From: cyril@ti•com (Cyril Chemparathy)
To: linux-arm-kernel@lists•infradead.org
Subject: [PATCH v7 01/10] ARM: davinci: move private EDMA API to arm/common
Date: Mon, 4 Feb 2013 16:54:45 -0500 [thread overview]
Message-ID: <51102E25.8040107@ti.com> (raw)
In-Reply-To: <CACRpkdbPyZt8=pLhz-5qcaSSAk6VBn61dPTNp6teU9HksBwN2w@mail.gmail.com>
On 02/04/2013 04:11 PM, Linus Walleij wrote:
> On Mon, Feb 4, 2013 at 9:33 PM, Mark Brown
> <broonie@opensource•wolfsonmicro.com> wrote:
>> On Mon, Feb 04, 2013 at 09:29:46PM +0100, Linus Walleij wrote:
>>> On Mon, Feb 4, 2013 at 8:22 PM, Cyril Chemparathy <cyril@ti•com> wrote:
>>
>>>> Based on our experience with fitting multiple subsystems on top of this
>>>> DMA-Engine driver, I must say that the DMA-Engine interface has proven
>>>> to be a less than ideal fit for the network driver use case.
>>
>>>> The first problem is that the DMA-Engine interface expects to "push"
>>>> completed traffic up into the upper layer as a part of its callback.
>>>> This doesn't fit cleanly with NAPI, which expects to "pull" completed
>>>> traffic from below in the NAPI poll. We've somehow kludged together a
>>>> solution around this, but it isn't very elegant.
>>
>>> I cannot understand the actual technical problem from the above
>>> paragraphs though. dmaengine doesn't have a concept of pushing
>>> nor polling, it basically copies streams of words from A to B, where
>>> A/B can be a device or a buffer, nothing else.
>>
>>> The thing you're looking for sounds more like an adapter on top
>>> of dmaengine, which can surely be constructed, some
>>> drivers/dma/dmaengine-napi.c or whatever.
>>
>> Broadly speaking what NAPI wants is to never get any callbacks from the
>> hardware (or DMAs). It wants to wake up periodically, take a look at
>> what packets have been read by the hardware and process them. The goal
>> is to have the DMAs sitting and running without disturbing the processor
>> at all after the first packet has been handled.
>
> OK we should definately be able to encompass that in dmaengine
> quite easily.
>
> So I think the above concerns are moot. The callback we can
> set on cookies is entirely optional, and it's even implemented by
> each DMA engine, and some may not even support it but *require*
> polling, and then it won't even be implemented by the driver.
>
> Which probably stems from the original design of the dmaengine
> API, which was for TCP networking acceleration, mainly.
>
> Cyril, just stack up the cookies and take a sweep over them to see
> which ones are baked when the NAPI poll comes in -> problem
> solved.
>
You're assuming that cookies complete in order. That is not necessarily
true.
Thanks
-- Cyril.
next prev parent reply other threads:[~2013-02-04 21:54 UTC|newest]
Thread overview: 82+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-02-01 18:22 [PATCH v7 00/10] DMA Engine support for AM33XX Matt Porter
2013-02-01 18:22 ` [PATCH v7 01/10] ARM: davinci: move private EDMA API to arm/common Matt Porter
2013-02-01 18:41 ` Tony Lindgren
2013-02-02 12:49 ` Russell King - ARM Linux
2013-02-02 14:44 ` Matt Porter
[not found] ` <5022f635a527470dbd0be932063e9cd2@DFLE72.ent.ti.com>
2013-02-01 18:49 ` Matt Porter
[not found] ` <2077c13e12314dc3adc8e5b653855da0@DFLE72.ent.ti.com>
2013-02-01 18:59 ` Matt Porter
2013-02-02 0:01 ` Sergei Shtylyov
2013-02-02 12:45 ` Russell King - ARM Linux
2013-02-02 17:27 ` Sergei Shtylyov
[not found] ` <e9be6668da8b4372a04687847daa1d8c@DFLE72.ent.ti.com>
