drm/amdgpu: Move to a per-IB secure flag (TMZ)

Move from a per-CS secure flag (TMZ) to a per-IB
secure flag.
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c

@@ -232,8 +232,6 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs
 	if (ret)
 		goto free_all_kdata;
 
-	p->job->secure = cs->in.flags & AMDGPU_CS_FLAGS_SECURE;
-
 	if (p->ctx->vram_lost_counter != p->job->vram_lost_counter) {
 		ret = -ECANCELED;
 		goto free_all_kdata;

drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c

@@ -133,6 +133,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	uint64_t fence_ctx;
 	uint32_t status = 0, alloc_size;
 	unsigned fence_flags = 0;
+	bool secure;
 
 	unsigned i;
 	int r = 0;
@@ -214,9 +215,10 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	if (job && ring->funcs->emit_cntxcntl) {
 		status |= job->preamble_status;
 		status |= job->preemption_status;
-		amdgpu_ring_emit_cntxcntl(ring, status, job->secure);
+		amdgpu_ring_emit_cntxcntl(ring, status);
 	}
 
+	secure = false;
 	for (i = 0; i < num_ibs; ++i) {
 		ib = &ibs[i];
 
@@ -228,12 +230,27 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		    !amdgpu_sriov_vf(adev)) /* for SRIOV preemption, Preamble CE ib must be inserted anyway */
 			continue;
 
+		/* If this IB is TMZ, add frame TMZ start packet,
+		 * else, turn off TMZ.
+		 */
+		if (ib->flags & AMDGPU_IB_FLAGS_SECURE && ring->funcs->emit_tmz) {
+			if (!secure) {
+				secure = true;
+				amdgpu_ring_emit_tmz(ring, true);
+			}
+		} else if (secure) {
+			secure = false;
+			amdgpu_ring_emit_tmz(ring, false);
+		}
+
 		amdgpu_ring_emit_ib(ring, job, ib, status);
 		status &= ~AMDGPU_HAVE_CTX_SWITCH;
 	}
 
-	if (ring->funcs->emit_tmz)
-		amdgpu_ring_emit_tmz(ring, false, job ? job->secure : false);
+	if (secure) {
+		secure = false;
+		amdgpu_ring_emit_tmz(ring, false);
+	}
 
 #ifdef CONFIG_X86_64
 	if (!(adev->flags & AMD_IS_APU))

drivers/gpu/drm/amd/amdgpu/amdgpu_job.h

@@ -62,9 +62,6 @@ struct amdgpu_job {
 	/* user fence handling */
 	uint64_t		uf_addr;
 	uint64_t		uf_sequence;
-
-	/* the job is due to a secure command submission */
-	bool			secure;
 };
 
