Closed Bug 2081642 Opened 1 month ago Closed 1 month ago

Firefox's interface (except a translation modal) does not invoke (on Wayland), unless via XWayland.

Categories

(Core :: Widget: Gtk, defect)

Firefox 160
x86_64
Linux
defect

Tracking

()

RESOLVED FIXED
Tracking Status
firefox-esr115 --- unaffected
firefox-esr153 --- unaffected
firefox158 --- unaffected
firefox159 --- unaffected
firefox160 --- fixed

People

(Reporter: B.RJL, Unassigned)

References

(Regression)

Details

(Keywords: regression)

Attachments

(2 files)

With MOZ_ENABLE_WAYLAND=0 prepended, it invokes as expected:

RokeJulianLockhart@beedell:~$ MOZ_ENABLE_WAYLAND=0 firefox-nightly
ATTENTION: default value of option mesa_glthread overridden by environment.
ATTENTION: default value of option mesa_glthread overridden by environment.
ATTENTION: default value of option mesa_glthread overridden by environment.
ATTENTION: default value of option mesa_glthread overridden by environment.
ATTENTION: default value of option mesa_glthread overridden by environment.
ATTENTION: default value of option mesa_glthread overridden by environment.
ATTENTION: default value of option mesa_glthread overridden by environment.
ATTENTION: default value of option mesa_glthread overridden by environment.
libva info: VA-API version 1.23.0
libva info: Trying to open /usr/lib64/dri-nonfree/radeonsi_drv_video.so
libva info: Trying to open /usr/lib64/dri-freeworld/radeonsi_drv_video.so
libva info: Found init function __vaDriverInit_1_23
libva info: va_openDriver() returns 0
ATTENTION: default value of option mesa_glthread overridden by environment.
ATTENTION: default value of option mesa_glthread overridden by environment.
ATTENTION: default value of option mesa_glthread overridden by environment.

Without it, all that I observe is a translation window:

RokeJulianLockhart@beedell:~$ firefox-nightly
libva info: VA-API version 1.23.0
libva info: Trying to open /usr/lib64/dri-nonfree/radeonsi_drv_video.so
libva info: Trying to open /usr/lib64/dri-freeworld/radeonsi_drv_video.so
libva info: Found init function __vaDriverInit_1_23
libva info: va_openDriver() returns 0
^CExiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.

My Environment

  • Firefox

    Python 3.14.8 (main, Oct  1 2026, 00:00:00) [GCC 16.2.1 20260819 (Red Hat 16.2.1-2)] on linux
    Type "help", "copyright", "credits" or "license" for more information.
    >>> #!/usr/bin/env python3
    ... import rpm
    ... import yaml
    ... from datetime import datetime, timezone
    ... package = "firefox-nightly"
    ... header = next(
    ...    rpm.TransactionSet().dbMatch(
    ...       rpm.RPMTAG_NAME,
    ...       package,
    ...    ),
    ...    None,
    ... )
    ... if header is None:
    ...    raise LookupError(
    ...       f"Package {package!r} is not installed"
    ...    )
    ... tags = (
    ...    "NAME",
    ...    "VERSION",
    ...    "RELEASE",
    ...    "ARCH",
    ...    "INSTALLTIME",
    ...    "SIZE",
    ...    "SIGPGP",
    ...    "SOURCERPM",
    ...    "BUILDTIME",
    ...    "BUILDHOST",
    ...    "PACKAGER",
    ...    "VENDOR",
    ... )
    ... properties = {}
    ... for name in tags:
    ...    value = header[
    ...       getattr(
    ...          rpm,
    ...          f"RPMTAG_{name}",
    ...       )
    ...    ]
    ...    if name in (
    ...       "INSTALLTIME",
    ...       "BUILDTIME",
    ...    ):
    ...       value = (
    ...          datetime.fromtimestamp(
    ...             value,
    ...             timezone.utc,
    ...          )
    ...          .isoformat(timespec="seconds")
    ...          .replace(
    ...             "+00:00",
    ...             "Z",
    ...          )
    ...       )
    ...    properties[name] = value
    ... print(
    ...    yaml.safe_dump(
    ...       properties,
    ...       sort_keys=False,
    ...       default_flow_style=False,
    ...    )
    ... )
    ... 
    NAME: firefox-nightly
    VERSION: 160.0a1
    RELEASE: '20261014215550'
    ARCH: x86_64
    INSTALLTIME: '2026-10-15T11:07:10Z'
    SIZE: 408800345
    SIGPGP: null
    SOURCERPM: firefox-nightly-160.0a1-20261014215550.src.rpm
    BUILDTIME: '2026-10-15T00:09:36Z'
    BUILDHOST: 8b84fb34dfdf
    PACKAGER: null
    VENDOR: Mozilla
    
    >>>
    
  • The OS

    PowerShell 7.6.3-0-ga1994c19a0badf524d714cacffe7986c5c8613c6
    PS /home/beedell.roke_julian_lockhart> hostnamectl status --json=short | yq -P
    Hostname: beedell.rokeJulianLockhart.tf81nk
    StaticHostname: beedell.rokeJulianLockhart.tf81nk
    PrettyHostname: beedell.roke_julian_lockhart.desktop.tf81nk
    DefaultHostname: fedora
    HostnameSource: static
    IconName: computer-desktop
    Chassis: desktop
    ChassisAssetTag: null
    Deployment: null
    Location: null
    KernelName: Linux
    KernelRelease: 7.2.10-200.fc44.x86_64
    KernelVersion: '#1 SMP PREEMPT_DYNAMIC Sun Oct 11 18:11:51 UTC 2026'
    OperatingSystemPrettyName: Fedora Linux 44 (KDE Plasma Desktop Edition)
    OperatingSystemCPEName: cpe:/o:fedoraproject:fedora:44
    OperatingSystemHomeURL: https://fedoraproject.org/
    OperatingSystemSupportEnd: 1810684800000000
    OperatingSystemReleaseData:
      - NAME=Fedora Linux
      - VERSION=44 (KDE Plasma Desktop Edition)
      - RELEASE_TYPE=stable
      - ID=fedora
      - VERSION_ID=44
      - VERSION_CODENAME=
      - PRETTY_NAME=Fedora Linux 44 (KDE Plasma Desktop Edition)
      - ANSI_COLOR=0;38;2;60;110;180
      - LOGO=fedora-logo-icon
      - CPE_NAME=cpe:/o:fedoraproject:fedora:44
      - DEFAULT_HOSTNAME=fedora
      - HOME_URL=https://fedoraproject.org/
      - DOCUMENTATION_URL=https://docs.fedoraproject.org/en-US/fedora/f44/
      - SUPPORT_URL=https://ask.fedoraproject.org/
      - BUG_REPORT_URL=https://bugzilla.redhat.com/
      - REDHAT_BUGZILLA_PRODUCT=Fedora
      - REDHAT_BUGZILLA_PRODUCT_VERSION=44
      - REDHAT_SUPPORT_PRODUCT=Fedora
      - REDHAT_SUPPORT_PRODUCT_VERSION=44
      - SUPPORT_END=2027-05-19
      - VARIANT=KDE Plasma Desktop Edition
      - VARIANT_ID=kde
    OperatingSystemImageID: null
    OperatingSystemImageVersion: null
    MachineInformationData:
      - PRETTY_HOSTNAME=beedell.roke_julian_lockhart.desktop.tf81nk
    HardwareVendor: ASRock
    HardwareModel: X670E Taichi
    HardwareSerial: null
    HardwareSKU: null
    HardwareVersion: null
    FirmwareVersion: "3.08"
    FirmwareVendor: American Megatrends International, LLC.
    FirmwareDate: 1726704000000000
    MachineID: cab57430e115444dae142931aa92a5c5
    BootID: 245809bfcdb54b55beeae73ff90a7ac4
    ProductUUID: null
    VSockCID: null
    
  • The Desktop Environment

    PowerShell 7.6.3-0-ga1994c19a0badf524d714cacffe7986c5c8613c6
    PS /home/beedell.roke_julian_lockhart> kinfo | yq -P
    Operating System: Fedora Linux 44
    KDE Plasma Version: 6.7.5
    KDE Frameworks Version: 6.31.0
    Qt Version: 6.11.2
    Kernel Version: 7.2.10-200.fc44.x86_64 (64-bit)
    Graphics Platform: Wayland
    Processors: 12 ร— AMD Ryzen 5 7600X 6-Core Processor
    Memory: 32 GiB of RAM (30.4 GiB usable)
    Graphics Processor 1: AMD Radeon RX 5700
    Graphics Processor 2: AMD Ryzen 5 7600X 6-Core Processor
    

This relates to id=2051153#c5, as the cause is identical.

See Also: → 2051153

The Bugbug bot thinks this bug should belong to the 'Core::Widget: Gtk' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Component: Untriaged → Widget: Gtk
Product: Firefox → Core
See Also: → 2081619

Thanks for the report. Could you try mozregression?
https://mozilla.github.io/mozregression/quickstart.html
This will help identify the bug that introduced the problem.

If you know when you first noticed it, then you can narrow down the run.
Date example
mozregression --good 2026-10-11 --bad 2026-10-14
Version example
mozregression --good 159 --bad 160

Flags: needinfo?(zn7esutb)
Severity: -- → S2

I reproduced it on a VM (Fedora 44 + KDE). And, it produced the same symptom of bug 2081619 if Firefox opening an English page without translation popup.

