Message-ID: <20150427153711.GA23428@kroah.com>
Date:	Mon, 27 Apr 2015 17:37:11 +0200
From:	Greg KH <greg@...ah.com>
To:	Beata Michalska <b.michalska@...sung.com>
Cc:	linux-kernel@...r.kernel.org, linux-fsdevel@...r.kernel.org,
	linux-api@...r.kernel.org, jack@...e.cz, tytso@....edu,
	adilger.kernel@...ger.ca, hughd@...gle.com, lczerner@...hat.com,
	hch@...radead.org, linux-ext4@...r.kernel.org, linux-mm@...ck.org,
	kyungmin.park@...sung.com, kmpark@...radead.org
Subject: Re: [RFC v2 1/4] fs: Add generic file system event notifications

On Mon, Apr 27, 2015 at 05:08:27PM +0200, Beata Michalska wrote:
> On 04/27/2015 04:24 PM, Greg KH wrote:
> > On Mon, Apr 27, 2015 at 01:51:41PM +0200, Beata Michalska wrote:
> >> Introduce configurable generic interface for file
> >> system-wide event notifications, to provide file
> >> systems with a common way of reporting any potential
> >> issues as they emerge.
> >>
> >> The notifications are to be issued through generic
> >> netlink interface by newly introduced multicast group.
> >>
> >> Threshold notifications have been included, allowing
> >> triggering an event whenever the amount of free space drops
> >> below a certain level - or levels to be more precise as two
> >> of them are being supported: the lower and the upper range.
> >> The notifications work both ways: once the threshold level
> >> has been reached, an event shall be generated whenever
> >> the number of available blocks goes up again re-activating
> >> the threshold.
> >>
> >> The interface has been exposed through a vfs. Once mounted,
> >> it serves as an entry point for the set-up where one can
> >> register for particular file system events.
> >>
> >> Signed-off-by: Beata Michalska <b.michalska@...sung.com>
> >> ---
> >>  Documentation/filesystems/events.txt |  231 ++++++++++
> >>  fs/Makefile                          |    1 +
> >>  fs/events/Makefile                   |    6 +
> >>  fs/events/fs_event.c                 |  770 ++++++++++++++++++++++++++++++++++
> >>  fs/events/fs_event.h                 |   25 ++
> >>  fs/events/fs_event_netlink.c         |   99 +++++
> >>  fs/namespace.c                       |    1 +
> >>  include/linux/fs.h                   |    6 +-
> >>  include/linux/fs_event.h             |   58 +++
> >>  include/uapi/linux/fs_event.h        |   54 +++
> >>  include/uapi/linux/genetlink.h       |    1 +
> >>  net/netlink/genetlink.c              |    7 +-
> >>  12 files changed, 1257 insertions(+), 2 deletions(-)
> >>  create mode 100644 Documentation/filesystems/events.txt
> >>  create mode 100644 fs/events/Makefile
> >>  create mode 100644 fs/events/fs_event.c
> >>  create mode 100644 fs/events/fs_event.h
> >>  create mode 100644 fs/events/fs_event_netlink.c
> >>  create mode 100644 include/linux/fs_event.h
> >>  create mode 100644 include/uapi/linux/fs_event.h
> > 
> > Any reason why you just don't do uevents for the block devices today,
> > and not create a new type of netlink message and userspace tool required
> > to read these?
> 
> The idea here is to have support for filesystems with no backing device as well.
> Parsing the message with libnl is really simple and requires few lines of code
> (sample application has been presented in the initial version of this RFC)

I'm not saying it's not "simple" to parse, just that now you are doing
something that requires a different tool.  If you have a block device,
you should be able to emit uevents for it, you don't need a backing
device, we handle virtual filesystems in /sys/block/ just fine :)

People already have tools that listen to libudev for system monitoring
and management, why require them to hook up to yet-another-library?  And
what is going to provide the ability for multiple userspace tools to
listen to these netlink messages in case you have more than one program
that wants to watch for these things (i.e. multiple desktop filesystem
monitoring tools, system-health checkers, etc.)?

> >> --- a/fs/Makefile
> >> +++ b/fs/Makefile
> >> @@ -126,3 +126,4 @@ obj-y				+= exofs/ # Multiple modules
> >>  obj-$(CONFIG_CEPH_FS)		+= ceph/
> >>  obj-$(CONFIG_PSTORE)		+= pstore/
> >>  obj-$(CONFIG_EFIVAR_FS)		+= efivarfs/
> >> +obj-y				+= events/
> > 
> > Always?
> > 
> >> diff --git a/fs/events/Makefile b/fs/events/Makefile
> >> new file mode 100644
> >> index 0000000..58d1454
> >> --- /dev/null
> >> +++ b/fs/events/Makefile
> >> @@ -0,0 +1,6 @@
> >> +#
> >> +# Makefile for the Linux Generic File System Event Interface
> >> +#
> >> +
> >> +obj-y := fs_event.o
> > 
> > Always?  Even if the option is not selected?  Why is everyone forced to
> > always use this code?  Can't you disable it for the "tiny" systems that
> > don't need it?
> > 
> 
> I was considering making it optional and I guess it's worth getting back
> to this idea.

The "linux-tiny" people will appreciate that :)

> >> +struct fs_trace_entry {
> >> +	atomic_t	 count;
> > 
> > Why not just use a 'struct kref' for your count, which will save a bunch
> > of open-coding of reference counting, and forcing us to audit your code
> > to verify you got all the corner cases correct?  :)
> > 
> >> +	atomic_t	 active;
> >> +	struct super_block *sb;
> 
> Not sure if using kref would change much here as the kref would not really
> make it easier to verify those corner cases, unfortunately.

Why not, that's the goal of a kref.  Yes, you already did the hard work,
but now you require everyone else to also do the hard work of trying to
audit your code.  That's why we have common functions/data structures in
the kernel, to make long-term maintenance easier.

Please switch to make it so that we "know" you are doing this correctly.

> > Are you properly reference counting this pointer?  I didn't see where
> > that was happening, so I must have missed it.
> > 
> > thanks,
> >
> 
> You haven't. And if I haven't missed anything, the sb is being used only
> as long as the super is alive.

How do you know that? :)

> Most of the code operates on sb only if it
> was explicitly asked to, through call from filesystem. There is also
> a callback notifying of mount being dropped (which proceeds the call to 
> kill_super) that invalidates the object that depends on it.
> Still, it should be explicitly stated that the sb is being used through
> bidding up the s_count counter, though that would require taking the
> sb_lock. AFAIK, one can get the reference to super block but for a particular
> device. Maybe it would be worth having it more generic (?).

Why not just grab a reference to the sb when you save the pointer, and
release it when you are done with it?  That should handle the lifecycle
properly.  It's always a very bad idea to have a pointer to a reference
counted object without actually grabbing the reference, as you have no
idea what is happening with it behind your back.

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists