Linux/AM3358: Kernel crash using GPIO0_10

Ben Taylor

Part Number: AM3358

Tool/software: Linux

I am currently trying to use pin gpio0_10 as a standard GPIO on the AM3358 processor and am experiencing issues. When I attempt to use the pin as an input it appears to cause a crash in the cpufreq_dt driver when it is initially loaded. I've seen a few different stack traces but I've attached an example. It also appears that something is sampling this pin during the boot, as once the system has booted the modprobe of cpufreq_dt will fail irrelevant of the state of the pin. Has anyone had any similar issues or are there any known issues with this pin?

Fullscreen stack_trace.txt Download

SD:~# modprobe cpufreq_dt
[  170.753986] Unable to handle kernel NULL pointer dereference at virtual address 000000d4
[  170.764300] pgd = c0004000
[  170.767696] [000000d4] *pgd=00000000
[  170.772245] Internal error: Oops: 805 [#1] ARM
[  170.777824] Modules linked in: cpufreq_dt bnep hci_uart bluetooth usb_serial_simple usbserial ipv6 tun ppp_async ppp_generic slhco
[  170.807489] CPU: 0 PID: 1951 Comm: cat Not tainted linux-4.9.57#2
[  170.815896] Hardware name: Generic AM33XX (Flattened Device Tree)
[  170.823541] task: ec8af080 task.stack: ec900000
[  170.829246] PC is at setup_arg_pages+0x18c/0x3dc
[  170.835045] LR is at lock_release+0x6c/0x4b8
[  170.840403] pc : [<c02280b8>]    lr : [<c01664f4>]    psr: 200a0013
[  170.840403] sp : ec901e28  ip : beb11000  fp : ec901e68
[  170.854807] r10: bf000000  r9 : beb10000  r8 : befff000
[  170.861361] r7 : edf0ddd4  r6 : 00000000  r5 : eeaa7d00  r4 : edf0cd68
[  170.869550] r3 : 00000000  r2 : beb10eb9  r1 : beaf0000  r0 : edf0cd68
[  170.877741] Flags: nzCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment none
[  170.886694] Control: 10c5387d  Table: ac918019  DAC: 00000051
[  170.893902] Process cat (pid: 1951, stack limit = 0xec900210)
[  170.901111] Stack: (0xec901e28 to 0xec902000)
[  170.906581] 1e20:                   00000000 00000000 edf0dd80 beb11000 eeaa7d00 edf0cd68
[  170.916845] 1e40: edf0dd80 00000000 00000000 befff000 bf000000 beeff000 ec8af080 00000000
[  170.927111] 1e60: 00000400 edf09000 00000001 eeafde84 ec901e90 c131e9c4 600a0013 c027ae3c
[  170.937375] 1e80: eeafde84 ec901e90 c131e9c4 ede9fe00 eeaa7d00 bf000000 00000001 eef2ea00
[  170.947639] 1ea0: ede9fe34 edfda800 ec901eb0 c027ae94 ede9fe34 00000034 00000002 edfed400
[  170.957903] 1ec0: edad90c0 c0166374 00000001 00000000 00000000 c02285d4 00000000 00000000
[  170.968168] 1ee0: 00000000 c0146ff0 00000000 c0b28270 600a0013 00000001 fffffff8 c0b28270
[  170.978432] 1f00: eeaa7d00 c0b2b80c eeaa7d00 c0b2828c fffffff8 c0b28270 c0b28270 00000001
[  170.988697] 1f20: edf0dd80 c02285c8 c0b7e19a edf12000 eeaa7d00 00000000 0000079f 00107370
[  170.998962] 1f40: 0000079f c0229ca0 00000000 00000000 c0229bc8 001072e0 0000079f c0b82148
[  171.009226] 1f60: 00000000 00000000 00000001 00107370 001021f8 000e761c 0000000b c0107584
[  171.019491] 1f80: ec900000 00000000 00000001 c022a1cc 00000000 c01074bc 001072e0 00107308
[  171.029756] 1fa0: 001072e0 c01073e0 00107308 001072e0 001072e0 00107370 001021f8 000e8a98
[  171.040020] 1fc0: 00107308 001072e0 000e761c 0000000b 001021f8 00000000 00000000 00000001
[  171.050284] 1fe0: b6e44d80 bedf42f4 0003b8f4 b6e44d8c 600a0010 001072e0 0a303831 303d3237
[  171.060571] [<c02280b8>] (setup_arg_pages) from [<c027ae94>] (load_elf_binary+0x404/0x11fc)
[  171.071061] [<c027ae94>] (load_elf_binary) from [<c02285c8>] (search_binary_handler+0x84/0x1e0)
[  171.081983] [<c02285c8>] (search_binary_handler) from [<c0229ca0>] (do_execveat_common+0x668/0x8f0)
[  171.093339] [<c0229ca0>] (do_execveat_common) from [<c022a1cc>] (SyS_execve+0x30/0x38)
[  171.103288] [<c022a1cc>] (SyS_execve) from [<c01073e0>] (ret_fast_syscall+0x0/0x1c)
[  171.112903] Code: 304c1003 e5903240 22411802 e1a00004 (e58320d4)
[  171.120709] ---[ end trace 8bdddcdab7343458 ]---
[  171.148256] Unable to handle kernel NULL pointer dereference at virtual address 0000003c
[  171.158448] pgd = ed920000
[  171.161840] [0000003c] *pgd=ac85a831, *pte=00000000, *ppte=00000000
[  171.169738] Internal error: Oops: 17 [#2] ARM
[  171.175205] Modules linked in: cpufreq_dt bnep hci_uart bluetooth usb_serial_simple usbserial ipv6 tun ppp_async ppp_generic slhco
[  171.204867] CPU: 0 PID: 1811 Comm: syslogd Tainted: G      D         4.9.57-eurotech-ti #2
[  171.215235] Hardware name: Generic AM33XX (Flattened Device Tree)
[  171.222879] task: edfdcd00 task.stack: ec80c000
[  171.228583] PC is at kmem_cache_alloc+0x44/0x1d4
[  171.234375] LR is at kmem_cache_alloc+0x40/0x1d4
[  171.240167] pc : [<c021a5c0>]    lr : [<c021a5bc>]    psr: 600f0093
[  171.240167] sp : ec80de70  ip : 00000000  fp : 00000000
[  171.254571] r10: ec80c000  r9 : c0257c6c  r8 : 024000c0
[  171.261125] r7 : a00f0013  r6 : ec888c80  r5 : 024000c0  r4 : ee800004
[  171.269314] r3 : ffffe000  r2 : 0000003c  r1 : 00000000  r0 : c021a5bc
[  171.277506] Flags: nZCv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment none
[  171.286567] Control: 10c5387d  Table: ad920019  DAC: 00000051
[  171.293774] Process syslogd (pid: 1811, stack limit = 0xec80c210)
[  171.301418] Stack: (0xec80de70 to 0xec80e000)
[  171.306887] de60:                                     eeaa7850 c06ecdc4 00000020 ec9711c0
[  171.317152] de80: ec888c80 ec888c80 c0b820c0 b6f81d88 00000000 ec80c000 00000000 c0257c6c
[  171.327417] dea0: ec9711c0 01200011 ec888c80 c0b820c0 b6f81d88 c0126010 00000000 00000000
[  171.337682] dec0: 00000000 ffffe000 ec971480 c0b7edb4 600f0013 c0166374 00000000 00000000
[  171.347946] dee0: 00000000 c0242c28 00000000 00000000 600f0013 00000101 d9446ac7 00000027
[  171.358211] df00: c0b7e1bd fffffff4 c0879d6c 01200011 00000000 b6f821e0 00000078 00000000
[  171.368475] df20: ec80c000 00000000 beeb25fc c01270b0 00000000 00000000 ffffffff edec78c8
[  171.378740] df40: fffff000 00000fff 00000900 00000000 00000011 edec78c0 edec78c0 00000051
[  171.389004] df60: 00000000 beeb25c0 00000008 00000000 beeb25c0 b6f81d88 00000000 b6f821e0
[  171.399270] df80: 00000078 c0107584 ec80c000 00000000 beeb25fc c01277cc b6f81d88 00000000
[  171.409533] dfa0: 00000000 c01073e0 b6f81d88 00000000 01200011 00000000 00000000 00000000
[  171.419798] dfc0: b6f81d88 00000000 b6f821e0 00000078 00000000 b6f50000 00000000 beeb25fc
[  171.430062] dfe0: b6f81d20 beeb25d0 00000713 b6eb2a04 600f0010 01200011 00000000 00000000
[  171.440348] [<c021a5c0>] (kmem_cache_alloc) from [<c0257c6c>] (copy_fs_struct+0x20/0xc4)
[  171.450516] [<c0257c6c>] (copy_fs_struct) from [<c0126010>] (copy_process.part.5+0x794/0x16f4)
[  171.461328] [<c0126010>] (copy_process.part.5) from [<c01270b0>] (_do_fork+0xa0/0x6f0)
[  171.471267] [<c01270b0>] (_do_fork) from [<c01277cc>] (SyS_clone+0x20/0x28)
[  171.480015] [<c01277cc>] (SyS_clone) from [<c01073e0>] (ret_fast_syscall+0x0/0x1c)
[  171.489520] Code: e10f7000 f10c0080 ebfd1a37 e5942000 (e5923000)
[  171.497174] ---[ end trace 8bdddcdab7343459 ]---

over 7 years ago

0 Biser Gatchev-XID over 7 years ago

TI__Guru**** 393215 points

What Linux version is this?

0 Ben Taylor over 7 years ago in reply to Biser Gatchev-XID

Prodigy 130 points

it's currently running 4.9.57.

0 Pavel Botev over 7 years ago in reply to Ben Taylor

TI__Guru**** 170625 points

Hi Ben,

This linux version (4.9.57) is not supported. I would suggest you to switch to some of the AM335x TI PSDK linux versions:

PSDK v4.01 - kernel 4.9.41
PSDK v4.02 - kernel 4.9.59
PSDK v4.03 - kernel 4.9.69
PSDK v5.00 - kernel 4.14.40

Please try with some of the TI supported kernels and let me know the result.

Regards,
Pavel

0 Ben Taylor over 7 years ago in reply to Pavel Botev

Prodigy 130 points

Hi Pavel,

I've upgraded the kernel to 4.9.59 and the problem persists, let me know if theres anything else which you think is worth a try.

Regards

Ben

0 Pavel Botev over 7 years ago in reply to Ben Taylor

TI__Guru**** 170625 points

Ben,

From what I understand you are using AM335x custom board. Is that custom board based on EVM, SK, ICE, BBB?

I see that cpufreq_dt driver is build-in the kernel per the default config file (tisdk_am335x-evm_defconfig) but you are using it as a external loadable mode. If you switch to build-in kernel cpufreq_dt, will you have the same issue?

AM335x pin R19 (ZCE) or M18 (ZCZ) is used in AM335x TI boards (EVM, SK, BBB) in LCD_DATA14 mode. And this pin is used as sysboot14 during device boot up (ROM Code stage). sysboot14 determines the main crystal osc frequency, thus make sure you are not modifying this pin early in the boot process. See AM335x TRM, section 26.1.5.3 Clocking Configuration.

Regards,
Pavel

0 Ben Taylor over 7 years ago in reply to Pavel Botev

Prodigy 130 points

Hi Pavel,

We managed to resolve the issue, it was relating to the state of sysboot14 device boot up. The issue was that we were modifying this pin during startup, which was causing the issues with the cpufreq_dt driver. Once we modified the board to place this pin in the correct state this then resolved the issue.

Many thanks for your help.

Regards

Ben

0 Pavel Botev over 7 years ago in reply to Ben Taylor

TI__Guru**** 170625 points

Ben,

Glad to see you have fix it. Please close/verify/resolve that thread.

Regards,
Pavel

Processors

Processors forum

Linux/AM3358: Kernel crash using GPIO0_10