I tracked this down to Bug 2080307 (2dd0b125686c). Reverting it on top of current central fixes the issue on Fedora 44 KDE. The same tree with the patch put back still fails.

What happens:

  1. With session restore (browser.startup.page=3), nsWindowWayland::NativeShow() doesn't show the main window until session store calls MoveToWorkspace(). GTK hasn't created the xdg_toplevel yet.
  2. Layout runs during that wait and calls UpdateOpaqueRegionInternal(). Since Bug 2080307 that goes through SetToplevelOpaqueRegion() -> ForceToplevelCommit(). Moving the helper subsurface makes GTK commit the toplevel wl_surface, which still has no role.
  3. GTK's on_frame_clock_after_paint() adds a wl_surface.frame() request to that commit and freezes the window's frame clock until the callback arrives.
  4. KWin sends frame callbacks only after it draws a mapped window's surface, and this surface isn't a window yet, so the callback never comes.
  5. When the window is finally shown, GTK creates the toplevel and acks the first configure. The paint after that is requested but never runs, because the clock is still frozen. So no buffer is ever attached to the toplevel surface, and KWin never maps the window.

This only happens with a GTK that has the xdg-toplevel-realized patch (Fedora 44's gtk3, upstream gtk-3-24 after 3.24.52), since IsSessionRestoreSupported() needs that signal. GNOME has session restore disabled.

From WAYLAND_DEBUG=1, on the toplevel surface:

-> wl_surface#71.set_opaque_region(wl_region#61)
-> wl_surface#71.frame(new id wl_callback#80)
-> wl_surface#71.commit()                         <- before get_xdg_surface
-> xdg_wm_base#28.get_xdg_surface(new id xdg_surface#61, wl_surface#71)
-> xdg_surface#61.get_toplevel(new id xdg_toplevel#84)
   xdg_surface#61.configure(41)
-> xdg_surface#61.ack_configure(41)
   (no wl_callback#80.done, no wl_surface#71.attach afterwards)

The translation panel still shows because popups don't wait for session restore. This is also the cause of Bug 2081619. As a workaround, widget.wayland.session-management.enabled=false fixes it.

Duplicate of this bug: 2081619

Ah, patched gtk3 wasn't on my bingo card. Ok, we should probably back out bug 2080307 for now.

Flags: needinfo?(aryx.bugmail)
Keywords: regression
Regressed by: 2080307
Flags: needinfo?(stransky)

Set release status flags based on info from the regressing bug 2080307

regressor has been backed out

Flags: needinfo?(aryx.bugmail)

Setting Fx160 to fixed since the regressor was backed out for Fx160 nightly

Flags: needinfo?(zn7esutb)
Flags: needinfo?(stransky)

:Davis.pies1, since you are the author of the regressor, bug 2080307, could you take a look?

For more information, please visit BugBot documentation.

Flags: needinfo?(Davis.pies1)

Thanks for tracking this down, and sorry for the breakage. The analysis in comment 5 matches the code: SetToplevelOpaqueRegion() calls ForceToplevelCommit() without checking whether the toplevel has been shown, so with session restore it commits a wl_surface that has no xdg role yet. My testing was on Arch, whose gtk3 3.24.52 doesn't carry the xdg-toplevel-realized patch, so IsSessionRestoreSupported() was false and this path never ran.

I've updated D332610 so SetToplevelOpaqueRegion() returns early while the window isn't mapped (mIsMapped). gdk_window_set_opaque_region() is still called, so GTK sends the cached region with the first paint after the window is shown. The existing ForceToplevelCommit() callers (MozContainerWayland, IMContextWrapper) are unchanged.

I haven't tried to reproduce this myself. :sinker, if you have time, could you check whether the updated patch fixes it on your Fedora 44 KDE VM?

Flags: needinfo?(Davis.pies1) → needinfo?(thinker.li)

yes, it fixes on the VM.

Flags: needinfo?(thinker.li)

Could you try mozregression?

Unfortunately, --good 2026-10-09 --bad 2026-10-15 was useless.

An updated version for Bug 2080307 landed so can you re-test please on latest nightly?
https://fedoraproject.org/wiki/How_to_debug_Firefox_problems#Testing_Mozilla_Nightly_binaries
Thanks.

can you re-test please on latest nightly?

If you're asking for me to, it has been resolved (in nightly-0:160.0a1-20261015161046), so I shall close this.

Status: NEW → RESOLVED
Closed: 1 month ago
Resolution: --- → FIXED
See Also: 2081619 →
QA Whiteboard: [qa-triage-done-c161/b160]
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: