[dm-devel] dm: fix congested_fn for request-based device

Hou Tao houtao1 at huawei.com
Wed Mar 4 00:54:50 UTC 2020


Hi Mike,

On 2020/3/4 0:15, Mike Snitzer wrote:
> On Tue, Mar 03 2020 at  3:45am -0500,
> Hou Tao <houtao1 at huawei.com> wrote:
> 
>> We neither assign congested_fn for requested-based blk-mq device
>> nor implement it correctly. So fix both.
>>
>> Fixes: 4aa9c692e052 ("bdi: separate out congested state into a separate struct")
>> Signed-off-by: Hou Tao <houtao1 at huawei.com>
>> ---
>>  drivers/md/dm.c | 5 ++++-
>>  1 file changed, 4 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/md/dm.c b/drivers/md/dm.c
>> index b89f07ee2eff..80b95e21db43 100644
>> --- a/drivers/md/dm.c
>> +++ b/drivers/md/dm.c
>> @@ -1784,11 +1784,13 @@ static int dm_any_congested(void *congested_data, int bdi_bits)
>>  
>>  	if (!test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags)) {
>>  		if (dm_request_based(md)) {
>> +			struct backing_dev_info *bdi =
>> +				md->queue->backing_dev_info;
>>  			/*
>>  			 * With request-based DM we only need to check the
>>  			 * top-level queue for congestion.
>>  			 */
>> -			r = md->queue->backing_dev_info->wb.state & bdi_bits;
>> +			r = bdi->wb.congested->state & bdi_bits;
>>  		} else {
>>  			map = dm_get_live_table_fast(md);
>>  			if (map)
> 
> The above change does indeed fix missing conversion that should've
> occurred in commit 4aa9c692e052.
> 
>> @@ -2265,6 +2267,7 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t)
>>  			DMERR("Cannot initialize queue for request-based dm-mq mapped device");
>>  			return r;
>>  		}
>> +		dm_init_normal_md_queue(md);
>>  		break;
>>  	case DM_TYPE_BIO_BASED:
>>  	case DM_TYPE_DAX_BIO_BASED:
> 
> I've renamed dm_init_normal_md_queue to dm_init_congested_fn and removed
> it's incorrect comment.
> 
Thanks for that.

Regards,
Tao
> Here is the final fix I just staged:
> https://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm.git/commit/?h=dm-5.6&id=974f51e8633f0f3f33e8f86bbb5ae66758aa63c7
> 
> Thanks,
> Mike
> 
> 
> .
> 





More information about the dm-devel mailing list