[PATCH 0/4] Cleanups and fixes

[PATCH 0/4] Cleanups and fixes

[Liviu Dudau]

From: Liviu Dudau <Liviu.Dudau@arm.com>

Hello,

I believe Brian has sent these patches before, but he did not had any
luck getting them accepted, so it is now my turn :)

These fixes were found while developing code for memory writeback testing
and general use of igt with Mali DP KMS driver, which does not have a
cursor plane.

Best regards,
Liviu

Brian Starkey (4):
  lib/igt_kms: Fix erroneous assert
  lib/igt_kms: Fix override_mode handling
  lib/igt_debugfs: Remove igt_debugfs_t
  lib/igt_debugfs: Only use valid values in igt_crc_to_str()

 lib/igt_debugfs.c | 12 ++++--------
 lib/igt_kms.c     | 19 +++++++++----------
 2 files changed, 13 insertions(+), 18 deletions(-)

-- 
2.13.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 1/4] lib/igt_kms: Fix erroneous assert

[Liviu Dudau]

From: Brian Starkey <brian.starkey@arm.com>

In trying to fix igt_display_init() for devices without cursors, I
actually made matters worse. Fix the assert.

Fixes: 545aa3398223 lib/igt_kms: Remove redundant cursor code
Signed-off-by: Brian Starkey <brian.starkey@arm.com>
---
 lib/igt_kms.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/lib/igt_kms.c b/lib/igt_kms.c
index c77716b4..5e2ef97b 100644
--- a/lib/igt_kms.c
+++ b/lib/igt_kms.c
@@ -1794,11 +1794,11 @@ void igt_display_init(igt_display_t *display, int drm_fd)
 		 */
 		igt_assert_eq(pipe->plane_primary, 0);
 
-		/*
-		 * There should be no gaps. If there is, something happened
-		 * which we can't handle (e.g. all planes are cursors).
-		 */
-		igt_assert_eq(p, last_plane);
+		/* Check that we filled every slot exactly once */
+		if (display->has_cursor_plane)
+			igt_assert_eq(p, last_plane);
+		else
+			igt_assert_eq(p, n_planes);
 
 		pipe->n_planes = n_planes;
 
-- 
2.13.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 2/4] lib/igt_kms: Fix override_mode handling

[Liviu Dudau]

From: Brian Starkey <brian.starkey@arm.com>

igt_display_commit isn't refreshing all outputs anymore, which means
that an override mode may never get picked up.

Instead of forcing a reprobe to handle copying the override_mode into
default_mode, just change igt_output_get_mode() to return the
override_mode if it's been set, and remove the old code which would
directly overwrite default_mode.

This should be more robust, as igt_output_get_mode() will always return
the correct mode, without needing the output to be reprobed.

This change means that output->config.default_mode always contains the
"non-overridden" default mode.

Signed-off-by: Brian Starkey <brian.starkey@arm.com>
---
 lib/igt_kms.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/lib/igt_kms.c b/lib/igt_kms.c
index 5e2ef97b..6d60a14f 100644
--- a/lib/igt_kms.c
+++ b/lib/igt_kms.c
@@ -2821,7 +2821,10 @@ const char *igt_output_name(igt_output_t *output)
 
 drmModeModeInfo *igt_output_get_mode(igt_output_t *output)
 {
-	return &output->config.default_mode;
+	if (output->use_override_mode)
+		return &output->override_mode;
+	else
+		return &output->config.default_mode;
 }
 
 /**
@@ -2839,10 +2842,6 @@ void igt_output_override_mode(igt_output_t *output, drmModeModeInfo *mode)
 
 	if (mode)
 		output->override_mode = *mode;
-	else /* restore default_mode, may have been overwritten in igt_output_refresh */
-		kmstest_get_connector_default_mode(output->display->drm_fd,
-						   output->config.connector,
-						   &output->config.default_mode);
 
 	output->use_override_mode = !!mode;
 
-- 
2.13.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 3/4] lib/igt_debugfs: Remove igt_debugfs_t

[Liviu Dudau]

From: Brian Starkey <brian.starkey@arm.com>

It's not used anymore, so remove it.

Signed-off-by: Brian Starkey <brian.starkey@arm.com>
---
 lib/igt_debugfs.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/lib/igt_debugfs.c b/lib/igt_debugfs.c
index 7584be5f..f5ed3daf 100644
--- a/lib/igt_debugfs.c
+++ b/lib/igt_debugfs.c
@@ -84,11 +84,6 @@
  * General debugfs helpers
  */
 
-typedef struct {
-	char root[128];
-	char dri_path[128];
-} igt_debugfs_t;
-
 static bool is_mountpoint(const char *path)
 {
 	char buf[strlen(path) + 4];
-- 
2.13.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 4/4] lib/igt_debugfs: Only use valid values in igt_crc_to_str()

[Liviu Dudau]

From: Brian Starkey <brian.starkey@arm.com>

Not all elements in the crc array may be valid, so only use the valid
ones to generate the string.

Signed-off-by: Brian Starkey <brian.starkey@arm.com>
---
 lib/igt_debugfs.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/lib/igt_debugfs.c b/lib/igt_debugfs.c
index f5ed3daf..80f25c61 100644
--- a/lib/igt_debugfs.c
+++ b/lib/igt_debugfs.c
@@ -312,10 +312,11 @@ void igt_assert_crc_equal(const igt_crc_t *a, const igt_crc_t *b)
  */
 char *igt_crc_to_string(igt_crc_t *crc)
 {
-	char buf[128];
+	int i;
+	char buf[128] = { 0 };
 
-	sprintf(buf, "%08x %08x %08x %08x %08x", crc->crc[0],
-		crc->crc[1], crc->crc[2], crc->crc[3], crc->crc[4]);
+	for (i = 0; i < crc->n_words; i++)
+		sprintf(buf + strlen(buf), "%08x ", crc->crc[i]);
 
 	return strdup(buf);
 }
-- 
2.13.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 2/4] lib/igt_kms: Fix override_mode handling

[Arkadiusz Hiler]

On Mon, Jun 05, 2017 at 02:28:38PM +0100, Liviu Dudau wrote:
> From: Brian Starkey <brian.starkey@arm.com>
> 
> igt_display_commit isn't refreshing all outputs anymore, which means
> that an override mode may never get picked up.
> 
> Instead of forcing a reprobe to handle copying the override_mode into
> default_mode, just change igt_output_get_mode() to return the
> override_mode if it's been set, and remove the old code which would
> directly overwrite default_mode.
> 
> This should be more robust, as igt_output_get_mode() will always return
> the correct mode, without needing the output to be reprobed.
> 
> This change means that output->config.default_mode always contains the
> "non-overridden" default mode.
> 
> Signed-off-by: Brian Starkey <brian.starkey@arm.com>
> ---
>  lib/igt_kms.c | 9 ++++-----
>  1 file changed, 4 insertions(+), 5 deletions(-)
> 
> diff --git a/lib/igt_kms.c b/lib/igt_kms.c
> index 5e2ef97b..6d60a14f 100644
> --- a/lib/igt_kms.c
> +++ b/lib/igt_kms.c
> @@ -2821,7 +2821,10 @@ const char *igt_output_name(igt_output_t *output)
>  
>  drmModeModeInfo *igt_output_get_mode(igt_output_t *output)
>  {
> -	return &output->config.default_mode;
> +	if (output->use_override_mode)
> +		return &output->override_mode;
> +	else
> +		return &output->config.default_mode;
>  }
>  
>  /**
> @@ -2839,10 +2842,6 @@ void igt_output_override_mode(igt_output_t *output, drmModeModeInfo *mode)
>  
>  	if (mode)
>  		output->override_mode = *mode;
> -	else /* restore default_mode, may have been overwritten in igt_output_refresh */

The warning caught my eye.

From what I understand, the idea for this change is that:
 * we should not ever go for output->cofnig.default_mode if the default
   mode is not what we really want.
 * we should use the igt_output_get_mode() to get the current mode (i.e.
   either the default or the overriden)

And yet igt_output_refresh() has this:

	if (output->use_override_mode)
		output->config.default_mode = output->override_mode;


Am I missing something or is this an oversight?

