This patch adds the process module and three helper modules:
- kfd_process, which handles process which open /dev/kfd
- kfd_doorbell, which provides helper functions for doorbell allocation,
release and mapping to userspace
- kfd_pasid, which provides helper functions for pasid allocation and release
- kfd_aperture, which provides helper functions for managing the LDS, Local GPU
memory and Scratch memory apertures of the process
This patch only contains the basic kfd_process module, which doesn't contain
the reference to the queue scheduler. This was done to allow easier code review.
Also, this patch doesn't contain the calls to the IOMMU driver for binding the
pasid to the device. Again, this was done to allow easier code review
The kfd_process object is created when a process opens /dev/kfd and is closed
when the mm_struct of that process is teared-down.
v3:
Removed kfd_vidmem.c file
Replaced direct mmput call to mmu_notifier release
Removed typedefs
Moved bool field to end of the structure
Added new kernel params for gart usage limitation
Added initialization of sa manager
Fixed debug messages
Remove support for LDS in 32 bit
Changed code to support mmap of doorbell pages from userspace
Added documentation for apertures
v4: Replaced RCU by SRCU for kfd_process list management
v5:
Move amdkfd from drm/radeon/ to drm/amd/
Rename kfd_aperture.c to kfd_flat_memory.c
Protect against multiple init calls
MQD size is H/W dependent so moved it to device info structure
Rename kfd_mem_obj structure's members
Use delayed function for process tear-down
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
This patch adds the topology module to the driver. The topology is exposed to
userspace through the sysfs.
The calls to add and remove a device to/from topology are done by the radeon
driver.
v3:
The CPU information, that is provided in the topology section of the amdkfd
driver, is extracted from the CRAT table. Unlike the CPU information located
in /sys/devices/system/cpu/cpu*, which is extracted from the SRAT table.
While the CPU information provided by the CRAT and the SRAT tables might be
identical, the node topology might be different. The SRAT table contains the
topology of CPU nodes only. The CRAT table contains the topology of CPU and GPU
nodes together (and can be interleaved). For example CPU node 1 in SRAT can be
CPU node 3 in CRAT. Furthermore it's worth to mention that the CRAT table
contains only HSA compatible nodes (nodes which are compliant with the HSA
spec).
To recap, amdkfd exposes a different kind of topology than the one exposed by
/sys/devices/system/cpu/cpu even though it may contain similar information.
v4:
The topology module doesn't support uevent handling and doesn't notify the
userspace about runtime modifications. It is up to the userspace to acquire
snapshots of the topology information created by the amdkfd and exposed
in sysfs.
The following is an example of how the topology looks on a Kaveri A10-7850K
system with amdkfd installed:
/sys/devices/virtual/kfd/kfd/
|
--- topology/
|
|--- generation_id
|--- system_properties
|--- nodes/
|
|--- 0/
|
|--- gpu_id
|--- name
|--- properties
|--- caches/
|
|--- 0/
|
|--- properties
|--- 1/
|
|--- properties
|--- 2/
|
|--- properties
|--- io_links/
|
|--- mem_banks/
|
|--- 0/
|
|--- properties
|--- 1/
|
|--- properties
|--- 2/
|
|--- properties
|--- 3/
|
|--- properties
v5:
Move amdkfd from drm/radeon/ to drm/amd/
Add a check if dev->gpu pointer is null before accessing it in the
node_show function in kfd_topology.c
This situation may occur when amdkfd is loaded and there is a GPU with a CRAT
table, but that GPU isn't supported by amdkfd
Signed-off-by: Evgeny Pinchuk <evgeny.pinchuk@amd.com>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
This patch adds the amdkfd skeleton driver. The driver does nothing except
define a /dev/kfd device.
It returns -ENODEV on all amdkfd IOCTLs.
v3: Move bool field to the end of structure, removed the pmc ioctls and added
a meaningful error message for ioctl error.
v5:
Create a new folder drm/amd and move amdkfd from drm/radeon/ to drm/amd/
Remove scheduler_class from kfd_priv.h as it was never used
Add skeleton implementation of the Get Version IOCTL
v6:
Update module version to the correct number and remove the "default m" from the
Kconfig file
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
- KFD_IOC_GET_VERSION:
Retrieves the interface version of amdkfd
- KFD_IOC_CREATE_QUEUE:
Creates a usermode queue that runs on a specific GPU device
- KFD_IOC_DESTROY_QUEUE:
Destroys an existing usermode queue
- KFD_IOC_SET_MEMORY_POLICY:
Sets the memory policy of the default and alternate aperture of the
calling process
- KFD_IOC_GET_CLOCK_COUNTERS:
Retrieves counters (timestamps) of CPU and GPU
- KFD_IOC_GET_PROCESS_APERTURES:
Retrieves information about process apertures that were initialized
during the open() call of the amdkfd device
- KFD_IOC_UPDATE_QUEUE:
Updates configuration of an existing usermode queue
v3: Remove pragma pack and pmc ioctls. Added parameter for doorbell offset and
a comment on counters
v5:
Add define for AQL queues.
Fix arguments of Get Version IOCTL
Make IOCTL's structures to be the same size on 32/64 bit
v6: Change the version of the amdkfd-thunk interface
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
This patch adds the interface between the radeon driver and the amdkfd driver.
The interface implementation is contained in radeon_kfd.c and radeon_kfd.h.
The interface itself is represented by a pointer to struct
kfd_dev. The pointer is located inside radeon_device structure.
All the register accesses that amdkfd need are done using this interface. This
allows us to avoid direct register accesses in amdkfd proper, while also
avoiding locking between amdkfd and radeon.
The single exception is the doorbells that are used in both of the drivers.
However, because they are located in separate pci bar pages, the danger of
sharing registers between the drivers is minimal.
Having said that, we are planning to move the doorbells as well to radeon.
v3:
Add interface for sa manager init and fini. The init function will allocate a
buffer on system memory and pin it to the GART address space via the radeon sa
manager.
All mappings of buffers to GART address space are done via the radeon sa
manager. The interface of allocate memory will use the radeon sa manager to sub
allocate from the single buffer that was allocated during the init function.
Change lower_32/upper_32 calls to use linux macros
Add documentation for the interface
v4:
Change ptr field type in kgd_mem from uint32_t* to void* to match to type that
is returned by radeon_sa_bo_cpu_addr
v5:
Change format of mqd structure to work with latest KV firmware
Add support for AQL queues creation to enable working with open-source HSA
runtime.
Move generic kfd-->kgd interface and other generic kgd definitions to a generic
header file that will be used by AMD's radeon and amdgpu drivers
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
Implementing a lock for selecting and accessing shader engines and arrays.
This lock will make sure that radeon and amdkfd are not colliding when
accessing shader engines and arrays with GRBM_GFX_INDEX register.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
radeon and amdkfd share the doorbell aperture.
radeon sets it up, takes the doorbells required for its own rings
and reports the setup to amdkfd.
radeon reserved doorbells are at the start of the doorbell aperture.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
To support HSA on KV, we need to limit the number of vmids and pipes
that are available for radeon's use with KV.
This patch reserves VMIDs 8-15 for amdkfd (so radeon can only use VMIDs
0-7) and also makes radeon thinks that KV has only a single MEC with a single
pipe in it
v3: Use define for static vmid allocation in radeon
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
This patch fixes a bug in the accounting of the device_state.
In the current code, the device_state was put (decremented) too many times,
which sometimes lead to the driver getting stuck permanently in
put_device_state_wait(). That happen because the device_state->count would go
below zero, which is never supposed to happen.
The root cause is that the device_state was decremented in put_pasid_state()
and put_pasid_state_wait() but also in all the functions that call those
functions. Therefore, the device_state was decremented twice in each of these
code paths.
The fix is to decouple the device_state accounting from the pasid_state
accounting - remove the call to put_device_state() from the
put_pasid_state() and the put_pasid_state_wait())
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
Make use of the new invalidate_range mmu_notifier call-back and remove the
old logic of assigning an empty page-table between invalidate_range_start
and invalidate_range_end.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Tested-by: Oded Gabbay <oded.gabbay@amd.com>
Reviewed-by: Andrea Arcangeli <aarcange@redhat.com>
Reviewed-by: Jérôme Glisse <jglisse@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Jay Cornwall <Jay.Cornwall@amd.com>
Cc: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
Now that the mmu_notifier_invalidate_range() calls are in place, add the
callback to allow subsystems to register against it.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Reviewed-by: Andrea Arcangeli <aarcange@redhat.com>
Reviewed-by: Jérôme Glisse <jglisse@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Jay Cornwall <Jay.Cornwall@amd.com>
Cc: Oded Gabbay <Oded.Gabbay@amd.com>
Cc: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
Add calls to the new mmu_notifier_invalidate_range() function to all
places in the VMM that need it.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Reviewed-by: Andrea Arcangeli <aarcange@redhat.com>
Reviewed-by: Jérôme Glisse <jglisse@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Jay Cornwall <Jay.Cornwall@amd.com>
Cc: Oded Gabbay <Oded.Gabbay@amd.com>
Cc: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
This notifier closes an important gap in the current mmu_notifier
implementation, the existing callbacks are called too early or too late to
reliably manage a non-CPU TLB. Specifically, invalidate_range_start() is
called when all pages are still mapped and invalidate_range_end() when all
pages are unmapped and potentially freed.
This is fine when the users of the mmu_notifiers manage their own SoftTLB,
like KVM does. When the TLB is managed in software it is easy to wipe out
entries for a given range and prevent new entries to be established until
invalidate_range_end is called.
But when the user of mmu_notifiers has to manage a hardware TLB it can
still wipe out TLB entries in invalidate_range_start, but it can't make
sure that no new TLB entries in the given range are established between
invalidate_range_start and invalidate_range_end.
To avoid silent data corruption the entries in the non-CPU TLB need to be
flushed when the pages are unmapped (at this point in time no _new_ TLB
entries can be established in the non-CPU TLB) but not yet freed (as the
non-CPU TLB may still have _existing_ entries pointing to the pages about
to be freed).
To fix this problem we need to catch the moment when the Linux VMM flushes
remote TLBs (as a non-CPU TLB is not very CPU TLB), as this is the point
in time when the pages are unmapped but _not_ yet freed.
The mmu_notifier_invalidate_range() function aims to catch that moment.
IOMMU code will be one user of the notifier-callback. Currently this is
only the AMD IOMMUv2 driver, but its code is about to be more generalized
and converted to a generic IOMMU-API extension to fit the needs of similar
functionality in other IOMMUs as well.
The current attempt in the AMD IOMMUv2 driver to work around the
invalidate_range_start/end() shortcoming is to assign an empty page table
to the non-CPU TLB between any invalidata_range_start/end calls. With the
empty page-table assigned, every page-table walk to re-fill the non-CPU
TLB will cause a page-fault reported to the IOMMU driver via an interrupt,
possibly causing interrupt storms.
The page-fault handler in the AMD IOMMUv2 driver doesn't handle the fault
if an invalidate_range_start/end pair is active, it just reports back
SUCCESS to the device and let it refault the page. But existing hardware
(newer Radeon GPUs) that makes use of this feature don't re-fault
indefinitly, after a certain number of faults for the same address the
device enters a failure state and needs to be resetted.
To avoid the GPUs entering a failure state we need to get rid of the
empty-page-table workaround and use the mmu_notifier_invalidate_range()
function introduced with this patch.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Reviewed-by: Andrea Arcangeli <aarcange@redhat.com>
Reviewed-by: Jérôme Glisse <jglisse@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Jay Cornwall <Jay.Cornwall@amd.com>
Cc: Oded Gabbay <Oded.Gabbay@amd.com>
Cc: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
Radeon patches for 3.19. Christian has a number of GPUVM improvements
slated as well, but I'd like to wait until he gets back to work next week
to pull those in. Highlights of this pull:
- ttm performance improvements
- CI dpm fixes
* 'drm-next-3.19' of git://people.freedesktop.org/~agd5f/linux: (26 commits)
drm/radeon/si/ci: make u8 static arrays constant
drm/radeon: set power control in ci dpm enable
drm/radeon: powertune fixes for hawaii
drm/radeon: fix dpm mc init for certain hawaii boards
drm/radeon: set bootup pcie level to max for ci dpm
drm/radeon: fix default dpm state setup
drm/radeon: workaround a hw bug in bonaire pcie dpm
drm/radeon: fix mclk vddc configuration for cards for hawaii
drm/radeon: fix sclk DS enablement
drm/radeon: fix activity settings for sclk and mclk for CI
drm/radeon: improve mclk param calcuations for ci dpm
drm/radeon: fix dram timing for certain hawaii boards
drm/radeon: switch force state commands for CI
drm/radeon: fix for memory training on bonaire 0x6649
drm/radeon/ci: handle gpio controlled dpm features properly
drm/radeon: store the gpio shift as well
drm/radeon: export radeon_atombios_lookup_gpio
drm/radeon: fix typo in CI dpm disable
drm/radeon: rework CI dpm thermal setup
drm/radeon: rework SI dpm thermal setup
...
These two arrays don't change, just make them constant,
reduces data segment by a few bytes.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
DRM_MM_SEARCH_BEST gets the smallest hole which can fit the BO. That seems
against the idea of TTM_PL_FLAG_TOPDOWN:
* The smallest hole may be in the overall bottom of the area
* If the hole isn't much larger than the BO, it doesn't make much
difference whether the BO is placed at the bottom or at the top of the
hole
Reviewed-by: Lauri Kasanen <cand@gmx.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If the BO should be placed at the top of the area, we should start looking
for holes from the top.
Reviewed-by: Lauri Kasanen <cand@gmx.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
I wasn't sure if TTM_PL_FLAG_TOPDOWN works correctly with non-0 lpfn, but
AFAICT it does.
Reviewed-by: Lauri Kasanen <cand@gmx.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This avoids them getting in the way of BOs which might be accessed by
the CPU. They can still go to the CPU accessible part of VRAM though if
there's no space outside of it.
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Motivated by the per-plane locking I've gone through all the get*
ioctls and reduced the locking to the bare minimum required.
v2: Rebase and make it compile ...
v3: Review from Sean:
- Simplify return handling in getplane_res.
- Add a comment to getplane_res that the plane list is invariant and
can be walked locklessly.
v4: Actually git add.
Cc: Sean Paul <seanpaul@chromium.org>
Reviewed-by: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Turned out to be much simpler on top of my latest atomic stuff than
what I've feared. Some details:
- Drop the modeset_lock_all snakeoil in drm_plane_init. Same
justification as for the equivalent change in drm_crtc_init done in
commit d0fa1af40e
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon Sep 8 09:02:49 2014 +0200
drm: Drop modeset locking from crtc init function
Without these the drm_modeset_lock_init would fall over the exact
same way.
- Since the atomic core code wraps the locking switching it to
per-plane locks was a one-line change.
- For the legacy ioctls add a plane argument to the locking helper so
that we can grab the right plane lock (cursor or primary). Since the
universal cursor plane might not be there, or someone really crazy
might forgoe the primary plane even accept NULL.
- Add some locking WARN_ON to the atomic helpers for good paranoid
measure and to check that it all works out.
Tested on my exynos atomic hackfest with full lockdep checks and ww
backoff injection.
v2: I've forgotten about the load-detect code in i915.
v3: Thierry reported that in latest 3.18-rc vmwgfx doesn't compile any
more due to
commit 21e88620aa
Author: Rob Clark <robdclark@gmail.com>
Date: Thu Oct 30 13:39:04 2014 -0400
drm/vmwgfx: fix lock breakage
Rebased and fix this up.
Cc: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Reviewed-by: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
v1: original
v2: danvet's kerneldoc nitpicks
Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJUX/DqAAoJEHm+PkMAQRiGLtQH/iAt3fRHlYDXjaJian/KG1Cb
wVP0I+HWZmvVmmd0PzyaxCZLgRNwdmmYHEH4QLy2JwZ3jZfFHlxhy+hDWCgz+67t
bIzkLs0Pf1T4kJ2+r8qW2kBEz9PWJHGTQw7NTqZ++Ts3rPptBA6Fg4mEJ6fQigXy
qRIY68DpipUkXV9BWBWijnTmrvP5tt7JtPzBr4DC8frMjvWct8+XwYhc2k2tEv2j
LwLYb1OW6PUpPv2BQBfWjqqH77vYNQVhJwuwGcDe2YZdI0UFkDheL24+RbbPcZ4f
OnrLjJSSgzv6lBWkAaXZK7/WJ/JZbXxEqHzWZQ3xXoQov97bm7lEYJqqi5gDasQ=
=6Qpa
-----END PGP SIGNATURE-----
Merge tag 'v3.18-rc4' into drm-next
backmerge to get vmwgfx locking changes into next as the
conflict with per-plane locking.
These two didn't get documented properly, do so.
Pointed out by Daniel.
v1.1: add missing boilerplate (Daniel)
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
So here's my atomic series, finally all debugged&reviewed. Sean Paul has
done a full detailed pass over it all, and a lot of other people have
commented and provided feedback on some parts. Rob Clark also converted
msm over the w/e and seems happy. The only small thing is that Rob wants
to export the wait_for_vblank, which imo makes sense. Since there's other
stuff still to do I think we should apply Rob's patch (once it has grown
appropriate kerneldoc) later on top of this.
This is just the core<->driver interface plus a big pile of helpers. Short
recap of the main ideas:
- There are essentially three helper libraries in this patch set:
* Transitional helpers to use the new plane callbacks for legacy plane
updates and in the crtc helper's ->mode_set callback. These helpers are
only temporarily used to convert drivers to atomic, but they allow a
nice separation between changing the driver backend and switching to
the atomic commit logic.
* Legacy helpers to implement all the legacy driver entry points
(page_flip, set_config, plane vfuncs) on top of the new atomic driver
interface. These are completely driver agnostic. The reason for having
the legacy support as helpers is that drivers can switch step-by-step.
And they could e.g. even keep the legacy page_flip code around for some
old platforms where converting to full-blown atomic isn't worth it.
* Atomic helpers which implement the various new ->atomic_* driver
interfaces in terms of the revised crtc helper and new plane helper
hooks.
- The revised crtc helper implemenation essentially implements all the
lessons learned in the i915 modeset rework (when using the atomic helpers
only):
* Enable/disable sequence for a given config are always the same and
callbacks are always called in the same order. This contrast starkly
with the crtc helpers, where the sequence of operations is heavily
dependent on the previous config.
One corollary of this is that if the configuration of a crtc only
partially changes (e.g. a connector moves in a cloned config) the
helper code will still disable/enable the full display pipeline. This
is the only way to ensure that the enable/disable sequence is always
the same.
* It won't call disable or enable hooks more than once any more because
it lost track of state, thanks to the atomic state tracking. And if
drivers implement the ->reset hook properly (by either resetting the hw
or reading out the hw state into the atomic structures) this even
extends to the hardware state. So no more disable-me-harder kind of
nonsense.
* The only thing missing is the hw state readout/cross-check support, but
if drivers have hw state readout support in their ->reset handlers it's
simple to extend that to cross-check the hw state.