2013-02-02 18:07 ` Matt Porter
2013-02-02 18:16 ` Tony Lindgren
2013-02-02 19:48 ` Matt Porter
2013-02-02 21:02 ` Tony Lindgren
2013-02-02 19:06 ` Sergei Shtylyov
[not found] ` <3245316d7aa94b2e823f98b69497547d@DLEE74.ent.ti.com>
2013-02-02 19:55 ` Matt Porter
2013-02-02 20:18 ` Sergei Shtylyov
2013-02-01 19:52 ` Sergei Shtylyov
2013-02-01 18:58 ` Felipe Balbi
2013-02-01 20:49 ` Sergei Shtylyov
2013-02-01 20:56 ` Felipe Balbi
2013-02-01 21:30 ` Russell King - ARM Linux
2013-02-02 0:07 ` Sergei Shtylyov
2013-02-02 0:44 ` Russell King - ARM Linux
2013-02-02 2:09 ` Sergei Shtylyov
2013-02-02 10:18 ` Russell King - ARM Linux
2013-02-02 12:17 ` Russell King - ARM Linux
2013-02-02 17:02 ` Sergei Shtylyov
2013-02-02 16:27 ` Sergei Shtylyov
2013-02-02 16:45 ` Russell King - ARM Linux
2013-02-02 17:17 ` Sergei Shtylyov
2013-02-04 14:27 ` Arnd Bergmann
2013-02-02 0:13 ` Sergei Shtylyov
2013-02-04 15:41 ` Felipe Balbi
2013-02-04 15:45 ` Russell King - ARM Linux
2013-02-04 17:36 ` Sergei Shtylyov
2013-02-04 16:47 ` Felipe Balbi
2013-02-04 17:10 ` Russell King - ARM Linux
2013-02-04 17:54 ` Sergei Shtylyov
2013-02-04 17:02 ` Felipe Balbi
2013-02-04 18:22 ` Sergei Shtylyov
2013-02-04 19:22 ` Cyril Chemparathy
2013-02-04 20:29 ` Linus Walleij
2013-02-04 20:33 ` Mark Brown
2013-02-04 21:11 ` Linus Walleij
2013-02-04 21:47 ` Arnd Bergmann
2013-02-05 12:38 ` Russell King - ARM Linux
2013-02-05 15:37 ` Cyril Chemparathy
2013-02-04 21:54 ` Cyril Chemparathy [this message]
2013-02-05 12:41 ` Russell King - ARM Linux
2013-02-05 15:42 ` Cyril Chemparathy
2013-02-05 15:30 ` Linus Walleij
2013-02-05 17:14 ` Russell King - ARM Linux
2013-02-05 18:33 ` Linus Walleij
2013-02-04 22:30 ` Cyril Chemparathy
2013-02-05 16:21 ` Linus Walleij
2013-02-05 16:47 ` Mark Brown
2013-02-05 17:06 ` Russell King - ARM Linux
2013-02-05 17:41 ` Mark Brown
2013-02-05 18:29 ` Linus Walleij
2013-02-05 19:45 ` Cyril Chemparathy
2013-02-05 18:28 ` Tony Lindgren
2013-02-05 22:26 ` Arnd Bergmann
2013-02-06 7:45 ` Felipe Balbi
2013-02-01 23:10 ` Sergei Shtylyov
2013-02-09 16:05 ` Sekhar Nori
2013-02-09 20:08 ` Russell King - ARM Linux
2013-03-04 22:05 ` Matt Porter
[not found] ` <e92425fefcc04bb4ab739ec8d4e82672@DLEE74.ent.ti.com>
2013-03-04 22:12 ` Matt Porter
2013-02-01 18:22 ` [PATCH v7 02/10] ARM: edma: remove unused transfer controller handlers Matt Porter
2013-02-01 18:22 ` [PATCH v7 03/10] ARM: edma: add AM33XX support to the private EDMA API Matt Porter
2013-02-01 18:22 ` [PATCH v7 04/10] dmaengine: edma: enable build for AM33XX Matt Porter
2013-02-01 18:22 ` [PATCH v7 05/10] dmaengine: edma: Add TI EDMA device tree binding Matt Porter
2013-02-01 18:26 ` Matt Porter
2013-02-01 18:22 ` [PATCH v7 06/10] ARM: dts: add AM33XX EDMA support Matt Porter
2013-02-01 18:22 ` [PATCH v7 07/10] dmaengine: add dma_request_slave_channel_compat() Matt Porter
2013-02-01 18:28 ` Matt Porter
2013-02-12 16:38 ` Vinod Koul
2013-02-01 18:22 ` [PATCH v7 08/10] spi: omap2-mcspi: convert to dma_request_slave_channel_compat() Matt Porter
2013-02-01 18:22 ` [PATCH v7 09/10] spi: omap2-mcspi: add generic DMA request support to the DT binding Matt Porter
2013-02-01 18:22 ` [PATCH v7 10/10] ARM: dts: add AM33XX SPI DMA support Matt Porter
2013-02-01 18:32 ` [PATCH v7 00/10] DMA Engine support for AM33XX Matt Porter
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=51102E25.8040107@ti.com \
--to=cyril@ti$(echo .)com \
--cc=linux-arm-kernel@lists$(echo .)infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox