Firefox 63 vs. sendmsg()

[Jan Beich]

Can someone help debug https://bugzilla.mozilla.org/show_bug.cgi?id=1475970#c7 ?
I'm not familar with socket code.

Steps to reproduce:
$ pkg install python27
$ hash git 2>/dev/null || pkg install mercurial
$ hg clone https://hg.mozilla.org/mozilla-unified firefox ||
git clone https://github.com/mozilla/gecko-dev firefox
$ cd firefox
$ hg update central || git checkout origin/master
$ fetch -qo- https://hg.mozilla.org/projects/nspr/raw-rev/776db96f834c | patch -Efsp1 -d nsprpub/
$ export CC=clang60 CXX=clang++60
$ ./mach bootstrap # select Firefox for Desktop
$ ./mach build
$ ./mach run

--
if ./mach bootstrap fails due to rust try updating rust package instead,
see https://clbin.com/n7tS9

[Jan Kokemüller]

Post by Jan Beich
Can someone help debug https://bugzilla.mozilla.org/show_bug.cgi?id=1475970#c7 ?
I'm not familar with socket code.

Could it be related to this bug[1]?
There is a demo program in another bug report[2] that still fails for me
(running a recent -current, r334337).

[1]: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=181741
[2]: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=215933

[Jan Beich]

Post by Jan Kokemüller

Post by Jan Beich
Can someone help debug https://bugzilla.mozilla.org/show_bug.cgi?id=1475970#c7 ?
I'm not familar with socket code.

Could it be related to this bug[1]?
There is a demo program in another bug report[2] that still fails for me
(running a recent -current, r334337).
[1]: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=181741
[2]: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=215933

Indeed. Firefox 63 works fine after applying the patch in bug 181741.
I wonder, if "packet loss" issue is also responsible for IPC instability
in Firefox < 63 and Chromium.

[Don Lewis]

Post by Jan Beich

Post by Jan Kokemüller

Post by Jan Beich
Can someone help debug https://bugzilla.mozilla.org/show_bug.cgi?id=1475970#c7 ?
I'm not familar with socket code.

Could it be related to this bug[1]?
There is a demo program in another bug report[2] that still fails for me
(running a recent -current, r334337).
[1]: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=181741
[2]: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=215933

Indeed. Firefox 63 works fine after applying the patch in bug 181741.
I wonder, if "packet loss" issue is also responsible for IPC instability
in Firefox < 63 and Chromium.

I've been having terrible stability problems with Firefox on my
11.2-STABLE amd64 machine since upgrading to version 61. See:
https://bugzilla.mozilla.org/show_bug.cgi?id=1476130
I applied this patch:
https://bugs.freebsd.org/bugzilla/attachment.cgi?id=168943&action=diff
after tweaking it a bit so that all chunks would apply. It's too soon
to tell for sure, but this fix appears promising.

[Jan Beich]

Post by Don Lewis

Post by Jan Beich

Post by Jan Kokemüller

Post by Jan Beich
Can someone help debug https://bugzilla.mozilla.org/show_bug.cgi?id=1475970#c7 ?
I'm not familar with socket code.

Could it be related to this bug[1]?
There is a demo program in another bug report[2] that still fails for me
(running a recent -current, r334337).
[1]: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=181741
[2]: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=215933

Indeed. Firefox 63 works fine after applying the patch in bug 181741.
I wonder, if "packet loss" issue is also responsible for IPC instability
in Firefox < 63 and Chromium.

I've been having terrible stability problems with Firefox on my
https://bugzilla.mozilla.org/show_bug.cgi?id=1476130
https://bugs.freebsd.org/bugzilla/attachment.cgi?id=168943&action=diff
after tweaking it a bit so that all chunks would apply. It's too soon
to tell for sure, but this fix appears promising.

Can you attach the rebased version to bug 181741? I think, we may want
to ask a few Chromium users to check if it has an impact on bug 212812.

For comparison, attached to this mail is my own (blind) rebase.

[Yuri]

Post by Jan Beich
Can you attach the rebased version to bug 181741? I think, we may want
to ask a few Chromium users to check if it has an impact on bug 212812.

Bug 212812 is a cross-OS bug. It also exists on Windows and Linux. It is
unlikely that any FreeBSD fix will cure the Chrome problem.


Yuri