[Bug binutils/30150] addr2line returns wrong results after several thousands of requests in pipe mode

[amodra at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=30150

--- Comment #26 from Alan Modra <amodra at gmail dot com> ---
Created attachment 14710
  --> https://sourceware.org/bugzilla/attachment.cgi?id=14710&action=edit
Tidied patch

(In reply to Nick Clifton from comment #16)
> I have found a simpler reproducer for the problem, assuming that a copy of
> vmlinux can be obtained:
> 
>   % addr2line -fipae vmlinux 0xffffffff82e21f2d 0xffffffff82e21f38
> 0xffffffff82e21f2d

Yes, I found a similar reproducer late last night too, and came to the same
conclusion as Nick, that the CU range info was broken.  I hadn't properly
investigated so didn't comment at the time.  FWIW the second address I used to
expose the int3_magic oddity was 0xffffffff819e9ab0.

Regarding the patch, can I suggest moving the extra conditions into
comp_unit_contains_address and renaming it to comp_unit_may_contain_address? 
Note that I haven't looked into why it is also necessary to test line_offset so
this patch is *not* a replacement for Nick's patch.  I also don't want to take
over the bug..

-- 
You are receiving this mail because:
You are on the CC list for the bug.