Hi, 2012/11/5 Huang Changming-R66093 <r66093@xxxxxxxxxxxxx>: >> >> 2012/11/2 Huang Changming-R66093 <r66093@xxxxxxxxxxxxx>: >> > Hi, Johan, >> > When quicks SDHCI_QUIRK_BROKEN_CARD_DETECTION is set, the driver will >> poll the card status with the command CMD13 periodically, then many >> interrupts will be generated, which affect the performance. >> >> Ok, just to see if I understand the scenario correctly: >> SDHCI_QUIRK_BROKEN_CARD_DETECTION causes the MMC_CAP_NEEDS_POLL cap to be >> set, which will cause mmc_rescan to be run at a one second interval. >> mmc_rescan calls bus_ops->detect (mmc_sd_detect) and this in turn calls >> _mmc_detect_card_removed, which will do the bus_ops->alive and send CMD13. >> So in this case, one CMD13 is sent per second, right? >> Is this the cause of the performance issue? > > Yes, You are right. Ok, it's a bit hard to understand how this can lead to a noticeable performance impact, but OK. :) >> A thought: if the host hardware does have a GPIO card detect pin hooked >> up, would it not be possible to make this pin generate an IRQ whenever a >> card is inserted or removed? This is what we do in the MMCI driver, for >> instance (mmci_cd_irq). > > Our silicones has this pin to do card detect, but some boards don't generate the interrupt when card is inserted or removed. So I have to use the poll mode. > >> > Yes, some cases to detect GPIO can't be trusted, so I only just >> implement this callback in eSDHC-of driver. that is to say, just when the >> platform support it, this callback can be implement, if not, continue to >> send the command CMD13. >> >> I'm just wondering how this will affect our system, where we use the cd >> GPIO to generate detect jobs on card insert/remove, but where the cd pin >> is not 100% synchronized with the electrical connection of the power and >> cmd line of the SD-card. So if I remember correctly, the cd pin may >> report that the card has been removed, but there is still an electric >> connection and the card is alive... >> > I don't see this on our board, when card is inserted or removed, the register field can reflect this state correctly. We see a difference on our boards with this patch, but no functional change (actually a removed card is detected earlier now), so from my perspective: Reviewed-By: Johan Rudholm <johan.rudholm@xxxxxxxxxxxxxx> //Johan -- 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
