Felix Kuehling [Wed, 16 Aug 2017 03:00:11 +0000 (23:00 -0400)]
drm/amdkfd: Allocate gtt_sa_bitmap in long units
gtt_sa_bitmap is accessed by bitmap functions, which operate on longs.
Therefore the array should be allocated in long units. Also round up
in case the number of bits is not a multiple of BITS_PER_LONG.
Felix Kuehling [Wed, 16 Aug 2017 03:00:09 +0000 (23:00 -0400)]
drm/amdkfd: Remove BUG_ONs for NULL pointer arguments
Remove BUG_ONs that check for NULL pointer arguments that are
dereferenced in the same function. Dereferencing the NULL pointer
will generate a BUG anyway, so the explicit check is redundant and
unnecessary overhead.
Kent Russell [Wed, 16 Aug 2017 03:00:08 +0000 (23:00 -0400)]
drm/amdkfd: Remove usage of alloc(sizeof(struct...
See https://kernel.org/doc/html/latest/process/coding-style.html
under "14) Allocating Memory" for rationale behind removing the
x=alloc(sizeof(struct) style and using x=alloc(sizeof(*x) instead
Kent Russell [Wed, 16 Aug 2017 03:00:07 +0000 (23:00 -0400)]
drm/amdkfd: Fix goto usage v2
Remove gotos that do not feature any common cleanup, and use gotos
instead of repeating cleanup commands.
According to kernel.org: "The goto statement comes in handy when a
function exits from multiple locations and some common work such as
cleanup has to be done. If there is no cleanup needed then just return
directly."
v2: Applied review suggestions in create_queue_nocpsch
Kent Russell [Wed, 16 Aug 2017 03:00:06 +0000 (23:00 -0400)]
drm/amdkfd: Change x==NULL/false references to !x
Upstream prefers the !x notation to x==NULL or x==false. Along those lines
change the ==true or !=NULL references as well. Also make the references
to !x the same, excluding () for readability.
Kent Russell [Wed, 16 Aug 2017 03:00:05 +0000 (23:00 -0400)]
drm/amdkfd: Consolidate and clean up log commands
Consolidate log commands so that dev_info(NULL, "Error...") uses the more
accurate pr_err, remove the module name from the log (can be seen via
dynamic debugging with +m), and the function name (can be seen via
dynamic debugging with +f). We also don't need debug messages saying
what function we're in. Those can be added by devs when needed
Don't print vendor and device ID in error messages. They are typically
the same for all GPUs in a multi-GPU system. So this doesn't add any
value to the message.
Lastly, remove parentheses around %d, %i and 0x%llX.
According to kernel.org:
"Printing numbers in parentheses (%d) adds no value and should be
avoided."
Kent Russell [Wed, 16 Aug 2017 03:00:04 +0000 (23:00 -0400)]
drm/amdkfd: Clean up KFD style errors and warnings v2
Using checkpatch.pl -f <file> showed a number of style issues. This
patch addresses as many of them as possible. Some long lines have been
left for readability, but attempts to minimize them have been made.
The most recent merge commit in airlied/drm-next has problems with
confusing old_crtc_state and new_crtc_state. Use the
for_each_oldnew_crtc_in_state macros to clean up the confusion,
and explicitly look at the correct state instead of looking at
asyh->state.
With these fixes it becomes more obvious what the code is trying to do,
which will hopefully prevent future confusion.
Jim Bride [Wed, 9 Aug 2017 19:48:53 +0000 (12:48 -0700)]
drm/i915/edp: Allow alternate fixed mode for eDP if available.
Some fixed resolution panels actually support more than one mode,
with the only thing different being the refresh rate. Having this
alternate mode available to us is desirable, because it allows us to
test PSR on panels whose setup time at the preferred mode is too long.
With this patch we allow the use of the alternate mode if it's
available and it was specifically requested.
v2 and v3: Rebase
v4: * Fix up some leaky mode stuff (Chris)
* Rebase
v5: * Fix a NULL pointer derefrence (David Weinehall)
v6: * Whitespace / spelling / checkpatch clean-up; no functional
change. (David)
* Rebase
Alex Deucher [Tue, 15 Aug 2017 14:34:55 +0000 (10:34 -0400)]
drm/amdgpu/gfx7: fix function name
Was using the wrong prefix (gmc rather than gfx). The function
is related to the gfx hw, not gmc. This also makes it consistent
with the naming in gfx8.
Dan Carpenter [Wed, 9 Aug 2017 10:30:46 +0000 (13:30 +0300)]
drm/amdgpu: Uninitialized variable in amdgpu_ttm_backend_bind()
My static checker complains that it's possible for "r" to be
uninitialized. It used to be set to zero so this returns it to the old
behavior.
Fixes: 98a7f88ce9a9 ("drm/amdgpu: bind BOs with GTT space allocated directly v2") Reviewed-by: Christian König <[email protected]> Reviewed-by: Michel Dänzer <[email protected]> Signed-off-by: Dan Carpenter <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
Trigger Huang [Tue, 8 Aug 2017 10:42:51 +0000 (06:42 -0400)]
drm/amdgpu: don't finish the ring if not initialized
If a ring is not initialized, it also should not be finished.
For example, in Vega10's SR-IOV environment, UVD's decode ring is not
initialized, but will be finnished in amdgpu_uvd_sw_fini, because UVD
driver put all the uvd decode ring's finish operation into
amdgpu_uvd_sw_fini function, while not uvd_vXXX_0_sw_fini. This will
lead to amdgpu module unloading failure.
Cihangir Akturk [Thu, 3 Aug 2017 11:58:35 +0000 (14:58 +0300)]
drm/radeon: switch to drm_*{get,put} helpers
drm_*_reference() and drm_*_unreference() functions are just
compatibility alias for drm_*_get() and drm_*_put() adn should not be
used by new code. So convert all users of compatibility functions to use
the new APIs.
Cihangir Akturk [Thu, 3 Aug 2017 11:58:16 +0000 (14:58 +0300)]
drm/amdgpu: switch to drm_*{get,put} helpers
drm_*_reference() and drm_*_unreference() functions are just
compatibility alias for drm_*_get() and drm_*_put() and should not be
used by new code. So convert all users of compatibility functions to use
the new APIs.
Jean Delvare [Sun, 30 Jul 2017 08:18:25 +0000 (10:18 +0200)]
drm/amdgpu: Fix undue fallthroughs in golden registers initialization
As I was staring at the si_init_golden_registers code, I noticed that
the Pitcairn initialization silently falls through the Cape Verde
initialization, and the Oland initialization falls through the Hainan
initialization. However there is no comment stating that this is
intentional, and the radeon driver doesn't have any such fallthrough,
so I suspect this is not supposed to happen.
Felix Kuehling [Fri, 28 Jul 2017 20:54:54 +0000 (16:54 -0400)]
drm/amdgpu: Fix KFD initialization for multi-GPU systems
kfd2kgd is device-specific, so it should not be a global variable.
Merge amdgpu_amdkfd_load_interface and amdgpu_amdkfd_device_probe
so that it's only needed as a local variable in one function.
We achieved that by setting S(SYSTEM) and P(PDE as PTE) bit to 1 for
PDEs and setting S bit to 1 for PTEs when the corresponding addresses
are not occupied by gpu driver allocated buffers.
Frank Min [Wed, 28 Jun 2017 12:02:04 +0000 (20:02 +0800)]
drm/amdgpu/sdma4: Enable sdma poll mem addr on vega10 for SRIOV
While doing flr on VFs, there is possibility to lost the doorbell
writing for sdma, so enable poll mem for sdma, then sdma fw would
check the pollmem holding wptr.
Frank Min [Mon, 12 Jun 2017 02:53:19 +0000 (10:53 +0800)]
drm/amdgpu: Clear vce&uvd ring wptr for SRIOV
MMSCH FW need to get the wptr from 0 after it get the mailbox request
from driver, since every time kick the mailbox, mmsch thinks that it
is the first time engine start to initialize.
Leo Liu [Thu, 27 Jul 2017 15:50:31 +0000 (11:50 -0400)]
drm/amdgpu: disable vcn power control for now
The dpm control for vcn has been moved to firmware, kernel always spins
"amdgpu: [powerplay] pp_dpm_powergate_uvd was not implemented", each
time when application runs, disable it for now till pg/cg. Also remove
clock manual setting when dpm disabled, which was inherited from uvd.
Jason Ekstrand [Tue, 15 Aug 2017 14:57:33 +0000 (15:57 +0100)]
drm/i915: Add support for drm syncobjs
This commit adds support for waiting on or signaling DRM syncobjs as
part of execbuf. It does so by hijacking the currently unused cliprects
pointer to instead point to an array of i915_gem_exec_fence structs
which containe a DRM syncobj and a flags parameter which specifies
whether to wait on it or to signal it. This implementation
theoretically allows for both flags to be set in which case it waits on
the dma_fence that was in the syncobj and then immediately replaces it
with the dma_fence from the current execbuf.
v2:
- Rebase on new syncobj API
v3:
- Pull everything out into helpers
- Do all allocation in gem_execbuffer2
- Pack the flags in the bottom 2 bits of the drm_syncobj*
v4:
- Prevent a potential race on syncobj->fence
Chris Wilson [Fri, 11 Aug 2017 10:57:31 +0000 (11:57 +0100)]
drm/i915: Handle full s64 precision for wait-ioctl
The wait-ioctl is optionally supplied a timeout with nanosecond
precision in a s64 field. We use nsecs_to_jiffies64() to convert that
into the jiffies consumed by the scheduler, but internally
nsecs_to_jiffies64() does not guard against overflow (as it's purpose is
for use by the scheduler and not drivers!). So we must guard against the
overflow ourselves, and in the process note that we may then return
much earlier than the timeout selected by the user, so don't report
ETIME unless we do hit the timeout. (Woe betold us though if the user
waits for a year (32bit) and the request is still not complete!)
v2: Refine overflow detection (to not include an overffow itself)
Chris Wilson [Fri, 11 Aug 2017 11:11:16 +0000 (12:11 +0100)]
drm/i915: Split obj->cache_coherent to track r/w
Another month, another story in the cache coherency saga. This time, we
come to the realisation that i915_gem_object_is_coherent() has been
reporting whether we can read from the target without requiring a cache
invalidate; but we were using it in places for testing whether we could
write into the object without requiring a cache flush. So split the
tracking into two, one to decide before reads, one after writes.
See commit e27ab73d17ef ("drm/i915: Mark CPU cache as dirty on every
transition for CPU writes") for the previous entry in this saga.
v2: Be verbose
v3: Remove unused function (i915_gem_object_is_coherent)
v4: Fix inverted coherency check prior to execbuf (from v2)
v5: Add comment for nasty code where we are optimising on gcc's behalf.
Lucas Stach [Tue, 8 Aug 2017 13:28:25 +0000 (15:28 +0200)]
drm/etnaviv: don't fail GPU bind when CONFIG_THERMAL isn't enabled
The stub functions returns -ENODEV when trying to register the cooling device,
thus failing the GPU bind, rendering the GPU subsystem unusable when
CONFIG_THERMAL isn't enabled.
Imre Deak [Mon, 14 Aug 2017 15:15:30 +0000 (18:15 +0300)]
drm/i915/hsw+: Add support for multiple power well regs
Future platforms increase the number of power wells which require
additional control registers. A convenient way to select the correct
register is to use the high bits of the power well ID as index. This
patch only prepares for this, while upcoming platform enabling patches
will add the actual new power well IDs and corresponding power well
control registers.
drm: omapdrm: Move PCD, LINEWIDTH and DOWNSCALE features to dispc driver
The FEAT_PARAM_DSS_PCD, FEAT_PARAM_LINEWIDTH and FEAT_PARAM_DOWNSCALE
features are specific to the DISPC, move them from the omap_dss_features
structure to the dispc driver.
projects / linux.git / log