presence module with in-memory presentity records

[miconda]

Track the issues discovered by different developers while implementing in-memory only presentity records for presence module.

[miconda]

A report and discussions were done at:

• f917318#commitcomment-38645006

[eschmidbauer]

segfault when modparam("presence", "subs_db_mode", 0)

(gdb)
(gdb) bt full
#0  0x00007f12607da547 in atomic_cmpxchg_int (var=0x8, old=0, new_v=1) at ../../core/mem/../atomic/atomic_x86.h:224
        ret = 32530
#1  0x00007f12607da57b in futex_get (lock=0x8) at ../../core/mem/../futexlock.h:99
        v = 21864
        i = 1024
#2  0x00007f12607fe0d0 in update_db_subs_timer_dbnone (no_lock=0) at subscribe.c:2079
        i = 0
        now = 1587487071
        s = 0x0
        prev_s = 0x0
        del_s = 0x5568ae3160b0
        __func__ = "update_db_subs_timer_dbnone"
#3  0x00007f1260802121 in timer_db_update (ticks=61825161, param=0x0) at subscribe.c:2441
        no_lock = 0
        __func__ = "timer_db_update"
#4  0x00005568ade05b2e in sr_wtimer_exec (ticks=61825161, param=0x0) at core/timer_proc.c:433
        wt = 0x7f126821dae8
        wn = 0x0
        wp = 0x0
        cs = 4
        __func__ = "sr_wtimer_exec"
#5  0x00005568ade04b58 in fork_sync_timer (child_id=-1, desc=0x5568ae01a3a1 "secondary timer", make_sock=1, f=0x5568ade057a8 <sr_wtimer_exec>, param=0x0, interval=1000) at core/timer_proc.c:267
        pid = 0
        ts1 = 989202582
        ts2 = 1000
#6  0x00005568ade05ece in sr_wtimer_start () at core/timer_proc.c:459
        __func__ = "sr_wtimer_start"
#7  0x00005568adc05e24 in main_loop () at main.c:1452
        i = 21864
        pid = 6405
        si = 0x0
        si_desc = "H4\037h\022\177\000\000\b\201\026h\022\177\000\000\000\000\200\000\003\000\000\000`.\022h\022\177", '\000' <repeats 13 times>, "\004\000\000\000\000H9\037h\022\177\000\000\323u\000\256hU\000\000\200\243\001\225\375\177\000\000\207ыh\022\177\000\000`\243\001\225\375\177\000\000 \000\000\000\060\000\000\000p\243\001\225\375\177\000\000\260\242\001\225\375\177\000\000\323u\000\256hU\000\000Pw\037h\022\177\000"
        nrprocs = 0
        woneinit = 21864
        __func__ = "main_loop"
#8  0x00005568adc123e7 in main (argc=9, argv=0x7ffd9501a8a8) at main.c:2845
        cfg_stream = 0x5568b01ad010
        c = -1
        r = 0
---Type <return> to continue, or q <return> to quit---
        tmp = 0x7f126976e708 ""
        tmp_len = 0
        port = 0
        proto = 32765
        ahost = 0x0
        aport = 0
        options = 0x5568adfd3918 ":f:cm:M:dVIhEeb:l:L:n:vKrRDTN:W:w:t:u:g:P:G:SQ:O:a:A:x:X:Y:"
        ret = -1
        seed = 4119354055
        rfd = 4
        debug_save = 4
        debug_flag = 1
        dont_fork_cnt = 1
        n_lst = 0x7f126976d948 <_rtld_global+2312>
        p = 0x7f1268907010 <handle_intel+144> "H\205\300uTD9t$\fA\215V\001v0A\211ָ\002"
        st = {st_dev = 18, st_ino = 63494, st_nlink = 2, st_mode = 16832, st_uid = 0, st_gid = 0, __pad0 = 0, st_rdev = 0, st_size = 60, st_blksize = 4096, st_blocks = 0, st_atim = {tv_sec = 1587484687, tv_nsec = 26720353}, st_mtim = {
            tv_sec = 1587486933, tv_nsec = 570847206}, st_ctim = {tv_sec = 1587486933, tv_nsec = 570847206}, __glibc_reserved = {0, 0, 0}}
        tbuf = "\330\331vi\022\177\000\000\220Q\342h\022\177\000\000H\331vi\022\177\000\000\370\250\001\225\375\177\000\000p\341vi\022\177\000\000\000\000\000\000\000\000\000\000@\320vi\022\177\000\000\224{Ui\022\177\000\000\001", '\000' <repeats 15 times>, "\350\244\001\225\375\177\000\000\360,\207h\022\177\000\000 \247\001\225\375\177\000\000\332\361Ui\022\177\000\000\300@\301h\022\177\000\000\327\352\224h\022\177\000\000\240K\342h\022\177\000\000\200\071\301h\022\177\000\000(\224\342h\022\177\000\000\000\020", '\000' <repeats 14 times>, "\310\344vi\022\177\000\000\000\000\000\000\000\000\000\000ؗTi\022\177\000\000\220\223Ti\022\177\000\000\200"...
        option_index = 0
        long_options = {{name = 0x5568adfd5916 "help", has_arg = 0, flag = 0x0, val = 104}, {name = 0x5568adfd10e9 "version", has_arg = 0, flag = 0x0, val = 118}, {name = 0x5568adfd591b "alias", has_arg = 1, flag = 0x0, val = 1024}, {
            name = 0x5568adfd5921 "subst", has_arg = 1, flag = 0x0, val = 1025}, {name = 0x5568adfd5927 "substdef", has_arg = 1, flag = 0x0, val = 1026}, {name = 0x5568adfd5930 "substdefs", has_arg = 1, flag = 0x0, val = 1027}, {
            name = 0x5568adfd593a "server-id", has_arg = 1, flag = 0x0, val = 1028}, {name = 0x5568adfd5944 "loadmodule", has_arg = 1, flag = 0x0, val = 1029}, {name = 0x5568adfd594f "modparam", has_arg = 1, flag = 0x0, val = 1030}, {
            name = 0x5568adfd5958 "log-engine", has_arg = 1, flag = 0x0, val = 1031}, {name = 0x0, has_arg = 0, flag = 0x0, val = 0}}
        __func__ = "main"
(gdb)

[miconda]

Can you get the output for following commands:

frame 2
p subs_htable
p subs_htable[0]
p shtable_size

[eschmidbauer]

Program terminated with signal SIGSEGV, Segmentation fault.
#0  0x00007f12607da547 in atomic_cmpxchg_int (var=0x8, old=0, new_v=1) at ../../core/mem/../atomic/atomic_x86.h:224
224	ATOMIC_FUNC_CMPXCHG(cmpxchg, "cmpxchgl %2, %1", int , int)
(gdb) frame 2
#2  0x00007f12607fe0d0 in update_db_subs_timer_dbnone (no_lock=0) at subscribe.c:2079
2079				lock_get(&subs_htable[i].lock);
(gdb) p subs_htable
$1 = (shtable_t) 0x0
(gdb) p subs_htable[0]
Cannot access memory at address 0x0
(gdb) p shtable_size
$2 = 9
(gdb)

[miconda]

Apparently I broke initialization of some structures when I added conditions on db connectivity. I just pushed another set of commits in master. Can you try again?

[eschmidbauer]

trying again on latest master branch. i'll post results.

[eschmidbauer]

more coredumps. is it okay to email the bt to you as they may contain sensitive data at this point?

[miconda]

Yes, send them to me.

I am at the stage of testing with default config file that still uses database for subscriptions and watchers, but with presentity records kept in memory. Tests are ok so far with two devices watching each other the presence status.

[miconda]

Closing based on a few days running without issues. From now on newly discovered bugs can get a dedicated issue on this tracker.