diff options
author | Nicolas Saenz Julienne <nsaenzjulienne@suse.de> | 2019-11-21 10:26:44 +0100 |
---|---|---|
committer | Christoph Hellwig <hch@lst.de> | 2019-11-21 18:14:35 +0100 |
commit | a7ba70f1787f977f970cd116076c6fce4b9e01cc (patch) | |
tree | 474b2c0bc2201b3d2adde4c7887d4f76d50ac753 /kernel/dma | |
parent | d7293f79caea45c50c0ab4294847e7af96501ced (diff) | |
download | linux-a7ba70f1787f977f970cd116076c6fce4b9e01cc.tar.bz2 |
dma-mapping: treat dev->bus_dma_mask as a DMA limit
Using a mask to represent bus DMA constraints has a set of limitations.
The biggest one being it can only hold a power of two (minus one). The
DMA mapping code is already aware of this and treats dev->bus_dma_mask
as a limit. This quirk is already used by some architectures although
still rare.
With the introduction of the Raspberry Pi 4 we've found a new contender
for the use of bus DMA limits, as its PCIe bus can only address the
lower 3GB of memory (of a total of 4GB). This is impossible to represent
with a mask. To make things worse the device-tree code rounds non power
of two bus DMA limits to the next power of two, which is unacceptable in
this case.
In the light of this, rename dev->bus_dma_mask to dev->bus_dma_limit all
over the tree and treat it as such. Note that dev->bus_dma_limit should
contain the higher accessible DMA address.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Diffstat (limited to 'kernel/dma')
-rw-r--r-- | kernel/dma/direct.c | 27 |
1 files changed, 13 insertions, 14 deletions
diff --git a/kernel/dma/direct.c b/kernel/dma/direct.c index 267b23a13b69..6af7ae83c4ad 100644 --- a/kernel/dma/direct.c +++ b/kernel/dma/direct.c @@ -27,10 +27,10 @@ static void report_addr(struct device *dev, dma_addr_t dma_addr, size_t size) { if (!dev->dma_mask) { dev_err_once(dev, "DMA map on device without dma_mask\n"); - } else if (*dev->dma_mask >= DMA_BIT_MASK(32) || dev->bus_dma_mask) { + } else if (*dev->dma_mask >= DMA_BIT_MASK(32) || dev->bus_dma_limit) { dev_err_once(dev, - "overflow %pad+%zu of DMA mask %llx bus mask %llx\n", - &dma_addr, size, *dev->dma_mask, dev->bus_dma_mask); + "overflow %pad+%zu of DMA mask %llx bus limit %llx\n", + &dma_addr, size, *dev->dma_mask, dev->bus_dma_limit); } WARN_ON_ONCE(1); } @@ -57,15 +57,14 @@ u64 dma_direct_get_required_mask(struct device *dev) } static gfp_t __dma_direct_optimal_gfp_mask(struct device *dev, u64 dma_mask, - u64 *phys_mask) + u64 *phys_limit) { - if (dev->bus_dma_mask && dev->bus_dma_mask < dma_mask) - dma_mask = dev->bus_dma_mask; + u64 dma_limit = min_not_zero(dma_mask, dev->bus_dma_limit); if (force_dma_unencrypted(dev)) - *phys_mask = __dma_to_phys(dev, dma_mask); + *phys_limit = __dma_to_phys(dev, dma_limit); else - *phys_mask = dma_to_phys(dev, dma_mask); + *phys_limit = dma_to_phys(dev, dma_limit); /* * Optimistically try the zone that the physical address mask falls @@ -75,9 +74,9 @@ static gfp_t __dma_direct_optimal_gfp_mask(struct device *dev, u64 dma_mask, * Note that GFP_DMA32 and GFP_DMA are no ops without the corresponding * zones. */ - if (*phys_mask <= DMA_BIT_MASK(zone_dma_bits)) + if (*phys_limit <= DMA_BIT_MASK(zone_dma_bits)) return GFP_DMA; - if (*phys_mask <= DMA_BIT_MASK(32)) + if (*phys_limit <= DMA_BIT_MASK(32)) return GFP_DMA32; return 0; } @@ -85,7 +84,7 @@ static gfp_t __dma_direct_optimal_gfp_mask(struct device *dev, u64 dma_mask, static bool dma_coherent_ok(struct device *dev, phys_addr_t phys, size_t size) { return phys_to_dma_direct(dev, phys) + size - 1 <= - min_not_zero(dev->coherent_dma_mask, dev->bus_dma_mask); + min_not_zero(dev->coherent_dma_mask, dev->bus_dma_limit); } struct page *__dma_direct_alloc_pages(struct device *dev, size_t size, @@ -94,7 +93,7 @@ struct page *__dma_direct_alloc_pages(struct device *dev, size_t size, size_t alloc_size = PAGE_ALIGN(size); int node = dev_to_node(dev); struct page *page = NULL; - u64 phys_mask; + u64 phys_limit; if (attrs & DMA_ATTR_NO_WARN) gfp |= __GFP_NOWARN; @@ -102,7 +101,7 @@ struct page *__dma_direct_alloc_pages(struct device *dev, size_t size, /* we always manually zero the memory once we are done: */ gfp &= ~__GFP_ZERO; gfp |= __dma_direct_optimal_gfp_mask(dev, dev->coherent_dma_mask, - &phys_mask); + &phys_limit); page = dma_alloc_contiguous(dev, alloc_size, gfp); if (page && !dma_coherent_ok(dev, page_to_phys(page), size)) { dma_free_contiguous(dev, page, alloc_size); @@ -116,7 +115,7 @@ again: page = NULL; if (IS_ENABLED(CONFIG_ZONE_DMA32) && - phys_mask < DMA_BIT_MASK(64) && + phys_limit < DMA_BIT_MASK(64) && !(gfp & (GFP_DMA32 | GFP_DMA))) { gfp |= GFP_DMA32; goto again; |