linux-rdma.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Leon Romanovsky <leon@kernel.org>
To: Jinpu Wang <jinpu.wang@ionos.com>
Cc: Md Haris Iqbal <haris.iqbal@ionos.com>,
	linux-rdma@vger.kernel.org, bvanassche@acm.org, jgg@ziepe.ca,
	Grzegorz Prajsner <grzegorz.prajsner@ionos.com>
Subject: Re: [PATCH for-next 02/13] RDMA/rtrs-srv: Fix use-after-free during session establishment
Date: Mon, 12 Aug 2024 14:00:59 +0300	[thread overview]
Message-ID: <20240812110059.GE12060@unreal> (raw)
In-Reply-To: <CAMGffEmvWLy4-ugYVPBLYAAe43xZ73=bp7sH5ackf6M7w2zdZg@mail.gmail.com>

On Mon, Aug 12, 2024 at 12:52:25PM +0200, Jinpu Wang wrote:
> Hi,
> 
> On Sun, Aug 11, 2024 at 10:41 AM Leon Romanovsky <leon@kernel.org> wrote:
> >
> > On Fri, Aug 09, 2024 at 03:15:27PM +0200, Md Haris Iqbal wrote:
> > > From: Jack Wang <jinpu.wang@ionos.com>
> > >
> > > In case of error happening during session stablishment, close_work is
> > > running. A new RDMA CM event may arrive since we don't destroy cm_id
> > > before destroying qp. To fix this, we first destroy cm_id after drain_qp,
> > > so no new RDMA CM event will arrive afterwards.
> > >
> > > Fixes: 9cb837480424 ("RDMA/rtrs: server: main functionality")
> > > Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
> > > Signed-off-by: Grzegorz Prajsner <grzegorz.prajsner@ionos.com>
> > > Signed-off-by: Md Haris Iqbal <haris.iqbal@ionos.com>
> > > ---
> > >  drivers/infiniband/ulp/rtrs/rtrs-srv.c | 2 +-
> > >  drivers/infiniband/ulp/rtrs/rtrs.c     | 2 +-
> > >  2 files changed, 2 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/drivers/infiniband/ulp/rtrs/rtrs-srv.c b/drivers/infiniband/ulp/rtrs/rtrs-srv.c
> > > index fb67b58a7f62..90ea25ad6720 100644
> > > --- a/drivers/infiniband/ulp/rtrs/rtrs-srv.c
> > > +++ b/drivers/infiniband/ulp/rtrs/rtrs-srv.c
> > > @@ -1540,6 +1540,7 @@ static void rtrs_srv_close_work(struct work_struct *work)
> > >               con = to_srv_con(srv_path->s.con[i]);
> > >               rdma_disconnect(con->c.cm_id);
> > >               ib_drain_qp(con->c.qp);
> > > +             rdma_destroy_id(con->c.cm_id);
> > >       }
> > >
> > >       /*
> > > @@ -1564,7 +1565,6 @@ static void rtrs_srv_close_work(struct work_struct *work)
> > >                       continue;
> > >               con = to_srv_con(srv_path->s.con[i]);
> > >               rtrs_cq_qp_destroy(&con->c);
> > > -             rdma_destroy_id(con->c.cm_id);
> > >               kfree(con);
> > >       }
> > >       rtrs_ib_dev_put(srv_path->s.dev);
> > > diff --git a/drivers/infiniband/ulp/rtrs/rtrs.c b/drivers/infiniband/ulp/rtrs/rtrs.c
> > > index 4e17d546d4cc..44167fd1c958 100644
> > > --- a/drivers/infiniband/ulp/rtrs/rtrs.c
> > > +++ b/drivers/infiniband/ulp/rtrs/rtrs.c
> > > @@ -318,7 +318,7 @@ EXPORT_SYMBOL_GPL(rtrs_cq_qp_create);
> > >  void rtrs_cq_qp_destroy(struct rtrs_con *con)
> > >  {
> > >       if (con->qp) {
> > > -             rdma_destroy_qp(con->cm_id);
> > > +             ib_destroy_qp(con->qp);
> >
> > You created that QP with rdma_create_qp() and you should destroy it with rdma_destroy_qp().
> We can't do it, as we move rdma_destroy_id before rtrs_cq_qp_destroy,
> if we still call rdma_destroy_qp, which will lead to use after free as
> cm_id could already be free-ed.

It is a hint that you are doing something wrong.

Thanks

> 
> >
> > Thanks
> Thx!
> >
> > >               con->qp = NULL;
> > >       }
> > >       destroy_cq(con);
> > > --
> > > 2.25.1
> > >
> 

  reply	other threads:[~2024-08-12 11:01 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-09 13:15 [PATCH for-next 00/13] Misc patches for RTRS Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 01/13] RDMA/rtrs-srv: Make rtrs_srv_open fail if its a second call Md Haris Iqbal
2024-08-11  8:38   ` Leon Romanovsky
2024-08-12 10:16     ` Haris Iqbal
2024-08-12 10:34       ` Leon Romanovsky
2024-08-12 10:39         ` Haris Iqbal
2024-08-12 10:59           ` Leon Romanovsky
2024-08-12 11:17             ` Haris Iqbal
2024-08-12 12:15               ` Leon Romanovsky
2024-08-13  9:42                 ` Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 02/13] RDMA/rtrs-srv: Fix use-after-free during session establishment Md Haris Iqbal
2024-08-11  8:41   ` Leon Romanovsky
2024-08-12 10:52     ` Jinpu Wang
2024-08-12 11:00       ` Leon Romanovsky [this message]
2024-08-13 10:54         ` Jinpu Wang
2024-08-09 13:15 ` [PATCH for-next 03/13] RDMA/rtrs: For HB error add additional clt/srv specific logging Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 04/13] RDMA/rtrs-clt: Fix need_inv setting in error case Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 05/13] RDMA/rtrs-clt: Rate limit errors in IO path Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 06/13] RDMA/rtrs: Reset hb_missed_cnt after receiving other traffic from peer Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 07/13] RDMA/rtrs-clt: Reuse need_inval from mr Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 08/13] RDMA/rtrs-clt: Reset cid to con_num - 1 to stay in bounds Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 09/13] RDMA/rtrs-clt: Print request type for errors Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 10/13] RDMA/rtrs-srv: Avoid null pointer deref during path establishment Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 11/13] RDMA/rtrs: register ib event handler Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 12/13] RDMA/rtrs-clt: Do local invalidate after write io completion Md Haris Iqbal
2024-08-09 13:15 ` [PATCH for-next 13/13] RDMA/rtrs-clt: Remove an extra space Md Haris Iqbal
2024-08-11  8:43   ` Leon Romanovsky
2024-08-12  4:10     ` Jinpu Wang
2024-08-12  9:48       ` Zhu Yanjun

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=20240812110059.GE12060@unreal \
    --to=leon@kernel.org \
    --cc=bvanassche@acm.org \
    --cc=grzegorz.prajsner@ionos.com \
    --cc=haris.iqbal@ionos.com \
    --cc=jgg@ziepe.ca \
    --cc=jinpu.wang@ionos.com \
    --cc=linux-rdma@vger.kernel.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).