* The crtc->mode_set callback is gone and its replacement only sets crtc
timings and no longer updates the primary plane state. This way we can
finally implement primary planes properly.
- The new plane helpers should be suitable enough for pretty much
everything, and a perfect fit for hardware with GO bits. Even if they
don't fit the atomic helper library is rather flexible and exports all
the functions for the individual steps to drivers. So drivers can pick
what matches and implement their own magic for everything else.
- A big difference compared to all previous atomic series is that this one
doesn't implement async commit in a generic way. Imo driver requirements
for that are too diverse to create anything reasonable sane which would
actually work on a reasonable amount of different drivers. Also, we've
never had a helper library for page_flips even, so it's really hard to
know what might work and what's stupid without a bit of experience in the form
of a few driver implementations.
I think with the current flexibility for drivers to pick individual
stages and existing helpers like drm_flip_queue it's rather easy though
to implement proper async commit.
- There's a few other differences of minor importance to earlier atomic
series:
* Common/generic properties are parsed in the callers/core and not in
drivers, and passed to drivers by directly setting the right members in
atomic state structures. That greatly simplifies all the transitional
and legacy helpers an removes a lot of boilerplate code.
* There's no crazy trylock mode used for the async commit since these
helpers don't do async commit. A simple ordered flip queue of atomic
state updates should be sufficient for preventing concurrent hw access
anyway, as long as synchronous updates stall correctly with e.g.
flush_work_queue or similar function. Abusing locks to enforce ordering
isn't a good idea imo anyway.
* These helpers reuse the existing ->mode_fixup hooks in the atomic_check
callback. Which means that drivers need to adapat and move a lot less code
into their atomic_check callbacks.
Now this isn't everything needed in the drm core and helpers for full
atomic support. But it's enough to start with converting drivers, and
except for actually testing multiplane and multicrtc updates also enough to
implement full atomic updates. Still missing are:
- Per-plane locking. Since these helpers here encapsulate the locking
completely this should be fairly easy to implement.
- fbdev support for atomic_check/commit, so that multi-pipe finally works
sanely in fbcon.
- Adding and decoding shared/core properties. That just needs to be rebased
from Rob's latest patch series, with minor adjustments so that the
decoding happens in the core instead of in drivers.
- Actually adding the atomic ioctl. Again just rebasing Rob's latest patch
should be all that's needed.
- Resolving how to deal with DPMS in atomic. Atomic is a good excuse to fix up
the crazy semantics dpms currently has. I'm floating an RFC about this topic
already.
- Finally I couldn't test connector/encoder stealing properly since my test
vehicle here doesn't allow a connector on different crtcs. So drivers
which support this might see some surprises in that area. There is no semantic
change though in how encoder stealing and assignment works (or at least no
intended one), so I think the risk is minimal.
As just mentioned I've done a fake conversion of an existing driver using
crtc helpers to debug the helper code and validate the smooth transition
approach. And that smooth transition was the really big motivation for
this. It seems to actually work and consists of 3 phases:
Phase 1: Rework driver backend for crtc/plane helpers
The requirement here is that universal plane support is already implement. If
universal plane support isn't implement yet it might be better though to just do
it as part of this phase, directly using the new plane helpers. There are two
big things to do:
- Split up the existing ->update/disable_plane hooks into check/commit
hooks and extract the crtc-wide prep/flush parts (like setting/clearing
GO bits).
- The other big change is to split the crtc->mode_set hook into the plane
update (done using the plane helpers) and the crtc setup in a new
->mode_set_nofb hook.
When phase 1 is complete the driver implements all the new callbacks which
push the software state into hardware, but still using all the legacy entry
points and crtc helpers. The transitional helpers serve as impendance
mismatch here.
Phase 2: Rework state handling
This consists of rolling out the state handling helpers for planes, crtcs
and connectors and reviewing all ->mode_fixup and similar hooks to make
sure they don't depend upon implicit global state which might change in the
atomic world. Any such code must be moved into ->atomic_check functions which
just rely on the free-standing atomic state update structures.
This phase also adds a few small pieces of fixup code to make sure the
atomic state doesn't get out of sync in the legacy driver callbacks.
Phase 3: Roll out atomic support
Now it's just about replacing vfuncs with the ones provided by the helper
and filling out the small missing pieces (like atomic_check logic or async
commit support needed for page_flips). Due to the prep work in phase 1 no
changes to the driver backend functions should be required, and because of
the prep work in phase 2 atomic implementations can be rolled out
step-by-step. So if async commit ins't implemented yet page_flip can be
implemented with the legacy functions without wreaking havoc in the other
operations.
* tag 'topic/atomic-helpers-2014-11-09' of git://anongit.freedesktop.org/drm-intel:
drm/atomic: Refcounting for plane_state->fb
drm: Docbook integration and over sections for all the new helpers
drm/atomic-helpers: functions for state duplicate/destroy/reset
drm/atomic-helper: implement ->page_flip
drm/atomic-helpers: document how to implement async commit
drm/atomic: Integrate fence support
drm/atomic-helper: implementatations for legacy interfaces
drm: Atomic crtc/connector updates using crtc/plane helper interfaces
drm/crtc-helper: Transitional functions using atomic plane helpers
drm/plane-helper: transitional atomic plane helpers
drm: Add atomic/plane helpers
drm: Global atomic state handling
drm: Add atomic driver interface definitions for objects
drm/modeset_lock: document trylock_only in kerneldoc
drm: fixup kerneldoc in drm_crtc.h
drm: Pull drm_crtc.h into the kerneldoc template
drm: Move drm_crtc_init from drm_crtc.h to drm_plane_helper.h
- enable bpf syscall for compat
- cpu_suspend fix when checking the idle state type
- defconfig update
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQIcBAABAgAGBQJUXzboAAoJEGvWsS0AyF7xSP8P/1IGrDeWgT5+J9edv1B4QYm1
hw+X+T6Wywl4yPpkh7vljL64ajDa0wPFKu3t83xLpuFE2TEKT+EdABB9TJHAb8RN
hebRGYQXMCfDowTXS2QGK058rtHFVsdH3j8NWoUNwIBVW6Wq2Jcaw87V3E+K3TLL
FlYnxKen40N/IUJRe+M5nn7VUo7lXPbIiVGHb3gfPBq4yeLccIsVYWYNIYUEDMTQ
/BJTmbY9Hte0uZK5+q4cXqmNKiDgODskIFIOQUowJoKPR0cuhBraC61GIMSSbqRA
sbOs8yjadRcHsyjnFL3dlrhbRKzV9Vy4EDJ7l15m9j67MSs2UHiZ65AKHHyUbS1C
vIYMNc2gcTQH3kUyoWs+Dc5lrBhQ/T3rBz9bPnRu9SXmr0RjwnFkKcgOsm19B6cv
hqYuj8Pbjg5+Qoq0gRznQRzM3S3RN2WhzXSHt/gmHIKfcd8mtpHQDNRKayqlIWBS
paiCKkl7fn8uuZVuuzn4gSK50k3kvpJV6YddiTfLa7bYslIEpQyIAJXHurCY+y96
tTVIyO2vh1RzJ0b1HT2hHeDSDHmAXuHMQoc3Q4TNkHap+TTcI70+/TTCw23qswV9
dMuz7iN0/U96O2EiQlmWFpGhC1VZnjbzdIalnnEnfEfMTFA2vJqZr5eyzID4TfR0
b3PkGNf1/qgPiibwbJ3G
=CBMO
-----END PGP SIGNATURE-----
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 fixes from Catalin Marinas:
- enable bpf syscall for compat
- cpu_suspend fix when checking the idle state type
- defconfig update
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
arm64: defconfig: update defconfig for 3.18
arm64: compat: Enable bpf syscall
arm64: psci: fix cpu_suspend to check idle state type for index
Oded Gabbay