--- /dev/null
+From f232d9ec029ce3e2543b05213e2979e01e503408 Mon Sep 17 00:00:00 2001
+From: Lucas Stach <l.stach@pengutronix.de>
+Date: Wed, 26 Feb 2020 16:27:08 +0100
+Subject: [PATCH] drm/etnaviv: fix TS cache flushing on GPUs with BLT engine
+
+As seen in the Vivante kernel driver, most GPUs with the BLT engine have
+a broken TS cache flush. The workaround is to temporarily set the BLT
+command to CLEAR_IMAGE, without actually executing the clear. Apparently
+this state change is enough to trigger the required TS cache flush. As
+the BLT engine is completely asychronous, we also need a few more stall
+states to synchronize the flush with the frontend.
+
+Root-caused-by: Jonathan Marek <jonathan@marek.ca>
+Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
+---
+ drivers/gpu/drm/etnaviv/etnaviv_buffer.c | 60 ++++++++++++++++++++++--
+ drivers/gpu/drm/etnaviv/state_blt.xml.h | 2 +
+ 2 files changed, 57 insertions(+), 5 deletions(-)
+
+diff --git a/drivers/gpu/drm/etnaviv/etnaviv_buffer.c b/drivers/gpu/drm/etnaviv/etnaviv_buffer.c
+index 32d9fac587f9..76d38561c910 100644
+--- a/drivers/gpu/drm/etnaviv/etnaviv_buffer.c
++++ b/drivers/gpu/drm/etnaviv/etnaviv_buffer.c
+@@ -12,6 +12,7 @@
+
+ #include "common.xml.h"
+ #include "state.xml.h"
++#include "state_blt.xml.h"
+ #include "state_hi.xml.h"
+ #include "state_3d.xml.h"
+ #include "cmdstream.xml.h"
+@@ -233,6 +234,8 @@ void etnaviv_buffer_end(struct etnaviv_gpu *gpu)
+ struct etnaviv_cmdbuf *buffer = &gpu->buffer;
+ unsigned int waitlink_offset = buffer->user_size - 16;
+ u32 link_target, flush = 0;
++ bool has_blt = !!(gpu->identity.minor_features5 &
++ chipMinorFeatures5_BLT_ENGINE);
+
+ lockdep_assert_held(&gpu->lock);
+
+@@ -248,16 +251,38 @@ void etnaviv_buffer_end(struct etnaviv_gpu *gpu)
+ if (flush) {
+ unsigned int dwords = 7;
+
++ if (has_blt)
++ dwords += 10;
++
+ link_target = etnaviv_buffer_reserve(gpu, buffer, dwords);
+
+ CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
+ CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
++ if (has_blt) {
++ CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x1);
++ CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
++ CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
++ CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x0);
++ }
+ CMD_LOAD_STATE(buffer, VIVS_GL_FLUSH_CACHE, flush);
+- if (gpu->exec_state == ETNA_PIPE_3D)
+- CMD_LOAD_STATE(buffer, VIVS_TS_FLUSH_CACHE,
+- VIVS_TS_FLUSH_CACHE_FLUSH);
++ if (gpu->exec_state == ETNA_PIPE_3D) {
++ if (has_blt) {
++ CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x1);
++ CMD_LOAD_STATE(buffer, VIVS_BLT_SET_COMMAND, 0x1);
++ CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x0);
++ } else {
++ CMD_LOAD_STATE(buffer, VIVS_TS_FLUSH_CACHE,
++ VIVS_TS_FLUSH_CACHE_FLUSH);
++ }
++ }
+ CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
+ CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
++ if (has_blt) {
++ CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x1);
++ CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
++ CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
++ CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x0);
++ }
+ CMD_END(buffer);
+
+ etnaviv_buffer_replace_wait(buffer, waitlink_offset,
+@@ -323,6 +348,8 @@ void etnaviv_buffer_queue(struct etnaviv_gpu *gpu, u32 exec_state,
+ bool switch_mmu_context = gpu->mmu_context != mmu_context;
+ unsigned int new_flush_seq = READ_ONCE(gpu->mmu_context->flush_seq);
+ bool need_flush = switch_mmu_context || gpu->flush_seq != new_flush_seq;
++ bool has_blt = !!(gpu->identity.minor_features5 &
++ chipMinorFeatures5_BLT_ENGINE);
+
+ lockdep_assert_held(&gpu->lock);
+
+@@ -433,6 +460,15 @@ void etnaviv_buffer_queue(struct etnaviv_gpu *gpu, u32 exec_state,
+ * 2 semaphore stall + 1 event + 1 wait + 1 link.
+ */
+ return_dwords = 7;
++
++ /*
++ * When the BLT engine is present we need 6 more dwords in the return
++ * target: 3 enable/flush/disable + 4 enable/semaphore stall/disable,
++ * but we don't need the normal TS flush state.
++ */
++ if (has_blt)
++ return_dwords += 6;
++
+ return_target = etnaviv_buffer_reserve(gpu, buffer, return_dwords);
+ CMD_LINK(cmdbuf, return_dwords, return_target);
+
+@@ -447,11 +483,25 @@ void etnaviv_buffer_queue(struct etnaviv_gpu *gpu, u32 exec_state,
+ CMD_LOAD_STATE(buffer, VIVS_GL_FLUSH_CACHE,
+ VIVS_GL_FLUSH_CACHE_DEPTH |
+ VIVS_GL_FLUSH_CACHE_COLOR);
+- CMD_LOAD_STATE(buffer, VIVS_TS_FLUSH_CACHE,
+- VIVS_TS_FLUSH_CACHE_FLUSH);
++ if (has_blt) {
++ CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x1);
++ CMD_LOAD_STATE(buffer, VIVS_BLT_SET_COMMAND, 0x1);
++ CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x0);
++ } else {
++ CMD_LOAD_STATE(buffer, VIVS_TS_FLUSH_CACHE,
++ VIVS_TS_FLUSH_CACHE_FLUSH);
++ }
+ }
+ CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
+ CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_PE);
++
++ if (has_blt) {
++ CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x1);
++ CMD_SEM(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
++ CMD_STALL(buffer, SYNC_RECIPIENT_FE, SYNC_RECIPIENT_BLT);
++ CMD_LOAD_STATE(buffer, VIVS_BLT_ENABLE, 0x0);
++ }
++
+ CMD_LOAD_STATE(buffer, VIVS_GL_EVENT, VIVS_GL_EVENT_EVENT_ID(event) |
+ VIVS_GL_EVENT_FROM_PE);
+ CMD_WAIT(buffer);
+diff --git a/drivers/gpu/drm/etnaviv/state_blt.xml.h b/drivers/gpu/drm/etnaviv/state_blt.xml.h
+index daae55995def..0e8bcf9dcc93 100644
+--- a/drivers/gpu/drm/etnaviv/state_blt.xml.h
++++ b/drivers/gpu/drm/etnaviv/state_blt.xml.h
+@@ -46,6 +46,8 @@ DEALINGS IN THE SOFTWARE.
+
+ /* This is a cut-down version of the state_blt.xml.h file */
+
++#define VIVS_BLT_SET_COMMAND 0x000140ac
++
+ #define VIVS_BLT_ENABLE 0x000140b8
+ #define VIVS_BLT_ENABLE_ENABLE 0x00000001
+
+--
+2.20.1
+
--- /dev/null
+From d7c5782acd354bdb5ed0fa10e1e397eaed558390 Mon Sep 17 00:00:00 2001
+From: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
+Date: Tue, 29 Oct 2019 11:03:05 -0400
+Subject: [PATCH] drm/sched: Fix passing zero to 'PTR_ERR' warning v2
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+Fix a static code checker warning.
+
+v2: Drop PTR_ERR_OR_ZERO.
+
+Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
+Reviewed-by: Emily Deng <Emily.Deng@amd.com>
+Reviewed-by: Christian König <christian.koenig@amd.com>
+Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
+---
+ drivers/gpu/drm/scheduler/sched_main.c | 7 +++++--
+ 1 file changed, 5 insertions(+), 2 deletions(-)
+
+diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c
+index 2af64459b3d7..fbb6446c4f69 100644
+--- a/drivers/gpu/drm/scheduler/sched_main.c
++++ b/drivers/gpu/drm/scheduler/sched_main.c
+@@ -496,8 +496,10 @@ void drm_sched_resubmit_jobs(struct drm_gpu_scheduler *sched)
+ fence = sched->ops->run_job(s_job);
+
+ if (IS_ERR_OR_NULL(fence)) {
++ if (IS_ERR(fence))
++ dma_fence_set_error(&s_fence->finished, PTR_ERR(fence));
++
+ s_job->s_fence->parent = NULL;
+- dma_fence_set_error(&s_fence->finished, PTR_ERR(fence));
+ } else {
+ s_job->s_fence->parent = fence;
+ }
+@@ -746,8 +748,9 @@ static int drm_sched_main(void *param)
+ r);
+ dma_fence_put(fence);
+ } else {
++ if (IS_ERR(fence))
++ dma_fence_set_error(&s_fence->finished, PTR_ERR(fence));
+
+- dma_fence_set_error(&s_fence->finished, PTR_ERR(fence));
+ drm_sched_process_job(NULL, &sched_job->cb);
+ }
+
+--
+2.20.1
+
Code Review / AGL / meta-agl.git / commitdiff