[BUG] buggy ali5451 driver

[BUG] buggy ali5451 driver

[Thierry Vignaud]

regarding ali5451 problems that show up on kml, i just had access to
oliver's laptop which has the infamous ali5451.

both oss' trident and alsa's snd-ali5451 cannot 

# modprobe snd-ali5451
PCI: Found IRQ 4 for device 00:08:0
URQ routing conflict for 00:08:0, have irq 5, want irq 4

then kernel cries:
hda: status timeout: status=0x80 { Busy }
hda: DMA disabled
hda: drive not ready for command.

disks cannot anymore be accessed. the system is still fine (no oops,
no freeze).
the vfs & ext3 then ended in I/O error since the disks answered
themselves busy to the eide controller.


the bugs ali5451 chips suffer may be related to :

- either the poor ressources allocation on laptios, which is poorer &
  poorer since acpi "invade" laptops

- or driver misconception regarding irq sharing


the cars is "M5451 PCI South Bridge Audio" manufactured by Acer
Laboratories Inc. [ALi] (with pci ids : 0x10b9 0x5451)


irq 5 has been allocated to ali snd card, pcmcia cardbus controller
(managed by yenta_socket).


olivier <olivier.ricordeau@wanadoo.fr> is the owner of that fscking
chip.

this is a compaq laptop.

i have access to another compaq laptop that works well with oss
trident.  this is a compaq laptop too.



kernel is 2.4.18 based and alsa was tested with 0.9.0rc3 & rc5



-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[Olivier Ricordeau]

Hi folks!
I am Olivier, the owner of this laptop.
If there's is anything I can do to get my sound card supported by alsa, 
I'd be glad to help! So if you need me form some testing, just mail me! 
(I've also subscribed the list)

I'm in deep shit if I don't manage to get the sound working on my laptop...

Olivier

Thierry Vignaud wrote:
> regarding ali5451 problems that show up on kml, i just had access to
> oliver's laptop which has the infamous ali5451.
> 
> both oss' trident and alsa's snd-ali5451 cannot 
> 
> # modprobe snd-ali5451
> PCI: Found IRQ 4 for device 00:08:0
> URQ routing conflict for 00:08:0, have irq 5, want irq 4
> 
> then kernel cries:
> hda: status timeout: status=0x80 { Busy }
> hda: DMA disabled
> hda: drive not ready for command.
> 
> disks cannot anymore be accessed. the system is still fine (no oops,
> no freeze).
> the vfs & ext3 then ended in I/O error since the disks answered
> themselves busy to the eide controller.
> 
> 
> the bugs ali5451 chips suffer may be related to :
> 
> - either the poor ressources allocation on laptios, which is poorer &
>   poorer since acpi "invade" laptops
> 
> - or driver misconception regarding irq sharing
> 
> 
> the cars is "M5451 PCI South Bridge Audio" manufactured by Acer
> Laboratories Inc. [ALi] (with pci ids : 0x10b9 0x5451)
> 
> 
> irq 5 has been allocated to ali snd card, pcmcia cardbus controller
> (managed by yenta_socket).
> 
> 
> olivier <olivier.ricordeau@wanadoo.fr> is the owner of that fscking
> chip.
> 
> this is a compaq laptop.
> 
> i have access to another compaq laptop that works well with oss
> trident.  this is a compaq laptop too.
> 
> 
> 
> kernel is 2.4.18 based and alsa was tested with 0.9.0rc3 & rc5
> 
> 


-- 

  -= *Olivier RICORDEAU* =-   |  ICQ: 65452304
  Ing1 EPITA Millesime 2005   | Phone: 0686720852
           olivier.ricordeau@wanadoo.fr
          _Might the source be with you_



-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[Takashi Iwai]

At Mon, 04 Nov 2002 12:47:52 +0100,
Thierry Vignaud wrote:
> 
> regarding ali5451 problems that show up on kml, i just had access to
> oliver's laptop which has the infamous ali5451.
> 
> both oss' trident and alsa's snd-ali5451 cannot 
> 
> # modprobe snd-ali5451
> PCI: Found IRQ 4 for device 00:08:0
> URQ routing conflict for 00:08:0, have irq 5, want irq 4
> 
> then kernel cries:
> hda: status timeout: status=0x80 { Busy }
> hda: DMA disabled
> hda: drive not ready for command.
> 
> disks cannot anymore be accessed. the system is still fine (no oops,
> no freeze).
> the vfs & ext3 then ended in I/O error since the disks answered
> themselves busy to the eide controller.

i know of similar problem with this.  they were much severe, caused
kernel panic eventually.

>
> the bugs ali5451 chips suffer may be related to :
> 
> - either the poor ressources allocation on laptios, which is poorer &
>   poorer since acpi "invade" laptops
 
i suspect this.  can he disable acpi support?


Takashi


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[Olivier Ricordeau]

Takashi Iwai wrote:
> At Mon, 04 Nov 2002 12:47:52 +0100,
> Thierry Vignaud wrote:
> 
>>regarding ali5451 problems that show up on kml, i just had access to
>>oliver's laptop which has the infamous ali5451.
>>
>>both oss' trident and alsa's snd-ali5451 cannot 
>>
>># modprobe snd-ali5451
>>PCI: Found IRQ 4 for device 00:08:0
>>URQ routing conflict for 00:08:0, have irq 5, want irq 4
>>
>>then kernel cries:
>>hda: status timeout: status=0x80 { Busy }
>>hda: DMA disabled
>>hda: drive not ready for command.
>>
>>disks cannot anymore be accessed. the system is still fine (no oops,
>>no freeze).
>>the vfs & ext3 then ended in I/O error since the disks answered
>>themselves busy to the eide controller.
> 
> 
> i know of similar problem with this.  they were much severe, caused
> kernel panic eventually.
> 
> 
>>the bugs ali5451 chips suffer may be related to :
>>
>>- either the poor ressources allocation on laptios, which is poorer &
>>  poorer since acpi "invade" laptops
> 
>  
> i suspect this.  can he disable acpi support?
> 

I can... if you explain me how to do this... A kernel option at boot 
time ?. Though I've used Linux for a few years, I'm quite a newbie 
regarding kernel stuffs.

> 
> Takashi
> 


-- 

  -= *Olivier RICORDEAU* =-   |  ICQ: 65452304
  Ing1 EPITA Millesime 2005   | Phone: 0686720852
           olivier.ricordeau@wanadoo.fr
          _Might the source be with you_



-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[Olivier Ricordeau]

Olivier Ricordeau wrote:
> I can... if you explain me how to do this... A kernel option at boot 
> time ?. Though I've used Linux for a few years, I'm quite a newbie 
> regarding kernel stuffs.

Ok I've searched the web, and I've tried to boot my standard MDK kernel 
with the "acpi=off", and it's the same: both "modprobe trident" and 
"modprobe snd-ali5451" make the laptop freeze (I/O errors on hard drive).

-- 

  -= *Olivier RICORDEAU* =-   |  ICQ: 65452304
  Ing1 EPITA Millesime 2005   | Phone: 0686720852
           olivier.ricordeau@wanadoo.fr
          _Might the source be with you_



-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[drclaw]

Try disabling "use dma by default" in the ide section.  If that then
fixes the problem I have a patch that can get you dma back.  I've got an
ali southbridge in my laptop as well and everything works fine.  

784 - Michael C. Piantedosi - drclaw@dogsolitude.org


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[Olivier Ricordeau]

drclaw@dogsolitude.org wrote:
> Try disabling "use dma by default" in the ide section.  If that then
> fixes the problem I have a patch that can get you dma back.  I've got an
> ali southbridge in my laptop as well and everything works fine.  
> 
> 784 - Michael C. Piantedosi - drclaw@dogsolitude.org

OK...
How can I disable dma in a shell??? Or do I need to recompile the kernel 
or something?

-- 

  -= *Olivier RICORDEAU* =-   |  ICQ: 65452304
  Ing1 EPITA Millesime 2005   | Phone: 0686720852
           olivier.ricordeau@wanadoo.fr
          _Might the source be with you_



-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[drclaw]

You could try using hdparm, but I'd just suggest recompiling and seeing
if that helps.  My machine won't even boot is dma is enabled by default.  

784 - Michael C. Piantedosi - drclaw@dogsolitude.org


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[Olivier Ricordeau]

Olivier Ricordeau wrote:
> drclaw@dogsolitude.org wrote:
> 
>> Try disabling "use dma by default" in the ide section.  If that then
>> fixes the problem I have a patch that can get you dma back.  I've got an
>> ali southbridge in my laptop as well and everything works fine. 
>> 784 - Michael C. Piantedosi - drclaw@dogsolitude.org
> 
> 
> OK...
> How can I disable dma in a shell??? Or do I need to recompile the kernel 
> or something?
> 

I've booted with the "nodma" option (I guess that what you wanted me to 
do), and now here is the new output for "modprobe ali5451":

"hda: status timeout: status=0x80 { Busy }
hda: drive not ready for command
ali mixer creating error.
Ali pci audio not found or device busy.
/lib/modules/2.4.19-18mdk/kernel/sound/pci/ali5451/snd-ali5451.o.gz: 
init_module: No such device
Hint: insmod errors can be caused by incorrect module parameters, 
including invalid IO or IRQ parameters.
	You may find more information in syslog or on the output from dmesg
mosprobe: insmod 
/lib/modules/2.4.19-18mdk/kernel/sound/pci/ali5451/snd-ali5451.o.gz failed
modprobe: insmod snd-ali5451 failed"

After that, all my commands result in:
"hda: status timeout: status 8x80 { Busy }
hda: DMA disabled
hda: drive not ready for command"

I've also tried the same modprobe in single-user mode after a "hdparm 
-d0 /dev/hda", and no error message is displayed, but all my command 
result in the same error message as above.

-- 

  -= *Olivier RICORDEAU* =-   |  ICQ: 65452304
  Ing1 EPITA Millesime 2005   | Phone: 0686720852
           olivier.ricordeau@wanadoo.fr
          _Might the source be with you_



-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[drclaw]

Ick, that's not what I was hoping would happen.  Well, I guess my
solution won't work.  Just out of curiosity, can you post your
/proc/interrupts and /proc/ioports?  I'm interested in just exactly what
is being shared by what, and what your ide chipset is supposed to be.  

784 - Michael C. Piantedosi - drclaw@dogsolitude.org


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[Olivier Ricordeau]

drclaw@dogsolitude.org wrote:
> Ick, that's not what I was hoping would happen.  Well, I guess my
> solution won't work.  Just out of curiosity, can you post your
> /proc/interrupts and /proc/ioports?  I'm interested in just exactly what
> is being shared by what, and what your ide chipset is supposed to be.  

Here is what you've asked for:

cat /proc/interrupts:
            CPU0
   0:     150359          XT-PIC  timer
   1:        197          XT-PIC  keyboard
   2:          0          XT-PIC  cascade
   5:          0          XT-PIC  Texas Instruments PCI1410 PC card 
Cardbus Controller
   8:          1          XT-PIC  rtc
   9:          3          XT-PIC  acpi
  11:          3          XT-PIC  eth0
  12:        280          XT-PIC  PS/2 Mouse
  14:       5506          XT-PIC  ide0
  15:          4          XT-PIC  ide1
NMI:          0
LOC:     150304
ERR:        170
MIS:          0



cat /proc/ioports:

0000-001f : dma1
0020-003f : pic1
0040-005f : timer
0060-006f : keyboard
0070-007f : rtc
0080-008f : dma page reg
00a0-00bf : pic2
00c0-00df : dma2
00f0-00ff : fpu
0170-0177 : ide1
01f0-01f7 : ide0
0376-0376 : ide1
03c0-03df : vesafb
03f6-03f6 : ide0
0cf8-0cff : PCI conf1
4000-40ff : PCI CardBus #02
4400-44ff : PCI CardBus #02
8000-803f : Acer Laboratories Inc. [ALi] M7101 PMU
8040-805f : Acer Laboratories Inc. [ALi] M7101 PMU
8080-808f : Acer Laboratories Inc. [ALi] M5229 IDE
   8080-8087 : ide0
   8088-808f : ide1
8090-8093 : PCI device 1002:cab0 (ATI Technologies Inc)
8098-809f : Conexant HSF 56k HSFi Modem
8400-84ff : Acer Laboratories Inc. [ALi] M5451 PCI AC-Link Controller 
Audio Devi
ce
8800-88ff : Realtek Semiconductor Co., Ltd. RTL-8139/8139C
   8800-88ff : 8139too
9000-9fff : PCI Bus #01
   9000-90ff : PCI device 1002:4336 (ATI Technologies Inc)



Hope it will help...
Note that this is wat I got after a "normal" boot (no kernel option, dma 
still active).

It would be just great not to be obliged to install a window$ on this 
laptop... I've not used Microsoft's products for several months and I 
feel just fine with it! ;-)

Oliv

-- 

  -= *Olivier RICORDEAU* =-   |  ICQ: 65452304
  Ing1 EPITA Millesime 2005   | Phone: 0686720852
           olivier.ricordeau@wanadoo.fr
          _Might the source be with you_



-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: [BUG] buggy ali5451 driver

[Thierry Vignaud]

Olivier Ricordeau <olivier.ricordeau@wanadoo.fr> writes:

> Hope it will help...
> Note that this is wat I got after a "normal" boot (no kernel option,
> dma still active).
> 
> It would be just great not to be obliged to install a window$ on this
> laptop... I've not used Microsoft's products for several months and I
> feel just fine with it! ;-)

since some bad interaction happens with eide disks, one thing that
could help would be to try 2.4.20-rc1 + alsa-0.9.0rc5 since alan cox
fixed ali eide controller driver.



-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en

Re: [BUG] buggy ali5451 driver

[Olivier Ricordeau]

Thierry Vignaud wrote:
> Olivier Ricordeau <olivier.ricordeau@wanadoo.fr> writes:
> 
> 
>>Hope it will help...
>>Note that this is wat I got after a "normal" boot (no kernel option,
>>dma still active).
>>
>>It would be just great not to be obliged to install a window$ on this
>>laptop... I've not used Microsoft's products for several months and I
>>feel just fine with it! ;-)
> 
> 
> since some bad interaction happens with eide disks, one thing that
> could help would be to try 2.4.20-rc1 + alsa-0.9.0rc5 since alan cox
> fixed ali eide controller driver.

Sounds good!
I'll try it next week when I have more time.

Thanks!

-- 

  -= *Olivier RICORDEAU* =-   |  ICQ: 65452304
  Ing1 EPITA Millesime 2005   | Phone: 0686720852
           olivier.ricordeau@wanadoo.fr
          _Might the source be with you_



-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en