DPDK patches and discussions
 help / color / mirror / Atom feed
From: Bruce Richardson <bruce.richardson@intel.com>
To: Tom Jones <thj@freebsd.org>
Cc: <dev@dpdk.org>
Subject: Re: [PATCH] freebsd: Add support for multiple dpdk instances on FreeBSD
Date: Fri, 3 May 2024 17:52:03 +0100	[thread overview]
Message-ID: <ZjUWM_tH7wg0bg95@bricha3-mobl1.ger.corp.intel.com> (raw)
In-Reply-To: <ZjUVRXNZ5Kh8-7Px@bricha3-mobl1.ger.corp.intel.com>

On Fri, May 03, 2024 at 05:48:05PM +0100, Bruce Richardson wrote:
> On Fri, May 03, 2024 at 09:46:15AM +0000, Tom Jones wrote:
> > Add support to the contigmem module on FreeBSD for multiple concurrent
> > files, this enables running multiple dpdk instances with the nic_uio
> > driver.
> > 
> > Add relevant parts in dpdk to support this.
> > 
> > Signed-off-by: Tom Jones <thj@freebsd.org>
> > ---
> >  config/rte_config.h                  |   2 +
> >  kernel/freebsd/contigmem/contigmem.c | 225 ++++++++++++++++++---------
> >  lib/eal/common/eal_common_options.c  |   2 +
> >  lib/eal/freebsd/eal.c                |  12 ++
> >  lib/eal/freebsd/eal_hugepage_info.c  |  12 +-
> >  lib/eal/freebsd/eal_memory.c         |   3 +-
> >  6 files changed, 180 insertions(+), 76 deletions(-)
> > 
> Hi Tom,
> 
> on testing this out the biggest gap I see is that we need a documentation
> update for it. In V3 can you please update the FreeBSD GSG doc with some
> details about how to use this feature.
> 
> Also, on first run using testpmd, the application failed to start, as I was
> missing a file-prefix option. At minimum I think the code needs to default
> the file-prefix on BSD to "contigmem0", so that when running a standalone
> app as before, it "just works". An idea for backward compatibility here
> that may help us, is to omit the "0" for contigmem0 device. That would mean
> that an older DPDK app will work with a newer kernel driver, and vice
> versa.
> 
> Beyond that, I'm not sure how to get multiple processes working right now
> (hence the doc update request!). When I run a second instance of DPDK on my
> system with "--file-prefix=contigmem1", I get errors about the shared
> config file in "/var/run/dpdk/rte/config".
> 
Inspiration struck me just after I hit send on this email. Removing the
#ifdefs in eal_common_options.c allows me to run two processes in
parallel just fine.

I think we need to consider how to rework this to avoid using the
file-prefix argument for the contigmem files. Not sure what a better scheme
is, though. Do you have any ideas or suggestions on this from your
experience doing up this patch?

Thanks,
/Bruce

  reply	other threads:[~2024-05-03 16:52 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-05-02 13:55 Tom Jones
2024-05-03  9:46 ` Tom Jones
2024-05-03 13:03   ` Bruce Richardson
2024-05-03 13:12     ` Tom Jones
2024-05-03 13:24       ` Bruce Richardson
2024-05-03 16:25   ` Bruce Richardson
2024-05-03 16:48   ` Bruce Richardson
2024-05-03 16:52     ` Bruce Richardson [this message]
2024-05-06 15:34       ` Tom Jones

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=ZjUWM_tH7wg0bg95@bricha3-mobl1.ger.corp.intel.com \
    --to=bruce.richardson@intel.com \
    --cc=dev@dpdk.org \
    --cc=thj@freebsd.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).