-- 
Cheers,
Arek

> -		kmstest_get_connector_default_mode(output->display->drm_fd,
> -						   output->config.connector,
> -						   &output->config.default_mode);
>  
>  	output->use_override_mode = !!mode;
>  
> -- 
> 2.13.0
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 1/4] lib/igt_kms: Fix erroneous assert

[Arkadiusz Hiler]

On Mon, Jun 05, 2017 at 02:28:37PM +0100, Liviu Dudau wrote:
> From: Brian Starkey <brian.starkey@arm.com>
> 
> In trying to fix igt_display_init() for devices without cursors, I
> actually made matters worse. Fix the assert.
> 
> Fixes: 545aa3398223 lib/igt_kms: Remove redundant cursor code
> Signed-off-by: Brian Starkey <brian.starkey@arm.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 3/4] lib/igt_debugfs: Remove igt_debugfs_t

[Arkadiusz Hiler]

On Mon, Jun 05, 2017 at 02:28:39PM +0100, Liviu Dudau wrote:
> From: Brian Starkey <brian.starkey@arm.com>
> 
> It's not used anymore, so remove it.
> 
> Signed-off-by: Brian Starkey <brian.starkey@arm.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 2/4] lib/igt_kms: Fix override_mode handling

[Liviu Dudau]

On Tue, Jun 06, 2017 at 01:30:28PM +0300, Arkadiusz Hiler wrote:
> On Mon, Jun 05, 2017 at 02:28:38PM +0100, Liviu Dudau wrote:
> > From: Brian Starkey <brian.starkey@arm.com>
> > 
> > igt_display_commit isn't refreshing all outputs anymore, which means
> > that an override mode may never get picked up.
> > 
> > Instead of forcing a reprobe to handle copying the override_mode into
> > default_mode, just change igt_output_get_mode() to return the
> > override_mode if it's been set, and remove the old code which would
> > directly overwrite default_mode.
> > 
> > This should be more robust, as igt_output_get_mode() will always return
> > the correct mode, without needing the output to be reprobed.
> > 
> > This change means that output->config.default_mode always contains the
> > "non-overridden" default mode.
> > 
> > Signed-off-by: Brian Starkey <brian.starkey@arm.com>
> > ---
> >  lib/igt_kms.c | 9 ++++-----
> >  1 file changed, 4 insertions(+), 5 deletions(-)
> > 
> > diff --git a/lib/igt_kms.c b/lib/igt_kms.c
> > index 5e2ef97b..6d60a14f 100644
> > --- a/lib/igt_kms.c
> > +++ b/lib/igt_kms.c
> > @@ -2821,7 +2821,10 @@ const char *igt_output_name(igt_output_t *output)
> >  
> >  drmModeModeInfo *igt_output_get_mode(igt_output_t *output)
> >  {
> > -	return &output->config.default_mode;
> > +	if (output->use_override_mode)
> > +		return &output->override_mode;
> > +	else
> > +		return &output->config.default_mode;
> >  }
> >  
> >  /**
> > @@ -2839,10 +2842,6 @@ void igt_output_override_mode(igt_output_t *output, drmModeModeInfo *mode)
> >  
> >  	if (mode)
> >  		output->override_mode = *mode;
> > -	else /* restore default_mode, may have been overwritten in igt_output_refresh */
> 
> The warning caught my eye.
> 
> From what I understand, the idea for this change is that:
>  * we should not ever go for output->cofnig.default_mode if the default
>    mode is not what we really want.
>  * we should use the igt_output_get_mode() to get the current mode (i.e.
>    either the default or the overriden)
> 
> And yet igt_output_refresh() has this:
> 
> 	if (output->use_override_mode)
> 		output->config.default_mode = output->override_mode;
> 
> 
> Am I missing something or is this an oversight?

Hi Arek,

An oversight, will fix in the next version.

