On Tue, Apr 24, 2007 at 05:55:39PM -0700, Cory Petkovsek wrote: > Thanks for the suggestions guys. Looking through the NDIS wrapper package > Makefile I see that KERNEL would be defined if the macros were working > properly. CFLAGS is set differently depending on the target, but for whatever > reason the macros aren't running properly so CFLAGS is never set at all. > Since > I'm not proficient with make, I resolved the problem by manually changing the > CFLAGS variable to the right setting before running each target. >
hm. you might want to work out these problems first since there may be other flags your missing that could cause more subtle failure modes that simple compile failures. > I got the package from this page: > http://www.opensolaris.org/os/community/laptop/wireless/ndis/ > > Again I'm running nexenta with opensolaris build 61 and GNU Make 3.81beta4. I > compiled the driver with gcc 4.0.3 and though I saw warnings it compiled > properly. I had to clean up my driver inf file to remove unicode, but > otherwise it seemed ok. > > The driver for my card (wli-cb-g54hp) came from here: > http://ndiswrapper.sourceforge.net/mediawiki/index.php/List > > It loaded but the "driver failed to attach". So now that I can compile I'll > try a few drivers and compiling with gcc3.4 as was suggested. Is there a way > I > can get more details about why it failed to attach? Or does it simply mean > that the driver and device do not match up? > well, the device framework tried to attach the driver to the device, but the problem is that the driver's attach entry point is returning failure. i've never looked at the ndis wrapper, but i'm assuming that the attach routine is defined in there and for some reason it's failing. pulling up that source for the attach routine and using dtrace would be a good way to figure it out what exactly is going wrong. ed > Thanks! > Cory > > On Sun, Apr 22, 2007 at 11:01:10PM -0700, Edward Pilatowicz wrote: > > well, given to the error message below i'm guessing that the problem is > > not with your usage of ddi_acc_handle_t, but whatever comes before it in > > the source file. probably a struct or union that doesn't have a semicolon > > after it? > > > > also, if make sure you compile with KERNEL defined, otherwise you > > won't get the definition for ddi_acc_handle_t even if you include > > sys/dditypes.h. > > > > ed > > > > On Sat, Apr 21, 2007 at 09:26:23PM -0700, Cory Petkovsek wrote: > > > Guys, I'm attempting to compile the NDIS driver for a buffalo wifi card > > > (wli-cb-g54hp) and have run accross simple definition errors compiling. > > > Gcc > > > complains about > > > > > > ../if_ndisvar.h:60: error: syntax error before 'ddi_acc_handle_t' > > > ../if_ndisvar.h:60: warning: no semicolon at end of struct or union > > > > > > I've found ddi_acc_handle_t is defined here as such: > > > > > > /usr/include/sys/dditypes.h: > > > typedef struct __ddi_acc_handle *ddi_acc_handle_t; > > > > > > The results don't change if I manually include <sys/dditypes.h>. Also in > > > this > > > same struct are ddi_dma_handle_t and ddi_dma_cookie_t which are both > > > defined in > > > dditypes.h. Gcc never complains about them. If I remove the > > > ddi_acc_handle_t > > > line it has no further problems with that struct, but I'm sure I don't > > > want to > > > do that. > > > > > > How can I get this to compile? > > > > > > I'm running nexenta with build 61 of opensolaris, gcc 4.0.3, make > > > 3.81beta4. > > > > > > Thanks, > > > Cory > > > > > -- > Cory Petkovsek Adapting Information > Adaptable IT Consulting Technology to Your > (858) 705-1655 Business > cory at AdaptableIT.com www.AdaptableIT.com > _______________________________________________ > laptop-discuss mailing list > laptop-discuss at opensolaris.org
