Hello ksi, k...@koi8.net wrote: > On Mon, 16 Feb 2009, Wolfgang Denk wrote: > >> Dear k...@koi8.net, >> >> In message <pine.lnx.4.64ksi.0902152224290.17...@home-gw.koi8.net> you wrote: >>>> Yes, good point. But do we need more then one i2c adapter when running >>>> from flash? I see only one reason to use i2c when running from flash: >>>> accessing SPD EEprom ... and this "bus" could always be the first >>>> hw adapter. All other accesses to i2c should be moved to run when >>>> we are in ram. >>> You can have, e.g. TWO SPD EEPROMs on different busses. And please remember >>> that infamous "640K ought to be enough for anybody..." >> This is a very unlikely hypothetical case. Based on such assumptions >> you can design any arbitrary complexity in any code. >> >> For now, please lets assume that there is no such usage mode, and >> that we do not add complexity to support it. >> >> >>> I don't think there is a viable reason to object unless somebody came up >>> with something better. >> Well, complex code that is difficult to read and difficult to >> understand is a very viable reason to object, me thinks. > > The code is not all that complex. Most of the code is the same, there is > just a thin HAL over it. And it is _NOT_ for reading two SPDs on different > busses, it is for consistency and uniformity. Reading two SPDs is a free > bonus, there is no a single line of additional code for this.
If the current pointer is writeable, this is also possible ... bye Heiko -- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
