ALSA fixes for non-coherent ppc32 again

ALSA fixes for non-coherent ppc32 again

[Gerhard Pircher]

Hi,

Takashi Iwai posted patches to make ALSA work on non-coherent PPC32
systems (almost exactly) a year ago. See here:
http://www.nabble.com/-PATCH-0-3--ALSA-fixes-for-non-coherent-ppc32-to17980027.html#a17980027

As far as I can see these patches never went upstream. Where there any
objections or did we just forget about them? It would be cool, if the
patches could be merged now, as at least two platforms need this bugfix
(namely Sam440 and AmigaOne).

Thanks!

Gerhard

-- 
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

Re: ALSA fixes for non-coherent ppc32 again

[Benjamin Herrenschmidt]

On Sun, 2009-06-21 at 20:18 +0200, Gerhard Pircher wrote:
> Hi,
> 
> Takashi Iwai posted patches to make ALSA work on non-coherent PPC32
> systems (almost exactly) a year ago. See here:
> http://www.nabble.com/-PATCH-0-3--ALSA-fixes-for-non-coherent-ppc32-to17980027.html#a17980027
> 
> As far as I can see these patches never went upstream. Where there any
> objections or did we just forget about them? It would be cool, if the
> patches could be merged now, as at least two platforms need this bugfix
> (namely Sam440 and AmigaOne).

I definitely forgot about those... But I'm fine with what Takashi did
for now, I can always make the powerpc helper for dma_mmap_coherent()
smarter later on if necessary.

Cheers,
Ben.

Re: ALSA fixes for non-coherent ppc32 again

[Takashi Iwai]

At Mon, 22 Jun 2009 08:34:38 +1000,
Benjamin Herrenschmidt wrote:
> 
> On Sun, 2009-06-21 at 20:18 +0200, Gerhard Pircher wrote:
> > Hi,
> > 
> > Takashi Iwai posted patches to make ALSA work on non-coherent PPC32
> > systems (almost exactly) a year ago. See here:
> > http://www.nabble.com/-PATCH-0-3--ALSA-fixes-for-non-coherent-ppc32-to17980027.html#a17980027
> > 
> > As far as I can see these patches never went upstream. Where there any
> > objections or did we just forget about them? It would be cool, if the
> > patches could be merged now, as at least two platforms need this bugfix
> > (namely Sam440 and AmigaOne).
> 
> I definitely forgot about those...

Me, too, almost... :)

> But I'm fine with what Takashi did
> for now, I can always make the powerpc helper for dma_mmap_coherent()
> smarter later on if necessary.

I updated the patch series for 2.6.31, including sparc32, parisc, mips
and sh support.  The patches are found in test/dma-fix branch of sound
git tree:
 git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git test/dma-fix

The old patches for 2.6.30 is found at history/dma-fix-2.6.30 branch
there, too.


For merging to the upstream, we'll need definitely discussions on
linux-arch ML or so, as once James Bottomley suggested.  I'll try to
make it up once after the merge window.

But, it'd be helpful if someone can test the patches above beforehand,
of course :)


thanks,

Takashi

Re: ALSA fixes for non-coherent ppc32 again

[Gerhard Pircher]

-------- Original-Nachricht --------
> Datum: Mon, 22 Jun 2009 09:12:35 +0200
> Von: Takashi Iwai <tiwai@suse.de>
> An: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> CC: Gerhard Pircher <gerhard_pircher@gmx.net>, linuxppc-dev@ozlabs.org
> Betreff: Re: ALSA fixes for non-coherent ppc32 again

