Issue in blackfin toolchain

[Mike Frysinger]

On Fri, Jun 10, 2011 at 14:38, Joel Sherrill wrote:
> On 06/10/2011 01:23 PM, Mike Frysinger wrote:
>> On Thu, Jun 9, 2011 at 19:54, Rohan Kangralkar wrote:
>>> I am still stuck up with the debugging issue.
>>>
>>> I try to build the system with the debug flags with no optimization. I
>>> have
>>> used the pre-built 4.11 toolchain for blackfin.
>>>
>>> Is this an issue with rtems port of the toolchain?  Can anyone suggest
>>> more
>>> test  that i should be performing?
>>>
>>> If this is not the group that deals with toolchain issues to which group
>>> should this issue be reported to as a bug?
>>>
>>> To which mailing list do you want me to move this thread to the rtems-gdb
>>> mailing list, gnu-gdb or adi-gdb?
>>
>> if the issue is only showing up in the rtems builds, i'm not sure the
>> ADI guys will spend time on it
>
> If you are using ADI supplied toolchains and not having issues
> and the RTEMS ones have issues, then it is likely because
> the ADI toolchains have bug fixes which have not been submitted
> to the FSF.  We stick very close to the FSF tool distributions.
>
> If you have a small test case and can narrow it down, then
> maybe the mailing list for adi is a starting spot.  But only if
> you ask the question: "I have this issue with the standard
> FSF version X.Y of Tool Z.  Do you have any patches related
> to this which are not submitted upstream?"
>
> The other route is to invert the question and ask on the
> gnu gdb list where this might be broken in the FSF source
> so you can dig a patch out of the ADI source.

the binutils/gdb trees should be in sync.  i can see gcc being out of
date as we fell behind (and that's being worked on now), but i would
initially suspect corrupted dwarf information coming out of common gcc
code.
-mike