From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jon Medhurst (Tixy) Date: Wed, 30 Nov 2011 15:19:06 +0000 Subject: [U-Boot] ARM: vexpress: Extend default boot sequence to load script from MMC In-Reply-To: References: <1322574376.4500.34.camel@linaro1> <1322641942.2321.13.camel@linaro1> Message-ID: <1322666346.3969.26.camel@linaro1> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de On Wed, 2011-11-30 at 07:51 -0700, Tom Rini wrote: > On Wed, Nov 30, 2011 at 1:32 AM, Jon Medhurst (Tixy) > wrote: > > On Tue, 2011-11-29 at 13:46 -0700, Tom Rini wrote: > >> On Tue, Nov 29, 2011 at 6:46 AM, Jon Medhurst (Tixy) > >> wrote: > >> > Extend the default boot sequence on Versatile Express to load a boot > >> > script from MMC. > >> > > >> > Signed-off-by: Jon Medhurst > >> > >> Is there any interest in defining a common 'fancy' boot command? Over > >> on beagleboard (and a few other eval boards) we are (or will be) > >> doing, roughly: > >> Is there mmc? { > >> Can we load a bootscript? { > >> load it, do it > >> } > >> Can we load uEnv.txt? { > >> Did it set 'uenvcmd'? { > >> Run it > >> } > >> } > >> Did we load the kernel? { > >> bootm it > >> } > >> } > >> Try nand. > >> > >> And it's not hard to replace NAND with "whatever flash the board sets" > > > > It could make sense. However, the combinations of boot methods could get > > quite big, e.g. do we include pxeboot, tftp? > > Well, maybe we define some blocks (TRY_MMC, TRY_ENET, TRY_NAND, > TRY_SPI, ...) ... > > > If everything was included in the common 'fancy boot' then some boards > > would have to define boot methods (or stubs) for things they weren't > > interested in. And, conversely, if fancy boot didn't include everything, > > then other people would need to extend it. > > ... provide a few examples of the blocks strung together and have > people string up what they need themselves, so long as it's an > otherwise good idea to have complex default boot methods? If something like this was planned for BeagleBoard anyway, then perhaps it could written in terms of these TRY_MMC etc. macros (which seem like a reasonable idea to me). Then a working prototype could be posted as an RFC for others to comment on? I suspect at the moment there's not many people reading this 'vexpress' thread ;-) -- Tixy