[PATCH] SMM BIOS on Dell i8100

[PATCH] SMM BIOS on Dell i8100

[Stephane Jourdois]

[-- Attachment #1: Type: text/plain, Size: 4196 bytes --]

Hello,

First, a very big thanx to Massimo for this great piece of code :-)
I've been trying to catch those events with no sucess for weeks.

I've got a Dell Inspiron 8100, which seems to differ slightly from
i8000. Here is a patch that fixes that. Please do not hesitate to ask me
to test some new code or anything on my laptop.

You should also replace your printk("string") with printk(KERN_INFO "string")


PS : Massimo, which version of setmixer do you use ? mine doesn't
support {+|-}val syntax... (debian sid, setmixer 27DEC94-4)

Thanks,



--- /home/kwisatz/i8kutils-1.1/i8k.c	Fri Nov  2 16:02:24 2001
+++ drivers/char/i8k.c	Mon Nov  5 01:06:31 2001
@@ -1,5 +1,5 @@
 /*
- * i8k.c -- Linux driver for accessing the SMM BIOS on Dell I8000 laptops
+ * i8k.c -- Linux driver for accessing the SMM BIOS on Dell I8x00 laptops
  *
  * Copyright (C) 2001  Massimo Dal Zotto <dz@debian.org>
  *
@@ -25,7 +25,7 @@
 #include <linux/i8k.h>
=20
 #define I8K_VERSION		"1.1 02/11/2001"
-#define I8K_BIOS_SIGNATURE	"Dell System Inspiron 8000"
+#define I8K_BIOS_SIGNATURE	"Dell System Inspiron 8"
 #define I8K_BIOS_SIGNATURE_ADDR 0x000ec000
 #define I8K_BIOS_VERSION_OFFSET	32
=20
@@ -40,10 +40,10 @@
 #define I8K_FAN_MULT		30
 #define I8K_MAX_TEMP		127
=20
-#define I8K_FN_NONE		0x08
-#define I8K_FN_UP		0x09
-#define I8K_FN_DOWN		0x0a
-#define I8K_FN_MUTE		0x0c
+#define I8K_FN_NONE		0x00
+#define I8K_FN_UP		0x01
+#define I8K_FN_DOWN		0x02
+#define I8K_FN_MUTE		0x04
=20
 #define I8K_POWER_AC		0x05
 #define I8K_POWER_BATTERY	0x01
@@ -56,10 +56,10 @@
 int force =3D 0;
=20
 MODULE_AUTHOR("Massimo Dal Zotto (dz@debian.org)");
-MODULE_DESCRIPTION("Driver for accessing the SMM BIOS on Dell I8000 laptop=
s");
+MODULE_DESCRIPTION("Driver for accessing the SMM BIOS on Dell I8x00 laptop=
s");
 MODULE_LICENSE("GPL");
 MODULE_PARM(force, "i");
-MODULE_PARM_DESC(force, "Force loading without checking for an Inspiron 80=
00");
+MODULE_PARM_DESC(force, "Force loading without checking for an Inspiron 8x=
00");
=20
 static ssize_t i8k_read(struct file *, char *, size_t, loff_t *);
 static int i8k_ioctl(struct inode *, struct file *, unsigned int,
@@ -157,7 +157,7 @@
 	return rc;
     }
=20
-    switch ((regs.eax & 0xff00) >> 8) {
+    switch ((regs.eax & 0x0700) >> 8) {
     case I8K_FN_UP:
 	return I8K_VOL_UP;
     case I8K_FN_DOWN:
@@ -439,7 +439,7 @@
 }
=20
 /*
- * Probe for the presence of an Inspiron I8000.
+ * Probe for the presence of an Inspiron I8x00.
  */
 static int i8k_probe(void)
 {
@@ -458,8 +458,9 @@
 	    printk("i8k: ioremap failed\n");
 	    return -ENODEV;
 	}
-	if (strncmp(buff,I8K_BIOS_SIGNATURE,sizeof(I8K_BIOS_SIGNATURE)) !=3D 0) {
-	    printk("i8k: Inspiron 8000 BIOS signature not found\n");
+	if (strncmp(buff,I8K_BIOS_SIGNATURE,sizeof(I8K_BIOS_SIGNATURE) - 1) !=3D =
0) {
+	    // -1 because we don't wan't to check for \0 at the end
+	    printk("i8k: Inspiron 8x00 BIOS signature not found\n");
 	    iounmap(buff);
 	    return -ENODEV;
 	}
@@ -475,7 +476,7 @@
 	}
 	for (p=3Dbuff; (p-buff)<(0x00100000-0x000c0000); p+=3D16) {
 	    if (strncmp(p,I8K_BIOS_SIGNATURE,sizeof(I8K_BIOS_SIGNATURE))=3D=3D0) {
-		printk("i8k: Inspiron 8000 BIOS signature found at %08x\n",
+		printk("i8k: Inspiron 8x00 BIOS signature found at %08x\n",
 		       0x000c0000+(p-buff));
 		break;
 	    }
@@ -518,7 +519,7 @@
 {
     struct proc_dir_entry *proc_i8k;
=20
-    /* Are we running on an Inspiron 8000 laptop? */
+    /* Are we running on an Inspiron 8x00 laptop? */
     if (i8k_probe() !=3D 0) {
 	return -ENODEV;
     }
@@ -532,7 +533,7 @@
     SET_MODULE_OWNER(proc_i8k);
=20
     printk(KERN_INFO
-	   "Inspiron 8000 SMM driver v%s Massimo Dal Zotto (dz@debian.org)\n",
+	   "Inspiron 8x00 SMM driver v%s Massimo Dal Zotto (dz@debian.org)\n",
 	   I8K_VERSION);
=20
     return 0;


-- 
 ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
(((    Ing?nieur d?veloppement 	\ /    AGAINST HTML MAIL    )))
 \\\   6, av. de la Belle Image	 X                         ///
  \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///

[-- Attachment #2: Type: application/pgp-signature, Size: 240 bytes --]

Re: [PATCH] SMM BIOS on Dell i8100

[Jeff Garzik]

Stephane Jourdois wrote:
> 
> Hello,
> 
> First, a very big thanx to Massimo for this great piece of code :-)
> I've been trying to catch those events with no sucess for weeks.
> 
> I've got a Dell Inspiron 8100, which seems to differ slightly from
> i8000. Here is a patch that fixes that. Please do not hesitate to ask me
> to test some new code or anything on my laptop.
> 
> You should also replace your printk("string") with printk(KERN_INFO "string")

Has this been tested in I8000?  You are changing a lot of magic numbers
in the code, and noone but you/Massimo know whether that is ok or not...

-- 
Jeff Garzik      | Only so many songs can be sung
Building 1024    | with two lips, two lungs, and one tongue.
MandrakeSoft     |         - nomeansno

Re: [PATCH] SMM BIOS on Dell i8100

[Stephane Jourdois]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=unknown-8bit, Size: 2142 bytes --]

On Mon, Nov 05, 2001 at 11:03:53AM -0500, Jeff Garzik wrote:
> Stephane Jourdois wrote:
> > I've got a Dell Inspiron 8100, which seems to differ slightly from
> > i8000. Here is a patch that fixes that. Please do not hesitate to ask me
> > to test some new code or anything on my laptop.
> Has this been tested in I8000?  You are changing a lot of magic numbers
> in the code, and noone but you/Massimo know whether that is ok or not...

I presume it works.
Let me explain why while looking at parts of my patch :

-#define I8K_FN_NONE		0x08
-#define I8K_FN_UP		0x09
-#define I8K_FN_DOWN		0x0a
-#define I8K_FN_MUTE		0x0c
+#define I8K_FN_NONE		0x00
+#define I8K_FN_UP		0x01
+#define I8K_FN_DOWN		0x02
+#define I8K_FN_MUTE		0x04

In fact, i8100 returns respectively 1, 2 and 4.
i8100 returns 0x09, 0x0a, and 0x0c (see Massimo's code).

Then we can take the 4 lighter bits, and test them without testing
the 4 greater bits which value are 0x08 on i8000 and 0x00 on i8100.

hence :
-    switch ((regs.eax & 0xff00) >> 8) {
+    switch ((regs.eax & 0x0700) >> 8) {

a different version (which I didn't choose), would be to keep Massimo's
defines, and test (((regs.eax | 0x0800 ) & 0xff00) >> 8). But that's
more operations, then not the better solution :-)

As the drivers disks provided by Dell (windows drivers of course) are
the same which works on i8000 and i8100, I suppose that they test only
the 4 lighter bits here


Of course, as Massismo's told me a few hours ago, this is a beta version
of this code, and it supports at least Massimo's i8000 and now my i8100.
We should wait a few days for feed-back from lklm readers, then
Massimo's intents are to implement a more modular code. That's a good
intent, isn't it ? :-)


Anyway, you'll agree with me that 2 laptops is better than one :-)


Thanks for your answer.
Stéphane.

-- 
 ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
(((    Ingénieur développement 	\ /    AGAINST HTML MAIL    )))
 \\\   6, av. de la Belle Image	 X                         ///
  \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///

[-- Attachment #2: Type: application/pgp-signature, Size: 240 bytes --]

Re: [PATCH] SMM BIOS on Dell i8100

[Juri Haberland]

In article <3BE6B869.D79E93B1@mandrakesoft.com> you wrote:
> Stephane Jourdois wrote:
>> 
>> Hello,
>> 
>> First, a very big thanx to Massimo for this great piece of code :-)
>> I've been trying to catch those events with no sucess for weeks.
>> 
>> I've got a Dell Inspiron 8100, which seems to differ slightly from
>> i8000. Here is a patch that fixes that. Please do not hesitate to ask me
>> to test some new code or anything on my laptop.
>> 
>> You should also replace your printk("string") with printk(KERN_INFO "string")
> 
> Has this been tested in I8000?  You are changing a lot of magic numbers
> in the code, and noone but you/Massimo know whether that is ok or not...

Actually, I just tried plain 2.4.14-pre8 and the i8k-module *didn't*
work with my i8000, but with the patch from Stephane it *does* ;)

Happy happy, joy joy...

Juri

PS: BIOS verion A17 if that matters

-- 
Juri Haberland  <juri@koschikode.com> 

Re: [PATCH] SMM BIOS on Dell i8100

[Stephane Jourdois]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=unknown-8bit, Size: 1298 bytes --]

On Tue, Nov 06, 2001 at 12:17:59AM +0100, Juri Haberland wrote:
> In article <3BE6B869.D79E93B1@mandrakesoft.com> you wrote:
> > Stephane Jourdois wrote:
> >> I've got a Dell Inspiron 8100, which seems to differ slightly from
> >> i8000. Here is a patch that fixes that. Please do not hesitate to ask me
> >> to test some new code or anything on my laptop.
> > Has this been tested in I8000?  You are changing a lot of magic numbers
> > in the code, and noone but you/Massimo know whether that is ok or not...
> Actually, I just tried plain 2.4.14-pre8 and the i8k-module *didn't*
> work with my i8000, but with the patch from Stephane it *does* ;)

> PS: BIOS verion A17 if that matters

mmm I have a i8100 with A04 BIOS...

But that confirms that multiple versions of SMM BIOS exists, and that we
have to test only 4 lighter bits, even on i8000.

Does anybody have a Dell Inspiron around there that still doesn't work ?
i8x00 should work, but perhaps others Inspirons have quite the same SMM
BIOS ?

Stephane.

-- 
 ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
(((    Ingénieur développement 	\ /    AGAINST HTML MAIL    )))
 \\\   6, av. de la Belle Image	 X                         ///
  \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///

[-- Attachment #2: Type: application/pgp-signature, Size: 240 bytes --]

Re: [PATCH] SMM BIOS on Dell i8100

[Massimo Dal Zotto]

> 
> Actually, I just tried plain 2.4.14-pre8 and the i8k-module *didn't*
> work with my i8000, but with the patch from Stephane it *does* ;)
> 
> Happy happy, joy joy...
> 
> Juri
> 
> PS: BIOS verion A17 if that matters
> 
> -- 
> Juri Haberland  <juri@koschikode.com> 
> 

Hi,

I have released version 1.2 of the driver. It contains Stephane's patches
for the I8100, a new i8kmon and some documentation. You can download from:

    http://www.debian.org/~dz/i8k/

Could you please explain what doesn't work with your I8000? Does the
module load? Can you read /proc/i8k?

