On 04/12/2016 06:39 PM, Felix Kuehling wrote:
> This is the implementation of ttm_round_pot:
>
> size_t ttm_round_pot(size_t size)
> {
> if ((size & (size - 1)) == 0)
> return size;
> else if (size > PAGE_SIZE)
> return PAGE_ALIGN(size);
> els
On 04/12/2016 05:57 PM, Alex Deucher wrote:
> From: Felix Kuehling
>
> TTM BO accounting is out of sync with how memory is really allocated
> in ttm[_dma]_tt_alloc_page_directory. This resulted in excessive
> estimated overhead with many small allocations.
>
> ttm_dma_tt_alloc_page_directory makes
This is the implementation of ttm_round_pot:
size_t ttm_round_pot(size_t size)
{
if ((size & (size - 1)) == 0)
return size;
else if (size > PAGE_SIZE)
return PAGE_ALIGN(size);
else {
size_t tmp_size = 4;
while
From: Felix Kuehling
TTM BO accounting is out of sync with how memory is really allocated
in ttm[_dma]_tt_alloc_page_directory. This resulted in excessive
estimated overhead with many small allocations.
ttm_dma_tt_alloc_page_directory makes a single allocation for three
arrays: pages, DMA and CP
