On Wed, Dec 11, 2013 at 11:03:31AM +0800, Dong Aisheng wrote:

quoted

quoted

+static void sdhci_prepare_data(struct sdhci_host *host, struct mmc_command *cmd)
+{
      u8 ctrl;
      struct mmc_data *data = cmd->data;
      int ret;

      WARN_ON(host->data);

-     if (data || (cmd->flags & MMC_RSP_BUSY)) {
-             count = sdhci_calc_timeout(host, cmd);

From what I read the commit log, I think it might be more appropriate to
patch sdhci_calc_timeout() like the following?

        if (host->quirks & SDHCI_QUIRK_BROKEN_TIMEOUT_VAL)
                if (host->ops->get_max_timeout)
                        return host->ops->get_max_timeout(host);
                else
                        return 0xE;

The return val of sdhci_calc_timeout is the register value to be
written into timeout counter register.
host->ops->get_max_timeout returns the max timeout in miliseconds directly.
So we can not do like that here.
They're two concepts.

I did not make my comment clear.  The .get_max_timeout() is not the one
you defined in patch #1 any more, but something like the following for
esdhc driver, which returns correct timeout value not milliseconds.

unsigned int esdhc_get_max_timeout(struct sdhci_host *host)
{
	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
	struct pltfm_imx_data *imx_data = pltfm_host->priv;

	return esdhc_is_usdhc(imx_data) ? 0xF : 0xE;

}

Does that match the problem you described in the commit log better?

Shawn