Re: frame unwinding patches

[Mark Wielaard <mark at klomp dot org>]

On Thu, 2017-04-20 at 11:26 +0200, Ulf Hermann wrote:
> The x86_64 case already works with the test case I sent. Maybe we can
> accept that one before the others. The aarch64 case almost works, but
> seems to generally duplicate the first entry it unwinds by frame
> pointer after unwinding anything by CFI. That should be fixable. I
> will research it and post a follow up patch. The 32bit arm case is a
> horrible mess and we may indeed need to lower our expectations for
> that one. Or maybe I can find a raise() that follows the same frame
> conventions as the gcc I'm using ...

I tweaked the testcases a little to test more things. I reverted the
"Optionally allow unknown symbols in the backtrace tests" and added an
explicit check for a frame in the middle, the backtracegen function.
That showed the ppc32 core testfiles were generated without CFI for the
main executable, so I regenerated those.

Then I added your x86_64 frame pointer unwinder, testcases, my i386
frame pointer unwinder and your aarch64 frame pointer unwinder. I
dropped the arm32 frame pointer unwinder for now (maybe we need a less
demanding testcase for that or, more awesome, add code to translate the
exidx section for that).

I do have a question about the aarch64 frame pointer unwinder (and the
initial frame), but I'll reply to the patch email for that.

I'll post the patches as reply to this message (stripping the binaries)
and have pushed them all to the (rebased) mjw/fpunwind branch:
https://sourceware.org/git/?p=elfutils.git;a=shortlog;h=refs/heads/mjw/fp-unwind
Could you take a look and see if that looks good?

Thanks,

Mark