diff options
author | Alex Deucher <alexander.deucher@amd.com> | 2012-07-17 14:02:29 -0400 |
---|---|---|
committer | Christian König <deathsimple@vodafone.de> | 2012-07-18 13:33:39 +0200 |
commit | a85a7da4c5922e6621526946d5a56b2194692fdd (patch) | |
tree | 68f23a3706ed03b194910c5beec0bf0ac8e211cf /drivers/gpu/drm/radeon/si.c | |
parent | 4ef72566f1035fef5abd0913100d13746e066ee3 (diff) | |
download | op-kernel-dev-a85a7da4c5922e6621526946d5a56b2194692fdd.zip op-kernel-dev-a85a7da4c5922e6621526946d5a56b2194692fdd.tar.gz |
drm/radeon: update ib_execute for SI (v2)
When submitting a CONST_IB, emit a SWITCH_BUFFER
packet before the CONST_IB. This isn't strictly necessary
(the driver will work fine without it), but is good practice
and allows for more flexible DE/CE sychronization options
in the future. Current userspace drivers do not take
advantage of the CE yet.
v2: - clean up code flow a bit
- no need to flush caches for CONST IB
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Diffstat (limited to 'drivers/gpu/drm/radeon/si.c')
-rw-r--r-- | drivers/gpu/drm/radeon/si.c | 49 |
1 files changed, 28 insertions, 21 deletions
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index 53e313b..2b12cae 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -1765,18 +1765,23 @@ void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) struct radeon_ring *ring = &rdev->ring[ib->ring]; u32 header; - if (ring->rptr_save_reg) { - uint32_t next_rptr = ring->wptr + 3 + 4 + 8; - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(ring, ((ring->rptr_save_reg - - PACKET3_SET_CONFIG_REG_START) >> 2)); - radeon_ring_write(ring, next_rptr); - } + if (ib->is_const_ib) { + /* set switch buffer packet before const IB */ + radeon_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0)); + radeon_ring_write(ring, 0); - if (ib->is_const_ib) header = PACKET3(PACKET3_INDIRECT_BUFFER_CONST, 2); - else + } else { + if (ring->rptr_save_reg) { + uint32_t next_rptr = ring->wptr + 3 + 4 + 8; + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, ((ring->rptr_save_reg - + PACKET3_SET_CONFIG_REG_START) >> 2)); + radeon_ring_write(ring, next_rptr); + } + header = PACKET3(PACKET3_INDIRECT_BUFFER, 2); + } radeon_ring_write(ring, header); radeon_ring_write(ring, @@ -1787,18 +1792,20 @@ void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFFFF); radeon_ring_write(ring, ib->length_dw | (ib->vm_id << 24)); - /* flush read cache over gart for this vmid */ - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(ring, ib->vm_id); - radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(ring, PACKET3_TCL1_ACTION_ENA | - PACKET3_TC_ACTION_ENA | - PACKET3_SH_KCACHE_ACTION_ENA | - PACKET3_SH_ICACHE_ACTION_ENA); - radeon_ring_write(ring, 0xFFFFFFFF); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 10); /* poll interval */ + if (!ib->is_const_ib) { + /* flush read cache over gart for this vmid */ + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2); + radeon_ring_write(ring, ib->vm_id); + radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + radeon_ring_write(ring, PACKET3_TCL1_ACTION_ENA | + PACKET3_TC_ACTION_ENA | + PACKET3_SH_KCACHE_ACTION_ENA | + PACKET3_SH_ICACHE_ACTION_ENA); + radeon_ring_write(ring, 0xFFFFFFFF); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 10); /* poll interval */ + } } /* |