DPDK usage discussions
 help / color / mirror / Atom feed
From: Hui Liu <onlyflyer@gmail.com>
To: "Wiles, Keith" <keith.wiles@intel.com>
Cc: "users@dpdk.org" <users@dpdk.org>
Subject: Re: [dpdk-users] DPDK library link issue with __attribute__((constructor))
Date: Tue, 6 Mar 2018 17:02:15 -0800	[thread overview]
Message-ID: <CAF=f732u52HyXm7zUeNhFvXYncYC8iyjca-1m8i_D4O1FoHA6w@mail.gmail.com> (raw)
In-Reply-To: <CB89733A-C08B-4B5E-89B7-6AD9DDAF8B91@intel.com>

Hi Keith,

Thanks a lot, that resolved my issue.

The key is to use --whole-archive instead of --no-as-needed.

--no-as-needed doesn't work.

Thanks,
Hui

On Mon, Mar 5, 2018 at 5:28 AM, Wiles, Keith <keith.wiles@intel.com> wrote:

>
>
> > On Mar 5, 2018, at 3:15 AM, Hui Liu <onlyflyer@gmail.com> wrote:
> >
> > I tried to put all the library into the command, but it still doesn't
> work.
> > That's wired...
> >
> > x86_64-poky-linux-gcc -m64 -march=core2 -mtune=core2 -msse3 -mfpmath=sse
> > --sysroot=/home/hliu/atomos/build-temp/build/tmp/work/
> core2-64-poky-linux/unittest/0.0-r0/recipe-sysroot
> > -Wl,--hash-style=gnu -Wl,--no-as-needed -o l2_ut_dpdk l2_ut_dpdk.c -Wall
> > -Werror -g -lrte_mempool_stack -lrte_pmd_i40e -lrte_pmd_ark -lrte_ethdev
> > -lrte_pmd_null -lrte_pmd_null_crypto -lrte_lpm -lrte_pmd_ena -lrte_kvargs
> > -lrte_cmdline -lrte_pmd_tap -lrte_pmd_sw_event -lrte_latencystats
> > -lrte_bitratestats -lrte_efd -lrte_sched -lrte_pmd_kni -lrte_pmd_qede
> > -lrte_hash -lrte_pmd_enic -lrte_pdump -lrte_pmd_lio -lrte_acl
> > -lrte_pmd_af_packet -lrte_eventdev -lrte_port -lrte_ip_frag
> -lrte_pmd_bond
> > -lrte_pmd_sfc_efx -lrte_pmd_e1000 -lrte_mbuf -lrte_pmd_virtio
> -lrte_metrics
> > -lrte_cryptodev -lrte_reorder -lrte_pmd_ring -lrte_eal -lrte_distributor
> > -lrte_vhost -lrte_pmd_bnxt -lrte_timer -lrte_pmd_avp
> > -lrte_pmd_crypto_scheduler -lrte_pmd_thunderx_nicvf
> > -lrte_pmd_skeleton_event -lrte_pmd_nfp -lrte_jobstats -lrte_net
> > -lrte_mempool -lrte_pmd_ixgbe -lrte_mempool_ring -lrte_pipeline
> > -lrte_pmd_octeontx_ssovf -lrte_pmd_vhost -lrte_meter -lrte_pmd_cxgbe
> > -lrte_kni -lrte_pmd_vmxnet3_uio -lrte_ring -lrte_pmd_fm10k -lrte_table
> > -lrte_cfgfile -lpthread -ldl
>
> One way with the above is to use —whole-archive … —no-whole-archive around
> the archives you want to force include. The other way to is include the .o
> files as part of the link line instead of the .a file. The problem with
> —whole-archive is it pulls in the complete archive and increases footprint.
> Using the .o method is normally used here and they include these .o files
> in the application Makefile.
>
> >
>
> > On Sun, Mar 4, 2018 at 9:08 PM, Hui Liu <onlyflyer@gmail.com> wrote:
> >
> >> Hi Guys,
> >>
> >> I am using DPDK with the yocto build system.
> >>
> >> I found some static library is not linked because of the linker's
> >> as-needed decision.
> >>
> >> Some .a file use __attribute__((constructor)) to register their APIs,
> >> we only know if we use these APIs at runtime, but linker make the
> decision
> >> early and then discarded them.
> >>
> >> I tried a lot to link the library I need:
> >>
> >> 1, use -Wl, --no-as-needed, for example:
> >> gcc a.c *-Wl, --no-as-needed* -ldl -ldpdk
> >>
> >> 2, Put the library explicitly to the linker like this:
> >> gcc a.c -Wl, --no-as-needed -ldl -ldpdk *-lrte_mempool*
> >>
> >> At last, I still didn't find my library in the elf file.
> >>
> >> Is there anything else I can try to link these library?
> >>
> >> --
> >> Thanks & Best Regards
> >> Liu Hui
> >> --
> >>
> >
> >
> >
> > --
> > Thanks & Best Regards
> > Liu Hui
> > --
>
> Regards,
> Keith
>
>


-- 
Thanks & Best Regards
Liu Hui
--

      reply	other threads:[~2018-03-07  1:02 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-03-05  5:08 Hui Liu
2018-03-05  9:15 ` Hui Liu
2018-03-05 13:28   ` Wiles, Keith
2018-03-07  1:02     ` Hui Liu [this message]

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='CAF=f732u52HyXm7zUeNhFvXYncYC8iyjca-1m8i_D4O1FoHA6w@mail.gmail.com' \
    --to=onlyflyer@gmail.com \
    --cc=keith.wiles@intel.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).