2015-05-18 14:23:27

by Prabu Thangamuthu

Subject: [PATCH 1/1] dw_mmc: insmod followed by rmmod will hung for eMMC

Removing dw_mmc driver immediately after inserting the dw_mmc driver is
getting hung for eMMC device. Root cause for this issue is, dw_mci_remove
will disable all the interrupts then it will call dw_mci_cleanup_slot.
dw_mci_cleanup_slot is issuing CMD6 to disable boot partition access and
it's waiting for command complete interrupt. Since INTMASK was already
cleared by dw_mci_remove, command complete interrupt is not reaching
the system. This leads to process hung.

Signed-off-by: Prabu Thangamuthu <[email protected]>
---
drivers/mmc/host/dw_mmc.c | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
index 5f5adaf..f0a0aff 100644
--- a/drivers/mmc/host/dw_mmc.c
+++ b/drivers/mmc/host/dw_mmc.c
@@ -2941,15 +2941,15 @@ void dw_mci_remove(struct dw_mci *host)
{
int i;

- mci_writel(host, RINTSTS, 0xFFFFFFFF);
- mci_writel(host, INTMASK, 0); /* disable all mmc interrupt first */
-
for (i = 0; i < host->num_slots; i++) {
dev_dbg(host->dev, "remove slot %d\n", i);
if (host->slot[i])
dw_mci_cleanup_slot(host->slot[i], i);
}

+ mci_writel(host, RINTSTS, 0xFFFFFFFF);
+ mci_writel(host, INTMASK, 0); /* disable all mmc interrupt first */
+
/* disable clock to CIU */
mci_writel(host, CLKENA, 0);
mci_writel(host, CLKSRC, 0);
--
1.7.6.5

2015-05-22 13:21:30

by Ulf Hansson

Subject: Re: [PATCH 1/1] dw_mmc: insmod followed by rmmod will hung for eMMC

On 18 May 2015 at 16:23, Prabu Thangamuthu <[email protected]> wrote:
> Removing dw_mmc driver immediately after inserting the dw_mmc driver is

I guess it hangs even if you remove it after a couple of days? :-)

Perhaps makes this a bit more clear?

> getting hung for eMMC device. Root cause for this issue is, dw_mci_remove
> will disable all the interrupts then it will call dw_mci_cleanup_slot.
> dw_mci_cleanup_slot is issuing CMD6 to disable boot partition access and
> it's waiting for command complete interrupt. Since INTMASK was already
> cleared by dw_mci_remove, command complete interrupt is not reaching
> the system. This leads to process hung.

/s dw_mci_remove / dw_mci_remove()

>
> Signed-off-by: Prabu Thangamuthu <[email protected]>

This patch looks good overall, but please send a new version with
updated changelog. Moreover, please use "mmc: dw_mmc:" as prefix for
the commit message header.

Kind regards
Uffe

> ---
> drivers/mmc/host/dw_mmc.c | 6 +++---
> 1 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
> index 5f5adaf..f0a0aff 100644
> --- a/drivers/mmc/host/dw_mmc.c
> +++ b/drivers/mmc/host/dw_mmc.c
> @@ -2941,15 +2941,15 @@ void dw_mci_remove(struct dw_mci *host)
> {
> int i;
>
> - mci_writel(host, RINTSTS, 0xFFFFFFFF);
> - mci_writel(host, INTMASK, 0); /* disable all mmc interrupt first */
> -
> for (i = 0; i < host->num_slots; i++) {
> dev_dbg(host->dev, "remove slot %d\n", i);
> if (host->slot[i])
> dw_mci_cleanup_slot(host->slot[i], i);
> }
>
> + mci_writel(host, RINTSTS, 0xFFFFFFFF);
> + mci_writel(host, INTMASK, 0); /* disable all mmc interrupt first */
> +
> /* disable clock to CIU */
> mci_writel(host, CLKENA, 0);
> mci_writel(host, CLKSRC, 0);
> --
> 1.7.6.5

2015-05-27 01:50:07

by Jaehoon Chung

Subject: Re: [PATCH 1/1] dw_mmc: insmod followed by rmmod will hung for eMMC

Hi, Pradu.

Sorry for late.
I will wait for your next version than i will check yours.

To Ulf.

Thanks for review!

Best Regards,
Jaehoon Chung

On 05/22/2015 10:21 PM, Ulf Hansson wrote:
> On 18 May 2015 at 16:23, Prabu Thangamuthu <[email protected]> wrote:
>> Removing dw_mmc driver immediately after inserting the dw_mmc driver is
>
> I guess it hangs even if you remove it after a couple of days? :-)
>
> Perhaps makes this a bit more clear?
>
>> getting hung for eMMC device. Root cause for this issue is, dw_mci_remove
>> will disable all the interrupts then it will call dw_mci_cleanup_slot.
>> dw_mci_cleanup_slot is issuing CMD6 to disable boot partition access and
>> it's waiting for command complete interrupt. Since INTMASK was already
>> cleared by dw_mci_remove, command complete interrupt is not reaching
>> the system. This leads to process hung.
>
> /s dw_mci_remove / dw_mci_remove()
>
>>
>> Signed-off-by: Prabu Thangamuthu <[email protected]>
>
> This patch looks good overall, but please send a new version with
> updated changelog. Moreover, please use "mmc: dw_mmc:" as prefix for
> the commit message header.
>
> Kind regards
> Uffe
>
>> ---
>> drivers/mmc/host/dw_mmc.c | 6 +++---
>> 1 files changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
>> index 5f5adaf..f0a0aff 100644
>> --- a/drivers/mmc/host/dw_mmc.c
>> +++ b/drivers/mmc/host/dw_mmc.c
>> @@ -2941,15 +2941,15 @@ void dw_mci_remove(struct dw_mci *host)
>> {
>> int i;
>>
>> - mci_writel(host, RINTSTS, 0xFFFFFFFF);
>> - mci_writel(host, INTMASK, 0); /* disable all mmc interrupt first */
>> -
>> for (i = 0; i < host->num_slots; i++) {
>> dev_dbg(host->dev, "remove slot %d\n", i);
>> if (host->slot[i])
>> dw_mci_cleanup_slot(host->slot[i], i);
>> }
>>
>> + mci_writel(host, RINTSTS, 0xFFFFFFFF);
>> + mci_writel(host, INTMASK, 0); /* disable all mmc interrupt first */
>> +
>> /* disable clock to CIU */
>> mci_writel(host, CLKENA, 0);
>> mci_writel(host, CLKSRC, 0);
>> --
>> 1.7.6.5
>