| Message ID | 1308566325-11600-2-git-send-email-shawn.guo@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Jun 20, 2011, at 3:38 AM, Shawn Guo wrote: > The issue was initially found by Eric Benard as below. > > http://permalink.gmane.org/gmane.linux.ports.arm.kernel/108031 > > Not sure about other SDHCI based controller, but on Freescale eSDHC, > the SDHCI_INT_CARD_INSERT bits will be immediately set again when it > gets cleared, if a card is inserted. The driver need to mask the irq > to prevent interrupt storm which will freeze the system. And the > SDHCI_INT_CARD_REMOVE gets the same situation. > > The patch fixes the problem based on the initial idea from > Eric Benard. > > Signed-off-by: Shawn Guo <shawn.guo@linaro.org> > Cc: Eric Benard <eric@eukrea.com> > --- > drivers/mmc/host/sdhci.c | 27 +++++++++++++++++++++++---- > 1 files changed, 23 insertions(+), 4 deletions(-) > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c > index 91d9892..d94e2b4 100644 > --- a/drivers/mmc/host/sdhci.c > +++ b/drivers/mmc/host/sdhci.c > @@ -127,7 +127,9 @@ static void sdhci_mask_irqs(struct sdhci_host *host, u32 irqs) > > static void sdhci_set_card_detection(struct sdhci_host *host, bool enable) > { > - u32 irqs = SDHCI_INT_CARD_REMOVE | SDHCI_INT_CARD_INSERT; > + u32 present = sdhci_readl(host, SDHCI_PRESENT_STATE) & > + SDHCI_CARD_PRESENT; > + u32 irqs = present ? SDHCI_INT_CARD_REMOVE : SDHCI_INT_CARD_INSERT; > > if (host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) > return; Accessing off chip registers is slow. It would be better to check for the quirk first and then read the present state register. eg u32 present; u32 irqs; > if (host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) > return; > present = sdhci_readl(host, SDHCI_PRESENT_STATE) & > SDHCI_CARD_PRESENT; > irqs = present ? SDHCI_INT_CARD_REMOVE : SDHCI_INT_CARD_INSERT; Philip -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Jun 20, 2011 at 08:59:38AM -0700, Philip Rakity wrote: > > On Jun 20, 2011, at 3:38 AM, Shawn Guo wrote: > > > The issue was initially found by Eric Benard as below. > > > > http://permalink.gmane.org/gmane.linux.ports.arm.kernel/108031 > > > > Not sure about other SDHCI based controller, but on Freescale eSDHC, > > the SDHCI_INT_CARD_INSERT bits will be immediately set again when it > > gets cleared, if a card is inserted. The driver need to mask the irq > > to prevent interrupt storm which will freeze the system. And the > > SDHCI_INT_CARD_REMOVE gets the same situation. > > > > The patch fixes the problem based on the initial idea from > > Eric Benard. > > > > Signed-off-by: Shawn Guo <shawn.guo@linaro.org> > > Cc: Eric Benard <eric@eukrea.com> > > --- > > drivers/mmc/host/sdhci.c | 27 +++++++++++++++++++++++---- > > 1 files changed, 23 insertions(+), 4 deletions(-) > > > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c > > index 91d9892..d94e2b4 100644 > > --- a/drivers/mmc/host/sdhci.c > > +++ b/drivers/mmc/host/sdhci.c > > @@ -127,7 +127,9 @@ static void sdhci_mask_irqs(struct sdhci_host *host, u32 irqs) > > > > static void sdhci_set_card_detection(struct sdhci_host *host, bool enable) > > { > > - u32 irqs = SDHCI_INT_CARD_REMOVE | SDHCI_INT_CARD_INSERT; > > + u32 present = sdhci_readl(host, SDHCI_PRESENT_STATE) & > > + SDHCI_CARD_PRESENT; > > + u32 irqs = present ? SDHCI_INT_CARD_REMOVE : SDHCI_INT_CARD_INSERT; > > > > if (host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) > > return; > > > Accessing off chip registers is slow. It would be better to check for the quirk first > and then read the present state register. > > eg > > u32 present; > u32 irqs; > > > if (host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) > > return; > > > present = sdhci_readl(host, SDHCI_PRESENT_STATE) & > > SDHCI_CARD_PRESENT; > > > > irqs = present ? SDHCI_INT_CARD_REMOVE : SDHCI_INT_CARD_INSERT; > Makes sense. Will do.
diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index 91d9892..d94e2b4 100644 --- a/drivers/mmc/host/sdhci.c +++ b/drivers/mmc/host/sdhci.c @@ -127,7 +127,9 @@ static void sdhci_mask_irqs(struct sdhci_host *host, u32 irqs) static void sdhci_set_card_detection(struct sdhci_host *host, bool enable) { - u32 irqs = SDHCI_INT_CARD_REMOVE | SDHCI_INT_CARD_INSERT; + u32 present = sdhci_readl(host, SDHCI_PRESENT_STATE) & + SDHCI_CARD_PRESENT; + u32 irqs = present ? SDHCI_INT_CARD_REMOVE : SDHCI_INT_CARD_INSERT; if (host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) return; @@ -2154,13 +2156,30 @@ static irqreturn_t sdhci_irq(int irq, void *dev_id) mmc_hostname(host->mmc), intmask); if (intmask & (SDHCI_INT_CARD_INSERT | SDHCI_INT_CARD_REMOVE)) { + u32 present = sdhci_readl(host, SDHCI_PRESENT_STATE) & + SDHCI_CARD_PRESENT; + + /* + * There is a observation on i.mx esdhc. INSERT bit will be + * immediately set again when it gets cleared, if a card is + * inserted. We have to mask the irq to prevent interrupt + * storm which will freeze the system. And the REMOVE gets + * the same situation. + * + * More testing are needed here to ensure it works for other + * platforms though. + */ + sdhci_mask_irqs(host, present ? SDHCI_INT_CARD_INSERT : + SDHCI_INT_CARD_REMOVE); + sdhci_unmask_irqs(host, present ? SDHCI_INT_CARD_REMOVE : + SDHCI_INT_CARD_INSERT); + sdhci_writel(host, intmask & (SDHCI_INT_CARD_INSERT | - SDHCI_INT_CARD_REMOVE), SDHCI_INT_STATUS); + SDHCI_INT_CARD_REMOVE), SDHCI_INT_STATUS); + intmask &= ~(SDHCI_INT_CARD_INSERT | SDHCI_INT_CARD_REMOVE); tasklet_schedule(&host->card_tasklet); } - intmask &= ~(SDHCI_INT_CARD_INSERT | SDHCI_INT_CARD_REMOVE); - if (intmask & SDHCI_INT_CMD_MASK) { sdhci_writel(host, intmask & SDHCI_INT_CMD_MASK, SDHCI_INT_STATUS);
The issue was initially found by Eric Benard as below. http://permalink.gmane.org/gmane.linux.ports.arm.kernel/108031 Not sure about other SDHCI based controller, but on Freescale eSDHC, the SDHCI_INT_CARD_INSERT bits will be immediately set again when it gets cleared, if a card is inserted. The driver need to mask the irq to prevent interrupt storm which will freeze the system. And the SDHCI_INT_CARD_REMOVE gets the same situation. The patch fixes the problem based on the initial idea from Eric Benard. Signed-off-by: Shawn Guo <shawn.guo@linaro.org> Cc: Eric Benard <eric@eukrea.com> --- drivers/mmc/host/sdhci.c | 27 +++++++++++++++++++++++---- 1 files changed, 23 insertions(+), 4 deletions(-)