On Sun, Jul 14, 2013 at 12:05 +0200, Lars-Peter Clausen wrote: > > On 07/12/2013 05:26 PM, Gerhard Sittig wrote: > [...] > ] > > + if (mchan->tcd_nunits) > > + tcd->nbytes = mchan->tcd_nunits * 4; > > + else > > + tcd->nbytes = 64; > > Just wondering where does the magic 64 come from?
"Inherited" from Alexander's submission. By coincidence the SDHC client happened to work with this constraint, as did Alexander's LPB test suite. I'm not aware of other DMA clients as of now (except for the memory transfer, which runs a different prep routine). It looks like there is more potential for generalization and auto adapting to user input (adjusting configurable inner-loop parameters to exactly fullfill the overall job from the user's spec yet meet the chunking constraints of the hardware). virtually yours Gerhard Sittig -- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr. 5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: off...@denx.de _______________________________________________ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
