Commit 1b673ef breaks riotboard boot

Commit 1b673ef breaks riotboard boot

[Nikolay Dimitrov]

Hi Fabio,

Commit 1b673ef (linux-fslc, branch "patch-4.0") breaks booting of
riotboard. Reverting the patch allows the board to boot properly. It
looks like the patch causes some issue with the sdcard detection, as
shown below in the logs (only the emmc is detected):

[    1.447707] sdhci: Secure Digital Host Controller Interface driver
[    1.453974] sdhci: Copyright(c) Pierre Ossman
[    1.458488] sdhci-pltfm: SDHCI platform and OF driver helper
[    1.465423] sdhci-esdhc-imx 2194000.usdhc: could not get ultra high 
speed state, work on normal mode
[    1.474725] sdhci-esdhc-imx 2194000.usdhc: Got CD GPIO
[    1.479985] sdhci-esdhc-imx 2194000.usdhc: Got WP GPIO
[    1.485574] sdhci-esdhc-imx 2194000.usdhc: No vqmmc regulator found
[    1.539993] mmc0: SDHCI controller on 2194000.usdhc [2194000.usdhc] 
using ADMA
[    1.551185] sdhci-esdhc-imx 2198000.usdhc: could not get ultra high 
speed state, work on normal mode
[    1.574222] sdhci-esdhc-imx 2198000.usdhc: Got CD GPIO
[    1.589766] sdhci-esdhc-imx 2198000.usdhc: Got WP GPIO
[    1.606144] sdhci-esdhc-imx 2198000.usdhc: No vqmmc regulator found
...
[    2.408926] VFS: Cannot open root device "mmcblk0p2" or 
unknown-block(179,2): error -6
[    2.416860] Please append a correct "root=" boot option; here are the 
available partitions:
...
[    2.500580] b300         3751936 mmcblk0  driver: mmcblk
[    2.505927] b318             128 mmcblk0rpmb  (driver?)
[    2.511211] b310            2048 mmcblk0boot1  (driver?)
[    2.516554] b308            2048 mmcblk0boot0  (driver?)
[    2.522003] Kernel panic - not syncing: VFS: Unable to mount root fs 
on unknown-block(179,2)
[    2.530452] drm_kms_helper: panic occurred, switching back to text 
console
[    2.537649] ---[ end Kernel panic - not syncing: VFS: Unable to mount 
root fs on unknown-block(179,2)

Thanks in advance for looking into this. Regards,
Nikolay

Re: Commit 1b673ef breaks riotboard boot

[Fabio Estevam]

Hi Nikolay,

On Mon, Apr 27, 2015 at 1:33 AM, Nikolay Dimitrov <picmaster@mail.bg> wrote:
> Hi Fabio,
>
> Commit 1b673ef (linux-fslc, branch "patch-4.0") breaks booting of
> riotboard. Reverting the patch allows the board to boot properly. It
> looks like the patch causes some issue with the sdcard detection, as
> shown below in the logs (only the emmc is detected):

Thanks for reporting it. I don't have access to a riotboard, so will
need some help to debugging this problem.

The commit you pointed out has reached 4.1-rc1 now as 8d86e4fcccf61bafe.

This commit has been in linux-next for sometime and I haven't seen
such behaviour in the boards I have access to. Also, Olof's boot farm
contains hummingboard and wandboard and they are booting fine in
linux-next and 4.1-rc1. Please check:
http://arm-soc.lixom.net/bootlogs/mainline/v4.1-rc1/wandboard-arm-imx_v6_v7_defconfig.html

Could you please try booting 4.1-rc1 on riotboard? I would like to
understand if this issue only happens on linux-fslc branch or not.

>
> [    1.447707] sdhci: Secure Digital Host Controller Interface driver
> [    1.453974] sdhci: Copyright(c) Pierre Ossman
> [    1.458488] sdhci-pltfm: SDHCI platform and OF driver helper
> [    1.465423] sdhci-esdhc-imx 2194000.usdhc: could not get ultra high speed
> state, work on normal mode
> [    1.474725] sdhci-esdhc-imx 2194000.usdhc: Got CD GPIO
> [    1.479985] sdhci-esdhc-imx 2194000.usdhc: Got WP GPIO

CD and WP GPIO were retrieved here.

> [    1.485574] sdhci-esdhc-imx 2194000.usdhc: No vqmmc regulator found
> [    1.539993] mmc0: SDHCI controller on 2194000.usdhc [2194000.usdhc] using
> ADMA
> [    1.551185] sdhci-esdhc-imx 2198000.usdhc: could not get ultra high speed
> state, work on normal mode
> [    1.574222] sdhci-esdhc-imx 2198000.usdhc: Got CD GPIO
> [    1.589766] sdhci-esdhc-imx 2198000.usdhc: Got WP GPIO

Same here.

