Re: [Intel-gfx] [PATCH 7/8] drm/i915: use pat_index instead of cache_level

["Yang, Fei" <fei.yang@xxxxxxxxx>]

> On 20/04/2023 11:13, Andrzej Hajda wrote:

>> On 20.04.2023 01:00, fei.yang@xxxxxxxxx wrote:

>>> From: Fei Yang <fei.yang@xxxxxxxxx>

>>>

>>> Currently the KMD is using enum i915_cache_level to set caching policy

>>> for

>>> buffer objects. This is flaky because the PAT index which really controls

>>> the caching behavior in PTE has far more levels than what's defined in

>>> the

>>> enum. In addition, the PAT index is platform dependent, having to

>>> translate

>>> between i915_cache_level and PAT index is not reliable, and makes the

>>> code

>>> more complicated.

>

> How it is flaky and not reliable - yet the series proposed to leave it in

> place and even claims using cache levels simplifies the code (lower in the

> commit message). Maybe just the commit message needs work.

If you look into the PTE encode functions, using cache_level there is not even

correct. There is no way to map the 4 possible cache levels to all available PAT

indices, let alone the number of PAT indices varies from platofrm to platform.

The architectural design allows UMD's to directly set PAT index, this is not

possible with the use of cache_level in KMD. PAT index from UMD -> cache_level

-> PAT index bits in PTE? Such translation is too much of a headache.

However in the kernel space, cache_level is still a good enough abstraction for

KMD objects. Removing that would require code like this,

if (need to set uncached mode) {

      if (MTL)

            pat_index = 2;

      else if (PVC)

            pat_index = 0;

      else if (GEN12)

            pat_index = 3;

}

>>>  From UMD's perspective there is also a necessity to set caching

>>> policy for

>>> performance fine tuning. It's much easier for the UMD to directly use PAT

>>> index because the behavior of each PAT index is clearly defined in Bspec.

>>> Having the abstracted i915_cache_level sitting in between would only

>>> cause

>>> more ambiguity.

>>>

>>> For these reasons this patch replaces i915_cache_level with PAT index.

>>> Also

>>> note, the cache_level is not completely removed yet, because the KMD

>>> still

>>> has the need of creating buffer objects with simple cache settings

>>> such as

>>> cached, uncached, or writethrough. For such simple cases, using

>>> cache_level

>>> would help simplify the code.

>>

>> It seems quite fundamental change to me. Does this "not completely

>> removed yet" means that in some future we will not have support for

>> generic cache levels at all?

I think further simplification is possible if there is a consistent PAT

table across all platforms. Looking at Xe 2/3, that seems to be the trend.

>>Seems strange to me. Even looking at the

>> number of users of i915_gem_get_pat_index below it seem very unlikely.

>>

>> And if the support for generic level will stay, maybe it would be better

>> to make usage of it more convienient. All conversion of

>>      f(..., cache_level, ...)

>> to

>>      f(..., i915_gem_get_pat_index(i915, cache_level), ...)

>> looks quite ugly to me.

>>

>> Maybe extending cache level to support pat index somehow, for example:

>> enum i915_cache_level {

>>      I915_CACHE_NONE = 0,

>>      I915_CACHE_...,

>>      ...

>>      I915_CACHE_1ST_PAT_INDEX = 0x100,

>> }

>>

>> so real_pat_index = cache_level - I915_CACHE_1ST_PAT_INDEX

>>

>> and in case of generic level there will be platform dependend conversion

>> to real_pat_index?

>>

>> I do not know the whole picture so maybe this is all wrong for some

>> reason, just asking :)

>

> It looks a bit unsightly to me too so yes please, brain storm on whether

> it can be made more elegant and less intrusive would be appreciated.

I still think using pat_index without abstraction is better...

>>> Cc: Chris Wilson <chris.p.wilson@xxxxxxxxxxxxxxx>

>>> Cc: Matt Roper <matthew.d.roper@xxxxxxxxx>

>>> Signed-off-by: Fei Yang <fei.yang@xxxxxxxxx>

>>> Reviewed-by: Andi Shyti <andi.shyti@xxxxxxxxxxxxxxx>

>>> ---

>>>   drivers/gpu/drm/i915/display/intel_dpt.c      | 12 +--

>>>   drivers/gpu/drm/i915/gem/i915_gem_domain.c    | 27 ++----

>>>   .../gpu/drm/i915/gem/i915_gem_execbuffer.c    | 10 ++-

>>>   drivers/gpu/drm/i915/gem/i915_gem_mman.c      |  3 +-

>>>   drivers/gpu/drm/i915/gem/i915_gem_object.c    | 52 +++++++++++-

>>>   drivers/gpu/drm/i915/gem/i915_gem_object.h    |  4 +

>>>   .../gpu/drm/i915/gem/i915_gem_object_types.h  | 25 +++++-

>>>   drivers/gpu/drm/i915/gem/i915_gem_stolen.c    |  4 +-

>>>   drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c  | 16 ++--

>>>   .../gpu/drm/i915/gem/selftests/huge_pages.c   |  2 +-

>>>   .../drm/i915/gem/selftests/i915_gem_migrate.c |  2 +-

>>>   .../drm/i915/gem/selftests/i915_gem_mman.c    |  2 +-

>>>   drivers/gpu/drm/i915/gt/gen6_ppgtt.c          | 10 ++-

>>>   drivers/gpu/drm/i915/gt/gen8_ppgtt.c          | 71 ++++++++--------

>>>   drivers/gpu/drm/i915/gt/gen8_ppgtt.h          |  3 +-

>>>   drivers/gpu/drm/i915/gt/intel_ggtt.c          | 82 +++++++++----------

>>>   drivers/gpu/drm/i915/gt/intel_gtt.h           | 20 ++---

>>>   drivers/gpu/drm/i915/gt/intel_migrate.c       | 47 ++++++-----

>>>   drivers/gpu/drm/i915/gt/intel_migrate.h       | 13 ++-

>>>   drivers/gpu/drm/i915/gt/intel_ppgtt.c         |  6 +-

>>>   drivers/gpu/drm/i915/gt/selftest_migrate.c    | 47 ++++++-----

>>>   drivers/gpu/drm/i915/gt/selftest_reset.c      |  8 +-

>>>   drivers/gpu/drm/i915/gt/selftest_timeline.c   |  2 +-

>>>   drivers/gpu/drm/i915/gt/selftest_tlb.c        |  4 +-

>>>   drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c      | 10 ++-

>>>   drivers/gpu/drm/i915/i915_debugfs.c           | 55 ++++++++++---

>>>   drivers/gpu/drm/i915/i915_gem.c               | 16 +++-

>>>   drivers/gpu/drm/i915/i915_gpu_error.c         |  8 +-

>>>   drivers/gpu/drm/i915/i915_vma.c               | 16 ++--

>>>   drivers/gpu/drm/i915/i915_vma.h               |  2 +-

>>>   drivers/gpu/drm/i915/i915_vma_types.h         |  2 -

>>>   drivers/gpu/drm/i915/selftests/i915_gem.c     |  5 +-

>>>   .../gpu/drm/i915/selftests/i915_gem_evict.c   |  4 +-

>>>   drivers/gpu/drm/i915/selftests/i915_gem_gtt.c | 15 ++--

>>>   .../drm/i915/selftests/intel_memory_region.c  |  4 +-

>>>   drivers/gpu/drm/i915/selftests/mock_gtt.c     |  8 +-

>>>   36 files changed, 378 insertions(+), 239 deletions(-)

>>>

>>> diff --git a/drivers/gpu/drm/i915/display/intel_dpt.c

>>> b/drivers/gpu/drm/i915/display/intel_dpt.c

>>> index c5eacfdba1a5..7c5fddb203ba 100644

>>> --- a/drivers/gpu/drm/i915/display/intel_dpt.c

>>> +++ b/drivers/gpu/drm/i915/display/intel_dpt.c

>>> @@ -43,24 +43,24 @@ static void gen8_set_pte(void __iomem *addr,

>>> gen8_pte_t pte)

>>>   static void dpt_insert_page(struct i915_address_space *vm,

>>>                   dma_addr_t addr,

>>>                   u64 offset,

>>> -                enum i915_cache_level level,

>>> +                unsigned int pat_index,

>>>                   u32 flags)

>>>   {

>>>       struct i915_dpt *dpt = i915_vm_to_dpt(vm);

>>>       gen8_pte_t __iomem *base = dpt->iomem;

>>>       gen8_set_pte(base + offset / I915_GTT_PAGE_SIZE,

>>> -             vm->pte_encode(addr, level, flags));

>>> +             vm->pte_encode(addr, pat_index, flags));

>>>   }

>>>   static void dpt_insert_entries(struct i915_address_space *vm,

>>>                      struct i915_vma_resource *vma_res,

>>> -                   enum i915_cache_level level,

>>> +                   unsigned int pat_index,

>>>                      u32 flags)

>>>   {

>>>       struct i915_dpt *dpt = i915_vm_to_dpt(vm);

>>>       gen8_pte_t __iomem *base = dpt->iomem;

>>> -    const gen8_pte_t pte_encode = vm->pte_encode(0, level, flags);

>>> +    const gen8_pte_t pte_encode = vm->pte_encode(0, pat_index, flags);

>>>       struct sgt_iter sgt_iter;

>>>       dma_addr_t addr;

>>>       int i;

>>> @@ -83,7 +83,7 @@ static void dpt_clear_range(struct

>>> i915_address_space *vm,

>>>   static void dpt_bind_vma(struct i915_address_space *vm,

>>>                struct i915_vm_pt_stash *stash,

>>>                struct i915_vma_resource *vma_res,

>>> -             enum i915_cache_level cache_level,

>>> +             unsigned int pat_index,

>>>                u32 flags)

>>>   {

>>>       u32 pte_flags;

>>> @@ -98,7 +98,7 @@ static void dpt_bind_vma(struct i915_address_space *vm,

>>>       if (vma_res->bi.lmem)

>>>           pte_flags |= PTE_LM;

>>> -    vm->insert_entries(vm, vma_res, cache_level, pte_flags);

>>> +    vm->insert_entries(vm, vma_res, pat_index, pte_flags);

>>>       vma_res->page_sizes_gtt = I915_GTT_PAGE_SIZE;

>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_domain.c

>>> b/drivers/gpu/drm/i915/gem/i915_gem_domain.c

>>> index bb3575b1479f..d5fd4c9cd9f8 100644

>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_domain.c

>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_domain.c

>>> @@ -27,8 +27,8 @@ static bool gpu_write_needs_clflush(struct

>>> drm_i915_gem_object *obj)

>>>       if (IS_DGFX(i915))

>>>           return false;

>>> -    return !(obj->cache_level == I915_CACHE_NONE ||

>>> -         obj->cache_level == I915_CACHE_WT);

>>> +    return !(i915_gem_object_has_cache_level(obj, I915_CACHE_NONE) ||

>>> +         i915_gem_object_has_cache_level(obj, I915_CACHE_WT));

>>>   }

>>>   bool i915_gem_cpu_write_needs_clflush(struct drm_i915_gem_object *obj)

>>> @@ -267,7 +267,7 @@ int i915_gem_object_set_cache_level(struct

>>> drm_i915_gem_object *obj,

>>>   {

>>>       int ret;

>>> -    if (obj->cache_level == cache_level)

>>> +    if (i915_gem_object_has_cache_level(obj, cache_level))

>>>           return 0;

>>>       ret = i915_gem_object_wait(obj,

>>> @@ -278,10 +278,8 @@ int i915_gem_object_set_cache_level(struct

>>> drm_i915_gem_object *obj,

>>>           return ret;

>>>       /* Always invalidate stale cachelines */

>>> -    if (obj->cache_level != cache_level) {

>>> -        i915_gem_object_set_cache_coherency(obj, cache_level);

>>> -        obj->cache_dirty = true;

>>> -    }

>>> +    i915_gem_object_set_cache_coherency(obj, cache_level);

>>> +    obj->cache_dirty = true;

>>>       /* The cache-level will be applied when each vma is rebound. */

>>>       return i915_gem_object_unbind(obj,

>>> @@ -306,20 +304,13 @@ int i915_gem_get_caching_ioctl(struct drm_device

>>> *dev, void *data,

>>>           goto out;

>>>       }

>>> -    switch (obj->cache_level) {

>>> -    case I915_CACHE_LLC:

>>> -    case I915_CACHE_L3_LLC:

>>> +    if (i915_gem_object_has_cache_level(obj, I915_CACHE_LLC) ||

>>> +        i915_gem_object_has_cache_level(obj, I915_CACHE_L3_LLC))

>>>           args->caching = I915_CACHING_CACHED;

>>> -        break;

>>> -

>>> -    case I915_CACHE_WT:

>>> +    else if (i915_gem_object_has_cache_level(obj, I915_CACHE_WT))

>>>           args->caching = I915_CACHING_DISPLAY;

>>> -        break;

>>> -

>>> -    default:

>>> +    else

>>>           args->caching = I915_CACHING_NONE;

>>> -        break;

>>> -    }

>>>   out:

>>>       rcu_read_unlock();

>>>       return err;

>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c

>>> b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c

>>> index 3aeede6aee4d..d42915516636 100644

>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c

>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c

>>> @@ -642,7 +642,7 @@ static inline int use_cpu_reloc(const struct

>>> reloc_cache *cache,

>>>       return (cache->has_llc ||

>>>           obj->cache_dirty ||

>>> -        obj->cache_level != I915_CACHE_NONE);

>>> +        !i915_gem_object_has_cache_level(obj, I915_CACHE_NONE));

>>>   }

>>>   static int eb_reserve_vma(struct i915_execbuffer *eb,

>>> @@ -1323,8 +1323,10 @@ static void *reloc_iomap(struct i915_vma *batch,

>>>       offset = cache->node.start;

>>>       if (drm_mm_node_allocated(&cache->node)) {

>>>           ggtt->vm.insert_page(&ggtt->vm,

>>> -                     i915_gem_object_get_dma_address(obj, page),

>>> -                     offset, I915_CACHE_NONE, 0);

