VBLANK callbacks in struct drm_driver are deprecated in favor of their
equivalents in struct drm_crtc_funcs. Convert i915 over.
The callback struct drm_driver.get_scanout_position() is deprecated
in favor of struct drm_crtc_helper_funcs.get_scanout_position().
i915 doesn't use CRTC helpers. Instead pass i915's implementation of
get_scanout_position() to DRM core's
drm_crtc_vblank_helper_get_vblank_timestamp_internal().
v3:
* rename dcrtc to _crtc
* use intel_ prefix for i915_crtc_get_vblank_timestamp()
* update for drm_crtc_vblank_helper_get_vblank_timestamp_internal()
v2:
* use DRM's implementation of get_vblank_timestamp()
* simplify function names
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200123135943.24140-8-tzimmermann@suse.de
The callback get_vblank_timestamp() is currently located in struct
drm_driver, but really belongs into struct drm_crtc_funcs. Add an
equivalent there. Driver will be converted in separate patches.
The default implementation is drm_calc_vbltimestamp_from_scanoutpos().
The patch adds drm_crtc_vblank_helper_get_vblank_timestamp(), which is
an implementation for the CRTC callback.
v4:
* more readable code for setting high_prec (Ville, Jani)
v3:
* use refactored timestamp calculation to minimize duplicated code
* do more checks for crtc != NULL to support legacy drivers
v2:
* rename helper to drm_crtc_vblank_helper_get_vblank_timestamp()
* replace drm_calc_vbltimestamp_from_scanoutpos() with
drm_crtc_vblank_helper_get_vblank_timestamp() in docs
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200123135943.24140-4-tzimmermann@suse.de
The new callback get_scanout_position() reads the current location
of the scanout process. The operation is currently located in struct
drm_driver, but really belongs to the CRTC. Drivers will be converted
in separate patches.
To help with the conversion, the timestamp calculation has been
moved from drm_calc_vbltimestamp_from_scanoutpos() to
drm_crtc_vblank_helper_get_vblank_timestamp_internal(). The helper
function supports the new and old interface of get_scanout_position().
drm_calc_vbltimestamp_from_scanoutpos() remains as a wrapper around
the new function.
Callback functions return the scanout position from the CRTC. The
legacy version of the interface receives the device and pipe index,
the modern version receives a pointer to the CRTC. We keep the
legacy version until all drivers have been converted.
v4:
* 80-character line fixes
v3:
* refactor drm_calc_vbltimestamp_from_scanoutpos() to minimize
code duplication
* define types for get_scanout_position() callbacks
v2:
* fix logical op in drm_calc_vbltimestamp_from_scanoutpos()
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Tested-by: Yannick Fertré <yannick.fertre@st.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200123135943.24140-3-tzimmermann@suse.de
VBLANK interrupts can be disabled immediately or with a delay, where the
latter is the default. The former option can be selected by setting
get_vblank_timestamp and enabling vblank_disable_immediate in struct
drm_device. Simplify the code in preparation of the removal of struct
drm_device.get_vblank_timestamp.
v3:
* remove internal setup of vblank_disable_immediate
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200123135943.24140-2-tzimmermann@suse.de
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertenly introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 7649773293 ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20200212193344.GA27929@embeddedor
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
If we fail training at a lower DP link rate let's now keep trying
until we run out of rates to try. Basically the algorithm here is to
start at the link rate that is the theoretical minimum and then slowly
bump up until we run out of rates or hit the max rate of the sink. We
query the sink using a DPCD read.
This is, in fact, important in practice. Specifically at least one
panel hooked up to the bridge (AUO B116XAK01) had a theoretical min
rate more than 1.62 GHz (if run at 24 bpp) and fails to train at the
next rate (2.16 GHz). It would train at 2.7 GHz, though.
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Tested-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191218143416.v3.8.I251add713bc5c97225200894ab110ea9183434fd@changeid
The current bridge driver always forced us to use 24 bits per pixel
over the DP link. This is a waste if you are hooked up to a panel
that only supports 6 bits per color or fewer, since in that case you
can run at 18 bits per pixel and thus end up at a lower DP clock rate.
Let's support this.
While at it, let's clean up the math in the function to avoid rounding
errors (and round in the correct direction when we have to round).
Numbers are sufficiently small (because mode->clock is in kHz) that we
don't need to worry about integer overflow.
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Tested-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
[narmstrong: s/ran/can/]
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191218143416.v3.6.Iaf8d698f4e5253d658ae283d2fd07268076a7c27@changeid
The ti-sn65dsi86 is a bridge from MIPI to DP and thus has two links:
the MIPI link and the DP link. The two links do not need to have the
same format or number of lanes. Stop using MIPI variables when
talking about the DP link.
