Hi,
we're using mainline 4.12 kernel (bit too old I know) on am335x board.
We have case where we receive files over g_serial using zmodem
protocol. We send small files every second and after ~20 minutes or so
we hit this issue:
Workqueue: events_unbound flush_to_ldisc
task: ee32d100 task.stack: ef132000
PC is at n_tty_receive_buf_common+0x50/0x8d4
LR is at n_tty_receive_buf_common+0x40/0x8d4
pc : [<c0551e54>] lr : [<c0551e44>] psr: 200e0013
sp : ef133e38 ip : ef133e38 fp : ef133e94
r10: c0fc4378 r9 : 00000000 r8 : c05553fc
r7 : 00000014 r6 : 00000000 r5 : c05526d8 r4 : ee066200
r3 : 00002000 r2 : 00000001 r1 : 00000000 r0 : ee066258
Flags: nzCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none
Control: 10c53c7d Table: ae728059 DAC: 00000051
Process kworker/u2:0 (pid: 7235, stack limit = 0xef132210)
Stack: (0xef133e38 to 0xef134000)
3e20: ef133e6c ef133e48
3e40: 00002cc3 ee066258 ee32d130 00000000 ee32d130 00000014 00000000 ee08d4a8
3e60: 00000000 ef133e70 c01519a8 00000014 c05526d8 00000000 00000014 c05553fc
3e80: 00000000 c0fc4378 ef133eac ef133e98 c05526fc c0551e10 00000001 ef133ea8
3ea0: ef133ec4 ef133eb0 c0554a1c c05526e4 ee08d4a8 ee0da040 ef133ee4 ef133ec8
3ec0: c0555450 c05549f8 ee08d400 ee6cfc04 ee6cfc14 ee6cfc00 ef133f0c ef133ee8
3ee0: c0554c2c c0555408 ee540980 ee6cfc04 ef002400 00000000 ef02d100 00000000
3f00: ef133f44 ef133f10 c014a1a8 c0554ba4 c0f28d80 ef002414 ee540998 ee540980
3f20: ef002400 ef002400 c0f28d80 ef002414 ee540998 00000088 ef133f74 ef133f48
3f40: c014b000 c0149f44 00000000 ee302340 ee302ac0 ef132000 00000000 ee540980
3f60: c014acf4 ee4d7eac ef133fac ef133f78 c014ff90 c014ad00 ee302358 ee302358
3f80: ef133fac ee302ac0 c014fe44 00000000 00000000 00000000 00000000 00000000
3fa0: 00000000 ef133fb0 c0107b08 c014fe50 00000000 00000000 00000000 00000000
3fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
3fe0: 00000000 00000000 00000000 00000000 00000013 00000000 2ecc2ecc 2ecc2ecc
[<c0551e54>] (n_tty_receive_buf_common) from [<c05526fc>]
(n_tty_receive_buf2+0x24/0x2c)
[<c05526fc>] (n_tty_receive_buf2) from [<c0554a1c>]
(tty_ldisc_receive_buf+0x30/0x68)
[<c0554a1c>] (tty_ldisc_receive_buf) from [<c0555450>]
(tty_port_default_receive_buf+0x54/0x64)
[<c0555450>] (tty_port_default_receive_buf) from [<c0554c2c>]
(flush_to_ldisc+0x94/0xb4)
[<c0554c2c>] (flush_to_ldisc) from [<c014a1a8>] (process_one_work+0x270/0x440)
[<c014a1a8>] (process_one_work) from [<c014b000>] (worker_thread+0x30c/0x498)
[<c014b000>] (worker_thread) from [<c014ff90>] (kthread+0x14c/0x164)
[<c014ff90>] (kthread) from [<c0107b08>] (ret_from_fork+0x14/0x2c)
Code: e3a03000 e50b3048 e51b303c e2833a02 (e5930244)
I disassemble function call and area is:
(gdb) l * n_tty_receive_buf_common+0x50
0xc054d064 is in n_tty_receive_buf_common (drivers/tty/n_tty.c:1692).
1687 * the consumer has loaded the data in read_buf up to the new
1688 * read_tail (so this producer will not overwrite unread data)
1689 */
1690 size_t tail = smp_load_acquire(&ldata->read_tail);
1691
1692 room = N_TTY_BUF_SIZE - (ldata->read_head - tail);
1693 if (I_PARMRK(tty))
1694 room = (room + 2) / 3;
1695 room--;
1696 if (room <= 0) {
so maybe ldata is not initialized and thus we got crash? Any ideas?
Thanks a lot for any pointers.
BR,
marek
--
as simple and primitive as possible
-------------------------------------------------
Marek Belisko - OPEN-NANDRA
Freelance Developer
Ruska Nova Ves 219 | Presov, 08005 Slovak Republic
Tel: +421 915 052 184
skype: marekwhite
twitter: #opennandra
web: http://open-nandra.com
