* [igt-dev] [PATCH i-g-t v2 1/4] include: bump drm uAPI headers
2018-10-15 16:20 [igt-dev] [PATCH i-g-t v2 0/4] tests/perf: add support for new oa buffer size parameter Lionel Landwerlin
@ 2018-10-15 16:20 ` Lionel Landwerlin
2018-10-22 18:36 ` Matthew Auld
2018-10-15 16:20 ` [igt-dev] [PATCH i-g-t v2 2/4] tests/perf: buffer-fill: drop load helper usage Lionel Landwerlin
` (4 subsequent siblings)
5 siblings, 1 reply; 11+ messages in thread
From: Lionel Landwerlin @ 2018-10-15 16:20 UTC (permalink / raw)
To: igt-dev
To be updated with a proper release.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
---
include/drm-uapi/amdgpu_drm.h | 52 +++-
include/drm-uapi/drm.h | 16 ++
include/drm-uapi/drm_fourcc.h | 215 ++++++++++++++
include/drm-uapi/drm_mode.h | 26 +-
include/drm-uapi/etnaviv_drm.h | 6 +
include/drm-uapi/exynos_drm.h | 240 ++++++++++++++++
include/drm-uapi/i915_drm.h | 56 ++++
include/drm-uapi/msm_drm.h | 2 +
include/drm-uapi/tegra_drm.h | 492 ++++++++++++++++++++++++++++++++-
include/drm-uapi/vc4_drm.h | 13 +-
include/drm-uapi/virtgpu_drm.h | 1 +
include/drm-uapi/vmwgfx_drm.h | 166 ++++++++---
12 files changed, 1229 insertions(+), 56 deletions(-)
diff --git a/include/drm-uapi/amdgpu_drm.h b/include/drm-uapi/amdgpu_drm.h
index 1816bd82..370e9a55 100644
--- a/include/drm-uapi/amdgpu_drm.h
+++ b/include/drm-uapi/amdgpu_drm.h
@@ -72,12 +72,41 @@ extern "C" {
#define DRM_IOCTL_AMDGPU_FENCE_TO_HANDLE DRM_IOWR(DRM_COMMAND_BASE + DRM_AMDGPU_FENCE_TO_HANDLE, union drm_amdgpu_fence_to_handle)
#define DRM_IOCTL_AMDGPU_SCHED DRM_IOW(DRM_COMMAND_BASE + DRM_AMDGPU_SCHED, union drm_amdgpu_sched)
+/**
+ * DOC: memory domains
+ *
+ * %AMDGPU_GEM_DOMAIN_CPU System memory that is not GPU accessible.
+ * Memory in this pool could be swapped out to disk if there is pressure.
+ *
+ * %AMDGPU_GEM_DOMAIN_GTT GPU accessible system memory, mapped into the
+ * GPU's virtual address space via gart. Gart memory linearizes non-contiguous
+ * pages of system memory, allows GPU access system memory in a linezrized
+ * fashion.
+ *
+ * %AMDGPU_GEM_DOMAIN_VRAM Local video memory. For APUs, it is memory
+ * carved out by the BIOS.
+ *
+ * %AMDGPU_GEM_DOMAIN_GDS Global on-chip data storage used to share data
+ * across shader threads.
+ *
+ * %AMDGPU_GEM_DOMAIN_GWS Global wave sync, used to synchronize the
+ * execution of all the waves on a device.
+ *
+ * %AMDGPU_GEM_DOMAIN_OA Ordered append, used by 3D or Compute engines
+ * for appending data.
+ */
#define AMDGPU_GEM_DOMAIN_CPU 0x1
#define AMDGPU_GEM_DOMAIN_GTT 0x2
#define AMDGPU_GEM_DOMAIN_VRAM 0x4
#define AMDGPU_GEM_DOMAIN_GDS 0x8
#define AMDGPU_GEM_DOMAIN_GWS 0x10
#define AMDGPU_GEM_DOMAIN_OA 0x20
+#define AMDGPU_GEM_DOMAIN_MASK (AMDGPU_GEM_DOMAIN_CPU | \
+ AMDGPU_GEM_DOMAIN_GTT | \
+ AMDGPU_GEM_DOMAIN_VRAM | \
+ AMDGPU_GEM_DOMAIN_GDS | \
+ AMDGPU_GEM_DOMAIN_GWS | \
+ AMDGPU_GEM_DOMAIN_OA)
/* Flag that CPU access will be required for the case of VRAM domain */
#define AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED (1 << 0)
@@ -95,6 +124,10 @@ extern "C" {
#define AMDGPU_GEM_CREATE_VM_ALWAYS_VALID (1 << 6)
/* Flag that BO sharing will be explicitly synchronized */
#define AMDGPU_GEM_CREATE_EXPLICIT_SYNC (1 << 7)
+/* Flag that indicates allocating MQD gart on GFX9, where the mtype
+ * for the second page onward should be set to NC.
+ */
+#define AMDGPU_GEM_CREATE_MQD_GFX9 (1 << 8)
struct drm_amdgpu_gem_create_in {
/** the requested memory size */
@@ -473,7 +506,8 @@ struct drm_amdgpu_gem_va {
#define AMDGPU_HW_IP_UVD_ENC 5
#define AMDGPU_HW_IP_VCN_DEC 6
#define AMDGPU_HW_IP_VCN_ENC 7
-#define AMDGPU_HW_IP_NUM 8
+#define AMDGPU_HW_IP_VCN_JPEG 8
+#define AMDGPU_HW_IP_NUM 9
#define AMDGPU_HW_IP_INSTANCE_MAX_COUNT 1
@@ -482,6 +516,7 @@ struct drm_amdgpu_gem_va {
#define AMDGPU_CHUNK_ID_DEPENDENCIES 0x03
#define AMDGPU_CHUNK_ID_SYNCOBJ_IN 0x04
#define AMDGPU_CHUNK_ID_SYNCOBJ_OUT 0x05
+#define AMDGPU_CHUNK_ID_BO_HANDLES 0x06
struct drm_amdgpu_cs_chunk {
__u32 chunk_id;
@@ -520,6 +555,10 @@ union drm_amdgpu_cs {
/* Preempt flag, IB should set Pre_enb bit if PREEMPT flag detected */
#define AMDGPU_IB_FLAG_PREEMPT (1<<2)
+/* The IB fence should do the L2 writeback but not invalidate any shader
+ * caches (L2/vL1/sL1/I$). */
+#define AMDGPU_IB_FLAG_TC_WB_NOT_INVALIDATE (1 << 3)
+
struct drm_amdgpu_cs_chunk_ib {
__u32 _pad;
/** AMDGPU_IB_FLAG_* */
@@ -618,6 +657,16 @@ struct drm_amdgpu_cs_chunk_data {
#define AMDGPU_INFO_FW_SOS 0x0c
/* Subquery id: Query PSP ASD firmware version */
#define AMDGPU_INFO_FW_ASD 0x0d
+ /* Subquery id: Query VCN firmware version */
+ #define AMDGPU_INFO_FW_VCN 0x0e
+ /* Subquery id: Query GFX RLC SRLC firmware version */
+ #define AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_CNTL 0x0f
+ /* Subquery id: Query GFX RLC SRLG firmware version */
+ #define AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_GPM_MEM 0x10
+ /* Subquery id: Query GFX RLC SRLS firmware version */
+ #define AMDGPU_INFO_FW_GFX_RLC_RESTORE_LIST_SRM_MEM 0x11
+ /* Subquery id: Query DMCU firmware version */
+ #define AMDGPU_INFO_FW_DMCU 0x12
/* number of bytes moved for TTM migration */
#define AMDGPU_INFO_NUM_BYTES_MOVED 0x0f
/* the used VRAM size */
@@ -806,6 +855,7 @@ struct drm_amdgpu_info_firmware {
#define AMDGPU_VRAM_TYPE_GDDR5 5
#define AMDGPU_VRAM_TYPE_HBM 6
#define AMDGPU_VRAM_TYPE_DDR3 7
+#define AMDGPU_VRAM_TYPE_DDR4 8
struct drm_amdgpu_info_device {
/** PCI Device ID */
diff --git a/include/drm-uapi/drm.h b/include/drm-uapi/drm.h
index f0bd91de..85c685a2 100644
--- a/include/drm-uapi/drm.h
+++ b/include/drm-uapi/drm.h
@@ -674,6 +674,22 @@ struct drm_get_cap {
*/
#define DRM_CLIENT_CAP_ATOMIC 3
+/**
+ * DRM_CLIENT_CAP_ASPECT_RATIO
+ *
+ * If set to 1, the DRM core will provide aspect ratio information in modes.
+ */
+#define DRM_CLIENT_CAP_ASPECT_RATIO 4
+
+/**
+ * DRM_CLIENT_CAP_WRITEBACK_CONNECTORS
+ *
+ * If set to 1, the DRM core will expose special connectors to be used for
+ * writing back to memory the scene setup in the commit. Depends on client
+ * also supporting DRM_CLIENT_CAP_ATOMIC
+ */
+#define DRM_CLIENT_CAP_WRITEBACK_CONNECTORS 5
+
/** DRM_IOCTL_SET_CLIENT_CAP ioctl argument type */
struct drm_set_client_cap {
__u64 capability;
diff --git a/include/drm-uapi/drm_fourcc.h b/include/drm-uapi/drm_fourcc.h
index e04613d3..139632b8 100644
--- a/include/drm-uapi/drm_fourcc.h
+++ b/include/drm-uapi/drm_fourcc.h
@@ -30,11 +30,50 @@
extern "C" {
#endif
+/**
+ * DOC: overview
+ *
+ * In the DRM subsystem, framebuffer pixel formats are described using the
+ * fourcc codes defined in `include/uapi/drm/drm_fourcc.h`. In addition to the
+ * fourcc code, a Format Modifier may optionally be provided, in order to
+ * further describe the buffer's format - for example tiling or compression.
+ *
+ * Format Modifiers
+ * ----------------
+ *
+ * Format modifiers are used in conjunction with a fourcc code, forming a
+ * unique fourcc:modifier pair. This format:modifier pair must fully define the
+ * format and data layout of the buffer, and should be the only way to describe
+ * that particular buffer.
+ *
+ * Having multiple fourcc:modifier pairs which describe the same layout should
+ * be avoided, as such aliases run the risk of different drivers exposing
+ * different names for the same data format, forcing userspace to understand
+ * that they are aliases.
+ *
+ * Format modifiers may change any property of the buffer, including the number
+ * of planes and/or the required allocation size. Format modifiers are
+ * vendor-namespaced, and as such the relationship between a fourcc code and a
+ * modifier is specific to the modifer being used. For example, some modifiers
+ * may preserve meaning - such as number of planes - from the fourcc code,
+ * whereas others may not.
+ *
+ * Vendors should document their modifier usage in as much detail as
+ * possible, to ensure maximum compatibility across devices, drivers and
+ * applications.
+ *
+ * The authoritative list of format modifier codes is found in
+ * `include/uapi/drm/drm_fourcc.h`
+ */
+
#define fourcc_code(a, b, c, d) ((__u32)(a) | ((__u32)(b) << 8) | \
((__u32)(c) << 16) | ((__u32)(d) << 24))
#define DRM_FORMAT_BIG_ENDIAN (1<<31) /* format is big endian instead of little endian */
+/* Reserve 0 for the invalid format specifier */
+#define DRM_FORMAT_INVALID 0
+
/* color index */
#define DRM_FORMAT_C8 fourcc_code('C', '8', ' ', ' ') /* [7:0] C */
@@ -183,6 +222,7 @@ extern "C" {
#define DRM_FORMAT_MOD_VENDOR_QCOM 0x05
#define DRM_FORMAT_MOD_VENDOR_VIVANTE 0x06
#define DRM_FORMAT_MOD_VENDOR_BROADCOM 0x07
+#define DRM_FORMAT_MOD_VENDOR_ARM 0x08
/* add more to the end as needed */
#define DRM_FORMAT_RESERVED ((1ULL << 56) - 1)
@@ -298,6 +338,19 @@ extern "C" {
*/
#define DRM_FORMAT_MOD_SAMSUNG_64_32_TILE fourcc_mod_code(SAMSUNG, 1)
+/*
+ * Qualcomm Compressed Format
+ *
+ * Refers to a compressed variant of the base format that is compressed.
+ * Implementation may be platform and base-format specific.
+ *
+ * Each macrotile consists of m x n (mostly 4 x 4) tiles.
+ * Pixel data pitch/stride is aligned with macrotile width.
+ * Pixel data height is aligned with macrotile height.
+ * Entire pixel data buffer is aligned with 4k(bytes).
+ */
+#define DRM_FORMAT_MOD_QCOM_COMPRESSED fourcc_mod_code(QCOM, 1)
+
/* Vivante framebuffer modifiers */
/*
@@ -384,6 +437,23 @@ extern "C" {
#define DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK_THIRTYTWO_GOB \
fourcc_mod_code(NVIDIA, 0x15)
+/*
+ * Some Broadcom modifiers take parameters, for example the number of
+ * vertical lines in the image. Reserve the lower 32 bits for modifier
+ * type, and the next 24 bits for parameters. Top 8 bits are the
+ * vendor code.
+ */
+#define __fourcc_mod_broadcom_param_shift 8
+#define __fourcc_mod_broadcom_param_bits 48
+#define fourcc_mod_broadcom_code(val, params) \
+ fourcc_mod_code(BROADCOM, ((((__u64)params) << __fourcc_mod_broadcom_param_shift) | val))
+#define fourcc_mod_broadcom_param(m) \
+ ((int)(((m) >> __fourcc_mod_broadcom_param_shift) & \
+ ((1ULL << __fourcc_mod_broadcom_param_bits) - 1)))
+#define fourcc_mod_broadcom_mod(m) \
+ ((m) & ~(((1ULL << __fourcc_mod_broadcom_param_bits) - 1) << \
+ __fourcc_mod_broadcom_param_shift))
+
/*
* Broadcom VC4 "T" format
*
@@ -405,6 +475,151 @@ extern "C" {
*/
#define DRM_FORMAT_MOD_BROADCOM_VC4_T_TILED fourcc_mod_code(BROADCOM, 1)
+/*
+ * Broadcom SAND format
+ *
+ * This is the native format that the H.264 codec block uses. For VC4
+ * HVS, it is only valid for H.264 (NV12/21) and RGBA modes.
+ *
+ * The image can be considered to be split into columns, and the
+ * columns are placed consecutively into memory. The width of those
+ * columns can be either 32, 64, 128, or 256 pixels, but in practice
+ * only 128 pixel columns are used.
+ *
+ * The pitch between the start of each column is set to optimally
+ * switch between SDRAM banks. This is passed as the number of lines
+ * of column width in the modifier (we can't use the stride value due
+ * to various core checks that look at it , so you should set the
+ * stride to width*cpp).
+ *
+ * Note that the column height for this format modifier is the same
+ * for all of the planes, assuming that each column contains both Y
+ * and UV. Some SAND-using hardware stores UV in a separate tiled
+ * image from Y to reduce the column height, which is not supported
+ * with these modifiers.
+ */
+
+#define DRM_FORMAT_MOD_BROADCOM_SAND32_COL_HEIGHT(v) \
+ fourcc_mod_broadcom_code(2, v)
+#define DRM_FORMAT_MOD_BROADCOM_SAND64_COL_HEIGHT(v) \
+ fourcc_mod_broadcom_code(3, v)
+#define DRM_FORMAT_MOD_BROADCOM_SAND128_COL_HEIGHT(v) \
+ fourcc_mod_broadcom_code(4, v)
+#define DRM_FORMAT_MOD_BROADCOM_SAND256_COL_HEIGHT(v) \
+ fourcc_mod_broadcom_code(5, v)
+
+#define DRM_FORMAT_MOD_BROADCOM_SAND32 \
+ DRM_FORMAT_MOD_BROADCOM_SAND32_COL_HEIGHT(0)
+#define DRM_FORMAT_MOD_BROADCOM_SAND64 \
+ DRM_FORMAT_MOD_BROADCOM_SAND64_COL_HEIGHT(0)
+#define DRM_FORMAT_MOD_BROADCOM_SAND128 \
+ DRM_FORMAT_MOD_BROADCOM_SAND128_COL_HEIGHT(0)
+#define DRM_FORMAT_MOD_BROADCOM_SAND256 \
+ DRM_FORMAT_MOD_BROADCOM_SAND256_COL_HEIGHT(0)
+
+/* Broadcom UIF format
+ *
+ * This is the common format for the current Broadcom multimedia
+ * blocks, including V3D 3.x and newer, newer video codecs, and
+ * displays.
+ *
+ * The image consists of utiles (64b blocks), UIF blocks (2x2 utiles),
+ * and macroblocks (4x4 UIF blocks). Those 4x4 UIF block groups are
+ * stored in columns, with padding between the columns to ensure that
+ * moving from one column to the next doesn't hit the same SDRAM page
+ * bank.
+ *
+ * To calculate the padding, it is assumed that each hardware block
+ * and the software driving it knows the platform's SDRAM page size,
+ * number of banks, and XOR address, and that it's identical between
+ * all blocks using the format. This tiling modifier will use XOR as
+ * necessary to reduce the padding. If a hardware block can't do XOR,
+ * the assumption is that a no-XOR tiling modifier will be created.
+ */
+#define DRM_FORMAT_MOD_BROADCOM_UIF fourcc_mod_code(BROADCOM, 6)
+
+/*
+ * Arm Framebuffer Compression (AFBC) modifiers
+ *
+ * AFBC is a proprietary lossless image compression protocol and format.
+ * It provides fine-grained random access and minimizes the amount of data
+ * transferred between IP blocks.
+ *
+ * AFBC has several features which may be supported and/or used, which are
+ * represented using bits in the modifier. Not all combinations are valid,
+ * and different devices or use-cases may support different combinations.
+ */
+#define DRM_FORMAT_MOD_ARM_AFBC(__afbc_mode) fourcc_mod_code(ARM, __afbc_mode)
+
+/*
+ * AFBC superblock size
+ *
+ * Indicates the superblock size(s) used for the AFBC buffer. The buffer
+ * size (in pixels) must be aligned to a multiple of the superblock size.
+ * Four lowest significant bits(LSBs) are reserved for block size.
+ */
+#define AFBC_FORMAT_MOD_BLOCK_SIZE_MASK 0xf
+#define AFBC_FORMAT_MOD_BLOCK_SIZE_16x16 (1ULL)
+#define AFBC_FORMAT_MOD_BLOCK_SIZE_32x8 (2ULL)
+
+/*
+ * AFBC lossless colorspace transform
+ *
+ * Indicates that the buffer makes use of the AFBC lossless colorspace
+ * transform.
+ */
+#define AFBC_FORMAT_MOD_YTR (1ULL << 4)
+
+/*
+ * AFBC block-split
+ *
+ * Indicates that the payload of each superblock is split. The second
+ * half of the payload is positioned at a predefined offset from the start
+ * of the superblock payload.
+ */
+#define AFBC_FORMAT_MOD_SPLIT (1ULL << 5)
+
+/*
+ * AFBC sparse layout
+ *
+ * This flag indicates that the payload of each superblock must be stored at a
+ * predefined position relative to the other superblocks in the same AFBC
+ * buffer. This order is the same order used by the header buffer. In this mode
+ * each superblock is given the same amount of space as an uncompressed
+ * superblock of the particular format would require, rounding up to the next
+ * multiple of 128 bytes in size.
+ */
+#define AFBC_FORMAT_MOD_SPARSE (1ULL << 6)
+
+/*
+ * AFBC copy-block restrict
+ *
+ * Buffers with this flag must obey the copy-block restriction. The restriction
+ * is such that there are no copy-blocks referring across the border of 8x8
+ * blocks. For the subsampled data the 8x8 limitation is also subsampled.
+ */
+#define AFBC_FORMAT_MOD_CBR (1ULL << 7)
+
+/*
+ * AFBC tiled layout
+ *
+ * The tiled layout groups superblocks in 8x8 or 4x4 tiles, where all
+ * superblocks inside a tile are stored together in memory. 8x8 tiles are used
+ * for pixel formats up to and including 32 bpp while 4x4 tiles are used for
+ * larger bpp formats. The order between the tiles is scan line.
+ * When the tiled layout is used, the buffer size (in pixels) must be aligned
+ * to the tile size.
+ */
+#define AFBC_FORMAT_MOD_TILED (1ULL << 8)
+
+/*
+ * AFBC solid color blocks
+ *
+ * Indicates that the buffer makes use of solid-color blocks, whereby bandwidth
+ * can be reduced if a whole superblock is a single color.
+ */
+#define AFBC_FORMAT_MOD_SC (1ULL << 9)
+
#if defined(__cplusplus)
}
#endif
diff --git a/include/drm-uapi/drm_mode.h b/include/drm-uapi/drm_mode.h
index 2c575794..d3e0fe31 100644
--- a/include/drm-uapi/drm_mode.h
+++ b/include/drm-uapi/drm_mode.h
@@ -93,6 +93,15 @@ extern "C" {
#define DRM_MODE_PICTURE_ASPECT_NONE 0
#define DRM_MODE_PICTURE_ASPECT_4_3 1
#define DRM_MODE_PICTURE_ASPECT_16_9 2
+#define DRM_MODE_PICTURE_ASPECT_64_27 3
+#define DRM_MODE_PICTURE_ASPECT_256_135 4
+
+/* Content type options */
+#define DRM_MODE_CONTENT_TYPE_NO_DATA 0
+#define DRM_MODE_CONTENT_TYPE_GRAPHICS 1
+#define DRM_MODE_CONTENT_TYPE_PHOTO 2
+#define DRM_MODE_CONTENT_TYPE_CINEMA 3
+#define DRM_MODE_CONTENT_TYPE_GAME 4
/* Aspect ratio flag bitmask (4 bits 22:19) */
#define DRM_MODE_FLAG_PIC_AR_MASK (0x0F<<19)
@@ -102,6 +111,10 @@ extern "C" {
(DRM_MODE_PICTURE_ASPECT_4_3<<19)
#define DRM_MODE_FLAG_PIC_AR_16_9 \
(DRM_MODE_PICTURE_ASPECT_16_9<<19)
+#define DRM_MODE_FLAG_PIC_AR_64_27 \
+ (DRM_MODE_PICTURE_ASPECT_64_27<<19)
+#define DRM_MODE_FLAG_PIC_AR_256_135 \
+ (DRM_MODE_PICTURE_ASPECT_256_135<<19)
#define DRM_MODE_FLAG_ALL (DRM_MODE_FLAG_PHSYNC | \
DRM_MODE_FLAG_NHSYNC | \
@@ -173,8 +186,9 @@ extern "C" {
/*
* DRM_MODE_REFLECT_<axis>
*
- * Signals that the contents of a drm plane is reflected in the <axis> axis,
+ * Signals that the contents of a drm plane is reflected along the <axis> axis,
* in the same way as mirroring.
+ * See kerneldoc chapter "Plane Composition Properties" for more details.
*
* This define is provided as a convenience, looking up the property id
* using the name->prop id lookup is the preferred method.
@@ -338,6 +352,7 @@ enum drm_mode_subconnector {
#define DRM_MODE_CONNECTOR_VIRTUAL 15
#define DRM_MODE_CONNECTOR_DSI 16
#define DRM_MODE_CONNECTOR_DPI 17
+#define DRM_MODE_CONNECTOR_WRITEBACK 18
struct drm_mode_get_connector {
@@ -363,7 +378,7 @@ struct drm_mode_get_connector {
__u32 pad;
};
-#define DRM_MODE_PROP_PENDING (1<<0)
+#define DRM_MODE_PROP_PENDING (1<<0) /* deprecated, do not use */
#define DRM_MODE_PROP_RANGE (1<<1)
#define DRM_MODE_PROP_IMMUTABLE (1<<2)
#define DRM_MODE_PROP_ENUM (1<<3) /* enumerated type with text strings */
@@ -598,8 +613,11 @@ struct drm_mode_crtc_lut {
};
struct drm_color_ctm {
- /* Conversion matrix in S31.32 format. */
- __s64 matrix[9];
+ /*
+ * Conversion matrix in S31.32 sign-magnitude
+ * (not two's complement!) format.
+ */
+ __u64 matrix[9];
};
struct drm_color_lut {
diff --git a/include/drm-uapi/etnaviv_drm.h b/include/drm-uapi/etnaviv_drm.h
index e9b997a0..0d5c49dc 100644
--- a/include/drm-uapi/etnaviv_drm.h
+++ b/include/drm-uapi/etnaviv_drm.h
@@ -55,6 +55,12 @@ struct drm_etnaviv_timespec {
#define ETNAVIV_PARAM_GPU_FEATURES_4 0x07
#define ETNAVIV_PARAM_GPU_FEATURES_5 0x08
#define ETNAVIV_PARAM_GPU_FEATURES_6 0x09
+#define ETNAVIV_PARAM_GPU_FEATURES_7 0x0a
+#define ETNAVIV_PARAM_GPU_FEATURES_8 0x0b
+#define ETNAVIV_PARAM_GPU_FEATURES_9 0x0c
+#define ETNAVIV_PARAM_GPU_FEATURES_10 0x0d
+#define ETNAVIV_PARAM_GPU_FEATURES_11 0x0e
+#define ETNAVIV_PARAM_GPU_FEATURES_12 0x0f
#define ETNAVIV_PARAM_GPU_STREAM_COUNT 0x10
#define ETNAVIV_PARAM_GPU_REGISTER_MAX 0x11
diff --git a/include/drm-uapi/exynos_drm.h b/include/drm-uapi/exynos_drm.h
index a00116b5..7414cfd7 100644
--- a/include/drm-uapi/exynos_drm.h
+++ b/include/drm-uapi/exynos_drm.h
@@ -135,6 +135,219 @@ struct drm_exynos_g2d_exec {
__u64 async;
};
+/* Exynos DRM IPP v2 API */
+
+/**
+ * Enumerate available IPP hardware modules.
+ *
+ * @count_ipps: size of ipp_id array / number of ipp modules (set by driver)
+ * @reserved: padding
+ * @ipp_id_ptr: pointer to ipp_id array or NULL
+ */
+struct drm_exynos_ioctl_ipp_get_res {
+ __u32 count_ipps;
+ __u32 reserved;
+ __u64 ipp_id_ptr;
+};
+
+enum drm_exynos_ipp_format_type {
+ DRM_EXYNOS_IPP_FORMAT_SOURCE = 0x01,
+ DRM_EXYNOS_IPP_FORMAT_DESTINATION = 0x02,
+};
+
+struct drm_exynos_ipp_format {
+ __u32 fourcc;
+ __u32 type;
+ __u64 modifier;
+};
+
+enum drm_exynos_ipp_capability {
+ DRM_EXYNOS_IPP_CAP_CROP = 0x01,
+ DRM_EXYNOS_IPP_CAP_ROTATE = 0x02,
+ DRM_EXYNOS_IPP_CAP_SCALE = 0x04,
+ DRM_EXYNOS_IPP_CAP_CONVERT = 0x08,
+};
+
+/**
+ * Get IPP hardware capabilities and supported image formats.
+ *
+ * @ipp_id: id of IPP module to query
+ * @capabilities: bitmask of drm_exynos_ipp_capability (set by driver)
+ * @reserved: padding
+ * @formats_count: size of formats array (in entries) / number of filled
+ * formats (set by driver)
+ * @formats_ptr: pointer to formats array or NULL
+ */
+struct drm_exynos_ioctl_ipp_get_caps {
+ __u32 ipp_id;
+ __u32 capabilities;
+ __u32 reserved;
+ __u32 formats_count;
+ __u64 formats_ptr;
+};
+
+enum drm_exynos_ipp_limit_type {
+ /* size (horizontal/vertial) limits, in pixels (min, max, alignment) */
+ DRM_EXYNOS_IPP_LIMIT_TYPE_SIZE = 0x0001,
+ /* scale ratio (horizonta/vertial), 16.16 fixed point (min, max) */
+ DRM_EXYNOS_IPP_LIMIT_TYPE_SCALE = 0x0002,
+
+ /* image buffer area */
+ DRM_EXYNOS_IPP_LIMIT_SIZE_BUFFER = 0x0001 << 16,
+ /* src/dst rectangle area */
+ DRM_EXYNOS_IPP_LIMIT_SIZE_AREA = 0x0002 << 16,
+ /* src/dst rectangle area when rotation enabled */
+ DRM_EXYNOS_IPP_LIMIT_SIZE_ROTATED = 0x0003 << 16,
+
+ DRM_EXYNOS_IPP_LIMIT_TYPE_MASK = 0x000f,
+ DRM_EXYNOS_IPP_LIMIT_SIZE_MASK = 0x000f << 16,
+};
+
+struct drm_exynos_ipp_limit_val {
+ __u32 min;
+ __u32 max;
+ __u32 align;
+ __u32 reserved;
+};
+
+/**
+ * IPP module limitation.
+ *
+ * @type: limit type (see drm_exynos_ipp_limit_type enum)
+ * @reserved: padding
+ * @h: horizontal limits
+ * @v: vertical limits
+ */
+struct drm_exynos_ipp_limit {
+ __u32 type;
+ __u32 reserved;
+ struct drm_exynos_ipp_limit_val h;
+ struct drm_exynos_ipp_limit_val v;
+};
+
+/**
+ * Get IPP limits for given image format.
+ *
+ * @ipp_id: id of IPP module to query
+ * @fourcc: image format code (see DRM_FORMAT_* in drm_fourcc.h)
+ * @modifier: image format modifier (see DRM_FORMAT_MOD_* in drm_fourcc.h)
+ * @type: source/destination identifier (drm_exynos_ipp_format_flag enum)
+ * @limits_count: size of limits array (in entries) / number of filled entries
+ * (set by driver)
+ * @limits_ptr: pointer to limits array or NULL
+ */
+struct drm_exynos_ioctl_ipp_get_limits {
+ __u32 ipp_id;
+ __u32 fourcc;
+ __u64 modifier;
+ __u32 type;
+ __u32 limits_count;
+ __u64 limits_ptr;
+};
+
+enum drm_exynos_ipp_task_id {
+ /* buffer described by struct drm_exynos_ipp_task_buffer */
+ DRM_EXYNOS_IPP_TASK_BUFFER = 0x0001,
+ /* rectangle described by struct drm_exynos_ipp_task_rect */
+ DRM_EXYNOS_IPP_TASK_RECTANGLE = 0x0002,
+ /* transformation described by struct drm_exynos_ipp_task_transform */
+ DRM_EXYNOS_IPP_TASK_TRANSFORM = 0x0003,
+ /* alpha configuration described by struct drm_exynos_ipp_task_alpha */
+ DRM_EXYNOS_IPP_TASK_ALPHA = 0x0004,
+
+ /* source image data (for buffer and rectangle chunks) */
+ DRM_EXYNOS_IPP_TASK_TYPE_SOURCE = 0x0001 << 16,
+ /* destination image data (for buffer and rectangle chunks) */
+ DRM_EXYNOS_IPP_TASK_TYPE_DESTINATION = 0x0002 << 16,
+};
+
+/**
+ * Memory buffer with image data.
+ *
+ * @id: must be DRM_EXYNOS_IPP_TASK_BUFFER
+ * other parameters are same as for AddFB2 generic DRM ioctl
+ */
+struct drm_exynos_ipp_task_buffer {
+ __u32 id;
+ __u32 fourcc;
+ __u32 width, height;
+ __u32 gem_id[4];
+ __u32 offset[4];
+ __u32 pitch[4];
+ __u64 modifier;
+};
+
+/**
+ * Rectangle for processing.
+ *
+ * @id: must be DRM_EXYNOS_IPP_TASK_RECTANGLE
+ * @reserved: padding
+ * @x,@y: left corner in pixels
+ * @w,@h: width/height in pixels
+ */
+struct drm_exynos_ipp_task_rect {
+ __u32 id;
+ __u32 reserved;
+ __u32 x;
+ __u32 y;
+ __u32 w;
+ __u32 h;
+};
+
+/**
+ * Image tranformation description.
+ *
+ * @id: must be DRM_EXYNOS_IPP_TASK_TRANSFORM
+ * @rotation: DRM_MODE_ROTATE_* and DRM_MODE_REFLECT_* values
+ */
+struct drm_exynos_ipp_task_transform {
+ __u32 id;
+ __u32 rotation;
+};
+
+/**
+ * Image global alpha configuration for formats without alpha values.
+ *
+ * @id: must be DRM_EXYNOS_IPP_TASK_ALPHA
+ * @value: global alpha value (0-255)
+ */
+struct drm_exynos_ipp_task_alpha {
+ __u32 id;
+ __u32 value;
+};
+
+enum drm_exynos_ipp_flag {
+ /* generate DRM event after processing */
+ DRM_EXYNOS_IPP_FLAG_EVENT = 0x01,
+ /* dry run, only check task parameters */
+ DRM_EXYNOS_IPP_FLAG_TEST_ONLY = 0x02,
+ /* non-blocking processing */
+ DRM_EXYNOS_IPP_FLAG_NONBLOCK = 0x04,
+};
+
+#define DRM_EXYNOS_IPP_FLAGS (DRM_EXYNOS_IPP_FLAG_EVENT |\
+ DRM_EXYNOS_IPP_FLAG_TEST_ONLY | DRM_EXYNOS_IPP_FLAG_NONBLOCK)
+
+/**
+ * Perform image processing described by array of drm_exynos_ipp_task_*
+ * structures (parameters array).
+ *
+ * @ipp_id: id of IPP module to run the task
+ * @flags: bitmask of drm_exynos_ipp_flag values
+ * @reserved: padding
+ * @params_size: size of parameters array (in bytes)
+ * @params_ptr: pointer to parameters array or NULL
+ * @user_data: (optional) data for drm event
+ */
+struct drm_exynos_ioctl_ipp_commit {
+ __u32 ipp_id;
+ __u32 flags;
+ __u32 reserved;
+ __u32 params_size;
+ __u64 params_ptr;
+ __u64 user_data;
+};
+
#define DRM_EXYNOS_GEM_CREATE 0x00
#define DRM_EXYNOS_GEM_MAP 0x01
/* Reserved 0x03 ~ 0x05 for exynos specific gem ioctl */
@@ -147,6 +360,11 @@ struct drm_exynos_g2d_exec {
#define DRM_EXYNOS_G2D_EXEC 0x22
/* Reserved 0x30 ~ 0x33 for obsolete Exynos IPP ioctls */
+/* IPP - Image Post Processing */
+#define DRM_EXYNOS_IPP_GET_RESOURCES 0x40
+#define DRM_EXYNOS_IPP_GET_CAPS 0x41
+#define DRM_EXYNOS_IPP_GET_LIMITS 0x42
+#define DRM_EXYNOS_IPP_COMMIT 0x43
#define DRM_IOCTL_EXYNOS_GEM_CREATE DRM_IOWR(DRM_COMMAND_BASE + \
DRM_EXYNOS_GEM_CREATE, struct drm_exynos_gem_create)
@@ -165,8 +383,20 @@ struct drm_exynos_g2d_exec {
#define DRM_IOCTL_EXYNOS_G2D_EXEC DRM_IOWR(DRM_COMMAND_BASE + \
DRM_EXYNOS_G2D_EXEC, struct drm_exynos_g2d_exec)
+#define DRM_IOCTL_EXYNOS_IPP_GET_RESOURCES DRM_IOWR(DRM_COMMAND_BASE + \
+ DRM_EXYNOS_IPP_GET_RESOURCES, \
+ struct drm_exynos_ioctl_ipp_get_res)
+#define DRM_IOCTL_EXYNOS_IPP_GET_CAPS DRM_IOWR(DRM_COMMAND_BASE + \
+ DRM_EXYNOS_IPP_GET_CAPS, struct drm_exynos_ioctl_ipp_get_caps)
+#define DRM_IOCTL_EXYNOS_IPP_GET_LIMITS DRM_IOWR(DRM_COMMAND_BASE + \
+ DRM_EXYNOS_IPP_GET_LIMITS, \
+ struct drm_exynos_ioctl_ipp_get_limits)
+#define DRM_IOCTL_EXYNOS_IPP_COMMIT DRM_IOWR(DRM_COMMAND_BASE + \
+ DRM_EXYNOS_IPP_COMMIT, struct drm_exynos_ioctl_ipp_commit)
+
/* EXYNOS specific events */
#define DRM_EXYNOS_G2D_EVENT 0x80000000
+#define DRM_EXYNOS_IPP_EVENT 0x80000002
struct drm_exynos_g2d_event {
struct drm_event base;
@@ -177,6 +407,16 @@ struct drm_exynos_g2d_event {
__u32 reserved;
};
+struct drm_exynos_ipp_event {
+ struct drm_event base;
+ __u64 user_data;
+ __u32 tv_sec;
+ __u32 tv_usec;
+ __u32 ipp_id;
+ __u32 sequence;
+ __u64 reserved;
+};
+
#if defined(__cplusplus)
}
#endif
diff --git a/include/drm-uapi/i915_drm.h b/include/drm-uapi/i915_drm.h
index 16e452aa..bfdeec61 100644
--- a/include/drm-uapi/i915_drm.h
+++ b/include/drm-uapi/i915_drm.h
@@ -98,6 +98,7 @@ enum drm_i915_gem_engine_class {
I915_ENGINE_CLASS_COPY = 1,
I915_ENGINE_CLASS_VIDEO = 2,
I915_ENGINE_CLASS_VIDEO_ENHANCE = 3,
+ I915_ENGINE_CLASS_COMPUTE = 4,
I915_ENGINE_CLASS_INVALID = -1
};
@@ -319,6 +320,7 @@ typedef struct _drm_i915_sarea {
#define DRM_I915_PERF_ADD_CONFIG 0x37
#define DRM_I915_PERF_REMOVE_CONFIG 0x38
#define DRM_I915_QUERY 0x39
+#define DRM_I915_GEM_MMAP2 0x3a
#define DRM_IOCTL_I915_INIT DRM_IOW( DRM_COMMAND_BASE + DRM_I915_INIT, drm_i915_init_t)
#define DRM_IOCTL_I915_FLUSH DRM_IO ( DRM_COMMAND_BASE + DRM_I915_FLUSH)
@@ -377,6 +379,7 @@ typedef struct _drm_i915_sarea {
#define DRM_IOCTL_I915_PERF_ADD_CONFIG DRM_IOW(DRM_COMMAND_BASE + DRM_I915_PERF_ADD_CONFIG, struct drm_i915_perf_oa_config)
#define DRM_IOCTL_I915_PERF_REMOVE_CONFIG DRM_IOW(DRM_COMMAND_BASE + DRM_I915_PERF_REMOVE_CONFIG, __u64)
#define DRM_IOCTL_I915_QUERY DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_QUERY, struct drm_i915_query)
+#define DRM_IOCTL_I915_GEM_MMAP2 DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_GEM_MMAP2, struct drm_i915_gem_mmap2)
/* Allow drivers to submit batchbuffers directly to hardware, relying
* on the security mechanisms provided by hardware.
@@ -529,6 +532,37 @@ typedef struct drm_i915_irq_wait {
*/
#define I915_PARAM_CS_TIMESTAMP_FREQUENCY 51
+/*
+ * Once upon a time we supposed that writes through the GGTT would be
+ * immediately in physical memory (once flushed out of the CPU path). However,
+ * on a few different processors and chipsets, this is not necessarily the case
+ * as the writes appear to be buffered internally. Thus a read of the backing
+ * storage (physical memory) via a different path (with different physical tags
+ * to the indirect write via the GGTT) will see stale values from before
+ * the GGTT write. Inside the kernel, we can for the most part keep track of
+ * the different read/write domains in use (e.g. set-domain), but the assumption
+ * of coherency is baked into the ABI, hence reporting its true state in this
+ * parameter.
+ *
+ * Reports true when writes via mmap_gtt are immediately visible following an
+ * lfence to flush the WCB.
+ *
+ * Reports false when writes via mmap_gtt are indeterminately delayed in an in
+ * internal buffer and are _not_ immediately visible to third parties accessing
+ * directly via mmap_cpu/mmap_wc. Use of mmap_gtt as part of an IPC
+ * communications channel when reporting false is strongly disadvised.
+ */
+#define I915_PARAM_MMAP_GTT_COHERENT 52
+
+/* Dual Context / Compute-only Engine */
+#define I915_PARAM_HAS_CCS 53
+
+/* Total local memory in bytes */
+#define I915_PARAM_LMEM_TOTAL_BYTES 54
+
+/* Available local memory in bytes */
+#define I915_PARAM_LMEM_AVAIL_BYTES 55
+
typedef struct drm_i915_getparam {
__s32 param;
/*
@@ -701,6 +735,19 @@ struct drm_i915_gem_mmap_gtt {
__u64 offset;
};
+struct drm_i915_gem_mmap2 {
+ /** Handle for the object being mapped. */
+ __u32 handle;
+ __u32 pad;
+ /**
+ * Fake offset to use for subsequent mmap call
+ *
+ * This is a fixed-size type for 32/64 compatibility.
+ */
+ __u64 offset;
+ __u64 flags;
+};
+
struct drm_i915_gem_set_domain {
/** Handle for the object */
__u32 handle;
@@ -948,6 +995,7 @@ struct drm_i915_gem_execbuffer2 {
#define I915_EXEC_BSD (2<<0)
#define I915_EXEC_BLT (3<<0)
#define I915_EXEC_VEBOX (4<<0)
+#define I915_EXEC_COMPUTE (5<<0)
/* Used for switching the constants addressing mode on gen4+ RENDER ring.
* Gen6+ only supports relative addressing to dynamic state (default) and
@@ -1510,6 +1558,14 @@ enum drm_i915_perf_property_id {
*/
DRM_I915_PERF_PROP_OA_EXPONENT,
+ /**
+ * Specify a global OA buffer size to be allocated in bytes.
+ * The driver will allocate a HW supported size that is at
+ * least as large as specified by this property. Larger sizes
+ * than what the HW supports will fail.
+ */
+ DRM_I915_PERF_PROP_OA_BUFFER_SIZE,
+
DRM_I915_PERF_PROP_MAX /* non-ABI */
};
diff --git a/include/drm-uapi/msm_drm.h b/include/drm-uapi/msm_drm.h
index bbbaffad..c06d0a5b 100644
--- a/include/drm-uapi/msm_drm.h
+++ b/include/drm-uapi/msm_drm.h
@@ -201,10 +201,12 @@ struct drm_msm_gem_submit_bo {
#define MSM_SUBMIT_NO_IMPLICIT 0x80000000 /* disable implicit sync */
#define MSM_SUBMIT_FENCE_FD_IN 0x40000000 /* enable input fence_fd */
#define MSM_SUBMIT_FENCE_FD_OUT 0x20000000 /* enable output fence_fd */
+#define MSM_SUBMIT_SUDO 0x10000000 /* run submitted cmds from RB */
#define MSM_SUBMIT_FLAGS ( \
MSM_SUBMIT_NO_IMPLICIT | \
MSM_SUBMIT_FENCE_FD_IN | \
MSM_SUBMIT_FENCE_FD_OUT | \
+ MSM_SUBMIT_SUDO | \
0)
/* Each cmdstream submit consists of a table of buffers involved, and
diff --git a/include/drm-uapi/tegra_drm.h b/include/drm-uapi/tegra_drm.h
index 12f9bf84..6c07919c 100644
--- a/include/drm-uapi/tegra_drm.h
+++ b/include/drm-uapi/tegra_drm.h
@@ -32,143 +32,615 @@ extern "C" {
#define DRM_TEGRA_GEM_CREATE_TILED (1 << 0)
#define DRM_TEGRA_GEM_CREATE_BOTTOM_UP (1 << 1)
+/**
+ * struct drm_tegra_gem_create - parameters for the GEM object creation IOCTL
+ */
struct drm_tegra_gem_create {
+ /**
+ * @size:
+ *
+ * The size, in bytes, of the buffer object to be created.
+ */
__u64 size;
+
+ /**
+ * @flags:
+ *
+ * A bitmask of flags that influence the creation of GEM objects:
+ *
+ * DRM_TEGRA_GEM_CREATE_TILED
+ * Use the 16x16 tiling format for this buffer.
+ *
+ * DRM_TEGRA_GEM_CREATE_BOTTOM_UP
+ * The buffer has a bottom-up layout.
+ */
__u32 flags;
+
+ /**
+ * @handle:
+ *
+ * The handle of the created GEM object. Set by the kernel upon
+ * successful completion of the IOCTL.
+ */
__u32 handle;
};
+/**
+ * struct drm_tegra_gem_mmap - parameters for the GEM mmap IOCTL
+ */
struct drm_tegra_gem_mmap {
+ /**
+ * @handle:
+ *
+ * Handle of the GEM object to obtain an mmap offset for.
+ */
__u32 handle;
+
+ /**
+ * @pad:
+ *
+ * Structure padding that may be used in the future. Must be 0.
+ */
__u32 pad;
+
+ /**
+ * @offset:
+ *
+ * The mmap offset for the given GEM object. Set by the kernel upon
+ * successful completion of the IOCTL.
+ */
__u64 offset;
};
+/**
+ * struct drm_tegra_syncpt_read - parameters for the read syncpoint IOCTL
+ */
struct drm_tegra_syncpt_read {
+ /**
+ * @id:
+ *
+ * ID of the syncpoint to read the current value from.
+ */
__u32 id;
+
+ /**
+ * @value:
+ *
+ * The current syncpoint value. Set by the kernel upon successful
+ * completion of the IOCTL.
+ */
__u32 value;
};
+/**
+ * struct drm_tegra_syncpt_incr - parameters for the increment syncpoint IOCTL
+ */
struct drm_tegra_syncpt_incr {
+ /**
+ * @id:
+ *
+ * ID of the syncpoint to increment.
+ */
__u32 id;
+
+ /**
+ * @pad:
+ *
+ * Structure padding that may be used in the future. Must be 0.
+ */
__u32 pad;
};
+/**
+ * struct drm_tegra_syncpt_wait - parameters for the wait syncpoint IOCTL
+ */
struct drm_tegra_syncpt_wait {
+ /**
+ * @id:
+ *
+ * ID of the syncpoint to wait on.
+ */
__u32 id;
+
+ /**
+ * @thresh:
+ *
+ * Threshold value for which to wait.
+ */
__u32 thresh;
+
+ /**
+ * @timeout:
+ *
+ * Timeout, in milliseconds, to wait.
+ */
__u32 timeout;
+
+ /**
+ * @value:
+ *
+ * The new syncpoint value after the wait. Set by the kernel upon
+ * successful completion of the IOCTL.
+ */
__u32 value;
};
#define DRM_TEGRA_NO_TIMEOUT (0xffffffff)
+/**
+ * struct drm_tegra_open_channel - parameters for the open channel IOCTL
+ */
struct drm_tegra_open_channel {
+ /**
+ * @client:
+ *
+ * The client ID for this channel.
+ */
__u32 client;
+
+ /**
+ * @pad:
+ *
+ * Structure padding that may be used in the future. Must be 0.
+ */
__u32 pad;
+
+ /**
+ * @context:
+ *
+ * The application context of this channel. Set by the kernel upon
+ * successful completion of the IOCTL. This context needs to be passed
+ * to the DRM_TEGRA_CHANNEL_CLOSE or the DRM_TEGRA_SUBMIT IOCTLs.
+ */
__u64 context;
};
+/**
+ * struct drm_tegra_close_channel - parameters for the close channel IOCTL
+ */
struct drm_tegra_close_channel {
+ /**
+ * @context:
+ *
+ * The application context of this channel. This is obtained from the
+ * DRM_TEGRA_OPEN_CHANNEL IOCTL.
+ */
__u64 context;
};
+/**
+ * struct drm_tegra_get_syncpt - parameters for the get syncpoint IOCTL
+ */
struct drm_tegra_get_syncpt {
+ /**
+ * @context:
+ *
+ * The application context identifying the channel for which to obtain
+ * the syncpoint ID.
+ */
__u64 context;
+
+ /**
+ * @index:
+ *
+ * Index of the client syncpoint for which to obtain the ID.
+ */
__u32 index;
+
+ /**
+ * @id:
+ *
+ * The ID of the given syncpoint. Set by the kernel upon successful
+ * completion of the IOCTL.
+ */
__u32 id;
};
+/**
+ * struct drm_tegra_get_syncpt_base - parameters for the get wait base IOCTL
+ */
struct drm_tegra_get_syncpt_base {
+ /**
+ * @context:
+ *
+ * The application context identifying for which channel to obtain the
+ * wait base.
+ */
__u64 context;
+
+ /**
+ * @syncpt:
+ *
+ * ID of the syncpoint for which to obtain the wait base.
+ */
__u32 syncpt;
+
+ /**
+ * @id:
+ *
+ * The ID of the wait base corresponding to the client syncpoint. Set
+ * by the kernel upon successful completion of the IOCTL.
+ */
__u32 id;
};
+/**
+ * struct drm_tegra_syncpt - syncpoint increment operation
+ */
struct drm_tegra_syncpt {
+ /**
+ * @id:
+ *
+ * ID of the syncpoint to operate on.
+ */
__u32 id;
+
+ /**
+ * @incrs:
+ *
+ * Number of increments to perform for the syncpoint.
+ */
__u32 incrs;
};
+/**
+ * struct drm_tegra_cmdbuf - structure describing a command buffer
+ */
struct drm_tegra_cmdbuf {
+ /**
+ * @handle:
+ *
+ * Handle to a GEM object containing the command buffer.
+ */
__u32 handle;
+
+ /**
+ * @offset:
+ *
+ * Offset, in bytes, into the GEM object identified by @handle at
+ * which the command buffer starts.
+ */
__u32 offset;
+
+ /**
+ * @words:
+ *
+ * Number of 32-bit words in this command buffer.
+ */
__u32 words;
+
+ /**
+ * @pad:
+ *
+ * Structure padding that may be used in the future. Must be 0.
+ */
__u32 pad;
};
+/**
+ * struct drm_tegra_reloc - GEM object relocation structure
+ */
struct drm_tegra_reloc {
struct {
+ /**
+ * @cmdbuf.handle:
+ *
+ * Handle to the GEM object containing the command buffer for
+ * which to perform this GEM object relocation.
+ */
__u32 handle;
+
+ /**
+ * @cmdbuf.offset:
+ *
+ * Offset, in bytes, into the command buffer at which to
+ * insert the relocated address.
+ */
__u32 offset;
} cmdbuf;
struct {
+ /**
+ * @target.handle:
+ *
+ * Handle to the GEM object to be relocated.
+ */
__u32 handle;
+
+ /**
+ * @target.offset:
+ *
+ * Offset, in bytes, into the target GEM object at which the
+ * relocated data starts.
+ */
__u32 offset;
} target;
+
+ /**
+ * @shift:
+ *
+ * The number of bits by which to shift relocated addresses.
+ */
__u32 shift;
+
+ /**
+ * @pad:
+ *
+ * Structure padding that may be used in the future. Must be 0.
+ */
__u32 pad;
};
+/**
+ * struct drm_tegra_waitchk - wait check structure
+ */
struct drm_tegra_waitchk {
+ /**
+ * @handle:
+ *
+ * Handle to the GEM object containing a command stream on which to
+ * perform the wait check.
+ */
__u32 handle;
+
+ /**
+ * @offset:
+ *
+ * Offset, in bytes, of the location in the command stream to perform
+ * the wait check on.
+ */
__u32 offset;
+
+ /**
+ * @syncpt:
+ *
+ * ID of the syncpoint to wait check.
+ */
__u32 syncpt;
+
+ /**
+ * @thresh:
+ *
+ * Threshold value for which to check.
+ */
__u32 thresh;
};
+/**
+ * struct drm_tegra_submit - job submission structure
+ */
struct drm_tegra_submit {
+ /**
+ * @context:
+ *
+ * The application context identifying the channel to use for the
+ * execution of this job.
+ */
__u64 context;
+
+ /**
+ * @num_syncpts:
+ *
+ * The number of syncpoints operated on by this job. This defines the
+ * length of the array pointed to by @syncpts.
+ */
__u32 num_syncpts;
+
+ /**
+ * @num_cmdbufs:
+ *
+ * The number of command buffers to execute as part of this job. This
+ * defines the length of the array pointed to by @cmdbufs.
+ */
__u32 num_cmdbufs;
+
+ /**
+ * @num_relocs:
+ *
+ * The number of relocations to perform before executing this job.
+ * This defines the length of the array pointed to by @relocs.
+ */
__u32 num_relocs;
+
+ /**
+ * @num_waitchks:
+ *
+ * The number of wait checks to perform as part of this job. This
+ * defines the length of the array pointed to by @waitchks.
+ */
__u32 num_waitchks;
+
+ /**
+ * @waitchk_mask:
+ *
+ * Bitmask of valid wait checks.
+ */
__u32 waitchk_mask;
+
+ /**
+ * @timeout:
+ *
+ * Timeout, in milliseconds, before this job is cancelled.
+ */
__u32 timeout;
+
+ /**
+ * @syncpts:
+ *
+ * A pointer to an array of &struct drm_tegra_syncpt structures that
+ * specify the syncpoint operations performed as part of this job.
+ * The number of elements in the array must be equal to the value
+ * given by @num_syncpts.
+ */
__u64 syncpts;
+
+ /**
+ * @cmdbufs:
+ *
+ * A pointer to an array of &struct drm_tegra_cmdbuf structures that
+ * define the command buffers to execute as part of this job. The
+ * number of elements in the array must be equal to the value given
+ * by @num_syncpts.
+ */
__u64 cmdbufs;
+
+ /**
+ * @relocs:
+ *
+ * A pointer to an array of &struct drm_tegra_reloc structures that
+ * specify the relocations that need to be performed before executing
+ * this job. The number of elements in the array must be equal to the
+ * value given by @num_relocs.
+ */
__u64 relocs;
+
+ /**
+ * @waitchks:
+ *
+ * A pointer to an array of &struct drm_tegra_waitchk structures that
+ * specify the wait checks to be performed while executing this job.
+ * The number of elements in the array must be equal to the value
+ * given by @num_waitchks.
+ */
__u64 waitchks;
- __u32 fence; /* Return value */
- __u32 reserved[5]; /* future expansion */
+ /**
+ * @fence:
+ *
+ * The threshold of the syncpoint associated with this job after it
+ * has been completed. Set by the kernel upon successful completion of
+ * the IOCTL. This can be used with the DRM_TEGRA_SYNCPT_WAIT IOCTL to
+ * wait for this job to be finished.
+ */
+ __u32 fence;
+
+ /**
+ * @reserved:
+ *
+ * This field is reserved for future use. Must be 0.
+ */
+ __u32 reserved[5];
};
#define DRM_TEGRA_GEM_TILING_MODE_PITCH 0
#define DRM_TEGRA_GEM_TILING_MODE_TILED 1
#define DRM_TEGRA_GEM_TILING_MODE_BLOCK 2
+/**
+ * struct drm_tegra_gem_set_tiling - parameters for the set tiling IOCTL
+ */
struct drm_tegra_gem_set_tiling {
- /* input */
+ /**
+ * @handle:
+ *
+ * Handle to the GEM object for which to set the tiling parameters.
+ */
__u32 handle;
+
+ /**
+ * @mode:
+ *
+ * The tiling mode to set. Must be one of:
+ *
+ * DRM_TEGRA_GEM_TILING_MODE_PITCH
+ * pitch linear format
+ *
+ * DRM_TEGRA_GEM_TILING_MODE_TILED
+ * 16x16 tiling format
+ *
+ * DRM_TEGRA_GEM_TILING_MODE_BLOCK
+ * 16Bx2 tiling format
+ */
__u32 mode;
+
+ /**
+ * @value:
+ *
+ * The value to set for the tiling mode parameter.
+ */
__u32 value;
+
+ /**
+ * @pad:
+ *
+ * Structure padding that may be used in the future. Must be 0.
+ */
__u32 pad;
};
+/**
+ * struct drm_tegra_gem_get_tiling - parameters for the get tiling IOCTL
+ */
struct drm_tegra_gem_get_tiling {
- /* input */
+ /**
+ * @handle:
+ *
+ * Handle to the GEM object for which to query the tiling parameters.
+ */
__u32 handle;
- /* output */
+
+ /**
+ * @mode:
+ *
+ * The tiling mode currently associated with the GEM object. Set by
+ * the kernel upon successful completion of the IOCTL.
+ */
__u32 mode;
+
+ /**
+ * @value:
+ *
+ * The tiling mode parameter currently associated with the GEM object.
+ * Set by the kernel upon successful completion of the IOCTL.
+ */
__u32 value;
+
+ /**
+ * @pad:
+ *
+ * Structure padding that may be used in the future. Must be 0.
+ */
__u32 pad;
};
#define DRM_TEGRA_GEM_BOTTOM_UP (1 << 0)
#define DRM_TEGRA_GEM_FLAGS (DRM_TEGRA_GEM_BOTTOM_UP)
+/**
+ * struct drm_tegra_gem_set_flags - parameters for the set flags IOCTL
+ */
struct drm_tegra_gem_set_flags {
- /* input */
+ /**
+ * @handle:
+ *
+ * Handle to the GEM object for which to set the flags.
+ */
__u32 handle;
- /* output */
+
+ /**
+ * @flags:
+ *
+ * The flags to set for the GEM object.
+ */
__u32 flags;
};
+/**
+ * struct drm_tegra_gem_get_flags - parameters for the get flags IOCTL
+ */
struct drm_tegra_gem_get_flags {
- /* input */
+ /**
+ * @handle:
+ *
+ * Handle to the GEM object for which to query the flags.
+ */
__u32 handle;
- /* output */
+
+ /**
+ * @flags:
+ *
+ * The flags currently associated with the GEM object. Set by the
+ * kernel upon successful completion of the IOCTL.
+ */
__u32 flags;
};
@@ -193,7 +665,7 @@ struct drm_tegra_gem_get_flags {
#define DRM_IOCTL_TEGRA_SYNCPT_INCR DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_SYNCPT_INCR, struct drm_tegra_syncpt_incr)
#define DRM_IOCTL_TEGRA_SYNCPT_WAIT DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_SYNCPT_WAIT, struct drm_tegra_syncpt_wait)
#define DRM_IOCTL_TEGRA_OPEN_CHANNEL DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_OPEN_CHANNEL, struct drm_tegra_open_channel)
-#define DRM_IOCTL_TEGRA_CLOSE_CHANNEL DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_CLOSE_CHANNEL, struct drm_tegra_open_channel)
+#define DRM_IOCTL_TEGRA_CLOSE_CHANNEL DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_CLOSE_CHANNEL, struct drm_tegra_close_channel)
#define DRM_IOCTL_TEGRA_GET_SYNCPT DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GET_SYNCPT, struct drm_tegra_get_syncpt)
#define DRM_IOCTL_TEGRA_SUBMIT DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_SUBMIT, struct drm_tegra_submit)
#define DRM_IOCTL_TEGRA_GET_SYNCPT_BASE DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GET_SYNCPT_BASE, struct drm_tegra_get_syncpt_base)
diff --git a/include/drm-uapi/vc4_drm.h b/include/drm-uapi/vc4_drm.h
index 4117117b..31f50de3 100644
--- a/include/drm-uapi/vc4_drm.h
+++ b/include/drm-uapi/vc4_drm.h
@@ -183,10 +183,17 @@ struct drm_vc4_submit_cl {
/* ID of the perfmon to attach to this job. 0 means no perfmon. */
__u32 perfmonid;
- /* Unused field to align this struct on 64 bits. Must be set to 0.
- * If one ever needs to add an u32 field to this struct, this field
- * can be used.
+ /* Syncobj handle to wait on. If set, processing of this render job
+ * will not start until the syncobj is signaled. 0 means ignore.
*/
+ __u32 in_sync;
+
+ /* Syncobj handle to export fence to. If set, the fence in the syncobj
+ * will be replaced with a fence that signals upon completion of this
+ * render job. 0 means ignore.
