From: santosh <santosh.shukla@caviumnetworks.com>
To: Olivier MATZ <olivier.matz@6wind.com>
Cc: dev@dpdk.org, thomas@monjalon.net,
jerin.jacob@caviumnetworks.com, hemant.agrawal@nxp.com
Subject: Re: [dpdk-dev] [PATCH v4 6/7] mempool: introduce block size align flag
Date: Thu, 7 Sep 2017 13:07:19 +0530 [thread overview]
Message-ID: <6586999c-9e80-14b6-f9de-b6bc5683debb@caviumnetworks.com> (raw)
In-Reply-To: <20170907072724.fixbpyxby4eprzt5@neon>
On Thursday 07 September 2017 12:57 PM, Olivier MATZ wrote:
> On Mon, Sep 04, 2017 at 11:15:50PM +0530, santosh wrote:
>> On Monday 04 September 2017 09:50 PM, Olivier MATZ wrote:
>>> On Tue, Aug 15, 2017 at 11:37:42AM +0530, Santosh Shukla wrote:
>>>> Some mempool hw like octeontx/fpa block, demands block size
>>>> (/total_elem_sz) aligned object start address.
>>>>
>>>> Introducing an MEMPOOL_F_POOL_BLK_SZ_ALIGNED flag.
>>>> If this flag is set:
>>>> - Align object start address to a multiple of total_elt_sz.
>>> Please specify if it's virtual or physical address.
>> virtual address. Yes will mention in v5. Thanks.
>>
>>> What do you think about MEMPOOL_F_BLK_ALIGNED_OBJECTS instead?
>>>
>>> I don't really like BLK because the word "block" is not used anywhere
>>> else in the mempool code. But I cannot find any good replacement for
>>> it. If you have another idea, please suggest.
>>>
>> Ok with renaming to MEMPOOL_F_BLK_ALIGNED_OBJECTS
>>
>>>> - Allocate one additional object. Additional object is needed to make
>>>> sure that requested 'n' object gets correctly populated. Example:
>>>>
>>>> - Let's say that we get 'x' size of memory chunk from memzone.
>>>> - And application has requested 'n' object from mempool.
>>>> - Ideally, we start using objects at start address 0 to...(x-block_sz)
>>>> for n obj.
>>>> - Not necessarily first object address i.e. 0 is aligned to block_sz.
>>>> - So we derive 'offset' value for block_sz alignment purpose i.e..'off'.
>>>> - That 'off' makes sure that start address of object is blk_sz
>>>> aligned.
>>>> - Calculating 'off' may end up sacrificing first block_sz area of
>>>> memzone area x. So total number of the object which can fit in the
>>>> pool area is n-1, Which is incorrect behavior.
>>>>
>>>> Therefore we request one additional object (/block_sz area) from memzone
>>>> when F_BLK_SZ_ALIGNED flag is set.
>>>>
>>>> Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
>>>> Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
>>>> ---
>>>> lib/librte_mempool/rte_mempool.c | 16 +++++++++++++---
>>>> lib/librte_mempool/rte_mempool.h | 1 +
>>>> 2 files changed, 14 insertions(+), 3 deletions(-)
>>>>
>>>> diff --git a/lib/librte_mempool/rte_mempool.c b/lib/librte_mempool/rte_mempool.c
>>>> index 19e5e6ddf..7610f0d1f 100644
>>>> --- a/lib/librte_mempool/rte_mempool.c
>>>> +++ b/lib/librte_mempool/rte_mempool.c
>>>> @@ -239,10 +239,14 @@ rte_mempool_calc_obj_size(uint32_t elt_size, uint32_t flags,
>>>> */
>>>> size_t
>>>> rte_mempool_xmem_size(uint32_t elt_num, size_t total_elt_sz, uint32_t pg_shift,
>>>> - __rte_unused const struct rte_mempool *mp)
>>>> + const struct rte_mempool *mp)
>>>> {
>>>> size_t obj_per_page, pg_num, pg_sz;
>>>>
>>>> + if (mp && mp->flags & MEMPOOL_F_POOL_BLK_SZ_ALIGNED)
>>>> + /* alignment need one additional object */
>>>> + elt_num += 1;
>>>> +
>>>> if (total_elt_sz == 0)
>>>> return 0;
>>> I'm wondering if it's correct if the mempool area is not contiguous.
