Segfault generating red screen

[badosu]

Recently I've been having some issues with my amd gpu driver, I'm not sure if it's related, I found this in my journal:

set 03 15:14:33 badosu-pangolin systemd-coredump[18085]: [🡕] Process 16572 (swaylock) of user 1000 dumped core.
                                                         
                                                         Stack trace of thread 16572:
                                                         #0  0x0000559b50b91551 n/a (/usr/bin/swaylock + 0x8551)
                                                         #1  0x00007ffa6be1bac6 n/a (libffi.so.8 + 0x7ac6)
                                                         #2  0x00007ffa6be1876b n/a (libffi.so.8 + 0x476b)
                                                         #3  0x00007ffa6be1b06e ffi_call (libffi.so.8 + 0x706e)
                                                         #4  0x00007ffa6c86948d n/a (libwayland-client.so.0 + 0x448d)
                                                         #5  0x00007ffa6c86a2e9 n/a (libwayland-client.so.0 + 0x52e9)
                                                         #6  0x00007ffa6c86a6f3 wl_display_dispatch_queue_pending (libwayland-client.so.0 + 0x56f3)
                                                         #7  0x00007ffa6c86df34 wl_display_dispatch_queue_timeout (libwayland-client.so.0 + 0x8f34)
                                                         #8  0x00007ffa6c86e010 wl_display_dispatch_queue (libwayland-client.so.0 + 0x9010)
                                                         #9  0x0000559b50b90b95 n/a (/usr/bin/swaylock + 0x7b95)
                                                         #10 0x0000559b50b8d991 n/a (/usr/bin/swaylock + 0x4991)
                                                         #11 0x00007ffa6c627675 n/a (libc.so.6 + 0x27675)
                                                         #12 0x00007ffa6c627729 __libc_start_main (libc.so.6 + 0x27729)
                                                         #13 0x0000559b50b8e6b5 n/a (/usr/bin/swaylock + 0x56b5)
                                                         ELF object binary architecture: AMD x86-64

Maybe this issue on poweralertd could also help?

Let me know how I can help diagnose further.

[kennylevinsen]

Is it reproducible? The stack trace here isn't useful on its own as it lacks all debug info, and just says "swaylock crashed while handling some Wayland message".

It would take the full back trace from a core dump with debug symbols (not "n/a") and description of the crash situation/reproduction to debug.

You can try and see with coredumpctl debug swaylock if gdb can download the symbols. If so, bt full should give something better than n/a.

[badosu]

I'll try and get a backtrace this weekend

[badosu]

I've since updated my system and have not experienced this issue again, since I can't reproduce and provide better data I'll close it. Thank you.

[badosu]

So, I had this issue again on 1.8.3 so I put some more effort in debugging it.

Since arch does not have a dbgsym package for swaylock, I downloaded the one from ubuntu for 1.8.2 and used it against my old coredump for 1.8.2. I'm not sure I ran all the dbg commands correctly, I hope this is useful:

bt full for 1.8.2

(gdb) bt full
#0  0x0000559b50b91551 in parse_options (argc=32762, argv=0x7ffe1cc09d90, state=0xe1cd0000633b, line_mode=0xe3db00007370, config_path=0xe3db00007e8d) at ../main.c:820
        opt_idx = 56196
        usage = "M4\000\000V\313\000\000\\\247\300\034\376\177\000\000p\250\222\201\233U\000\000@\354\270l\372\177\000\000\300\235\300\034\376\177\000\000\306\362\251l\372\177\000\000\000\t\000\000\000\b\000\000\333\353\251l\372\177\000\000\b\032\000\000y\253\000\000\\\247\300\034\376\177\000\0000\236\300\034\376\177\000\000j\025\252l\372\177\000\000\332\021\000\000\340\000\000\000\362\000\000\000\340\000\000\000\t\000\000\000\b\000\000\000\340\000\000\000\340\000\000\000\216\r\000\000\n\203\000\000\000ή\216\233a\344s\217\002\000\000\000\000\000\000\240̻l\372\177\000\000\000\000\000\000\000\000\000\000\300\242\300\034\376\177\000\000Ģ\300\034\376\177\000\000 \247\300\034\376\177\000\000`\242\300\034\376\177\000\000㌮l\372"...
        c = 
        long_options = Python Exception : Cannot access memory at address 0x110e0
#1  0x00007ffa6be1bac6 in ??? () at /usr/lib/libffi.so.8
#2  0x00007ffa6be1876b in ??? () at /usr/lib/libffi.so.8
#3  0x00007ffa6be1b06e in ffi_call () at /usr/lib/libffi.so.8
#4  0x00007ffa6c86948d in ??? () at /usr/lib/libwayland-client.so.0
#5  0x00007ffa6c86a2e9 in ??? () at /usr/lib/libwayland-client.so.0
#6  0x00007ffa6c86a6f3 in wl_display_dispatch_queue_pending () at /usr/lib/libwayland-client.so.0
#7  0x00007ffa6c86df34 in wl_display_dispatch_queue_timeout () at /usr/lib/libwayland-client.so.0
#8  0x00007ffa6c86e010 in wl_display_dispatch_queue () at /usr/lib/libwayland-client.so.0
#9  0x0000559b50b90b95 in handle_global (version=, data=0x559b81937f60, registry=0x431bde82d7b634db, name=2173906880, interface=0x559b81932fa0 "") at ../main.c:270
        state = 
        state = 
        seat = 
        swaylock_seat = 
        surface = 
#10 handle_global (data=0x559b81937f60, registry=0x431bde82d7b634db, name=2173906880, interface=0x559b81932fa0 "", version=) at ../main.c:251
        state = 0x559b81937f60
        seat = 
        swaylock_seat = 
        surface = 
#11 0x0000559b50b8d991 in ??? ()
#12 0x00007ffa6c627675 in ??? () at /usr/lib/libc.so.6
#13 0x00007ffa6c627729 in __libc_start_main () at /usr/lib/libc.so.6
#14 0x0000559b50b8e6b5 in ??? ()