> At Mon, 22 Jun 2009 08:34:38 +1000,
> Benjamin Herrenschmidt wrote:
> > 
> > On Sun, 2009-06-21 at 20:18 +0200, Gerhard Pircher wrote:
> > > Hi,
> > > 
> > > Takashi Iwai posted patches to make ALSA work on non-coherent PPC32
> > > systems (almost exactly) a year ago. See here:
> > >
> > > http://www.nabble.com/-PATCH-0-3--ALSA-fixes-for-non-coherent-ppc32-to17980027.html#a17980027
> > > 
> > > As far as I can see these patches never went upstream. Where there
> > > any objections or did we just forget about them? It would be cool,
> > > if the patches could be merged now, as at least two platforms need
> > > this bugfix (namely Sam440 and AmigaOne).
> > 
> > I definitely forgot about those...
> 
> Me, too, almost... :)
:)

> > But I'm fine with what Takashi did
> > for now, I can always make the powerpc helper for dma_mmap_coherent()
> > smarter later on if necessary.
> 
> I updated the patch series for 2.6.31, including sparc32, parisc, mips
> and sh support.  The patches are found in test/dma-fix branch of sound
> git tree:
>  git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git
> test/dma-fix
Thanks!

> The old patches for 2.6.30 is found at history/dma-fix-2.6.30 branch
> there, too.
> 
> 
> For merging to the upstream, we'll need definitely discussions on
> linux-arch ML or so, as once James Bottomley suggested.  I'll try to
> make it up once after the merge window.
> 
> But, it'd be helpful if someone can test the patches above beforehand,
> of course :)
Sure, I'll give it a try until tomorrow.

Gerhard

-- 
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

Re: ALSA fixes for non-coherent ppc32 again

[Gerhard Pircher]

-------- Original-Nachricht --------
> Datum: Mon, 22 Jun 2009 09:12:35 +0200
> Von: Takashi Iwai <tiwai@suse.de>
> An: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> CC: Gerhard Pircher <gerhard_pircher@gmx.net>, linuxppc-dev@ozlabs.org
> Betreff: Re: ALSA fixes for non-coherent ppc32 again

> But, it'd be helpful if someone can test the patches above beforehand,
> of course :)
Okay, I checked out your test/dma-fix branch and reformatted your
dma_mmap_coherent for powerpc patch ( http://www.nabble.com/-PATCH-0-3--ALSA-fixes-for-non-coherent-ppc32-to17980027.html#a17980027 ) to
adapt it for dma_mapping_ops (please take a look at the patch below).
I also had to change "def_bool n" to "def_bool y" for SND_NONCOHERENT_DMA
to actually enable it.

Unfortunately the build process stops with these error messages here
(but compiles fine, if SND_COHERENT_DMA is not selected):

  CC [M]  sound/core/memalloc.o
  CC [M]  sound/core/sgbuf.o
sound/core/sgbuf.c: In function ‘snd_free_sgbuf_pages’:
sound/core/sgbuf.c:46: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:47: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:48: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:50: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:51: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:52: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:56: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:57: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c: In function ‘snd_malloc_sgbuf_pages’:
sound/core/sgbuf.c:78: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:81: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:82: error: implicit declaration of function ‘snd_sgbuf_aligned_pages’
sound/core/sgbuf.c:83: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:84: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:84: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:87: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:88: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:91: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:103: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:107: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:112: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:113: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:115: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:116: error: increment of pointer to unknown structure
sound/core/sgbuf.c:116: error: arithmetic on pointer to an incomplete type
sound/core/sgbuf.c:121: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:127: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:128: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:128: error: dereferencing pointer to incomplete type
sound/core/sgbuf.c:132: error: dereferencing pointer to incomplete type

I also tried to compile it with the orginal dma_mmap_coherent for
powerpc patch, but that doesn't make a difference.

As the next step I applied the reformatted dma_mmap_coherent patch and
the following patches from your test/dma-fix branch to a 2.6.30-rc8
branch:
- ALSA: Remove old DMA-mmap code from arm/devdma.c
- ALSA: Fix SG-buffer DMA with non-coherent architectures
- ALSA: Fix mapping of DMA buffers

This one compiled fine, but ALSA didn't work. No kernel oops, just the
sound of silence. :)

Any idea what's wrong here or if I did something wrong?

Thanks!

