June 2023 - openSUSE Factory Mozilla - openSUSE Mailing Lists

[Bug 1212101] Firefox 114.0 and 114.0.1 immediately crashes at startup.
by bugzilla_noreply＠suse.com 20 Jun '23

20 Jun '23

https://bugzilla.suse.com/show_bug.cgi?id=1212101 https://bugzilla.suse.com/show_bug.cgi?id=1212101#c62 Jiri Slaby <jslaby(a)suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo? CC| |jh(a)suse.com, | |mjambor(a)suse.com, | |rguenther(a)suse.com --- Comment #62 from Jiri Slaby <jslaby(a)suse.com> --- (In reply to Andreas Stieger from comment #56) > -ac_add_options --enable-lto > +#ac_add_options --enable-lto > > that did the trick. Guys, any idea how to debug this LTO issue? It looks like without any -m for gfx/skia/skia/src/core/SkM44.cpp, the final ELF still receives AVX instructions through immintrin.h. And that crashes on older CPUs like Nehalem, of course. Note that there are some objects where AVX instructions are expected (like gfx/skia/SkOpts_avx.o). So maybe the flags are inherited from that (or other) file by mistake when linking the final ELF? See also comment #53. -- You are receiving this mail because: You are the assignee for the bug.

1 0

[Bug 1212101] Firefox 114.0 and 114.0.1 immediately crashes at startup.
by bugzilla_noreply＠suse.com 20 Jun '23

20 Jun '23

https://bugzilla.suse.com/show_bug.cgi?id=1212101 https://bugzilla.suse.com/show_bug.cgi?id=1212101#c60 --- Comment #60 from Martin Sirringhaus <martin.sirringhaus(a)suse.com> --- Then my hunch was correct. On the one hand good, at least we now have a functioning binary again. On the other hand not so good, as upstream suspects a gcc-bug, which might be difficult to pinpoint. -- You are receiving this mail because: You are the assignee for the bug.

1 0

[Bug 1212101] Firefox 114.0 and 114.0.1 immediately crashes at startup.
by bugzilla_noreply＠suse.com 20 Jun '23

20 Jun '23

https://bugzilla.suse.com/show_bug.cgi?id=1212101 https://bugzilla.suse.com/show_bug.cgi?id=1212101#c58 Wolfgang Rosenauer <wolfgang(a)rosenauer.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|CONFIRMED |IN_PROGRESS --- Comment #58 from Wolfgang Rosenauer <wolfgang(a)rosenauer.org> --- The workaround has been submitted for Tumbleweed. I hope our gcc experts (or mozilla folks) find the underlying reason so we can eventually turn it on again. -- You are receiving this mail because: You are the assignee for the bug.

1 0

[Bug 1212101] Firefox 114.0 and 114.0.1 immediately crashes at startup.
by bugzilla_noreply＠suse.com 20 Jun '23

20 Jun '23

https://bugzilla.suse.com/show_bug.cgi?id=1212101 https://bugzilla.suse.com/show_bug.cgi?id=1212101#c57 --- Comment #57 from Gerhard Maier <Gerhard(a)maier-espino.de> --- MozillaFirefox-114.0.1-5.1 works again. For me, the problem is solved! -- You are receiving this mail because: You are the assignee for the bug.

1 0

[Bug 1212101] Firefox 114.0 and 114.0.1 immediately crashes at startup.
by bugzilla_noreply＠suse.com 19 Jun '23

19 Jun '23

https://bugzilla.suse.com/show_bug.cgi?id=1212101 https://bugzilla.suse.com/show_bug.cgi?id=1212101#c54 --- Comment #54 from Wolfgang Rosenauer <wolfgang(a)rosenauer.org> --- There is another build with a different build change: https://download.opensuse.org/repositories/mozilla/openSUSE_Tumbleweed/x86_… Please try and report if it changes anything. -- You are receiving this mail because: You are the assignee for the bug.

