On Tue, Sep 25, 2012 at 02:13:10AM -0700, Yong Ding wrote: [...] > So, in all, u are right if with my current patch, some host drivers need > some improvement to add MMC_CAP_NEEDS_POLL when it is actually needed. > But I think this shall be the right way to follow. Or, we might enable > polling for some cases in which it is unnecessary, and maybe this is a > potential issue-bomb. How do u think? I think if you carefully review and fixup all the drivers, it will be fine. But you'd have to add MMC_CAP_NEEDS_POLL into drivers' code (while w/ the quirk it's less lines of code for drivers). So, here is another idea: how about something like this #define SDHCI_QUIRK_NEEDS_POLL \ (SDHCI_QUIRK_BROKEN_CARD_DETECTION | (1 << NN)) And changing the logic to: if ((host->quirks & SDHCI_QUIRK_NEEDS_POLL) && !(host->mmc->caps & MMC_CAP_NONREMOVABLE)) mmc->caps |= MMC_CAP_NEEDS_POLL; And then, you'd just convert all the current drivers to SDHCI_QUIRK_NEEDS_POLL (which would be 100% safe), and for your driver, you'd only set SDHCI_QUIRK_BROKEN_CARD_DETECTION. Thanks, Anton. -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html