On Mon, 2006-01-30 at 10:13 -0800, Jean Tourrilhes wrote: > On Fri, Jan 27, 2006 at 05:01:44PM -0800, Jouni Malinen wrote: > > On Thu, Jan 12, 2006 at 03:00:58PM -0500, Dan Williams wrote: > > > > > ESSIDs can technically include NULL characters. Drivers should not be > > > adjusting the length of the ESSID before reporting it in their > > > SIOCGIWESSID handlers. Breaks stuff like wpa_supplicant. Note that ipw > > > drivers, which seem to currently be the "most correct", don't have this > > > problem. > > > > I'm still trying to go through the huge amount of email I've received > > during two-week time travelling and noticed this one just now. I did not > > see any discussion about this on netdev list, but I believe the change > > here is changes the WE user space interface and this has not been done > > previously in order to not break user space programs. > > > > I agree that the choice of SSID len+1 is unfortunate for a data > > structure that is not really a string but array of octets. However, that > > was the way SIOCGIWESSID/SIOCSIWESSID was designed if I have understood > > correctly (Jean cc'ed just in case). > > Yes. I admit that when I designed the API (a long time ago), I > made this mistake. It could have been worse, I could have defined it > as a pure C string (i.e. without any lenght). As it's a minor defect > and pretty easy to workaround (just add 1), I decided to leave it as > it is, to not break API compatibility. > I also made all the wtools accept both forms form > SIOCGIWESSID, so that any future change would work. > > > Is this change, if applied (or was it already applied?), an indication > > that we are now changing the WE interface in a way that is not > > backwards compatible and we are willing to break an existing kernel-user > > space interface? > > Good old API practices seem to be out of the window. Yes, I > would have prefered to be consulted, to change the WE number and make > sure all the bits fall in place.
I'll post a revert patch for the patch I originally sent so that we go back to the original behavior. Dan - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
