On 29/11/2021 13:10, Prathamesh Shete wrote:
> When CMD13 is sent after switching to HS400 mode, the bus
> is operating at either MMC_HIGH_26_MAX_DTR or MMC_HIGH_52_MAX_DTR.
> To meet Tegra SDHCI requirement at HS400 mode, force SDHCI
> interface clock to MMC_HS200_MAX_DTR (200 MHz) so that host
> controller CAR clock and the interface clock are rate matched.
This sounds like an issue with or without enhanced strobe?
Also it seems more like a fix than "adding support".
i.e. is this really "mmc: sdhci-tegra: Fix switch to HS400ES mode" ?
>
> Signed-off-by: Prathamesh Shete <pshete@xxxxxxxxxx>
> ---
> drivers/mmc/host/sdhci-tegra.c | 12 ++++++++++++
> 1 file changed, 12 insertions(+)
>
> diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
> index 387ce9cdbd7c..d800396d1112 100644
> --- a/drivers/mmc/host/sdhci-tegra.c
> +++ b/drivers/mmc/host/sdhci-tegra.c
> @@ -123,6 +123,8 @@
> SDHCI_TRNS_BLK_CNT_EN | \
> SDHCI_TRNS_DMA)
>
> +static void tegra_sdhci_set_clock(struct sdhci_host *host, unsigned int clock);
Please move one of the functions so that the forward declaration is not necessary.
> +
> struct sdhci_tegra_soc_data {
> const struct sdhci_pltfm_data *pdata;
> u64 dma_mask;
> @@ -369,6 +371,16 @@ static void tegra_sdhci_hs400_enhanced_strobe(struct mmc_host *mmc,
>
> sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_SYS_SW_CTRL);
>
> + /*
> + * When CMD13 is sent after switching to HS400 mode, the bus
> + * is operating at either MMC_HIGH_26_MAX_DTR or
> + * MMC_HIGH_52_MAX_DTR.
> + * To meet Tegra SDHCI requirement at HS400 mode, force SDHCI
> + * interface clock to MMC_HS200_MAX_DTR (200 MHz) so that host
> + * controller CAR clock and the interface clock are rate matched.
> + */
> + tegra_sdhci_set_clock(host, MMC_HS200_MAX_DTR);
This being done irrespective of ios->enhanced_strobe, which doesn't seem
right based on the comment?
> +
Please remove the blank line.
> }
>
> static void tegra_sdhci_reset(struct sdhci_host *host, u8 mask)
>
