Remove the socket process
Categories
(Core :: Networking, task, P2)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox161 | --- | fixed |
People
(Reporter: leggert, Assigned: leggert)
References
(Blocks 2 open bugs)
Details
(Keywords: perf-alert, Whiteboard: [necko-triaged])
Attachments
(31 external files, 3 obsolete files)
| Assignee | ||
Updated•1 month ago
|
Updated•1 month ago
|
| Assignee | ||
Updated•1 month ago
|
| Assignee | ||
Comment 1•1 month ago
|
||
| Assignee | ||
Comment 2•1 month ago
|
||
| Assignee | ||
Comment 3•1 month ago
|
||
Covers observer forwarding, the netwerk/base callers and the SSLTokensCache
load and clear forwarding.
| Assignee | ||
Comment 4•1 month ago
|
||
Removes the connection manager, HTTP transaction, WebSocket connection,
alt-svc, AltService, activity distributor, throttle queue and data bridge
actors, stops creating nsHttpHandler in the socket process, and drops the
SocketToParent PerfStats metrics.
| Assignee | ||
Comment 5•1 month ago
|
||
| Assignee | ||
Comment 6•1 month ago
|
||
Removes the TRR service, native resolver override and proxy config lookup
actors, makes PDNSRequest messages one-directional, and drops the dead parent
branch of ChildDNSService::GetMyHostName.
| Assignee | ||
Comment 7•1 month ago
|
||
Removes the PSM actors, ipcclientcerts, PSMIPCUtils.h and the PSM IPC types
they used.
| Assignee | ||
Comment 8•1 month ago
|
||
Also collapses ProxyAutoConfigBase into ProxyAutoConfig.
| Assignee | ||
Comment 9•1 month ago
|
||
Removes isLoadedBySocketProcess and resolvedInSocketProcess, forwards only the
prefs the socket process reads, and stops nsIOService implementing
nsIObserverService.
| Assignee | ||
Comment 10•1 month ago
|
||
| Assignee | ||
Comment 11•1 month ago
|
||
| Assignee | ||
Comment 12•1 month ago
|
||
| Assignee | ||
Comment 13•1 month ago
|
||
| Assignee | ||
Comment 14•1 month ago
|
||
Makes HttpActivityArgs a plain C++ type and removes the nsHttpTransaction
observer.
| Assignee | ||
Comment 15•1 month ago
|
||
| Assignee | ||
Comment 16•1 month ago
|
||
| Assignee | ||
Comment 17•1 month ago
|
||
| Assignee | ||
Comment 18•1 month ago
|
||
| Assignee | ||
Comment 19•1 month ago
|
||
| Assignee | ||
Comment 20•1 month ago
|
||
| Assignee | ||
Comment 21•1 month ago
|
||
| Assignee | ||
Comment 22•1 month ago
|
||
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
| Assignee | ||
Comment 23•1 month ago
|
||
Comment 24•1 month ago
|
||
:lars - you may already know this, but while looking into another bug, I found that the mochitest and xpcshell suites run with network.process.enabled=false and media.peerconnection.mtransport_process=false (from ensure_spi_disabled_on_all_but_spi).
It looks like the new test sets the pref itself, but the shipped socket process configuration is effectively only exercised by wpt and mochitest-media-spi, so a regression in what still runs there might not be caught.
(NI - just so you don't miss the comment.)
| Assignee | ||
Updated•1 month ago
|
Updated•1 month ago
|
Updated•1 month ago
|
| Assignee | ||
Comment 25•1 month ago
|
||
| Assignee | ||
Comment 26•1 month ago
|
||
| Assignee | ||
Comment 27•1 month ago
|
||
| Assignee | ||
Comment 28•1 month ago
|
||
| Assignee | ||
Comment 29•1 month ago
|
||
| Assignee | ||
Comment 30•1 month ago
|
||
| Assignee | ||
Comment 31•1 month ago
|
||
| Assignee | ||
Comment 32•1 month ago
|
||
| Assignee | ||
Comment 33•1 month ago
|
||
| Assignee | ||
Comment 34•1 month ago
|
||
| Assignee | ||
Comment 35•1 month ago
|
||
| Assignee | ||
Updated•1 month ago
|
| Assignee | ||
Comment 36•1 month ago
|
||
Thanks Bob!
- D333260 makes WebRTC media transport always run in the content process.
- D333262 removes the socketprocess test variant, which includes mochitest-media-spi.
- D333263 removes ensure_spi_disabled_on_all_but_spi, plus network.process.enabled and media.peerconnection.mtransport_process.
- D331034 is now abandoned with the latest changes
Updated•26 days ago
|
Updated•18 days ago
|
Comment 37•17 days ago
|
||
Comment 38•17 days ago
|
||
| bugherder | ||
https://hg.mozilla.org/mozilla-central/rev/20d66feb77d9
https://hg.mozilla.org/mozilla-central/rev/881ebb71fc68
https://hg.mozilla.org/mozilla-central/rev/9186fd0b7cad
https://hg.mozilla.org/mozilla-central/rev/ab24ed9e9162
https://hg.mozilla.org/mozilla-central/rev/3f5ff2880265
https://hg.mozilla.org/mozilla-central/rev/6824cfb1e200
https://hg.mozilla.org/mozilla-central/rev/e5d7fea34bcf
https://hg.mozilla.org/mozilla-central/rev/f5a36c6b6759
https://hg.mozilla.org/mozilla-central/rev/343123327402
https://hg.mozilla.org/mozilla-central/rev/72e8657c5b7a
https://hg.mozilla.org/mozilla-central/rev/b06cad832e75
https://hg.mozilla.org/mozilla-central/rev/9be1f7cd140e
https://hg.mozilla.org/mozilla-central/rev/eb4c4484c6a4
https://hg.mozilla.org/mozilla-central/rev/9471c2312e4c
https://hg.mozilla.org/mozilla-central/rev/a35b5bca693e
https://hg.mozilla.org/mozilla-central/rev/caca0d950a92
https://hg.mozilla.org/mozilla-central/rev/b53083ac7057
https://hg.mozilla.org/mozilla-central/rev/a305eb99ca83
https://hg.mozilla.org/mozilla-central/rev/b88ede8e89c5
https://hg.mozilla.org/mozilla-central/rev/c63200393e11
https://hg.mozilla.org/mozilla-central/rev/e124eaad5796
https://hg.mozilla.org/mozilla-central/rev/930746805f84
https://hg.mozilla.org/mozilla-central/rev/a9278b676ed7
https://hg.mozilla.org/mozilla-central/rev/02d1bdc452ee
https://hg.mozilla.org/mozilla-central/rev/d30f0b8928a0
https://hg.mozilla.org/mozilla-central/rev/b49188db8ce9
https://hg.mozilla.org/mozilla-central/rev/6067c2f1c5ea
https://hg.mozilla.org/mozilla-central/rev/da75c3698a30
https://hg.mozilla.org/mozilla-central/rev/a68446036367
https://hg.mozilla.org/mozilla-central/rev/018c55895d48
https://hg.mozilla.org/mozilla-central/rev/2c032e405d58
Comment 39•11 days ago
|
||
Perfherder has detected a build_metrics performance change from push 2c032e405d58954cd819179bfac6466b7f8034ed.
No action is required from the author; this comment is provided for informational purposes only.
| Improvement | Test | Platform | Options | Absolute values [old vs new] |
|---|---|---|---|---|
| 0.21% | installer size | osx-aarch64-nightlyasrelease | nightly nightly-as-release | 123,179,373.67 -> 122,923,331.33 |
| 0.15% | installer size | win64-nightlyasrelease | nightly nightly-as-release | 144,654,988.92 -> 144,442,087.83 |
| 0.12% | installer size | windows2012-64-shippable | nightly | 146,934,624.17 -> 146,756,168.58 |
Need Help or Information?
If you have any questions, please reach out to aesanu@mozilla.com. Alternatively, you can find help on Slack by joining #perf-help, and on Matrix you can find help by joining #perftest.
Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests.
Updated•6 days ago
|
Description
•