[v4,4/4] drm: allow to use mmuless SoC

Some SoC without MMU have display driver where a drm/kms driver
could be implemented.

Before doing such kind of thing drm/kms must allow to use mmuless devices.
This patch propose to remove MMU configuration flag and add a cma helper
function to help implementing mmuless display driver

version 4:
- add documentation about drm_gem_cma_get_unmapped_area()
- stub it MMU case

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@linaro.org>
---
 Documentation/gpu/drm-mm.rst         | 11 ++++++
 drivers/gpu/drm/Kconfig              |  4 +--
 drivers/gpu/drm/drm_gem_cma_helper.c | 69 ++++++++++++++++++++++++++++++++++++
 include/drm/drm_gem_cma_helper.h     | 17 +++++++++
 4 files changed, 99 insertions(+), 2 deletions(-)

On Wed, Dec 07, 2016 at 11:06:51AM +0100, Benjamin Gaignard wrote:
> Some SoC without MMU have display driver where a drm/kms driver
> could be implemented.
> 
> Before doing such kind of thing drm/kms must allow to use mmuless devices.
> This patch propose to remove MMU configuration flag and add a cma helper
> function to help implementing mmuless display driver
> 
> version 4:
> - add documentation about drm_gem_cma_get_unmapped_area()
> - stub it MMU case
> 
> Signed-off-by: Benjamin Gaignard <benjamin.gaignard@linaro.org>

Shouldn't we also update DRM_FB_HELPER_DEFAULT_OPS? With that addressed
this all looks good to me.
-Daniel

> ---
>  Documentation/gpu/drm-mm.rst         | 11 ++++++
>  drivers/gpu/drm/Kconfig              |  4 +--
>  drivers/gpu/drm/drm_gem_cma_helper.c | 69 ++++++++++++++++++++++++++++++++++++
>  include/drm/drm_gem_cma_helper.h     | 17 +++++++++
>  4 files changed, 99 insertions(+), 2 deletions(-)
> 
> diff --git a/Documentation/gpu/drm-mm.rst b/Documentation/gpu/drm-mm.rst
> index bca8085..6e7efab 100644
> --- a/Documentation/gpu/drm-mm.rst
> +++ b/Documentation/gpu/drm-mm.rst
> @@ -303,6 +303,17 @@ created.
>  Drivers that want to map the GEM object upfront instead of handling page
>  faults can implement their own mmap file operation handler.
>  
> +For platforms without MMU GEM care provides a helper method
> +:c:func:`drm_gem_cma_get_unmapped_area`. The mmap() routines will call
> +this to get a proposed address for the mapping.
> +
> +To use :c:func:`drm_gem_cma_get_unmapped_area`, drivers must fill the
> +struct :c:type:`struct file_operations <file_operations>` get_unmapped_area
> +field with a pointer on :c:func:`drm_gem_cma_get_unmapped_area`.
> +
> +More detailed information about get_unmapped_area can be found in
> +Documentation/nommu-mmap.txt
> +
>  Memory Coherency
>  ----------------
>  
> diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> index 83ac815..0eae4ad 100644
> --- a/drivers/gpu/drm/Kconfig
> +++ b/drivers/gpu/drm/Kconfig
> @@ -6,7 +6,7 @@
>  #
>  menuconfig DRM
>  	tristate "Direct Rendering Manager (XFree86 4.1.0 and higher DRI support)"
> -	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && MMU && HAS_DMA
> +	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && HAS_DMA
>  	select HDMI
>  	select FB_CMDLINE
>  	select I2C
> @@ -98,7 +98,7 @@ config DRM_LOAD_EDID_FIRMWARE
>  
>  config DRM_TTM
>  	tristate
> -	depends on DRM
> +	depends on DRM && MMU
>  	help
>  	  GPU memory management subsystem for devices with multiple
>  	  GPU memory types. Will be enabled automatically if a device driver
> diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c
> index 1d6c335..01016b1 100644
> --- a/drivers/gpu/drm/drm_gem_cma_helper.c
> +++ b/drivers/gpu/drm/drm_gem_cma_helper.c
> @@ -358,6 +358,75 @@ int drm_gem_cma_mmap(struct file *filp, struct vm_area_struct *vma)
>  }
>  EXPORT_SYMBOL_GPL(drm_gem_cma_mmap);
>  
> +#ifndef CONFIG_MMU
> +/**
> + * drm_gem_cma_get_unmapped_area - propose address for mapping in noMMU cases
> + * @filp: file object
> + * @addr: memory address
> + * @len: buffer size
> + * @pgoff: page offset
> + * @flags: memory flags
> + *
> + * This function is used in noMMU platforms to propose address mapping
> + * for a given buffer.
> + *
> + * Returns:
> + * mapping address on success or a negative error code on failure
> + */
> +unsigned long drm_gem_cma_get_unmapped_area(struct file *filp,
> +					    unsigned long addr,
> +					    unsigned long len,
> +					    unsigned long pgoff,
> +					    unsigned long flags)
> +{
> +	struct drm_gem_cma_object *cma_obj;
> +	struct drm_gem_object *obj = NULL;
> +	struct drm_file *priv = filp->private_data;
> +	struct drm_device *dev = priv->minor->dev;
> +	struct drm_vma_offset_node *node;
> +
> +	if (drm_device_is_unplugged(dev))
> +		return -ENODEV;
> +
> +	drm_vma_offset_lock_lookup(dev->vma_offset_manager);
> +	node = drm_vma_offset_exact_lookup_locked(dev->vma_offset_manager,
> +						  pgoff,
> +						  len >> PAGE_SHIFT);
> +	if (likely(node)) {
> +		obj = container_of(node, struct drm_gem_object, vma_node);
> +		/*
> +		 * When the object is being freed, after it hits 0-refcnt it
> +		 * proceeds to tear down the object. In the process it will
> +		 * attempt to remove the VMA offset and so acquire this
> +		 * mgr->vm_lock.  Therefore if we find an object with a 0-refcnt
> +		 * that matches our range, we know it is in the process of being
> +		 * destroyed and will be freed as soon as we release the lock -
> +		 * so we have to check for the 0-refcnted object and treat it as
> +		 * invalid.
> +		 */
> +		if (!kref_get_unless_zero(&obj->refcount))
> +			obj = NULL;
> +	}
> +
> +	drm_vma_offset_unlock_lookup(dev->vma_offset_manager);
> +
> +	if (!obj)
> +		return -EINVAL;
> +
> +	if (!drm_vma_node_is_allowed(node, priv)) {
> +		drm_gem_object_unreference_unlocked(obj);
> +		return -EACCES;
> +	}
> +
> +	cma_obj = to_drm_gem_cma_obj(obj);
> +
> +	drm_gem_object_unreference_unlocked(obj);
> +
> +	return cma_obj->vaddr ? (unsigned long)cma_obj->vaddr : -EINVAL;
> +}
> +EXPORT_SYMBOL_GPL(drm_gem_cma_get_unmapped_area);
> +#endif
> +
>  #ifdef CONFIG_DEBUG_FS
>  /**
>   * drm_gem_cma_describe - describe a CMA GEM object for debugfs
> diff --git a/include/drm/drm_gem_cma_helper.h b/include/drm/drm_gem_cma_helper.h
> index acd6af8..180b586 100644
> --- a/include/drm/drm_gem_cma_helper.h
> +++ b/include/drm/drm_gem_cma_helper.h
> @@ -53,6 +53,23 @@ struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm,
>  
>  extern const struct vm_operations_struct drm_gem_cma_vm_ops;
>  
> +#ifndef CONFIG_MMU
> +unsigned long drm_gem_cma_get_unmapped_area(struct file *filp,
> +					    unsigned long addr,
> +					    unsigned long len,
> +					    unsigned long pgoff,
> +					    unsigned long flags);
> +#else
> +unsigned long drm_gem_cma_get_unmapped_area(struct file *filp,
> +					    unsigned long addr,
> +					    unsigned long len,
> +					    unsigned long pgoff,
> +					    unsigned long flags)
> +{
> +	return -EINVAL;
> +}
> +#endif
> +
>  #ifdef CONFIG_DEBUG_FS
>  void drm_gem_cma_describe(struct drm_gem_cma_object *obj, struct seq_file *m);
>  #endif
> -- 
> 1.9.1
>

2016-12-07 11:27 GMT+01:00 Daniel Vetter <daniel@ffwll.ch>:
> On Wed, Dec 07, 2016 at 11:06:51AM +0100, Benjamin Gaignard wrote:
>> Some SoC without MMU have display driver where a drm/kms driver
>> could be implemented.
>>
>> Before doing such kind of thing drm/kms must allow to use mmuless devices.
>> This patch propose to remove MMU configuration flag and add a cma helper
>> function to help implementing mmuless display driver
>>
>> version 4:
>> - add documentation about drm_gem_cma_get_unmapped_area()
>> - stub it MMU case
>>
>> Signed-off-by: Benjamin Gaignard <benjamin.gaignard@linaro.org>
>
> Shouldn't we also update DRM_FB_HELPER_DEFAULT_OPS? With that addressed
> this all looks good to me.

Those changes shouldn't impact fb_ops but only fbdev file_operations
like it is done in fbmem patch

> -Daniel
>
[snip]

Hi Benjamin,

Thank you for the patch.

On Wednesday 07 Dec 2016 11:06:51 Benjamin Gaignard wrote:
> Some SoC without MMU have display driver where a drm/kms driver
> could be implemented.
> 
> Before doing such kind of thing drm/kms must allow to use mmuless devices.
> This patch propose to remove MMU configuration flag and add a cma helper
> function to help implementing mmuless display driver
> 
> version 4:
> - add documentation about drm_gem_cma_get_unmapped_area()
> - stub it MMU case
> 
> Signed-off-by: Benjamin Gaignard <benjamin.gaignard@linaro.org>
> ---
>  Documentation/gpu/drm-mm.rst         | 11 ++++++
>  drivers/gpu/drm/Kconfig              |  4 +--
>  drivers/gpu/drm/drm_gem_cma_helper.c | 69 +++++++++++++++++++++++++++++++++
>  include/drm/drm_gem_cma_helper.h     | 17 +++++++++
>  4 files changed, 99 insertions(+), 2 deletions(-)
> 
> diff --git a/Documentation/gpu/drm-mm.rst b/Documentation/gpu/drm-mm.rst
> index bca8085..6e7efab 100644
> --- a/Documentation/gpu/drm-mm.rst
> +++ b/Documentation/gpu/drm-mm.rst
> @@ -303,6 +303,17 @@ created.
>  Drivers that want to map the GEM object upfront instead of handling page
>  faults can implement their own mmap file operation handler.
> 
> +For platforms without MMU GEM care provides a helper method

s/GEM care/the GEM core/

> +:c:func:`drm_gem_cma_get_unmapped_area`. The mmap() routines will call
> +this to get a proposed address for the mapping.
> +
> +To use :c:func:`drm_gem_cma_get_unmapped_area`, drivers must fill the
> +struct :c:type:`struct file_operations <file_operations>` get_unmapped_area
> +field with a pointer on :c:func:`drm_gem_cma_get_unmapped_area`.
> +
> +More detailed information about get_unmapped_area can be found in
> +Documentation/nommu-mmap.txt
> +
>  Memory Coherency
>  ----------------
> 
> diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> index 83ac815..0eae4ad 100644
> --- a/drivers/gpu/drm/Kconfig
> +++ b/drivers/gpu/drm/Kconfig
> @@ -6,7 +6,7 @@
>  #
>  menuconfig DRM
>  	tristate "Direct Rendering Manager (XFree86 4.1.0 and higher DRI 
support)"
> -	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && MMU && HAS_DMA
> +	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && HAS_DMA
>  	select HDMI
>  	select FB_CMDLINE
>  	select I2C
> @@ -98,7 +98,7 @@ config DRM_LOAD_EDID_FIRMWARE
> 
>  config DRM_TTM
>  	tristate
> -	depends on DRM
> +	depends on DRM && MMU
>  	help
>  	  GPU memory management subsystem for devices with multiple
>  	  GPU memory types. Will be enabled automatically if a device driver
> diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c
> b/drivers/gpu/drm/drm_gem_cma_helper.c index 1d6c335..01016b1 100644
> --- a/drivers/gpu/drm/drm_gem_cma_helper.c
> +++ b/drivers/gpu/drm/drm_gem_cma_helper.c
> @@ -358,6 +358,75 @@ int drm_gem_cma_mmap(struct file *filp, struct
> vm_area_struct *vma) }
>  EXPORT_SYMBOL_GPL(drm_gem_cma_mmap);
> 
> +#ifndef CONFIG_MMU
> +/**
> + * drm_gem_cma_get_unmapped_area - propose address for mapping in noMMU
> cases
> + * @filp: file object
> + * @addr: memory address
> + * @len: buffer size
> + * @pgoff: page offset
> + * @flags: memory flags
> + *
> + * This function is used in noMMU platforms to propose address mapping
> + * for a given buffer.