-- 
Massimo Dal Zotto

+----------------------------------------------------------------------+
|  Massimo Dal Zotto               email: massimo.dalzotto@libero.it   |
|  Via Marconi, 141                phone: ++39-461534251               |
|  38057 Pergine Valsugana (TN)      www: http://www.cs.unitn.it/~dz/  |
|  Italy                                  http://www.debian.org/~dz/   |
|  gpg:   2DB65596  3CED BDC6 4F23 BEDA F489 2445 147F 1AEA 2DB6 5596  |
+----------------------------------------------------------------------+

Re: [PATCH] SMM BIOS on Dell i8100

[Juri Haberland]

In article <200111061645.RAA02115@fandango.cs.unitn.it> you wrote:
>> 
>> Actually, I just tried plain 2.4.14-pre8 and the i8k-module *didn't*
>> work with my i8000, but with the patch from Stephane it *does* ;)
>> 
>> Happy happy, joy joy...
>> 
>> Juri
>> 
>> PS: BIOS verion A17 if that matters

> Hi,
> 
> I have released version 1.2 of the driver. It contains Stephane's patches
> for the I8100, a new i8kmon and some documentation. You can download from:
> 
>     http://www.debian.org/~dz/i8k/
> 
> Could you please explain what doesn't work with your I8000? Does the
> module load? Can you read /proc/i8k?

Yes, sure, sorry. The module (from 2.4.14pre8) loaded perfectly and I could
read /proc/i8k. Also controlling the fans worked. The only thing that didn't
work was the key-thingy. Pressing any of the five keys (those two near the
power button and the Fn-keys) didn't do anything with the i8kbuttons script.

Btw Thank You Very Much!

Juri

-- 
Juri Haberland  <juri@koschikode.com> 

Re: [PATCH] SMM BIOS on Dell i8100

[Marcel J.E. Mol]

Massimo,
Here is a small patch to sort of implement autorepeat for the volume buttons.
It just adds a -r | --repeat <count> option to i8kbuttons. With the
default timeout and -r 1 I get a reasonable autorepeat function.
The name/docs might not be optimal but I leave that to you if you choose
to include the patch. It is agains i8kutils-1.2.

-Marcel
--- i8kbuttons.1.org	Tue Nov  6 22:30:56 2001
+++ i8kbuttons.1	Tue Nov  6 22:33:47 2001
@@ -36,6 +36,11 @@
 \fB\-t\fR, \fB\-\-timeout\fR <\fImilliseconds\fP>
 Specifies the interval in milliseconds at which the daemon checks the
 button status. Useful values are in the range of 50\-200. Default is 100.
+.TP 
+\fB\-r\fR, \fB\-\-repeat\fR <\fIcount\fP>
+Enables autorepeat (0 is off is default). While holding down one of the volume
+buttons, the volume is changed every count timeoutintervals. So with -t 100 -r 2
+it will change every 200 milliseconds.
 .LP 
 .SH "FILES"
 .LP 
--- i8kbuttons.c.org	Tue Nov  6 22:18:09 2001
+++ i8kbuttons.c	Tue Nov  6 22:24:08 2001
@@ -89,6 +89,7 @@
     printf("	-d|--down <command>\n");
     printf("	-m|--mute <command>\n");
     printf("	-t|--timeout <milliseconds>\n");
+    printf("	-r|--repeat <count>\n");
     printf("\n");
 }
 
@@ -97,6 +98,8 @@
 {
     int fd, status;
     int timeout = POLL_TIMEOUT;
+    int repeat = 0;
+    int autorpt = 0;
     int last = 0;
     char *progname;
     struct timespec req;
@@ -137,6 +140,11 @@
 	    }
 	    continue;
 	}
+	if ((strcmp(argv[0],"-r")==0) || (strcmp(argv[0],"--repeat")==0)) {
+	    if (argc < 2) break;
+	    autorpt = atoi(argv[1]);
+	    continue;
+	}
 	break;
     }
     if (argc > 0) {
@@ -162,8 +170,10 @@
 	DPRINTF("%d\n", status);
 
 	if (status == last) {
-	    continue;
+	    if (!autorpt || (repeat++ < autorpt))
+	        continue;
 	}
+        repeat = 0;
 	last = status;
 
 	switch (status) {



-- 
     ======--------         Marcel J.E. Mol                MESA Consulting B.V.
    =======---------        ph. +31-(0)6-54724868          P.O. Box 112
    =======---------        marcel@mesa.nl                 2630 AC  Nootdorp
__==== www.mesa.nl ---____U_n_i_x______I_n_t_e_r_n_e_t____ The Netherlands ____
 They couldn't think of a number,           Linux user 1148  --  counter.li.org
    so they gave me a name!  -- Rupert Hine  --  www.ruperthine.com

Re: [PATCH] SMM BIOS on Dell i8100

[Stephane Jourdois]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=unknown-8bit, Size: 1846 bytes --]

On Tue, Nov 06, 2001 at 05:45:13PM +0100, Massimo Dal Zotto wrote:
> I have released version 1.2 of the driver. It contains Stephane's patches
> for the I8100, a new i8kmon and some documentation. You can download from:
>     http://www.debian.org/~dz/i8k/

Hello,

Here is my patch to i8kutils-1.2 :

======
diff -u i8kutils-1.2.orig/i8kbuttons.c i8kutils-1.2/i8kbuttons.c
--- i8kutils-1.2.orig/i8kbuttons.c      Tue Nov  6 20:07:27 2001
+++ i8kutils-1.2/i8kbuttons.c   Tue Nov  6 20:11:19 2001
@@ -53,15 +53,7 @@

     DPRINTF("exec_cmd: %s\n", cmd);

-    if ((rc=fork()) < 0) {
-       perror("fork failed");
-       return;
-    }
-
-    if (rc == 0) {
-       execl("/bin/sh", "sh", "-c", cmd, NULL);
-       exit(0);
-    }
+    system(cmd);
 }
 
 static int
======

Without that, I get as much zombies processes as I have pressed the
volume buttons :-) I know system() is not great, but as security is not
a problem here...
I use debian/sid, and aumix as the mixer.


Now a fundamental question :
Does the load of the i8k module inhibits the fans start ? I can see my
processor temp increasing (I saw 80°C ...) without the fans start. Then
I started i8kmon to avoid an explosion. If the modules inhibits material
protections, then if that can be modified, it would be great ; if not,
i8kmon needs to get included in the kernel as a daemon. The i8kmon
should be a funny tool, not a system critical tool.

Also, I couldn't understand why sometimes the left fan is printed on red
color in i8kmon...


Thank you,

	Stephane


-- 
 ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
(((    Ingénieur développement 	\ /    AGAINST HTML MAIL    )))
 \\\   6, av. de la Belle Image	 X                         ///
  \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///

[-- Attachment #2: Type: application/pgp-signature, Size: 240 bytes --]

Re: [PATCH] SMM BIOS on Dell i8100

[Marcel J.E. Mol]

Maybe this is better:

diff -u i8kutils-1.2.orig/i8kbuttons.c i8kutils-1.2/i8kbuttons.c
--- i8kutils-1.2.orig/i8kbuttons.c      Tue Nov  6 20:07:27 2001
+++ i8kutils-1.2/i8kbuttons.c   Tue Nov  6 20:11:19 2001
@@ -53,15 +53,17 @@

     DPRINTF("exec_cmd: %s\n", cmd);

     if ((rc=fork()) < 0) {
        perror("fork failed");
        return;
     }
 
     if (rc == 0) {
        execl("/bin/sh", "sh", "-c", cmd, NULL);
        exit(0);
     }
+    else
+       wait(&rc);
 }
 
It get rid of the zombies and allows only one setmixer/auimix command to be
active at a time...


On Wed, Nov 07, 2001 at 10:44:05AM +0100, Stephane Jourdois wrote:
> On Tue, Nov 06, 2001 at 05:45:13PM +0100, Massimo Dal Zotto wrote:
> > I have released version 1.2 of the driver. It contains Stephane's patches
> > for the I8100, a new i8kmon and some documentation. You can download from:
> >     http://www.debian.org/~dz/i8k/
> 
> Hello,
> 
> Here is my patch to i8kutils-1.2 :
> 
> ======
> diff -u i8kutils-1.2.orig/i8kbuttons.c i8kutils-1.2/i8kbuttons.c
> --- i8kutils-1.2.orig/i8kbuttons.c      Tue Nov  6 20:07:27 2001
> +++ i8kutils-1.2/i8kbuttons.c   Tue Nov  6 20:11:19 2001
> @@ -53,15 +53,7 @@
> 
>      DPRINTF("exec_cmd: %s\n", cmd);
> 
> -    if ((rc=fork()) < 0) {
> -       perror("fork failed");
> -       return;
> -    }
> -
> -    if (rc == 0) {
> -       execl("/bin/sh", "sh", "-c", cmd, NULL);
> -       exit(0);
> -    }
> +    system(cmd);
>  }
>  
>  static int
> ======
> 
> Without that, I get as much zombies processes as I have pressed the
> volume buttons :-) I know system() is not great, but as security is not
> a problem here...
> I use debian/sid, and aumix as the mixer.
> 
> 
> Now a fundamental question :
> Does the load of the i8k module inhibits the fans start ? I can see my
> processor temp increasing (I saw 80°C ...) without the fans start. Then
> I started i8kmon to avoid an explosion. If the modules inhibits material
> protections, then if that can be modified, it would be great ; if not,
> i8kmon needs to get included in the kernel as a daemon. The i8kmon
> should be a funny tool, not a system critical tool.
> 
> Also, I couldn't understand why sometimes the left fan is printed on red
> color in i8kmon...
> 
> 
> Thank you,
> 
> 	Stephane
> 
> 
> -- 
>  ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
> (((    Ingénieur développement 	\ /    AGAINST HTML MAIL    )))
>  \\\   6, av. de la Belle Image	 X                         ///
>   \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///



-- 
     ======--------         Marcel J.E. Mol                MESA Consulting B.V.
    =======---------        ph. +31-(0)6-54724868          P.O. Box 112
    =======---------        marcel@mesa.nl                 2630 AC  Nootdorp
__==== www.mesa.nl ---____U_n_i_x______I_n_t_e_r_n_e_t____ The Netherlands ____
 They couldn't think of a number,           Linux user 1148  --  counter.li.org
    so they gave me a name!  -- Rupert Hine  --  www.ruperthine.com

Re: [PATCH] SMM BIOS on Dell i8100

[Stephane Jourdois]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=unknown-8bit, Size: 1303 bytes --]

On Wed, Nov 07, 2001 at 11:01:41AM +0100, Marcel J.E. Mol wrote:
> Maybe this is better:
> 
> diff -u i8kutils-1.2.orig/i8kbuttons.c i8kutils-1.2/i8kbuttons.c
> --- i8kutils-1.2.orig/i8kbuttons.c      Tue Nov  6 20:07:27 2001
> +++ i8kutils-1.2/i8kbuttons.c   Tue Nov  6 20:11:19 2001
> @@ -53,15 +53,17 @@
> 
>      DPRINTF("exec_cmd: %s\n", cmd);
> 
>      if ((rc=fork()) < 0) {
>         perror("fork failed");
>         return;
>      }
>  
>      if (rc == 0) {
>         execl("/bin/sh", "sh", "-c", cmd, NULL);
>         exit(0);
>      }
> +    else
> +       wait(&rc);
>  }
>  
> It get rid of the zombies and allows only one setmixer/auimix command
> to be active at a time...

Yup I did that on i8kutils-1.1, then reverted to system(), because I
removed the anti-repeat system, and used a slow mixer application. Then
the repeat was bad (sometimes slow, sometimes quick).
In fact, I don't really mind, as now I use aumix (and it is fast
enought).

Massimo, choose one please :-)

	Stephane


-- 
 ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
(((    Ingénieur développement 	\ /    AGAINST HTML MAIL    )))
 \\\   6, av. de la Belle Image	 X                         ///
  \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///

[-- Attachment #2: Type: application/pgp-signature, Size: 240 bytes --]

Re: [PATCH] SMM BIOS on Dell i8100

[Massimo Dal Zotto]

> 
> Yup I did that on i8kutils-1.1, then reverted to system(), because I
> removed the anti-repeat system, and used a slow mixer application. Then
> the repeat was bad (sometimes slow, sometimes quick).
> In fact, I don't really mind, as now I use aumix (and it is fast
> enought).
> 
> Massimo, choose one please :-)
> 
>         Stephane
> 

I have released version 1.3 of the package. It includes the autorepeat
feature and solves the zombie problem. You can download it from:

  http://www.debian.org/~dz/i8k/

There is also a new smm-test script which can be used to test the SMM bios
on Inspiron laptops. This command can be very dangerous, use with caution!

-- 
Massimo Dal Zotto

+----------------------------------------------------------------------+
|  Massimo Dal Zotto               email: massimo.dalzotto@libero.it   |
|  Via Marconi, 141                phone: ++39-461534251               |
|  38057 Pergine Valsugana (TN)      www: http://www.cs.unitn.it/~dz/  |
|  Italy                                  http://www.debian.org/~dz/   |
|  gpg:   2DB65596  3CED BDC6 4F23 BEDA F489 2445 147F 1AEA 2DB6 5596  |
+----------------------------------------------------------------------+

Re: [PATCH] SMM BIOS on Dell i8100

[Massimo Dal Zotto]

> Now a fundamental question :
> Does the load of the i8k module inhibits the fans start ? I can see my
> processor temp increasing (I saw 80°C ...) without the fans start. Then
> I started i8kmon to avoid an explosion. If the modules inhibits material
> protections, then if that can be modified, it would be great ; if not,
> i8kmon needs to get included in the kernel as a daemon. The i8kmon
> should be a funny tool, not a system critical tool.

The i8k module itself doesn't interfere with the BIOS fan control.
It just adds some ioctl to allow fan control from user-space. The
real problem is that the BIOS fan control is broken, as everybody
who own an I8K knows very well. If you run i8kmon with the --noauto
option you will just see the temperature and fan status as managed
by the BIOS. The i8kmon does the same job as the BIOS but uses lower
thresholds. The only difference is is that my program works.

> Also, I couldn't understand why sometimes the left fan is printed on red
> color in i8kmon...

If it becomes red for a short time it may be that the fan is slow at
turning on. If the button stays red for long time it means that your fan
is broken or stuck. Try starting it manually with a bent paper clip and
see if the red disappears. This is explained in the i8kmon manpage.

-- 
Massimo Dal Zotto

+----------------------------------------------------------------------+
|  Massimo Dal Zotto               email: massimo.dalzotto@libero.it   |
|  Via Marconi, 141                phone: ++39-461534251               |
|  38057 Pergine Valsugana (TN)      www: http://www.cs.unitn.it/~dz/  |
|  Italy                                  http://www.debian.org/~dz/   |
|  gpg:   2DB65596  3CED BDC6 4F23 BEDA F489 2445 147F 1AEA 2DB6 5596  |
+----------------------------------------------------------------------+

Re: [PATCH] SMM BIOS on Dell i8100

[David Woodhouse]

stephane@tuxfinder.org said:
>  Without that, I get as much zombies processes as I have pressed the
> volume buttons :-) I know system() is not great, but as security is
> not a problem here...

signal(SIGCHLD, SIG_IGN);


--
dwmw2

Re: [PATCH] SMM BIOS on Dell i8100

[Stephane Jourdois]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=unknown-8bit, Size: 3212 bytes --]

On Mon, Nov 05, 2001 at 01:09:54PM +0100, Marcel J.E. Mol wrote:
> I have an i8000 too. I tried some info to use the 'multimedia' buttons
> next to the power button. Do you know if these are also managed through
> smm?  Or are thre other ways/tools to access these buttons?
In fact, the two buttons next to the power button only are supported by
the SMM BIOS. The 4 others buttons send e0 keyboard scancodes, and are
fully supported by every kernel (Respectively e0 01, e0 02, e0 03, and
e0 04, from left to right).
Here is a patch to prove what I'm saying :


--- drivers/char/pc_keyb.c.orig      Mon Nov  5 17:41:54 2001
+++ drivers/char/pc_keyb.c   Mon Nov  5 17:45:56 2001
@@ -227,9 +227,18 @@
 #define E0_MSLW        125
 #define E0_MSRW        126
 #define E0_MSTM        127
+/*
+ * Multimedia Keys on Dell Inspiron i8x00
+ * (from left to right : play, stop, previous, next)
+ */
+#define E0_DELL_PLAY           120
+#define E0_DELL_STOP           121
+#define E0_DELL_PREVIOUS       122
+#define E0_DELL_NEXT           123

 static unsigned char e0_keys[128] = {
-  0, 0, 0, 0, 0, 0, 0, 0,                            /* 0x00-0x07 */
+  E0_DELL_PLAY, E0_DELL_STOP, E0_DELL_PREVIOUS, E0_DELL_NEXT, /*
0x00-0x03 */
+  0, 0, 0, 0,                                        /* 0x04-0x07 */
   0, 0, 0, 0, 0, 0, 0, 0,                            /* 0x08-0x0f */
   0, 0, 0, 0, 0, 0, 0, 0,                            /* 0x10-0x17 */
   0, 0, 0, 0, E0_KPENTER, E0_RCTRL, 0, 0,            /* 0x18-0x1f */



Then edit your /etc/console/bootime.kmap.gz (package console-common in
debian) to add for example :
----
keycode 120 = F70 F71
keycode 121 = Last_Console
keycode 122 = Decr_Console
keycode 123 = Incr_Console
string F70 = "Hello World !"
string F71 = "Dell Inspiron Powah !"
----


then issue the command /etc/init.d/keymap.sh (or loadkeys if you do not
use debian) *under console*, and you'll then believe me :-)


To support those keys under X, that's more complicated, because AFAIK we
have to create X keysyms before assigning them to keycodes.
but try that in your ~/.xmodmap :
----
keycode 129 = a b c d
keycode 130 = a b c d
keycode 131 = a b c d
keycode 132 = a b c d
----
then xmodmap ~/.xmodmap, and try <PLAY>, SHIFT-<PLAY>, ALT-<PLAY>,
CTRL-<PLAY>, etc.
(In fact, I really don't remember why I used 129-132 here... try with
120-123 if it doesn't work :-) I think I used those values to test if
keycodes are 8bits clean, they are not.)


Hope this helps, and if anybody knows how to implement keysyms, I'm
interested... as long as I don't have to patch XFree86 !

Please do not hesitate to contact me if you have any other questions.

To finish, SMM BIOS implements supoprt for the two big buttons next to
the power button (try to press them both at the same time to have mute
;-), and the three blue Fn-keys for sound (PgUp, PgDown, and End)


Stephane.

-- 
 ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
(((    Ingénieur développement 	\ /    AGAINST HTML MAIL    )))
 \\\   6, av. de la Belle Image	 X                         ///
  \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///

[-- Attachment #2: Type: application/pgp-signature, Size: 240 bytes --]

Re: [PATCH] SMM BIOS on Dell i8100

[Marcel J.E. Mol]

Hi Stephane,

Thanks for the info. I really believe you :-)
The X keycodes are indeed 129-132. I checked them using xev.
I checked /proc/i8k while pressing the buttons next to the
power button and they work fine. 
So what's left is to tie these buttons to some function so that
I can control the soundvolume/dvd player wherever I am i X...

