| Message ID | 20231005203030.223489-1-vgoyal@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | virtiofs: Export filesystem tags through sysfs | expand |
On Thu, 5 Oct 2023 at 22:30, Vivek Goyal <vgoyal@redhat.com> wrote: > > virtiofs filesystem is mounted using a "tag" which is exported by the > virtiofs device. virtiofs driver knows about all the available tags but > these are not exported to user space. > > People have asked these tags to be exported to user space. Most recently > Lennart Poettering has asked for it as he wants to scan the tags and mount > virtiofs automatically in certain cases. > > https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 > > This patch exports tags through sysfs. One tag is associated with each > virtiofs device. A new "tag" file appears under virtiofs device dir. > Actual filesystem tag can be obtained by reading this "tag" file. > > For example, if a virtiofs device exports tag "myfs", a new file "tag" > will show up here. > > /sys/bus/virtio/devices/virtio<N>/tag > > # cat /sys/bus/virtio/devices/virtio<N>/tag > myfs > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com> Hi Vivek, This needs something under Documentation/. While the interface looks good to me, I think we need an ack on that from the virtio maintainer. Thanks, Miklos
On Mon, Oct 09, 2023 at 11:53:42AM +0200, Miklos Szeredi wrote: > On Thu, 5 Oct 2023 at 22:30, Vivek Goyal <vgoyal@redhat.com> wrote: > > > > virtiofs filesystem is mounted using a "tag" which is exported by the > > virtiofs device. virtiofs driver knows about all the available tags but > > these are not exported to user space. > > > > People have asked these tags to be exported to user space. Most recently > > Lennart Poettering has asked for it as he wants to scan the tags and mount > > virtiofs automatically in certain cases. > > > > https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 > > > > This patch exports tags through sysfs. One tag is associated with each > > virtiofs device. A new "tag" file appears under virtiofs device dir. > > Actual filesystem tag can be obtained by reading this "tag" file. > > > > For example, if a virtiofs device exports tag "myfs", a new file "tag" > > will show up here. > > > > /sys/bus/virtio/devices/virtio<N>/tag > > > > # cat /sys/bus/virtio/devices/virtio<N>/tag > > myfs > > > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com> > > Hi Vivek, > > This needs something under Documentation/. Hi Miklos, Hmm.., I can easily update the virtiofs documentation. Initially I was thinking to put something in Documentation/ABI/testing/ as well. But I don't see any virtio related. In fact can't find any files related to "virtio" at all. So I will just update the Documentation/filesystems/virtiofs.rst for now. > > While the interface looks good to me, I think we need an ack on that > from the virtio maintainer. I am assuming Stefan should be able to provide an ACK. But I will also add Michael and Paolo in the V2 of the patch and hoping we should be able to get atleast 1 ACK. Thanks Vivek
On Thu, Oct 05, 2023 at 04:30:30PM -0400, Vivek Goyal wrote: > virtiofs filesystem is mounted using a "tag" which is exported by the > virtiofs device. virtiofs driver knows about all the available tags but > these are not exported to user space. > > People have asked these tags to be exported to user space. Most recently > Lennart Poettering has asked for it as he wants to scan the tags and mount > virtiofs automatically in certain cases. > > https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 > > This patch exports tags through sysfs. One tag is associated with each > virtiofs device. A new "tag" file appears under virtiofs device dir. > Actual filesystem tag can be obtained by reading this "tag" file. > > For example, if a virtiofs device exports tag "myfs", a new file "tag" > will show up here. > > /sys/bus/virtio/devices/virtio<N>/tag > > # cat /sys/bus/virtio/devices/virtio<N>/tag > myfs If you respin this series, please mention that the tag is available at KOBJ_BIND time, but not KOBJ_ADD. Just a sentence or two is enough to help someone trying to figure out how to use this new sysfs attr with udev. > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com> > --- > fs/fuse/virtio_fs.c | 34 ++++++++++++++++++++++++++++++++++ > 1 file changed, 34 insertions(+) Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
On Tue, Oct 10, 2023 at 8:38 PM Stefan Hajnoczi <stefanha@redhat.com> wrote: > > On Thu, Oct 05, 2023 at 04:30:30PM -0400, Vivek Goyal wrote: > > virtiofs filesystem is mounted using a "tag" which is exported by the > > virtiofs device. virtiofs driver knows about all the available tags but > > these are not exported to user space. > > > > People have asked these tags to be exported to user space. Most recently > > Lennart Poettering has asked for it as he wants to scan the tags and mount > > virtiofs automatically in certain cases. > > > > https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 > > > > This patch exports tags through sysfs. One tag is associated with each > > virtiofs device. A new "tag" file appears under virtiofs device dir. > > Actual filesystem tag can be obtained by reading this "tag" file. > > > > For example, if a virtiofs device exports tag "myfs", a new file "tag" > > will show up here. > > > > /sys/bus/virtio/devices/virtio<N>/tag > > > > # cat /sys/bus/virtio/devices/virtio<N>/tag > > myfs > > If you respin this series, please mention that the tag is available at > KOBJ_BIND time, but not KOBJ_ADD. Just a sentence or two is enough to > help someone trying to figure out how to use this new sysfs attr with > udev. Maybe it's also worth mention, that the tag file is created after a successful probe, so the tag should be a valid one: non-empty and unique > > > > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com> > > --- > > fs/fuse/virtio_fs.c | 34 ++++++++++++++++++++++++++++++++++ > > 1 file changed, 34 insertions(+) > > Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
On Thu, Oct 05, 2023 at 04:30:30PM -0400, Vivek Goyal wrote: > virtiofs filesystem is mounted using a "tag" which is exported by the > virtiofs device. virtiofs driver knows about all the available tags but > these are not exported to user space. > > People have asked these tags to be exported to user space. Most recently > Lennart Poettering has asked for it as he wants to scan the tags and mount > virtiofs automatically in certain cases. > > https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 Hi, I was one of those people. :) > This patch exports tags through sysfs. One tag is associated with each > virtiofs device. A new "tag" file appears under virtiofs device dir. > Actual filesystem tag can be obtained by reading this "tag" file. > > For example, if a virtiofs device exports tag "myfs", a new file "tag" > will show up here. > > /sys/bus/virtio/devices/virtio<N>/tag > > # cat /sys/bus/virtio/devices/virtio<N>/tag > myfs > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com> Are you still thinking about exposing this in the uevent as well? That would be much more convenient for me, because with this approach by the time the "remove" uevent arrives, it's no longer possible to check what tag was associated with the device — you have to store it somewhere when the device appears, so you can look it up again when the device is removed. (Not everybody uses udev.) Regardless, Tested-by: Alyssa Ross <hi@alyssa.is> … and a review comment below. > --- > fs/fuse/virtio_fs.c | 34 ++++++++++++++++++++++++++++++++++ > 1 file changed, 34 insertions(+) > > diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c > index 5f1be1da92ce..a5b11e18f331 100644 > --- a/fs/fuse/virtio_fs.c > +++ b/fs/fuse/virtio_fs.c > @@ -107,6 +107,21 @@ static const struct fs_parameter_spec virtio_fs_parameters[] = { > {} > }; > > +/* Forward Declarations */ > +static void virtio_fs_stop_all_queues(struct virtio_fs *fs); > + > +/* sysfs related */ > +static ssize_t tag_show(struct device *dev, struct device_attribute *attr, > + char *buf) > +{ > + struct virtio_device *vdev = container_of(dev, struct virtio_device, > + dev); > + struct virtio_fs *fs = vdev->priv; > + > + return sysfs_emit(buf, "%s", fs->tag); All of the other files in the device directory end with trailing newlines. Should this one be an exception? > +} > +static DEVICE_ATTR_RO(tag);
On Sat, Oct 21, 2023 at 04:10:21PM +0000, Alyssa Ross wrote: > On Thu, Oct 05, 2023 at 04:30:30PM -0400, Vivek Goyal wrote: > > virtiofs filesystem is mounted using a "tag" which is exported by the > > virtiofs device. virtiofs driver knows about all the available tags but > > these are not exported to user space. > > > > People have asked these tags to be exported to user space. Most recently > > Lennart Poettering has asked for it as he wants to scan the tags and mount > > virtiofs automatically in certain cases. > > > > https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 > > Hi, I was one of those people. :) Hi Alyssa, Aha... Thanks. I was not able to remember who had requested this. > > > This patch exports tags through sysfs. One tag is associated with each > > virtiofs device. A new "tag" file appears under virtiofs device dir. > > Actual filesystem tag can be obtained by reading this "tag" file. > > > > For example, if a virtiofs device exports tag "myfs", a new file "tag" > > will show up here. > > > > /sys/bus/virtio/devices/virtio<N>/tag > > > > # cat /sys/bus/virtio/devices/virtio<N>/tag > > myfs > > > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com> > > Are you still thinking about exposing this in the uevent as well? > That would be much more convenient for me, because with this approach > by the time the "remove" uevent arrives, it's no longer possible to > check what tag was associated with the device — you have to store it > somewhere when the device appears, so you can look it up again when the > device is removed. (Not everybody uses udev.) Looks like systemd + udev combination can already take care of it. I just had to specify "StopWhenUnneeded=true" in my systemd .mount unit file. And that made sure that when device goes away, virtiofs is unmounted and service is deactivated. Following is my mount unit file. $ cat /etc/systemd/system/mnt-virtiofs.mount [Unit] Description=Virtiofs mount myfs DefaultDependencies=no ConditionPathExists=/mnt/virtiofs ConditionCapability=CAP_SYS_ADMIN Before=sysinit.target StopWhenUnneeded=true [Mount] What=myfs Where=/mnt/virtiofs Type=virtiofs And following is the udev rule I used. $ cat /etc/udev/rules.d/80-local.rules SUBSYSTEM=="virtio", DRIVER=="virtiofs", ATTR{tag}=="myfs", TAG+="systemd", ENV{SYSTEMD_WANTS}+="mnt-virtiofs.mount" And a combination of above two seems to work. virtiofs is automatically mounted when device is hotplugged and it is unmounted when device is hot unplugged. > > Regardless, > > Tested-by: Alyssa Ross <hi@alyssa.is> > > … and a review comment below. > > > --- > > fs/fuse/virtio_fs.c | 34 ++++++++++++++++++++++++++++++++++ > > 1 file changed, 34 insertions(+) > > > > diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c > > index 5f1be1da92ce..a5b11e18f331 100644 > > --- a/fs/fuse/virtio_fs.c > > +++ b/fs/fuse/virtio_fs.c > > @@ -107,6 +107,21 @@ static const struct fs_parameter_spec virtio_fs_parameters[] = { > > {} > > }; > > > > +/* Forward Declarations */ > > +static void virtio_fs_stop_all_queues(struct virtio_fs *fs); > > + > > +/* sysfs related */ > > +static ssize_t tag_show(struct device *dev, struct device_attribute *attr, > > + char *buf) > > +{ > > + struct virtio_device *vdev = container_of(dev, struct virtio_device, > > + dev); > > + struct virtio_fs *fs = vdev->priv; > > + > > + return sysfs_emit(buf, "%s", fs->tag); > > All of the other files in the device directory end with trailing > newlines. Should this one be an exception? I did not think much about it but you are right that other files seem to have a newline at the end. I will add one as well. Thanks Vivek
Vivek Goyal <vgoyal@redhat.com> writes: > On Sat, Oct 21, 2023 at 04:10:21PM +0000, Alyssa Ross wrote: >> Are you still thinking about exposing this in the uevent as well? >> That would be much more convenient for me, because with this approach >> by the time the "remove" uevent arrives, it's no longer possible to >> check what tag was associated with the device — you have to store it >> somewhere when the device appears, so you can look it up again when the >> device is removed. (Not everybody uses udev.) > > Looks like systemd + udev combination can already take care of it. I just > had to specify "StopWhenUnneeded=true" in my systemd .mount unit file. And > that made sure that when device goes away, virtiofs is unmounted and > service is deactivated. My point is that, if it's not exposed in the uevent, the tag information has to be tracked somehow. systemd/udev may do that already, but every other system people might be using (mine uses mdevd) also has to track that state. Whereas if the uevent did contain that information, userspace would be able to do the unmount directly, without needing to look up some information it has previously saved. Relying on tracking state from previous events also introduces potential reliability problems — it's possible to miss uevents if the netlink queue fills up. Suppose I have a system where virtiofs filesystems should always be unmounted when the device goes away. I miss the uevent for the device being added, the user mounts the filesystem anyway, and then when the device removal uevent comes in, unless that uevent contains the filesystem tag, I'm not going to know which filesystem to unmount. > Following is my mount unit file. > > $ cat /etc/systemd/system/mnt-virtiofs.mount > > [Unit] > Description=Virtiofs mount myfs > DefaultDependencies=no > ConditionPathExists=/mnt/virtiofs > ConditionCapability=CAP_SYS_ADMIN > Before=sysinit.target > StopWhenUnneeded=true > > [Mount] > What=myfs > Where=/mnt/virtiofs > Type=virtiofs > > And following is the udev rule I used. > > $ cat /etc/udev/rules.d/80-local.rules > SUBSYSTEM=="virtio", DRIVER=="virtiofs", ATTR{tag}=="myfs", TAG+="systemd", ENV{SYSTEMD_WANTS}+="mnt-virtiofs.mount" > > And a combination of above two seems to work. virtiofs is automatically > mounted when device is hotplugged and it is unmounted when device is > hot unplugged.
On Mon, Oct 09, 2023 at 04:21:46PM -0400, Vivek Goyal wrote: > On Mon, Oct 09, 2023 at 11:53:42AM +0200, Miklos Szeredi wrote: > > On Thu, 5 Oct 2023 at 22:30, Vivek Goyal <vgoyal@redhat.com> wrote: > > > > > > virtiofs filesystem is mounted using a "tag" which is exported by the > > > virtiofs device. virtiofs driver knows about all the available tags but > > > these are not exported to user space. > > > > > > People have asked these tags to be exported to user space. Most recently > > > Lennart Poettering has asked for it as he wants to scan the tags and mount > > > virtiofs automatically in certain cases. > > > > > > https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 > > > > > > This patch exports tags through sysfs. One tag is associated with each > > > virtiofs device. A new "tag" file appears under virtiofs device dir. > > > Actual filesystem tag can be obtained by reading this "tag" file. > > > > > > For example, if a virtiofs device exports tag "myfs", a new file "tag" > > > will show up here. > > > > > > /sys/bus/virtio/devices/virtio<N>/tag > > > > > > # cat /sys/bus/virtio/devices/virtio<N>/tag > > > myfs > > > > > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com> > > > > Hi Vivek, > > > > This needs something under Documentation/. > > Hi Miklos, > > Hmm.., I can easily update the virtiofs documentation. > > Initially I was thinking to put something in Documentation/ABI/testing/ > as well. But I don't see any virtio related. In fact can't find any > files related to "virtio" at all. > > So I will just update the Documentation/filesystems/virtiofs.rst for now. No, please put it in Documentation/ABI/ that is where sysfs file information goes, and where our tools actually look for it (and test for it!) thanks, greg k-h
On Thu, Oct 05, 2023 at 04:30:30PM -0400, Vivek Goyal wrote: > virtiofs filesystem is mounted using a "tag" which is exported by the > virtiofs device. virtiofs driver knows about all the available tags but > these are not exported to user space. > > People have asked these tags to be exported to user space. Most recently > Lennart Poettering has asked for it as he wants to scan the tags and mount > virtiofs automatically in certain cases. > > https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 > > This patch exports tags through sysfs. One tag is associated with each > virtiofs device. A new "tag" file appears under virtiofs device dir. > Actual filesystem tag can be obtained by reading this "tag" file. > > For example, if a virtiofs device exports tag "myfs", a new file "tag" > will show up here. > > /sys/bus/virtio/devices/virtio<N>/tag > > # cat /sys/bus/virtio/devices/virtio<N>/tag > myfs > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com> > --- > fs/fuse/virtio_fs.c | 34 ++++++++++++++++++++++++++++++++++ > 1 file changed, 34 insertions(+) > > diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c > index 5f1be1da92ce..a5b11e18f331 100644 > --- a/fs/fuse/virtio_fs.c > +++ b/fs/fuse/virtio_fs.c > @@ -107,6 +107,21 @@ static const struct fs_parameter_spec virtio_fs_parameters[] = { > {} > }; > > +/* Forward Declarations */ > +static void virtio_fs_stop_all_queues(struct virtio_fs *fs); > + > +/* sysfs related */ > +static ssize_t tag_show(struct device *dev, struct device_attribute *attr, > + char *buf) > +{ > + struct virtio_device *vdev = container_of(dev, struct virtio_device, > + dev); > + struct virtio_fs *fs = vdev->priv; > + > + return sysfs_emit(buf, "%s", fs->tag); > +} > +static DEVICE_ATTR_RO(tag); > + > static int virtio_fs_parse_param(struct fs_context *fsc, > struct fs_parameter *param) > { > @@ -265,6 +280,15 @@ static int virtio_fs_add_instance(struct virtio_fs *fs) > return 0; > } > > +static void virtio_fs_remove_instance(struct virtio_fs *fs) > +{ > + mutex_lock(&virtio_fs_mutex); > + list_del_init(&fs->list); > + virtio_fs_stop_all_queues(fs); > + virtio_fs_drain_all_queues_locked(fs); > + mutex_unlock(&virtio_fs_mutex); > +} > + > /* Return the virtio_fs with a given tag, or NULL */ > static struct virtio_fs *virtio_fs_find_instance(const char *tag) > { > @@ -891,8 +915,15 @@ static int virtio_fs_probe(struct virtio_device *vdev) > if (ret < 0) > goto out_vqs; > > + /* Export tag through sysfs */ > + ret = device_create_file(&vdev->dev, &dev_attr_tag); > + if (ret < 0) > + goto out_sysfs_attr; You just raced with userspace and lost :( Please use default groups for your device/bus, that is what they are there for. thanks, greg k-h
On Thu, Nov 09, 2023 at 04:57:01PM +0100, Alyssa Ross wrote: > Vivek Goyal <vgoyal@redhat.com> writes: > > > On Sat, Oct 21, 2023 at 04:10:21PM +0000, Alyssa Ross wrote: > >> Are you still thinking about exposing this in the uevent as well? > >> That would be much more convenient for me, because with this approach > >> by the time the "remove" uevent arrives, it's no longer possible to > >> check what tag was associated with the device — you have to store it > >> somewhere when the device appears, so you can look it up again when the > >> device is removed. (Not everybody uses udev.) > > > > Looks like systemd + udev combination can already take care of it. I just > > had to specify "StopWhenUnneeded=true" in my systemd .mount unit file. And > > that made sure that when device goes away, virtiofs is unmounted and > > service is deactivated. > > My point is that, if it's not exposed in the uevent, the tag information > has to be tracked somehow. systemd/udev may do that already, but every > other system people might be using (mine uses mdevd) also has to track > that state. Whereas if the uevent did contain that information, > userspace would be able to do the unmount directly, without needing to > look up some information it has previously saved. > > Relying on tracking state from previous events also introduces potential > reliability problems — it's possible to miss uevents if the netlink > queue fills up. Suppose I have a system where virtiofs filesystems > should always be unmounted when the device goes away. I miss the uevent > for the device being added, the user mounts the filesystem anyway, and > then when the device removal uevent comes in, unless that uevent > contains the filesystem tag, I'm not going to know which filesystem to > unmount. I agree that it's hard for userspace to react when a virtiofs device is removed. Looking at the Linux device remove code, it appears the sysfs attr is already gone when the UNBIND event is emitted. My naive idea as a sysfs newbie is that the driver's ->remove() should be able to produce an envp[] argument to the later kobject_uevent_env() call. This would allow the driver to pass additional information to userspace. Then the virtiofs driver could include the tag string in the UNBIND event. I'm a bit surprised that this functionality doesn't already exist. Maybe there is another way of solving this? (A different approach that feels wrong is to leave the "tag" attr on the virtio device and just make sure it gets cleaned up when the virtio bus driver removes the device kobj. That way unbind udev rules could continue to use the sysfs attr.) Stefan > > > Following is my mount unit file. > > > > $ cat /etc/systemd/system/mnt-virtiofs.mount > > > > [Unit] > > Description=Virtiofs mount myfs > > DefaultDependencies=no > > ConditionPathExists=/mnt/virtiofs > > ConditionCapability=CAP_SYS_ADMIN > > Before=sysinit.target > > StopWhenUnneeded=true > > > > [Mount] > > What=myfs > > Where=/mnt/virtiofs > > Type=virtiofs > > > > And following is the udev rule I used. > > > > $ cat /etc/udev/rules.d/80-local.rules > > SUBSYSTEM=="virtio", DRIVER=="virtiofs", ATTR{tag}=="myfs", TAG+="systemd", ENV{SYSTEMD_WANTS}+="mnt-virtiofs.mount" > > > > And a combination of above two seems to work. virtiofs is automatically > > mounted when device is hotplugged and it is unmounted when device is > > hot unplugged.
On Sat, Nov 11, 2023 at 06:53:40AM -0500, Greg KH wrote: > On Thu, Oct 05, 2023 at 04:30:30PM -0400, Vivek Goyal wrote: > > virtiofs filesystem is mounted using a "tag" which is exported by the > > virtiofs device. virtiofs driver knows about all the available tags but > > these are not exported to user space. > > > > People have asked these tags to be exported to user space. Most recently > > Lennart Poettering has asked for it as he wants to scan the tags and mount > > virtiofs automatically in certain cases. > > > > https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 > > > > This patch exports tags through sysfs. One tag is associated with each > > virtiofs device. A new "tag" file appears under virtiofs device dir. > > Actual filesystem tag can be obtained by reading this "tag" file. > > > > For example, if a virtiofs device exports tag "myfs", a new file "tag" > > will show up here. > > > > /sys/bus/virtio/devices/virtio<N>/tag > > > > # cat /sys/bus/virtio/devices/virtio<N>/tag > > myfs > > > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com> > > --- > > fs/fuse/virtio_fs.c | 34 ++++++++++++++++++++++++++++++++++ > > 1 file changed, 34 insertions(+) > > > > diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c > > index 5f1be1da92ce..a5b11e18f331 100644 > > --- a/fs/fuse/virtio_fs.c > > +++ b/fs/fuse/virtio_fs.c > > @@ -107,6 +107,21 @@ static const struct fs_parameter_spec virtio_fs_parameters[] = { > > {} > > }; > > > > +/* Forward Declarations */ > > +static void virtio_fs_stop_all_queues(struct virtio_fs *fs); > > + > > +/* sysfs related */ > > +static ssize_t tag_show(struct device *dev, struct device_attribute *attr, > > + char *buf) > > +{ > > + struct virtio_device *vdev = container_of(dev, struct virtio_device, > > + dev); > > + struct virtio_fs *fs = vdev->priv; > > + > > + return sysfs_emit(buf, "%s", fs->tag); > > +} > > +static DEVICE_ATTR_RO(tag); > > + > > static int virtio_fs_parse_param(struct fs_context *fsc, > > struct fs_parameter *param) > > { > > @@ -265,6 +280,15 @@ static int virtio_fs_add_instance(struct virtio_fs *fs) > > return 0; > > } > > > > +static void virtio_fs_remove_instance(struct virtio_fs *fs) > > +{ > > + mutex_lock(&virtio_fs_mutex); > > + list_del_init(&fs->list); > > + virtio_fs_stop_all_queues(fs); > > + virtio_fs_drain_all_queues_locked(fs); > > + mutex_unlock(&virtio_fs_mutex); > > +} > > + > > /* Return the virtio_fs with a given tag, or NULL */ > > static struct virtio_fs *virtio_fs_find_instance(const char *tag) > > { > > @@ -891,8 +915,15 @@ static int virtio_fs_probe(struct virtio_device *vdev) > > if (ret < 0) > > goto out_vqs; > > > > + /* Export tag through sysfs */ > > + ret = device_create_file(&vdev->dev, &dev_attr_tag); > > + if (ret < 0) > > + goto out_sysfs_attr; > > You just raced with userspace and lost :( > > Please use default groups for your device/bus, that is what they are > there for. Hi Greg, Getting back to this thread after a long time. Sorry, got busy in other things. Trying to understand the races with user space. IIUC, your primary concern is that kobject has already been published to the user space and I am trying to add/remove attributes to the object after the fact and that makes it racy? And suggestion is to first fully create kobject (including attrs) before publishing it to user space. Do I understand it correctly or I missed the point completely? Thanks Vivek
On Fri, Jan 05, 2024 at 03:44:55PM -0500, Vivek Goyal wrote: > On Sat, Nov 11, 2023 at 06:53:40AM -0500, Greg KH wrote: > > On Thu, Oct 05, 2023 at 04:30:30PM -0400, Vivek Goyal wrote: > > > virtiofs filesystem is mounted using a "tag" which is exported by the > > > virtiofs device. virtiofs driver knows about all the available tags but > > > these are not exported to user space. > > > > > > People have asked these tags to be exported to user space. Most recently > > > Lennart Poettering has asked for it as he wants to scan the tags and mount > > > virtiofs automatically in certain cases. > > > > > > https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 > > > > > > This patch exports tags through sysfs. One tag is associated with each > > > virtiofs device. A new "tag" file appears under virtiofs device dir. > > > Actual filesystem tag can be obtained by reading this "tag" file. > > > > > > For example, if a virtiofs device exports tag "myfs", a new file "tag" > > > will show up here. > > > > > > /sys/bus/virtio/devices/virtio<N>/tag > > > > > > # cat /sys/bus/virtio/devices/virtio<N>/tag > > > myfs > > > > > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com> > > > --- > > > fs/fuse/virtio_fs.c | 34 ++++++++++++++++++++++++++++++++++ > > > 1 file changed, 34 insertions(+) > > > > > > diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c > > > index 5f1be1da92ce..a5b11e18f331 100644 > > > --- a/fs/fuse/virtio_fs.c > > > +++ b/fs/fuse/virtio_fs.c > > > @@ -107,6 +107,21 @@ static const struct fs_parameter_spec virtio_fs_parameters[] = { > > > {} > > > }; > > > > > > +/* Forward Declarations */ > > > +static void virtio_fs_stop_all_queues(struct virtio_fs *fs); > > > + > > > +/* sysfs related */ > > > +static ssize_t tag_show(struct device *dev, struct device_attribute *attr, > > > + char *buf) > > > +{ > > > + struct virtio_device *vdev = container_of(dev, struct virtio_device, > > > + dev); > > > + struct virtio_fs *fs = vdev->priv; > > > + > > > + return sysfs_emit(buf, "%s", fs->tag); > > > +} > > > +static DEVICE_ATTR_RO(tag); > > > + > > > static int virtio_fs_parse_param(struct fs_context *fsc, > > > struct fs_parameter *param) > > > { > > > @@ -265,6 +280,15 @@ static int virtio_fs_add_instance(struct virtio_fs *fs) > > > return 0; > > > } > > > > > > +static void virtio_fs_remove_instance(struct virtio_fs *fs) > > > +{ > > > + mutex_lock(&virtio_fs_mutex); > > > + list_del_init(&fs->list); > > > + virtio_fs_stop_all_queues(fs); > > > + virtio_fs_drain_all_queues_locked(fs); > > > + mutex_unlock(&virtio_fs_mutex); > > > +} > > > + > > > /* Return the virtio_fs with a given tag, or NULL */ > > > static struct virtio_fs *virtio_fs_find_instance(const char *tag) > > > { > > > @@ -891,8 +915,15 @@ static int virtio_fs_probe(struct virtio_device *vdev) > > > if (ret < 0) > > > goto out_vqs; > > > > > > + /* Export tag through sysfs */ > > > + ret = device_create_file(&vdev->dev, &dev_attr_tag); > > > + if (ret < 0) > > > + goto out_sysfs_attr; > > > > You just raced with userspace and lost :( > > > > Please use default groups for your device/bus, that is what they are > > there for. > > Hi Greg, > > Getting back to this thread after a long time. Sorry, got busy in other > things. > > Trying to understand the races with user space. IIUC, your primary concern > is that kobject has already been published to the user space and I am > trying to add/remove attributes to the object after the fact and that > makes it racy? Yes. > And suggestion is to first fully create kobject (including attrs) before > publishing it to user space. Do I understand it correctly or I missed > the point completely? Use the default groups pointer in the device instead which makes the driver core create/remove the files automatically for you. Here's something I wrote over a decade ago as to how to prevent this: http://www.kroah.com/log/blog/2013/06/26/how-to-create-a-sysfs-file-correctly/ thanks, greg k-h
diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c index 5f1be1da92ce..a5b11e18f331 100644 --- a/fs/fuse/virtio_fs.c +++ b/fs/fuse/virtio_fs.c @@ -107,6 +107,21 @@ static const struct fs_parameter_spec virtio_fs_parameters[] = { {} }; +/* Forward Declarations */ +static void virtio_fs_stop_all_queues(struct virtio_fs *fs); + +/* sysfs related */ +static ssize_t tag_show(struct device *dev, struct device_attribute *attr, + char *buf) +{ + struct virtio_device *vdev = container_of(dev, struct virtio_device, + dev); + struct virtio_fs *fs = vdev->priv; + + return sysfs_emit(buf, "%s", fs->tag); +} +static DEVICE_ATTR_RO(tag); + static int virtio_fs_parse_param(struct fs_context *fsc, struct fs_parameter *param) { @@ -265,6 +280,15 @@ static int virtio_fs_add_instance(struct virtio_fs *fs) return 0; } +static void virtio_fs_remove_instance(struct virtio_fs *fs) +{ + mutex_lock(&virtio_fs_mutex); + list_del_init(&fs->list); + virtio_fs_stop_all_queues(fs); + virtio_fs_drain_all_queues_locked(fs); + mutex_unlock(&virtio_fs_mutex); +} + /* Return the virtio_fs with a given tag, or NULL */ static struct virtio_fs *virtio_fs_find_instance(const char *tag) { @@ -891,8 +915,15 @@ static int virtio_fs_probe(struct virtio_device *vdev) if (ret < 0) goto out_vqs; + /* Export tag through sysfs */ + ret = device_create_file(&vdev->dev, &dev_attr_tag); + if (ret < 0) + goto out_sysfs_attr; + return 0; +out_sysfs_attr: + virtio_fs_remove_instance(fs); out_vqs: virtio_reset_device(vdev); virtio_fs_cleanup_vqs(vdev); @@ -922,6 +953,9 @@ static void virtio_fs_remove(struct virtio_device *vdev) struct virtio_fs *fs = vdev->priv; mutex_lock(&virtio_fs_mutex); + /* Remove tag attr from sysfs */ + device_remove_file(&vdev->dev, &dev_attr_tag); + /* This device is going away. No one should get new reference */ list_del_init(&fs->list); virtio_fs_stop_all_queues(fs);
virtiofs filesystem is mounted using a "tag" which is exported by the virtiofs device. virtiofs driver knows about all the available tags but these are not exported to user space. People have asked these tags to be exported to user space. Most recently Lennart Poettering has asked for it as he wants to scan the tags and mount virtiofs automatically in certain cases. https://gitlab.com/virtio-fs/virtiofsd/-/issues/128 This patch exports tags through sysfs. One tag is associated with each virtiofs device. A new "tag" file appears under virtiofs device dir. Actual filesystem tag can be obtained by reading this "tag" file. For example, if a virtiofs device exports tag "myfs", a new file "tag" will show up here. /sys/bus/virtio/devices/virtio<N>/tag # cat /sys/bus/virtio/devices/virtio<N>/tag myfs Signed-off-by: Vivek Goyal <vgoyal@redhat.com> --- fs/fuse/virtio_fs.c | 34 ++++++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+)