We consistently use the _irq_handler postfix for functions called in
hardirq context. Especially when it's a non-static function hardirq is
a crazy enough calling context to warrant this level of ocd. So rename
it.

Cc: Thomas Daniel <thomas.daniel@intel.com>
Reviewed-by: NThomas Daniel <thomas.daniel@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

Up until now, we have pinned every logical ring context backing object
during creation, and left it pinned until destruction. This made my life
easier, but it's a harmful thing to do, because we cause fragmentation
of the GGTT (and, eventually, we would run out of space).

This patch makes the pinning on-demand: the backing objects of the two
contexts that are written to the ELSP are pinned right before submission
and unpinned once the hardware is done with them. The only context that
is still pinned regardless is the global default one, so that the HWS can
still be accessed in the same way (ring->status_page).

v2: In the early version of this patch, we were pinning the context as
we put it into the ELSP: on the one hand, this is very efficient because
only a maximum two contexts are pinned at any given time, but on the other
hand, we cannot really pin in interrupt time :(

v3: Use a mutex rather than atomic_t to protect pin count to avoid races.
Do not unpin default context in free_request.

v4: Break out pin and unpin into functions.  Fix style problems reported
by checkpatch

v5: Remove unpin_lock as all pinning and unpinning is done with the struct
mutex already locked.  Add WARN_ONs to make sure this is the case in future.

Issue: VIZ-4277
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Signed-off-by: NThomas Daniel <thomas.daniel@intel.com>
Reviewed-by: NAkash Goel <akash.goels@gmail.com>
Reviewed-by: Deepak S<deepak.s@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

No longer create a work item to clean each execlist queue item.
Instead, move retired execlist requests to a queue and clean up the
items during retire_requests.

v2: Fix legacy ring path broken during overzealous cleanup

v3: Update idle detection to take execlists queue into account

v4: Grab execlist lock when checking queue state

v5: Fix leaking requests by freeing in execlists_retire_requests.

Issue: VIZ-4274
Signed-off-by: NThomas Daniel <thomas.daniel@intel.com>
Reviewed-by: NDeepak S <deepak.s@linux.intel.com>
Reviewed-by: NAkash Goel <akash.goels@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The batchbuffer that sets the render context state is submitted
in a different way, and from different places.

We needed to make both the render state preparation and free functions
outside accesible, and namespace accordingly. This mess is so that all
LR, LRC and Execlists functionality can go together in intel_lrc.c: we
can fix all of this later on, once the interfaces are clear.

v2: Create a separate ctx->rcs_initialized for the Execlists case, as
suggested by Chris Wilson.
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>

v3: Setup ring status page in lr_context_deferred_create when the
default context is being created. This means that the render state
init for the default context is no longer a special case.  Execute
deferred creation of the default context at the end of
logical_ring_init to allow the render state commands to be submitted.
Fix style errors reported by checkpatch. Rebased.
Signed-off-by: NThomas Daniel <thomas.daniel@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Add theory of operation notes to intel_lrc.c and comments to externally
visible functions.

v2: Add notes on logical ring context creation.

v3: Use kerneldoc.

v4: Integrate it in the DocBook template.

Signed-off-by: Thomas Daniel <thomas.daniel@intel.com> (v1)
Signed-off-by: Oscar Mateo <oscar.mateo@intel.com> (v2, v3)
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
[danvet: Drop hunk about render ring init function since that's not
yet merged.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

v2: Warn and return if LRCs are not enabled.

v3: Grab the Execlists spinlock (noticed by Daniel Vetter).
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>

v4: Lock the struct mutex for atomic state capture
Signed-off-by: NThomas Daniel <thomas.daniel@intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
[danvet: Checkpatch.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

In the current Execlists feeding mechanism, full preemption is not
supported yet: only lite-restores are allowed (this is: the GPU
simply samples a new tail pointer for the context currently in
execution).

But we have identified an scenario in which a full preemption occurs:
1) We submit two contexts for execution (A & B).
2) The GPU finishes with the first one (A), switches to the second one
(B) and informs us.
3) We submit B again (hoping to cause a lite restore) together with C,
but in the time we spend writing to the ELSP, the GPU finishes B.
4) The GPU start executing B again (since we told it so).
5) We receive a B finished interrupt and, mistakenly, we submit C (again)
and D, causing a full preemption of B.

The race is avoided by keeping track of how many times a context has been
submitted to the hardware and by better discriminating the received context
switch interrupts: in the example, when we have submitted B twice, we won´t
submit C and D as soon as we receive the notification that B is completed
because we were expecting to get a LITE_RESTORE and we didn´t, so we know a
second completion will be received shortly.

Without this explicit checking, somehow, the batch buffer execution order
gets messed with. This can be verified with the IGT test I sent together with
the series. I don´t know the exact mechanism by which the pre-emption messes
with the execution order but, since other people is working on the Scheduler
+ Preemption on Execlists, I didn´t try to fix it. In these series, only Lite
Restores are supported (other kind of preemptions WARN).

v2: elsp_submitted belongs in the new intel_ctx_submit_request. Several
rebase changes.

v3: Clarify how the race is avoided, as requested by Daniel.
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
[danvet: Align function parameters ...]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Handle all context status events in the context status buffer on every
context switch interrupt. We only remove work from the execlist queue
after a context status buffer reports that it has completed and we only
attempt to schedule new contexts on interrupt when a previously submitted
context completes (unless no contexts are queued, which means the GPU is
free).

We canot call intel_runtime_pm_get() in an interrupt (or with a spinlock
grabbed, FWIW), because it might sleep, which is not a nice thing to do.
Instead, do the runtime_pm get/put together with the create/destroy request,
and handle the forcewake get/put directly.
Signed-off-by: NThomas Daniel <thomas.daniel@intel.com>

v2: Unreferencing the context when we are freeing the request might free
the backing bo, which requires the struct_mutex to be grabbed, so defer
unreferencing and freeing to a bottom half.

v3:
- Ack the interrupt inmediately, before trying to handle it (fix for
missing interrupts by Bob Beckett <robert.beckett@intel.com>).
- Update the Context Status Buffer Read Pointer, just in case (spotted
by Damien Lespiau).

v4: New namespace and multiple rebase changes.

v5: Squash with "drm/i915/bdw: Do not call intel_runtime_pm_get() in an
interrupt", as suggested by Daniel.
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
[danvet: Checkpatch ...]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Context switch (and execlist submission) should happen only when
other contexts are not active, otherwise pre-emption occurs.

To assure this, we place context switch requests in a queue and those
request are later consumed when the right context switch interrupt is
received (still TODO).

v2: Use a spinlock, do not remove the requests on unqueue (wait for
context switch completion).
Signed-off-by: NThomas Daniel <thomas.daniel@intel.com>

v3: Several rebases and code changes. Use unique ID.

v4:
- Move the queue/lock init to the late ring initialization.
- Damien's kmalloc review comments: check return, use sizeof(*req),
do not cast.

v5:
- Do not reuse drm_i915_gem_request. Instead, create our own.
- New namespace.

Signed-off-by: Michel Thierry <michel.thierry@intel.com> (v1)
Signed-off-by: Oscar Mateo <oscar.mateo@intel.com> (v2-v5)
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
[davnet: Checkpatch + wash-up s/BUG_ON/WARN_ON/.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

A context switch occurs by submitting a context descriptor to the
ExecList Submission Port. Given that we can now initialize a context,
it's possible to begin implementing the context switch by creating the
descriptor and submitting it to ELSP (actually two, since the ELSP
has two ports).

The context object must be mapped in the GGTT, which means it must exist
in the 0-4GB graphics VA range.
Signed-off-by: NBen Widawsky <ben@bwidawsk.net>

v2: This code has changed quite a lot in various rebases. Of particular
importance is that now we use the globally unique Submission ID to send
to the hardware. Also, context pages are now pinned unconditionally to
GGTT, so there is no need to bind them.

v3: Use LRCA[31:12] as hwCtxId[19:0]. This guarantees that the HW context
ID we submit to the ELSP is globally unique and != 0 (Bspec requirements
of the software use-only bits of the Context ID in the Context Descriptor
Format) without the hassle of the previous submission Id construction.
Also, re-add the ELSP porting read (it was dropped somewhere during the
rebases).

v4:
- Squash with "drm/i915/bdw: Add forcewake lock around ELSP writes" (BSPEC
  says: "SW must set Force Wakeup bit to prevent GT from entering C6 while
  ELSP writes are in progress") as noted by Thomas Daniel
  (thomas.daniel@intel.com).
- Rename functions and use an execlists/intel_execlists_ namespace.
- The BUG_ON only checked that the LRCA was <32 bits, but it didn't make
  sure that it was properly aligned. Spotted by Alistair Mcaulay
  <alistair.mcaulay@intel.com>.

v5:
- Improved source code comments as suggested by Chris Wilson.
- No need to abstract submit_ctx away, as pointed by Brad Volkin.
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
[danvet: Checkpatch. Sigh.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

On a previous iteration of this patch, I created an Execlists
version of __i915_add_request and asbtracted it away as a
vfunc. Daniel Vetter wondered then why that was needed:

"with the clean split in command submission I expect every
function to know wether it'll submit to an lrc (everything in
intel_lrc.c) or wether it'll submit to a legacy ring (existing
code), so I don't see a need for an add_request vfunc."

The honest, hairy truth is that this patch is the glue keeping
the whole logical ring puzzle together:

- i915_add_request is used by intel_ring_idle, which in turn is
  used by i915_gpu_idle, which in turn is used in several places
  inside the eviction and gtt codes.
- Also, it is used by i915_gem_check_olr, which is littered all
  over i915_gem.c
- ...

