Foundary27. Plan to post the patch this week after testing it.
• Friday had meeting with Kerry on alpha/beta plans for Trinity2.
• Sada is currently setting up a tinderbox build/test of Trinity2 using gcc 4.2.1 on tinderbox. He’s currently working through some build issues. Brian is assisting with resolution.
• Had a problem booting some targets in the lab but turned out to be due to conflicting IP address on terminal server.

Mosa Yeung

• Mostly complete with benchmarking of interrupt latency on Trinity 2. Will review results with Robert Craig before sending to the product manager.
• Fixed a number of warnings that were reported with the lint tool against Trinity 2. Still quite a few to complete. Would like someone to take a look to determine if the results are what was expected. Chris or Attilla will take a look.

Brian Stecher

• Finished the implementation for the MIPS wired TLB fix for Trinity 2. Completed code reviews and testing. Provided a patch on Tuesday for a customer to use with performance testing of Trinity 2.
• Working on a number of Trinity 2 PRs.
  • For PR 51820 that was reported by a customer working with Trinity 2, was not able to reproduce it here. Gave the test to the customer and it passed for them as well. Currently waiting on some additional info from them.

Mike Kisel

• Involved with issue reported for keeping external and internal repositories in sync for Foundary27. The IS team has cleaned up the ones that have been reported and are working on a script to fix the overall problem.
• Largest part of last week was spent reworking interrupt locking threads for Memory Partitioning on SMP. Having some problems debugging the SMP issues but have given some fixes to Marian to test.
• For Memory Partitioning interface, currently waiting on feedback from code reviews.
• This week will continue working on Memory Partitioning and resolving issues.

Shiv Nagarajan

• Covering for Colin this week. Working with a customer that is doing early development with Trinity 2.
• Have been primarily working on 3 customer reported issues for Trinity 2.
  • Pidin failure. (PR 52308) This does not occur here but did finally manage to reproduce it at customer end.
  • Kernel crash. (PR 51824) Only seems to happen if use what the customer describes as a hack to adjust the size of the image and use a custom dumpifs. Took awhile to get a proper setup to reproduce this.
  • SMP kernel crash. This is similar to a problem seen earlier for which a fix was provided.
• Looking at a Posix failure PR for Trinity 2. Talked to Thomas about how to resolve it. He made some suggestions but need to look at it more. (PR 45384)

Sunil Kittur (status reported by email)

• Working on the POSIX Conformance Statement Questionnaire required for submission for certification on Trinity 2. Requires reviewing our implementation of a large number of functions and identifying any cancellation points. Also need to confirm that various optional error codes are correct.
• There is also a Conformance Document to complete. Making good progress on this.
• Both docs need to be reviewed prior to submission.
• Investigating two POSIX failures that were passing previously. One other is being investigated by Shiv. When a clean test run is achieved, will start the official registration.
• In addition:
  • Currently assisting with a customer support issue.
  • Providing pre-sales consultation.
  • Looking at a Trinity 2 M2 regression for a crash on an ARM board.
  • Also need to look at some Trinity 2 ARMv6 PRs.

Gordon Rice

• Received a fix from Doug on Thursday for PR 51782 so will take a look at what is required to port this to 6.2. The code has significantly changed between 6.2 and Head so it’s probably half a day job to complete this plus additional time to test.
• Working on a number of other high priority fixes for a customer modified kernel.

Marian Kos

• Worked on general code cleanup for test cases for Memory Partitioning in Trinity 2. Fixed some tests that had failed due to problems with the test case.
• Completed review of Memory Partitioning test cases and have them checked into svn.
• Ran the Memory Partitioning tests over the weekend. Submitted failures to Mike for resolution.
• Spent some time on a torture test called head-thread. Will file a PR against Trinity 2 for an issues found with threads being left in stop state

Linda Deng

• Created and added a test case for PR 52390, freemem info fix in Trinity 2.
• Working on file system test plan for 6.4.
• Ran into a compiler problem with the file system. Colin/Sunil are looking at it.
• Running file system tests on Trinity 2 M2. See some new errors so going through those now.
• Working with Sunil on completing POSIX certification applications. Seeing two POSIX failures that fail on our test environment but pass on Sunil’s. Sunil is investigating. There is a third failure that Shiv is investigating.
• Working on Variable page size performance tests.

Angelina Gomes

• Last week was spent working on Memory Partitioning documentation. Received some feedback from Marian so will update the version tomorrow morning. Need feedback from Attilla and Mike.
• Fixed some doc PRs for Trinity 2.
• Will finalize the ARMv6 tech notes this week.
• Plan to work on Dinkum doc updates this week. Will talk to Sunil to get more info on what needs to be done.
• Will continue to work on Memory Partitioning doc updates based on feedback.
• For variable page size Tech Note, need some info from Brian. Will send an email reminder later this week.