On Tue, Nov 29, 2011 at 1:02 AM, Thierry Reding <thierry.red...@avionic-design.de> wrote: > * Andy Fleming wrote: > [...] >> Well, while this isn't the case in any systems I know of now, it is >> quite possible for there to be more than one *type* of SD/MMC >> controller on an SoC, and there's always the possibility that an SoC >> provides a non-controller-specific card-detect mechanism. The idea is >> that, lacking a board-specific card-detect mechanism, the SoC might be >> able to direct the query to the right place. >> >> But I'm talking very theoretically, here. I wouldn't object to a >> mechanism that was just: >> >> cd = board_mmc_getcd(mmc); >> >> if (cd < 0 && mmc->getcd) >> cd = mmc->getcd(mmc); >> >> If we ever ran into a case where an SoC had better knowledge than the >> driver, then it's easy to fix the code. > > That makes sense. The code can always be extended when new hardware requires > it. No need to over-engineer at this point. > > Do you want me to prepare a patch or should we rather wait for some more > input from others?
I'll happily take a patch. I may wait to apply it, but probably, as the merge window is *long* closed, so it won't get pulled into mainline until mid-December (I'll apply it to my -next branch). Andy _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