This has zero functional change because:
* currently we are hardcoding the MIPI link as unpacked RGB888 which
requires 24 bits and currently we are not changing the DP link rate
from the bridge's default of 8 bits per pixel.
* currently we are hardcoding both the MIPI and DP as being 4 lanes.
This is all in prep for fixing some of the above.
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Tested-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191218143416.v3.3.Ia6e05f4961adb0d4a0d32ba769dd7781ee8db431@changeid
This patch reworks the whole delayed deletion of BOs which aren't idle.
Instead of having two counters for the BO structure we resurrect the BO
when we find that a deleted BO is not idle yet.
This has many advantages, especially that we don't need to
increment/decrement the BOs reference counter any more when it
moves on the LRUs.
v2: remove duplicate ttm_tt_destroy, fix holde lock for LRU move
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: xinhui pan <xinhui.pan@amd.com>
Link: https://patchwork.freedesktop.org/patch/352912/
There is no real reason to require drivers to set and use
dev->dev_private. Indeed, the current recommendation, as documented in
drm_device.h, is to embed struct drm_device in the per-device struct
instead of using dev_private.
Remove the requirement for dev_private to have been set to indicate
driver initialization.
For background, quoting Daniel Vetter:
Now there might be some hilarious races this papers over, but:
- Proper drivers should only call drm_dev_register once everything is
set up, including this stuff here. No race possible with anything else
really.
- Slightly more wobbly drivers, including the legacy ones, all use
drm_global_mutex. This was the former BKL, which means that it was
impossible for soeone to go through the load/unload/reload (between
lastclose and firstopen) paths and also run the ioctl. But the ioctl
had to be made unlocked because blocking there killed X:
commit 8f4ff2b06a
Author: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Mon Oct 31 17:46:18 2011 -0400
drm: do not sleep on vblank while holding a mutex
The even more legacy DRM_CONTROL ioctl stayed fully locked. But the
file open/close paths are still fully locked, and that's the only
place legacy drivers should call drm_irq_install/uninstall, so should
all still be fully ordered and protected and happy.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel@ffwll.ch>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200211144753.3175-1-jani.nikula@intel.com
This catches the majority of drivers (unfortunately not if we take
users into account, because all the big drivers have at least a
lastclose hook).
With the prep patches out of the way all drm state is fully protected
and either prevents or can deal with the races from dropping the BKL
around open/close. The only thing left to audit are the various driver
hooks - by keeping the BKL around if any of them are set we have a
very simple cop-out!
Note that one of the biggest prep pieces to get here was making
dev->open_count atomic, which was done in
commit 7e13ad8964
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Fri Jan 24 13:01:07 2020 +0000
drm: Avoid drm_global_mutex for simple inc/dec of dev->open_count
v2:
- Rebase and fix locking in drm_open() (Chris)
- Indentation fix in drm_release
- Typo fix in the commit message (Sam)
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Sam Ravnborg <sam@ravnborg.org>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200204150146.2006481-6-daniel.vetter@ffwll.ch
We want to only take the BKL on crap drivers, but to know whether
we have a crap driver we first need to look it up. Split this shuffle
out from the main BKL-disabling patch, for more clarity. Historical
aside: When the kernel-wide BKL was removed, it was replaced by
drm_global_mutex within the scope of the drm subsystem hence why these
two things are (almost) interchangeable as concepts here.
Since the minors are refcounted drm_minor_acquire is purely internal
and this does not have a driver visible effect.
v2: Push the locking even further into drm_open(), suggested by Chris.
This gives us more symmetry with drm_release(), and maybe a futuer
avenue where we make drm_global_mutex locking (partially) opt-in like
with drm_release_noglobal().
v3:
- Actually push this stuff correctly, don't unlock twice (Chris)
- Fix typo on commit message, plus explain why BKL = drm_global_mutex
(Sam)
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200204150146.2006481-5-daniel.vetter@ffwll.ch
Instead check for master status, in case we've raced.
This is the last exception to the general rule that we restore fbcon
only when there's no master active. Compositors are supposed to drop
their master status before they switch to a different console back to
text mode (or just switch to text mode directly, without a vt switch).
This is known to break some subtests of kms_fbcon_fbt in igt, but they're
just wrong - it does a graphics/text mode switch for the vt without
updating the master status.
Also add a comment to the drm_client->restore hook that this is expected
going forward from all clients (there's currently just one).
v2: Also drop the force in pan_display
v3: Restore the _force to pan_display, this actually means _locked in that
path. Spotted by Noralf.
Cc: Noralf Trønnes <noralf@tronnes.org>
Reviewed-by: Noralf Trønnes <noralf@tronnes.org>
Tested-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200204150146.2006481-3-daniel.vetter@ffwll.ch
