[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20181014181447.22783-1-hch@lst.de>
Date: Sun, 14 Oct 2018 20:14:47 +0200
From: Christoph Hellwig <hch@....de>
To: iommu@...ts.linux-foundation.org
Cc: linux-kernel@...r.kernel.org
Subject: [PATCH] dma-direct: reject highmem pages from dma_alloc_from_contiguous
dma_alloc_from_contiguous can return highmem pages depending on the
setup, which a plain non-remapping DMA allocator can't handle. Detect
this case and try the normal page allocator instead.
Signed-off-by: Christoph Hellwig <hch@....de>
---
kernel/dma/direct.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/kernel/dma/direct.c b/kernel/dma/direct.c
index 87a6bc2a96c0..46fbaa49125b 100644
--- a/kernel/dma/direct.c
+++ b/kernel/dma/direct.c
@@ -126,6 +126,18 @@ void *dma_direct_alloc_pages(struct device *dev, size_t size,
if (gfpflags_allow_blocking(gfp)) {
page = dma_alloc_from_contiguous(dev, count, page_order,
gfp & __GFP_NOWARN);
+ if (page && PageHighMem(page)) {
+ /*
+ * Depending on the cma= arguments and per-arch setup
+ * dma_alloc_from_contiguous could return highmem
+ * pages. Without remapping there is no way to return
+ * them here, so log an error and fail.
+ */
+ dev_info(dev, "Ignoring highmem page from CMA.\n");
+ dma_release_from_contiguous(dev, page, count);
+ page = NULL;
+ }
+
if (page && !dma_coherent_ok(dev, page_to_phys(page), size)) {
dma_release_from_contiguous(dev, page, count);
page = NULL;
--
2.19.1
Powered by blists - more mailing lists