From: Linus Walleij Date: Mon, 25 Jan 2010 06:14:46 +0000 (+0100) Subject: ARM: 5896/1: MMCI: work around a hardware bug in U300 X-Git-Tag: v2.6.34-rc1~241^2~4^8~9 X-Git-Url: https://git.karo-electronics.de/?a=commitdiff_plain;h=f28e8a4d027e4e21c3d0a52706527bb87397bea0;p=karo-tx-linux.git ARM: 5896/1: MMCI: work around a hardware bug in U300 In the U300 some hardware bug makes the status flag not come up signalling a successful write (or anything else, like an error, for that matter) on write requests. This little quirk makes the writes work on U300. Signed-off-by: Linus Walleij Signed-off-by: Russell King --- diff --git a/drivers/mmc/host/mmci.c b/drivers/mmc/host/mmci.c index 90d168ad03b6..643818a5ac45 100644 --- a/drivers/mmc/host/mmci.c +++ b/drivers/mmc/host/mmci.c @@ -184,6 +184,17 @@ mmci_data_irq(struct mmci_host *host, struct mmc_data *data, { if (status & MCI_DATABLOCKEND) { host->data_xfered += data->blksz; +#ifdef CONFIG_ARCH_U300 + /* + * On the U300 some signal or other is + * badly routed so that a data write does + * not properly terminate with a MCI_DATAEND + * status flag. This quirk will make writes + * work again. + */ + if (data->flags & MMC_DATA_WRITE) + status |= MCI_DATAEND; +#endif } if (status & (MCI_DATACRCFAIL|MCI_DATATIMEOUT|MCI_TXUNDERRUN|MCI_RXOVERRUN)) { if (status & MCI_DATACRCFAIL)