Re: [PATCH 6/6] UBI: Fastmap: Make ubi_refill_pools() fair

[Richard Weinberger]

Tanya,

Am 05.12.2014 um 18:55 schrieb Tanya Brokhman:
> Hi Richard,
> 
> On 11/24/2014 3:20 PM, Richard Weinberger wrote:
>> Currently ubi_refill_pools() first fills the first and then
>> the second one.
>> If only very few free PEBs are available the second pool can get
>> zero PEBs.
>> Change ubi_refill_pools() to distribute free PEBs fair between
>> all pools.
>>
>> Signed-off-by: Richard Weinberger <richard@xxxxxx>
>> ---
>>   drivers/mtd/ubi/wl.c | 77 +++++++++++++++++++++++++++-------------------------
>>   1 file changed, 40 insertions(+), 37 deletions(-)
>>
>> diff --git a/drivers/mtd/ubi/wl.c b/drivers/mtd/ubi/wl.c
>> index f028b68..c2822f7 100644
>> --- a/drivers/mtd/ubi/wl.c
>> +++ b/drivers/mtd/ubi/wl.c
>> @@ -583,59 +583,62 @@ static void return_unused_pool_pebs(struct ubi_device *ubi,
>>   }
>>
>>   /**
>> - * refill_wl_pool - refills all the fastmap pool used by the
>> - * WL sub-system.
>> + * ubi_refill_pools - refills all fastmap PEB pools.
>>    * @ubi: UBI device description object
>>    */
>> -static void refill_wl_pool(struct ubi_device *ubi)
>> +void ubi_refill_pools(struct ubi_device *ubi)
>>   {
>> +    struct ubi_fm_pool *wl_pool = &ubi->fm_wl_pool;
>> +    struct ubi_fm_pool *pool = &ubi->fm_pool;
>>       struct ubi_wl_entry *e;
>> -    struct ubi_fm_pool *pool = &ubi->fm_wl_pool;
>> +    int enough;
>>
>> +    spin_lock(&ubi->wl_lock);
>> +
>> +    return_unused_pool_pebs(ubi, wl_pool);
>>       return_unused_pool_pebs(ubi, pool);
>>
>> -    for (pool->size = 0; pool->size < pool->max_size; pool->size++) {
>> -        if (!ubi->free.rb_node ||
>> -           (ubi->free_count - ubi->beb_rsvd_pebs < 5))
>> -            break;
>> +    wl_pool->size = 0;
>> +    pool->size = 0;
>>
>> -        e = find_wl_entry(ubi, &ubi->free, WL_FREE_MAX_DIFF);
>> -        self_check_in_wl_tree(ubi, e, &ubi->free);
>> -        rb_erase(&e->u.rb, &ubi->free);
>> -        ubi->free_count--;
>> +    for (;;) {
> 
> You loop for max(pool->max_size, wl_pool->max_size) itterations. IMO, the code will be more clear if you use for(i=0; i<max(pool->max_size, wl_pool->max_size); i++) instead of "int
> enough".
> This is just coding style preference of course. I personally don't like for(;;) that much.... Just a suggestion. :)

I agree that it's a matter of taste. :)

>> +        enough = 0;
>> +        if (pool->size < pool->max_size) {
>> +            if (!ubi->free.rb_node ||
>> +               (ubi->free_count - ubi->beb_rsvd_pebs < 5))
>> +                break;
>>
>> -        pool->pebs[pool->size] = e->pnum;
>> -    }
>> -    pool->used = 0;
>> -}
>> +            e = wl_get_wle(ubi);
>> +            if (!e)
>> +                break;
>>
>> -/**
>> - * refill_wl_user_pool - refills all the fastmap pool used by ubi_wl_get_peb.
>> - * @ubi: UBI device description object
>> - */
>> -static void refill_wl_user_pool(struct ubi_device *ubi)
>> -{
>> -    struct ubi_fm_pool *pool = &ubi->fm_pool;
>> +            pool->pebs[pool->size] = e->pnum;
>> +            pool->size++;
>> +        } else
>> +            enough++;
>>
>> -    return_unused_pool_pebs(ubi, pool);
>> +        if (wl_pool->size < wl_pool->max_size) {
>> +            if (!ubi->free.rb_node ||
>> +               (ubi->free_count - ubi->beb_rsvd_pebs < 5))
>> +                break;
>>
>> -    for (pool->size = 0; pool->size < pool->max_size; pool->size++) {
>> -        pool->pebs[pool->size] = __wl_get_peb(ubi);
>> -        if (pool->pebs[pool->size] < 0)
>> +            e = find_wl_entry(ubi, &ubi->free, WL_FREE_MAX_DIFF);
>> +            self_check_in_wl_tree(ubi, e, &ubi->free);
>> +            rb_erase(&e->u.rb, &ubi->free);
>> +            ubi->free_count--;
> 
> why don't you use wl_get_peb() here?

Because wl_get_peb() is not equivalent to the above code.
We want a PEB to be used for wear-leveling not for "end users" like UBIFS.

Thanks,
//richard
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/