DPDK patches and discussions
 help / color / mirror / Atom feed
From: Zhangfei Gao <zhangfei.gao@linaro.org>
To: Thomas Monjalon <thomas@monjalon.net>,
	Zhangfei Gao <zhangfei.gao@linaro.org>,
	Akhil Goyal <gakhil@marvell.com>
Cc: "techboard@dpdk.org" <techboard@dpdk.org>,
	"dev@dpdk.org" <dev@dpdk.org>,
	 "acc@openeuler.org" <acc@openeuler.org>
Subject: Re: crypto/uadk: introduce uadk crypto driver
Date: Wed, 19 Oct 2022 15:24:41 +0800	[thread overview]
Message-ID: <75c0d590-0711-4e8c-8637-e928a89d9800@linaro.org> (raw)
In-Reply-To: <5877730.88bMQJbFj6@thomas>



On 2022/10/19 下午2:55, Thomas Monjalon wrote:
> 18/10/2022 12:52, Zhangfei Gao:
>> On 2022/10/18 下午6:07, Zhangfei Gao wrote:
>>> Hi, Akhil
>>>
>>>> On 2022/10/17 下午3:02, Akhil Goyal wrote:
>>>>> Hi everyone,
>>>>>
>>>>> We have a new crypto PMD submitted on ML
>>>>> (http://patches.dpdk.org/project/dpdk/cover/20221008083747.6559-1-zhangfei.gao@linaro.org/)
>>>>>
>>>>> This PMD is dependent on an external library which needs to be cross
>>>>> compiled for ARM on x86.
>>>>>
>>>>> It mandatorily need “numactl” also cross-compiled and instead of
>>>>> using pkg-config of libnuma, it need the library and headers to be
>>>>> copied manually to aarch toolchain.
>>>>>
>>>>> And if that is also done, it is not generating pkg-config which DPDK
>>>>> can use to compile the PMD.
>>>>>
>>>> Thanks for the guidance
>>>>
>>>> We are trying to provide pkgconfig file for uadk libs
>>>> https://github.com/Linaro/uadk/pull/513
>>>>
>>>> And have fixed the x86 local build.
>>>> https://github.com/Linaro/uadk/pull/512
>>> Under this git pull, I have verified building and installing dpdk on
>>> x86 with this diff to switch to pkg-config
>>>
>>> diff --git a/drivers/crypto/uadk/meson.build
>>> b/drivers/crypto/uadk/meson.build
>>> index dda6c6b257..f6fae0a239 100644
>>> --- a/drivers/crypto/uadk/meson.build
>>> +++ b/drivers/crypto/uadk/meson.build
>>> @@ -8,18 +8,12 @@ if not is_linux
>>>       subdir_done()
>>>   endif
>>>
>>> -if arch_subdir != 'arm' or not dpdk_conf.get('RTE_ARCH_64')
>>> -    build = false
>>> -    reason = 'only supported on aarch64'
>>> -    subdir_done()
>>> -endif
>>> -
>>>   sources = files(
>>>           'uadk_crypto_pmd.c',
>>>   )
>>>
>>>   deps += 'bus_vdev'
>>> -dep = cc.find_library('libwd_crypto', required: false)
>>> +dep = dependency('libwd_crypto', required: false, method: 'pkg-config')
>>>   if not dep.found()
>>>          build = false
>>>          reason = 'missing dependency, "libwd_crypto"'
>>> @@ -27,7 +21,7 @@ else
>>>          ext_deps += dep
>>>   endif
>>>
>>> -dep = cc.find_library('libwd', required: false)
>>> +dep = dependency('libwd', required: false, method: 'pkg-config')
>>>   if not dep.found()
>>>          build = false
>>>          reason = 'missing dependency, "libwd"'
>>>
>>>
>>> One uncertainty is numa,  "Requires: numa",
>>> This requires numa.pc has to be installed, so numa only can be
>>> installed from source code (make install).
>>> While apt-get install libnuma-dev does not install numa.pc, so system
>>> can not search numa.
>>> So do we have to build numa from source code?
>> By the way.
>>
>> When trying to use +Libs.Private: -lnuma,
>> I found Even remove building uadk, the dpdk fails if no numa.pc, though
>> apt-get insall libnuma-dev
>>
>> /usr/bin/ld: lib/librte_eal.so.23.0: version node not found for symbol
>> numa_run_on_node_mask@@libnuma_1.2
>> /usr/bin/ld: failed to set dynamic section sizes: bad value
>> collect2: error: ld returned 1 exit status
>> [2813/3314] Compiling C object
>> 'drivers/a715181@@tmp_rte_event_cnxk@sta/event_cnxk_tx_cn10k_tx_112_127_seg.c.o'.
>> ninja: build stopped: subcommand failed.
> libnuma is a general requirement of DPDK.
> numa.pc should be installed by the NUMA package.
> Maybe a problem with your distribution.

It looks it is the numa source code issue.

Only if we do not install numa by source code,
apt-get install libnuma-dev always works.

But if install numa from source code once via make install,
apt-get seems not workable anymore, same on x86 or arm.
It only install symbol (libnuma.so), but no real library (libnuma.so.1.0.0)

$ sudo apt-get install libnuma-dev
$ numactl
numactl: error while loading shared libraries: libnuma.so.1: cannot open 
shared object file: No such file or directory
$ ll /usr/lib/aarch64-linux-gnu/libnuma*
-rw-r--r-- 1 root root 77642 Sep 8 2018 /usr/lib/aarch64-linux-gnu/libnuma.a
lrwxrwxrwx 1 root root 16 Sep 8 2018 
/usr/lib/aarch64-linux-gnu/libnuma.so -> libnuma.so.1.0.0

More details, https://github.com/Linaro/uadk/issues/515

But never mind, will set numa as private, then both case can work, so 
will ignore this issue first.

+Libs.Private: -lnuma

Thanks

>
>
>


  reply	other threads:[~2022-10-19  7:24 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-17  7:02 Akhil Goyal
2022-10-17 14:02 ` Zhangfei Gao
2022-10-18 10:07   ` Zhangfei Gao
2022-10-18 10:52     ` Zhangfei Gao
2022-10-19  6:55       ` Thomas Monjalon
2022-10-19  7:24         ` Zhangfei Gao [this message]
2022-10-19 13:19 ` Zhangfei Gao
2022-10-19 14:25   ` [EXT] " Akhil Goyal
2022-10-19 15:27     ` Zhangfei Gao
2022-10-20  2:48       ` Zhangfei Gao

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=75c0d590-0711-4e8c-8637-e928a89d9800@linaro.org \
    --to=zhangfei.gao@linaro.org \
    --cc=acc@openeuler.org \
    --cc=dev@dpdk.org \
    --cc=gakhil@marvell.com \
    --cc=techboard@dpdk.org \
    --cc=thomas@monjalon.net \
    /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).