https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126348
--- Comment #4 from Longjun Luo <luolongjuna at gmail dot com> --- v3 posted: https://inbox.sourceware.org/gcc-patches/[email protected]/ The compiler change is unchanged from v2. v3 only pins --param=max-inline-insns-auto in the regression test so that function splitting is exercised independently of target-dependent size estimates; Linaro's precommit CI had reported the test failing on arm-linux-gnueabihf for that reason.
