Closed Bug 1378071 Opened 7 years ago Closed 3 years ago

Flash applications are slow

Categories

(Core Graveyard :: Plug-ins, defect, P5)

54 Branch
defect

Tracking

(Not tracked)

RESOLVED WONTFIX

People

(Reporter: sworddragon2, Unassigned)

Details

(Keywords: flashplayer, perf, regression)

User Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:54.0) Gecko/20100101 Firefox/54.0
Build ID: 20170616154447



Actual results:

For example on the game http://www.kongregate.com/games/kupo707/epic-battle-fantasy-4 I'm noticing performance issues (for example if the battle menu is redrawn a freeze of a few hundred milliseconds appears) while this was not the case some time before.


Expected results:

Additional information:

It could be the last time I have played the game (a few months before) I was still on the Flash Player version 11.2.202.x while I'm currently on version 26.x as the Flash Player for the NPAPI got a major update more or less recently on Linux. Also I'm not able anymore to reproduce bug #1312899 not matter if dom.ipc.plugins.asyncdrawing.enabled is set to false or true (and also the performance issue is the same no matter of its setting).
Component: Untriaged → Plug-ins
Keywords: flashplayer, perf
Product: Firefox → Core
As pointed out in bug #1212813 comment #39 I was able to reproduce bug #1312899 still on Windows and maybe the performance issue is just a Flash Player issue but I can't tell this for sure.
Keywords: regression
Priority: -- → P5
This is still an issue with Firefox 59.0.1 and the Flash Player 29.0.0.113. I noticed this issue also on other flash applications in the past and figured out that the performance impact does not come from drawing things but from processing mouse events. Mainly this happens in several situations (not all) if I'm making a mouseclick as the application visually freezes then for a few hundred milliseconds but this seems to be only an optical issue as the processing of the applications still always continued during the freeze. If I'm searching for input lag issues for the Flash Player I always find issues related to dom.ipc.plugins.asyncdrawing.enabled but as already mentioned nothing changes here at my end no matter of its setting.

However, the interesting part is that this issue was fully resolved/heavily mitigated at the time I started a Windows 10 VM with VirtualBox. I was wondering what a VM running on the host does actually change that also changes the behavior of Firefox/Flash on the host. One of my guesses was that due to the additional used CPU time from the VM maybe Firefox/Flash was restricted temporary to less cores causing a mitigation of this issue. But testing this by limiting Firefox/Flash to only 1 core did not confirm this and also additionally testing if disabling e10s does help turned to be out false. Another guess was that the IO operations from the VM influenced this somehow but this turned out to be false too on a test.

If I'm making a look at the terminal I'm seeing always 2 lines when I load a site that causes the plugin-container for a flash application to start in the form of:

[Parent 37806, Gecko_IOThread] WARNING: pipe error (126): Connection reset by peer: file /build/firefox-_TqoHc/firefox-59.0.1+build1/ipc/chromium/src/chrome/common/ipc_channel_posix.cc, line 353
[Parent 37806, Gecko_IOThread] WARNING: pipe error (116): Connection reset by peer: file /build/firefox-_TqoHc/firefox-59.0.1+build1/ipc/chromium/src/chrome/common/ipc_channel_posix.cc, line 353

No idea if this is related to this issue or if those lines are an issue at all. Also I'm really wondering what did actually change that this issue was heavily mitigated for a few minutes at the time the VM was started the first time.

Flash applications are still slow in Firefox 67. I understand Chrome bundles Flash but why do Flash applications run so smoothly in Chrome?

Status: UNCONFIRMED → RESOLVED
Closed: 3 years ago
Resolution: --- → WONTFIX
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.