On 04/22/2016 05:59 PM, Waldemar Brodkorb wrote:
Hi Jeff,
Jeff Law wrote,
On 03/19/2016 02:50 AM, Waldemar Brodkorb wrote:
Hi GCC devs,
I like to enable m68k support in Buildroot. They use
a two-stage toolchain building method.
But libgcc contains a header which is not available at this time.
Removing the reference make compilation work fine.
It seems the inclusion of asm/unistd.h is unnecessary.
diff -Nur gcc-5.3.0.orig/libgcc/config/m68k/linux-atomic.c
gcc-5.3.0/libgcc/config/m68k/linux-atomic.c
--- gcc-5.3.0.orig/libgcc/config/m68k/linux-atomic.c 2015-01-05
13:33:28.000000000 +0100
+++ gcc-5.3.0/libgcc/config/m68k/linux-atomic.c 2016-03-19
09:25:07.000000000 +0100
@@ -33,7 +33,6 @@
using the kernel helper defined below. There is no support for
64-bit operations yet. */
-#include <asm/unistd.h>
#include <stdbool.h>
#ifndef __NR_atomic_cmpxchg_32
Should I open a bug report for such a change request or problem?
Thanks. I've installed this with a suitable ChangeLog on the trunk (gcc-7).
Great, thanks.
Do you have an idea, why this bug can't be fixed as suggested?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=53833
Haven't really looked at it yet. m68k is not a priority platform. I
usually try to flush out as many m68k issues as I can each fall, but
even that is probably more time than I can rationally justify.
jeff