1 0

[Bug 1212101] Firefox 114.0 and 114.0.1 immediately crashes at startup.
by bugzilla_noreply＠suse.com 19 Jun '23

19 Jun '23

https://bugzilla.suse.com/show_bug.cgi?id=1212101 https://bugzilla.suse.com/show_bug.cgi?id=1212101#c53 --- Comment #53 from Martin Sirringhaus <martin.sirringhaus(a)suse.com> --- Created attachment 867672 --> https://bugzilla.suse.com/attachment.cgi?id=867672&action=edit save-temps Copied the gcc-call of OBS and added `-save-temps`: /usr/bin/g++ -save-temps -c -o Unified_cpp_gfx_skia5.o -flto -flifetime-dse=1 -I/home/abuild/rpmbuild/BUILD/obj/dist/stl_wrappers -I/home/abuild/rpmbuild/BUILD/obj/dist/system_wrappers -include /home/abuild/rpmbuild/BUILD/firefox-114.0.1/config/gcc_hidden.h -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=2 -fstack-protector-strong -DNDEBUG=1 -DTRIMMED=1 -DSKIA_IMPLEMENTATION=1 -DSK_PDF_USE_HARFBUZZ_SUBSETTING=1 -DMOZ_HAS_MOZGLUE -DMOZILLA_INTERNAL_API -DIMPL_LIBXUL -DSTATIC_EXPORTABLE_JS_API -I/home/abuild/rpmbuild/BUILD/firefox-114.0.1/gfx/skia -I/home/abuild/rpmbuild/BUILD/obj/gfx/skia -I/home/abuild/rpmbuild/BUILD/firefox-114.0.1/gfx/skia/skia -I/home/abuild/rpmbuild/BUILD/firefox-114.0.1/gfx/cairo/cairo/src -I/home/abuild/rpmbuild/BUILD/obj/dist/include -I/usr/include/nspr4 -I/usr/include/nss3 -I/usr/include/nspr4 -I/home/abuild/rpmbuild/BUILD/obj/dist/include/nss -DMOZILLA_CLIENT -include /home/abuild/rpmbuild/BUILD/obj/mozilla-config.h -fno-sized-deallocation -fno-aligned-new -O2 -Wall -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=3 -fstack-protector-strong -funwind-tables -fasynchronous-unwind-tables -fstack-clash-protection -Werror=return-type -g -fimplicit-constexpr -fno-exceptions -fPIC -fno-rtti -ffunction-sections -fdata-sections -fno-exceptions -fno-math-errno -pthread -pipe -g1 -freorder-blocks -O2 -fomit-frame-pointer -funwind-tables -Wall -Wempty-body -Wignored-qualifiers -Wpointer-arith -Wsign-compare -Wtype-limits -Wunreachable-code -Wno-invalid-offsetof -Wc++2a-compat -Wcomma-subscript -Wvolatile -Wno-error=deprecated -Wno-error=deprecated-enum-enum-conversion -Wduplicated-cond -Wimplicit-fallthrough -Wlogical-op -Wno-error=maybe-uninitialized -Wno-error=deprecated-declarations -Wno-error=array-bounds -Wno-error=free-nonheap-object -Wno-multistatement-macros -Wno-error=class-memaccess -Wformat -Wformat-overflow=2 -Wno-psabi -Wno-deprecated-declarations -Wno-overloaded-virtual -Wno-sign-compare -Wno-unreachable-code -Wno-unused-function -Wno-logical-op -Wno-maybe-uninitialized -I/usr/include/freetype2 -I/usr/include/freetype2 -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include -I/usr/include/harfbuzz -I/usr/include/freetype2 -pthread -I/usr/include/libmount -I/usr/include/blkid -I/usr/include/fribidi -I/usr/include/cairo -I/usr/include/libpng16 -I/usr/include/pixman-1 -fno-strict-aliasing -ffp-contract=off -MD -MP Unified_cpp_gfx_skia5.cpp All files are in the attachment. If the stack traces posted here are correct, it points to the default-constructor of VecStorage. The preprocessed file doesn't show more than that. -- You are receiving this mail because: You are the assignee for the bug.

