* [PATCH] AArch64 compile dbl-64/e_sqrt.c with -ffp-contract=off
@ 2013-12-17 13:17 Marcus Shawcroft
2013-12-17 14:20 ` Joseph S. Myers
0 siblings, 1 reply; 2+ messages in thread
From: Marcus Shawcroft @ 2013-12-17 13:17 UTC (permalink / raw)
To: libc-ports
[-- Attachment #1: Type: text/plain, Size: 365 bytes --]
This patch fixes the numerical errors in aarch64 sqrt() due to the
compilers use of fnmsub (negated fused multiply-subtract).
Previous thread:
https://sourceware.org/ml/libc-alpha/2013-12/msg00473.html
I'll leave this on the list for 24 hours before committing to give
folks the opportunity to comment on the (in)appropriateness of my
solution...
Cheers
/Marcus
[-- Attachment #2: glibc-e_sqrt-flags.txt --]
[-- Type: text/plain, Size: 769 bytes --]
diff --git a/ports/ChangeLog.aarch64 b/ports/ChangeLog.aarch64
index 09645c3..e65e87a 100644
--- a/ports/ChangeLog.aarch64
+++ b/ports/ChangeLog.aarch64
@@ -1,3 +1,8 @@
+2013-12-17 Marcus Shawcroft <marcus.shawcroft@linaro.org>
+
+ * sysdeps/aarch64/Makefile (CFLAGS-e_sqrt.c): Add
+ $(config-cflags-nofma).
+
2013-12-17 Marcus Shawcroft <marcus.shawcroft@linaro.org>
* sysdeps/aarch64/libm-test-ulps: Regenerated.
diff --git a/ports/sysdeps/aarch64/Makefile b/ports/sysdeps/aarch64/Makefile
index 5af7a16..43341c1 100644
--- a/ports/sysdeps/aarch64/Makefile
+++ b/ports/sysdeps/aarch64/Makefile
@@ -14,3 +14,7 @@ endif
ifeq ($(subdir),csu)
gen-as-const-headers += tlsdesc.sym
endif
+
+ifeq ($(subdir),math)
+CFLAGS-e_sqrt.c = $(config-cflags-nofma)
+endif
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] AArch64 compile dbl-64/e_sqrt.c with -ffp-contract=off
2013-12-17 13:17 [PATCH] AArch64 compile dbl-64/e_sqrt.c with -ffp-contract=off Marcus Shawcroft
@ 2013-12-17 14:20 ` Joseph S. Myers
0 siblings, 0 replies; 2+ messages in thread
From: Joseph S. Myers @ 2013-12-17 14:20 UTC (permalink / raw)
To: Marcus Shawcroft; +Cc: libc-ports
On Tue, 17 Dec 2013, Marcus Shawcroft wrote:
> This patch fixes the numerical errors in aarch64 sqrt() due to the
> compilers use of fnmsub (negated fused multiply-subtract).
>
> Previous thread:
> https://sourceware.org/ml/libc-alpha/2013-12/msg00473.html
>
> I'll leave this on the list for 24 hours before committing to give
> folks the opportunity to comment on the (in)appropriateness of my
> solution...
The source file in question is architecture-independent, so the makefile
change should also be architecture-independent (i.e. in
sysdeps/ieee754/dbl-64/Makefile).
--
Joseph S. Myers
joseph@codesourcery.com
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2013-12-17 14:20 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-12-17 13:17 [PATCH] AArch64 compile dbl-64/e_sqrt.c with -ffp-contract=off Marcus Shawcroft
2013-12-17 14:20 ` Joseph S. Myers
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).