| Message ID | 1459937403-22657-2-git-send-email-james.greenhalgh@arm.com |
|---|---|
| State | New |
| Headers | show |
On 6 April 2016 at 12:10, James Greenhalgh <james.greenhalgh@arm.com> wrote: > > Hi, > > This change reflects binutils support for CRC, where it is always enabled > for armv8.1-a. > Does v8.1 always enable CRC? If not, then don't you want to change the binutils default instead? Christophe. > OK? > > Thanks, > James > > --- > 2016-04-06 James Greenhalgh <james.greenhalgh@arm.com> > > * config/aarch64/aarch64.h (AARCH64_FL_FOR_ARCH8_1): Also add > AARCH64_FL_CRC. >
On Thu, Apr 07, 2016 at 05:23:59PM +0200, Christophe Lyon wrote: > On 6 April 2016 at 12:10, James Greenhalgh <james.greenhalgh@arm.com> wrote: > > > > Hi, > > > > This change reflects binutils support for CRC, where it is always enabled > > for armv8.1-a. > > > > Does v8.1 always enable CRC? Yes. -march=armv8.1-a should always enable CRC. Unfortunately some binutils versions do not honour this, which is why in the next patch we must always put +crc out. > If not, then don't you want to change the binutils default instead? No, this patch is the correct thing to do - regadless of what binutils does, GCC should enable access to the CRC intrinsics with -march=armv8.1-a, so we want this patch. Thanks, James > > --- > > 2016-04-06 James Greenhalgh <james.greenhalgh@arm.com> > > > > * config/aarch64/aarch64.h (AARCH64_FL_FOR_ARCH8_1): Also add > > AARCH64_FL_CRC. > >
diff --git a/gcc/config/aarch64/aarch64.h b/gcc/config/aarch64/aarch64.h index 7750d1c..15d7e40 100644 --- a/gcc/config/aarch64/aarch64.h +++ b/gcc/config/aarch64/aarch64.h @@ -145,7 +145,7 @@ extern unsigned aarch64_architecture_version; /* Architecture flags that effect instruction selection. */ #define AARCH64_FL_FOR_ARCH8 (AARCH64_FL_FPSIMD) #define AARCH64_FL_FOR_ARCH8_1 \ - (AARCH64_FL_FOR_ARCH8 | AARCH64_FL_LSE | AARCH64_FL_V8_1) + (AARCH64_FL_FOR_ARCH8 | AARCH64_FL_LSE | AARCH64_FL_CRC | AARCH64_FL_V8_1) /* Macros to test ISA flags. */