A new thought: I'm not ruling out that my corporate AnyConnect VPN somehow interfered with the driver, since "OAS" sounds like an acronym that my company might use. So if you find no evidence for "tapoas", I wouldn't be too concerned. On 24 April 2014 09:25, David Woodhouse <dwmw2 at infradead.org> wrote: > On Thu, 2014-04-17 at 17:01 -0400, Greg Toombs wrote: >> Found the problem. tun-win32.c:45 - >> #define TAP_COMPONENT_ID "tap0901" >> >> This is only valid for the most recent version of the TAP adapter. For >> other versions, this should actually be "tapoas". So openconnect >> saying that there are no TAP adapters is incorrect. Instead, it should >> check all known component IDs, and then check for the version, so that >> instead of reporting no TAP adapters, it detects the following: >> >> Error: TAP-Windows driver v9.9 or greater is required (found 9.7) >> Set up tun device failed > > Hm, really "tapoas"? Where did that driver come from? > > As far as I can tell from the OpenVPN revision history, they've used > 'tap0901' since February 2007? and it was 'tap0801' before that. > > Adding the openvon-devel list to Cc since the clue we seek is probably > there... > > -- > dwmw2 > > ? https://github.com/OpenVPN/tap-windows/commit/ee313a978f7aa9 >
