Re: ruby : A bit simpler keyboard adding for multiple locale X servers

[Aivils S. <Aiv...@un...>]

>> Here is dumb console driver created by me for X.
>
>Hm. I see dummycon has come back to tease me.

Now these is kids tested and mother aproved.
I couldn't find other way start "unlimited" count of X.
Of course current only two X is started - no more monitors.
Large tuxracing campaign. :))

>> Dumb console create VT and then do nothig.
>> All work do linuxconsole vt-manager.
>> You can configure dumb with kernel command line "dumbcon=x",
>> where x is number of dumb consoles.
>> Every console is binded with /dev/tty's depended from
>> MAX _NR_USER_CONSOLES.
>
>Hm. Go add about the number of VTs.
>

My coredistro.sourceforge.net has only /dev/tty1-tty10.
Therefore I use MAX_NR_USER_CONSOLES==4.

>> Sample:
>> file vt_kern.h
>> #define MAX_NR_USER_CONSOLES 4
>
>MAX_NR_USER_CONSOLES is how many VCs per VT device. This is always 16.
>MAX_NR_CONSOLES represents how many total VCs you can have. Dividing the
>two you see wwe can have at most 4 VTs.

So I have 4 VCs per VT and 4 /dev/tty's per VT.
Any have enough /dev/tty files can use different
MAX_NR_USER_CONSOLES definition.


>> This are tested by me with xdm, triple graphic login succesfully run.
>> Edit /etc/X11/xdm/Xservers .
>> Seems kdm, gdm don't like multiple X or wrong configuration on my end.
>> X patch is necessary.
>
>Can you post the patch :-)
>
I couldn't explain this patch.
Patch was posted by Mr. svetljo to linuxconsole mailing list.
May be he has different results. This work for me.
Seems X try to find all video cards and stop their PCI access.
We remove this feature ;-).
IMHO Complete patch must create different X0.lock file,
then X1 check X0.lock and do not disable X0 PCI access and vice versa.

Know bugs:
hardware GLX may be unstable , unsure about this , any other app can
hung up system.
halt sometimes will not stop one X - system hung up.
RedHat xfs will grab 30-60% CPU ~10min definitely slow down system - I stop
it.
RedHat fam will grab 99% CPU ~5min or more - I stop xinetd.

--- xc\programs\Xserver\hw\xfree86\common\xf86pciBus.c.original   Fri Nov
30 14:11:55 2001
+++ xc\programs\Xserver\hw\xfree86\common\xf86pciBus.c  Mon Apr 15 10:10:41
2002
@@ -584,9 +584,11 @@
 #ifdef DEBUG
     ErrorF("pciIoAccessDisable: 0x%05lx\n", *(PCITAG *)arg);
 #endif
+#if 0
     ((pciArg*)arg)->ctrl &= ~SETBITS;
     ((pciArg*)arg)->func(((pciArg*)arg)->tag, PCI_CMD_STAT_REG,
                ((pciArg*)arg)->ctrl);
+#endif
 }

 #undef SETBITS
@@ -608,9 +610,11 @@
 #ifdef DEBUG
     ErrorF("pciIo_MemAccessDisable: 0x%05lx\n", *(PCITAG *)arg);
 #endif
+#if 0
     ((pciArg*)arg)->ctrl &= ~SETBITS;
     ((pciArg*)arg)->func(((pciArg*)arg)->tag, PCI_CMD_STAT_REG,
                ((pciArg*)arg)->ctrl);
+#endif
 }

 #undef SETBITS
@@ -632,9 +636,11 @@
 #ifdef DEBUG
     ErrorF("pciMemAccessDisable: 0x%05lx\n", *(PCITAG *)arg);
 #endif
+#if 0
     ((pciArg*)arg)->ctrl &= ~SETBITS;
     ((pciArg*)arg)->func(((pciArg*)arg)->tag, PCI_CMD_STAT_REG,
                ((pciArg*)arg)->ctrl);
+#endif
 }
 #undef SETBITS


>> Q: Why linuxconsole vt-manager is better as separate keyboard driver for
X.
>> A: Linuxconsole project already have tons of keyboard drivers. All You
can
>> add to X
>>      via /dev/tty.
>
> Well XFree86 is currently based on using the VT. This will go away in the
> future. I already have spoken to Jim Getty about this and he already is
> working on making XFree86 independent of VT switching and to use the
input
> api!!! Yeah!!!!


Very far future ?