Re: [PATCH v1 07/11] mmc: mmci: move ST specific register extensions access under condition.

[Linus Walleij]

On Tue, Apr 29, 2014 at 10:20 AM,  <srinivas.kandagatla@xxxxxxxxxx> wrote:

> From: Srinivas Kandagatla <srinivas.kandagatla@xxxxxxxxxx>
>
> This patch moves some of the ST specific register extensions access under
> condition, so that other SOCs like Qualcomm or ARM would not a side effect of
> writing to those reserved/different purpose bits.
>
> Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@xxxxxxxxxx>
(...)

>         /* Keep ST Micro busy mode if enabled */
> -       datactrl |= host->datactrl_reg & MCI_ST_DPSM_BUSYMODE;
> +       if (host->hw_designer == AMBA_VENDOR_ST)
> +               datactrl |= host->datactrl_reg & MCI_ST_DPSM_BUSYMODE;

Do not hard-check the hw_designer everywhere, follow the pattern
if storing special stuff in the variant data.

struct variant_data {
        u32 datactrl_mask_busymode;
(...)

static struct variant_data variant_u300 = {
        .datactrl_mask_busymode = MCI_ST_DPSM_BUSYMODE,
(...)
static struct variant_data variant_nomadik = {
        .datactrl_mask_busymode = MCI_ST_DPSM_BUSYMODE,
(...)
static struct variant_data variant_ux500 = {
        .datactrl_mask_busymode = MCI_ST_DPSM_BUSYMODE,
(...)
static struct variant_data variant_ux500v2 = {
        .datactrl_mask_busymode = MCI_ST_DPSM_BUSYMODE,
(...)

Then end up like this:

>         /* Keep ST Micro busy mode if enabled */
> -       datactrl |= host->datactrl_reg & MCI_ST_DPSM_BUSYMODE;
> +      datactrl |= host->datactrl_reg & host->vendor->datactrl_mask_busymode;

OK we know this should have been done like this from the beginning
but that is not an excuse not to fix it up now.

> -       if (host->mmc->ios.timing == MMC_TIMING_UHS_DDR50)
> +       if (host->mmc->ios.timing == MMC_TIMING_UHS_DDR50 &&
> +           host->hw_designer == AMBA_VENDOR_ST)
>                 datactrl |= MCI_ST_DPSM_DDRMODE;

Same pattern here.

Actually I think this is only available on Ux500v2 (Ulf? Can you verify
this) so it is probably plain wrong to do it for other variants.

struct variant_data {
        u32 datactrl_ddrmode;
(...)
static struct variant_data variant_ux500v2 = {
        .datactrl_ddrmode = MCI_ST_DPSM_DDRMODE,
(...)

Then end up like this:

>        if (host->mmc->ios.timing == MMC_TIMING_UHS_DDR50)
> -                datactrl |= MCI_ST_DPSM_DDRMODE;
> +               datactrl |= host->vendor->datactrl_ddrmode;

Yours,
Linus Walleij
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/