This is an automated email from the ASF dual-hosted git repository.

xiaoxiang781216 pushed a commit to branch releases/13.0
in repository https://gitbox.apache.org/repos/asf/nuttx.git


The following commit(s) were added to refs/heads/releases/13.0 by this push:
     new ec9891c03a5 drivers/mmcsd: fix eMMC bus width switch sequencing
ec9891c03a5 is described below

commit ec9891c03a57eca577cd4bb6210ecca2b07bd881
Author: DuoYuWang <[email protected]>
AuthorDate: Wed Jul 22 11:56:31 2026 +0800

    drivers/mmcsd: fix eMMC bus width switch sequencing
    
    Two sequencing problems in the MMC wide bus path break eMMC 4-bit
    operation on hosts that program the bus width in the SDIO widebus /
    clock callbacks (e.g. STM32H7):
    
    1. The SWITCH command (CMD6) is issued before the host has switched
       to wide bus operation.  When the card completes the switch while
       the host is still in 1-bit mode the switch never takes effect and
       every following data transfer times out.  Switch the host to wide
       bus operation before issuing CMD6.
    
    2. The transfer clock is selected only at the end of mmcsd_widebus(),
       so the whole switch sequence runs at ID-mode clock and, on the
       affected hosts, the final clock update does not take effect either,
       leaving the bus at ~400 kHz.  Select the MMC transfer clock before
       calling mmcsd_widebus(), and pick CLOCK_MMC_TRANSFER_4BIT when wide
       bus operation is active (mirroring the SD card path) so a later
       clock selection cannot revert the host to 1-bit.
    
    No behavior change for SD cards, and no change on hosts whose widebus
    callback only records the requested state.
    
    Tested on a custom STM32H743 board with eMMC: sd_bench sequential
    write ~4.1 MB/s, sequential read ~6.3 MB/s (previously all data
    transfers timed out).
    
    Signed-off-by: DuoYuWang <[email protected]>
---
 drivers/mmcsd/mmcsd_sdio.c | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

diff --git a/drivers/mmcsd/mmcsd_sdio.c b/drivers/mmcsd/mmcsd_sdio.c
index c034f48fef5..f81412704d4 100644
--- a/drivers/mmcsd/mmcsd_sdio.c
+++ b/drivers/mmcsd/mmcsd_sdio.c
@@ -2848,8 +2848,17 @@ static int mmcsd_widebus(FAR struct mmcsd_state_s *priv)
     {
       /* Configuring MMC - Use MMC_SWITCH access modes.
        * Select 8-bit if host supports it, otherwise 4-bit.
+       *
+       * Switch the host to wide bus operation before issuing the
+       * SWITCH command: on hosts that program the bus width in the
+       * widebus callback, switching the card first leaves the switch
+       * unfinished and all following transfers fail.
        */
 
+      SDIO_WIDEBUS(priv->dev, true);
+      priv->widebus = true;
+      MMCSD_USLEEP(MMCSD_CLK_DELAY);
+
       if (priv->caps & SDIO_CAPS_8BIT)
         {
           mmcsd_sendcmdpoll(priv, MMCSD_CMD6,
@@ -2936,7 +2945,13 @@ static int mmcsd_widebus(FAR struct mmcsd_state_s *priv)
           priv->mode = EXT_CSD_HS_TIMING_HS;
         }
 
-      SDIO_CLOCK(priv->dev, CLOCK_MMC_TRANSFER);
+      /* Select the MMC transfer clocking according to the negotiated
+       * bus width, mirroring the SD card path above, so that a later
+       * clock selection cannot revert the host to 1-bit operation.
+       */
+
+      SDIO_CLOCK(priv->dev, priv->widebus ? CLOCK_MMC_TRANSFER_4BIT :
+                                            CLOCK_MMC_TRANSFER);
     }
 #endif /* #ifdef CONFIG_MMCSD_MMCSUPPORT */
 
@@ -3171,6 +3186,15 @@ static int mmcsd_mmcinitialize(FAR struct mmcsd_state_s 
*priv)
 
   mmcsd_decode_csd(priv, priv->csd);
 
+  /* Select high speed MMC clocking (which may depend on the DSR setting)
+   * before switching the bus width: on hosts that program the bus width
+   * in the clock callback, the transfer clock must already be in place
+   * before the switch sequence starts.
+   */
+
+  SDIO_CLOCK(priv->dev, CLOCK_MMC_TRANSFER);
+  MMCSD_USLEEP(MMCSD_CLK_DELAY);
+
   /* It's up to the driver to act on the widebus request.  mmcsd_widebus()
    * enables the CLOCK_MMC_TRANSFER, so call it here always.
    */

Reply via email to