DPDK patches and discussions
 help / color / mirror / Atom feed
From: Rajesh Ravi <rajesh.ravi@broadcom.com>
To: "Burakov, Anatoly" <anatoly.burakov@intel.com>
Cc: Ajit Khaparde <ajit.khaparde@broadcom.com>,
	dev@dpdk.org,
	 Jonathan Richardson <jonathan.richardson@broadcom.com>,
	 Scott Branden <scott.branden@broadcom.com>,
	 Vikram Mysore Prakash <vikram.prakash@broadcom.com>,
	Srinath Mannam <srinath.mannam@broadcom.com>
Subject: Re: [dpdk-dev] [PATCH] eal: add option --iso-cmem for external custom memory
Date: Fri, 18 Oct 2019 16:24:11 +0530	[thread overview]
Message-ID: <CAAr5zNeKyCBCEu2YFXMMDeTtxAfGa9UiT0xVVnDfrVfF+op+tg@mail.gmail.com> (raw)
In-Reply-To: <da983f2b-0ad6-fe2e-c83a-599b3168e1fb@intel.com>

+ Srinath

Thanks Anatoly for reviewing this. Please find my reply inline  below:

[Anatoly] First of all, what is "iso-cmem"? It doesn't seem to have any
defined
meaning nor any relation to any existing functionality, and it's not
explained anywhere what is "isolated cmem".
[Rajesh] I 'll correct commit message to include clearer explanation.

*Context & purpose*
We 're using this to improve SPDK performance. When NVMe completion queues
are allocated from a certain memory range,
out of order competions completions  are enabled with our PCIe and it
improves performance.

*Usage*
spdk_env_init()-->(calls rte_eal_init(), and then calls
)spdk_env_dpdk_post_init() [file: spdk/lib/env_dpdk/init.c]
--> spdk_mem_map_init() [lib/env_dpdk/memory.c]-->map_cmem_virtual_area()
[this is our custom function]
In  map_cmem_virtual_area(): we 're mmaping anonymous &  private region and
then creating iova table which makes iova addresses which fall in  custom
memory region.
Then we call rte_malloc_heap_memory_add() and then allocate NVMe completion
queues from this heap.

[Anatoly] More importantly, why is this necessary? Type1 map only bypasses
external segments when adding memory at startup - it doesn't stop you
from calling rte_vfio_dma_map() to map the memory with VFIO when you
create the segment.
[Rajesh] Please correct me if I'm wrong or  missing some thing here.
  A) We wanted to create a heap from which we can allocate dynamically

  B) I see that type1_map()
[file:dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c] getting called once
rte_malloc_heap_memory_add()() was invoked.
       SPDK uses type1 dma map [spdk/lib/env_dpdk/memory.c]
        vfio_type1_dma_map() is registered for .dma_map_func member
[dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c]
        So type1_map is called. Not sure whether we can change this.

Regards,
Rajesh












On Thu, Oct 17, 2019 at 9:15 PM Burakov, Anatoly <anatoly.burakov@intel.com>
wrote:

> On 15-Oct-19 6:30 AM, Ajit Khaparde wrote:
> > From: Rajesh Ravi <rajesh.ravi@broadcom.com>
> >
> > Support external custom memory added to heap to be used with vfio with
> > --iso-cmem option.Type1 memory mapping was by passed for external memory.
> > But an exception is added to allow external custom memory to be used with
> > vfio.
> >
> > Signed-off-by: Rajesh Ravi <rajesh.ravi@broadcom.com>
> > Reviewed-by: Jonathan Richardson <jonathan.richardson@broadcom.com>
> > Reviewed-by: Scott Branden <scott.branden@broadcom.com>
> > Reviewed-by: Vikram Mysore Prakash <vikram.prakash@broadcom.com>
> > Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
> > ---
>
> Hi,
>
> First of all, what is "iso-cmem"? It doesn't seem to have any defined
> meaning nor any relation to any existing functionality, and it's not
> explained anywhere what is "isolated cmem".
>
> More importantly, why is this necessary? Type1 map only bypasses
> external segments when adding memory at startup - it doesn't stop you
> from calling rte_vfio_dma_map() to map the memory with VFIO when you
> create the segment.
>
> --
> Thanks,
> Anatoly
>


-- 
Regards,
Rajesh

  reply	other threads:[~2019-10-21  7:52 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-15  5:30 Ajit Khaparde
2019-10-17 15:45 ` Burakov, Anatoly
2019-10-18 10:54   ` Rajesh Ravi [this message]
2019-10-18 16:53     ` Burakov, Anatoly
2019-10-21 15:46       ` Rajesh Ravi
2019-10-22  7:56         ` Rajesh Ravi
2019-10-24 11:43           ` Burakov, Anatoly
2019-10-25 12:53             ` Rajesh Ravi
2019-10-25 15:02               ` Burakov, Anatoly
2019-10-30 19:50                 ` Rajesh Ravi
2019-11-04 10:25                   ` Burakov, Anatoly
2019-11-05 11:41                     ` Burakov, Anatoly
2019-11-05 14:10                       ` Rajesh Ravi
2019-11-05 15:18                         ` Burakov, Anatoly
2019-11-05 17:13                           ` Burakov, Anatoly
2019-11-06 13:55                             ` David Marchand
2019-11-07 15:51                               ` David Marchand
2019-11-07 16:16                                 ` Rajesh Ravi

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=CAAr5zNeKyCBCEu2YFXMMDeTtxAfGa9UiT0xVVnDfrVfF+op+tg@mail.gmail.com \
    --to=rajesh.ravi@broadcom.com \
    --cc=ajit.khaparde@broadcom.com \
    --cc=anatoly.burakov@intel.com \
    --cc=dev@dpdk.org \
    --cc=jonathan.richardson@broadcom.com \
    --cc=scott.branden@broadcom.com \
    --cc=srinath.mannam@broadcom.com \
    --cc=vikram.prakash@broadcom.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).