[U-Boot] error: sdhci_send_command: Timeout for status update!

[U-Boot] error: sdhci_send_command: Timeout for status update!

[Steve Rae]

Stefan,

I am asking you about this code, because of this:
  commit 29905a4 - mmc: sdhci: Use timer based timeout detection in
sdhci_send_command()

Occasionally, I am seeing failures when writing to flash on my device:

  Flashing sparse image at offset 2078720
  Flashing Sparse Image
  sdhci_send_command: Timeout for status update!
  mmc fail to send stop cmd
  write_sparse_image: Write failed, block #2181088 [0]

I can eliminate this error if I kludge the following code:

diff --git a/drivers/mmc/sdhci.c b/drivers/mmc/sdhci.c
index 5c71ab8..854f7ce 100644
--- a/drivers/mmc/sdhci.c
+++ b/drivers/mmc/sdhci.c
@@ -243,9 +243,9 @@ static int sdhci_send_command(struct mmc *mmc,
struct mmc_cmd *cmd,
                if (stat & SDHCI_INT_ERROR)
                        break;
        } while (((stat & mask) != mask) &&
-                (get_timer(start) < CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT));
+                (get_timer(start) < 600 /*CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT*/));

-       if (get_timer(start) >= CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT) {
+       if (get_timer(start) >= 600 /*CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT*/) {
                if (host->quirks & SDHCI_QUIRK_BROKEN_R1B)
                        return 0;
                else {

I don't know much about this - I arrived at 600 by trial and error
(500 fails...)
Any ideas?

Thanks, Steve
steve.rae at raedomain.com  (previously: srae at broadcom.com)

[U-Boot] error: sdhci_send_command: Timeout for status update!

[Stefan Roese]

Steve,

On 24.06.2016 00:06, Steve Rae wrote:
> I am asking you about this code, because of this:
>    commit 29905a4 - mmc: sdhci: Use timer based timeout detection in
> sdhci_send_command()
>
> Occasionally, I am seeing failures when writing to flash on my device:
>
>    Flashing sparse image at offset 2078720
>    Flashing Sparse Image
>    sdhci_send_command: Timeout for status update!
>    mmc fail to send stop cmd
>    write_sparse_image: Write failed, block #2181088 [0]
>
> I can eliminate this error if I kludge the following code:
>
> diff --git a/drivers/mmc/sdhci.c b/drivers/mmc/sdhci.c
> index 5c71ab8..854f7ce 100644
> --- a/drivers/mmc/sdhci.c
> +++ b/drivers/mmc/sdhci.c
> @@ -243,9 +243,9 @@ static int sdhci_send_command(struct mmc *mmc,
> struct mmc_cmd *cmd,
>                  if (stat & SDHCI_INT_ERROR)
>                          break;
>          } while (((stat & mask) != mask) &&
> -                (get_timer(start) < CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT));
> +                (get_timer(start) < 600 /*CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT*/));
>
> -       if (get_timer(start) >= CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT) {
> +       if (get_timer(start) >= 600 /*CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT*/) {
>                  if (host->quirks & SDHCI_QUIRK_BROKEN_R1B)
>                          return 0;
>                  else {
>
> I don't know much about this - I arrived at 600 by trial and error
> (500 fails...)
> Any ideas?

Perhaps your platform needs a bigger timeout value. Did you check
what the Linux driver does here? You could move
CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT to Kconfig with a default of 100
and select a higher value for your platform.

Thanks,
Stefan

[U-Boot] error: sdhci_send_command: Timeout for status update!

