This adds the description of the following eight function:
dma_sync_single_for_cpu
pci_dma_sync_single_for_cpuvoid
dma_sync_single_for_device
pci_dma_sync_single_for_device
dma_sync_sg_for_cpu
pci_dma_sync_sg_for_cpu
dma_sync_sg_for_device
pci_dma_sync_sg_for_device
It was unclear that the API permits a partial sync (some network
drivers already do though). I made it clear that the sync_single API
can do a partial sync but the sync_sg API can't.
We could do a partial sync with the sync_sg too, however, it's
difficult for driver writers to correctly use the sync_sg API for a
partial sync since the scatterlists passed in to the mapping API can
be modified. It's unlikely that driver writers want to do a partial
sync with the sync_sg API (because the sync_sg API are usually used
for block drivers). So I think that it's better to forbid a partial
sync with the sync_sg API.
Signed-off-by: FUJITA Tomonori <[email protected]>
Cc: David Miller <[email protected]>
Cc: James Bottomley <[email protected]>
---
Documentation/DMA-API.txt | 42 ++++++++++++++++++++++++++++++++++++++++++
1 files changed, 42 insertions(+), 0 deletions(-)
diff --git a/Documentation/DMA-API.txt b/Documentation/DMA-API.txt
index bb0e75e..c8db3d0 100644
--- a/Documentation/DMA-API.txt
+++ b/Documentation/DMA-API.txt
@@ -364,6 +364,48 @@ API.
Note: <nents> must be the number you passed in, *not* the number of
physical entries returned.
+void
+dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle, size_t size,
+ enum dma_data_direction direction)
+void
+pci_dma_sync_single_for_cpu(struct pci_dev *hwdev, dma_addr_t dma_handle,
+ size_t size, int direction)
+void
+dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle, size_t size,
+ enum dma_data_direction direction)
+void
+pci_dma_sync_single_for_device(struct pci_dev *hwdev, dma_addr_t dma_handle,
+ size_t size, int direction)
+void
+dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg, int nelems,
+ enum dma_data_direction direction)
+void
+pci_dma_sync_sg_for_cpu(struct pci_dev *hwdev, struct scatterlist *sg,
+ int nelems, int direction)
+void
+dma_sync_sg_for_device(struct device *dev, struct scatterlist *sg, int nelems,
+ enum dma_data_direction direction)
+void
+pci_dma_sync_sg_for_device(struct pci_dev *hwdev, struct scatterlist *sg,
+ int nelems, int direction)
+
+Synchronise a single contiguous or scatter/gather mapping for the cpu
+and device. With the sync_sg API, all the parameters must be the same
+as those passed into the single mapping API. With the sync_single API,
+you can use dma_handle and size parameters that aren't identical to
+those passed into the single mapping API to do a partial sync.
+
+Notes: You must do this:
+
+- Before reading values that have been written by DMA from the device
+ (use the DMA_FROM_DEVICE direction)
+- After writing values that will be written to the device using DMA
+ (use the DMA_TO_DEVICE) direction
+- before *and* after handing memory to the device if the memory is
+ DMA_BIDIRECTIONAL
+
+See also dma_map_single().
+
dma_addr_t
dma_map_single_attrs(struct device *dev, void *cpu_addr, size_t size,
enum dma_data_direction dir,
--
1.5.6.5
dma_sync_single_for_cpu/for_device supports a partial sync so there is
no point to have dma_sync_single_range (also dma_sync_single was
obsoleted long ago, replaced with dma_sync_single_for_cpu/for_device).
There is no user of dma_sync_single_range() in mainline and only Alpha
architecture supports dma_sync_single_range(). So it's unlikely that
someone out of the tree uses it.
