Fedora Account System
Red Hat Associate
Red Hat Customer
Description of problem: Malcontent prevents launch of flatpak applications even when not enabled, and cannot be removed. Attempts to configure malcontent via UI crash Gnome Control Center Version-Release number of selected component (if applicable): * Flatpak 1.13 * Fedora Linux Rawhide.20220715.n.0 (Silverblue Prerelease) * Gnome 42.2 How reproducible: 100% Steps to Reproduce: 1. Try to launch an older Flatpak or any Flatpak not configured to support Malcontent on current rawhide, such as Thunderbird. 2. Get: "error: Not allowed to query parental controls data for user 1000" 3. Look into configuring parental controls by going to Gnome Control Panel --User OR: launch malcontent-control 4. Control panel UI locks up and crashes Currently, I have no way to use thunderbird or several other flatpaks on Fedora. Expected results: A. Malcontent should not be blocking flatpaks if it's not configured B. Attempts to configure malcontent shouldn't crash OR: given the number of terrible bugs in malcontent, maybe we should take it out of Fedora? It doesn't seem to work for actual parental control, and it does seem to break a lot of things.
(filed against gnome-shell because I've been told that malcontent is now part of gnome-shell)
(In reply to Josh Berkus from comment #0) > 4. Control panel UI locks up and crashes Let's start with the crash. Please provide a backtrace using 'coredumpctl gdb'.
And, now, not crashing. Instead I get some actual error output: (gnome-control-center:134682): user-accounts-cc-panel-WARNING **: 09:01:49.509: Error retrieving app filter for user (null): User 4294967295 does not exist (gnome-control-center:134682): user-accounts-cc-panel-CRITICAL **: 09:02:01.361: cc_avatar_chooser_set_user: assertion 'self != NULL' failed I would attach a screenshot, but the screenshot tool also is not launching. Previously, control panel was freezing and crashing when I did this. Now it seems to be operating ... just inaccessible.
More info: I rebased to Fedora 36 release, and all of this worked fine. So the problems are due to some change introduced since 36.0.
(In reply to Josh Berkus from comment #3) > And, now, not crashing. Instead I get some actual error output: > > > (gnome-control-center:134682): user-accounts-cc-panel-WARNING **: > 09:01:49.509: Error retrieving app filter for user (null): User 4294967295 > does not exist > > (gnome-control-center:134682): user-accounts-cc-panel-CRITICAL **: > 09:02:01.361: cc_avatar_chooser_set_user: assertion 'self != NULL' failed You can still get backtraces. Use G_DEBUG=fatal-warnings to catch the first one, which looks relevant here. You can also use G_DEBUG=fatal-criticals to report the second one in a separate bug report.
Will do. This will be a little slow, though, because this is my work machine and I need to reboot it into the unusable configuration in order to test things. In the meantime, I'm going to split this bug and file the "Not allowed to query parental controls" as a new bug, since it looks like they're not related after all.
Hey! Hitting the same bug on rawhide, it is still present, and quite concerning since branching is well... today. Also, maybe related, but I am unable to unlock any setting in gnome-control-center.
This bug appears to have been reported against 'rawhide' during the Fedora Linux 37 development cycle. Changing version to 37.
Huh, a restart solved this for me.
The blocking, or the control panel part?
Still need to see a backtrace for that first warning.
I hit it again, was able to "crash it", and I am now reporting it through ABRT. Although, my internet is painfully slow where I am right now, so you'll have to wait for a bit.
Shouldn't this be reported to malcontent instead, btw? It seems for me more like malcontent failing at some point, and blocking everything by default.
Created attachment 1905467 [details] backtrace So, I wasn't able to generate an ABRT ticket, but, here's the report : https://retrace.fedoraproject.org/faf/reports/494617/ Also, I attached the backtrace.
