Closed Bug 1376842 Opened 7 years ago Closed 7 years ago

gdb support code still using asm.js SIGSEGV handler

Categories

(Core :: JavaScript Engine, enhancement)

enhancement
Not set
normal

Tracking

()

RESOLVED FIXED
mozilla56
Tracking Status
firefox56 --- fixed

People

(Reporter: sfink, Assigned: sfink)

Details

Attachments

(1 file)

It confused me because it was "just working" for so long.
r? tromey for the long(...) casts, which I'm totally unsure of. I was trying to get the spidermonkey stack unwinder working under rr (I don't know if that's possible or not), and this *seemed* to resolve one error? Maybe? But I'm not sure, because something strange is happening -- it complains about the per_tls_context error not existing, but it seems to be initializing more than once, and the other initializations are ok and make it past there -- but if I record the self.d that failed in a global variable and later check it, per_tls_context is there. /me is totally confused, which lowers my confidence of these changes.

The signal handler stuff seems to work fine, fwiw.
Attachment #8881832 - Flags: review?(ttromey)
Comment on attachment 8881832 [details] [diff] [review]
Update gdb support code to new wasm SIGSEGV handler, and some other updates

Review of attachment 8881832 [details] [diff] [review]:
-----------------------------------------------------------------

I think those casts to long should not strictly be necessary, but I'm not completely sure, and anyway they won't hurt.

> I was trying to get the spidermonkey stack unwinder working under rr (I don't know if that's possible or not)

I think it should be possible.  I think I must have tried it at some point in the past, because the unwinder
is why I added the libthread_db stuff to rr.  But, I haven't tried it since then.

> it complains about the per_tls_context error not existing

No theories here but I wonder if the lookup might not be happening with the language set to C++.
I feel like I saw this once before.  Maybe this is why I ended up doing that __getattr__ hack, hard to recall.
Attachment #8881832 - Flags: review?(ttromey) → review+
Pushed by sfink@mozilla.com:
https://hg.mozilla.org/integration/mozilla-inbound/rev/157445a8c76b
Update gdb support code to new wasm SIGSEGV handler, and some other updates, r=tromey
https://hg.mozilla.org/mozilla-central/rev/157445a8c76b
Status: ASSIGNED → RESOLVED
Closed: 7 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla56
You need to log in before you can comment on or make changes to this bug.