Am 26.06.20 um 18:12 schrieb Alex Jivin:
Adding a delay between writing to UVD control register and reading from it.
This is to allow the HW to process the write command.
Signed-off-by: Alex Jivin <[email protected]>
Suggested-By: Luben Tukov <[email protected]>
---
drivers/gpu/drm/amd/amdgpu/si.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/si.c b/drivers/gpu/drm/amd/amdgpu/si.c
index 9d7b4ccd17b8..42cdc14fb79d 100644
--- a/drivers/gpu/drm/amd/amdgpu/si.c
+++ b/drivers/gpu/drm/amd/amdgpu/si.c
@@ -1435,6 +1435,12 @@ static int si_uvd_send_upll_ctlreq(struct amdgpu_device
*adev,
/* Assert UPLL_CTLREQ */
WREG32_P(cg_upll_func_cntl, UPLL_CTLREQ_MASK, ~UPLL_CTLREQ_MASK);
+ /* Waiting for HW to process the previous write.
+ * This is to give a chance to HW to act before
+ * the first read is done.
+ */
+ mdelay(1);
+
Mhm, that is most likely not a good idea.
We need to issue a read after the write to make sure that the stuff is
send out to the hardware.
Adding a delay here is probably just postponing that. Do we have some
note in the documentation that this is necessary?
Christian.
/* Wait for CTLACK and CTLACK2 to get asserted */
for (i = 0; i < SI_MAX_CTLACKS_ASSERTION_WAIT; ++i) {
uint32_t mask = UPLL_CTLACK_MASK | UPLL_CTLACK2_MASK;
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
