On Sun, Apr 29 2018, Christian Lütke-Stetzkamp wrote:
> The function msdc_set_blknum consists of one (real) line of code and
> is only called once, so inline it makes the code shorter and more
> readable.
>
> Signed-off-by: Christian Lütke-Stetzkamp <christ...@lkamp.de>
> ---
> drivers/staging/mt7621-mmc/sd.c | 10 +---------
> 1 file changed, 1 insertion(+), 9 deletions(-)
>
> diff --git a/drivers/staging/mt7621-mmc/sd.c b/drivers/staging/mt7621-mmc/sd.c
> index c9f0180f3220..00c862b02419 100644
> --- a/drivers/staging/mt7621-mmc/sd.c
> +++ b/drivers/staging/mt7621-mmc/sd.c
> @@ -1182,14 +1182,6 @@ static void msdc_dma_setup(struct msdc_host *host,
> struct msdc_dma *dma,
> msdc_dma_config(host, dma);
> }
>
> -/* set block number before send command */
> -static void msdc_set_blknum(struct msdc_host *host, u32 blknum)
> -{
> - void __iomem *base = host->base;
> -
> - sdr_write32(SDC_BLK_NUM, blknum);
> -}
> -
This change is correct, but the reason isn't obvious.
In mt6575_sd.h are:
#define SDC_BLK_NUM REG_ADDR(SDC_BLK_NUM)
#define REG_ADDR(x) (base + OFFSET_##x)
So the above becomes
sdr_write32((base + OFFSET_SDC_BLK_NUM), blknum);
so the base is important.
msdc_do_request() starts with
void __iomem *base = host->base;
and never changes host. So that "base" that is used in the new code is
the same as the "base" used in the old code.
Ugh.. what a horrible way to write code!
Thanks,
NeilBrown
> static int msdc_do_request(struct mmc_host *mmc, struct mmc_request *mrq)
> __must_hold(&host->lock)
> {
> @@ -1239,7 +1231,7 @@ static int msdc_do_request(struct mmc_host *mmc, struct
> mmc_request *mrq)
> }
> }
>
> - msdc_set_blknum(host, data->blocks);
> + sdr_write32(SDC_BLK_NUM, data->blocks);
> //msdc_clr_fifo(); /* no need */
>
> msdc_dma_on(); /* enable DMA mode first!! */
> --
> 2.16.1
signature.asc
Description: PGP signature
_______________________________________________ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
