On Thu, Mar 18, 2010 at 02:59:00PM +0100, Milan Broz wrote: > For some node operation is userspace we need to know > if previous suspend state of device really changed. > Patch add new flag to indicate suspend state change. I'd like more discussion of this one. (a) Why userspace needs it and the alternatives. - Is it just an optimisation to save an info ioctl? (b) A precise definition of the flag in the header file, mentioning 'corner cases' explicitly, and consideration of alternative ways of getting the information out of the kernel. (E.g. Different flag definitions, alternative or generic ways of returning 'previous' state if that is what's needed.) Alasdair -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
