Re: unify sysfs device tree

[Kay Sievers]

On Mon, Jan 16, 2006 at 10:03:06AM -0500, Mike Snitzer wrote:
> On 1/16/06, Kay Sievers <kay.sievers@xxxxxxxx> wrote:
> 
> > Here is an updated patch, that:
> >   o moves the devices in /sys/block to /sys/devices to match the
> >     class layout. Block devices will be childs of their physical
> >     device chain like every other class device too. Partitions
> >     will be childs of the disk device. A usual DEVPATH looks like:
> >       /devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/sda/sda1
> >
> >   o flattens the block class view and moves the block symlinks to
> >     /sys/class/block. Disks and partitons like /sys/class/block/sda
> >     and /sys/class/block/sda1 will be at the same level. /sys/block
> >     does not longer exist.
> 
> What is the problem with maintaining compatibility by having
> /sys/block be a symlink to /sys/class/block?  Userspace applications
> shouldn't have to now conditionalize the path to block devices
> (/sys/block/... vs /sys/class/block/...).  Forcing this kind of change
> is what taints Linux for use in hardened applications. 
> Conditionalizing code for 2.4 vs 2.6 is understandable but having to
> do so for minor 2.6.x revisions is rediculous.  <insert ref to Linus'
> views in the recent 'userspace breakage' thread here>.

Please read the whole mail before wasting other peoples time:

> I've removed all historical stuff like the "device" and "bus" link,
> cause they don't make sense anymore. What we will actually remove with
> the real conversion, does nobody know at this point.

Thanks,
Kay
-
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/