Re: [PATCH dwarves 0/2] clang true signature follow-on

On 14/07/2026 14:50, Alan Maguire wrote:
> Though the true signature series [1] has not landed yet, this
> is intended as a follow-on series to resolve an issue it raises;
> by sharing the optimization detection code it can result in
> the non true signature case being overly aggressive in omitting
> functions from BTF that have optimized parameters.
> 
> See patch 1 for the details; patch 2 tests the change for gcc.
> 
> This series applies on top of [1]; the danger is we do not
> want to land the true signature series without resolving this
> issue first so hopefully folks can test with this series on top to
> see if issues are resolved.  CI passed with this series on top
> of the true signature series at [2].
> 

Last call on this before I land Yonghong's series with these
patches on top; in my own testing it looks like it resolves
issues Jiri ran into (with functions with optimized params
making it into BTF when true signature was disabled), so unless
anyone spots any additional issues I think Yonghong's series
and this one are safe to land. CI looks good with the combination.

Thanks!

Alan

> [1] https://lore.kernel.org/dwarves/[email protected]/
> [2] https://github.com/alan-maguire/dwarves/actions/runs/29270441984
> 
> Alan Maguire (2):
>   btf_encoder: Only skip optimized parms when ABI changed
>   tests: Add GCC default-BTF optimized parameter coverage
> 
>  btf_encoder.c                  | 97 +++++++++++++++++++++++++---------
>  tests/gcc_optimized_default.sh | 69 ++++++++++++++++++++++++
>  2 files changed, 142 insertions(+), 24 deletions(-)
>  create mode 100755 tests/gcc_optimized_default.sh
> 
> -- 
> 2.43.5
> 
>