Commit f1d2736c8156 (mmc: dw_mmc: control card read threshold) added dw_mci_ctrl_rd_thld() with an unconditional write to the CDTHRCTL register at offset 0x100. However before version 240a, the FIFO region started at 0x100, so the write messes with the FIFO and completely breaks the driver. If the version id < 240A, return early from dw_mci_ctl_rd_thld() so as not to hit this problem. Fixes: f1d2736c8156 (mmc: dw_mmc: control card read threshold) Signed-off-by: James Hogan <james.hogan@xxxxxxxxxx> Cc: <stable@xxxxxxxxxxxxxxx> # v3.13+ Cc: Seungwon Jeon <tgih.jun@xxxxxxxxxxx> Cc: Jaehoon Chung <jh80.chung@xxxxxxxxxxx> Cc: Chris Ball <chris@xxxxxxxxxx> Cc: Ulf Hansson <ulf.hansson@xxxxxxxxxx> Cc: linux-mmc@xxxxxxxxxxxxxxx --- Note I've completely guessed at 240A being the point where that register was introduced since the FIFO region moved then. --- drivers/mmc/host/dw_mmc.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c index 69f0cc68d5b2..f7c95abc8c11 100644 --- a/drivers/mmc/host/dw_mmc.c +++ b/drivers/mmc/host/dw_mmc.c @@ -626,6 +626,13 @@ static void dw_mci_ctrl_rd_thld(struct dw_mci *host, struct mmc_data *data) WARN_ON(!(data->flags & MMC_DATA_READ)); + /* + * CDTHRCTL doesn't exist prior to 240A (in fact that register offset is + * in the FIFO region, so we really shouldn't access it). + */ + if (host->verid < DW_MMC_240A) + return; + if (host->timing != MMC_TIMING_MMC_HS200 && host->timing != MMC_TIMING_UHS_SDR104) goto disable; -- 2.0.4 -- To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html