``` # emerge net-im/zoom -av These are the packages that would be merged, in order: Calculating dependencies... done! [ebuild R ] net-im/zoom-5.10.0.2450-r2::gentoo USE="bundled-qt -bundled-libjpeg-turbo -opencl -pulseaudio -wayland" 0 KiB ``` pulseaudio disabled... but fails to launch looking for `pactl`: ``` $ zoom sh: line 1: pactl: command not found Class App Lib Possible Culprit Flags resip::Connection 648 648 resip::Data 36 36 resip::DnsResult 1080 1080 resip::Headers 1 1 resip::MsgHeaderScanner 40 40 resip::SipMessage 5208 5208 resip::TransportSelector 864 864 resip::Tuple 128 128 resip::UdpTransport 1144 1144 resip::GenericIPAddress 28 28 zoom started. sh: line 1: pactl: command not found sh: line 1: pactl: command not found Class App Lib Possible Culprit Flags resip::Connection 648 648 resip::Data 36 36 resip::DnsResult 1080 1080 resip::Headers 1 1 resip::MsgHeaderScanner 40 40 resip::SipMessage 5208 5208 resip::TransportSelector 864 864 resip::Tuple 128 128 resip::UdpTransport 1144 1144 resip::GenericIPAddress 28 28 Class App Lib Possible Culprit Flags resip::Connection 648 648 resip::Data 36 36 resip::DnsResult 1080 1080 resip::Headers 1 1 resip::MsgHeaderScanner 40 40 resip::SipMessage 5208 5208 resip::TransportSelector 864 864 resip::Tuple 128 128 resip::UdpTransport 1144 1144 resip::GenericIPAddress 28 28 zoom started. zoom started. Client: Breakpad is using Single Client Mode! client fd = -1 sh: line 1: pactl: command not found sh: line 1: pactl: command not found sh: line 1: pactl: command not found Class App Lib Possible Culprit Flags resip::Connection 648 648 resip::Data 36 36 resip::DnsResult 1080 1080 resip::Headers 1 1 resip::MsgHeaderScanner 40 40 resip::SipMessage 5208 5208 resip::TransportSelector 864 864 resip::Tuple 128 128 resip::UdpTransport 1144 1144 resip::GenericIPAddress 28 28 zoom started. [CZPClientLogMgr::LogClientEnvironment] [MacAddr: 04:D9:F5:21:86:7C][client: Linux][OS: Gentoo Linux x64][Hardware: CPU Core:8 Frenquency:3.6 G Memory size:32030MB CPU Brand:AMD Ryzen 7 3700X 8-Core Processor GPU Brand:][Req ID: ] Linux Client Version is 5.10.0 (2450) QSG_RENDER_LOOP is XDG_CURRENT_DESKTOP = KDE; GDMSESSION = ; XDG_SESSION_TYPE = x11 Graphics Card Info:: 0a:00.0 VGA compatible controller: NVIDIA Corporation GA104 [GeForce RTX 3070] (rev a1) Zoom package arch is 64bit, runing OS arch is x86_64 AppIconMgr::systemDesktopName log Desktop Name: /usr/share/xsessions/plasma [0324/091407.973757:FATAL:gpu_data_manager_impl_private.cc(414)] GPU process isn't usable. Goodbye. sh: line 1: pactl: command not found sh: line 1: pactl: command not found [1] 28351 illegal hardware instruction zoom ``` Reproducible: Always Steps to Reproduce: 1. Install latest zoom with -pulseaudio 2. launch zoom from cli 3. zoom crashes with `sh: line 1: pactl: command not found` Actual Results: zoom crashes / fails to start looking for pactl Expected Results: zoom should start emerge --info ``` # emerge --info Portage 3.0.30 (python 3.9.12-final-0, default/linux/amd64/17.1/desktop/plasma, gcc-11.2.1, glibc-2.34-r10, 5.16.14-gentoo-x86_64 x86_64) ================================================================= System uname: Linux-5.16.14-gentoo-x86_64-x86_64-AMD_Ryzen_7_3700X_8-Core_Processor-with-glibc2.34 KiB Mem: 32798928 total, 21068384 free KiB Swap: 0 total, 0 free Head commit of repository gentoo: 24eea31e5e563c645b2227a7f25290f51fbdd3de Head commit of repository nightdragon_layman: ec6c2cc7e26fbccdc98f30416d0e8134a728a056 Head commit of repository steam-overlay: 263b24d7b2640268a5b689add162d1dbfb80f660 sh bash 5.1_p16 ld GNU ld (Gentoo 2.37_p1 p2) 2.37 app-misc/pax-utils: 1.3.3::gentoo app-shells/bash: 5.1_p16::gentoo dev-java/java-config: 2.3.1::gentoo dev-lang/perl: 5.34.1::gentoo dev-lang/python: 2.7.18_p15::gentoo, 3.9.12::gentoo, 3.10.4::gentoo dev-lang/rust: 1.59.0::gentoo dev-util/cmake: 3.22.3::gentoo dev-util/meson: 0.61.3::gentoo sys-apps/baselayout: 2.8::gentoo sys-apps/openrc: 0.44.10::gentoo sys-apps/sandbox: 2.29::gentoo sys-devel/autoconf: 2.13-r1::gentoo, 2.71-r1::gentoo sys-devel/automake: 1.16.5::gentoo sys-devel/binutils: 2.37_p1-r2::gentoo, 2.38-r1::gentoo sys-devel/binutils-config: 5.4.1::gentoo sys-devel/clang: 13.0.1::gentoo, 14.0.0::gentoo sys-devel/gcc: 11.2.1_p20220115::gentoo sys-devel/gcc-config: 2.5-r1::gentoo sys-devel/libtool: 2.4.6-r6::gentoo sys-devel/lld: 13.0.1::gentoo sys-devel/llvm: 13.0.1::gentoo, 14.0.0::gentoo sys-devel/make: 4.3::gentoo sys-kernel/linux-headers: 5.16::gentoo (virtual/os-headers) sys-libs/glibc: 2.34-r10::gentoo Repositories: gentoo location: /var/db/repos/gentoo sync-type: git sync-uri: https://anongit.gentoo.org/git/repo/gentoo.git priority: -1000 nightdragon_layman location: /var/db/repos/nightdragon_layman sync-type: git sync-uri: https://github.com/NightDragon1/nightdragon_layman.git masters: gentoo steam-overlay location: /usr/local/portage/steam-overlay sync-type: git sync-uri: https://github.com/anyc/steam-overlay.git masters: gentoo priority: 50 ACCEPT_KEYWORDS="amd64 ~amd64" ACCEPT_LICENSE="@FREE" CBUILD="x86_64-pc-linux-gnu" CFLAGS="-pipe -O2 -march=native" CHOST="x86_64-pc-linux-gnu" CONFIG_PROTECT="/etc /usr/share/config /usr/share/gnupg/qualified.txt /usr/share/maven-bin-3.8/conf /usr/share/themes/oxygen-gtk/gtk-2.0 /var/bind" CONFIG_PROTECT_MASK="/etc/ca-certificates.conf /etc/dconf /etc/env.d /etc/fonts/fonts.conf /etc/gconf /etc/gentoo-release /etc/revdep-rebuild /etc/sandbox.d /etc/terminfo" CXXFLAGS="-pipe -O2 -march=native" DISTDIR="/var/cache/distfiles" ENV_UNSET="CARGO_HOME DBUS_SESSION_BUS_ADDRESS DISPLAY GOBIN GOPATH PERL5LIB PERL5OPT PERLPREFIX PERL_CORE PERL_MB_OPT PERL_MM_OPT XAUTHORITY XDG_CACHE_HOME XDG_CONFIG_HOME XDG_DATA_HOME XDG_RUNTIME_DIR" FCFLAGS="-O2 -pipe" FEATURES="assume-digests binpkg-docompress binpkg-dostrip binpkg-logs binpkg-multi-instance buildpkg-live config-protect-if-modified distlocks ebuild-locks fixlafiles ipc-sandbox merge-sync multilib-strict network-sandbox news parallel-fetch pid-sandbox preserve-libs protect-owned qa-unresolved-soname-deps sandbox sfperms strict unknown-features-warn unmerge-logs unmerge-orphans userfetch userpriv usersandbox usersync xattr" FFLAGS="-O2 -pipe" GENTOO_MIRRORS="http://distfiles.gentoo.org" LANG="en_US.UTF8" LDFLAGS="-Wl,-O1 -Wl,--as-needed" MAKEOPTS="-j8" PKGDIR="/var/cache/binpkgs" PORTAGE_CONFIGROOT="/" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --omit-dir-times --compress --force --whole-file --delete --stats --human-readable --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages --exclude=/.git" PORTAGE_TMPDIR="/var/tmp" SHELL="/bin/bash" USE="X a52 aac aacs acl acpi activities alsa amd64 bdplus bluetooth branding bzip2 cairo cdda cddb cdr cli client crypt cups dbus declarative dga dri dts dvd dvdr elogind encode exif farstream firmware flac fortran gdbm gif gles2 gpg gpm gstreamer gtk gui harfbuzz hddtemp http-proxy iconv icu ipv6 java jpeg kde kwallet lcms libglvnd libnotify libtirpc libuv lm-sensors mad mng mount mp3 mp4 mpeg mtp multilib multimedia musicbrainz ncurses nls nocd normalize nptl odbc offensive ogg openal openexr opengl openmp opus osc pam pango pcre pda pdf perl pie pim plasma png policykit postgres postproc ppds python qml qt5 raw rdp readline recode ruby samba scanner sctp sdl seccomp semantic-desktop session smp sockets socks5 sox spell split-usr ssl startup-notification svg syslog taglib theora threads tiff truetype udev udisks unicode upower usb utils v4l vorbis widgets wxwidgets x264 xattr xcb xml xv xvid zip zlib zsh-completion" ABI_X86="64" ADA_TARGET="gnat_2020" APACHE2_MODULES="authn_core authz_core socache_shmcb unixd actions alias auth_basic authn_alias authn_anon authn_dbm authn_default authn_file authz_dbm authz_default authz_groupfile authz_host authz_owner authz_user autoindex cache cgi cgid dav dav_fs dav_lock deflate dir disk_cache env expires ext_filter file_cache filter headers include info log_config logio mem_cache mime mime_magic negotiation rewrite setenvif speling status unique_id userdir usertrack vhost_alias" CALLIGRA_FEATURES="karbon sheets words" COLLECTD_PLUGINS="df interface irq load memory rrdtool swap syslog" CPU_FLAGS_X86="aes avx avx2 f16c fma3 mmx mmxext pclmul popcnt rdrand sha sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3" ELIBC="glibc" GPSD_PROTOCOLS="ashtech aivdm earthmate evermore fv18 garmin garmintxt gpsclock greis isync itrax mtk3301 nmea ntrip navcom oceanserver oldstyle oncore rtcm104v2 rtcm104v3 sirf skytraq superstar2 timing tsip tripmate tnt ublox ubx" INPUT_DEVICES="libinput" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" LIBREOFFICE_EXTENSIONS="presenter-console presenter-minimizer" LUA_SINGLE_TARGET="lua5-1" LUA_TARGETS="lua5-1" OFFICE_IMPLEMENTATION="libreoffice" PHP_TARGETS="php7-4 php8-0" POSTGRES_TARGETS="postgres12 postgres13" PYTHON_SINGLE_TARGET="python3_9" PYTHON_TARGETS="python3_9" RUBY_TARGETS="ruby27 ruby30" USERLAND="GNU" VIDEO_CARDS="nouveau" XTABLES_ADDONS="quota2 psd pknock lscan length2 ipv4options ipset ipp2p iface geoip fuzzy condition tee tarpit sysrq proto steal rawnat logmark ipmark dhcpmac delude chaos account" Unset: ADDR2LINE, AR, ARFLAGS, AS, ASFLAGS, CC, CCLD, CONFIG_SHELL, CPP, CPPFLAGS, CTARGET, CXX, CXXFILT, ELFEDIT, EMERGE_DEFAULT_OPTS, EXTRA_ECONF, F77FLAGS, FC, GCOV, GPROF, INSTALL_MASK, LC_ALL, LD, LEX, LFLAGS, LIBTOOL, LINGUAS, MAKE, MAKEFLAGS, NM, OBJCOPY, OBJDUMP, PORTAGE_BINHOST, PORTAGE_BUNZIP2_COMMAND, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS, RANLIB, READELF, RUSTFLAGS, SIZE, STRINGS, STRIP, YACC, YFLAGS ```
Maybe pusleaudio is a red herring. This actually seems to be caused by: [0324/091407.973757:FATAL:gpu_data_manager_impl_private.cc(414)] GPU process isn't usable. Goodbye.
I'm also getting "Illegal instruction" with zoom-5.10.0.2450 dmesg says traps: Chrome_IOThread[17625] trap invalid opcode ip:7f31fe0aa5a5 sp:7f31ed78e5e0 error:0 in libcef.so[7f31fb083000+8451000] So it seems to be an issue with the bundled libcef.so (chromium embedded framework) zoom-5.9.6.2225 does not contain this library and doesn't crash. Title should be edited, the pactl stuff is definitely a red herring. zoom-5.9 and many previous versions have also complained about pactl's absence yet work just fine.
I'm also getting the "GPU process isn't usable" message - google suggests that's a common issue with chromium embedded, affecting many electron apps as well as something called JCEF that some java things use. google chrome works fine on the same system btw. Not sure if this message and the 'illegal instruction' dump are related though, those seem like two separate fatal issues.
After finding https://www.reddit.com/r/openSUSE/comments/srbswv/comment/hwth03z/ it seems that adding --disable-gpu-sandbox commandline option seems to make it work - ie /usr/bin/zoom --disable-gpu-sandbox or edit /usr/bin/zoom (it's a bash script) to always supply this option. Perhaps the zoom ebuild should edit this file?
(In reply to Michael Moon from comment #2) Could you report this upstream please?
(In reply to Ulrich Müller from comment #5) > Could you report this upstream please? Yep, I've already reported it upstream, but unfortunately zoom's support tickets are not public so there's no point in me linking it here.
Just wanna say I am super impressed by the quick response here! I love gentoo!
That's an odd problem, because this specific zoom version (and ebuild) worked for me still on 2022-03-25 (see also #835822) but since today not anymore. I also have an NVidia-GPU (in case this has anything to do with it, but a downgrade to nvidia-drivers-510.54 didn't work).
Some reports on a similar problem (same error message), points to newer versions of glibc and the use of clone3.
Glibc eh? I'm on sys-libs/glibc-2.34-r10 fwiw
(In reply to Michael Moon from comment #10) > Glibc eh? I'm on sys-libs/glibc-2.34-r10 fwiw Thanks.. saved me from having to setup a system with a down level version of glibc.. (I'm in 2.35). But regardless.. it might still be an issue with glibc and zoom.. The IT team at work is reporting the problem to zoom as well.
I don't think it's specific to recent glibc, as I've found various posts reaching back as far as 2019 with chromium embedded frameworks throwing this error in a very wide range of different software, from intellij to Minecraft and including a mountain of electron apps. It seems like zoom 5.10 added CEF and is running straight into this issue, while zoom 5.9 doesn't contain libcef.so and thus isn't affected
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=173ec8a43fd6b7c9bc1dc0f81fcb5e7c755bf96c commit 173ec8a43fd6b7c9bc1dc0f81fcb5e7c755bf96c Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-03-30 16:16:15 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-03-30 16:16:15 +0000 profiles: Package mask net-im/zoom-5.10.0.2450 Bug: https://bugs.gentoo.org/835930 Signed-off-by: Ulrich Müller <ulm@gentoo.org> profiles/package.mask | 5 +++++ 1 file changed, 5 insertions(+)
It works absolutely fine here with glibc-2.34. Whatever problem you have is not related to clone3.
2 out of 2 reports are on nvidia-drivers. (which makes sense since this is about gpu acceleration...) don't rely on closed-source binary garbage.
I think it may be due to a different code path with nvidia-drivers or something, but it does fail on clone3 for me: >trace.9493-clone3({flags=CLONE_VM|CLONE_FS|CLONE_FILES|CLONE_SIGHAND|CLONE_THREAD|CLONE_SYSVSEM|CLONE_SETTLS|CLONE_PARENT_SETTID|CLONE_CHILD_CLEARTID, child_tid=0x7fffe4acf910, parent_tid=0x7fffe4acf910, exit_signal=0, stack=0x7fffe42cf000, stack_size=0x7ffbc0, tls=0x7fffe4acf640} => {parent_tid=[-1]}, 88) = 435 >trace.9493---- SIGSYS {si_signo=SIGSYS, si_code=SYS_SECCOMP, si_errno=EPERM, si_call_addr=0x7fffec358eb9, si_syscall=__NR_clone3, si_arch=AUDIT_ARCH_X86_64} --- >trace.9493:+++ killed by SIGSYS +++ Applying the old clone3 patch to glibc-2.35 (haven't tried 2.34) gave me a seemingly working zoom without needing --disable-gpu-sandbox
(In reply to Ionen Wolkens from comment #16) > >trace.9493:+++ killed by SIGSYS +++ To clarify, this seems to spiral and other threads die on SIGILL, leading to that Illegal instruction error.
(In reply to Andreas K. Hüttel from comment #15) > 2 out of 2 reports are on nvidia-drivers. (which makes sense since this is > about gpu acceleration...) > > don't rely on closed-source binary garbage. ... of which we have two instances here in combination. :( In any case, there's little that can be done on the distro side. Therefore >=zoom-5.10 will stay package.masked until there is confirmation that the problem has been fixed upstream (e.g. in their release notes).
Downgrading to glibc-2.33-r13 fixed the problem for me. I DO NOT RECOMMEND THIS WAY, GREAT RISK OF BREAKING YOUR SYSTEM!! I checked with genlop which packages I had installed over the weekend, given this specific, same zoom-version still worked for me on Friday: Sun Mar 27 14:45:20 2022 >>> sys-libs/glibc-2.34-r10 Sun Mar 27 14:45:45 2022 >>> sys-apps/sandbox-2.29 Sun Mar 27 14:45:53 2022 >>> dev-python/flit_core-3.7.1 Sun Mar 27 14:46:37 2022 >>> app-arch/libarchive-3.6.0 Sun Mar 27 14:46:49 2022 >>> x11-libs/libvdpau-1.5 Sun Mar 27 14:47:00 2022 >>> dev-python/click-8.0.4 Sun Mar 27 14:47:06 2022 >>> sys-devel/crossdev-20220205 Sun Mar 27 14:47:13 2022 >>> www-client/google-chrome-99.0.4844.84 Sun Mar 27 14:47:22 2022 >>> net-misc/teamviewer-15.28.6 Sun Mar 27 14:47:29 2022 >>> app-portage/eix-0.36.1 Sun Mar 27 14:49:22 2022 >>> media-video/ffmpeg-4.4.1-r4 Sun Mar 27 14:49:46 2022 >>> dev-python/jupyter_client-7.1.2 Sun Mar 27 14:49:52 2022 >>> dev-python/matplotlib-inline-0.1.3-r1 Sun Mar 27 14:50:02 2022 >>> net-im/zoom-5.10.0.2450-r2 Sun Mar 27 14:50:13 2022 >>> dev-python/ipywidgets-7.6.5-r2 Sun Mar 27 14:50:30 2022 >>> app-i18n/man-pages-l10n-4.12.1-r1 Mon Mar 28 19:50:00 2022 >>> net-analyzer/mtr-0.95 Mon Mar 28 19:57:16 2022 >>> sys-devel/binutils-config-5.4.1 Mon Mar 28 19:57:37 2022 >>> dev-python/bottleneck-1.3.4 Mon Mar 28 19:57:45 2022 >>> dev-python/cryptography-3.4.7-r3 Mon Mar 28 19:57:55 2022 >>> dev-python/reportlab-3.6.7 Mon Mar 28 19:58:30 2022 >>> sys-kernel/gentoo-sources-5.17.1 Mon Mar 28 19:58:36 2022 >>> media-libs/libv4l-1.22.1 Mon Mar 28 19:58:45 2022 >>> dev-vcs/git-2.35.1 Mon Mar 28 19:58:57 2022 >>> cross-arm-xilinx-linux-gnueabi/glibc-2.35-r1 Wed Mar 30 11:51:26 2022 >>> net-im/zoom-5.10.0.2450-r2 Wed Mar 30 12:07:53 2022 >>> app-admin/ryzen_smu-0.1.2_p20211205 Wed Mar 30 12:07:58 2022 >>> media-video/v4l2loopback-0.12.5-r1 Wed Mar 30 12:08:04 2022 >>> sys-power/acpi_call-1.2.2 Wed Mar 30 12:08:36 2022 >>> x11-drivers/nvidia-drivers-510.54 Previously, I was probably on 2.33-r10, so downgrading to 2.33-r13 wouldn't be much of an issue I thought, given it's an upgrade for most of the packages on my system, apart from the ones listed above. The only critical one would be sys-apps/sandbox. After downgrading, the emerge failed to execute postinst. Emerging sys-apps/sandbox again with sandboxing turned off in FEATURES brought everything back to life. I currently re-emerge all of the above packets, but everything seems to be normal and ok. For gentoo-sources I had to clean the tools-folder specifically (after distclean) with make -C tools/objtool clean.
Can confirm that running with a patched glibc (backing out the patch 0049-Revert-clone-internal.c-make-clone3-syscall-optional.patch) does allow the latest level of zoom to work again. I'm okay with leaving the latest zoom version masked until the zoom folks fix the problem. I've relayed some of the information from this ticket to my IT person who has reported the problem to zoom. Thanks
We'll (reluctantly) restore USE=clone3.
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=bfd4644d464949179a652fa6c2cbfd1fd2fd4b3e commit bfd4644d464949179a652fa6c2cbfd1fd2fd4b3e Author: Andreas K. Hüttel <dilfridge@gentoo.org> AuthorDate: 2022-03-30 23:18:57 +0000 Commit: Andreas K. Hüttel <dilfridge@gentoo.org> CommitDate: 2022-03-30 23:19:22 +0000 sys-libs/glibc: Revbumps to reintroduce clone3 useflag Unkeyworded until minimal testing is done Bug: https://bugs.gentoo.org/835930 Package-Manager: Portage-3.0.30, Repoman-3.0.3 Signed-off-by: Andreas K. Hüttel <dilfridge@gentoo.org> sys-libs/glibc/Manifest | 2 + sys-libs/glibc/glibc-2.34-r11.ebuild | 1619 ++++++++++++++++++++++++++++++++++ sys-libs/glibc/glibc-2.35-r2.ebuild | 1590 +++++++++++++++++++++++++++++++++ sys-libs/glibc/metadata.xml | 1 + 4 files changed, 3212 insertions(+)
(In reply to Andreas K. Hüttel from comment #15) > 2 out of 2 reports are on nvidia-drivers. (which makes sense since this is > about gpu acceleration...) > > don't rely on closed-source binary garbage. I am using nouveau, what are you talking about? burzmali@usul ~ $ lsmod|grep nvidia burzmali@usul ~ $ lsmod|grep nouveau nouveau 1855488 4 video 49152 1 nouveau drm_ttm_helper 16384 1 nouveau ttm 57344 2 drm_ttm_helper,nouveau drm_kms_helper 204800 1 nouveau drm 409600 8 drm_kms_helper,drm_ttm_helper,ttm,nouveau backlight 16384 4 video,drm_kms_helper,nouveau,drm i2c_algo_bit 16384 2 igb,nouveau hwmon 28672 4 nct6775,igb,k10temp,nouveau i2c_core 65536 6 drm_kms_helper,i2c_algo_bit,igb,i2c_piix4,nouveau,drm wmi 24576 4 nct6775,wmi_bmof,mxm_wmi,nouveau button 20480 1 nouveau burzmali@usul ~ $
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=157fbebc9dbb9466dd7597a9ffcf545ad2706ad4 commit 157fbebc9dbb9466dd7597a9ffcf545ad2706ad4 Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-03-31 06:55:59 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-03-31 06:55:59 +0000 net-im/zoom: Depend on glibc[-clone3] Package will stay masked for now, waiting for upstream fix. Bug: https://bugs.gentoo.org/835930 Signed-off-by: Ulrich Müller <ulm@gentoo.org> net-im/zoom/{zoom-5.10.0.2450-r2.ebuild => zoom-5.10.0.2450-r3.ebuild} | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=143da360ff5c803d89dce81ee5458832c1d7b564 commit 143da360ff5c803d89dce81ee5458832c1d7b564 Author: Andreas K. Hüttel <dilfridge@gentoo.org> AuthorDate: 2022-03-31 20:55:30 +0000 Commit: Andreas K. Hüttel <dilfridge@gentoo.org> CommitDate: 2022-03-31 20:55:30 +0000 package.mask: drop net-im/zoom mask, glibc[-clone3] is back Closes: https://bugs.gentoo.org/835930 Signed-off-by: Andreas K. Hüttel <dilfridge@gentoo.org> profiles/package.mask | 5 ----- 1 file changed, 5 deletions(-)
Please test zoom-5.10.0.2450-r3.
I will test 5.10 again when the glibc update doesn't require a keyword mask. I'm happy to keyword-unmask many non-crucial packages, but I'm not willing to use an even theoretically unstable glibc on this system. I love that Gentoo gives me this choice. I've mentioned this bug report in my Zoom ticket about this issue - although for some reason Zoom staff suddenly decided that it was a feature request, as if "not crashing immediately" was a nice feature rather than a crucial one.
(In reply to Michael Moon from comment #27) > I've mentioned this bug report in my Zoom ticket about this issue - although > for some reason Zoom staff suddenly decided that it was a feature request, > as if "not crashing immediately" was a nice feature rather than a crucial > one. Well, sooner or later upstream will encounter this issue also on other distros with >=glibc-2.34.
(In reply to Ulrich Müller from comment #28) > Well, sooner or later upstream will encounter this issue also on other > distros with >=glibc-2.34. If and when that happens, perhaps they'll take us Gentoo folk a little more seriously - that's the dream at least. Well actually the dream would be a published thorough API so folks can make their own clients, but that's a huge back-and-forth these days since the high water mark and subsequent graveyarding of Google Talk's use of XMPP.
Hi, I've just tested following package combination and it seems like it works fine: =sys-libs/glibc-2.34-r11 =net-im/zoom-5.10.0.2450-r3 =x11-drivers/nvidia-drivers-470.103.01 Feel free to ask me if any additional checks are required.
(In reply to Mike Limansky from comment #30) > I've just tested following package combination and it seems like it works > fine: > > =sys-libs/glibc-2.34-r11 > =net-im/zoom-5.10.0.2450-r3 > =x11-drivers/nvidia-drivers-470.103.01 Thank you. glibc is with USE="-clone3" I suppose?
I tested with: =sys-libs/glibc-2.34-r11 (-clone3) =net-im/zoom-5.10.0.2450-r3 =x11-drivers/xf86-video-nouveau-1.0.17 (i don't use the binary nvidia drivers) and it is working for me.
(In reply to Ulrich Müller from comment #31) > Thank you. glibc is with USE="-clone3" I suppose? Sure, cause it's a dependency of =net-im/zoom-5.10.0.2450-r3.
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=176c9d26cb48763fb341aae2c51207b51c7d0343 commit 176c9d26cb48763fb341aae2c51207b51c7d0343 Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-04-20 05:56:59 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-04-20 05:56:59 +0000 net-im/zoom: Version bump to 5.10.3.2778 Bug: https://bugs.gentoo.org/835930 Signed-off-by: Ulrich Müller <ulm@gentoo.org> net-im/zoom/Manifest | 1 + net-im/zoom/zoom-5.10.3.2778.ebuild | 179 ++++++++++++++++++++++++++++++++++++ 2 files changed, 180 insertions(+)
Could someone please test if the issue is fixed with zoom-5.10.3.2778 and glibc-2.35-r2[+clone3]? If it is, that zoom version could be unmasked again.
(In reply to Ulrich Müller from comment #35) > Could someone please test if the issue is fixed with zoom-5.10.3.2778 and > glibc-2.35-r2[+clone3]? If it is, that zoom version could be unmasked again. No change for me. $ emerge -pvq sys-libs/glibc net-im/zoom [ebuild R ] sys-libs/glibc-2.35-r2 USE="clone3 ... [ebuild R ] net-im/zoom-5.10.3.2778 [0420/022446.556702:FATAL:gpu_data_manager_impl_private.cc(414)] GPU process isn't usable. Goodbye. Illegal instruction
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=13f54c3176d492028990e7a2bb9515ad61337cb1 commit 13f54c3176d492028990e7a2bb9515ad61337cb1 Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-04-20 06:44:39 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-04-20 06:45:13 +0000 net-im/zoom: Version 5.10.3.2778 needs to depend on glibc[-clone3] Bug: https://bugs.gentoo.org/835930#c36 Signed-off-by: Ulrich Müller <ulm@gentoo.org> net-im/zoom/zoom-5.10.3.2778.ebuild | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=cddb736f31ff4ad70a2d270ae6e36bef7c0ac8d6 commit cddb736f31ff4ad70a2d270ae6e36bef7c0ac8d6 Author: Andreas K. Hüttel <dilfridge@gentoo.org> AuthorDate: 2022-04-27 09:26:01 +0000 Commit: Andreas K. Hüttel <dilfridge@gentoo.org> CommitDate: 2022-04-27 09:27:52 +0000 package.mask: Drop mask for >=zoom-5.10 Reason: zoom-5.9 can't sign in to the network anymore. Nothing regarding bug 835930 has actually changed. Please disable the clone3 useflag of glibc if you encounter problems. Bug: https://bugs.gentoo.org/835930 Signed-off-by: Andreas K. Hüttel <dilfridge@gentoo.org> profiles/package.mask | 5 ----- 1 file changed, 5 deletions(-)
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=22544225350a4985f624cb46c270cddc0af7c60f commit 22544225350a4985f624cb46c270cddc0af7c60f Author: Andreas K. Hüttel <dilfridge@gentoo.org> AuthorDate: 2022-04-27 09:29:24 +0000 Commit: Andreas K. Hüttel <dilfridge@gentoo.org> CommitDate: 2022-04-27 09:29:24 +0000 net-im/zoom: remove old Zoom 5.9 can't sign in to the network anymore. Bug: https://bugs.gentoo.org/835930 Package-Manager: Portage-3.0.30, Repoman-3.0.3 Signed-off-by: Andreas K. Hüttel <dilfridge@gentoo.org> net-im/zoom/Manifest | 1 - net-im/zoom/zoom-5.9.6.2225-r1.ebuild | 165 ---------------------------------- 2 files changed, 166 deletions(-)
(In reply to Andreas K. Hüttel from comment #38) > Reason: zoom-5.9 can't sign in to the network anymore. Uhh zoom-5.9 works just fine here, no need to drop it yet…
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=8e83d5804fe006f3da794768b288b4091536bd1e commit 8e83d5804fe006f3da794768b288b4091536bd1e Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-04-27 16:31:33 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-04-27 16:38:20 +0000 net-im/zoom: Don't depend on glibc[-clone3] Display an explanatory message in README.gentoo instead. Bug: https://bugs.gentoo.org/835930 Signed-off-by: Ulrich Müller <ulm@gentoo.org> net-im/zoom/files/README.gentoo | 8 -------- net-im/zoom/zoom-5.10.3.2778.ebuild | 21 +++++++++++++++------ ....10.4.2845.ebuild => zoom-5.10.4.2845-r1.ebuild} | 21 +++++++++++++++------ 3 files changed, 30 insertions(+), 20 deletions(-)
Closing. Not much more that we can do on the distro level, so please report any further related issues upstream. Alternatively, consider using a free alternative. :)
(In reply to Ulrich Müller from comment #35) > Could someone please test if the issue is fixed with zoom-5.10.3.2778 and > glibc-2.35-r2[+clone3]? If it is, that zoom version could be unmasked again. I've just checked net-im/zoom-5.10.4.2845-r1 doesn't work with glibc[+clone3], and works fine with glibc[-clone3]. I use nvidia-drivers (looks like it might be related).
(In reply to Andreas K. Hüttel from comment #39) Can we reinstate the zoom-5.9 ebuild? It doesn't suffer from this bug, and it signs into the zoom network just fine for me.
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=030998bee4664b6c5b2fd65e6cb8c077164a3f2d commit 030998bee4664b6c5b2fd65e6cb8c077164a3f2d Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-05-02 10:52:13 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-05-02 10:52:13 +0000 net-im/zoom: Restore version 5.6.9.2225 Bug: https://bugs.gentoo.org/835930#c44 Signed-off-by: Ulrich Müller <ulm@gentoo.org> net-im/zoom/Manifest | 1 + net-im/zoom/zoom-5.9.6.2225-r2.ebuild | 171 ++++++++++++++++++++++++++++++++++ 2 files changed, 172 insertions(+)
(In reply to Ulrich Müller from comment #45) > net-im/zoom: Restore version 5.6.9.2225 Thanks! This'll stop portage complaining about 5.10 being masked on my system but no other versions available, until either upstream fixes this issue or ≥glibc-2.34-r11[clone3] is stabilised.
It's not just you, Michael Moon. Something recent broke Zoom for me. I'm running glibc-2.34-r10, where the clone3 USE flag does not exist yet, and zoom-5.10.4.2845-r1. I get the "GPU process isn't usable. Goodbye." and "19502 Illegal instruction" messages. So, what does this mean? Does Zoom 5.10 require glibc-2.35? If so, shouldn't that be a dependency?
(In reply to Hypoon from comment #47) > Does Zoom 5.10 require glibc-2.35? Zoom is closed source, so you should really ask this question upstream. Alternatively, check their release notes: https://support.zoom.us/hc/en-us/articles/205759689
(In reply to Hypoon from comment #47) > It's not just you, Michael Moon. Something recent broke Zoom for me. > > I'm running glibc-2.34-r10, where the clone3 USE flag does not exist yet, > and zoom-5.10.4.2845-r1. > > I get the "GPU process isn't usable. Goodbye." and "19502 Illegal > instruction" messages. > > So, what does this mean? Does Zoom 5.10 require glibc-2.35? If so, shouldn't > that be a dependency? If you want to use zoom-5.10 with glibc-2.34-r10, see my comment #4 for a workaround. The issue seems to be something weird happening in chromium embedded frameworks (CEF), which zoom-5.10 added (zoom-5.9 doesn't contain libcef.so). The CEF issue affects multiple different projects, including a laundry list of Electron/AppImage apps, a bunch of Java things, and a few others. I wouldn't be surprised if other bugs start getting reported against seemingly unrelated ebuilds that use CEF (or gentoo packages supporting non-gentoo things), hopefully the bug wranglers can spot the relevant issues and mark them as dupes of this one. Ideally, CEF should fix its issue with glibc's clone3, and all these binary blobs should update their CEF library version to the fixed one - but who knows if or when that'll happen.
I've opened a ticket upstream. We'll see if it goes anywhere useful.
(In reply to Michael Moon from comment #49) > Ideally, CEF should fix its issue with glibc's clone3, and all these binary > blobs should update their CEF library version to the fixed one - but who > knows if or when that'll happen. My impression is that either it's already fixed in newer cef, e.g. I do not have that problem with obs-studio[browser], or other packages are disabling the gpu sandbox by default. My "guess" is that this is what zoom does for other cards (would use clone3 all the same, it's not nvidia that's doing anything here) but for some reason the sandbox doesn't get disabled with nvidia-drivers. If upstream does nothing and this is going to be a long-standing issue, my personal recommendation to have least amount of surprise (rather than just postinst message), while still not forcing it for everyone, would be to rely on IUSE=video_cards_nvidia to request adequate [-clone3] version.
(In reply to Ionen Wolkens from comment #51) > If upstream does nothing and this is going to be a long-standing issue, my > personal recommendation to have least amount of surprise (rather than just > postinst message), while still not forcing it for everyone, would be to rely > on IUSE=video_cards_nvidia to request adequate [-clone3] version. Sounds good to me - so stable can have zoom-5.9 and glibc-2.34-r10, and ~arch can have zoom-5.10 with DEPEND="video_cards_nvidia? ( sys-libs/glibc[-clone3] )" ? Either that or the ebuild could inject --disable-gpu-sandbox into the /usr/bin/zoom shell script.. I'm curious what's going on with comment #23 though, jharveysmith reports encountering this issue with nouveau rather than nvidia-drivers
(In reply to Ionen Wolkens from comment #51) > If upstream does nothing and this is going to be a long-standing issue, my > personal recommendation to have least amount of surprise (rather than just > postinst message), while still not forcing it for everyone, would be to rely > on IUSE=video_cards_nvidia to request adequate [-clone3] version. Nope. If it would be a long-standing issue and upstream wouldn't fix it (and we cannot patch it), then IMHO the logical thing to do is to last-rite such a broken package, instead of adding more dirty workarounds. That's because we already have more than enough of those. For example, some systems require the bundled libjpeg-turbo while others require a bundled Qt. And again, it's closed source, so we don't even know *why* these are needed.
(In reply to Ulrich Müller from comment #53) > (In reply to Ionen Wolkens from comment #51) > > If upstream does nothing and this is going to be a long-standing issue, my > > personal recommendation to have least amount of surprise (rather than just > > postinst message), while still not forcing it for everyone, would be to rely > > on IUSE=video_cards_nvidia to request adequate [-clone3] version. > > Nope. If it would be a long-standing issue and upstream wouldn't fix it (and > we cannot patch it), then IMHO the logical thing to do is to last-rite such > a broken package, instead of adding more dirty workarounds. > > That's because we already have more than enough of those. For example, some > systems require the bundled libjpeg-turbo while others require a bundled Qt. > And again, it's closed source, so we don't even know *why* these are needed. Well, it is closed source, and it's a thing people want, so we have to live with it. That's why toolchain@ added the flag. I suggest you keep using it until it's no longer necessary or until such a time that toolchain@ wants all consumers gone and has had enough. THere's no need to be pre-emptively dropping it. Also, it's silly, given we restored it specifically for Zoom, and now it's been dropped again, but users are going to need it. Proprietary software is never fun or easy to package. If you don't want to maintain it, that's completely understandable, but this is quite an easy tweak to make to allow people to continue using it.
(... also, what tends to happen with these is, they just magically fixed one day after a few months when upstream do their regular sync with CEF. Happened with Teams, Skype, all the various binary browsers ...)
(In reply to Hypoon from comment #50) > I've opened a ticket upstream. We'll see if it goes anywhere useful. Good luck. Please report back here if you get any (even a negative) reaction.
The other option which doesn't require modifying glibc, is just passing --disable-gpu-sandbox, could also be managed through video_cards_nvidia.
(In reply to Ionen Wolkens from comment #57) > The other option which doesn't require modifying glibc, is just passing > --disable-gpu-sandbox, could also be managed through video_cards_nvidia. ... and as mentioned before, my impression is that it is disabling it for other cards already. Otherwise I don't see why it's not using clone3 with everything else. Although [-clone3] is nice to keep sandbox and the option is there.
(In reply to Ionen Wolkens from comment #57) > The other option which doesn't require modifying glibc, is just passing > --disable-gpu-sandbox, could also be managed through video_cards_nvidia. We don't know for certain if it is related (only) to Nvidia cards, so I'd rather not link it to that use-expand. We could have something like the following though: IUSE="... +clone3 ..." RDEPEND="... || ( >=sys-libs/glibc-2.34[!clone3?(+)] <sys-libs/glibc-2.34 ) ..." src_install() { ... make_wrapper zoom \ "/opt/zoom/zoom $(usev clone3 --disable-gpu-sandbox)" \ /opt/zoom /opt/zoom:/opt/zoom/cef ... }
I assume the sandbox exists for a reason, so "--disable-gpu-sandbox" does not seem like a good solution to me. I'll defer to someone else who understands the purpose of the gpu sandbox and the consequences of disabling it. It's easy to put all the burden on upstream for this issue, but there's something important to remember here: upstream has clear installation instructions (and implied support) for more than a couple of Linux distributions: Debian, Ubuntu, Linux Mint, openSUSE, Oracle Linux, CentOS, RedHat, Fedora, Arch, and Manjaro. If it works on all of those, but it doesn't work in Gentoo... that sounds like a Gentoo problem.
Well, I wouldn't say Gentoo-exclusive problem: The Ubuntu (and Debian?)-devs simply disabled clone3 until the recent version 22.04 (see https://git.launchpad.net/~ubuntu-core-dev/ubuntu/+source/glibc/commit/debian?id=bcf2f127422f9cad5a51191152fa2ba8758ba65e) and with 22.04 they seem to have the same problem (see https://askubuntu.com/a/1405420). On a (slightly related) sidenote: I seem to be missing the clone3 use-flag from 2.34-r10. Has it been removed? With 2.34-r11 it's still there. Does the (now unconditionally applied) patch fix the problem with/for zoom?
(In reply to Frederik Pfautsch from comment #61) > I seem to be missing the clone3 use-flag from 2.34-r10. 2.43-r10 doesn't have it. I think @toolchain set it to always enabled, but reinstated the flag for ≥2.34-r11 due to this issue
If other distributions are seeing the same issue, then I think it's pretty likely that upstream will issue a fix. Out of curiosity, I tried replacing the libcef.so that ships with Zoom by using this binary release of CEF: cef_binary_101.0.15+gca159c5+chromium-101.0.4951.54_linux64_minimal.tar.bz2 ... and created a new /opt/zoom/cef directory and copied the contents of Release/ and Resources/ to it. Here's the result: /opt/zoom/zoom: symbol lookup error: /opt/zoom/zoom: undefined symbol: cef_unregister_internal_web_plugin Cleanup successful. I didn't dive any deeper.
