diff --git a/drivers/gpu/host1x/Makefile b/drivers/gpu/host1x/Makefile
index 5b6bfcae..34344be2 100644
--- a/drivers/gpu/host1x/Makefile
+++ b/drivers/gpu/host1x/Makefile
@@ -14,7 +14,6 @@ host1x-next-y = \
 	job.o \
 	debug.o \
 	mipi.o \
-	uapi.o \
 	fence.o \
 	hw/host1x01.o \
 	hw/host1x02.o \
diff --git a/drivers/gpu/host1x/dev.c b/drivers/gpu/host1x/dev.c
index a89f2350..2aed4592 100644
--- a/drivers/gpu/host1x/dev.c
+++ b/drivers/gpu/host1x/dev.c
@@ -528,12 +528,6 @@ static int host1x_probe(struct platform_device *pdev)
 		goto deinit_syncpt;
 	}
 
-	err = host1x_uapi_init(&host->uapi, host);
-	if (err) {
-		dev_err(&pdev->dev, "failed to initialize uapi\n");
-		goto deinit_intr;
-	}
-
 	host1x_debug_init(host);
 
 	if (host->info->has_hypervisor)
@@ -553,8 +547,6 @@ unregister:
 	host1x_unregister(host);
 deinit_debugfs:
 	host1x_debug_deinit(host);
-	host1x_uapi_deinit(&host->uapi);
-deinit_intr:
 	host1x_intr_deinit(host);
 deinit_syncpt:
 	host1x_syncpt_deinit(host);
@@ -576,7 +568,6 @@ static int host1x_remove(struct platform_device *pdev)
 
 	host1x_unregister(host);
 	host1x_debug_deinit(host);
-	host1x_uapi_deinit(&host->uapi);
 	host1x_intr_deinit(host);
 	host1x_syncpt_deinit(host);
 	reset_control_assert(host->rst);
diff --git a/drivers/gpu/host1x/dev.h b/drivers/gpu/host1x/dev.h
index 95c78451..5b7fdea5 100644
--- a/drivers/gpu/host1x/dev.h
+++ b/drivers/gpu/host1x/dev.h
@@ -17,7 +17,6 @@
 #include "intr.h"
 #include "job.h"
 #include "syncpt.h"
-#include "uapi.h"
 
 struct host1x_syncpt;
 struct host1x_syncpt_base;
@@ -152,7 +151,6 @@ struct host1x {
 	struct device_dma_parameters dma_parms;
 
 	struct host1x_bo_cache cache;
-	struct host1x_uapi uapi;
 };
 
 void host1x_hypervisor_writel(struct host1x *host1x, u32 r, u32 v);
diff --git a/drivers/gpu/host1x/fence.c b/drivers/gpu/host1x/fence.c
index ea4a0c7d..4a8fdbfb 100644
--- a/drivers/gpu/host1x/fence.c
+++ b/drivers/gpu/host1x/fence.c
@@ -31,20 +31,24 @@ struct host1x_syncpt_fence {
 	struct delayed_work timeout_work;
 };
 
-static const char *syncpt_fence_get_driver_name(struct dma_fence *f)
+static const char *host1x_syncpt_fence_get_driver_name(struct dma_fence *f)
 {
 	return "host1x";
 }
 
-static const char *syncpt_fence_get_timeline_name(struct dma_fence *f)
+static const char *host1x_syncpt_fence_get_timeline_name(struct dma_fence *f)
 {
 	return "syncpoint";
 }
 
-static bool syncpt_fence_enable_signaling(struct dma_fence *f)
+static struct host1x_syncpt_fence *to_host1x_fence(struct dma_fence *f)
 {
-	struct host1x_syncpt_fence *sf =
-		container_of(f, struct host1x_syncpt_fence, base);
+	return container_of(f, struct host1x_syncpt_fence, base);
+}
+
+static bool host1x_syncpt_fence_enable_signaling(struct dma_fence *f)
+{
+	struct host1x_syncpt_fence *sf = to_host1x_fence(f);
 	int err;
 
 	if (host1x_syncpt_is_expired(sf->sp, sf->threshold))
@@ -56,7 +60,7 @@ static bool syncpt_fence_enable_signaling(struct dma_fence *f)
 	 * The dma_fence framework requires the fence driver to keep a
 	 * reference to any fences for which 'enable_signaling' has been
 	 * called (and that have not been signalled).
-	 * 
+	 *
 	 * We provide a userspace API to create arbitrary syncpoint fences,
 	 * so we cannot normally guarantee that all fences get signalled.
 	 * As such, setup a timeout, so that long-lasting fences will get
@@ -85,10 +89,9 @@ static bool syncpt_fence_enable_signaling(struct dma_fence *f)
 	return true;
 }
 
-static void syncpt_fence_release(struct dma_fence *f)
+static void host1x_syncpt_fence_release(struct dma_fence *f)
 {
-	struct host1x_syncpt_fence *sf =
-		container_of(f, struct host1x_syncpt_fence, base);
+	struct host1x_syncpt_fence *sf = to_host1x_fence(f);
 
 	if (sf->waiter)
 		kfree(sf->waiter);
@@ -96,11 +99,11 @@ static void syncpt_fence_release(struct dma_fence *f)
 	dma_fence_free(f);
 }
 
-const struct dma_fence_ops syncpt_fence_ops = {
-	.get_driver_name = syncpt_fence_get_driver_name,
-	.get_timeline_name = syncpt_fence_get_timeline_name,
-	.enable_signaling = syncpt_fence_enable_signaling,
-	.release = syncpt_fence_release,
+const struct dma_fence_ops host1x_syncpt_fence_ops = {
+	.get_driver_name = host1x_syncpt_fence_get_driver_name,
+	.get_timeline_name = host1x_syncpt_fence_get_timeline_name,
+	.enable_signaling = host1x_syncpt_fence_enable_signaling,
+	.release = host1x_syncpt_fence_release,
 };
 
 void host1x_fence_signal(struct host1x_syncpt_fence *f)
@@ -157,7 +160,7 @@ struct dma_fence *host1x_fence_create(struct host1x_syncpt *sp, u32 threshold)
 	fence->sp = sp;
 	fence->threshold = threshold;
 
-	dma_fence_init(&fence->base, &syncpt_fence_ops, &lock,
+	dma_fence_init(&fence->base, &host1x_syncpt_fence_ops, &lock,
 		       dma_fence_context_alloc(1), 0);
 
 	INIT_DELAYED_WORK(&fence->timeout_work, do_fence_timeout);
@@ -165,46 +168,3 @@ struct dma_fence *host1x_fence_create(struct host1x_syncpt *sp, u32 threshold)
 	return &fence->base;
 }
 EXPORT_SYMBOL(host1x_fence_create);
-
-int host1x_fence_create_fd(struct host1x_syncpt *sp, u32 threshold)
-{
-	struct sync_file *file;
-	struct dma_fence *f;
-	int fd;
-
-	f = host1x_fence_create(sp, threshold);
-	if (IS_ERR(f))
-		return PTR_ERR(f);
-
-	fd = get_unused_fd_flags(O_CLOEXEC);
-	if (fd < 0) {
-		dma_fence_put(f);
-		return fd;
-	}
-
-	file = sync_file_create(f);
-	dma_fence_put(f);
-	if (!file)
-		return -ENOMEM;
-
-	fd_install(fd, file->file);
-
-	return fd;
-}
-EXPORT_SYMBOL(host1x_fence_create_fd);
-
-int host1x_fence_extract(struct dma_fence *fence, u32 *id, u32 *threshold)
-{
-	struct host1x_syncpt_fence *f;
-
-	if (fence->ops != &syncpt_fence_ops)
-		return -EINVAL;
-
-	f = container_of(fence, struct host1x_syncpt_fence, base);
-
-	*id = f->sp->id;
-	*threshold = f->threshold;
-
-	return 0;
-}
-EXPORT_SYMBOL(host1x_fence_extract);
diff --git a/drivers/gpu/host1x/hw/channel_hw.c b/drivers/gpu/host1x/hw/channel_hw.c
index 1917e760..4eda10a3 100644
--- a/drivers/gpu/host1x/hw/channel_hw.c
+++ b/drivers/gpu/host1x/hw/channel_hw.c
@@ -47,27 +47,57 @@ static void trace_write_gather(struct host1x_cdma *cdma, struct host1x_bo *bo,
 	}
 }
 
-static void submit_gathers(struct host1x_job *job)
+static void submit_wait(struct host1x_cdma *cdma, u32 id, u32 threshold,
+			u32 next_class)
+{
+#if HOST1X_HW >= 2
+	host1x_cdma_push_wide(cdma,
+		host1x_opcode_setclass(
+			HOST1X_CLASS_HOST1X,
+			HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32,
+			/* WAIT_SYNCPT_32 is at SYNCPT_PAYLOAD_32+2 */
+			BIT(0) | BIT(2)
+		),
+		threshold,
+		id,
+		host1x_opcode_setclass(next_class, 0, 0)
+	);
+#else
+	/* TODO add waitchk or use waitbases or other mitigation */
+	host1x_cdma_push(cdma,
+		host1x_opcode_setclass(
+			HOST1X_CLASS_HOST1X,
+			host1x_uclass_wait_syncpt_r(),
+			BIT(0)
+		),
+		host1x_class_host_wait_syncpt(id, threshold)
+	);
+	host1x_cdma_push(cdma,
+		host1x_opcode_setclass(next_class, 0, 0),
+		HOST1X_OPCODE_NOP
+	);
+#endif
+}
+
+static void submit_gathers(struct host1x_job *job, u32 job_syncpt_base)
 {
 	struct host1x_cdma *cdma = &job->channel->cdma;
 #if HOST1X_HW < 6
 	struct device *dev = job->channel->dev;
 #endif
 	unsigned int i;
+	u32 threshold;
 
 	for (i = 0; i < job->num_cmds; i++) {
 		struct host1x_job_cmd *cmd = &job->cmds[i];
 
 		if (cmd->is_wait) {
-			/* TODO use modern wait */
-			host1x_cdma_push(cdma,
-				 host1x_opcode_setclass(HOST1X_CLASS_HOST1X,
-					host1x_uclass_wait_syncpt_r(), 1),
-				 host1x_class_host_wait_syncpt(cmd->wait.id,
-					cmd->wait.threshold));
-			host1x_cdma_push(
-				cdma, host1x_opcode_setclass(job->class, 0, 0),
-				HOST1X_OPCODE_NOP);
+			if (cmd->wait.relative)
+				threshold = job_syncpt_base + cmd->wait.threshold;
+			else
+				threshold = cmd->wait.threshold;
+
+			submit_wait(cdma, cmd->wait.id, threshold, cmd->wait.next_class);
 		} else {
 			struct host1x_job_gather *g = &cmd->gather;
 
@@ -196,7 +226,7 @@ static int channel_submit(struct host1x_job *job)
 				 host1x_opcode_setclass(job->class, 0, 0),
 				 HOST1X_OPCODE_NOP);
 
-	submit_gathers(job);
+	submit_gathers(job, syncval - user_syncpt_incrs);
 
 	/* end CDMA submit & stash pinned hMems into sync queue */
 	host1x_cdma_end(&ch->cdma, job);
diff --git a/drivers/gpu/host1x/hw/debug_hw.c b/drivers/gpu/host1x/hw/debug_hw.c
index 35952fd5..54e31d81 100644
--- a/drivers/gpu/host1x/hw/debug_hw.c
+++ b/drivers/gpu/host1x/hw/debug_hw.c
@@ -156,9 +156,9 @@ static unsigned int show_channel_command(struct output *o, u32 val,
 	}
 }
 
-static void show_gather(struct output *o, phys_addr_t phys_addr,
+static void show_gather(struct output *o, dma_addr_t phys_addr,
 			unsigned int words, struct host1x_cdma *cdma,
-			phys_addr_t pin_addr, u32 *map_addr)
+			dma_addr_t pin_addr, u32 *map_addr)
 {
 	/* Map dmaget cursor to corresponding mem handle */
 	u32 offset = phys_addr - pin_addr;
@@ -176,11 +176,11 @@ static void show_gather(struct output *o, phys_addr_t phys_addr,
 	}
 
 	for (i = 0; i < words; i++) {
-		u32 addr = phys_addr + i * 4;
+		dma_addr_t addr = phys_addr + i * 4;
 		u32 val = *(map_addr + offset / 4 + i);
 
 		if (!data_count) {
-			host1x_debug_output(o, "%08x: %08x: ", addr, val);
+			host1x_debug_output(o, "    %pad: %08x: ", &addr, val);
 			data_count = show_channel_command(o, val, &payload);
 		} else {
 			host1x_debug_cont(o, "%08x%s", val,
@@ -195,19 +195,16 @@ static void show_channel_gathers(struct output *o, struct host1x_cdma *cdma)
 	struct push_buffer *pb = &cdma->push_buffer;
 	struct host1x_job *job;
 
-	host1x_debug_output(o, "PUSHBUF at %pad, %u words\n",
-			    &pb->dma, pb->size / 4);
-
-	show_gather(o, pb->dma, pb->size / 4, cdma, pb->dma, pb->mapped);
-
 	list_for_each_entry(job, &cdma->sync_queue, list) {
 		unsigned int i;
 
-		host1x_debug_output(o, "\n%p: JOB, syncpt_id=%d, syncpt_val=%d, first_get=%08x, timeout=%d num_slots=%d, num_handles=%d\n",
-				    job, job->syncpt->id, job->syncpt_end,
-				    job->first_get, job->timeout,
+		host1x_debug_output(o, "JOB, syncpt %u: %u timeout: %u num_slots: %u num_handles: %u\n",
+				    job->syncpt->id, job->syncpt_end, job->timeout,
 				    job->num_slots, job->num_unpins);
 
+		show_gather(o, pb->dma + job->first_get, job->num_slots * 2, cdma,
+			    pb->dma + job->first_get, pb->mapped + job->first_get);
+
 		for (i = 0; i < job->num_cmds; i++) {
 			struct host1x_job_gather *g;
 			u32 *mapped;
@@ -227,7 +224,7 @@ static void show_channel_gathers(struct output *o, struct host1x_cdma *cdma)
 				continue;
 			}
 
-			host1x_debug_output(o, "    GATHER at %pad+%#x, %d words\n",
+			host1x_debug_output(o, "  GATHER at %pad+%#x, %d words\n",
 					    &g->base, g->offset, g->words);
 
 			show_gather(o, g->base + g->offset, g->words, cdma,
diff --git a/drivers/gpu/host1x/hw/debug_hw_1x01.c b/drivers/gpu/host1x/hw/debug_hw_1x01.c
index 02a93305..85242a59 100644
--- a/drivers/gpu/host1x/hw/debug_hw_1x01.c
+++ b/drivers/gpu/host1x/hw/debug_hw_1x01.c
@@ -16,10 +16,13 @@ static void host1x_debug_show_channel_cdma(struct host1x *host,
 					   struct output *o)
 {
 	struct host1x_cdma *cdma = &ch->cdma;
+	dma_addr_t dmastart, dmaend;
 	u32 dmaput, dmaget, dmactrl;
 	u32 cbstat, cbread;
 	u32 val, base, baseval;
 
+	dmastart = host1x_ch_readl(ch, HOST1X_CHANNEL_DMASTART);
+	dmaend = host1x_ch_readl(ch, HOST1X_CHANNEL_DMAEND);
 	dmaput = host1x_ch_readl(ch, HOST1X_CHANNEL_DMAPUT);
 	dmaget = host1x_ch_readl(ch, HOST1X_CHANNEL_DMAGET);
 	dmactrl = host1x_ch_readl(ch, HOST1X_CHANNEL_DMACTRL);
@@ -56,9 +59,10 @@ static void host1x_debug_show_channel_cdma(struct host1x *host,
 				    HOST1X_SYNC_CBSTAT_CBOFFSET_V(cbstat),
 				    cbread);
 
-	host1x_debug_output(o, "DMAPUT %08x, DMAGET %08x, DMACTL %08x\n",
+	host1x_debug_output(o, "DMASTART %pad, DMAEND %pad\n", &dmastart, &dmaend);
+	host1x_debug_output(o, "DMAPUT %08x DMAGET %08x DMACTL %08x\n",
 			    dmaput, dmaget, dmactrl);
-	host1x_debug_output(o, "CBREAD %08x, CBSTAT %08x\n", cbread, cbstat);
+	host1x_debug_output(o, "CBREAD %08x CBSTAT %08x\n", cbread, cbstat);
 
 	show_channel_gathers(o, cdma);
 	host1x_debug_output(o, "\n");
diff --git a/drivers/gpu/host1x/hw/debug_hw_1x06.c b/drivers/gpu/host1x/hw/debug_hw_1x06.c
index 6d1b583a..9d066787 100644
--- a/drivers/gpu/host1x/hw/debug_hw_1x06.c
+++ b/drivers/gpu/host1x/hw/debug_hw_1x06.c
@@ -16,10 +16,23 @@ static void host1x_debug_show_channel_cdma(struct host1x *host,
 					   struct output *o)
 {
 	struct host1x_cdma *cdma = &ch->cdma;
+	dma_addr_t dmastart = 0, dmaend = 0;
 	u32 dmaput, dmaget, dmactrl;
 	u32 offset, class;
 	u32 ch_stat;
 
+#if defined(CONFIG_ARCH_DMA_ADDR_T_64BIT) && HOST1X_HW >= 6
+	dmastart = host1x_ch_readl(ch, HOST1X_CHANNEL_DMASTART_HI);
+	dmastart <<= 32;
+#endif
+	dmastart |= host1x_ch_readl(ch, HOST1X_CHANNEL_DMASTART);
+
+#if defined(CONFIG_ARCH_DMA_ADDR_T_64BIT) && HOST1X_HW >= 6
+	dmaend = host1x_ch_readl(ch, HOST1X_CHANNEL_DMAEND_HI);
+	dmaend <<= 32;
+#endif
+	dmaend |= host1x_ch_readl(ch, HOST1X_CHANNEL_DMAEND);
+
 	dmaput = host1x_ch_readl(ch, HOST1X_CHANNEL_DMAPUT);
 	dmaget = host1x_ch_readl(ch, HOST1X_CHANNEL_DMAGET);
 	dmactrl = host1x_ch_readl(ch, HOST1X_CHANNEL_DMACTRL);
@@ -41,7 +54,8 @@ static void host1x_debug_show_channel_cdma(struct host1x *host,
 		host1x_debug_output(o, "active class %02x, offset %04x\n",
 				    class, offset);
 
-	host1x_debug_output(o, "DMAPUT %08x, DMAGET %08x, DMACTL %08x\n",
+	host1x_debug_output(o, "DMASTART %pad, DMAEND %pad\n", &dmastart, &dmaend);
+	host1x_debug_output(o, "DMAPUT %08x DMAGET %08x DMACTL %08x\n",
 			    dmaput, dmaget, dmactrl);
 	host1x_debug_output(o, "CHANNELSTAT %02x\n", ch_stat);
 
diff --git a/drivers/gpu/host1x/hw/hw_host1x02_uclass.h b/drivers/gpu/host1x/hw/hw_host1x02_uclass.h
index 4fc51f70..0a2ab8f1 100644
--- a/drivers/gpu/host1x/hw/hw_host1x02_uclass.h
+++ b/drivers/gpu/host1x/hw/hw_host1x02_uclass.h
@@ -165,5 +165,17 @@ static inline u32 host1x_uclass_indoff_rwn_read_v(void)
 }
 #define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \
 	host1x_uclass_indoff_indroffset_f(v)
+static inline u32 host1x_uclass_load_syncpt_payload_32_r(void)
+{
+	return 0x4e;
+}
+#define HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32 \
+	host1x_uclass_load_syncpt_payload_32_r()
+static inline u32 host1x_uclass_wait_syncpt_32_r(void)
+{
+	return 0x50;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_32 \
+	host1x_uclass_wait_syncpt_32_r()
 
 #endif
diff --git a/drivers/gpu/host1x/hw/hw_host1x04_uclass.h b/drivers/gpu/host1x/hw/hw_host1x04_uclass.h
index 9e84a4ad..60c692b9 100644
--- a/drivers/gpu/host1x/hw/hw_host1x04_uclass.h
+++ b/drivers/gpu/host1x/hw/hw_host1x04_uclass.h
@@ -165,5 +165,17 @@ static inline u32 host1x_uclass_indoff_rwn_read_v(void)
 }
 #define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \
 	host1x_uclass_indoff_indroffset_f(v)
+static inline u32 host1x_uclass_load_syncpt_payload_32_r(void)
+{
+	return 0x4e;
+}
+#define HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32 \
+	host1x_uclass_load_syncpt_payload_32_r()
+static inline u32 host1x_uclass_wait_syncpt_32_r(void)
+{
+	return 0x50;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_32 \
+	host1x_uclass_wait_syncpt_32_r()
 
 #endif
diff --git a/drivers/gpu/host1x/hw/hw_host1x05_uclass.h b/drivers/gpu/host1x/hw/hw_host1x05_uclass.h
index aee5a4e3..2fcc9a2a 100644
--- a/drivers/gpu/host1x/hw/hw_host1x05_uclass.h
+++ b/drivers/gpu/host1x/hw/hw_host1x05_uclass.h
@@ -165,5 +165,17 @@ static inline u32 host1x_uclass_indoff_rwn_read_v(void)
 }
 #define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \
 	host1x_uclass_indoff_indroffset_f(v)
+static inline u32 host1x_uclass_load_syncpt_payload_32_r(void)
+{
+	return 0x4e;
+}
+#define HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32 \
+	host1x_uclass_load_syncpt_payload_32_r()
+static inline u32 host1x_uclass_wait_syncpt_32_r(void)
+{
+	return 0x50;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_32 \
+	host1x_uclass_wait_syncpt_32_r()
 
 #endif
diff --git a/drivers/gpu/host1x/hw/hw_host1x06_uclass.h b/drivers/gpu/host1x/hw/hw_host1x06_uclass.h
index c4bacdb7..5f831438 100644
--- a/drivers/gpu/host1x/hw/hw_host1x06_uclass.h
+++ b/drivers/gpu/host1x/hw/hw_host1x06_uclass.h
@@ -165,5 +165,17 @@ static inline u32 host1x_uclass_indoff_rwn_read_v(void)
 }
 #define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \
 	host1x_uclass_indoff_indroffset_f(v)
+static inline u32 host1x_uclass_load_syncpt_payload_32_r(void)
+{
+	return 0x4e;
+}
+#define HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32 \
+	host1x_uclass_load_syncpt_payload_32_r()
+static inline u32 host1x_uclass_wait_syncpt_32_r(void)
+{
+	return 0x50;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_32 \
+	host1x_uclass_wait_syncpt_32_r()
 
 #endif
diff --git a/drivers/gpu/host1x/hw/hw_host1x07_uclass.h b/drivers/gpu/host1x/hw/hw_host1x07_uclass.h
index c74070f3..8cd2ef08 100644
--- a/drivers/gpu/host1x/hw/hw_host1x07_uclass.h
+++ b/drivers/gpu/host1x/hw/hw_host1x07_uclass.h
@@ -165,5 +165,17 @@ static inline u32 host1x_uclass_indoff_rwn_read_v(void)
 }
 #define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \
 	host1x_uclass_indoff_indroffset_f(v)
+static inline u32 host1x_uclass_load_syncpt_payload_32_r(void)
+{
+	return 0x4e;
+}
+#define HOST1X_UCLASS_LOAD_SYNCPT_PAYLOAD_32 \
+	host1x_uclass_load_syncpt_payload_32_r()
+static inline u32 host1x_uclass_wait_syncpt_32_r(void)
+{
+	return 0x50;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_32 \
+	host1x_uclass_wait_syncpt_32_r()
 
 #endif
diff --git a/drivers/gpu/host1x/include/linux/host1x-next.h b/drivers/gpu/host1x/include/linux/host1x-next.h
index 45e0a712..c9f47fad 100644
--- a/drivers/gpu/host1x/include/linux/host1x-next.h
+++ b/drivers/gpu/host1x/include/linux/host1x-next.h
@@ -218,11 +218,7 @@ u32 host1x_syncpt_base_id(struct host1x_syncpt_base *base);
 void host1x_syncpt_release_vblank_reservation(struct host1x_client *client,
 					      u32 syncpt_id);
 
-struct host1x_syncpt *host1x_syncpt_fd_get(int fd);
-
 struct dma_fence *host1x_fence_create(struct host1x_syncpt *sp, u32 threshold);
-int host1x_fence_create_fd(struct host1x_syncpt *sp, u32 threshold);
-int host1x_fence_extract(struct dma_fence *fence, u32 *id, u32 *threshold);
 
 /*
  * host1x channel
@@ -324,13 +320,18 @@ struct host1x_job {
 	/* Callback called when job is freed */
 	void (*release)(struct host1x_job *job);
 	void *user_data;
+
+	/* Whether host1x-side firewall should be ran for this job or not */
+	bool enable_firewall;
 };
 
 struct host1x_job *host1x_job_alloc(struct host1x_channel *ch,
-				    u32 num_cmdbufs, u32 num_relocs);
+				    u32 num_cmdbufs, u32 num_relocs,
+				    bool skip_firewall);
 void host1x_job_add_gather(struct host1x_job *job, struct host1x_bo *bo,
 			   unsigned int words, unsigned int offset);
-void host1x_job_add_wait(struct host1x_job *job, u32 id, u32 thresh);
+void host1x_job_add_wait(struct host1x_job *job, u32 id, u32 thresh,
+			 bool relative, u32 next_class);
 struct host1x_job *host1x_job_get(struct host1x_job *job);
 void host1x_job_put(struct host1x_job *job);
 int host1x_job_pin(struct host1x_job *job, struct device *dev);
diff --git a/drivers/gpu/host1x/include/uapi/linux/host1x-next.h b/drivers/gpu/host1x/include/uapi/linux/host1x-next.h
deleted file mode 100644
index 9c8fb942..00000000
--- a/drivers/gpu/host1x/include/uapi/linux/host1x-next.h
+++ /dev/null
@@ -1,134 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
-/* Copyright (c) 2020 NVIDIA Corporation */
-
-#ifndef _UAPI__LINUX_HOST1X_H
-#define _UAPI__LINUX_HOST1X_H
-
-#include <linux/ioctl.h>
-#include <linux/types.h>
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-struct host1x_allocate_syncpoint {
-	/**
-	 * @fd: [out]
-	 *
-	 * New file descriptor representing the allocated syncpoint.
-	 */
-	__s32 fd;
-
-	__u32 reserved[3];
-};
-
-struct host1x_syncpoint_info {
-	/**
-	 * @id: [out]
-	 *
-	 * System-global ID of the syncpoint.
-	 */
-	__u32 id;
-
-	__u32 reserved[3];
-};
-
-struct host1x_syncpoint_increment {
-	/**
-	 * @count: [in]
-	 *
-	 * Number of times to increment the syncpoint. The syncpoint can
-	 * be observed at in-between values, but each increment is atomic.
-	 */
-	__u32 count;
-};
-
-struct host1x_read_syncpoint {
-	/**
-	 * @id: [in]
-	 *
-	 * ID of the syncpoint to read.
-	 */
-	__u32 id;
-
-	/**
-	 * @value: [out]
-	 *
-	 * Current value of the syncpoint.
-	 */
-	__u32 value;
-};
-
-struct host1x_create_fence {
-	/**
-	 * @id: [in]
-	 *
-	 * ID of the syncpoint to create a fence for.
-	 */
-	__u32 id;
-
-	/**
-	 * @threshold: [in]
-	 *
-	 * When the syncpoint reaches this value, the fence will be signaled.
-	 * The syncpoint is considered to have reached the threshold when the
-	 * following condition is true:
-	 *
-	 * 	((value - threshold) & 0x80000000U) == 0U
-	 *
-	 */
-	__u32 threshold;
-
-	/**
-	 * @fence_fd: [out]
-	 *
-	 * New sync_file file descriptor containing the created fence.
-	 */
-	__s32 fence_fd;
-
-	__u32 reserved[1];
-};
-
-struct host1x_fence_extract_fence {
-	__u32 id;
-	__u32 threshold;
-};
-
-struct host1x_fence_extract {
-	/**
-	 * @fence_fd: [in]
-	 *
-	 * sync_file file descriptor
-	 */
-	__s32 fence_fd;
-
-	/**
-	 * @num_fences: [in,out]
-	 *
-	 * In: size of the `fences_ptr` array counted in elements.
-	 * Out: required size of the `fences_ptr` array counted in elements.
-	 */
-	__u32 num_fences;
-
-	/**
-	 * @fences_ptr: [in]
-	 *
-	 * Pointer to array of `struct host1x_fence_extract_fence`.
-	 */
-	__u64 fences_ptr;
-
-	__u32 reserved[2];
-};
-
-#define HOST1X_IOCTL_ALLOCATE_SYNCPOINT  _IOWR('X', 0x00, struct host1x_allocate_syncpoint)
-#define HOST1X_IOCTL_READ_SYNCPOINT      _IOR ('X', 0x01, struct host1x_read_syncpoint)
-#define HOST1X_IOCTL_CREATE_FENCE        _IOWR('X', 0x02, struct host1x_create_fence)
-#define HOST1X_IOCTL_SYNCPOINT_INFO      _IOWR('X', 0x03, struct host1x_syncpoint_info)
-#define HOST1X_IOCTL_SYNCPOINT_INCREMENT _IOWR('X', 0x04, struct host1x_syncpoint_increment)
-#define HOST1X_IOCTL_FENCE_EXTRACT       _IOWR('X', 0x05, struct host1x_fence_extract)
-
-#if defined(__cplusplus)
-}
-#endif
-
-#endif
diff --git a/drivers/gpu/host1x/intr.c b/drivers/gpu/host1x/intr.c
index 7a847545..45b6be92 100644
--- a/drivers/gpu/host1x/intr.c
+++ b/drivers/gpu/host1x/intr.c
@@ -273,7 +273,7 @@ void host1x_intr_put_ref(struct host1x *host, unsigned int id, void *ref,
 	if (flush) {
 		/* Wait until any concurrently executing handler has finished. */
 		while (atomic_read(&waiter->state) != WLS_HANDLED)
-			cpu_relax();
+			schedule();
 	}
 
 	kref_put(&waiter->refcount, waiter_release);
diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c
index b5896749..ee47735a 100644
--- a/drivers/gpu/host1x/job.c
+++ b/drivers/gpu/host1x/job.c
@@ -24,14 +24,18 @@
 #define HOST1X_WAIT_SYNCPT_OFFSET 0x8
 
 struct host1x_job *host1x_job_alloc(struct host1x_channel *ch,
-				    u32 num_cmdbufs, u32 num_relocs)
+				    u32 num_cmdbufs, u32 num_relocs,
+				    bool skip_firewall)
 {
 	struct host1x_job *job = NULL;
 	unsigned int num_unpins = num_relocs;
+	bool enable_firewall;
 	u64 total;
 	void *mem;
 
-	if (!IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL))
+	enable_firewall = IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL) && !skip_firewall;
+
+	if (!enable_firewall)
 		num_unpins += num_cmdbufs;
 
 	/* Check that we're not going to overflow */
@@ -48,6 +52,8 @@ struct host1x_job *host1x_job_alloc(struct host1x_channel *ch,
 	if (!job)
 		return NULL;
 
+	job->enable_firewall = enable_firewall;
+
 	kref_init(&job->ref);
 	job->channel = ch;
 
@@ -111,13 +117,16 @@ void host1x_job_add_gather(struct host1x_job *job, struct host1x_bo *bo,
 }
 EXPORT_SYMBOL(host1x_job_add_gather);
 
-void host1x_job_add_wait(struct host1x_job *job, u32 id, u32 thresh)
+void host1x_job_add_wait(struct host1x_job *job, u32 id, u32 thresh,
+			 bool relative, u32 next_class)
 {
 	struct host1x_job_cmd *cmd = &job->cmds[job->num_cmds];
 
 	cmd->is_wait = true;
 	cmd->wait.id = id;
 	cmd->wait.threshold = thresh;
+	cmd->wait.next_class = next_class;
+	cmd->wait.relative = relative;
 
 	job->num_cmds++;
 }
@@ -193,7 +202,7 @@ static unsigned int pin_job(struct host1x *host, struct host1x_job *job)
 	 * We will copy gathers BO content later, so there is no need to
 	 * hold and pin them.
 	 */
-	if (IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL))
+	if (job->enable_firewall)
 		return 0;
 
 	for (i = 0; i < job->num_cmds; i++) {
@@ -280,7 +289,7 @@ static int do_relocs(struct host1x_job *job, struct host1x_job_gather *g)
 		if (cmdbuf != reloc->cmdbuf.bo)
 			continue;
 
-		if (IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL)) {
+		if (job->enable_firewall) {
 			target = (u32 *)job->gather_copy_mapped +
 					reloc->cmdbuf.offset / sizeof(u32) +
 						g->offset / sizeof(u32);
@@ -593,7 +602,7 @@ int host1x_job_pin(struct host1x_job *job, struct device *dev)
 	if (err)
 		goto out;
 
-	if (IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL)) {
+	if (job->enable_firewall) {
 		err = copy_gathers(host->dev, job, dev);
 		if (err)
 			goto out;
@@ -612,7 +621,7 @@ int host1x_job_pin(struct host1x_job *job, struct device *dev)
 			continue;
 
 		/* copy_gathers() sets gathers base if firewall is enabled */
-		if (!IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL))
+		if (!job->enable_firewall)
 			g->base = job->gather_addr_phys[i];
 
 		for (j = i + 1; j < job->num_cmds; j++) {
@@ -646,7 +655,7 @@ void host1x_job_unpin(struct host1x_job *job)
 		struct host1x_bo_mapping *map = job->unpins[i].map;
 		struct host1x_bo *bo = map->bo;
 
-		if (!IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL) && map->size && host->domain) {
+		if (!job->enable_firewall && map->size && host->domain) {
 			iommu_unmap(host->domain, job->addr_phys[i], map->size);
 			free_iova(&host->iova, iova_pfn(&host->iova, job->addr_phys[i]));
 		}
diff --git a/drivers/gpu/host1x/job.h b/drivers/gpu/host1x/job.h
index 657e2400..dad5a194 100644
--- a/drivers/gpu/host1x/job.h
+++ b/drivers/gpu/host1x/job.h
@@ -21,6 +21,8 @@ struct host1x_job_gather {
 struct host1x_job_wait {
 	u32 id;
 	u32 threshold;
+	u32 next_class;
+	bool relative;
 };
 
 struct host1x_job_cmd {
diff --git a/drivers/gpu/host1x/syncpt.c b/drivers/gpu/host1x/syncpt.c
index 100270ac..d198a108 100644
--- a/drivers/gpu/host1x/syncpt.c
+++ b/drivers/gpu/host1x/syncpt.c
@@ -42,6 +42,18 @@ static void host1x_syncpt_base_free(struct host1x_syncpt_base *base)
 		base->requested = false;
 }
 
+/**
+ * host1x_syncpt_alloc() - allocate a syncpoint
+ * @host: host1x device data
+ * @flags: bitfield of HOST1X_SYNCPT_* flags
+ * @name: name for the syncpoint for use in debug prints
+ *
+ * Allocates a hardware syncpoint for the caller's use. The caller then has
+ * the sole authority to mutate the syncpoint's value until it is freed again.
+ *
+ * If no free syncpoints are available, or a NULL name was specified, returns
+ * NULL.
+ */
 struct host1x_syncpt *host1x_syncpt_alloc(struct host1x *host,
 					  unsigned long flags,
 					  const char *name)
@@ -50,6 +62,9 @@ struct host1x_syncpt *host1x_syncpt_alloc(struct host1x *host,
 	char *full_name;
 	unsigned int i;
 
+	if (!name)
+		return NULL;
+
 	mutex_lock(&host->syncpt_mutex);
 
 	for (i = 0; i < host->info->nb_pts && kref_read(&sp->ref); i++, sp++)
@@ -391,6 +406,7 @@ static void syncpt_release(struct kref *ref)
 	struct host1x_syncpt *sp = container_of(ref, struct host1x_syncpt, ref);
 
 	atomic_set(&sp->max_val, host1x_syncpt_read(sp));
+
 	sp->locked = false;
 
 	mutex_lock(&sp->host->syncpt_mutex);
@@ -560,6 +576,7 @@ static void do_nothing(struct kref *ref)
  *   available for allocation
  *
  * @client: host1x bus client
+ * @syncpt_id: syncpoint ID to make available
  *
  * Makes VBLANK<i> syncpoint available for allocatation if it was
  * reserved at initialization time. This should be called by the display
diff --git a/drivers/gpu/host1x/syncpt.h b/drivers/gpu/host1x/syncpt.h
index b42b4254..d6a4e3fb 100644
--- a/drivers/gpu/host1x/syncpt.h
+++ b/drivers/gpu/host1x/syncpt.h
@@ -41,7 +41,7 @@ struct host1x_syncpt {
 	/* interrupt data */
 	struct host1x_syncpt_intr intr;
 
-	/* 
+	/*
 	 * If a submission incrementing this syncpoint fails, lock it so that
 	 * further submission cannot be made until application has handled the
 	 * failure.
diff --git a/drivers/gpu/host1x/uapi.c b/drivers/gpu/host1x/uapi.c
deleted file mode 100644
index 6a12c183..00000000
--- a/drivers/gpu/host1x/uapi.c
+++ /dev/null
@@ -1,379 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-only
-/*
- * /dev/host1x syncpoint interface
- *
- * Copyright (c) 2020, NVIDIA Corporation.
- */
-
-#include <linux/anon_inodes.h>
-#include <linux/cdev.h>
-#include <linux/file.h>
-#include <linux/fs.h>
-#include <linux/host1x-next.h>
-#include <linux/nospec.h>
-#include <linux/sync_file.h>
-
-#include "dev.h"
-#include "fence.h"
-#include "syncpt.h"
-#include "uapi.h"
-
-#include <uapi/linux/host1x-next.h>
-
-static int syncpt_file_release(struct inode *inode, struct file *file)
-{
-	struct host1x_syncpt *sp = file->private_data;
-
-	host1x_syncpt_put(sp);
-
-	return 0;
-}
-
-static int syncpt_file_ioctl_info(struct host1x_syncpt *sp, void __user *data)
-{
-	struct host1x_syncpoint_info args;
-	unsigned long copy_err;
-
-	copy_err = copy_from_user(&args, data, sizeof(args));
-	if (copy_err)
-		return -EFAULT;
-
-	if (args.reserved[0] || args.reserved[1] || args.reserved[2])
-		return -EINVAL;
-
-	args.id = sp->id;
-
-	copy_err = copy_to_user(data, &args, sizeof(args));
-	if (copy_err)
-		return -EFAULT;
-
-	return 0;
-}
-
-static int syncpt_file_ioctl_incr(struct host1x_syncpt *sp, void __user *data)
-{
-	struct host1x_syncpoint_increment args;
-	unsigned long copy_err;
-	u32 i;
-
-	copy_err = copy_from_user(&args, data, sizeof(args));
-	if (copy_err)
-		return -EFAULT;
-
-	for (i = 0; i < args.count; i++) {
-		host1x_syncpt_incr(sp);
-		if (signal_pending(current))
-			return -EINTR;
-	}
-
-	return 0;
-}
-
-static long syncpt_file_ioctl(struct file *file, unsigned int cmd,
-			      unsigned long arg)
-{
-	void __user *data = (void __user *)arg;
-	long err;
-
-	switch (cmd) {
-	case HOST1X_IOCTL_SYNCPOINT_INFO:
-		err = syncpt_file_ioctl_info(file->private_data, data);
-		break;
-
-	case HOST1X_IOCTL_SYNCPOINT_INCREMENT:
-		err = syncpt_file_ioctl_incr(file->private_data, data);
-		break;
-
-	default:
-		err = -ENOTTY;
-	}
-
-	return err;
-}
-
-static const struct file_operations syncpt_file_fops = {
-	.owner = THIS_MODULE,
-	.release = syncpt_file_release,
-	.unlocked_ioctl = syncpt_file_ioctl,
-	.compat_ioctl = syncpt_file_ioctl,
-};
-
-struct host1x_syncpt *host1x_syncpt_fd_get(int fd)
-{
-	struct host1x_syncpt *sp;
-	struct file *file = fget(fd);
-
-	if (!file)
-		return ERR_PTR(-EINVAL);
-
-	if (file->f_op != &syncpt_file_fops) {
-		fput(file);
-		return ERR_PTR(-EINVAL);
-	}
-
-	sp = file->private_data;
-
-	host1x_syncpt_get(sp);
-
-	fput(file);
-
-	return sp;
-}
-EXPORT_SYMBOL(host1x_syncpt_fd_get);
-
-static int dev_file_open(struct inode *inode, struct file *file)
-{
-	struct host1x_uapi *uapi =
-		container_of(inode->i_cdev, struct host1x_uapi, cdev);
-
-	file->private_data = container_of(uapi, struct host1x, uapi);
-
-	return 0;
-}
-
-static int dev_file_ioctl_read_syncpoint(struct host1x *host1x,
-					 void __user *data)
-{
-	struct host1x_read_syncpoint args;
-	unsigned long copy_err;
-
-	copy_err = copy_from_user(&args, data, sizeof(args));
-	if (copy_err)
-		return -EFAULT;
-
-	if (args.id >= host1x_syncpt_nb_pts(host1x))
-		return -EINVAL;
-
-	args.id = array_index_nospec(args.id, host1x_syncpt_nb_pts(host1x));
-	args.value = host1x_syncpt_read(&host1x->syncpt[args.id]);
-
-	copy_err = copy_to_user(data, &args, sizeof(args));
-	if (copy_err)
-		return -EFAULT;
-
-	return 0;
-}
-
-static int dev_file_ioctl_alloc_syncpoint(struct host1x *host1x,
-					  void __user *data)
-{
-	struct host1x_allocate_syncpoint args;
-	struct host1x_syncpt *sp;
-	unsigned long copy_err;
-	int err;
-
-	copy_err = copy_from_user(&args, data, sizeof(args));
-	if (copy_err)
-		return -EFAULT;
-
-	if (args.reserved[0] || args.reserved[1] || args.reserved[2])
-		return -EINVAL;
-
-	sp = host1x_syncpt_alloc(host1x, HOST1X_SYNCPT_CLIENT_MANAGED,
-				 current->comm);
-	if (!sp)
-		return -EBUSY;
-
-	err = anon_inode_getfd("host1x_syncpt", &syncpt_file_fops, sp,
-			       O_CLOEXEC);
-	if (err < 0)
-		goto free_syncpt;
-
-	args.fd = err;
-
-	copy_err = copy_to_user(data, &args, sizeof(args));
-	if (copy_err) {
-		err = -EFAULT;
-		goto put_fd;
-	}
-
-	return 0;
-
-put_fd:
-	put_unused_fd(args.fd);
-free_syncpt:
-	host1x_syncpt_put(sp);
-
-	return err;
-}
-
