Convert assert to if

[tsdgeos]

ddc_open_display has this code

TRACED_ASSERT(bus_info); // need to convert to a test?

On KDE we are getting asserts with this trace

Thread 1270936 Crashed:
0 libc.so.6 0x7f48d3881e9c __pthread_kill_implementation (source-5f134746-#usr#src#debug#glibc-2.41-11.fc42.x86_64#nptl#pthread_kill.c:45)
1 libc.so.6 0x7f48d3881f63 __pthread_kill_internal (source-5f134746-#usr#src#debug#glibc-2.41-11.fc42.x86_64#nptl#pthread_kill.c:90)
2 libc.so.6 0x7f48d3827f3e __GI_raise (source-5b0e7cae-#usr#src#debug#glibc-2.41-11.fc42.x86_64#signal#..#sysdeps#posix#raise.c:27)
3 libc.so.6 0x7f48d380f6d0 __GI_abort (source-cde0a3ef-#usr#src#debug#glibc-2.41-11.fc42.x86_64#stdlib#abort.c:78)
4 libc.so.6 0x7f48d380f639 __assert_fail_base (source-41a13c79-#usr#src#debug#glibc-2.41-11.fc42.x86_64#assert#assert.c:119)
5 libc.so.6 0x7f48d38200af __assert_fail (source-41a13c79-#usr#src#debug#glibc-2.41-11.fc42.x86_64#assert#assert.c:128)
6 libddcutil.so.5 0x7f48d34b0b6f ddc_open_display (source-63f8c033-#usr#src#debug#ddcutil-2.2.1-1.fc42.x86_64#src#ddc#ddc_packet_io.c:333)
7 libddcutil.so.5 0x7f48d352d9de ddci_open_display3 (source-b50cbc0d-#usr#src#debug#ddcutil-2.2.1-1.fc42.x86_64#src#libmain#api_displays.c:621)
8 libddcutil.so.5 0x7f48d3533f81 ddca_open_display2 (source-b50cbc0d-#usr#src#debug#ddcutil-2.2.1-1.fc42.x86_64#src#libmain#api_displays.c:653)
9 libpowerdevilcore.so.2 0x7f48d528becc BrightnessWorker::ddcSetBrightness (source-77a70433-#usr#src#debug#powerdevil-6.4.5-1.fc42.x86_64#daemon#controllers#ddcutildisplay.cpp:241)

[1270936] Assertion failed: "bus_info" in file ddc_packet_io.c at line 332

So it would seem that yes, it makes sense to turn that assert into a check.

[rockowitz]

What version of ddcutil does this occur with? At first glance it looks like a bug introduced in 2.2.0 that was fixed in 2.2.1. See issue #528.

[tsdgeos]

As per the line numbers it seems to be 2.2.1 (ddcutil-2.2.1-1.fc42.x86_64)

[rockowitz]

It should be in the system log. journalctl -t org_kde_powerdevil, look for line containing "Initializing libddcutil".

[rockowitz]

Also, can you build from branch 2.2.2-dev and confirm the problem exists with that branch?

[rockowitz]

Also if possible:

• output of journalctl -t org_kde_powerdevil, starting with the "Initializing libddcutil" line, that show the error
• instructions on how to reproduce the error

[tsdgeos]

This is not a crash I have, this is a crash our users are having, sadly i can not provide any of that.

[rockowitz]

Can you ask users to submit the output of journalctl -t org_kde_powerdevil described above? Github users can submit it (as an attachment) to this issue; anyone can send it to my email address rockowitz@minsoftcom. Without knowing what is going on, all I can do is convert the assert() to an "Internal Error" status code returned (ultimately) to PowerDevil.

[tsdgeos]

Unfortunately we can't contact the user. The backtrace i pasted is all we have, the crash was reported automatically via the crash reporter, we don't have a way to contact the user.

all I can do is convert the assert() to an "Internal Error" status code returned (ultimately) to PowerDevil.

Better than crashing i guess?