-Marcel

On Mon, Nov 05, 2001 at 06:01:24PM +0100, Stephane Jourdois wrote:
> On Mon, Nov 05, 2001 at 01:09:54PM +0100, Marcel J.E. Mol wrote:
> > I have an i8000 too. I tried some info to use the 'multimedia' buttons
> > next to the power button. Do you know if these are also managed through
> > smm?  Or are thre other ways/tools to access these buttons?
> In fact, the two buttons next to the power button only are supported by
> the SMM BIOS. The 4 others buttons send e0 keyboard scancodes, and are
> fully supported by every kernel (Respectively e0 01, e0 02, e0 03, and
> e0 04, from left to right).
> Here is a patch to prove what I'm saying :
> 
> 
> --- drivers/char/pc_keyb.c.orig      Mon Nov  5 17:41:54 2001
> +++ drivers/char/pc_keyb.c   Mon Nov  5 17:45:56 2001
> @@ -227,9 +227,18 @@
>  #define E0_MSLW        125
>  #define E0_MSRW        126
>  #define E0_MSTM        127
> +/*
> + * Multimedia Keys on Dell Inspiron i8x00
> + * (from left to right : play, stop, previous, next)
> + */
> +#define E0_DELL_PLAY           120
> +#define E0_DELL_STOP           121
> +#define E0_DELL_PREVIOUS       122
> +#define E0_DELL_NEXT           123
> 
>  static unsigned char e0_keys[128] = {
> -  0, 0, 0, 0, 0, 0, 0, 0,                            /* 0x00-0x07 */
> +  E0_DELL_PLAY, E0_DELL_STOP, E0_DELL_PREVIOUS, E0_DELL_NEXT, /*
> 0x00-0x03 */
> +  0, 0, 0, 0,                                        /* 0x04-0x07 */
>    0, 0, 0, 0, 0, 0, 0, 0,                            /* 0x08-0x0f */
>    0, 0, 0, 0, 0, 0, 0, 0,                            /* 0x10-0x17 */
>    0, 0, 0, 0, E0_KPENTER, E0_RCTRL, 0, 0,            /* 0x18-0x1f */
> 
> 
> 
> Then edit your /etc/console/bootime.kmap.gz (package console-common in
> debian) to add for example :
> ----
> keycode 120 = F70 F71
> keycode 121 = Last_Console
> keycode 122 = Decr_Console
> keycode 123 = Incr_Console
> string F70 = "Hello World !"
> string F71 = "Dell Inspiron Powah !"
> ----
> 
> 
> then issue the command /etc/init.d/keymap.sh (or loadkeys if you do not
> use debian) *under console*, and you'll then believe me :-)
> 
> 
> To support those keys under X, that's more complicated, because AFAIK we
> have to create X keysyms before assigning them to keycodes.
> but try that in your ~/.xmodmap :
> ----
> keycode 129 = a b c d
> keycode 130 = a b c d
> keycode 131 = a b c d
> keycode 132 = a b c d
> ----
> then xmodmap ~/.xmodmap, and try <PLAY>, SHIFT-<PLAY>, ALT-<PLAY>,
> CTRL-<PLAY>, etc.
> (In fact, I really don't remember why I used 129-132 here... try with
> 120-123 if it doesn't work :-) I think I used those values to test if
> keycodes are 8bits clean, they are not.)
> 
> 
> Hope this helps, and if anybody knows how to implement keysyms, I'm
> interested... as long as I don't have to patch XFree86 !
> 
> Please do not hesitate to contact me if you have any other questions.
> 
> To finish, SMM BIOS implements supoprt for the two big buttons next to
> the power button (try to press them both at the same time to have mute
> ;-), and the three blue Fn-keys for sound (PgUp, PgDown, and End)
> 
> 
> Stephane.
> 
> -- 
>  ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
> (((    Ingénieur développement 	\ /    AGAINST HTML MAIL    )))
>  \\\   6, av. de la Belle Image	 X                         ///
>   \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///



-- 
     ======--------         Marcel J.E. Mol                MESA Consulting B.V.
    =======---------        ph. +31-(0)6-54724868          P.O. Box 112
    =======---------        marcel@mesa.nl                 2630 AC  Nootdorp
__==== www.mesa.nl ---____U_n_i_x______I_n_t_e_r_n_e_t____ The Netherlands ____
 They couldn't think of a number,           Linux user 1148  --  counter.li.org
    so they gave me a name!  -- Rupert Hine  --  www.ruperthine.com

Re: [PATCH] SMM BIOS on Dell i8100

[Stephane Jourdois]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=unknown-8bit, Size: 1931 bytes --]

On Mon, Nov 05, 2001 at 06:20:29PM +0100, Marcel J.E. Mol wrote:
> Thanks for the info. I really believe you :-)
> The X keycodes are indeed 129-132. I checked them using xev.
Yep, I did that just after mutt quitting. If someone could tell me
why those values... are e0 00-8f scancodes binded to 129-255 under
XFree86 ? (Not important, but just to increase my poor knowledge)

> I checked /proc/i8k while pressing the buttons next to the
> power button and they work fine. 
Did you use my patch ? I think that Jeff Garzik would appreciate
to have a proof that it still works on i8000 ;-)

> So what's left is to tie these buttons to some function so that
> I can control the soundvolume/dvd player wherever I am i X...
Then now you should download i8kutils-1.1 from :
	http://people.debian.org/~dz/i8k/
Marcello's utils are just what you are looking for now.
in fact, i8kbuttons.c will enable volume controll, i8kmon will
display a very usefull fans status (and *YES* ! you can start them
with a click :-)).

Just a little hint for i8kbuttons :
I suggest that you run :
./i8kbuttons --up "aumix -v +5" --down "aumix -v -5"

default bindings do not work, they are Massimo's aliases :-)


I use both tools since yesterday... and I watched a DVD just for
that occasion (I hate * and / in mplayer ;-)

Finally, as I told Massimo this morning, I just *love* to start my
fans just to here the little noise and see the proc temp decreasing.
(try a cat /dev/zero to increase quickly your cpu temp and enjoy !)


