On Mon, Feb 7, 2011 at 9:25 AM, Enrico Scholz <[email protected]> wrote: > Fix for http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43999 and is > required to build gcc for Thumb2 only architectures (e.g. Cortex-M3). > > Signed-off-by: Enrico Scholz <[email protected]> > --- > recipes/gcc/gcc-4.5.inc | 3 ++- > recipes/gcc/gcc-4.5/thumb2-build.patch | 16 ++++++++++++++++ > 2 files changed, 18 insertions(+), 1 deletions(-) > create mode 100644 recipes/gcc/gcc-4.5/thumb2-build.patch > > diff --git a/recipes/gcc/gcc-4.5.inc b/recipes/gcc/gcc-4.5.inc > index b630528..9ba2e22 100644 > --- a/recipes/gcc/gcc-4.5.inc > +++ b/recipes/gcc/gcc-4.5.inc > @@ -8,7 +8,7 @@ DEPENDS = "mpfr gmp libmpc libelf" > NATIVEDEPS = "mpfr-native gmp-native libmpc-native" > > > -INC_PR = "r31" > +INC_PR = "r32" > > SRCREV = "168622" > PV = "4.5" > @@ -166,6 +166,7 @@ SRC_URI = > "svn://gcc.gnu.org/svn/gcc/branches;module=${BRANCH} \ > # file://linaro/gcc-4.5-linaro-r99457.patch \ > file://gcc-scalar-widening-pr45847.patch \ > file://gcc-arm-volatile-bitfield-fix.patch \ > + file://thumb2-build.patch \ > " > > SRC_URI_append_mips64 = " file://mips64-nomultilib.patch " > diff --git a/recipes/gcc/gcc-4.5/thumb2-build.patch > b/recipes/gcc/gcc-4.5/thumb2-build.patch > new file mode 100644 > index 0000000..75e765e > --- /dev/null > +++ b/recipes/gcc/gcc-4.5/thumb2-build.patch > @@ -0,0 +1,16 @@ > +Fix for http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43999 > + > +Patch was taken from comment #3 (Jing Yu) there; the patch attached > +to the named bugreport is wrong. > + > +Index: gcc/config/arm/lib1funcs.asm > +=================================================================== > +--- a/gcc/config/arm/lib1funcs.asm (revision 163231) > ++++ b/gcc/config/arm/lib1funcs.asm (working copy) > +@@ -642,5 +642,5 @@ > + cmp \dividend, #0 @ Early termination? > +- do_it hs, t > ++ do_it ne, t
This needs to be tested on all arms before being taken in. I would suggest test it on beagleboard and qemuarm that will cover other two case arm/thumb1 if gcc testsuites show no new regressions then we can consider it. > + movnes \curbit, \curbit, lsr #4 @ No, any more bits to do? > + movne \divisor, \divisor, lsr #4 > + bne 1b > -- > 1.7.4 > > > _______________________________________________ > Openembedded-devel mailing list > [email protected] > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel > _______________________________________________ Openembedded-devel mailing list [email protected] http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