Gerhard

---
 arch/powerpc/include/asm/dma-mapping.h |   14 ++++++++++++++
 arch/powerpc/kernel/dma.c              |   21 +++++++++++++++++++++
 2 files changed, 35 insertions(+), 0 deletions(-)

diff --git a/arch/powerpc/include/asm/dma-mapping.h b/arch/powerpc/include/asm/dma-mapping.h
index 3d9e887..6fbeafe 100644
--- a/arch/powerpc/include/asm/dma-mapping.h
+++ b/arch/powerpc/include/asm/dma-mapping.h
@@ -89,6 +89,8 @@ struct dma_mapping_ops {
                 struct dma_attrs *attrs);
     int        (*addr_needs_map)(struct device *dev, dma_addr_t addr,
                 size_t size);
+    int        (*mmap_coherent)(struct device *dev, struct vm_area_struct *vma,
+                void *cpu_addr, dma_addr_t handle, size_t size);
 #ifdef CONFIG_PPC_NEED_DMA_SYNC_OPS
     void            (*sync_single_range_for_cpu)(struct device *hwdev,
                 dma_addr_t dma_handle, unsigned long offset,
@@ -301,6 +303,18 @@ static inline void dma_unmap_sg(struct device *dev, struct scatterlist *sg,
     dma_unmap_sg_attrs(dev, sg, nhwentries, direction, NULL);
 }
 
+static inline int dma_mmap_coherent(struct device *dev,
+                    struct vm_area_struct *vma,
+                    void *cpu_addr, dma_addr_t handle,
+                    size_t size)
+{
+    struct dma_mapping_ops *dma_ops = get_dma_ops(dev);
+
+    BUG_ON(!dma_ops);
+
+    return dma_ops->mmap_coherent(dev, vma, cpu_addr, handle, size);
+}
+
 #ifdef CONFIG_PPC_NEED_DMA_SYNC_OPS
 static inline void dma_sync_single_for_cpu(struct device *dev,
         dma_addr_t dma_handle, size_t size,
diff --git a/arch/powerpc/kernel/dma.c b/arch/powerpc/kernel/dma.c
index 20a60d6..7ff3772 100644
--- a/arch/powerpc/kernel/dma.c
+++ b/arch/powerpc/kernel/dma.c
@@ -120,6 +120,26 @@ static inline void dma_direct_unmap_page(struct device *dev,
 {
 }
 
+/*
+ * A helper to mmap the pages allocated via dma_alloc_coherent()
+ */
+static inline int dma_direct_mmap_coherent(struct device *dev,
+                       struct vm_area_struct *vma,
+                       void *cpu_addr, dma_addr_t handle,
+                       size_t size)
+{
+    struct page *pg;
+
+#ifdef CONFIG_NOT_COHERENT_CACHE
+    /* I'm too lazy and can't stop using bus_to_virt() here... */
+    cpu_addr = bus_to_virt(handle);
+#endif
+    pg = virt_to_page(cpu_addr);
+    return remap_pfn_range(vma, vma->vm_start,
+                   page_to_pfn(pg) + vma->vm_pgoff,
+                   size, vma->vm_page_prot);
+}
+
 #ifdef CONFIG_NOT_COHERENT_CACHE
 static inline void dma_direct_sync_sg(struct device *dev,
         struct scatterlist *sgl, int nents,
@@ -148,6 +168,7 @@ struct dma_mapping_ops dma_direct_ops = {
     .dma_supported    = dma_direct_dma_supported,
     .map_page    = dma_direct_map_page,
     .unmap_page    = dma_direct_unmap_page,
+    .mmap_coherent    = dma_direct_mmap_coherent,
 #ifdef CONFIG_NOT_COHERENT_CACHE
     .sync_single_range_for_cpu     = dma_direct_sync_single_range,
     .sync_single_range_for_device     = dma_direct_sync_single_range,
-- 
1.5.6.5
-- 
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

