Re: imx-usb-image write to usb error

From: Sascha Hauer <s.hauer@pengutronix.de>
To: Jason Cooper <barebox@lakedaemon.net>
Cc: barebox@lists.infradead.org
Subject: Re: imx-usb-image write to usb error
Date: Wed, 11 Sep 2013 09:00:04 +0200	[thread overview]
Message-ID: <20130911070004.GJ30088@pengutronix.de> (raw)
In-Reply-To: <20130909174117.GA12927@titan.lakedaemon.net>

On Mon, Sep 09, 2013 at 01:41:17PM -0400, Jason Cooper wrote:
> On Thu, Sep 05, 2013 at 08:46:39AM +0200, Sascha Hauer wrote:
> > Hi Jason,
> > 
> > On Wed, Sep 04, 2013 at 09:04:36PM -0400, Jason Cooper wrote:
> > > All,
> > > 
> > > This is the first I've messed with i.MX, so I'm a little lost. :)
> > > 
> > > I'm putting together a patch to add support for the SanDisk Connect
> > > Wireless Media Drive [1], code dump from SanDisk is here [2].  It turns
> > > out when you unplug the battery, it'll show up as the Freescale hid.
> > > 
> > > Here's what I get:
> > > 
> > > $ ./scripts/imx/imx-image -c \
> > > arch/arm/boards/sandisk-connect-wmd/flash-header-lpddr1.imxcfg -f \
> > > barebox.bin -o barebox.imx
> > > $ sudo ./scripts/imx/imx-usb-loader -vvvvv ./barebox.imx 
> > > found i.MX50 USB device [15a2:0052]
> > 
> > I never tested this on an i.MX50 due to lack of hardware. Have a look
> > at scripts/imx/imx-usb-loader.c at imx_ids. i.MX50 is very similar to
> > i.MX51. The i.MX51 has:
> > 
> > 	{
> > 		.vid = 0x15a2,
> > 		.pid = 0x0041,
> > 		.name = "i.MX51",
> > 		.header_type = HDR_MX51,
> > 		.mode = MODE_BULK,
> > 		.max_transfer = 64,
> > 	},
> > 
> > The header_type, mode and max_transfer fields are missing for i.MX50. I
> > assume that's the problem.
> 
> Well, it looks like it works, but I'm unable to get a barebox prompt on
> the tty.  (patches to come once I get a prompt).  So I decided to try
> tackling the problem from the other end for a bit.  Namely booting a
> mainline kernel from the installed u-boot.
> 
> Something is confusing me though.  The part number on the chip is
> MCIMX502CVK8B.  Which, iiuc, is an imx51.  The provided kernel spits out
> this:
> 
> [    0.000000] MXC_Early serial console at MMIO 0x53fbc000 (options '115200')
> 
> However, the only imx SoCs using that address range are imx53 and imx25,
> according to arch/arm/boot/dts/ in Linux, at least.
> 
> In addition, the docs for the imx50 clearly state that it uses DCD v2
> (version = 0x40) [1] (Table 6-27, DCD Header), but the code in barebox's
> imx-image.c says imx51 uses DCD v1...

The i.MX53 is not much more than a moderately improved i.MX51. The i.MX50
is somewhere in between. Maybe it already uses the v2 header format.
Have you tried copying the values from the i.MX53?

> 
> Anyone with more experience with Freescale SoCs care to fill in the
> blanks?  The first hint I'm on the wrong path is that I feel like I need
> to write a whole dtsi file from scratch. :-/

The i.MX51 and i.MX53 have a very similar memory map. Only the AIPS1/2
base addresses are different to allow for bigger SDRAM on the i.MX53.
The i.MX50 memory map seems to be a subset of the i.MX53.

What I'm afraid is really different between i.MX50 and i.MX53 is the
clock controller.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox