Calling drmm_vram_helper_init() sets up a managed instance of
VRAM MM. Releasing the DRM device also frees the memory manager.
The patch also updates the DRM documentation for VRAM helpers. The
tutorial now describes the new managed interface. The old interfaces
are deprecated and should not be used in new code.
v2:
* rename init function to drmm_vram_helper_init()
* return errno code from init function; caller does not
need vram_mm anyway
* update documentation and remove docs for deprecated
un-managed functions
Silence compiler warning about used but uninitialized 'ipu_state'
variable. In practice, the variable would never be used when
uninitialized, but the compiler cannot know that 'priv->ipu_plane' will
always be NULL if CONFIG_INGENIC_IPU is disabled.
Silence the warning by initializing the value to NULL.
Paul Cercueil [Thu, 16 Jul 2020 16:38:45 +0000 (18:38 +0200)]
drm/ingenic: Support multiple panels/bridges
Support multiple panels or bridges connected to the same DPI output of
the SoC. This setup can be found for instance on the GCW Zero, where the
same DPI output interfaces the internal 320x240 TFT panel, and the ITE
IT6610 HDMI chip.
Paul Cercueil [Thu, 16 Jul 2020 16:38:44 +0000 (18:38 +0200)]
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Paul Cercueil [Thu, 16 Jul 2020 16:38:43 +0000 (18:38 +0200)]
drm/ingenic: Add support for OSD mode
All Ingenic SoCs starting from the JZ4725B support OSD mode.
In this mode, two separate planes can be used. They can have different
positions and sizes, and one can be overlayed on top of the other.
v2: Use fallthrough; instead of /* fall-through */
v3: - Add custom atomic_tail function to handle case where HW gives no
VBLANK
- Use regmap_set_bits() / regmap_clear_bits() when possible
- Use dma_hwdesc_f{0,1} fields in priv structure instead of array
- Use dmam_alloc_coherent() instead of dma_alloc_coherent()
- Use more meaningful 0xf0 / 0xf1 values as DMA descriptors IDs
- Add a bit more code comments
Paul Cercueil [Thu, 16 Jul 2020 16:38:39 +0000 (18:38 +0200)]
drm/ingenic: Add missing CR in debug strings
If you pass a string that is not terminated with a carriage return to
dev_err(), it will eventually be printed with a carriage return, but
not right away, since the kernel will wait for a pr_cont().
Paul Cercueil [Thu, 16 Jul 2020 16:38:38 +0000 (18:38 +0200)]
drm/ingenic: Rename ingenic-drm.c to ingenic-drm-drv.c
Full rename without any modification, except to the Makefile.
Renaming ingenic-drm.c to ingenic-drm-drv.c allow to decouple the module
name from the source file name in the Makefile. This will be useful
later when more source files are added.
Paul Cercueil [Thu, 16 Jul 2020 16:38:35 +0000 (18:38 +0200)]
drm/ingenic: Fix incorrect assumption about plane->index
plane->index is NOT the index of the color plane in a YUV frame.
Actually, a YUV frame is represented by a single drm_plane, even though
it contains three Y, U, V planes.
Lyude Paul [Mon, 29 Jun 2020 22:36:25 +0000 (18:36 -0400)]
drm/nouveau/kms/nvd9-: Fix disabling CRCs alongside OR reprogramming
While I had thought I'd tested this before, it looks like this one issue
slipped by my original CRC patches. Basically, there seem to be a few
rules we need to follow when sending CRC commands to the display
controller:
* CRCs cannot be both disabled and enabled for a single head in the same
flush
* If a head with CRC reporting enabled switches from one OR to another,
there must be a flush before the OR is re-enabled regardless of the
final state of CRC reporting.
So, split nv50_crc_atomic_prepare_notifier_contexts() into two
functions:
* nv_crc_atomic_release_notifier_contexts() - checks whether the CRC
notifier contexts were released successfully after the first flush
* nv_crc_atomic_init_notifier_contexts() - prepares any CRC notifier
contexts for use before enabling reporting
Additionally, in order to force a flush when we re-assign ORs with heads
that have CRCs enabled we split our atomic check function into two:
* nv50_crc_atomic_check_head() - called from our heads' atomic checks,
determines whether a state needs to set or clear CRC reporting
* nv50_crc_atomic_check_outp() - called at the end of the atomic check
after all ORs have been added to the atomic state, and sets
nv50_atom->flush_disable if needed
We expose all available CRC sources. SF, SOR, PIOR, and DAC are exposed
through a single set of "outp" sources: outp-active/auto for a CRC of
the scanout region, outp-complete for a CRC of both the scanout and
blanking/sync region combined, and outp-inactive for a CRC of only the
blanking/sync region. For each source, nouveau selects the appropriate
tap point based on the output path in use. We also expose an "rg"
source, which allows for capturing CRCs of the scanout raster before
it's encoded into a video signal in the output path. This tap point is
referred to as the raster generator.
Note that while there's some other neat features that can be used with
CRC capture on nvidia hardware, like capturing from two CRC sources
simultaneously, I couldn't see any usecase for them and did not
implement them.
