Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

MemShrink Triage (12 July 2017)

9 views
Skip to first unread message

Eric Rahm

unread,
Jul 12, 2017, 7:30:51 PM7/12/17
to dev-m...@lists.mozilla.org

MemShrink triage: 2017-07-12

Triage URL: http://mzl.la/1yYeaGL

Agenda

Note: We're going to just triage via email this week.

Bug List

Vote for:

  • P1 High importance, will follow up periodically
  • P2 Important issue, possibly already being worked on
  • P3 Real issue, someone will get to it if they have time
  • moreinfo We need logs, or clarifying information
  • invalid Misclassified, remove the MemShrink tag

17 bugs to triage

  • 1357598 - Firefox :: General - Poor performance and high memory usage in Firefox after waking up computer from sleep

    Votes:

    erahm, what do you think?

  • 1362767 - Core :: Untriaged - Firefox doesn't release memory after closing tabs

    Votes:

  • 1367999 - Core :: General - Got leak messages when close firefox

    Votes:

  • 1372092 - Firefox :: Untriaged - Long running Firefox eat all memory

    Votes:

  • 1373089 - Core :: JavaScript: Internationalization API - Use Sets rather than objects to store available locales in Intl API

    Votes:

  • 1373340 - Core :: IPC - Intermittent LeakSanitizer | leak at Malloc, nsTArray_base, AppendElements, IPC::ParamTraits

    Votes:

    mccr8, what do you think?

  • 1373418 - Toolkit :: WebExtensions: General - Reduce the amount of compartment overhead in the WebExtension framework

    Votes:

  • 1373825 - Core :: JavaScript: GC - Nightly memory use running away over time

    Votes:

    erahm, mccr8, what do you think?

  • 1373895 - Firefox :: Activity Streams: General - Activity Stream leaks a ton of memory

    Votes:

  • 1374503 - Core :: General - 7+ GB of heap-unclassified on nightly

    Votes:

  • 1376013 - Firefox :: Untriaged - Using Firefox for a long time causes it to lag, use over 2.5GB of RAM, and freeze.

    Votes:

  • 1376105 - Toolkit :: Async Tooling - 2.57 - 4.76% Explicit Memory / JS / Resident Memory (linux64, windows10-64-vm, windows7-32-vm) regression on push 6d121d4c2033db893f0c21951f761822d87df960 (Fri Jun 23 2017)

    Votes:

    erahm, what do you think?

  • 1376883 - Core :: CSS Parsing and Computation - stylo: Drop thread pool stack size to 100k

    Votes:

  • 1377999 - Core :: DOM - Try out an arena based allocator for DOM nodes again

    Votes:

    erahm, what do you think?

  • 1378526 - Core :: CSS Parsing and Computation - Measure Stylo memory usage using AWSY tests

    Votes:

  • 1379043 - Firefox :: Untriaged - HIGH RAM/DISK USAGE

    Votes:

  • 1379132 - Firefox :: Untriaged - very high memory usage (2GB after 4 hours usage)

    Votes:

Eric Rahm

unread,
Jul 12, 2017, 9:54:03 PM7/12/17
to dev-m...@lists.mozilla.org
On Wed, Jul 12, 2017 at 4:30 PM, Eric Rahm <er...@mozilla.com> wrote:

MemShrink triage: 2017-07-12

Triage URL: http://mzl.la/1yYeaGL

Agenda

Note: We're going to just triage via email this week.

Bug List

Vote for:

  • P1 High importance, will follow up periodically
  • P2 Important issue, possibly already being worked on
  • P3 Real issue, someone will get to it if they have time
  • moreinfo We need logs, or clarifying information
  • invalid Misclassified, remove the MemShrink tag

17 bugs to triage

  • 1357598 - Firefox :: General - Poor performance and high memory usage in Firefox after waking up computer from sleep

    Votes:

Close as incomplete, no feedback from reporter in 4 weeks.

  • erahm, what do you think?

  • 1362767 - Core :: Untriaged - Firefox doesn't release memory after closing tabs

    Votes:

ni? to clarify what the expected behavior is.

  • 1367999 - Core :: General - Got leak messages when close firefox

    Votes:


P2, maybe P1? Taipei folks are seeing shutdown leaks of google.com. mccr8, ting's on it but maybe you can take a look as well?
 

  • 1372092 - Firefox :: Untriaged - Long running Firefox eat all memory

    Votes:

 ni? For testing with sane prefs and add-ons disabled.

  • 1373089 - Core :: JavaScript: Internationalization API - Use Sets rather than objects to store available locales in Intl API

    Votes:

P2, has a patch, ideally a nice savings (48k?) per window.

  • 1373340 - Core :: IPC - Intermittent LeakSanitizer | leak at Malloc, nsTArray_base, AppendElements, IPC::ParamTraits

    Votes:

Unsure, I could see it being P1, P2, or P3. mcc8, wdyt?

  • mccr8, what do you think?

  • 1373418 - Toolkit :: WebExtensions: General - Reduce the amount of compartment overhead in the WebExtension framework

    Votes:

 ni kmag for more details on impact, I'm leaning P1.

  • 1373825 - Core :: JavaScript: GC - Nightly memory use running away over time

    Votes:

    erahm, mccr8, what do you think?

 Still waiting on Jon, tempted to P1 considering this is on dropbox.

  • 1373895 - Firefox :: Activity Streams: General - Activity Stream leaks a ton of memory

    Votes:

P1. It should definitely block actively streams from be pref'd on.

  • 1374503 - Core :: General - 7+ GB of heap-unclassified on nightly

    Votes:

 P1. 7 GB of heap-unclassified after resuming laptop.

  • 1376013 - Firefox :: Untriaged - Using Firefox for a long time causes it to lag, use over 2.5GB of RAM, and freeze.

    Votes:

 ni? for testing w/ add-ons disabled.

  • 1376105 - Toolkit :: Async Tooling - 2.57 - 4.76% Explicit Memory / JS / Resident Memory (linux64, windows10-64-vm, windows7-32-vm) regression on push 6d121d4c2033db893f0c21951f761822d87df960 (Fri Jun 23 2017)

    Votes:

    erahm, what do you think?

This is a regression from switching to native promises from promise.jsm. It's a pretty big regression, I'm inclined to P1 but would like more feedback.

  • 1376883 - Core :: CSS Parsing and Computation - stylo: Drop thread pool stack size to 100k

    Votes:

P1. This is one of the largest stylo regressions.

  • 1377999 - Core :: DOM - Try out an arena based allocator for DOM nodes again

    Votes:

    erahm, what do you think?

This is really a perf thing, I think maybe there's a concern it could increase memory but in the long run hopefully it would decrease fragmentation. We need to see some test data to get a feel for what kind of tradeoffs we'll see. I suggest deferring until Ehsan is back.

  • 1378526 - Core :: CSS Parsing and Computation - Measure Stylo memory usage using AWSY tests

    Votes:

P1, we really want to measure this sooner than later. It sounds like this is going to just be some build configuration stuff.

  • 1379043 - Firefox :: Untriaged - HIGH RAM/DISK USAGE

    Votes:

P2? It's high heap unclassified on facebook, but not sure about add-ons etc. Maybe just ni? for now.

  • 1379132 - Firefox :: Untriaged - very high memory usage (2GB after 4 hours usage)

    Votes:

ni?

Andrew McCreight

unread,
Jul 14, 2017, 7:21:55 PM7/14/17
to dev-m...@lists.mozilla.org
On Wed, Jul 12, 2017 at 4:30 PM, Eric Rahm <er...@mozilla.com> wrote:

MemShrink triage: 2017-07-12

Triage URL: http://mzl.la/1yYeaGL

Agenda

Note: We're going to just triage via email this week.

Bug List

Vote for:

  • P1 High importance, will follow up periodically
  • P2 Important issue, possibly already being worked on
  • P3 Real issue, someone will get to it if they have time
  • moreinfo We need logs, or clarifying information
  • invalid Misclassified, remove the MemShrink tag

17 bugs to triage

  • 1357598 - Firefox :: General - Poor performance and high memory usage in Firefox after waking up computer from sleep

    Votes:

  • erahm, what do you think?

  • 1362767 - Core :: Untriaged - Firefox doesn't release memory after closing tabs

    Votes:

  • 1367999 - Core :: General - Got leak messages when close firefox


P2 or P3? This is a long-standing issue, which isn't great, but I think it doesn't happen if you close the tab first before shutting down, so I don't think it is a leak we need to worry much about.

  • Votes:

  • 1372092 - Firefox :: Untriaged - Long running Firefox eat all memory


    Sounds bad, but maybe a bad addon or site?

    • Votes:

    • 1373089 - Core :: JavaScript: Internationalization API - Use Sets rather than objects to store available locales in Intl API

      Votes:

    • 1373340 - Core :: IPC - Intermittent LeakSanitizer | leak at Malloc, nsTArray_base, AppendElements, IPC::ParamTraits

      Votes:

      mccr8, what do you think?


    Looks like a fluke. I think it can be P3.

    • 1373418 - Toolkit :: WebExtensions: General - Reduce the amount of compartment overhead in the WebExtension framework


    P1 maybe, given that we're going to turn on WebExtensions for all Firefox users? Hopefully JSM merging will fix this.

    • Votes:

    • 1373825 - Core :: JavaScript: GC - Nightly memory use running away over time


      Sounds bad, but I don't know if we'll be able to do anything without STR. Maybe one of the other GC issues will fix this.

        • Votes:

          erahm, mccr8, what do you think?

        • 1373895 - Firefox :: Activity Streams: General - Activity Stream leaks a ton of memory


        Sounds like this can get marked WFM.

        • Votes:

        • 1374503 - Core :: General - 7+ GB of heap-unclassified on nightly


          This sounds vaguely similar to bug 1349991. We really need some better memory reporting for IPC. I filed bug 1381173for that.

          • Votes:

          • 1376013 - Firefox :: Untriaged - Using Firefox for a long time causes it to lag, use over 2.5GB of RAM, and freeze.

            Votes:

          • 1376105 - Toolkit :: Async Tooling - 2.57 - 4.76% Explicit Memory / JS / Resident Memory (linux64, windows10-64-vm, windows7-32-vm) regression on push 6d121d4c2033db893f0c21951f761822d87df960 (Fri Jun 23 2017)


          P1 maybe? That is a large regression, and we should at least understand what is happening. It is a very large change so maybe we can't really avoid it.

            • Votes:

              erahm, what do you think?

            • 1376883 - Core :: CSS Parsing and Computation - stylo: Drop thread pool stack size to 100k

              Votes:

            • 1377999 - Core :: DOM - Try out an arena based allocator for DOM nodes again

              Votes:

            I don't know if this will really help our memory usage. P2 maybe? I feel like mostly we don't want to regress it.
             

              • erahm, what do you think?

              • 1378526 - Core :: CSS Parsing and Computation - Measure Stylo memory usage using AWSY tests

               
              dev-memory mailing list
              dev-m...@lists.mozilla.org
              https://lists.mozilla.org/listinfo/dev-memory

              0 new messages