I think you'll appreciate those tools...
I can imagine your joy, I'm just thinking to me yesterday evening ;-)


Stephane.

-- 
 ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
(((    Ingénieur développement 	\ /    AGAINST HTML MAIL    )))
 \\\   6, av. de la Belle Image	 X                         ///
  \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///

[-- Attachment #2: Type: application/pgp-signature, Size: 240 bytes --]

Re: [PATCH] SMM BIOS on Dell i8100

[Marcel J.E. Mol]

On Mon, Nov 05, 2001 at 06:37:24PM +0100, Stephane Jourdois wrote:
> On Mon, Nov 05, 2001 at 06:20:29PM +0100, Marcel J.E. Mol wrote:
> > I checked /proc/i8k while pressing the buttons next to the
> > power button and they work fine. 
> Did you use my patch ? I think that Jeff Garzik would appreciate
> to have a proof that it still works on i8000 ;-)

Yup, just did it. I see no difference in funtionality with your
patch added.
 
> > So what's left is to tie these buttons to some function so that
> > I can control the soundvolume/dvd player wherever I am i X...
> Then now you should download i8kutils-1.1 from :
> 	http://people.debian.org/~dz/i8k/
> Marcello's utils are just what you are looking for now.
> in fact, i8kbuttons.c will enable volume controll, i8kmon will
> display a very usefull fans status (and *YES* ! you can start them
> with a click :-)).

Arg... I already downloaded this and made the module from it as i did
not want to compile the pre8 kernel yet.  (I only had to run the i8kmon
using 'wish i8kmon' as it complained about 'Tk'; this is latest Rawhide)

> Just a little hint for i8kbuttons :
> I suggest that you run :
> ./i8kbuttons --up "aumix -v +5" --down "aumix -v -5"

Great, Just add "--mute "aumix -v 0" to mute, but I do prefer the gmix
way that restores original volume level when the mutebutton is pressed
again but that probaly only works as gmix is running all the time...
Getting OT but does anyone know if you can sent events to gmix to do
volume control?
 
> Finally, as I told Massimo this morning, I just *love* to start my
> fans just to here the little noise and see the proc temp decreasing.
> (try a cat /dev/zero to increase quickly your cpu temp and enjoy !)
> I think you'll appreciate those tools...
> I can imagine your joy, I'm just thinking to me yesterday evening ;-)

I sure do! It is so nice to have a choice to do things quickly/easy with
the keyboard or with the mouse...

-Marcel
-- 
     ======--------         Marcel J.E. Mol                MESA Consulting B.V.
    =======---------        ph. +31-(0)6-54724868          P.O. Box 112
    =======---------        marcel@mesa.nl                 2630 AC  Nootdorp
__==== www.mesa.nl ---____U_n_i_x______I_n_t_e_r_n_e_t____ The Netherlands ____
 They couldn't think of a number,           Linux user 1148  --  counter.li.org
    so they gave me a name!  -- Rupert Hine  --  www.ruperthine.com

Re: [PATCH] SMM BIOS on Dell i8100

[Georg Nikodym]

On Mon, 2001-11-05 at 12:01, Stephane Jourdois wrote:

> Hope this helps, and if anybody knows how to implement keysyms, I'm
> interested... as long as I don't have to patch XFree86 !

No need.  Here's what I do with xmodmap:

!
! Dell 8000
! I have sawfish map these to
! xmms -u, -s, -r and -f respectively
!
keycode 129 = XF86AudioPlay
keycode 130 = XF86AudioStop
keycode 131 = XF86AudioPrev
keycode 132 = XF86AudioNext

Re: [PATCH] SMM BIOS on Dell i8100

[Stephane Jourdois]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=unknown-8bit, Size: 1862 bytes --]

On Mon, Nov 05, 2001 at 02:00:22PM -0500, Georg Nikodym wrote:
> On Mon, 2001-11-05 at 12:01, Stephane Jourdois wrote:
> > Hope this helps, and if anybody knows how to implement keysyms, I'm
> > interested... as long as I don't have to patch XFree86 !
> 
> No need.  Here's what I do with xmodmap:
> 
> !
> ! Dell 8000
> ! I have sawfish map these to
> ! xmms -u, -s, -r and -f respectively
> !
> keycode 129 = XF86AudioPlay
> keycode 130 = XF86AudioStop
> keycode 131 = XF86AudioPrev
> keycode 132 = XF86AudioNext

And that works indeed !

Try :
	[ `ps -C xmms | wc -l` = 7 ] && xmms --pause || xmms --play
instead of xmms -u. Then you really have Play/Pause...



For those who use enlightenment... here are the keybindings for it
(in your .enlightenment/keybindings.cfg, that you copyied from
/usr/share/enlightenment/config/) :

  __NEXT_ACTION
    __MODIFIER_KEY __NONE
    __KEY XF86AudioPlay
    __EVENT __KEY_PRESS
    __ACTION __A_EXEC /home/kwisatz/bin/xmms_play
  __NEXT_ACTION
    __MODIFIER_KEY __NONE
    __KEY XF86AudioStop
    __EVENT __KEY_PRESS
    __ACTION __A_EXEC xmms --stop
  __NEXT_ACTION
    __MODIFIER_KEY __NONE
    __KEY XF86AudioPrev
    __EVENT __KEY_PRESS
    __ACTION __A_EXEC xmms --rew
  __NEXT_ACTION
    __MODIFIER_KEY __NONE
    __KEY XF86AudioNext
    __EVENT __KEY_PRESS
    __ACTION __A_EXEC xmms --fwd

/home/kwisatz/bin/xmms_play is a onle-liner (Cf. above)


PS : I know this is OT, but as the only mean of those i8k buttons
is multimedia, I think it's great to learn how to bind them :-)

Stephane

-- 
 ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
(((    Ingénieur développement 	\ /    AGAINST HTML MAIL    )))
 \\\   6, av. de la Belle Image	 X                         ///
  \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///

[-- Attachment #2: Type: application/pgp-signature, Size: 240 bytes --]

Re: [PATCH] SMM BIOS on Dell i8100

[Massimo Dal Zotto]

Hi,

I have released version 1.4 of my package with a new kernel module and
some enhancements to the i8kmon utility.

It should now work on more Dell laptops (Inspiron and Latitude) and report
correctly the bios version and machine id. If it doesn't load try forcing
it with:

    insmod ./i8k.o force=1

Please test the new version and send me a report. Don't forget to include
the following information:

    laptop model
    bios version
    driver version
    i8kmon version
    kernel messages while loading the driver
    output of "cat /proc/i8k"

and a detailed description of any problem you have found.

-- 
Massimo Dal Zotto

+----------------------------------------------------------------------+
|  Massimo Dal Zotto               email: massimo.dalzotto@libero.it   |
|  Via Marconi, 141                phone: ++39-461534251               |
|  38057 Pergine Valsugana (TN)      www: http://www.cs.unitn.it/~dz/  |
|  Italy                                  http://www.debian.org/~dz/   |
|  gpg:   2DB65596  3CED BDC6 4F23 BEDA F489 2445 147F 1AEA 2DB6 5596  |
+----------------------------------------------------------------------+

Re: [PATCH] SMM BIOS on Dell i8100

[Stephane Jourdois]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=unknown-8bit, Size: 968 bytes --]

On Thu, Nov 08, 2001 at 07:14:03PM +0100, Massimo Dal Zotto wrote:
> I have released version 1.4 of my package with a new kernel module and
> some enhancements to the i8kmon utility.

hello,

[no patch today :-)]

here is what top gives me :
  PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
 8102 kwisatz   15   0  2612 2612  1804 S     1.3  0.5   0:01 tclsh
 1663 kwisatz   10   0  3172 2436  2028 S     0.1  0.4   0:25 gkrellm

It seems that gkrellm (which monitors more than 20 things on my laptop)
takes only 0.1% of cpu, whereas i8kmon takes 1.3%...
I have absolutely no idea how to improve that, but I think that should
be the first thing on the TODO :-)

++

-- 
 ///  Stephane Jourdois        	/"\  ASCII RIBBON CAMPAIGN \\\
(((    Ingénieur développement 	\ /    AGAINST HTML MAIL    )))
 \\\   6, av. de la Belle Image	 X                         ///
  \\\  94440 Marolles en Brie  	/ \    +33 6 8643 3085    ///

[-- Attachment #2: Type: application/pgp-signature, Size: 240 bytes --]

Re: [PATCH] SMM BIOS on Dell i8100

[Massimo Dal Zotto]

> On Thu, Nov 08, 2001 at 07:14:03PM +0100, Massimo Dal Zotto wrote:
> > I have released version 1.4 of my package with a new kernel module and
> > some enhancements to the i8kmon utility.
> 
> hello,
> 
> [no patch today :-)]
> 
> here is what top gives me :
>   PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
>  8102 kwisatz   15   0  2612 2612  1804 S     1.3  0.5   0:01 tclsh
>  1663 kwisatz   10   0  3172 2436  2028 S     0.1  0.4   0:25 gkrellm
> 
> It seems that gkrellm (which monitors more than 20 things on my laptop)
> takes only 0.1% of cpu, whereas i8kmon takes 1.3%...
> I have absolutely no idea how to improve that, but I think that should
> be the first thing on the TODO :-)
> 

Hi Stephane,

first question, why is your mailer using a charset=unknown-8bit instead
of the standard iso8859-1?

> Content-Type: text/plain; charset=unknown-8bit
> Content-Disposition: inline
> Content-Transfer-Encoding: quoted-printable

This gives an error in my editor elm and forces me to use cut-and-paste:

> [Charset unknown-8bit unsupported, skipping...]


Now back to i8kutils. No patch today because I was debugging exactly this
problem. I discovered that the SMM calls GET_CPU_TEMP and GET_POWER_STATUS
take a very long time compared to the other calls and this slows down the
whole program. Here is what I discovered:

    cpu_temp    = i8k_get_cpu_temp();			/* 11100 µs */
    left_fan    = i8k_get_fan_status(I8K_FAN_LEFT);	/*   580 µs */
    right_fan   = i8k_get_fan_status(I8K_FAN_RIGHT);	/*   580 µs */
    left_speed  = i8k_get_fan_speed(I8K_FAN_LEFT);	/*   580 µs */
    right_speed = i8k_get_fan_speed(I8K_FAN_RIGHT);	/*   580 µs */
    ac_power    = i8k_get_power_status();		/* 14700 µs */
    fn_key      = i8k_get_fn_status();			/*   750 µs */

Since I can't fix the SMM BIOS I think there is very little I can do,
except avoiding the GET_POWER_STATUS call which uses half of the time.
I will try to get the same information from /proc/apm.

Did you try the latest version (1.4)? Does it work on your I8100?

-- 
Massimo Dal Zotto

+----------------------------------------------------------------------+
|  Massimo Dal Zotto               email: massimo.dalzotto@libero.it   |
|  Via Marconi, 141                phone: ++39-461534251               |
|  38057 Pergine Valsugana (TN)      www: http://www.cs.unitn.it/~dz/  |
|  Italy                                  http://www.debian.org/~dz/   |
|  gpg:   2DB65596  3CED BDC6 4F23 BEDA F489 2445 147F 1AEA 2DB6 5596  |
+----------------------------------------------------------------------+

Re: [PATCH] SMM BIOS on Dell i8100

[Massimo Dal Zotto]

> 
> dz@cs.unitn.it said:
> >  Since I can't fix the SMM BIOS I think there is very little I can do,
> > except avoiding the GET_POWER_STATUS call which uses half of the time.
> > I will try to get the same information from /proc/apm.
> 
> APM will probably take the same amount of time, and indeed be the same code.
> It'd be best to work out how to talk to the stuff on the i2c bus directly
> from Linux.
> 
> --
> dwmw2

Yes, but since GET_POWER_STATUS is not really needed to control the fans
I plan to remove it from /proc/i8k and use /proc/apm which provides the
same information but can be read less frequently. Once per minute would
be fine if it is neede only to use different temperature thresholds.
This would save half of the cpu time spent in the SMM bios.
Also a higher timeout would help. Probably 5 seconds would be ok.

-- 
Massimo Dal Zotto

+----------------------------------------------------------------------+
|  Massimo Dal Zotto               email: massimo.dalzotto@libero.it   |
|  Via Marconi, 141                phone: ++39-461534251               |
|  38057 Pergine Valsugana (TN)      www: http://www.cs.unitn.it/~dz/  |
|  Italy                                  http://www.debian.org/~dz/   |
|  gpg:   2DB65596  3CED BDC6 4F23 BEDA F489 2445 147F 1AEA 2DB6 5596  |
+----------------------------------------------------------------------+