linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Alex Williamson <alex.williamson@redhat.com>
To: Cornelia Huck <cohuck@redhat.com>
Cc: kvm@vger.kernel.org, linux-pci@vger.kernel.org,
	linux-kernel@vger.kernel.org, dev@dpdk.org, mtosatti@redhat.com,
	thomas@monjalon.net, bluca@debian.org, jerinjacobk@gmail.com,
	bruce.richardson@intel.com
Subject: Re: [RFC PATCH 1/7] vfio: Include optional device match in vfio_device_ops callbacks
Date: Thu, 6 Feb 2020 11:18:42 -0700	[thread overview]
Message-ID: <20200206111842.705bf58a@w520.home> (raw)
In-Reply-To: <20200206121419.69997326.cohuck@redhat.com>

On Thu, 6 Feb 2020 12:14:19 +0100
Cornelia Huck <cohuck@redhat.com> wrote:

> On Tue, 04 Feb 2020 16:05:43 -0700
> Alex Williamson <alex.williamson@redhat.com> wrote:
> 
> > Allow bus drivers to provide their own callback to match a device to
> > the user provided string.
> > 
> > Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
> > ---
> >  drivers/vfio/vfio.c  |   19 +++++++++++++++----
> >  include/linux/vfio.h |    3 +++
> >  2 files changed, 18 insertions(+), 4 deletions(-)
> > 
> > diff --git a/drivers/vfio/vfio.c b/drivers/vfio/vfio.c
> > index 388597930b64..dda1726adda8 100644
> > --- a/drivers/vfio/vfio.c
> > +++ b/drivers/vfio/vfio.c
> > @@ -875,11 +875,22 @@ EXPORT_SYMBOL_GPL(vfio_device_get_from_dev);
> >  static struct vfio_device *vfio_device_get_from_name(struct vfio_group *group,
> >  						     char *buf)
> >  {
> > -	struct vfio_device *it, *device = NULL;
> > +	struct vfio_device *it, *device = ERR_PTR(-ENODEV);
> >  
> >  	mutex_lock(&group->device_lock);
> >  	list_for_each_entry(it, &group->device_list, group_next) {
> > -		if (!strcmp(dev_name(it->dev), buf)) {
> > +		int ret;
> > +
> > +		if (it->ops->match) {
> > +			ret = it->ops->match(it->device_data, buf);
> > +			if (ret < 0 && ret != -ENODEV) {
> > +				device = ERR_PTR(ret);
> > +				break;
> > +			}
> > +		} else
> > +			ret = strcmp(dev_name(it->dev), buf);  
> 
> The asymmetric braces look a bit odd.

Ok

> > +
> > +		if (!ret) {
> >  			device = it;
> >  			vfio_device_get(device);
> >  			break;
> > @@ -1441,8 +1452,8 @@ static int vfio_group_get_device_fd(struct vfio_group *group, char *buf)
> >  		return -EPERM;
> >  
> >  	device = vfio_device_get_from_name(group, buf);
> > -	if (!device)
> > -		return -ENODEV;
> > +	if (IS_ERR(device))
> > +		return PTR_ERR(device);
> >  
> >  	ret = device->ops->open(device->device_data);
> >  	if (ret) {
> > diff --git a/include/linux/vfio.h b/include/linux/vfio.h
> > index e42a711a2800..755e0f0e2900 100644
> > --- a/include/linux/vfio.h
> > +++ b/include/linux/vfio.h
> > @@ -26,6 +26,8 @@
> >   *         operations documented below
> >   * @mmap: Perform mmap(2) on a region of the device file descriptor
> >   * @request: Request for the bus driver to release the device
> > + * @match: Optional device name match callback (return: 0 for match, -ENODEV
> > + *         (or >0) for no match and continue, other -errno: no match and stop)  
> 
> I'm wondering about these conventions.
> 
> If you basically want a tri-state return (match, don't match/continue,
> don't match/stop), putting -ENODEV and >0 together feels odd. I would
> rather expect either
> - < 0 == don't match/stop, 0 == don't match/continue, > 0 == match, or

So sort of a bool + errno.  I shied away from this because returning
zero for success, or match, is such a common semantic, especially when
we're replacing a simple strcmp().  I suppose it's logically just
!strcmp() though, which avoids the abort case for a simple
implementation like patch 2/7.

> - 0 == match, -ENODEV (or some other defined error) == don't
>   match/continue, all other values == don't match/abort?

This is closest to what I arrived at in this version, but I found it
necessary to exclude positive values from the no-match/abort and
consider them as no-match/continue because I didn't want to assume the
errno for that case.

I think with your first option we arrive at something like this:

diff --git a/drivers/vfio/vfio.c b/drivers/vfio/vfio.c
index dda1726adda8..b5609a411139 100644
--- a/drivers/vfio/vfio.c
+++ b/drivers/vfio/vfio.c
@@ -883,14 +883,15 @@ static struct vfio_device *vfio_device_get_from_name(struct vfio_group *group,
 
 		if (it->ops->match) {
 			ret = it->ops->match(it->device_data, buf);
-			if (ret < 0 && ret != -ENODEV) {
+			if (ret < 0) {
 				device = ERR_PTR(ret);
 				break;
 			}
-		} else
-			ret = strcmp(dev_name(it->dev), buf);
+		} else {
+			ret = !strcmp(dev_name(it->dev), buf);
+		}
 
-		if (!ret) {
+		if (ret) {
 			device = it;
 			vfio_device_get(device);
 			break;
diff --git a/include/linux/vfio.h b/include/linux/vfio.h
index 755e0f0e2900..029694b977f2 100644
--- a/include/linux/vfio.h
+++ b/include/linux/vfio.h
@@ -26,8 +26,9 @@
  *         operations documented below
  * @mmap: Perform mmap(2) on a region of the device file descriptor
  * @request: Request for the bus driver to release the device
- * @match: Optional device name match callback (return: 0 for match, -ENODEV
- *         (or >0) for no match and continue, other -errno: no match and stop)
+ * @match: Optional device name match callback (return: 0 for no-match, >0 for
+ *         match, -errno for abort (ex. match with insufficient or incorrect
+ *         additional args)
  */
 struct vfio_device_ops {
 	char	*name;


I like that.  Thanks,

Alex


  reply	other threads:[~2020-02-06 18:18 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-04 23:05 [RFC PATCH 0/7] vfio/pci: SR-IOV support Alex Williamson
2020-02-04 23:05 ` [RFC PATCH 1/7] vfio: Include optional device match in vfio_device_ops callbacks Alex Williamson
2020-02-06 11:14   ` Cornelia Huck
2020-02-06 18:18     ` Alex Williamson [this message]
2020-02-07  9:33       ` Cornelia Huck
2020-02-04 23:05 ` [RFC PATCH 2/7] vfio/pci: Implement match ops Alex Williamson
2020-02-04 23:06 ` [RFC PATCH 3/7] vfio/pci: Introduce VF token Alex Williamson
2020-02-05  7:57   ` Liu, Yi L
2020-02-05 14:13     ` Alex Williamson
2020-02-04 23:06 ` [RFC PATCH 4/7] vfio: Introduce VFIO_DEVICE_FEATURE ioctl and first user Alex Williamson
2020-02-04 23:06 ` [RFC PATCH 5/7] vfio/pci: Add sriov_configure support Alex Williamson
2020-02-04 23:06 ` [RFC PATCH 6/7] vfio/pci: Remove dev_fmt definition Alex Williamson
2020-02-06 13:45   ` Cornelia Huck
2020-02-04 23:06 ` [RFC PATCH 7/7] vfio/pci: Cleanup .probe() exit paths Alex Williamson
2020-02-04 23:17 ` [RFC PATCH 0/7] vfio/pci: SR-IOV support Alex Williamson
2020-02-05  7:57   ` Liu, Yi L
2020-02-05 14:18     ` Alex Williamson
2020-02-05  7:01 ` Christoph Hellwig
2020-02-05 13:58   ` Alex Williamson
2020-02-05  7:57 ` Liu, Yi L
2020-02-05 14:10   ` Alex Williamson
2020-02-11 11:18 ` Jerin Jacob
2020-02-11 13:57   ` [dpdk-dev] " Thomas Monjalon
2020-02-11 17:06   ` Alex Williamson
2020-02-11 18:03     ` Jerin Jacob

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=20200206111842.705bf58a@w520.home \
    --to=alex.williamson@redhat.com \
    --cc=bluca@debian.org \
    --cc=bruce.richardson@intel.com \
    --cc=cohuck@redhat.com \
    --cc=dev@dpdk.org \
    --cc=jerinjacobk@gmail.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=mtosatti@redhat.com \
    --cc=thomas@monjalon.net \
    /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).