diff mbox series

[1/2] drm/nouveau: Disable atomic support on a per-device basis

Message ID 20180913163147.27900-1-ville.syrjala@linux.intel.com (mailing list archive)
State New, archived
Headers show
Series [1/2] drm/nouveau: Disable atomic support on a per-device basis | expand

Commit Message

Ville Syrjälä Sept. 13, 2018, 4:31 p.m. UTC 
From: Ville Syrjälä <ville.syrjala@linux.intel.com>

We now have per-device driver_features, so let's use that
to disable atomic only for pre-nv50.

Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: Lyude Paul <lyude@redhat.com>
Cc: nouveau@lists.freedesktop.org
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Suggested-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/nouveau/dispnv04/disp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Lyude Paul Sept. 13, 2018, 9:02 p.m. UTC | #1 
Hm, one nitpick here. Since /sys/kernel/debug/dri/*/state creation depends on
the driver supporting atomic, maybe it would be good to make it so that we set
DRIVER_ATOMIC in the driver_stub structure, then disable it per-device depending
on the nouveau_atomic setting + hw support. That way we can always have the
state debugfs file while maintaining nouveau's current behavior with exposing
atomic ioctls. Assuming that wouldn't have any unintended side-effects of course

On Thu, 2018-09-13 at 19:31 +0300, Ville Syrjala wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> We now have per-device driver_features, so let's use that
> to disable atomic only for pre-nv50.
> 
> Cc: Ben Skeggs <bskeggs@redhat.com>
> Cc: Lyude Paul <lyude@redhat.com>
> Cc: nouveau@lists.freedesktop.org
> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> Suggested-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> ---
>  drivers/gpu/drm/nouveau/dispnv04/disp.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/nouveau/dispnv04/disp.c
> b/drivers/gpu/drm/nouveau/dispnv04/disp.c
> index 70dce544984e..670535a68d3b 100644
> --- a/drivers/gpu/drm/nouveau/dispnv04/disp.c
> +++ b/drivers/gpu/drm/nouveau/dispnv04/disp.c
> @@ -56,7 +56,7 @@ nv04_display_create(struct drm_device *dev)
>  	nouveau_display(dev)->fini = nv04_display_fini;
>  
>  	/* Pre-nv50 doesn't support atomic, so don't expose the ioctls */
> -	dev->driver->driver_features &= ~DRIVER_ATOMIC;
> +	dev->driver_features &= ~DRIVER_ATOMIC;
>  
>  	nouveau_hw_save_vga_fonts(dev, 1);
>
kernel test robot Sept. 14, 2018, 7:45 a.m. UTC | #2 
Hi Ville,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on v4.19-rc3 next-20180913]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url:    https://github.com/0day-ci/linux/commits/Ville-Syrjala/drm-nouveau-Disable-atomic-support-on-a-per-device-basis/20180914-111059
config: i386-allmodconfig (attached as .config)
compiler: gcc-7 (Debian 7.3.0-1) 7.3.0
reproduce:
        # save the attached .config to linux build tree
        make ARCH=i386 

All errors (new ones prefixed by >>):

   drivers/gpu/drm/nouveau/dispnv04/disp.c: In function 'nv04_display_create':
>> drivers/gpu/drm/nouveau/dispnv04/disp.c:59:5: error: 'struct drm_device' has no member named 'driver_features'
     dev->driver_features &= ~DRIVER_ATOMIC;
        ^~

vim +59 drivers/gpu/drm/nouveau/dispnv04/disp.c

    33	
    34	int
    35	nv04_display_create(struct drm_device *dev)
    36	{
    37		struct nouveau_drm *drm = nouveau_drm(dev);
    38		struct nvkm_i2c *i2c = nvxx_i2c(&drm->client.device);
    39		struct dcb_table *dcb = &drm->vbios.dcb;
    40		struct drm_connector *connector, *ct;
    41		struct drm_encoder *encoder;
    42		struct nouveau_encoder *nv_encoder;
    43		struct nouveau_crtc *crtc;
    44		struct nv04_display *disp;
    45		int i, ret;
    46	
    47		disp = kzalloc(sizeof(*disp), GFP_KERNEL);
    48		if (!disp)
    49			return -ENOMEM;
    50	
    51		nvif_object_map(&drm->client.device.object, NULL, 0);
    52	
    53		nouveau_display(dev)->priv = disp;
    54		nouveau_display(dev)->dtor = nv04_display_destroy;
    55		nouveau_display(dev)->init = nv04_display_init;
    56		nouveau_display(dev)->fini = nv04_display_fini;
    57	
    58		/* Pre-nv50 doesn't support atomic, so don't expose the ioctls */
  > 59		dev->driver_features &= ~DRIVER_ATOMIC;
    60	
    61		nouveau_hw_save_vga_fonts(dev, 1);
    62	
    63		nv04_crtc_create(dev, 0);
    64		if (nv_two_heads(dev))
    65			nv04_crtc_create(dev, 1);
    66	
    67		for (i = 0; i < dcb->entries; i++) {
    68			struct dcb_output *dcbent = &dcb->entry[i];
    69	
    70			connector = nouveau_connector_create(dev, dcbent->connector);
    71			if (IS_ERR(connector))
    72				continue;
    73	
    74			switch (dcbent->type) {
    75			case DCB_OUTPUT_ANALOG:
    76				ret = nv04_dac_create(connector, dcbent);
    77				break;
    78			case DCB_OUTPUT_LVDS:
    79			case DCB_OUTPUT_TMDS:
    80				ret = nv04_dfp_create(connector, dcbent);
    81				break;
    82			case DCB_OUTPUT_TV:
    83				if (dcbent->location == DCB_LOC_ON_CHIP)
    84					ret = nv17_tv_create(connector, dcbent);
    85				else
    86					ret = nv04_tv_create(connector, dcbent);
    87				break;
    88			default:
    89				NV_WARN(drm, "DCB type %d not known\n", dcbent->type);
    90				continue;
    91			}
    92	
    93			if (ret)
    94				continue;
    95		}
    96	
    97		list_for_each_entry_safe(connector, ct,
    98					 &dev->mode_config.connector_list, head) {
    99			if (!connector->encoder_ids[0]) {
   100				NV_WARN(drm, "%s has no encoders, removing\n",
   101					connector->name);
   102				connector->funcs->destroy(connector);
   103			}
   104		}
   105	
   106		list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
   107			struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
   108			struct nvkm_i2c_bus *bus =
   109				nvkm_i2c_bus_find(i2c, nv_encoder->dcb->i2c_index);
   110			nv_encoder->i2c = bus ? &bus->i2c : NULL;
   111		}
   112	
   113		/* Save previous state */
   114		list_for_each_entry(crtc, &dev->mode_config.crtc_list, base.head)
   115			crtc->save(&crtc->base);
   116	
   117		list_for_each_entry(nv_encoder, &dev->mode_config.encoder_list, base.base.head)
   118			nv_encoder->enc_save(&nv_encoder->base.base);
   119	
   120		nouveau_overlay_init(dev);
   121	
   122		return 0;
   123	}
   124	

---
0-DAY kernel test infrastructure                Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all                   Intel Corporation
Daniel Vetter Sept. 14, 2018, 8:11 a.m. UTC | #3 
On Thu, Sep 13, 2018 at 11:02 PM, Lyude Paul <lyude@redhat.com> wrote:
> Hm, one nitpick here. Since /sys/kernel/debug/dri/*/state creation depends on
> the driver supporting atomic, maybe it would be good to make it so that we set
> DRIVER_ATOMIC in the driver_stub structure, then disable it per-device depending
> on the nouveau_atomic setting + hw support. That way we can always have the
> state debugfs file while maintaining nouveau's current behavior with exposing
> atomic ioctls. Assuming that wouldn't have any unintended side-effects of course

dri/*/state only works with atomic drivers. There's no explicit state
with legacy drivers at all, it's all just implicit in hw and some
random driver structures.

