DPDK patches and discussions
 help / color / mirror / Atom feed
From: Panu Matilainen <pmatilai@redhat.com>
To: Ferruh Yigit <ferruh.yigit@intel.com>, dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH v3 1/3] mk: clear up libm and librt linkage confusion
Date: Tue, 15 Mar 2016 08:24:09 +0200	[thread overview]
Message-ID: <56E7AA89.9000107@redhat.com> (raw)
In-Reply-To: <56E6EA59.2010300@intel.com>

On 03/14/2016 06:44 PM, Ferruh Yigit wrote:
> On 3/10/2016 1:15 PM, Panu Matilainen wrote:
>> There are two places that need -lm (test app and librte_sched) and
>> exactly one that needs -lrt (librte_sched). Add the relevant
>> DT_NEEDED entries to both, and eliminate the bogus discrepancy
>> between Linux and BSD EXECENV_LDLIBS wrt these libs.
>>
>> Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
>> ---
>>   app/test/Makefile                | 2 ++
>>   lib/librte_sched/Makefile        | 3 +++
>>   mk/exec-env/linuxapp/rte.vars.mk | 2 +-
>>   mk/rte.app.mk                    | 6 ++----
>>   4 files changed, 8 insertions(+), 5 deletions(-)
>>
>> diff --git a/app/test/Makefile b/app/test/Makefile
>> index ec33e1a..00e4df2 100644
>> --- a/app/test/Makefile
>> +++ b/app/test/Makefile
>> @@ -160,6 +160,8 @@ CFLAGS += $(WERROR_FLAGS)
>>
>>   CFLAGS += -D_GNU_SOURCE
>>
>> +LDLIBS += -lm
>> +
>>   # Disable VTA for memcpy test
>>   ifeq ($(CC), gcc)
>>   ifeq ($(shell test $(GCC_VERSION) -ge 44 && echo 1), 1)
>> diff --git a/lib/librte_sched/Makefile b/lib/librte_sched/Makefile
>> index b1cb285..4d631f6 100644
>> --- a/lib/librte_sched/Makefile
>> +++ b/lib/librte_sched/Makefile
>> @@ -41,6 +41,9 @@ CFLAGS += $(WERROR_FLAGS)
>>
>>   CFLAGS_rte_red.o := -D_GNU_SOURCE
>>
>> +LDLIBS += -lm
>> +LDLIBS += -lrt
>> +
>>   EXPORT_MAP := rte_sched_version.map
>>
>>   LIBABIVER := 1
>> diff --git a/mk/exec-env/linuxapp/rte.vars.mk b/mk/exec-env/linuxapp/rte.vars.mk
>> index 5fd7d85..d51bd17 100644
>> --- a/mk/exec-env/linuxapp/rte.vars.mk
>> +++ b/mk/exec-env/linuxapp/rte.vars.mk
>> @@ -48,7 +48,7 @@ endif
>>   # Workaround lack of DT_NEEDED entry
>>   EXECENV_LDFLAGS = --no-as-needed
>>
>> -EXECENV_LDLIBS  = -lrt -lm
>> +EXECENV_LDLIBS  =
>>   EXECENV_ASFLAGS =
>>
>>   ifeq ($(CONFIG_RTE_BUILD_SHARED_LIB),y)
>> diff --git a/mk/rte.app.mk b/mk/rte.app.mk
>> index daac09f..cadc7ab 100644
>> --- a/mk/rte.app.mk
>> +++ b/mk/rte.app.mk
>> @@ -77,11 +77,7 @@ _LDLIBS-$(CONFIG_RTE_LIBRTE_LPM)            += -lrte_lpm
>>   _LDLIBS-$(CONFIG_RTE_LIBRTE_POWER)          += -lrte_power
>>   _LDLIBS-$(CONFIG_RTE_LIBRTE_ACL)            += -lrte_acl
>>   _LDLIBS-$(CONFIG_RTE_LIBRTE_METER)          += -lrte_meter
>> -
>>   _LDLIBS-$(CONFIG_RTE_LIBRTE_SCHED)          += -lrte_sched
>> -_LDLIBS-$(CONFIG_RTE_LIBRTE_SCHED)          += -lm
>> -_LDLIBS-$(CONFIG_RTE_LIBRTE_SCHED)          += -lrt
>> -
>>   _LDLIBS-$(CONFIG_RTE_LIBRTE_VHOST)          += -lrte_vhost
>>
>>   ifeq ($(CONFIG_RTE_LIBRTE_VHOST_NUMA),y)
>> @@ -104,6 +100,8 @@ _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_XENVIRT)    += -lxenstore
>>   _LDLIBS-$(CONFIG_RTE_LIBRTE_MPIPE_PMD)      += -lgxio
>>   # QAT PMD has a dependency on libcrypto (from openssl) for calculating HMAC precomputes
>>   _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_QAT)        += -lcrypto
>> +_LDLIBS-$(CONFIG_RTE_LIBRTE_SCHED)          += -lm
>> +_LDLIBS-$(CONFIG_RTE_LIBRTE_SCHED)          += -lrt
>>   endif # !CONFIG_RTE_BUILD_SHARED_LIBS
>>
>>   _LDLIBS-y += --start-group
>>
> This is causing a shared library compilation error with gcc:
>
> "
> == Build app/test-acl
>    LD testacl
> /root/dpdk/build/lib/librte_meter.so: error: undefined reference to 'ceil'
> collect2: error: ld returned 1 exit status
> "
>
> There is an indirect libm dependency from test-acl. Adding -lm fixes the
> issue.
>
> But this issue not seen by everybody, not sure why I am getting this but
> not others.
>
> Also clang compiles fine, only fails with gcc.
> I am using Fedora 23, gcc version:
> gcc (GCC) 5.3.1 20151207 (Red Hat 5.3.1-2)
>
> I will dig some more.

Right, librte_meter indeed has a dependency on libm that I've missed.
The curious thing is why is it not failing everywhere - I cannot 
reproduce that at all on Fedora 23. Poking around in math.h leads to 
bits/mathlinline.h where the answer probably lies: ceil() is typically 
provided by inline code, but for one reason or another that is not 
available in your setup.

Anyway, I'll send a patch to add the missing dependency, thanks for 
spotting and reporting!

	- Panu -

> Regards,
> ferruh
>
>
>
>

  reply	other threads:[~2016-03-15  6:24 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-08 11:47 [dpdk-dev] [PATCH v2] mk: fix external shared library dependencies of libraries Panu Matilainen
2015-12-08 16:28 ` Sergio Gonzalez Monroy
2015-12-09 12:09   ` Panu Matilainen
2016-03-10 13:15 ` [dpdk-dev] [PATCH v3 0/3] mk: add DT_NEEDED entries for external library deps Panu Matilainen
2016-03-13 19:28   ` Thomas Monjalon
2016-03-10 13:15 ` [dpdk-dev] [PATCH v3 1/3] mk: clear up libm and librt linkage confusion Panu Matilainen
2016-03-14 16:44   ` Ferruh Yigit
2016-03-15  6:24     ` Panu Matilainen [this message]
2016-03-10 13:16 ` [dpdk-dev] [PATCH v3 2/3] mk: add DT_NEEDED entries for librte_vhost external dependencies Panu Matilainen
2016-03-10 13:16 ` [dpdk-dev] [PATCH v3 3/3] mk: add DT_NEEDED entries for librte_eal " Panu Matilainen

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=56E7AA89.9000107@redhat.com \
    --to=pmatilai@redhat.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.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).