[Masahiro Yamada]

2016-06-24 7:06 GMT+09:00 Steve Rae <steve.rae@raedomain.com>:
> Stefan,
>
> I am asking you about this code, because of this:
>   commit 29905a4 - mmc: sdhci: Use timer based timeout detection in
> sdhci_send_command()
>
> Occasionally, I am seeing failures when writing to flash on my device:
>
>   Flashing sparse image at offset 2078720
>   Flashing Sparse Image
>   sdhci_send_command: Timeout for status update!
>   mmc fail to send stop cmd
>   write_sparse_image: Write failed, block #2181088 [0]
>
> I can eliminate this error if I kludge the following code:
>
> diff --git a/drivers/mmc/sdhci.c b/drivers/mmc/sdhci.c
> index 5c71ab8..854f7ce 100644
> --- a/drivers/mmc/sdhci.c
> +++ b/drivers/mmc/sdhci.c
> @@ -243,9 +243,9 @@ static int sdhci_send_command(struct mmc *mmc,
> struct mmc_cmd *cmd,
>                 if (stat & SDHCI_INT_ERROR)
>                         break;
>         } while (((stat & mask) != mask) &&
> -                (get_timer(start) < CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT));
> +                (get_timer(start) < 600 /*CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT*/));
>
> -       if (get_timer(start) >= CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT) {
> +       if (get_timer(start) >= 600 /*CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT*/) {
>                 if (host->quirks & SDHCI_QUIRK_BROKEN_R1B)
>                         return 0;
>                 else {
>
> I don't know much about this - I arrived at 600 by trial and error
> (500 fails...)
> Any ideas?
>
>

Hi Steve,

You are not alone.

I am also suffering this too short timeout value.

I sent this patch about two months ago.
http://patchwork.ozlabs.org/patch/615994/

But, nothing happened.




-- 
Best Regards
Masahiro Yamada

[U-Boot] error: sdhci_send_command: Timeout for status update!

[Steve Rae]

Hi Stefan, Masahiro

On Thu, Jun 23, 2016 at 9:56 PM, Stefan Roese <sr@denx.de> wrote:
> Steve,
>
>
> On 24.06.2016 00:06, Steve Rae wrote:
>>
>> I am asking you about this code, because of this:
>>    commit 29905a4 - mmc: sdhci: Use timer based timeout detection in
>> sdhci_send_command()
>>
>> Occasionally, I am seeing failures when writing to flash on my device:
>>
>>    Flashing sparse image at offset 2078720
>>    Flashing Sparse Image
>>    sdhci_send_command: Timeout for status update!
>>    mmc fail to send stop cmd
>>    write_sparse_image: Write failed, block #2181088 [0]
>>
>> I can eliminate this error if I kludge the following code:
>>
>> diff --git a/drivers/mmc/sdhci.c b/drivers/mmc/sdhci.c
>> index 5c71ab8..854f7ce 100644
>> --- a/drivers/mmc/sdhci.c
>> +++ b/drivers/mmc/sdhci.c
>> @@ -243,9 +243,9 @@ static int sdhci_send_command(struct mmc *mmc,
>> struct mmc_cmd *cmd,
>>                  if (stat & SDHCI_INT_ERROR)
>>                          break;
>>          } while (((stat & mask) != mask) &&
>> -                (get_timer(start) < CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT));
>> +                (get_timer(start) < 600
>> /*CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT*/));
>>
>> -       if (get_timer(start) >= CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT) {
>> +       if (get_timer(start) >= 600 /*CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT*/)
>> {
>>                  if (host->quirks & SDHCI_QUIRK_BROKEN_R1B)
>>                          return 0;
>>                  else {
>>
>> I don't know much about this - I arrived at 600 by trial and error
>> (500 fails...)
>> Any ideas?
>
>
> Perhaps your platform needs a bigger timeout value. Did you check
> what the Linux driver does here?

It looks like (v.4.6) the code loops for max_loops=16,
and it looks like the loop delay is created by a write (which does not
exist in the U-Boot code):
     sdhci_writel(host, mask, SDHCI_INT_STATUS);

> You could move
> CONFIG_SDHCI_CMD_DEFAULT_TIMEOUT to Kconfig with a default of 100
> and select a higher value for your platform.
>

Would it be wise to create a new CONFIG for the reading of the status,
rather than overloading the CONFIG used for sending the command?
maybe: CONFIG_SDHCI_STATUS_DEFAULT_TIMEOUT ?

Thanks, Steve

> Thanks,
> Stefan