Signed-off-by: FUJITA Tomonori <[email protected]>
Cc: Jeff Dike <[email protected]>
---
arch/um/include/asm/dma-mapping.h | 8 --------
1 files changed, 0 insertions(+), 8 deletions(-)
diff --git a/arch/um/include/asm/dma-mapping.h b/arch/um/include/asm/dma-mapping.h
index 378de4b..b948c14 100644
--- a/arch/um/include/asm/dma-mapping.h
+++ b/arch/um/include/asm/dma-mapping.h
@@ -104,14 +104,6 @@ dma_get_cache_alignment(void)
}
static inline void
-dma_sync_single_range(struct device *dev, dma_addr_t dma_handle,
- unsigned long offset, size_t size,
- enum dma_data_direction direction)
-{
- BUG();
-}
-
-static inline void
dma_cache_sync(struct device *dev, void *vaddr, size_t size,
enum dma_data_direction direction)
{
--
1.5.6.5
dma_sync_single_for_cpu/for_device supports a partial sync so there is
no point to have dma_sync_single_range (also dma_sync_single was
obsoleted long ago, replaced with dma_sync_single_for_cpu/for_device).
There is no user of dma_sync_single_range() in mainline and only Alpha
architecture supports dma_sync_single_range(). So it's unlikely that
someone out of the tree uses it.
Signed-off-by: FUJITA Tomonori <[email protected]>
Cc: David Miller <[email protected]>
Cc: James Bottomley <[email protected]>
---
Documentation/DMA-API.txt | 10 ----------
1 files changed, 0 insertions(+), 10 deletions(-)
diff --git a/Documentation/DMA-API.txt b/Documentation/DMA-API.txt
index c8db3d0..364a6cb 100644
--- a/Documentation/DMA-API.txt
+++ b/Documentation/DMA-API.txt
@@ -528,16 +528,6 @@ into the width returned by this call. It will also always be a power
of two for easy alignment.
void
-dma_sync_single_range(struct device *dev, dma_addr_t dma_handle,
- unsigned long offset, size_t size,
- enum dma_data_direction direction)
-
-Does a partial sync, starting at offset and continuing for size. You
-must be careful to observe the cache alignment and width when doing
-anything like this. You must also be extra careful about accessing
-memory you intend to sync partially.
-
-void
dma_cache_sync(struct device *dev, void *vaddr, size_t size,
enum dma_data_direction direction)
--
1.5.6.5
dma_sync_single_for_cpu/for_device supports a partial sync so there is
no point to have dma_sync_single_range (also dma_sync_single was
obsoleted long ago, replaced with dma_sync_single_for_cpu/for_device).
There is no user of dma_sync_single_range() in mainline and only Alpha
architecture supports dma_sync_single_range(). So it's unlikely that
someone out of the tree uses it.
Signed-off-by: FUJITA Tomonori <[email protected]>
Cc: Richard Henderson <[email protected]>
Cc: Ivan Kokshaysky <[email protected]>
Cc: Matt Turner <[email protected]>
---
arch/alpha/include/asm/dma-mapping.h | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/arch/alpha/include/asm/dma-mapping.h b/arch/alpha/include/asm/dma-mapping.h
index 04eb568..f514ede 100644
--- a/arch/alpha/include/asm/dma-mapping.h
+++ b/arch/alpha/include/asm/dma-mapping.h
@@ -59,7 +59,6 @@ int dma_set_mask(struct device *dev, u64 mask);
#define dma_sync_single_for_cpu(dev, addr, size, dir) ((void)0)
#define dma_sync_single_for_device(dev, addr, size, dir) ((void)0)
-#define dma_sync_single_range(dev, addr, off, size, dir) ((void)0)
#define dma_sync_sg_for_cpu(dev, sg, nents, dir) ((void)0)
#define dma_sync_sg_for_device(dev, sg, nents, dir) ((void)0)
#define dma_cache_sync(dev, va, size, dir) ((void)0)
--
1.5.6.5
From: FUJITA Tomonori <[email protected]>
Date: Thu, 4 Feb 2010 12:13:22 +0900
> This adds the description of the following eight function:
>
> dma_sync_single_for_cpu
> pci_dma_sync_single_for_cpuvoid
^^^^
That "void" is probably some cut and paste error :-)
> dma_sync_single_for_device
> pci_dma_sync_single_for_device
> dma_sync_sg_for_cpu
> pci_dma_sync_sg_for_cpu
> dma_sync_sg_for_device
> pci_dma_sync_sg_for_device
>
> It was unclear that the API permits a partial sync (some network
> drivers already do though). I made it clear that the sync_single API
> can do a partial sync but the sync_sg API can't.
>
> We could do a partial sync with the sync_sg too, however, it's
> difficult for driver writers to correctly use the sync_sg API for a
> partial sync since the scatterlists passed in to the mapping API can
> be modified. It's unlikely that driver writers want to do a partial
> sync with the sync_sg API (because the sync_sg API are usually used
> for block drivers). So I think that it's better to forbid a partial
> sync with the sync_sg API.
>
> Signed-off-by: FUJITA Tomonori <[email protected]>
Acked-by: David S. Miller <[email protected]>
From: FUJITA Tomonori <[email protected]>
Date: Thu, 4 Feb 2010 12:13:23 +0900
> dma_sync_single_for_cpu/for_device supports a partial sync so there is
> no point to have dma_sync_single_range (also dma_sync_single was
> obsoleted long ago, replaced with dma_sync_single_for_cpu/for_device).
>
> There is no user of dma_sync_single_range() in mainline and only Alpha
> architecture supports dma_sync_single_range(). So it's unlikely that
> someone out of the tree uses it.
>
> Signed-off-by: FUJITA Tomonori <[email protected]>
Acked-by: David S. Miller <[email protected]>
From: FUJITA Tomonori <[email protected]>
Date: Thu, 4 Feb 2010 12:13:24 +0900
> dma_sync_single_for_cpu/for_device supports a partial sync so there is
> no point to have dma_sync_single_range (also dma_sync_single was
> obsoleted long ago, replaced with dma_sync_single_for_cpu/for_device).
>
> There is no user of dma_sync_single_range() in mainline and only Alpha
> architecture supports dma_sync_single_range(). So it's unlikely that
> someone out of the tree uses it.
>
> Signed-off-by: FUJITA Tomonori <[email protected]>
Acked-by: David S. Miller <[email protected]>
From: FUJITA Tomonori <[email protected]>
Date: Thu, 4 Feb 2010 12:13:25 +0900
> dma_sync_single_for_cpu/for_device supports a partial sync so there is
> no point to have dma_sync_single_range (also dma_sync_single was
> obsoleted long ago, replaced with dma_sync_single_for_cpu/for_device).
>
> There is no user of dma_sync_single_range() in mainline and only Alpha
> architecture supports dma_sync_single_range(). So it's unlikely that
> someone out of the tree uses it.
>
> Signed-off-by: FUJITA Tomonori <[email protected]>
Acked-by: David S. Miller <[email protected]>
On Wed, 03 Feb 2010 19:21:23 -0800 (PST)
David Miller <[email protected]> wrote:
> From: FUJITA Tomonori <[email protected]>
> Date: Thu, 4 Feb 2010 12:13:22 +0900
>
> > This adds the description of the following eight function:
> >
> > dma_sync_single_for_cpu
> > pci_dma_sync_single_for_cpuvoid
> ^^^^
>
> That "void" is probably some cut and paste error :-)
Oops.
> > dma_sync_single_for_device
> > pci_dma_sync_single_for_device
> > dma_sync_sg_for_cpu
> > pci_dma_sync_sg_for_cpu
> > dma_sync_sg_for_device
> > pci_dma_sync_sg_for_device
> >
> > It was unclear that the API permits a partial sync (some network
> > drivers already do though). I made it clear that the sync_single API
> > can do a partial sync but the sync_sg API can't.
> >
> > We could do a partial sync with the sync_sg too, however, it's
> > difficult for driver writers to correctly use the sync_sg API for a
> > partial sync since the scatterlists passed in to the mapping API can
> > be modified. It's unlikely that driver writers want to do a partial
> > sync with the sync_sg API (because the sync_sg API are usually used
> > for block drivers). So I think that it's better to forbid a partial
> > sync with the sync_sg API.
> >
> > Signed-off-by: FUJITA Tomonori <[email protected]>
>
> Acked-by: David S. Miller <[email protected]>
Thanks,
I found other two typos in the commit log so here's the update.
=
From: FUJITA Tomonori <[email protected]>
Subject: [PATCH -mm 1/4] DMA-API.txt: add dma_sync_single/sg API description
This adds the description of the following eight function:
dma_sync_single_for_cpu
pci_dma_sync_single_for_cpu
dma_sync_single_for_device
pci_dma_sync_single_for_device
dma_sync_sg_for_cpu
pci_dma_sync_sg_for_cpu
dma_sync_sg_for_device
pci_dma_sync_sg_for_device
It was unclear that the API permits a partial sync (some network
drivers already do though). I made it clear that the sync_single API
can do a partial sync but the sync_sg API can't.