We should make sure though that the debugfs stuff looks at
drm_drv_uses_atomic_modsetting(), and not DRIVER_ATOMIC. Former is
about the internals (i915 is internally atomic everywhere), latter
about the uapi (some old platforms aren't properly validated for full
atomic features, hence why it's disabled).
-Daniel

> On Thu, 2018-09-13 at 19:31 +0300, Ville Syrjala wrote:
>> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
>>
>> We now have per-device driver_features, so let's use that
>> to disable atomic only for pre-nv50.
>>
>> Cc: Ben Skeggs <bskeggs@redhat.com>
>> Cc: Lyude Paul <lyude@redhat.com>
>> Cc: nouveau@lists.freedesktop.org
>> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
>> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
>> Suggested-by: Daniel Vetter <daniel.vetter@ffwll.ch>
>> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
>> ---
>>  drivers/gpu/drm/nouveau/dispnv04/disp.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/nouveau/dispnv04/disp.c
>> b/drivers/gpu/drm/nouveau/dispnv04/disp.c
>> index 70dce544984e..670535a68d3b 100644
>> --- a/drivers/gpu/drm/nouveau/dispnv04/disp.c
>> +++ b/drivers/gpu/drm/nouveau/dispnv04/disp.c
>> @@ -56,7 +56,7 @@ nv04_display_create(struct drm_device *dev)
>>       nouveau_display(dev)->fini = nv04_display_fini;
>>
>>       /* Pre-nv50 doesn't support atomic, so don't expose the ioctls */
>> -     dev->driver->driver_features &= ~DRIVER_ATOMIC;
>> +     dev->driver_features &= ~DRIVER_ATOMIC;
>>
>>       nouveau_hw_save_vga_fonts(dev, 1);
>>
>
Ville Syrjälä Sept. 14, 2018, 3:04 p.m. UTC | #4 
On Thu, Sep 13, 2018 at 05:02:05PM -0400, Lyude Paul wrote:
> Hm, one nitpick here. Since /sys/kernel/debug/dri/*/state creation depends on
> the driver supporting atomic, maybe it would be good to make it so that we set
> DRIVER_ATOMIC in the driver_stub structure, then disable it per-device depending
> on the nouveau_atomic setting + hw support. That way we can always have the
> state debugfs file while maintaining nouveau's current behavior with exposing
> atomic ioctls. Assuming that wouldn't have any unintended side-effects of course

I'm not sure why there are three driver structs in nouveau.
Maybe someone can just nuke two of them?