> [    1.606144] sdhci-esdhc-imx 2198000.usdhc: No vqmmc regulator found
> ...
> [    2.408926] VFS: Cannot open root device "mmcblk0p2" or
> unknown-block(179,2): error -6

Could this be related to the old mmc probe order issue? What is the
mmc port you want to mount the rootfs? Has it changed to mmcblk1?

Also, could you please remove 'wp-gpios' from the dts to see if this
makes any difference?

Thanks,

Fabio Estevam

Re: Commit 1b673ef breaks riotboard boot

[Gary Thomas]

On 2015-04-27 06:46, Fabio Estevam wrote:
> Hi Nikolay,
>
> On Mon, Apr 27, 2015 at 1:33 AM, Nikolay Dimitrov <picmaster@mail.bg> wrote:
>> Hi Fabio,
>>
>> Commit 1b673ef (linux-fslc, branch "patch-4.0") breaks booting of
>> riotboard. Reverting the patch allows the board to boot properly. It
>> looks like the patch causes some issue with the sdcard detection, as
>> shown below in the logs (only the emmc is detected):
>
> Thanks for reporting it. I don't have access to a riotboard, so will
> need some help to debugging this problem.
>
> The commit you pointed out has reached 4.1-rc1 now as 8d86e4fcccf61bafe.
>
> This commit has been in linux-next for sometime and I haven't seen
> such behaviour in the boards I have access to. Also, Olof's boot farm
> contains hummingboard and wandboard and they are booting fine in
> linux-next and 4.1-rc1. Please check:
> http://arm-soc.lixom.net/bootlogs/mainline/v4.1-rc1/wandboard-arm-imx_v6_v7_defconfig.html
>
> Could you please try booting 4.1-rc1 on riotboard? I would like to
> understand if this issue only happens on linux-fslc branch or not.

Linux mainline 4.1-rc1 booted fine out of the box on my RIotBoard, using mmc0 (full size
SD on bottom of board) as root.

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

Re: Commit 1b673ef breaks riotboard boot

[Fabio Estevam]

Hi Gary,

On Mon, Apr 27, 2015 at 11:45 AM, Gary Thomas <gary@mlbassoc.com> wrote:

>> Could you please try booting 4.1-rc1 on riotboard? I would like to
>> understand if this issue only happens on linux-fslc branch or not.
>
>
> Linux mainline 4.1-rc1 booted fine out of the box on my RIotBoard, using
> mmc0 (full size
> SD on bottom of board) as root.

Thanks for testing it. I think this is good news as it shows we do not
have such regression in mainline.

Would it be possible to send us the full boot logs when booting
4.1-rc1 and 4.0 linux-fslc, so that I can compare them?

Thanks,

Fabio Estevam

Re: Commit 1b673ef breaks riotboard boot

[Gary Thomas]

[-- Attachment #1: Type: text/plain, Size: 1011 bytes --]

On 2015-04-27 09:02, Fabio Estevam wrote:
> Hi Gary,
>
> On Mon, Apr 27, 2015 at 11:45 AM, Gary Thomas <gary@mlbassoc.com> wrote:
>
>>> Could you please try booting 4.1-rc1 on riotboard? I would like to
>>> understand if this issue only happens on linux-fslc branch or not.
>>
>>
>> Linux mainline 4.1-rc1 booted fine out of the box on my RIotBoard, using
>> mmc0 (full size
>> SD on bottom of board) as root.
>
> Thanks for testing it. I think this is good news as it shows we do not
> have such regression in mainline.
>
> Would it be possible to send us the full boot logs when booting
> 4.1-rc1 and 4.0 linux-fslc, so that I can compare them?
>
> Thanks,
>
> Fabio Estevam

Logs attached
   boot-4.0: linux-fslc
   boot-4.1: linux-mainline (4.1-rc1)

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

[-- Attachment #2: boot-4.0 --]
[-- Type: application/octet-stream, Size: 18732 bytes --]

[-- Attachment #3: boot-4.1 --]
[-- Type: application/octet-stream, Size: 21400 bytes --]

Re: Commit 1b673ef breaks riotboard boot

[Fabio Estevam]

Hi Gary / Nikolay,

On Mon, Apr 27, 2015 at 1:31 PM, Gary Thomas <gary@mlbassoc.com> wrote:

> Logs attached
>   boot-4.0: linux-fslc
>   boot-4.1: linux-mainline (4.1-rc1)

I was able to reproduce the same behaviour on a mx6qsabresd, so I hope
I can debug this tomorrow.

Regards,

Fabio Estevam

Re: Commit 1b673ef breaks riotboard boot

[Fabio Estevam]

[-- Attachment #1: Type: text/plain, Size: 358 bytes --]

Hi Gary / Nikolay,

On Tue, Apr 28, 2015 at 1:25 AM, Fabio Estevam <festevam@gmail.com> wrote:

> I was able to reproduce the same behaviour on a mx6qsabresd, so I hope
> I can debug this tomorrow.

