Merge lp:~ams-codesourcery/gcc-linaro/eabi-half-precision into lp:gcc-linaro/4.6

[Andrew Stubbs]

Add ARM EABI conforming half-precision conversion routine names.

This patch is awaiting upstream review here:
 http://<email address hidden>/msg03210.html

[Ira Rosen (irar)]

I am out of the office until 17/04/2011.

Note: This is an automated response to your message "[Merge]
lp:~ams-codesourcery/gcc-linaro/eabi-half-precision into
lp:gcc-linaro/4.6" sent on 14/4/2011 18:12:13.

This is the only notification you will receive while this person is away.

[Linaro Toolchain Builder (cbuild)]

cbuild has taken a snapshot of this branch at r106733 and queued it for build.

The snapshot is available at:
 http://ex.seabright.co.nz/snapshots/gcc-linaro-4.6+bzr106733~ams-codesourcery~eabi-half-precision.tar.xdelta3.xz

and will be built on the following builders:
 a8-builder a9-builder i686 x86_64

You can track the build queue at:
 http://ex.seabright.co.nz/helpers/scheduler

cbuild-snapshot: gcc-linaro-4.6+bzr106733~ams-codesourcery~eabi-half-precision
cbuild-ancestor: lp:gcc-linaro/4.6+bzr106732
cbuild-state: check

[Michael Hope (michaelh1)]

The return value of some externally visible functions has changed:

-unsigned short
+unsigned int
 __gnu_f2h_alternative(unsigned int x)

Is this change binary compatible? Does it matter?

Note that this patch is being reviewed on mainline and shouldn't be committed until the review is done and patch landed.

[Linaro Toolchain Builder (cbuild)]

cbuild successfully built this on i686-lucid-cbuild109-scorpius-i686r1.

The build results are available at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.6+bzr106733~ams-codesourcery~eabi-half-precision/logs/i686-lucid-cbuild109-scorpius-i686r1

The test suite results were unchanged compared to the branch point lp:gcc-linaro/4.6+bzr106732.

The full testsuite results are at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.6+bzr106733~ams-codesourcery~eabi-half-precision/logs/i686-lucid-cbuild109-scorpius-i686r1/gcc-testsuite.txt

cbuild-checked: i686-lucid-cbuild109-scorpius-i686r1

[Linaro Toolchain Builder (cbuild)]

cbuild successfully built this on x86_64-maverick-cbuild109-crucis-x86_64r1.

The build results are available at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.6+bzr106733~ams-codesourcery~eabi-half-precision/logs/x86_64-maverick-cbuild109-crucis-x86_64r1

The test suite results were unchanged compared to the branch point lp:gcc-linaro/4.6+bzr106732.

The full testsuite results are at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.6+bzr106733~ams-codesourcery~eabi-half-precision/logs/x86_64-maverick-cbuild109-crucis-x86_64r1/gcc-testsuite.txt

cbuild-checked: x86_64-maverick-cbuild109-crucis-x86_64r1

[Andrew Stubbs (ams-codesourcery)]

> The return value of some externally visible functions has changed:
>
> -unsigned short
> +unsigned int
> __gnu_f2h_alternative(unsigned int x)
>
> Is this change binary compatible? Does it matter?

Return values are always passed in a register, and are therefore always de facto int type; the unsigned types will have the same bit pattern in registers. I don't believe this change should have any impact.

[Richard Sandiford (rsandifo)]

For the record, although it's my week roster-wise, I don't
feel comfortable enough with the EABI to review this one.

Unmerged revisions

106733. By Andrew Stubbs on 2011-04-14

EABI half-precision function names.

Merged from Linaro GCC 4.5.

=== modified file 'ChangeLog.linaro'
--- ChangeLog.linaro	2011-04-07 10:52:12 +0000
+++ ChangeLog.linaro	2011-04-14 15:33:00 +0000
@@ -1,3 +1,34 @@
[phone]  Andrew Stubbs  <[e-mail]>
+
+	gcc/
+	* config/arm/arm.c (arm_init_libfuncs): Change __gnu_f2h_ieee to
+	__aeabi_f2h, __gnu_f2h_alternative to __aeabi_f2h_alt, __gnu_h2f_ieee
+	to __aeabi_h2f, and __gnu_h2f_alternative to __aeabi_h2f_alt.
+	* config/arm/fp16.c (__gnu_f2h_internal): Change return type to
+	unsigned int. Change 'sign' variable likewise.
+	(__gnu_h2f_internal): Set to static inline.
+	Change return type to unsigned int. Change 'sign' variable likewise.
+	(ALIAS): New define.
+	(__gnu_f2h_ieee): Change unsigned short to unsigned int.
+	(__gnu_h2f_ieee): Likewise.
+	(__gnu_f2h_alternative): Likewise.
+	(__gnu_h2f_alternative): Likewise.
+	(__aeabi_f2h, __aeabi_h2f): New aliases.
+	(__aeabi_f2h_alt, __aeabi_h2f_alt): Likewise.
+	* config/arm/sfp-machine.h (__extendhfsf2): Set to __aeabi_h2f.
+	(__truncsfhf2): Set to __aeabi_f2h.
+	* config/arm/t-bpabi (LIB2FUNCS_STATIC_EXTRA): Move fp16.c ...
+	(LIB2FUNCS_EXTRA): ... to here.
+	* config/arm/t-symbian (LIB2FUNCS_STATIC_EXTRA): Move fp16.c ...
+	(LIB2FUNCS_EXTRA): ... to here.
+
+	gcc/testsuite/
+	* g++.dg/ext/arm-fp16/arm-fp16-ops-5.C: Check for __aeabi_h2f
+	and __aeabi_f2h.
+	* g++.dg/ext/arm-fp16/arm-fp16-ops-6.C: Likewise.
+	* gcc.dg/torture/arm-fp16-ops-5.c: Likewise.
+	* gcc.dg/torture/arm-fp16-ops-6.c: Likewise.
+
 2011-04-07  Andrew Stubbs  <[e-mail]>
 
 	Merge from FSF GCC 4.6 (svn branches/gcc-4_6-branch 171921).
 
=== modified file 'gcc/config/arm/arm.c'
--- gcc/config/arm/arm.c	2011-03-02 11:38:43 +0000
+++ gcc/config/arm/arm.c	2011-04-14 15:33:00 +0000
@@ -1190,12 +1190,12 @@
       /* Conversions.  */
       set_conv_libfunc (trunc_optab, HFmode, SFmode,
 			(arm_fp16_format == ARM_FP16_FORMAT_IEEE
-			 ? "__gnu_f2h_ieee"
-			 : "__gnu_f2h_alternative"));
+			 ? "__aeabi_f2h"
+			 : "__aeabi_f2h_alt"));
       set_conv_libfunc (sext_optab, SFmode, HFmode, 
 			(arm_fp16_format == ARM_FP16_FORMAT_IEEE
-			 ? "__gnu_h2f_ieee"
-			 : "__gnu_h2f_alternative"));
+			 ? "__aeabi_h2f"
+			 : "__aeabi_h2f_alt"));
       
       /* Arithmetic.  */
       set_optab_libfunc (add_optab, HFmode, NULL);
 
=== modified file 'gcc/config/arm/fp16.c'
--- gcc/config/arm/fp16.c	2009-06-18 11:26:37 +0000
+++ gcc/config/arm/fp16.c	2011-04-14 15:33:00 +0000
@@ -22,10 +22,10 @@
    see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
    <http://www.gnu.org/licenses/>.  */
 
-static inline unsigned short
+static inline unsigned int
 __gnu_f2h_internal(unsigned int a, int ieee)
 {
-  unsigned short sign = (a >> 16) & 0x8000;
+  unsigned int sign = (a >> 16) & 0x8000;
   int aexp = (a >> 23) & 0xff;
   unsigned int mantissa = a & 0x007fffff;
   unsigned int mask;
@@ -95,10 +95,10 @@
   return sign | (((aexp + 14) << 10) + (mantissa >> 13));
 }
 
-unsigned int
-__gnu_h2f_internal(unsigned short a, int ieee)
+static inline unsigned int
+__gnu_h2f_internal(unsigned int a, int ieee)
 {
-  unsigned int sign = (unsigned int)(a & 0x8000) << 16;
+  unsigned int sign = (a & 0x00008000) << 16;
   int aexp = (a >> 10) & 0x1f;
   unsigned int mantissa = a & 0x3ff;
 
@@ -120,26 +120,33 @@
   return sign | (((aexp + 0x70) << 23) + (mantissa << 13));
 }
 
-unsigned short
+#define ALIAS(src, dst) \
+  typeof (src) dst __attribute__ ((alias (#src)));
+
+unsigned int
 __gnu_f2h_ieee(unsigned int a)
 {
   return __gnu_f2h_internal(a, 1);
 }
+ALIAS (__gnu_f2h_ieee, __aeabi_f2h)
 
 unsigned int
-__gnu_h2f_ieee(unsigned short a)
+__gnu_h2f_ieee(unsigned int a)
 {
   return __gnu_h2f_internal(a, 1);
 }
+ALIAS (__gnu_h2f_ieee, __aeabi_h2f)
 
-unsigned short
+unsigned int
 __gnu_f2h_alternative(unsigned int x)
 {
   return __gnu_f2h_internal(x, 0);
 }
+ALIAS (__gnu_f2h_alternative, __aeabi_f2h_alt)
 
 unsigned int
-__gnu_h2f_alternative(unsigned short a)
+__gnu_h2f_alternative(unsigned int a)
 {
   return __gnu_h2f_internal(a, 0);
 }
+ALIAS (__gnu_h2f_alternative, __aeabi_h2f_alt)
 
=== modified file 'gcc/config/arm/libgcc-bpabi.ver'
--- gcc/config/arm/libgcc-bpabi.ver	2009-04-21 18:03:23 +0000
+++ gcc/config/arm/libgcc-bpabi.ver	2011-04-14 15:33:00 +0000
@@ -106,3 +106,10 @@
 GCC_4.3.0 {
   _Unwind_Backtrace
 }
+
+GCC_4.7.0 {
+  __aeabi_f2h
+  __aeabi_f2h_alt
+  __aeabi_h2f
+  __aeabi_h2f_alt
+}
 
=== modified file 'gcc/config/arm/sfp-machine.h'
--- gcc/config/arm/sfp-machine.h	2009-06-18 11:26:37 +0000
+++ gcc/config/arm/sfp-machine.h	2011-04-14 15:33:00 +0000
@@ -99,7 +99,7 @@
 #define __fixdfdi	__aeabi_d2lz
 #define __fixunsdfdi	__aeabi_d2ulz
 #define __floatdidf	__aeabi_l2d
-#define __extendhfsf2	__gnu_h2f_ieee
-#define __truncsfhf2	__gnu_f2h_ieee
+#define __extendhfsf2	__aeabi_h2f
+#define __truncsfhf2	__aeabi_f2h
 
 #endif /* __ARM_EABI__ */
 
=== modified file 'gcc/config/arm/t-bpabi'
--- gcc/config/arm/t-bpabi	2009-06-18 11:26:37 +0000
+++ gcc/config/arm/t-bpabi	2011-04-14 15:33:00 +0000
@@ -21,9 +21,8 @@
 
 # Add the BPABI C functions.
 LIB2FUNCS_EXTRA = $(srcdir)/config/arm/bpabi.c \
-		  $(srcdir)/config/arm/unaligned-funcs.c
-
-LIB2FUNCS_STATIC_EXTRA = $(srcdir)/config/arm/fp16.c
+		  $(srcdir)/config/arm/unaligned-funcs.c \
+		  $(srcdir)/config/arm/fp16.c
 
