On 10/03/2018 04:17 AM, Masahiro Yamada wrote:
On Wed, Oct 3, 2018 at 7:54 AM Marek Vasut marek.vasut@gmail.com wrote:
Pass the entire source data pointer to tmio_sd_addr_is_dmaable() to avoid losing top 32 bits on 64bit systems.
Really?
sizeof(long) is 8 on 64bit systems.
Isn't that long long that's guaranteed to be 64bit ? Anyway, this looks cleaner to me and removes one cast from the code.
(In other words, long and (void *) have the same size)
Why is the top 32-bits lost?
Signed-off-by: Marek Vasut marek.vasut+renesas@gmail.com Cc: Masahiro Yamada yamada.masahiro@socionext.com
drivers/mmc/tmio-common.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/mmc/tmio-common.c b/drivers/mmc/tmio-common.c index b311b80be8..6b21941991 100644 --- a/drivers/mmc/tmio-common.c +++ b/drivers/mmc/tmio-common.c @@ -372,8 +372,10 @@ static int tmio_sd_dma_xfer(struct udevice *dev, struct mmc_data *data) }
/* check if the address is DMA'able */ -static bool tmio_sd_addr_is_dmaable(unsigned long addr) +static bool tmio_sd_addr_is_dmaable(const char *src) {
uintptr_t addr = (uintptr_t)src;if (!IS_ALIGNED(addr, TMIO_SD_DMA_MINALIGN)) return false;@@ -486,7 +488,7 @@ int tmio_sd_send_cmd(struct udevice *dev, struct mmc_cmd *cmd, if (data) { /* use DMA if the HW supports it and the buffer is aligned */ if (priv->caps & TMIO_SD_CAP_DMA_INTERNAL &&
tmio_sd_addr_is_dmaable((long)data->src))
tmio_sd_addr_is_dmaable(data->src)) ret = tmio_sd_dma_xfer(dev, data); else ret = tmio_sd_pio_xfer(dev, data);-- 2.18.0
U-Boot mailing list U-Boot@lists.denx.de https://lists.denx.de/listinfo/u-boot