On 29/03/17 01:41 PM, Christian König wrote:
> Am 29.03.2017 um 19:07 schrieb Tom St Denis:
>> On 29/03/17 01:07 PM, Tom St Denis wrote:
>>> On gfx9 hardware the value is not wrapped and is a 64-bit value. So
>>> we reduce it modulo the ring size.
>>>
>>> Signed-off-by: Tom St Denis <tom.stdenis at amd.com>
>>> ---
>>> drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c | 4 ++--
>>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
>>> index ac0ce3f27f87..b6d2c0b2a501 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
>>> @@ -320,8 +320,8 @@ static ssize_t amdgpu_debugfs_ring_read(struct
>>> file *f, char __user *buf,
>>>
>>> if (*pos < 12) {
>>> early[0] = amdgpu_ring_get_rptr(ring);
>>> - early[1] = amdgpu_ring_get_wptr(ring);
>>> - early[2] = ring->wptr;
>>> + early[1] = amdgpu_ring_get_wptr(ring) & ((ring->ring_size >>
>>> 2) - 1);
>>> + early[2] = ring->wptr & ((ring->ring_size >> 2) - 1);
>>> for (i = *pos / 4; i < 3 && size; i++) {
>>> r = put_user(early[i], (uint32_t *)buf);
>>> if (r)
>>>
>>
>> Particularly if reviewers could make sure I got the dword/byte
>> conversion correct here. It's my understanding ring_size is bytes and
>> the get_*ptr() macros is in dwords right?
>
> Actually you should use ptr_mask here instead, this way you don't need
> to mess with the ring size at all.
ring->buf_mask = (ring->ring_size / 4) - 1;
ring->ptr_mask = ring->funcs->support_64bit_ptrs ?
0xffffffffffffffff : ring->buf_mask;
This won't wrap the value like debugfs needs though.
Tom
