https://sourceware.org/bugzilla/show_bug.cgi?id=25110
--- Comment #2 from dave.anglin at bell dot net --- On 2019-10-15 10:23 p.m., amodra at gmail dot com wrote: > Works for me with hppa-linux-gcc (GCC) 8.1.1 20180502 It's probably because I recently changed the optimization of indirect calls in gcc-9 and gcc-10. Gcc no longer converts indirect calls to an inline version of $$dyncall if $$dyncall can be reached with a pc-relative or "ble" branch. Only the long PIC branch to $$dyncall is now optimized into inline versions of $$dyncall. The reason is $$dyncall is slightly shorter and the inline sequences got slightly longer in order to load the branch target before the new PIC register value. The above changes are preliminary to fixing a glibc race in lazy binding. Final executables always need to be linked against the millicode library on hppa to provide $$dyncall and various arithmetic operations that are implemented in assembly. Seems we have lost that in test. -- You are receiving this mail because: You are on the CC list for the bug. _______________________________________________ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
