Re: [RESEND PATCH V3] mmc: core: Detect card removal on I/O error

[Ulf Hansson <ulf.hansson@stericsson.com>]

Adrian Hunter wrote:
> On 03/02/12 11:33, Ulf Hansson wrote:
>> To prevent I/O as soon as possible at card removal, a new
>> detect work is re-scheduled without a delay to let a rescan
>> remove the card device as soon a possible.
>>
>> Additionally, MMC_CAP2_DETECT_ON_ERR can now be used to handle
>> "slowly" removed cards that a scheduled detect work did not
>> detect as removed. To prevent further I/O requests for these
>> lingering removed cards, check if card has been removed and then
>> schedule a detect work to properly remove it.
>>
>> Signed-off-by: Ulf Hansson <ulf.hansson@stericsson.com>
>> ---
>>
>> Changes in v3:
>> 	- Check for card is NULL and minor code simplifications.
>>
>> Changes in v2:
>> 	- Updated according to review comments.
>> 	- Merging two patches for this feature into one.
>>
>> ---
>>  drivers/mmc/core/core.c  |   24 +++++++++++++++++++++---
>>  include/linux/mmc/host.h |    1 +
>>  2 files changed, 22 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
>> index bec0bf2..9645e8c 100644
>> --- a/drivers/mmc/core/core.c
>> +++ b/drivers/mmc/core/core.c
>> @@ -2077,18 +2077,36 @@ int _mmc_detect_card_removed(struct mmc_host *host)
>>  int mmc_detect_card_removed(struct mmc_host *host)
>>  {
>>  	struct mmc_card *card = host->card;
>> +	int ret;
>>  
>>  	WARN_ON(!host->claimed);
>> +
>> +	if (!card)
>> +		return 1;
>> +
>> +	ret = mmc_card_removed(card);
>>  	/*
>>  	 * The card will be considered unchanged unless we have been asked to
>>  	 * detect a change or host requires polling to provide card detection.
>>  	 */
>> -	if (card && !host->detect_change && !(host->caps & MMC_CAP_NEEDS_POLL))
>> -		return mmc_card_removed(card);
>> +	if (!host->detect_change && !(host->caps & MMC_CAP_NEEDS_POLL) &&
>> +	    !(host->caps2 & MMC_CAP2_DETECT_ON_ERR))
>> +		return ret;
>>  
>>  	host->detect_change = 0;
>> +	if (!ret) {
>> +		ret = _mmc_detect_card_removed(host);
>> +		if (ret) {
> 
> Please make this
> 
> 		if (ret && (host->caps2 & MMC_CAP2_DETECT_ON_ERR))
> 
> because if rescan is already running this will needlessly queue
> another one.
> 

It wont queue another one. It will cancel a work which likely has been 
scheduled to be executed within several ms later. Then it will 
re-schedule a new one without any delay since there is no need to wait, 
when we already know that the card has been removed.


>> +			/*
>> +			 * Schedule a detect work as soon as possible to let a
>> +			 * rescan handle the card removal.
>> +			 */
>> +			cancel_delayed_work(&host->detect);
>> +			mmc_detect_change(host, 0);
>> +		}
>> +	}
>>  
>> -	return _mmc_detect_card_removed(host);
>> +	return ret;
>>  }
>>  EXPORT_SYMBOL(mmc_detect_card_removed);
>>  
>> diff --git a/include/linux/mmc/host.h b/include/linux/mmc/host.h
>> index dd13e05..368a2b9 100644
>> --- a/include/linux/mmc/host.h
>> +++ b/include/linux/mmc/host.h
>> @@ -257,6 +257,7 @@ struct mmc_host {
>>  #define MMC_CAP2_HS200_1_2V_SDR	(1 << 6)        /* can support */
>>  #define MMC_CAP2_HS200		(MMC_CAP2_HS200_1_8V_SDR | \
>>  				 MMC_CAP2_HS200_1_2V_SDR)
>> +#define MMC_CAP2_DETECT_ON_ERR	(1 << 7)	/* On I/O err check card removal */
>>  
>>  	mmc_pm_flag_t		pm_caps;	/* supported pm features */
>>  	unsigned int        power_notify_type;
> 
>