Re: /dev/psaux
Michel Dänzer wrote:
On Wed, 2004-03-24 at 09:26, Ole-Egil Hvitmyren wrote:
Michel Dänzer wrote:
On Wed, 2004-03-24 at 03:02, Viktor Rosenfeld wrote:
Minor gotcha: I switched the X mouse device from /dev/gpmdata (I had gpm
repeat mouse events) to /dev/input/mice, because I remember reading that
with the new 2.6 kernel more than one processes can listen to mouse
events
/dev/input/mice could always do that, only broken devices like
/dev/psaux couldn't.
Uhm, on every system I have seen, gpm is connected to /dev/psaux when in
console and XFree86 is connected when in in X.
From X:
olegil@olegil:~$ sudo lsof | grep psaux
XFree86 10915 root 7u CHR 10,1 1016488
/dev/psaux
From console:
olegil:~# lsof | grep psaux
gpm 19563 root 0u CHR 10,1 1016488
/dev/psaux
Are you saying this isn't happening, and that my systems don't exist?
No. For all I know, several processes using /dev/psaux at the same time
doesn't work, and your output above doesn't show otherwise. Where does
gpm go when X is running?
It is running, it just releases control of /dev/psaux.
I'm not kidding here...
Here's from X:
olegil@olegil:~$ sudo lsof | grep gpm
gpm 19563 root cwd DIR 3,2 4096 2 /
gpm 19563 root rtd DIR 3,2 4096 2 /
gpm 19563 root txt REG 3,2 65112 1684485
/usr/sbin/gpm
gpm 19563 root mem REG 3,2 90048 443329
/lib/ld-2.3.2.so
gpm 19563 root mem REG 3,2 252272 441708
/lib/libncurses.so.5.4
gpm 19563 root mem REG 3,2 1244004 443339
/lib/libc-2.3.2.so
gpm 19563 root 1u unix 0xdb8cab00 4476181
/dev/gpmctl
And here's from console (X is still "running" in the background, I just
hop to console with ctrl-alt-F1):
gpm 19563 root cwd DIR 3,2 4096 2 /
gpm 19563 root rtd DIR 3,2 4096 2 /
gpm 19563 root txt REG 3,2 65112 1684485
/usr/sbin/gpm
gpm 19563 root mem REG 3,2 90048 443329
/lib/ld-2.3.2.so
gpm 19563 root mem REG 3,2 252272 441708
/lib/libncurses.so.5.4
gpm 19563 root mem REG 3,2 1244004 443339
/lib/libc-2.3.2.so
gpm 19563 root 0u CHR 10,1 1016488
/dev/psaux
gpm 19563 root 1u unix 0xdb8cab00 4476181
/dev/gpmctl
Seriously, gpm just gives up control when I go to tty7 (X), and X
releases control of it when I go someplace else. It might very well be
that /dev/psaux can only be read by one process, but I have never had a
problem with that, because the processes that read it ALWAYS work around
it... So my point isn't that /dev/psaux isn't braindead, but that _that
has never been a problem_ ;-)
Ole-Egil
--
AmigaOne dev list FAQ (when I say F, I mean F):
http://www.samfundet.no/~olegil/amiga/
Reply to: