linux-kernel - Re: [PATCH rdma-next] RDMA/core: Rely on vendors to set right IOVA

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <YXp0XSpXsarxrGqS@unreal>
Date:   Thu, 28 Oct 2021 12:58:53 +0300
From:   Leon Romanovsky <leon@...nel.org>
To:     Christoph Hellwig <hch@...radead.org>
Cc:     Doug Ledford <dledford@...hat.com>,
        Jason Gunthorpe <jgg@...dia.com>,
        Aharon Landau <aharonl@...dia.com>,
        linux-kernel@...r.kernel.org, linux-rdma@...r.kernel.org
Subject: Re: [PATCH rdma-next] RDMA/core: Rely on vendors to set right IOVA

On Thu, Oct 28, 2021 at 12:56:46AM -0700, Christoph Hellwig wrote:
> On Thu, Oct 28, 2021 at 08:55:22AM +0300, Leon Romanovsky wrote:
> > From: Aharon Landau <aharonl@...dia.com>
> > 
> > The vendors set the IOVA of newly created MRs in rereg_user_mr, so don't
> > overwrite it. That ensures that this field is set only if IB_MR_REREG_TRANS
> > flag is provided.
> 
> How is a "vendor" involved with this?  This should all be upstream code.

"vendor" is wrong word here.

I wanted to say that all drivers which support ".rereg_user_mr()"
callback and return new_mr should set everything. In case of IB_MR_REREG_TRANS
flow, it is IOVA which is not cmd.hca_va, but mr->iova.

Thanks