2018-11-21 16:02:14

by Robin Murphy

[permalink] [raw]
Subject: [PATCH v2 0/2] SWIOTLB fixes for 4.20

Here's a quick v2 addressing Christoph's nits and collecting the tags
given so far. I've reproduced my USB problem with a commit prior to the
dma-mapping pull, so it looks like the cache maintenance changes are off
the hook for that one (and I'll have to venture into USB/UAS territory).

No objection from me if Konrad wants to pick this up instead depending
on timing.

Robin.


Robin Murphy (2):
dma-direct: Make DIRECT_MAPPING_ERROR viable for SWIOTLB
swiotlb: Skip cache maintenance on map error

include/linux/dma-direct.h | 2 +-
kernel/dma/swiotlb.c | 3 ++-
2 files changed, 3 insertions(+), 2 deletions(-)

--
2.19.1.dirty



2018-11-21 16:02:04

by Robin Murphy

[permalink] [raw]
Subject: [PATCH v2 2/2] swiotlb: Skip cache maintenance on map error

If swiotlb_bounce_page() failed, calling arch_sync_dma_for_device() may
lead to such delights as performing cache maintenance on whatever
address phys_to_virt(SWIOTLB_MAP_ERROR) looks like, which is typically
outside the kernel memory map and goes about as well as expected.

Don't do that.

Fixes: a4a4330db46a ("swiotlb: add support for non-coherent DMA")
Tested-by: John Stultz <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Signed-off-by: Robin Murphy <[email protected]>
---

v2: Collect tags

kernel/dma/swiotlb.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index 5731daa09a32..045930e32c0e 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -679,7 +679,8 @@ dma_addr_t swiotlb_map_page(struct device *dev, struct page *page,
}

if (!dev_is_dma_coherent(dev) &&
- (attrs & DMA_ATTR_SKIP_CPU_SYNC) == 0)
+ (attrs & DMA_ATTR_SKIP_CPU_SYNC) == 0 &&
+ dev_addr != DIRECT_MAPPING_ERROR)
arch_sync_dma_for_device(dev, phys, size, dir);

return dev_addr;
--
2.19.1.dirty


2018-11-21 18:32:46

by Robin Murphy

[permalink] [raw]
Subject: [PATCH v2 1/2] dma-direct: Make DIRECT_MAPPING_ERROR viable for SWIOTLB

With the overflow buffer removed, we no longer have a unique address
which is guaranteed not to be a valid DMA target to use as an error
token. The DIRECT_MAPPING_ERROR value of 0 tries to at least represent
an unlikely DMA target, but unfortunately there are already SWIOTLB
users with DMA-able memory at physical address 0 which now gets falsely
treated as a mapping failure and leads to all manner of misbehaviour.

The best we can do to mitigate that is flip DIRECT_MAPPING_ERROR to the
other commonly-used error value of all-bits-set, since the last single
byte of memory is by far the least-likely-valid DMA target.

Fixes: dff8d6c1ed58 ("swiotlb: remove the overflow buffer")
Reported-by: John Stultz <[email protected]>
Tested-by: John Stultz <[email protected]>
Acked-by: Konrad Rzeszutek Wilk <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Signed-off-by: Robin Murphy <[email protected]>
---

v2: Add parentheses, tweak commit message, collect tags

include/linux/dma-direct.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/linux/dma-direct.h b/include/linux/dma-direct.h
index bd73e7a91410..9e66bfe369aa 100644
--- a/include/linux/dma-direct.h
+++ b/include/linux/dma-direct.h
@@ -5,7 +5,7 @@
#include <linux/dma-mapping.h>
#include <linux/mem_encrypt.h>

-#define DIRECT_MAPPING_ERROR 0
+#define DIRECT_MAPPING_ERROR (~(dma_addr_t)0)

#ifdef CONFIG_ARCH_HAS_PHYS_TO_DMA
#include <asm/dma-direct.h>
--
2.19.1.dirty


2018-11-21 20:08:09

by Christoph Hellwig

[permalink] [raw]
Subject: Re: [PATCH v2 0/2] SWIOTLB fixes for 4.20

Thanks,

aplied to the dma-mapping for-linus tree.

I'll wait a little longer for a review from Stefano, but I really want
it in linux-next asap and hopefully off to Linus for the weekend.

2018-11-27 20:40:36

by Stefano Stabellini

[permalink] [raw]
Subject: Re: [PATCH v2 2/2] swiotlb: Skip cache maintenance on map error

On Wed, 21 Nov 2018, Robin Murphy wrote:
> If swiotlb_bounce_page() failed, calling arch_sync_dma_for_device() may
> lead to such delights as performing cache maintenance on whatever
> address phys_to_virt(SWIOTLB_MAP_ERROR) looks like, which is typically
> outside the kernel memory map and goes about as well as expected.
>
> Don't do that.
>
> Fixes: a4a4330db46a ("swiotlb: add support for non-coherent DMA")
> Tested-by: John Stultz <[email protected]>
> Reviewed-by: Christoph Hellwig <[email protected]>
> Signed-off-by: Robin Murphy <[email protected]>

Acked-by: Stefano Stabellini <[email protected]>

> ---
>
> v2: Collect tags
>
> kernel/dma/swiotlb.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
> index 5731daa09a32..045930e32c0e 100644
> --- a/kernel/dma/swiotlb.c
> +++ b/kernel/dma/swiotlb.c
> @@ -679,7 +679,8 @@ dma_addr_t swiotlb_map_page(struct device *dev, struct page *page,
> }
>
> if (!dev_is_dma_coherent(dev) &&
> - (attrs & DMA_ATTR_SKIP_CPU_SYNC) == 0)
> + (attrs & DMA_ATTR_SKIP_CPU_SYNC) == 0 &&
> + dev_addr != DIRECT_MAPPING_ERROR)
> arch_sync_dma_for_device(dev, phys, size, dir);
>
> return dev_addr;
> --
> 2.19.1.dirty
>