Nvidia only allows for accessing CRCs through a shared DMA region that
we program through the core EVO/NvDisplay channel which is referred to
as the notifier context. The notifier context is limited to either 255
(for Fermi-Pascal) or 2047 (Volta+) entries to store CRCs in, and
unfortunately the hardware simply drops CRCs and reports an overflow
once all available entries in the notifier context are filled.
Since the DRM CRC API and igt-gpu-tools don't expect there to be a limit
on how many CRCs can be captured, we work around this in nouveau by
allocating two separate notifier contexts for each head instead of one.
We schedule a vblank worker ahead of time so that once we start getting
close to filling up all of the available entries in the notifier
context, we can swap the currently used notifier context out with
another pre-prepared notifier context in a manner similar to page
flipping.
Unfortunately, the hardware only allows us to this by flushing two
separate updates on the core channel: one to release the current
notifier context handle, and one to program the next notifier context's
handle. When the hardware processes the first update, the CRC for the
current frame is lost. However, the second update can be flushed
immediately without waiting for the first to complete so that CRC
generation resumes on the next frame. According to Nvidia's hardware
engineers, there isn't any cleaner way of flipping notifier contexts
that would avoid this.
Since using vblank workers to swap out the notifier context will ensure
we can usually flush both updates to hardware within the timespan of a
single frame, we can also ensure that there will only be exactly one
frame lost between the first and second update being executed by the
hardware. This gives us the guarantee that we're always correctly
matching each CRC entry with it's respective frame even after a context
flip. And since IGT will retrieve the CRC entry for a frame by waiting
until it receives a CRC for any subsequent frames, this doesn't cause an
issue with any tests and is much simpler than trying to change the
current DRM API to accommodate.
In order to facilitate testing of correct handling of this limitation,
we also expose a debugfs interface to manually control the threshold for
when we start trying to flip the notifier context. We will use this in
igt to trigger a context flip for testing purposes without needing to
wait for the notifier to completely fill up. This threshold is reset
to the default value set by nouveau after each capture, and is exposed
in a separate folder within each CRTC's debugfs directory labelled
"nv_crc".
Changes since v1:
* Forgot to finish saving crc.h before saving, whoops. This just adds
some corrections to the empty function declarations that we use if
CONFIG_DEBUG_FS isn't enabled.
Changes since v2:
* Don't check return code from debugfs_create_dir() or
debugfs_create_file() - Greg K-H
Changes since v3:
(no functional changes)
* Fix SPDX license identifiers (checkpatch)
* s/uint32_t/u32/ (checkpatch)
* Fix indenting in switch cases (checkpatch)
Changes since v4:
* Remove unneeded param changes with nv50_head_flush_clr/set
* Rebase
Changes since v5:
* Remove set but unused variable (outp) in nv50_crc_atomic_check() -
Kbuild bot
Lyude Paul [Tue, 21 Jan 2020 20:53:46 +0000 (15:53 -0500)]
drm/nouveau/kms/nv50-: Move hard-coded object handles into header
While most of the functionality on Nvidia GPUs doesn't require using an
explicit handle instead of the main VRAM handle + offset, there are a
couple of places that do require explicit handles, such as CRC
functionality. Since this means we're about to add another
nouveau-chosen handle, let's just go ahead and move any hard-coded
handles into a single header. This is just to keep things slightly
organized, and to make it a little bit easier if we need to add more
handles in the future.
This patch should contain no functional changes.
Changes since v3:
* Correct SPDX license identifier (checkpatch)
Lyude Paul [Fri, 7 Feb 2020 19:10:46 +0000 (14:10 -0500)]
drm/nouveau/kms/nv50-: Expose nv50_outp_atom in disp.h
In order to make sure that we flush disable updates at the right time
when disabling CRCs, we'll need to be able to look at the outp state to
see if we're changing it at the same time that we're disabling CRCs.
Lyude Paul [Thu, 6 Feb 2020 19:37:36 +0000 (14:37 -0500)]
drm/nouveau/kms/nv140-: Track wndw mappings in nv50_head_atom
While we're not quite ready yet to add support for flexible wndw
mappings, we are going to need to at least keep track of the static wndw
mappings we're currently using in each head's atomic state. We'll likely
use this in the future to implement real flexible window mapping, but
the primary reason we'll need this is for CRC support.
See: on nvidia hardware, each CRC entry in the CRC notifier dma context
has a "tag". This tag corresponds to the nth update on a specific
EVO/NvDisplay channel, which itself is referred to as the "controlling
channel". For gf119+ this can be the core channel, ovly channel, or base
channel. Since we don't expose CRC entry tags to userspace, we simply
ignore this feature and always use the core channel as the controlling
channel. Simple.
Things get a little bit more complicated on gv100+ though. GV100+ only
lets us set the controlling channel to a specific wndw channel, and that
wndw must be owned by the head that we're grabbing CRCs when we enable
CRC generation. Thus, we always need to make sure that each atomic head
state has at least one wndw that is mapped to the head, which will be
used as the controlling channel.
Note that since we don't have flexible wndw mappings yet, we don't
expect to run into any scenarios yet where we'd have a head with no
mapped wndws. When we do add support for flexible wndw mappings however,
we'll need to make sure that we handle reprogramming CRC capture if our
controlling wndw is moved to another head (and potentially reject the
new head state entirely if we can't find another available wndw to
replace it).
With that being said, nouveau currently tracks wndw visibility on heads.
It does not keep track of the actual ownership mappings, which are
(currently) statically programmed. To fix this, we introduce another
bitmask into nv50_head_atom.wndw to keep track of ownership separately
from visibility. We then introduce a nv50_head callback to handle
populating the wndw ownership map, and call it during the atomic check
phase when core->assign_windows is set to true.
Lyude Paul [Tue, 17 Mar 2020 18:54:06 +0000 (14:54 -0400)]
drm/nouveau/kms/nv50-: Fix disabling dithering
While we expose the ability to turn off hardware dithering for nouveau,
we actually make the mistake of turning it on anyway, due to
dithering_depth containing a non-zero value if our dithering depth isn't
also set to 6 bpc.
So, fix it by never enabling dithering when it's disabled.
Lyude Paul [Fri, 13 Dec 2019 22:06:59 +0000 (17:06 -0500)]
drm/nouveau/kms/nv140-: Don't modify depth in state during atomic commit
Currently, we modify the depth value stored in the atomic state when
performing a commit in order to workaround the fact we haven't
implemented support for depths higher then 10 yet. This isn't idempotent
though, as it will happen every atomic commit where we modify the OR
state even if the head's depth in the atomic state hasn't been modified.
Normally this wouldn't matter, since we don't modify OR state outside of
modesets, but since the CRC capture region is implemented as part of the
OR state in hardware we'll want to make sure all commits modifying OR
state are idempotent so as to avoid changing the depth unexpectedly.
So, fix this by simply not writing the reduced depth value we come up
with to the atomic state.
Lyude Paul [Fri, 17 Apr 2020 19:33:13 +0000 (15:33 -0400)]
drm/vblank: Add vblank works
Add some kind of vblank workers. The interface is similar to regular
delayed works, and is mostly based off kthread_work. It allows for
scheduling delayed works that execute once a particular vblank sequence
has passed. It also allows for accurate flushing of scheduled vblank
works - in that flushing waits for both the vblank sequence and job
execution to complete, or for the work to get cancelled - whichever
comes first.
Whatever hardware programming we do in the work must be fast (must at
least complete during the vblank or scanout period, sometimes during the
first few scanlines of the vblank). As such we use a high-priority
per-CRTC thread to accomplish this.
Changes since v7:
* Stuff drm_vblank_internal.h and drm_vblank_work_internal.h contents
into drm_internal.h
* Get rid of unnecessary spinlock in drm_crtc_vblank_on()
* Remove !vblank->worker check
* Grab vbl_lock in drm_vblank_work_schedule()
* Mention self-rearming work items in drm_vblank_work_schedule() kdocs
* Return 1 from drm_vblank_work_schedule() if the work was scheduled
successfully, 0 or error code otherwise
* Use drm_dbg_core() instead of DRM_DEV_ERROR() in
drm_vblank_work_schedule()
* Remove vblank->worker checks in drm_vblank_destroy_worker() and
drm_vblank_flush_worker()
Changes since v6:
* Get rid of ->pending and seqcounts, and implement flushing through
simpler means - danvet
* Get rid of work_lock, just use drm_device->event_lock
* Move drm_vblank_work item cleanup into drm_crtc_vblank_off() so that
we ensure that all vblank work has finished before disabling vblanks
* Add checks into drm_crtc_vblank_reset() so we yell if it gets called
while there's vblank workers active
* Grab event_lock in both drm_crtc_vblank_on()/drm_crtc_vblank_off(),
the main reason for this is so that other threads calling
drm_vblank_work_schedule() are blocked from attempting to schedule
while we're in the middle of enabling/disabling vblanks.
* Move drm_handle_vblank_works() call below drm_handle_vblank_events()
* Simplify drm_vblank_work_cancel_sync()
* Fix drm_vblank_work_cancel_sync() documentation
* Move wake_up_all() calls out of spinlock where we can. The only one I
left was the call to wake_up_all() in drm_vblank_handle_works() as
this seemed like it made more sense just living in that function
(which is all technically under lock)
* Move drm_vblank_work related functions into their own source files
* Add drm_vblank_internal.h so we can export some functions we don't
want drivers using, but that we do need to use in drm_vblank_work.c
* Add a bunch of documentation
Changes since v4:
* Get rid of kthread interfaces we tried adding and move all of the
locking into drm_vblank.c. For implementing drm_vblank_work_flush(),
we now use a wait_queue and sequence counters in order to
differentiate between multiple work item executions.
* Get rid of drm_vblank_work_cancel() - this would have been pretty
difficult to actually reimplement and it occurred to me that neither
nouveau or i915 are even planning to use this function. Since there's
also no async cancel function for most of the work interfaces in the
kernel, it seems a bit unnecessary anyway.
* Get rid of to_drm_vblank_work() since we now are also able to just
pass the struct drm_vblank_work to work item callbacks anyway
Changes since v3:
* Use our own spinlocks, don't integrate so tightly with kthread_works
Changes since v2:
* Use kthread_workers instead of reinventing the wheel.
Cc: Tejun Heo <tj@kernel.org> Cc: dri-devel@lists.freedesktop.org Cc: nouveau@lists.freedesktop.org Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Co-developed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Lyude Paul <lyude@redhat.com> Acked-by: Dave Airlie <airlied@gmail.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200627194657.156514-4-lyude@redhat.com
Lyude Paul [Tue, 23 Jun 2020 20:24:00 +0000 (16:24 -0400)]
drm/vblank: Use spin_(un)lock_irq() in drm_crtc_vblank_off()
This got me confused for a bit while looking over this code: I had been
planning on adding some blocking function calls into this function, but
seeing the irqsave/irqrestore variants of spin_(un)lock() didn't make it
very clear whether or not that would actually be safe.
So I went ahead and reviewed every single driver in the kernel that uses
this function, and they all fall into three categories:
All of these will be guaranteed to have IRQs enabled, which means it's
perfectly safe to block here. Just to make things a little less
confusing to others in the future, let's switch over to
spin_lock_irq()/spin_unlock_irq() to make that fact a little more
obvious.
Lyude Paul [Tue, 14 Apr 2020 16:34:52 +0000 (12:34 -0400)]
drm/vblank: Register drmm cleanup action once per drm_vblank_crtc
Since we'll be allocating resources for kthread_create_worker() in the
next commit (which could fail and require us to clean up the mess),
let's simplify the cleanup process a bit by registering a
drm_vblank_init_release() action for each drm_vblank_crtc so they're
still cleaned up if we fail to initialize one of them.
Changes since v3:
* Use drmm_add_action_or_reset() - Daniel Vetter
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: dri-devel@lists.freedesktop.org Cc: nouveau@lists.freedesktop.org Reviewed-by: Daniel Vetter <daniel@ffwll.ch> Signed-off-by: Lyude Paul <lyude@redhat.com> Acked-by: Dave Airlie <airlied@gmail.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200627194657.156514-2-lyude@redhat.com
Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.
Deterministic algorithm:
For each file:
If not .svg:
For each line:
If doesn't contain `\bxmlns\b`:
For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
If both the HTTP and HTTPS versions
return 200 OK and serve the same content:
Replace HTTP with HTTPS.
Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.
Deterministic algorithm:
For each file:
If not .svg:
For each line:
If doesn't contain `\bxmlns\b`:
For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
If both the HTTP and HTTPS versions
return 200 OK and serve the same content:
Replace HTTP with HTTPS.
Paul Cercueil [Thu, 16 Jul 2020 12:56:46 +0000 (14:56 +0200)]
drm/panel-simple: Fix inverted V/H SYNC for Frida FRD350H54004 panel
The FRD350H54004 panel was marked as having active-high VSYNC and HSYNC
signals, which sorts-of worked, but resulted in the picture fading out
under certain circumstances.
Fix this issue by marking VSYNC and HSYNC signals active-low.
Daniel Vetter [Wed, 8 Jul 2020 14:20:50 +0000 (16:20 +0200)]
drm/tilcdc: Use standard drm_atomic_helper_commit
Gives us proper nonblocking support for free, and a pile of other
things. The tilcdc code is simply old enough that it was never
converted over, but was stuck forever with the copypasta from when it
was initially merged.
The riskiest thing with this conversion is maybe that there's an issue
with the vblank handling or vblank event handling, which will upset
the modern commit support in atomic helpers. But from a cursory review
drm_crtc_vblank_on/off is called in the right places, and the event
handling also seems to exist (albeit with much hand-rolling and
probably some races, could perhaps be converted over to
drm_crtc_arm_vblank_event without any real loss).
Motivated by me not having to hand-roll the dma-fence annotations for
this.
v2: Clear out crtc_state->event when we're handling the event, to
avoid upsetting the helpers (reported by Jyri).
v3: Also send out even whent the crtc is getting disabled. Tilcdc looks a
bit like conversion to simple display helpers would work out really
nice.
drm/mgag200: Rename G200WB prepare/commit function
The prepare and commit helpers for G200WB devices control the BMC.
Rename them accordingly. While at it, also change the passed value's
type to struct mga_device and remove some type upcasting.
drm/mgag200: Set/clear <syncrst> field in display enable/disable helpers
Modifying the <syncrst> field in mgag200_{enable,disable}_display()
makes the code more readable. Also clear the <asyncrst> field to enable
the display. The other bits in SEQ0 are unused, so no functional changes
are made.
Of the DPMS code, only ON and OFF states are used. Simplify mode setting
by moving both into separate functions and removing the rest.
The original code busy waited in the middle of updating the screen state
in SEQ1. To simplify the procedure, the new code busy waits first and then
updates SEQ1 in one chunk.
The DPMS code also set the LUT before enabling the screen. The patch moves
this code into the simple-display pipe's enable function.
drm/mgag200: Don't set or clear <scroff> field during modeset
The simple pipe's disable function disables the screen by calling
mgag200_disable_screen(). The simple pipe's enable function enables the
screen by calling mgag200_enable_display().
During modeset operations the screen is off and remains off. It's only
enabled after the modeset has been completed. Therefore remove all code
that sets or clears the <scroff> field while in modeset.
The related code also modifies the <syncrst> field in SEQ0. For now, keep
this code in place.
Marek Szyprowski [Fri, 19 Jun 2020 10:36:17 +0000 (12:36 +0200)]
drm: radeon: fix common struct sg_table related issues
The Documentation/DMA-API-HOWTO.txt states that the dma_map_sg() function
returns the number of the created entries in the DMA address space.
However the subsequent calls to the dma_sync_sg_for_{device,cpu}() and
dma_unmap_sg must be called with the original number of the entries
passed to the dma_map_sg().
struct sg_table is a common structure used for describing a non-contiguous
memory buffer, used commonly in the DRM and graphics subsystems. It
consists of a scatterlist with memory pages and DMA addresses (sgl entry),
as well as the number of scatterlist entries: CPU pages (orig_nents entry)
and DMA mapped pages (nents entry).
It turned out that it was a common mistake to misuse nents and orig_nents
entries, calling DMA-mapping functions with a wrong number of entries or
ignoring the number of mapped entries returned by the dma_map_sg()
function.
To avoid such issues, lets use a common dma-mapping wrappers operating
directly on the struct sg_table objects and use scatterlist page
iterators where possible. This, almost always, hides references to the
nents and orig_nents entries, making the code robust, easier to follow
and copy/paste safe.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Reviewed-by: Christian König <christian.koenig@amd.com> Link: https://patchwork.freedesktop.org/patch/371172/ Signed-off-by: Christian König <christian.koenig@amd.com>
Marek Szyprowski [Fri, 19 Jun 2020 10:36:04 +0000 (12:36 +0200)]
drm: amdgpu: fix common struct sg_table related issues
The Documentation/DMA-API-HOWTO.txt states that the dma_map_sg() function
returns the number of the created entries in the DMA address space.
However the subsequent calls to the dma_sync_sg_for_{device,cpu}() and
dma_unmap_sg must be called with the original number of the entries
passed to the dma_map_sg().
struct sg_table is a common structure used for describing a non-contiguous
memory buffer, used commonly in the DRM and graphics subsystems. It
consists of a scatterlist with memory pages and DMA addresses (sgl entry),
as well as the number of scatterlist entries: CPU pages (orig_nents entry)
and DMA mapped pages (nents entry).
It turned out that it was a common mistake to misuse nents and orig_nents
entries, calling DMA-mapping functions with a wrong number of entries or
ignoring the number of mapped entries returned by the dma_map_sg()
function.
To avoid such issues, lets use a common dma-mapping wrappers operating
directly on the struct sg_table objects and use scatterlist page
iterators where possible. This, almost always, hides references to the
nents and orig_nents entries, making the code robust, easier to follow
and copy/paste safe.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Reviewed-by: Christian König <christian.koenig@amd.com> Link: https://patchwork.freedesktop.org/patch/371142/ Signed-off-by: Christian König <christian.koenig@amd.com>
Andrey Lebedev [Fri, 19 Jun 2020 07:58:59 +0000 (10:58 +0300)]
drm/lima: Expose job_hang_limit module parameter
Some pp or gp jobs can be successfully repeated even after they time outs.
Introduce lima module parameter to specify number of times a job can hang
before being dropped.
This change expands the coverage for the IGT kms_cursor_crc test, where
the size varies between 64 and 512 for a square cursor. With this, in
addition to the cursor 64x64, this patch enables the test of cursors with
sizes: 128x128, 256x256, and 512x512.
Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.
Deterministic algorithm:
For each file:
If not .svg:
For each line:
If doesn't contain `\bxmlns\b`:
For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
If both the HTTP and HTTPS versions
return 200 OK and serve the same content:
Replace HTTP with HTTPS.
Liu Ying [Thu, 9 Jul 2020 02:02:36 +0000 (10:02 +0800)]
drm/bridge: dw-hdmi: Always add the bridge in the global bridge list
It doesn't hurt to add the bridge in the global bridge list also for
platform specific dw-hdmi drivers which are based on the component
framework. This can be achieved by moving the drm_bridge_add() function
call from dw_hdmi_probe() to __dw_hdmi_probe(). A counterpart movement
for drm_bridge_remove() is also needed then. Moreover, since drm_bridge_add()
initializes &bridge->hpd_mutex, this may help those platform specific
dw-hdmi drivers(based on the component framework) avoid accessing the
uninitialized mutex in drm_bridge_hpd_notify() which is called in
dw_hdmi_irq(). Putting drm_bridge_add() in __dw_hdmi_probe() just before
it returns successfully should bring no logic change for platforms based
on the DRM bridge API, which is a good choice from safety point of view.
Also, __dw_hdmi_probe() is renamed to dw_hdmi_probe() since dw_hdmi_probe()
does nothing else but calling __dw_hdmi_probe(). Similar renaming applies
to the __dw_hdmi_remove()/dw_hdmi_remove() pair.
Fixes: ec971aaa6775 ("drm: bridge: dw-hdmi: Make connector creation optional") Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Neil Armstrong <narmstrong@baylibre.com> Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> Cc: Jonas Karlman <jonas@kwiboo.se> Cc: Jernej Skrabec <jernej.skrabec@siol.net> Cc: David Airlie <airlied@linux.ie> Cc: Daniel Vetter <daniel@ffwll.ch> Cc: Boris Brezillon <boris.brezillon@collabora.com> Cc: Jerome Brunet <jbrunet@baylibre.com> Cc: Cheng-Yi Chiang <cychiang@chromium.org> Cc: Dariusz Marcinkiewicz <darekm@google.com> Cc: Archit Taneja <architt@codeaurora.org> Cc: Jose Abreu <joabreu@synopsys.com> Cc: Sam Ravnborg <sam@ravnborg.org> Cc: dri-devel@lists.freedesktop.org Cc: NXP Linux Team <linux-imx@nxp.com> Signed-off-by: Liu Ying <victor.liu@nxp.com> Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Link: https://patchwork.freedesktop.org/patch/msgid/1594260156-8316-2-git-send-email-victor.liu@nxp.com
Liu Ying [Thu, 9 Jul 2020 02:02:35 +0000 (10:02 +0800)]
drm/bridge: dw-hdmi: Don't cleanup i2c adapter and ddc ptr in __dw_hdmi_probe() bailout path
It's unnecessary to cleanup the i2c adapter and the ddc pointer in
the bailout path of __dw_hdmi_probe(), since the adapter is not
added and the ddc pointer is not set.
Fixes: a23d6265f033 ("drm: bridge: dw-hdmi: Extract PHY interrupt setup to a function") Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Neil Armstrong <narmstrong@baylibre.com> Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> Cc: Jonas Karlman <jonas@kwiboo.se> Cc: Jernej Skrabec <jernej.skrabec@siol.net> Cc: David Airlie <airlied@linux.ie> Cc: Daniel Vetter <daniel@ffwll.ch> Cc: Boris Brezillon <boris.brezillon@collabora.com> Cc: Jerome Brunet <jbrunet@baylibre.com> Cc: Cheng-Yi Chiang <cychiang@chromium.org> Cc: Dariusz Marcinkiewicz <darekm@google.com> Cc: Archit Taneja <architt@codeaurora.org> Cc: Jose Abreu <joabreu@synopsys.com> Cc: dri-devel@lists.freedesktop.org Cc: NXP Linux Team <linux-imx@nxp.com> Signed-off-by: Liu Ying <victor.liu@nxp.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Link: https://patchwork.freedesktop.org/patch/msgid/1594260156-8316-1-git-send-email-victor.liu@nxp.com
Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.
Deterministic algorithm:
For each file:
If not .svg:
For each line:
If doesn't contain `\bxmlns\b`:
For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
If both the HTTP and HTTPS versions
return 200 OK and serve the same content:
Replace HTTP with HTTPS.
If smtcfb_pci_probe() does not detect a valid chip it cleans up
everything and returns 0. This can result in various bad things later.
The patch sets the error code on the corresponding path.
Found by Linux Driver Verification project (linuxtesting.org).
fbdev: da8xx-fb: go to proper label on error handling paths in probe
fb_probe() can successfully allocate a new frame buffer, but then fail
to perform some operations with regulator. In these cases fb_probe()
goes to label err_pm_runtime_disable where the frame buffer is not
released. The patch makes fb_probe() to go to label err_release_fb on
corresponding error handling paths.
Found by Linux Driver Verification project (linuxtesting.org).
Signed-off-by: Evgeny Novikov <novikov@ispras.ru> Cc: Sekhar Nori <nsekhar@ti.com> Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200702160540.24546-1-novikov@ispras.ru
Adam Ford [Tue, 30 Jun 2020 18:26:36 +0000 (13:26 -0500)]
omapfb: dss: Fix max fclk divider for omap36xx
The drm/omap driver was fixed to correct an issue where using a
divider of 32 breaks the DSS despite the TRM stating 32 is a valid
number. Through experimentation, it appears that 31 works, and
it is consistent with the value used by the drm/omap driver.
This patch fixes the divider for fbdev driver instead of the drm.
Fixes: f76ee892a99e ("omapfb: copy omapdss & displays for omapfb") Cc: <stable@vger.kernel.org> #4.5+ Signed-off-by: Adam Ford <aford173@gmail.com> Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ti.com> Cc: Dave Airlie <airlied@gmail.com> Cc: Rob Clark <robdclark@gmail.com>
[b.zolnierkie: mark patch as applicable to stable 4.5+ (was 4.9+)] Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200630182636.439015-1-aford173@gmail.com
Aditya Pakki [Sun, 14 Jun 2020 03:05:18 +0000 (22:05 -0500)]
omapfb: fix multiple reference count leaks due to pm_runtime_get_sync
On calling pm_runtime_get_sync() the reference count of the device
is incremented. In case of failure, decrement the
reference count before returning the error.
Evgeny Novikov [Tue, 30 Jun 2020 19:54:51 +0000 (22:54 +0300)]
video: fbdev: neofb: fix memory leak in neo_scan_monitor()
neofb_probe() calls neo_scan_monitor() that can successfully allocate a
memory for info->monspecs.modedb and proceed to case 0x03. There it does
not free the memory and returns -1. neofb_probe() goes to label
err_scan_monitor, thus, it does not free this memory through calling
fb_destroy_modedb() as well. We can not go to label err_init_hw since
neo_scan_monitor() can fail during memory allocation. So, the patch frees
the memory directly for case 0x03.
Found by Linux Driver Verification project (linuxtesting.org).
Signed-off-by: Evgeny Novikov <novikov@ispras.ru> Cc: Jani Nikula <jani.nikula@intel.com> Cc: Mike Rapoport <rppt@linux.ibm.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200630195451.18675-1-novikov@ispras.ru
Evgeny Novikov [Fri, 19 Jun 2020 16:21:36 +0000 (19:21 +0300)]
video: fbdev: savage: fix memory leak on error handling path in probe
savagefb_probe() calls savage_init_fb_info() that can successfully
allocate memory for info->pixmap.addr but then fail when
fb_alloc_cmap() fails. savagefb_probe() goes to label failed_init and
does not free allocated memory. It is not valid to go to label
failed_mmio since savage_init_fb_info() can fail during memory
allocation as well. So, the patch free allocated memory on the error
handling path in savage_init_fb_info() itself.
Found by Linux Driver Verification project (linuxtesting.org).
Use array3_size() helper instead of the open-coded version in scr_memcpyw()
and scr_memsetw(). These sorts of multiplication factors need to be wrapped
in array3_size().
This issue was found with the help of Coccinelle and, audited and fixed
manually.
video: fbdev: amifb: add FIXMEs about {put,get}_user() failures
Since we lack the hardware (or proper emulator setup) for
testing needed changes add FIXMEs to document the issues
(so at least they are not forgotten).
video: fbdev: amifb: add FIXME about dead APUS support
On 5/14/20 10:21 PM, Geert Uytterhoeven wrote:
> These #ifdefs are relics from APUS (Amiga Power-Up System), which
> added a PPC board. APUS support was killed off a long time ago,
> when arch/ppc/ was still king, but these #ifdefs were missed, because
> they didn't test for CONFIG_APUS.
Add FIXME about using the C code variants (APUS ones) in the future.
Ville Syrjälä [Wed, 27 May 2020 13:03:09 +0000 (16:03 +0300)]
drm/edid: Iterate through all DispID ext blocks
Apparently there are EDIDs in the wild with multiple DispID extension
blocks. Iterate through them all.
In one particular case the tile information is specicied in the
second DispID ext block, and since the current parser only looks
at the first DispID ext block we don't notice that we're dealing
with a tiled display.
While at it change a few functions to return void since we have
no use for the errno.
Ville Syrjälä [Wed, 27 May 2020 13:03:08 +0000 (16:03 +0300)]
drm/edid: Allow looking for ext blocks starting from a specified index
Apparently EDIDs with multiple DispID ext blocks is a thing, so prepare
for iterating through multiple ext blocks of the same type by
passing the starting ext block index to drm_find_edid_extension(). Well
also have drm_find_edid_extension() update the index to point to the
next ext block on success. Thus we should be able to call
drm_find_edid_extension() in loop.
Chris Wilson [Wed, 8 Jul 2020 15:49:11 +0000 (16:49 +0100)]
drm/vgem: Replace opencoded version of drm_gem_dumb_map_offset()
drm_gem_dumb_map_offset() now exists and does everything
vgem_gem_dump_map does and *ought* to do.
In particular, vgem_gem_dumb_map() was trying to reject mmapping an
imported dmabuf by checking the existence of obj->filp. Unfortunately,
we always allocated an obj->filp, even if unused for an imported dmabuf.
Instead, the drm_gem_dumb_map_offset(), since commit 90378e589192
("drm/gem: drm_gem_dumb_map_offset(): reject dma-buf"), uses the
obj->import_attach to reject such invalid mmaps.
This prevents vgem from allowing userspace mmapping the dumb handle and
attempting to incorrectly fault in remote pages belonging to another
device, where there may not even be a struct page.
v2: Use the default drm_gem_dumb_map_offset() callback
Marek Vasut [Sat, 29 Feb 2020 22:16:49 +0000 (23:16 +0100)]
drm/stm: repair runtime power management
Add missing pm_runtime_get_sync() into ltdc_crtc_atomic_enable() to
match pm_runtime_put_sync() in ltdc_crtc_atomic_disable(), otherwise
the LTDC might suspend via runtime PM, disable clock, and then fail
to resume later on.
The test which triggers it is roughly -- run qt5 application which
uses eglfs platform and etnaviv, stop the application, sleep for 15
minutes, run the application again. This leads to a timeout waiting
for vsync, because the LTDC has suspended, but did not resume.
