Re: [PATCH v2] Add frame pointer unwinding for aarch64

[Mark Wielaard <mark@klomp.org>]

On Thu, 2017-04-20 at 14:41 +0200, Ulf Hermann wrote:
> diff --git a/tests/backtrace.aarch64.fp.core.bz2 b/tests/backtrace.aarch64.fp.core.bz2
> new file mode 100644
> index 0000000..ef628af
> Binary files /dev/null and b/tests/backtrace.aarch64.fp.core.bz2 differ
> diff --git a/tests/backtrace.aarch64.fp.exec.bz2 b/tests/backtrace.aarch64.fp.exec.bz2
> new file mode 100755
> index 0000000..69ba887
> Binary files /dev/null and b/tests/backtrace.aarch64.fp.exec.bz2 differ

I got these separately. I assume they are as in the email you sent on
Mon, 10 Apr 2017 14:48:06 +0200 (which didn't hit the list because it
had the binaries attached...)

I'll but them on my mjw/fp-unwind branch (which I am rewriting
currently, not pushed yet).

> diff --git a/tests/run-backtrace-fp-core-aarch64.sh b/tests/run-backtrace-fp-core-aarch64.sh

> +# The binary is generated by compiling backtrace-child without debug
> +# information, but with -fno-omit-frame-pointer.
> +#
> +# gcc -static -O2 -fno-omit-frame-pointer -D_GNU_SOURCE \
> +#     -pthread -o backtrace.aarch64.fp.exec backtrace-child.c \
> +#
> +# Then strip the .eh_frame and .eh_frame_hdr sections from the binary:
> +#
> +# strip -R .eh_frame backtrace.aarch64.fp.exec
> +# strip -R .eh_frame_hdr backtrace.aarch64.fp.exec
> +#
> +# The core is generated by calling the binary with --gencore

This description doesn't seem to match anymore. Should this be:

# The binary is generated by compiling with eh_frame CFI, but with frame
# pointers.
#
# gcc -static -O2 -fno-omit-frame-pointer -fno-asynchronous-unwind-tables \
#     -D_GNU_SOURCE -pthread -o tests/backtrace.x86_64.fp.exec -I. -Ilib \
#     tests/backtrace-child.c
#
# The core is generated by calling the binary with --gencore

Thanks,

Mark