On Wed, 10 Sep 2014, Jason Cooper wrote:

Christoph,

On Wed, Sep 10, 2014 at 09:15:30AM -0500, Christoph Lameter wrote:

quoted

On Tue, 9 Sep 2014, Tejun Heo wrote:

quoted

On Fri, Sep 05, 2014 at 02:11:23AM +0900, Tejun Heo wrote:

quoted

On Tue, Sep 02, 2014 at 10:00:07AM -0500, Christoph Lameter wrote:

quoted

Subject: irqchip: Properly fetch the per cpu offset

The raw_cpu_read() conversion dropped the fetch of the offset
from base->percpu_base in gic_get_percpu_base.

Signed-off-by: Christoph Lameter <redacted>

Index: linux/drivers/irqchip/irq-gic.c
===================================================================

--- linux.orig/drivers/irqchip/irq-gic.c
+++ linux/drivers/irqchip/irq-gic.c

@@ -102,7 +102,7 @@ static struct gic_chip_data gic_data[MAX
 #ifdef CONFIG_GIC_NON_BANKED
 static void __iomem *gic_get_percpu_base(union gic_base *base)
 {
-	return raw_cpu_read(base->percpu_base);
+	return raw_cpu_read(*base->percpu_base);

I suppose this should go through percpu/for-3.18-consistent-ops?  Can
we please cc irq folks and get acks?

Christoph, please drive this forward.

CCing IRQ subsystem and IRQCHIP drivers maintainers.

Thanks Chris.  I'm a bit behind on irqchip stuff, so I may have missed
something here...  Did we get Cc'd on the original patch that caused the
regression?  I'm fairly certain I haven't seen it to Ack it.

It wouldn't be an issue worth mentioning if it were a vendor-specific
driver, however, the gic is used by damn near every major ARM SoC.

In fact the problematic code lives in a code path that very few 
platform use i.e. CONFIG_GIC_NON_BANKED set by CONFIG_ARCH_EXYNOS4.

And I did ask for a clarification on the proposed fix which produced no 
answer so far.


Nicolas