(In reply to Lyes Saadi from comment #13) > Shouldn't this be reported to malcontent instead, btw? It seems for me more > like malcontent failing at some point, and blocking everything by default. I'm not sure what's wrong yet, but I don't see evidence of a problem with malcontent. Instead, it looks like gnome-control-center is going a bit crazy. It's trying to look up the malcontent app filter for uid 2^32 - 1, i.e. UINT_MAX. Surely it shouldn't be doing that. The backtrace you attached gives us enough for an upstream bug report. Would you mind reporting this to https://gitlab.gnome.org/GNOME/gnome-control-center/issues?
Ok! Will report this upstream! But, since flatpak has the same issue, I guess it's either some shared code or malcontent? Should we file for the flatpak bug a separate issue?
This does not look like code that would be shared. Is flatpak printing warnings or criticals? If so, can you post another backtrace?
Yeah, I guess before submitting anything, I should recapitulate what happens in this bug report since it goes a bit everywhere. What happens is that, at some point when normally using Fedora, a lot of things stop working at once, and they all point to malcontent somehow. And the only fix is to reboot. 1. It becomes impossible to run _any_ flatpak app (the most urgent IMO, and maybe a blocker bug): ``` error: Not allowed to query parental controls data for user 1000 ``` That's the only error. 2. In Settings, a lot of things stop working, the user is unable to unlock settings in multiple panels, like the user panel, the printer panel, the date and time panels, unable to fetch apps for the applications panel. Here are all the workings I got travelling through the Settings app: ``` bluetooth-cc-panel-Message: 17:33:38.750: BluetoothHardwareAirplaneMode: 0 (gnome-control-center:119039): cc-applications-panel-WARNING **: 17:33:43.342: Error retrieving app filter: Not allowed to query parental controls data for user 1000 (gnome-control-center:119039): GLib-CRITICAL **: 17:33:47.287: g_atomic_ref_count_dec: assertion 'old_value > 0' failed (gnome-control-center:119039): microphone-cc-panel-WARNING **: 17:33:48.190: Failed fetch permissions from flatpak permission store: GDBus.Error:org.freedesktop.portal.Error.NotFound: No entry for microphone GLib-GIO: _g_io_module_get_default: Found default implementation dconf (DConfSettingsBackend) for ?gsettings-backend?GoaBackend: Loading all providers: GoaBackend: - googleGoaBackend: - owncloudGoaBackend: - windows_liveGoaBackend: - exchangeGoaBackend: - fedoraGoaBackend: activated Fedora providerGoaBackend: - imap_smtpGoaBackend: - kerberosGoaBackend: activated kerberos provider (gnome-control-center:119039): user-accounts-cc-panel-CRITICAL **: 17:34:11.734: cc_avatar_chooser_set_user: assertion 'self != NULL' failed (gnome-control-center:119039): user-accounts-cc-panel-WARNING **: 17:34:11.925: Impossible to update fingerprint manager state: GDBus.Error:net.reactivated.Fprint.Error.PermissionDenied: Not Authorized: net.reactivated.fprint.device.verify (gnome-control-center:119039): datetime-cc-panel-WARNING **: 17:34:13.622: Could not set system time: GDBus.Error:org.freedesktop.timedate1.AutomaticTimeSyncEnabled: AutomaticTimeSyncEnabled: Automatic time synchronization is enabled ``` Some are unrelated, but most are permission issues. Other panels do not work but do not print a warning either, like the printer panel, where I am unable to unlock settings. And, both things happen at the same time, and both seem related to malcontent or permissions. I could try and force remove malcontent through RPM (but breaking my system). Want me to try?
