https://sourceware.org/bugzilla/show_bug.cgi?id=32816

--- Comment #11 from H.J. Lu <hjl.tools at gmail dot com> ---
(In reply to Alan Modra from comment #10)
> Yes, I knew about the ld.so _start.  It should not matter as normally
> *crt1.o will supply _start and thus _start won't cause ld.so to be pulled
> in.  The testcase that failed linked Yebefore pr26580-a.o
> (which defined _start).  If the order was changed the testcase would pass
> without removing _start from the library.

Yes, the pr26580-b.s change looks reasonable.  We should add a test for this.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

Reply via email to