We could do a partial sync with the sync_sg API too, however, it's
difficult for driver writers to correctly use the sync_sg API for a
partial sync since the scatterlists passed in to the mapping API can't
be modified. It's unlikely that driver writers want to do a partial
sync with the sync_sg API (because the sync_sg API are usually used
for block drivers). So I think that it's better to forbid a partial
sync with the sync_sg API.
Signed-off-by: FUJITA Tomonori <[email protected]>
Acked-by: David S. Miller <[email protected]>
Cc: James Bottomley <[email protected]>
On Thu, 2010-02-04 at 12:31 +0900, FUJITA Tomonori wrote:
> On Wed, 03 Feb 2010 19:21:23 -0800 (PST)
> David Miller <[email protected]> wrote:
>
> > From: FUJITA Tomonori <[email protected]>
> > Date: Thu, 4 Feb 2010 12:13:22 +0900
> >
> > > This adds the description of the following eight function:
> > >
> > > dma_sync_single_for_cpu
> > > pci_dma_sync_single_for_cpuvoid
> > ^^^^
> >
> > That "void" is probably some cut and paste error :-)
>
> Oops.
>
>
> > > dma_sync_single_for_device
> > > pci_dma_sync_single_for_device
> > > dma_sync_sg_for_cpu
> > > pci_dma_sync_sg_for_cpu
> > > dma_sync_sg_for_device
> > > pci_dma_sync_sg_for_device
> > >
> > > It was unclear that the API permits a partial sync (some network
> > > drivers already do though). I made it clear that the sync_single API
> > > can do a partial sync but the sync_sg API can't.
> > >
> > > We could do a partial sync with the sync_sg too, however, it's
> > > difficult for driver writers to correctly use the sync_sg API for a
> > > partial sync since the scatterlists passed in to the mapping API can
> > > be modified. It's unlikely that driver writers want to do a partial
> > > sync with the sync_sg API (because the sync_sg API are usually used
> > > for block drivers). So I think that it's better to forbid a partial
> > > sync with the sync_sg API.
> > >
> > > Signed-off-by: FUJITA Tomonori <[email protected]>
> >
> > Acked-by: David S. Miller <[email protected]>
>
> Thanks,
>
> I found other two typos in the commit log so here's the update.
>
> =
> From: FUJITA Tomonori <[email protected]>
> Subject: [PATCH -mm 1/4] DMA-API.txt: add dma_sync_single/sg API description
>
> This adds the description of the following eight function:
>
> dma_sync_single_for_cpu
> pci_dma_sync_single_for_cpu
> dma_sync_single_for_device
> pci_dma_sync_single_for_device
> dma_sync_sg_for_cpu
> pci_dma_sync_sg_for_cpu
> dma_sync_sg_for_device
> pci_dma_sync_sg_for_device
>
> It was unclear that the API permits a partial sync (some network
> drivers already do though). I made it clear that the sync_single API
> can do a partial sync but the sync_sg API can't.
>
> We could do a partial sync with the sync_sg API too, however, it's
> difficult for driver writers to correctly use the sync_sg API for a
> partial sync since the scatterlists passed in to the mapping API can't
> be modified. It's unlikely that driver writers want to do a partial
> sync with the sync_sg API (because the sync_sg API are usually used
> for block drivers). So I think that it's better to forbid a partial
> sync with the sync_sg API.
>
> Signed-off-by: FUJITA Tomonori <[email protected]>
> Acked-by: David S. Miller <[email protected]>
> Cc: James Bottomley <[email protected]>
For what it's worth, you can add my acked by to both patches too.
James
On Wed, Feb 3, 2010 at 10:13 PM, FUJITA Tomonori
<[email protected]> wrote:
> dma_sync_single_for_cpu/for_device supports a partial sync so there is
> no point to have dma_sync_single_range (also dma_sync_single was
> obsoleted long ago, replaced with dma_sync_single_for_cpu/for_device).
>
> There is no user of dma_sync_single_range() in mainline and only Alpha
> architecture supports dma_sync_single_range(). So it's unlikely that
> someone out of the tree uses it.
>
> Signed-off-by: FUJITA Tomonori <[email protected]>
> Cc: Richard Henderson <[email protected]>
> Cc: Ivan Kokshaysky <[email protected]>
> Cc: Matt Turner <[email protected]>
> ---
> ?arch/alpha/include/asm/dma-mapping.h | ? ?1 -
> ?1 files changed, 0 insertions(+), 1 deletions(-)
>
> diff --git a/arch/alpha/include/asm/dma-mapping.h b/arch/alpha/include/asm/dma-mapping.h
> index 04eb568..f514ede 100644
> --- a/arch/alpha/include/asm/dma-mapping.h
> +++ b/arch/alpha/include/asm/dma-mapping.h
> @@ -59,7 +59,6 @@ int dma_set_mask(struct device *dev, u64 mask);
>
> ?#define dma_sync_single_for_cpu(dev, addr, size, dir) ? ?((void)0)
> ?#define dma_sync_single_for_device(dev, addr, size, dir) ?((void)0)
> -#define dma_sync_single_range(dev, addr, off, size, dir) ?((void)0)
> ?#define dma_sync_sg_for_cpu(dev, sg, nents, dir) ? ? ? ? ((void)0)
> ?#define dma_sync_sg_for_device(dev, sg, nents, dir) ? ? ?((void)0)
> ?#define dma_cache_sync(dev, va, size, dir) ? ? ? ? ? ? ? ((void)0)
> --
> 1.5.6.5
Acked-by: Matt Turner <[email protected]>
Should I vacuum this one up, or should it go in the -mm tree? I don't
guess there's much point in sending Linus pull requests for single
patches.
Matt
On Wed, 3 Feb 2010 23:06:37 -0500 Matt Turner <[email protected]> wrote:
> On Wed, Feb 3, 2010 at 10:13 PM, FUJITA Tomonori
> <[email protected]> wrote:
> > dma_sync_single_for_cpu/for_device supports a partial sync so there is
> > no point to have dma_sync_single_range (also dma_sync_single was
> > obsoleted long ago, replaced with dma_sync_single_for_cpu/for_device).
> >
> > There is no user of dma_sync_single_range() in mainline and only Alpha
> > architecture supports dma_sync_single_range(). So it's unlikely that
> > someone out of the tree uses it.
> >
> > Signed-off-by: FUJITA Tomonori <[email protected]>
> > Cc: Richard Henderson <[email protected]>
> > Cc: Ivan Kokshaysky <[email protected]>
> > Cc: Matt Turner <[email protected]>
> > ---
> > ?arch/alpha/include/asm/dma-mapping.h | ? ?1 -
> > ?1 files changed, 0 insertions(+), 1 deletions(-)
> >
> > diff --git a/arch/alpha/include/asm/dma-mapping.h b/arch/alpha/include/asm/dma-mapping.h
> > index 04eb568..f514ede 100644
> > --- a/arch/alpha/include/asm/dma-mapping.h
> > +++ b/arch/alpha/include/asm/dma-mapping.h
> > @@ -59,7 +59,6 @@ int dma_set_mask(struct device *dev, u64 mask);
> >
> > ?#define dma_sync_single_for_cpu(dev, addr, size, dir) ? ?((void)0)
> > ?#define dma_sync_single_for_device(dev, addr, size, dir) ?((void)0)
> > -#define dma_sync_single_range(dev, addr, off, size, dir) ?((void)0)
> > ?#define dma_sync_sg_for_cpu(dev, sg, nents, dir) ? ? ? ? ((void)0)
> > ?#define dma_sync_sg_for_device(dev, sg, nents, dir) ? ? ?((void)0)
> > ?#define dma_cache_sync(dev, va, size, dir) ? ? ? ? ? ? ? ((void)0)
> > --
> > 1.5.6.5
>
> Acked-by: Matt Turner <[email protected]>
>
> Should I vacuum this one up, or should it go in the -mm tree? I don't
> guess there's much point in sending Linus pull requests for single
> patches.
>
I'll grab it.