Kernel panic on ARM

Kernel panic on ARM

[Alexandre Karlov]

Hello,=20

I'm currently evaluting the deployment of wireguard on an ARM-based =
32-bit module (AM335x Cortex-A8). After running for some time (5, 10, 20 =
minutes) with some packets regularly going through wg interface, the =
kernel goes in panic mode. I have tried to compile wg as a module as =
well as a patch to the kernel with the end result being the same. The =
kernel version is 4.4.13.

According to netconsole traces it seems to segfault in different places, =
but related to to some irq stuff. I'm joining two netconsole traces and =
will be happy to provide anything useful for debugging.=20

Thanks for the great work on wireguard BTW.=20

Cheers,
Alex


[ 2571.967115] ------------[ cut here ]------------
[ 2571.972045] WARNING: CPU: -844340024 PID: -870604168 at =
kernel/sched/core.c:2960 preempt_count_sub+0xa4/0x130()
[ 2571.982580] DEBUG_LOCKS_WARN_ON(val > preempt_count())
[ 2571.987774] Modules linked in: wireguard(O) xt_hashlimit udp_tunnel =
ip6_udp_tunnel netconsole
[ 2571.997006] CPU: 0 PID: 1480 Comm: i2_app Tainted: G        W  O    =
4.4.13tdma335x #10
[ 2572.005277] Hardware name: Generic AM33XX (Flattened Device Tree)
[ 2572.011673] ---[ end trace 21fe04c5c6b5534c ]---
[ 2572.016522] Alignment trap: not handling instruction e1921f9f at =
[<c0097630>]
[ 2572.023977] Unhandled fault: alignment exception (0x001) at =
0xc098e3cb
[ 2572.030809] Unable to handle kernel paging request at virtual address =
1b000014
[ 2572.038353] pgd =3D cc150000
[ 2572.041187] [1b000014] *pgd=3D00000000
[ 2572.044950] Internal error: Oops: 5 [#1] PREEMPT SMP ARM
[ 2572.050504] Modules linked in: wireguard(O) xt_hashlimit udp_tunnel =
ip6_udp_tunnel netconsole
[ 2572.059533] CPU: 0 PID: 1480 Comm: i2_app Tainted: G        W  O    =
4.4.13tdma335x #10
[ 2572.067798] Hardware name: Generic AM33XX (Flattened Device Tree)
[ 2572.074166] task: cdde9f80 ti: cc1b8000 task.ti: cc1b8000
[ 2572.079818] PC is at show_pte+0x2c/0xfc
[ 2572.083834] LR is at do_DataAbort+0x90/0xc8
[ 2572.088210] pc : [<c0026ab4>]    lr : [<c00092a0>]    psr: 200d0193
[ 2572.088210] sp : cc1ba020  ip : cc1ba040  fp : cc1ba03c
[ 2572.100194] r10: cc1ba1d0  r9 : cd82e000  r8 : cc1ba0f0
[ 2572.105652] r7 : c098e3cb  r6 : c0028b8c  r5 : c098e3cb  r4 : =
1afffff4
[ 2572.112474] r3 : c09c3fb8  r2 : 00000002  r1 : c098e3cb  r0 : =
1afffff4
[ 2572.119295] Flags: nzCv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  =
Segment none
[ 2572.126837] Control: 10c5387d  Table: 8c160019  DAC: 00000051
[ 2572.132842] Process i2_app (pid: 1480, stack limit =3D 0xcc1b8218)
[ 2572.139115] Stack: (0xcc1ba020 to 0xcc1ba000)
[ 2572.143691] [<c0026ab4>] (show_pte) from [<c00092a0>] =
(do_DataAbort+0x90/0xc8)
[ 2572.151250] [<c00092a0>] (do_DataAbort) from [<c0664a24>] =
(__dabt_svc+0x44/0x80)
[ 2572.158976] Exception stack(0xcc1ba0f0 to 0xcc1ba138)
[ 2572.164265] a0e0:                                     c07c9578 =
c07da9cc c098e3cb c098e3bf
[ 2572.172810] a100: c098e3c0 a00d0193 00000001 00000000 00000001 =
cd82e000 cc1ba1d0 cc1ba154
[ 2572.181353] a120: cc1ba130 cc1ba140 c0300ad0 c0097634 a00d0193 =
ffffffff
[ 2572.188280] [<c0664a24>] (__dabt_svc) from [<c0097634>] =
(rcu_eqs_exit_common.constprop.24+0x40/0x5c)
[ 2572.197824] [<c0097634>] (rcu_eqs_exit_common.constprop.24) from =
[<c00983f0>] (rcu_irq_enter+0x70/0x78)
[ 2572.207651] [<c00983f0>] (rcu_irq_enter) from [<c00463d0>] =
(irq_enter+0x18/0x90)
[ 2572.215388] [<c00463d0>] (irq_enter) from [<c008a830>] =
(__handle_domain_irq+0x5c/0xfc)
[ 2572.223668] [<c008a830>] (__handle_domain_irq) from [<c00094f4>] =
(omap_intc_handle_irq+0x44/0xa0)
[ 2572.232943] [<c00094f4>] (omap_intc_handle_irq) from [<c0664ab8>] =
(__irq_svc+0x58/0x98)
[ 2572.241301] Exception stack(0xcc1ba1d0 to 0xcc1ba218)
[ 2572.246583] a1c0:                                     cdfebdb0 =
400d0113 00000000 c0998498
[ 2572.255130] a1e0: 400d0113 cc654d00 cdfeb800 0000008a cdfebcc0 =
cdfebdb0 400d0113 cc1ba234
[ 2572.263672] a200: cc1ba220 cc1ba220 c06643a8 c06643ac 400d0113 =
ffffffff
[ 2572.270588] [<c0664ab8>] (__irq_svc) from [<c06643ac>] =
(_raw_spin_unlock_irqrestore+0x7c/0x80)
[ 2572.279593] [<c06643ac>] (_raw_spin_unlock_irqrestore) from =
[<c03fd348>] (usbnet_start_xmit+0x2b0/0x6d4)
[ 2572.289509] [<c03fd348>] (usbnet_start_xmit) from [<c053b270>] =
(dev_hard_start_xmit+0x2b8/0x500)
[ 2572.298709] [<c053b270>] (dev_hard_start_xmit) from [<c055d24c>] =
(sch_direct_xmit+0xd4/0x1fc)
[ 2572.307623] [<c055d24c>] (sch_direct_xmit) from [<c053b7f8>] =
(__dev_queue_xmit+0x1d4/0x598)
[ 2572.316352] [<c053b7f8>] (__dev_queue_xmit) from [<c053bbd8>] =
(dev_queue_xmit+0x1c/0x20)
[ 2572.324822] [<c053bbd8>] (dev_queue_xmit) from [<c057bd68>] =
(ip_finish_output2+0x18c/0x3e0)
[ 2572.333553] [<c057bd68>] (ip_finish_output2) from [<c057d92c>] =
(ip_finish_output+0x170/0x21c)
[ 2572.342465] [<c057d92c>] (ip_finish_output) from [<c057e4c8>] =
(ip_output+0x120/0x130)
[ 2572.350649] [<c057e4c8>] (ip_output) from [<c057db6c>] =
(ip_local_out+0x48/0x4c)
[ 2572.358301] [<c057db6c>] (ip_local_out) from [<c05c3b88>] =
(iptunnel_xmit+0x120/0x134)
[ 2572.366518] [<c05c3b88>] (iptunnel_xmit) from [<bf00923c>] =
(udp_tunnel_xmit_skb+0xf0/0xf8 [udp_tunnel])
[ 2572.376388] [<bf00923c>] (udp_tunnel_xmit_skb [udp_tunnel]) from =
[<bf018d6c>] (socket_send_skb_to_peer+0x2b8/0x684 [wireguard])
[ 2572.388412] [<bf018d6c>] (socket_send_skb_to_peer [wireguard]) from =
[<bf0178a8>] (packet_create_data_done+0x60/0x1bc [wireguard])
[ 2572.400609] [<bf0178a8>] (packet_create_data_done [wireguard]) from =
[<bf016c20>] (packet_create_data+0x15c/0x4d0 [wireguard])
[ 2572.412446] [<bf016c20>] (packet_create_data [wireguard]) from =
[<bf0176dc>] (packet_send_queue+0xa8/0x214 [wireguard])
[ 2572.423645] [<bf0176dc>] (packet_send_queue [wireguard]) from =
[<bf0155b4>] (xmit+0x178/0x2d0 [wireguard])
[ 2572.433657] [<bf0155b4>] (xmit [wireguard]) from [<c053b270>] =
(dev_hard_start_xmit+0x2b8/0x500)
[ 2572.442753] [<c053b270>] (dev_hard_start_xmit) from [<c053bb28>] =
(__dev_queue_xmit+0x504/0x598)
[ 2572.451849] [<c053bb28>] (__dev_queue_xmit) from [<c053bbd8>] =
(dev_queue_xmit+0x1c/0x20)
[ 2572.460308] [<c053bbd8>] (dev_queue_xmit) from [<c0544fe0>] =
(neigh_direct_output+0x1c/0x20)
[ 2572.469038] [<c0544fe0>] (neigh_direct_output) from [<c057bd80>] =
(ip_finish_output2+0x1a4/0x3e0)
[ 2572.478223] [<c057bd80>] (ip_finish_output2) from [<c057d92c>] =
(ip_finish_output+0x170/0x21c)
[ 2572.487128] [<c057d92c>] (ip_finish_output) from [<c057e4c8>] =
(ip_output+0x120/0x130)
[ 2572.495312] [<c057e4c8>] (ip_output) from [<c057db6c>] =
(ip_local_out+0x48/0x4c)
[ 2572.502956] [<c057db6c>] (ip_local_out) from [<c057de7c>] =
(ip_queue_xmit+0x158/0x3d0)
[ 2572.511152] [<c057de7c>] (ip_queue_xmit) from [<c0596c78>] =
(tcp_transmit_skb+0x454/0x8e8)
[ 2572.519708] [<c0596c78>] (tcp_transmit_skb) from [<c0598df0>] =
(tcp_send_ack+0x150/0x1a0)
[ 2572.528170] [<c0598df0>] (tcp_send_ack) from [<c058bfe4>] =
(__tcp_ack_snd_check+0x60/0xbc)
[ 2572.536718] [<c058bfe4>] (__tcp_ack_snd_check) from [<c0593a08>] =
(tcp_rcv_established+0x498/0x760)
[ 2572.546082] [<c0593a08>] (tcp_rcv_established) from [<c059cdf8>] =
(tcp_v4_do_rcv+0x154/0x1b4)
[ 2572.554901] [<c059cdf8>] (tcp_v4_do_rcv) from [<c059efc4>] =
(tcp_v4_rcv+0xb8c/0xbe0)
[ 2572.562908] [<c059efc4>] (tcp_v4_rcv) from [<c0578240>] =
(ip_local_deliver_finish+0xbc/0x284)
[ 2572.571732] [<c0578240>] (ip_local_deliver_finish) from [<c05789bc>] =
(ip_local_deliver+0xe8/0xf4)
[ 2572.581005] [<c05789bc>] (ip_local_deliver) from [<c0578580>] =
(ip_rcv_finish+0x178/0x38c)
[ 2572.589557] [<c0578580>] (ip_rcv_finish) from [<c0578d90>] =
(ip_rcv+0x3c8/0x4ac)
[ 2572.597199] [<c0578d90>] (ip_rcv) from [<c0535624>] =
(__netif_receive_skb_core+0x2f8/0x9a0)
[ 2572.605839] [<c0535624>] (__netif_receive_skb_core) from [<c0537d04>] =
(__netif_receive_skb+0x38/0x94)
[ 2572.615473] [<c0537d04>] (__netif_receive_skb) from [<c0537dfc>] =
(process_backlog+0x9c/0x158)
[ 2572.624384] [<c0537dfc>] (process_backlog) from [<c0538568>] =
(net_rx_action+0x284/0x430)
[ 2572.632844] [<c0538568>] (net_rx_action) from [<c0045ec4>] =
(__do_softirq+0x180/0x438)
[ 2572.641029] [<c0045ec4>] (__do_softirq) from [<c00464f0>] =
(irq_exit+0xa8/0xdc)
[ 2572.648579] [<c00464f0>] (irq_exit) from [<c008a864>] =
(__handle_domain_irq+0x90/0xfc)
[ 2572.656772] [<c008a864>] (__handle_domain_irq) from [<c00094f4>] =
(omap_intc_handle_irq+0x44/0xa0)
[ 2572.666047] [<c00094f4>] (omap_intc_handle_irq) from [<c0664ab8>] =
(__irq_svc+0x58/0x98)
[ 2572.674406] Exception stack(0xcc1baee8 to 0xcc1baf30)
[ 2572.679689] aee0:                   cc1bb220 cc1bb658 cc1bb820 =
00000000 00000000 00000000
[ 2572.688235] af00: 036d5439 00c34f61 0168a24a 00000000 cc1bb788 =
cc1baf5c 00000000 cc1baf38
[ 2572.696776] af20: 036d5439 bf01d7d0 400d0013 ffffffff
[ 2572.702084] [<c0664ab8>] (__irq_svc) from [<bf01d7d0>] =
(fproduct+0x4ec/0x58c [wireguard])
[ 2572.710662] [<bf01d7d0>] (fproduct [wireguard]) from [<bf01f3f8>] =
(fmonty+0x638/0xec8 [wireguard])
[ 2572.720058] [<bf01f3f8>] (fmonty [wireguard]) from [<bf020620>] =
(cmult+0x998/0x10d4 [wireguard])
[ 2572.729265] [<bf020620>] (cmult [wireguard]) from [<bf0216a8>] =
(curve25519+0x688/0xc38 [wireguard])
[ 2572.738741] [<bf0216a8>] (curve25519 [wireguard]) from [<bf01479c>] =
(noise_handshake_consume_response+0x1a8/0x42c [wireguard])
[ 2572.750665] [<bf01479c>] (noise_handshake_consume_response =
[wireguard]) from [<bf017ccc>] (receive_handshake_packet+0x2c8/0x52c =
[wireguard])
[ 2572.763856] [<bf017ccc>] (receive_handshake_packet [wireguard]) from =
[<bf017f70>] (packet_process_queued_handshake_packets+0x40/0x68 =
[wireguard])
[ 2572.777496] [<bf017f70>] (packet_process_queued_handshake_packets =
[wireguard]) from [<c005a028>] (process_one_work+0x1fc/0x578)
[ 2572.789487] [<c005a028>] (process_one_work) from [<c005aed8>] =
(worker_thread+0x54/0x52c)
[ 2572.797949] [<c005aed8>] (worker_thread) from [<c00606b4>] =
(kthread+0xec/0x104)
[ 2572.805613] [<c00606b4>] (kthread) from [<c0016850>] =
(ret_from_fork+0x14/0x24)
[ 2572.813168] Code: e34c309c e1a05001 11a04000 01a04003 (e5941020)=20
[ 2572.819555] ---[ end trace 21fe04c5c6b5534d ]---



  836.681511] ------------[ cut here ]------------
[  836.686699] Unable to handle kernel NULL pointer dereference at =
virtual address 000001c8
[  836.695164] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.703617] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.712078] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.720527] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.728975] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.737424] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.745873] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.754322] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.762774] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.771224] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.779676] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.788126] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.796575] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.805024] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.813472] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.821921] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.830369] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.838820] Unable to handle kernel NULL pointer dereference at =
virtual address 00000025
[  836.847325] Unable to handle kernel NULL pointer dereference at =
virtual address 000001c8
[  836.855777] pgd =3D cb928000
[  836.858610] [000001c8] *pgd=3D8b9c9831, *pte=3D00000000, =
*ppte=3D00000000
[  836.865252] Internal error: Oops: 17 [#1] PREEMPT SMP ARM
[  836.870896] Modules linked in: wireguard(O) xt_hashlimit udp_tunnel =
ip6_udp_tunnel netconsole
[  836.879932] CPU: 0 PID: 3898 Comm: i2_app Tainted: G        W  O    =
4.4.13tdma335x #10
[  836.888198] Hardware name: Generic AM33XX (Flattened Device Tree)
[  836.894566] task: cb983480 ti: cbe68000 task.ti: cbe68000
[  836.900224] PC is at do_page_fault+0x48/0x44c
[  836.904782] LR is at do_page_fault+0x1f8/0x44c
[  836.909429] pc : [<c0665230>]    lr : [<c06653e0>]    psr: 20010193
[  836.909429] sp : cbe6a108  ip : cbe6a108  fp : cbe6a15c
[  836.921416] r10: cbe7c000  r9 : ffffe000  r8 : cbe6a210
[  836.926875] r7 : 000001c8  r6 : 00000017  r5 : 00000000  r4 : =
cbe6a210
[  836.933693] r3 : 20010193  r2 : cbe6a000  r1 : c07db0b4  r0 : =
00000001
[  836.940511] Flags: nzCv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  =
Segment none
[  836.948053] Control: 10c5387d  Table: 8bd88019  DAC: 00000051
[  836.954054] Process i2_app (pid: 3898, stack limit =3D 0xcbe68218)
[  836.960326] Stack: (0xcbe6a108 to 0xcbe6a000)
[  836.964906] [<c0665230>] (do_page_fault) from [<c000925c>] =
(do_DataAbort+0x4c/0xc8)
[  836.972933] [<c000925c>] (do_DataAbort) from [<c0664a24>] =
(__dabt_svc+0x44/0x80)
[  836.980658] Exception stack(0xcbe6a210 to 0xcbe6a258)
[  836.985939] a200:                                     00000001 =
c07db0b4 cbe6a000 20010193
[  836.994486] a220: cbe6a368 00000000 00000017 000001c8 cbe6a368 =
ffffe000 cbe7c000 cbe6a2b4
[  837.003029] a240: cbe6a260 cbe6a260 c06653e0 c0665230 20010193 =
ffffffff
[  837.009944] [<c0664a24>] (__dabt_svc) from [<c0665230>] =
(do_page_fault+0x48/0x44c)
[  837.017860] [<c0665230>] (do_page_fault) from [<c000925c>] =
(do_DataAbort+0x4c/0xc8)
[  837.025865] [<c000925c>] (do_DataAbort) from [<c0664a24>] =
(__dabt_svc+0x44/0x80)
[  837.033592] Exception stack(0xcbe6a368 to 0xcbe6a3b0)
[  837.038875] a360:                   00000001 c07db0b4 cbe6a000 =
20010193 cbe6a4c0 00000000
[  837.047420] a380: 00000017 000001c8 cbe6a4c0 ffffe000 cbe7c000 =
cbe6a40c cbe6a3b8 cbe6a3b8
[  837.055958] a3a0: c06653e0 c0665230 20010193 ffffffff
[  837.061243] [<c0664a24>] (__dabt_svc) from [<c0665230>] =
(do_page_fault+0x48/0x44c)
[  837.069155] [<c0665230>] (do_page_fault) from [<c000925c>] =
(do_DataAbort+0x4c/0xc8)
[  837.077158] [<c000925c>] (do_DataAbort) from [<c0664a24>] =
(__dabt_svc+0x44/0x80)
[  837.084879] Exception stack(0xcbe6a4c0 to 0xcbe6a508)
[  837.090168] a4c0: 00000001 c07db0b4 cbe6a000 20010193 cbe6a618 =
00000000 00000017 000001c8
[  837.098713] a4e0: cbe6a618 ffffe000 cbe7c000 cbe6a564 cbe6a510 =
cbe6a510 c06653e0 c0665230
[  837.107253] a500: 20010193 ffffffff
[  837.110908] [<c0664a24>] (__dabt_svc) from [<c0665230>] =
(do_page_fault+0x48/0x44c)
[  837.118818] [<c0665230>] (do_page_fault) from [<c000925c>] =
(do_DataAbort+0x4c/0xc8)
[  837.126820] [<c000925c>] (do_DataAbort) from [<c0664a24>] =
(__dabt_svc+0x44/0x80)
[  837.134543] Exception stack(0xcbe6a618 to 0xcbe6a660)
[  837.139822] a600:                                                     =
  00000001 c07db0b4
[  837.148369] a620: cbe6a000 20010193 cbe6a770 00000000 00000017 =
000001c8 cbe6a770 ffffe000
[  837.156916] a640: cbe7c000 cbe6a6bc cbe6a668 cbe6a668 c06653e0 =
c0665230 20010193 ffffffff
[  837.165461] [<c0664a24>] (__dabt_svc) from [<c0665230>] =
(do_page_fault+0x48/0x44c)
[  837.173375] [<c0665230>] (do_page_fault) from [<c000925c>] =
(do_DataAbort+0x4c/0xc8)
[  837.181376] [<c000925c>] (do_DataAbort) from [<c0664a24>] =
(__dabt_svc+0x44/0x80)
[  837.189098] Exception stack(0xcbe6a770 to 0xcbe6a7b8)
[  837.194380] a760:                                     00000001 =
c07db0b4 cbe6a000 20010193
[  837.202926] a780: cbe6a8c8 00000000 00000017 000001c8 cbe6a8c8 =
ffffe000 cbe7c000 cbe6a814
[  837.211474] a7a0: cbe6a7c0 cbe6a7c0 c06653e0 c0665230 20010193 =
ffffffff
[  837.218392] [<c0664a24>] (__dabt_svc) from [<c0665230>] =
(do_page_fault+0x48/0x44c)
[  837.226307] [<c0665230>] (do_page_fault) from [<c000925c>] =
(do_DataAbort+0x4c/0xc8)
[  837.234310] [<c000925c>] (do_DataAbort) from [<c0664a24>] =
(__dabt_svc+0x44/0x80)
[  837.242033] Exception stack(0xcbe6a8c8 to 0xcbe6a910)
[  837.247316] a8c0:                   00000001 c07db0b4 cbe6a000 =
20010193 cbe6aa20 00000000
[  837.255862] a8e0: 00000017 000001c8 cbe6aa20 ffffe000 cbe7c000 =
cbe6a96c cbe6a918 cbe6a918
[  837.264402] a900: c06653e0 c0665230 20010193 ffffffff
[  837.269692] [<c0664a24>] (__dabt_svc) from [<c0665230>] =
(do_page_fault+0x48/0x44c)
[  837.277606] [<c0665230>] (do_page_fault) from [<c000925c>] =
(do_DataAbort+0x4c/0xc8)
[  837.285611] [<c000925c>] (do_DataAbort) from [<c0664a24>] =
(__dabt_svc+0x44/0x80)
[  837.293336] Exception stack(0xcbe6aa20 to 0xcbe6aa68)
[  837.298623] aa20: 00000001 c07db0b4 cbe6a000 20010193 cbe6ab78 =
00000000 00000017 000001c8
[  837.307166] aa40: cbe6ab78 ffffe000 cbe7c000 cbe6aac4 cbe6aa70 =
cbe6aa70 c06653e0 c0665230
[  837.315703] aa60: 20010193 ffffffff
[  837.319364] [<c0664a24>] (__dabt_svc) from [<c0665230>] =
(do_page_fault+0x48/0x44c)
[  837.327281] [<c0665230>] (do_page_fault) from [<c000925c>] =
(do_DataAbort+0x4c/0xc8)
[  837.335282] [<c000925c>] (do_DataAbort) from [<c0664a24>] =
(__dabt_svc+0x44/0x80)
[  837.343004] Exception stack(0xcbe6ab78 to 0xcbe6abc0)
[  837.348281] ab60:                                                     =
  00000001 c07db0b4
[  837.356826] ab80: cbe6a000 20010193 cbe6acd0 00000000 00000017 =
000001c8 cbe6acd0 ffffe000
[  837.365368] aba0: cbe7c000 cbe6ac1c cbe6abc8 cbe6abc8 c06653e0 =
c0665230 20010193 ffffffff
[  837.373911] [<c0664a24>] (__dabt_svc) from [<c0665230>] =
(do_page_fault+0x48/0x44c)
[  837.381822] Code: e1a0200d e0022009 e3130080 e592500c (e59581c8)=20
[  837.388213] ---[ end trace 21fe04c5c6b5534c ]---=

Re: Kernel panic on ARM

[Jason A. Donenfeld]

Hey Alexandre,

Thanks for the detailed debug logs. Is this with the latest snapshot
(0.0.20170517) or with the master git repo? I ask because I recently
added some unreleased-and-potentially-unstable ARM-specific changes to
the master branch, which I'm still trying to get right.

Jason

Re: Kernel panic on ARM

[Alexandre Karlov]

Hi Jason,=20

Thanks for quick reply and sorry I forgot to specify - it is the =
20170517 release.=20

Alex

> On 30 May 2017, at 13:48 , Jason A. Donenfeld <Jason@zx2c4.com> wrote:
>=20
> Hey Alexandre,
>=20
> Thanks for the detailed debug logs. Is this with the latest snapshot
> (0.0.20170517) or with the master git repo? I ask because I recently
> added some unreleased-and-potentially-unstable ARM-specific changes to
> the master branch, which I'm still trying to get right.
>=20
> Jason

Re: Kernel panic on ARM

[Jason A. Donenfeld]

Hi Alexandre,

Could you compile with `make debug`, so that dmesg gets plenty of
debugging output? That will let me see if this is related to
handshakes or to transport messages. An initial theory, based on the
stacktrace there involving fmonty and such, is that the 32-bit
implementation of curve25519-donna uses a lot of stack space, and if
you're on a kernel that puts the softirq stack and the process stack
in the same place, this could be an issue. Also, would you mind
sending your kernel's .config?

Jason

Re: Kernel panic on ARM

[Jason A. Donenfeld]

Hey Alexandre,

Can you try this patch, and tell me if it makes the problem go away?

http://ix.io/vMh

If it works, I'll include this in the next snapshot.

Thanks,
Jason

Re: Kernel panic on ARM

[Alexandre Karlov]

Hi Jason,=20

Thanks for the patch! I compiled it in as well as the extended debugging =
for wg.=20

As I=E2=80=99ve said, the kernel panics in a rather erratical fashion, =
so I let it run for a while and if the problem reappears (let=E2=80=99s =
cross the fingers it wouldn=E2=80=99t), I=E2=80=99ll send the trace and =
kernel .config.=20

Cheers,
Alex

> On 30 May 2017, at 14:45 , Jason A. Donenfeld <Jason@zx2c4.com> wrote:
>=20
> Hey Alexandre,
>=20
> Can you try this patch, and tell me if it makes the problem go away?
>=20
> http://ix.io/vMh
>=20
> If it works, I'll include this in the next snapshot.
>=20
> Thanks,
> Jason

Re: Kernel panic on ARM

[Jason A. Donenfeld]

Hi Alexandre,

As 2 hours have gone by, I assume things are working somewhat decently. Great!

If that patch is stable for you, you might be interested in trying out
the latest `master` branch in the git repo:

$ git clone https://git.zx2c4.com/WireGuard
$ cd WireGuard/src
$ make debug -j$(nproc) && make install

I'll probably make a new snapshot in a few days, but if you'd like to
try things out in advance, that'd be welcome. The master branch
contains some NEON accelerations for the hardware you're using, so if
all goes well, you should see an increase in performance.

Regards,
Jason

Re: Kernel panic on ARM

[Alexandre Karlov]

Hey Jason,=20

I confirm, the hardware is now running for almost 4 hours without any =
issues, so let=E2=80=99s be optimistic and say the patch solved it.=20

I will continue to run the patched version on one device and try the =
latest master branch on another one. I=E2=80=99ll report there is =
something weird.=20

Thanks for your quick help and keep up the great work!

Cheers,
Alexandre=20

> On 30 May 2017, at 18:04 , Jason A. Donenfeld <Jason@zx2c4.com> wrote:
>=20
> Hi Alexandre,
>=20
> As 2 hours have gone by, I assume things are working somewhat =
decently. Great!
>=20
> If that patch is stable for you, you might be interested in trying out
> the latest `master` branch in the git repo:
>=20
> $ git clone https://git.zx2c4.com/WireGuard
> $ cd WireGuard/src
> $ make debug -j$(nproc) && make install
>=20
> I'll probably make a new snapshot in a few days, but if you'd like to
> try things out in advance, that'd be welcome. The master branch
> contains some NEON accelerations for the hardware you're using, so if
> all goes well, you should see an increase in performance.
>=20
> Regards,
> Jason