| Message ID | 1573816075-26390-2-git-send-email-haibo.chen@nxp.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [01/14] mmc: sdhci: do not enable card detect interrupt for gpio cd type | expand |
On 15/11/19 1:07 PM, haibo.chen@nxp.com wrote: > From: Haibo Chen <haibo.chen@nxp.com> > > Currently, when use standard tuning, driver default disable DMA just before > send tuning command. But on i.MX8 usdhc, this is not enough. Need also clear > DMA_SEL. If not, once the DMA_SEL select AMDA2 before, even dma already disabled, > when send tuning command, usdhc will still prefetch the ADMA script from wrong > DMA address, then we will see IOMMU report some error which show lack of TLB > mapping. > > Signed-off-by: Haibo Chen <haibo.chen@nxp.com> > --- > drivers/mmc/host/sdhci.c | 12 ++++++++++++ > 1 file changed, 12 insertions(+) > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c > index 68db86c1b4c9..1436cc9c5f82 100644 > --- a/drivers/mmc/host/sdhci.c > +++ b/drivers/mmc/host/sdhci.c > @@ -2361,6 +2361,7 @@ void sdhci_send_tuning(struct sdhci_host *host, u32 opcode) > struct mmc_request mrq = {}; > unsigned long flags; > u32 b = host->sdma_boundary; > + u8 ctrl; > > spin_lock_irqsave(&host->lock, flags); > > @@ -2388,6 +2389,17 @@ void sdhci_send_tuning(struct sdhci_host *host, u32 opcode) > */ > sdhci_writew(host, SDHCI_TRNS_READ, SDHCI_TRANSFER_MODE); > > + > + /* DMA already disabled, so clear the DMA Select here. > + * Otherwise, if use ADMA2, even disable DMA, some > + * controllers like i.MX usdhc will still prefetch the > + * ADMA script when send tuning command, which will cause > + * IOMMU report lack of TLB mapping error > + */ > + ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL); > + ctrl &= ~SDHCI_CTRL_DMA_MASK; Value zero is SDMA, so this does not look like a generic sdhci change. What about doing it in sdhci-esdhc-imx.c before executing tuning? > + sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL); > + > sdhci_send_command(host, &cmd); > > host->cmd = NULL; >
> -----Original Message----- > From: Adrian Hunter <adrian.hunter@intel.com> > Sent: 2019年11月26日 19:58 > To: BOUGH CHEN <haibo.chen@nxp.com>; ulf.hansson@linaro.org > Cc: linux-mmc@vger.kernel.org; dl-linux-imx <linux-imx@nxp.com> > Subject: Re: [PATCH 02/14] mmc: sdhci: clear DMA_SEL before sending tuning > command for Standard tuning > > On 15/11/19 1:07 PM, haibo.chen@nxp.com wrote: > > From: Haibo Chen <haibo.chen@nxp.com> > > > > Currently, when use standard tuning, driver default disable DMA just > > before send tuning command. But on i.MX8 usdhc, this is not enough. > > Need also clear DMA_SEL. If not, once the DMA_SEL select AMDA2 before, > > even dma already disabled, when send tuning command, usdhc will still > > prefetch the ADMA script from wrong DMA address, then we will see > > IOMMU report some error which show lack of TLB mapping. > > > > Signed-off-by: Haibo Chen <haibo.chen@nxp.com> > > --- > > drivers/mmc/host/sdhci.c | 12 ++++++++++++ > > 1 file changed, 12 insertions(+) > > > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index > > 68db86c1b4c9..1436cc9c5f82 100644 > > --- a/drivers/mmc/host/sdhci.c > > +++ b/drivers/mmc/host/sdhci.c > > @@ -2361,6 +2361,7 @@ void sdhci_send_tuning(struct sdhci_host *host, > u32 opcode) > > struct mmc_request mrq = {}; > > unsigned long flags; > > u32 b = host->sdma_boundary; > > + u8 ctrl; > > > > spin_lock_irqsave(&host->lock, flags); > > > > @@ -2388,6 +2389,17 @@ void sdhci_send_tuning(struct sdhci_host *host, > u32 opcode) > > */ > > sdhci_writew(host, SDHCI_TRNS_READ, SDHCI_TRANSFER_MODE); > > > > + > > + /* DMA already disabled, so clear the DMA Select here. > > + * Otherwise, if use ADMA2, even disable DMA, some > > + * controllers like i.MX usdhc will still prefetch the > > + * ADMA script when send tuning command, which will cause > > + * IOMMU report lack of TLB mapping error > > + */ > > + ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL); > > + ctrl &= ~SDHCI_CTRL_DMA_MASK; > > Value zero is SDMA, so this does not look like a generic sdhci change. What > about doing it in sdhci-esdhc-imx.c before executing tuning? > Okay, the value zero for i.mx usdhc means "No DMA or Simple DMA is selected", a little bit different with generic sdhci definition. I will add this change in sdhci-esdhc-imx.c in V2 patch. Thanks! Haibo Chen > > + sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL); > > + > > sdhci_send_command(host, &cmd); > > > > host->cmd = NULL; > >
diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index 68db86c1b4c9..1436cc9c5f82 100644 --- a/drivers/mmc/host/sdhci.c +++ b/drivers/mmc/host/sdhci.c @@ -2361,6 +2361,7 @@ void sdhci_send_tuning(struct sdhci_host *host, u32 opcode) struct mmc_request mrq = {}; unsigned long flags; u32 b = host->sdma_boundary; + u8 ctrl; spin_lock_irqsave(&host->lock, flags); @@ -2388,6 +2389,17 @@ void sdhci_send_tuning(struct sdhci_host *host, u32 opcode) */ sdhci_writew(host, SDHCI_TRNS_READ, SDHCI_TRANSFER_MODE); + + /* DMA already disabled, so clear the DMA Select here. + * Otherwise, if use ADMA2, even disable DMA, some + * controllers like i.MX usdhc will still prefetch the + * ADMA script when send tuning command, which will cause + * IOMMU report lack of TLB mapping error + */ + ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL); + ctrl &= ~SDHCI_CTRL_DMA_MASK; + sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL); + sdhci_send_command(host, &cmd); host->cmd = NULL;