On Thu, 2014-09-18 at 11:39 +0200, Lee Duncan wrote: > On 09/18/2014 03:05 AM, Nicholas A. Bellinger wrote: > > On Wed, 2014-09-17 at 14:01 -0600, Lee Duncan wrote: > >> Hi Nicholas: > >> > >> > >> Apologies for the top-post, but my current mail agent sucks. :) > >> > >> > >> No problem on the delay. > >> > >> > >> What about handling "NULL" and "" as meaning "clear the current value? > >> I could supply a patch ... > >> > >> > > > > While writing the earlier response, I was trying to remember why "" was > > deemed insufficient in the first place for clearing the current value, > > but alas the original reasoning escapes me atm. > > Perhaps the reason for this is that writing an empty string from the > shell is not easy. Trying to do an 'echo -n "" > some_file" gets > optimized into not writing anything out. I had to use something like: > 'echo -n "\0" > some_file" to get it to write an empty string. > Mmmm, yes this sounds familiar.. > > > > It could very well have to do with some early user-space (eg: lio-utils) > > during state re-creation had a distinction between an empty value and > > explicitly clearing a value, but I still don't recall exactly why this > > mattered. > > > > That said, there was some reasoning behind this, and I'd like to > > remember why first before making a change that could negatively effect > > user-space. > > If you remember please let me know, as I've patched our kernel to accept > these NULL values. Ok, in that case I'm OK with entertaining a patch as long no regressions are introduced that break existing user-space. Adding Jerome to the CC' on the user-space side for good measure. Thanks Lee! --nab -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
