734685 - Support IPC Thread::SetName on Mac

Status: RESOLVED FIXED

(Core :: IPC, defect)

Description

[Benoit Girard (:BenWa)]

Attachment: patch

Useful for gdb, this code is an adaption from chromium/src/base/threading/platform_thread_posix.cc.

Comment 1

[Benoit Girard (:BenWa)]

Attachment: Screenshot (IPC thread 5+21 fixed by this patch)

Comment 2

[Josh Matthews [:jdm]]

Is this related to bug 720778 at all?

Comment 3

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Comment on attachment 604687 [details] [diff] [review]
patch

Did you check if this was fixed in upstream chromium?  I'd be happy to pull a change from there, too.

>+  std::string shortened_name = std::string(name).substr(0, kMaxNameLength);
>+  // pthread_setname() fails (harmlessly) in the sandbox, ignore when it does.
>+  // See http://crbug.com/47058
>+  dynamic_pthread_setname_np(shortened_name.c_str());

What's the ownership model for the pointer here?  Please doc.

r=me with ownership clarified (/ proven correct ;) ).

Comment 4

[Benoit Girard (:BenWa)]

I got the code from 'chromium/src/base/threading/platform_thread_posix.cc', wouldn't that be upstream? The change doesn't apply cleanly so I had to make a patch manually.

Comment 5

[Benoit Girard (:BenWa)]

I looked in bionic and it's implemented by copying name:
bionic/libc/bionic/pthread.c

So it must certainly be the same elsewhere. I'll add a comment when checking in.

Comment 6

[Benoit Girard (:BenWa)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/192785313581

Comment 7

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Pulling from upstream is good.  Would prefer you check the darwin docs, but r+ from upstream (hopefully) means someone else already checked docs.

Comment 8

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/192785313581