If I were to duplicate all the code that directly or indirectly
uses __i915_add_request, I'll end up creating a separate driver.

To show the differences between the existing legacy version and
the new Execlists one, this time I have special-cased
__i915_add_request instead of adding an add_request vfunc. I
hope this helps to untangle this Gordian knot.
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
[danvet: Adjust to ringbuf->FIXME_lrc_ctx per the discussion with
Thomas Daniel.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Well, new-ish: if all this code looks familiar, that's because it's
a clone of the existing submission mechanism (with some modifications
here and there to adapt it to LRCs and Execlists).

And why did we do this instead of reusing code, one might wonder?
Well, there are some fears that the differences are big enough that
they will end up breaking all platforms.

Also, Execlists offer several advantages, like control over when the
GPU is done with a given workload, that can help simplify the
submission mechanism, no doubt. I am interested in getting Execlists
to work first and foremost, but in the future this parallel submission
mechanism will help us to fine tune the mechanism without affecting
old gens.

v2: Pass the ringbuffer only (whenever possible).
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
[danvet: Appease checkpatch. Again. And drop the legacy sarea gunk
that somehow crept in.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Execlists are indeed a brave new world with respect to workload
submission to the GPU.

In previous version of these series, I have tried to impact the
legacy ringbuffer submission path as little as possible (mostly,
passing the context around and using the correct ringbuffer when I
needed one) but Daniel is afraid (probably with a reason) that
these changes and, especially, future ones, will end up breaking
older gens.

This commit and some others coming next will try to limit the
damage by creating an alternative path for workload submission.
The first step is here: laying out a new ring init/fini.
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

For the moment this is just a placeholder, but it shows one of the
main differences between the good ol' HW contexts and the shiny
new Logical Ring Contexts: LR contexts allocate  and free their
own backing objects. Another difference is that the allocation is
deferred (as the create function name suggests), but that does not
happen in this patch yet, because for the moment we are only dealing
with the default context.

Early in the series we had our own gen8_gem_context_init/fini
functions, but the truth is they now look almost the same as the
legacy hw context init/fini functions. We can always split them
later if this ceases to be the case.

Also, we do not fall back to legacy ringbuffers when logical ring
context initialization fails (not very likely to happen and, even
if it does, hw contexts would probably fail as well).

v2: Daniel says "explain, do not showcase".
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
[danvet: s/BUG_ON/WARN_ON/.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

GEN8 brings an expansion of the HW contexts: "Logical Ring Contexts".
These expanded contexts enable a number of new abilities, especially
"Execlists".

The macro is defined to off until we have things in place to hope to
work.

v2: Rename "advanced contexts" to the more correct "logical ring
contexts".

v3: Add a module parameter to enable execlists. Execlist are relatively
new, and so it'd be wise to be able to switch back to ring submission
to debug subtle problems that will inevitably arise.

v4: Add an intel_enable_execlists function.

v5: Sanitize early, as suggested by Daniel. Remove lrc_enabled.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v1)
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> (v3)
Signed-off-by: Oscar Mateo <oscar.mateo@intel.com> (v2, v4 & v5)
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Some legacy HW context code assumptions don't make sense for this new
submission method, so we will place this stuff in a separate file.

Note for reviewers: I've carefully considered the best name for this file
and this was my best option (other possibilities were intel_lr_context.c
or intel_execlist.c). I am open to a certain bikeshedding on this matter,
anyway.

And some point in time, it would be a good idea to split intel_lrc.c/.h
even further, but for the moment just shove everything together.

v2: Change to intel_lrc.c

v3: Squash together with the header file addition
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Allows us to use the 3D engine for memory management
and allows us to use vram beyond the BAR aperture.

v2: fix copy paste typo
Reported-by: NNils Wallménius <nils.wallmenius@gmail.com>
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Cayman asics have 3 ring buffers:
ring 0 supports both gfx and compute
rings 1 and 2 are compute only

At the moment we only support ring 0.
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

This patch implements blit support for bo moves using
the 3D engine.  It uses the same method as r6xx/r7xx:
- store the base state in an IB
- emit variable state and vertex buffers to do the blit

This allows the hw to move bos using the 3D engine and allows
full use of vram beyond the pci aperture size.
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Noticed while adding evergreen blit support.
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

This adds the r600 KMS + CS support to the Linux kernel.

The r600 TTM support is quite basic and still needs more
work esp around using interrupts, but the polled fencing
should work okay for now.

Also currently TTM is using memcpy to do VRAM moves,
the code is here to use a 3D blit to do this, but
isn't fully debugged yet.

Authors:
Alex Deucher <alexdeucher@gmail.com>
Dave Airlie <airlied@redhat.com>
Jerome Glisse <jglisse@redhat.com>
Signed-off-by: NJerome Glisse <jglisse@redhat.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>