On CIK, the compute rings work slightly differently than
on previous asics, however the basic concepts are the same.
The main differences:
- New MEC engines for compute queues
- Multiple queues per MEC:
- CI/KB: 1 MEC, 4 pipes per MEC, 8 queues per pipe = 32 queues
- KV: 2 MEC, 4 pipes per MEC, 8 queues per pipe = 64 queues
- Queues can be allocated and scheduled by another queue
- New doorbell aperture allows you to assign space in the aperture
for the wptr which allows for userspace access to queues
v2: add wptr shadow, fix eop setup
v3: fix comment
v4: switch to new callback method
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Sets up the GFX ring and loads ucode for GFX and Compute.
Todo:
- handle compute queue setup.
v2: add documentation
v3: integrate with latest reset changes
v4: additional init fixes
v5: scratch reg write back no longer supported on CIK
v6: properly set CP_RB0_BASE_HI
v7: rebase
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
That not only saves some power, but also solves problems with
older chips where an idle UVD block on higher clocks can
cause problems.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Just everything needed to decode videos using UVD.
v6: just all the bugfixes and support for R7xx-SI merged in one patch
v7: UVD_CGC_GATE is a write only register, lockup detection fix
v8: split out VRAM fallback changes, remove support for RV770,
add support for HEMLOCK, add buffer sizes checks
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Let the CS module decide if we can fall back to VRAM or not.
v2: remove unintended change
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
v2: update error message and comment
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
For very large page table updates, we can exceed the
size of the ring. To avoid this, use an IB to perform
the page table update.
v2(ck): cleanup the IB infrastructure and the use it instead
of filling the struct ourself.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
next_reloc function does the same thing in all ASICs with
the exception of R600 which has a special case in legacy mode.
Pull out the common function in preparation for refactoring.
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This function is not limited to r100, but it can dump a
(raw) packet for any ASIC. Rename it accordingly and move
its declaration to radeon.h
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Once we factored out radeon_cs_packet_parse function,
evergreen_cs_next_is_pkt3_nop and r600_cs_next_is_pkt3_nop
functions became identical, so they can be factored out
into a common function.
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
CS packet parse functions have a lot of in common across
all ASICs. Implement a common function and take care of
small differences between families inside the function.
This patch is a prep for major refactoring and consolidation
of CS parsing code.
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
length_dw field was assigned twice. While at it, move user_ptr
assignment together with all other assignments to p->chunks[i]
structure to make the code more readable.
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If one (but not both) allocations of p->chunks[].kpage[]
in radeon_cs_parser_init fail, the error path will free
the successfully allocated page, but leave a stale pointer
value in the kpage[] field. This will later cause a
double-free when radeon_cs_parser_fini is called.
This patch fixes the issue by forcing both pointers to NULL
after kfree in the error path.
The circumstances under which the problem happens are very
rare. The card must be AGP and the system must run out of
kmalloc area just at the right time so that one allocation
succeeds, while the other fails.
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Cc: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Index into chunks[] array doesn't look right.
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
In UMS mode parser->rdev is NULL, so dereferencing
will cause an oops.
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This enables the functionality added in the previous
patches. Userspace acceleration drivers can use the
CS ioctl to submit command buffers to the async DMA
rings.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Allows us to use the DMA ring from userspace.
DMA doesn't have a good NOP packet in which to embed the
reloc idx, so userspace has to add a reloc for each
buffer used and order them to match the command stream.
v2: fix address bounds checking, reloc indexing
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Make it possible to allocate a persistent page table.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Pull drm merge (part 1) from Dave Airlie:
"So first of all my tree and uapi stuff has a conflict mess, its my
fault as the nouveau stuff didn't hit -next as were trying to rebase
regressions out of it before we merged.
Highlights:
- SH mobile modesetting driver and associated helpers
- some DRM core documentation
- i915 modesetting rework, haswell hdmi, haswell and vlv fixes, write
combined pte writing, ilk rc6 support,
- nouveau: major driver rework into a hw core driver, makes features
like SLI a lot saner to implement,
- psb: add eDP/DP support for Cedarview
- radeon: 2 layer page tables, async VM pte updates, better PLL
selection for > 2 screens, better ACPI interactions
The rest is general grab bag of fixes.
So why part 1? well I have the exynos pull req which came in a bit
late but was waiting for me to do something they shouldn't have and it
looks fairly safe, and David Howells has some more header cleanups
he'd like me to pull, that seem like a good idea, but I'd like to get
this merge out of the way so -next dosen't get blocked."
Tons of conflicts mostly due to silly include line changes, but mostly
mindless. A few other small semantic conflicts too, noted from Dave's
pre-merged branch.
* 'drm-next' of git://people.freedesktop.org/~airlied/linux: (447 commits)
drm/nv98/crypt: fix fuc build with latest envyas
drm/nouveau/devinit: fixup various issues with subdev ctor/init ordering
drm/nv41/vm: fix and enable use of "real" pciegart
drm/nv44/vm: fix and enable use of "real" pciegart
drm/nv04/dmaobj: fixup vm target handling in preparation for nv4x pcie
drm/nouveau: store supported dma mask in vmmgr
drm/nvc0/ibus: initial implementation of subdev
drm/nouveau/therm: add support for fan-control modes
drm/nouveau/hwmon: rename pwm0* to pmw1* to follow hwmon's rules
drm/nouveau/therm: calculate the pwm divisor on nv50+
drm/nouveau/fan: rewrite the fan tachometer driver to get more precision, faster
drm/nouveau/therm: move thermal-related functions to the therm subdev
drm/nouveau/bios: parse the pwm divisor from the perf table
drm/nouveau/therm: use the EXTDEV table to detect i2c monitoring devices
drm/nouveau/therm: rework thermal table parsing
drm/nouveau/gpio: expose the PWM/TOGGLE parameter found in the gpio vbios table
drm/nouveau: fix pm initialization order
drm/nouveau/bios: check that fixed tvdac gpio data is valid before using it
drm/nouveau: log channel debug/error messages from client object rather than drm client
drm/nouveau: have drm debugging macros build on top of core macros
...
Convert #include "..." to #include <path/...> in drivers/gpu/.
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Dave Jones <davej@redhat.com>
When a VM is used on more than one ring we need to
sync to the last user.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Let's allow GCC to optimize better.
This exposed some five unused functions, but this patch doesn't remove them.
Signed-off-by: Lauri Kasanen <cand@gmx.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Make sure that the ib bo is bound and is page table is up to date
in the virtual address space.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Removing the need to wait for anything.
Still not ideal, since we need to free pt on va remove.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Move binding onto the ring, simplifying handling a bit.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Move flushing the VMs as function into the rings.
First step to make VM operations async.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Store a reference to the VM into the IB structure, that
makes calculating the IBs address a bit less complicated.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Virtual address need to be fenced to know when we can safely remove it.
This patch also properly clear the pagetable. Previously it was
serouisly broken.
Kernel 3.5/3.4 need a similar patch but adapted for difference in mutex locking.
v2: For to update pagetable when unbinding bo (don't bailout if
bo_va->valid is true).
v3: Add kernel 3.5/3.4 comment.
v4: Fix compilation warnings.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Const IBs are executed on the CE not the CP, so we can't
fence them in the normal way.
So submit them directly before the IB instead, just as
the documentation says.
v2: keep the extra documentation
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
GPU reset need to be exclusive, one happening at a time. For this
add a rw semaphore so that any path that trigger GPU activities
have to take the semaphore as a reader thus allowing concurency.
The GPU reset path take the semaphore as a writer ensuring that
no concurrent reset take place.
v2: init rw semaphore
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Don't return success if scheduling the IB fails, otherwise
we end up with an oops in ttm_eu_fence_buffer_objects.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Try to remove or replace the cs_mutex with a
vm_mutex where it is still needed.
v2: fix locking order
v3: rebased on drm-next
Signed-off-by: Christian König <deathsimple@vodafone.de>
Move inter ring syncing with semaphores into the
existing ring allocations, with that we need to
lock the ring mutex only once.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
radeon_cs_parser_init is called by both the legacy UMS
CS ioctl and the KMS CS ioctl. Protect KMS specific
pieces of the code by checking that rdev is not NULL.
Reported-by: Michael Burian <michael.burian@sbg.at>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
Just move its only caller into the same file as it and make it static.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
No need to malloc it any more.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
It never really belonged there in the first place.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Instead of offset + size keep start and end offset directly.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
We need to sync with the GFX ring as ttm might have schedule bo move
on it and new command scheduled for other ring need to wait for bo
data to be in place.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Instead of all this humpy pumpy with recursive
mutex (which also fixes only halve of the problem)
move the actual gpu reset out of the fence code,
return -EDEADLK and then reset the gpu in the
calling ioctl function.
v2: Split removal of radeon_mutex into separate patch.
Return -EAGAIN if reset is successful.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Rings need to lock in order, otherwise
the ring subsystem can deadlock.
v2: fix error handling and number of locked doublewords.
v3: stop creating unneeded semaphores.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
On coherent systems (not-AGP) the IB should be in cached memory so should
be just as fast, so we can avoid copying to temporary pages and just use it
directly.
provides minor speedups on rv530: gears ~1820->1860, ipers: 29.9->30.6,
but always good to use less CPU if we can.
v3: cleanup unneeded bits.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Very basic implementation for picking the ring priority.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This adds a new chunk id to the CS ioctl to support the
INDIRECT_BUFFER_CONST packet.
On SI, the CP adds a new engine called the CE (Constant Engine)
which runs simulatenously with the DE (Drawing Engine, formerly
called the ME). This allows the CP to process two related IBs
simultaneously. The CE is tasked with loading the constant data
(constant buffers, resource descriptors, samplers, etc.) while
the DE loads context register state and issues drawing commands.
It's up to the userspace application to sychronize the CE and the
DE using special synchronization packets.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
The function radeon_cs_parser_init is only called from two places, in
drivers/gpu/drm/radeon/radeon_cs.c and drivers/gpu/drm/radeon/r600_cs.c.
In each case, if the call fails another function is called that frees all
of the kdata and dpage information in the chunks array. So this
information should not be freed in radeon_cs_parser_init as well.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