Re: ALSA fixes for non-coherent ppc32 again

[Gerhard Pircher]

-------- Original-Nachricht --------
> Datum: Tue, 23 Jun 2009 10:55:54 +0200
> Von: "Gerhard Pircher" <gerhard_pircher@gmx.net>
> An: Takashi Iwai <tiwai@suse.de>, benh@kernel.crashing.org
> CC: linuxppc-dev@ozlabs.org
> Betreff: Re: ALSA fixes for non-coherent ppc32 again

> 
> -------- Original-Nachricht --------
> > Datum: Mon, 22 Jun 2009 09:12:35 +0200
> > Von: Takashi Iwai <tiwai@suse.de>
> > An: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> > CC: Gerhard Pircher <gerhard_pircher@gmx.net>, linuxppc-dev@ozlabs.org
> > Betreff: Re: ALSA fixes for non-coherent ppc32 again
> 
> > But, it'd be helpful if someone can test the patches above beforehand,
> > of course :)
> Okay, I checked out your test/dma-fix branch and reformatted your
> dma_mmap_coherent for powerpc patch (
> http://www.nabble.com/-PATCH-0-3--ALSA-fixes-for-non-coherent-ppc32-to17980027.html#a17980027 ) to
> adapt it for dma_mapping_ops (please take a look at the patch below).
> I also had to change "def_bool n" to "def_bool y" for SND_NONCOHERENT_DMA
> to actually enable it.
> 
> Unfortunately the build process stops with these error messages here
> (but compiles fine, if SND_COHERENT_DMA is not selected):
> 
>   CC [M]  sound/core/memalloc.o
>   CC [M]  sound/core/sgbuf.o
> sound/core/sgbuf.c: In function ‘snd_free_sgbuf_pages’:
> sound/core/sgbuf.c:46: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:47: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:48: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:50: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:51: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:52: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:56: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:57: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c: In function ‘snd_malloc_sgbuf_pages’:
> sound/core/sgbuf.c:78: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:81: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:82: error: implicit declaration of function
> ‘snd_sgbuf_aligned_pages’
> sound/core/sgbuf.c:83: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:84: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:84: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:87: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:88: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:91: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:103: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:107: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:112: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:113: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:115: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:116: error: increment of pointer to unknown structure
> sound/core/sgbuf.c:116: error: arithmetic on pointer to an incomplete type
> sound/core/sgbuf.c:121: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:127: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:128: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:128: error: dereferencing pointer to incomplete type
> sound/core/sgbuf.c:132: error: dereferencing pointer to incomplete type
> 
> I also tried to compile it with the orginal dma_mmap_coherent for
> powerpc patch, but that doesn't make a difference.
> 
> As the next step I applied the reformatted dma_mmap_coherent patch and
> the following patches from your test/dma-fix branch to a 2.6.30-rc8
> branch:
> - ALSA: Remove old DMA-mmap code from arm/devdma.c
> - ALSA: Fix SG-buffer DMA with non-coherent architectures
> - ALSA: Fix mapping of DMA buffers
> 
> This one compiled fine, but ALSA didn't work. No kernel oops, just the
> sound of silence. :)
Okay, that's wrong. I somehow messed up the .config file. It doesn't
compile, too.

Gerhard

-- 
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

Re: ALSA fixes for non-coherent ppc32 again

[Gerhard Pircher]

-------- Original-Nachricht --------
> Datum: Tue, 23 Jun 2009 23:42:24 +0200
> Von: "Gerhard Pircher" <gerhard_pircher@gmx.net>
> An: benh@kernel.crashing.org, tiwai@suse.de
> CC: linuxppc-dev@ozlabs.org
> Betreff: Re: ALSA fixes for non-coherent ppc32 again

> Okay, that's wrong. I somehow messed up the .config file. It doesn't
> compile, too.
I got it to compile now and it seems to work fine. I overlooked a typo
in sound/core/Makefile first (ifndef CONFIG_SND_NONCOHERNT_DMA.)
                                                        ^
I vote for a merge, if the config option and the typo is fixed. Ben,
should I resend a properly formatted patch for the dma_mmap_coherent
implementation or can we take the patch I already sent?

Thanks!

Gerhard

-- 
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

Re: ALSA fixes for non-coherent ppc32 again

[Takashi Iwai]

At Wed, 24 Jun 2009 10:46:01 +0200,
Gerhard Pircher wrote:
> 
> 
> -------- Original-Nachricht --------
> > Datum: Tue, 23 Jun 2009 23:42:24 +0200
> > Von: "Gerhard Pircher" <gerhard_pircher@gmx.net>
> > An: benh@kernel.crashing.org, tiwai@suse.de
> > CC: linuxppc-dev@ozlabs.org
> > Betreff: Re: ALSA fixes for non-coherent ppc32 again
> 
> > Okay, that's wrong. I somehow messed up the .config file. It doesn't
> > compile, too.
> I got it to compile now and it seems to work fine. I overlooked a typo
> in sound/core/Makefile first (ifndef CONFIG_SND_NONCOHERNT_DMA.)

Gah, thanks, fixed now on the git tree too.


Takashi

Re: ALSA fixes for non-coherent ppc32 again

[Gerhard Pircher]

-------- Original-Nachricht --------
> Datum: Wed, 24 Jun 2009 11:47:13 +0200
> Von: Takashi Iwai <tiwai@suse.de>
> An: "Gerhard Pircher" <gerhard_pircher@gmx.net>
> CC: benh@kernel.crashing.org, linuxppc-dev@ozlabs.org
> Betreff: Re: ALSA fixes for non-coherent ppc32 again

> At Wed, 24 Jun 2009 10:46:01 +0200,
> Gerhard Pircher wrote:
> > 
> > 
> > -------- Original-Nachricht --------
> > > Datum: Tue, 23 Jun 2009 23:42:24 +0200
> > > Von: "Gerhard Pircher" <gerhard_pircher@gmx.net>
> > > An: benh@kernel.crashing.org, tiwai@suse.de
> > > CC: linuxppc-dev@ozlabs.org
> > > Betreff: Re: ALSA fixes for non-coherent ppc32 again
> > 
> > > Okay, that's wrong. I somehow messed up the .config file. It doesn't
> > > compile, too.
> > I got it to compile now and it seems to work fine. I overlooked a typo
> > in sound/core/Makefile first (ifndef CONFIG_SND_NONCOHERNT_DMA.)
> 
> Gah, thanks, fixed now on the git tree too.

The Kconfig option still needs to be fixed, otherwise SND_COHERENT_DMA
isn't selected for my "PPC32 && NOT_COHERENT_CACHE" machine.

config SND_NONCOHERENT_DMA
        def_bool n  <-- should be "y"
        depends on (PPC32 && NOT_COHERENT_CACHE)...

BTW: Can you put me on CC: please, if you bring up this topic on
linux-arch or so?

Gerhard

-- 
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

Re: ALSA fixes for non-coherent ppc32 again

[Benjamin Herrenschmidt]

On Mon, 2009-06-22 at 08:34 +1000, Benjamin Herrenschmidt wrote:
> On Sun, 2009-06-21 at 20:18 +0200, Gerhard Pircher wrote:
> > Hi,
> > 
> > Takashi Iwai posted patches to make ALSA work on non-coherent PPC32
> > systems (almost exactly) a year ago. See here:
> > http://www.nabble.com/-PATCH-0-3--ALSA-fixes-for-non-coherent-ppc32-to17980027.html#a17980027
> > 
> > As far as I can see these patches never went upstream. Where there any
> > objections or did we just forget about them? It would be cool, if the
> > patches could be merged now, as at least two platforms need this bugfix
> > (namely Sam440 and AmigaOne).
> 
> I definitely forgot about those... But I'm fine with what Takashi did
> for now, I can always make the powerpc helper for dma_mmap_coherent()
> smarter later on if necessary.

BTW. Can you guys send a "final" patch for adding mmap_coherent to
powerpc ? Please make so that the dma_mmap_coherent() function doesn't
explose if dma_ops->mmap_coherent is NULL though (either fail gracefully
or fallback to some standard mmap).

Note that we probably need to add the virtual address too. IE. On
platforms with an iommu, we cannot easily go back from the dma_addr_t to
the memory address, it's easier to do that from the virtual address in
fact.

Cheers,
Ben.

Re: ALSA fixes for non-coherent ppc32 again

[Takashi Iwai]

At Wed, 08 Jul 2009 13:01:50 +1000,
Benjamin Herrenschmidt wrote:
>=20
> On Mon, 2009-06-22 at 08:34 +1000, Benjamin Herrenschmidt wrote:
> > On Sun, 2009-06-21 at 20:18 +0200, Gerhard Pircher wrote:
> > > Hi,
> > >=20
> > > Takashi Iwai posted patches to make ALSA work on non-coherent PPC32
> > > systems (almost exactly) a year ago. See here:
> > > http://www.nabble.com/-PATCH-0-3--ALSA-fixes-for-non-coherent-ppc32-t=
o17980027.html#a17980027
> > >=20
> > > As far as I can see these patches never went upstream. Where there any
> > > objections or did we just forget about them? It would be cool, if the
> > > patches could be merged now, as at least two platforms need this bugf=
ix
> > > (namely Sam440 and AmigaOne).
> >=20
> > I definitely forgot about those... But I'm fine with what Takashi did
> > for now, I can always make the powerpc helper for dma_mmap_coherent()
> > smarter later on if necessary.
>=20
> BTW. Can you guys send a "final" patch for adding mmap_coherent to
> powerpc ? Please make so that the dma_mmap_coherent() function doesn't
> explose if dma_ops->mmap_coherent is NULL though (either fail gracefully
> or fallback to some standard mmap).

The attached is the revised patch.  It falls backs to the standard
mmap.

The whole patch series are found in test/dma-fix branch of sound git
tree below:
  git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git test/dm=
a-fix

I had no time to play cross-build yet, so it'd be helpful if someone
can test it...
(BTW, any good minimal gcc and binutils package for kernel builds for
 PPC32?)

> Note that we probably need to add the virtual address too. IE. On
> platforms with an iommu, we cannot easily go back from the dma_addr_t to
> the memory address, it's easier to do that from the virtual address in
> fact.

Yes, the virtual address is passed as well.  Which parameter to be
used is just a matter of arch-specific implementation.


thanks,

Takashi

=3D=3D=3D
=46rom a8cb7ed04595785d902d3ad20ba2fe5ccbe31690 Mon Sep 17 00:00:00 2001
From: Takashi Iwai <tiwai@suse.de>
Date: Wed, 8 Jul 2009 11:35:56 +0200
Subject: [PATCH] powerpc: implement dma_mmap_coherent()

A lazy version of dma_mmap_coherent() implementation for powerpc.
The standard mmap is used as a fallback.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
---
 arch/powerpc/include/asm/dma-mapping.h |   22 ++++++++++++++++++++++
 arch/powerpc/kernel/dma.c              |   21 +++++++++++++++++++++
 2 files changed, 43 insertions(+), 0 deletions(-)

diff --git a/arch/powerpc/include/asm/dma-mapping.h b/arch/powerpc/include/=
asm/dma-mapping.h
index b44aaab..030a4b1 100644
--- a/arch/powerpc/include/asm/dma-mapping.h
+++ b/arch/powerpc/include/asm/dma-mapping.h
@@ -105,6 +105,10 @@ struct dma_mapping_ops {
 				struct scatterlist *sg, int nelems,
 				enum dma_data_direction direction);
 #endif
+	int		(*mmap_coherent)(struct device *hwdev,
+					 struct vm_area_struct *vma,
+					 void *cpu_addr, dma_addr_t dma_handle,
+					 size_t size);
 };
=20
 /*
@@ -415,6 +419,24 @@ static inline void dma_sync_single_range_for_device(st=
ruct device *dev,
 }
 #endif
=20
+#define ARCH_HAS_DMA_MMAP_COHERENT
+static inline int dma_mmap_coherent(struct device *dev,
+				    struct vm_area_struct *vma,
+				    void *cpu_addr, dma_addr_t dma_handle,
+				    size_t size)
+{
+	struct dma_mapping_ops *dma_ops =3D get_dma_ops(dev);
+
+	BUG_ON(!dma_ops);
+
+	if (dma_ops->mmap_coherent)
+		return dma_ops->mmap_coherent(dev, vma, cpu_addr, dma_handle,
+					      size);
+	else
+		return dma_direct_ops.mmap_coherent(dev, vma, cpu_addr,
+						    dma_handle, size);
+}
+
 static inline int dma_mapping_error(struct device *dev, dma_addr_t dma_add=
r)
 {
 #ifdef CONFIG_PPC64
diff --git a/arch/powerpc/kernel/dma.c b/arch/powerpc/kernel/dma.c
index 20a60d6..d11db99 100644
--- a/arch/powerpc/kernel/dma.c
+++ b/arch/powerpc/kernel/dma.c
@@ -140,6 +140,26 @@ static inline void dma_direct_sync_single_range(struct=
 device *dev,
 }
 #endif
=20
+static int dma_direct_mmap_coherent(struct device *dev,
+				    struct vm_area_struct *vma,
+				    void *cpu_addr, dma_addr_t dma_handle,
+				    size_t size)
+{
+	unsigned long pfn;
+#ifdef CONFIG_NOT_COHERENT_CACHE
+	dma_handle -=3D get_dma_direct_offset(dev);
+	/* assume dma_handle set via pfn_to_phys() in
+	 * mm/dma-noncoherent.c
+	 */
+	pfn =3D dma_handle >> PAGE_SHIFT;
+#else
+	pfn =3D page_to_pfn(virt_to_page(cpu_addr));
+#endif
+	vma->vm_page_prot =3D pgprot_noncached(vma->vm_page_prot);
+	return remap_pfn_range(vma, vma->vm_start, pfn + vma->vm_pgoff,
+			       size, vma->vm_page_prot);
+}
+
 struct dma_mapping_ops dma_direct_ops =3D {
 	.alloc_coherent	=3D dma_direct_alloc_coherent,
 	.free_coherent	=3D dma_direct_free_coherent,
@@ -154,5 +174,6 @@ struct dma_mapping_ops dma_direct_ops =3D {
 	.sync_sg_for_cpu 		=3D dma_direct_sync_sg,
 	.sync_sg_for_device 		=3D dma_direct_sync_sg,
 #endif
+	.mmap_coherent	=3D dma_direct_mmap_coherent,
 };
 EXPORT_SYMBOL(dma_direct_ops);
--=20
1.6.3.2

Re: ALSA fixes for non-coherent ppc32 again

[Takashi Iwai]

