exynos5 : cpu_suspend for cpuidle broken

It has two reasons:
1: flush_cache_all has been replaced by flush_cache_louis in "suspend.c". This is not working on exynos5250. So when cpu0 enters cpuidle it hangs.
Fix: The fix for this is to call flush_cache_all in "cpuidle.c".
diff --git a/arch/arm/mach-exynos/cpuidle.c b/arch/arm/mach-exynos/cpuidle.c
index fcfe025..a70b53a 100644
--- a/arch/arm/mach-exynos/cpuidle.c
+++ b/arch/arm/mach-exynos/cpuidle.c
@@ -21,6 +21,7 @@
 #include <asm/suspend.h>
 #include <asm/unified.h>
 #include <asm/cpuidle.h>
+#include <asm/cacheflush.h>
 #include <mach/regs-clock.h>
 #include <mach/regs-pmu.h>

@@ -89,6 +90,8 @@ static void restore_cpu_arch_register(void)

 static int idle_finisher(unsigned long flags)
 {
+ if (soc_is_exynos5250())
+ flush_cache_all();
        cpu_do_idle();
        return 1;
 }

2: u-boot in arndaleboard with ubuntu package is booting cpu in HYP mode. so while coming out of cpuidle cpu0 in not in HYP mode or some other misbehaviour.
Fix: use mainline u-boot which is not booting cpu in HYP mode.
As of now diden't found solution to fix cpuidle with cpus in HYP mode.

I did the modification in the cpuidle driver and used the u-boot before the commit introducing the cpu booting in hypervisor mode [1] from the u-boot for arndale [2].

I confirm the AFTR state works with this workaround.

IIUC, the cpu_suspend is not complete as it does not handle the restore for when the cpu is in Hypervisor mode, right ?

I will keep this bug opened until I see a fix for that.

[1] https://git.linaro.org/gitweb?p=people/ronynandy/u-boot-arndale.git;a=commit;h=affeaad51746944e8c356091b12ff9c9c6183c4a

[2] https://git.linaro.org/gitweb?p=people/ronynandy/u-boot-arndale.git;a=summary

cpuidle on exynos is breking after enabling HYP mode in u-boot.