linux-toradex.git/drivers/gpu/drm/v3d/v3d_debugfs.c, branch v5.3.2

drm/v3d: Dump V3D error debug registers in debugfs, and one at reset.

2019-05-16T16:24:43+00:00

Looking at a hang recently, I noticed these registers that might tell
me if something obvious was wrong.  They didn't help in this case, but
keep it around for the future.

Signed-off-by: Eric Anholt 
Link: https://patchwork.freedesktop.org/patch/msgid/20190419001014.23579-3-eric@anholt.net
Reviewed-by: Paul Kocialkowski

drm/v3d: Set the correct DMA mask according to the MMU's limits.

2019-05-16T16:24:34+00:00

On 7278, we've got 40 bits to work with.

Signed-off-by: Eric Anholt 
Link: https://patchwork.freedesktop.org/patch/msgid/20190419001014.23579-2-eric@anholt.net
Reviewed-by: Paul Kocialkowski

drm/v3d: Fix debugfs reads of MMU regs.

2019-05-16T16:24:19+00:00

They're in the hub, not the individual cores.

Signed-off-by: Eric Anholt 
Link: https://patchwork.freedesktop.org/patch/msgid/20190419001014.23579-1-eric@anholt.net
Reviewed-by: Paul Kocialkowski

drm/v3d: Add support for compute shader dispatch.

2019-04-18T16:54:10+00:00

The compute shader dispatch interface is pretty simple -- just pass in
the regs that userspace has passed us, with no CLs to run.  However,
with no CL to run it means that we need to do manual cache flushing of
the L2 after the HW execution completes (for SSBO, atomic, and
image_load_store writes that are the output of compute shaders).

This doesn't yet expose the L2 cache's ability to have a region of the
address space not write back to memory (which could be used for
shared_var storage).

So far, the Mesa side has been tested on V3D v4.2 simpenrose (passing
the ES31 tests), and on the kernel side on 7278 (failing atomic
compswap tests in a way that doesn't reproduce on simpenrose).

v2: Fix excessive allocation for the clean_job (reported by Dan
    Carpenter).  Keep refs on jobs until clean_job is finished, to
    avoid spurious MMU errors if the output BOs are freed by userspace
    before L2 cleaning is finished.

Signed-off-by: Eric Anholt 
Link: https://patchwork.freedesktop.org/patch/msgid/20190416225856.20264-4-eric@anholt.net
Acked-by: Rob Clark

drm/v3d: Make sure the GPU is on when measuring clocks.

2019-03-08T16:59:35+00:00

You'll get garbage measurements if the registers always read back
0xdeadbeef

Signed-off-by: Eric Anholt 
Link: https://patchwork.freedesktop.org/patch/msgid/20190220233658.986-3-eric@anholt.net
Reviewed-by: Dave Emett

drm/v3d: Skip debugfs dumping GCA on platforms without GCA.

2018-10-15T20:10:54+00:00

Fixes an oops reading this debugfs entry on BCM7278.

Signed-off-by: Eric Anholt 
Link: https://patchwork.freedesktop.org/patch/msgid/20180928232126.4332-4-eric@anholt.net
Fixes: 57692c94dcbe ("drm/v3d: Introduce a new DRM driver for Broadcom V3D V3.x+")
Cc: 
Reviewed-by: Boris Brezillon

drm/v3d: Add a little debugfs entry for measuring the core clock.

2018-10-15T20:10:47+00:00

This adds just enough performance counter support to measure the
clock.  We don't have linux kernel drivers for the clock driving the
HW, and this was useful for determining that the V3D HW is running on
a slow clock, not that the driver was slow.

Signed-off-by: Eric Anholt 
Link: https://patchwork.freedesktop.org/patch/msgid/20180928232126.4332-2-eric@anholt.net
Reviewed-by: Boris Brezillon

drm/v3d: Introduce a new DRM driver for Broadcom V3D V3.x+

2018-05-03T23:26:30+00:00

This driver will be used to support Mesa on the Broadcom 7268 and 7278
platforms.

V3D 3.3 introduces an MMU, which means we no longer need CMA or vc4's
complicated CL/shader validation scheme.  This massively changes the
GEM behavior, so I've forked off to a new driver.

v2: Mark SUBMIT_CL as needing DRM_AUTH.  coccinelle fixes from kbuild
    test robot. Drop personal git link from MAINTAINERS.  Don't
    double-map dma-buf imported BOs.  Add kerneldoc about needing MMU
    eviction.  Drop prime vmap/unmap stubs.  Delay mmap offset setup
    to mmap time.  Use drm_dev_init instead of _alloc.  Use
    ktime_get() for wait_bo timeouts.  Drop drm_can_sleep() usage,
    since we don't modeset.  Switch page tables back to WC (debug
    change to coherent had slipped in).  Switch
    drm_gem_object_unreference_unlocked() to
    drm_gem_object_put_unlocked().  Simplify overflow mem handling by
    not sharing overflow mem between jobs.
v3: no changes
v4: align submit_cl to 64 bits (review by airlied), check zero flags in
    other ioctls.

Signed-off-by: Eric Anholt 
Acked-by: Daniel Vetter  (v4)
Acked-by: Dave Airlie  (v3, requested submit_cl change)
Link: https://patchwork.freedesktop.org/patch/msgid/20180430181058.30181-3-eric@anholt.net