Re: (driver model) bus kset list manipulation bug

[Hollis Blanchard]

On Jan 27, 2004, at 5:31 PM, Hollis Blanchard wrote:
> devices_subsys looks like it's only used for two things: global 
> hotplug policy and suspend. Of the 3 hotplug functions it provides 
> (dev_hotplug_filter, dev_hotplug_name, and dev_hotplug), 2 of them 
> refer to bus data or code anyways.
>
> I'm very surprised to see it's used by device_shutdown(). I thought 
> one of the points of the device tree was to do depth-first-suspend, so 
> e.g we don't try to suspend a PCI bridge and *then* try to suspend 
> children of that bridge. Instead we're walking a global list in the 
> reverse order they were registered. I guess this works because busses 
> are discovered from the root down, so going backwards will give you 
> the deepest first.

To reply to myself again (starting to get the hint...), I wonder how 
long the global devices_subsys list will work for power-suspend once we 
start hotplugging devices and busses? Seems to me that a cascading bus 
power-down message is what has to happen...

--
Hollis Blanchard
IBM Linux Technology Center

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