Re: SUMMARY: GGI/X : the other way??

MenTaLboY (mentalboy@geocities.com)
Tue, 7 Apr 1998 01:42:50 -0400 (EDT)


On Mon, 6 Apr 1998, Albert D. Cahalan wrote:
> If revoke() can kill a mmap on a file, users can make servers crash.
> It might be better to convert mappings to /dev/zero or /dev/null
> so that servers won't get SIGSEGV or SIGBUS.

Or alternatively, if you're not using a trusted server, but using a library
like LibGGI with KGI, where the application is notified when the framebuffer
is unmapped could be a solution. Because it's a library and not a shared
server, it only shoots itself (that particular instance of the application)
in the foot, not the whole graphics system. (and the kernel portion can set
up a sane graphics state again on that VT, without disturbing the rest of
the graphics apps).

In any case, most apps' response to the unmapping would be simply to stop
themselves until the framebuffer was remapped to them, and they were
notified with appropriate signals. Others might save the framebuffer, and
continue to work with a copy until it was mapped back, and still others
might continue processing, but forgo drawing.

-=MenTaLboY=-

Btw ... LibGGI is independent of KGI, and does not require it at all. LibGGI
even runs on top of X.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu