On Tue, Jul 28, 2015 at 9:30 AM, Shawn Guo [off-list ref] wrote:

On Tue, Jul 28, 2015 at 02:16:20PM +0000, Shenwei Wang wrote:

quoted

quoted

The files are named with "-imx7".  It's a clear sign that the build of the files
should be controlled by something like related to "imx7".
Ideally, it should be CONFIG_SOC_IMX7.  Since imx7d is the only supported imx7
soc and there is no CONFIG_SOC_IMX7 available so far, it falls on
CONFIG_SOC_IMX7D.

We can also interpret "-imx7" is the first SoC which has the GPCv2 block.

In that case you should name the block gpc-imx7d rather than gpcv2 and
symbol as GPC_IMX7D rather than IMX_GPCV2.

So far, only IMX7D use this GPC.
i.MX8 will use system controller to control power.

best regards
Frank Li

quoted

quoted

quoted

quoted

quoted

+extern struct imx_gpcv2_irq *gpcv2_irq_instance;

Will this give a checkpatch warning?

Yes. Any suggestion for that? Move it to a header file?

Get rid of it.

Then you may resolve the following compile error:

You didn't get my point.  Rather than defining it as a global variable,
we should try to keep imx_gpcv2_irq as an internal data structure to
irqchip driver, and use well-defined interface to access the data from
pm driver when necessary.

Shawn

quoted

CC      arch/arm/mach-imx/pm-imx7.o
arch/arm/mach-imx/pm-imx7.c: In function ?imx_gpcv2_pm_init?:
arch/arm/mach-imx/pm-imx7.c:836:7: error: ?gpcv2_irq_instance? undeclared (first use in this function)
  cd = gpcv2_irq_instance;
       ^
arch/arm/mach-imx/pm-imx7.c:836:7: note: each undeclared identifier is reported only once for each function it appears in
make[1]: *** [arch/arm/mach-imx/pm-imx7.o] Error 1

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel