This will eventually be shared with newer asics to
reduce code duplication.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

No longer used now that we use the async dma engines or
CP DMA for bo copies.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

For r6xx+ asics.  This mirrors the behavior of pre-r6xx
asics.  We need to program the MC even if something
else in startup() fails.  Failure to do so results in
an unusable GPU.

Based on a fix from: Mark Kettenis <kettenis@openbsd.org>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Removing the clock/power or resetting the VCPU can cause
hangs if that happens in the middle of a register write.

Stall the memory and register bus before putting the VCPU
into reset. Keep it in reset when unloading the module or
suspending.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Can cause hangs when enabled in certain motherboards.
Set radeon.aspm=0 to disable aspm.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Helpful for debugging GPUVM errors as we can see what
hw block and page generated the fault in the log.

v2: simplify fault decoding
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>

Fix patching of vddc values for SI and enable manually forcing
clocks to default levels as per NI.

This improves the out of the box performance with SI asics.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

They cause problems with dynamic clocking.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Required for SI.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

TN has some additional powergating features beyond what is
supported on ON/LN.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Enables PCIE ASPM (Active State Power Management) on
evergreen-cayman asics.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This adds dpm support for trinity asics.  This includes:
- clockgating
- powergating
- dynamic engine clock scaling
- dynamic voltage scaling

set radeon.dpm=1 to enable it.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This adds dpm support for evergreen asics.  This includes:
- clockgating
- dynamic engine clock scaling
- dynamic memory clock scaling
- dynamic voltage scaling
- dynamic pcie gen1/gen2 switching (requires additional acpi support)

Set radeon.dpm=1 to enable.

v2: reduce stack usage, rename ulv struct
v3: fix thermal interrupt check notices by Jerome
v4: fix state enable
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Calculate the low and high watermarks based on the low and high
clocks for the current power state.  The dynamic pm hw will select
the appropriate watermark based on the internal dpm state.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This is required for certain advanced functionality.

v2: save/restore list takes dword offsets
v3: rebase on gpu reset changes
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Avoids confusion and duplication.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The current radeon driver initialization routines, when using KMS, are written
so that the IRQ installation routine is called before initializing the WB buffer
and the CP rings. With some ASICs, though, the IRQ routine tries to access the
GFX_INDEX ring causing a call to RREG32 with the value of -1 in
radeon_fence_read. This, in turn causes the system to completely hang with some
cards, requiring a hard reset.

A call stack that can cause such a hang looks like this (using rv515 ASIC for the
example here):
 * rv515_init (rv515.c)
 * radeon_irq_kms_init (radeon_irq_kms.c)
 * drm_irq_install (drm_irq.c)
 * radeon_driver_irq_preinstall_kms (radeon_irq_kms.c)
 * rs600_irq_process (rs600.c)
 * radeon_fence_process - due to SW interrupt (radeon_fence.c)
 * radeon_fence_read (radeon_fence.c)
 * hang due to RREG32(-1)

The patch moves the IRQ installation to the card startup routine, after the ring
has been initialized, but before the IRQ has been set. This fixes the issue, but
requires a check to see if the IRQ is already installed, as is the case in the
system resume codepath.
I have tested the patch on three machines using the rv515, the rv770 and the
evergreen ASIC. They worked without issues.

This seems to be a known issue and has been reported on several bug tracking
sites by various distributions (see links below). Most of reports recommend
booting the system with KMS disabled and then enabling KMS by reloading the
radeon module. For some reason, this was indeed a usable workaround, however,
UMS is now deprecated and disabled by default.

Bug reports:
https://bugzilla.redhat.com/show_bug.cgi?id=845745
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/561789
https://bbs.archlinux.org/viewtopic.php?id=156964Signed-off-by: NAdis Hamzić <adis@hamzadis.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

radeon currently uses a drm function to get the speed capabilities for
the bus, drm_pcie_get_speed_cap_mask. However, this is a non-standard
method of performing this detection and this patch changes it to use
the max_bus_speed attribute.

From: Lucas Kannebley Tavares <lucaskt@linux.vnet.ibm.com>
Signed-off-by: NKleber Sacilotto de Souza <klebers@linux.vnet.ibm.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Hainan has no display hardware:
- no DCE (crtc, uniphy, dac, etc.)
- no VGA

v2: fix bios fetch
v3: fix interrupts
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>

Add ULL prefix to avoid overflow.
Signed-off-by: NNiels Ole Salscheider <niels_ole@salscheider-online.de>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Instead of duplicating the code over and over again, just use a single
function to handle the clock calculations.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Just power down the PLL when we get a VCLK or DCLK of zero.
Enabling the bypass mode early should also allow us to
switch UVD clocks on the fly.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Just disabling the mem requests should be enough, but
that doesn't seem to work correctly on efi systems.

May fix:
https://bugs.freedesktop.org/show_bug.cgi?id=57567
https://bugs.freedesktop.org/show_bug.cgi?id=43655
https://bugzilla.kernel.org/show_bug.cgi?id=56441

v2: blank displays first, then disable.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Need to wait for the new addresses to take affect before
re-enabling the MC.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Properly wait for the next vblank region.  The previous
code didn't always wait long enough depending on the timing.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Avoids potential interrupt storms when the display is disabled.

May fix:
https://bugzilla.kernel.org/show_bug.cgi?id=56041Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

If the disabled rb mask register is not properly initialized
program a sane default based on the number of RBs for the
asic.  This avoids a potential divide by 0 when calculating
the backend mask.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Avoid confusion with the *REG32_P mask macro.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

v2: set UVD tiling config for rv730
Signed-off-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NJerome Glisse <jglisse@redhat.com>

v2: remove unneeded register definitions
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NJerome Glisse <jglisse@redhat.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

v2: write clk registers only once!
v3: update cg scratch register properly
v4: add TN support
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NJerome Glisse <jglisse@redhat.com>
Signed-off-by: NAlex 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: NChristian König <christian.koenig@amd.com>
Reviewed-by: NJerome Glisse <jglisse@redhat.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The MC is mostly likely busy (e.g., display requests), not hung
so no need to reset it.  Doing an MC reset is tricky and not
particularly reliable.  Fixes hangs in certain cases.
Reported-by: NJosh Boyer <jwboyer@gmail.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

vddci needs to track mclk for multi-head.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

The MC isn't part of the GPU per se.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

fetch the reset mask and check if the relevant ring flags
are set to determine whether the ring is hung or not.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

It's better to halt the engines before we disable the MC.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

When we attempt the reset the GPU, look at the status registers
to determine what blocks need to be reset.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>