DPDK usage discussions
 help / color / mirror / Atom feed
From: Robert Hable <Robert.Hable@spusu.at>
To: Dmitry Kozlyuk <dmitry.kozliuk@gmail.com>
Cc: "users@dpdk.org" <users@dpdk.org>
Subject: AW: rte_eth_dev_count_avail returns different value in library
Date: Fri, 11 Nov 2022 13:51:45 +0000	[thread overview]
Message-ID: <PAXP190MB17411989CEAE70A4DEE39C67E1009@PAXP190MB1741.EURP190.PROD.OUTLOOK.COM> (raw)
In-Reply-To: <20221028134808.78435c9d@sovereign>

Hello Dimitry,

Yes, I am building DPDK as static library. As far as I have seen it should also be the default option on windows builds?

I tried linking the following libraries to my application, which is using my own dpdk-library, but I had no success in using any dpdk functionality (rte_eth_dev_count_avail() returns 0).
In the linker output I found out that the libraries are currently unused:

1>Unused libraries:
1>  C:\Dev\Extern\dpdk\build\lib\librte_eal.a
1>  C:\Dev\Extern\dpdk\build\lib\librte_mbuf.a
1>  C:\Dev\Extern\dpdk\build\lib\librte_mempool.a
1>  C:\Dev\Extern\dpdk\build\lib\librte_ethdev.a
1>  C:\Dev\Extern\dpdk\build\drivers\librte_common_mlx5.a
1>  C:\Dev\Extern\dpdk\build\drivers\librte_net_mlx5.a

Do you have any other ideas what I am missing there and why they are not linked to it? Is it may be possible to include all the PMD libraries already in my own library, so I only need to link against one library?
Yes, rte_eth_dev_count_avail() is called directly after rte_eal_init().

By the way I am using the visual studio 2022 compiler to build my own application and clang for compiling dpdk if that matters.

Kind regards,
Robert

-----Ursprüngliche Nachricht-----
Von: Dmitry Kozlyuk <dmitry.kozliuk@gmail.com> 
Gesendet: Freitag, 28. Oktober 2022 12:48
An: Robert Hable <Robert.Hable@spusu.at>
Cc: users@dpdk.org
Betreff: Re: rte_eth_dev_count_avail returns different value in library

2022-10-28 09:45 (UTC+0000), Robert Hable:
> Hello,
> 
> I am trying to build a static library based on the skeleton example. My problem is when calling rte_eth_dev_count_avail() I always receive 0 in the static library. When I start the skeleton example app as a regular exe file, then it returns 2.
> What could be the reason that rte_eth_dev_count_avail() returns different values depending on if it is an exe application or a static library?
> 
> rte_eal_init() is called in both applications with the default parameters.
> Both processes are running under the same administrator user.
> 
> I am using Windows Server 2022 with a Mellanox ConnectX-4 on DPDK V 22.11.0-rc0.
> 
> Kind regards,
> Robert Hable
> 

Hi Robert,

Are you building DPDK as static libraries?
Shared build is currently non-functional on Windows.

If so, are you linking PMD libraries (librte_common_mlx5, librte_net_mlx5) to the application that is using your static library?
If relevant PMD libraries are not linked, HW will not be discovered (same for mempools BTW).

It is irrelevant from where rte_eth_dev_count_avail() is called, but please check that in both cases it is done after rte_eal_init().

  reply	other threads:[~2022-11-11 13:52 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-28  9:45 Robert Hable
2022-10-28 10:48 ` Dmitry Kozlyuk
2022-11-11 13:51   ` Robert Hable [this message]
2022-11-11 14:48     ` Dmitry Kozlyuk
2022-11-17  9:17       ` AW: " Robert Hable

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=PAXP190MB17411989CEAE70A4DEE39C67E1009@PAXP190MB1741.EURP190.PROD.OUTLOOK.COM \
    --to=robert.hable@spusu.at \
    --cc=dmitry.kozliuk@gmail.com \
    --cc=users@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).