On Friday, 15 September 2006 23:28, Tim Dijkstra wrote:
> On Fri, 15 Sep 2006 22:15:09 +0200
> "Rafael J. Wysocki" <[EMAIL PROTECTED]> wrote:
>
> > On Friday, 15 September 2006 21:39, Tim Dijkstra wrote:
> > > On Fri, 15 Sep 2006 11:06:36 +0200
> > > Pavel Machek <[EMAIL PROTECTED]> wrote:
> > >
> > > > Hi!
> > > >
> > > > > This patch will lock the VT that is active just after splash is
> > > > > initialized and release it just before the splash system is
> > > > > stopped.
> > > > >
> > > > > I haven't tested it extensively with bootsplash.org, but it
> > > > > seems to work with that (in vmware) too.
> > > > >
> > > > > Comments?
> > > > >
> > > >
> > > > > @@ -1027,6 +1030,75 @@
> > > > > }
> > > > > #endif
> > > > >
> > > > > +
> > > > > +static void release_vt(int signo)
> > > > > +{
> > > > > + printf("Ignoring request to release VT\n");
> > > > > + ioctl(vfd, VT_RELDISP, 0);
> > > > > +}
> > > > > +
> > > >
> > > > Is printf safe to call from signal handler?
> > >
> > > It seems to work. But these were there mainly for testing, so
> > > they can go.
> > >
> > > > > +static int lock_vt()
> > > >
> > > > (void).
> > > >
> > > > > + error = ioctl(fd, VT_GETSTATE, &vtstat);
> > > > > + close(fd);
> > > > > +
> > > > > + if ( error < 0 )
> > > >
> > > > (error < 0), (no spaces) please.
> > > >
> > > > > + /* Setting vt mode to VT_PROCESS means this process
> > > > > + * will handle vt switching requests.
> > > > > + * We installed handlers for release and acquire
> > > > > requests.
> > > > > + */
> > > >
> > > > Is it possible to just ignore those signals? Kernel will be
> > > > unable to switch, anyway, no?
> > >
> > > I think we can drop the handler for the acquire event. That should
> > > never happen anyway.
> > >
> > > Hmm, I could have sworn that I tested with SIG_IGN as a handler, and
> > > that it did not work. After testing it again, it does seem to do the
> > > trick ...
> > >
> > > But installing that handler can't hurt, right? It does explicitly
> > > dismiss the switch. But indeed, after reading the linux source, I
> > > think ignoring the signal works also.
> > >
> > > > > + printf("Locked against VT switching\n");
> > > >
> > > > This should be gone in final version.
> > > >
> > > > Otherwise looks okay to me...
> > > >
> > > > Pavel
> > >
> > > Rafael:
> > > > It looks good to me too except for one thing: could we please
> > > > return well-defined error codes from lock_vt() instead of -1?
> > >
> > > What do you propose? Return -errno?
> >
> > For example:
> >
> > error = ioctl(vfd, VT_GETMODE, &vtm);
> > if (error)
> > return error;
>
> AFAIK ioctl will on error always return -1
Yes, but the above looks better than "if (ioctl(...) < 0)) return -1" IMHO. ;-)
Greetings,
Rafael
--
You never change things by fighting the existing reality.
R. Buckminster Fuller
-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Suspend-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/suspend-devel
