Hi Shimoda-san,

> We should set the needs_adjust_hs400 when HS400EN=1 is set.
> In other words, we should set it in renesas_sdhi_hs400_complete(),
> not renesas_sdhi_prepare_hs400_tuning().

Oh, thanks for pointing this out! I simply put it into the same function
as the BSP, but due to massive refactoring in the upstream driver we
really need to put it where tuning is completed. I used the following
patch on top of this patch:

diff --git b/drivers/mmc/host/renesas_sdhi_core.c a/drivers/mmc/host/renesas_sdhi_core.c
index 2834b30c3ba6..b54dd8a7ecaa 100644
--- b/drivers/mmc/host/renesas_sdhi_core.c
+++ a/drivers/mmc/host/renesas_sdhi_core.c
@@ -419,6 +419,9 @@ static void renesas_sdhi_hs400_complete(struct mmc_host *mmc)
 
 	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, CLK_CTL_SCLKEN |
 			sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+	if (priv->adjust_hs400_calib_table)
+		priv->needs_adjust_hs400 = true;
 }
 
 static void renesas_sdhi_reset_scc(struct tmio_mmc_host *host,
@@ -544,12 +547,8 @@ static void renesas_sdhi_reset_hs400_mode(struct tmio_mmc_host *host,
 static int renesas_sdhi_prepare_hs400_tuning(struct mmc_host *mmc, struct mmc_ios *ios)
 {
 	struct tmio_mmc_host *host = mmc_priv(mmc);
-	struct renesas_sdhi *priv = host_to_priv(host);
 
-	if (priv->adjust_hs400_calib_table)
-		priv->needs_adjust_hs400 = true;
-
-	renesas_sdhi_reset_hs400_mode(host, priv);
+	renesas_sdhi_reset_hs400_mode(host, host_to_priv(host));
 	return 0;
 }

and got these debug reports from my M3-N:

     kworker/0:1-21    [000] ....     4.171435: renesas_sdhi_fixup_request: code 17 replacement 18

which was different without the fix:

     kworker/0:1-21    [000] ....     4.171435: renesas_sdhi_fixup_request: code 31 replacement 31

> When we read the calib_code at room temperature, the value will be around 0x10.
> However, we will get 0x00 here. So, we need to fix it.

So, my new debug output looks good.

I will fold the fixup patch into the next version of this series. I
still need to get the stalled-SCC issue tackled first. But I am now
working on it.

Thanks again,

   Wolfram