From: "Ananyev, Konstantin" <konstantin.ananyev@intel.com>
To: yang_y_yi <yang_y_yi@163.com>, "Hu, Jiayu" <jiayu.hu@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>,
"olivier.matz@6wind.com" <olivier.matz@6wind.com>,
"thomas@monjalon.net" <thomas@monjalon.net>,
"yangyi01@inspur.com" <yangyi01@inspur.com>
Subject: Re: [dpdk-dev] [PATCH] gso: fix free issue of mbuf gso segments attach to
Date: Wed, 14 Oct 2020 12:05:44 +0000 [thread overview]
Message-ID: <SN6PR11MB3310BA07C362EEAA2235DD899A050@SN6PR11MB3310.namprd11.prod.outlook.com> (raw)
In-Reply-To: <9239b2e.2116.17525095531.Coremail.yang_y_yi@163.com>
> From: yang_y_yi <yang_y_yi@163.com>
> Sent: Wednesday, October 14, 2020 3:56 AM
> To: Hu, Jiayu <jiayu.hu@intel.com>
> Cc: Ananyev, Konstantin <konstantin.ananyev@intel.com>; dev@dpdk.org; olivier.matz@6wind.com; thomas@monjalon.net;
> yangyi01@inspur.com
> Subject: Re:RE: [PATCH] gso: fix free issue of mbuf gso segments attach to
>
> I think it isn't a good idea to free it in rte_gso_segment, maybe application will continue to use this pkt for other purpose, rte_gso_segment
> can't make decision for application without any notice, it is better to return this decision right backt to application.
>
I think, if user wants to keep original packet, he can always call rte_pktmbuf_refcnt_update(pkt, 1)
just before calling gso function.
Also, as I remember in some cases it is not safe to do free() for input packet
(as pkt_out[] can contain input pkt itself). Would it also be user responsibility to determine
such situations?
P.S. Please don't reply on the top.
Konstantin
>
>
>
>
> At 2020-10-14 09:00:12, "Hu, Jiayu" <mailto:jiayu.hu@intel.com> wrote:
> >
> >
> >> -----Original Message-----
> >> From: Ananyev, Konstantin <mailto:konstantin.ananyev@intel.com>
> >> Sent: Tuesday, October 13, 2020 11:39 PM
> >> To: Hu, Jiayu <mailto:jiayu.hu@intel.com>; mailto:yang_y_yi@163.com; mailto:dev@dpdk.org
> >> Cc: mailto:mark.b.kavanagh@intel.com; mailto:olivier.matz@6wind.com;
> >> mailto:thomas@monjalon.net; mailto:yangyi01@inspur.com
> >> Subject: RE: [PATCH] gso: fix free issue of mbuf gso segments attach to
> >>
> >>
> >> > > rte_gso_segment decreased refcnt of pkt by one, but
> >> > > it is wrong if pkt is external mbuf, pkt won't be
> >> > > freed because of incorrect refcnt, the result is
> >> > > application can't allocate mbuf from mempool because
> >> > > mbufs in mempool are run out of.
> >> > >
> >> > > One correct way is application should call
> >> > > rte_pktmbuf_free after calling rte_gso_segment to free
> >> > > pkt explicitly. rte_gso_segment mustn't handle it, this
> >> > > should be responsibility of application.
> >> >
> >> > GSO doesn't support the input pktmbuf has external buffer.
> >> > Indeed, requiring users to free the input pktmbuf can avoid
> >> > memory leak, but I'm afraid that it also changes the semantic
> >> > of rte_gso_segment() which is defined in rte_gso.h.
> >> >
> >> > @Konstantin, any suggestions?
> >>
> >> Probably, a stupid question, but why can't we call mbuf_free()
> >> here instead fo decrementing refcnt manually:
> >> if (ret > 1)
> >> rte_pktmbuf_free(pkt);
> >> else if ...
> >> ?
> >
> >You are right. Freeing mbuf inside GSO is a better way to solve
> >the problem.
> >
> >Thanks,
> >Jiayu
> >>
> >>
> >>
> >> >
> >> > Thanks,
> >> > Jiayu
> >> > >
> >> > > Fixes: 119583797b6a ("gso: support TCP/IPv4 GSO")
> >> > > Signed-off-by: Yi Yang <mailto:yangyi01@inspur.com>
> >> > > ---
> >> > > doc/guides/prog_guide/generic_segmentation_offload_lib.rst | 7 +++++-
> >> -
> >> > > lib/librte_gso/rte_gso.c | 9 +--------
> >> > > 2 files changed, 6 insertions(+), 10 deletions(-)
> >> > >
> >> > > diff --git a/doc/guides/prog_guide/generic_segmentation_offload_lib.rst
> >> > > b/doc/guides/prog_guide/generic_segmentation_offload_lib.rst
> >> > > index 205cb8a..8577572 100644
> >> > > --- a/doc/guides/prog_guide/generic_segmentation_offload_lib.rst
> >> > > +++ b/doc/guides/prog_guide/generic_segmentation_offload_lib.rst
> >> > > @@ -25,8 +25,9 @@ Bearing that in mind, the GSO library enables DPDK
> >> > > applications to segment
> >> > > packets in software. Note however, that GSO is implemented as a
> >> > > standalone
> >> > > library, and not via a 'fallback' mechanism (i.e. for when TSO is
> >> unsupported
> >> > > in the underlying hardware); that is, applications must explicitly invoke
> >> the
> >> > > -GSO library to segment packets. The size of GSO segments ``(segsz)`` is
> >> > > -configurable by the application.
> >> > > +GSO library to segment packets, they also must call
> >> ``rte_pktmbuf_free()`` to
> >> > > +free mbuf GSO segments attach to after calling ``rte_gso_segment()``.
> >> The
> >> > > size
> >> > > +of GSO segments ``(segsz)`` is configurable by the application.
> >> > >
> >> > > Limitations
> >> > > -----------
> >> > > @@ -233,6 +234,8 @@ To segment an outgoing packet, an application
> >> must:
> >> > >
> >> > > #. Invoke the GSO segmentation API, ``rte_gso_segment()``.
> >> > >
> >> > > +#. Call ``rte_pktmbuf_free()`` to free mbuf ``rte_gso_segment()``
> >> segments.
> >> > > +
> >> > > #. If required, update the L3 and L4 checksums of the newly-created
> >> > > segments.
> >> > > For tunneled packets, the outer IPv4 headers' checksums should also
> >> be
> >> > > updated. Alternatively, the application may offload checksum
> >> calculation
> >> > > diff --git a/lib/librte_gso/rte_gso.c b/lib/librte_gso/rte_gso.c
> >> > > index 751b5b6..0d6cae5 100644
> >> > > --- a/lib/librte_gso/rte_gso.c
> >> > > +++ b/lib/librte_gso/rte_gso.c
> >> > > @@ -30,7 +30,6 @@
> >> > > uint16_t nb_pkts_out)
> >> > > {
> >> > > struct rte_mempool *direct_pool, *indirect_pool;
> >> > > -struct rte_mbuf *pkt_seg;
> >> > > uint64_t ol_flags;
> >> > > uint16_t gso_size;
> >> > > uint8_t ipid_delta;
> >> > > @@ -80,13 +79,7 @@
> >> > > return 1;
> >> > > }
> >> > >
> >> > > -if (ret > 1) {
> >> > > -pkt_seg = pkt;
> >> > > -while (pkt_seg) {
> >> > > -rte_mbuf_refcnt_update(pkt_seg, -1);
> >> > > -pkt_seg = pkt_seg->next;
> >> > > -}
> >> > > -} else if (ret < 0) {
> >> > > +if (ret < 0) {
> >> > > /* Revert the ol_flags in the event of failure. */
> >> > > pkt->ol_flags = ol_flags;
> >> > > }
> >> > > --
> >> > > 1.8.3.1
> >> >
> >>
>
>
next prev parent reply other threads:[~2020-10-14 12:05 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-10-10 3:10 yang_y_yi
2020-10-13 7:28 ` Hu, Jiayu
2020-10-13 15:39 ` Ananyev, Konstantin
2020-10-14 1:00 ` Hu, Jiayu
2020-10-14 2:56 ` yang_y_yi
2020-10-14 12:05 ` Ananyev, Konstantin [this message]
2020-10-15 5:14 ` Hu, Jiayu
2020-10-15 16:16 ` Ananyev, Konstantin
2020-10-16 0:53 ` Hu, Jiayu
2020-10-16 8:31 ` Ananyev, Konstantin
2020-10-19 3:17 ` Hu, Jiayu
2020-10-19 6:44 ` yang_y_yi
2020-10-19 8:47 ` Ananyev, Konstantin
2020-10-20 1:16 ` yang_y_yi
2020-10-19 2:29 ` yang_y_yi
2020-10-19 2:20 ` yang_y_yi
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=SN6PR11MB3310BA07C362EEAA2235DD899A050@SN6PR11MB3310.namprd11.prod.outlook.com \
--to=konstantin.ananyev@intel.com \
--cc=dev@dpdk.org \
--cc=jiayu.hu@intel.com \
--cc=olivier.matz@6wind.com \
--cc=thomas@monjalon.net \
--cc=yang_y_yi@163.com \
--cc=yangyi01@inspur.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).