[PATCH RESEND 2/6] clk: exynos: Fix incorrect usage of IS_ERR_OR_NULL

[PATCH RESEND 2/6] clk: exynos: Fix incorrect usage of IS_ERR_OR_NULL

[Tony Prisk]

Resend to include mailing lists.

Replace IS_ERR_OR_NULL with IS_ERR on clk_get results.

In the fail: path of mixer_resources_init() and vp_resources_init()
the first clk tested cannot be NULL either, so IS_ERR_OR_NULL is
removed from these as well. Other clocks may still be NULL as they
haven't been clk_get'd yet.

Signed-off-by: Tony Prisk <linux@prisktech.co.nz>
CC: Inki Dae <inki.dae@samsung.com>
CC: Joonyoung Shim <jy0922.shim@samsung.com>
CC: Seung-Woo Kim <sw0312.kim@samsung.com>
CC: Kyungmin Park <kyungmin.park@samsung.com>
CC: dri-devel@lists.freedesktop.org
---
 drivers/gpu/drm/exynos/exynos_mixer.c |   14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index e7fbb82..dbd97c0 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -972,14 +972,14 @@ static int __devinit mixer_resources_init(struct exynos_drm_hdmi_context *ctx,
 	spin_lock_init(&mixer_res->reg_slock);
 
 	mixer_res->mixer = clk_get(dev, "mixer");
-	if (IS_ERR_OR_NULL(mixer_res->mixer)) {
+	if (IS_ERR(mixer_res->mixer)) {
 		dev_err(dev, "failed to get clock 'mixer'\n");
 		ret = -ENODEV;
 		goto fail;
 	}
 
 	mixer_res->sclk_hdmi = clk_get(dev, "sclk_hdmi");
-	if (IS_ERR_OR_NULL(mixer_res->sclk_hdmi)) {
+	if (IS_ERR(mixer_res->sclk_hdmi)) {
 		dev_err(dev, "failed to get clock 'sclk_hdmi'\n");
 		ret = -ENODEV;
 		goto fail;
@@ -1019,7 +1019,7 @@ static int __devinit mixer_resources_init(struct exynos_drm_hdmi_context *ctx,
 fail:
 	if (!IS_ERR_OR_NULL(mixer_res->sclk_hdmi))
 		clk_put(mixer_res->sclk_hdmi);
-	if (!IS_ERR_OR_NULL(mixer_res->mixer))
+	if (!IS_ERR(mixer_res->mixer))
 		clk_put(mixer_res->mixer);
 	return ret;
 }
@@ -1034,19 +1034,19 @@ static int __devinit vp_resources_init(struct exynos_drm_hdmi_context *ctx,
 	int ret;
 
 	mixer_res->vp = clk_get(dev, "vp");
-	if (IS_ERR_OR_NULL(mixer_res->vp)) {
+	if (IS_ERR(mixer_res->vp)) {
 		dev_err(dev, "failed to get clock 'vp'\n");
 		ret = -ENODEV;
 		goto fail;
 	}
 	mixer_res->sclk_mixer = clk_get(dev, "sclk_mixer");
-	if (IS_ERR_OR_NULL(mixer_res->sclk_mixer)) {
+	if (IS_ERR(mixer_res->sclk_mixer)) {
 		dev_err(dev, "failed to get clock 'sclk_mixer'\n");
 		ret = -ENODEV;
 		goto fail;
 	}
 	mixer_res->sclk_dac = clk_get(dev, "sclk_dac");
-	if (IS_ERR_OR_NULL(mixer_res->sclk_dac)) {
+	if (IS_ERR(mixer_res->sclk_dac)) {
 		dev_err(dev, "failed to get clock 'sclk_dac'\n");
 		ret = -ENODEV;
 		goto fail;
@@ -1077,7 +1077,7 @@ fail:
 		clk_put(mixer_res->sclk_dac);
 	if (!IS_ERR_OR_NULL(mixer_res->sclk_mixer))
 		clk_put(mixer_res->sclk_mixer);
-	if (!IS_ERR_OR_NULL(mixer_res->vp))
+	if (!IS_ERR(mixer_res->vp))
 		clk_put(mixer_res->vp);
 	return ret;
 }
-- 
1.7.9.5

Re: [PATCH RESEND 2/6] clk: exynos: Fix incorrect usage of IS_ERR_OR_NULL

[Dan Carpenter]

On Wed, Dec 19, 2012 at 06:34:04AM +1300, Tony Prisk wrote:
> Resend to include mailing lists.

These kind of comments should go under the Signed of by line under
a "---" line.  They will be removed by git-am instead of being
preserved in the git log.

Signed-off-by bla bla blah
---
Commments...

<patch>

regards,
dan carpenter

[PATCH RESEND 3/6] clk: exynos: Fix incorrect usage of IS_ERR_OR_NULL

[Tony Prisk]

Resend to include mailing lists.

Replace IS_ERR_OR_NULL with IS_ERR on clk_get results.

Signed-off-by: Tony Prisk <linux@prisktech.co.nz>
CC: Inki Dae <inki.dae@samsung.com>
CC: Joonyoung Shim <jy0922.shim@samsung.com>
CC: Seung-Woo Kim <sw0312.kim@samsung.com>
CC: Kyungmin Park <kyungmin.park@samsung.com>
CC: dri-devel@lists.freedesktop.org
---
 drivers/gpu/drm/exynos/exynos_hdmi.c |   10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
index 2c115f8..557ef2f 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
@@ -2167,27 +2167,27 @@ static int __devinit hdmi_resources_init(struct hdmi_context *hdata)
 
 	/* get clocks, power */
 	res->hdmi = clk_get(dev, "hdmi");
-	if (IS_ERR_OR_NULL(res->hdmi)) {
+	if (IS_ERR(res->hdmi)) {
 		DRM_ERROR("failed to get clock 'hdmi'\n");
 		goto fail;
 	}
 	res->sclk_hdmi = clk_get(dev, "sclk_hdmi");
-	if (IS_ERR_OR_NULL(res->sclk_hdmi)) {
+	if (IS_ERR(res->sclk_hdmi)) {
 		DRM_ERROR("failed to get clock 'sclk_hdmi'\n");
 		goto fail;
 	}
 	res->sclk_pixel = clk_get(dev, "sclk_pixel");
-	if (IS_ERR_OR_NULL(res->sclk_pixel)) {
+	if (IS_ERR(res->sclk_pixel)) {
 		DRM_ERROR("failed to get clock 'sclk_pixel'\n");
 		goto fail;
 	}
 	res->sclk_hdmiphy = clk_get(dev, "sclk_hdmiphy");
-	if (IS_ERR_OR_NULL(res->sclk_hdmiphy)) {
+	if (IS_ERR(res->sclk_hdmiphy)) {
 		DRM_ERROR("failed to get clock 'sclk_hdmiphy'\n");
 		goto fail;
 	}
 	res->hdmiphy = clk_get(dev, "hdmiphy");
-	if (IS_ERR_OR_NULL(res->hdmiphy)) {
+	if (IS_ERR(res->hdmiphy)) {
 		DRM_ERROR("failed to get clock 'hdmiphy'\n");
 		goto fail;
 	}
-- 
1.7.9.5

Re: [PATCH RESEND 3/6] clk: exynos: Fix incorrect usage of IS_ERR_OR_NULL

[Dan Carpenter]

On Wed, Dec 19, 2012 at 06:34:05AM +1300, Tony Prisk wrote:
> Resend to include mailing lists.
> 
> Replace IS_ERR_OR_NULL with IS_ERR on clk_get results.
> 

The original code is correct.  clk_get() can return NULL depending
on the .config.

regards,
dan carpenter

Re: [PATCH RESEND 3/6] clk: exynos: Fix incorrect usage of IS_ERR_OR_NULL

[Tony Prisk]

On Tue, 2012-12-18 at 21:39 +0300, Dan Carpenter wrote:
> On Wed, Dec 19, 2012 at 06:34:05AM +1300, Tony Prisk wrote:
> > Resend to include mailing lists.
> > 
> > Replace IS_ERR_OR_NULL with IS_ERR on clk_get results.
> > 
> 
> The original code is correct.  clk_get() can return NULL depending
> on the .config.
> 
> regards,
> dan carpenter

Thanks for than Dan,

Arguably that seems like an incorrect behaviour on the part of the clock
subsystem given that the 'proper' function has kernel doc:

* Returns a struct clk corresponding to the clock producer, or
* valid IS_ERR() condition containing errno.

Therefore the 'empty' version should adhere to the same rules, and not
return NULL.

I've cc'd Mike Turquette as well for his thoughts.

Regards
Tony P

Re: [PATCH RESEND 3/6] clk: exynos: Fix incorrect usage of IS_ERR_OR_NULL

[Dan Carpenter]

I don't care either way, but being different from the documentation
is less bad than crashing which is what your patch does.  Please
be more careful in the future.

regards,
dan carpenter

Re: [PATCH RESEND 3/6] clk: exynos: Fix incorrect usage of IS_ERR_OR_NULL

[Tony Prisk]

On Tue, 2012-12-18 at 22:03 +0300, Dan Carpenter wrote:
> I don't care either way, but being different from the documentation
> is less bad than crashing which is what your patch does.  Please
> be more careful in the future.
> 
> regards,
> dan carpenter

Critism accepted.

Given that the driver depends on (PLAT_SAMSUNG || ARCH_MULTIPLATFORM),
and multiplatform select COMMON_CLK and PLAT_SAMSUNG is selected only by
ARCH_S3C64XX which has HAVE_CLK I didn't see a problem here.

Your point is still valid and I will sort it out with Mike first.

Regards
Tony P

Re: [PATCH RESEND 3/6] clk: exynos: Fix incorrect usage of IS_ERR_OR_NULL

[Tony Prisk]

On Wed, 2012-12-19 at 08:11 +1300, Tony Prisk wrote:
> On Tue, 2012-12-18 at 22:03 +0300, Dan Carpenter wrote:
> > I don't care either way, but being different from the documentation
> > is less bad than crashing which is what your patch does.  Please
> > be more careful in the future.
> > 
> > regards,
> > dan carpenter
> 
> Critism accepted.
> 
> Given that the driver depends on (PLAT_SAMSUNG || ARCH_MULTIPLATFORM),
> and multiplatform select COMMON_CLK and PLAT_SAMSUNG is selected only by
> ARCH_S3C64XX which has HAVE_CLK I didn't see a problem here.
> 
> Your point is still valid and I will sort it out with Mike first.
> 
> Regards
> Tony P

Actually, I was wrong here - PLAT_SAMSUNG is selectable via

PLAT_S3C24XX || ARCH_S3C64XX || PLAT_S5P

but all these options (or the options that select them) seem to select a
clock system as well, hence still no reason for a crash.

Regards
Tony P