Re: [PATCH 1/2] mmc: core: Prevent I/O as soon as possible at card removal

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

Adrian Hunter wrote:
> On 19/01/12 18:39, Ulf Hansson wrote:
>> Once the card has been detected to be removed by the
>> mmc_detect_card_removed function, schedule a new detect work
>> immediately and without a delay to let a rescan remove the
>> card device as soon a possible. This will sooner prevent
>> further I/O requests.
>>
>> Signed-off-by: Ulf Hansson <ulf.hansson@stericsson.com>
>> ---
>>  drivers/mmc/core/core.c |   16 ++++++++++++++--
>>  1 files changed, 14 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
>> index bec0bf2..265dfd8 100644
>> --- a/drivers/mmc/core/core.c
>> +++ b/drivers/mmc/core/core.c
>> @@ -2077,6 +2077,7 @@ 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);
>>  	/*
>> @@ -2086,9 +2087,20 @@ int mmc_detect_card_removed(struct mmc_host *host)
>>  	if (card && !host->detect_change && !(host->caps & MMC_CAP_NEEDS_POLL))
>>  		return mmc_card_removed(card);
>>  
>> -	host->detect_change = 0;
> 
> That line should not be removed.  It is not related to your change.

I think it is. Since my patch is trying to make it possible to "prevent 
I/O as soon as possible..."

Clearing the detect_change flag here will prevent the I/O layer from 
doing further tests to see if the card is removed by using 
"mmc_detect_card_removed -> _mmc_detect_card_removed" due to the upper 
if sentence.

I think this flag should only be cleared from the mmc_rescan function.

> 
>> +	ret = mmc_card_removed(card);
> 
> Calling mmc_card_removed() is not needed here since
> _mmc_detect_card_removed() does it anyway.
> 
>> +	if (!ret) {
>> +		ret = _mmc_detect_card_removed(host);
>> +		if (ret) {
>> +			/*
>> +			 * Schedule a detect work as soon as possible to let a
>> +			 * rescan handle the card removal.
>> +			 */
>> +			cancel_delayed_work(&host->detect);
> 
> Why cancel the detect work?

To "prevent I/O as soon as possible...".

The detect work could have been scheduled to be run at several ms later. 
There is no need to wait for it since we already now that card will be 
removed when the rescan function will execute.

> 
>> +			mmc_detect_change(host, 0);
>> +		}
>> +	}
>>  
>> -	return _mmc_detect_card_removed(host);
>> +	return ret;
>>  }
>>  EXPORT_SYMBOL(mmc_detect_card_removed);
>>  
> 
> 

Br
Ulf Hansson