(In reply to Lyes Saadi from comment #18) > Yeah, I guess before submitting anything, I should recapitulate what happens > in this bug report since it goes a bit everywhere. > > What happens is that, at some point when normally using Fedora, a lot of > things stop working at once, and they all point to malcontent somehow. And > the only fix is to reboot. > > 1. It becomes impossible to run _any_ flatpak app (the most urgent IMO, and > maybe a blocker bug): > > ``` > error: Not allowed to query parental controls data for user 1000 > ``` > > That's the only error. Can you provide a backtrace for this error, please? > 2. In Settings, a lot of things stop working, the user is unable to unlock > settings in multiple panels, like the user panel, the printer panel, the > date and time panels, unable to fetch apps for the applications panel. Here > are all the workings I got travelling through the Settings app: > > ``` > bluetooth-cc-panel-Message: 17:33:38.750: BluetoothHardwareAirplaneMode: 0 > > (gnome-control-center:119039): cc-applications-panel-WARNING **: > 17:33:43.342: Error retrieving app filter: Not allowed to query parental > controls data for user 1000 Let's focus on the first errors. This one is *almost* the same as what you posted a backtrace for, except this time it shows a failure for user 1000 rather than user 4294967295. So that's very different. The other errors look like all D-Bus services have become unresponsive. And you're saying this happens not just for the gnome-control-center process, but also for the flatpak process too? It kinda seems like your entire session bus is nonresponsive, which probably has nothing to do with malcontent? Does this happen randomly, or are you able to consistently reproduce it?
> Can you provide a backtrace for this error, please? Well, how? > It kinda seems like your entire session bus is nonresponsive, which probably has nothing to do with malcontent? How could I check that? Honestly, this bug just reminded me of another one, but I am unable to find it... Mainly a bug with nvidia-powerd spamming dbus. I had removed it, but, rebasing to rawhide seems to have brought it back? Would it be the same bug by any chance? John Berkus, do you have the xorg-x11-drv-nvidia-power package by any chance? > Does this happen randomly, or are you able to consistently reproduce it? Randomly, but consistently after some time. By that I mean it activates at a random time, but always activate after some time.
> Honestly, this bug just reminded me of another one, but I am unable to find it... Mainly a bug with nvidia-powerd spamming dbus. I had removed it, but, rebasing to rawhide seems to have brought it back? Would it be the same bug by any chance? John Berkus, do you have the xorg-x11-drv-nvidia-power package by any chance? I doesn't seem to be like this. I do not have the same circumstances of the bug, and dbus seems to still be working (at least I see activity with dbus-monitor)?
Ok, so, I had to restart the session, and I noticed that actually it was GNOME entirely which was breaking down. Restarting the session resulted in an entirely non-functional gnome-shell. I rebooted and removed nvidia-power just in case, I will see if I still hit this bug or not.
Ok! It's back, so, it's not nvidia! (and the nvidia-powerd bug seems to have been fixed anyway). Ready to do any other test needed.
Another bug I've discovered : the inability to reboot/shutdown properly from the user session directly. On the power button, there's no restart/shutdown button, and if I try through the command line, I have this message : ``` Failed to reboot system via logind: Access denied ``` The only way to reboot is to use sudo.
I think your entire session and/or system bus is hosed. I guess you can try running dbus-monitor to maybe see what is going on? That is going to be a *huge* amount of spam, though... but maybe you'll spot something suspicious? If nothing is obviously wrong, we might need to attach gdb to dbus-broker.
I rebooted, but when I tried running dbus-monitor, it wasn't spamming. There were new messages every second or so. So, to be expected for a system bus.
Anyway, even though we have not found the underlying cause of what's going wrong for you, there are still multiple gnome-control-center bugs here. gnome-control-center shouldn't freeze. gnome-control-center shouldn't emit criticals. Criticals are always programmer error: a D-Bus service being down should not be enough to cause them.
I think it would be useful to see a full journal from around the time that this breakage happens.
The issue is that it is hard to know exactly when the breakage happens. But, I could send the entire file of my last boot, which is quite "short" (only ~26 000 lines). The issue is that it is kinda revealing a huge chunk of home folder content due to some tracker-miner-fs logs (which represents 11 000 of the lines of the file, btw). Should we make this bug report private or create a new private one for the sake of this bug report?
Comment on attachment 1905615 [details] journald logs I'll make the attachment private so that it's only visible to Red Hat. If any other Fedora contributor wants to see it, just ask. I don't think there's any reason to restrict access to the rest of this bug.