[rockowitz]

Per commit 7fe3ff5, the assert()s in question have been converted to if tests that will ultimately cause the API call in question to write some information to the system log and return status DDCRC_INTERNAL_ERROR. Not very satisfactory, but absent more information as to what is going on that's the best I can do for now. How common is this failure? I suspect it occurs somehow as a result of display disconnection/connection.

[arch-btw]

@tsdgeos thank you for reporting this and @rockowitz thank you for looking into it.

I think this might be a crash that I'm experiencing myself too. Here the requested information:

journalctl -t org_kde_powerdevil :

Oct 28 19:37:24 arch org_kde_powerdevil[2241]: [  2241] libddcutil recheck thread 0x55aa07250c50 started
Oct 28 19:37:24 arch org_kde_powerdevil[2241]: [  2241] libddcutil watch thread 0x55aa0718c3b0 started
Oct 28 19:37:24 arch org_kde_powerdevil[2241]: [  2241] Display redetection finished.
Oct 28 19:37:24 arch org_kde_powerdevil[2241]: [  2241] Unquiescing libddcutil API...
Oct 28 19:37:24 arch org_kde_powerdevil[2241]: [  4959] (dw_recheck_displays_func) Recheck interval: Sleeping for 200 milliseconds
Oct 28 23:34:00 arch org_kde_powerdevil[2241]: There are no outputs - creating placeholder screen
Oct 28 23:34:01 arch org_kde_powerdevil[2241]: [  4961] (ddc_open_display) Delay before rechecking attribute status: Sleeping for 1000 milliseconds
Oct 28 23:34:01 arch org_kde_powerdevil[2241]: [  4960] Removing connected display on bus 1
Oct 28 23:34:01 arch org_kde_powerdevil[2241]: [  4960] Emitting DDCA_Display_Status_Event[14980.110:  DDCA_EVENT_DISPLAY_DISCONNECTED, card1-HDMI-A-1, dref: DDCA_Display_Ref[3], io_pa>
Oct 28 23:34:01 arch org_kde_powerdevil[2241]: [  4960] Starting 1 callback threads
Oct 28 23:34:01 arch org_kde_powerdevil[2241]: [  4960] libddcutil callback thread 0x7fd438014190 started
Oct 28 23:34:01 arch org_kde_powerdevil[2241]: [  4960] Started 1 event callback thread(s)
Oct 28 23:34:01 arch org_kde_powerdevil[2241]: [  5640] Invoking callback function 0x7fd4624eca00 for event DDCA_Display_Status_Event[14980.110:  DDCA_EVENT_DISPLAY_DISCONNECTED, card1>
Oct 28 23:34:01 arch org_kde_powerdevil[2241]: [  5640] Callback function 0x7fd4624eca00 for event DDCA_Display_Status_Event[14980.110:  DDCA_EVENT_DISPLAY_DISCONNECTED, card1-HDMI-A-1>
Oct 28 23:34:02 arch org_kde_powerdevil[2241]: [  4960] Adding connected display with bus 1
Oct 28 23:34:02 arch org_kde_powerdevil[2241]: [  4960] Emitting DDCA_Display_Status_Event[14980.794:  DDCA_EVENT_DISPLAY_CONNECTED, card1-HDMI-A-1, dref: DDCA_Display_Ref[4], io_path:>
Oct 28 23:34:02 arch org_kde_powerdevil[2241]: [  4960] Starting 1 callback threads
Oct 28 23:34:02 arch org_kde_powerdevil[2241]: [  4960] libddcutil callback thread 0x7fd4380403e0 started
Oct 28 23:34:02 arch org_kde_powerdevil[2241]: [  4960] Started 1 event callback thread(s)
Oct 28 23:34:02 arch org_kde_powerdevil[2241]: [  5707] Invoking callback function 0x7fd4624eca00 for event DDCA_Display_Status_Event[14980.794:  DDCA_EVENT_DISPLAY_CONNECTED, card1-HD>
Oct 28 23:34:02 arch org_kde_powerdevil[2241]: [  5707] Callback function 0x7fd4624eca00 for event DDCA_Display_Status_Event[14980.794:  DDCA_EVENT_DISPLAY_CONNECTED, card1-HDMI-A-1, d>
Oct 28 23:34:02 arch org_kde_powerdevil[2241]: [  4961] Assertion failed: "bus_info" in file ddc_packet_io.c at line 332
Oct 28 23:34:02 arch org_kde_powerdevil[2241]: org_kde_powerdevil: ddc_packet_io.c:332: ddc_open_display: Assertion `bus_info' failed.
Oct 28 23:34:03 arch org_kde_powerdevil[2241]: KCrash: Attempting to start /usr/lib/org_kde_powerdevil
Oct 28 23:34:03 arch org_kde_powerdevil[2241]: KCrash: Application 'org_kde_powerdevil' crashing... crashRecursionCounter = 2
Oct 28 23:34:03 arch org_kde_powerdevil[5708]: Starting libddcutil. library built Not set at Not set. stdout_stderr_redirected=true
Oct 28 23:34:03 arch org_kde_powerdevil[5759]: Starting libddcutil. library built Not set at Not set. stdout_stderr_redirected=true
Oct 28 23:34:03 arch org_kde_powerdevil[5759]: Initializing libddcutil.  ddcutil version: 2.2.1, shared library: /usr/lib/libddcutil.so.5.3.0
Oct 28 23:34:03 arch org_kde_powerdevil[5759]: [  5759] Options passed from client:
Oct 28 23:34:03 arch org_kde_powerdevil[5759]: [  5759] Applying combined libddcutil options:
Oct 28 23:34:04 arch org_kde_powerdevil[5759]: Library initialization complete.
Oct 28 23:34:04 arch org_kde_powerdevil[5759]: Watching for DPMS state changes unimplemented
Oct 28 23:34:04 arch org_kde_powerdevil[5759]: [  5759] Watching for DPMS state changes unimplemented
Oct 28 23:34:04 arch org_kde_powerdevil[5759]: [  5759] Watching for display connection changes, resolved watch mode = Watch_Mode_Xevent, poll loop interval = 100 millisec
Oct 28 23:34:04 arch org_kde_powerdevil[5759]: [  5759]                                          extra_stabilization_millisec: 0,  stabilization_poll_millisec: 100
Oct 28 23:34:04 arch org_kde_powerdevil[5759]: [  5759] libddcutil recheck thread 0x55b5b1d54f70 started
Oct 28 23:34:04 arch org_kde_powerdevil[5759]: [  5759] libddcutil watch thread 0x55b5b1d54490 started
Oct 28 23:34:04 arch org_kde_powerdevil[5759]: [  5777] (dw_recheck_displays_func) Recheck interval: Sleeping for 200 milliseconds
Oct 28 23:34:05 arch org_kde_powerdevil[5759]: org.kde.powerdevil.chargethresholdhelper.getthreshold failed "Charge thresholds are not supported by the kernel for this hardware"
Oct 28 23:34:05 arch org_kde_powerdevil[5759]: org.kde.powerdevil.backlighthelper.brightness failed
Oct 28 23:34:05 arch org_kde_powerdevil[5759]: Handle button events action could not check for screen configuration
Oct 28 23:34:05 arch org_kde_powerdevil[5759]: org.kde.powerdevil.chargethresholdhelper.getthreshold failed "Charge thresholds are not supported by the kernel for this hardware"

Instructions on how to reproduce the error:

Arch Linux - KDE Plasma 6 - Wayland.

It happens for me when the monitor (HDMI) is turned off due to power management settings. Although, this also causes me to have to login again, so I'm not sure if it's login related or monitor related. Suspend is disabled.

Power management settings:

[rockowitz]

@arch-btw Thank you! That is immensely helpful.

Does the failure happen when the monitor is turned off or when it is turned on again? I'd like to better understand your configuration. Please run ddcutil interrogate and submit the output as an attachment.

[arch-btw]

@rockowitz no problem at all, thank you too 👍

I have emailed you the log file for: ddcutil interrogate

I'm actually not sure if it happens during the monitor's shutdown or when I'm turning it back on. The message is there when I turn it back on but I'm not sure how long it's been there for.

Also, I wanted to mention that when I build ddcutil from the current dev branch, then the crash is gone. So it does appear to be fixed.

Thank you!