[PATCH v2 28/42] Remove dwarf2_per_cu_data::objfile ()

[Simon Marchi]

On 2020-05-27 5:08 p.m., Tom de Vries wrote:
>> I don't see this failure.  Is it with a particular board file?  Or maybe
>> gcc-version specific?
>>
> 
> No, this is with native, and reproduced with gcc 7.5.0, gcc 4.8.5, gcc
> 10.1.1 and clang 5.0.2.
> 
>> It looks related to what we have discussed here:
>>
>> https://sourceware.org/pipermail/gdb-patches/2020-May/168906.html
>>
>> I think the fix would be to temporarily override this->per_objfile to be
>> caller_per_objfile (with a scoped_restore).  But I'd like to be able to
>> reproduce it to understand what's happening.
> 
> I think maybe installing debug info for glibc and building with
> --with-separate-debug-dir=/usr/lib/debug may do the trick.
> 
> Thanks,
> - Tom
> 

I do have debug info for glibc and it is found:

(gdb) info shared
>From                To                  Syms Read   Shared Object Library
0x00007ffff7dd7ac0  0x00007ffff7df5790  Yes         /lib64/ld-linux-x86-64.so.2
0x00007ffff7bd3da0  0x00007ffff7bd498e  Yes         /lib/x86_64-linux-gnu/libdl.so.2
0x00007ffff78288b0  0x00007ffff797bb04  Yes         /lib/x86_64-linux-gnu/libc.so.6

I tested with these two gccs:

$ gcc --version
gcc (Ubuntu 5.5.0-12ubuntu1~16.04) 5.5.0 20171010
$ gcc-9 --version
gcc-9 (Ubuntu 9.3.0-10ubuntu2~16.04) 9.3.0

I will give a shot at building GDB in an opensuse container [1], we'll see.

Simon

[1] https://hub.docker.com/r/opensuse/leap