1 0

[Bug 1212101] Firefox 114.0 and 114.0.1 immediately crashes at startup.
by bugzilla_noreply＠suse.com 19 Jun '23

19 Jun '23

https://bugzilla.suse.com/show_bug.cgi?id=1212101 https://bugzilla.suse.com/show_bug.cgi?id=1212101#c51 --- Comment #51 from Jiri Slaby <jslaby(a)suse.com> --- (In reply to Wolfgang Rosenauer from comment #49) > Since it's still quite some guesswork I have built Firefox for TW using > gcc12 (instead of 13). Easily reproducible with an old "CPU": qemu-kvm ... -cpu Nehalem Thread 1 "firefox" received signal SIGILL, Illegal instruction. (gdb) x/i $rip => 0x7ffff006a850: vshufps $0x0,%xmm0,%xmm0,%xmm0 (In reply to Martin Sirringhaus from comment #50) > Should be inside one of the `Unified_cpp_gfx_skia*`-files. In my local > build-dir, it's Unified_cpp_gfx_skia5.cpp, but build-flags are identical for > all of them, as far as I can see. And I also fail to see anything that could > specify -mavx directly or indirectly for those calls. Can you paste here a preprocessed SkM44.cpp (generated with -E instead of -c)? And the gcc command line? -- You are receiving this mail because: You are the assignee for the bug.

1 0

[Bug 1212101] Firefox 114.0 and 114.0.1 immediately crashes at startup.
by bugzilla_noreply＠suse.com 19 Jun '23

19 Jun '23

https://bugzilla.suse.com/show_bug.cgi?id=1212101 https://bugzilla.suse.com/show_bug.cgi?id=1212101#c50 --- Comment #50 from Martin Sirringhaus <martin.sirringhaus(a)suse.com> --- (In reply to Jiri Slaby from comment #48) > > Sure, I assume -march=native (or similar) is used somewhere (and maybe some > of the build hosts started exposing AVX and friends to the build VMs). > > But I failed to find where as the build scripts are non-standard and the > build log is not helping (e.g. there is no build line for SkM44.cpp -- it's > likely one of those Unified_cpp's). I hope Wolfgang knows better... Should be inside one of the `Unified_cpp_gfx_skia*`-files. In my local build-dir, it's Unified_cpp_gfx_skia5.cpp, but build-flags are identical for all of them, as far as I can see. And I also fail to see anything that could specify -mavx directly or indirectly for those calls. -- You are receiving this mail because: You are the assignee for the bug.

1 0

[Bug 1212101] Firefox 114.0 and 114.0.1 immediately crashes at startup.
by bugzilla_noreply＠suse.com 19 Jun '23

19 Jun '23

https://bugzilla.suse.com/show_bug.cgi?id=1212101 https://bugzilla.suse.com/show_bug.cgi?id=1212101#c49 --- Comment #49 from Wolfgang Rosenauer <wolfgang(a)rosenauer.org> --- Since it's still quite some guesswork I have built Firefox for TW using gcc12 (instead of 13). If someone wants to test it's MozillaFirefox-114.0.1-4.1.x86_64.rpm https://download.opensuse.org/repositories/mozilla/openSUSE_Tumbleweed/x86_… in the mozilla repository. -- You are receiving this mail because: You are the assignee for the bug.

1 0

[Bug 1212101] Firefox 114.0 and 114.0.1 immediately crashes at startup.
by bugzilla_noreply＠suse.com 19 Jun '23

19 Jun '23

https://bugzilla.suse.com/show_bug.cgi?id=1212101 https://bugzilla.suse.com/show_bug.cgi?id=1212101#c48 --- Comment #48 from Jiri Slaby <jslaby(a)suse.com> --- (In reply to Paul Tannington from comment #47) > Official builds on x86 are only built with SSE2, with higher SIMD levels > only explicitly enabled selectively in various files. I would encourage the > distro build maintainers to fix this on their end by not supplying AVX flags > to the build. Acceleration for higher levels of SIMD is still used in > Mozilla builds, but again, we only selectively enable that in various parts > of the moz.build where it is actually safe and properly gated by CPU > feature-level checks. Sure, I assume -march=native (or similar) is used somewhere (and maybe some of the build hosts started exposing AVX and friends to the build VMs). But I failed to find where as the build scripts are non-standard and the build log is not helping (e.g. there is no build line for SkM44.cpp -- it's likely one of those Unified_cpp's). I hope Wolfgang knows better... -- You are receiving this mail because: You are the assignee for the bug.

1 0