Richard L. Hamilton wrote: >> Yes, it's implemented. The streamio.c strioctl() >> implementation is the >> real one. The gentty.c module is a red herring. >> It's part of the >> bscure "sy" driver ... which has no man page and >> dates back to the USL >> implementation. I have no idea what it does. :-/ > > At least on Solaris 10 (don't have anything newer > handy at the moment): > # modinfo|grep ' sy ' > 190 7b7f5570 59c 22 1 sy (Indirect driver for tty 'sy' 1.) > # ls -lL /dev/tty > crw-rw-rw- 1 root tty 22, 0 Aug 6 11:44 /dev/tty > > In other words, sy implements /dev/tty...and probably > predates STREAMS-based ttys even. > > (I'm having SVR2 flashbacks, having read a lot of that code > way back when. In fact, I think back in even earlier code, > the sy driver handled something else in addition, maybe > the console, as another minor device. But maybe my memory > is playing tricks on me.)
Ah, ok. > So I'm presuming that under normal use, the streamio code > takes care of the TIOCNOTTY ioctl, and it never gets down to > the sy driver (which by the comments, couldn't safely handle it). It's not a STREAMS driver (look at the cb_ops), so, no, the streamio code has nothing to do what that. If 'sy' is somehow your controlling terminal, I don't believe that TIOCNOTTY will do what you expect. But then since /dev/tty is supposed to be a representation of your controlling tty (which is in turn some other device; see tty(7D)), you'd have a chicken-and-egg problem in trying to assign /dev/tty _as_ your controlling tty in the first place. -- James Carlson 42.703N 71.076W <carls...@workingcode.com> _______________________________________________ opensolaris-code mailing list opensolaris-code@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