>>> +            i915_gem_object_get_dma_address(obj, page),

>>> +            offset,

>>> +            i915_gem_get_pat_index(ggtt->vm.i915, I915_CACHE_NONE),

>>> +            0);

>>>       } else {

>>>           offset += page << PAGE_SHIFT;

>>>       }

>>> @@ -1464,7 +1466,7 @@ eb_relocate_entry(struct i915_execbuffer *eb,

>>>               reloc_cache_unmap(&eb->reloc_cache);

>>>               mutex_lock(&vma->vm->mutex);

>>>               err = i915_vma_bind(target->vma,

>>> -                        target->vma->obj->cache_level,

>>> +                        target->vma->obj->pat_index,

>>>                           PIN_GLOBAL, NULL, NULL);

>>>               mutex_unlock(&vma->vm->mutex);

>>>               reloc_cache_remap(&eb->reloc_cache, ev->vma->obj);

>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.c

>>> b/drivers/gpu/drm/i915/gem/i915_gem_mman.c

>>> index 3dbacdf0911a..50c30efa08a3 100644

>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_mman.c

>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.c

>>> @@ -383,7 +383,8 @@ static vm_fault_t vm_fault_gtt(struct vm_fault *vmf)

>>>       }

>>>       /* Access to snoopable pages through the GTT is incoherent. */

>>> -    if (obj->cache_level != I915_CACHE_NONE && !HAS_LLC(i915)) {

>>> +    if (!(i915_gem_object_has_cache_level(obj, I915_CACHE_NONE) ||

>>> +          HAS_LLC(i915))) {

>>>           ret = -EFAULT;

>>>           goto err_unpin;

>>>       }

>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c

>>> b/drivers/gpu/drm/i915/gem/i915_gem_object.c

>>> index 8c70a0ec7d2f..27c948350b5b 100644

>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_object.c

>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c

>>> @@ -54,6 +54,25 @@ unsigned int i915_gem_get_pat_index(struct

>>> drm_i915_private *i915,

>>>       return INTEL_INFO(i915)->cachelevel_to_pat[level];

>>>   }

>>> +bool i915_gem_object_has_cache_level(const struct drm_i915_gem_object

>>> *obj,

>>> +                     enum i915_cache_level lvl)

>>

>> The name suggest object can have more cache levels, maybe only my

>> impression, up to you.

>>

>>> +{

>>> +    /*

>>> +     * cache_level == I915_CACHE_INVAL indicates the UMD's have set the

>>> +     * caching policy through pat_index, in which case the KMD should

>>> +     * leave the coherency to be managed by user space, simply return

>>> +     * true here.

>>> +     */

>>> +    if (obj->cache_level == I915_CACHE_INVAL)

>>> +        return true;

>

> It's a "bit" counter intuitive that answer "has cache level" is yes when

> cache level is set to invalid!

The only case for this condition to be true is an object created by UMD through

GEM_CREATE with PAT index specified by set_pat extension. In this case the KMD

is not supposed to touch the setting.

> I worry we don't create an impenetrable code base so I hope this can be improved.

>

>>> +

>>> +    /*

>>> +     * Otherwise the pat_index should have been converted from

>>> cache_level

>>> +     * so that the following comparison is valid.

>>> +     */

>>> +    return obj->pat_index == i915_gem_get_pat_index(obj_to_i915(obj),

>>> lvl);

>>> +}

>>> +

>>>   struct drm_i915_gem_object *i915_gem_object_alloc(void)

>>>   {

>>>       struct drm_i915_gem_object *obj;

>>> @@ -133,7 +152,7 @@ void i915_gem_object_set_cache_coherency(struct

>>> drm_i915_gem_object *obj,

>>>   {

>>>       struct drm_i915_private *i915 = to_i915(obj->base.dev);

>>> -    obj->cache_level = cache_level;

>>> +    obj->pat_index = i915_gem_get_pat_index(i915, cache_level);

>>>       if (cache_level != I915_CACHE_NONE)

>>>           obj->cache_coherent = (I915_BO_CACHE_COHERENT_FOR_READ |

>>> @@ -148,6 +167,37 @@ void i915_gem_object_set_cache_coherency(struct

>>> drm_i915_gem_object *obj,

>>>           !IS_DGFX(i915);

>>>   }

>>> +/**

>>> + * i915_gem_object_set_pat_index - set PAT index to be used in PTE

>>> encode

>>> + * @obj: #drm_i915_gem_object

>>> + * @pat_index: PAT index

>>> + *

>>> + * This is a clone of i915_gem_object_set_cache_coherency taking pat

>>> index

>>> + * instead of cache_level as its second argument.

>>> + */

>>> +void i915_gem_object_set_pat_index(struct drm_i915_gem_object *obj,

>>> +                   unsigned int pat_index)

>>> +{

>>> +    struct drm_i915_private *i915 = to_i915(obj->base.dev);

>>> +

>>> +    if (obj->pat_index == pat_index)

>>> +        return;

>>> +

>>> +    obj->pat_index = pat_index;

>>> +

>>> +    if (pat_index != i915_gem_get_pat_index(i915, I915_CACHE_NONE))

>>> +        obj->cache_coherent = (I915_BO_CACHE_COHERENT_FOR_READ |

>>> +                       I915_BO_CACHE_COHERENT_FOR_WRITE);

>>> +    else if (HAS_LLC(i915))

>>> +        obj->cache_coherent = I915_BO_CACHE_COHERENT_FOR_READ;

>>> +    else

>>> +        obj->cache_coherent = 0;

>>> +

>>> +    obj->cache_dirty =

>>> +        !(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_WRITE) &&

>>> +        !IS_DGFX(i915);

>>> +}

>>> +

>>>   bool i915_gem_object_can_bypass_llc(struct drm_i915_gem_object *obj)

>>>   {

>>>       struct drm_i915_private *i915 = to_i915(obj->base.dev);

>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.h

>>> b/drivers/gpu/drm/i915/gem/i915_gem_object.h

>>> index 4c92e17b4337..6f00aab10015 100644

>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_object.h

>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.h

>>> @@ -34,6 +34,8 @@ static inline bool i915_gem_object_size_2big(u64 size)

>>>   unsigned int i915_gem_get_pat_index(struct drm_i915_private *i915,

>>>                       enum i915_cache_level level);

>>> +bool i915_gem_object_has_cache_level(const struct drm_i915_gem_object

>>> *obj,

>>> +                     enum i915_cache_level lvl);

>>>   void i915_gem_init__objects(struct drm_i915_private *i915);

>>>   void i915_objects_module_exit(void);

>>> @@ -764,6 +766,8 @@ bool i915_gem_object_has_unknown_state(struct

>>> drm_i915_gem_object *obj);

>>>   void i915_gem_object_set_cache_coherency(struct drm_i915_gem_object

>>> *obj,

>>>                        unsigned int cache_level);

>>> +void i915_gem_object_set_pat_index(struct drm_i915_gem_object *obj,

>>> +                   unsigned int pat_index);

>>>   bool i915_gem_object_can_bypass_llc(struct drm_i915_gem_object *obj);

>>>   void i915_gem_object_flush_if_display(struct drm_i915_gem_object *obj);

>>>   void i915_gem_object_flush_if_display_locked(struct

>>> drm_i915_gem_object *obj);

>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h

>>> b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h

>>> index 41b35abccf88..132ce01dee9f 100644

>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h

>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h

>>> @@ -195,6 +195,7 @@ enum i915_cache_level {

>>>        */

>>>       I915_CACHE_WT,

>>>       I915_MAX_CACHE_LEVEL,

>>> +    I915_CACHE_INVAL = I915_MAX_CACHE_LEVEL,

>>>   };

>>>   enum i915_map_type {

>>> @@ -358,10 +359,28 @@ struct drm_i915_gem_object {

>>>   #define I915_BO_FLAG_STRUCT_PAGE BIT(0) /* Object backed by struct

>>> pages */

>>>   #define I915_BO_FLAG_IOMEM       BIT(1) /* Object backed by IO

>>> memory */

>>>       /**

>>> -     * @cache_level: The desired GTT caching level.

>>> +     * @pat_index: The desired PAT index.

>>> +     *

>>> +     * See hardware specification for valid PAT indices for each

>>> platform.

>

> Side note for the last patch in the series - the UAPI blurb next to u32

> index needs to at least point to some public PRM

That would be the Bspec.

As far as the design goes, the UMD's should be aware of the PAT index and

its platform dependent nature. The KMD is doing a sanity check so that the

PAT index won't go pass the boundary (e.g. 4 for MTL, 7 for PVC, 3 for TGL/ADL).

I believe the UMD's must be maintaining a list for each platform...

>  which lists the PATs

> and their configuration I would think. Otherwise it's not fully

> transparent how to use the feature.

>

>>> +     * This field used to contain a value of enum i915_cache_level.

>

> What does this mean? Nothing is changed to unsigned here but just new field added.

This comment needs update. Will do in the next version.

>It's

>>> +     * changed to an unsigned int because PAT indices are being used by

>>> +     * both UMD and KMD for caching policy control after GEN12.

>>> +     * For backward compatibility, this field will continue to contain

>>> +     * value of i915_cache_level for pre-GEN12 platforms so that the PTE

>

> Pat_index:6 is a copy of cache_level:3 pre-Gen12?

There wasn't PAT index defined in pre-GEN12 platforms, so this is just to say

that pat_index would behave the same as the cache_level for these platforms.

> But when I look at changes like:

>

>@@ -1302,7 +1298,9 @@ bool i915_ggtt_resume_vm(struct i915_address_space *vm)

>                  */

>                 vma->resource->bound_flags = 0;

>                 vma->ops->bind_vma(vm, NULL, vma->resource,

>-                                  obj ? obj->cache_level : 0,

>+                                  obj ? obj->pat_index :

>+                                        i915_gem_get_pat_index(vm->i915,

>+                                                               I915_CACHE_NONE),

>                                    was_bound);

>

> That suggests it is not a copy but that obj->pat_index is always

> valid and directly a PAT index.

>

> In which case new cache_level enum to say "use pat instead" may

> indeed be nicer as Andrzej suggested.

I'm not sure I understand the concern here. I915_CACHE_NONE is 0, this is

just trying to select between obj->pat_index and pat_index for UC, depending

on whether the obj is valid.

> Although it is not clear to me for a glance that we need both. Maybe all

> in driver object creation can use cache_level but immediately convert to

> PAT internally and just don't store cache_level? I haven't looked in detail

> is my disclaimer though.. I guess it may boil down to does i915 ever need

> to read back cache_level, other than on the top entry points like setting

> it or so.

>

>>> +     * encode functions for these legacy platforms can stay the same.

>>> +     * In the meantime platform specific tables are created to translate

>>> +     * i915_cache_level into pat index, for more details check the

>>> macros

>>> +     * defined i915/i915_pci.c, e.g. PVC_CACHELEVEL.

>>> +     */

>>> +    unsigned int pat_index:6;

>

> Existing bitfield takes up 7 bits. I'd check here with pahole if making

> pat_index a full u8 and changing the existing ones to u8 field:bits maybe

> ends up better overall.

>

>>> +    /**

>>> +     * @cache_level: Indicate whether pat_index is set by UMD

>>>        *

>>> -     * See enum i915_cache_level for possible values, along with what

>>> -     * each does.

>>> +     * This used to hold desired GTT caching level, but is now

>>> replaced by

>>> +     * pat_index. It's kept here for KMD to tell whether the

>>> pat_index is

>>> +     * set by UMD or converted from enum i915_cache_level.

>>> +     * This field should be 0 by default, but I915_CACHE_INVAL if the

>>> +     * pat_index is set by UMD.

>>>        */

>>>       unsigned int cache_level:3;

>>>       /**

>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_stolen.c

>>> b/drivers/gpu/drm/i915/gem/i915_gem_stolen.c

>>> index ee492d823f1b..3b094d36a0b0 100644

>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_stolen.c

>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_stolen.c

>>> @@ -565,7 +565,9 @@ static void dbg_poison(struct i915_ggtt *ggtt,

>>>           ggtt->vm.insert_page(&ggtt->vm, addr,

>>>                        ggtt->error_capture.start,

>>> -                     I915_CACHE_NONE, 0);

>>> +                     i915_gem_get_pat_index(ggtt->vm.i915,

>>> +                                I915_CACHE_NONE),

>>> +                     0);

>>>           mb();

>>>           s = io_mapping_map_wc(&ggtt->iomap,

>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c

>>> b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c

>>> index 69eb20ed4d47..e40761e13c2a 100644

>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c

>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c

>>> @@ -214,7 +214,8 @@ static struct dma_fence

>>> *i915_ttm_accel_move(struct ttm_buffer_object *bo,

>>>           intel_engine_pm_get(to_gt(i915)->migrate.context->engine);

>>>           ret =

>>> intel_context_migrate_clear(to_gt(i915)->migrate.context, deps,

>>> -                          dst_st->sgl, dst_level,

>>> +                          dst_st->sgl,

>>> +                          i915_gem_get_pat_index(i915, dst_level),

>>>                             i915_ttm_gtt_binds_lmem(dst_mem),

>>>                             0, &rq);

>>>       } else {

>>> @@ -227,12 +228,13 @@ static struct dma_fence

>>> *i915_ttm_accel_move(struct ttm_buffer_object *bo,

>>>           src_level = i915_ttm_cache_level(i915, bo->resource, src_ttm);

>>>           intel_engine_pm_get(to_gt(i915)->migrate.context->engine);

>>>           ret = intel_context_migrate_copy(to_gt(i915)->migrate.context,

>>> -                         deps, src_rsgt->table.sgl,

>>> -                         src_level,

>>> -                         i915_ttm_gtt_binds_lmem(bo->resource),

>>> -                         dst_st->sgl, dst_level,

>>> -                         i915_ttm_gtt_binds_lmem(dst_mem),

>>> -                         &rq);

>>> +                    deps, src_rsgt->table.sgl,

>>> +                    i915_gem_get_pat_index(i915, src_level),

>>> +                    i915_ttm_gtt_binds_lmem(bo->resource),

>>> +                    dst_st->sgl,

>>> +                    i915_gem_get_pat_index(i915, dst_level),

>>> +                    i915_ttm_gtt_binds_lmem(dst_mem),

>>> +                    &rq);

>>>           i915_refct_sgt_put(src_rsgt);

>>>       }

>>> diff --git a/drivers/gpu/drm/i915/gem/selftests/huge_pages.c

>>> b/drivers/gpu/drm/i915/gem/selftests/huge_pages.c

>>> index defece0bcb81..ebb68ac9cd5e 100644

>>> --- a/drivers/gpu/drm/i915/gem/selftests/huge_pages.c

>>> +++ b/drivers/gpu/drm/i915/gem/selftests/huge_pages.c

>>> @@ -354,7 +354,7 @@ fake_huge_pages_object(struct drm_i915_private

>>> *i915, u64 size, bool single)

>>>       obj->write_domain = I915_GEM_DOMAIN_CPU;

>>>       obj->read_domains = I915_GEM_DOMAIN_CPU;

>>> -    obj->cache_level = I915_CACHE_NONE;

>>> +    obj->pat_index = i915_gem_get_pat_index(i915, I915_CACHE_NONE);

>>>       return obj;

>>>   }

>>> diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_migrate.c

>>> b/drivers/gpu/drm/i915/gem/selftests/i915_gem_migrate.c

>>> index fe6c37fd7859..a93a90b15907 100644

>>> --- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_migrate.c

>>> +++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_migrate.c

>>> @@ -219,7 +219,7 @@ static int __igt_lmem_pages_migrate(struct

>>> intel_gt *gt,

>>>               continue;

>>>           err = intel_migrate_clear(&gt->migrate, &ww, deps,

>>> -                      obj->mm.pages->sgl, obj->cache_level,

>>> +                      obj->mm.pages->sgl, obj->pat_index,

>>>                         i915_gem_object_is_lmem(obj),

>>>                         0xdeadbeaf, &rq);

>>>           if (rq) {

>>> diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c

>>> b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c

>>> index 56279908ed30..a93d8f9f8bc1 100644

>>> --- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c

>>> +++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c

>>> @@ -1222,7 +1222,7 @@ static int __igt_mmap_migrate(struct

>>> intel_memory_region **placements,

>>>       }

>>>       err = intel_context_migrate_clear(to_gt(i915)->migrate.context,

>>> NULL,

>>> -                      obj->mm.pages->sgl, obj->cache_level,

>>> +                      obj->mm.pages->sgl, obj->pat_index,

>>>                         i915_gem_object_is_lmem(obj),

>>>                         expand32(POISON_INUSE), &rq);

>>>       i915_gem_object_unpin_pages(obj);

>>> diff --git a/drivers/gpu/drm/i915/gt/gen6_ppgtt.c

>>> b/drivers/gpu/drm/i915/gt/gen6_ppgtt.c

>>> index 5aaacc53fa4c..c2bdc133c89a 100644

>>> --- a/drivers/gpu/drm/i915/gt/gen6_ppgtt.c

>>> +++ b/drivers/gpu/drm/i915/gt/gen6_ppgtt.c

>>> @@ -109,7 +109,7 @@ static void gen6_ppgtt_clear_range(struct

>>> i915_address_space *vm,

>>>   static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,

>>>                         struct i915_vma_resource *vma_res,

>>> -                      enum i915_cache_level cache_level,

>>> +                      unsigned int pat_index,

>>>                         u32 flags)

>>>   {

>>>       struct i915_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);

>>> @@ -117,7 +117,7 @@ static void gen6_ppgtt_insert_entries(struct

>>> i915_address_space *vm,

>>>       unsigned int first_entry = vma_res->start / I915_GTT_PAGE_SIZE;

>>>       unsigned int act_pt = first_entry / GEN6_PTES;

>>>       unsigned int act_pte = first_entry % GEN6_PTES;

>>> -    const u32 pte_encode = vm->pte_encode(0, cache_level, flags);

>>> +    const u32 pte_encode = vm->pte_encode(0, pat_index, flags);

>>>       struct sgt_dma iter = sgt_dma(vma_res);

>>>       gen6_pte_t *vaddr;

>>> @@ -227,7 +227,9 @@ static int gen6_ppgtt_init_scratch(struct

>>> gen6_ppgtt *ppgtt)

>>>       vm->scratch[0]->encode =

>>>           vm->pte_encode(px_dma(vm->scratch[0]),

>>> -                   I915_CACHE_NONE, PTE_READ_ONLY);

>>> +                   i915_gem_get_pat_index(vm->i915,

>>> +                              I915_CACHE_NONE),

>>> +                   PTE_READ_ONLY);

>>>       vm->scratch[1] = vm->alloc_pt_dma(vm, I915_GTT_PAGE_SIZE_4K);

>>>       if (IS_ERR(vm->scratch[1])) {

>>> @@ -278,7 +280,7 @@ static void gen6_ppgtt_cleanup(struct

>>> i915_address_space *vm)

>>>   static void pd_vma_bind(struct i915_address_space *vm,

>>>               struct i915_vm_pt_stash *stash,

>>>               struct i915_vma_resource *vma_res,

>>> -            enum i915_cache_level cache_level,

>>> +            unsigned int pat_index,

>>>               u32 unused)

>>>   {

>>>       struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);

>>> diff --git a/drivers/gpu/drm/i915/gt/gen8_ppgtt.c

>>> b/drivers/gpu/drm/i915/gt/gen8_ppgtt.c

>>> index 7a4b1d1afce9..c046813514f4 100644

>>> --- a/drivers/gpu/drm/i915/gt/gen8_ppgtt.c

>>> +++ b/drivers/gpu/drm/i915/gt/gen8_ppgtt.c

>>> @@ -56,7 +56,7 @@ static u64 gen8_pte_encode(dma_addr_t addr,

>>>   }

>>>   static u64 mtl_pte_encode(dma_addr_t addr,

>>> -              enum i915_cache_level level,

>>> +              unsigned int pat_index,

>>>                 u32 flags)

>>>   {

>>>       gen8_pte_t pte = addr | GEN8_PAGE_PRESENT | GEN8_PAGE_RW;

>>> @@ -67,24 +67,17 @@ static u64 mtl_pte_encode(dma_addr_t addr,

>>>       if (flags & PTE_LM)

>>>           pte |= GEN12_PPGTT_PTE_LM | GEN12_PPGTT_PTE_NC;

>>> -    switch (level) {

>>> -    case I915_CACHE_NONE:

>>> -        pte |= GEN12_PPGTT_PTE_PAT1;

>>> -        break;

>>> -    case I915_CACHE_LLC:

>>> -    case I915_CACHE_L3_LLC:

>>> -        pte |= GEN12_PPGTT_PTE_PAT0 | GEN12_PPGTT_PTE_PAT1;

>>> -        break;

>>> -    case I915_CACHE_WT:

>>> +    if (pat_index & BIT(0))

>>>           pte |= GEN12_PPGTT_PTE_PAT0;

>>> -        break;

>>> -    default:

>>> -        /* This should never happen. Added to deal with the compile

>>> -         * error due to the addition of I915_MAX_CACHE_LEVEL. Will

>>> -         * be removed by the pat_index patch.

>>> -         */

>>> -        break;

>>> -    }

>>> +

>>> +    if (pat_index & BIT(1))

>>> +        pte |= GEN12_PPGTT_PTE_PAT1;

>>> +

>>> +    if (pat_index & BIT(2))

>>> +        pte |= GEN12_PPGTT_PTE_PAT2;

>>> +

>>> +    if (pat_index & BIT(3))

>>> +        pte |= MTL_PPGTT_PTE_PAT3;

>>>       return pte;

>>>   }

>>> @@ -457,11 +450,11 @@ gen8_ppgtt_insert_pte(struct i915_ppgtt *ppgtt,

>>>                 struct i915_page_directory *pdp,

>>>                 struct sgt_dma *iter,

>>>                 u64 idx,

>>> -              enum i915_cache_level cache_level,

>>> +              unsigned int pat_index,

>>>                 u32 flags)

>>>   {

>>>       struct i915_page_directory *pd;

>>> -    const gen8_pte_t pte_encode = ppgtt->vm.pte_encode(0,

>>> cache_level, flags);

>>> +    const gen8_pte_t pte_encode = ppgtt->vm.pte_encode(0, pat_index,

>>> flags);

>>>       gen8_pte_t *vaddr;

>>>       pd = i915_pd_entry(pdp, gen8_pd_index(idx, 2));

>>> @@ -504,10 +497,10 @@ static void

>>>   xehpsdv_ppgtt_insert_huge(struct i915_address_space *vm,

>>>                 struct i915_vma_resource *vma_res,

>>>                 struct sgt_dma *iter,

>>> -              enum i915_cache_level cache_level,

>>> +              unsigned int pat_index,

>>>                 u32 flags)

>>>   {

>>> -    const gen8_pte_t pte_encode = vm->pte_encode(0, cache_level, flags);

>>> +    const gen8_pte_t pte_encode = vm->pte_encode(0, pat_index, flags);

>>>       unsigned int rem = sg_dma_len(iter->sg);

>>>       u64 start = vma_res->start;

>>>       u64 end = start + vma_res->vma_size;

>>> @@ -611,10 +604,10 @@ xehpsdv_ppgtt_insert_huge(struct

>>> i915_address_space *vm,

>>>   static void gen8_ppgtt_insert_huge(struct i915_address_space *vm,

>>>                      struct i915_vma_resource *vma_res,

>>>                      struct sgt_dma *iter,

>>> -                   enum i915_cache_level cache_level,

>>> +                   unsigned int pat_index,

>>>                      u32 flags)

>>>   {

>>> -    const gen8_pte_t pte_encode = vm->pte_encode(0, cache_level, flags);

>>> +    const gen8_pte_t pte_encode = vm->pte_encode(0, pat_index, flags);

>>>       unsigned int rem = sg_dma_len(iter->sg);

>>>       u64 start = vma_res->start;

>>> @@ -734,7 +727,7 @@ static void gen8_ppgtt_insert_huge(struct

>>> i915_address_space *vm,

>>>   static void gen8_ppgtt_insert(struct i915_address_space *vm,

>>>                     struct i915_vma_resource *vma_res,

>>> -                  enum i915_cache_level cache_level,

>>> +                  unsigned int pat_index,

>>>                     u32 flags)

>>>   {

>>>       struct i915_ppgtt * const ppgtt = i915_vm_to_ppgtt(vm);

>>> @@ -742,9 +735,9 @@ static void gen8_ppgtt_insert(struct

>>> i915_address_space *vm,

>>>       if (vma_res->bi.page_sizes.sg > I915_GTT_PAGE_SIZE) {

>>>           if (HAS_64K_PAGES(vm->i915))

>>> -            xehpsdv_ppgtt_insert_huge(vm, vma_res, &iter,

>>> cache_level, flags);

>>> +            xehpsdv_ppgtt_insert_huge(vm, vma_res, &iter, pat_index,

>>> flags);

>>>           else

>>> -            gen8_ppgtt_insert_huge(vm, vma_res, &iter, cache_level,

>>> flags);

>>> +            gen8_ppgtt_insert_huge(vm, vma_res, &iter, pat_index,

>>> flags);

>>>       } else  {

>>>           u64 idx = vma_res->start >> GEN8_PTE_SHIFT;

>>> @@ -753,7 +746,7 @@ static void gen8_ppgtt_insert(struct

>>> i915_address_space *vm,

>>>                   gen8_pdp_for_page_index(vm, idx);

>>>               idx = gen8_ppgtt_insert_pte(ppgtt, pdp, &iter, idx,

>>> -                            cache_level, flags);

>>> +                            pat_index, flags);

>>>           } while (idx);

>>>           vma_res->page_sizes_gtt = I915_GTT_PAGE_SIZE;

>>> @@ -763,7 +756,7 @@ static void gen8_ppgtt_insert(struct

>>> i915_address_space *vm,

>>>   static void gen8_ppgtt_insert_entry(struct i915_address_space *vm,

>>>                       dma_addr_t addr,

>>>                       u64 offset,

>>> -                    enum i915_cache_level level,

>>> +                    unsigned int pat_index,

>>>                       u32 flags)

>>>   {

>>>       u64 idx = offset >> GEN8_PTE_SHIFT;

>>> @@ -777,14 +770,14 @@ static void gen8_ppgtt_insert_entry(struct

>>> i915_address_space *vm,

>>>       GEM_BUG_ON(pt->is_compact);

>>>       vaddr = px_vaddr(pt);

>>> -    vaddr[gen8_pd_index(idx, 0)] = vm->pte_encode(addr, level, flags);

>>> +    vaddr[gen8_pd_index(idx, 0)] = vm->pte_encode(addr, pat_index,

>>> flags);

>>>       drm_clflush_virt_range(&vaddr[gen8_pd_index(idx, 0)],

>>> sizeof(*vaddr));

>>>   }

>>>   static void __xehpsdv_ppgtt_insert_entry_lm(struct

>>> i915_address_space *vm,

>>>                           dma_addr_t addr,

>>>                           u64 offset,

>>> -                        enum i915_cache_level level,

>>> +                        unsigned int pat_index,

>>>                           u32 flags)

>>>   {

>>>       u64 idx = offset >> GEN8_PTE_SHIFT;

>>> @@ -807,20 +800,20 @@ static void

>>> __xehpsdv_ppgtt_insert_entry_lm(struct i915_address_space *vm,

>>>       }

>>>       vaddr = px_vaddr(pt);

>>> -    vaddr[gen8_pd_index(idx, 0) / 16] = vm->pte_encode(addr, level,

>>> flags);

>>> +    vaddr[gen8_pd_index(idx, 0) / 16] = vm->pte_encode(addr,

>>> pat_index, flags);

>>>   }

>>>   static void xehpsdv_ppgtt_insert_entry(struct i915_address_space *vm,

>>>                          dma_addr_t addr,

>>>                          u64 offset,

>>> -                       enum i915_cache_level level,

>>> +                       unsigned int pat_index,

>>>                          u32 flags)

>>>   {

>>>       if (flags & PTE_LM)

>>>           return __xehpsdv_ppgtt_insert_entry_lm(vm, addr, offset,

>>> -                               level, flags);

>>> +                               pat_index, flags);

>>> -    return gen8_ppgtt_insert_entry(vm, addr, offset, level, flags);

>>> +    return gen8_ppgtt_insert_entry(vm, addr, offset, pat_index, flags);

>>>   }

>>>   static int gen8_init_scratch(struct i915_address_space *vm)

>>> @@ -855,7 +848,9 @@ static int gen8_init_scratch(struct

>>> i915_address_space *vm)

>>>       vm->scratch[0]->encode =

>>>           vm->pte_encode(px_dma(vm->scratch[0]),

>>> -                   I915_CACHE_NONE, pte_flags);

>>> +                   i915_gem_get_pat_index(vm->i915,

>>> +                              I915_CACHE_NONE),

>>> +                   pte_flags);

>>>       for (i = 1; i <= vm->top; i++) {

>>>           struct drm_i915_gem_object *obj;

>>> @@ -873,7 +868,9 @@ static int gen8_init_scratch(struct

>>> i915_address_space *vm)

>>>           }

>>>           fill_px(obj, vm->scratch[i - 1]->encode);

>>> -        obj->encode = gen8_pde_encode(px_dma(obj), I915_CACHE_NONE);

>>> +        obj->encode = gen8_pde_encode(px_dma(obj),

>>> +                          i915_gem_get_pat_index(vm->i915,

>>> +                                     I915_CACHE_NONE));

>>>           vm->scratch[i] = obj;

>>>       }

>>> diff --git a/drivers/gpu/drm/i915/gt/gen8_ppgtt.h

>>> b/drivers/gpu/drm/i915/gt/gen8_ppgtt.h

>>> index f541d19264b4..19c635441642 100644

>>> --- a/drivers/gpu/drm/i915/gt/gen8_ppgtt.h

>>> +++ b/drivers/gpu/drm/i915/gt/gen8_ppgtt.h

>>> @@ -10,13 +10,12 @@

>>>   struct i915_address_space;

>>>   struct intel_gt;

>>> -enum i915_cache_level;

>>>   struct i915_ppgtt *gen8_ppgtt_create(struct intel_gt *gt,

>>>                        unsigned long lmem_pt_obj_flags);

>>>   u64 gen8_ggtt_pte_encode(dma_addr_t addr,

>>> -             enum i915_cache_level level,

>>> +             unsigned int pat_index,

>>>                u32 flags);

>>>   #endif

>>> diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt.c

>>> b/drivers/gpu/drm/i915/gt/intel_ggtt.c

>>> index c8390d03fce2..2a7942fac798 100644

>>> --- a/drivers/gpu/drm/i915/gt/intel_ggtt.c

>>> +++ b/drivers/gpu/drm/i915/gt/intel_ggtt.c

>>> @@ -221,7 +221,7 @@ static void guc_ggtt_invalidate(struct i915_ggtt

>>> *ggtt)

>>>   }

>>>   static u64 mtl_ggtt_pte_encode(dma_addr_t addr,

>>> -                   enum i915_cache_level level,

>>> +                   unsigned int pat_index,

>>>                      u32 flags)

>>>   {

>>>       gen8_pte_t pte = addr | GEN8_PAGE_PRESENT;

>>> @@ -231,30 +231,17 @@ static u64 mtl_ggtt_pte_encode(dma_addr_t addr,

>>>       if (flags & PTE_LM)

>>>           pte |= GEN12_GGTT_PTE_LM;

>>> -    switch (level) {

>>> -    case I915_CACHE_NONE:

>>> -        pte |= MTL_GGTT_PTE_PAT1;

>>> -        break;

>>> -    case I915_CACHE_LLC:

>>> -    case I915_CACHE_L3_LLC:

>>> -        pte |= MTL_GGTT_PTE_PAT0 | MTL_GGTT_PTE_PAT1;

>>> -        break;

>>> -    case I915_CACHE_WT:

>>> +    if (pat_index & BIT(0))

>>>           pte |= MTL_GGTT_PTE_PAT0;

>>> -        break;

>>> -    default:

>>> -        /* This should never happen. Added to deal with the compile

>>> -         * error due to the addition of I915_MAX_CACHE_LEVEL. Will

>>> -         * be removed by the pat_index patch.

>>> -         */

>>> -        break;

>>> -    }

>>> +

>>> +    if (pat_index & BIT(1))

>>> +        pte |= MTL_GGTT_PTE_PAT1;

>>>       return pte;

>>>   }

>>>   u64 gen8_ggtt_pte_encode(dma_addr_t addr,

>>> -             enum i915_cache_level level,

>>> +             unsigned int pat_index,

>>>                u32 flags)

>>>   {

>>>       gen8_pte_t pte = addr | GEN8_PAGE_PRESENT;

>>> @@ -273,25 +260,25 @@ static void gen8_set_pte(void __iomem *addr,

>>> gen8_pte_t pte)

>>>   static void gen8_ggtt_insert_page(struct i915_address_space *vm,

>>>                     dma_addr_t addr,

>>>                     u64 offset,

>>> -                  enum i915_cache_level level,

>>> +                  unsigned int pat_index,

>>>                     u32 flags)

>>>   {

>>>       struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);

>>>       gen8_pte_t __iomem *pte =

>>>           (gen8_pte_t __iomem *)ggtt->gsm + offset / I915_GTT_PAGE_SIZE;

>>> -    gen8_set_pte(pte, ggtt->vm.pte_encode(addr, level, flags));

>>> +    gen8_set_pte(pte, ggtt->vm.pte_encode(addr, pat_index, flags));

>>>       ggtt->invalidate(ggtt);

>>>   }

>>>   static void gen8_ggtt_insert_entries(struct i915_address_space *vm,

>>>                        struct i915_vma_resource *vma_res,

>>> -                     enum i915_cache_level level,

>>> +                     unsigned int pat_index,

>>>                        u32 flags)

>>>   {

>>>       struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);

>>> -    const gen8_pte_t pte_encode = ggtt->vm.pte_encode(0, level, flags);

>>> +    const gen8_pte_t pte_encode = ggtt->vm.pte_encode(0, pat_index,

>>> flags);

>>>       gen8_pte_t __iomem *gte;

>>>       gen8_pte_t __iomem *end;

>>>       struct sgt_iter iter;

>>> @@ -348,14 +335,14 @@ static void gen8_ggtt_clear_range(struct

>>> i915_address_space *vm,

>>>   static void gen6_ggtt_insert_page(struct i915_address_space *vm,

>>>                     dma_addr_t addr,

>>>                     u64 offset,

>>> -                  enum i915_cache_level level,

>>> +                  unsigned int pat_index,

>>>                     u32 flags)

>>>   {

>>>       struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);

>>>       gen6_pte_t __iomem *pte =

>>>           (gen6_pte_t __iomem *)ggtt->gsm + offset / I915_GTT_PAGE_SIZE;

>>> -    iowrite32(vm->pte_encode(addr, level, flags), pte);

>>> +    iowrite32(vm->pte_encode(addr, pat_index, flags), pte);

>>>       ggtt->invalidate(ggtt);

>>>   }

>>> @@ -368,7 +355,7 @@ static void gen6_ggtt_insert_page(struct

>>> i915_address_space *vm,

>>>    */

>>>   static void gen6_ggtt_insert_entries(struct i915_address_space *vm,

>>>                        struct i915_vma_resource *vma_res,

>>> -                     enum i915_cache_level level,

>>> +                     unsigned int pat_index,

>>>                        u32 flags)

>>>   {

>>>       struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);

>>> @@ -385,7 +372,7 @@ static void gen6_ggtt_insert_entries(struct

>>> i915_address_space *vm,

>>>           iowrite32(vm->scratch[0]->encode, gte++);

>>>       end += (vma_res->node_size + vma_res->guard) / I915_GTT_PAGE_SIZE;

>>>       for_each_sgt_daddr(addr, iter, vma_res->bi.pages)

>>> -        iowrite32(vm->pte_encode(addr, level, flags), gte++);

>>> +        iowrite32(vm->pte_encode(addr, pat_index, flags), gte++);

>>>       GEM_BUG_ON(gte > end);

>>>       /* Fill the allocated but "unused" space beyond the end of the

>>> buffer */

>>> @@ -420,14 +407,15 @@ struct insert_page {

>>>       struct i915_address_space *vm;

>>>       dma_addr_t addr;

>>>       u64 offset;

>>> -    enum i915_cache_level level;

>>> +    unsigned int pat_index;

>>>   };

>>>   static int bxt_vtd_ggtt_insert_page__cb(void *_arg)

>>>   {

>>>       struct insert_page *arg = _arg;

>>> -    gen8_ggtt_insert_page(arg->vm, arg->addr, arg->offset,

>>> arg->level, 0);

>>> +    gen8_ggtt_insert_page(arg->vm, arg->addr, arg->offset,

>>> +                  arg->pat_index, 0);

>>>       bxt_vtd_ggtt_wa(arg->vm);

>>>       return 0;

>>> @@ -436,10 +424,10 @@ static int bxt_vtd_ggtt_insert_page__cb(void *_arg)

>>>   static void bxt_vtd_ggtt_insert_page__BKL(struct i915_address_space

>>> *vm,

>>>                         dma_addr_t addr,

>>>                         u64 offset,

>>> -                      enum i915_cache_level level,

>>> +                      unsigned int pat_index,

>>>                         u32 unused)

>>>   {

>>> -    struct insert_page arg = { vm, addr, offset, level };

>>> +    struct insert_page arg = { vm, addr, offset, pat_index };

>>>       stop_machine(bxt_vtd_ggtt_insert_page__cb, &arg, NULL);

>>>   }

>>> @@ -447,7 +435,7 @@ static void bxt_vtd_ggtt_insert_page__BKL(struct

>>> i915_address_space *vm,

>>>   struct insert_entries {

>>>       struct i915_address_space *vm;

>>>       struct i915_vma_resource *vma_res;

>>> -    enum i915_cache_level level;

>>> +    unsigned int pat_index;

>>>       u32 flags;

>>>   };

>>> @@ -455,7 +443,8 @@ static int bxt_vtd_ggtt_insert_entries__cb(void

>>> *_arg)

>>>   {

>>>       struct insert_entries *arg = _arg;

>>> -    gen8_ggtt_insert_entries(arg->vm, arg->vma_res, arg->level,

>>> arg->flags);

>>> +    gen8_ggtt_insert_entries(arg->vm, arg->vma_res,

>>> +                 arg->pat_index, arg->flags);

>>>       bxt_vtd_ggtt_wa(arg->vm);

>>>       return 0;

>>> @@ -463,10 +452,10 @@ static int bxt_vtd_ggtt_insert_entries__cb(void

>>> *_arg)

>>>   static void bxt_vtd_ggtt_insert_entries__BKL(struct

>>> i915_address_space *vm,

>>>                            struct i915_vma_resource *vma_res,

>>> -                         enum i915_cache_level level,

>>> +                         unsigned int pat_index,

>>>                            u32 flags)

>>>   {

>>> -    struct insert_entries arg = { vm, vma_res, level, flags };

>>> +    struct insert_entries arg = { vm, vma_res, pat_index, flags };

>>>       stop_machine(bxt_vtd_ggtt_insert_entries__cb, &arg, NULL);

>>>   }

>>> @@ -495,7 +484,7 @@ static void gen6_ggtt_clear_range(struct

>>> i915_address_space *vm,

>>>   void intel_ggtt_bind_vma(struct i915_address_space *vm,

>>>                struct i915_vm_pt_stash *stash,

>>>                struct i915_vma_resource *vma_res,

>>> -             enum i915_cache_level cache_level,

>>> +             unsigned int pat_index,

>>>                u32 flags)

>>>   {

>>>       u32 pte_flags;

>>> @@ -512,7 +501,7 @@ void intel_ggtt_bind_vma(struct i915_address_space

>>> *vm,

>>>       if (vma_res->bi.lmem)

>>>           pte_flags |= PTE_LM;

>>> -    vm->insert_entries(vm, vma_res, cache_level, pte_flags);

>>> +    vm->insert_entries(vm, vma_res, pat_index, pte_flags);

>>>       vma_res->page_sizes_gtt = I915_GTT_PAGE_SIZE;

>>>   }

>>> @@ -661,7 +650,7 @@ static int init_ggtt(struct i915_ggtt *ggtt)

>>>   static void aliasing_gtt_bind_vma(struct i915_address_space *vm,

>>>                     struct i915_vm_pt_stash *stash,

>>>                     struct i915_vma_resource *vma_res,

>>> -                  enum i915_cache_level cache_level,

>>> +                  unsigned int pat_index,

>>>                     u32 flags)

>>>   {

>>>       u32 pte_flags;

>>> @@ -673,10 +662,10 @@ static void aliasing_gtt_bind_vma(struct

>>> i915_address_space *vm,

>>>       if (flags & I915_VMA_LOCAL_BIND)

>>>           ppgtt_bind_vma(&i915_vm_to_ggtt(vm)->alias->vm,

>>> -                   stash, vma_res, cache_level, flags);

>>> +                   stash, vma_res, pat_index, flags);

>>>       if (flags & I915_VMA_GLOBAL_BIND)

>>> -        vm->insert_entries(vm, vma_res, cache_level, pte_flags);

>>> +        vm->insert_entries(vm, vma_res, pat_index, pte_flags);

>>>       vma_res->bound_flags |= flags;

>>>   }

>>> @@ -933,7 +922,9 @@ static int ggtt_probe_common(struct i915_ggtt

>>> *ggtt, u64 size)

>>>       ggtt->vm.scratch[0]->encode =

>>>           ggtt->vm.pte_encode(px_dma(ggtt->vm.scratch[0]),

>>> -                    I915_CACHE_NONE, pte_flags);

>>> +                    i915_gem_get_pat_index(i915,

>>> +                               I915_CACHE_NONE),

>>> +                    pte_flags);

>>>       return 0;

>>>   }

>>> @@ -1022,6 +1013,11 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt)

>>>       return ggtt_probe_common(ggtt, size);

>>>   }

>>> +/*

>>> + * For pre-gen8 platforms pat_index is the same as enum

>>> i915_cache_level,

>>> + * so these PTE encode functions are left with using cache_level.

>>> + * See translation table LEGACY_CACHELEVEL.

>>> + */

>>>   static u64 snb_pte_encode(dma_addr_t addr,

>>>                 enum i915_cache_level level,

>>>                 u32 flags)

>>> @@ -1302,7 +1298,9 @@ bool i915_ggtt_resume_vm(struct

>>> i915_address_space *vm)

>>>            */

>>>           vma->resource->bound_flags = 0;

>>>           vma->ops->bind_vma(vm, NULL, vma->resource,

>>> -                   obj ? obj->cache_level : 0,

>>> +                   obj ? obj->pat_index :

>>> +                     i915_gem_get_pat_index(vm->i915,

>>> +                                I915_CACHE_NONE),

>>>                      was_bound);

>>>           if (obj) { /* only used during resume => exclusive access */

>>> diff --git a/drivers/gpu/drm/i915/gt/intel_gtt.h

>>> b/drivers/gpu/drm/i915/gt/intel_gtt.h

>>> index 854ec09fd588..be767e13b1e5 100644

>>> --- a/drivers/gpu/drm/i915/gt/intel_gtt.h

>>> +++ b/drivers/gpu/drm/i915/gt/intel_gtt.h

>>> @@ -165,8 +165,6 @@ typedef u64 gen8_pte_t;

>>>   #define MTL_2_COH_1W    REG_FIELD_PREP(MTL_PAT_INDEX_COH_MODE_MASK, 2)

>>>   #define MTL_0_COH_NON    REG_FIELD_PREP(MTL_PAT_INDEX_COH_MODE_MASK, 0)

>>> -enum i915_cache_level;

>>> -

>>>   struct drm_i915_gem_object;

>>>   struct i915_fence_reg;

>>>   struct i915_vma;

>>> @@ -234,7 +232,7 @@ struct i915_vma_ops {

>>>       void (*bind_vma)(struct i915_address_space *vm,

>>>                struct i915_vm_pt_stash *stash,

>>>                struct i915_vma_resource *vma_res,

>>> -             enum i915_cache_level cache_level,

>>> +             unsigned int pat_index,

>>>                u32 flags);

>>>       /*

>>>        * Unmap an object from an address space. This usually consists of

>>> @@ -306,7 +304,7 @@ struct i915_address_space {

>>>           (*alloc_scratch_dma)(struct i915_address_space *vm, int sz);

>>>       u64 (*pte_encode)(dma_addr_t addr,

>>> -              enum i915_cache_level level,

>>> +              unsigned int pat_index,

>>>                 u32 flags); /* Create a valid PTE */

>>>   #define PTE_READ_ONLY    BIT(0)

>>>   #define PTE_LM        BIT(1)

>>> @@ -321,20 +319,20 @@ struct i915_address_space {

>>>       void (*insert_page)(struct i915_address_space *vm,

>>>                   dma_addr_t addr,

>>>                   u64 offset,

>>> -                enum i915_cache_level cache_level,

>>> +                unsigned int pat_index,

>>>                   u32 flags);

>>>       void (*insert_entries)(struct i915_address_space *vm,

>>>                      struct i915_vma_resource *vma_res,

>>> -                   enum i915_cache_level cache_level,

>>> +                   unsigned int pat_index,

>>>                      u32 flags);

>>>       void (*raw_insert_page)(struct i915_address_space *vm,

>>>                   dma_addr_t addr,

>>>                   u64 offset,

>>> -                enum i915_cache_level cache_level,

>>> +                unsigned int pat_index,

>>>                   u32 flags);

>>>       void (*raw_insert_entries)(struct i915_address_space *vm,

>>>                      struct i915_vma_resource *vma_res,

>>> -                   enum i915_cache_level cache_level,

>>> +                   unsigned int pat_index,

>>>                      u32 flags);

>>>       void (*cleanup)(struct i915_address_space *vm);

>>> @@ -581,7 +579,7 @@ void ppgtt_init(struct i915_ppgtt *ppgtt, struct

>>> intel_gt *gt,

>>>   void intel_ggtt_bind_vma(struct i915_address_space *vm,

>>>                struct i915_vm_pt_stash *stash,

>>>                struct i915_vma_resource *vma_res,

>>> -             enum i915_cache_level cache_level,

>>> +             unsigned int pat_index,

>>>                u32 flags);

>>>   void intel_ggtt_unbind_vma(struct i915_address_space *vm,

>>>                  struct i915_vma_resource *vma_res);

>>> @@ -639,7 +637,7 @@ void

>>>   __set_pd_entry(struct i915_page_directory * const pd,

>>>              const unsigned short idx,

>>>              struct i915_page_table *pt,

>>> -           u64 (*encode)(const dma_addr_t, const enum

>>> i915_cache_level));

>>> +           u64 (*encode)(const dma_addr_t, const unsigned int

>>> pat_index));

>>>   #define set_pd_entry(pd, idx, to) \

>>>       __set_pd_entry((pd), (idx), px_pt(to), gen8_pde_encode)

>>> @@ -659,7 +657,7 @@ void gen6_ggtt_invalidate(struct i915_ggtt *ggtt);

>>>   void ppgtt_bind_vma(struct i915_address_space *vm,

>>>               struct i915_vm_pt_stash *stash,

>>>               struct i915_vma_resource *vma_res,

>>> -            enum i915_cache_level cache_level,

>>> +            unsigned int pat_index,

>>>               u32 flags);

>>>   void ppgtt_unbind_vma(struct i915_address_space *vm,

>>>                 struct i915_vma_resource *vma_res);

>>> diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.c

>>> b/drivers/gpu/drm/i915/gt/intel_migrate.c

>>> index 3f638f198796..117c3d05af3e 100644

>>> --- a/drivers/gpu/drm/i915/gt/intel_migrate.c

>>> +++ b/drivers/gpu/drm/i915/gt/intel_migrate.c

>>> @@ -45,7 +45,9 @@ static void xehpsdv_toggle_pdes(struct

>>> i915_address_space *vm,

>>>        * Insert a dummy PTE into every PT that will map to LMEM to ensure

>>>        * we have a correctly setup PDE structure for later use.

>>>        */

>>> -    vm->insert_page(vm, 0, d->offset, I915_CACHE_NONE, PTE_LM);

>>> +    vm->insert_page(vm, 0, d->offset,

>>> +            i915_gem_get_pat_index(vm->i915, I915_CACHE_NONE),

>>> +            PTE_LM);

>>>       GEM_BUG_ON(!pt->is_compact);

>>>       d->offset += SZ_2M;

>>>   }

>>> @@ -63,7 +65,9 @@ static void xehpsdv_insert_pte(struct

>>> i915_address_space *vm,

>>>        * alignment is 64K underneath for the pt, and we are careful

>>>        * not to access the space in the void.

>>>        */

>>> -    vm->insert_page(vm, px_dma(pt), d->offset, I915_CACHE_NONE, PTE_LM);

>>> +    vm->insert_page(vm, px_dma(pt), d->offset,

>>> +            i915_gem_get_pat_index(vm->i915, I915_CACHE_NONE),

>>> +            PTE_LM);

>>>       d->offset += SZ_64K;

>>>   }

>>> @@ -73,7 +77,8 @@ static void insert_pte(struct i915_address_space *vm,

>>>   {

>>>       struct insert_pte_data *d = data;

>>> -    vm->insert_page(vm, px_dma(pt), d->offset, I915_CACHE_NONE,

>>> +    vm->insert_page(vm, px_dma(pt), d->offset,

>>> +            i915_gem_get_pat_index(vm->i915, I915_CACHE_NONE),

>>>               i915_gem_object_is_lmem(pt->base) ? PTE_LM : 0);

>>>       d->offset += PAGE_SIZE;

>>>   }

>>> @@ -356,13 +361,13 @@ static int max_pte_pkt_size(struct i915_request

>>> *rq, int pkt)

>>>   static int emit_pte(struct i915_request *rq,

>>>               struct sgt_dma *it,

>>> -            enum i915_cache_level cache_level,

>>> +            unsigned int pat_index,

>>>               bool is_lmem,

>>>               u64 offset,

>>>               int length)

>>>   {

>>>       bool has_64K_pages = HAS_64K_PAGES(rq->engine->i915);

>>> -    const u64 encode = rq->context->vm->pte_encode(0, cache_level,

>>> +    const u64 encode = rq->context->vm->pte_encode(0, pat_index,

>>>                                  is_lmem ? PTE_LM : 0);

>>>       struct intel_ring *ring = rq->ring;

>>>       int pkt, dword_length;

>>> @@ -673,17 +678,17 @@ int

>>>   intel_context_migrate_copy(struct intel_context *ce,

>>>                  const struct i915_deps *deps,

>>>                  struct scatterlist *src,

>>> -               enum i915_cache_level src_cache_level,

>>> +               unsigned int src_pat_index,

>>>                  bool src_is_lmem,

>>>                  struct scatterlist *dst,

>>> -               enum i915_cache_level dst_cache_level,

>>> +               unsigned int dst_pat_index,

>>>                  bool dst_is_lmem,

>>>                  struct i915_request **out)

>>>   {

>>>       struct sgt_dma it_src = sg_sgt(src), it_dst = sg_sgt(dst), it_ccs;

>>>       struct drm_i915_private *i915 = ce->engine->i915;

>>>       u64 ccs_bytes_to_cpy = 0, bytes_to_cpy;

>>> -    enum i915_cache_level ccs_cache_level;

>>> +    unsigned int ccs_pat_index;

>>>       u32 src_offset, dst_offset;

>>>       u8 src_access, dst_access;

>>>       struct i915_request *rq;

>>> @@ -707,12 +712,12 @@ intel_context_migrate_copy(struct intel_context

>>> *ce,

>>>           dst_sz = scatter_list_length(dst);

>>>           if (src_is_lmem) {

>>>               it_ccs = it_dst;

>>> -            ccs_cache_level = dst_cache_level;

>>> +            ccs_pat_index = dst_pat_index;

>>>               ccs_is_src = false;

>>>           } else if (dst_is_lmem) {

>>>               bytes_to_cpy = dst_sz;

>>>               it_ccs = it_src;

>>> -            ccs_cache_level = src_cache_level;

>>> +            ccs_pat_index = src_pat_index;

>>>               ccs_is_src = true;

>>>           }

>>> @@ -773,7 +778,7 @@ intel_context_migrate_copy(struct intel_context *ce,

>>>           src_sz = calculate_chunk_sz(i915, src_is_lmem,

>>>                           bytes_to_cpy, ccs_bytes_to_cpy);

>>> -        len = emit_pte(rq, &it_src, src_cache_level, src_is_lmem,

>>> +        len = emit_pte(rq, &it_src, src_pat_index, src_is_lmem,

>>>                      src_offset, src_sz);

>>>           if (!len) {

>>>               err = -EINVAL;

>>> @@ -784,7 +789,7 @@ intel_context_migrate_copy(struct intel_context *ce,

>>>               goto out_rq;

>>>           }

>>> -        err = emit_pte(rq, &it_dst, dst_cache_level, dst_is_lmem,

>>> +        err = emit_pte(rq, &it_dst, dst_pat_index, dst_is_lmem,

>>>                      dst_offset, len);

>>>           if (err < 0)

>>>               goto out_rq;

>>> @@ -811,7 +816,7 @@ intel_context_migrate_copy(struct intel_context *ce,

>>>                   goto out_rq;

>>>               ccs_sz = GET_CCS_BYTES(i915, len);

>>> -            err = emit_pte(rq, &it_ccs, ccs_cache_level, false,

>>> +            err = emit_pte(rq, &it_ccs, ccs_pat_index, false,

>>>                          ccs_is_src ? src_offset : dst_offset,

>>>                          ccs_sz);

>>>               if (err < 0)

>>> @@ -979,7 +984,7 @@ int

>>>   intel_context_migrate_clear(struct intel_context *ce,

>>>                   const struct i915_deps *deps,

>>>                   struct scatterlist *sg,

>>> -                enum i915_cache_level cache_level,

>>> +                unsigned int pat_index,

>>>                   bool is_lmem,

>>>                   u32 value,

>>>                   struct i915_request **out)

>>> @@ -1027,7 +1032,7 @@ intel_context_migrate_clear(struct intel_context

>>> *ce,

>>>           if (err)

>>>               goto out_rq;

>>> -        len = emit_pte(rq, &it, cache_level, is_lmem, offset, CHUNK_SZ);

>>> +        len = emit_pte(rq, &it, pat_index, is_lmem, offset, CHUNK_SZ);

>>>           if (len <= 0) {

>>>               err = len;

>>>               goto out_rq;

>>> @@ -1074,10 +1079,10 @@ int intel_migrate_copy(struct intel_migrate *m,

>>>                  struct i915_gem_ww_ctx *ww,

>>>                  const struct i915_deps *deps,

>>>                  struct scatterlist *src,

>>> -               enum i915_cache_level src_cache_level,

>>> +               unsigned int src_pat_index,

>>>                  bool src_is_lmem,

>>>                  struct scatterlist *dst,

>>> -               enum i915_cache_level dst_cache_level,

>>> +               unsigned int dst_pat_index,

>>>                  bool dst_is_lmem,

>>>                  struct i915_request **out)

>>>   {

>>> @@ -1098,8 +1103,8 @@ int intel_migrate_copy(struct intel_migrate *m,

>>>           goto out;

>>>       err = intel_context_migrate_copy(ce, deps,

>>> -                     src, src_cache_level, src_is_lmem,

>>> -                     dst, dst_cache_level, dst_is_lmem,

>>> +                     src, src_pat_index, src_is_lmem,

>>> +                     dst, dst_pat_index, dst_is_lmem,

>>>                        out);

>>>       intel_context_unpin(ce);

>>> @@ -1113,7 +1118,7 @@ intel_migrate_clear(struct intel_migrate *m,

>>>               struct i915_gem_ww_ctx *ww,

>>>               const struct i915_deps *deps,

>>>               struct scatterlist *sg,

>>> -            enum i915_cache_level cache_level,

>>> +            unsigned int pat_index,

>>>               bool is_lmem,

>>>               u32 value,

>>>               struct i915_request **out)

>>> @@ -1134,7 +1139,7 @@ intel_migrate_clear(struct intel_migrate *m,

>>>       if (err)

>>>           goto out;

>>> -    err = intel_context_migrate_clear(ce, deps, sg, cache_level,

>>> +    err = intel_context_migrate_clear(ce, deps, sg, pat_index,

>>>                         is_lmem, value, out);

>>>       intel_context_unpin(ce);

>>> diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.h

>>> b/drivers/gpu/drm/i915/gt/intel_migrate.h

>>> index ccc677ec4aa3..11fc09a00c4b 100644

>>> --- a/drivers/gpu/drm/i915/gt/intel_migrate.h

>>> +++ b/drivers/gpu/drm/i915/gt/intel_migrate.h

>>> @@ -16,7 +16,6 @@ struct i915_request;

>>>   struct i915_gem_ww_ctx;

>>>   struct intel_gt;

>>>   struct scatterlist;

>>> -enum i915_cache_level;

>>>   int intel_migrate_init(struct intel_migrate *m, struct intel_gt *gt);

>>> @@ -26,20 +25,20 @@ int intel_migrate_copy(struct intel_migrate *m,

>>>                  struct i915_gem_ww_ctx *ww,

>>>                  const struct i915_deps *deps,

>>>                  struct scatterlist *src,

>>> -               enum i915_cache_level src_cache_level,

>>> +               unsigned int src_pat_index,

>>>                  bool src_is_lmem,

>>>                  struct scatterlist *dst,

>>> -               enum i915_cache_level dst_cache_level,

>>> +               unsigned int dst_pat_index,

>>>                  bool dst_is_lmem,

>>>                  struct i915_request **out);

>>>   int intel_context_migrate_copy(struct intel_context *ce,

>>>                      const struct i915_deps *deps,

>>>                      struct scatterlist *src,

>>> -                   enum i915_cache_level src_cache_level,

>>> +                   unsigned int src_pat_index,

>>>                      bool src_is_lmem,

>>>                      struct scatterlist *dst,

>>> -                   enum i915_cache_level dst_cache_level,

>>> +                   unsigned int dst_pat_index,

>>>                      bool dst_is_lmem,

>>>                      struct i915_request **out);

>>> @@ -48,7 +47,7 @@ intel_migrate_clear(struct intel_migrate *m,

>>>               struct i915_gem_ww_ctx *ww,

>>>               const struct i915_deps *deps,

>>>               struct scatterlist *sg,

>>> -            enum i915_cache_level cache_level,

>>> +            unsigned int pat_index,

>>>               bool is_lmem,

>>>               u32 value,

>>>               struct i915_request **out);

>>> @@ -56,7 +55,7 @@ int

>>>   intel_context_migrate_clear(struct intel_context *ce,

>>>                   const struct i915_deps *deps,

>>>                   struct scatterlist *sg,

>>> -                enum i915_cache_level cache_level,

>>> +                unsigned int pat_index,

>>>                   bool is_lmem,

>>>                   u32 value,

>>>                   struct i915_request **out);

>>> diff --git a/drivers/gpu/drm/i915/gt/intel_ppgtt.c

>>> b/drivers/gpu/drm/i915/gt/intel_ppgtt.c

>>> index 7ecfa672f738..f0da3555c6db 100644

>>> --- a/drivers/gpu/drm/i915/gt/intel_ppgtt.c

>>> +++ b/drivers/gpu/drm/i915/gt/intel_ppgtt.c

>>> @@ -98,7 +98,7 @@ void

>>>   __set_pd_entry(struct i915_page_directory * const pd,

>>>              const unsigned short idx,

>>>              struct i915_page_table * const to,

>>> -           u64 (*encode)(const dma_addr_t, const enum i915_cache_level))

>>> +           u64 (*encode)(const dma_addr_t, const unsigned int))

>>>   {

>>>       /* Each thread pre-pins the pd, and we may have a thread per

>>> pde. */

>>>       GEM_BUG_ON(atomic_read(px_used(pd)) > NALLOC * I915_PDES);

>>> @@ -181,7 +181,7 @@ struct i915_ppgtt *i915_ppgtt_create(struct

>>> intel_gt *gt,

>>>   void ppgtt_bind_vma(struct i915_address_space *vm,

>>>               struct i915_vm_pt_stash *stash,

>>>               struct i915_vma_resource *vma_res,

>>> -            enum i915_cache_level cache_level,

>>> +            unsigned int pat_index,

>>>               u32 flags)

>>>   {

>>>       u32 pte_flags;

>>> @@ -199,7 +199,7 @@ void ppgtt_bind_vma(struct i915_address_space *vm,

>>>       if (vma_res->bi.lmem)

>>>           pte_flags |= PTE_LM;

>>> -    vm->insert_entries(vm, vma_res, cache_level, pte_flags);

>>> +    vm->insert_entries(vm, vma_res, pat_index, pte_flags);

>>>       wmb();

>>>   }

>>> diff --git a/drivers/gpu/drm/i915/gt/selftest_migrate.c

>>> b/drivers/gpu/drm/i915/gt/selftest_migrate.c

>>> index e677f2da093d..3def5ca72dec 100644

>>> --- a/drivers/gpu/drm/i915/gt/selftest_migrate.c

>>> +++ b/drivers/gpu/drm/i915/gt/selftest_migrate.c

>>> @@ -137,7 +137,7 @@ static int copy(struct intel_migrate *migrate,

>>>   static int intel_context_copy_ccs(struct intel_context *ce,

>>>                     const struct i915_deps *deps,

>>>                     struct scatterlist *sg,

>>> -                  enum i915_cache_level cache_level,

>>> +                  unsigned int pat_index,

>>>                     bool write_to_ccs,

>>>                     struct i915_request **out)

>>>   {

>>> @@ -185,7 +185,7 @@ static int intel_context_copy_ccs(struct

>>> intel_context *ce,

>>>           if (err)

>>>               goto out_rq;

>>> -        len = emit_pte(rq, &it, cache_level, true, offset, CHUNK_SZ);

>>> +        len = emit_pte(rq, &it, pat_index, true, offset, CHUNK_SZ);

>>>           if (len <= 0) {

>>>               err = len;

>>>               goto out_rq;

>>> @@ -223,7 +223,7 @@ intel_migrate_ccs_copy(struct intel_migrate *m,

>>>                  struct i915_gem_ww_ctx *ww,

>>>                  const struct i915_deps *deps,

>>>                  struct scatterlist *sg,

>>> -               enum i915_cache_level cache_level,

>>> +               unsigned int pat_index,

>>>                  bool write_to_ccs,

>>>                  struct i915_request **out)

>>>   {

>>> @@ -243,7 +243,7 @@ intel_migrate_ccs_copy(struct intel_migrate *m,

>>>       if (err)

>>>           goto out;

>>> -    err = intel_context_copy_ccs(ce, deps, sg, cache_level,

>>> +    err = intel_context_copy_ccs(ce, deps, sg, pat_index,

>>>                        write_to_ccs, out);

>>>       intel_context_unpin(ce);

>>> @@ -300,7 +300,7 @@ static int clear(struct intel_migrate *migrate,

>>>               /* Write the obj data into ccs surface */

>>>               err = intel_migrate_ccs_copy(migrate, &ww, NULL,

>>>                                obj->mm.pages->sgl,

>>> -                             obj->cache_level,

>>> +                             obj->pat_index,

>>>                                true, &rq);

>>>               if (rq && !err) {

>>>                   if (i915_request_wait(rq, 0, HZ) < 0) {

>>> @@ -351,7 +351,7 @@ static int clear(struct intel_migrate *migrate,

>>>               err = intel_migrate_ccs_copy(migrate, &ww, NULL,

>>>                                obj->mm.pages->sgl,

>>> -                             obj->cache_level,

>>> +                             obj->pat_index,

>>>                                false, &rq);

>>>               if (rq && !err) {

>>>                   if (i915_request_wait(rq, 0, HZ) < 0) {

>>> @@ -414,9 +414,9 @@ static int __migrate_copy(struct intel_migrate

>>> *migrate,

>>>                 struct i915_request **out)

>>>   {

>>>       return intel_migrate_copy(migrate, ww, NULL,

>>> -                  src->mm.pages->sgl, src->cache_level,

>>> +                  src->mm.pages->sgl, src->pat_index,

>>>                     i915_gem_object_is_lmem(src),

>>> -                  dst->mm.pages->sgl, dst->cache_level,

>>> +                  dst->mm.pages->sgl, dst->pat_index,

>>>                     i915_gem_object_is_lmem(dst),

>>>                     out);

>>>   }

>>> @@ -428,9 +428,9 @@ static int __global_copy(struct intel_migrate

>>> *migrate,

>>>                struct i915_request **out)

>>>   {

>>>       return intel_context_migrate_copy(migrate->context, NULL,

>>> -                      src->mm.pages->sgl, src->cache_level,

>>> +                      src->mm.pages->sgl, src->pat_index,

>>>                         i915_gem_object_is_lmem(src),

>>> -                      dst->mm.pages->sgl, dst->cache_level,

>>> +                      dst->mm.pages->sgl, dst->pat_index,

>>>                         i915_gem_object_is_lmem(dst),

>>>                         out);

>>>   }

>>> @@ -455,7 +455,7 @@ static int __migrate_clear(struct intel_migrate

>>> *migrate,

>>>   {

>>>       return intel_migrate_clear(migrate, ww, NULL,

>>>                      obj->mm.pages->sgl,

>>> -                   obj->cache_level,

>>> +                   obj->pat_index,

>>>                      i915_gem_object_is_lmem(obj),

>>>                      value, out);

>>>   }

>>> @@ -468,7 +468,7 @@ static int __global_clear(struct intel_migrate

>>> *migrate,

>>>   {

>>>       return intel_context_migrate_clear(migrate->context, NULL,

>>>                          obj->mm.pages->sgl,

>>> -                       obj->cache_level,

>>> +                       obj->pat_index,

>>>                          i915_gem_object_is_lmem(obj),

>>>                          value, out);

>>>   }

>>> @@ -648,7 +648,7 @@ static int live_emit_pte_full_ring(void *arg)

>>>        */

>>>       pr_info("%s emite_pte ring space=%u\n", __func__, rq->ring->space);

>>>       it = sg_sgt(obj->mm.pages->sgl);

>>> -    len = emit_pte(rq, &it, obj->cache_level, false, 0, CHUNK_SZ);

>>> +    len = emit_pte(rq, &it, obj->pat_index, false, 0, CHUNK_SZ);

>>>       if (!len) {

>>>           err = -EINVAL;

>>>           goto out_rq;

>>> @@ -844,7 +844,7 @@ static int wrap_ktime_compare(const void *A, const

>>> void *B)

>>>   static int __perf_clear_blt(struct intel_context *ce,

>>>                   struct scatterlist *sg,

>>> -                enum i915_cache_level cache_level,

>>> +                unsigned int pat_index,

>>>                   bool is_lmem,

>>>                   size_t sz)

>>>   {

>>> @@ -858,7 +858,7 @@ static int __perf_clear_blt(struct intel_context *ce,

>>>           t0 = ktime_get();

>>> -        err = intel_context_migrate_clear(ce, NULL, sg, cache_level,

>>> +        err = intel_context_migrate_clear(ce, NULL, sg, pat_index,

>>>                             is_lmem, 0, &rq);

>>>           if (rq) {

>>>               if (i915_request_wait(rq, 0, MAX_SCHEDULE_TIMEOUT) < 0)

>>> @@ -904,7 +904,8 @@ static int perf_clear_blt(void *arg)

>>>           err = __perf_clear_blt(gt->migrate.context,

>>>                          dst->mm.pages->sgl,

>>> -                       I915_CACHE_NONE,

>>> +                       i915_gem_get_pat_index(gt->i915,

>>> +                                  I915_CACHE_NONE),

>>>                          i915_gem_object_is_lmem(dst),

>>>                          sizes[i]);

>>> @@ -919,10 +920,10 @@ static int perf_clear_blt(void *arg)

>>>   static int __perf_copy_blt(struct intel_context *ce,

>>>                  struct scatterlist *src,

>>> -               enum i915_cache_level src_cache_level,

>>> +               unsigned int src_pat_index,

>>>                  bool src_is_lmem,

>>>                  struct scatterlist *dst,

>>> -               enum i915_cache_level dst_cache_level,

>>> +               unsigned int dst_pat_index,

>>>                  bool dst_is_lmem,

>>>                  size_t sz)

>>>   {

>>> @@ -937,9 +938,9 @@ static int __perf_copy_blt(struct intel_context *ce,

>>>           t0 = ktime_get();

>>>           err = intel_context_migrate_copy(ce, NULL,

>>> -                         src, src_cache_level,

>>> +                         src, src_pat_index,

>>>                            src_is_lmem,

>>> -                         dst, dst_cache_level,

>>> +                         dst, dst_pat_index,

>>>                            dst_is_lmem,

>>>                            &rq);

>>>           if (rq) {

>>> @@ -994,10 +995,12 @@ static int perf_copy_blt(void *arg)

>>>           err = __perf_copy_blt(gt->migrate.context,

>>>                         src->mm.pages->sgl,

>>> -                      I915_CACHE_NONE,

>>> +                      i915_gem_get_pat_index(gt->i915,

>>> +                                 I915_CACHE_NONE),

>>>                         i915_gem_object_is_lmem(src),

>>>                         dst->mm.pages->sgl,

>>> -                      I915_CACHE_NONE,

>>> +                      i915_gem_get_pat_index(gt->i915,

>>> +                                 I915_CACHE_NONE),

>>>                         i915_gem_object_is_lmem(dst),

>>>                         sz);

>>> diff --git a/drivers/gpu/drm/i915/gt/selftest_reset.c

>>> b/drivers/gpu/drm/i915/gt/selftest_reset.c

>>> index a9e0a91bc0e0..79aa6ac66ad2 100644

>>> --- a/drivers/gpu/drm/i915/gt/selftest_reset.c

>>> +++ b/drivers/gpu/drm/i915/gt/selftest_reset.c

>>> @@ -86,7 +86,9 @@ __igt_reset_stolen(struct intel_gt *gt,

>>>           ggtt->vm.insert_page(&ggtt->vm, dma,

>>>                        ggtt->error_capture.start,

>>> -                     I915_CACHE_NONE, 0);

>>> +                     i915_gem_get_pat_index(gt->i915,

>>> +                                I915_CACHE_NONE),

>>> +                     0);

>>>           mb();

>>>           s = io_mapping_map_wc(&ggtt->iomap,

>>> @@ -127,7 +129,9 @@ __igt_reset_stolen(struct intel_gt *gt,

>>>           ggtt->vm.insert_page(&ggtt->vm, dma,

>>>                        ggtt->error_capture.start,

>>> -                     I915_CACHE_NONE, 0);

>>> +                     i915_gem_get_pat_index(gt->i915,

>>> +                                I915_CACHE_NONE),

>>> +                     0);

>>>           mb();

>>>           s = io_mapping_map_wc(&ggtt->iomap,

>>> diff --git a/drivers/gpu/drm/i915/gt/selftest_timeline.c

>>> b/drivers/gpu/drm/i915/gt/selftest_timeline.c

>>> index 9f536c251179..39c3ec12df1a 100644

>>> --- a/drivers/gpu/drm/i915/gt/selftest_timeline.c

>>> +++ b/drivers/gpu/drm/i915/gt/selftest_timeline.c

>>> @@ -836,7 +836,7 @@ static int setup_watcher(struct hwsp_watcher *w,

>>> struct intel_gt *gt,

>>>           return PTR_ERR(obj);

>>>       /* keep the same cache settings as timeline */

>>> -    i915_gem_object_set_cache_coherency(obj,

>>> tl->hwsp_ggtt->obj->cache_level);

>>> +    i915_gem_object_set_pat_index(obj, tl->hwsp_ggtt->obj->pat_index);

>>>       w->map = i915_gem_object_pin_map_unlocked(obj,

>>>                            

>>> page_unmask_bits(tl->hwsp_ggtt->obj->mm.mapping));

>>>       if (IS_ERR(w->map)) {

>>> diff --git a/drivers/gpu/drm/i915/gt/selftest_tlb.c

>>> b/drivers/gpu/drm/i915/gt/selftest_tlb.c

>>> index e6cac1f15d6e..4493c8518e91 100644

>>> --- a/drivers/gpu/drm/i915/gt/selftest_tlb.c

>>> +++ b/drivers/gpu/drm/i915/gt/selftest_tlb.c

>>> @@ -36,6 +36,8 @@ pte_tlbinv(struct intel_context *ce,

>>>          u64 length,

>>>          struct rnd_state *prng)

>>>   {

>>> +    const unsigned int pat_index =

>>> +        i915_gem_get_pat_index(ce->vm->i915, I915_CACHE_NONE);

>>>       struct drm_i915_gem_object *batch;

>>>       struct drm_mm_node vb_node;

>>>       struct i915_request *rq;

>>> @@ -155,7 +157,7 @@ pte_tlbinv(struct intel_context *ce,

>>>           /* Flip the PTE between A and B */

>>>           if (i915_gem_object_is_lmem(vb->obj))

>>>               pte_flags |= PTE_LM;

>>> -        ce->vm->insert_entries(ce->vm, &vb_res, 0, pte_flags);

>>> +        ce->vm->insert_entries(ce->vm, &vb_res, pat_index, pte_flags);

>>>           /* Flush the PTE update to concurrent HW */

>>>           tlbinv(ce->vm, addr & -length, length);

>>> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c

>>> b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c

>>> index a82a53dbbc86..145681ae20a5 100644

>>> --- a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c

>>> +++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c

>>> @@ -890,9 +890,15 @@ static void uc_fw_bind_ggtt(struct intel_uc_fw

>>> *uc_fw)

>>>           pte_flags |= PTE_LM;

>>>       if (ggtt->vm.raw_insert_entries)

>>> -        ggtt->vm.raw_insert_entries(&ggtt->vm, dummy,

>>> I915_CACHE_NONE, pte_flags);

>>> +        ggtt->vm.raw_insert_entries(&ggtt->vm, dummy,

>>> +                        i915_gem_get_pat_index(ggtt->vm.i915,

>>> +                                   I915_CACHE_NONE),

>>> +                        pte_flags);

>>>       else

>>> -        ggtt->vm.insert_entries(&ggtt->vm, dummy, I915_CACHE_NONE,

>>> pte_flags);

>>> +        ggtt->vm.insert_entries(&ggtt->vm, dummy,

>>> +                    i915_gem_get_pat_index(ggtt->vm.i915,

>>> +                                   I915_CACHE_NONE),

>>> +                    pte_flags);

>>>   }

>>>   static void uc_fw_unbind_ggtt(struct intel_uc_fw *uc_fw)

>>> diff --git a/drivers/gpu/drm/i915/i915_debugfs.c

>>> b/drivers/gpu/drm/i915/i915_debugfs.c

>>> index 41389a32e998..9a4922da3a71 100644

>>> --- a/drivers/gpu/drm/i915/i915_debugfs.c

>>> +++ b/drivers/gpu/drm/i915/i915_debugfs.c

>>> @@ -139,21 +139,56 @@ static const char *stringify_vma_type(const

>>> struct i915_vma *vma)

>>>       return "ppgtt";

>>>   }

>>> -static const char *i915_cache_level_str(struct drm_i915_private

>>> *i915, int type)

>>> -{

>>> -    switch (type) {

>>> -    case I915_CACHE_NONE: return " uncached";

>>> -    case I915_CACHE_LLC: return HAS_LLC(i915) ? " LLC" : " snooped";

>>> -    case I915_CACHE_L3_LLC: return " L3+LLC";

>>> -    case I915_CACHE_WT: return " WT";

>>> -    default: return "";

>>> +static const char *i915_cache_level_str(struct drm_i915_gem_object *obj)

>>> +{

>>> +    struct drm_i915_private *i915 = obj_to_i915(obj);

>>> +

>>> +    if (IS_METEORLAKE(i915)) {

>>> +        switch (obj->pat_index) {

>>> +        case 0: return " WB";

>>> +        case 1: return " WT";

>>> +        case 2: return " UC";

>>> +        case 3: return " WB (1-Way Coh)";

>>> +        case 4: return " WB (2-Way Coh)";

>>> +        default: return " not defined";

>

> Is not defined possible?

Not possible because there is a sanity check on pat_index, but I think I got

a compilation warning without default, then due to the treat-warning-as-error

flag, I actually failed the compilation.

> Also, it may be nicer to handle the leading space in the caller.

will check...

>>> +        }

>>> +    } else if (IS_PONTEVECCHIO(i915)) {

>>> +        switch (obj->pat_index) {

>>> +        case 0: return " UC";

>>> +        case 1: return " WC";

>>> +        case 2: return " WT";

>>> +        case 3: return " WB";

>>> +        case 4: return " WT (CLOS1)";

>>> +        case 5: return " WB (CLOS1)";

>>> +        case 6: return " WT (CLOS2)";

>>> +        case 7: return " WT (CLOS2)";

>>> +        default: return " not defined";

>>> +        }

>>> +    } else if (GRAPHICS_VER(i915) >= 12) {

>>> +        switch (obj->pat_index) {

>>> +        case 0: return " WB";

>>> +        case 1: return " WC";

>>> +        case 2: return " WT";

>>> +        case 3: return " UC";

>>> +        default: return " not defined";

>>> +        }

>>> +    } else {

>

> Is this correct if a legacy platform used the set pat extension?

> I don't see that it is disallowed.

That is allowed. For legacy platforms pat_index is the same as cache_level.

> Would it simplify things to add a reverse table to device info, so like

> cachelevel_to_pat, just for pat_index to names? I guess it depends what

> names PRMs use for PATs on legacy platforms. Is it consistend with the

> above UC/WC/WB/... or with the below names.

I was trying to avoid adding tables, but couldn't get away with cachelevel_to_pat

translation. Well, I guess having constant tables defined locally here might

not be too bad...

>>> +        if (i915_gem_object_has_cache_level(obj, I915_CACHE_NONE))

>>> +            return " uncached";

>

> UC for consistency?

Will update.

>>> +        else if (i915_gem_object_has_cache_level(obj, I915_CACHE_LLC))

>>> +            return HAS_LLC(i915) ? " LLC" : " snooped";

>>> +        else if (i915_gem_object_has_cache_level(obj,

>>> I915_CACHE_L3_LLC))

>>> +            return " L3+LLC";

>

> Is this correct if !HAS_LLC?

Nice catch. Will update.

>>> +        else if (i915_gem_object_has_cache_level(obj, I915_CACHE_WT))

>>> +            return " WT";

>>> +        else

>>> +            return " not defined";

>

> Current code prints nothing for the default switch statement.

>

> But is this even reachable or should it be MISSING_CASE warning?

hmm... I guess it's possible, will re-examine the code.

>>>       }

>>>   }

>>>   void

>>>   i915_debugfs_describe_obj(struct seq_file *m, struct

>>> drm_i915_gem_object *obj)

>>>   {

>>> -    struct drm_i915_private *dev_priv = to_i915(obj->base.dev);

>>>       struct i915_vma *vma;

>>>       int pin_count = 0;

>>> @@ -165,7 +200,7 @@ i915_debugfs_describe_obj(struct seq_file *m,

>>> struct drm_i915_gem_object *obj)

>>>              obj->base.size / 1024,

>>>              obj->read_domains,

>>>              obj->write_domain,

>>> -           i915_cache_level_str(dev_priv, obj->cache_level),

>>> +           i915_cache_level_str(obj),

>>>              obj->mm.dirty ? " dirty" : "",

>>>              obj->mm.madv == I915_MADV_DONTNEED ? " purgeable" : "");

>>>       if (obj->base.name)

>>> diff --git a/drivers/gpu/drm/i915/i915_gem.c

>>> b/drivers/gpu/drm/i915/i915_gem.c

>>> index 0a78bdbd36b1..63207b0740b3 100644

>>> --- a/drivers/gpu/drm/i915/i915_gem.c

>>> +++ b/drivers/gpu/drm/i915/i915_gem.c

>>> @@ -420,8 +420,12 @@ i915_gem_gtt_pread(struct drm_i915_gem_object *obj,

>>>           page_length = remain < page_length ? remain : page_length;

>>>           if (drm_mm_node_allocated(&node)) {

>>>               ggtt->vm.insert_page(&ggtt->vm,

>>> -                         i915_gem_object_get_dma_address(obj, offset

>>> >> PAGE_SHIFT),

>>> -                         node.start, I915_CACHE_NONE, 0);

>>> +                    i915_gem_object_get_dma_address(obj,

>>> +                                    offset >> PAGE_SHIFT),

>>> +                    node.start,

>>> +                    i915_gem_get_pat_index(i915,

>>> +                                   I915_CACHE_NONE),

>

> For the callsites which use const levels you could at least do something

> like i915->pat_cache_none, or I know the not very popular static inline

> i915_gem_get_pat_index so it can be evaluated at runtime. Not sure really,

> throwing out ideas which may be invalid if a more elegant refactoring is

> possible.

You meant defining each cache_level in the i915 structure in stead of having

a constant table to do the translation?

