peter64m
commented
1 month ago I can confirm that I have the same problem. I can reproduce the panic in the BMC with unicast or broadcast ping using a 9000 byte buffer and no fragmentation. Windows:
PS C:\Users\peter> ping -f -l 9000 turingpi.lan
It has happened every time
I am building v2.1.0-RC2 to compare the behaviour. Network Interfaces file:
auto lo
iface lo inet loopback
auto eth0
iface eth0 inet dhcp
pre-up /etc/network/nfs_check
wait-delay 15
hostname $(hostname)Console output after receiving a jumbo ping:
_____ _ _ ____ ___ _ _ ____
|_ _| | | | _ \|_ _| \ | |/ ___|
| | | | | | |_) || || \| | | _
| | | |_| | _ < | || |\ | |_| |
|_| \___/|_| \_\___|_| \_|\____|
Welcome to Turing Pi
turingpi login: 2024-08-29T10:25:29.395Z INFO [bmcd] Turing Pi 2 BMC Daemon v2.0.5
[ 188.690107] skbuff: skb_over_panic: text:0913898d len:9042 put:9042 head:f1867692 data:14d1e50c tail:0xc5d8fc94 end:0xc5d8e180 dev:eth0
[ 188.703817] ------------[ cut here ]------------
[ 188.708998] kernel BUG at net/core/skbuff.c:109!
[ 188.714174] Internal error: Oops - BUG: 0 [#1] PREEMPT SMP ARM
[ 188.720714] Modules linked in:
[ 188.724143] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.4.61 #1
[ 188.730780] Hardware name: Generic DT based system
[ 188.736162] PC is at skb_panic+0x40/0x4c
[ 188.740560] LR is at skb_panic+0x40/0x4c
[ 188.744957] pc : [<c04eb2d8>] lr : [<c04eb2d8>] psr: 60070113
[ 188.751985] sp : c0901e28 ip : 00000001 fp : 000000ec
[ 188.757843] r10: 00000000 r9 : 00000040 r8 : 00002352
[ 188.763701] r7 : c5d8e180 r6 : c5d8fc94 r5 : c5d8d942 r4 : c5d8d900
[ 188.771024] r3 : 00000000 r2 : 00000000 r1 : 00000004 r0 : 0000007b
[ 188.778347] Flags: nZCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none
[ 188.786350] Control: 10c5387d Table: 45f0c06a DAC: 00000051
[ 188.792798] Process swapper/0 (pid: 0, stack limit = 0xb719506f)
[ 188.799532] Stack: (0xc0901e28 to 0xc0902000)
[ 188.804421] 1e20: 00002352 c5d8d900 c5d8d942 c5d8fc94 c5d8e180 c6961000
[ 188.813599] 1e40: 00000005 c04e4088 c6961638 c5ce2c80 c7447ec0 c0419a84 c6961738 00000040
[ 188.822776] 1e60: c0901e94 c6961638 c7ea74c0 00000001 c083e4c0 07669000 c0901e94 0000012c
[ 188.831954] 1e80: c0901e9c c04fab48 00000040 ffffd485 c074c3eb c0901e94 c0901e94 c0901e9c
[ 188.841131] 1ea0: c0901e9c c73048aa 00000029 c090208c 0000000c 00000008 c0901ed0 c0900000
[ 188.850312] 1ec0: c083d780 c08372fc 00000101 c0102ad8 10c5387d c0157fd0 00200002 0000000a
[ 188.859488] 1ee0: ffffd484 00000004 40000003 c0902080 c074c3eb c083d780 00000000 c7020400
[ 188.868665] 1f00: 00000001 c082d238 c0900000 10c5387d 00000000 c011db08 00000000 c01577c8
[ 188.877843] 1f20: c0901f48 c8802000 c0904138 c0901f7c c082d238 c0361ff4 c0107fd8 60070013
[ 188.887020] 1f40: ffffffff c01021cc 0006c898 00000000 0006c898 c0114c40 c0900000 c0903de8
[ 188.896197] 1f60: 00000001 c0903e24 c082d238 410fc075 10c5387d 00000000 00000064 c0901f98
[ 188.905375] 1f80: c0107fe8 c0107fd8 60070013 ffffffff 00000051 00000000 c0900000 c013e684
[ 188.914553] 1fa0: 000000cf c0941000 c0903dc0 00000000 c082d238 c013e974 c0941048 c0800ca8
[ 188.923730] 1fc0: ffffffff ffffffff 00000000 c0800568 00000000 c082d238 c7354c97 00000000
[ 188.932907] 1fe0: c0800330 00000051 10c0387d 00000000 47d33000 00000000 00000000 00000000
[ 188.942092] [<c04eb2d8>] (skb_panic) from [<c04e4088>] (skb_put+0x44/0x50)
[ 188.949812] [<c04e4088>] (skb_put) from [<c0419a84>] (geth_poll+0x10c/0x190)
[ 188.957725] [<c0419a84>] (geth_poll) from [<c04fab48>] (net_rx_action+0x118/0x30c)
[ 188.966225] [<c04fab48>] (net_rx_action) from [<c0102ad8>] (__do_softirq+0x1f0/0x278)
[ 188.975018] [<c0102ad8>] (__do_softirq) from [<c011db08>] (irq_exit+0x78/0xd0)
[ 188.983129] [<c011db08>] (irq_exit) from [<c01577c8>] (__handle_domain_irq+0x74/0xa0)
[ 188.991921] [<c01577c8>] (__handle_domain_irq) from [<c0361ff4>] (gic_handle_irq+0x44/0x74)
[ 189.001294] [<c0361ff4>] (gic_handle_irq) from [<c01021cc>] (__irq_svc+0x6c/0xa8)
[ 189.009687] Exception stack(0xc0901f48 to 0xc0901f90)
[ 189.015354] 1f40: 0006c898 00000000 0006c898 c0114c40 c0900000 c0903de8
[ 189.024531] 1f60: 00000001 c0903e24 c082d238 410fc075 10c5387d 00000000 00000064 c0901f98
[ 189.033705] 1f80: c0107fe8 c0107fd8 60070013 ffffffff
[ 189.039377] [<c01021cc>] (__irq_svc) from [<c0107fd8>] (arch_cpu_idle+0x1c/0x38)
[ 189.047681] [<c0107fd8>] (arch_cpu_idle) from [<c013e684>] (do_idle+0xd0/0x120)
[ 189.055886] [<c013e684>] (do_idle) from [<c013e974>] (cpu_startup_entry+0x18/0x1c)
[ 189.064387] [<c013e974>] (cpu_startup_entry) from [<c0800ca8>] (start_kernel+0x38c/0x438)
[ 189.073566] Code: e58d3014 e5903050 e59f0008 ebf1ad82 (e7f001f2)
[ 189.080405] ---[ end trace bdf74f70f70f33c1 ]---
[ 189.085586] Kernel panic - not syncing: Fatal exception in interrupt
[ 189.092728] CPU1: stopping
[ 189.095767] CPU: 1 PID: 0 Comm: swapper/1 Tainted: G D 5.4.61 #1
[ 189.103963] Hardware name: Generic DT based system
[ 189.109348] [<c010e390>] (unwind_backtrace) from [<c010a850>] (show_stack+0x10/0x14)
[ 189.118044] [<c010a850>] (show_stack) from [<c05e432c>] (dump_stack+0x78/0x94)
[ 189.126154] [<c05e432c>] (dump_stack) from [<c010c6d0>] (handle_IPI+0xcc/0x168)
[ 189.134360] [<c010c6d0>] (handle_IPI) from [<c036201c>] (gic_handle_irq+0x6c/0x74)
[ 189.142857] [<c036201c>] (gic_handle_irq) from [<c01021cc>] (__irq_svc+0x6c/0xa8)
[ 189.151250] Exception stack(0xc7097f80 to 0xc7097fc8)
[ 189.156919] 7f80: 0003eb70 00000000 0003eb70 c0114c40 c7096000 c0903de8 00000002 c0903e24
[ 189.166096] 7fa0: 4000406a 410fc075 00000000 00000000 00000000 c7097fd0 c0107fe8 c0107fd8
[ 189.175268] 7fc0: 600d0013 ffffffff
[ 189.179187] [<c01021cc>] (__irq_svc) from [<c0107fd8>] (arch_cpu_idle+0x1c/0x38)
[ 189.187490] [<c0107fd8>] (arch_cpu_idle) from [<c013e684>] (do_idle+0xd0/0x120)
[ 189.195694] [<c013e684>] (do_idle) from [<c013e974>] (cpu_startup_entry+0x18/0x1c)
[ 189.204188] [<c013e974>] (cpu_startup_entry) from [<40102bec>] (0x40102bec)
[ 189.212004] ---[ end Kernel panic - not syncing: Fatal exception in interrupt ]---
Describe the bug The BMC on my TuringPI2 becomes unresponsive as soon as jumbo frames are enabled on my network. I kept seeing ~10 to a maximum of ~40 seconds of pings that would go through before I can't see it onthe network anymore. On some tries I managed to login, but it still crashed after roughly the same amount of time.
To Reproduce Steps to reproduce the behavior:
Expected behavior It would ba nice if the BMC also worked on a network with jumbo frames enabled.
Versions tested on bmc version=2.0.5
Additional context As an easy workaround I disabled jumbo frames on my network and rebooted the TuringPi2 board, which worked. But as soon as I turn on jumbo frames on my network again the BMC crashes.