On 04/22/2013 02:01 PM, MORITA Kazutaka wrote: > It looks better to use BDRV_SECTOR_SIZE . I'd suggest preparing > another patch to replace all the SECTOR_SIZE with BDRV_SECTOR_SIZE. >
Okay >> > + end = start + (nb_sectors * SECTOR_SIZE) / >> > SD_DATA_OBJ_SIZE; > Using 'start' to calculate 'end' is wrong because 'start' may be > rounded down. > Good catch. Will updated it in v2 >> > + >> > + for (idx = start; idx <= end; idx++) { >> > + if (inode->data_vdi_id[idx] == 0) { >> > + break; >> > + } >> > + } >> > + if (idx == start) { >> > + *pnum = SD_DATA_OBJ_SIZE / SECTOR_SIZE; > Isn't it better to set the longest length of the unallocated sectors? > Good idea. Thanks, Yuan