Remove voodoo code from i2c
PATCH 1/1 Remove voodoo code from i2c
Hi,radeon_probe_i2c_connector contains very strange code that just send empty start-stop sequencies on the I2C bus. I copied the code from XFree86, but speaking with I2C gurus the general feeling is that this code can go away without any harm. I tested it on 3 different setup (all x86...) and it works well. I don't have any "old" monitor around to fiddle with though.
Oh, if you think that this code should stay I'll send you another patch to make it more readable.
===== drivers/video/aty/radeon_i2c.c 1.3 vs edited =====
drivers/video/aty/radeon_i2c.c +0 -50
@@ -180,59 +180,9 @@OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUTPUT |VGA_DDC_CLK_OUTPUT )); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); -for (i =0 ; i <3 ; i++) { -/* For some old monitors we need the -* following process to initialize/stop DDC -*/ -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); -wait_ms (13 ); - -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); -for (j =0 ; j <5 ; j++) { -wait_ms (10 ); -if (INREG (reg) &VGA_DDC_CLK_INPUT ) -break ; - } -if (j ==5 ) -continue ; - -OUTREG (reg,INREG (reg) |VGA_DDC_DATA_OUT_EN ); - (void )INREG (reg); -wait_ms (15 ); -OUTREG (reg,INREG (reg) |VGA_DDC_CLK_OUT_EN ); - (void )INREG (reg); -wait_ms (15 ); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); -wait_ms (15 ); -/* Do the real work */ edid =radeon_do_probe_i2c_edid (&rinfo->i2c[conn-1 ]); - -OUTREG (reg,INREG (reg) | - (VGA_DDC_DATA_OUT_EN |VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); -wait_ms (15 ); - -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); -for (j =0 ; j <10 ; j++) { -wait_ms (10 ); -if (INREG (reg) &VGA_DDC_CLK_INPUT ) -break ; - } - -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); -wait_ms (15 ); -OUTREG (reg,INREG (reg) | - (VGA_DDC_DATA_OUT_EN |VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg);if (edid)break ; }
Luca
Signature · 9 lines
Home: http://kronoz.cjb.net Runtime error 6D at f000:a12f : user incompetente ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
Quoted 10 lines · Kronos
Hi,
radeon_probe_i2c_connectorcontains very strange code that just send empty start-stop sequencies on the I2C bus. I copied the code from XFree86, but speaking with I2C gurus the general feeling is that this code can go away without any harm. I tested it on 3 different setup (all x86...) and it works well. I don't have any "old" monitor around to fiddle with though.Oh, if you think that this code should stay I'll send you another patch to make it more readable.
I'd rather keep the code as close to XFree as possible at this point..
BTW... I'm still looking for an EDID spec if that exist ;)
Ben.
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
Il Sat, Mar 06, 2004 at 10:55:14AM +1100, Benjamin Herrenschmidt ha scritto:
Quoted 9 lines · Kronos
Quoted 8 lines
Hi,
radeon_probe_i2c_connectorcontains very strange code that just send empty start-stop sequencies on the I2C bus. I copied the code from XFree86, but speaking with I2C gurus the general feeling is that this code can go away without any harm. I tested it on 3 different setup (all x86...) and it works well. I don't have any "old" monitor around to fiddle with though.
Quoted 4 lines
Quoted 2 lines
Oh, if you think that this code should stay I'll send you another patch to make it more readable.
I'd rather keep the code as close to XFree as possible at this point..
Ok, what about this (cosmetic) patch:
drivers/video/aty/radeon_i2c.c renamed from drivers/video/aty/radeon_i2c.c.orig +20 -26
@@ -174,65 +174,59 @@int radeon_probe_i2c_connector (struct radeonfb_info *rinfo,int conn,u8 **out_edid) {u32 reg = rinfo->i2c[conn-1 ].ddc_reg; +struct radeon_i2c_chan *chan = &rinfo->i2c[conn-1 ];u8 *edid =NULL ;int i, j;OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUTPUT |VGA_DDC_CLK_OUTPUT )); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setscl (chan,1 );for (i =0 ; i <3 ; i++) {/* For some old monitors we need the * following process to initialize/stop DDC */ -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,1 );wait_ms (13 ); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setscl (chan,1 );for (j =0 ; j <5 ; j++) { +/* Wait for ack */ wait_ms (10 ); -if (INREG (reg) &VGA_DDC_CLK_INPUT ) +if (radeon_gpio_getscl (chan))break ; }if (j ==5 )continue ; -OUTREG (reg,INREG (reg) |VGA_DDC_DATA_OUT_EN ); - (void )INREG (reg); +radeon_gpio_setsda (chan,0 );wait_ms (15 ); -OUTREG (reg,INREG (reg) |VGA_DDC_CLK_OUT_EN ); - (void )INREG (reg); +radeon_gpio_setscl (chan,0 );wait_ms (15 ); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,1 );wait_ms (15 );/* Do the real work */ - edid =radeon_do_probe_i2c_edid (&rinfo->i2c[conn-1 ]); + edid =radeon_do_probe_i2c_edid (chan); -OUTREG (reg,INREG (reg) | - (VGA_DDC_DATA_OUT_EN |VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,0 ); +radeon_gpio_setscl (chan,0 );wait_ms (15 ); - -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); + +radeon_gpio_setscl (chan,1 );for (j =0 ; j <10 ; j++) { +/* Wait for ack */ wait_ms (10 ); -if (INREG (reg) &VGA_DDC_CLK_INPUT ) +if (radeon_gpio_getscl (chan))break ; } -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,1 );wait_ms (15 ); -OUTREG (reg,INREG (reg) | - (VGA_DDC_DATA_OUT_EN |VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); + +radeon_gpio_setsda (chan,0 ); +radeon_gpio_setscl (chan,0 );if (edid)break ; }
In this way it clearer what's going on.
Luca
Signature · 10 lines
Home: http://kronoz.cjb.net "Sei l'unica donna della mia vita". (Adamo) ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
Can this be written as a generic i2c probe function by calling i2c chanel functions instead of directly calling radeon_gpio_setscl, etc? The i2c chanel functions will then in turn call radeon_gpio_setscl, etc.
I was under the impression that this code is used to wake up old style DDC chips in the monitor. Waking them up is a function of the monitor, not the radeon card. If that is right, this is a monitor problem and this code needs to be in every DDC driver, not just the radeon one. On the other hand I've tried every old monitor I have around here and none of them need this code.
Quoted 107 lines · --- Kronos
Il Sat, Mar 06, 2004 at 10:55:14AM +1100, Benjamin Herrenschmidt ha scritto:
Quoted 9 lines · Kronos
Quoted 8 lines
Hi,
radeon_probe_i2c_connectorcontains very strange code that just send empty start-stop sequencies on the I2C bus. I copied the code from XFree86, but speaking with I2C gurus the general feeling is that this code can go away without any harm. I tested it on 3 different setup (all x86...) and it works well. I don't have any "old" monitor around to fiddle with though.Quoted 4 lines
Quoted 2 lines
Oh, if you think that this code should stay I'll send you another patch to make it more readable.
I'd rather keep the code as close to XFree as possible at this point..
Ok, what about this (cosmetic) patch:
drivers/video/aty/radeon_i2c.c renamed from drivers/video/aty/radeon_i2c.c.orig +20
-26@@ -174,65 +174,59 @@int radeon_probe_i2c_connector (struct radeonfb_info *rinfo,int conn,u8 **out_edid) {u32 reg = rinfo->i2c[conn-1 ].ddc_reg; +struct radeon_i2c_chan *chan = &rinfo->i2c[conn-1 ];u8 *edid =NULL ;int i, j;OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUTPUT |VGA_DDC_CLK_OUTPUT )); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setscl (chan,1 );for (i =0 ; i <3 ; i++) {/* For some old monitors we need the * following process to initialize/stop DDC */ -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,1 );wait_ms (13 ); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setscl (chan,1 );for (j =0 ; j <5 ; j++) { +/* Wait for ack */ wait_ms (10 ); -if (INREG (reg) &VGA_DDC_CLK_INPUT ) +if (radeon_gpio_getscl (chan))break ; }if (j ==5 )continue ; -OUTREG (reg,INREG (reg) |VGA_DDC_DATA_OUT_EN ); - (void )INREG (reg); +radeon_gpio_setsda (chan,0 );wait_ms (15 ); -OUTREG (reg,INREG (reg) |VGA_DDC_CLK_OUT_EN ); - (void )INREG (reg); +radeon_gpio_setscl (chan,0 );wait_ms (15 ); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,1 );wait_ms (15 );/* Do the real work */ - edid =radeon_do_probe_i2c_edid (&rinfo->i2c[conn-1 ]); + edid =radeon_do_probe_i2c_edid (chan); -OUTREG (reg,INREG (reg) | - (VGA_DDC_DATA_OUT_EN |VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,0 ); +radeon_gpio_setscl (chan,0 );wait_ms (15 ); - -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); + +radeon_gpio_setscl (chan,1 );for (j =0 ; j <10 ; j++) { +/* Wait for ack */ wait_ms (10 ); -if (INREG (reg) &VGA_DDC_CLK_INPUT ) +if (radeon_gpio_getscl (chan))break ; } -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,1 );wait_ms (15 ); -OUTREG (reg,INREG (reg) | - (VGA_DDC_DATA_OUT_EN |VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); + +radeon_gpio_setsda (chan,0 ); +radeon_gpio_setscl (chan,0 );if (edid)break ;}
Quoted 18 lines
In this way it clearer what's going on.