Matthias noticed: août 15 11:01:31 lyes-pc polkitd[1296]: Error evaluating authorization rules août 15 11:01:31 lyes-pc polkitd[1296]: Error evaluating authorization rules which is bad. Can you please show us what files you have in /etc/polkit-1/rules.d and /usr/share/polkit-1/rules.d? Maybe something suspicious will be there? Also 'rpm -qi polkit' to show what version and release you have installed.
> I'll make the attachment private so that it's only visible to Red Hat. If any other Fedora contributor wants to see it, just ask. Thanks, didn't know that was possible. > Can you please show us what files you have in /etc/polkit-1/rules.d and /usr/share/polkit-1/rules.d? ``` [root@lyes-pc lyes]# cd /etc/polkit-1/rules.d/ [root@lyes-pc rules.d]# ls 49-polkit-pkla-compat.rules 50-default.rules [root@lyes-pc rules.d]# cat * polkit.addAdminRule(function(action, subject) { //polkit.log('Starting pkla-admin-identities\n'); // Let exception, if any, propagate to the JS authority var res = polkit.spawn(['/usr/bin/pkla-admin-identities']); //polkit.log('Got "' + res.replace(/\n/g, '\\n') + '"\n'); if (res == '') return null; var identities = res.split('\n'); //polkit.log('Identities: ' + identities.join(',') + '\n'); if (identities[identities.length - 1] == '') identities.pop() //polkit.log('Returning: ' + identities.join(',') + '\n'); return identities; }); polkit.addRule(function(action, subject) { var params = ['/usr/bin/pkla-check-authorization', subject.user, subject.local ? 'true' : 'false', subject.active ? 'true' : 'false', action.id]; //polkit.log('Starting ' + params.join(' ') + '\n'); var res = polkit.spawn(params); //polkit.log('Got "' + res.replace(/\n/g, '\\n') + '"\n'); if (res == '') return null; return res.replace(/\n$/, ''); }); /* -*- mode: js; js-indent-level: 4; indent-tabs-mode: nil -*- */ // DO NOT EDIT THIS FILE, it will be overwritten on update // // Default rules for polkit // // See the polkit(8) man page for more information // about configuring polkit. polkit.addAdminRule(function(action, subject) { return ["unix-group:wheel"]; }); [root@lyes-pc rules.d]# cd /usr/share/polkit-1/rules.d/ [root@lyes-pc rules.d]# ls 20-gnome-initial-setup.rules 50-libvirt.rules com.endlessm.ParentalControls.rules gnome-control-center.rules org.a11y.brlapi.rules org.fedoraproject.FirewallD1.desktop.rules.choice org.fedoraproject.FirewallD1.rules org.fedoraproject.thirdparty.rules org.freedesktop.bolt.rules org.freedesktop.Flatpak.rules org.freedesktop.fwupd.rules org.freedesktop.GeoClue2.rules org.freedesktop.packagekit.rules org.gtk.vfs.file-operations.rules systemd-networkd.rules [root@lyes-pc rules.d]# cat * // -*- mode: js; js-indent-level: 4; indent-tabs-mode: nil -*- // // DO NOT EDIT THIS FILE, it will be overwritten on update. // // Allow the gnome-initial-setup user to do certain actions without // being interrupted by password dialogs polkit.addRule(function(action, subject) { if (subject.user !== 'gnome-initial-setup') return undefined; var actionMatches = (action.id.indexOf('org.freedesktop.hostname1.') === 0 || action.id.indexOf('org.freedesktop.NetworkManager.') === 0 || action.id.indexOf('org.freedesktop.locale1.') === 0 || action.id.indexOf('org.freedesktop.accounts.') === 0 || action.id.indexOf('org.freedesktop.timedate1.') === 0 || action.id.indexOf('org.freedesktop.realmd.') === 0 || action.id.indexOf('com.endlessm.ParentalControls.') === 0 || action.id.indexOf('org.fedoraproject.thirdparty.') === 0); if (actionMatches) { if (subject.local) return 'yes'; else return 'auth_admin'; } return undefined; }); // Allow any user in the 'libvirt' group to connect to system libvirtd // without entering a password. polkit.addRule(function(action, subject) { if (action.id == "org.libvirt.unix.manage" && subject.isInGroup("libvirt")) { return polkit.Result.YES; } }); /* * Copyright © 2019 Endless Mobile, Inc. * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * * This library is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public * License along with this library; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA * * Authors: * - Philip Withnall <withnall> */ polkit.addRule(function(action, subject) { /* Allow administrators to read parental controls (for any account) without * needing an additional polkit authorisation dialogue. */ if ((action.id == "com.endlessm.ParentalControls.AppFilter.ReadOwn" || action.id == "com.endlessm.ParentalControls.AppFilter.ReadAny" || action.id == "com.endlessm.ParentalControls.SessionLimits.ReadOwn" || action.id == "com.endlessm.ParentalControls.SessionLimits.ReadAny") && subject.active && subject.local && subject.isInGroup("wheel")) { return polkit.Result.YES; } return polkit.Result.NOT_HANDLED; }); polkit.addRule(function(action, subject) { if ((action.id == "org.freedesktop.locale1.set-locale" || action.id == "org.freedesktop.locale1.set-keyboard" || action.id == "org.freedesktop.ModemManager1.Device.Control" || action.id == "org.freedesktop.hostname1.set-static-hostname" || action.id == "org.freedesktop.hostname1.set-hostname" || action.id == "org.gnome.controlcenter.datetime.configure") && subject.local && subject.active && subject.isInGroup ("wheel")) { return polkit.Result.YES; } }); polkit.addRule(function(action, subject) { if (action.id == "org.a11y.brlapi.write-display") { if (subject.isInGroup("brlapi")) { return polkit.Result.YES; } } }); // firewalld authorizations/policy for the wheel group. // // DO NOT EDIT THIS FILE, it will be overwritten on update. // // Allow users in the wheel group to use firewalld without being // interrupted by a password dialog polkit.addRule(function(action, subject) { if ((action.id == "org.fedoraproject.FirewallD1.config" || action.id == "org.fedoraproject.FirewallD1.direct" || action.id == "org.fedoraproject.FirewallD1.ipset" || action.id == "org.fedoraproject.FirewallD1.policy" || action.id == "org.fedoraproject.FirewallD1.zone") && subject.active == true && subject.local == true && subject.isInGroup("wheel")) { return polkit.Result.YES; } }); // firewalld authorizations/policy for the wheel group. // // DO NOT EDIT THIS FILE, it will be overwritten on update. // // Allow users in the wheel group to use firewalld without being // interrupted by a password dialog polkit.addRule(function(action, subject) { if ((action.id == "org.fedoraproject.FirewallD1.config" || action.id == "org.fedoraproject.FirewallD1.direct" || action.id == "org.fedoraproject.FirewallD1.ipset" || action.id == "org.fedoraproject.FirewallD1.policy" || action.id == "org.fedoraproject.FirewallD1.zone") && subject.active == true && subject.local == true && subject.isInGroup("wheel")) { return polkit.Result.YES; } }); // Allow the initial opt-out from the "unset" state to happen without an // authentication dialog. polkit.addRule(function(action, subject) { if (action.id == "org.fedoraproject.thirdparty.opt-out" && subject.active == true && subject.local == true && subject.isInGroup("wheel")) { return polkit.Result.YES; } return polkit.Result.NOT_HANDLED; }); // -*- mode: js2 -*- polkit.addRule(function(action, subject) { if ((action.id === "org.freedesktop.bolt.enroll" || action.id === "org.freedesktop.bolt.authorize" || action.id === "org.freedesktop.bolt.manage") && subject.active === true && subject.local === true && subject.isInGroup("wheel")) { return polkit.Result.YES; } }); polkit.addRule(function(action, subject) { if ((action.id == "org.freedesktop.Flatpak.app-install" || action.id == "org.freedesktop.Flatpak.runtime-install"|| action.id == "org.freedesktop.Flatpak.app-uninstall" || action.id == "org.freedesktop.Flatpak.runtime-uninstall" || action.id == "org.freedesktop.Flatpak.modify-repo") && subject.active == true && subject.local == true && subject.isInGroup("wheel")) { return polkit.Result.YES; } return polkit.Result.NOT_HANDLED; }); polkit.addRule(function(action, subject) { if (action.id == "org.freedesktop.Flatpak.override-parental-controls") { return polkit.Result.AUTH_ADMIN; } return polkit.Result.NOT_HANDLED; }); polkit.addRule(function(action, subject) { if (action.id == "org.freedesktop.fwupd.update-internal" && subject.active == true && subject.local == true && subject.isInGroup("wheel")) { return polkit.Result.YES; } }); polkit.addRule(function(action, subject) { if ((action.id == "org.freedesktop.ModemManager1.Device.Control" || action.id == "org.freedesktop.ModemManager1.Location") && subject.user == "geoclue") { return polkit.Result.YES; } }); polkit.addRule(function(action, subject) { if ((action.id == "org.freedesktop.packagekit.package-install" || action.id == "org.freedesktop.packagekit.package-remove") && subject.active == true && subject.local == true && subject.isInGroup("wheel")) { return polkit.Result.YES; } }); // Allows users belonging to privileged group to start gvfsd-admin without // authorization. This prevents redundant password prompt when starting // gvfsd-admin. The gvfsd-admin causes another password prompt to be shown // for each client process using the different action id and for the subject // based on the client process. polkit.addRule(function(action, subject) { if ((action.id == "org.gtk.vfs.file-operations-helper") && subject.local && subject.active && subject.isInGroup ("wheel")) { return polkit.Result.YES; } }); // This file is part of systemd. // See systemd-networkd.service(8) and polkit(8) for more information. // Allow systemd-networkd to set timezone, get product UUID, // and transient hostname polkit.addRule(function(action, subject) { if ((action.id == "org.freedesktop.hostname1.set-hostname" || action.id == "org.freedesktop.hostname1.get-product-uuid" || action.id == "org.freedesktop.timedate1.set-timezone") && subject.user == "systemd-network") { return polkit.Result.YES; } }); ``` > Also 'rpm -qi polkit' to show what version and release you have installed. ``` ❯ rpm -qi polkit Name : polkit Version : 121 Release : 3.fc37 Architecture: x86_64 Install Date: lun. 08 août 2022 19:37:55 Group : Unspecified Size : 463123 License : LGPLv2+ Signature : RSA/SHA256, mar. 02 août 2022 16:25:01, Key ID f55ad3fb5323552a Source RPM : polkit-121-3.fc37.src.rpm Build Date : mar. 02 août 2022 16:22:32 Build Host : buildvm-x86-17.iad2.fedoraproject.org Packager : Fedora Project ```
I've also noticed: polkitd[1296]: Terminating runaway script and that seems to be the cause for the subsequent polkit errors. I have no idea how this could happen though. Unless polkit gets confused about timestamps, because of: chronyd[1455]: Forward time jump detected!
At first I suspected mozjs -> duktape, but 121-3.fc37 actually has the duktape backend disabled. It's not reenabled until 121-4.fc37. Also, this bug was reported one day before Fedora switched to duktape. I wonder if it's a red herring. :/ If polkit needs to perform any D-Bus calls when executing JavaScript, then those will time out and it would look just like runaway JS. Hard to know if this is the actual problem or just another symptom.
> polkit needs to perform any D-Bus calls when executing JavaScript I don't think there is anything in polkit rules that makes dbus calls though.
I suspect this may be caused by the same fd leak as https://bugzilla.redhat.com/show_bug.cgi?id=2109145 . Could you test the fixed glib2 build, and see if it helps with this issue? https://koji.fedoraproject.org/koji/buildinfo?buildID=2046028
Hello Kalev! That description is indeed very similar to what's happening. My guess is that they just didn't test Flatpaks and gnome-control-center xD!
Never hit that bug since. I'd say it's safe to close this :)!
Heh, that was a tricky one.... *** This bug has been marked as a duplicate of bug 2109145 ***
I can no longer reproduce this bug either, just to confirm.