On 6/25/19 08:59, Christoph Hellwig wrote:
On Tue, Jun 25, 2019 at 08:54:11AM +0200, John Paul Adrian Glaubitz wrote:Okay, thanks. I'll whip up a patch for Frank to test.The one below should do it, but from looking at the ia64 zone initialization I'm not sure this will be the culprit. diff --git a/kernel/dma/direct.c b/kernel/dma/direct.c index 2c2772e9702a..3e802f4580b3 100644 --- a/kernel/dma/direct.c +++ b/kernel/dma/direct.c @@ -82,9 +82,7 @@ static gfp_t __dma_direct_optimal_gfp_mask(struct device *dev, u64 dma_mask, */ if (*phys_mask <= DMA_BIT_MASK(ARCH_ZONE_DMA_BITS)) return GFP_DMA; - if (*phys_mask <= DMA_BIT_MASK(32)) - return GFP_DMA32; - return 0; + return GFP_DMA32; } static bool dma_coherent_ok(struct device *dev, phys_addr_t phys, size_t size)
Ok, will apply that to the most recent non-rc kernel source and give it a try. Should take about 45 mins or so. Cheers, Frank
|