radeonfb blanks my monitor

radeonfb blanks my monitor

[Frédéric L. W. Meunier]

I just replaced my Matrox G400 with a Jetway Radeon 9600LE 
(256Mb). If I run 'modprobe radeonfb', the monitor blanks out 
and the power on light keeps flashing.

What may be wrong ? Using 2.6.11.

-- 
How to contact me - http://www.pervalidus.net/contact.html

Re: radeonfb blanks my monitor

[Benjamin Herrenschmidt]

On Wed, 2005-03-02 at 23:51 -0300, Frédéric L. W. Meunier wrote:
> I just replaced my Matrox G400 with a Jetway Radeon 9600LE 
> (256Mb). If I run 'modprobe radeonfb', the monitor blanks out 
> and the power on light keeps flashing.
> 
> What may be wrong ? Using 2.6.11.

Do you have a way to capture the dmesg log produced ?

Also, does it work if radeonfb is built-in ?

Re: radeonfb blanks my monitor

[Frédéric L. W. Meunier]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 1161 bytes --]

On Thu, 3 Mar 2005, Benjamin Herrenschmidt wrote:

> On Wed, 2005-03-02 at 23:51 -0300, Frédéric L. W. Meunier wrote:
>> I just replaced my Matrox G400 with a Jetway Radeon 9600LE
>> (256Mb). If I run 'modprobe radeonfb', the monitor blanks out
>> and the power on light keeps flashing.
>>
>> What may be wrong ? Using 2.6.11.
>
> Do you have a way to capture the dmesg log produced ?

These are the lines before I have to use SysRq.

Mar  2 15:16:45 darkstar kernel: radeonfb: Found Intel x86 BIOS ROM Image
Mar  2 15:16:45 darkstar kernel: radeonfb: Retreived PLL infos from BIOS
Mar  2 15:16:45 darkstar kernel: i2c-algo-bit.o: dvi passed test.
Mar  2 15:16:45 darkstar kernel: i2c-algo-bit.o: vga passed test.
Mar  2 15:16:46 darkstar kernel: radeonfb: Monitor 1 type CRT found
Mar  2 15:16:46 darkstar kernel: radeonfb: EDID probed
Mar  2 15:16:46 darkstar kernel: radeonfb: Monitor 2 type no found

BTW, I don't know if it could be related, but my motherboard 
only supports AGP 4x.

> Also, does it work if radeonfb is built-in ?

I'll try later. Time to sleep.

-- 
How to contact me - http://www.pervalidus.net/contact.html

Re: radeonfb blanks my monitor

[Benjamin Herrenschmidt]

On Thu, 2005-03-03 at 01:39 -0300, Frédéric L. W. Meunier wrote:
> On Thu, 3 Mar 2005, Benjamin Herrenschmidt wrote:
> 
> > On Wed, 2005-03-02 at 23:51 -0300, Frédéric L. W. Meunier wrote:
> >> I just replaced my Matrox G400 with a Jetway Radeon 9600LE
> >> (256Mb). If I run 'modprobe radeonfb', the monitor blanks out
> >> and the power on light keeps flashing.
> >>
> >> What may be wrong ? Using 2.6.11.
> >
> > Do you have a way to capture the dmesg log produced ?
> 
> These are the lines before I have to use SysRq.
> 
> Mar  2 15:16:45 darkstar kernel: radeonfb: Found Intel x86 BIOS ROM Image
> Mar  2 15:16:45 darkstar kernel: radeonfb: Retreived PLL infos from BIOS
> Mar  2 15:16:45 darkstar kernel: i2c-algo-bit.o: dvi passed test.
> Mar  2 15:16:45 darkstar kernel: i2c-algo-bit.o: vga passed test.
> Mar  2 15:16:46 darkstar kernel: radeonfb: Monitor 1 type CRT found
> Mar  2 15:16:46 darkstar kernel: radeonfb: EDID probed
> Mar  2 15:16:46 darkstar kernel: radeonfb: Monitor 2 type no found
> 
> BTW, I don't know if it could be related, but my motherboard 
> only supports AGP 4x

There should be more than these... Does it continue booting afte the
screen goes blank or not at all ? Can you send the full dmesg log too ?
Also, enable radeonfb verbose debug in the config.

> > Also, does it work if radeonfb is built-in ?
> 
> I'll try later. Time to sleep.

Ok, thanks.

Re: radeonfb blanks my monitor

[Frédéric L. W. Meunier]

On Thu, 3 Mar 2005, Benjamin Herrenschmidt wrote:

> There should be more than these... Does it continue booting 
> afte the screen goes blank or not at all ? Can you send the 
> full dmesg log too ? Also, enable radeonfb verbose debug in 
> the config.

Yes, there were more in /var/log/syslog:

Mar  2 15:16:45 darkstar kernel: radeonfb: Reference=27.00 MHz (RefDiv=12) Memory=325.00 Mhz, System=200.00 MHz
Mar  2 15:16:45 darkstar kernel: radeonfb: PLL min 20000 max 40000
Mar  2 15:16:45 darkstar kernel: i2c-algo-bit.o: monid seems to be busy.
Mar  2 15:16:45 darkstar kernel: radeonfb 0000:01:00.0: Failed to register I2C bus monid.
Mar  2 15:16:45 darkstar kernel: i2c-algo-bit.o: crt2 seems to be busy.
Mar  2 15:16:46 darkstar kernel: radeonfb 0000:01:00.0: Failed to register I2C bus crt2.
Mar  2 15:16:46 darkstar kernel: Console: switching to colour frame buffer device 90x25
Mar  2 15:16:46 darkstar kernel: radeonfb (0000:01:00.0): ATI Radeon AP

Do the "seems to be busy." and/or "Failed to register I2C bus" 
indicate a problem ?

There's nothing about radeonfb in dmesg because I manually 
loaded the modules.

I now compiled built-in with debug enabled and got the same 
problem. Nothing got logged. Everything seems to stop when it 
blanks, but SysRq works.

-- 
How to contact me - http://www.pervalidus.net/contact.html

Re: radeonfb blanks my monitor

[Benjamin Herrenschmidt]

On Thu, 2005-03-03 at 12:38 -0300, Frédéric L. W. Meunier wrote:
> On Thu, 3 Mar 2005, Benjamin Herrenschmidt wrote:
> 
> > There should be more than these... Does it continue booting 
> > afte the screen goes blank or not at all ? Can you send the 
> > full dmesg log too ? Also, enable radeonfb verbose debug in 
> > the config.
> 
> Yes, there were more in /var/log/syslog:
> 
> Mar  2 15:16:45 darkstar kernel: radeonfb: Reference=27.00 MHz (RefDiv=12) Memory=325.00 Mhz, System=200.00 MHz
> Mar  2 15:16:45 darkstar kernel: radeonfb: PLL min 20000 max 40000
> Mar  2 15:16:45 darkstar kernel: i2c-algo-bit.o: monid seems to be busy.
> Mar  2 15:16:45 darkstar kernel: radeonfb 0000:01:00.0: Failed to register I2C bus monid.
> Mar  2 15:16:45 darkstar kernel: i2c-algo-bit.o: crt2 seems to be busy.
> Mar  2 15:16:46 darkstar kernel: radeonfb 0000:01:00.0: Failed to register I2C bus crt2.
> Mar  2 15:16:46 darkstar kernel: Console: switching to colour frame buffer device 90x25
> Mar  2 15:16:46 darkstar kernel: radeonfb (0000:01:00.0): ATI Radeon AP
> 
> Do the "seems to be busy." and/or "Failed to register I2C bus" 
> indicate a problem ?

Maybe we are having conflicting bus names between radeonfb and matroxfb,
or 2 instances of radeonfb ? Can you send the entire log please ?

> There's nothing about radeonfb in dmesg because I manually 
> loaded the modules.

And ? You should have it in dmesg after the module load...

> I now compiled built-in with debug enabled and got the same 
> problem. Nothing got logged. Everything seems to stop when it 
> blanks, but SysRq works.


Well, I would need the full log, and with radeonfb verbose debug enabled
in the config.

Ben.

Re: radeonfb blanks my monitor

[Frédéric L. W. Meunier]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 748 bytes --]

On Fri, 4 Mar 2005, Benjamin Herrenschmidt wrote:

> On Thu, 2005-03-03 at 12:38 -0300, Frédéric L. W. Meunier wrote:
> 
> Maybe we are having conflicting bus names between radeonfb 
> and matroxfb, or 2 instances of radeonfb ? Can you send the 
> entire log please ?

I don't have matroxfb and there aren't 2 instances of radeonfb.

>> There's nothing about radeonfb in dmesg because I manually
>> loaded the modules.
>
> And ? You should have it in dmesg after the module load...

Not in /var/log/dmesg, and I can't type dmesg.

> Well, I would need the full log, and with radeonfb verbose 
> debug enabled in the config.

I'll later try as module with debug.

-- 
How to contact me - http://www.pervalidus.net/contact.html

Re: radeonfb blanks my monitor

[Frédéric L. W. Meunier]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 3238 bytes --]

On Thu, 3 Mar 2005, Frédéric L. W. Meunier wrote:

> On Fri, 4 Mar 2005, Benjamin Herrenschmidt wrote:
>
>>  Well, I would need the full log, and with radeonfb verbose debug enabled 
>>  in the config.
>
> I'll later try as module with debug.

Here's:

/var/log/messages:

Mar  4 14:00:29 pervalidus kernel: radeonfb: Found Intel x86 BIOS ROM Image
Mar  4 14:00:29 pervalidus kernel: radeonfb: Retreived PLL infos from BIOS
Mar  4 14:00:30 pervalidus kernel: radeonfb: Monitor 1 type CRT found
Mar  4 14:00:30 pervalidus kernel: radeonfb: EDID probed
Mar  4 14:00:30 pervalidus kernel: radeonfb: Monitor 2 type no found

/var/log/syslog:

Mar  4 14:00:28 pervalidus kernel: radeonfb_pci_register BEGIN
Mar  4 14:00:28 pervalidus kernel: radeonfb (0000:01:00.0): Found 262144k of DDR 128 bits wide videoram
Mar  4 14:00:29 pervalidus kernel: radeonfb (0000:01:00.0): mapped 16384k videoram
Mar  4 14:00:29 pervalidus kernel: radeonfb: Reference=27.00 MHz (RefDiv=12) Memory=325.00 Mhz, System=200.00 MHz
Mar  4 14:00:29 pervalidus kernel: radeonfb: PLL min 20000 max 40000
Mar  4 14:00:29 pervalidus kernel: 1 chips in connector info
Mar  4 14:00:29 pervalidus kernel:  - chip 1 has 2 connectors
Mar  4 14:00:29 pervalidus kernel:   * connector 0 of type 2 (CRT) : 2300
Mar  4 14:00:29 pervalidus kernel:   * connector 1 of type 3 (DVI-I) : 3221
Mar  4 14:00:29 pervalidus kernel: Starting monitor auto detection...
Mar  4 14:00:29 pervalidus kernel: radeonfb: I2C (port 1) ... not found
Mar  4 14:00:29 pervalidus kernel: radeonfb: I2C (port 2) ... not found
Mar  4 14:00:29 pervalidus kernel: radeonfb: I2C (port 3) ... found CRT display
Mar  4 14:00:30 pervalidus kernel: radeonfb: I2C (port 4) ... not found
Mar  4 14:00:30 pervalidus kernel: radeonfb: I2C (port 2) ... not found
Mar  4 14:00:30 pervalidus kernel: radeonfb: I2C (port 4) ... not found
Mar  4 14:00:30 pervalidus kernel: radeonfb: I2C (port 3) ... found CRT display
Mar  4 14:00:30 pervalidus kernel: hStart = 694, hEnd = 757, hTotal = 795
Mar  4 14:00:30 pervalidus kernel: vStart = 402, vEnd = 408, vTotal = 418
Mar  4 14:00:30 pervalidus kernel: h_total_disp = 0x590062^I   hsync_strt_wid =0x8702c0
Mar  4 14:00:30 pervalidus kernel: v_total_disp = 0x18f01a1^I   vsync_strt_wid = 0x860191
Mar  4 14:00:30 pervalidus kernel: pixclock = 85925
Mar  4 14:00:30 pervalidus kernel: freq = 1163
Mar  4 14:00:30 pervalidus kernel: freq = 1666, PLL min = 20000, PLL max = 40000
Mar  4 14:00:30 pervalidus kernel: ref_div = 12, ref_clk = 2700, output_freq = 26656
Mar  4 14:00:30 pervalidus kernel: ref_div = 12, ref_clk = 2700, output_freq = 26656
Mar  4 14:00:30 pervalidus kernel: post div = 0x5
Mar  4 14:00:30 pervalidus kernel: fb_div = 0x76
Mar  4 14:00:30 pervalidus kernel: ppll_div_3 = 0x50076
Mar  4 14:00:30 pervalidus kernel: Console: switching to colour frame buffer device 90x25
Mar  4 14:00:30 pervalidus kernel: radeonfb (0000:01:00.0): ATI Radeon AP
Mar  4 14:00:30 pervalidus kernel: radeonfb_pci_register END

