Odroid XU4 SDL Problem Emulationstation

aaronouthier

FWIW, speaking of sources, I was able to download the sources with 'apt source python3-uinput' from current repos. Not too sure how I can incorporate that into the existing build.

I've compiled stuff before, but never via apt-source.

mitu

@aaronouthier said in Odroid XU4 SDL Problem Emulationstation:

Do I still need python3-sdl2, ...

Yes, only the SDL library is compiled into a package, the rest are expected to be provided by the distribution's repostories.

aaronouthier

@mitu I was pretty sure this was the case, but wanted to be certain.

aaronouthier

Dang it!

running build
running build_py
running build_ext
building '_libsuinput' extension
arm-linux-gnueabihf-gcc -fno-strict-overflow -Wsign-compare -DNDEBUG -g -O2 -Wall -fPIC -I/usr/include/python3.12 -c libsuinput/src/suinput.c -o build/temp.linux-armv7l-cpython-312/libsuinput/src/suinput.o
libsuinput/src/suinput.c: In function ‘suinput_emit’:
libsuinput/src/suinput.c:48:28: error: ‘struct input_event’ has no member named ‘time’
   48 |         gettimeofday(&event.time, 0);
      |                            ^
error: command '/usr/bin/arm-linux-gnueabihf-gcc' failed with exit code 1

Trying to build the source package.

Was logged-in as root. I don't think that would make a difference though.

aaronouthier

@mitu Not able to install the 22.04 version either. Depends on Python3.10 or below, but Python3.12 is installed.

At this point, I think I'm going to wipe the SD card and try again with 22.04 .

aaronouthier

Currently compiling the system for xu4 under Ubuntu 22.04 Operating System. No problems yet. (Need to go knock some wood now...)

aaronouthier

Well, about 1.5 hours later, compilation finished. No errors reported during compile process that I could see, however, emulationstation crashes when run.

I tried with --debug flag, however the screen is cleared by the crash dialog that comes up, so I can't read. I'll try tee'ing it to a text file.

aaronouthier

@aaronouthier I watched a bit closer. Nothing out of the ordinary displayed when running with --debug flag. Just shows loading some xml files.

On a related note, however, Xserver does NOT seem to be running. I am attaching a tar.gz file with all of the build logs, a log of installed packages, and a log of running processes.

Hmm. Can't seem to upload it directly.
Link to files on my Google drive: https://drive.google.com/file/d/1ro_ZYzHFcI6TfbSUjzsLksvwf6rc9vMy/view?usp=sharing

mitu

@aaronouthier said in Odroid XU4 SDL Problem Emulationstation:

I tried with --debug flag, however the screen is cleared by the crash dialog that comes up, so I can't read. I'll try tee'ing it to a text file.

That would have been useful to look, the other logs not so much. You can find the same log file in $HOME/.emulationstation/es_log.txt, there's no need to tee it.
Does detection works better on 22.04 - is the SBC detected as odroid-xu ?

aaronouthier

@mitu

I checked those logs. 3 lines about mame-related .xml files.

No, still detects as "Armv7l". I used your workaround when I ran RetroPie-Setup.

I mentioned twice now that XWindows doesn't seem to be running. Very few X-related packages installed as well. This might be normal, but I don't have a frame of reference here. This was the case in 24.04 also.

mitu

@aaronouthier said in Odroid XU4 SDL Problem Emulationstation:

I checked those logs. 3 lines about mame-related .xml files.

Even when running with --debug ? There should be more information printed when debug is enabled.

No, still detects as "Armv7l". I used your workaround when I ran RetroPie-Setup.

I added some detection here, but it may not be enough (?) or you're using an earlier commit.

I mentioned twice now that XWindows doesn't seem to be running.

I think that's ok.

aaronouthier

@mitu

Even when running with --debug ? There should be more information printed when debug is enabled.

There might have been another line or two. I'm not in front of the system right now. Yes, with --debug. My point was that no errors were displayed.

I added some detection here, but it may not be enough (?) or you're using an earlier commit.

Probably hadn't pulled the latest commit then. This was 2 days ago.

I created an image file from the SD card, then tried to install Lakka just to test that out. That one doesn't boot at all, but hey, I tried. I haven't taken the time to restore the RetroPie Disk image just yet.

aaronouthier

@aaronouthier
Re-wrote the disk-image. Originally, when I cat'd the file mentioned, it showed the exact same contents listed in post 19 in this thread.

I then checked for package updates. After updating them, I again ran 'emulationstation --debug'. There was no output, but it still crashed. The es_log.txt is now empty.

I'm now recompiling everything. Also, your new commit to retropie_setup.sh still shows arm7l, with no mention of Odroid XU or XU4, unless I run your workaround.

aaronouthier

So I did some more digging. Realized that /usr/bin/emulationstation is a shell script. Found the actual executable, and ran it:

Feb 26 19:43:02 lvl2:   EmulationStation - v2.11.2rp, built Feb 20 2025 - 11:49:42
Feb 26 19:43:02 lvl2:   Parsing XML file "/opt/retropie/supplementary/emulationstation/resources/mamenames.xml"...
Feb 26 19:43:03 lvl2:   Parsing XML file "/opt/retropie/supplementary/emulationstation/resources/mamebioses.xml"...
Feb 26 19:43:03 lvl2:   Parsing XML file "/opt/retropie/supplementary/emulationstation/resources/mamedevices.xml"...
Feb 26 19:43:03 lvl2:   Creating window...
Segmentation fault (core dumped)

Where are core dumps stored? I'm willing to send it to you for investigation.

mitu

@aaronouthier said in Odroid XU4 SDL Problem Emulationstation:

Where are core dumps stored? I'm willing to send it to you for investigation.

Check the folder where you ran the command ($HOME ?) for a core file or verify in cat /proc/sys/kernel/core_pattern what should be the core file name.

aaronouthier

@mitu
It's amazing what a Google search can turn up, isn't it?
It was in /var/crashes/

I have removed dep and core packages, and am compiling again.. Strangely, that was finding some missing packages while building deps.

I am just noticing the following when compiling emulationstation:

= = = = = = = = = = = = = = = = = = = = =
Building 'emulationstation' : EmulationStation - Frontend used by RetroPie for launching emulators
= = = = = = = = = = = = = = = = = = = = =

Removing additional swap
CMake Deprecation Warning at CMakeLists.txt:1 (cmake_minimum_required):
  Compatibility with CMake < 2.8.12 will be removed from a future version of
  CMake.

  Update the VERSION argument <min> value or use a ...<max> suffix to tell
  CMake that the project does not need compatibility with older versions.


CMake Deprecation Warning at CMakeLists.txt:20 (cmake_policy):
  The OLD behavior for policy CMP0072 will be removed from a future version
  of CMake.

  The cmake-policies(7) manual explains that the OLD behaviors of all
  policies are deprecated and that a policy should be set to OLD only under
  specific short-term circumstances.  Projects should be ported to the NEW
  behavior and not rely on setting a policy to OLD.

Probably of no consequence, but thought I'd mention it. Compilation is proceeding despite this.

I'll be back shortly.

aaronouthier

Latest crash dump

https://drive.google.com/file/d/1N7r_waoGE_rRo_SlUUsNLIrNlBNF7kDs/view?usp=sharing

mitu

@aaronouthier That's more than a crash dump - can you upload just the core file ? You can also analyze it by loading it into gdb and running backtrace to see where it crashes.

aaronouthier

@mitu That was the only file I found. There are no files in current directory, only 2 folders. EmulationStation was run from the folder in question (my home folder).

The full path and filename was originally:
/var/crash/_opt_retropie_supplementary_emulationstation_emulationstation.1000.crash

I shortened it.

I believe you are looking for the base64 encoded part at the bottom of the file? I will see what I can find out.

aaronouthier

@mitu said in Odroid XU4 SDL Problem Emulationstation:

verify in cat /proc/sys/kernel/core_pattern what should be the core file name

That command outputs

|/usr/share/apport/apport -p%p -s%s -c%c -d%d -P%P -u%u -g%g -- %E

gdb run with --core arg pointing to correct core dump file:

Reading symbols from /opt/retropie/supplementary/emulationstation/emulationstation...
(No debugging symbols found in /opt/retropie/supplementary/emulationstation/emulationstation)

warning: core file may not match specified executable file.
[New LWP 1114]
[New LWP 1115]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/arm-linux-gnueabihf/libthread_db.so.1".
Core was generated by `/opt/retropie/supplementary/emulationstation/emulationstation'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  __libc_do_syscall () at ../sysdeps/unix/sysv/linux/arm/libc-do-syscall.S:47
47      ../sysdeps/unix/sysv/linux/arm/libc-do-syscall.S: No such file or directory.
[Current thread is 1 (Thread 0xb6cea020 (LWP 1114))]
(gdb) backtrace
#0  __libc_do_syscall () at ../sysdeps/unix/sysv/linux/arm/libc-do-syscall.S:47
#1  0xb544d292 in __pthread_kill_implementation (threadid=3066994720, signo=11, no_tid=<optimized out>) at pthread_kill.c:43
#2  0xb541c840 in __GI_raise (sig=11) at ../sysdeps/posix/raise.c:26
#3  0xb6e21be4 in ?? () from /lib/arm-linux-gnueabihf/libSDL2-2.0.so.0
Backtrace stopped: previous frame identical to this frame (corrupt stack?)