+ */
+ __u32 out_sync;
+
__u32 pad2;
};
diff --git a/include/drm-uapi/virtgpu_drm.h b/include/drm-uapi/virtgpu_drm.h
index 91a31ffe..9a781f06 100644
--- a/include/drm-uapi/virtgpu_drm.h
+++ b/include/drm-uapi/virtgpu_drm.h
@@ -63,6 +63,7 @@ struct drm_virtgpu_execbuffer {
};
#define VIRTGPU_PARAM_3D_FEATURES 1 /* do we have 3D features in the hw */
+#define VIRTGPU_PARAM_CAPSET_QUERY_FIX 2 /* do we have the capset fix */
struct drm_virtgpu_getparam {
__u64 param;
diff --git a/include/drm-uapi/vmwgfx_drm.h b/include/drm-uapi/vmwgfx_drm.h
index 0bc784f5..399f5831 100644
--- a/include/drm-uapi/vmwgfx_drm.h
+++ b/include/drm-uapi/vmwgfx_drm.h
@@ -40,6 +40,7 @@ extern "C" {
#define DRM_VMW_GET_PARAM 0
#define DRM_VMW_ALLOC_DMABUF 1
+#define DRM_VMW_ALLOC_BO 1
#define DRM_VMW_UNREF_DMABUF 2
#define DRM_VMW_HANDLE_CLOSE 2
#define DRM_VMW_CURSOR_BYPASS 3
@@ -68,6 +69,8 @@ extern "C" {
#define DRM_VMW_GB_SURFACE_REF 24
#define DRM_VMW_SYNCCPU 25
#define DRM_VMW_CREATE_EXTENDED_CONTEXT 26
+#define DRM_VMW_GB_SURFACE_CREATE_EXT 27
+#define DRM_VMW_GB_SURFACE_REF_EXT 28
/*************************************************************************/
/**
@@ -79,6 +82,9 @@ extern "C" {
*
* DRM_VMW_PARAM_OVERLAY_IOCTL:
* Does the driver support the overlay ioctl.
+ *
+ * DRM_VMW_PARAM_SM4_1
+ * SM4_1 support is enabled.
*/
#define DRM_VMW_PARAM_NUM_STREAMS 0
@@ -94,6 +100,8 @@ extern "C" {
#define DRM_VMW_PARAM_MAX_MOB_SIZE 10
#define DRM_VMW_PARAM_SCREEN_TARGET 11
#define DRM_VMW_PARAM_DX 12
+#define DRM_VMW_PARAM_HW_CAPS2 13
+#define DRM_VMW_PARAM_SM4_1 14
/**
* enum drm_vmw_handle_type - handle type for ref ioctls
@@ -356,9 +364,9 @@ struct drm_vmw_fence_rep {
/*************************************************************************/
/**
- * DRM_VMW_ALLOC_DMABUF
+ * DRM_VMW_ALLOC_BO
*
- * Allocate a DMA buffer that is visible also to the host.
+ * Allocate a buffer object that is visible also to the host.
* NOTE: The buffer is
* identified by a handle and an offset, which are private to the guest, but
* useable in the command stream. The guest kernel may translate these
@@ -366,27 +374,28 @@ struct drm_vmw_fence_rep {
* be zero at all times, or it may disappear from the interface before it is
* fixed.
*
- * The DMA buffer may stay user-space mapped in the guest at all times,
+ * The buffer object may stay user-space mapped in the guest at all times,
* and is thus suitable for sub-allocation.
*
- * DMA buffers are mapped using the mmap() syscall on the drm device.
+ * Buffer objects are mapped using the mmap() syscall on the drm device.
*/
/**
- * struct drm_vmw_alloc_dmabuf_req
+ * struct drm_vmw_alloc_bo_req
*
* @size: Required minimum size of the buffer.
*
- * Input data to the DRM_VMW_ALLOC_DMABUF Ioctl.
+ * Input data to the DRM_VMW_ALLOC_BO Ioctl.
*/
-struct drm_vmw_alloc_dmabuf_req {
+struct drm_vmw_alloc_bo_req {
__u32 size;
__u32 pad64;
};
+#define drm_vmw_alloc_dmabuf_req drm_vmw_alloc_bo_req
/**
- * struct drm_vmw_dmabuf_rep
+ * struct drm_vmw_bo_rep
*
* @map_handle: Offset to use in the mmap() call used to map the buffer.
* @handle: Handle unique to this buffer. Used for unreferencing.
@@ -395,50 +404,32 @@ struct drm_vmw_alloc_dmabuf_req {
* @cur_gmr_offset: Offset to use in the command stream when this buffer is
* referenced. See note above.
*
- * Output data from the DRM_VMW_ALLOC_DMABUF Ioctl.
+ * Output data from the DRM_VMW_ALLOC_BO Ioctl.
*/
-struct drm_vmw_dmabuf_rep {
+struct drm_vmw_bo_rep {
__u64 map_handle;
__u32 handle;
__u32 cur_gmr_id;
__u32 cur_gmr_offset;
__u32 pad64;
};
+#define drm_vmw_dmabuf_rep drm_vmw_bo_rep
/**
- * union drm_vmw_dmabuf_arg
+ * union drm_vmw_alloc_bo_arg
*
* @req: Input data as described above.
* @rep: Output data as described above.
*
- * Argument to the DRM_VMW_ALLOC_DMABUF Ioctl.
+ * Argument to the DRM_VMW_ALLOC_BO Ioctl.
*/
-union drm_vmw_alloc_dmabuf_arg {
- struct drm_vmw_alloc_dmabuf_req req;
- struct drm_vmw_dmabuf_rep rep;
-};
-
-/*************************************************************************/
-/**
- * DRM_VMW_UNREF_DMABUF - Free a DMA buffer.
- *
- */
-
-/**
- * struct drm_vmw_unref_dmabuf_arg
- *
- * @handle: Handle indicating what buffer to free. Obtained from the
- * DRM_VMW_ALLOC_DMABUF Ioctl.
- *
- * Argument to the DRM_VMW_UNREF_DMABUF Ioctl.
- */
-
-struct drm_vmw_unref_dmabuf_arg {
- __u32 handle;
- __u32 pad64;
+union drm_vmw_alloc_bo_arg {
+ struct drm_vmw_alloc_bo_req req;
+ struct drm_vmw_bo_rep rep;
};
+#define drm_vmw_alloc_dmabuf_arg drm_vmw_alloc_bo_arg
/*************************************************************************/
/**
@@ -1103,9 +1094,8 @@ union drm_vmw_extended_context_arg {
* DRM_VMW_HANDLE_CLOSE - Close a user-space handle and release its
* underlying resource.
*
- * Note that this ioctl is overlaid on the DRM_VMW_UNREF_DMABUF Ioctl.
- * The ioctl arguments therefore need to be identical in layout.
- *
+ * Note that this ioctl is overlaid on the deprecated DRM_VMW_UNREF_DMABUF
+ * Ioctl.
*/
/**
@@ -1119,7 +1109,107 @@ struct drm_vmw_handle_close_arg {
__u32 handle;
__u32 pad64;
};
+#define drm_vmw_unref_dmabuf_arg drm_vmw_handle_close_arg
+
+/*************************************************************************/
+/**
+ * DRM_VMW_GB_SURFACE_CREATE_EXT - Create a host guest-backed surface.
+ *
+ * Allocates a surface handle and queues a create surface command
+ * for the host on the first use of the surface. The surface ID can
+ * be used as the surface ID in commands referencing the surface.
+ *
+ * This new command extends DRM_VMW_GB_SURFACE_CREATE by adding version
+ * parameter and 64 bit svga flag.
+ */
+
+/**
+ * enum drm_vmw_surface_version
+ *
+ * @drm_vmw_surface_gb_v1: Corresponds to current gb surface format with
+ * svga3d surface flags split into 2, upper half and lower half.
+ */
+enum drm_vmw_surface_version {
+ drm_vmw_gb_surface_v1
+};
+
+/**
+ * struct drm_vmw_gb_surface_create_ext_req
+ *
+ * @base: Surface create parameters.
+ * @version: Version of surface create ioctl.
+ * @svga3d_flags_upper_32_bits: Upper 32 bits of svga3d flags.
+ * @multisample_pattern: Multisampling pattern when msaa is supported.
+ * @quality_level: Precision settings for each sample.
+ * @must_be_zero: Reserved for future usage.
+ *
+ * Input argument to the DRM_VMW_GB_SURFACE_CREATE_EXT Ioctl.
+ * Part of output argument for the DRM_VMW_GB_SURFACE_REF_EXT Ioctl.
+ */
+struct drm_vmw_gb_surface_create_ext_req {
+ struct drm_vmw_gb_surface_create_req base;
+ enum drm_vmw_surface_version version;
+ uint32_t svga3d_flags_upper_32_bits;
+ SVGA3dMSPattern multisample_pattern;
+ SVGA3dMSQualityLevel quality_level;
+ uint64_t must_be_zero;
+};
+
+/**
+ * union drm_vmw_gb_surface_create_ext_arg
+ *
+ * @req: Input argument as described above.
+ * @rep: Output argument as described above.
+ *
+ * Argument to the DRM_VMW_GB_SURFACE_CREATE_EXT ioctl.
+ */
+union drm_vmw_gb_surface_create_ext_arg {
+ struct drm_vmw_gb_surface_create_rep rep;
+ struct drm_vmw_gb_surface_create_ext_req req;
+};
+
+/*************************************************************************/
+/**
+ * DRM_VMW_GB_SURFACE_REF_EXT - Reference a host surface.
+ *
+ * Puts a reference on a host surface with a given handle, as previously
+ * returned by the DRM_VMW_GB_SURFACE_CREATE_EXT ioctl.
+ * A reference will make sure the surface isn't destroyed while we hold
+ * it and will allow the calling client to use the surface handle in
+ * the command stream.
+ *
+ * On successful return, the Ioctl returns the surface information given
+ * to and returned from the DRM_VMW_GB_SURFACE_CREATE_EXT ioctl.
+ */
+/**
+ * struct drm_vmw_gb_surface_ref_ext_rep
+ *
+ * @creq: The data used as input when the surface was created, as described
+ * above at "struct drm_vmw_gb_surface_create_ext_req"
+ * @crep: Additional data output when the surface was created, as described
+ * above at "struct drm_vmw_gb_surface_create_rep"
+ *
+ * Output Argument to the DRM_VMW_GB_SURFACE_REF_EXT ioctl.
+ */
+struct drm_vmw_gb_surface_ref_ext_rep {
+ struct drm_vmw_gb_surface_create_ext_req creq;
+ struct drm_vmw_gb_surface_create_rep crep;
+};
+
+/**
+ * union drm_vmw_gb_surface_reference_ext_arg
+ *
+ * @req: Input data as described above at "struct drm_vmw_surface_arg"
+ * @rep: Output data as described above at
+ * "struct drm_vmw_gb_surface_ref_ext_rep"
+ *
+ * Argument to the DRM_VMW_GB_SURFACE_REF Ioctl.
+ */
+union drm_vmw_gb_surface_reference_ext_arg {
+ struct drm_vmw_gb_surface_ref_ext_rep rep;
+ struct drm_vmw_surface_arg req;
+};
#if defined(__cplusplus)
}
--
2.19.1
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply related [flat|nested] 11+ messages in thread* [igt-dev] [PATCH i-g-t v2 4/4] tests/perf: add new tests for configurable OA buffer size
2018-10-15 16:20 [igt-dev] [PATCH i-g-t v2 0/4] tests/perf: add support for new oa buffer size parameter Lionel Landwerlin
` (2 preceding siblings ...)
2018-10-15 16:20 ` [igt-dev] [PATCH i-g-t v2 3/4] tests/perf: rename MAX_OA_BUFFER_SIZE Lionel Landwerlin
@ 2018-10-15 16:20 ` Lionel Landwerlin
2018-10-22 18:33 ` Matthew Auld
2018-10-15 17:53 ` [igt-dev] ✓ Fi.CI.BAT: success for tests/perf: add support for new oa buffer size parameter (rev2) Patchwork
2018-10-15 23:12 ` [igt-dev] ✓ Fi.CI.IGT: " Patchwork
5 siblings, 1 reply; 11+ messages in thread
From: Lionel Landwerlin @ 2018-10-15 16:20 UTC (permalink / raw)
To: igt-dev
Verify that the buffer fills up within the expected time based of a
given size and also that invalid sizes are rejected.
v2: Update tests to reflect that we have to match the exact HW buffer
size (Lionel)
Bump the minimum buffer fill duration requirement to 40ms
(unreliable otherwise on small cores & ICL) (Lionel)
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
---
tests/perf.c | 120 +++++++++++++++++++++++++++++++++++++++++++++++----
1 file changed, 111 insertions(+), 9 deletions(-)
diff --git a/tests/perf.c b/tests/perf.c
index 6c448036..4b7bb689 100644
--- a/tests/perf.c
+++ b/tests/perf.c
@@ -2348,11 +2348,31 @@ test_polling(void)
__perf_close(stream_fd);
}
+static int
+find_oa_exponent_for_buffer_fill_time(size_t oa_buf_size, size_t report_size, uint64_t fill_time_ns)
+{
+ size_t n_reports = oa_buf_size / report_size;
+
+ for (int e = 1; e < 32; e++) {
+ if (fill_time_ns < oa_exponent_to_ns(e) * n_reports)
+ return e;
+ }
+
+ igt_assert(!"reached");
+ return -1;
+}
+
static void
-test_buffer_fill(void)
+test_buffer_fill(bool program_size, size_t oa_buf_size)
{
- /* ~5 micro second period */
- int oa_exponent = max_oa_exponent_for_period_lte(5000);
+ size_t report_size = get_oa_format(test_oa_format).size;
+ /*
+ * Select an exponent that guarantees that the buffer fills up
+ * in about 40ms. Otherwise the nanosleep might be too short
+ * which makes the test unreliable.
+ */
+ int oa_exponent = find_oa_exponent_for_buffer_fill_time(oa_buf_size, report_size,
+ 40 * 1000 * 1000UL);
uint64_t oa_period = oa_exponent_to_ns(oa_exponent);
uint64_t properties[] = {
/* Include OA reports in samples */
@@ -2362,18 +2382,18 @@ test_buffer_fill(void)
DRM_I915_PERF_PROP_OA_METRICS_SET, test_metric_set_id,
DRM_I915_PERF_PROP_OA_FORMAT, test_oa_format,
DRM_I915_PERF_PROP_OA_EXPONENT, oa_exponent,
+ DRM_I915_PERF_PROP_OA_BUFFER_SIZE, oa_buf_size,
};
struct drm_i915_perf_open_param param = {
.flags = I915_PERF_FLAG_FD_CLOEXEC,
- .num_properties = sizeof(properties) / 16,
+ .num_properties = program_size ?
+ (ARRAY_SIZE(properties) / 2) : ((ARRAY_SIZE(properties) / 2) - 1),
.properties_ptr = to_user_pointer(properties),
};
struct drm_i915_perf_record_header *header;
int buf_size = 65536 * (256 + sizeof(struct drm_i915_perf_record_header));
uint8_t *buf = malloc(buf_size);
int len;
- size_t oa_buf_size = DEFAULT_OA_BUF_SIZE;
- size_t report_size = get_oa_format(test_oa_format).size;
int n_full_oa_reports = oa_buf_size / report_size;
uint64_t fill_duration = n_full_oa_reports * oa_period;
@@ -2410,8 +2430,9 @@ test_buffer_fill(void)
do_ioctl(stream_fd, I915_PERF_IOCTL_DISABLE, 0);
- igt_debug("fill_duration = %"PRIu64"ns, oa_exponent = %u\n",
- fill_duration, oa_exponent);
+ igt_debug("fill_duration = %"PRIu64"ns/%"PRIu64"us/%"PRIu64"ms, oa_exponent = %u, buf_size = %lu\n",
+ fill_duration, fill_duration / 1000, fill_duration / 1000000,
+ oa_exponent, oa_buf_size);
do_ioctl(stream_fd, I915_PERF_IOCTL_ENABLE, 0);
@@ -2491,6 +2512,43 @@ test_buffer_fill(void)
__perf_close(stream_fd);
}
+static void
+test_invalid_buffer_size()
+{
+ uint64_t properties[] = {
+ /* Include OA reports in samples */
+ DRM_I915_PERF_PROP_SAMPLE_OA, true,
+
+ /* OA unit configuration */
+ DRM_I915_PERF_PROP_OA_METRICS_SET, test_metric_set_id,
+ DRM_I915_PERF_PROP_OA_FORMAT, test_oa_format,
+ DRM_I915_PERF_PROP_OA_EXPONENT, max_oa_exponent_for_period_lte(5000),
+ DRM_I915_PERF_PROP_OA_BUFFER_SIZE, 0 /* updated below */,
+ };
+ struct drm_i915_perf_open_param param = {
+ .flags = I915_PERF_FLAG_FD_CLOEXEC,
+ .num_properties = ARRAY_SIZE(properties) / 2,
+ .properties_ptr = to_user_pointer(properties),
+ };
+ size_t size = 128 * 1024;
+
+ while (size <= DEFAULT_OA_BUF_SIZE) {
+ properties[ARRAY_SIZE(properties) - 1] = size + 2;
+ do_ioctl_err(drm_fd, DRM_IOCTL_I915_PERF_OPEN, ¶m, EINVAL);
+
+ size *= 2;
+ }
+
+ properties[ARRAY_SIZE(properties) - 1] = DEFAULT_OA_BUF_SIZE * 2;
+ do_ioctl_err(drm_fd, DRM_IOCTL_I915_PERF_OPEN, ¶m, EINVAL);
+
+ properties[ARRAY_SIZE(properties) - 1] = 0xffffffff;
+ do_ioctl_err(drm_fd, DRM_IOCTL_I915_PERF_OPEN, ¶m, EINVAL);
+
+ properties[ARRAY_SIZE(properties) - 1] = 0xffffffffffffffff;
+ do_ioctl_err(drm_fd, DRM_IOCTL_I915_PERF_OPEN, ¶m, EINVAL);
+}
+
static void
test_enable_disable(void)
{
@@ -4093,6 +4151,32 @@ test_sysctl_defaults(void)
igt_assert_eq(max_freq, 100000);
}
+static bool
+kernel_supports_open_option(int fd, uint64_t option, uint64_t value)
+{
+ uint64_t properties[] = {
+ /* Intentionally wrong handle */
+ DRM_I915_PERF_PROP_CTX_HANDLE, UINT64_MAX,
+
+ DRM_I915_PERF_PROP_SAMPLE_OA, true,
+ DRM_I915_PERF_PROP_OA_METRICS_SET, test_metric_set_id,
+ DRM_I915_PERF_PROP_OA_FORMAT, test_oa_format,
+ DRM_I915_PERF_PROP_OA_EXPONENT, max_oa_exponent_for_period_lte(5000),
+ option, value,
+ };
+ struct drm_i915_perf_open_param param = {
+ .flags = I915_PERF_FLAG_FD_CLOEXEC,
+ .num_properties = ARRAY_SIZE(properties) / 2,
+ .properties_ptr = to_user_pointer(properties),
+ };
+ int ret;
+
+ ret = igt_ioctl(fd, DRM_IOCTL_I915_PERF_OPEN, ¶m);
+ assert(ret == -1);
+
+ return errno == ENOENT;
+}
+
igt_main
{
igt_skip_on_simulation();
@@ -4166,7 +4250,25 @@ igt_main
}
igt_subtest("buffer-fill")
- test_buffer_fill();
+ test_buffer_fill(false, DEFAULT_OA_BUF_SIZE);
+
+ igt_subtest("buffer-fill-sized") {
+ size_t size = 128 * 1024; /* Smallest HW supported size. */
+ const size_t max_size = DEFAULT_OA_BUF_SIZE;
+
+ igt_require(kernel_supports_open_option(drm_fd, DRM_I915_PERF_PROP_OA_BUFFER_SIZE, 0));
+
+ while (size <= max_size) {
+ test_buffer_fill(true, size);
+ size *= 2;
+ }
+ }
+
+ igt_subtest("invalid-buffer-size") {
+ igt_require(kernel_supports_open_option(drm_fd, DRM_I915_PERF_PROP_OA_BUFFER_SIZE, 0));
+
+ test_invalid_buffer_size();
+ }
igt_subtest("disabled-read-error")
test_disabled_read_error();
--
2.19.1
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply related [flat|nested] 11+ messages in thread