-static int dev_file_ioctl_create_fence(struct host1x *host1x, void __user *data)
-{
-	struct host1x_create_fence args;
-	unsigned long copy_err;
-	int fd;
-
-	copy_err = copy_from_user(&args, data, sizeof(args));
-	if (copy_err)
-		return -EFAULT;
-
-	if (args.reserved[0])
-		return -EINVAL;
-
-	if (args.id >= host1x_syncpt_nb_pts(host1x))
-		return -EINVAL;
-
-	args.id = array_index_nospec(args.id, host1x_syncpt_nb_pts(host1x));
-
-	fd = host1x_fence_create_fd(&host1x->syncpt[args.id], args.threshold);
-	if (fd < 0)
-		return fd;
-
-	args.fence_fd = fd;
-
-	copy_err = copy_to_user(data, &args, sizeof(args));
-	if (copy_err)
-		return -EFAULT;
-
-	return 0;
-}
-
-static int dev_file_ioctl_fence_extract(struct host1x *host1x, void __user *data)
-{
-	struct host1x_fence_extract_fence __user *fences_user_ptr;
-	struct dma_fence *fence, **fences;
-	struct host1x_fence_extract args;
-	struct dma_fence_array *array;
-	unsigned int num_fences, i;
-	unsigned long copy_err;
-	int err;
-
-	copy_err = copy_from_user(&args, data, sizeof(args));
-	if (copy_err)
-		return -EFAULT;
-
-	fences_user_ptr = u64_to_user_ptr(args.fences_ptr);
-
-	if (args.reserved[0] || args.reserved[1])
-		return -EINVAL;
-
-	fence = sync_file_get_fence(args.fence_fd);
-	if (!fence)
-		return -EINVAL;
-
-	array = to_dma_fence_array(fence);
-	if (array) {
-		fences = array->fences;
-		num_fences = array->num_fences;
-	} else {
-		fences = &fence;
-		num_fences = 1;
-	}
-
-	for (i = 0; i < min(num_fences, args.num_fences); i++) {
-		struct host1x_fence_extract_fence f;
-
-		err = host1x_fence_extract(fences[i], &f.id, &f.threshold);
-		if (err)
-			goto put_fence;
-
-		copy_err = copy_to_user(fences_user_ptr + i, &f, sizeof(f));
-		if (copy_err) {
-			err = -EFAULT;
-			goto put_fence;
-		}
-	}
-
-	args.num_fences = i+1;
-
-	copy_err = copy_to_user(data, &args, sizeof(args));
-	if (copy_err) {
-		err = -EFAULT;
-		goto put_fence;
-	}
-
-	return 0;
-
-put_fence:
-	dma_fence_put(fence);
-
-	return err;
-}
-
-static long dev_file_ioctl(struct file *file, unsigned int cmd,
-			   unsigned long arg)
-{
-	void __user *data = (void __user *)arg;
-	long err;
-
-	switch (cmd) {
-	case HOST1X_IOCTL_READ_SYNCPOINT:
-		err = dev_file_ioctl_read_syncpoint(file->private_data, data);
-		break;
-
-	case HOST1X_IOCTL_ALLOCATE_SYNCPOINT:
-		err = dev_file_ioctl_alloc_syncpoint(file->private_data, data);
-		break;
-
-	case HOST1X_IOCTL_CREATE_FENCE:
-		err = dev_file_ioctl_create_fence(file->private_data, data);
-		break;
-
-	case HOST1X_IOCTL_FENCE_EXTRACT:
-		err = dev_file_ioctl_fence_extract(file->private_data, data);
-		break;
-
-	default:
-		err = -ENOTTY;
-	}
-
-	return err;
-}
-
-static const struct file_operations dev_file_fops = {
-	.owner = THIS_MODULE,
-	.open = dev_file_open,
-	.unlocked_ioctl = dev_file_ioctl,
-	.compat_ioctl = dev_file_ioctl,
-};
-
-int host1x_uapi_init(struct host1x_uapi *uapi, struct host1x *host1x)
-{
-	int err;
-	dev_t dev_num;
-
-	err = alloc_chrdev_region(&dev_num, 0, 1, "host1x");
-	if (err)
-		return err;
-
-	uapi->class = class_create(THIS_MODULE, "host1x");
-	if (IS_ERR(uapi->class)) {
-		err = PTR_ERR(uapi->class);
-		goto unregister_chrdev_region;
-	}
-
-	cdev_init(&uapi->cdev, &dev_file_fops);
-	err = cdev_add(&uapi->cdev, dev_num, 1);
-	if (err)
-		goto destroy_class;
-
-	uapi->dev = device_create(uapi->class, host1x->dev,
-				  dev_num, NULL, "host1x");
-	if (IS_ERR(uapi->dev)) {
-		err = PTR_ERR(uapi->dev);
-		goto del_cdev;
-	}
-
-	cdev_add(&uapi->cdev, dev_num, 1);
-
-	uapi->dev_num = dev_num;
-
-	return 0;
-
-del_cdev:
-	cdev_del(&uapi->cdev);
-destroy_class:
-	class_destroy(uapi->class);
-unregister_chrdev_region:
-	unregister_chrdev_region(dev_num, 1);
-
-	return err;
-}
-
-void host1x_uapi_deinit(struct host1x_uapi *uapi)
-{
-	device_destroy(uapi->class, uapi->dev_num);
-	cdev_del(&uapi->cdev);
-	class_destroy(uapi->class);
-	unregister_chrdev_region(uapi->dev_num, 1);
-}
diff --git a/drivers/gpu/host1x/uapi.h b/drivers/gpu/host1x/uapi.h
deleted file mode 100644
index 7beb5e44..00000000
--- a/drivers/gpu/host1x/uapi.h
+++ /dev/null
@@ -1,22 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-only */
-/*
- * Copyright (c) 2020, NVIDIA Corporation.
- */
-
-#ifndef HOST1X_UAPI_H
-#define HOST1X_UAPI_H
-
-#include <linux/cdev.h>
-
-struct host1x_uapi {
-	struct class *class;
-
-	struct cdev cdev;
-	struct device *dev;
-	dev_t dev_num;
-};
-
-int host1x_uapi_init(struct host1x_uapi *uapi, struct host1x *host1x);
-void host1x_uapi_deinit(struct host1x_uapi *uapi);
-
-#endif