RDB3-v1.1-pfe-calltrace-when-CONFIG_FUNCTION_TRACER-enabled-based-on-bsp41

[hittzt]

Hi,

I noticed that, the S32G's SDK BSP41 has been released, and I tested some features recently.

But when I tested pfe with v6.6 kernel, I caught the following calltrace when the kernel config FUNCTION_TRACER enabled:

[ 8.043360] ------------[ cut here ]------------
[ 8.047582] Misaligned patch-site 0xffffffc0793f59f4
[ 8.052811] WARNING: CPU: 2 PID: 198 at ftrace_call_adjust+0x64/0x70
[ 8.052833] Modules linked in: sja1110(O)
[ 8.052851] CPU: 2 PID: 198 Comm: systemd-udevd Tainted: G O 6.6.25-rt29-g8c8864ffb5db #1
[ 8.052859] Hardware name: NXP S32G3 Reference Design Board (S32G-VNP-RDB3) (DT)
[ 8.052864] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[ 8.052871] pc : ftrace_call_adjust+0x64/0x70
[ 8.052878] lr : ftrace_call_adjust+0x64/0x70
[ 8.052884] sp : ffffffc081d9ba80
[ 8.052887] x29: ffffffc081d9ba80 x28: 00000000000000ad x27: 0000000000001000
[ 8.052898] x26: ffffff8800ffb240 x25: ffffffc079445f00 x24: ffffffc079443010
[ 8.052908] x23: ffffff8800ffb240 x22: 0000000000000000 x21: ffffffc0813b7898
[ 8.052920] x20: 0000000000000000 x19: ffffffc0793f59ec x18: ffffffc081debd58
[ 8.052930] x17: 0000000000000000 x16: ffffffc080bf7d88 x15: 0000007f9e1f7fff
[ 8.052940] x14: 0000000000000000 x13: ffffffc080d83580 x12: ffffffc081258718
[ 8.052950] x11: 0000000000000001 x10: 0000000000000920 x9 : ffffffc0800840b0
[ 8.052959] x8 : ffffff8803922480 x7 : 0000000000000004 x6 : 00000000000000d0
[ 8.052969] x5 : 00000000410fd030 x4 : 0000000000f0000f x3 : 0000000000000004
[ 8.052980] x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffffff8803921b00
[ 8.052990] Call trace:
[ 8.052994] ftrace_call_adjust+0x64/0x70
[ 8.053001] ftrace_process_locs+0x22c/0x400
[ 8.053011] ftrace_module_init+0x44/0x70
[ 8.053018] load_module+0x1404/0x1c70
[ 8.053027] init_module_from_file+0x90/0xe0
[ 8.053035] __arm64_sys_finit_module+0x1c8/0x298
[ 8.053042] invoke_syscall+0x50/0x128
[ 8.053052] el0_svc_common.constprop.0+0xd0/0xf8
[ 8.053061] do_el0_svc+0x28/0x40
[ 8.053069] el0_svc+0x40/0xf8
[ 8.053079] el0t_64_sync_handler+0x13c/0x158
[ 8.053088] el0t_64_sync+0x190/0x198
[ 8.053096] ---[ end trace 0000000000000000 ]---
[ 8.224811] ------------[ cut here ]------------

The whole log is attached.

The reason seems that the pfe is an out of tree modue. I am working to find the root reason. 

Would you please help to confirm the issue and give a workaroud?

Thanks,

Zhantao

[chenyin_h]

Thanks for your reminding. I have reproduced the issues when enabled the PFE.

Seems it is triggered when loading PFE driver, may I know if you had seen any similar prints when loading other kind of out-of-tree modules?

By the way, I did not see similar issue on BSP41 with kernel 5.15.x.

Best Regards

Chenyin

[hittzt]

Hi @chenyin_h ,

No, I did not see similar issue with other driver, like ipc-shm.

It seems the 6.6 kernel update codes related with the out of tree module.

But the root reason is still not clear.

Thanks,

Zhantao

[chenyin_h]

Hi， @hittzt 

Thanks for your findings.

Sorry that I have not reproduced the issue locally on a RDB3 with BSP41, the log is attached for your reference, if anything missed from the test settings, feel free to contact me.

Best Regards

Chenyin

[hittzt]

Hi @chenyin_h ,

Thanks for your reply.

The pfe is enabled in your project, right? Would you please check the the pfe driver probing log and share the whole dmesg?

Thanks,

Zhantao