> 
> On Thu, 2018-09-13 at 19:31 +0300, Ville Syrjala wrote:
> > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > 
> > We now have per-device driver_features, so let's use that
> > to disable atomic only for pre-nv50.
> > 
> > Cc: Ben Skeggs <bskeggs@redhat.com>
> > Cc: Lyude Paul <lyude@redhat.com>
> > Cc: nouveau@lists.freedesktop.org
> > Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> > Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> > Suggested-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > ---
> >  drivers/gpu/drm/nouveau/dispnv04/disp.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/nouveau/dispnv04/disp.c
> > b/drivers/gpu/drm/nouveau/dispnv04/disp.c
> > index 70dce544984e..670535a68d3b 100644
> > --- a/drivers/gpu/drm/nouveau/dispnv04/disp.c
> > +++ b/drivers/gpu/drm/nouveau/dispnv04/disp.c
> > @@ -56,7 +56,7 @@ nv04_display_create(struct drm_device *dev)
> >  	nouveau_display(dev)->fini = nv04_display_fini;
> >  
> >  	/* Pre-nv50 doesn't support atomic, so don't expose the ioctls */
> > -	dev->driver->driver_features &= ~DRIVER_ATOMIC;
> > +	dev->driver_features &= ~DRIVER_ATOMIC;
> >  
> >  	nouveau_hw_save_vga_fonts(dev, 1);
> >
Lyude Paul Sept. 17, 2018, 5:34 p.m. UTC | #5 
On Fri, 2018-09-14 at 10:11 +0200, Daniel Vetter wrote:
> On Thu, Sep 13, 2018 at 11:02 PM, Lyude Paul <lyude@redhat.com> wrote:
> > Hm, one nitpick here. Since /sys/kernel/debug/dri/*/state creation depends
> > on
> > the driver supporting atomic, maybe it would be good to make it so that we
> > set
> > DRIVER_ATOMIC in the driver_stub structure, then disable it per-device
> > depending
> > on the nouveau_atomic setting + hw support. That way we can always have the
> > state debugfs file while maintaining nouveau's current behavior with
> > exposing
> > atomic ioctls. Assuming that wouldn't have any unintended side-effects of
> > course
> 
> dri/*/state only works with atomic drivers. There's no explicit state
> with legacy drivers at all, it's all just implicit in hw and some
> random driver structures.
> 
> We should make sure though that the debugfs stuff looks at
> drm_drv_uses_atomic_modsetting(), and not DRIVER_ATOMIC. Former is
> about the internals (i915 is internally atomic everywhere), latter
> about the uapi (some old platforms aren't properly validated for full
> atomic features, hence why it's disabled).
> -Daniel
Makes sense, it seems doing that results in exactly what I wanted with nouveau!
As for this patch:

Reviewed-by: Lyude Paul <lyude@redhat.com>

> 
> > On Thu, 2018-09-13 at 19:31 +0300, Ville Syrjala wrote:
> > > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > 
> > > We now have per-device driver_features, so let's use that
> > > to disable atomic only for pre-nv50.
> > > 
> > > Cc: Ben Skeggs <bskeggs@redhat.com>
> > > Cc: Lyude Paul <lyude@redhat.com>
> > > Cc: nouveau@lists.freedesktop.org
> > > Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> > > Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> > > Suggested-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > ---
> > >  drivers/gpu/drm/nouveau/dispnv04/disp.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/gpu/drm/nouveau/dispnv04/disp.c
> > > b/drivers/gpu/drm/nouveau/dispnv04/disp.c
> > > index 70dce544984e..670535a68d3b 100644
> > > --- a/drivers/gpu/drm/nouveau/dispnv04/disp.c
> > > +++ b/drivers/gpu/drm/nouveau/dispnv04/disp.c
> > > @@ -56,7 +56,7 @@ nv04_display_create(struct drm_device *dev)
> > >       nouveau_display(dev)->fini = nv04_display_fini;
> > > 
> > >       /* Pre-nv50 doesn't support atomic, so don't expose the ioctls */
> > > -     dev->driver->driver_features &= ~DRIVER_ATOMIC;
> > > +     dev->driver_features &= ~DRIVER_ATOMIC;
> > > 
> > >       nouveau_hw_save_vga_fonts(dev, 1);
> > > 
> 
> 
>
diff mbox series

Patch

diff --git a/drivers/gpu/drm/nouveau/dispnv04/disp.c b/drivers/gpu/drm/nouveau/dispnv04/disp.c
index 70dce544984e..670535a68d3b 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/disp.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/disp.c
@@ -56,7 +56,7 @@  nv04_display_create(struct drm_device *dev)
 	nouveau_display(dev)->fini = nv04_display_fini;
 
 	/* Pre-nv50 doesn't support atomic, so don't expose the ioctls */
-	dev->driver->driver_features &= ~DRIVER_ATOMIC;
+	dev->driver_features &= ~DRIVER_ATOMIC;
 
 	nouveau_hw_save_vga_fonts(dev, 1);