On 11/17, Timur Tabi wrote:

On 11/17/2017 11:46 AM, Stephen Boyd wrote:

quoted

On 11/16, Timur Tabi wrote:

quoted

On 11/16/17 8:43 PM, Stephen Boyd wrote:

quoted

quoted

Most access to unavailable GPIOs can be blocked via the gpio_chip.request
function.  The one exception is when gpiochip_add_data() scans all of

quoted

If patch 1 is applied is this statement still true?

Nope.

Ok. So what's the point of this patch then? Put another way, is
there some path that doesn't request the gpio still even with
patch 1 applied?

This patch is how the pinctrl-msm driver actually reports the sparse
GPIO map.

Patch 1 prevents gpiolib from accessing GPIOs that haven't been
properly requested.

Patch 2 adds an infrastructure for sparse GPIO maps

Patch 3 updates pinctrl-msm to use that expose that new infrastructure

Patch 4 updates pinctrl-qdf2xxx to work with the new feature in pinctrl-msm

Ok. What path doesn't request the gpio though? If we don't have a
path that fails to request the gpio before using it then I don't
see the need for this patch.

-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project