Hi, the patch below fixes PR 48613 which is an ICE with -O0 -findirect-inlining. Rather than adding "&& optimize" here and there, at this place we can easily see whether there is something to do or not by testing ipa_node_params_vector for NULL. And the flag-triggering combinations can -and are - dealt elsewhere.
Bootstrapped and tested on trunk on x86_64-linux. OK for trunk and subsequently for the 4.6 branch too? Thanks, Martin 2011-06-13 Martin Jambor <mjam...@suse.cz> PR tree-optimization/48613 * ipa-prop.c (ipa_prop_write_jump_functions): Return immediately if ipa_node_params_vector is NULL. Index: src/gcc/ipa-prop.c =================================================================== --- src.orig/gcc/ipa-prop.c +++ src/gcc/ipa-prop.c @@ -2900,12 +2900,15 @@ void ipa_prop_write_jump_functions (cgraph_node_set set) { struct cgraph_node *node; - struct output_block *ob = create_output_block (LTO_section_jump_functions); + struct output_block *ob; unsigned int count = 0; cgraph_node_set_iterator csi; - ob->cgraph_node = NULL; + if (!ipa_node_params_vector) + return; + ob = create_output_block (LTO_section_jump_functions); + ob->cgraph_node = NULL; for (csi = csi_start (set); !csi_end_p (csi); csi_next (&csi)) { node = csi_node (csi);