drm/amd/display: implement workaround for riommu related hang

[Why]
During S4/S5/reboot, sometimes riommu invalidation request arrive too
early, DCN may be unable to respond to the invalidation request
resulting in pstate hang.

[How]
VBIOS will force allow pstate for riommu invalidation and driver will
clear it after powering down display pipes.
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Eric Yang <Eric.Yang2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h

@@ -636,6 +636,7 @@ struct dce_hwseq_registers {
 	uint32_t ODM_MEM_PWR_CTRL3;
 	uint32_t DMU_MEM_PWR_CNTL;
 	uint32_t MMHUBBUB_MEM_PWR_CNTL;
+	uint32_t DCHUBBUB_ARB_HOSTVM_CNTL;
 };
  /* set field name */
 #define HWS_SF(blk_name, reg_name, field_name, post_fix)\
@@ -1110,7 +1111,8 @@ struct dce_hwseq_registers {
 	type DOMAIN_POWER_FORCEON;\
 	type DOMAIN_POWER_GATE;\
 	type DOMAIN_PGFSM_PWR_STATUS;\
-	type HPO_HDMISTREAMCLK_G_GATE_DIS;
+	type HPO_HDMISTREAMCLK_G_GATE_DIS;\
+	type DISABLE_HOSTVM_FORCE_ALLOW_PSTATE;
 
 struct dce_hwseq_shift {
 	HWSEQ_REG_FIELD_LIST(uint8_t)

drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c

@@ -47,6 +47,7 @@
 #include "dce/dmub_outbox.h"
 #include "dc_link_dp.h"
 #include "inc/link_dpcd.h"
+#include "dcn10/dcn10_hw_sequencer.h"
 
 #define DC_LOGGER_INIT(logger)
 
@@ -594,3 +595,20 @@ bool dcn31_is_abm_supported(struct dc *dc,
 	}
 	return false;
 }
+
+static void apply_riommu_invalidation_wa(struct dc *dc)
+{
+	struct dce_hwseq *hws = dc->hwseq;
+
+	if (!hws->wa.early_riommu_invalidation)
+		return;
+
+	REG_UPDATE(DCHUBBUB_ARB_HOSTVM_CNTL, DISABLE_HOSTVM_FORCE_ALLOW_PSTATE, 0);
+}
+
+void dcn31_init_pipes(struct dc *dc, struct dc_state *context)
+{
+	dcn10_init_pipes(dc, context);
+	apply_riommu_invalidation_wa(dc);
+
+}

drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.h

@@ -52,5 +52,6 @@ void dcn31_reset_hw_ctx_wrap(
 		struct dc_state *context);
 bool dcn31_is_abm_supported(struct dc *dc,
 		struct dc_state *context, struct dc_stream_state *stream);
+void dcn31_init_pipes(struct dc *dc, struct dc_state *context);
 
 #endif /* __DC_HWSS_DCN31_H__ */

drivers/gpu/drm/amd/display/dc/dcn31/dcn31_init.c

@@ -104,7 +104,7 @@ static const struct hw_sequencer_funcs dcn31_funcs = {
 };
 
 static const struct hwseq_private_funcs dcn31_private_funcs = {
-	.init_pipes = dcn10_init_pipes,
+	.init_pipes = dcn31_init_pipes,
 	.update_plane_addr = dcn20_update_plane_addr,
 	.plane_atomic_disconnect = dcn10_plane_atomic_disconnect,
 	.update_mpcc = dcn20_update_mpcc,

drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c

@@ -741,6 +741,7 @@ static const struct dccg_mask dccg_mask = {
 
 #define HWSEQ_DCN31_REG_LIST()\
 	SR(DCHUBBUB_GLOBAL_TIMER_CNTL), \
+	SR(DCHUBBUB_ARB_HOSTVM_CNTL), \
 	SR(DIO_MEM_PWR_CTRL), \
 	SR(ODM_MEM_PWR_CTRL3), \
 	SR(DMU_MEM_PWR_CNTL), \
@@ -801,6 +802,7 @@ static const struct dce_hwseq_registers hwseq_reg = {
 #define HWSEQ_DCN31_MASK_SH_LIST(mask_sh)\
 	HWSEQ_DCN_MASK_SH_LIST(mask_sh), \
 	HWS_SF(, DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_REFDIV, mask_sh), \
+	HWS_SF(, DCHUBBUB_ARB_HOSTVM_CNTL, DISABLE_HOSTVM_FORCE_ALLOW_PSTATE, mask_sh), \
 	HWS_SF(, DOMAIN0_PG_CONFIG, DOMAIN_POWER_FORCEON, mask_sh), \
 	HWS_SF(, DOMAIN0_PG_CONFIG, DOMAIN_POWER_GATE, mask_sh), \
 	HWS_SF(, DOMAIN1_PG_CONFIG, DOMAIN_POWER_FORCEON, mask_sh), \
@@ -1299,6 +1301,7 @@ static struct dce_hwseq *dcn31_hwseq_create(
 		hws->regs = &hwseq_reg;
 		hws->shifts = &hwseq_shift;
 		hws->masks = &hwseq_mask;
+		hws->wa.early_riommu_invalidation = true;
 	}
 	return hws;
 }

drivers/gpu/drm/amd/display/dc/inc/hw_sequencer_private.h

@@ -41,6 +41,7 @@ struct dce_hwseq_wa {
 	bool DEGVIDCN10_254;
 	bool DEGVIDCN21;
 	bool disallow_self_refresh_during_multi_plane_transition;
+	bool early_riommu_invalidation;
 };
 
 struct hwseq_wa_state {