Some SoCs (as seen on A20) seem to misreport the MMC FIFO level if the
FIFO is completely full: the level size reads as zero, but the FIFO_FULL
bit is set. We won't do a single iteration of the read loop in this
case, so will be stuck forever.

Check for this situation and use a safe minimal FIFO size instead when
we hit this case.

This fixes MMC boot on A20 devices after the MMC FIFO optimisation
(9faae5457f52).

Signed-off-by: Andre Przywara <andre.przyw...@arm.com>
---
 drivers/mmc/sunxi_mmc.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/drivers/mmc/sunxi_mmc.c b/drivers/mmc/sunxi_mmc.c
index 178b8cf106..aaab0cf866 100644
--- a/drivers/mmc/sunxi_mmc.c
+++ b/drivers/mmc/sunxi_mmc.c
@@ -349,10 +349,14 @@ static int mmc_trans_data_by_cpu(struct sunxi_mmc_priv 
*priv, struct mmc *mmc,
                 * register without checking the status register after every
                 * read. That saves half of the costly MMIO reads, effectively
                 * doubling the read performance.
+                * Some SoCs (A20) report a level of 0 if the FIFO is
+                * completely full (value masked out?). Use a safe minimal
+                * FIFO size in this case.
                 */
-               for (in_fifo = SUNXI_MMC_STATUS_FIFO_LEVEL(status);
-                    in_fifo > 0;
-                    in_fifo--)
+               in_fifo = SUNXI_MMC_STATUS_FIFO_LEVEL(status);
+               if (in_fifo == 0 && (status & SUNXI_MMC_STATUS_FIFO_FULL))
+                       in_fifo = 32;
+               for (; in_fifo > 0; in_fifo--)
                        buff[i++] = readl_relaxed(&priv->reg->fifo);
                dmb();
        }
-- 
2.17.1

Reply via email to