social.kernel.org

Conversation

Richard Weinberger

Today I finally carved out some time to implement return value capture for the Linux kernel function tracer. As of 6.17-rc1, this is done via the HAVE_FUNCTION_GRAPH_FREGS architecture feature. On ARM32, that basically means stashing function call specific registers in pt_regs inside the ftrace arch code.

Once I wrapped my head around the intended semantics, it came together surprisingly smoothly.

Then I enabled the ftrace self-test… and hit this splat:

[    3.299111] ------------[ cut here ]------------
[    3.299173] WARNING: CPU: 0 PID: 1 at kernel/trace/trace.c:2215 run_tracer_selftest+0x110/0x148
[    3.299441] Modules linked in:
[    3.299725] CPU: 0 UID: 0 PID: 1 Comm: swapper/0 Not tainted 6.17.0-rc1-gd7ee5bdce789 #141 NONE 
[    3.299806] Hardware name: Generic DT based system
[    3.299887] Call trace: 
[    3.299932]  unwind_backtrace from show_stack+0x18/0x1c
[    3.299987]  show_stack from dump_stack_lvl+0x54/0x68
[    3.299999]  dump_stack_lvl from __warn+0x88/0x12c
[    3.300013]  __warn from warn_slowpath_fmt+0x194/0x19c
[    3.300026]  warn_slowpath_fmt from run_tracer_selftest+0x110/0x148
[    3.300039]  run_tracer_selftest from register_tracer+0x11c/0x1cc
[    3.300054]  register_tracer from do_one_initcall+0x60/0x210
[    3.300066]  do_one_initcall from kernel_init_freeable+0x1d4/0x240
[    3.300079]  kernel_init_freeable from kernel_init+0x20/0x140
[    3.300091]  kernel_init from ret_from_fork+0x14/0x28
[    3.300131] Exception stack(0xe080dfb0 to 0xe080dff8)
[    3.300204] dfa0:                                     00000000 00000000 00000000 00000000
[    3.300217] dfc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[    3.300224] dfe0: 00000000 00000000 00000000 00000000 00000013 00000000
[    3.300282] ---[ end trace 0000000000000000 ]---

I double- and triple-checked my changes, still no clue. Then it dawned on me: the test fails even without any of my patches. 🫠

Alison Chaiken

alison@burningboard.net

11 days ago

Reply to @rw

@rw I struggle about whether to report bugs in such cases sometimes. At least your code is all GPL (presumably). When proprietary userspace code triggered schedule_timeout recently, I thought about trying to write a minimum reproducer to post. Naturally $JOB has no time in the schedule to perform such tasks. I couldn't write the code at home on the weekend without using the work HW . . . ugh.

Richard Weinberger

11 days ago

Reply to @alison@burningboard.net

@alison Huh? Do you fear that the splat contains confidential information?

About social.kernel.org

Terms of service

Please do not use this service in violation of the Linux Kernel Code of Conduct. Doing so will result in your account suspension with the referral of the matter to the CoC committee.
"Repeating"/"boosting" someone else's status on this platform will be treated as endorsement and will fall under rule #1.
You are encouraged to use this platform to promote your work on the Linux Kernel, but there is no restriction on permitted topics (with the exception of anything covered by #1 above).
There is no requirement to post in English, but it should be considered the primary language of communication on this platform.

Privacy notice

The admins of this service have access to all posted statuses. They aren't looking, but if it's something they shouldn't know about, then you should not post it on this platform.

Please see the Linux Foundation Privacy Policy, which applies to this platform as well.

Getting your own account

If you would like an account on this instance, please check that the following applies to you:

You are listed in MAINTAINERS or CREDITS
OR: You have a kernel.org account or email address
OR: You have a long and established history of involvement with the Linux Kernel

If the above is true and you agree with the Terms of Service and Privacy Notice listed above, please use these instructions to request an account:

How to request an account on social.kernel.org