Fixes: 35ab6cfbf211 ("drm/stm: support runtime power management") Signed-off-by: Marek Vasut <marex@denx.de> Cc: Yannick Fertré <yannick.fertre@st.com> Cc: Philippe Cornu <philippe.cornu@st.com> Cc: Benjamin Gaignard <benjamin.gaignard@linaro.org> Cc: Vincent Abriou <vincent.abriou@st.com> Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com> Cc: Alexandre Torgue <alexandre.torgue@st.com>
To: dri-devel@lists.freedesktop.org Cc: linux-stm32@st-md-mailman.stormreply.com Cc: linux-arm-kernel@lists.infradead.org Acked-by: Philippe Cornu <philippe.cornu@st.com> Tested-by: Yannick Fertre <yannick.fertre@st.com> Signed-off-by: Benjamin Gaignard <benjamin.gaignard@st.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200229221649.90813-1-marex@denx.de
Neil Armstrong [Tue, 7 Jul 2020 13:50:09 +0000 (15:50 +0200)]
drm/meson: overlay: fix build failure
The recent GCC compiler is very picky with the VD_H_START() and
AFBC_DEC_PIXEL_BGN_H() macros, triggering a runtime assert error as:
In function 'meson_overlay_setup_scaler_params',
inlined from 'meson_overlay_atomic_update' at
drivers/gpu/drm/meson/meson_overlay.c:542:2:
./include/linux/compiler.h:392:38: error: call to
'__compiletime_assert_341' declared with attribute error: FIELD_PREP:
value too large for the field
drivers/gpu/drm/meson/meson_overlay.c:413:4: note: in expansion of macro
'AFBC_DEC_PIXEL_BGN_H'
413 | AFBC_DEC_PIXEL_BGN_H(hd_start_lines - afbc_left) |
| ^~~~~~~~~~~~~~~~~~~~
./include/linux/compiler.h:392:38: error: call to
'__compiletime_assert_401' declared with attribute error: FIELD_PREP:
value too large for the field
It's not expected to overflow these fields, but the compiler did
find a case where it overflows.
We can safely ignore this, so mask the value with the field width.
drm/ast: Initialize mode setting in ast_mode_config_init()
There's modesetting init code in ast_main.c. Move it to ast_mode.c and
merge it with the modesetting init code in ast_mode_init(). The result
is ast_mode_config_init(), which initalizes the whole modesetting.
drm/ast: Replace struct ast_crtc with struct drm_crtc
Struct ast_crtc has been cleaned up and it's now a wrapper around the
DRM CRTC structure struct drm_crtc. This patch converts the driver to
struct drm_crtc and removes struct ast_crtc.
drm/ast: Init cursors before creating modesetting structures
The cursor helpers reserve buffer objects in VRAM and update their
content. So although tied to modesetting, cursor helpers are more
of a memory manager. The modesetting's cursor plane requires this
functionality, so initialize cursors before modesetting.
While at it, also add an error check for ast_cursor_init().
Updating the image in a cursor's HW BO requires a mapping of the BO's
buffer in the kernel's address space. Cursor image updates can happen
frequently and create CPU overhead.
As cursor HW BOs are small and never move, they are now map exactly
once during the initialization and the mapping is used throughout the
driver's lifetime.
This change also removes a possible source of failures from
ast_cursor_show(). As the helper does not establish mappings, it cannot
fail. As a result, the cursor plane's atomic-update helper does not
call any failable interfaces. All failures are detected before trying
to update the cursor plane.
drm/ast: Replace ast_cursor_move() with ast_cursor_show()
Having a cursor move function is misleading, as it actually enables the
cursor's image for displaying. So rename it to ast_cursor_show(). It's
semantics is to show a cursor at the specified location on the screen.
The displayed cursor is always the image in the cursor front BO.
This change also simplifies struct ast_crtc to being a mere wrapper around
around struct drm_crtc. It will be removed by a later patch.
drm/ast: Move cursor fb pinning and mapping into helper
The new helper ast_cursor_blit() updates a cursor's backbuffer HW
BO from a framebuffer structure. The cursor plane's prepare_fb()
function now uses the new interface.
Pinning and mapping of BOs is done automatically by the helper. This
includes the source BO, which was not pinned by the original code in
prepare_fb().
Maxime Ripard [Thu, 11 Jun 2020 13:36:54 +0000 (15:36 +0200)]
drm/vc4: crtc: Remove the feed_txp tests
Now that the code in vc4_crtc accessing registers is only meant for the
pixelvalve, it doesn't make sense anymore to test whether we're accessing
the TXP or not and we can safely remove those checks.
Maxime Ripard [Thu, 11 Jun 2020 13:36:53 +0000 (15:36 +0200)]
drm/vc4: txp: Turn the TXP into a CRTC of its own
The TXP so far has been leveraging the PixelValve infrastructure in the
driver, that was really two things: the interaction with DRM's CRTC
concept, the setup of the underlying pixelvalve and the setup of the shared
HVS, the pixelvalve part being irrelevant to the TXP since it accesses the
HVS directly.
Now that we have a clear separation between the three parts, we can
represent the TXP as a CRTC of its own, leveraging the common CRTC and HVS
code, but leaving aside the pixelvalve setup.
projects / mirror_ubuntu-kernels.git / log