OS Status Meeting Minutes October 29, 2007

Steve Bergwerff

• Shiv is here in Ottawa for the next 2 weeks
• Important Trinity 2 dates to be aware of:
  • M2 build was completed on Oct 22
  • M2 sanity testing was completed on Oct 24.
  • M2 testing by other groups started October 24
  • M2 test results review is scheduled for November 9

Doug Bailey

• Ran ARM benchmarks for a special Custom Engineering drop. The results were sent to the customer on Friday
• Working on a fix for a 6.2.1 based customer issue. Have a good handle on what needs to be done. Solution is to add a new flag in thread control block. Expect to have it coded for mid week and regression at end of week. This does not affect post 6.2.1 releases.(PR 51782)
• Working on an issue that Kevin was seeing in graphics lab on Trinity 2. Looks like it could be a generic problem with way memory maps are working. Problem with CPUs with multiple cache colors resulting in conflicts. Have a good idea and submitted a proposed change to Brian. Should be resolved this week. (PR 51940)
• On Friday, worked on a qnet problem with remote spawn. It was reported by a customer on a 6.2.1 based release. Alex did some investiagaion and believes it is a resource leak. (PR 21182)\
• Fixed a Trinity 2 problem with Adaptive Partitioning on ARM.
• Attended FAE meeting last week

Mosa Yeung

• Working on benchmarking interrupt latency on Trinity 2. Read through some documentation for generating interrupts. Robert provided some sample code. Will be working on it this week and expect to finish it this week.

Angelina Gomes

• Working on Adaptvive Partitioning guide for Trinity 2. Hoping to send a version to Marian, Mike, and Attilla for review on Thursday.
• Will talk to Sunil about updates to the Trinity 2 Dinkum documentation.

Marian Kos

• Working on general code cleanup for test cases for Memory Partitioning in Trinity 2. Making sure that any test failures are not problems with the test cases.
• The MP tests run for about 6 hours. All requirements are covered. Need to determine the percent of code coverage.
• Filed a bug report on Tuesday for a Trinity 2 problem with increasing number of threads. Looks like Shiv had already fixed it. Don't see the failure in the latest code. (PR 52206)
• Had planned to put the test cases into svn but need a CI first. Expect to have this done this week.

Chris Travis

• Swapped in new terminal server in the labs. No problems were seen with it on the weekend.
• Tested a customer PSP last week. Results good.
• Have another customer PSP to run tonight on the OMAP.
• For M2, saw some issues with building test utilities with some of the POSIX changes that were done for large page support. Colin has looked at one of the issues but it's not completely fixed yet. Colin is going to get Sunil to check it.
• There are some other Trintiy 2 issues with header file changes seen with mtree, dtmgr, and memory statistics for free memory not going down. Will request an updated proc build on Wednesday with the fixes for these in order to run it on the failed tests.
• M1 was posted on Foundary27 this morning. Will post M2 shortly.

Attilla Danko

• Working on async messaging for Trinity 2. Posted a safe fix for async messaging problem with SMP. Brian/Colin to review. (PR 42643) There's another issue with it losing messages on uni-processor as well which is not fixed and could require a complete rewrite to resolve.
• Spent some time reading mapping code for message transfer
• Small amount of time on power management

Shiv Nagarajan

• Looking at a Posix failure PR for Trinity 2. Need to talk to Thomas about how to resolve it.
• Looking at Pidin failure seen by a customer that is using M1. (PR 52308)
• Also a Kernel crash reported by the same customer. (PR 51824)
• Helped with a special M2 source drop on Friday for a customer.

Mike Kisel

• Last week fixed some build problems when checking in the memory partitioning code for M2.
• Spent week going thru issues raised during the memory partioning code inspect. Have one change out for review.
• Also resolving problems raised by Marian's memory partitioning testing.

Peter Mitsis

• Spent time fixing up problems with Trintiy 2 test cases that I wrote.
• Fixed a problem from a check-in that was done for IOFS code
• Working on PRs for fdisk and USB mass storage.
• Working on test cases for a fix for queuing signals on smp systems.

Linda Deng

• M2 testing. Raised a POSIX failure PR.
• Working on file system test plan for 6.4
• Need to get ok from legal on Posix compliance documentation
• Working on Variable page size performance tests. Talked to Stephane, he has some numbers as well.

Brian Stecher

• Working on fixing MIPS message passing issue with Trinity 2 wired tlb fix. It was a subtle problem. Completed a regress over the weekend which resulted in a couple of additional issues to investigate. Hopefully will have this completed today. Waiting on acknowledgement from the code inspect.
• Spent some time on the special customer drop for Trinity 2 M2.
• This week will finish MIPS transfer and clean up backlog of miscellaneous items that have been on hold since focussing on tlb fix.

Colin Burgess

• Had successul laser eye surgery on Friday.
• Last week finishing off last few items on Trinity 2 special customer drop. Completed the code for watchpoints. Completed for MIPS but the PPC changes are more difficult.
• This week will continue with making sure the Friday drop works.

Gordon Rice

• Working on 16 high priority fixes for a customer modified kernel. Most are in progress and should have majority done this week