From: Jerin Jacob <jerin.jacob@caviumnetworks.com>
To: Ophir Munk <ophirmu@mellanox.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>,
Thomas Monjalon <thomas@monjalon.net>,
Pascal Mazon <pascal.mazon@6wind.com>,
Jan Viktorin <viktorin@rehivetech.com>,
Jianbo Liu <jianbo.liu@arm.com>
Subject: Re: [dpdk-dev] [PATCH] net/tap: fix build issue with ARM32
Date: Wed, 31 Jan 2018 17:00:50 +0530 [thread overview]
Message-ID: <20180131113049.GA30737@jerin> (raw)
In-Reply-To: <AM0PR0502MB3875A2B143A6D122471D431BD1FB0@AM0PR0502MB3875.eurprd05.prod.outlook.com>
-----Original Message-----
> Date: Wed, 31 Jan 2018 10:38:32 +0000
> From: Ophir Munk <ophirmu@mellanox.com>
> To: Jerin Jacob <jerin.jacob@caviumnetworks.com>
> CC: "dev@dpdk.org" <dev@dpdk.org>, Thomas Monjalon <thomas@monjalon.net>,
> Pascal Mazon <pascal.mazon@6wind.com>, Jan Viktorin
> <viktorin@rehivetech.com>, Jianbo Liu <jianbo.liu@arm.com>
> Subject: RE: [dpdk-dev] [PATCH] net/tap: fix build issue with ARM32
>
> Hi,
> Please see inline
>
> > -----Original Message-----
> > From: Jerin Jacob [mailto:jerin.jacob@caviumnetworks.com]
> > Sent: Wednesday, January 31, 2018 12:24 PM
> > To: Ophir Munk <ophirmu@mellanox.com>
> > Cc: dev@dpdk.org; Thomas Monjalon <thomas@monjalon.net>; Pascal
> > Mazon <pascal.mazon@6wind.com>; Jan Viktorin
> > <viktorin@rehivetech.com>; Jianbo Liu <jianbo.liu@arm.com>
> > Subject: Re: [dpdk-dev] [PATCH] net/tap: fix build issue with ARM32
> >
> > -----Original Message-----
> > > Date: Wed, 31 Jan 2018 09:59:45 +0000
> > > From: Ophir Munk <ophirmu@mellanox.com>
> > > To: Jerin Jacob <jerin.jacob@caviumnetworks.com>, "dev@dpdk.org"
> > > <dev@dpdk.org>
> > > CC: Thomas Monjalon <thomas@monjalon.net>, Pascal Mazon
> > > <pascal.mazon@6wind.com>, Jan Viktorin <viktorin@rehivetech.com>,
> > > Jianbo Liu <jianbo.liu@arm.com>
> > > Subject: RE: [dpdk-dev] [PATCH] net/tap: fix build issue with ARM32
> > >
> > > Hi,
> > > Please see comments inline
> > >
> > > > -----Original Message-----
> > > > From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Jerin Jacob
> > > > Sent: Wednesday, January 31, 2018 7:14 AM
> > > > To: dev@dpdk.org
> > > > Cc: Thomas Monjalon <thomas@monjalon.net>; Jerin Jacob
> > > > <jerin.jacob@caviumnetworks.com>; Pascal Mazon
> > > > <pascal.mazon@6wind.com>; Jan Viktorin <viktorin@rehivetech.com>;
> > > > Jianbo Liu <jianbo.liu@arm.com>
> > > > Subject: [dpdk-dev] [PATCH] net/tap: fix build issue with ARM32
> > > >
> > > > arm32 Linux does not support eBPF interface.
> > > > Since tap PMD driver has a dependency on eBPF(the symbol __NR_bpf),
> > > > disabling it armv7 config.
> > > >
> > > > compilation error log:
> > > > In file included from /tmp/dpdk/drivers/net/tap/tap_bpf_api.c:15:0:
> > > > /tmp/dpdk/drivers/net/tap/tap_bpf.h:103:4: error: #error __NR_bpf
> > > > not defined # error __NR_bpf not defined
> > > > ^
> > >
> > > arm architecture is missing is file drivers/tap/tap_bpf.h I suggest
> > > adding it, for example
> > >
> > > # elif defined(__arm__)
> > > # define __NR_bpf 386
> >
> > By looking at Linux source code
> > https://emea01.safelinks.protection.outlook.com/?url=http%3A%2F%2Felixi
> > r.free-
> > electrons.com%2Flinux%2Fv4.15%2Fsource%2Ftools%2Fbuild%2Ffeature%2
> > Ftest-
> > bpf.c%23L6&data=02%7C01%7Cophirmu%40mellanox.com%7C19f50ad7f27
> > 34d00173808d56894c997%7Ca652971c7d2e4d9ba6a4d149256f461b%7C0%
> > 7C0%7C636529910613736017&sdata=ynsxHG9fvao4LyfWLo4GVJjlKNmcF0q
> > 0BkFocJyarAE%3D&reserved=0
> > I am not sure, Is eBPF support available for arm32 support?
>
> TAP eBPF has a graceful approach: it must successfully compile on all Linux
> distributions. If a specific kernel cannot support eBPF it will gracefully
> refuse the eBPF netlink message sent to it.
> Thus, there is no harm even if there is no eBPF support available for arm32.
OK. But, To enable grace full refuse, we should not any random system call number like #define __NR_bpf 386.
It should at least unused for arm32.
>
> You should enable TAP compilation for cases where you only need TAP without eBPF.
Yes. It would have been better if new eBPF depended features would have
under conditional compilation flag so it it can be turned off selectively.
>
> >
> > >
> > > It should work for you. Can you please try it?
> >
> > I don't have access to an arm32 board. I was facing cross compilation build
> > issue with arm32.
> >
> > > If it doesn't work for you, please have a look in this link:
> > >
> > >
> > https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsou
> > >
> > rceforge.net%2Fp%2Fpredef%2Fwiki%2FArchitectures%2F&data=02%7C01%
> > 7Coph
> > >
> > irmu%40mellanox.com%7C19f50ad7f2734d00173808d56894c997%7Ca6529
> > 71c7d2e4
> > >
> > d9ba6a4d149256f461b%7C0%7C0%7C636529910613736017&sdata=g5fEkG
> > Gwm%2BJgR
> > > KyDrxJgbFII0%2B1%2BxtbT%2FE4Qy1HACEg%3D&reserved=0
> > > where you can see more arm architecture definitions.
> > > You should find the one suitable for you.
> >
> > I guess, it is not just matter of enabling the build for arm32. It should work at
> > runtime too, If some can confirm following addition in the
> > drivers/net/tap/tap_bpf.h works then we can add it.
> >
>
> As mentioned above it must not work at runtime too.
>
> > # elif defined(__arm__)
> > # define __NR_bpf 386
> >
> > >
> > > > /tmp/dpdk/drivers/net/tap/tap_bpf_api.c: In function ‘sys_bpf’:
> > > > /tmp/dpdk/drivers/net/tap/tap_bpf_api.c:92:17: error: ‘__NR_bpf’
> > > > undeclared (first use in this function)
> > > > return syscall(__NR_bpf, cmd, attr, size);
> > > > ^
> > > > /tmp/dpdk/drivers/net/tap/tap_bpf_api.c:92:17: note: each undeclared
> > > > identifier is reported only once for each function it appears in
> > > > /tmp/dpdk/drivers/net/tap/tap_bpf_api.c:93:1: error: control reaches
> > > > end of non-void function [-Werror=return-type] } ^
> > > > cc1: all warnings being treated as errors
> > > >
> > > > Fixes: b02d85e1 ("net/tap: add eBPF API")
> > > >
> > > > Cc: Pascal Mazon <pascal.mazon@6wind.com>
> > > > Cc: Jan Viktorin <viktorin@rehivetech.com>
> > > > Cc: Jianbo Liu <jianbo.liu@arm.com>
> > > >
> > > > Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
> > > > ---
> > > > config/defconfig_arm-armv7a-linuxapp-gcc | 1 +
> > > > 1 file changed, 1 insertion(+)
> > > >
> > > > diff --git a/config/defconfig_arm-armv7a-linuxapp-gcc
> > > > b/config/defconfig_arm-armv7a-linuxapp-gcc
> > > > index a20b7a85f..4a55bb397 100644
> > > > --- a/config/defconfig_arm-armv7a-linuxapp-gcc
> > > > +++ b/config/defconfig_arm-armv7a-linuxapp-gcc
> > > > @@ -80,3 +80,4 @@ CONFIG_RTE_LIBRTE_BNX2X_PMD=n
> > > > CONFIG_RTE_LIBRTE_QEDE_PMD=n
> > CONFIG_RTE_LIBRTE_SFC_EFX_PMD=n
> > > > CONFIG_RTE_LIBRTE_AVP_PMD=n
> > > > +CONFIG_RTE_LIBRTE_PMD_TAP=n
> > > > --
> > >
> > > I suggest fixing the compilation issue by adding the missing
> > > architecture system call number (as suggested above) instead of
> > > disabling TAP compilation by default
> > >
> > > > 2.16.1
> > >
next prev parent reply other threads:[~2018-01-31 11:31 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-01-31 5:14 Jerin Jacob
2018-01-31 9:59 ` Ophir Munk
2018-01-31 10:23 ` Jerin Jacob
2018-01-31 10:38 ` Ophir Munk
2018-01-31 11:19 ` Thomas Monjalon
2018-01-31 11:30 ` Jerin Jacob [this message]
2018-01-31 11:57 ` Ophir Munk
2018-01-31 12:40 ` Jerin Jacob
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=20180131113049.GA30737@jerin \
--to=jerin.jacob@caviumnetworks.com \
--cc=dev@dpdk.org \
--cc=jianbo.liu@arm.com \
--cc=ophirmu@mellanox.com \
--cc=pascal.mazon@6wind.com \
--cc=thomas@monjalon.net \
--cc=viktorin@rehivetech.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).