DHD: 4398 - Do not allocate coherent memory using GFP_ATOMIC
DHD init was failing due to dhd_dma_buf_alloc trying to allocate
coherent memory for DMA using GFP_ATOMIC flag.
This was due to problems on older customer platforms where
allocating coherent memory using GFP_KERNEL was leading to
cache coherence problems.
Google confirmed that on their platforms this problem should
not be there.
Hence disabling compile time macro DHD_ALLOC_COHERENT_MEM_FROM_ATOMIC_POOL
Bug: 300578859
Change-Id: If7a9811867acc940d5e2794e8d10451e9627b55d
Signed-off-by: Yadukishore Kodicherla <yadukishore.kodicherla@broadcom.corp-partner.google.com>
1 file changed