Thanks for reviewing this!
Liviu

> 
> -- 
> Cheers,
> Arek
> 
> > -		kmstest_get_connector_default_mode(output->display->drm_fd,
> > -						   output->config.connector,
> > -						   &output->config.default_mode);
> >  
> >  	output->use_override_mode = !!mode;
> >  
> > -- 
> > 2.13.0
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
====================
| I would like to |
| fix the world,  |
| but they're not |
| giving me the   |
 \ source code!  /
  ---------------
    ¯\_(ツ)_/¯
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 4/4] lib/igt_debugfs: Only use valid values in igt_crc_to_str()

[Arkadiusz Hiler]

On Mon, Jun 05, 2017 at 02:28:40PM +0100, Liviu Dudau wrote:
> From: Brian Starkey <brian.starkey@arm.com>
> 
> Not all elements in the crc array may be valid, so only use the valid
> ones to generate the string.
> 
> Signed-off-by: Brian Starkey <brian.starkey@arm.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>

> ---
>  lib/igt_debugfs.c | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/lib/igt_debugfs.c b/lib/igt_debugfs.c
> index f5ed3daf..80f25c61 100644
> --- a/lib/igt_debugfs.c
> +++ b/lib/igt_debugfs.c
> @@ -312,10 +312,11 @@ void igt_assert_crc_equal(const igt_crc_t *a, const igt_crc_t *b)
>   */
>  char *igt_crc_to_string(igt_crc_t *crc)
>  {
> -	char buf[128];
> +	int i;
> +	char buf[128] = { 0 };
>  
> -	sprintf(buf, "%08x %08x %08x %08x %08x", crc->crc[0],
> -		crc->crc[1], crc->crc[2], crc->crc[3], crc->crc[4]);
> +	for (i = 0; i < crc->n_words; i++)
> +		sprintf(buf + strlen(buf), "%08x ", crc->crc[i]);

Although it is worth noting that this will result in an excessive space
at the end and the format will change.

This should be fine for the few debug messages from the tests...
but may brake something that parses output of intel_display_crc tool.

-- 
Cheers,
Arek

>  
>  	return strdup(buf);
>  }
> -- 
> 2.13.0
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 0/4] Cleanups and fixes

[Arkadiusz Hiler]

On Mon, Jun 05, 2017 at 02:28:36PM +0100, Liviu Dudau wrote:
> From: Liviu Dudau <Liviu.Dudau@arm.com>
> 
> Hello,
> 
> I believe Brian has sent these patches before, but he did not had any
> luck getting them accepted, so it is now my turn :)
> 
> These fixes were found while developing code for memory writeback testing
> and general use of igt with Mali DP KMS driver, which does not have a
> cursor plane.
> 
> Best regards,
> Liviu

I've pushed patches 1, 2 and 4.

Please Cc me on a fixed version of patch 3 :-)

Thanks for the contribution!

-- 
Cheers,
Arek

> 
> Brian Starkey (4):
>   lib/igt_kms: Fix erroneous assert
>   lib/igt_kms: Fix override_mode handling
>   lib/igt_debugfs: Remove igt_debugfs_t
>   lib/igt_debugfs: Only use valid values in igt_crc_to_str()
> 
>  lib/igt_debugfs.c | 12 ++++--------
>  lib/igt_kms.c     | 19 +++++++++----------
>  2 files changed, 13 insertions(+), 18 deletions(-)
> 
> -- 
> 2.13.0
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 0/4] Cleanups and fixes

[Arkadiusz Hiler]

On Fri, Jun 09, 2017 at 03:34:42PM +0300, Arkadiusz Hiler wrote:
> On Mon, Jun 05, 2017 at 02:28:36PM +0100, Liviu Dudau wrote:
> > From: Liviu Dudau <Liviu.Dudau@arm.com>
> > 
> > Hello,
> > 
> > I believe Brian has sent these patches before, but he did not had any
> > luck getting them accepted, so it is now my turn :)
> > 
> > These fixes were found while developing code for memory writeback testing
> > and general use of igt with Mali DP KMS driver, which does not have a
> > cursor plane.
> > 
> > Best regards,
> > Liviu
> 
> I've pushed patches 1, 2 and 4.

Whoops. Mixed up patch numbers.

