Messages in this thread | | | Date | Tue, 3 Jun 2008 09:35:15 -0700 | From | Greg KH <> | Subject | Re: [PATCH 09/10] netns: Enable tagging for net_class directories in sysfs |
| |
On Tue, Jun 03, 2008 at 05:24:21PM +0200, Benjamin Thery wrote: > Greg KH wrote: >> On Mon, Jun 02, 2008 at 03:46:08PM +0200, Benjamin Thery wrote: >>> net: Enable tagging for net_class directories in sysfs >>> >>> The problem. Network devices show up in sysfs and with the network >>> namespace active multiple devices with the same name can show up in >>> the same directory, ouch! >>> >>> To avoid that problem and allow existing applications in network >>> namespaces >>> to see the same interface that is currently presented in sysfs, this >>> patch enables the tagging directory support in sysfs. >>> >>> By using the network namespace pointers as tags to separate out the >>> the sysfs directory entries we ensure that we don't have conflicts >>> in the directories and applications only see a limited set of >>> the network devices. >> I don't like it how the network subsystem is starting to leach into the >> sysfs core here. What happens when the next subsystem wants to do the >> same thing? And then the next one? Will they all have to do this kind >> of intrusive changes to sysfs? > > > > Can't this be done only in the network subsystem? > > I'm not sure to understand exactly what you mean. > > What you don't like is seeing these hunks of network code in > fs/sysfs/mount.c? And you prefer to see these bits of code resides in > the network subsystem instead and see only "generic" sysfs services in > fs/sysfs/mount.c?
Yes, exactly.
I don't want the problem that if more subsystems want to implement something like this, they too need to modify the sysfs core.
And the mess with the #ifdef, that's not nice either :)
> If this is it, I have some idea to implement a less intrusive > sysfs_net_exit(), which can be shared with the other namespaces. > Serge introduces the same kind of changes in patch 10 to fix an issue in > user namespace. I think we can share a bit of code and move the parts > specific to each namespace in their own subsystems.
I think that would be a good idea. Care to redo the series?
thanks,
greg k-h
| |