-Fei

> Regards,

>

> Tvrtko

>

>>> +                    0);

>>>           } else {

>>>               page_base += offset & PAGE_MASK;

>>>           }

>>> @@ -598,8 +602,12 @@ i915_gem_gtt_pwrite_fast(struct

>>> drm_i915_gem_object *obj,

>>>               /* flush the write before we modify the GGTT */

>>>               intel_gt_flush_ggtt_writes(ggtt->vm.gt);

>>>               ggtt->vm.insert_page(&ggtt->vm,

>>> -                         i915_gem_object_get_dma_address(obj, offset

>>> >> PAGE_SHIFT),

>>> -                         node.start, I915_CACHE_NONE, 0);

>>> +                    i915_gem_object_get_dma_address(obj,

>>> +                                    offset >> PAGE_SHIFT),

>>> +                    node.start,

>>> +                    i915_gem_get_pat_index(i915,

>>> +                                   I915_CACHE_NONE),

>>> +                    0);

>>>               wmb(); /* flush modifications to the GGTT (insert_page) */

>>>           } else {

>>>               page_base += offset & PAGE_MASK;

>>> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c

>>> b/drivers/gpu/drm/i915/i915_gpu_error.c

>>> index f020c0086fbc..2556cabea02c 100644

>>> --- a/drivers/gpu/drm/i915/i915_gpu_error.c

>>> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c

>>> @@ -1117,10 +1117,14 @@ i915_vma_coredump_create(const struct intel_gt

>>> *gt,

>>>               mutex_lock(&ggtt->error_mutex);

>>>               if (ggtt->vm.raw_insert_page)

>>>                   ggtt->vm.raw_insert_page(&ggtt->vm, dma, slot,

>>> -                             I915_CACHE_NONE, 0);

>>> +                        i915_gem_get_pat_index(gt->i915,

>>> +                                       I915_CACHE_NONE),

>>> +                        0);

>>>               else

>>>                   ggtt->vm.insert_page(&ggtt->vm, dma, slot,

>>> -                             I915_CACHE_NONE, 0);

>>> +                        i915_gem_get_pat_index(gt->i915,

>>> +                                       I915_CACHE_NONE),

>>> +                        0);

>>>               mb();

>>>               s = io_mapping_map_wc(&ggtt->iomap, slot, PAGE_SIZE);

>>> diff --git a/drivers/gpu/drm/i915/i915_vma.c

>>> b/drivers/gpu/drm/i915/i915_vma.c

>>> index 20a44788999e..a814775a363d 100644

>>> --- a/drivers/gpu/drm/i915/i915_vma.c

>>> +++ b/drivers/gpu/drm/i915/i915_vma.c

>>> @@ -315,7 +315,7 @@ struct i915_vma_work {

>>>       struct i915_vma_resource *vma_res;

>>>       struct drm_i915_gem_object *obj;

>>>       struct i915_sw_dma_fence_cb cb;

>>> -    enum i915_cache_level cache_level;

>>> +    unsigned int pat_index;

>>>       unsigned int flags;

>>>   };

>>> @@ -334,7 +334,7 @@ static void __vma_bind(struct dma_fence_work *work)

>>>           return;

>>>       vma_res->ops->bind_vma(vma_res->vm, &vw->stash,

>>> -                   vma_res, vw->cache_level, vw->flags);

>>> +                   vma_res, vw->pat_index, vw->flags);

>>>   }

>>>   static void __vma_release(struct dma_fence_work *work)

>>> @@ -426,7 +426,7 @@ i915_vma_resource_init_from_vma(struct

>>> i915_vma_resource *vma_res,

>>>   /**

>>>    * i915_vma_bind - Sets up PTEs for an VMA in it's corresponding

>>> address space.

>>>    * @vma: VMA to map

>>> - * @cache_level: mapping cache level

>>> + * @pat_index: PAT index to set in PTE

>>>    * @flags: flags like global or local mapping

>>>    * @work: preallocated worker for allocating and binding the PTE

>>>    * @vma_res: pointer to a preallocated vma resource. The resource is

>>> either

>>> @@ -437,7 +437,7 @@ i915_vma_resource_init_from_vma(struct

>>> i915_vma_resource *vma_res,

>>>    * Note that DMA addresses are also the only part of the SG table we

>>> care about.

>>>    */

>>>   int i915_vma_bind(struct i915_vma *vma,

>>> -          enum i915_cache_level cache_level,

>>> +          unsigned int pat_index,

>>>             u32 flags,

>>>             struct i915_vma_work *work,

>>>             struct i915_vma_resource *vma_res)

>>> @@ -507,7 +507,7 @@ int i915_vma_bind(struct i915_vma *vma,

>>>           struct dma_fence *prev;

>>>           work->vma_res = i915_vma_resource_get(vma->resource);

>>> -        work->cache_level = cache_level;

>>> +        work->pat_index = pat_index;

>>>           work->flags = bind_flags;

>>>           /*

>>> @@ -537,7 +537,7 @@ int i915_vma_bind(struct i915_vma *vma,

>>>               return ret;

>>>           }

>>> -        vma->ops->bind_vma(vma->vm, NULL, vma->resource, cache_level,

>>> +        vma->ops->bind_vma(vma->vm, NULL, vma->resource, pat_index,

>>>                      bind_flags);

>>>       }

>>> @@ -814,7 +814,7 @@ i915_vma_insert(struct i915_vma *vma, struct

>>> i915_gem_ww_ctx *ww,

>>>       color = 0;

>>>       if (i915_vm_has_cache_coloring(vma->vm))

>>> -        color = vma->obj->cache_level;

>>> +        color = vma->obj->pat_index;

>>>       if (flags & PIN_OFFSET_FIXED) {

>>>           u64 offset = flags & PIN_OFFSET_MASK;

>>> @@ -1518,7 +1518,7 @@ int i915_vma_pin_ww(struct i915_vma *vma, struct

>>> i915_gem_ww_ctx *ww,

>>>       GEM_BUG_ON(!vma->pages);

>>>       err = i915_vma_bind(vma,

>>> -                vma->obj->cache_level,

>>> +                vma->obj->pat_index,

>>>                   flags, work, vma_res);

>>>       vma_res = NULL;

>>>       if (err)

>>> diff --git a/drivers/gpu/drm/i915/i915_vma.h

>>> b/drivers/gpu/drm/i915/i915_vma.h

>>> index ed5c9d682a1b..31a8f8aa5558 100644

>>> --- a/drivers/gpu/drm/i915/i915_vma.h

>>> +++ b/drivers/gpu/drm/i915/i915_vma.h

>>> @@ -250,7 +250,7 @@ i915_vma_compare(struct i915_vma *vma,

>>>   struct i915_vma_work *i915_vma_work(void);

>>>   int i915_vma_bind(struct i915_vma *vma,

>>> -          enum i915_cache_level cache_level,

>>> +          unsigned int pat_index,

>>>             u32 flags,

>>>             struct i915_vma_work *work,

>>>             struct i915_vma_resource *vma_res);

>>> diff --git a/drivers/gpu/drm/i915/i915_vma_types.h

>>> b/drivers/gpu/drm/i915/i915_vma_types.h

>>> index 77fda2244d16..64472b7f0e77 100644

>>> --- a/drivers/gpu/drm/i915/i915_vma_types.h

>>> +++ b/drivers/gpu/drm/i915/i915_vma_types.h

>>> @@ -32,8 +32,6 @@

>>>   #include "gem/i915_gem_object_types.h"

>>> -enum i915_cache_level;

>>> -

>>>   /**

>>>    * DOC: Global GTT views

>>>    *

>>> diff --git a/drivers/gpu/drm/i915/selftests/i915_gem.c

>>> b/drivers/gpu/drm/i915/selftests/i915_gem.c

>>> index d91d0ade8abd..61da4ed9d521 100644

>>> --- a/drivers/gpu/drm/i915/selftests/i915_gem.c

>>> +++ b/drivers/gpu/drm/i915/selftests/i915_gem.c

>>> @@ -57,7 +57,10 @@ static void trash_stolen(struct drm_i915_private

>>> *i915)

>>>           u32 __iomem *s;

>>>           int x;

>>> -        ggtt->vm.insert_page(&ggtt->vm, dma, slot, I915_CACHE_NONE, 0);

>>> +        ggtt->vm.insert_page(&ggtt->vm, dma, slot,

>>> +                     i915_gem_get_pat_index(i915,

>>> +                                I915_CACHE_NONE),

>>> +                     0);

>>>           s = io_mapping_map_atomic_wc(&ggtt->iomap, slot);

>>>           for (x = 0; x < PAGE_SIZE / sizeof(u32); x++) {

>>> diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_evict.c

>>> b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c

>>> index 37068542aafe..f13a4d265814 100644

>>> --- a/drivers/gpu/drm/i915/selftests/i915_gem_evict.c

>>> +++ b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c

>>> @@ -245,7 +245,7 @@ static int igt_evict_for_cache_color(void *arg)

>>>       struct drm_mm_node target = {

>>>           .start = I915_GTT_PAGE_SIZE * 2,

>>>           .size = I915_GTT_PAGE_SIZE,

>>> -        .color = I915_CACHE_LLC,

>>> +        .color = i915_gem_get_pat_index(gt->i915, I915_CACHE_LLC),

>>>       };

>>>       struct drm_i915_gem_object *obj;

>>>       struct i915_vma *vma;

>>> @@ -308,7 +308,7 @@ static int igt_evict_for_cache_color(void *arg)

>>>       /* Attempt to remove the first *pinned* vma, by removing the

>>> (empty)

>>>        * neighbour -- this should fail.

>>>        */

>>> -    target.color = I915_CACHE_L3_LLC;

>>> +    target.color = i915_gem_get_pat_index(gt->i915, I915_CACHE_L3_LLC);

>>>       mutex_lock(&ggtt->vm.mutex);

>>>       err = i915_gem_evict_for_node(&ggtt->vm, NULL, &target, 0);

>>> diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c

>>> b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c

>>> index 154801f1c468..36940ef10108 100644

>>> --- a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c

>>> +++ b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c

>>> @@ -135,7 +135,7 @@ fake_dma_object(struct drm_i915_private *i915, u64

>>> size)

>>>       obj->write_domain = I915_GEM_DOMAIN_CPU;

>>>       obj->read_domains = I915_GEM_DOMAIN_CPU;

>>> -    obj->cache_level = I915_CACHE_NONE;

>>> +    obj->pat_index = i915_gem_get_pat_index(i915, I915_CACHE_NONE);

>>>       /* Preallocate the "backing storage" */

>>>       if (i915_gem_object_pin_pages_unlocked(obj))

>>> @@ -359,7 +359,9 @@ static int lowlevel_hole(struct i915_address_space

>>> *vm,

>>>               with_intel_runtime_pm(vm->gt->uncore->rpm, wakeref)

>>>                 vm->insert_entries(vm, mock_vma_res,

>>> -                           I915_CACHE_NONE, 0);

>>> +                         i915_gem_get_pat_index(vm->i915,

>>> +                                    I915_CACHE_NONE),

>>> +                         0);

>>>           }

>>>           count = n;

>>> @@ -1377,7 +1379,10 @@ static int igt_ggtt_page(void *arg)

>>>           ggtt->vm.insert_page(&ggtt->vm,

>>>                        i915_gem_object_get_dma_address(obj, 0),

>>> -                     offset, I915_CACHE_NONE, 0);

>>> +                     offset,

>>> +                     i915_gem_get_pat_index(i915,

>>> +                                I915_CACHE_NONE),

>>> +                     0);

>>>       }

>>>       order = i915_random_order(count, &prng);

>>> @@ -1510,7 +1515,7 @@ static int reserve_gtt_with_resource(struct

>>> i915_vma *vma, u64 offset)

>>>       mutex_lock(&vm->mutex);

>>>       err = i915_gem_gtt_reserve(vm, NULL, &vma->node, obj->base.size,

>>>                      offset,

>>> -                   obj->cache_level,

>>> +                   obj->pat_index,

>>>                      0);

>>>       if (!err) {

>>>           i915_vma_resource_init_from_vma(vma_res, vma);

>>> @@ -1690,7 +1695,7 @@ static int insert_gtt_with_resource(struct

>>> i915_vma *vma)

>>>       mutex_lock(&vm->mutex);

>>>       err = i915_gem_gtt_insert(vm, NULL, &vma->node, obj->base.size, 0,

>>> -                  obj->cache_level, 0, vm->total, 0);

>>> +                  obj->pat_index, 0, vm->total, 0);

>>>       if (!err) {

>>>           i915_vma_resource_init_from_vma(vma_res, vma);

>>>           vma->resource = vma_res;

>>> diff --git a/drivers/gpu/drm/i915/selftests/intel_memory_region.c

>>> b/drivers/gpu/drm/i915/selftests/intel_memory_region.c

>>> index 3b18e5905c86..d985d9bae2e8 100644

>>> --- a/drivers/gpu/drm/i915/selftests/intel_memory_region.c

>>> +++ b/drivers/gpu/drm/i915/selftests/intel_memory_region.c

>>> @@ -1070,7 +1070,9 @@ static int igt_lmem_write_cpu(void *arg)

>>>       /* Put the pages into a known state -- from the gpu for added

>>> fun */

>>>       intel_engine_pm_get(engine);

>>>       err = intel_context_migrate_clear(engine->gt->migrate.context,

>>> NULL,

>>> -                      obj->mm.pages->sgl, I915_CACHE_NONE,

>>> +                      obj->mm.pages->sgl,

>>> +                      i915_gem_get_pat_index(i915,

>>> +                                 I915_CACHE_NONE),

>>>                         true, 0xdeadbeaf, &rq);

>>>       if (rq) {

>>>           dma_resv_add_fence(obj->base.resv, &rq->fence,

>>> diff --git a/drivers/gpu/drm/i915/selftests/mock_gtt.c

>>> b/drivers/gpu/drm/i915/selftests/mock_gtt.c

>>> index ece97e4faacb..a516c0aa88fd 100644

>>> --- a/drivers/gpu/drm/i915/selftests/mock_gtt.c

>>> +++ b/drivers/gpu/drm/i915/selftests/mock_gtt.c

>>> @@ -27,21 +27,21 @@

>>>   static void mock_insert_page(struct i915_address_space *vm,

>>>                    dma_addr_t addr,

>>>                    u64 offset,

>>> -                 enum i915_cache_level level,

>>> +                 unsigned int pat_index,

>>>                    u32 flags)

>>>   {

>>>   }

>>>   static void mock_insert_entries(struct i915_address_space *vm,

>>>                   struct i915_vma_resource *vma_res,

>>> -                enum i915_cache_level level, u32 flags)

>>> +                unsigned int pat_index, u32 flags)

>>>   {

>>>   }

>>>   static void mock_bind_ppgtt(struct i915_address_space *vm,

>>>                   struct i915_vm_pt_stash *stash,

>>>                   struct i915_vma_resource *vma_res,

>>> -                enum i915_cache_level cache_level,

>>> +                unsigned int pat_index,

>>>                   u32 flags)

>>>   {

>>>       GEM_BUG_ON(flags & I915_VMA_GLOBAL_BIND);

>>> @@ -94,7 +94,7 @@ struct i915_ppgtt *mock_ppgtt(struct

>>> drm_i915_private *i915, const char *name)

>>>   static void mock_bind_ggtt(struct i915_address_space *vm,

>>>                  struct i915_vm_pt_stash *stash,

>>>                  struct i915_vma_resource *vma_res,

>>> -               enum i915_cache_level cache_level,

>>> +               unsigned int pat_index,

>>>                  u32 flags)

>>>   {

>>>   }

>>