I've meant 1, 3 and 4.
> 
> Please Cc me on a fixed version of patch 3 :-)

patch 2

> Thanks for the contribution!
> 
> -- 
> Cheers,
> Arek
> 
> > 
> > Brian Starkey (4):
> >   lib/igt_kms: Fix erroneous assert
> >   lib/igt_kms: Fix override_mode handling
> >   lib/igt_debugfs: Remove igt_debugfs_t
> >   lib/igt_debugfs: Only use valid values in igt_crc_to_str()
> > 
> >  lib/igt_debugfs.c | 12 ++++--------
> >  lib/igt_kms.c     | 19 +++++++++----------
> >  2 files changed, 13 insertions(+), 18 deletions(-)
> > 
> > -- 
> > 2.13.0
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 0/4] Cleanups and fixes

[Liviu Dudau]

On Fri, Jun 09, 2017 at 03:37:54PM +0300, Arkadiusz Hiler wrote:
> On Fri, Jun 09, 2017 at 03:34:42PM +0300, Arkadiusz Hiler wrote:
> > On Mon, Jun 05, 2017 at 02:28:36PM +0100, Liviu Dudau wrote:
> > > From: Liviu Dudau <Liviu.Dudau@arm.com>
> > > 
> > > Hello,
> > > 
> > > I believe Brian has sent these patches before, but he did not had any
> > > luck getting them accepted, so it is now my turn :)
> > > 
> > > These fixes were found while developing code for memory writeback testing
> > > and general use of igt with Mali DP KMS driver, which does not have a
> > > cursor plane.
> > > 
> > > Best regards,
> > > Liviu
> > 
> > I've pushed patches 1, 2 and 4.
> 
> Whoops. Mixed up patch numbers.
> 
> I've meant 1, 3 and 4.
> > 
> > Please Cc me on a fixed version of patch 3 :-)
> 
> patch 2
> 
> > Thanks for the contribution!

Thanks, will send updated patch today.

Liviu

> > 
> > -- 
> > Cheers,
> > Arek
> > 
> > > 
> > > Brian Starkey (4):
> > >   lib/igt_kms: Fix erroneous assert
> > >   lib/igt_kms: Fix override_mode handling
> > >   lib/igt_debugfs: Remove igt_debugfs_t
> > >   lib/igt_debugfs: Only use valid values in igt_crc_to_str()
> > > 
> > >  lib/igt_debugfs.c | 12 ++++--------
> > >  lib/igt_kms.c     | 19 +++++++++----------
> > >  2 files changed, 13 insertions(+), 18 deletions(-)
> > > 
> > > -- 
> > > 2.13.0
> > > 
> > > _______________________________________________
> > > Intel-gfx mailing list
> > > Intel-gfx@lists.freedesktop.org
> > > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
====================
| I would like to |
| fix the world,  |
| but they're not |
| giving me the   |
 \ source code!  /
  ---------------
    ¯\_(ツ)_/¯
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 0/4] Cleanups and fixes

[Liviu Dudau]

On Fri, Jun 09, 2017 at 03:37:54PM +0300, Arkadiusz Hiler wrote:
> On Fri, Jun 09, 2017 at 03:34:42PM +0300, Arkadiusz Hiler wrote:
> > On Mon, Jun 05, 2017 at 02:28:36PM +0100, Liviu Dudau wrote:
> > > From: Liviu Dudau <Liviu.Dudau@arm.com>
> > > 
> > > Hello,
> > > 
> > > I believe Brian has sent these patches before, but he did not had any
> > > luck getting them accepted, so it is now my turn :)
> > > 
> > > These fixes were found while developing code for memory writeback testing
> > > and general use of igt with Mali DP KMS driver, which does not have a
> > > cursor plane.
> > > 
> > > Best regards,
> > > Liviu
> > 
> > I've pushed patches 1, 2 and 4.
> 
> Whoops. Mixed up patch numbers.
> 
> I've meant 1, 3 and 4.
> > 
> > Please Cc me on a fixed version of patch 3 :-)
> 
> patch 2

Done, with the stupid mistake of using the root account due to the way I've setup
the testing on my dev board.

Thanks,
Liviu

> 
> > Thanks for the contribution!
> > 
> > -- 
> > Cheers,
> > Arek
> > 
> > > 
> > > Brian Starkey (4):
> > >   lib/igt_kms: Fix erroneous assert
> > >   lib/igt_kms: Fix override_mode handling
> > >   lib/igt_debugfs: Remove igt_debugfs_t
> > >   lib/igt_debugfs: Only use valid values in igt_crc_to_str()
> > > 
> > >  lib/igt_debugfs.c | 12 ++++--------
> > >  lib/igt_kms.c     | 19 +++++++++----------
> > >  2 files changed, 13 insertions(+), 18 deletions(-)
> > > 
> > > -- 
> > > 2.13.0
> > > 
> > > _______________________________________________
> > > Intel-gfx mailing list
> > > Intel-gfx@lists.freedesktop.org
> > > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
====================
| I would like to |
| fix the world,  |
| but they're not |
| giving me the   |
 \ source code!  /
  ---------------
    ¯\_(ツ)_/¯
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 0/4] Cleanups and fixes

[Arkadiusz Hiler]

On Mon, Jun 12, 2017 at 11:33:13AM +0100, Liviu Dudau wrote:
> On Fri, Jun 09, 2017 at 03:37:54PM +0300, Arkadiusz Hiler wrote:
> > On Fri, Jun 09, 2017 at 03:34:42PM +0300, Arkadiusz Hiler wrote:
> > > On Mon, Jun 05, 2017 at 02:28:36PM +0100, Liviu Dudau wrote:
> > > > From: Liviu Dudau <Liviu.Dudau@arm.com>
> > > > 
> > > > Hello,
> > > > 
> > > > I believe Brian has sent these patches before, but he did not had any
> > > > luck getting them accepted, so it is now my turn :)
> > > > 
> > > > These fixes were found while developing code for memory writeback testing
> > > > and general use of igt with Mali DP KMS driver, which does not have a
> > > > cursor plane.
> > > > 
> > > > Best regards,
> > > > Liviu
> > > 
> > > I've pushed patches 1, 2 and 4.
> > 
> > Whoops. Mixed up patch numbers.
> > 
> > I've meant 1, 3 and 4.
> > > 
> > > Please Cc me on a fixed version of patch 3 :-)
> > 
> > patch 2
> 
> Done, with the stupid mistake of using the root account due to the way I've setup
> the testing on my dev board.
> 
> Thanks,
> Liviu


No sight of them on the ML. I think those might have stuck in the
moderation queue because of the root account thing.

Jani?

-- 
Cheers,
Arek

> 
> > 
> > > Thanks for the contribution!
> > > 
> > > -- 
> > > Cheers,
> > > Arek
> > > 
> > > > 
> > > > Brian Starkey (4):
> > > >   lib/igt_kms: Fix erroneous assert
> > > >   lib/igt_kms: Fix override_mode handling
> > > >   lib/igt_debugfs: Remove igt_debugfs_t
> > > >   lib/igt_debugfs: Only use valid values in igt_crc_to_str()
> > > > 
> > > >  lib/igt_debugfs.c | 12 ++++--------
> > > >  lib/igt_kms.c     | 19 +++++++++----------
> > > >  2 files changed, 13 insertions(+), 18 deletions(-)
> > > > 
> > > > -- 
> > > > 2.13.0
> > > > 
> > > > _______________________________________________
> > > > Intel-gfx mailing list
> > > > Intel-gfx@lists.freedesktop.org
> > > > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> > > _______________________________________________
> > > Intel-gfx mailing list
> > > Intel-gfx@lists.freedesktop.org
> > > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 
> -- 
> ====================
> | I would like to |
> | fix the world,  |
> | but they're not |
> | giving me the   |
>  \ source code!  /
>   ---------------
>     ¯\_(ツ)_/¯
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 0/4] Cleanups and fixes

[Liviu Dudau]

On Fri, Jun 16, 2017 at 02:08:25PM +0300, Arkadiusz Hiler wrote:
> On Mon, Jun 12, 2017 at 11:33:13AM +0100, Liviu Dudau wrote:
> > On Fri, Jun 09, 2017 at 03:37:54PM +0300, Arkadiusz Hiler wrote:
> > > On Fri, Jun 09, 2017 at 03:34:42PM +0300, Arkadiusz Hiler wrote:
> > > > On Mon, Jun 05, 2017 at 02:28:36PM +0100, Liviu Dudau wrote:
> > > > > From: Liviu Dudau <Liviu.Dudau@arm.com>
> > > > > 
> > > > > Hello,
> > > > > 
> > > > > I believe Brian has sent these patches before, but he did not had any
> > > > > luck getting them accepted, so it is now my turn :)
> > > > > 
> > > > > These fixes were found while developing code for memory writeback testing
> > > > > and general use of igt with Mali DP KMS driver, which does not have a
> > > > > cursor plane.
> > > > > 
> > > > > Best regards,
> > > > > Liviu
> > > > 
> > > > I've pushed patches 1, 2 and 4.
> > > 
> > > Whoops. Mixed up patch numbers.
> > > 
> > > I've meant 1, 3 and 4.
> > > > 
> > > > Please Cc me on a fixed version of patch 3 :-)
> > > 
> > > patch 2
> > 
> > Done, with the stupid mistake of using the root account due to the way I've setup
> > the testing on my dev board.
> > 
> > Thanks,
> > Liviu
> 
> 
> No sight of them on the ML. I think those might have stuck in the
> moderation queue because of the root account thing.