 int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,

drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h

@@ -168,8 +168,7 @@ struct amdgpu_ring_funcs {
 	void (*begin_use)(struct amdgpu_ring *ring);
 	void (*end_use)(struct amdgpu_ring *ring);
 	void (*emit_switch_buffer) (struct amdgpu_ring *ring);
-	void (*emit_cntxcntl) (struct amdgpu_ring *ring, uint32_t flags,
-			       bool trusted);
+	void (*emit_cntxcntl) (struct amdgpu_ring *ring, uint32_t flags);
 	void (*emit_rreg)(struct amdgpu_ring *ring, uint32_t reg,
 			  uint32_t reg_val_offs);
 	void (*emit_wreg)(struct amdgpu_ring *ring, uint32_t reg, uint32_t val);
@@ -178,7 +177,7 @@ struct amdgpu_ring_funcs {
 	void (*emit_reg_write_reg_wait)(struct amdgpu_ring *ring,
 					uint32_t reg0, uint32_t reg1,
 					uint32_t ref, uint32_t mask);
-	void (*emit_tmz)(struct amdgpu_ring *ring, bool start, bool trusted);
+	void (*emit_tmz)(struct amdgpu_ring *ring, bool start);
 	/* Try to soft recover the ring to make the fence signal */
 	void (*soft_recovery)(struct amdgpu_ring *ring, unsigned vmid);
 	int (*preempt_ib)(struct amdgpu_ring *ring);
@@ -252,12 +251,12 @@ struct amdgpu_ring {
 #define amdgpu_ring_emit_gds_switch(r, v, db, ds, wb, ws, ab, as) (r)->funcs->emit_gds_switch((r), (v), (db), (ds), (wb), (ws), (ab), (as))
 #define amdgpu_ring_emit_hdp_flush(r) (r)->funcs->emit_hdp_flush((r))
 #define amdgpu_ring_emit_switch_buffer(r) (r)->funcs->emit_switch_buffer((r))
-#define amdgpu_ring_emit_cntxcntl(r, d, s) (r)->funcs->emit_cntxcntl((r), (d), (s))
+#define amdgpu_ring_emit_cntxcntl(r, d) (r)->funcs->emit_cntxcntl((r), (d))
 #define amdgpu_ring_emit_rreg(r, d, o) (r)->funcs->emit_rreg((r), (d), (o))
 #define amdgpu_ring_emit_wreg(r, d, v) (r)->funcs->emit_wreg((r), (d), (v))
 #define amdgpu_ring_emit_reg_wait(r, d, v, m) (r)->funcs->emit_reg_wait((r), (d), (v), (m))
 #define amdgpu_ring_emit_reg_write_reg_wait(r, d0, d1, v, m) (r)->funcs->emit_reg_write_reg_wait((r), (d0), (d1), (v), (m))
-#define amdgpu_ring_emit_tmz(r, b, s) (r)->funcs->emit_tmz((r), (b), (s))
+#define amdgpu_ring_emit_tmz(r, b) (r)->funcs->emit_tmz((r), (b))
 #define amdgpu_ring_pad_ib(r, ib) ((r)->funcs->pad_ib((r), (ib)))
 #define amdgpu_ring_init_cond_exec(r) (r)->funcs->init_cond_exec((r))
 #define amdgpu_ring_patch_cond_exec(r,o) (r)->funcs->patch_cond_exec((r),(o))

drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c

@@ -3037,8 +3037,7 @@ static int gfx_v10_0_rlc_backdoor_autoload_enable(struct amdgpu_device *adev);
 static int gfx_v10_0_wait_for_rlc_autoload_complete(struct amdgpu_device *adev);
 static void gfx_v10_0_ring_emit_ce_meta(struct amdgpu_ring *ring, bool resume);
 static void gfx_v10_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume);
-static void gfx_v10_0_ring_emit_tmz(struct amdgpu_ring *ring, bool start,
-				    bool trusted);
+static void gfx_v10_0_ring_emit_tmz(struct amdgpu_ring *ring, bool start);
 
 static void gfx10_kiq_set_resources(struct amdgpu_ring *kiq_ring, uint64_t queue_mask)
 {
@@ -7436,8 +7435,7 @@ static void gfx_v10_0_ring_emit_sb(struct amdgpu_ring *ring)
 }
 
 static void gfx_v10_0_ring_emit_cntxcntl(struct amdgpu_ring *ring,
-					 uint32_t flags,
-					 bool trusted)
+					 uint32_t flags)
 {
 	uint32_t dw2 = 0;
 
@@ -7445,8 +7443,6 @@ static void gfx_v10_0_ring_emit_cntxcntl(struct amdgpu_ring *ring,
 		gfx_v10_0_ring_emit_ce_meta(ring,
 				    (!amdgpu_sriov_vf(ring->adev) && flags & AMDGPU_IB_PREEMPTED) ? true : false);
 
-	gfx_v10_0_ring_emit_tmz(ring, true, trusted);
-
 	dw2 |= 0x80000000; /* set load_enable otherwise this package is just NOPs */
 	if (flags & AMDGPU_HAVE_CTX_SWITCH) {
 		/* set load_global_config & load_global_uconfig */
@@ -7603,17 +7599,12 @@ static void gfx_v10_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume)
 					   sizeof(de_payload) >> 2);
 }
 