Please try the attached patch and let us know if it works on riotboard.

I will also test it on warp board later today.

Regards,

Fabio Estevam

[-- Attachment #2: 0001-sdhcfixfor4.0.patch --]
[-- Type: text/plain, Size: 2652 bytes --]

From 3bda6d618a6bab3c0d7836a8f8d4f7573eec3e36 Mon Sep 17 00:00:00 2001
From: Fabio Estevam <fabio.estevam@freescale.com>
Date: Tue, 28 Apr 2015 02:07:30 -0300
Subject: [PATCH] mmc: sdhci-esdhc-imx: Avoid calling mmc_of_parse()

This is a partial revert from 1b673ef39236("mmc: sdhci-esdhc-imx: Call 
mmc_of_parse()"), which calls mmc_pwrseq_alloc() instead of the recommeded
mmc_of_parse() to allow powering up Wifi/BT gpios connected via esdhc.

Such commit is causing SD cards to not getting detected.

This is a quick workaround for kernel 4.0, as 4.1-rc1 does not show this
problem.

Tested on a imx6q-sabresd.

Reported-by: Nikolay Dimitrov <picmaster@mail.bg>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
---
 drivers/mmc/host/sdhci-esdhc-imx.c | 39 +++++++++++++++++++++++++++++++++++---
 1 file changed, 36 insertions(+), 3 deletions(-)

diff --git a/drivers/mmc/host/sdhci-esdhc-imx.c b/drivers/mmc/host/sdhci-esdhc-imx.c
index 1690d0b..59882b2 100644
--- a/drivers/mmc/host/sdhci-esdhc-imx.c
+++ b/drivers/mmc/host/sdhci-esdhc-imx.c
@@ -31,6 +31,8 @@
 #include "sdhci-pltfm.h"
 #include "sdhci-esdhc.h"
 
+#include "../core/pwrseq.h"
+
 #define	ESDHC_CTRL_D3CD			0x08
 /* VENDOR SPEC register */
 #define ESDHC_VENDOR_SPEC		0xc0
@@ -1009,9 +1011,40 @@ static int sdhci_esdhc_imx_probe(struct platform_device *pdev)
 					host->mmc->parent->platform_data);
 	}
 
+	/* write_protect */
+	if (boarddata->wp_type == ESDHC_WP_GPIO) {
+		err = mmc_gpio_request_ro(host->mmc, boarddata->wp_gpio);
+		if (err) {
+			dev_err(mmc_dev(host->mmc),
+				"failed to request write-protect gpio!\n");
+			goto disable_clk;
+		}
+		host->mmc->caps2 |= MMC_CAP2_RO_ACTIVE_HIGH;
+	}
+
 	/* card_detect */
-	if (boarddata->cd_type == ESDHC_CD_CONTROLLER)
+	switch (boarddata->cd_type) {
+	case ESDHC_CD_GPIO:
+		err = mmc_gpio_request_cd(host->mmc, boarddata->cd_gpio, 0);
+		if (err) {
+			dev_err(mmc_dev(host->mmc),
+				"failed to request card-detect gpio!\n");
+			goto disable_clk;
+		}
+		/* fall through */
+
+	case ESDHC_CD_CONTROLLER:
+		/* we have a working card_detect back */
 		host->quirks &= ~SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+		break;
+
+	case ESDHC_CD_PERMANENT:
+		host->mmc->caps |= MMC_CAP_NONREMOVABLE;
+		break;
+
+	case ESDHC_CD_NONE:
+		break;
+	}
 
 	switch (boarddata->max_bus_width) {
 	case 8:
@@ -1044,8 +1077,8 @@ static int sdhci_esdhc_imx_probe(struct platform_device *pdev)
 		host->quirks2 |= SDHCI_QUIRK2_NO_1_8_V;
 	}
 
-	/* call to generic mmc_of_parse to support additional capabilities */
-	err = mmc_of_parse(host->mmc);
+
+	err = mmc_pwrseq_alloc(host->mmc);
 	if (err)
 		goto disable_clk;
 
-- 
1.9.1

Re: Commit 1b673ef breaks riotboard boot

[Gary Thomas]

On 2015-04-28 04:42, Fabio Estevam wrote:
> Hi Gary / Nikolay,
>
> On Tue, Apr 28, 2015 at 1:25 AM, Fabio Estevam <festevam@gmail.com> wrote:
>
>> I was able to reproduce the same behaviour on a mx6qsabresd, so I hope
>> I can debug this tomorrow.
>
> Please try the attached patch and let us know if it works on riotboard.
>
> I will also test it on warp board later today.

Yes, this patch allows the RIoTboard to boot.

Thanks

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------