Date: Tue, 5 Jun 2007 18:59:09 +0200 From: Bernd Walter <ticso@cicely12.cicely.de> To: "M. Warner Losh" <imp@bsdimp.com> Cc: ticso@cicely12.cicely.de, arm@freebsd.org, ticso@cicely.de Subject: Re: timeout while detecting SD card Message-ID: <20070605165908.GP16463@cicely12.cicely.de> In-Reply-To: <20070605.100615.-1384053623.imp@bsdimp.com> References: <53067.2001:6f8:101e:0:20e:cff:fe6d:6adb.1181056352.squirrel@webmail.alpha-tierchen.de> <20070605154250.GN16463@cicely12.cicely.de> <20070605.100615.-1384053623.imp@bsdimp.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Jun 05, 2007 at 10:06:15AM -0600, M. Warner Losh wrote: > In message: <20070605154250.GN16463@cicely12.cicely.de> > Bernd Walter <ticso@cicely12.cicely.de> writes: > : On Tue, Jun 05, 2007 at 05:12:32PM +0200, Björn König wrote: > : > Hello, > : > > : > yet another problem. I can't access the SD card. I googled a bit and > : > noticed that I'm not the only one with this problem, but I haven't found a > : > solution yet. Here are some details: > : > > : > Everything seems to work fine until sending the app command in > : > mmc_wait_for_app_cmd. The driver gets an interrupt with a "response > : > time-out error" set in status register. That's it. > > I think this is what I fixed by increasing the timeout... > > : > I tried to find the problem and executed an Atmel MCI demo programm in > : > kernel shortly after mmc_scan. It does basically the same and detects the > : > card in the SD card bay properly. There is one obvious difference: the > : > demo doesn't use an interrupt service routine. > : > > : > I hope someone has a hint for me, once again. ;-) > : > : All drivers expect the bootcode to setup the io-lines. > : I also saw the effect that when booting via bootspi (MCI init added) > : then the first boot may not find the card. > : Booting via boot2 always succeed. > : You may want to check about what redboot does about MCI init. > > The theory I read somewhere in linux land was that the boot loader was > responsible for setting up the various I/O lines. I coded full speed > ahead with this assumption. If it isn't actually true, then we can > reevaluate. It isn't that hard to add a board init function. But > then we need to add a board id to boot2 and have the kernel use it. I > believe that redboot already passes one in... I personally like the current situation is best. I can have the same image for variuous boards, whichout having to define a new board name everytime. Just a few points in respect to hardcoded clock rates may be a problem. E.g. MCK can be higher than 60MHz if you accept to reduce the CPU clock, which may or may not be faster depending on the io usage. Not to speak about overclocking, which, if I got the RM9200 datasheet right, allows higher clock rates by just overclocking the PLL. -- B.Walter http://www.bwct.de http://www.fizon.de bernd@bwct.de info@bwct.de support@fizon.de
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070605165908.GP16463>