On Fri, 2 Aug 2024 10:37:38 +0100 Catalin Marinas <catalin.marinas@xxxxxxx> wrote: > On Fri, Aug 02, 2024 at 09:03:47AM +0300, Baruch Siach wrote: > > diff --git a/kernel/dma/direct.c b/kernel/dma/direct.c > > index 3b4be4ca3b08..62b36fda44c9 100644 > > --- a/kernel/dma/direct.c > > +++ b/kernel/dma/direct.c > > @@ -20,7 +20,7 @@ > > * it for entirely different regions. In that case the arch code needs to > > * override the variable below for dma-direct to work properly. > > */ > > -unsigned int zone_dma_bits __ro_after_init = 24; > > +u64 zone_dma_limit __ro_after_init = DMA_BIT_MASK(24); > > u64 here makes sense even if it may be larger than phys_addr_t. It > matches the phys_limit type in the swiotlb code. The compilers should no > longer complain. FTR I have never quite understood why phys_limit is u64, but u64 was already used all around the place when I first looked into swiotlb. > > diff --git a/kernel/dma/pool.c b/kernel/dma/pool.c > > index d10613eb0f63..7b04f7575796 100644 > > --- a/kernel/dma/pool.c > > +++ b/kernel/dma/pool.c > > @@ -70,9 +70,9 @@ static bool cma_in_zone(gfp_t gfp) > > /* CMA can't cross zone boundaries, see cma_activate_area() */ > > end = cma_get_base(cma) + size - 1; > > if (IS_ENABLED(CONFIG_ZONE_DMA) && (gfp & GFP_DMA)) > > - return end <= DMA_BIT_MASK(zone_dma_bits); > > + return end <= zone_dma_limit; > > if (IS_ENABLED(CONFIG_ZONE_DMA32) && (gfp & GFP_DMA32)) > > - return end <= DMA_BIT_MASK(32); > > + return end <= max(DMA_BIT_MASK(32), zone_dma_limit); > > return true; > > } > > > > diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c > > index 043b0ecd3e8d..bb51bd5335ad 100644 > > --- a/kernel/dma/swiotlb.c > > +++ b/kernel/dma/swiotlb.c > > @@ -450,9 +450,9 @@ int swiotlb_init_late(size_t size, gfp_t gfp_mask, > > if (!remap) > > io_tlb_default_mem.can_grow = true; > > if (IS_ENABLED(CONFIG_ZONE_DMA) && (gfp_mask & __GFP_DMA)) > > - io_tlb_default_mem.phys_limit = DMA_BIT_MASK(zone_dma_bits); > > + io_tlb_default_mem.phys_limit = zone_dma_limit; > > else if (IS_ENABLED(CONFIG_ZONE_DMA32) && (gfp_mask & __GFP_DMA32)) > > - io_tlb_default_mem.phys_limit = DMA_BIT_MASK(32); > > + io_tlb_default_mem.phys_limit = max(DMA_BIT_MASK(32), zone_dma_limit); > > else > > io_tlb_default_mem.phys_limit = virt_to_phys(high_memory - 1); > > #endif > > These two look correct to me now and it's the least intrusive (the > alternative would have been a zone_dma32_limit). The arch code, however, > needs to ensure that zone_dma_limit can always support 32-bit devices > even if it is above 4GB (with the relevant dma offsets in place for such > devices). Just to make sure, the DMA zone (if present) must map to at most 32-bit bus address space (possibly behind a bridge). Is that what you're saying? Petr T
