"Card stuck in programming state! __mmc_switch" problem on i.MX6 + eMMC

Hello

I have a custom i.MX6Q based board with a Hynix eMMC chip attached to USDHC3. OS is a vanilla Linux 3.16 with some patches.

Loading the iMX SDHC driver results in the following error:

modprobe sdhci-esdhc-imx
sdhci: Secure Digital Host Controller Interface driver
sdhci: Copyright(c) Pierre Ossman
sdhci-pltfm: SDHCI platform and OF driver helper
mmc0: no vqmmc regulator found
mmc0: SDHCI controller on 2198000.usdhc [2198000.usdhc] using ADMA
mmc0: BKOPS_EN bit is not set
mmc0: Card stuck in programming state! __mmc_switch
mmc0: error -110 whilst initialising MMC card
mmc0: BKOPS_EN bit is not set
mmc0: Card stuck in programming state! __mmc_switch
mmc0: error -110 whilst initialising MMC card
mmc0: BKOPS_EN bit is not set
mmc0: Card stuck in programming state! __mmc_switch
mmc0: error -110 whilst initialising MMC card
mmc0: BKOPS_EN bit is not set
mmc0: Card stuck in programming state! __mmc_switch
mmc0: error -110 whilst initialising MMC card

Pin muxing looks ok nevertheless I played around with it to see if it makes a difference. I also tried the "Add delay between CMD6 and CMD13 for Sandisk eMMC cards" workaround previously mentioned on this list.

As the whole thing is working under U-Boot as shown below, I tend to rule out soldering problems and non-functional hardware.

=> ver
U-Boot 2014.07-00045-g0955ecc-dirty (Aug 26 2014 - 12:47:21) => mmc dev 1
switch to partitions #0, OK
mmc1(part 0) is current device
=> mmc info
Device: FSL_SDHC
Manufacturer ID: 90
OEM: 14a
Name: H8G2d
Tran Speed: 52000000
Rd Block Len: 512
MMC version 4.41
High Capacity: Yes
Capacity: 7.2 GiB
Bus Width: 8-bit

Any idea what is going wrong?
Has anybody a Hynix eMMC working under Linux?

Hi David,

2014-09-16 16:37 GMT+02:00 David Müller (ELSOFT AG) <[email protected]>:

Can you show your DT (at least pinmux part and usdhc3 node) ? -110 is a timeout error and I know there is some quirks in Freescale repository I don't know if this is your best bet...

JM
PS: Sorry for double posting, forgot linux-mmc in cc

Hi Jean-Michel

	pinctrl_usdhc3: usdhc3grp {
		fsl,pins = <
			MX6QDL_PAD_SD3_CMD__SD3_CMD    0x1b0b1
			MX6QDL_PAD_SD3_CLK__SD3_CLK    0x100b1
			MX6QDL_PAD_SD3_DAT0__SD3_DATA0 0x1b0b1
			MX6QDL_PAD_SD3_DAT1__SD3_DATA1 0x1b0b1
			MX6QDL_PAD_SD3_DAT2__SD3_DATA2 0x1b0b1
			MX6QDL_PAD_SD3_DAT3__SD3_DATA3 0x1b0b1
			MX6QDL_PAD_SD3_DAT4__SD3_DATA4 0x1b0b1
			MX6QDL_PAD_SD3_DAT5__SD3_DATA5 0x1b0b1
			MX6QDL_PAD_SD3_DAT6__SD3_DATA6 0x1b0b1
			MX6QDL_PAD_SD3_DAT7__SD3_DATA7 0x1b0b1
			MX6QDL_PAD_SD3_RST__SD3_RESET  0x1b070
		>;
	};

&usdhc3 { /* eMMC */

	pinctrl-names = "default";
	pinctrl-0 = <&pinctrl_usdhc3>;
	vmmc-supply = <&reg_3v15>;
	non-removable;
	no-1-8-v;
	bus-width = <8>;
	status = "okay";
};

I've tracked the problem down to the point were the MMC code tries to switch the eMMC to 8bit bus width (mmc_select_bus_width()).

Hi David,

I remember having similar issues and they were fixed by this series:
http://www.spinics.net/lists/linux-mmc/msg28260.html

You can try linux-next as this series is already applied there already.