BTW, I noticed that it locking the system. I was able to issue 
a shutdown. The commands still work, but beep is disabled.

-- 
How to contact me - http://www.pervalidus.net/contact.html

Re: radeonfb blanks my monitor

[Jean Delvare]

Hi Frédéric, hi Benjamin,

> Mar  2 15:16:45 darkstar kernel: radeonfb: Reference=27.00 MHz 
> (RefDiv=12) Memory=325.00 Mhz, System=200.00 MHz
> Mar  2 15:16:45 darkstar kernel: radeonfb: PLL min 20000 max 40000
> Mar  2 15:16:45 darkstar kernel: i2c-algo-bit.o: monid seems to be busy.
> Mar  2 15:16:45 darkstar kernel: radeonfb 0000:01:00.0: Failed to 
> register I2C bus monid.
> Mar  2 15:16:45 darkstar kernel: i2c-algo-bit.o: crt2 seems to be busy.
> Mar  2 15:16:46 darkstar kernel: radeonfb 0000:01:00.0: Failed to 
> register I2C bus crt2.
> Mar  2 15:16:46 darkstar kernel: Console: switching to colour frame 
> buffer device 90x25
> Mar  2 15:16:46 darkstar kernel: radeonfb (0000:01:00.0): ATI Radeon AP
> 
> Do the "seems to be busy." and/or "Failed to register I2C bus" indicate 
> a problem ?

The i2c-algo-bit driver has an optional bus check function. It can be 
used to detect stalled, disconnected or improperly configured busses. 
The check seems to be enabled in your case, which is kind of strange 
since it isn't by default.

Frédéric, can you check in /etc/modprobe.conf if you have a line like:
options i2c-algo-bit bit_test=1
If you do, please comment it out and see if it changes anything.

Benjamin, i2c-algo-bit with bit_test=1 may cause I2C bus creation to 
fail. As far as I can see, it is even the only way the bus creation 
function can return a negative value (which I consider a bug because 
there are other possible failure causes). I will test the effect of 
bit_test=1 on my two radeon systems this evening and see what happens. I 
will also work on a patch fixing the return value if it needs to be.

What I wanted to underline is the fact that it might be the first time 
bus registration ever fails, and although it looks like the radeonfb 
code handles the case, that code path had certainly not received much 
testing so far.

-- 
Jean Delvare

Re: radeonfb blanks my monitor

[Frédéric L. W. Meunier]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 350 bytes --]

On Fri, 4 Mar 2005, Jean Delvare wrote:

> Frédéric, can you check in /etc/modprobe.conf if you have a 
> line like: options i2c-algo-bit bit_test=1 If you do, please 
> comment it out and see if it changes anything.

Yes, I had, but commenting it out didn't change anything.

-- 
How to contact me - http://www.pervalidus.net/contact.html

Re: radeonfb blanks my monitor

[Jean Delvare]

> > Frédéric, can you check in /etc/modprobe.conf if you have a 
> > line like: options i2c-algo-bit bit_test=1 If you do, please 
> > comment it out and see if it changes anything.
> 
> Yes, I had, but commenting it out didn't change anything.

I can confirm that. Adding bit_test=1 here did NOT cause any problem on
my two Radeon systems, so the problem you have isn't related to this
option. Sorry for the noise (but I guess it was still worth trying).

-- 
Jean Delvare