Home Original page

[PATCH v3 3/3] sframe: FDE function start address relative to SFrame section

Jens Remus jremus@linux.ibm.com
Fri Mar 7 09:42:56 GMT 2025 
On 05.03.2025 15:24, Indu Bhagat wrote:
> On 2/28/25 5:32 AM, Jens Remus wrote:
>> The SFrame FDE sfde_func_start_address field is specified to contain
>> the function offset from .sframe section.  This is true in executables
>> and shared libraries after final link.  In assembler generated objects
>> it contained the function offset from SFrame FDE (using a PC-relative
>> relocation) and the linker performed a respective fixup when merging
>> the .sframe sections.
>>
>> Change the assembler to emit sfde_func_start_address relative from the
>> .sframe section.  Change the linker to only apply the output .sframe
>> section offset as fixup when merging the sections.
>>
>> While at it enhance the comments in the code, fix a typo in SFD_INFO,
>> and remove commented out debugging code.
>>
> 
> Thanks for the patch.
> 
> I would like to consider this patch along with some other patches
> that I have for fixing the outstanding SFrame issues (PR 32589) and
> its repercussions on relocatable links (PR 32666) and --gc-sections.
> 
> I think it is better to review this change along with the fixes for
> the above issues.  I will include this patch in a separate series I
> plan to post soon.

Makes sense. Thank you!

What about the rest of the series?  Ok to commit to mainline?  I think
any further refinement of patch 1 can be done with SFrame V3, once the
requirements are defined.

Thanks and regards,
Jens
-- 
Jens Remus
Linux on Z Development (D3303)
+49-7031-16-1128 Office
jremus@de.ibm.com

IBM

IBM Deutschland Research & Development GmbH; Vorsitzender des Aufsichtsrats: Wolfgang Wendt; Geschäftsführung: David Faller; Sitz der Gesellschaft: Böblingen; Registergericht: Amtsgericht Stuttgart, HRB 243294
IBM Data Privacy Statement: https://www.ibm.com/privacy/
More information about the Binutils mailing list