Applied, thanks.
Best regards,
Nobuhiro
2016-04-04 18:49 GMT+09:00 Yannick Gicquel <yannick.gicq...@iot.bzh>:
> Building w/ GCC v5.2, the SD card access is broken due to invalid data
> in the response command reconstructed at the end of
> sh_sdhci_get_response().
>
> Add a memory barrier between the two main steps of this function to
> ensure the resp[] table content is consistent before bits reordering.
>
> This fix has been tested Ok on Porter board rev1.0 using v2016.03
> release.
>
> Signed-off-by: Yannick Gicquel <yannick.gicq...@iot.bzh>
> ---
> drivers/mmc/sh_sdhi.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/mmc/sh_sdhi.c b/drivers/mmc/sh_sdhi.c
> index 30e538c..dc4c686 100644
> --- a/drivers/mmc/sh_sdhi.c
> +++ b/drivers/mmc/sh_sdhi.c
> @@ -430,6 +430,9 @@ static void sh_sdhi_get_response(struct sh_sdhi_host
> *host, struct mmc_cmd *cmd)
> }
>
> p2 = (unsigned long *)cmd->response;
> +
> + barrier();
> +
> #if defined(__BIG_ENDIAN_BITFIELD)
> for (i = 0; i < cnt; i++) {
> *p2++ = ((*p1 >> 16) & 0x0000ffff) |
> --
> 1.9.1
>
--
Nobuhiro Iwamatsu
iwamatsu at {nigauri.org / debian.org}
GPG ID: 40AD1FA6
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