>>>
>>> For instance:
>>> page size = 4096
>>> object size = 1900
>>> elt_num = 10
>>>
>>> With your calculation, you will request (11+2-1)/2 = 6 pages.
>>> But actually you may need 10 pages (max), since the number of object per
>>> page matching the alignement constraint is 1, not 2.
>>>
>> In our case, we set PMD flag MEMPOOL_F_CAPA_PHYS_CONTIG to detect contiguity,
>> would fail at pool creation time, as HW don't support.
> Yes but here it's generic code. If MEMPOOL_F_POOL_BLK_SZ_ALIGNED implies
> MEMPOOL_F_CAPA_PHYS_CONTIG, it should be enforced somewhere.
>
Right,
Approach:
We agree to keep _F_CAPA_PHYS_CONTIG as flag not set by application,
i'm thinking to keep that way for _ALIGNED flag too, both setted by mempool
handler, such that mempool handler sets both the flag before pool creation time.
And above condition will check for both the flags? are you fine with that approach?
Or pl. suggest an alternative? As because right now octeontx mempool block _need_
aligned blocks and cares for contiguity. Having said that ext-mempool in particular
and mempool in general doesn't facilitate anything like that..
Thanks.
next prev parent reply other threads:[~2017-09-07 7:37 UTC|newest]
Thread overview: 116+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-06-21 17:32 [dpdk-dev] [PATCH 0/4] Infrastructure to support octeontx HW mempool manager Santosh Shukla
2017-06-21 17:32 ` [dpdk-dev] [PATCH 1/4] mempool: get the external mempool capability Santosh Shukla
2017-07-03 16:37 ` Olivier Matz
2017-07-05 6:41 ` santosh
2017-07-10 13:55 ` Olivier Matz
2017-07-10 16:09 ` santosh
2017-06-21 17:32 ` [dpdk-dev] [PATCH 2/4] mempool: detect physical contiguous object in pool Santosh Shukla
2017-07-03 16:37 ` Olivier Matz
2017-07-05 7:07 ` santosh
2017-06-21 17:32 ` [dpdk-dev] [PATCH 3/4] mempool: introduce block size align flag Santosh Shukla
2017-07-03 16:37 ` Olivier Matz
2017-07-05 7:35 ` santosh
2017-07-10 13:15 ` Olivier Matz
2017-07-10 16:22 ` santosh
2017-06-21 17:32 ` [dpdk-dev] [PATCH 4/4] mempool: update range info to pool Santosh Shukla
2017-07-13 9:32 ` [dpdk-dev] [PATCH v2 0/6] Infrastructure to support octeontx HW mempool manager Santosh Shukla
2017-07-13 9:32 ` [dpdk-dev] [PATCH v2 1/6] mempool: fix flags data type Santosh Shukla
2017-07-13 9:32 ` [dpdk-dev] [PATCH v2 2/6] mempool: get the mempool capability Santosh Shukla
2017-07-13 9:32 ` [dpdk-dev] [PATCH v2 3/6] mempool: detect physical contiguous object in pool Santosh Shukla
2017-07-13 9:32 ` [dpdk-dev] [PATCH v2 4/6] mempool: add mempool arg in xmem size and usage Santosh Shukla
2017-07-13 9:32 ` [dpdk-dev] [PATCH v2 5/6] mempool: introduce block size align flag Santosh Shukla
2017-07-13 9:32 ` [dpdk-dev] [PATCH v2 6/6] mempool: update range info to pool Santosh Shukla
2017-07-18 6:07 ` [dpdk-dev] [PATCH v2 0/6] Infrastructure to support octeontx HW mempool manager santosh
2017-07-20 13:47 ` [dpdk-dev] [PATCH v3 " Santosh Shukla
2017-07-20 13:47 ` [dpdk-dev] [PATCH v3 1/6] mempool: fix flags data type Santosh Shukla
2017-07-20 13:47 ` [dpdk-dev] [PATCH v3 2/6] mempool: get the mempool capability Santosh Shukla
2017-07-20 13:47 ` [dpdk-dev] [PATCH v3 3/6] mempool: detect physical contiguous object in pool Santosh Shukla
2017-07-20 13:47 ` [dpdk-dev] [PATCH v3 4/6] mempool: add mempool arg in xmem size and usage Santosh Shukla
2017-07-20 13:47 ` [dpdk-dev] [PATCH v3 5/6] mempool: introduce block size align flag Santosh Shukla
2017-07-20 13:47 ` [dpdk-dev] [PATCH v3 6/6] mempool: update range info to pool Santosh Shukla
2017-08-15 6:07 ` [dpdk-dev] [PATCH v4 0/7] Infrastructure to support octeontx HW mempool manager Santosh Shukla
2017-08-15 6:07 ` [dpdk-dev] [PATCH v4 1/7] mempool: fix flags data type Santosh Shukla
2017-09-04 14:11 ` Olivier MATZ
2017-09-04 14:18 ` santosh
2017-08-15 6:07 ` [dpdk-dev] [PATCH v4 2/7] mempool: add mempool arg in xmem size and usage Santosh Shukla
2017-09-04 14:22 ` Olivier MATZ
2017-09-04 14:33 ` santosh
2017-09-04 14:46 ` Olivier MATZ
2017-09-04 14:58 ` santosh
2017-09-04 15:23 ` Olivier MATZ
2017-09-04 15:52 ` santosh
2017-08-15 6:07 ` [dpdk-dev] [PATCH v4 3/7] doc: remove mempool api change notice Santosh Shukla
2017-08-15 6:07 ` [dpdk-dev] [PATCH v4 4/7] mempool: get the mempool capability Santosh Shukla
2017-09-04 14:32 ` Olivier MATZ
2017-09-04 14:44 ` santosh
2017-09-04 15:56 ` Olivier MATZ
2017-09-04 16:29 ` santosh
2017-08-15 6:07 ` [dpdk-dev] [PATCH v4 5/7] mempool: detect physical contiguous object in pool Santosh Shukla
2017-09-04 14:43 ` Olivier MATZ
2017-09-04 14:47 ` santosh
2017-09-04 16:00 ` Olivier MATZ
2017-08-15 6:07 ` [dpdk-dev] [PATCH v4 6/7] mempool: introduce block size align flag Santosh Shukla
2017-09-04 16:20 ` Olivier MATZ
2017-09-04 17:45 ` santosh
2017-09-07 7:27 ` Olivier MATZ
2017-09-07 7:37 ` santosh [this message]
2017-08-15 6:07 ` [dpdk-dev] [PATCH v4 7/7] mempool: update range info to pool Santosh Shukla
2017-09-06 11:28 ` [dpdk-dev] [PATCH v5 0/8] Infrastructure to support octeontx HW mempool manager Santosh Shukla
2017-09-06 11:28 ` [dpdk-dev] [PATCH v5 1/8] mempool: remove unused flags argument Santosh Shukla
2017-09-07 7:41 ` Olivier MATZ
2017-09-06 11:28 ` [dpdk-dev] [PATCH v5 2/8] mempool: change flags from int to unsigned int Santosh Shukla
2017-09-07 7:43 ` Olivier MATZ
2017-09-06 11:28 ` [dpdk-dev] [PATCH v5 3/8] mempool: add flags arg in xmem size and usage Santosh Shukla
2017-09-07 7:46 ` Olivier MATZ
2017-09-07 7:49 ` santosh
2017-09-06 11:28 ` [dpdk-dev] [PATCH v5 4/8] doc: remove mempool notice Santosh Shukla
2017-09-07 7:47 ` Olivier MATZ
2017-09-06 11:28 ` [dpdk-dev] [PATCH v5 5/8] mempool: get the mempool capability Santosh Shukla
2017-09-07 7:59 ` Olivier MATZ
2017-09-07 8:15 ` santosh
2017-09-07 8:39 ` Olivier MATZ
2017-09-06 11:28 ` [dpdk-dev] [PATCH v5 6/8] mempool: detect physical contiguous object in pool Santosh Shukla
2017-09-07 8:05 ` Olivier MATZ
2017-09-06 11:28 ` [dpdk-dev] [PATCH v5 7/8] mempool: introduce block size align flag Santosh Shukla
2017-09-07 8:13 ` Olivier MATZ
2017-09-07 8:27 ` santosh
2017-09-07 8:57 ` Olivier MATZ
2017-09-06 11:28 ` [dpdk-dev] [PATCH v5 8/8] mempool: update range info to pool Santosh Shukla
2017-09-07 8:30 ` Olivier MATZ
2017-09-07 8:56 ` santosh
2017-09-07 9:09 ` Olivier MATZ
2017-09-07 15:30 ` [dpdk-dev] [PATCH v6 0/8] Infrastructure to support octeontx HW mempool manager Santosh Shukla
2017-09-07 15:30 ` [dpdk-dev] [PATCH v6 1/8] mempool: remove unused flags argument Santosh Shukla
2017-09-07 15:30 ` [dpdk-dev] [PATCH v6 2/8] mempool: change flags from int to unsigned int Santosh Shukla
2017-09-07 15:30 ` [dpdk-dev] [PATCH v6 3/8] mempool: add flags arg in xmem size and usage Santosh Shukla
2017-09-25 11:24 ` Olivier MATZ
2017-09-07 15:30 ` [dpdk-dev] [PATCH v6 4/8] doc: remove mempool notice Santosh Shukla
2017-09-07 15:30 ` [dpdk-dev] [PATCH v6 5/8] mempool: get the mempool capability Santosh Shukla
2017-09-25 11:26 ` Olivier MATZ
2017-09-07 15:30 ` [dpdk-dev] [PATCH v6 6/8] mempool: detect physical contiguous object in pool Santosh Shukla
2017-09-07 15:30 ` [dpdk-dev] [PATCH v6 7/8] mempool: introduce block size align flag Santosh Shukla
2017-09-22 12:59 ` Hemant Agrawal
2017-09-25 11:32 ` Olivier MATZ
2017-09-25 22:08 ` santosh
2017-09-07 15:30 ` [dpdk-dev] [PATCH v6 8/8] mempool: notify memory area to pool Santosh Shukla
2017-09-25 11:41 ` Olivier MATZ
2017-09-25 22:18 ` santosh
2017-09-29 4:53 ` santosh
2017-09-29 8:20 ` Olivier MATZ
2017-09-29 8:25 ` santosh
2017-09-13 9:58 ` [dpdk-dev] [PATCH v6 0/8] Infrastructure to support octeontx HW mempool manager santosh
2017-09-19 8:26 ` santosh
2017-10-01 9:28 ` [dpdk-dev] [PATCH v7 " Santosh Shukla
2017-10-01 9:28 ` [dpdk-dev] [PATCH v7 1/8] mempool: remove unused flags argument Santosh Shukla
2017-10-01 9:28 ` [dpdk-dev] [PATCH v7 2/8] mempool: change flags from int to unsigned int Santosh Shukla
2017-10-01 9:28 ` [dpdk-dev] [PATCH v7 3/8] mempool: add flags arg in xmem size and usage Santosh Shukla
2017-10-01 9:28 ` [dpdk-dev] [PATCH v7 4/8] doc: remove mempool notice Santosh Shukla
2017-10-01 9:28 ` [dpdk-dev] [PATCH v7 5/8] mempool: get the mempool capability Santosh Shukla
2017-10-01 9:29 ` [dpdk-dev] [PATCH v7 6/8] mempool: detect physical contiguous object in pool Santosh Shukla
2017-10-01 9:29 ` [dpdk-dev] [PATCH v7 7/8] mempool: introduce block size align flag Santosh Shukla
2017-10-02 8:35 ` santosh
2017-10-02 14:26 ` Olivier MATZ
2017-10-01 9:29 ` [dpdk-dev] [PATCH v7 8/8] mempool: notify memory area to pool Santosh Shukla
2017-10-02 8:36 ` santosh
2017-10-02 14:27 ` Olivier MATZ
2017-10-06 20:00 ` [dpdk-dev] [PATCH v7 0/8] Infrastructure to support octeontx HW mempool manager Thomas Monjalon
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=6586999c-9e80-14b6-f9de-b6bc5683debb@caviumnetworks.com \
--to=santosh.shukla@caviumnetworks.com \
--cc=dev@dpdk.org \
--cc=hemant.agrawal@nxp.com \
--cc=jerin.jacob@caviumnetworks.com \
--cc=olivier.matz@6wind.com \
--cc=thomas@monjalon.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).