 UNWIND_H = $(srcdir)/config/arm/unwind-arm.h
 LIB2ADDEH = $(srcdir)/config/arm/unwind-arm.c \
 
=== modified file 'gcc/config/arm/t-symbian'
--- gcc/config/arm/t-symbian	2009-06-21 19:48:15 +0000
+++ gcc/config/arm/t-symbian	2011-04-14 15:33:00 +0000
@@ -36,7 +36,7 @@
 LIB2ADDEHDEP = $(UNWIND_H)
 
 # Include half-float helpers.
-LIB2FUNCS_STATIC_EXTRA = $(srcdir)/config/arm/fp16.c
+LIB2FUNCS_EXTRA += $(srcdir)/config/arm/fp16.c
 
 # Create a multilib for processors with VFP floating-point, and a
 # multilib for those without -- using the soft-float ABI in both
 
=== modified file 'gcc/testsuite/g++.dg/ext/arm-fp16/arm-fp16-ops-5.C'
--- gcc/testsuite/g++.dg/ext/arm-fp16/arm-fp16-ops-5.C	2010-05-24 18:36:31 +0000
+++ gcc/testsuite/g++.dg/ext/arm-fp16/arm-fp16-ops-5.C	2011-04-14 15:33:00 +0000
@@ -13,3 +13,5 @@
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_h\[a-z\]*_ieee" } } */
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_h2f_ieee" } } */
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_f2h_ieee" } } */
+/* { dg-final { scan-assembler-not "\tbl\t__aeabi_h2f" } } */
+/* { dg-final { scan-assembler-not "\tbl\t__aeabi_f2h" } } */
 
=== modified file 'gcc/testsuite/g++.dg/ext/arm-fp16/arm-fp16-ops-6.C'
--- gcc/testsuite/g++.dg/ext/arm-fp16/arm-fp16-ops-6.C	2010-05-24 18:36:31 +0000
+++ gcc/testsuite/g++.dg/ext/arm-fp16/arm-fp16-ops-6.C	2011-04-14 15:33:00 +0000
@@ -13,3 +13,5 @@
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_h\[a-z\]*_ieee" } } */
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_h2f_ieee" } } */
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_f2h_ieee" } } */
+/* { dg-final { scan-assembler-not "\tbl\t__aeabi_h2f" } } */
+/* { dg-final { scan-assembler-not "\tbl\t__aeabi_f2h" } } */
 
=== modified file 'gcc/testsuite/gcc.dg/torture/arm-fp16-ops-5.c'
--- gcc/testsuite/gcc.dg/torture/arm-fp16-ops-5.c	2010-05-24 18:36:31 +0000
+++ gcc/testsuite/gcc.dg/torture/arm-fp16-ops-5.c	2011-04-14 15:33:00 +0000
@@ -13,3 +13,5 @@
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_h\[a-z\]*_ieee" } } */
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_h2f_ieee" } } */
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_f2h_ieee" } } */
+/* { dg-final { scan-assembler-not "\tbl\t__aeabi_h2f" } } */
+/* { dg-final { scan-assembler-not "\tbl\t__aeabi_f2h" } } */
 
=== modified file 'gcc/testsuite/gcc.dg/torture/arm-fp16-ops-6.c'
--- gcc/testsuite/gcc.dg/torture/arm-fp16-ops-6.c	2010-05-24 18:36:31 +0000
+++ gcc/testsuite/gcc.dg/torture/arm-fp16-ops-6.c	2011-04-14 15:33:00 +0000
@@ -13,3 +13,5 @@
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_h\[a-z\]*_ieee" } } */
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_h2f_ieee" } } */
 /* { dg-final { scan-assembler-not "\tbl\t__gnu_f2h_ieee" } } */
+/* { dg-final { scan-assembler-not "\tbl\t__aeabi_h2f" } } */
+/* { dg-final { scan-assembler-not "\tbl\t__aeabi_f2h" } } */