On Fri, Jun 03, 2016 at 11:56:40AM +0100, Robin Murphy wrote:

Hi Russell,

On 03/06/16 08:58, Russell King wrote:

quoted

Convert DT component matching to use component_match_add_release().

Signed-off-by: Russell King <redacted>
---
 drivers/gpu/drm/arm/hdlcd_drv.c             | 10 ++++++++--
 drivers/gpu/drm/armada/armada_drv.c         |  9 +++++++--
 drivers/gpu/drm/drm_of.c                    | 13 +++++++++----
 drivers/gpu/drm/msm/msm_drv.c               |  8 +++++++-
 drivers/gpu/drm/rockchip/rockchip_drm_drv.c | 13 ++++++++++---
 drivers/gpu/drm/sti/sti_drv.c               |  9 +++++++--
 drivers/gpu/drm/tilcdc/tilcdc_external.c    |  9 +++++++--
 7 files changed, 55 insertions(+), 16 deletions(-)

diff --git a/drivers/gpu/drm/arm/hdlcd_drv.c b/drivers/gpu/drm/arm/hdlcd_drv.c
index b987c63ba8d6..bbde48c4f550 100644
--- a/drivers/gpu/drm/arm/hdlcd_drv.c
+++ b/drivers/gpu/drm/arm/hdlcd_drv.c

@@ -443,11 +443,16 @@ static const struct component_master_ops hdlcd_master_ops = {
 	.unbind		= hdlcd_drm_unbind,
 };

-static int compare_dev(struct device *dev, void *data)
+static int compare_of(struct device *dev, void *data)
 {
 	return dev->of_node == data;
 }

+static void release_of(struct device *dev, void *data)
+{
+	of_node_put(data);
+}

Considering that there's 7 identical copies of this function in this patch
alone, perhaps there's some mileage in defining it commonly as a static
__maybe_unused default_release_of() in component.h or drm_of.h (and maybe
default_compare_of() similarly)?

(Apologies if there's already been some strong argument against that which
I've not seen, but it seems like a reasonable thing to do.)

What we could do is extract out the common bits of OF-component matching
into drivers/of/of_component.c and make the whole thing a tad better.
I'll send a v2 series threaded to this message.

-- 
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
according to speedtest.net.