firefox is used in openqa tests for ppc64le and ppc64 since Fedora-27-20171001.n.2 there is segmentation fault when running firefox I reproduced the problem manually to get some traces (see attachments) I installed a ppc64le VM with the Fedora-Server-dvd-ppc64le-27-20171003.n.0.iso Then I start firefox thru a ssh -X session firefox start but then scratch with a panel "Gah. Your tab just crashed." and the window disappears. Note that firefox was running correctly for last time with Fedora-27-20170929.n.0
Created attachment 1334363 [details] msg at screen
Created attachment 1334364 [details] dmesg
Created attachment 1334365 [details] journalctl
Created attachment 1334366 [details] coredump thru gdb
Created attachment 1334367 [details] firefox with strace
for info, problem is still there in firefox-57.0-0.4
most probably a duplicate of previous bug #1486737 identified before for fc26 update tests.
*** Bug 1486737 has been marked as a duplicate of this bug. ***
Created attachment 1437274 [details] start of ppc64le patch which avoid the crash I started to test this patch sometimes ago but get blocked because of different other problems that crash firefox before I reach this code. The problem blocking my test is currently the bug #1578429 So my patch could not be completely satisfactory as is working ok in normal mode but I still have error when I put firefox in debug mode.
*** Bug 1578429 has been marked as a duplicate of this bug. ***
I added your patch to firefox-60.0-6.* builds at https://koji.fedoraproject.org/koji/packageinfo?packageID=37
firefox-60.0-6 seems working ok so far for ppc64le. I update my openqa environment to have next tests running on rawhide using this new firefox and I will update this bug if tests are ok. Also, jemalloc has been disable when building for powerpc, is there a plan to make jemalloc working for powerpc ?
(In reply to Menanteau Guy from comment #12) > Also, jemalloc has been disable when building for powerpc, is there a plan > to make jemalloc working for powerpc ? I don't have such intention.
Do you mind to file upstream (at bugzilla.mozilla.org) bug or should I do so? Will be good to have this patch upstreamed.
Done, see https://bugzilla.mozilla.org/show_bug.cgi?id=1462566 Please, feel free to add any comments. Thanks