(In reply to Hypoon from comment #63) > If other distributions are seeing the same issue, then I think it's pretty > likely that upstream will issue a fix. I have reported Ubuntu LTS being affected to my (non-public) upstream zoom issue. I would like to think that this spurs them to consider Gentoo bug reports as predictive, but (sadly) I've little hope that this will actually occur.
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=b4a4bdb0c8fe2a795c48adab25b001754d289f88 commit b4a4bdb0c8fe2a795c48adab25b001754d289f88 Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-05-07 21:00:07 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-05-07 21:03:26 +0000 net-im/zoom: Be more noisy when warning about libcef breakage Bug: https://bugs.gentoo.org/835930 Suggested-by: Andreas K. Hüttel <dilfridge@gentoo.org> Signed-off-by: Ulrich Müller <ulm@gentoo.org> .../{zoom-5.10.4.2845-r1.ebuild => zoom-5.10.4.2845-r2.ebuild} | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
Downgrading to 5.9 worked around this problem. *Not* resolved. I also use nvidia-drivers. My comment is that zoom-5.10 should *not* have been stabilized with this critical issue unresolved. That was definitely a nasty Monday-morning issue for me.
(In reply to Jonathan Snow from comment #66) > Downgrading to 5.9 worked around this problem. *Not* resolved. > > I also use nvidia-drivers. > > My comment is that zoom-5.10 should *not* have been stabilized with this > critical issue unresolved. That was definitely a nasty Monday-morning issue > for me. I apologize: deleting the zoom keyword also worked around this problem. It was not stabilized
Jonathan Snow, both versions of Zoom currently in the tree are keyworded ~amd64, so deleting the Zoom keyword solves nothing. Right now, there's no logic (related to this problem, at least) that informs Portage to keep Zoom at version 5.9. I agree that this should change. Portage should know that upgrading to 5.10 may not be safe yet. A printed warning isn't enough... the role of a package manager is to ensure compatibility with dependencies, not to inform users about (in)compatibilities so users can manage them manually. One way to do this would be to have Zoom 5.10 depend on a version of glibc which has the clone3 USE flag (e.g., 5.34-r11) and which has that flag disabled. Portage would be unable to satisfy this dependency until 5.34-r11 becomes stabilized, so it would fall back to Zoom 5.9. Another option would be to mask this version of 5.10 so that users who want to upgrade have to manually unmask it. Either way, Portage would issue downgrades to 5.9, and that's acceptable since it's currently unknown why Zoom 5.10 + glibc-with-clone3 works for some users and not others. I assume that stabilizing Zoom 5.9 is inappropriate? Zoom 5.9 works, but I'm not sure what other requirements there are for stabilizing a package. Now some good news. Upstream has confirmed the bug and says that a fix will come with version 5.10.6 (Reminder: at time of writing, we're on 5.10.4). There's no timeline for when 5.10.6 will be available. I'll report back with any additional information I get regarding the specific combination of dependencies that cause this problem.
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=0a1218e187b8e3edd8aa9d2daf6567c6b90e36f5 commit 0a1218e187b8e3edd8aa9d2daf6567c6b90e36f5 Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-05-10 11:24:09 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-05-11 05:51:01 +0000 Revert "package.mask: Drop mask for >=zoom-5.10" This reverts commit cddb736f31ff4ad70a2d270ae6e36bef7c0ac8d6. Bug: https://bugs.gentoo.org/835930 Signed-off-by: Ulrich Müller <ulm@gentoo.org> profiles/package.mask | 5 +++++ 1 file changed, 5 insertions(+)
Relevant to this issue: glibc-2.34-r13 has been stabilized, so the clone3 USE flag is present once again without having to use ~arch glibc.
(In reply to Hypoon from comment #68) > Now some good news. Upstream has confirmed the bug and says that a fix will > come with version 5.10.6 (Reminder: at time of writing, we're on 5.10.4). > There's no timeline for when 5.10.6 will be available. I'll report back with > any additional information I get regarding the specific combination of > dependencies that cause this problem. zoom-5.10.6.3192 is in the tree. However, there's nothing in upstream release notes from which I would conclude that this bug is fixed. So, could somebody please test if the problem is gone in 5.10.6? That is, does it work with glibc[clone3] and nvidia-drivers (and 5.10.4 didn't)? If yes, then we can remove the warning message from the ebuild and the package mask.
Actually, the contents of the cef/ subdirectory (containing libcef.so and other files) is identical for versions 5.10.4.2845 and 5.10.6.3192. So I'm not very optimistic.
Just ran a sync, but it didn't get pulled in. I probably need to wait for the mirrors to synchronize first. I'll try again later (breaking Gentoo etiquette running multiple syncs in the same day, sorry!).
(In reply to Hypoon from comment #73) > Just ran a sync, but it didn't get pulled in. I probably need to wait for > the mirrors to synchronize first. I'll try again later (breaking Gentoo > etiquette running multiple syncs in the same day, sorry!). Should be on rsync mirrors now, with this timestamp.x: 1653414001 Tue 24 May 2022 05:40:01 PM UTC
(In reply to Ulrich Müller from comment #71) > So, could somebody please test if the problem is gone in 5.10.6? That is, > does it work with glibc[clone3] and nvidia-drivers (and 5.10.4 didn't)? If > yes, then we can remove the warning message from the ebuild and the package > mask. Hi, I've just tested net-im/zoom-5.10.6.3192, with sys-libs/glibc-2.34-r13[clone3] and x11-drivers/nvidia-drivers-510.68.02. It looks like it works fine.
I can also confirm that =net-im/zoom-5.10.6.3192 works with =sys-libs/glibc-2.34-r13[clone3]. Thanks, everyone! Part of me thinks it's safe to unmask 5.10.6.3192. The other part of me thinks we should get more confirmations first. It seems wrong to test the fix by just unmasking and waiting to see if anybody complains.
I can also confirm that =net-im/zoom-5.10.6.3192 seems to work with =sys-libs/glibc-2.34-r13[clone3] - at least with regards to the issue this bug tracks, ie CTD on startup with complaints about GPU processes and illegal instructions.
This might be related https://github.com/NVIDIA/open-gpu-kernel-modules/issues/302
