Skip to contentSkip to conversation

XFree86 seems to be being wrongly accused of doing the wrong thing

DISCUSSION Peter Williams 4 participants 10 messages Mar 18, 2004 – Mar 19, 2004

lore
Share & tools
lore
#0 lore raw

With 2.6.4 I'm getting the following messages very early in the boot long before XFree86 is started:

Mar 18 16:05:31 mudlark kernel: atkbd.c: Unknown key released (translated set 2, code 0x7a on isa0060/serio0).
Mar 18 16:05:31 mudlark kernel: atkbd.c: This is an XFree86 bug. It shouldn't access hardware directly.

They are repeated 6 times and are NOT the result of any keys being pressed or released.

Peter

Signature · 4 lines
Dr Peter Williams, Chief Scientist                [email protected]
Aurema Pty Limited                                Tel:+61 2 9698 2322
PO Box 305, Strawberry Hills NSW 2012, Australia  Fax:+61 2 9699 9174
79 Myrtle Street, Chippendale NSW 2008, Australia http://www.aurema.com
#1 lore raw

Wrongly ? I don't think so, as it has presumably been fixed in XFree86 after 4.4.0.

http://www.xfree86.org/cvs/changes.html:

6. Do the Linux KDKBDREP ioctl on the correct fd. This prevents the fallback that tries to directly program the keyboard repeat rate, and the related warning messages that recent Linux kernels generate (David Dawes).

I'm attaching the patch I extracted from CVS.

Vojtech, what about adding such information to your HOWTO ? And better, adding the URL to atkbd.c, so people stop reporting it.

Quoted 10 lines · Peter Williams

With 2.6.4 I'm getting the following messages very early in the boot long before XFree86 is started:

Mar 18 16:05:31 mudlark kernel: atkbd.c: Unknown key released
(translated set 2, code 0x7a on isa0060/serio0).
Mar 18 16:05:31 mudlark kernel: atkbd.c: This is an XFree86 bug. It

shouldn't access hardware directly.

They are repeated 6 times and are NOT the result of any keys being pressed or released.

Signature · 1 line
http://www.pervalidus.net/contact.html

Index: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c

programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c renamed from programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c:3.26 +5 -5
diff -u xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c:3.26 xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c:3.27
@@ -1,4 +1,4 @@
-/* $XFree86: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c,v 3.26 2003/11/17 22:20:41 dawes Exp $ */
+/* $XFree86: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c,v 3.27 2004/03/03 18:53:41 dawes Exp $ */
 /*
  * Copyright 1992 by Orest Zborowski <[email protected]>
  * Copyright 1993 by David Dawes <[email protected]>
@@ -81,7 +81,7 @@
 #endif
 
 static int
-KDKBDREP_ioctl_ok(int rate, int delay) {
+KDKBDREP_ioctl_ok(int fd, int rate, int delay) {
 #if defined(KDKBDREP) && !defined(__sparc__)
      /* This ioctl is defined in <linux/kd.h> but is not
 	implemented anywhere - must be in some m68k patches. */
@@ -90,7 +90,7 @@
    /* don't change, just test */
    kbdrep_s.rate = -1;
    kbdrep_s.delay = -1;
-   if (ioctl( 0, KDKBDREP, &kbdrep_s )) {
+   if (ioctl( fd, KDKBDREP, &kbdrep_s )) {
        return 0;
    }
 
@@ -105,7 +105,7 @@
    if (kbdrep_s.delay < 1)
      kbdrep_s.delay = 1;
    
-   if (ioctl( 0, KDKBDREP, &kbdrep_s )) {
+   if (ioctl( fd, KDKBDREP, &kbdrep_s )) {
      return 0;
    }
 
@@ -174,7 +174,7 @@
     delay = xf86Info.kbdDelay;
 
 
-  if(KDKBDREP_ioctl_ok(rate, delay)) 	/* m68k? */
+  if(KDKBDREP_ioctl_ok(xf86Info.consoleFd, rate, delay)) 	/* m68k? */
     return;
 
   if(KIOCSRATE_ioctl_ok(rate, delay))	/* sparc? */

Index: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c

programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c renamed from programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c:1.5 +5 -5
diff -u xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c:1.5 xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c:1.6
@@ -1,4 +1,4 @@
-/* $XFree86: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c,v 1.5 2003/11/04 03:14:39 tsi Exp $ */
+/* $XFree86: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c,v 1.6 2004/03/03 18:53:41 dawes Exp $ */
 
 /*
  * Copyright (c) 2002 by The XFree86 Project, Inc.
@@ -108,7 +108,7 @@
 #endif
 
 static int
-KDKBDREP_ioctl_ok(int rate, int delay) {
+KDKBDREP_ioctl_ok(int fd, int rate, int delay) {
 #if defined(KDKBDREP) && !defined(__sparc__)
      /* This ioctl is defined in <linux/kd.h> but is not
 	implemented anywhere - must be in some m68k patches. */
@@ -117,7 +117,7 @@
    /* don't change, just test */
    kbdrep_s.rate = -1;
    kbdrep_s.delay = -1;
-   if (ioctl( 0, KDKBDREP, &kbdrep_s )) {
+   if (ioctl( fd, KDKBDREP, &kbdrep_s )) {
        return 0;
    }
 
@@ -132,7 +132,7 @@
    if (kbdrep_s.delay < 1)
      kbdrep_s.delay = 1;
    
-   if (ioctl( 0, KDKBDREP, &kbdrep_s )) {
+   if (ioctl( fd, KDKBDREP, &kbdrep_s )) {
      return 0;
    }
 
@@ -200,7 +200,7 @@
   if (pKbd->delay >= 0)
     delay = pKbd->delay;
 
-  if(KDKBDREP_ioctl_ok(rate, delay)) 	/* m68k? */
+  if(KDKBDREP_ioctl_ok(pInfo->fd, rate, delay)) 	/* m68k? */
     return;
 
   if(KIOCSRATE_ioctl_ok(rate, delay))	/* sparc? */
#2 lore raw
Quoted 1 line · Frédéric L. W. Meunier

Wrongly ?

Yes, wrongly. XFree86 wasn't even running when the messages appeared so there's no way that it could be to blame. Also no keys had been pressed or released.

Peter

Signature · 4 lines
Dr Peter Williams, Chief Scientist                [email protected]
Aurema Pty Limited                                Tel:+61 2 9698 2322
PO Box 305, Strawberry Hills NSW 2012, Australia  Fax:+61 2 9699 9174
79 Myrtle Street, Chippendale NSW 2008, Australia http://www.aurema.com
#3 lore raw
Quoted 6 lines · Peter Williams
Quoted 1 line · Frédéric L. W. Meunier

Wrongly ?

Yes, wrongly. XFree86 wasn't even running when the messages appeared so there's no way that it could be to blame. Also no keys had been pressed or released.

XFree86 wasn't the only one doing that. Old versions of 'kbdrate' have did direct HW accesses, too.

Signature · 2 lines
Vojtech Pavlik
SuSE Labs, SuSE CR
#4 lore raw
In reply to Peter Williams
Quoted 6 lines · Peter Williams
Quoted 1 line · Frédéric L. W. Meunier

Wrongly ?

Yes, wrongly. XFree86 wasn't even running when the messages appeared so there's no way that it could be to blame. Also no keys had been pressed or released.

I have a machine here I see that message on before the init scripts even load.

Gerhard
--
Gerhard Mack

[email protected]

<>< As a computer I find your faith in technology amusing.

#5 lore raw
Quoted 11 lines · Gerhard Mack
Quoted 6 lines · Peter Williams
Quoted 1 line · Frédéric L. W. Meunier

Wrongly ?

Yes, wrongly. XFree86 wasn't even running when the messages appeared so there's no way that it could be to blame. Also no keys had been pressed or released.

I have a machine here I see that message on before the init scripts even load.

Quick question: Does it go away if you compile USB support into the kernel statically?

Signature · 2 lines
Vojtech Pavlik
SuSE Labs, SuSE CR
#6 lore raw
Quoted 19 lines · Vojtech Pavlik

Date: Fri, 19 Mar 2004 20:03:55 +0100
From: Vojtech Pavlik <[email protected]>
To: Gerhard Mack <[email protected]>
Cc: Peter Williams <[email protected]>,

     "[iso-8859-1] \"Frédéric L. W. Meunier\"" <[email protected]>,
     [email protected]

Subject: Re: XFree86 seems to be being wrongly accused of doing the wrong
thing

Quoted 9 lines · Gerhard Mack
Quoted 6 lines · Peter Williams
Quoted 1 line · Frédéric L. W. Meunier

Wrongly ?

Yes, wrongly. XFree86 wasn't even running when the messages appeared so there's no way that it could be to blame. Also no keys had been pressed or released.

Quoted 6 lines
Quoted 2 lines

I have a machine here I see that message on before the init scripts even load.

Quick question: Does it go away if you compile USB support into the kernel statically?

I have USB compiled in and it's still there. I should also note that the keyboard is plugged into the ps/2 port.

The mouse on the other hand is plugged into the USB port.

Gerhard

--
Gerhard Mack

[email protected]

<>< As a computer I find your faith in technology amusing.

#7 lore raw
Quoted 24 lines · Gerhard Mack
Quoted 22 lines · Vojtech Pavlik

Date: Fri, 19 Mar 2004 20:03:55 +0100
From: Vojtech Pavlik <[email protected]>
To: Gerhard Mack <[email protected]>
Cc: Peter Williams <[email protected]>,

     "[iso-8859-1] \"Frédéric L. W. Meunier\"" <[email protected]>,
     [email protected]

Subject: Re: XFree86 seems to be being wrongly accused of doing the wrong
thing

Quoted 11 lines · Gerhard Mack
Quoted 6 lines · Peter Williams
Quoted 1 line · Frédéric L. W. Meunier

Wrongly ?

Yes, wrongly. XFree86 wasn't even running when the messages appeared so there's no way that it could be to blame. Also no keys had been pressed or released.

I have a machine here I see that message on before the init scripts even load.

Quoted 6 lines
Quoted 3 lines

Quick question: Does it go away if you compile USB support into the kernel statically?

I have USB compiled in and it's still there. I should also note that the keyboard is plugged into the ps/2 port.

Yes, if you had the keyboard in USB, then i8042 wouldn't need to be used at all.

Quoted 1 line

The mouse on the other hand is plugged into the USB port.

Signature · 2 lines
Vojtech Pavlik
SuSE Labs, SuSE CR
#8 lore raw
In reply to Frédéric L. W. Meunier
Quoted 10 lines · Frédéric L. W. Meunier

Wrongly ? I don't think so, as it has presumably been fixed in XFree86 after 4.4.0.

http://www.xfree86.org/cvs/changes.html:

6. Do the Linux KDKBDREP ioctl on the correct fd. This prevents the fallback that tries to directly program the keyboard repeat rate, and the related warning messages that recent Linux kernels generate (David Dawes).

Quoted 4 lines

I'm attaching the patch I extracted from CVS.

Vojtech, what about adding such information to your HOWTO ? And better, adding the URL to atkbd.c, so people stop reporting it.

I'll add the info and the URL into the HOWTO and kill the message.

Quoted 107 lines · Peter Williams
Quoted 11 lines

With 2.6.4 I'm getting the following messages very early in the boot long before XFree86 is started:

Mar 18 16:05:31 mudlark kernel: atkbd.c: Unknown key released
(translated set 2, code 0x7a on isa0060/serio0).
Mar 18 16:05:31 mudlark kernel: atkbd.c: This is an XFree86 bug. It

shouldn't access hardware directly.

They are repeated 6 times and are NOT the result of any keys being pressed or released.

Signature · 3 lines
http://www.pervalidus.net/contact.html
Index: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c
diff -u xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c:3.26 xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c:3.27
programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c renamed from programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c:3.26 +5 -5
@@ -1,4 +1,4 @@
-/* $XFree86: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c,v 3.26 2003/11/17 22:20:41 dawes Exp $ */
+/* $XFree86: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c,v 3.27 2004/03/03 18:53:41 dawes Exp $ */
 /*
  * Copyright 1992 by Orest Zborowski <[email protected]>
  * Copyright 1993 by David Dawes <[email protected]>
@@ -81,7 +81,7 @@
 #endif
 
 static int
-KDKBDREP_ioctl_ok(int rate, int delay) {
+KDKBDREP_ioctl_ok(int fd, int rate, int delay) {
 #if defined(KDKBDREP) && !defined(__sparc__)
      /* This ioctl is defined in <linux/kd.h> but is not
 	implemented anywhere - must be in some m68k patches. */
@@ -90,7 +90,7 @@
    /* don't change, just test */
    kbdrep_s.rate = -1;
    kbdrep_s.delay = -1;
-   if (ioctl( 0, KDKBDREP, &kbdrep_s )) {
+   if (ioctl( fd, KDKBDREP, &kbdrep_s )) {
        return 0;
    }
 
@@ -105,7 +105,7 @@
    if (kbdrep_s.delay < 1)
      kbdrep_s.delay = 1;
 
-   if (ioctl( 0, KDKBDREP, &kbdrep_s )) {
+   if (ioctl( fd, KDKBDREP, &kbdrep_s )) {
      return 0;
    }
 
@@ -174,7 +174,7 @@
     delay = xf86Info.kbdDelay;
 
 
-  if(KDKBDREP_ioctl_ok(rate, delay)) 	/* m68k? */
+  if(KDKBDREP_ioctl_ok(xf86Info.consoleFd, rate, delay)) 	/* m68k? */
     return;
 
   if(KIOCSRATE_ioctl_ok(rate, delay))	/* sparc? */

Index: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c
diff -u xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c:1.5 xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c:1.6

programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c renamed from programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c:1.5 +5 -5
@@ -1,4 +1,4 @@
-/* $XFree86: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c,v 1.5 2003/11/04 03:14:39 tsi Exp $ */
+/* $XFree86: xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c,v 1.6 2004/03/03 18:53:41 dawes Exp $ */
 
 /*
  * Copyright (c) 2002 by The XFree86 Project, Inc.
@@ -108,7 +108,7 @@
 #endif
 
 static int
-KDKBDREP_ioctl_ok(int rate, int delay) {
+KDKBDREP_ioctl_ok(int fd, int rate, int delay) {
 #if defined(KDKBDREP) && !defined(__sparc__)
      /* This ioctl is defined in <linux/kd.h> but is not
 	implemented anywhere - must be in some m68k patches. */
@@ -117,7 +117,7 @@
    /* don't change, just test */
    kbdrep_s.rate = -1;
    kbdrep_s.delay = -1;
-   if (ioctl( 0, KDKBDREP, &kbdrep_s )) {
+   if (ioctl( fd, KDKBDREP, &kbdrep_s )) {
        return 0;
    }
 
@@ -132,7 +132,7 @@
    if (kbdrep_s.delay < 1)
      kbdrep_s.delay = 1;
 
-   if (ioctl( 0, KDKBDREP, &kbdrep_s )) {
+   if (ioctl( fd, KDKBDREP, &kbdrep_s )) {
      return 0;
    }
 
@@ -200,7 +200,7 @@
   if (pKbd->delay >= 0)
     delay = pKbd->delay;
 
-  if(KDKBDREP_ioctl_ok(rate, delay)) 	/* m68k? */
+  if(KDKBDREP_ioctl_ok(pInfo->fd, rate, delay)) 	/* m68k? */
     return;
 
   if(KIOCSRATE_ioctl_ok(rate, delay))	/* sparc? */
Signature · 2 lines
Vojtech Pavlik
SuSE Labs, SuSE CR
#9 lore raw
Quoted 12 lines · Vojtech Pavlik
Quoted 10 lines · Frédéric L. W. Meunier

Wrongly ? I don't think so, as it has presumably been fixed in XFree86 after 4.4.0.

http://www.xfree86.org/cvs/changes.html:

6. Do the Linux KDKBDREP ioctl on the correct fd. This prevents the fallback that tries to directly program the keyboard repeat rate, and the related warning messages that recent Linux kernels generate (David Dawes).

Quoted 6 lines
Quoted 4 lines

I'm attaching the patch I extracted from CVS.

Vojtech, what about adding such information to your HOWTO ? And better, adding the URL to atkbd.c, so people stop reporting it.

I'll add the info and the URL into the HOWTO and kill the message.

Here are the URLs directly from CVS (just in case someone doesn't trust the diff I sent):

http://cvsweb.xfree86.org/cvsweb/xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c.diff?r1=3.26&r2=3.27
http://cvsweb.xfree86.org/cvsweb/xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c.diff?r1=1.5&r2=1.6

Signature · 1 line
http://www.pervalidus.net/contact.html