[v4,3/3] Cleanup __ieee754_sqrt(f/l)

Message ID DB6PR0801MB2053AAAEC4607DC14F9905D683D70@DB6PR0801MB2053.eurprd08.prod.outlook.com
State New, archived
Headers

Commit Message

Wilco Dijkstra March 16, 2018, 2:36 p.m. UTC
  Joseph Myers wrote:
  
> This appears to have broken the build of sqrtl for m68k, at least with GCC
> 6.

Yes it seems like it still needs the __inline_mathop for some complex macro
manupulations. I've committed this:

Revert m68k __ieee754_sqrt change as it causes a build failure in one
m68k configuration.  m68k-linux-gnu now passes again.

ChangeLog:

2018-03-16  Wilco Dijkstra  <wdijkstr@arm.com>

        * sysdeps/m68k/m680x0/fpu/mathimpl.h (__ieee754_sqrt): Revert previous
        commit.
--
  

Patch

diff --git a/sysdeps/m68k/m680x0/fpu/mathimpl.h b/sysdeps/m68k/m680x0/fpu/mathimpl.h
index d39094cccbd3712dfae4601967ee695722b446a0..848f9cffaadad09443cd539ceb68da5712a93b39 100644
--- a/sysdeps/m68k/m680x0/fpu/mathimpl.h
+++ b/sysdeps/m68k/m680x0/fpu/mathimpl.h
@@ -30,6 +30,7 @@  __inline_mathop	(__ieee754_exp10, tentox,)
 __inline_mathop	(__ieee754_log10, log10,)
 __inline_mathop	(__ieee754_log2, log2,)
 __inline_mathop	(__ieee754_log, logn,)
+__inline_mathop	(__ieee754_sqrt, sqrt,)
 __inline_mathop	(__ieee754_atanh, atanh,)
 
 __m81_defun (double, __ieee754_remainder, (double __x, double __y),)