On Thu, Nov 13, 2025 at 6:01 AM Richard Biener <[email protected]> wrote:
>
> We currently remove stmts inside of a FOR_EACH_IMM_USE_STMT iteration
> which can be problematical. The following adjusts purge_all_uses
> to gather all stmts to remove and remove them in reverse order
> afterwards which also better deals with debug stmt generation.
>
> Bootstrapped and tested on x86_64-unknown-linux-gnu.
>
> OK?
LGTM but I can't approve it.
>
> Thanks,
> Richard.
>
> PR ipa/122663
> * ipa-param-manipulation.cc (purge_all_uses): Collect
> stmts to remove and process that list in reverse.
> ---
> gcc/ipa-param-manipulation.cc | 13 +++++++++++--
> 1 file changed, 11 insertions(+), 2 deletions(-)
>
> diff --git a/gcc/ipa-param-manipulation.cc b/gcc/ipa-param-manipulation.cc
> index 11f843c48d1..96ab125dee1 100644
> --- a/gcc/ipa-param-manipulation.cc
> +++ b/gcc/ipa-param-manipulation.cc
> @@ -636,6 +636,7 @@ purge_all_uses (tree name, hash_set <tree> *killed_ssas)
> imm_use_iterator imm_iter;
> gimple *stmt;
> auto_vec <tree, 4> worklist;
> + auto_vec <gimple *, 4> kill_list;
>
> worklist.safe_push (name);
> while (!worklist.is_empty ())
> @@ -664,11 +665,19 @@ purge_all_uses (tree name, hash_set <tree> *killed_ssas)
> if (!killed_ssas->add (lhs))
> {
> worklist.safe_push (lhs);
> - gimple_stmt_iterator gsi = gsi_for_stmt (stmt);
> - gsi_remove (&gsi, true);
> + kill_list.safe_push (stmt);
> }
> }
> }
> +
> + /* Remove stmts in reverse and afterwards to properly handle debug stmt
> + generation and to not interfere with immediate use iteration. */
> + while (!kill_list.is_empty ())
> + {
> + stmt = kill_list.pop ();
> + gimple_stmt_iterator gsi = gsi_for_stmt (stmt);
> + gsi_remove (&gsi, true);
> + }
> }
>
> /* Modify actual arguments of a function call in statement currently
> belonging
> --
> 2.51.0
