Message ID | 20180525090401.10496-1-abrodkin@synopsys.com |
---|---|
State | Rejected |
Delegated to: | Tom Rini |
Headers | show |
Series | [U-Boot] arm: Use mtune for Xscale | expand |
Hello, Just a short update. On Fri, 2018-05-25 at 12:04 +0300, Alexey Brodkin wrote: > Otherwise with GCC 8 we see this warning [which is rendered to an error > with -Werror, which happens for buildman/travisci], see [1]: > ------------------------------->8------------------------------- > cc1: warning: switch -mcpu=xscale conflicts with -march=armv5te switch > ------------------------------->8------------------------------- Even though problem listed above got solved with this patch but it introduces another issue I missed initially with GCC 8.1.0, see [3]: ------------------------------->8------------------------------- arm: + h2200 +arch/arm/cpu/pxa/start.S: Assembler messages: +arch/arm/cpu/pxa/start.S:179: Error: selected processor does not support `strd r4,[r1],#8' in ARM mode ------------------------------->8------------------------------- [3] https://travis-ci.org/abrodkin/u-boot/jobs/383423699 So that doesn't seem to be the easy fix :( -Alexey > Interesting enough with GCC 7.3.0 (which is current default in buildman) > that change leads to another failure, see [2]: > ------------------------------->8------------------------------- > Error: selected processor does not support `pld [lr]' in ARM mode > ------------------------------->8------------------------------- > > So I guess this change might be put in series with buildman toolchain > bump. > > [1] https://travis-ci.org/abrodkin/u-boot/jobs/383219348 > [2] https://travis-ci.org/abrodkin/u-boot/jobs/383594006 > > Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com> > Cc: Albert Aribaud <albert.u.boot@aribaud.net> > --- > arch/arm/Makefile | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/arm/Makefile b/arch/arm/Makefile > index 680c6e8516dc..18e0193c43c0 100644 > --- a/arch/arm/Makefile > +++ b/arch/arm/Makefile > @@ -37,7 +37,7 @@ tune-$(CONFIG_CPU_ARM920T) = > tune-$(CONFIG_CPU_ARM926EJS) = > tune-$(CONFIG_CPU_ARM946ES) = > tune-$(CONFIG_CPU_SA1100) =-mtune=strongarm1100 > -tune-$(CONFIG_CPU_PXA) =-mcpu=xscale > +tune-$(CONFIG_CPU_PXA) =-mtune=xscale > tune-$(CONFIG_CPU_ARM1136) = > tune-$(CONFIG_CPU_ARM1176) = > tune-$(CONFIG_CPU_V7A) =
diff --git a/arch/arm/Makefile b/arch/arm/Makefile index 680c6e8516dc..18e0193c43c0 100644 --- a/arch/arm/Makefile +++ b/arch/arm/Makefile @@ -37,7 +37,7 @@ tune-$(CONFIG_CPU_ARM920T) = tune-$(CONFIG_CPU_ARM926EJS) = tune-$(CONFIG_CPU_ARM946ES) = tune-$(CONFIG_CPU_SA1100) =-mtune=strongarm1100 -tune-$(CONFIG_CPU_PXA) =-mcpu=xscale +tune-$(CONFIG_CPU_PXA) =-mtune=xscale tune-$(CONFIG_CPU_ARM1136) = tune-$(CONFIG_CPU_ARM1176) = tune-$(CONFIG_CPU_V7A) =
Otherwise with GCC 8 we see this warning [which is rendered to an error with -Werror, which happens for buildman/travisci], see [1]: ------------------------------->8------------------------------- cc1: warning: switch -mcpu=xscale conflicts with -march=armv5te switch ------------------------------->8------------------------------- Interesting enough with GCC 7.3.0 (which is current default in buildman) that change leads to another failure, see [2]: ------------------------------->8------------------------------- Error: selected processor does not support `pld [lr]' in ARM mode ------------------------------->8------------------------------- So I guess this change might be put in series with buildman toolchain bump. [1] https://travis-ci.org/abrodkin/u-boot/jobs/383219348 [2] https://travis-ci.org/abrodkin/u-boot/jobs/383594006 Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com> Cc: Albert Aribaud <albert.u.boot@aribaud.net> --- arch/arm/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)