Luca
Signature · 14 lines
Home: http://kronoz.cjb.net "Sei l'unica donna della mia vita". (Adamo) ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click _______________________________________________ Linux-fbdev-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linux-fbdev-devel
=====
Jon Smirl
[email protected]
__________________________________
Do you Yahoo!?
Yahoo! Search - Find what you’re looking for faster
http://search.yahoo.com
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
Quoted 3 lines · Jon Smirl
Can this be written as a generic i2c probe function by calling i2c chanel functions instead of directly calling
radeon_gpio_setscl, etc? The i2c chanel functions will then in turn callradeon_gpio_setscl, etc.
Why ? we aren't doing normal i2c big banging here, just hacking around.
Quoted 5 lines
I was under the impression that this code is used to wake up old style DDC chips in the monitor. Waking them up is a function of the monitor, not the radeon card. If that is right, this is a monitor problem and this code needs to be in every DDC driver, not just the radeon one. On the other hand I've tried every old monitor I have around here and none of them need this code.
Why the hell are you folks trying to change that perfectly working code?
We have much more important things to do at this point.
Ben.
Quoted 136 lines
Quoted 107 lines · --- Kronos
Il Sat, Mar 06, 2004 at 10:55:14AM +1100, Benjamin Herrenschmidt ha scritto:
Quoted 9 lines · Kronos
Quoted 8 lines
Hi,
radeon_probe_i2c_connectorcontains very strange code that just send empty start-stop sequencies on the I2C bus. I copied the code from XFree86, but speaking with I2C gurus the general feeling is that this code can go away without any harm. I tested it on 3 different setup (all x86...) and it works well. I don't have any "old" monitor around to fiddle with though.Quoted 4 lines
Quoted 2 lines
Oh, if you think that this code should stay I'll send you another patch to make it more readable.
I'd rather keep the code as close to XFree as possible at this point..
Ok, what about this (cosmetic) patch:
drivers/video/aty/radeon_i2c.c renamed from drivers/video/aty/radeon_i2c.c.orig +20
-26@@ -174,65 +174,59 @@int radeon_probe_i2c_connector (struct radeonfb_info *rinfo,int conn,u8 **out_edid) {u32 reg = rinfo->i2c[conn-1 ].ddc_reg; +struct radeon_i2c_chan *chan = &rinfo->i2c[conn-1 ];u8 *edid =NULL ;int i, j;OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUTPUT |VGA_DDC_CLK_OUTPUT )); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setscl (chan,1 );for (i =0 ; i <3 ; i++) {/* For some old monitors we need the * following process to initialize/stop DDC */ -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,1 );wait_ms (13 ); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setscl (chan,1 );for (j =0 ; j <5 ; j++) { +/* Wait for ack */ wait_ms (10 ); -if (INREG (reg) &VGA_DDC_CLK_INPUT ) +if (radeon_gpio_getscl (chan))break ; }if (j ==5 )continue ; -OUTREG (reg,INREG (reg) |VGA_DDC_DATA_OUT_EN ); - (void )INREG (reg); +radeon_gpio_setsda (chan,0 );wait_ms (15 ); -OUTREG (reg,INREG (reg) |VGA_DDC_CLK_OUT_EN ); - (void )INREG (reg); +radeon_gpio_setscl (chan,0 );wait_ms (15 ); -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,1 );wait_ms (15 );/* Do the real work */ - edid =radeon_do_probe_i2c_edid (&rinfo->i2c[conn-1 ]); + edid =radeon_do_probe_i2c_edid (chan); -OUTREG (reg,INREG (reg) | - (VGA_DDC_DATA_OUT_EN |VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,0 ); +radeon_gpio_setscl (chan,0 );wait_ms (15 ); - -OUTREG (reg,INREG (reg) & ~(VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); + +radeon_gpio_setscl (chan,1 );for (j =0 ; j <10 ; j++) { +/* Wait for ack */ wait_ms (10 ); -if (INREG (reg) &VGA_DDC_CLK_INPUT ) +if (radeon_gpio_getscl (chan))break ; } -OUTREG (reg,INREG (reg) & ~(VGA_DDC_DATA_OUT_EN )); - (void )INREG (reg); +radeon_gpio_setsda (chan,1 );wait_ms (15 ); -OUTREG (reg,INREG (reg) | - (VGA_DDC_DATA_OUT_EN |VGA_DDC_CLK_OUT_EN )); - (void )INREG (reg); + +radeon_gpio_setsda (chan,0 ); +radeon_gpio_setscl (chan,0 );if (edid)break ;}
Quoted 18 lines
In this way it clearer what's going on.
Luca
Signature · 14 lines
Home: http://kronoz.cjb.net "Sei l'unica donna della mia vita". (Adamo) ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click _______________________________________________ Linux-fbdev-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linux-fbdev-devel=====
Jon Smirl
[email protected]__________________________________
Do you Yahoo!?
Yahoo! Search - Find what youre looking for faster
http://search.yahoo.com
Signature · 9 lines
Benjamin Herrenschmidt <[email protected]> ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
Quoted 8 lines · --- Benjamin Herrenschmidt
Quoted 2 lines · Jon Smirl
Can this be written as a generic i2c probe function by calling i2c chanel functions instead of directly calling
radeon_gpio_setscl, etc? The i2cchanel
Quoted 1 line
functions will then in turn call
radeon_gpio_setscl, etc.Why ? we aren't doing normal i2c big banging here, just hacking around.
If I understand correctly, this code is needed to enable DDC on older monitors. If this is the case it is a generic problem and it needs to be corrected in all of the framebuffer drivers. Correcting it in all of the drivers is easier if the code is written in the generic i2c form.
Also, how do we know this is working? Nobody can come up with one of the monitors that needs this or even a spec describing it.
The discussion about this got started around a port of 2.4 ddcmon (i2c device driver like eeprom) to 2.6 and sysfs. The code in the radeon driver is not in ddcmon and the question is, should it be added? If it should be added it needs to be converted to the i2c generic form, not the radeon specific one.
=====
Jon Smirl
[email protected]
__________________________________
Do you Yahoo!?
Yahoo! Search - Find what you’re looking for faster
http://search.yahoo.com
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
Il Fri, Mar 05, 2004 at 06:10:16PM -0800, Jon Smirl ha scritto:
Quoted 3 lines
Can this be written as a generic i2c probe function by calling i2c chanel functions instead of directly calling
radeon_gpio_setscl, etc? The i2c chanel functions will then in turn callradeon_gpio_setscl, etc.
You can't use the i2c layer because you just send empty start-stop sequencies on the bus, no real data. You must use bit banging functions:
struct i2c_algo_bit_data *bits = adapter->algo_data;
Then you use ->{s,g}et{scl,sda}. Remeber to pass bits->data to those functions.
Luca
Signature · 10 lines
Home: http://kronoz.cjb.net
"It is more complicated than you think"
-- The Eighth Networking Truth from RFC 1925
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=clickQuoted 3 lines · Benjamin Herrenschmidt
BTW... I'm still looking for an EDID spec if that exist ;)
One exists but you have to pay for it. We have the spec.. What info do you need?
Signature · 11 lines
Richard A. Smith [email protected] ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
Quoted 8 lines · Richard Smith
Quoted 3 lines · Benjamin Herrenschmidt
BTW... I'm still looking for an EDID spec if that exist ;)
One exists but you have to pay for it. We have the spec.. What info do you need?
Thanks. I'll have those soon I suppose, I think I found out who within IBM is doing the VESA membership stuff.
Ben.
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
Quoted 7 lines · Kronos
Hi,
radeon_probe_i2c_connectorcontains very strange code that just send empty start-stop sequencies on the I2C bus. I copied the code from XFree86, but speaking with I2C gurus the general feeling is that this code can go away without any harm. I tested it on 3 different setup (all x86...) and it works well. I don't have any "old" monitor around to fiddle with though.
Have you asked Hui Yu, assuming the code originates from him? He might be able to explain what it's all about. There are many kinds of Radeon cards out there, some very weird...
Signature · 10 lines
Earthling Michel Dänzer | Debian (powerpc), X and DRI developer Libre software enthusiast | http://svcs.affero.net/rm.php?r=daenzer ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click