Let me know if there's anything else I can help with. If x86_64 debug symbols are available for swaylock 1.8.3 I can run it again on them.

[francma]

Same issue for me, it happens when I put laptop to sleep (without any external output connected) and then wake-up with external output connected.

Screen is fully red, doesn't respond to any input. Putting the laptop to sleep again (via power button) and then waking it up fixes the issue.

#0  render_frame (surface=0x5616de3c2310) at ../swaylock/render.c:132
        state = 0xab192a935174a34d
        attempts = "\375\177\000"
        text = 0x0
        layout_text = 0x0
        draw_indicator = <optimized out>
        arc_radius = <optimized out>
        arc_thickness = <optimized out>
        buffer_diameter = <optimized out>
        buffer_width = <optimized out>
        buffer_height = <optimized out>
        subsurf_xpos = <optimized out>
        subsurf_ypos = <optimized out>
        buffer = <optimized out>
        cairo = <optimized out>
#1  render (surface=0x5616de3c2310) at ../swaylock/render.c:103
        state = 0xab192a935174a34d
        buffer_width = 1920
        buffer_height = 1080
        need_destroy = false
        buffer = {buffer = 0x5616de4281a0, surface = 0x5616de3c2388, cairo = 0x7ffdfc9ba2a0, width = 3728461256, height = 22038, data = 0x5616de4281a0, size = 140728841512248, busy = 80}
#2  0x00007f2553c7aac6 in ffi_call_unix64 () at ../src/x86/unix64.S:104
#3  0x00007f2553c7776b in ffi_call_int (cif=cif@entry=0x7ffdfc9ba450, fn=fn@entry=0x5616acadb8e0 <surface_frame_handle_done>, rvalue=<optimized out>, rvalue@entry=0x0, avalue=avalue@entry=0x7ffdfc9ba520, closure=closure@entry=0x0)
    at ../src/x86/ffi64.c:676
        classes = {X86_64_INTEGERSI_CLASS, X86_64_NO_CLASS, 1405600320, 32549}
        stack = <optimized out>
        argp = 0x7ffdfc9ba2a0 ""
        arg_types = <optimized out>
        gprcount = 3
        ssecount = <optimized out>
        ngpr = 1
        nsse = 0
        i = <optimized out>
        avn = <optimized out>
        flags = <optimized out>
        reg_args = <optimized out>
#4  0x00007f2553c7a06e in ffi_call (cif=cif@entry=0x7ffdfc9ba450, fn=0x5616acadb8e0 <surface_frame_handle_done>, rvalue=rvalue@entry=0x0, avalue=avalue@entry=0x7ffdfc9ba520) at ../src/x86/ffi64.c:713
        arg_types = <optimized out>
        i = <optimized out>
        nargs = <optimized out>
        max_reg_struct_size = <optimized out>
#5  0x00007f25549d348d in wl_closure_invoke (closure=closure@entry=0x5616de3c2140, target=<optimized out>, target@entry=0x5616de4281a0, opcode=opcode@entry=0, data=<optimized out>, flags=1) at ../wayland-1.24.0/src/connection.c:1241
        count = 1
        cif = {abi = FFI_UNIX64, nargs = 3, arg_types = 0x7ffdfc9ba470, rtype = 0x7f2553c7c1a0 <ffi_type_void>, bytes = 0, flags = 0}
        ffi_types = {0x7f2553c7c2a0 <ffi_type_pointer>, 0x7f2553c7c2a0 <ffi_type_pointer>, 0x7f2553c7c240 <ffi_type_uint32>, 0x5616de3bcca0, 0x7f25549dda98 <wl_display_events+24>, 0x0, 0x7ffdfc9ba500, 0x7f25549d26e8 <wl_closure_init+200>, 0xdc53c7c2a0, 0xdc, 0x75, 0xff64faca11041100, 0x2f, 0x1, 0x5616de3bcca0, 0x5616de3bcdc8, 0x7ffdfc9ba5a0, 0x5616de3bcca0, 0x7ffdfc9ba6c0, 0x7f25549d67eb <wl_display_read_events+155>, 0x10000006f, 0xff64faca11041100}
        ffi_args = {0x7ffdfc9ba430, 0x7ffdfc9ba438, 0x5616de3c2158, 0x5616de3bcca0, 0xc, 0x5616de3bcca0, 0x5616de3bce30, 0x5616de3da6c8, 0x7ffdfc9ba600, 0x56160000000c, 0x7f25549d9bb5, 0x7f250000000c, 0x0, 0x5616de3bcdc8, 0x0, 0x0, 0x75, 0xff64faca00000000, 0x7ffdfc9ba5e0, 0x7f255484cb38 <_int_free_create_chunk+136>, 0x90, 0xf0}
        implementation = <optimized out>
#6  0x00007f25549d42e9 in dispatch_event (display=display@entry=0x5616de3bcca0, queue=queue@entry=0x5616de3bcd98) at ../wayland-1.24.0/src/wayland-client.c:1707
        closure = 0x5616de3c2140
        proxy = 0x5616de4281a0
        opcode = 0
        proxy_destroyed = <optimized out>
#7  0x00007f25549d46f3 in dispatch_queue (display=0x5616de3bcca0, queue=0x5616de3bcd98) at ../wayland-1.24.0/src/wayland-client.c:1853
        count = 8
#8  wl_display_dispatch_queue_pending (display=display@entry=0x5616de3bcca0, queue=queue@entry=0x5616de3bcd98) at ../wayland-1.24.0/src/wayland-client.c:2190
        ret = <optimized out>
#9  0x00007f25549d7f34 in wl_display_dispatch_queue_timeout (display=0x5616de3bcca0, queue=0x5616de3bcd98, timeout=timeout@entry=0x0) at ../wayland-1.24.0/src/wayland-client.c:2096
        ret = <optimized out>
        now = {tv_sec = 140728841512752, tv_nsec = -43634350591045376}
        deadline = {tv_sec = 0, tv_nsec = <optimized out>}
--Type <RET> for more, q to quit, c to continue without paging--
        result = {tv_sec = 46032, tv_nsec = 94655386354264}
        remaining_timeout = 0x0
#10 0x00007f25549d8010 in wl_display_dispatch_queue (display=<optimized out>, queue=<optimized out>) at ../wayland-1.24.0/src/wayland-client.c:2163
        ret = <optimized out>
#11 0x00007f25549d8171 in wl_display_dispatch (display=<optimized out>) at ../wayland-1.24.0/src/wayland-client.c:2233
#12 0x00005616acad8b95 in display_in (fd=<optimized out>, mask=<optimized out>, data=<optimized out>) at ../swaylock/main.c:1031
#13 0x00005616acad5991 in loop_poll (loop=0x5616de3bbfa0) at ../swaylock/loop.c:99
        pfd = {fd = <optimized out>, events = <optimized out>, revents = <optimized out>}
        events = <optimized out>
        ms = <optimized out>
        ret = <optimized out>
        fd_index = <optimized out>
        event = 0x5616de3c0f60
#14 main (argc=<optimized out>, argv=<optimized out>) at ../swaylock/main.c:1263
        line_mode = LM_LINE
        config_path = 0x0
        result = <optimized out>
        registry = <optimized out>
        surface = <optimized out>
        sa = {__sigaction_handler = {sa_handler = 0x5616acad6a50 <do_sigusr>, sa_sigaction = 0x5616acad6a50 <do_sigusr>}, sa_mask = {__val = {0, 140728841513488, 139798309972672, 2064, 94656217713840, 95056, 139798308517996, 140728841513104, 139798308517996, 8, 0, 0, 0, 2064, 139798309972672, 2064}}, sa_flags = 268435456, sa_restorer = 0x7ffdfc9ba8e0}

Looking at the stack trace, one thing that comes as unusual is that the output is reported as 1920x1080. Which is true for internal output, but the external one is 2560x1440.

[kennylevinsen]

Same issue for me

Red screen is what sway shows any time a lock screen terminates without unlocking the session. The issue is swaylock crashing, and your crash does not appear related. Please open a separate issue.

[kennylevinsen]

So, I had this issue again on 1.8.3 so I put some more effort in debugging it.

Hmm, that crash is in parse_options, while swaylock is reading command-line options and configuration files, before it locks anything - so this if hit would just mean that you screen wouldn't lock in the first place.

The line of the crash is when parsing a color, but I don't see any immediate issue there. What config file/command-line arguments are you using when the crash occurs?

[badosu]

The line of the crash is when parsing a color, but I don't see any immediate issue there. What config file/command-line arguments are you using when the crash occurs?

Just swaylock --daemonize --show-failed-attempts without any swaylock config file in ~/.swaylock, ~/.config/swaylock or /etc/swaylock. Let me know if I can find any other places that could be sending problematic options.

[ebenali]

this happens when the program exits early and doesn't allow for the chance of this line to get called at the end of main.c:

ext_session_lock_v1_unlock_and_destroy(state.ext_session_lock_v1);