All of lore.kernel.org
 help / color / mirror / Atom feed
From: Daniel Vetter <daniel.vetter@ffwll.ch>
To: DRI Development <dri-devel@lists.freedesktop.org>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Subject: [PATCH 16/35] drm: add per-crtc locks
Date: Thu, 10 Jan 2013 21:47:57 +0100	[thread overview]
Message-ID: <1357850897-27102-17-git-send-email-daniel.vetter@ffwll.ch> (raw)
In-Reply-To: <1357850897-27102-1-git-send-email-daniel.vetter@ffwll.ch>

*drumroll*

The basic idea is to protect per-crtc state which can change without
touching the output configuration with separate mutexes, i.e.  all the
input side state to a crtc like framebuffers, cursor settings or plane
configuration. Holding such a crtc lock gives a read-lock on all the
other crtc state which can be changed by e.g. a modeset.

All non-crtc state is still protected by the mode_config mutex.
Callers that need to change modeset state of a crtc (e.g. dpms or
set_mode) need to grab both the mode_config lock and nested within any
crtc locks.

Note that since there can only ever be one holder of the mode_config
lock we can grab the subordinate crtc locks in any order (if we need
to grab more than one of them). Lockdep can handle such nesting with
the mutex_lock_nest_lock call correctly.

With this functions that only touch connectors/encoders but not crtcs
only need to take the mode_config lock. The biggest such case is the
output probing, which means that we can now pageflip and move cursors
while the output probe code is reading an edid.

Most cases neatly fall into the three buckets:
- Only touches connectors and similar output state and so only needs
  the mode_config lock.
- Touches the global configuration and so needs all locks.
- Only touches the crtc input side and so only needs the crtc lock.

But a few cases that need special consideration:

- Load detection which requires a crtc. The mode_config lock already
  prevents a modeset change, so we can use any unused crtc as we like
  to do load detection. The only thing to consider is that such
  temporary state changes don't leak out to userspace through ioctls
  that only take the crtc look (like a pageflip). Hence the load
  detect code needs to grab the crtc of any output pipes it touches
  (but only if it touches state used by the pageflip or cursor
  ioctls).

- Atomic pageflip when moving planes. The first case is sane hw, where
  planes have a fixed association with crtcs - nothing needs to be
  done there. More insane^Wflexible hw needs to have plane->crtc
  mapping which is separately protect with a lock that nests within
  the crtc lock. If the plane is unused we can just assign it to the
  current crtc and continue. But if a plane is already in use by
  another crtc we can't just reassign it.

  Two solution present themselves: Either go back to a slow-path which
  takes all modeset locks, potentially incure quite a hefty delay. Or
  simply disallowing such changes in one atomic pageflip - in general
  the vblanks of two crtcs are not synced, so there's no sane way to
  atomically flip such plane changes accross more than one crtc. I'd
  heavily favour the later approach, going as far as mandating it as
  part of the ABI of such a new a nuclear pageflip.

  And if we _really_ want such semantics, we can always get them by
  introducing another pageflip mutex between the mode_config.mutex and
  the individual crtc locks. Pageflips crossing more than one crtc
  would then need to take that lock first, to lock out concurrent
  multi-crtc pageflips.

- Optimized global modeset operations: We could just take the
  mode_config lock and then lazily lock all crtc which are affected by
  a modeset operation. This has the advantage that pageflip could
  continue unhampered on unaffected crtc. But if e.g. global resources
  like plls need to be reassigned and so affect unrelated crtcs we can
  still do that - nested locking works in any order.

This patch just adds the locks and takes them in drm_modeset_lock_all,
no real locking changes yet.

v2: Need to initialize the new lock in crtc_init and lock it righ
away, for otherwise the modeset_unlock_all below will try to unlock a
not-locked mutex.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/drm_crtc.c |   12 ++++++++++++
 include/drm/drm_crtc.h     |    9 +++++++++
 2 files changed, 21 insertions(+)

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 7c8c5cd..0494ff4 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -46,7 +46,12 @@
  */
 void drm_modeset_lock_all(struct drm_device *dev)
 {
+	struct drm_crtc *crtc;
+
 	mutex_lock(&dev->mode_config.mutex);
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+		mutex_lock_nest_lock(&crtc->mutex, &dev->mode_config.mutex);
 }
 EXPORT_SYMBOL(drm_modeset_lock_all);
 
@@ -56,6 +61,11 @@ EXPORT_SYMBOL(drm_modeset_lock_all);
  */
 void drm_modeset_unlock_all(struct drm_device *dev)
 {
+	struct drm_crtc *crtc;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+		mutex_unlock(&crtc->mutex);
+
 	mutex_unlock(&dev->mode_config.mutex);
 }
 EXPORT_SYMBOL(drm_modeset_unlock_all);
@@ -449,6 +459,8 @@ int drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
 	crtc->invert_dimensions = false;
 
 	drm_modeset_lock_all(dev);
+	mutex_init(&crtc->mutex);
+	mutex_lock_nest_lock(&crtc->mutex, &dev->mode_config.mutex);
 
 	ret = drm_mode_object_get(dev, &crtc->base, DRM_MODE_OBJECT_CRTC);
 	if (ret)
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index 9f0524d..c89b116 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -390,6 +390,15 @@ struct drm_crtc {
 	struct drm_device *dev;
 	struct list_head head;
 
+	/**
+	 * crtc mutex
+	 *
+	 * This provides a read lock for the overall crtc state (mode, dpms
+	 * state, ...) and a write lock for everything which can be update
+	 * without a full modeset (fb, cursor data, ...)
+	 */
+	struct mutex mutex;
+
 	struct drm_mode_object base;
 
 	/* framebuffer the connector is currently bound to */
-- 
1.7.10.4

  parent reply	other threads:[~2013-01-10 20:48 UTC|newest]

Thread overview: 45+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-01-10 20:47 [PATCH 00/36] kms locking rework Daniel Vetter
2013-01-10 20:47 ` [PATCH 01/35] drm: review locking rules in drm_crtc.c Daniel Vetter
2013-01-10 20:47 ` [PATCH 02/35] drm/doc: integrate drm_crtc.c kerneldoc Daniel Vetter
2013-01-10 20:47 ` [PATCH 03/35] drm/<drivers>: reorder framebuffer init sequence Daniel Vetter
2013-01-11 21:06   ` Rob Clark
2013-01-10 20:47 ` [PATCH 04/35] drm/vmwgfx: " Daniel Vetter
2013-01-10 20:47 ` [PATCH 05/35] drm/gma500: move fbcon restore to lastclose Daniel Vetter
2013-01-10 20:47 ` [PATCH 06/35] drm/nouveau: protect evo_wait/evo_kick sections with a channel mutex Daniel Vetter
2013-01-10 20:47 ` [PATCH 07/35] drm/nouveau: try to protect nbo->pin_refcount Daniel Vetter
2013-01-10 20:47 ` [PATCH 08/35] drm/<drivers>: Unified handling of unimplemented fb->create_handle Daniel Vetter
2013-01-18 15:00   ` Thierry Reding
2013-01-18 18:32     ` Daniel Vetter
2013-01-10 20:47 ` [PATCH 09/35] drm: encapsulate crtc->set_config calls Daniel Vetter
2013-01-10 20:47 ` [PATCH 10/35] drm: add drm_modeset_lock|unlock_all Daniel Vetter
2013-01-10 20:47 ` [PATCH 11/35] drm/i915: use drm_modeset_lock_all Daniel Vetter
2013-01-10 20:47 ` [PATCH 12/35] drm/gma500: " Daniel Vetter
2013-01-10 22:36   ` Alan Cox
2013-01-11 13:25     ` Daniel Vetter
2013-01-10 20:47 ` [PATCH 13/35] drm/ast: " Daniel Vetter
2013-01-10 20:47 ` [PATCH 14/35] drm/shmobile: " Daniel Vetter
2013-01-10 20:47 ` [PATCH 15/35] drm/vmgfx: " Daniel Vetter
2013-01-10 20:47 ` Daniel Vetter [this message]
2013-01-10 20:47 ` [PATCH 17/35] drm: only take the crtc lock for ->cursor_set Daniel Vetter
2013-01-10 20:47 ` [PATCH 18/35] drm: only take the crtc lock for ->cursor_move Daniel Vetter
2013-01-10 20:48 ` [PATCH 19/35] drm: revamp locking around fb creation/destruction Daniel Vetter
2013-01-10 20:48 ` [PATCH 20/35] drm: create drm_framebuffer_lookup Daniel Vetter
2013-01-10 20:48 ` [PATCH 21/35] drm: revamp framebuffer cleanup interfaces Daniel Vetter
2013-01-10 20:48 ` [PATCH 22/35] drm: reference framebuffers which are on the idr Daniel Vetter
2013-01-10 20:48 ` [PATCH 23/35] drm: nest modeset locks within fpriv->fbs_lock Daniel Vetter
2013-01-10 20:48 ` [PATCH 24/35] drm: push modeset_lock_all into ->fb_create driver callbacks Daniel Vetter
2013-01-10 20:48 ` [PATCH 25/35] drm: don't take modeset locks in getfb ioctl Daniel Vetter
2013-01-10 20:48 ` [PATCH 26/35] drm: fb refcounting for dirtyfb_ioctl Daniel Vetter
2013-01-10 20:48 ` [PATCH 27/35] drm: refcounting for sprite framebuffers Daniel Vetter
2013-01-10 20:48 ` [PATCH 28/35] drm: refcounting for crtc framebuffers Daniel Vetter
2013-01-10 20:48 ` [PATCH 29/35] drm/i915: dump refcount into framebuffer debugfs file Daniel Vetter
2013-01-11 22:20   ` Rob Clark
2013-01-10 20:48 ` [PATCH 30/35] drm/vmwgfx: add proper framebuffer refcounting Daniel Vetter
2013-01-10 20:48 ` [PATCH 31/35] drm: optimize drm_framebuffer_remove Daniel Vetter
2013-01-10 20:48 ` [PATCH 32/35] drm: only grab the crtc lock for pageflips Daniel Vetter
2013-01-10 20:48 ` [PATCH 33/35] drm: don't hold crtc mutexes for connector ->detect callbacks Daniel Vetter
2013-01-10 20:48 ` [PATCH 34/35] drm/doc: updates for new framebuffer lifetime rules Daniel Vetter
2013-01-10 20:48 ` [PATCH 35/35] drm/fb_helper: check whether fbcon is bound Daniel Vetter
2013-01-10 20:48 ` [PATCH 36/36] drm/i915: wake up all pageflip waiters Daniel Vetter
2013-01-10 20:50   ` Daniel Vetter
2013-01-11 23:17 ` [PATCH 00/36] kms locking rework Rob Clark

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1357850897-27102-17-git-send-email-daniel.vetter@ffwll.ch \
    --to=daniel.vetter@ffwll.ch \
    --cc=dri-devel@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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.