I would add

"It's intended to be used as a direct handler for the :c:type:`struct 
file_operations <file_operations>` .get_unmapped_area() operation."

Apart from that,

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

> + * Returns:
> + * mapping address on success or a negative error code on failure
> + */
> +unsigned long drm_gem_cma_get_unmapped_area(struct file *filp,
> +					    unsigned long addr,
> +					    unsigned long len,
> +					    unsigned long pgoff,
> +					    unsigned long flags)
> +{
> +	struct drm_gem_cma_object *cma_obj;
> +	struct drm_gem_object *obj = NULL;
> +	struct drm_file *priv = filp->private_data;
> +	struct drm_device *dev = priv->minor->dev;
> +	struct drm_vma_offset_node *node;
> +
> +	if (drm_device_is_unplugged(dev))
> +		return -ENODEV;
> +
> +	drm_vma_offset_lock_lookup(dev->vma_offset_manager);
> +	node = drm_vma_offset_exact_lookup_locked(dev->vma_offset_manager,
> +						  pgoff,
> +						  len >> PAGE_SHIFT);
> +	if (likely(node)) {
> +		obj = container_of(node, struct drm_gem_object, vma_node);
> +		/*
> +		 * When the object is being freed, after it hits 0-refcnt it
> +		 * proceeds to tear down the object. In the process it will
> +		 * attempt to remove the VMA offset and so acquire this
> +		 * mgr->vm_lock.  Therefore if we find an object with a 0-
refcnt
> +		 * that matches our range, we know it is in the process of 
being
> +		 * destroyed and will be freed as soon as we release the lock 
-
> +		 * so we have to check for the 0-refcnted object and treat it 
as
> +		 * invalid.
> +		 */
> +		if (!kref_get_unless_zero(&obj->refcount))
> +			obj = NULL;
> +	}
> +
> +	drm_vma_offset_unlock_lookup(dev->vma_offset_manager);
> +
> +	if (!obj)
> +		return -EINVAL;
> +
> +	if (!drm_vma_node_is_allowed(node, priv)) {
> +		drm_gem_object_unreference_unlocked(obj);
> +		return -EACCES;
> +	}
> +
> +	cma_obj = to_drm_gem_cma_obj(obj);
> +
> +	drm_gem_object_unreference_unlocked(obj);
> +
> +	return cma_obj->vaddr ? (unsigned long)cma_obj->vaddr : -EINVAL;
> +}
> +EXPORT_SYMBOL_GPL(drm_gem_cma_get_unmapped_area);
> +#endif
> +
>  #ifdef CONFIG_DEBUG_FS
>  /**
>   * drm_gem_cma_describe - describe a CMA GEM object for debugfs
> diff --git a/include/drm/drm_gem_cma_helper.h
> b/include/drm/drm_gem_cma_helper.h index acd6af8..180b586 100644
> --- a/include/drm/drm_gem_cma_helper.h
> +++ b/include/drm/drm_gem_cma_helper.h
> @@ -53,6 +53,23 @@ struct drm_gem_cma_object *drm_gem_cma_create(struct
> drm_device *drm,
> 
>  extern const struct vm_operations_struct drm_gem_cma_vm_ops;
> 
> +#ifndef CONFIG_MMU
> +unsigned long drm_gem_cma_get_unmapped_area(struct file *filp,
> +					    unsigned long addr,
> +					    unsigned long len,
> +					    unsigned long pgoff,
> +					    unsigned long flags);
> +#else
> +unsigned long drm_gem_cma_get_unmapped_area(struct file *filp,
> +					    unsigned long addr,
> +					    unsigned long len,
> +					    unsigned long pgoff,
> +					    unsigned long flags)
> +{
> +	return -EINVAL;
> +}
> +#endif
> +
>  #ifdef CONFIG_DEBUG_FS
>  void drm_gem_cma_describe(struct drm_gem_cma_object *obj, struct seq_file
> *m); #endif

On Wed, Dec 07, 2016 at 04:32:44PM +0200, Laurent Pinchart wrote:
> Hi Benjamin,
> 
> Thank you for the patch.
> 
> On Wednesday 07 Dec 2016 11:06:51 Benjamin Gaignard wrote:
> > Some SoC without MMU have display driver where a drm/kms driver
> > could be implemented.
> > 
> > Before doing such kind of thing drm/kms must allow to use mmuless devices.
> > This patch propose to remove MMU configuration flag and add a cma helper
> > function to help implementing mmuless display driver
> > 
> > version 4:
> > - add documentation about drm_gem_cma_get_unmapped_area()
> > - stub it MMU case
> > 
> > Signed-off-by: Benjamin Gaignard <benjamin.gaignard@linaro.org>
> > ---
> >  Documentation/gpu/drm-mm.rst         | 11 ++++++
> >  drivers/gpu/drm/Kconfig              |  4 +--
> >  drivers/gpu/drm/drm_gem_cma_helper.c | 69 +++++++++++++++++++++++++++++++++
> >  include/drm/drm_gem_cma_helper.h     | 17 +++++++++
> >  4 files changed, 99 insertions(+), 2 deletions(-)
> > 
> > diff --git a/Documentation/gpu/drm-mm.rst b/Documentation/gpu/drm-mm.rst
> > index bca8085..6e7efab 100644
> > --- a/Documentation/gpu/drm-mm.rst
> > +++ b/Documentation/gpu/drm-mm.rst
> > @@ -303,6 +303,17 @@ created.
> >  Drivers that want to map the GEM object upfront instead of handling page
> >  faults can implement their own mmap file operation handler.
> > 
> > +For platforms without MMU GEM care provides a helper method
> 
> s/GEM care/the GEM core/
> 
> > +:c:func:`drm_gem_cma_get_unmapped_area`. The mmap() routines will call
> > +this to get a proposed address for the mapping.
> > +
> > +To use :c:func:`drm_gem_cma_get_unmapped_area`, drivers must fill the
> > +struct :c:type:`struct file_operations <file_operations>` get_unmapped_area
> > +field with a pointer on :c:func:`drm_gem_cma_get_unmapped_area`.
> > +
> > +More detailed information about get_unmapped_area can be found in
> > +Documentation/nommu-mmap.txt
> > +
> >  Memory Coherency
> >  ----------------
> > 
> > diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> > index 83ac815..0eae4ad 100644
> > --- a/drivers/gpu/drm/Kconfig
> > +++ b/drivers/gpu/drm/Kconfig
> > @@ -6,7 +6,7 @@
> >  #
> >  menuconfig DRM
> >  	tristate "Direct Rendering Manager (XFree86 4.1.0 and higher DRI 
> support)"
> > -	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && MMU && HAS_DMA
> > +	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && HAS_DMA
> >  	select HDMI
> >  	select FB_CMDLINE
> >  	select I2C
> > @@ -98,7 +98,7 @@ config DRM_LOAD_EDID_FIRMWARE
> > 
> >  config DRM_TTM
> >  	tristate
> > -	depends on DRM
> > +	depends on DRM && MMU
> >  	help
> >  	  GPU memory management subsystem for devices with multiple
> >  	  GPU memory types. Will be enabled automatically if a device driver
> > diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c
> > b/drivers/gpu/drm/drm_gem_cma_helper.c index 1d6c335..01016b1 100644
> > --- a/drivers/gpu/drm/drm_gem_cma_helper.c
> > +++ b/drivers/gpu/drm/drm_gem_cma_helper.c
> > @@ -358,6 +358,75 @@ int drm_gem_cma_mmap(struct file *filp, struct
> > vm_area_struct *vma) }
> >  EXPORT_SYMBOL_GPL(drm_gem_cma_mmap);
> > 
> > +#ifndef CONFIG_MMU
> > +/**
> > + * drm_gem_cma_get_unmapped_area - propose address for mapping in noMMU
> > cases
> > + * @filp: file object
> > + * @addr: memory address
> > + * @len: buffer size
> > + * @pgoff: page offset
> > + * @flags: memory flags
> > + *
> > + * This function is used in noMMU platforms to propose address mapping
> > + * for a given buffer.
> 
> I would add
> 
> "It's intended to be used as a direct handler for the :c:type:`struct 
> file_operations <file_operations>` .get_unmapped_area() operation."

If' it's kernel-doc you can use shortcuts to autogenerate the links. Only
works with kernel-doc comments though, not with .rst in general:

"It's intended to be used as a direct handler for the struct 
&file_operations .get_unmapped_area() operation."

And +1 from me on sprinkling links all over docs. We might want to do that
for all the fbdev fops, for consistency.

Cheers, Daniel