All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH RFC 1/2] drm: extract closefb logic in separate function
@ 2021-10-07 13:15 Simon Ser
  2021-10-07 13:15 ` [PATCH RFC 2/2] drm: introduce CLOSEFB IOCTL Simon Ser
  0 siblings, 1 reply; 4+ messages in thread
From: Simon Ser @ 2021-10-07 13:15 UTC (permalink / raw)
  To: dri-devel
  Cc: Hans de Goede, Dennis Filder, Daniel Vetter, Pekka Paalanen,
	Rob Clark, Sean Paul

drm_mode_rmfb performs two operations: drop the FB from the
file_priv->fbs list, and make sure the FB is no longer used on a
plane.

In the next commit an IOCTL which only does so former will be
introduced, so let's split it into a separate function.

No functional change, only refactoring.

Signed-off-by: Simon Ser <contact@emersion.fr>
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Dennis Filder <d.filder@web.de>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Pekka Paalanen <ppaalanen@gmail.com>
Cc: Rob Clark <robdclark@gmail.com>
Cc: Sean Paul <seanpaul@chromium.org>
---
 drivers/gpu/drm/drm_framebuffer.c | 51 +++++++++++++++++++------------
 1 file changed, 31 insertions(+), 20 deletions(-)

diff --git a/drivers/gpu/drm/drm_framebuffer.c b/drivers/gpu/drm/drm_framebuffer.c
index 07f5abc875e9..2352972ba6ac 100644
--- a/drivers/gpu/drm/drm_framebuffer.c
+++ b/drivers/gpu/drm/drm_framebuffer.c
@@ -412,6 +412,31 @@ static void drm_mode_rmfb_work_fn(struct work_struct *w)
 	}
 }
 
+static int drm_mode_closefb(struct drm_framebuffer *fb,
+			    struct drm_file *file_priv)
+{
+	struct drm_framebuffer *fbl = NULL;
+	bool found = false;
+
+	mutex_lock(&file_priv->fbs_lock);
+	list_for_each_entry(fbl, &file_priv->fbs, filp_head)
+		if (fb == fbl)
+			found = true;
+
+	if (!found) {
+		mutex_unlock(&file_priv->fbs_lock);
+		return -ENOENT;
+	}
+
+	list_del_init(&fb->filp_head);
+	mutex_unlock(&file_priv->fbs_lock);
+
+	/* Drop the reference that was stored in the fbs list */
+	drm_framebuffer_put(fb);
+
+	return 0;
+}
+
 /**
  * drm_mode_rmfb - remove an FB from the configuration
  * @dev: drm device
@@ -429,8 +454,7 @@ int drm_mode_rmfb(struct drm_device *dev, u32 fb_id,
 		  struct drm_file *file_priv)
 {
 	struct drm_framebuffer *fb = NULL;
-	struct drm_framebuffer *fbl = NULL;
-	int found = 0;
+	int ret;
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EOPNOTSUPP;
@@ -439,23 +463,14 @@ int drm_mode_rmfb(struct drm_device *dev, u32 fb_id,
 	if (!fb)
 		return -ENOENT;
 
-	mutex_lock(&file_priv->fbs_lock);
-	list_for_each_entry(fbl, &file_priv->fbs, filp_head)
-		if (fb == fbl)
-			found = 1;
-	if (!found) {
-		mutex_unlock(&file_priv->fbs_lock);
-		goto fail_unref;
+	ret = drm_mode_closefb(fb, file_priv);
+	if (ret != 0) {
+		drm_framebuffer_put(fb);
+		return ret;
 	}
 
-	list_del_init(&fb->filp_head);
-	mutex_unlock(&file_priv->fbs_lock);
-
-	/* drop the reference we picked up in framebuffer lookup */
-	drm_framebuffer_put(fb);
-
 	/*
-	 * we now own the reference that was stored in the fbs list
+	 * We now own the reference we picked up in drm_framebuffer_lookup.
 	 *
 	 * drm_framebuffer_remove may fail with -EINTR on pending signals,
 	 * so run this in a separate stack as there's no way to correctly
@@ -475,10 +490,6 @@ int drm_mode_rmfb(struct drm_device *dev, u32 fb_id,
 		drm_framebuffer_put(fb);
 
 	return 0;
-
-fail_unref:
-	drm_framebuffer_put(fb);
-	return -ENOENT;
 }
 
 int drm_mode_rmfb_ioctl(struct drm_device *dev,
-- 
2.33.0



^ permalink raw reply related	[flat|nested] 4+ messages in thread

* [PATCH RFC 2/2] drm: introduce CLOSEFB IOCTL
  2021-10-07 13:15 [PATCH RFC 1/2] drm: extract closefb logic in separate function Simon Ser
@ 2021-10-07 13:15 ` Simon Ser
  2021-10-08  7:29   ` Pekka Paalanen
  0 siblings, 1 reply; 4+ messages in thread
From: Simon Ser @ 2021-10-07 13:15 UTC (permalink / raw)
  To: dri-devel
  Cc: Hans de Goede, Dennis Filder, Daniel Vetter, Pekka Paalanen,
	Rob Clark, Sean Paul

This new IOCTL allows callers to close a framebuffer without
disabling planes or CRTCs. This takes inspiration from Rob Clark's
unref_fb IOCTL [1] and DRM_MODE_FB_PERSIST [2].

[1]: https://lore.kernel.org/dri-devel/20170509153654.23464-1-robdclark@gmail.com/
[2]: https://lore.kernel.org/dri-devel/20211006151921.312714-1-contact@emersion.fr/

Signed-off-by: Simon Ser <contact@emersion.fr>
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Dennis Filder <d.filder@web.de>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Pekka Paalanen <ppaalanen@gmail.com>
Cc: Rob Clark <robdclark@gmail.com>
Cc: Sean Paul <seanpaul@chromium.org>
---
 drivers/gpu/drm/drm_crtc_internal.h |  2 ++
 drivers/gpu/drm/drm_framebuffer.c   | 19 +++++++++++++++++++
 drivers/gpu/drm/drm_ioctl.c         |  1 +
 include/uapi/drm/drm.h              | 19 +++++++++++++++++++
 4 files changed, 41 insertions(+)

diff --git a/drivers/gpu/drm/drm_crtc_internal.h b/drivers/gpu/drm/drm_crtc_internal.h
index 63279e984342..c6377b9ba981 100644
--- a/drivers/gpu/drm/drm_crtc_internal.h
+++ b/drivers/gpu/drm/drm_crtc_internal.h
@@ -221,6 +221,8 @@ int drm_mode_addfb2_ioctl(struct drm_device *dev,
 			  void *data, struct drm_file *file_priv);
 int drm_mode_rmfb_ioctl(struct drm_device *dev,
 			void *data, struct drm_file *file_priv);
+int drm_mode_closefb_ioctl(struct drm_device *dev,
+			   void *data, struct drm_file *file_priv);
 int drm_mode_getfb(struct drm_device *dev,
 		   void *data, struct drm_file *file_priv);
 int drm_mode_getfb2_ioctl(struct drm_device *dev,
diff --git a/drivers/gpu/drm/drm_framebuffer.c b/drivers/gpu/drm/drm_framebuffer.c
index 2352972ba6ac..15f7c22dbd4a 100644
--- a/drivers/gpu/drm/drm_framebuffer.c
+++ b/drivers/gpu/drm/drm_framebuffer.c
@@ -500,6 +500,25 @@ int drm_mode_rmfb_ioctl(struct drm_device *dev,
 	return drm_mode_rmfb(dev, *fb_id, file_priv);
 }
 
+int drm_mode_closefb_ioctl(struct drm_device *dev,
+			   void *data, struct drm_file *file_priv)
+{
+	uint32_t *fb_id = data;
+	struct drm_framebuffer *fb;
+	int ret;
+
+	if (!drm_core_check_feature(dev, DRIVER_MODESET))
+		return -EOPNOTSUPP;
+
+	fb = drm_framebuffer_lookup(dev, file_priv, *fb_id);
+	if (!fb)
+		return -ENOENT;
+
+	ret = drm_mode_closefb(fb, file_priv);
+	drm_framebuffer_put(fb);
+	return ret;
+}
+
 /**
  * drm_mode_getfb - get FB info
  * @dev: drm device for the ioctl
diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c
index 8b8744dcf691..545762bc16d0 100644
--- a/drivers/gpu/drm/drm_ioctl.c
+++ b/drivers/gpu/drm/drm_ioctl.c
@@ -670,6 +670,7 @@ static const struct drm_ioctl_desc drm_ioctls[] = {
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB, drm_mode_addfb_ioctl, 0),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB2, drm_mode_addfb2_ioctl, 0),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb_ioctl, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CLOSEFB, drm_mode_closefb_ioctl, 0),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CREATE_DUMB, drm_mode_create_dumb_ioctl, 0),
diff --git a/include/uapi/drm/drm.h b/include/uapi/drm/drm.h
index 3b810b53ba8b..8726f003f382 100644
--- a/include/uapi/drm/drm.h
+++ b/include/uapi/drm/drm.h
@@ -1097,6 +1097,25 @@ extern "C" {
 #define DRM_IOCTL_SYNCOBJ_TIMELINE_SIGNAL	DRM_IOWR(0xCD, struct drm_syncobj_timeline_array)
 
 #define DRM_IOCTL_MODE_GETFB2		DRM_IOWR(0xCE, struct drm_mode_fb_cmd2)
+/**
+ * DRM_IOCTL_MODE_CLOSEFB - Close a framebuffer.
+ *
+ * This closes a framebuffer previously added via ADDFB/ADDFB2. The IOCTL
+ * argument is a framebuffer object ID.
+ *
+ * This IOCTL is similar to &DRM_IOCTL_MODE_RMFB, except it doesn't disable
+ * planes and CRTCs. As long as the framebuffer is used by a plane, it's kept
+ * alive. When the plane no longer uses the framebuffer (because the
+ * framebuffer is replaced with another one, or the plane is disabled), the
+ * framebuffer is cleaned up.
+ *
+ * This is useful to implement flicker-free transitions between two processes.
+ *
+ * Depending on the threat model, user-space may want to ensure that the
+ * framebuffer doesn't expose any sensitive user information: closed
+ * framebuffers attached to a plane can be read back by the next DRM master.
+ */
+#define DRM_IOCTL_MODE_CLOSEFB		DRM_IOWR(0xCF, unsigned int)
 
 /*
  * Device specific ioctls should only be in their respective headers
-- 
2.33.0



^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH RFC 2/2] drm: introduce CLOSEFB IOCTL
  2021-10-07 13:15 ` [PATCH RFC 2/2] drm: introduce CLOSEFB IOCTL Simon Ser
@ 2021-10-08  7:29   ` Pekka Paalanen
  2021-10-18  8:52     ` Simon Ser
  0 siblings, 1 reply; 4+ messages in thread
From: Pekka Paalanen @ 2021-10-08  7:29 UTC (permalink / raw)
  To: Simon Ser
  Cc: dri-devel, Hans de Goede, Dennis Filder, Daniel Vetter,
	Rob Clark, Sean Paul

[-- Attachment #1: Type: text/plain, Size: 3534 bytes --]

On Thu, 07 Oct 2021 13:15:25 +0000
Simon Ser <contact@emersion.fr> wrote:

> This new IOCTL allows callers to close a framebuffer without
> disabling planes or CRTCs. This takes inspiration from Rob Clark's
> unref_fb IOCTL [1] and DRM_MODE_FB_PERSIST [2].
> 
> [1]: https://lore.kernel.org/dri-devel/20170509153654.23464-1-robdclark@gmail.com/
> [2]: https://lore.kernel.org/dri-devel/20211006151921.312714-1-contact@emersion.fr/
> 
> Signed-off-by: Simon Ser <contact@emersion.fr>
> Cc: Hans de Goede <hdegoede@redhat.com>
> Cc: Dennis Filder <d.filder@web.de>
> Cc: Daniel Vetter <daniel@ffwll.ch>
> Cc: Pekka Paalanen <ppaalanen@gmail.com>
> Cc: Rob Clark <robdclark@gmail.com>
> Cc: Sean Paul <seanpaul@chromium.org>
> ---
>  drivers/gpu/drm/drm_crtc_internal.h |  2 ++
>  drivers/gpu/drm/drm_framebuffer.c   | 19 +++++++++++++++++++
>  drivers/gpu/drm/drm_ioctl.c         |  1 +
>  include/uapi/drm/drm.h              | 19 +++++++++++++++++++
>  4 files changed, 41 insertions(+)

...

> diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c
> index 8b8744dcf691..545762bc16d0 100644
> --- a/drivers/gpu/drm/drm_ioctl.c
> +++ b/drivers/gpu/drm/drm_ioctl.c
> @@ -670,6 +670,7 @@ static const struct drm_ioctl_desc drm_ioctls[] = {
>  	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB, drm_mode_addfb_ioctl, 0),
>  	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB2, drm_mode_addfb2_ioctl, 0),
>  	DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb_ioctl, 0),
> +	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CLOSEFB, drm_mode_closefb_ioctl, 0),
>  	DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER),
>  	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER),
>  	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CREATE_DUMB, drm_mode_create_dumb_ioctl, 0),
> diff --git a/include/uapi/drm/drm.h b/include/uapi/drm/drm.h
> index 3b810b53ba8b..8726f003f382 100644
> --- a/include/uapi/drm/drm.h
> +++ b/include/uapi/drm/drm.h
> @@ -1097,6 +1097,25 @@ extern "C" {
>  #define DRM_IOCTL_SYNCOBJ_TIMELINE_SIGNAL	DRM_IOWR(0xCD, struct drm_syncobj_timeline_array)
>  
>  #define DRM_IOCTL_MODE_GETFB2		DRM_IOWR(0xCE, struct drm_mode_fb_cmd2)
> +/**
> + * DRM_IOCTL_MODE_CLOSEFB - Close a framebuffer.
> + *
> + * This closes a framebuffer previously added via ADDFB/ADDFB2. The IOCTL
> + * argument is a framebuffer object ID.
> + *
> + * This IOCTL is similar to &DRM_IOCTL_MODE_RMFB, except it doesn't disable
> + * planes and CRTCs. As long as the framebuffer is used by a plane, it's kept
> + * alive. When the plane no longer uses the framebuffer (because the
> + * framebuffer is replaced with another one, or the plane is disabled), the
> + * framebuffer is cleaned up.
> + *
> + * This is useful to implement flicker-free transitions between two processes.
> + *
> + * Depending on the threat model, user-space may want to ensure that the
> + * framebuffer doesn't expose any sensitive user information: closed
> + * framebuffers attached to a plane can be read back by the next DRM master.
> + */

LGTM!

Semantics
Acked-by: Pekka Paalanen <pekka.paalanen@collabora.com>

> +#define DRM_IOCTL_MODE_CLOSEFB		DRM_IOWR(0xCF, unsigned int)

Should it have a structure with 'flags' for future-proofing?

ISTR some rule of thumb that everything new must have 'flags' field
enforced to be zero just in case. I only now saw that RMFB cannot have
flags.


Thanks,
pq

>  
>  /*
>   * Device specific ioctls should only be in their respective headers


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH RFC 2/2] drm: introduce CLOSEFB IOCTL
  2021-10-08  7:29   ` Pekka Paalanen
@ 2021-10-18  8:52     ` Simon Ser
  0 siblings, 0 replies; 4+ messages in thread
From: Simon Ser @ 2021-10-18  8:52 UTC (permalink / raw)
  To: Pekka Paalanen
  Cc: dri-devel, Hans de Goede, Dennis Filder, Daniel Vetter,
	Rob Clark, Sean Paul

Thanks for having a look at this patch, Pekka!

On Friday, October 8th, 2021 at 09:29, Pekka Paalanen <ppaalanen@gmail.com> wrote:

> > +#define DRM_IOCTL_MODE_CLOSEFB		DRM_IOWR(0xCF, unsigned int)
>
> Should it have a structure with 'flags' for future-proofing?
>
> ISTR some rule of thumb that everything new must have 'flags' field
> enforced to be zero just in case. I only now saw that RMFB cannot have
> flags.

I wonder. I don't mind adding a struct with flags or padding, not sure what the
policy is for the DRM subsystem.

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2021-10-18  8:52 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-10-07 13:15 [PATCH RFC 1/2] drm: extract closefb logic in separate function Simon Ser
2021-10-07 13:15 ` [PATCH RFC 2/2] drm: introduce CLOSEFB IOCTL Simon Ser
2021-10-08  7:29   ` Pekka Paalanen
2021-10-18  8:52     ` Simon Ser

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.