[PATCH ]Re: hal doesn't recognize mounting of devices

David Zeuthen david at fubar.dk
Tue Oct 25 10:00:45 PDT 2005


On Tue, 2005-10-25 at 18:38 +0200, Schupp Roderich (extern) BenQ MD PD 
> What about the other points raised by Al Viro on LKML
> (_not_ the namespace issues - if namespaces are ever used
> in Linux there will bigger breakage to fix):

Sigh. Viro is being a pedant. News at 11.

> - there may be more than one mountpoint for any given device

Just chose the first one you see. For all desktop usage this is
sufficient. So right now the definition of volume.mount_point is

 A fully qualified path to the mount point of the volume

maybe we should change it to

 A fully qualified path to a mount point of the volume

> - hald may find the device in /proc/mounts, but the mountpoint 
>   may have been mount --move'd away (there's no uevent for that yet, right?)

If the kernel really supports crazy things like this we could build
support for it in HAL. 

Not to whine, but, really, I wish people would keep it real and work on
the issues that matter for 99.9% of desktop users [1] instead of
pathological examples like this. 

Cheers,
David

[1] : Like, for example, how about someone doing a patch to give us a
Format() method on volumes such that we can enable users to easily
format their USB pendrives and create e.g. LUKS cryptography volumes





More information about the hal mailing list