Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <CAPM=9txSKv_xwZJ6SndtqsdQm6aK1KJVF91dB5Odhc_Xv6Qdrw@mail.gmail.com>
 <CAFCwf10CsLgt+_qT7dT=8DVXsL0a=w=uXN6HC=CpP5EfitvLfQ@mail.gmail.com>
 <YuvctaLwRi+z0Gw4@nvidia.com> <CAFCwf12wD3uEhr+kxwN9ROXApHzGh_n1je5susZV5NgGR9fCcQ@mail.gmail.com>
 <Yuxi1eRHPN36Or+1@nvidia.com> <CAFCwf13QF_JdzNcpw==zzBoEQUYChMXfechotH31qmAfYZUGmg@mail.gmail.com>
 <CAFCwf107tLxHKxkPqSRsOHVVp5s2tDEFOOy2oDZUz_KGmv-rDg@mail.gmail.com>
 <YvCozvgodIY19LSr@kroah.com> <YvFOEPdC9r8QBH11@nvidia.com>
 <YvH9X5puer4jpzMX@kroah.com> <YvIU/wMdqFA1fYc6@infradead.org>
In-Reply-To: <YvIU/wMdqFA1fYc6@infradead.org>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Tue, 9 Aug 2022 10:32:27 +0200
Message-ID: <CAK8P3a3y2=FyzK4S6MRfZdrz=DdLat1ajdT_uPmN533mNYmF9w@mail.gmail.com>
Subject: Re: New subsystem for acceleration devices
To:     Christoph Hellwig <hch@infradead.org>
Cc:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Jason Gunthorpe <jgg@nvidia.com>,
        Oded Gabbay <oded.gabbay@gmail.com>,
        Dave Airlie <airlied@gmail.com>,
        dri-devel <dri-devel@lists.freedesktop.org>,
        Yuji Ishikawa <yuji2.ishikawa@toshiba.co.jp>,
        Jiho Chu <jiho.chu@samsung.com>, Arnd Bergmann <arnd@arndb.de>,
        "Linux-Kernel@Vger. Kernel. Org" <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Tue, Aug 9, 2022 at 10:04 AM Christoph Hellwig <hch@infradead.org> wrote:
> On Tue, Aug 09, 2022 at 08:23:27AM +0200, Greg Kroah-Hartman wrote:
> > > This is different from the number of FDs pointing at the struct file.
> > > Userpsace can open a HW state and point a lot of FDs at it, that is
> > > userspace's problem. From a kernel view they all share one struct file
> > > and thus one HW state.
> >
> > Yes, that's fine, if that is what is happening here, I have no
> > objection.
>
> It would be great if we could actually life that into a common
> layer (chardev or vfs) given just how common this, and drivers tend
> to get it wrong, do it suboptimal so often.

Totally agreed.

I think for devices with hardware MMU contexts you actually want to
bind the context to a 'mm_struct', and then ensure that the context
is only ever used from a process that shares this mm_struct,
regardless of who else has access to the same file or inode.

We did this in a somewhat hacky way in spufs a long time ago, and
I would expect this to be solved more cleanly in drivers/gpu/drm, but
I don't know where to look for that code.

        Arnd