[igt-dev] [PATCH i-g-t 1/2] lib/i915/intel_memory_region: Add gem_has_smallbar

[Andrzej Hajda]

On 23.01.2023 16:45, Das, Nirmoy wrote:
>
> On 1/23/2023 4:16 PM, Andrzej Hajda wrote:
>>
>>
>> On 20.01.2023 11:07, Nirmoy Das wrote:
>>> Add gem_has_smallbar() to detect a device with smallbar.
>>>
>>> Signed-off-by: Nirmoy Das <nirmoy.das at intel.com>
>>> ---
>>>   lib/i915/intel_memory_region.c | 32 ++++++++++++++++++++++++++++++++
>>>   lib/i915/intel_memory_region.h |  1 +
>>>   2 files changed, 33 insertions(+)
>>>
>>> diff --git a/lib/i915/intel_memory_region.c 
>>> b/lib/i915/intel_memory_region.c
>>> index 84e1bceb3..655720819 100644
>>> --- a/lib/i915/intel_memory_region.c
>>> +++ b/lib/i915/intel_memory_region.c
>>> @@ -183,6 +183,38 @@ bool gem_has_lmem(int fd)
>>>       return gem_get_lmem_region_count(fd) > 0;
>>>   }
>>>   +/**
>>> + * gem_has_smallbar:
>>> + * @fd: open i915 drm file descriptor
>>> + *
>>> + * Helper function to check if the device comes with small-bar.
>>> + *
>>> + * Returns: True if at least one lmem region was found.
>>> + */
>>> +bool gem_has_smallbar(int fd)
>>> +{
>>> +    struct drm_i915_query_memory_regions *info;
>>> +    struct drm_i915_memory_region_info *rf;
>>> +    bool ret = false;
>>> +
>>> +    info = gem_get_query_memory_regions(fd);
>>> +    igt_assert(info);
>>> +
>>> +    for (int i = 0; i < info->num_regions; i++) {
>>> +        rf = &info->regions[i];
>>> +        if (rf->region.memory_class == I915_MEMORY_CLASS_DEVICE) {
>>> +            if (rf->probed_size > rf->probed_cpu_visible_size) {
>>> +                ret = true;
>>> +                break;
>>> +            }
>>> +        }
>>> +    }
>>> +    free(info);
>>
>> Why not for_each_memory_region ?
>
>
> No particular reason, it is inspired by existing code from the same 
> file.  for_each_memory_region() though calls into 
> gem_get_query_memory_regions() eventually.

I though it could be shorter, sth like:
bool gem_has_smallbar(int fd)
{
     bool ret;

     for_each_memory_region(r, fd)
         if (r->ci.memory_class == I915_MEMORY_CLASS_DEVICE && r->size > 
r->cpu_size)
             ret = true;
     return ret;
}

Apparently 'break' leaks memory, so it is not so perfect :)
Up to you.

Reviewed-by: Andrzej Hajda <andrzej.hajda at intel.com>

Regards
Andrzej

>
>
> Regards,
>
> Nirmoy
>
>>
>> Regards
>> Andrzej
>>
>>> +
>>> +    return ret;
>>> +}
>>> +
>>> +
>>>   /* A version of gem_create_in_memory_region_list which can be 
>>> allowed to
>>>      fail so that the object creation can be retried */
>>>   int __gem_create_in_memory_region_list(int fd, uint32_t *handle, 
>>> uint64_t *size, uint32_t flags,
>>> diff --git a/lib/i915/intel_memory_region.h 
>>> b/lib/i915/intel_memory_region.h
>>> index 425bda0ec..9e24bd8fb 100644
>>> --- a/lib/i915/intel_memory_region.h
>>> +++ b/lib/i915/intel_memory_region.h
>>> @@ -62,6 +62,7 @@ struct drm_i915_query_memory_regions 
>>> *gem_get_query_memory_regions(int fd);
>>>   unsigned int gem_get_lmem_region_count(int fd);
>>>     bool gem_has_lmem(int fd);
>>> +bool gem_has_smallbar(int fd);
>>>     int __gem_create_in_memory_region_list(int fd, uint32_t *handle, 
>>> uint64_t *size, uint32_t flags,
>>>                          const struct 
>>> drm_i915_gem_memory_class_instance *mem_regions,
>>