TPEPerformance/Meetings/2017-02-13: Difference between revisions
< TPEPerformance | Meetings
Jump to navigation
Jump to search
(2017-02-13) |
(/* Project Status) |
||
Line 7: | Line 7: | ||
= Project Status = | = Project Status = | ||
== Web Painting - Improve DisplayList Memory Locality == | == Quantum Flow == | ||
== Remove Sync IPC Everywhere == | === Web Painting - Improve DisplayList Memory Locality === | ||
== Quantum Telemetry == | * WebRendering had a meeting in Toronto, will sync up with them this week. | ||
== TaskTracer integration to Gecko Profiler == | === Remove Sync IPC Everywhere === | ||
== Crash Client (Breakpad) Improvement == | * Will start to fix PScreenManager | ||
== Windows Dynamic Analysis Improvement == | * Will start to implement async returning MozPromise | ||
=== Quantum Telemetry === | |||
* WIP patch, verifying different event type timings | |||
* Need to clarify the exact telemetry we want to record | |||
=== TaskTracer integration to Gecko Profiler === | |||
* Fixed a key issue in patch, reviewing | |||
* Will discuss with Markus when to enable on central | |||
* Will run talos test to measure performance overhead | |||
== Uptime == | |||
=== Crash Client (Breakpad) Improvement === | |||
* Ted forked breakpad client to m-c, need to redo heap-in-minidump patches | |||
* Consider to enable heap-in-minidump by default | |||
* Thread-name patch got r+, still need Ted's review | |||
* Need to file socorro bug for thread-name | |||
* JS Stack - gdb unwinder is not really helpful | |||
* Risk: evaluating crashpad or reinvent crashpad, what's the advantage? | |||
=== Windows Dynamic Analysis Improvement === | |||
* Experimenting adding a new automation build type | |||
* Risk: only work on TaskCluster, will depends on the date we officially move to TC. | |||
= Previous Week = | = Previous Week = |
Revision as of 03:36, 13 February 2017
« previous week | index | next week »
Project Status
Quantum Flow
Web Painting - Improve DisplayList Memory Locality
- WebRendering had a meeting in Toronto, will sync up with them this week.
Remove Sync IPC Everywhere
- Will start to fix PScreenManager
- Will start to implement async returning MozPromise
Quantum Telemetry
- WIP patch, verifying different event type timings
- Need to clarify the exact telemetry we want to record
TaskTracer integration to Gecko Profiler
- Fixed a key issue in patch, reviewing
- Will discuss with Markus when to enable on central
- Will run talos test to measure performance overhead
Uptime
Crash Client (Breakpad) Improvement
- Ted forked breakpad client to m-c, need to redo heap-in-minidump patches
- Consider to enable heap-in-minidump by default
- Thread-name patch got r+, still need Ted's review
- Need to file socorro bug for thread-name
- JS Stack - gdb unwinder is not really helpful
- Risk: evaluating crashpad or reinvent crashpad, what's the advantage?
Windows Dynamic Analysis Improvement
- Experimenting adding a new automation build type
- Risk: only work on TaskCluster, will depends on the date we officially move to TC.
Previous Week
ting
- Triaged 0205 nightly/aurora crashes
- bug 1322554 NEW:: Startup crashes in BaseThreadInitThunk since 2016-12-03
- bug 1322554 am installing win7 32 bit
- bug 1322554 checked few more crash dumps for suspicious modules, but no luck
- bug 1322554 commented what i found for robosizer
- bug 1322554 it seems most crash are from 32bit win7, am double confirming and trying to get a 32bit win7 to install
- bug 1322554 tried to build easyhook dbg version, but vs2012 express is lack of atlbase.h
- bug 1335080 sorted out how robosizer causes the crash, but doesn't seem the main reason of bug 1322554
- bug 1334027 RESOLVED::FIXED Add unloaded modules and process/thread data to minidumps
- bug 1334027 pushed to autoland
- bug 1333003 ASSIGNED:: Add a test platform for win64 ASan
- Start to work on bug 1333003
- bug 1335080 NEW:: RoboSizer software causes startup crashes on Windows 10
- bug 1335080 sorted out how robosizer causes the crash, but doesn't seem the main reason of bug 1322554
cervantes
- Investigating getting JS stack in crash client
- Joined discussion about IPC fuzzer design
- bug 1328569 RESOLVED::FIXED [App Verifier] UnmapViewOfFile() called with null lpBaseAddress
- Requested uplift for bug 1328569
- bug 1024669 NEW:: Send the name of threads
- Requested review for bug 1024669
- Revising the patch and adding gtest for bug 1024669
- Updated the patch and test cases for bug 1024669
- bug 1320134 NEW:: Crash in xul.dll@0x4d768 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0x4d739 | xul.dll@0...
- Taking bug 1320134 - start working on a diagnostic patch for TLS slot exhaustion on Windows
thinker
- bug 1329929 NEW:: Fix memory leaking for TaskTracer
- Send review for bug 1329929 tasktracer memory leaking.
- bug 1313753 Invalid bug or security bug
- Checkin bug 1313753
- bug 1310611 NEW:: Erroneous transition when scale-transforming text with ellipsis
- bug 1310611 - ellipse's color is changed during animation.
kanru
- bug 1194751 NEW:: PScreenManager should not use sync messages
- Working on bug 1194751
- bug 1335122 Invalid bug or security bug
- Looking at bug 1335122