At Fri, 26 Jun 2009 15:14:54 +0200,
Gerhard Pircher wrote:
> 
> 
> -------- Original-Nachricht --------
> > Datum: Wed, 24 Jun 2009 11:47:13 +0200
> > Von: Takashi Iwai <tiwai@suse.de>
> > An: "Gerhard Pircher" <gerhard_pircher@gmx.net>
> > CC: benh@kernel.crashing.org, linuxppc-dev@ozlabs.org
> > Betreff: Re: ALSA fixes for non-coherent ppc32 again
> 
> > At Wed, 24 Jun 2009 10:46:01 +0200,
> > Gerhard Pircher wrote:
> > > 
> > > 
> > > -------- Original-Nachricht --------
> > > > Datum: Tue, 23 Jun 2009 23:42:24 +0200
> > > > Von: "Gerhard Pircher" <gerhard_pircher@gmx.net>
> > > > An: benh@kernel.crashing.org, tiwai@suse.de
> > > > CC: linuxppc-dev@ozlabs.org
> > > > Betreff: Re: ALSA fixes for non-coherent ppc32 again
> > > 
> > > > Okay, that's wrong. I somehow messed up the .config file. It doesn't
> > > > compile, too.
> > > I got it to compile now and it seems to work fine. I overlooked a typo
> > > in sound/core/Makefile first (ifndef CONFIG_SND_NONCOHERNT_DMA.)
> > 
> > Gah, thanks, fixed now on the git tree too.
> 
> The Kconfig option still needs to be fixed, otherwise SND_COHERENT_DMA
> isn't selected for my "PPC32 && NOT_COHERENT_CACHE" machine.
> 
> config SND_NONCOHERENT_DMA
>         def_bool n  <-- should be "y"
>         depends on (PPC32 && NOT_COHERENT_CACHE)...

Thanks.  I changed the patch rather to add a define instead of
Kconfig so that it can detect more easily.


> BTW: Can you put me on CC: please, if you bring up this topic on
> linux-arch or so?

Sure.
Hopefully I can bring up the stuff in this week...


Takashi

Re: ALSA fixes for non-coherent ppc32 again

[Gerhard Pircher]

-------- Original-Nachricht --------
> Datum: Wed, 08 Jul 2009 16:13:10 +0200
> Von: Takashi Iwai <tiwai@suse.de>
> An: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> CC: Gerhard Pircher <gerhard_pircher@gmx.net>, linuxppc-dev@ozlabs.org
> Betreff: Re: ALSA fixes for non-coherent ppc32 again

> At Wed, 08 Jul 2009 13:01:50 +1000,
> Benjamin Herrenschmidt wrote:
> > 
> > On Mon, 2009-06-22 at 08:34 +1000, Benjamin Herrenschmidt wrote:
> > > On Sun, 2009-06-21 at 20:18 +0200, Gerhard Pircher wrote:
> > > > Hi,
> > > > 
> > > > Takashi Iwai posted patches to make ALSA work on non-coherent
> > > > PPC32 systems (almost exactly) a year ago. See here:
> > > >
> > > > http://www.nabble.com/-PATCH-0-3--ALSA-fixes-for-non-coherent-ppc32-to17980027.html#a17980027
> > > > 
> > > > As far as I can see these patches never went upstream. Where
> > > > there any objections or did we just forget about them? It would
> > > > be cool, if the patches could be merged now, as at least two
> > > > platforms need this bugfix (namely Sam440 and AmigaOne).
> > > 
> > > I definitely forgot about those... But I'm fine with what Takashi
> > > did for now, I can always make the powerpc helper for
> > > dma_mmap_coherent() smarter later on if necessary.
> > 
> > BTW. Can you guys send a "final" patch for adding mmap_coherent to
> > powerpc ? Please make so that the dma_mmap_coherent() function doesn't
> > explose if dma_ops->mmap_coherent is NULL though (either fail
> > gracefully or fallback to some standard mmap).
> 
> The attached is the revised patch.  It falls backs to the standard
> mmap.
> 
> The whole patch series are found in test/dma-fix branch of sound git
> tree below:
>   git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git
> test/dma-fix
> 
> I had no time to play cross-build yet, so it'd be helpful if someone
> can test it...
> (BTW, any good minimal gcc and binutils package for kernel builds for
>  PPC32?)
Okay, I cross-build it on a x86 machine and tested it on my AmigaOne with
the VIA686 onboard sound and a Soundblaster 5.1. It worked fine after I
applied the "ALSA: Fix SG-buffer DMA with non-coherent architectures"
patch.

Thanks!

Gerhard

-- 
Neu: GMX Doppel-FLAT mit Internet-Flatrate + Telefon-Flatrate
für nur 19,99 Euro/mtl.!* http://portal.gmx.net/de/go/dsl02