Oops, I'll resend the patch. You were in the Cc line, I would've thought that
bypassed the ML.

Thanks for letting me know!
Liviu

> 
> Jani?
> 
> -- 
> Cheers,
> Arek
> 
> > 
> > > 
> > > > Thanks for the contribution!
> > > > 
> > > > -- 
> > > > Cheers,
> > > > Arek
> > > > 
> > > > > 
> > > > > Brian Starkey (4):
> > > > >   lib/igt_kms: Fix erroneous assert
> > > > >   lib/igt_kms: Fix override_mode handling
> > > > >   lib/igt_debugfs: Remove igt_debugfs_t
> > > > >   lib/igt_debugfs: Only use valid values in igt_crc_to_str()
> > > > > 
> > > > >  lib/igt_debugfs.c | 12 ++++--------
> > > > >  lib/igt_kms.c     | 19 +++++++++----------
> > > > >  2 files changed, 13 insertions(+), 18 deletions(-)
> > > > > 
> > > > > -- 
> > > > > 2.13.0
> > > > > 
> > > > > _______________________________________________
> > > > > Intel-gfx mailing list
> > > > > Intel-gfx@lists.freedesktop.org
> > > > > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> > > > _______________________________________________
> > > > Intel-gfx mailing list
> > > > Intel-gfx@lists.freedesktop.org
> > > > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> > 
> > -- 
> > ====================
> > | I would like to |
> > | fix the world,  |
> > | but they're not |
> > | giving me the   |
> >  \ source code!  /
> >   ---------------
> >     ¯\_(ツ)_/¯

-- 
====================
| I would like to |
| fix the world,  |
| but they're not |
| giving me the   |
 \ source code!  /
  ---------------
    ¯\_(ツ)_/¯
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH v2] lib/igt_kms: Fix override_mode handling

[Liviu Dudau]

From: Brian Starkey <brian.starkey@arm.com>

Changelog:
- v2: removed the forced overwrite of output->config.default_mode
- v1: original submission by Brian

igt_display_commit isn't refreshing all outputs anymore, which means
that an override mode may never get picked up.

Instead of forcing a reprobe to handle copying the override_mode into
default_mode, just change igt_output_get_mode() to return the
override_mode if it's been set, and remove the old code which would
directly overwrite default_mode.

This should be more robust, as igt_output_get_mode() will always return
the correct mode, without needing the output to be reprobed.

This change means that output->config.default_mode always contains the
"non-overridden" default mode.

Signed-off-by: Brian Starkey <brian.starkey@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
---
 lib/igt_kms.c | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)

diff --git a/lib/igt_kms.c b/lib/igt_kms.c
index 5e2ef97b..2abb6b4e 100644
--- a/lib/igt_kms.c
+++ b/lib/igt_kms.c
@@ -1580,9 +1580,6 @@ static void igt_output_refresh(igt_output_t *output)
 						    BROADCAST_RGB_FULL);
 	}
 
-	if (output->use_override_mode)
-		output->config.default_mode = output->override_mode;
-
 	if (output->config.pipe == PIPE_NONE)
 		return;
 
@@ -2821,7 +2818,10 @@ const char *igt_output_name(igt_output_t *output)
 
 drmModeModeInfo *igt_output_get_mode(igt_output_t *output)
 {
-	return &output->config.default_mode;
+	if (output->use_override_mode)
+		return &output->override_mode;
+	else
+		return &output->config.default_mode;
 }
 
 /**
@@ -2839,10 +2839,6 @@ void igt_output_override_mode(igt_output_t *output, drmModeModeInfo *mode)
 
 	if (mode)
 		output->override_mode = *mode;
-	else /* restore default_mode, may have been overwritten in igt_output_refresh */
-		kmstest_get_connector_default_mode(output->display->drm_fd,
-						   output->config.connector,
-						   &output->config.default_mode);
 
 	output->use_override_mode = !!mode;
 
-- 
2.13.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH v2] lib/igt_kms: Fix override_mode handling

[Arkadiusz Hiler]

On Fri, Jun 16, 2017 at 03:00:06PM +0100, Liviu Dudau wrote:
> From: Brian Starkey <brian.starkey@arm.com>
> 
> Changelog:
> - v2: removed the forced overwrite of output->config.default_mode
> - v1: original submission by Brian
> 
> igt_display_commit isn't refreshing all outputs anymore, which means
> that an override mode may never get picked up.
> 
> Instead of forcing a reprobe to handle copying the override_mode into
> default_mode, just change igt_output_get_mode() to return the
> override_mode if it's been set, and remove the old code which would
> directly overwrite default_mode.
> 
> This should be more robust, as igt_output_get_mode() will always return
> the correct mode, without needing the output to be reprobed.
> 
> This change means that output->config.default_mode always contains the
> "non-overridden" default mode.
> 
> Signed-off-by: Brian Starkey <brian.starkey@arm.com>
> Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>

... and pushed, thanks!


Also please remember of setting subject prefix to "PATCH i-g-t" as it
makes the changes easily trackable via filters and patchwork :-)

Excerpt from CONTRIBUTING:
-----------------------------------------------------------------------------
Please use --subject-prefix="PATCH i-g-t" so that i-g-t patches are easily
identified in the massive amount mails on intel-gfx. To ensure this is always
done, autogen.sh will run:

  git config format.subjectprefix "PATCH i-g-t"

on its first invocation.
-----------------------------------------------------------------------------

The autogen.sh thing is new addition.

-- 
Cheers,
Arek
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH v2] lib/igt_kms: Fix override_mode handling

[Liviu Dudau]

On Tue, Jun 20, 2017 at 04:28:14PM +0300, Arkadiusz Hiler wrote:
> On Fri, Jun 16, 2017 at 03:00:06PM +0100, Liviu Dudau wrote:
> > From: Brian Starkey <brian.starkey@arm.com>
> > 
> > Changelog:
> > - v2: removed the forced overwrite of output->config.default_mode
> > - v1: original submission by Brian
> > 
> > igt_display_commit isn't refreshing all outputs anymore, which means
> > that an override mode may never get picked up.
> > 
> > Instead of forcing a reprobe to handle copying the override_mode into
> > default_mode, just change igt_output_get_mode() to return the
> > override_mode if it's been set, and remove the old code which would
> > directly overwrite default_mode.
> > 
> > This should be more robust, as igt_output_get_mode() will always return
> > the correct mode, without needing the output to be reprobed.
> > 
> > This change means that output->config.default_mode always contains the
> > "non-overridden" default mode.
> > 
> > Signed-off-by: Brian Starkey <brian.starkey@arm.com>
> > Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
> Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
> 
> ... and pushed, thanks!

Many thanks! Now for the more important patches, the writeback support :)
Will send them this week (I'm having quite a busy time all of a sudden)

> 
> 
> Also please remember of setting subject prefix to "PATCH i-g-t" as it
> makes the changes easily trackable via filters and patchwork :-)

Will do and appologies for not reading the document before.

Best regards,
Liviu


> 
> Excerpt from CONTRIBUTING:
> -----------------------------------------------------------------------------
> Please use --subject-prefix="PATCH i-g-t" so that i-g-t patches are easily
> identified in the massive amount mails on intel-gfx. To ensure this is always
> done, autogen.sh will run:
> 
>   git config format.subjectprefix "PATCH i-g-t"
> 
> on its first invocation.
> -----------------------------------------------------------------------------
> 
> The autogen.sh thing is new addition.
> 
> -- 
> Cheers,
> Arek

-- 
====================
| I would like to |
| fix the world,  |
| but they're not |
| giving me the   |
 \ source code!  /
  ---------------
    ¯\_(ツ)_/¯
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx