Re: [RFC PATCH] perf/x86: make perf callchain work without CONFIG_FRAME_POINTER

[Josh Poimboeuf]

On Fri, Apr 05, 2019 at 01:25:45AM +0800, Kairui Song wrote:
> Currently perf callchain is not working properly with ORC unwinder,
> we'll get useless in kernel callchain like this:
> 
> perf  6429 [000]    22.498450:             kmem:mm_page_alloc: page=0x176a17 pfn=1534487 order=0 migratetype=0 gfp_flags=GFP_KERNEL
>         ffffffffbe23e32e __alloc_pages_nodemask+0x22e (/lib/modules/5.1.0-rc3+/build/vmlinux)
>             7efdf7f7d3e8 __poll+0x18 (/usr/lib64/libc-2.28.so)
>             5651468729c1 [unknown] (/usr/bin/perf)
>             5651467ee82a main+0x69a (/usr/bin/perf)
>             7efdf7eaf413 __libc_start_main+0xf3 (/usr/lib64/libc-2.28.so)
>         5541f689495641d7 [unknown] ([unknown])
> 
> Without CONFIG_FRAME_POINTER, bp is not reserved as frame pointer so
> can't get callers frame pointer, instead current frame pointer is
> returned when trying to fetch caller registers. The unwinder will error
> out early, and end the stacktrace early.
> 
> So instead of let the unwinder start with the dumped register, we start
> it right where the unwinding started when the stacktrace is triggered by
> trace event directly. And skip until the frame pointer is reached.
> 
> This makes the callchain get the full in kernel stacktrace again:
> 
> perf  6503 [000]  1567.570191:             kmem:mm_page_alloc: page=0x16c904 pfn=1493252 order=0 migratetype=0 gfp_flags=GFP_KERNEL
>         ffffffffb523e2ae __alloc_pages_nodemask+0x22e (/lib/modules/5.1.0-rc3+/build/vmlinux)
>         ffffffffb52383bd __get_free_pages+0xd (/lib/modules/5.1.0-rc3+/build/vmlinux)
>         ffffffffb52fd28a __pollwait+0x8a (/lib/modules/5.1.0-rc3+/build/vmlinux)
>         ffffffffb521426f perf_poll+0x2f (/lib/modules/5.1.0-rc3+/build/vmlinux)
>         ffffffffb52fe3e2 do_sys_poll+0x252 (/lib/modules/5.1.0-rc3+/build/vmlinux)
>         ffffffffb52ff027 __x64_sys_poll+0x37 (/lib/modules/5.1.0-rc3+/build/vmlinux)
>         ffffffffb500418b do_syscall_64+0x5b (/lib/modules/5.1.0-rc3+/build/vmlinux)
>         ffffffffb5a0008c entry_SYSCALL_64_after_hwframe+0x44 (/lib/modules/5.1.0-rc3+/build/vmlinux)
>             7f71e92d03e8 __poll+0x18 (/usr/lib64/libc-2.28.so)
>             55a22960d9c1 [unknown] (/usr/bin/perf)
>             55a22958982a main+0x69a (/usr/bin/perf)
>             7f71e9202413 __libc_start_main+0xf3 (/usr/lib64/libc-2.28.so)
>         5541f689495641d7 [unknown] ([unknown])
> 
> ----
> 
> Just found with ORC unwinder the perf callchain is unusable, and this
> seems fixes it well, any suggestion is welcome, thanks!

Hi Kairui,

Without CONFIG_FRAME_POINTER, the BP register has no meaning, so I don't
see how this patch could work.

Also, perf stack traces seem to work fine for me with ORC.  Can you give
some details on how to recreate the issue?

-- 
Josh