From d93f79376f210e0b19da57a3dc841ba332daa9d0 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Thu, 23 May 2013 12:10:04 +0200
Subject: drm/radeon: initial VCE support v4
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Only VCE 2.0 support so far.

v2: squashing multiple patches into this one
v3: add IRQ support for CIK, major cleanups,
    basic code documentation
v4: remove HAINAN from chipset list

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/radeon/radeon_kms.c | 1 +
 1 file changed, 1 insertion(+)

(limited to 'drivers/gpu/drm/radeon/radeon_kms.c')

diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 114d1672d616..0e078afc5db7 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -610,6 +610,7 @@ void radeon_driver_preclose_kms(struct drm_device *dev,
 	if (rdev->cmask_filp == file_priv)
 		rdev->cmask_filp = NULL;
 	radeon_uvd_free_handles(rdev, file_priv);
+	radeon_vce_free_handles(rdev, file_priv);
 }
 
 /*
-- 
cgit v1.2.3


From f7ba8b04b22d7c74898f53a0e118f31b9d40dcc2 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Mon, 27 Jan 2014 10:16:06 -0700
Subject: drm/radeon: add VCE ring query
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/radeon/radeon_kms.c | 3 +++
 1 file changed, 3 insertions(+)

(limited to 'drivers/gpu/drm/radeon/radeon_kms.c')

diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 0e078afc5db7..ea018d53a85e 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -433,6 +433,9 @@ static int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 		case RADEON_CS_RING_UVD:
 			*value = rdev->ring[R600_RING_TYPE_UVD_INDEX].ready;
 			break;
+		case RADEON_CS_RING_VCE:
+			*value = rdev->ring[TN_RING_TYPE_VCE1_INDEX].ready;
+			break;
 		default:
 			return -EINVAL;
 		}
-- 
cgit v1.2.3


From 98ccc291ffdc34ccb9b13f0c29cc51d6eab24022 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Thu, 23 Jan 2014 09:50:49 -0700
Subject: drm/radeon: add VCE version parsing and checking
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Also make the result available to userspace.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/radeon/radeon.h     |  2 ++
 drivers/gpu/drm/radeon/radeon_kms.c |  6 ++++
 drivers/gpu/drm/radeon/radeon_vce.c | 56 +++++++++++++++++++++++++++++++++----
 include/uapi/drm/radeon_drm.h       |  4 +++
 4 files changed, 62 insertions(+), 6 deletions(-)

(limited to 'drivers/gpu/drm/radeon/radeon_kms.c')

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index a58a38942c73..d1491d41fc2d 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1606,6 +1606,8 @@ struct radeon_vce {
 	struct radeon_bo	*vcpu_bo;
 	void			*cpu_addr;
 	uint64_t		gpu_addr;
+	unsigned		fw_version;
+	unsigned		fb_version;
 	atomic_t		handles[RADEON_MAX_VCE_HANDLES];
 	struct drm_file		*filp[RADEON_MAX_VCE_HANDLES];
 };
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index ea018d53a85e..baff98be65b1 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -480,6 +480,12 @@ static int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 		else
 			*value = rdev->pm.default_sclk * 10;
 		break;
+	case RADEON_INFO_VCE_FW_VERSION:
+		*value = rdev->vce.fw_version;
+		break;
+	case RADEON_INFO_VCE_FB_VERSION:
+		*value = rdev->vce.fb_version;
+		break;
 	default:
 		DRM_DEBUG_KMS("Invalid request %d\n", info->request);
 		return -EINVAL;
diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c
index 2547d8ea347a..f46563b60921 100644
--- a/drivers/gpu/drm/radeon/radeon_vce.c
+++ b/drivers/gpu/drm/radeon/radeon_vce.c
@@ -48,8 +48,11 @@ MODULE_FIRMWARE(FIRMWARE_BONAIRE);
  */
 int radeon_vce_init(struct radeon_device *rdev)
 {
-	unsigned long bo_size;
-	const char *fw_name;
+	static const char *fw_version = "[ATI LIB=VCEFW,";
+	static const char *fb_version = "[ATI LIB=VCEFWSTATS,";
+	unsigned long size;
+	const char *fw_name, *c;
+	uint8_t start, mid, end;
 	int i, r;
 
 	switch (rdev->family) {
@@ -70,9 +73,50 @@ int radeon_vce_init(struct radeon_device *rdev)
 		return r;
 	}
 
-	bo_size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size) +
-		  RADEON_VCE_STACK_SIZE + RADEON_VCE_HEAP_SIZE;
-	r = radeon_bo_create(rdev, bo_size, PAGE_SIZE, true,
+	/* search for firmware version */
+
+	size = rdev->vce_fw->size - strlen(fw_version) - 9;
+	c = rdev->vce_fw->data;
+	for (;size > 0; --size, ++c)
+		if (strncmp(c, fw_version, strlen(fw_version)) == 0)
+			break;
+
+	if (size == 0)
+		return -EINVAL;
+
+	c += strlen(fw_version);
+	if (sscanf(c, "%2hhd.%2hhd.%2hhd]", &start, &mid, &end) != 3)
+		return -EINVAL;
+
+	/* search for feedback version */
+
+	size = rdev->vce_fw->size - strlen(fb_version) - 3;
+	c = rdev->vce_fw->data;
+	for (;size > 0; --size, ++c)
+		if (strncmp(c, fb_version, strlen(fb_version)) == 0)
+			break;
+
+	if (size == 0)
+		return -EINVAL;
+
+	c += strlen(fb_version);
+	if (sscanf(c, "%2u]", &rdev->vce.fb_version) != 1)
+		return -EINVAL;
+
+	DRM_INFO("Found VCE firmware/feedback version %hhd.%hhd.%hhd / %d!\n",
+		 start, mid, end, rdev->vce.fb_version);
+
+	rdev->vce.fw_version = (start << 24) | (mid << 16) | (end << 8);
+
+	/* we can only work with this fw version for now */
+	if (rdev->vce.fw_version != ((40 << 24) | (2 << 16) | (2 << 8)))
+		return -EINVAL;
+
+	/* load firmware into VRAM */
+
+	size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size) +
+	       RADEON_VCE_STACK_SIZE + RADEON_VCE_HEAP_SIZE;
+	r = radeon_bo_create(rdev, size, PAGE_SIZE, true,
 			     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->vce.vcpu_bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate VCE bo\n", r);
@@ -83,7 +127,7 @@ int radeon_vce_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	memset(rdev->vce.cpu_addr, 0, bo_size);
+	memset(rdev->vce.cpu_addr, 0, size);
 	memcpy(rdev->vce.cpu_addr, rdev->vce_fw->data, rdev->vce_fw->size);
 
 	r = radeon_vce_suspend(rdev);
diff --git a/include/uapi/drm/radeon_drm.h b/include/uapi/drm/radeon_drm.h
index 6493ca55b463..1cf18b4a39ec 100644
--- a/include/uapi/drm/radeon_drm.h
+++ b/include/uapi/drm/radeon_drm.h
@@ -988,6 +988,10 @@ struct drm_radeon_cs {
 #define RADEON_INFO_SI_BACKEND_ENABLED_MASK	0x19
 /* max engine clock - needed for OpenCL */
 #define RADEON_INFO_MAX_SCLK		0x1a
+/* version of VCE firmware */
+#define RADEON_INFO_VCE_FW_VERSION	0x1b
+/* version of VCE feedback */
+#define RADEON_INFO_VCE_FB_VERSION	0x1c
 
 
 struct drm_radeon_info {
-- 
cgit v1.2.3


From bda72d58a20120aee1f78eb17d7eddb955d6696b Mon Sep 17 00:00:00 2001
From: Marek Olšák <marek.olsak@amd.com>
Date: Sun, 2 Mar 2014 00:56:17 +0100
Subject: drm/radeon: add a way to get and set initial buffer domains v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

When passing buffers between processes, the receiving process needs to know
the original buffer domain, so that it doesn't accidentally move the buffer.

v2: reserve the buffer

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/radeon/radeon.h        |  3 +++
 drivers/gpu/drm/radeon/radeon_drv.c    |  3 ++-
 drivers/gpu/drm/radeon/radeon_gem.c    | 36 ++++++++++++++++++++++++++++++++++
 drivers/gpu/drm/radeon/radeon_kms.c    |  1 +
 drivers/gpu/drm/radeon/radeon_object.c |  3 +++
 include/uapi/drm/radeon_drm.h          | 12 ++++++++++++
 6 files changed, 57 insertions(+), 1 deletion(-)

(limited to 'drivers/gpu/drm/radeon/radeon_kms.c')

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 1ac3393bab66..c20d88c93940 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -456,6 +456,7 @@ struct radeon_bo {
 	/* Protected by gem.mutex */
 	struct list_head		list;
 	/* Protected by tbo.reserved */
+	u32				initial_domain;
 	u32				placements[3];
 	struct ttm_placement		placement;
 	struct ttm_buffer_object	tbo;
@@ -2116,6 +2117,8 @@ int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 			      struct drm_file *filp);
 int radeon_gem_va_ioctl(struct drm_device *dev, void *data,
 			  struct drm_file *filp);
+int radeon_gem_op_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *filp);
 int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp);
 int radeon_gem_set_tiling_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *filp);
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index 84a1bbb75f91..4392b7c95ee6 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -79,9 +79,10 @@
  *   2.35.0 - Add CIK macrotile mode array query
  *   2.36.0 - Fix CIK DCE tiling setup
  *   2.37.0 - allow GS ring setup on r6xx/r7xx
+ *   2.38.0 - RADEON_GEM_OP (GET_INITIAL_DOMAIN, SET_INITIAL_DOMAIN)
  */
 #define KMS_DRIVER_MAJOR	2
-#define KMS_DRIVER_MINOR	37
+#define KMS_DRIVER_MINOR	38
 #define KMS_DRIVER_PATCHLEVEL	0
 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
 int radeon_driver_unload_kms(struct drm_device *dev);
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index b96c819024b3..9863ca742494 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -533,6 +533,42 @@ out:
 	return r;
 }
 
+int radeon_gem_op_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *filp)
+{
+	struct drm_radeon_gem_op *args = data;
+	struct drm_gem_object *gobj;
+	struct radeon_bo *robj;
+	int r;
+
+	gobj = drm_gem_object_lookup(dev, filp, args->handle);
+	if (gobj == NULL) {
+		return -ENOENT;
+	}
+	robj = gem_to_radeon_bo(gobj);
+	r = radeon_bo_reserve(robj, false);
+	if (unlikely(r))
+		goto out;
+
+	switch (args->op) {
+	case RADEON_GEM_OP_GET_INITIAL_DOMAIN:
+		args->value = robj->initial_domain;
+		break;
+	case RADEON_GEM_OP_SET_INITIAL_DOMAIN:
+		robj->initial_domain = args->value & (RADEON_GEM_DOMAIN_VRAM |
+						      RADEON_GEM_DOMAIN_GTT |
+						      RADEON_GEM_DOMAIN_CPU);
+		break;
+	default:
+		r = -EINVAL;
+	}
+
+	radeon_bo_unreserve(robj);
+out:
+	drm_gem_object_unreference_unlocked(gobj);
+	return r;
+}
+
 int radeon_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_device *dev,
 			    struct drm_mode_create_dumb *args)
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index baff98be65b1..0b631ebeeb18 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -814,5 +814,6 @@ const struct drm_ioctl_desc radeon_ioctls_kms[] = {
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_GET_TILING, radeon_gem_get_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_BUSY, radeon_gem_busy_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_VA, radeon_gem_va_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(RADEON_GEM_OP, radeon_gem_op_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
 };
 int radeon_max_kms_ioctl = DRM_ARRAY_SIZE(radeon_ioctls_kms);
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index 08595cf90b01..dd12bb4025ac 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -145,6 +145,9 @@ int radeon_bo_create(struct radeon_device *rdev,
 	bo->surface_reg = -1;
 	INIT_LIST_HEAD(&bo->list);
 	INIT_LIST_HEAD(&bo->va);
+	bo->initial_domain = domain & (RADEON_GEM_DOMAIN_VRAM |
+	                               RADEON_GEM_DOMAIN_GTT |
+	                               RADEON_GEM_DOMAIN_CPU);
 	radeon_ttm_placement_from_domain(bo, domain);
 	/* Kernel allocation are uninterruptible */
 	down_read(&rdev->pm.mclk_lock);
diff --git a/include/uapi/drm/radeon_drm.h b/include/uapi/drm/radeon_drm.h
index 1cf18b4a39ec..cb5c93a4c266 100644
--- a/include/uapi/drm/radeon_drm.h
+++ b/include/uapi/drm/radeon_drm.h
@@ -510,6 +510,7 @@ typedef struct {
 #define DRM_RADEON_GEM_GET_TILING	0x29
 #define DRM_RADEON_GEM_BUSY		0x2a
 #define DRM_RADEON_GEM_VA		0x2b
+#define DRM_RADEON_GEM_OP		0x2c
 
 #define DRM_IOCTL_RADEON_CP_INIT    DRM_IOW( DRM_COMMAND_BASE + DRM_RADEON_CP_INIT, drm_radeon_init_t)
 #define DRM_IOCTL_RADEON_CP_START   DRM_IO(  DRM_COMMAND_BASE + DRM_RADEON_CP_START)
@@ -552,6 +553,7 @@ typedef struct {
 #define DRM_IOCTL_RADEON_GEM_GET_TILING	DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_GET_TILING, struct drm_radeon_gem_get_tiling)
 #define DRM_IOCTL_RADEON_GEM_BUSY	DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_BUSY, struct drm_radeon_gem_busy)
 #define DRM_IOCTL_RADEON_GEM_VA		DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_VA, struct drm_radeon_gem_va)
+#define DRM_IOCTL_RADEON_GEM_OP		DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_OP, struct drm_radeon_gem_op)
 
 typedef struct drm_radeon_init {
 	enum {
@@ -884,6 +886,16 @@ struct drm_radeon_gem_pwrite {
 	uint64_t data_ptr;
 };
 
+/* Sets or returns a value associated with a buffer. */
+struct drm_radeon_gem_op {
+	uint32_t	handle; /* buffer */
+	uint32_t	op;     /* RADEON_GEM_OP_* */
+	uint64_t	value;  /* input or return value */
+};
+
+#define RADEON_GEM_OP_GET_INITIAL_DOMAIN	0
+#define RADEON_GEM_OP_SET_INITIAL_DOMAIN	1
+
 #define RADEON_VA_MAP			1
 #define RADEON_VA_UNMAP			2
 
-- 
cgit v1.2.3


From 67e8e3f970ad747d3c854fb40f8ec0cecedd9089 Mon Sep 17 00:00:00 2001
From: Marek Olšák <marek.olsak@amd.com>
Date: Sun, 2 Mar 2014 00:56:18 +0100
Subject: drm/radeon: track memory statistics about VRAM and GTT usage and
 buffer moves v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The statistics are:
- VRAM usage in bytes
- GTT usage in bytes
- number of bytes moved by TTM

The last one is actually a counter, so you need to sample it before and after
command submission and take the difference.

This is useful for finding performance bottlenecks. Userspace queries are
also added.

v2: use atomic64_t

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/radeon/radeon.h        |  4 ++++
 drivers/gpu/drm/radeon/radeon_kms.c    | 15 ++++++++++++++
 drivers/gpu/drm/radeon/radeon_object.c | 36 +++++++++++++++++++++++++++++++++-
 drivers/gpu/drm/radeon/radeon_object.h |  2 +-
 drivers/gpu/drm/radeon/radeon_ttm.c    |  8 +++++++-
 include/uapi/drm/radeon_drm.h          |  3 +++
 6 files changed, 65 insertions(+), 3 deletions(-)

(limited to 'drivers/gpu/drm/radeon/radeon_kms.c')

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index c20d88c93940..7bb8fd96f3ce 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -2309,6 +2309,10 @@ struct radeon_device {
 	/* virtual memory */
 	struct radeon_vm_manager	vm_manager;
 	struct mutex			gpu_clock_mutex;
+	/* memory stats */
+	atomic64_t			vram_usage;
+	atomic64_t			gtt_usage;
+	atomic64_t			num_bytes_moved;
 	/* ACPI interface */
 	struct radeon_atif		atif;
 	struct radeon_atcs		atcs;
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 0b631ebeeb18..806506c03a79 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -486,6 +486,21 @@ static int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 	case RADEON_INFO_VCE_FB_VERSION:
 		*value = rdev->vce.fb_version;
 		break;
+	case RADEON_INFO_NUM_BYTES_MOVED:
+		value = (uint32_t*)&value64;
+		value_size = sizeof(uint64_t);
+		value64 = atomic64_read(&rdev->num_bytes_moved);
+		break;
+	case RADEON_INFO_VRAM_USAGE:
+		value = (uint32_t*)&value64;
+		value_size = sizeof(uint64_t);
+		value64 = atomic64_read(&rdev->vram_usage);
+		break;
+	case RADEON_INFO_GTT_USAGE:
+		value = (uint32_t*)&value64;
+		value_size = sizeof(uint64_t);
+		value64 = atomic64_read(&rdev->gtt_usage);
+		break;
 	default:
 		DRM_DEBUG_KMS("Invalid request %d\n", info->request);
 		return -EINVAL;
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index dd12bb4025ac..282d6a248396 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -56,11 +56,36 @@ static void radeon_bo_clear_va(struct radeon_bo *bo)
 	}
 }
 
+static void radeon_update_memory_usage(struct radeon_bo *bo,
+				       unsigned mem_type, int sign)
+{
+	struct radeon_device *rdev = bo->rdev;
+	u64 size = (u64)bo->tbo.num_pages << PAGE_SHIFT;
+
+	switch (mem_type) {
+	case TTM_PL_TT:
+		if (sign > 0)
+			atomic64_add(size, &rdev->gtt_usage);
+		else
+			atomic64_sub(size, &rdev->gtt_usage);
+		break;
+	case TTM_PL_VRAM:
+		if (sign > 0)
+			atomic64_add(size, &rdev->vram_usage);
+		else
+			atomic64_sub(size, &rdev->vram_usage);
+		break;
+	}
+}
+
 static void radeon_ttm_bo_destroy(struct ttm_buffer_object *tbo)
 {
 	struct radeon_bo *bo;
 
 	bo = container_of(tbo, struct radeon_bo, tbo);
+
+	radeon_update_memory_usage(bo, bo->tbo.mem.mem_type, -1);
+
 	mutex_lock(&bo->rdev->gem.mutex);
 	list_del_init(&bo->list);
 	mutex_unlock(&bo->rdev->gem.mutex);
@@ -567,14 +592,23 @@ int radeon_bo_check_tiling(struct radeon_bo *bo, bool has_moved,
 }
 
 void radeon_bo_move_notify(struct ttm_buffer_object *bo,
-			   struct ttm_mem_reg *mem)
+			   struct ttm_mem_reg *new_mem)
 {
 	struct radeon_bo *rbo;
+
 	if (!radeon_ttm_bo_is_radeon_bo(bo))
 		return;
+
 	rbo = container_of(bo, struct radeon_bo, tbo);
 	radeon_bo_check_tiling(rbo, 0, 1);
 	radeon_vm_bo_invalidate(rbo->rdev, rbo);
+
+	/* update statistics */
+	if (!new_mem)
+		return;
+
+	radeon_update_memory_usage(rbo, bo->mem.mem_type, -1);
+	radeon_update_memory_usage(rbo, new_mem->mem_type, 1);
 }
 
 int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h
index 209b11150263..a9a8c11bd80d 100644
--- a/drivers/gpu/drm/radeon/radeon_object.h
+++ b/drivers/gpu/drm/radeon/radeon_object.h
@@ -151,7 +151,7 @@ extern void radeon_bo_get_tiling_flags(struct radeon_bo *bo,
 extern int radeon_bo_check_tiling(struct radeon_bo *bo, bool has_moved,
 				bool force_drop);
 extern void radeon_bo_move_notify(struct ttm_buffer_object *bo,
-					struct ttm_mem_reg *mem);
+				  struct ttm_mem_reg *new_mem);
 extern int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo);
 extern int radeon_bo_get_surface_reg(struct radeon_bo *bo);
 
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 77f5b0c3edb8..60dfce889ecf 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -406,8 +406,14 @@ static int radeon_bo_move(struct ttm_buffer_object *bo,
 	if (r) {
 memcpy:
 		r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
+		if (r) {
+			return r;
+		}
 	}
-	return r;
+
+	/* update statistics */
+	atomic64_add((u64)bo->num_pages << PAGE_SHIFT, &rdev->num_bytes_moved);
+	return 0;
 }
 
 static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
diff --git a/include/uapi/drm/radeon_drm.h b/include/uapi/drm/radeon_drm.h
index cb5c93a4c266..aefa2f6afa3b 100644
--- a/include/uapi/drm/radeon_drm.h
+++ b/include/uapi/drm/radeon_drm.h
@@ -1004,6 +1004,9 @@ struct drm_radeon_cs {
 #define RADEON_INFO_VCE_FW_VERSION	0x1b
 /* version of VCE feedback */
 #define RADEON_INFO_VCE_FB_VERSION	0x1c
+#define RADEON_INFO_NUM_BYTES_MOVED	0x1d
+#define RADEON_INFO_VRAM_USAGE		0x1e
+#define RADEON_INFO_GTT_USAGE		0x1f
 
 
 struct drm_radeon_info {
-- 
cgit v1.2.3


From f1e3dc708aaadb960b15ee40029f611475f14027 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Thu, 20 Feb 2014 17:34:06 +0100
Subject: drm/radeon: fix missing bo reservation
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/radeon/radeon_kms.c | 6 ++++++
 1 file changed, 6 insertions(+)

(limited to 'drivers/gpu/drm/radeon/radeon_kms.c')

diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 806506c03a79..7a810d0796a0 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -561,6 +561,10 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 
 		radeon_vm_init(rdev, &fpriv->vm);
 
+		r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
+		if (r)
+			return r;
+
 		/* map the ib pool buffer read only into
 		 * virtual address space */
 		bo_va = radeon_vm_bo_add(rdev, &fpriv->vm,
@@ -568,6 +572,8 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 		r = radeon_vm_bo_set_addr(rdev, bo_va, RADEON_VA_IB_OFFSET,
 					  RADEON_VM_PAGE_READABLE |
 					  RADEON_VM_PAGE_SNOOPED);
+
+		radeon_bo_unreserve(rdev->ring_tmp_bo.bo);
 		if (r) {
 			radeon_vm_fini(rdev, &fpriv->vm);
 			kfree(fpriv);
-- 
cgit v1.2.3


From 6d2f2944e95e504a7d33385eeeb9bb7fcca72592 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Thu, 20 Feb 2014 13:42:17 +0100
Subject: drm/radeon: use normal BOs for the page tables v4
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

No need to make it more complicated than necessary,
just allocate the page tables as normal BO and
flush whenever the address change.

v2: update comments and function name
v3: squash bug fixes, page directory and tables patch
v4: rebased on Mareks changes

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/radeon/radeon.h        |  23 +-
 drivers/gpu/drm/radeon/radeon_cs.c     |  44 +--
 drivers/gpu/drm/radeon/radeon_device.c |   1 -
 drivers/gpu/drm/radeon/radeon_kms.c    |   4 +-
 drivers/gpu/drm/radeon/radeon_vm.c     | 492 +++++++++++++++------------------
 5 files changed, 269 insertions(+), 295 deletions(-)

(limited to 'drivers/gpu/drm/radeon/radeon_kms.c')

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 644d922990ef..c31e3c2c1b3a 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -857,17 +857,22 @@ struct radeon_mec {
 #define R600_PTE_READABLE	(1 << 5)
 #define R600_PTE_WRITEABLE	(1 << 6)
 
+struct radeon_vm_pt {
+	struct radeon_bo		*bo;
+	uint64_t			addr;
+};
+
 struct radeon_vm {
-	struct list_head		list;
 	struct list_head		va;
 	unsigned			id;
 
 	/* contains the page directory */
-	struct radeon_sa_bo		*page_directory;
+	struct radeon_bo		*page_directory;
 	uint64_t			pd_gpu_addr;
+	unsigned			max_pde_used;
 
 	/* array of page tables, one for each page directory entry */
-	struct radeon_sa_bo		**page_tables;
+	struct radeon_vm_pt		*page_tables;
 
 	struct mutex			mutex;
 	/* last fence for cs using this vm */
@@ -880,9 +885,7 @@ struct radeon_vm {
 
 struct radeon_vm_manager {
 	struct mutex			lock;
-	struct list_head		lru_vm;
 	struct radeon_fence		*active[RADEON_NUM_VM];
-	struct radeon_sa_manager	sa_manager;
 	uint32_t			max_pfn;
 	/* number of VMIDs */
 	unsigned			nvm;
@@ -1011,6 +1014,7 @@ struct radeon_cs_parser {
 	unsigned		nrelocs;
 	struct radeon_cs_reloc	*relocs;
 	struct radeon_cs_reloc	**relocs_ptr;
+	struct radeon_bo_list	*vm_bos;
 	struct list_head	validated;
 	unsigned		dma_reloc_idx;
 	/* indices of various chunks */
@@ -2798,10 +2802,11 @@ extern void radeon_program_register_sequence(struct radeon_device *rdev,
  */
 int radeon_vm_manager_init(struct radeon_device *rdev);
 void radeon_vm_manager_fini(struct radeon_device *rdev);
-void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm);
+int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm);
 void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm);
-int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm);
-void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm);
+struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev,
+					 struct radeon_vm *vm,
+                                         struct list_head *head);
 struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
 				       struct radeon_vm *vm, int ring);
 void radeon_vm_flush(struct radeon_device *rdev,
@@ -2811,6 +2816,8 @@ void radeon_vm_fence(struct radeon_device *rdev,
 		     struct radeon_vm *vm,
 		     struct radeon_fence *fence);
 uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr);
+int radeon_vm_update_page_directory(struct radeon_device *rdev,
+				    struct radeon_vm *vm);
 int radeon_vm_bo_update(struct radeon_device *rdev,
 			struct radeon_vm *vm,
 			struct radeon_bo *bo,
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index f92df2e8ebdd..420c28dd6a1c 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -168,6 +168,10 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 
 	radeon_cs_buckets_get_list(&buckets, &p->validated);
 
+	if (p->cs_flags & RADEON_CS_USE_VM)
+		p->vm_bos = radeon_vm_get_bos(p->rdev, p->ib.vm,
+					      &p->validated);
+
 	return radeon_bo_list_validate(p->rdev, &p->ticket, &p->validated, p->ring);
 }
 
@@ -401,6 +405,7 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo
 	kfree(parser->track);
 	kfree(parser->relocs);
 	kfree(parser->relocs_ptr);
+	kfree(parser->vm_bos);
 	for (i = 0; i < parser->nchunks; i++)
 		drm_free_large(parser->chunks[i].kdata);
 	kfree(parser->chunks);
@@ -440,24 +445,32 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev,
 	return r;
 }
 
-static int radeon_bo_vm_update_pte(struct radeon_cs_parser *parser,
+static int radeon_bo_vm_update_pte(struct radeon_cs_parser *p,
 				   struct radeon_vm *vm)
 {
-	struct radeon_device *rdev = parser->rdev;
-	struct radeon_bo_list *lobj;
-	struct radeon_bo *bo;
-	int r;
+	struct radeon_device *rdev = p->rdev;
+	int i, r;
 
-	r = radeon_vm_bo_update(rdev, vm, rdev->ring_tmp_bo.bo, &rdev->ring_tmp_bo.bo->tbo.mem);
-	if (r) {
+	r = radeon_vm_update_page_directory(rdev, vm);
+	if (r)
 		return r;
-	}
-	list_for_each_entry(lobj, &parser->validated, tv.head) {
-		bo = lobj->bo;
-		r = radeon_vm_bo_update(parser->rdev, vm, bo, &bo->tbo.mem);
-		if (r) {
+
+	r = radeon_vm_bo_update(rdev, vm, rdev->ring_tmp_bo.bo,
+				&rdev->ring_tmp_bo.bo->tbo.mem);
+	if (r)
+		return r;
+
+	for (i = 0; i < p->nrelocs; i++) {
+		struct radeon_bo *bo;
+
+		/* ignore duplicates */
+		if (p->relocs_ptr[i] != &p->relocs[i])
+			continue;
+
+		bo = p->relocs[i].robj;
+		r = radeon_vm_bo_update(rdev, vm, bo, &bo->tbo.mem);
+		if (r)
 			return r;
-		}
 	}
 	return 0;
 }
@@ -491,10 +504,6 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 
 	mutex_lock(&rdev->vm_manager.lock);
 	mutex_lock(&vm->mutex);
-	r = radeon_vm_alloc_pt(rdev, vm);
-	if (r) {
-		goto out;
-	}
 	r = radeon_bo_vm_update_pte(parser, vm);
 	if (r) {
 		goto out;
@@ -512,7 +521,6 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	}
 
 out:
-	radeon_vm_add_to_lru(rdev, vm);
 	mutex_unlock(&vm->mutex);
 	mutex_unlock(&rdev->vm_manager.lock);
 	return r;
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index fa7841b9d7b6..e58dbabd7d7d 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -1198,7 +1198,6 @@ int radeon_device_init(struct radeon_device *rdev,
 	 * Max GPUVM size for cayman and SI is 40 bits.
 	 */
 	rdev->vm_manager.max_pfn = 1 << 20;
-	INIT_LIST_HEAD(&rdev->vm_manager.lru_vm);
 
 	/* Set asic functions */
 	r = radeon_asic_init(rdev);
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 7a810d0796a0..37b1deacb5b1 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -559,7 +559,9 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 			return -ENOMEM;
 		}
 
-		radeon_vm_init(rdev, &fpriv->vm);
+		r = radeon_vm_init(rdev, &fpriv->vm);
+		if (r)
+			return r;
 
 		r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
 		if (r)
diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c
index 516017689793..44b6918a87ba 100644
--- a/drivers/gpu/drm/radeon/radeon_vm.c
+++ b/drivers/gpu/drm/radeon/radeon_vm.c
@@ -84,84 +84,18 @@ static unsigned radeon_vm_directory_size(struct radeon_device *rdev)
  */
 int radeon_vm_manager_init(struct radeon_device *rdev)
 {
-	struct radeon_vm *vm;
-	struct radeon_bo_va *bo_va;
 	int r;
-	unsigned size;
 
 	if (!rdev->vm_manager.enabled) {
-		/* allocate enough for 2 full VM pts */
-		size = radeon_vm_directory_size(rdev);
-		size += rdev->vm_manager.max_pfn * 8;
-		size *= 2;
-		r = radeon_sa_bo_manager_init(rdev, &rdev->vm_manager.sa_manager,
-					      RADEON_GPU_PAGE_ALIGN(size),
-					      RADEON_VM_PTB_ALIGN_SIZE,
-					      RADEON_GEM_DOMAIN_VRAM);
-		if (r) {
-			dev_err(rdev->dev, "failed to allocate vm bo (%dKB)\n",
-				(rdev->vm_manager.max_pfn * 8) >> 10);
-			return r;
-		}
-
 		r = radeon_asic_vm_init(rdev);
 		if (r)
 			return r;
 
 		rdev->vm_manager.enabled = true;
-
-		r = radeon_sa_bo_manager_start(rdev, &rdev->vm_manager.sa_manager);
-		if (r)
-			return r;
-	}
-
-	/* restore page table */
-	list_for_each_entry(vm, &rdev->vm_manager.lru_vm, list) {
-		if (vm->page_directory == NULL)
-			continue;
-
-		list_for_each_entry(bo_va, &vm->va, vm_list) {
-			bo_va->valid = false;
-		}
 	}
 	return 0;
 }
 
-/**
- * radeon_vm_free_pt - free the page table for a specific vm
- *
- * @rdev: radeon_device pointer
- * @vm: vm to unbind
- *
- * Free the page table of a specific vm (cayman+).
- *
- * Global and local mutex must be lock!
- */
-static void radeon_vm_free_pt(struct radeon_device *rdev,
-				    struct radeon_vm *vm)
-{
-	struct radeon_bo_va *bo_va;
-	int i;
-
-	if (!vm->page_directory)
-		return;
-
-	list_del_init(&vm->list);
-	radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence);
-
-	list_for_each_entry(bo_va, &vm->va, vm_list) {
-		bo_va->valid = false;
-	}
-
-	if (vm->page_tables == NULL)
-		return;
-
-	for (i = 0; i < radeon_vm_num_pdes(rdev); i++)
-		radeon_sa_bo_free(rdev, &vm->page_tables[i], vm->fence);
-
-	kfree(vm->page_tables);
-}
-
 /**
  * radeon_vm_manager_fini - tear down the vm manager
  *
@@ -171,155 +105,59 @@ static void radeon_vm_free_pt(struct radeon_device *rdev,
  */
 void radeon_vm_manager_fini(struct radeon_device *rdev)
 {
-	struct radeon_vm *vm, *tmp;
 	int i;
 
 	if (!rdev->vm_manager.enabled)
 		return;
 
 	mutex_lock(&rdev->vm_manager.lock);
-	/* free all allocated page tables */
-	list_for_each_entry_safe(vm, tmp, &rdev->vm_manager.lru_vm, list) {
-		mutex_lock(&vm->mutex);
-		radeon_vm_free_pt(rdev, vm);
-		mutex_unlock(&vm->mutex);
-	}
-	for (i = 0; i < RADEON_NUM_VM; ++i) {
+	for (i = 0; i < RADEON_NUM_VM; ++i)
 		radeon_fence_unref(&rdev->vm_manager.active[i]);
-	}
 	radeon_asic_vm_fini(rdev);
-	mutex_unlock(&rdev->vm_manager.lock);
-
-	radeon_sa_bo_manager_suspend(rdev, &rdev->vm_manager.sa_manager);
-	radeon_sa_bo_manager_fini(rdev, &rdev->vm_manager.sa_manager);
 	rdev->vm_manager.enabled = false;
+	mutex_unlock(&rdev->vm_manager.lock);
 }
 
 /**
- * radeon_vm_evict - evict page table to make room for new one
- *
- * @rdev: radeon_device pointer
- * @vm: VM we want to allocate something for
+ * radeon_vm_get_bos - add the vm BOs to a validation list
  *
- * Evict a VM from the lru, making sure that it isn't @vm. (cayman+).
- * Returns 0 for success, -ENOMEM for failure.
+ * @vm: vm providing the BOs
+ * @head: head of validation list
  *
- * Global and local mutex must be locked!
+ * Add the page directory to the list of BOs to
+ * validate for command submission (cayman+).
  */
-static int radeon_vm_evict(struct radeon_device *rdev, struct radeon_vm *vm)
+struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev,
+					 struct radeon_vm *vm,
+					 struct list_head *head)
 {
-	struct radeon_vm *vm_evict;
+	struct radeon_bo_list *list;
+	unsigned i, idx, size;
 
-	if (list_empty(&rdev->vm_manager.lru_vm))
-		return -ENOMEM;
-
-	vm_evict = list_first_entry(&rdev->vm_manager.lru_vm,
-				    struct radeon_vm, list);
-	if (vm_evict == vm)
-		return -ENOMEM;
-
-	mutex_lock(&vm_evict->mutex);
-	radeon_vm_free_pt(rdev, vm_evict);
-	mutex_unlock(&vm_evict->mutex);
-	return 0;
-}
-
-/**
- * radeon_vm_alloc_pt - allocates a page table for a VM
- *
- * @rdev: radeon_device pointer
- * @vm: vm to bind
- *
- * Allocate a page table for the requested vm (cayman+).
- * Returns 0 for success, error for failure.
- *
- * Global and local mutex must be locked!
- */
-int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm)
-{
-	unsigned pd_size, pd_entries, pts_size;
-	struct radeon_ib ib;
-	int r;
-
-	if (vm == NULL) {
-		return -EINVAL;
-	}
-
-	if (vm->page_directory != NULL) {
-		return 0;
-	}
-
-	pd_size = radeon_vm_directory_size(rdev);
-	pd_entries = radeon_vm_num_pdes(rdev);
-
-retry:
-	r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager,
-			     &vm->page_directory, pd_size,
-			     RADEON_VM_PTB_ALIGN_SIZE, false);
-	if (r == -ENOMEM) {
-		r = radeon_vm_evict(rdev, vm);
-		if (r)
-			return r;
-		goto retry;
-
-	} else if (r) {
-		return r;
-	}
-
-	vm->pd_gpu_addr = radeon_sa_bo_gpu_addr(vm->page_directory);
-
-	/* Initially clear the page directory */
-	r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib,
-			  NULL, pd_entries * 2 + 64);
-	if (r) {
-		radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence);
-		return r;
-	}
-
-	ib.length_dw = 0;
+	size = (radeon_vm_num_pdes(rdev) + 1) * sizeof(struct radeon_bo_list);
+	list = kmalloc(size, GFP_KERNEL);
+	if (!list)
+		return NULL;
 
-	radeon_asic_vm_set_page(rdev, &ib, vm->pd_gpu_addr,
-				0, pd_entries, 0, 0);
+	/* add the vm page table to the list */
+	list[0].bo = vm->page_directory;
+	list[0].domain = RADEON_GEM_DOMAIN_VRAM;
+	list[0].alt_domain = RADEON_GEM_DOMAIN_VRAM;
+	list[0].tv.bo = &vm->page_directory->tbo;
+	list_add(&list[0].tv.head, head);
 
-	radeon_semaphore_sync_to(ib.semaphore, vm->fence);
-	r = radeon_ib_schedule(rdev, &ib, NULL);
-	if (r) {
-		radeon_ib_free(rdev, &ib);
-		radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence);
-		return r;
-	}
-	radeon_fence_unref(&vm->fence);
-	vm->fence = radeon_fence_ref(ib.fence);
-	radeon_ib_free(rdev, &ib);
-	radeon_fence_unref(&vm->last_flush);
-
-	/* allocate page table array */
-	pts_size = radeon_vm_num_pdes(rdev) * sizeof(struct radeon_sa_bo *);
-	vm->page_tables = kzalloc(pts_size, GFP_KERNEL);
+	for (i = 0, idx = 1; i <= vm->max_pde_used; i++) {
+		if (!vm->page_tables[i].bo)
+			continue;
 
-	if (vm->page_tables == NULL) {
-		DRM_ERROR("Cannot allocate memory for page table array\n");
-		radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence);
-		return -ENOMEM;
+		list[idx].bo = vm->page_tables[i].bo;
+		list[idx].domain = RADEON_GEM_DOMAIN_VRAM;
+		list[idx].alt_domain = RADEON_GEM_DOMAIN_VRAM;
+		list[idx].tv.bo = &list[idx].bo->tbo;
+		list_add(&list[idx++].tv.head, head);
 	}
 
-	return 0;
-}
-
-/**
- * radeon_vm_add_to_lru - add VMs page table to LRU list
- *
- * @rdev: radeon_device pointer
- * @vm: vm to add to LRU
- *
- * Add the allocated page table to the LRU list (cayman+).
- *
- * Global mutex must be locked!
- */
-void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm)
-{
-	list_del_init(&vm->list);
-	list_add_tail(&vm->list, &rdev->vm_manager.lru_vm);
+	return list;
 }
 
 /**
@@ -393,10 +231,14 @@ void radeon_vm_flush(struct radeon_device *rdev,
 		     struct radeon_vm *vm,
 		     int ring)
 {
+	uint64_t pd_addr = radeon_bo_gpu_offset(vm->page_directory);
+
 	/* if we can't remember our last VM flush then flush now! */
 	/* XXX figure out why we have to flush all the time */
-	if (!vm->last_flush || true)
+	if (!vm->last_flush || true || pd_addr != vm->pd_gpu_addr) {
+		vm->pd_gpu_addr = pd_addr;
 		radeon_ring_vm_flush(rdev, ring, vm);
+	}
 }
 
 /**
@@ -495,6 +337,63 @@ struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev,
 	return bo_va;
 }
 
+/**
+ * radeon_vm_clear_bo - initially clear the page dir/table
+ *
+ * @rdev: radeon_device pointer
+ * @bo: bo to clear
+ */
+static int radeon_vm_clear_bo(struct radeon_device *rdev,
+			      struct radeon_bo *bo)
+{
+        struct ttm_validate_buffer tv;
+        struct ww_acquire_ctx ticket;
+        struct list_head head;
+	struct radeon_ib ib;
+	unsigned entries;
+	uint64_t addr;
+	int r;
+
+        memset(&tv, 0, sizeof(tv));
+        tv.bo = &bo->tbo;
+
+        INIT_LIST_HEAD(&head);
+        list_add(&tv.head, &head);
+
+        r = ttm_eu_reserve_buffers(&ticket, &head);
+        if (r)
+		return r;
+
+        r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+        if (r)
+                goto error;
+
+	addr = radeon_bo_gpu_offset(bo);
+	entries = radeon_bo_size(bo) / 8;
+
+	r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib,
+			  NULL, entries * 2 + 64);
+	if (r)
+                goto error;
+
+	ib.length_dw = 0;
+
+	radeon_asic_vm_set_page(rdev, &ib, addr, 0, entries, 0, 0);
+
+	r = radeon_ib_schedule(rdev, &ib, NULL);
+	if (r)
+                goto error;
+
+	ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence);
+	radeon_ib_free(rdev, &ib);
+
+	return 0;
+
+error:
+	ttm_eu_backoff_reservation(&ticket, &head);
+	return r;
+}
+
 /**
  * radeon_vm_bo_set_addr - set bos virtual address inside a vm
  *
@@ -519,7 +418,8 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 	struct radeon_vm *vm = bo_va->vm;
 	struct radeon_bo_va *tmp;
 	struct list_head *head;
-	unsigned last_pfn;
+	unsigned last_pfn, pt_idx;
+	int r;
 
 	if (soffset) {
 		/* make sure object fit at this offset */
@@ -570,8 +470,53 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 	bo_va->valid = false;
 	list_move(&bo_va->vm_list, head);
 
+	soffset = (soffset / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE;
+	eoffset = (eoffset / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE;
+
+	if (eoffset > vm->max_pde_used)
+		vm->max_pde_used = eoffset;
+
+	radeon_bo_unreserve(bo_va->bo);
+
+	/* walk over the address space and allocate the page tables */
+	for (pt_idx = soffset; pt_idx <= eoffset; ++pt_idx) {
+		struct radeon_bo *pt;
+
+		if (vm->page_tables[pt_idx].bo)
+			continue;
+
+		/* drop mutex to allocate and clear page table */
+		mutex_unlock(&vm->mutex);
+
+		r = radeon_bo_create(rdev, RADEON_VM_PTE_COUNT * 8,
+				     RADEON_GPU_PAGE_SIZE, false, 
+				     RADEON_GEM_DOMAIN_VRAM, NULL, &pt);
+		if (r)
+			return r;
+
+		r = radeon_vm_clear_bo(rdev, pt);
+		if (r) {
+			radeon_bo_unref(&pt);
+			radeon_bo_reserve(bo_va->bo, false);
+			return r;
+		}
+
+		/* aquire mutex again */
+		mutex_lock(&vm->mutex);
+		if (vm->page_tables[pt_idx].bo) {
+			/* someone else allocated the pt in the meantime */
+			mutex_unlock(&vm->mutex);
+			radeon_bo_unref(&pt);
+			mutex_lock(&vm->mutex);
+			continue;
+		}
+
+		vm->page_tables[pt_idx].addr = 0;
+		vm->page_tables[pt_idx].bo = pt;
+	}
+
 	mutex_unlock(&vm->mutex);
-	return 0;
+	return radeon_bo_reserve(bo_va->bo, false);
 }
 
 /**
@@ -631,58 +576,53 @@ static uint32_t radeon_vm_page_flags(uint32_t flags)
  *
  * Global and local mutex must be locked!
  */
-static int radeon_vm_update_pdes(struct radeon_device *rdev,
-				 struct radeon_vm *vm,
-				 struct radeon_ib *ib,
-				 uint64_t start, uint64_t end)
+int radeon_vm_update_page_directory(struct radeon_device *rdev,
+				    struct radeon_vm *vm)
 {
 	static const uint32_t incr = RADEON_VM_PTE_COUNT * 8;
 
+	uint64_t pd_addr = radeon_bo_gpu_offset(vm->page_directory);
 	uint64_t last_pde = ~0, last_pt = ~0;
-	unsigned count = 0;
-	uint64_t pt_idx;
+	unsigned count = 0, pt_idx, ndw;
+	struct radeon_ib ib;
 	int r;
 
-	start = (start / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE;
-	end = (end / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE;
+	/* padding, etc. */
+	ndw = 64;
+
+	/* assume the worst case */
+	ndw += vm->max_pde_used * 12;
+
+	/* update too big for an IB */
+	if (ndw > 0xfffff)
+		return -ENOMEM;
+
+	r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, ndw * 4);
+	if (r)
+		return r;
+	ib.length_dw = 0;
 
 	/* walk over the address space and update the page directory */
-	for (pt_idx = start; pt_idx <= end; ++pt_idx) {
+	for (pt_idx = 0; pt_idx <= vm->max_pde_used; ++pt_idx) {
+		struct radeon_bo *bo = vm->page_tables[pt_idx].bo;
 		uint64_t pde, pt;
 
-		if (vm->page_tables[pt_idx])
+		if (bo == NULL)
 			continue;
 
-retry:
-		r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager,
-				     &vm->page_tables[pt_idx],
-				     RADEON_VM_PTE_COUNT * 8,
-				     RADEON_GPU_PAGE_SIZE, false);
-
-		if (r == -ENOMEM) {
-			r = radeon_vm_evict(rdev, vm);
-			if (r)
-				return r;
-			goto retry;
-		} else if (r) {
-			return r;
-		}
-
-		pde = vm->pd_gpu_addr + pt_idx * 8;
-
-		pt = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]);
+		pt = radeon_bo_gpu_offset(bo);
+		if (vm->page_tables[pt_idx].addr == pt)
+			continue;
+		vm->page_tables[pt_idx].addr = pt;
 
+		pde = pd_addr + pt_idx * 8;
 		if (((last_pde + 8 * count) != pde) ||
 		    ((last_pt + incr * count) != pt)) {
 
 			if (count) {
-				radeon_asic_vm_set_page(rdev, ib, last_pde,
+				radeon_asic_vm_set_page(rdev, &ib, last_pde,
 							last_pt, count, incr,
 							R600_PTE_VALID);
-
-				count *= RADEON_VM_PTE_COUNT;
-				radeon_asic_vm_set_page(rdev, ib, last_pt, 0,
-							count, 0, 0);
 			}
 
 			count = 1;
@@ -693,14 +633,22 @@ retry:
 		}
 	}
 
-	if (count) {
-		radeon_asic_vm_set_page(rdev, ib, last_pde, last_pt, count,
+	if (count)
+		radeon_asic_vm_set_page(rdev, &ib, last_pde, last_pt, count,
 					incr, R600_PTE_VALID);
 
-		count *= RADEON_VM_PTE_COUNT;
-		radeon_asic_vm_set_page(rdev, ib, last_pt, 0,
-					count, 0, 0);
+	if (ib.length_dw != 0) {
+		radeon_semaphore_sync_to(ib.semaphore, vm->last_id_use);
+		r = radeon_ib_schedule(rdev, &ib, NULL);
+		if (r) {
+			radeon_ib_free(rdev, &ib);
+			return r;
+		}
+		radeon_fence_unref(&vm->fence);
+		vm->fence = radeon_fence_ref(ib.fence);
+		radeon_fence_unref(&vm->last_flush);
 	}
+	radeon_ib_free(rdev, &ib);
 
 	return 0;
 }
@@ -745,7 +693,7 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
 		else
 			nptes = RADEON_VM_PTE_COUNT - (addr & mask);
 
-		pte = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]);
+		pte = radeon_bo_gpu_offset(vm->page_tables[pt_idx].bo);
 		pte += (addr & mask) * 8;
 
 		if ((last_pte + 8 * count) != pte) {
@@ -795,14 +743,10 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 {
 	struct radeon_ib ib;
 	struct radeon_bo_va *bo_va;
-	unsigned nptes, npdes, ndw;
+	unsigned nptes, ndw;
 	uint64_t addr;
 	int r;
 
-	/* nothing to do if vm isn't bound */
-	if (vm->page_directory == NULL)
-		return 0;
-
 	bo_va = radeon_vm_bo_find(vm, bo);
 	if (bo_va == NULL) {
 		dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm);
@@ -840,9 +784,6 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 
 	nptes = radeon_bo_ngpu_pages(bo);
 
-	/* assume two extra pdes in case the mapping overlaps the borders */
-	npdes = (nptes >> RADEON_VM_BLOCK_SIZE) + 2;
-
 	/* padding, etc. */
 	ndw = 64;
 
@@ -857,15 +798,6 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 	/* reserve space for pte addresses */
 	ndw += nptes * 2;
 
-	/* reserve space for one header for every 2k dwords */
-	ndw += (npdes >> 11) * 4;
-
-	/* reserve space for pde addresses */
-	ndw += npdes * 2;
-
-	/* reserve space for clearing new page tables */
-	ndw += npdes * 2 * RADEON_VM_PTE_COUNT;
-
 	/* update too big for an IB */
 	if (ndw > 0xfffff)
 		return -ENOMEM;
@@ -875,12 +807,6 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 		return r;
 	ib.length_dw = 0;
 
-	r = radeon_vm_update_pdes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset);
-	if (r) {
-		radeon_ib_free(rdev, &ib);
-		return r;
-	}
-
 	radeon_vm_update_ptes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset,
 			      addr, radeon_vm_page_flags(bo_va->flags));
 
@@ -957,15 +883,43 @@ void radeon_vm_bo_invalidate(struct radeon_device *rdev,
  *
  * Init @vm fields (cayman+).
  */
-void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
+int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
 {
+	unsigned pd_size, pd_entries, pts_size;
+	int r;
+
 	vm->id = 0;
 	vm->fence = NULL;
 	vm->last_flush = NULL;
 	vm->last_id_use = NULL;
 	mutex_init(&vm->mutex);
-	INIT_LIST_HEAD(&vm->list);
 	INIT_LIST_HEAD(&vm->va);
+
+	pd_size = radeon_vm_directory_size(rdev);
+	pd_entries = radeon_vm_num_pdes(rdev);
+
+	/* allocate page table array */
+	pts_size = pd_entries * sizeof(struct radeon_vm_pt);
+	vm->page_tables = kzalloc(pts_size, GFP_KERNEL);
+	if (vm->page_tables == NULL) {
+		DRM_ERROR("Cannot allocate memory for page table array\n");
+		return -ENOMEM;
+	}
+
+	r = radeon_bo_create(rdev, pd_size, RADEON_VM_PTB_ALIGN_SIZE, false,
+			     RADEON_GEM_DOMAIN_VRAM, NULL,
+			     &vm->page_directory);
+	if (r)
+		return r;
+
+	r = radeon_vm_clear_bo(rdev, vm->page_directory);
+	if (r) {
+		radeon_bo_unref(&vm->page_directory);
+		vm->page_directory = NULL;
+		return r;
+	}
+
+	return 0;
 }
 
 /**
@@ -980,12 +934,7 @@ void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
 void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
 {
 	struct radeon_bo_va *bo_va, *tmp;
-	int r;
-
-	mutex_lock(&rdev->vm_manager.lock);
-	mutex_lock(&vm->mutex);
-	radeon_vm_free_pt(rdev, vm);
-	mutex_unlock(&rdev->vm_manager.lock);
+	int i, r;
 
 	if (!list_empty(&vm->va)) {
 		dev_err(rdev->dev, "still active bo inside vm\n");
@@ -999,8 +948,17 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
 			kfree(bo_va);
 		}
 	}
+
+
+	for (i = 0; i < radeon_vm_num_pdes(rdev); i++)
+		radeon_bo_unref(&vm->page_tables[i].bo);
+	kfree(vm->page_tables);
+
+	radeon_bo_unref(&vm->page_directory);
+
 	radeon_fence_unref(&vm->fence);
 	radeon_fence_unref(&vm->last_flush);
 	radeon_fence_unref(&vm->last_id_use);
-	mutex_unlock(&vm->mutex);
+
+	mutex_destroy(&vm->mutex);
 }
-- 
cgit v1.2.3


From e84c20aff1ce7493bce26b75f1db363bb3f05979 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Thu, 20 Mar 2014 11:08:11 +1000
Subject: drm/radeon/kms: merge conflicted badly

Not sure why git didn't flag this, but the result of automerge
from 3.14-rc7 screwed up the radeon init procedure.

Reported-by: Fireburn on #radeon
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/radeon/radeon_kms.c | 4 ----
 1 file changed, 4 deletions(-)

(limited to 'drivers/gpu/drm/radeon/radeon_kms.c')

diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 6f1dfac17507..3e49342a20e6 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -575,10 +575,6 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 		if (r)
 			return r;
 
-		r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
-		if (r)
-			return r;
-
 		/* map the ib pool buffer read only into
 		 * virtual address space */
 		bo_va = radeon_vm_bo_add(rdev, &fpriv->vm,
-- 
cgit v1.2.3


From 90c4cde9d5a2bb6239cb3e253bb3832ed89dc75c Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexdeucher@gmail.com>
Date: Thu, 10 Apr 2014 22:29:01 -0400
Subject: drm/radeon: fix runpm handling on APUs (v4)

Don't try and runtime suspend the APU in PX systems.  We
only want to power down the dGPU.

v2: fix harder
v3: fix stupid typo
v4: consolidate runpm enablement to a single flag

bugs:
https://bugs.freedesktop.org/show_bug.cgi?id=75127
https://bugzilla.kernel.org/show_bug.cgi?id=72701

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/radeon/radeon.h              |  1 +
 drivers/gpu/drm/radeon/radeon_atpx_handler.c |  2 +-
 drivers/gpu/drm/radeon/radeon_device.c       | 19 ++++++++++---------
 drivers/gpu/drm/radeon/radeon_drv.c          | 24 ++++--------------------
 drivers/gpu/drm/radeon/radeon_family.h       |  1 +
 drivers/gpu/drm/radeon/radeon_kms.c          | 14 ++++++++++----
 6 files changed, 27 insertions(+), 34 deletions(-)

(limited to 'drivers/gpu/drm/radeon/radeon_kms.c')

