[Intel-gfx] [PATCH i-g-t v3 03/11] lib/i915/gem_mman: add fixed mode to gem_mmap_offset__cpu

[Maarten Lankhorst]

Op 02-08-2021 om 08:29 schreef Dixit, Ashutosh:
> On Fri, 30 Jul 2021 01:53:40 -0700, Matthew Auld wrote:
>> On discrete we only support the new fixed mode.
>>
>> Signed-off-by: Matthew Auld <matthew.auld at intel.com>
>> Cc: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
>> Cc: Ashutosh Dixit <ashutosh.dixit at intel.com>
>> Cc: Daniel Vetter <daniel.vetter at ffwll.ch>
>> Cc: Ramalingam C <ramalingam.c at intel.com>
>> ---
>>  lib/i915/gem_mman.c | 8 +++++++-
>>  1 file changed, 7 insertions(+), 1 deletion(-)
>>
>> diff --git a/lib/i915/gem_mman.c b/lib/i915/gem_mman.c
>> index c432bb16..563a7ccf 100644
>> --- a/lib/i915/gem_mman.c
>> +++ b/lib/i915/gem_mman.c
>> @@ -474,8 +474,14 @@ void *gem_mmap__cpu(int fd, uint32_t handle, uint64_t offset, uint64_t size, uns
>>  void *__gem_mmap_offset__cpu(int fd, uint32_t handle, uint64_t offset,
>> 			     uint64_t size, unsigned prot)
>>  {
>> -	return __gem_mmap_offset(fd, handle, offset, size, prot,
>> +	void *ptr;
>> +
>> +	ptr = __gem_mmap_offset(fd, handle, offset, size, prot,
>> 				 I915_MMAP_OFFSET_WB);
>> +	if (!ptr)
>> +		ptr = __gem_mmap_offset__fixed(fd, handle, offset, size, prot);
>> +
>> +	return ptr;
> Imo there's some asymmetry here. If we are adding fixed mode to
> mmap__device_coherent (in the previous patch) then we should also be adding
> it to mmap__cpu_coherent (as before). Or, if we are adding fixed mode to
> __gem_mmap_offset__cpu we should also be adding it to
> __gem_mmap_offset__wc. Thanks.

Why do we need it in __gem_mmap_offset_wc btw?

Rest of series doesn't seem to be blocked by it, so pushed for now.