Skip to contentSkip to conversation

mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

PATCH Superseded Fabio Estevam 5 participants 10 messages Nov 19, 2018 – Jan 14, 2019 MMC, SD and SDIO

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

The following hang is observed on a Hummingboard 2 MicroSOM i2eX iMX6D - rev 1.3 with no eMMC populated on board:

U-Boot SPL 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000) Trying to boot from MMC1

U-Boot 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)

CPU:   Freescale i.MX6D rev1.5 996 MHz (running at 792 MHz)
CPU:   Extended Commercial temperature grade (-20C to 105C) at 33C
Reset cause: POR
Board: MX6 HummingBoard2
DRAM:  1 GiB
MMC:   FSL_SDHC: 0, FSL_SDHC: 1

Loading Environment from MMC... *** Warning - bad CRC, using default environment

No panel detected: default to HDMI
Display: HDMI (1024x768)
In:    serial
Out:   serial
Err:   serial

---> hangs

which is caused by the following infinite loop inside esdhc_send_cmd_common()

	while (!(esdhc_read32(&regs->irqstat) & flags))
		;

Instead of looping forever, provide an exit path so that a timeout error can be propagated in the case irqstat does not report any interrupts, which may happen when no eMMC is populated on board.

---
Changes since v1:
- Jump to the label out on error path (Baruch).

 drivers/mmc/fsl_esdhc.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/mmc/fsl_esdhc.c +8 -2
index 3cdfa7f..b8d0b00 100644
@@ -396,6 +396,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,
 	uint	irqstat;
 	u32	flags = IRQSTAT_CC | IRQSTAT_CTOE;
 	struct fsl_esdhc *regs = priv->esdhc_regs;
+	unsigned long start;
 
 #ifdef CONFIG_SYS_FSL_ERRATUM_ESDHC111
 	if (cmd->cmdidx == MMC_CMD_STOP_TRANSMISSION)
@@ -453,8 +454,13 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,
 		flags = IRQSTAT_BRR;
 
 	/* Wait for the command to complete */
-	while (!(esdhc_read32(&regs->irqstat) & flags))
-		;
+	start = get_timer(0);
+	while (!(esdhc_read32(&regs->irqstat) & flags)) {
+		if (get_timer(start) > 1000) {
+			err = -ETIMEDOUT;
+			goto out;
+		}
+	}
 
 	irqstat = esdhc_read32(&regs->irqstat);
 
Signature · 1 line
2.7.4
#1 lore raw

PATCH v2 1/1 mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

Quoted 29 lines · Fabio Estevam

The following hang is observed on a Hummingboard 2 MicroSOM i2eX iMX6D - rev 1.3 with no eMMC populated on board:

U-Boot SPL 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000) Trying to boot from MMC1

U-Boot 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)

CPU:   Freescale i.MX6D rev1.5 996 MHz (running at 792 MHz)
CPU:   Extended Commercial temperature grade (-20C to 105C) at 33C
Reset cause: POR
Board: MX6 HummingBoard2
DRAM:  1 GiB
MMC:   FSL_SDHC: 0, FSL_SDHC: 1

Loading Environment from MMC... *** Warning - bad CRC, using default environment

No panel detected: default to HDMI
Display: HDMI (1024x768)
In:    serial
Out:   serial
Err:   serial

---> hangs

which is caused by the following infinite loop inside esdhc_send_cmd_common()

        while (!(esdhc_read32(&regs->irqstat) & flags))
                ;
Quoted 7 lines

Instead of looping forever, provide an exit path so that a timeout error can be propagated in the case irqstat does not report any interrupts, which may happen when no eMMC is populated on board.

Boots fine on my i.MX6 Quad hummingboard2 with SOM rev 1.5 which previously has the problem I reported back in 2018.09 series.

Peter

Quoted 38 lines

---
Changes since v1:
- Jump to the label out on error path (Baruch).

 drivers/mmc/fsl_esdhc.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/mmc/fsl_esdhc.c +8 -3
index 3cdfa7f..b8d0b00 100644
@@ -396,6 +396,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,
        uint    irqstat;
        u32     flags = IRQSTAT_CC | IRQSTAT_CTOE;
        struct fsl_esdhc *regs = priv->esdhc_regs;
+       unsigned long start;
 
 #ifdef CONFIG_SYS_FSL_ERRATUM_ESDHC111
        if (cmd->cmdidx == MMC_CMD_STOP_TRANSMISSION)
@@ -453,8 +454,13 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,
                flags = IRQSTAT_BRR;
 
        /* Wait for the command to complete */
-       while (!(esdhc_read32(&regs->irqstat) & flags))
-               ;
+       start = get_timer(0);
+       while (!(esdhc_read32(&regs->irqstat) & flags)) {
+               if (get_timer(start) > 1000) {
+                       err = -ETIMEDOUT;
+                       goto out;
+               }
+       }
 
        irqstat = esdhc_read32(&regs->irqstat);
 
--

2.7.4

#2 lore raw

PATCH v2 1/1 mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

In reply to Fabio Estevam

Hi Fabio,

Quoted 58 lines · Fabio Estevam

The following hang is observed on a Hummingboard 2 MicroSOM i2eX iMX6D - rev 1.3 with no eMMC populated on board:

U-Boot SPL 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000) Trying to boot from MMC1

U-Boot 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)

CPU:   Freescale i.MX6D rev1.5 996 MHz (running at 792 MHz)
CPU:   Extended Commercial temperature grade (-20C to 105C) at 33C
Reset cause: POR
Board: MX6 HummingBoard2
DRAM:  1 GiB
MMC:   FSL_SDHC: 0, FSL_SDHC: 1

Loading Environment from MMC... *** Warning - bad CRC, using default environment

No panel detected: default to HDMI
Display: HDMI (1024x768)
In:    serial
Out:   serial
Err:   serial

---> hangs

which is caused by the following infinite loop inside esdhc_send_cmd_common()

        while (!(esdhc_read32(&regs->irqstat) & flags))
                ;

Instead of looping forever, provide an exit path so that a timeout error can be propagated in the case irqstat does not report any interrupts, which may happen when no eMMC is populated on board.

---
Changes since v1:
- Jump to the label out on error path (Baruch).

 drivers/mmc/fsl_esdhc.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/mmc/fsl_esdhc.c +1 -0
index 3cdfa7f..b8d0b00 100644
@@ -396,6 +396,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,
        uint    irqstat;
        u32     flags = IRQSTAT_CC | IRQSTAT_CTOE;
        struct fsl_esdhc *regs = priv->esdhc_regs;
+       unsigned long start;
 
 #ifdef CONFIG_SYS_FSL_ERRATUM_ESDHC111
        if (cmd->cmdidx == MMC_CMD_STOP_TRANSMISSION)
@@ -453,8 +454,13 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,
                flags = IRQSTAT_BRR;
 

drivers/mmc/fsl_esdhc.c static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc, · line 463

Quoted 10 lines
drivers/mmc/fsl_esdhc.c +7 -2
        /* Wait for the command to complete */
-       while (!(esdhc_read32(&regs->irqstat) & flags))
-               ;
+       start = get_timer(0);
+       while (!(esdhc_read32(&regs->irqstat) & flags)) {
+               if (get_timer(start) > 1000) {
+                       err = -ETIMEDOUT;
+                       goto out;
+               }
+       }

This fixes the hang I was having, but adds a significant boot delay because it needs to wait until it gets the timeout (1000). Is there a better timeout value to be used here?

Thanks,

Signature · 1 line
Ricardo Salveti de Araujo
#3 lore raw

PATCH v2 1/1 mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

Quoted 62 lines · Ricardo Salveti

Hi Fabio,

Quoted 58 lines · Fabio Estevam

The following hang is observed on a Hummingboard 2 MicroSOM i2eX iMX6D - rev 1.3 with no eMMC populated on board:

U-Boot SPL 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000) Trying to boot from MMC1

U-Boot 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)

CPU:   Freescale i.MX6D rev1.5 996 MHz (running at 792 MHz)
CPU:   Extended Commercial temperature grade (-20C to 105C) at 33C
Reset cause: POR
Board: MX6 HummingBoard2
DRAM:  1 GiB
MMC:   FSL_SDHC: 0, FSL_SDHC: 1

Loading Environment from MMC... *** Warning - bad CRC, using default environment

No panel detected: default to HDMI
Display: HDMI (1024x768)
In:    serial
Out:   serial
Err:   serial

---> hangs

which is caused by the following infinite loop inside esdhc_send_cmd_common()

        while (!(esdhc_read32(&regs->irqstat) & flags))
                ;

Instead of looping forever, provide an exit path so that a timeout error can be propagated in the case irqstat does not report any interrupts, which may happen when no eMMC is populated on board.

---
Changes since v1:
- Jump to the label out on error path (Baruch).

 drivers/mmc/fsl_esdhc.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/mmc/fsl_esdhc.c +1 -0
index 3cdfa7f..b8d0b00 100644
@@ -396,6 +396,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,
        uint    irqstat;
        u32     flags = IRQSTAT_CC | IRQSTAT_CTOE;
        struct fsl_esdhc *regs = priv->esdhc_regs;
+       unsigned long start;
 
 #ifdef CONFIG_SYS_FSL_ERRATUM_ESDHC111
        if (cmd->cmdidx == MMC_CMD_STOP_TRANSMISSION)
@@ -453,8 +454,13 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,
                flags = IRQSTAT_BRR;
 
Quoted 14 lines

drivers/mmc/fsl_esdhc.c static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc, · line 463

Quoted 10 lines
drivers/mmc/fsl_esdhc.c +7 -2
        /* Wait for the command to complete */
-       while (!(esdhc_read32(&regs->irqstat) & flags))
-               ;
+       start = get_timer(0);
+       while (!(esdhc_read32(&regs->irqstat) & flags)) {
+               if (get_timer(start) > 1000) {
+                       err = -ETIMEDOUT;
+                       goto out;
+               }
+       }

This fixes the hang I was having, but adds a significant boot delay because it needs to wait until it gets the timeout (1000). Is there a better timeout value to be used here?

Looking a bit further, this ends up adding a significant boot delay because the mmc driver still tries to send another two additional commands, so it ends executing this timeout logic 3 times in total. Guess this can be improved at
http://git.denx.de/?p=u-boot.git;a=blob;f=drivers/mmc/mmc.c;h=d6b9cdc99229d8bfe0c8be3adfd6e5209a11601c;hb=HEAD#l2633
by either checking for mmc first or by simply just checking and handling the return code from mmc_send_if_cond, but this could be another patch.

Would still be nice to confirm if 1000 is the minimal timeout value to be used here.

Thanks,

Signature · 1 line
Ricardo Salveti de Araujo
#4 lore raw

PATCH v2 1/1 mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

Hi Ricardo,

Quoted 8 lines · Ricardo Salveti

Looking a bit further, this ends up adding a significant boot delay because the mmc driver still tries to send another two additional commands, so it ends executing this timeout logic 3 times in total. Guess this can be improved at
http://git.denx.de/?p=u-boot.git;a=blob;f=drivers/mmc/mmc.c;h=d6b9cdc99229d8bfe0c8be3adfd6e5209a11601c;hb=HEAD#l2633
by either checking for mmc first or by simply just checking and handling the return code from mmc_send_if_cond, but this could be another patch.

Yes, I think you are right.

Does only adding the error check below avoid the hang in your case?

        /* Test for SD version 2 */
         err = mmc_send_if_cond(mmc);
         if (err)
            return err;

Thanks

#5 lore raw

PATCH v2 1/1 mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

Hi Fabio,

Quoted 12 lines · Fabio Estevam
Quoted 8 lines · Ricardo Salveti

Looking a bit further, this ends up adding a significant boot delay because the mmc driver still tries to send another two additional commands, so it ends executing this timeout logic 3 times in total. Guess this can be improved at
http://git.denx.de/?p=u-boot.git;a=blob;f=drivers/mmc/mmc.c;h=d6b9cdc99229d8bfe0c8be3adfd6e5209a11601c;hb=HEAD#l2633
by either checking for mmc first or by simply just checking and handling the return code from mmc_send_if_cond, but this could be another patch.

Quoted 8 lines

Yes, I think you are right.

Does only adding the error check below avoid the hang in your case?

        /* Test for SD version 2 */
         err = mmc_send_if_cond(mmc);
         if (err)
            return err;

There used to be a similar check in the past. Commit f33cb34b3971dabe3 ("mmc:
Remove return from mmc_init for non SD 2.0 compatible cards.") removed this check.

baruch

Signature · 3 lines
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch at tkos.co.il - tel: +972.2.679.5364, http://www.tkos.co.il -
#6 lore raw

PATCH v2 1/1 mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

In reply to Fabio Estevam
Quoted 14 lines · Fabio Estevam

Hi Ricardo,

Quoted 8 lines · Ricardo Salveti

Looking a bit further, this ends up adding a significant boot delay because the mmc driver still tries to send another two additional commands, so it ends executing this timeout logic 3 times in total. Guess this can be improved at
http://git.denx.de/?p=u-boot.git;a=blob;f=drivers/mmc/mmc.c;h=d6b9cdc99229d8bfe0c8be3adfd6e5209a11601c;hb=HEAD#l2633
by either checking for mmc first or by simply just checking and handling the return code from mmc_send_if_cond, but this could be another patch.

Quoted 8 lines

Yes, I think you are right.

Does only adding the error check below avoid the hang in your case?

        /* Test for SD version 2 */
         err = mmc_send_if_cond(mmc);
         if (err)
            return err;

You patch is still required as the hang I was getting was happening in the while loop itself. Adding this change on top of your patch would just reduce the delay as the code path would avoid another 2 extra commands (both returning timeout).

The only issue, as pointed out by Baruch, is that we already had a check that was ignoring the timeout here before, so I guess the current behavior is correct/expected.

+1 for your patch as it is a good fix anyway and unblocks the boot process on my som/board, even if it adds a longer boot delay.

Thanks,

Signature · 1 line
Ricardo Salveti de Araujo
#7 lore raw

PATCH v2 1/1 mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

In reply to Fabio Estevam

Hi Jaehoon,

Quoted 29 lines · Fabio Estevam

The following hang is observed on a Hummingboard 2 MicroSOM i2eX iMX6D - rev 1.3 with no eMMC populated on board:

U-Boot SPL 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000) Trying to boot from MMC1

U-Boot 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)

CPU:   Freescale i.MX6D rev1.5 996 MHz (running at 792 MHz)
CPU:   Extended Commercial temperature grade (-20C to 105C) at 33C
Reset cause: POR
Board: MX6 HummingBoard2
DRAM:  1 GiB
MMC:   FSL_SDHC: 0, FSL_SDHC: 1

Loading Environment from MMC... *** Warning - bad CRC, using default environment

No panel detected: default to HDMI
Display: HDMI (1024x768)
In:    serial
Out:   serial
Err:   serial

---> hangs

which is caused by the following infinite loop inside esdhc_send_cmd_common()

        while (!(esdhc_read32(&regs->irqstat) & flags))
                ;
Quoted 7 lines

Instead of looping forever, provide an exit path so that a timeout error can be propagated in the case irqstat does not report any interrupts, which may happen when no eMMC is populated on board.

Could this one be applied to 2019.01?

#8 lore raw

PATCH v2 1/1 mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

Hi Tom/Stefano,

Quoted 38 lines · Fabio Estevam

Hi Jaehoon,

Quoted 34 lines · Fabio Estevam

The following hang is observed on a Hummingboard 2 MicroSOM i2eX iMX6D - rev 1.3 with no eMMC populated on board:

U-Boot SPL 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000) Trying to boot from MMC1

U-Boot 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)

CPU:   Freescale i.MX6D rev1.5 996 MHz (running at 792 MHz)
CPU:   Extended Commercial temperature grade (-20C to 105C) at 33C
Reset cause: POR
Board: MX6 HummingBoard2
DRAM:  1 GiB
MMC:   FSL_SDHC: 0, FSL_SDHC: 1

Loading Environment from MMC... *** Warning - bad CRC, using default environment

No panel detected: default to HDMI
Display: HDMI (1024x768)
In:    serial
Out:   serial
Err:   serial

---> hangs

which is caused by the following infinite loop inside esdhc_send_cmd_common()

        while (!(esdhc_read32(&regs->irqstat) & flags))
                ;

Instead of looping forever, provide an exit path so that a timeout error can be propagated in the case irqstat does not report any interrupts, which may happen when no eMMC is populated on board.

Quoted 4 lines
Quoted 2 lines

Could this one be applied to 2019.01?

I haven't heard back from Jaehoon.

Can we have this one applied for 2019.01 in order to fix a regression?

Thanks

#9 lore raw

PATCH v2 1/1 mmc: fsl_esdhc: Avoid infinite loop in esdhc_send_cmd_common()

In reply to Fabio Estevam
Quoted 14 lines · Fabio Estevam

The following hang is observed on a Hummingboard 2 MicroSOM i2eX iMX6D - rev 1.3 with no eMMC populated on board:

U-Boot SPL 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000) Trying to boot from MMC1

U-Boot 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)

CPU:   Freescale i.MX6D rev1.5 996 MHz (running at 792 MHz)
CPU:   Extended Commercial temperature grade (-20C to 105C) at 33C
Reset cause: POR
Board: MX6 HummingBoard2
DRAM:  1 GiB
MMC:   FSL_SDHC: 0, FSL_SDHC: 1
Quoted 7 lines

Loading Environment from MMC... *** Warning - bad CRC, using default environment

No panel detected: default to HDMI
Display: HDMI (1024x768)
In:    serial
Out:   serial
Err:   serial

Applied to u-boot/master, thanks!

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20190114/bc4637fd/attachment.sig>