From: David Miller on
From: Stephen Hemminger <shemminger(a)linux-foundation.org>
Date: Wed, 20 Jan 2010 12:30:33 -0800

> It looks like the size passed to sync_single has to match size of original
> mapping.

I think that's rediculious, and unreasonable.

All implementations of these APIs where the sync matters have the
ability to handle partial syncs correctly.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Jarek Poplawski on
On Wed, Jan 20, 2010 at 10:24:14PM +0000, Alan Cox wrote:
> > > Seems like an underlying bug in the DMA api. Maybe it just can't
> > > handle operations on partial mapping.
> > >
> > > Other drivers with same problem:
> > > bnx2, cassini, pcnet32, r8169, rrunner, skge, sungem, tg3,
> >
> > It seems using the same length (even without pci_unmap_len()) is
> > crucial here, but I hope maintainers (added to CC) will take care.
>
> The API needs fixing - if you've got a large mapping and you want to sync
> part of it then we need to support that. Now it might well be that the
> implementation on some braindead platform has to sync the entire thing,
> and some implementations entire pages or cache lines.
>
> You can't fix this in the drivers, they requested a service and they
> don't have enough information nor is it their job to know about all the
> platform specific rules.

Yes, the need to repeat some other values if there is a dedicated
structure/pointer could be misleading. Btw, it seems to be a trivial
overlooking since there is dma_sync_single_range() ready to use.

Jarek P.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: David Miller on
From: Jarek Poplawski <jarkao2(a)gmail.com>
Date: Wed, 20 Jan 2010 21:58:42 +0100

> Yes, and it's mainly for lib/dma-debug (until it's not verified dmar
> errors reported by Michael Breuer could be connected). So, I'm not
> sure for now how serious this warning could be. On the other hand,
> Documentation/(PCI-?)DMA-mapping, mentioned by Stephen, doesn't seem
> to allow or use in examples the "size" different than mapped.

Someone find me even one actual implemenetation of the DMA APIs, not
the debugging code, which can't handle a partial sync correctly, then
we can talk.

As the person who originally created the PCI DMA APIs I can tell
you the intention was definitely to allow arbitrary SYNC lengths
and there was no requirement that the length needed to be the
same as that used in the mapping call.

Nothing else makes sense. People sync when they have a partially used
buffer and want to copy the data out then recycle the big original
buffer back to the chip. I knew this when creating the PCI DMA APIs
because I first tried to use them in network drivers and that's what
so many of them do.

Currently all we have is an overly anal DMA debugging layer check, and
some speculation that it might or might not fix some real bug.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: David Miller on
From: Alan Cox <alan(a)lxorguk.ukuu.org.uk>
Date: Wed, 20 Jan 2010 22:24:14 +0000

>> > Seems like an underlying bug in the DMA api. Maybe it just can't
>> > handle operations on partial mapping.
>> >
>> > Other drivers with same problem:
>> > bnx2, cassini, pcnet32, r8169, rrunner, skge, sungem, tg3,
>>
>> It seems using the same length (even without pci_unmap_len()) is
>> crucial here, but I hope maintainers (added to CC) will take care.
>
> The API needs fixing - if you've got a large mapping and you want to sync
> part of it then we need to support that. Now it might well be that the
> implementation on some braindead platform has to sync the entire thing,
> and some implementations entire pages or cache lines.
>
> You can't fix this in the drivers, they requested a service and they
> don't have enough information nor is it their job to know about all the
> platform specific rules.

Absolutely and %100 agreed, the DMA debugging layer and documentation
are both buggy.

The intention from the beginning was always to allow partial SYNCs
exactly for the reasons Alan and myself are mentioning here.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: FUJITA Tomonori on
On Wed, 20 Jan 2010 23:53:22 +0100
Jarek Poplawski <jarkao2(a)gmail.com> wrote:

> On Wed, Jan 20, 2010 at 10:24:14PM +0000, Alan Cox wrote:
> > > > Seems like an underlying bug in the DMA api. Maybe it just can't
> > > > handle operations on partial mapping.
> > > >
> > > > Other drivers with same problem:
> > > > bnx2, cassini, pcnet32, r8169, rrunner, skge, sungem, tg3,
> > >
> > > It seems using the same length (even without pci_unmap_len()) is
> > > crucial here, but I hope maintainers (added to CC) will take care.
> >
> > The API needs fixing - if you've got a large mapping and you want to sync
> > part of it then we need to support that. Now it might well be that the
> > implementation on some braindead platform has to sync the entire thing,
> > and some implementations entire pages or cache lines.
> >
> > You can't fix this in the drivers, they requested a service and they
> > don't have enough information nor is it their job to know about all the
> > platform specific rules.
>
> Yes, the need to repeat some other values if there is a dedicated
> structure/pointer could be misleading. Btw, it seems to be a trivial
> overlooking since there is dma_sync_single_range() ready to use.

Yeah, dma_sync_single_range() enables you to do a partial sync. But
you must be really careful with a partial sync (as DMA-API.txt says).
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/