projects / platform / upstream / gcc.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: edbb255) | patch
IPA: support -flto + -flive-patching=inline-clone
authorMartin Liska <mliska@suse.cz>
Wed, 5 Oct 2022 09:15:36 +0000 (11:15 +0200)
committerMartin Liska <mliska@suse.cz>
Fri, 7 Oct 2022 13:46:30 +0000 (15:46 +0200)
commit1a308905c1baf64d0ea4d09d7d92b55e79a2a339
treee549294a5e42a24d75db24c9b32e25d17a03daddtree | snapshot
parentedbb2551d156d69a2e337dcd8daa69f2680d57eacommit | diff
IPA: support -flto + -flive-patching=inline-clone

There's no fundamental reason why -flive-patching=inline-clone can't
coexist with -flto. Yes, one can theoretically have many more clone
function that includes a live patch. It is pretty much the same
as in-module inlining.

gcc/ChangeLog:

* opts.cc (finish_options): Print sorry message only
for -flive-patching=inline-only-static.

gcc/testsuite/ChangeLog:

* gcc.dg/live-patching-2.c: Update scanned pattern.
* gcc.dg/live-patching-5.c: New test.
gcc/opts.cc diff | blob | history
gcc/testsuite/gcc.dg/live-patching-2.c diff | blob | history
gcc/testsuite/gcc.dg/live-patching-5.c [new file with mode: 0644] blob
Domain: System / Toolchain;