virtualization.lists.linux-foundation.org archive mirror
 help / color / mirror / Atom feed
From: "Michael S. Tsirkin" <mst@redhat.com>
To: Max Gurtovoy <mgurtovoy@nvidia.com>
Cc: stefanha@redhat.com, sgarzare@redhat.com,
	virtualization@lists.linux-foundation.org, kvm@vger.kernel.org,
	oren@nvidia.com, aevdaev@nvidia.com, aaptel@nvidia.com
Subject: Re: [PATCH v2 1/1] virtio: add driver_override support
Date: Tue, 9 Dec 2025 13:24:12 -0500	[thread overview]
Message-ID: <20251209132326-mutt-send-email-mst@kernel.org> (raw)
In-Reply-To: <df0d1e2e-a496-416a-8cfc-22d69ad4c37b@nvidia.com>

On Tue, Dec 09, 2025 at 08:20:45PM +0200, Max Gurtovoy wrote:
> 
> On 09/12/2025 19:48, Michael S. Tsirkin wrote:
> > On Tue, Dec 09, 2025 at 07:45:19PM +0200, Max Gurtovoy wrote:
> > > On 09/12/2025 18:34, Michael S. Tsirkin wrote:
> > > > On Tue, Dec 09, 2025 at 12:09:08AM +0200, Max Gurtovoy wrote:
> > > > > Add support for the 'driver_override' attribute to Virtio devices. This
> > > > > allows users to control which Virtio bus driver binds to a given Virtio
> > > > > device.
> > > > > 
> > > > > If 'driver_override' is not set, the existing behavior is preserved and
> > > > > devices will continue to auto-bind to the first matching Virtio bus
> > > > > driver.
> > > > oh, it's a device driver not the bus driver, actually.
> > > Yes. I'll fix the commit message.
> > > 
> > > 
> > > > > Tested with virtio blk device (virtio core and pci drivers are loaded):
> > > > > 
> > > > >     $ modprobe my_virtio_blk
> > > > > 
> > > > >     # automatically unbind from virtio_blk driver and override + bind to
> > > > >     # my_virtio_blk driver.
> > > > >     $ driverctl -v -b virtio set-override virtio0 my_virtio_blk
> > > > > 
> > > > > In addition, driverctl saves the configuration persistently under
> > > > > /etc/driverctl.d/.
> > > > what is this "mydriver" though? what are valid examples that
> > > > we want to support?
> > > This is an example for a custom virtio block driver.
> > > 
> > > It can be any custom virtio-XX driver (XX=FS/NET/..).
> > Is "custom" a way to say "out of tree"?
> 
> The driver_override mechanism fits both in-tree/out-of-tree drivers.

good to know.  which in tree drivers can use it?
this is the part that's missing in the commit log.

thanks


> > 
> > > > > Signed-off-by: Avraham Evdaev <aevdaev@nvidia.com>
> > > > > Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
> > > > > ---
> > > > > 
> > > > > changes from v1:
> > > > >    - use !strcmp() to compare strings (MST)
> > > > >    - extend commit msg with example (MST)
> > > > > 
> > > > > ---
> > > > >    drivers/virtio/virtio.c | 34 ++++++++++++++++++++++++++++++++++
> > > > >    include/linux/virtio.h  |  4 ++++
> > > > >    2 files changed, 38 insertions(+)
> > > > > 
> > > > > diff --git a/drivers/virtio/virtio.c b/drivers/virtio/virtio.c
> > > > > index a09eb4d62f82..993dc928be49 100644
> > > > > --- a/drivers/virtio/virtio.c
> > > > > +++ b/drivers/virtio/virtio.c
> > > > > @@ -61,12 +61,41 @@ static ssize_t features_show(struct device *_d,
> > > > >    }
> > > > >    static DEVICE_ATTR_RO(features);
> > > > > +static ssize_t driver_override_store(struct device *_d,
> > > > > +				     struct device_attribute *attr,
> > > > > +				     const char *buf, size_t count)
> > > > > +{
> > > > > +	struct virtio_device *dev = dev_to_virtio(_d);
> > > > > +	int ret;
> > > > > +
> > > > > +	ret = driver_set_override(_d, &dev->driver_override, buf, count);
> > > > > +	if (ret)
> > > > > +		return ret;
> > > > > +
> > > > > +	return count;
> > > > > +}
> > > > > +
> > > > > +static ssize_t driver_override_show(struct device *_d,
> > > > > +				    struct device_attribute *attr, char *buf)
> > > > > +{
> > > > > +	struct virtio_device *dev = dev_to_virtio(_d);
> > > > > +	ssize_t len;
> > > > > +
> > > > > +	device_lock(_d);
> > > > > +	len = sysfs_emit(buf, "%s\n", dev->driver_override);
> > > > > +	device_unlock(_d);
> > > > > +
> > > > > +	return len;
> > > > > +}
> > > > > +static DEVICE_ATTR_RW(driver_override);
> > > > > +
> > > > >    static struct attribute *virtio_dev_attrs[] = {
> > > > >    	&dev_attr_device.attr,
> > > > >    	&dev_attr_vendor.attr,
> > > > >    	&dev_attr_status.attr,
> > > > >    	&dev_attr_modalias.attr,
> > > > >    	&dev_attr_features.attr,
> > > > > +	&dev_attr_driver_override.attr,
> > > > >    	NULL,
> > > > >    };
> > > > >    ATTRIBUTE_GROUPS(virtio_dev);
> > > > > @@ -88,6 +117,10 @@ static int virtio_dev_match(struct device *_dv, const struct device_driver *_dr)
> > > > >    	struct virtio_device *dev = dev_to_virtio(_dv);
> > > > >    	const struct virtio_device_id *ids;
> > > > > +	/* Check override first, and if set, only use the named driver */
> > > > > +	if (dev->driver_override)
> > > > > +		return !strcmp(dev->driver_override, _dr->name);
> > > > > +
> > > > >    	ids = drv_to_virtio(_dr)->id_table;
> > > > >    	for (i = 0; ids[i].device; i++)
> > > > >    		if (virtio_id_match(dev, &ids[i]))
> > > > > @@ -582,6 +615,7 @@ void unregister_virtio_device(struct virtio_device *dev)
> > > > >    {
> > > > >    	int index = dev->index; /* save for after device release */
> > > > > +	kfree(dev->driver_override);
> > > > >    	device_unregister(&dev->dev);
> > > > >    	virtio_debug_device_exit(dev);
> > > > >    	ida_free(&virtio_index_ida, index);
> > > > > diff --git a/include/linux/virtio.h b/include/linux/virtio.h
> > > > > index db31fc6f4f1f..418bb490bdc6 100644
> > > > > --- a/include/linux/virtio.h
> > > > > +++ b/include/linux/virtio.h
> > > > > @@ -138,6 +138,9 @@ struct virtio_admin_cmd {
> > > > >     * @config_lock: protects configuration change reporting
> > > > >     * @vqs_list_lock: protects @vqs.
> > > > >     * @dev: underlying device.
> > > > > + * @driver_override: driver name to force a match; do not set directly,
> > > > > + *                   because core frees it; use driver_set_override() to
> > > > > + *                   set or clear it.
> > > > >     * @id: the device type identification (used to match it with a driver).
> > > > >     * @config: the configuration ops for this device.
> > > > >     * @vringh_config: configuration ops for host vrings.
> > > > > @@ -158,6 +161,7 @@ struct virtio_device {
> > > > >    	spinlock_t config_lock;
> > > > >    	spinlock_t vqs_list_lock;
> > > > >    	struct device dev;
> > > > > +	const char *driver_override;
> > > > >    	struct virtio_device_id id;
> > > > >    	const struct virtio_config_ops *config;
> > > > >    	const struct vringh_config_ops *vringh_config;
> > > > > -- 
> > > > > 2.18.1


      reply	other threads:[~2025-12-09 18:24 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-12-08 22:09 [PATCH v2 1/1] virtio: add driver_override support Max Gurtovoy
2025-12-09 15:48 ` Stefan Hajnoczi
2025-12-09 16:34 ` Michael S. Tsirkin
2025-12-09 17:45   ` Max Gurtovoy
2025-12-09 17:48     ` Michael S. Tsirkin
2025-12-09 18:20       ` Max Gurtovoy
2025-12-09 18:24         ` Michael S. Tsirkin [this message]

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=20251209132326-mutt-send-email-mst@kernel.org \
    --to=mst@redhat.com \
    --cc=aaptel@nvidia.com \
    --cc=aevdaev@nvidia.com \
    --cc=kvm@vger.kernel.org \
    --cc=mgurtovoy@nvidia.com \
    --cc=oren@nvidia.com \
    --cc=sgarzare@redhat.com \
    --cc=stefanha@redhat.com \
    --cc=virtualization@lists.linux-foundation.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).