On 14/1/07 11:05 am, "Jan Engelhardt" [off-list ref] wrote:

quoted

The block device frontend driver allows the kernel to access block
devices exported exported by a virtual machine containing a physical
block device driver.

Is this significantly different from ubd/hostfs that it actually warrants a
reinvention?

It is certainly unlike hostfs because hostfs provides file-level access to
host storage, not block-level. It's unlike both ubd and hostfs in that both
of those (I believe) make significant use of the syscall interface (and so
assume they run in a process on a Linux host). Also our driver appears to be
lower level, pushing more responsibility for features like CoW into the VMM.
Arguably that's a more generically reusable and flexible strategy although
it requires more VMM run-time support (which a Xen system provides).

quoted

+ (void)xenbus_switch_state(info->xbdev, XenbusStateConnected);

Cast remove, if xenbus_switch_state does not have __must_check.
Also elsewhere.

Okay, we should certainly follow the general rule here.

 Thanks,
 Keir