-static void gfx_v10_0_ring_emit_tmz(struct amdgpu_ring *ring, bool start,
-				    bool trusted)
+static void gfx_v10_0_ring_emit_tmz(struct amdgpu_ring *ring, bool start)
 {
-	amdgpu_ring_write(ring, PACKET3(PACKET3_FRAME_CONTROL, 0));
-	/*
-	 * cmd = 0: frame begin
-	 * cmd = 1: frame end
-	 */
-	amdgpu_ring_write(ring,
-			  ((amdgpu_is_tmz(ring->adev) && trusted) ? FRAME_TMZ : 0)
-			  | FRAME_CMD(start ? 0 : 1));
+	if (amdgpu_is_tmz(ring->adev)) {
+		amdgpu_ring_write(ring, PACKET3(PACKET3_FRAME_CONTROL, 0));
+		amdgpu_ring_write(ring, FRAME_TMZ | FRAME_CMD(start ? 0 : 1));
+	}
 }
 
 static void gfx_v10_0_ring_emit_rreg(struct amdgpu_ring *ring, uint32_t reg,

drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c

@@ -2969,8 +2969,7 @@ static uint64_t gfx_v6_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 	return clock;
 }
 
-static void gfx_v6_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags,
-				      bool trusted)
+static void gfx_v6_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 {
 	if (flags & AMDGPU_HAVE_CTX_SWITCH)
 		gfx_v6_0_ring_emit_vgt_flush(ring);

drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c

@@ -2320,8 +2320,7 @@ static void gfx_v7_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, control);
 }
 
-static void gfx_v7_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags,
-				      bool trusted)
+static void gfx_v7_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 {
 	uint32_t dw2 = 0;
 

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c

@@ -6329,8 +6329,7 @@ static void gfx_v8_ring_emit_sb(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, 0);
 }
 
-static void gfx_v8_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags,
-				      bool trusted)
+static void gfx_v8_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 {
 	uint32_t dw2 = 0;
 

drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c

@@ -5442,29 +5442,21 @@ static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring)
 	amdgpu_ring_write_multiple(ring, (void *)&de_payload, sizeof(de_payload) >> 2);
 }
 
-static void gfx_v9_0_ring_emit_tmz(struct amdgpu_ring *ring, bool start,
-				   bool trusted)
+static void gfx_v9_0_ring_emit_tmz(struct amdgpu_ring *ring, bool start)
 {
-	amdgpu_ring_write(ring, PACKET3(PACKET3_FRAME_CONTROL, 0));
-	/*
-	 * cmd = 0: frame begin
-	 * cmd = 1: frame end
-	 */
-	amdgpu_ring_write(ring,
-			  ((amdgpu_is_tmz(ring->adev) && trusted) ? FRAME_TMZ : 0)
-			  | FRAME_CMD(start ? 0 : 1));
+	if (amdgpu_is_tmz(ring->adev)) {
+		amdgpu_ring_write(ring, PACKET3(PACKET3_FRAME_CONTROL, 0));
+		amdgpu_ring_write(ring, FRAME_TMZ | FRAME_CMD(start ? 0 : 1));
+	}
 }
 
-static void gfx_v9_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags,
-				      bool trusted)
+static void gfx_v9_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 {
 	uint32_t dw2 = 0;
 
 	if (amdgpu_sriov_vf(ring->adev))
 		gfx_v9_0_ring_emit_ce_meta(ring);
 
-	gfx_v9_0_ring_emit_tmz(ring, true, trusted);
-
 	dw2 |= 0x80000000; /* set load_enable otherwise this package is just NOPs */
 	if (flags & AMDGPU_HAVE_CTX_SWITCH) {
 		/* set load_global_config & load_global_uconfig */

include/uapi/drm/amdgpu_drm.h

@@ -558,9 +558,6 @@ struct drm_amdgpu_cs_chunk {
 	__u64		chunk_data;
 };
 
-/* Flag the command submission as secure */
-#define AMDGPU_CS_FLAGS_SECURE          (1 << 0)
-
 struct drm_amdgpu_cs_in {
 	/** Rendering context id */
 	__u32		ctx_id;
@@ -601,6 +598,10 @@ union drm_amdgpu_cs {
  */
 #define AMDGPU_IB_FLAG_RESET_GDS_MAX_WAVE_ID (1 << 4)
 
+/* Flag the IB as secure (TMZ)
+ */
+#define AMDGPU_IB_FLAGS_SECURE  (1 << 5)
+
 struct drm_amdgpu_cs_chunk_ib {
 	__u32 _pad;
 	/** AMDGPU_IB_FLAG_* */