From mboxrd@z Thu Jan  1 00:00:00 1970
From: Marek Vasut <marex@denx.de>
Date: Tue, 22 Sep 2015 06:13:46 +0200
Subject: [U-Boot] [PATCH] spi : convert altera_spi to driver model
In-Reply-To: <5600D09A.2040107@wytron.com.tw>
References: <1442840311-12689-1-git-send-email-thomas@wytron.com.tw>
 <201509220210.26908.marex@denx.de> <5600D09A.2040107@wytron.com.tw>
Message-ID: <201509220613.46875.marex@denx.de>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

On Tuesday, September 22, 2015 at 05:52:58 AM, Thomas Chou wrote:
> Hi Marek,

Hi,

> On 09/22/2015 08:10 AM, Marek Vasut wrote:
> > I suspect you might want to tweak the core code to check if those
> > .set_speed and .set_mode are assigned in dm_spi_ops structure and
> > if not, don't call them.
> 
> I wonder if this worths, as altera_spi is the only one that cannot set
> speed and mode.

I'm expecting that EPCQ Wundercontroller to have the same problem (see [1]
section 42, Altera GQSPI). Altera tried to push this into mainline Linux
even [2], without much success, in particular because they duplicated half
of the spi-nor stack to do it. I suspect this driver is coming to U-Boot
sooner or later.

But maybe we should do this change to the SPI NOR stuff once this driver
hits the U-Boot ML instead of preemptively adding such checks. What do you
think ?

[1] https://www.altera.com/content/dam/altera-
www/global/en_US/pdfs/literature/ug/ug_embedded_ip.pdf
[2] https://lwn.net/Articles/636882/

> >> +	plat->regs = ioremap(dev_get_addr(bus),
> >> +		sizeof(struct altera_spi_regs));
> > 
> > I guess the same thing about ioremap() and ranges applies here?
> 
> Same thing about ioremap(), as almost every other arch does not include
> it in u-boot. You see that I just added it to nios2 lately.
> 
> Sometimes the address returned by dev_get_addr() may be used as memory,
> and does not need ioremap.

Gotcha.

Best regards,
Marek Vasut