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index f21db7a0b34d..7014bdd688ce 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -2321,6 +2321,7 @@ struct radeon_device {
 	bool have_disp_power_ref;
 };
 
+bool radeon_is_px(struct drm_device *dev);
 int radeon_device_init(struct radeon_device *rdev,
 		       struct drm_device *ddev,
 		       struct pci_dev *pdev,
diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
index fa9a9c02751e..dedea72f48c4 100644
--- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c
+++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
@@ -59,7 +59,7 @@ struct atpx_mux {
 	u16 mux;
 } __packed;
 
-bool radeon_is_px(void) {
+bool radeon_has_atpx(void) {
 	return radeon_atpx_priv.atpx_detected;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 835516d2d257..511fe26198e4 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -102,11 +102,14 @@ static const char radeon_family_name[][16] = {
 	"LAST",
 };
 
-#if defined(CONFIG_VGA_SWITCHEROO)
-bool radeon_is_px(void);
-#else
-static inline bool radeon_is_px(void) { return false; }
-#endif
+bool radeon_is_px(struct drm_device *dev)
+{
+	struct radeon_device *rdev = dev->dev_private;
+
+	if (rdev->flags & RADEON_IS_PX)
+		return true;
+	return false;
+}
 
 /**
  * radeon_program_register_sequence - program an array of registers.
@@ -1082,7 +1085,7 @@ static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero
 {
 	struct drm_device *dev = pci_get_drvdata(pdev);
 
-	if (radeon_is_px() && state == VGA_SWITCHEROO_OFF)
+	if (radeon_is_px(dev) && state == VGA_SWITCHEROO_OFF)
 		return;
 
 	if (state == VGA_SWITCHEROO_ON) {
@@ -1301,9 +1304,7 @@ int radeon_device_init(struct radeon_device *rdev,
 	 * ignore it */
 	vga_client_register(rdev->pdev, rdev, NULL, radeon_vga_set_decode);
 
-	if (radeon_runtime_pm == 1)
-		runtime = true;
-	if ((radeon_runtime_pm == -1) && radeon_is_px())
+	if (rdev->flags & RADEON_IS_PX)
 		runtime = true;
 	vga_switcheroo_register_client(rdev->pdev, &radeon_switcheroo_ops, runtime);
 	if (runtime)
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index d0eba48dd74e..25127ba44ed9 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -115,6 +115,7 @@ extern int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc,
 				      unsigned int flags,
 				      int *vpos, int *hpos, ktime_t *stime,
 				      ktime_t *etime);
+extern bool radeon_is_px(struct drm_device *dev);
 extern const struct drm_ioctl_desc radeon_ioctls_kms[];
 extern int radeon_max_kms_ioctl;
 int radeon_mmap(struct file *filp, struct vm_area_struct *vma);
@@ -144,11 +145,9 @@ void radeon_debugfs_cleanup(struct drm_minor *minor);
 #if defined(CONFIG_VGA_SWITCHEROO)
 void radeon_register_atpx_handler(void);
 void radeon_unregister_atpx_handler(void);
-bool radeon_is_px(void);
 #else
 static inline void radeon_register_atpx_handler(void) {}
 static inline void radeon_unregister_atpx_handler(void) {}
-static inline bool radeon_is_px(void) { return false; }
 #endif
 
 int radeon_no_wb;
@@ -405,12 +404,7 @@ static int radeon_pmops_runtime_suspend(struct device *dev)
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
 	int ret;
 
-	if (radeon_runtime_pm == 0) {
-		pm_runtime_forbid(dev);
-		return -EBUSY;
-	}
-
-	if (radeon_runtime_pm == -1 && !radeon_is_px()) {
+	if (!radeon_is_px(drm_dev)) {
 		pm_runtime_forbid(dev);
 		return -EBUSY;
 	}
@@ -434,10 +428,7 @@ static int radeon_pmops_runtime_resume(struct device *dev)
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
 	int ret;
 
-	if (radeon_runtime_pm == 0)
-		return -EINVAL;
-
-	if (radeon_runtime_pm == -1 && !radeon_is_px())
+	if (!radeon_is_px(drm_dev))
 		return -EINVAL;
 
 	drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
@@ -462,14 +453,7 @@ static int radeon_pmops_runtime_idle(struct device *dev)
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
 	struct drm_crtc *crtc;
 
-	if (radeon_runtime_pm == 0) {
-		pm_runtime_forbid(dev);
-		return -EBUSY;
-	}
-
-	/* are we PX enabled? */
-	if (radeon_runtime_pm == -1 && !radeon_is_px()) {
-		DRM_DEBUG_DRIVER("failing to power off - not px\n");
+	if (!radeon_is_px(drm_dev)) {
 		pm_runtime_forbid(dev);
 		return -EBUSY;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_family.h b/drivers/gpu/drm/radeon/radeon_family.h
index 614ad549297f..9da5da4ffd17 100644
--- a/drivers/gpu/drm/radeon/radeon_family.h
+++ b/drivers/gpu/drm/radeon/radeon_family.h
@@ -115,6 +115,7 @@ enum radeon_chip_flags {
 	RADEON_NEW_MEMMAP = 0x00400000UL,
 	RADEON_IS_PCI = 0x00800000UL,
 	RADEON_IS_IGPGART = 0x01000000UL,
+	RADEON_IS_PX = 0x02000000UL,
 };
 
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 3e49342a20e6..9337820e4d37 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -35,9 +35,9 @@
 #include <linux/pm_runtime.h>
 
 #if defined(CONFIG_VGA_SWITCHEROO)
-bool radeon_is_px(void);
+bool radeon_has_atpx(void);
 #else
-static inline bool radeon_is_px(void) { return false; }
+static inline bool radeon_has_atpx(void) { return false; }
 #endif
 
 /**
@@ -107,6 +107,13 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags)
 		flags |= RADEON_IS_PCI;
 	}
 
+	if (radeon_runtime_pm == 1)
+		flags |= RADEON_IS_PX;
+	else if ((radeon_runtime_pm == -1) &&
+		 radeon_has_atpx() &&
+		 ((flags & RADEON_IS_IGP) == 0))
+		flags |= RADEON_IS_PX;
+
 	/* radeon_device_init should report only fatal error
 	 * like memory allocation failure or iomapping failure,
 	 * or memory manager initialization failure, it must
@@ -137,8 +144,7 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags)
 				"Error during ACPI methods call\n");
 	}
 
-	if ((radeon_runtime_pm == 1) ||
-	    ((radeon_runtime_pm == -1) && radeon_is_px())) {
+	if (radeon_is_px(dev)) {
 		pm_runtime_use_autosuspend(dev->dev);
 		pm_runtime_set_autosuspend_delay(dev->dev, 5000);
 		pm_runtime_set_active(dev->dev);
-- 
cgit v1.2.3


From 74073c9dd29905645feb6dee03c144657a9844cd Mon Sep 17 00:00:00 2001
From: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Date: Tue, 18 Mar 2014 17:16:52 +0100
Subject: drm/radeon: memory leak on bo reservation failure. v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

On bo reservation failure, we end up leaking fpriv.

v2 (chk): rebased and added missing free on vm failure as well

Fixes: 5e386b574cf7e1 ("drm/radeon: fix missing bo reservation")
Cc: stable@vger.kernel.org
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/radeon/radeon_kms.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

(limited to 'drivers/gpu/drm/radeon/radeon_kms.c')

diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 9337820e4d37..fb3d13f693dd 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -574,12 +574,17 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 		}
 
 		r = radeon_vm_init(rdev, &fpriv->vm);
-		if (r)
+		if (r) {
+			kfree(fpriv);
 			return r;
+		}
 
 		r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
-		if (r)
+		if (r) {
+			radeon_vm_fini(rdev, &fpriv->vm);
+			kfree(fpriv);
 			return r;
+		}
 
 		/* map the ib pool buffer read only into
 		 * virtual address space */
-- 
cgit v1.2.3