Re: [PATCH v2 1/4] drm/i915/gt: Clear compress metadata for Xe_HP platforms

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, 2022-03-02 at 03:23 +0530, Ramalingam C wrote:
> From: Ayaz A Siddiqui <ayaz.siddiqui@xxxxxxxxx>
> 
> Xe-HP and latest devices support Flat CCS which reserved a portion of
> the device memory to store compression metadata, during the clearing
> of
> device memory buffer object we also need to clear the associated
> CCS buffer.
> 
> Flat CCS memory can not be directly accessed by S/W.
> Address of CCS buffer associated main BO is automatically calculated
> by device itself. KMD/UMD can only access this buffer indirectly
> using
> XY_CTRL_SURF_COPY_BLT cmd via the address of device memory buffer.
> 
> v2: Fixed issues with platform naming [Lucas]
> v3: Rebased [Ram]
>     Used the round_up funcs [Bob]
> v4: Fixed ccs blk calculation [Ram]
>     Added Kdoc on flat-ccs.
> v5: GENMASK is used [Matt]
>     mocs fix [Matt]
>     Comments Fix [Matt]
>     Flush address programming [Ram]
> v6: FLUSH_DW is fixed
>     Few coding style fix
> 
> Signed-off-by: Ayaz A Siddiqui <ayaz.siddiqui@xxxxxxxxx>
> Signed-off-by: Ramalingam C <ramalingam.c@xxxxxxxxx>
> ---
>  drivers/gpu/drm/i915/gt/intel_gpu_commands.h |  15 ++
>  drivers/gpu/drm/i915/gt/intel_migrate.c      | 143
> ++++++++++++++++++-
>  2 files changed, 154 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
> b/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
> index f8253012d166..237c1baccc64 100644
> --- a/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
> +++ b/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
> @@ -203,6 +203,21 @@
>  #define GFX_OP_DRAWRECT_INFO    
> ((0x3<<29)|(0x1d<<24)|(0x80<<16)|(0x3))
>  #define GFX_OP_DRAWRECT_INFO_I965  ((0x7900<<16)|0x2)
>  
> +#define XY_CTRL_SURF_INSTR_SIZE        5
> +#define MI_FLUSH_DW_SIZE               3
> +#define XY_CTRL_SURF_COPY_BLT          ((2 << 29) | (0x48 << 22) |
> 3)
> +#define   SRC_ACCESS_TYPE_SHIFT                21
> +#define   DST_ACCESS_TYPE_SHIFT                20
> +#define   CCS_SIZE_MASK                        GENMASK(17, 8)
> +#define   XY_CTRL_SURF_MOCS_MASK       GENMASK(31, 25)
> +#define   NUM_CCS_BYTES_PER_BLOCK      256
> +#define   NUM_BYTES_PER_CCS_BYTE       256
> +#define   NUM_CCS_BLKS_PER_XFER                1024
> +#define   INDIRECT_ACCESS              0
> +#define   DIRECT_ACCESS                        1
> +#define  MI_FLUSH_LLC                  BIT(9)
> +#define  MI_FLUSH_CCS                  BIT(16)
> +
>  #define COLOR_BLT_CMD                  (2 << 29 | 0x40 << 22 | (5 -
> 2))
>  #define XY_COLOR_BLT_CMD               (2 << 29 | 0x50 << 22)
>  #define SRC_COPY_BLT_CMD               (2 << 29 | 0x43 << 22)
> diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.c
> b/drivers/gpu/drm/i915/gt/intel_migrate.c
> index 20444d6ceb3c..330fcdc3e0cf 100644
> --- a/drivers/gpu/drm/i915/gt/intel_migrate.c
> +++ b/drivers/gpu/drm/i915/gt/intel_migrate.c
> @@ -16,6 +16,8 @@ struct insert_pte_data {
>  };
>  
>  #define CHUNK_SZ SZ_8M /* ~1ms at 8GiB/s preemption delay */
> +#define GET_CCS_BYTES(i915, size)      (HAS_FLAT_CCS(i915) ? \
> +                                        DIV_ROUND_UP(size,
> NUM_BYTES_PER_CCS_BYTE) : 0)
>  
>  static bool engine_supports_migration(struct intel_engine_cs
> *engine)
>  {
> @@ -467,6 +469,110 @@ static bool wa_1209644611_applies(int ver, u32
> size)
>         return height % 4 == 3 && height <= 8;
>  }
>  
> +/**
> + * DOC: Flat-CCS - Memory compression for Local memory
> + *
> + * On Xe-HP and later devices, we use dedicated compression control
> state (CCS)
> + * stored in local memory for each surface, to support the 3D and
> media
> + * compression formats.
> + *
> + * The memory required for the CCS of the entire local memory is
> 1/256 of the
> + * local memory size. So before the kernel boot, the required memory
> is reserved
> + * for the CCS data and a secure register will be programmed with
> the CCS base
> + * address.
> + *
> + * Flat CCS data needs to be cleared when a lmem object is
> allocated.
> + * And CCS data can be copied in and out of CCS region through
> + * XY_CTRL_SURF_COPY_BLT. CPU can't access the CCS data directly.
> + *
> + * When we exhaust the lmem, if the object's placements support
> smem, then we can
> + * directly decompress the compressed lmem object into smem and
> start using it
> + * from smem itself.
> + *
> + * But when we need to swapout the compressed lmem object into a
> smem region
> + * though objects' placement doesn't support smem, then we copy the
> lmem content
> + * as it is into smem region along with ccs data (using
> XY_CTRL_SURF_COPY_BLT).
> + * When the object is referred, lmem content will be swaped in along
> with
> + * restoration of the CCS data (using XY_CTRL_SURF_COPY_BLT) at
> corresponding
> + * location.
> + */
> +
> +static inline u32 *i915_flush_dw(u32 *cmd, u32 flags)
> +{
> +       *cmd++ = MI_FLUSH_DW | flags;
> +       *cmd++ = 0;
> +       *cmd++ = 0;
> +
> +       return cmd;
> +}
> +
> +static u32 calc_ctrl_surf_instr_size(struct drm_i915_private *i915,
> int size)
> +{
> +       u32 num_cmds, num_blks, total_size;
> +
> +       if (!GET_CCS_BYTES(i915, size))
> +               return 0;
> +
> +       /*
> +        * XY_CTRL_SURF_COPY_BLT transfers CCS in 256 byte
> +        * blocks. one XY_CTRL_SURF_COPY_BLT command can
> +        * transfer upto 1024 blocks.
> +        */
> +       num_blks = DIV_ROUND_UP(GET_CCS_BYTES(i915, size),
> +                               NUM_CCS_BYTES_PER_BLOCK);
> +       num_cmds = DIV_ROUND_UP(num_blks, NUM_CCS_BLKS_PER_XFER);
> +       total_size = XY_CTRL_SURF_INSTR_SIZE * num_cmds;
> +
> +       /*
> +        * Adding a flush before and after XY_CTRL_SURF_COPY_BLT
> +        */
> +       total_size += 2 * MI_FLUSH_DW_SIZE;
> +
> +       return total_size;
> +}
> +

Since we should always interleave the ctrl_surf_copy_blt() on max
CHUNK_SZ pieces of LMEM (See also patch 4/4), I figure we would never
need to split the command since it can do 64M worth of LMEM in a single
command vs a CHUNK_SZ of 8M. Instead perhaps an assert that CHUNK_SZ
never exceeds the capability of the XY_CTRL_SURF_COPY_BLT?

Also I think it's important that we try to figure out whether we can
use the XY_FAST_COLOR_BLT command to clear also CCS on DG2. Would save
us a lot of code, and also at least on DG1 (without CCS) it speeds
clearing up significantly.

/Thomas

> +static u32 *_i915_ctrl_surf_copy_blt(u32 *cmd, u64 src_addr, u64
> dst_addr,
> +                                    u8 src_mem_access, u8
> dst_mem_access,
> +                                    int src_mocs, int dst_mocs,
> +                                    u32 ccs_blocks)
> +{
> +       /*
> +        * The XY_CTRL_SURF_COPY_BLT instruction is used to copy the
> CCS
> +        * data in and out of the CCS region.
> +        *
> +        * We can copy at most 1024 blocks of 256 bytes using one
> +        * XY_CTRL_SURF_COPY_BLT instruction.
> +        *
> +        * In case we need to copy more than 1024 blocks, we need to
> add
> +        * another instruction to the same batch buffer.
> +        *
> +        * 1024 blocks of 256 bytes of CCS represent a total 256KB of
> CCS.
> +        *
> +        * 256 KB of CCS represents 256 * 256 KB = 64 MB of LMEM.
> +        */
> +       do {
> +               int blks_per_copy;
> +
> +               blks_per_copy = ccs_blocks >= NUM_CCS_BLKS_PER_XFER ?
> +                               NUM_CCS_BLKS_PER_XFER : ccs_blocks;
> +               *cmd++ = XY_CTRL_SURF_COPY_BLT |
> +                        src_mem_access << SRC_ACCESS_TYPE_SHIFT |
> +                        dst_mem_access << DST_ACCESS_TYPE_SHIFT |
> +                        FIELD_PREP(CCS_SIZE_MASK, blks_per_copy -
> 1);
> +               *cmd++ = lower_32_bits(src_addr);
> +               *cmd++ = (upper_32_bits(src_addr) & 0xFFFF) |
> +                         FIELD_PREP(XY_CTRL_SURF_MOCS_MASK,
> src_mocs);
> +               *cmd++ = lower_32_bits(dst_addr);
> +               *cmd++ = (upper_32_bits(dst_addr) & 0xFFFF) |
> +                         FIELD_PREP(XY_CTRL_SURF_MOCS_MASK,
> dst_mocs);
> +               src_addr += SZ_64M;
> +               dst_addr += SZ_64M;
> +               ccs_blocks -= blks_per_copy;
> +       } while (ccs_blocks > 0);
> +
> +       return cmd;
> +}
> +
>  static int emit_copy(struct i915_request *rq,
>                      u32 dst_offset, u32 src_offset, int size)
>  {
> @@ -614,16 +720,24 @@ intel_context_migrate_copy(struct intel_context
> *ce,
>         return err;
>  }
>  
> -static int emit_clear(struct i915_request *rq, u64 offset, int size,
> u32 value)
> +static int emit_clear(struct i915_request *rq, u64 offset, int size,
> +                     u32 value, bool is_lmem)
>  {
> -       const int ver = GRAPHICS_VER(rq->engine->i915);
> +       struct drm_i915_private *i915 = rq->engine->i915;
> +       const int ver = GRAPHICS_VER(i915);
> +       u32 num_ccs_blks, ccs_ring_size;
> +       int mocs = rq->engine->gt->mocs.uc_index << 1;
>         u32 *cs;
>  
>         GEM_BUG_ON(size >> PAGE_SHIFT > S16_MAX);
>  
>         offset += (u64)rq->engine->instance << 32;
>  
> -       cs = intel_ring_begin(rq, ver >= 8 ? 8 : 6);
> +       /* Clear CCS only when value is 0 */
> +       ccs_ring_size = (is_lmem && !value) ?
> +                        calc_ctrl_surf_instr_size(i915, size) : 0;
> +
> +       cs = intel_ring_begin(rq, round_up(ver >= 8 ? 8 +
> ccs_ring_size : 6, 2));
>         if (IS_ERR(cs))
>                 return PTR_ERR(cs);
>  
> @@ -646,6 +760,27 @@ static int emit_clear(struct i915_request *rq,
> u64 offset, int size, u32 value)
>                 *cs++ = value;
>         }
>  
> +       if (is_lmem && HAS_FLAT_CCS(i915) && !value) {
> +               num_ccs_blks = DIV_ROUND_UP(GET_CCS_BYTES(i915,
> size),
> +                                           NUM_CCS_BYTES_PER_BLOCK);
> +
> +               /*
> +                * Flat CCS surface can only be accessed via
> +                * XY_CTRL_SURF_COPY_BLT CMD and using indirect
> +                * mapping of associated LMEM.
> +                * We can clear ccs surface by writing all 0s,
> +                * so we will flush the previously cleared buffer
> +                * and use it as a source.
> +                */
> +               cs = i915_flush_dw(cs, MI_FLUSH_LLC | MI_FLUSH_CCS);
> +               cs = _i915_ctrl_surf_copy_blt(cs, offset, offset,
> +                                             DIRECT_ACCESS,
> INDIRECT_ACCESS,
> +                                             mocs, mocs,
> num_ccs_blks);
> +               cs = i915_flush_dw(cs, MI_FLUSH_LLC | MI_FLUSH_CCS);
> +
> +               if (ccs_ring_size & 1)
> +                       *cs++ = MI_NOOP;
> +       }
>         intel_ring_advance(rq, cs);
>         return 0;
>  }
> @@ -711,7 +846,7 @@ intel_context_migrate_clear(struct intel_context
> *ce,
>                 if (err)
>                         goto out_rq;
>  
> -               err = emit_clear(rq, offset, len, value);
> +               err = emit_clear(rq, offset, len, value, is_lmem);
>  
>                 /* Arbitration is re-enabled between requests. */
>  out_rq:

----------------------------------------------------------------------
Intel Sweden AB
Registered Office: Isafjordsgatan 30B, 164 40 Kista, Stockholm, Sweden
Registration Number: 556189-6027

This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.




[Index of Archives]     [AMD Graphics]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux