[PATCH] dw_mmc: ensure the ciu-reset

[PATCH] dw_mmc: ensure the ciu-reset

[Jaehoon Chung]

This patch ensured the CIU-reset.

If data0 line is low, bit[9] in status register is set to 1.
Then we can know the card is busy.

Using this bit in status register, we can ensure to reset the CIU correctly.
When card is busy and doing ciu-reset, seem like reset completed.
But actually didn't complete ciu-reset. because card is busy.

So i think good whether check card busy or not before reset the ciu.


Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Reported-by: Wonil Choi <wonil22.choi@samsung.com>
---
 drivers/mmc/host/dw_mmc.c |   15 +++++++++------
 drivers/mmc/host/dw_mmc.h |    3 +++
 2 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
index 0ed1d28..277f8ef 100644
--- a/drivers/mmc/host/dw_mmc.c
+++ b/drivers/mmc/host/dw_mmc.c
@@ -1785,17 +1785,20 @@ no_dma:
 static bool mci_wait_reset(struct device *dev, struct dw_mci *host)
 {
 	unsigned long timeout = jiffies + msecs_to_jiffies(500);
-	unsigned int ctrl;
+	unsigned int ctrl, status;
 
-	mci_writel(host, CTRL, (SDMMC_CTRL_RESET | SDMMC_CTRL_FIFO_RESET |
-				SDMMC_CTRL_DMA_RESET));
+	mci_writel(host, CTRL, SDMMC_CTRL_RESET_ALL);
 
 	/* wait till resets clear */
 	do {
 		ctrl = mci_readl(host, CTRL);
-		if (!(ctrl & (SDMMC_CTRL_RESET | SDMMC_CTRL_FIFO_RESET |
-			      SDMMC_CTRL_DMA_RESET)))
-			return true;
+		if (!(ctrl & SDMMC_CTRL_RESET_ALL)) {
+			status = mci_readl(host, STATUS);
+			if (!(status & SDMMC_DATA_BUSY))
+				return true;
+			else
+				mci_writel(host, CTRL, SDMMC_CTRL_RESET_ALL);
+		}
 	} while (time_before(jiffies, timeout));
 
 	dev_err(dev, "Timeout resetting block (ctrl %#x)\n", ctrl);
diff --git a/drivers/mmc/host/dw_mmc.h b/drivers/mmc/host/dw_mmc.h
index bfa3c1c..94e48ad 100644
--- a/drivers/mmc/host/dw_mmc.h
+++ b/drivers/mmc/host/dw_mmc.h
@@ -69,6 +69,8 @@
 #define SDMMC_CTRL_DMA_RESET		BIT(2)
 #define SDMMC_CTRL_FIFO_RESET		BIT(1)
 #define SDMMC_CTRL_RESET		BIT(0)
+#define SDMMC_CTRL_RESET_ALL	(SDMMC_CTRL_DMA_RESET |\
+			SDMMC_CTRL_FIFO_RESET | SDMMC_CTRL_RESET)
 /* Clock Enable register defines */
 #define SDMMC_CLKEN_LOW_PWR		BIT(16)
 #define SDMMC_CLKEN_ENABLE		BIT(0)
@@ -118,6 +120,7 @@
 #define SDMMC_CMD_INDX(n)		((n) & 0x1F)
 /* Status register defines */
 #define SDMMC_GET_FCNT(x)		(((x)>>17) & 0x1FF)
+#define SDMMC_DATA_BUSY			BIT(9)
 /* Internal DMAC interrupt defines */
 #define SDMMC_IDMAC_INT_AI		BIT(9)
 #define SDMMC_IDMAC_INT_NI		BIT(8)

Re: [PATCH] dw_mmc: ensure the ciu-reset

[Will Newton]

On Mon, Sep 5, 2011 at 1:39 AM, Jaehoon Chung <jh80.chung@samsung.com> wrote:

Hi Jaehoon,

> This patch ensured the CIU-reset.
>
> If data0 line is low, bit[9] in status register is set to 1.
> Then we can know the card is busy.
>
> Using this bit in status register, we can ensure to reset the CIU correctly.
> When card is busy and doing ciu-reset, seem like reset completed.
> But actually didn't complete ciu-reset. because card is busy.
>
> So i think good whether check card busy or not before reset the ciu.

It looks like this isn't exactly what the patch does - the reset
happens whether the card is busy or not, but if card data is busy then
the reset may be re-issued.

Is there some way to reproduce the issue in this patch? I can't find
anything in the TRM that suggests that we need to issue the reset more
than once.

Thanks,

Re: [PATCH] dw_mmc: ensure the ciu-reset

[Jaehoon Chung]

Hi Will

Will Newton wrote:

> On Mon, Sep 5, 2011 at 1:39 AM, Jaehoon Chung <jh80.chung@samsung.com> wrote:
> 
> Hi Jaehoon,
> 
>> This patch ensured the CIU-reset.
>>
>> If data0 line is low, bit[9] in status register is set to 1.
>> Then we can know the card is busy.
>>
>> Using this bit in status register, we can ensure to reset the CIU correctly.
>> When card is busy and doing ciu-reset, seem like reset completed.
>> But actually didn't complete ciu-reset. because card is busy.
>>
>> So i think good whether check card busy or not before reset the ciu.
> 
> It looks like this isn't exactly what the patch does - the reset
> happens whether the card is busy or not, but if card data is busy then
> the reset may be re-issued.

the reset is happen regardless of the card is busy or not.
Maybe looks like the reset is completed. and no more issued the reset.
But when card is bus and reset completed, maybe next sequence didn't run.
(next sequence is maybe CMD0 issued)
This patch should be prevent not to run next sequence.

> 
> Is there some way to reproduce the issue in this patch? I can't find
> anything in the TRM that suggests that we need to issue the reset more
> than once.


i think this problem is similar to Software/Hardware restrictions.
(in spec, mentioned that have to check the bit[9] in status register before clock change)
Also in this case, to ensure that card is correctly running,
i think that need the more reset processing than once.

in the most case, maybe card status is not busy. then perform to reset only one-time

Best regards,
Jaehoon Chung

> Thanks,
> --
> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>