From: Leon Romanovsky <leon@kernel.org>
To: Jason Gunthorpe <jgg@nvidia.com>
Cc: Doug Ledford <dledford@redhat.com>, linux-rdma@vger.kernel.org
Subject: Re: [PATCH rdma-next v2 01/14] RDMA/cma: Delete from restrack DB after successful destroy
Date: Thu, 17 Sep 2020 19:19:50 +0300 [thread overview]
Message-ID: <20200917161950.GF869610@unreal> (raw)
In-Reply-To: <20200917120636.GA103244@nvidia.com>
On Thu, Sep 17, 2020 at 09:06:36AM -0300, Jason Gunthorpe wrote:
> On Mon, Sep 07, 2020 at 03:21:43PM +0300, Leon Romanovsky wrote:
> > From: Leon Romanovsky <leonro@mellanox.com>
> >
> > Update the code to have similar destroy pattern like other IB objects.
> >
> > This change create asymmetry to the rdma_id_private create flow to make
> > sure that memory is managed by restrack.
> >
> > Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
> > drivers/infiniband/core/cma.c | 3 +--
> > 1 file changed, 1 insertion(+), 2 deletions(-)
> >
> > diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
> > index f9ff8b7f05e7..24e09416de4f 100644
> > +++ b/drivers/infiniband/core/cma.c
> > @@ -1821,7 +1821,6 @@ static void _destroy_id(struct rdma_id_private *id_priv,
> > {
> > cma_cancel_operation(id_priv, state);
> >
> > - rdma_restrack_del(&id_priv->res);
> > if (id_priv->cma_dev) {
> > if (rdma_cap_ib_cm(id_priv->id.device, 1)) {
> > if (id_priv->cm_id.ib)
> > @@ -1847,6 +1846,7 @@ static void _destroy_id(struct rdma_id_private *id_priv,
> > rdma_put_gid_attr(id_priv->id.route.addr.dev_addr.sgid_attr);
> >
> > put_net(id_priv->id.route.addr.dev_addr.net);
> > + rdma_restrack_del(&id_priv->res);
> > kfree(id_priv);
> > }
>
> This is wrong, rdma_restrack_del() has to be called before
> ib_destroy_cm_id() because restrack reaches into the cm_id and touches
> that memory:
>
> case RDMA_RESTRACK_CM_ID:
> return container_of(res, struct rdma_id_private,
> res)->id.device;
>
> Which will now be freed after this change.
We access "id" which is struct rdma_cm_id, while ib_destroy_cm_id()
releases something different struct ib_cm_id cm_id.ib.
"id" is embedded into id_priv released later in _destroy_id() function.
Thanks
>
> Jason
next prev parent reply other threads:[~2020-09-17 18:55 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-07 12:21 [PATCH rdma-next v2 00/14] Track memory allocation with restrack DB help Leon Romanovsky
2020-09-07 12:21 ` [PATCH rdma-next v2 01/14] RDMA/cma: Delete from restrack DB after successful destroy Leon Romanovsky
2020-09-17 12:06 ` Jason Gunthorpe
2020-09-17 16:19 ` Leon Romanovsky [this message]
2020-09-17 16:32 ` Jason Gunthorpe
2020-09-07 12:21 ` [PATCH rdma-next v2 02/14] RDMA/mlx5: Don't call to restrack recursively Leon Romanovsky
2020-09-07 12:21 ` [PATCH rdma-next v2 03/14] RDMA/mlx4: Provide port number for special QPs Leon Romanovsky
2020-09-07 12:21 ` [PATCH rdma-next v2 04/14] RDMA/restrack: Count references to the verbs objects Leon Romanovsky
2020-09-18 17:00 ` Jason Gunthorpe
2020-09-19 8:30 ` Leon Romanovsky
2020-09-07 12:21 ` [PATCH rdma-next v2 05/14] RDMA/restrack: Simplify restrack tracking in kernel flows Leon Romanovsky
2020-09-07 12:21 ` [PATCH rdma-next v2 06/14] RDMA/restrack: Improve readability in task name management Leon Romanovsky
2020-09-18 23:17 ` Jason Gunthorpe
2020-09-19 8:42 ` Leon Romanovsky
2020-09-07 12:21 ` [PATCH rdma-next v2 07/14] RDMA/cma: Be strict with attaching to CMA device Leon Romanovsky
2020-09-18 23:26 ` Jason Gunthorpe
2020-09-19 9:03 ` Leon Romanovsky
2020-09-24 19:46 ` Jason Gunthorpe
2020-09-07 12:21 ` [PATCH rdma-next v2 08/14] RDMA/core: Allow drivers to disable restrack DB Leon Romanovsky
2020-09-07 12:21 ` [PATCH rdma-next v2 09/14] RDMA/counter: Combine allocation and bind logic Leon Romanovsky
2020-09-07 12:21 ` [PATCH rdma-next v2 10/14] RDMA/restrack: Store all special QPs in restrack DB Leon Romanovsky
2020-09-18 23:30 ` Jason Gunthorpe
2020-09-19 8:27 ` Leon Romanovsky
2020-09-07 12:21 ` [PATCH rdma-next v2 11/14] RDMA/restrack: Make restrack DB mandatory for IB objects Leon Romanovsky
2020-09-18 23:31 ` Jason Gunthorpe
2020-09-19 9:09 ` Leon Romanovsky
2020-09-24 19:50 ` Jason Gunthorpe
2020-09-24 19:59 ` Jason Gunthorpe
2020-09-07 12:21 ` [PATCH rdma-next v2 12/14] RDMA/restrack: Support all QP types Leon Romanovsky
2020-09-07 12:21 ` [PATCH rdma-next v2 13/14] RDMA/core: Track device memory MRs Leon Romanovsky
2020-09-24 20:02 ` Jason Gunthorpe
2020-09-07 12:21 ` [PATCH rdma-next v2 14/14] RDMA/restrack: Drop valid restrack field as source of ambiguity Leon Romanovsky
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=20200917161950.GF869610@unreal \
--to=leon@kernel.org \
--cc=dledford@redhat.com \
--cc=jgg@nvidia.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).