1576168 - Update audio_thread_priority to 0.19.1

Status: RESOLVED FIXED

(Core :: Audio/Video: cubeb, enhancement, P3)

Description

[Paul Adenot (:padenot)]

This allows catching SIGXCPU.

Comment 1

[Paul Adenot (:padenot)]

Attachment: Bug 1576168 - Update audio_thread_priority to 0.19.1. r?pehrsons

This changes the hard-limit of RLIMIT_RTTIME to be the maximum available
(200ms on my system), and keep the soft limit to the same number.

Having different numbers allow catching SIGXCPU before getting SIGKILL.

Comment 2

[Paul Adenot (:padenot)]

Attachment: Bug 1576168 - mach vendor rust. r?pehrsons

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/4e8fc76cd529
Update audio_thread_priority to 0.19.0. r=pehrsons
https://hg.mozilla.org/integration/autoland/rev/3ea6795200b1
mach vendor rust. r=pehrsons

Comment 4

[Cristina Coroiu [:ccoroiu]]

Backed out 2 changesets (bug 1576168) for build bustages

Backout: https://hg.mozilla.org/integration/autoland/rev/f277c58a8bcb52c8d18dbdeb63de50dd8a8973d1

Failure push:
https://treeherder.mozilla.org/#/jobs?repo=autoland&revision=3ea6795200b155f75de68931f12671a0b3b0c397

Failure log:
https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=263524487&repo=autoland&lineNumber=18565

[task 2019-08-26T16:44:28.638Z] 16:44:28 ERROR - error[E0432]: unresolved import mach_sys
[task 2019-08-26T16:44:28.638Z] 16:44:28 INFO - --> /builds/worker/workspace/build/src/third_party/rust/audio_thread_priority/src/rt_mach.rs:8:5
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - |
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - 8 | use mach_sys::*;
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - | ^^^^^^^^ help: a similar path exists: crate::mach_sys
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - |
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - = note: use statements changed in Rust 2018; read more at https://doc.rust-lang.org/edition-guide/rust-2018/module-system/path-clarity.html
[task 2019-08-26T16:44:28.639Z] 16:44:28 ERROR - error[E0412]: cannot find type thread_t in this scope
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - --> /builds/worker/workspace/build/src/third_party/rust/audio_thread_priority/src/rt_mach.rs:16:34
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - |
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - 16 | fn thread_policy_set(thread: thread_t,
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - | ^^^^^^^^
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - help: a type alias with a similar name exists
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - |
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - 16 | fn thread_policy_set(thread: pthread_t,
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - | ^^^^^^^^^
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - help: possible candidates are found in other modules, you can import them into scope
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - |
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - 5 | use crate::mach_sys::thread_t;
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - |
[task 2019-08-26T16:44:28.639Z] 16:44:28 INFO - 5 | use mach::mach_types::thread_t;
[task 2019-08-26T16:44:28.640Z] 16:44:28 INFO - |
[task 2019-08-26T16:44:28.640Z] 16:44:28 ERROR - error[E0412]: cannot find type thread_policy_flavor_t in this scope
[task 2019-08-26T16:44:28.640Z] 16:44:28 INFO - --> /builds/worker/workspace/build/src/third_party/rust/audio_thread_priority/src/rt_mach.rs:17:34
[task 2019-08-26T16:44:28.640Z] 16:44:28 INFO - |
[task 2019-08-26T16:44:28.640Z] 16:44:28 INFO - 17 | flavor: thread_policy_flavor_t,
[task 2019-08-26T16:44:28.640Z] 16:44:28 INFO - | ^^^^^^^^^^^^^^^^^^^^^^ not found in this scope
[task 2019-08-26T16:44:28.640Z] 16:44:28 INFO - help: possible candidate is found in another module, you can import it into scope
[task 2019-08-26T16:44:28.640Z] 16:44:28 INFO - |
[task 2019-08-26T16:44:28.640Z] 16:44:28 INFO - 5 | use crate::mach_sys::thread_policy_flavor_t;
[task 2019-08-26T16:44:28.640Z] 16:44:28 INFO - |
[task 2019-08-26T16:44:28.640Z] 16:44:28 ERROR - error[E0412]: cannot find type thread_policy_t in this scope

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/6e85d1ee673a
Update audio_thread_priority to 0.19.1. r=pehrsons
https://hg.mozilla.org/integration/autoland/rev/1f4a5ebd5c7b
mach vendor rust. r=pehrsons

Comment 6

[Cosmin Sabou [:CosminS]]

https://hg.mozilla.org/mozilla-central/rev/6e85d1ee673a
https://hg.mozilla.org/mozilla-central/rev/1f4a5ebd5c7b

Comment 7

[Eric Wheeler]

Hi Paul,

Based on the result of a bisect, this change of audio_thread_priority to 0.19.1 appears to have caused bug#1538435 which has been open for 3 years now. However, that bug has been open longer than this fix so it may not be the whole story.

I can confirm from bisection that one of these two changesets caused the issue I am presently having which presents the same as bug#1538435:

(quote from bug#1538435)

The first bad revision is:

changeset:   490166:1f4a5ebd5c7b
user:        Paul Adenot <[e-mail]>
date:        Tue Aug 27 08:00:43 2019 +0000
summary:     Bug 1576168 - mach vendor rust. r=pehrsons

And there is a related commit just before on the same day:

changeset:   490165:6e85d1ee673a
user:        Paul Adenot <[e-mail]>
date:        Tue Aug 27 08:00:36 2019 +0000
summary:     Bug 1576168 - Update audio_thread_priority to 0.19.1. r=pehrsons

Should this bug be reopened and have the other bug marked as a regression of this bugfix?

Any idea how we can fix both issues?

Please let me know if I can test anything because I still have my bisect tree available and an environment that triggers bug#1538435.

Thanks for your help!

Comment 8

[Eric Wheeler]

(Please see bug#1538435 because it has all the bisect and build details there to prevent duplicate content.)

Comment 9

[Eric Wheeler]

It seems my bisect test case wasn't testing what I thought it was: it detected the change from SIGKILL to SIGXCPU, but this isn't the right commit. Please disregard my comments above.