From: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
To: Olivier Matz <olivier.matz@6wind.com>
Cc: dev@dpdk.org, "Morten Brørup" <mb@smartsharesystems.com>,
"Beilei Xing" <beilei.xing@intel.com>,
"Bruce Richardson" <bruce.richardson@intel.com>,
"Jerin Jacob Kollanukkaran" <jerinj@marvell.com>
Subject: Re: [PATCH v3] mempool: fix get objects from mempool with cache
Date: Wed, 5 Oct 2022 12:56:03 +0300 [thread overview]
Message-ID: <ea31a687-11be-6a3e-956d-2bbb980a624d@oktetlabs.ru> (raw)
In-Reply-To: <20221005095037.997006-2-andrew.rybchenko@oktetlabs.ru>
I'm sorry, below duplicate is sent my mistake.
On 10/5/22 12:50, Andrew Rybchenko wrote:
> From: Morten Brørup <mb@smartsharesystems.com>
>
> A flush threshold for the mempool cache was introduced in DPDK version
> 1.3, but rte_mempool_do_generic_get() was not completely updated back
> then, and some inefficiencies were introduced.
>
> Fix the following in rte_mempool_do_generic_get():
>
> 1. The code that initially screens the cache request was not updated
> with the change in DPDK version 1.3.
> The initial screening compared the request length to the cache size,
> which was correct before, but became irrelevant with the introduction of
> the flush threshold. E.g. the cache can hold up to flushthresh objects,
> which is more than its size, so some requests were not served from the
> cache, even though they could be.
> The initial screening has now been corrected to match the initial
> screening in rte_mempool_do_generic_put(), which verifies that a cache
> is present, and that the length of the request does not overflow the
> memory allocated for the cache.
>
> This bug caused a major performance degradation in scenarios where the
> application burst length is the same as the cache size. In such cases,
> the objects were not ever fetched from the mempool cache, regardless if
> they could have been.
> This scenario occurs e.g. if an application has configured a mempool
> with a size matching the application's burst size.
>
> 2. The function is a helper for rte_mempool_generic_get(), so it must
> behave according to the description of that function.
> Specifically, objects must first be returned from the cache,
> subsequently from the ring.
> After the change in DPDK version 1.3, this was not the behavior when
> the request was partially satisfied from the cache; instead, the objects
> from the ring were returned ahead of the objects from the cache.
> This bug degraded application performance on CPUs with a small L1 cache,
> which benefit from having the hot objects first in the returned array.
> (This is probably also the reason why the function returns the objects
> in reverse order, which it still does.)
> Now, all code paths first return objects from the cache, subsequently
> from the ring.
>
> The function was not behaving as described (by the function using it)
> and expected by applications using it. This in itself is also a bug.
>
> 3. If the cache could not be backfilled, the function would attempt
> to get all the requested objects from the ring (instead of only the
> number of requested objects minus the objects available in the ring),
> and the function would fail if that failed.
> Now, the first part of the request is always satisfied from the cache,
> and if the subsequent backfilling of the cache from the ring fails, only
> the remaining requested objects are retrieved from the ring.
>
> The function would fail despite there are enough objects in the cache
> plus the common pool.
>
> 4. The code flow for satisfying the request from the cache was slightly
> inefficient:
> The likely code path where the objects are simply served from the cache
> was treated as unlikely. Now it is treated as likely.
>
> Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
> Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
> ---
> v3 changes (Andrew Rybchenko)
> - Always get first objects from the cache even if request is bigger
> than cache size. Remove one corresponding condition from the path
> when request is fully served from cache.
> - Simplify code to avoid duplication:
> - Get objects directly from backend in single place only.
> - Share code which gets from the cache first regardless if
> everythihg is obtained from the cache or just the first part.
> - Rollback cache length in unlikely failure branch to avoid cache
> vs NULL check in success branch.
>
> v2 changes
> - Do not modify description of return value. This belongs in a separate
> doc fix.
> - Elaborate even more on which bugs the modifications fix.
>
> lib/mempool/rte_mempool.h | 74 +++++++++++++++++++++++++--------------
> 1 file changed, 48 insertions(+), 26 deletions(-)
>
next prev parent reply other threads:[~2022-10-05 9:56 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-06-24 10:23 [PATCH] net/af_xdp: make compatible with libbpf v0.8.0 Ciara Loftus
2022-06-24 11:45 ` Andrew Rybchenko
2022-06-27 14:17 ` Loftus, Ciara
2022-06-27 14:50 ` Andrew Rybchenko
2022-06-27 15:24 ` Loftus, Ciara
2022-06-28 9:15 ` Andrew Rybchenko
2022-06-28 10:07 ` Loftus, Ciara
2022-07-21 12:16 ` Loftus, Ciara
2022-06-28 12:18 ` [PATCH v2] " Ciara Loftus
2022-10-05 9:50 ` [PATCH v3 0/6] " Andrew Rybchenko
2022-10-05 9:50 ` [PATCH v3] mempool: fix get objects from mempool with cache Andrew Rybchenko
2022-10-05 9:56 ` Andrew Rybchenko [this message]
2022-10-05 9:50 ` [PATCH v3 1/6] net/af_xdp: move XDP library presence flag to right branch Andrew Rybchenko
2022-10-05 9:50 ` [PATCH v3 2/6] net/af_xdp: make it clear which libxdp version is required Andrew Rybchenko
2022-10-05 9:50 ` [PATCH v3 3/6] net/af_xdp: avoid version-based check for shared UMEM Andrew Rybchenko
2022-10-05 9:50 ` [PATCH v3 4/6] net/af_xdp: avoid version-based check for program load mech Andrew Rybchenko
2022-10-05 9:50 ` [PATCH v3 5/6] net/af_xdp: log errors on XDP program removal failures Andrew Rybchenko
2022-10-05 9:50 ` [PATCH v3 6/6] net/af_xdp: make compatible with libbpf v0.8.0 Andrew Rybchenko
2022-10-06 6:26 ` [PATCH v4 0/6] " Andrew Rybchenko
2022-10-06 6:26 ` [PATCH v4 1/6] net/af_xdp: move XDP library presence flag to right branch Andrew Rybchenko
2022-10-06 6:26 ` [PATCH v4 2/6] net/af_xdp: make it clear which libxdp version is required Andrew Rybchenko
2022-10-06 6:26 ` [PATCH v4 3/6] net/af_xdp: avoid version-based check for shared UMEM Andrew Rybchenko
2022-10-06 6:26 ` [PATCH v4 4/6] net/af_xdp: avoid version-based check for program load mech Andrew Rybchenko
2022-10-06 6:26 ` [PATCH v4 5/6] net/af_xdp: log errors on XDP program removal failures Andrew Rybchenko
2022-10-06 6:26 ` [PATCH v4 6/6] net/af_xdp: make compatible with libbpf v0.8.0 Andrew Rybchenko
2022-10-07 17:19 ` Ferruh Yigit
2022-10-07 17:28 ` Ferruh Yigit
2022-10-07 17:40 ` [PATCH v4 0/6] " Ferruh Yigit
2022-12-20 14:05 ` [PATCH] " Kevin Traynor
2022-12-21 6:09 ` Andrew Rybchenko
2022-12-21 9:28 ` Kevin Traynor
2023-03-15 11:47 ` Kevin Traynor
2023-03-16 13:31 ` Kevin Traynor
2023-03-23 10:23 ` Kevin Traynor
2023-04-04 15:51 ` Kevin Traynor
-- strict thread matches above, loose matches on Subject: below --
2021-12-26 15:34 [RFC] mempool: rte_mempool_do_generic_get optimizations Morten Brørup
2022-10-04 12:53 ` [PATCH v3] mempool: fix get objects from mempool with cache Andrew Rybchenko
2022-10-04 14:42 ` Morten Brørup
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=ea31a687-11be-6a3e-956d-2bbb980a624d@oktetlabs.ru \
--to=andrew.rybchenko@oktetlabs.ru \
--cc=beilei.xing@intel.com \
--cc=bruce.richardson@intel.com \
--cc=dev@dpdk.org \
--cc=jerinj@marvell.com \
--cc=mb@smartsharesystems.com \
--cc=olivier.matz@6wind.com \
/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).