DPDK patches and discussions
 help / color / mirror / Atom feed
From: Maxime Coquelin <maxime.coquelin@redhat.com>
To: David Marchand <david.marchand@redhat.com>, dev@dpdk.org
Cc: Chenbo Xia <chenbo.xia@intel.com>
Subject: Re: [PATCH v3 4/4] vhost: stop using mempool for IOTLB cache
Date: Tue, 26 Jul 2022 11:26:45 +0200	[thread overview]
Message-ID: <551fa8d0-9855-cd8f-7d5b-5118f3583b7e@redhat.com> (raw)
In-Reply-To: <20220725203206.427083-5-david.marchand@redhat.com>



On 7/25/22 22:32, David Marchand wrote:
> A mempool consumes 3 memzones (with the default ring mempool driver).
> The default DPDK configuration allows RTE_MAX_MEMZONE (2560) memzones.
> 
> Assuming there is no other memzones that means that we can have a
> maximum of 853 mempools.
> 
> In the vhost library, the IOTLB cache code so far was requesting a
> mempool per vq, which means that at the maximum, the vhost library
> could request mempools for 426 qps.
> 
> This limit was recently reached on big systems with a lot of virtio
> ports (and multiqueue in use).
> 
> While the limit on mempool count could be something we fix at the DPDK
> project level, there is no reason to use mempools for the IOTLB cache:
> - the IOTLB cache entries do not need to be DMA-able and are only used
>    by the current process (in multiprocess context),
> - getting/putting objects from/in the mempool is always associated with
>    some other locks, so some level of lock contention is already present,
> 
> We can convert to a malloc'd pool with objects put in a free list
> protected by a spinlock.
> 
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> ---
>   lib/vhost/iotlb.c | 102 ++++++++++++++++++++++++++++------------------
>   lib/vhost/iotlb.h |   1 +
>   lib/vhost/vhost.c |   2 +-
>   lib/vhost/vhost.h |   4 +-
>   4 files changed, 67 insertions(+), 42 deletions(-)
> 

Thanks for working on this, this is definitely not needed to use mempool
for this.

Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

Maxime


  reply	other threads:[~2022-07-26  9:26 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-07-22 13:53 [PATCH 1/2] vhost: keep a reference to virtqueue index David Marchand
2022-07-22 13:53 ` [PATCH 2/2] vhost: stop using mempool for IOTLB David Marchand
2022-07-22 14:00 ` [PATCH 1/2] vhost: keep a reference to virtqueue index David Marchand
2022-07-22 15:13 ` David Marchand
2022-07-25  7:11 ` [PATCH v2 " David Marchand
2022-07-25  7:11   ` [PATCH v2 2/2] vhost: stop using mempool for IOTLB cache David Marchand
2022-07-25 20:32 ` [PATCH v3 0/4] vHost IOTLB cache rework David Marchand
2022-07-25 20:32   ` [PATCH v3 1/4] vhost: fix vq use after free on NUMA reallocation David Marchand
2022-07-26  7:55     ` Maxime Coquelin
2022-09-13 15:02       ` Maxime Coquelin
2022-09-14  1:05         ` Xia, Chenbo
2022-09-14  7:14           ` Maxime Coquelin
2022-09-14  9:15             ` Thomas Monjalon
2022-09-14  9:34               ` Maxime Coquelin
2022-09-14  9:45                 ` Thomas Monjalon
2022-09-14 11:48                   ` Maxime Coquelin
2022-07-25 20:32   ` [PATCH v3 2/4] vhost: make NUMA reallocation code more robust David Marchand
2022-07-26  8:39     ` Maxime Coquelin
2022-07-25 20:32   ` [PATCH v3 3/4] vhost: keep a reference to virtqueue index David Marchand
2022-07-26  8:52     ` Maxime Coquelin
2022-07-26 10:00       ` David Marchand
2022-07-25 20:32   ` [PATCH v3 4/4] vhost: stop using mempool for IOTLB cache David Marchand
2022-07-26  9:26     ` Maxime Coquelin [this message]
2023-02-17  7:42       ` Yuan, DukaiX
2022-09-15 16:02   ` [PATCH v3 0/4] vHost IOTLB cache rework 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=551fa8d0-9855-cd8f-7d5b-5118f3583b7e@redhat.com \
    --to=maxime.coquelin@redhat.com \
    --cc=chenbo.xia@intel.com \
    --cc=david.marchand@redhat.com \
    --cc=dev@dpdk.org \
    /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).