[Bug lto/118790] [15 Regression] ICE when building fiat (crash in loc_list_from_tree_1)

2025-02-09 Thread law at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118790

Jeffrey A. Law  changed:

   What|Removed |Added

   Priority|P3  |P1

[Bug lto/118790] [15 Regression] ICE when building fiat (crash in loc_list_from_tree_1)

2025-02-07 Thread pinskia at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118790

--- Comment #9 from Andrew Pinski  ---
Created attachment 60420
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=60420&action=edit
Reduced testcase as much as I can do it

`-O2 -flto -flto-partition=none -shared -fallow-argument-mismatch -g -fopenmp`
is needed to reproduce the failure .

Note this reduced testcase needs -flto-partition=none now. Before we didn't
need it.

[Bug lto/118790] [15 Regression] ICE when building fiat (crash in loc_list_from_tree_1)

2025-02-07 Thread pinskia at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118790

Andrew Pinski  changed:

   What|Removed |Added

 Ever confirmed|0   |1
 Status|UNCONFIRMED |NEW
   Last reconfirmed||2025-02-07

--- Comment #11 from Andrew Pinski  ---
Confirmed. I suspect the bisection will point to some optmization which just
exposes the latent bug ...

[Bug lto/118790] [15 Regression] ICE when building fiat (crash in loc_list_from_tree_1)

2025-02-07 Thread pinskia at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118790

--- Comment #10 from Andrew Pinski  ---
But this is related to VLA type used from a nested function though.

[Bug lto/118790] [15 Regression] ICE when building fiat (crash in loc_list_from_tree_1)

2025-02-07 Thread pinskia at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118790

--- Comment #8 from Andrew Pinski  ---
(In reply to Andrew Pinski from comment #7)
> So I think it is a nested function that has been put in a different
> partition from its outer function. Still checking that theory.

Nope, because using -flto-partition=none still fails.

[Bug lto/118790] [15 Regression] ICE when building fiat (crash in loc_list_from_tree_1)

2025-02-07 Thread pinskia at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118790

Andrew Pinski  changed:

   What|Removed |Added

  Component|debug   |lto

--- Comment #7 from Andrew Pinski  ---
So I think it is a nested function that has been put in a different partition
from its outer function. Still checking that theory.