[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <438BB0150E931F4B9CE701519A44630104A3641362@bgsmsx502.gar.corp.intel.com>
Date: Wed, 21 Jul 2010 08:38:09 +0530
From: "Koul, Vinod" <vinod.koul@...el.com>
To: Linus Walleij <linus.ml.walleij@...il.com>
CC: "Williams, Dan J" <dan.j.williams@...el.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
Alan Cox <alan@...ux.intel.com>
Subject: RE: [PATCH 1/3] DMAENGINE: generic slave channel control
> ? Are you using memcpy() to talk to slaves?
Yes, I don't have sg support yet, that's something I need to add next.
> I was assuming all slave communication was to use sglists through
> the .device_prep_slave_sg() call. This is currently the design constraint,
> memcpy() will by design increase both source and destination address
> and also always operate on the memory bus.
>
> (If you need reconfiguration also for memcpy() I think will be a
> different issue, I'm only looking at slaves now.)
>
> With only the slave interface the dma_chan struct can be deferred
> by the DMA engine into a local struct which has this address configured
> from the platform, statically.
>
> The runtime configuration API is exactly about being able to reconfigure
> even the source/destination address at runtime. This is why
> these are on the interface.
Okay looking at the sg API, now I can understand why you need the address in
this structure, I would also need that in future.
One suggestion since we are giving io address here, how about naming the
variable as io_addr, and we can add comment to be used for sg operations as io
addr if anyone wants to use memcpy() they can ignore this
Thanks
Vinod
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists