Skip to contentSkip to conversation

PCMCIA and snd-hdsp

DISCUSSION Orm Finnendahl 5 participants 10 messages Aug 28, 2002 – Aug 29, 2002 2 messages not in the archive

lore
Share & tools
lore
#0 lore raw

Hi,

after recompiling a kernel on my laptop, my RME HDSP configuration is broken. Does anybody have a working setup with the RME PCMCIA card and would be willing to share some information about the process of initialization off list?

--
Orm

-------------------------------------------------------
This sf.net email is sponsored by: Jabber - The world's fastest growing real-time communications platform! Don't just IM. Build it in!
http://www.jabber.com/osdn/xim

#1 lore raw
Quoted 4 lines

after recompiling a kernel on my laptop, my RME HDSP configuration is broken. Does anybody have a working setup with the RME PCMCIA card and would be willing to share some information about the process of initialization off list?

i don't have the working setup (with pcmcia), but i would obviously like to hear more about whatever has broken. do you still have the old kernel, and does it still work?

--p

-------------------------------------------------------
This sf.net email is sponsored by: Jabber - The world's fastest growing real-time communications platform! Don't just IM. Build it in!
http://www.jabber.com/osdn/xim

#6 lore raw
In reply to a message not in the archive

Hi,

Am Mittwoch, den 28. August 2002 um 08:30:21 Uhr (-0400) schrieb Paul Davis:

Quoted 4 lines

i don't have the working setup (with pcmcia), but i would obviously like to hear more about whatever has broken. do you still have the old kernel, and does it still work?

Yes and no. This is what happened:

I compiled a new debian kernel package with a 2.4.19 kernel. My working kernel was 2.4.18-686 and in addition I had the original 2.4.18-bf2.4 (both available with debian). After I had compiled and installed the kernel, it didn't work until I had deinstalled and reinstalled the old 2.4.18-686 kernel and the appropriate pcmcia-cs and pcmcia-modules. After that, I recompiled the whole ALSA.

Now the snd-hammerfall-mem install o.k. but snd-hdsp segfaults (and I have to reboot in order to get rid of the module).

I found out that I got a new version of pcmcia-cs (3.2 instead of the old potato version 3.1.22) and the syslog looks different in that respect. I'll attach the syslog message with the relevant part. The Aug 27 8:07 h one is the working one, the other one the failing one. I even tried to reinstall the old potato pcmcia cardservices but that seems to be broken with the upgrade of debian to woody. It seems to be quite a mess and I try to get down to the problem. This is clear:

- pcmcia works fine with other cards with the new setup.

- my built-in soundcard (ali5451) works fine with the fresh compiled ALSA

When I try to get information about the snd-hdsp card (/proc/irq or io settings), nothing is displayed although snd-hammerfall-mem finds the inserted card.

I would like to know whether the problem is on the side of the cardservices (e.g. by not providing the necessary irq and io information for the hdsp driver) or how the whole mechanism works. I couldn't find the relvant information anywhere.

--
Orm

here are the logs:

startup (multiface connected, but the snd-hdsp driver not yet loaded)

Aug 27 08:07:06 grisey kernel: Linux Kernel Card Services 3.1.22
Aug 27 08:07:06 grisey kernel:   options:  [pci] [cardbus] [pm]
Aug 27 08:07:07 grisey kernel: isapnp: Scanning for PnP cards...
Aug 27 08:07:07 grisey kernel: isapnp: No Plug & Play device found
Aug 27 08:07:07 grisey kernel: Intel PCIC probe: not found.
Aug 27 08:07:07 grisey kernel: PCI: Guessed IRQ 11 for device 00:13.0
Aug 27 08:07:07 grisey kernel: PCI: Sharing IRQ 11 with 00:13.1
Aug 27 08:07:07 grisey kernel: PCI: Guessed IRQ 11 for device 00:13.1
Aug 27 08:07:07 grisey kernel: PCI: Sharing IRQ 11 with 00:13.0
Aug 27 08:07:07 grisey kernel: Yenta IRQ list 0238, PCI irq11
Aug 27 08:07:07 grisey kernel: Socket status: 30000826
Aug 27 08:07:07 grisey kernel: Yenta IRQ list 0238, PCI irq11
Aug 27 08:07:07 grisey kernel: Socket status: 30000006
Aug 27 08:07:08 grisey kernel: cs: cb_alloc(bus 2): vendor 0x10ee, device 0x3fc5
Aug 27 08:07:08 grisey kernel: PCI: Enabling device 02:00.0 (0080 -> 0082)
Aug 27 08:07:08 grisey cardmgr[290]: error in file './hdsp.conf' line 2: syntax error
Aug 27 08:07:08 grisey cardmgr[290]: watching 2 sockets
Aug 27 08:07:08 grisey kernel: cs: IO port probe 0x0c00-0x0cff: clean.
Aug 27 08:07:08 grisey kernel: cs: IO port probe 0x0800-0x08ff: clean.
Aug 27 08:07:08 grisey kernel: cs: IO port probe 0x0100-0x04ff: excluding 0x378-0x37f 0x408-0x40f 0x480-0x48f 0x4d0-0x4d7
Aug 27 08:07:08 grisey kernel: cs: IO port probe 0x0a00-0x0aff: clean.
Aug 27 08:07:08 grisey cardmgr[291]: starting, version is 3.2.0
Aug 27 08:07:08 grisey cardmgr[291]: socket 0: CardBus hotplug device
Aug 27 08:07:09 grisey kernel: Hammerfall memory allocator: buffers allocated for 1 cards

loading of the driver:

Aug 27 14:53:18 grisey kernel: cs: cb_free(bus 2)
Aug 27 14:53:22 grisey kernel: cs: cb_alloc(bus 2): vendor 0x10ee, device 0x3fc5
Aug 27 14:53:22 grisey kernel: PCI: Enabling device 02:00.0 (0080 -> 0082)
Aug 27 14:53:22 grisey cardmgr[291]: socket 0: CardBus hotplug device
Aug 27 14:53:28 grisey kernel: PCI: Setting latency timer of device 02:00.0 to 64
Aug 27 14:53:28 grisey kernel: ALSA ../../alsa-kernel/pci/rme9652/hdsp.c:453: wait for FIFO status <= 0 failed after 100 iterations
Aug 27 14:53:28 grisey kernel: looks like a multiface

This setup doesn't work:

Aug 28 18:16:37 grisey kernel: Linux PCMCIA Card Services 3.2.0
Aug 28 18:16:37 grisey kernel:   kernel build: 2.4.18-686 unknown
Aug 28 18:16:37 grisey kernel:   options:  [pci] [cardbus] [apm]
Aug 28 18:16:37 grisey kernel: Intel ISA/PCI/CardBus PCIC probe:
Aug 28 18:16:37 grisey kernel: PCI: Guessed IRQ 11 for device 00:13.0
Aug 28 18:16:37 grisey kernel: PCI: Sharing IRQ 11 with 00:13.1
Aug 28 18:16:37 grisey kernel: PCI: Guessed IRQ 11 for device 00:13.1
Aug 28 18:16:37 grisey kernel: PCI: Sharing IRQ 11 with 00:13.0
Aug 28 18:16:37 grisey kernel:   O2Micro OZ6933 rev 01 PCI-to-CardBus at slot 00:13, mem 0x10000000
Aug 28 18:16:37 grisey kernel:     host opts [0]: [pci/way] [pci irq 11] [lat 32/176] [bus 2/5]
Aug 28 18:16:37 grisey kernel:     host opts [1]: [pci/way] [pci irq 11] [lat 32/176] [bus 6/9]
Aug 28 18:16:37 grisey kernel:     ISA irqs (default) = 3,4,5,7,9,12 PCI status changes
Aug 28 18:16:38 grisey cardmgr[296]: watching 2 sockets
Aug 28 18:16:38 grisey cardmgr[297]: starting, version is 3.2.0
Aug 28 18:16:38 grisey kernel: cs: cb_alloc(bus 2): vendor 0x10ee, device 0x3fc5
Aug 28 18:16:38 grisey cardmgr[297]: unsupported card in socket 0
Aug 28 18:16:38 grisey cardmgr[297]:   product info: "Martin Kirst", "Hammerfall DSP", "CardBus Audio Card 1.00"
Aug 28 18:16:38 grisey cardmgr[297]:   PCI id: 0x10ee, 0x3fc5
Aug 28 18:16:38 grisey kernel: Hammerfall memory allocator: buffers allocated for 1 cards

Trying to modprobe snd-hdsp...

Aug 28 18:18:11 grisey kernel: PCI: No IRQ known for interrupt pin A of device . Please try using pci=biosirq.
Aug 28 18:18:11 grisey kernel: PCI: Setting latency timer of device  to 64
Aug 28 18:18:11 grisey kernel: ALSA ../../alsa-kernel/pci/rme9652/hdsp.c:2967: unable to grab memory region 0x0-0x1447
Aug 28 18:18:11 grisey kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000040
Aug 28 18:18:11 grisey kernel:  printing eip:
Aug 28 18:18:11 grisey kernel: d0a2741b
Aug 28 18:18:11 grisey kernel: *pde = 00000000
Aug 28 18:18:11 grisey kernel: Oops: 0002
Aug 28 18:18:11 grisey kernel: CPU:    0
Aug 28 18:18:11 grisey kernel: EIP:    0010:[ide-mod:ide_devfs_handle_Reb53c0a2+1796923/253010458]    Not tainted
Aug 28 18:18:11 grisey kernel: EFLAGS: 00010246
Aug 28 18:18:11 grisey kernel: eax: 00000000   ebx: cba581a8   ecx: cba58140   edx: 00000000
Aug 28 18:18:11 grisey kernel: esi: fffffff0   edi: d0ae7480   ebp: cba581a8   esp: cba5debc
Aug 28 18:18:11 grisey kernel: ds: 0018   es: 0018   ss: 0018
Aug 28 18:18:11 grisey kernel: Process modprobe (pid: 547, stackpage=cba5d000)
Aug 28 18:18:11 grisey kernel: Stack: cba58000 d0a2748c cba581a8 d0a05aa3 cba58000 cba58000 d0a27578 cba58000 
Aug 28 18:18:11 grisey kernel:        d0ae6f90 cf7aa41c d0ae7480 00000000 c017a12e cf7aa41c d0ae6f90 cf7aa41c 
Aug 28 18:18:11 grisey kernel:        d0ae7480 00000000 000c34ac c017a194 d0ae7480 cf7aa41c d0a24000 00000004 
Aug 28 18:18:11 grisey kernel: Call Trace: [ide-mod:ide_devfs_handle_Reb53c0a2+1797036/253010345] [ide-mod:ide_devfs_handle_Reb53c0a2+1659331/253148050] [ide-mod:ide_devfs_handle_Reb53c0a2+1797272/253010109] [ide-mod:ide_devfs_handle_Reb53c0a2+2582192/252225189] [ide-mod:ide_devfs_handle_Reb53c0a2+2583456/252223925] 
Aug 28 18:18:11 grisey kernel:    [pci_announce_device+54/84] [ide-mod:ide_devfs_handle_Reb53c0a2+2582192/252225189] [ide-mod:ide_devfs_handle_Reb53c0a2+2583456/252223925] [pci_register_driver+72/96] [ide-mod:ide_devfs_handle_Reb53c0a2+2583456/252223925] [ide-mod:ide_devfs_handle_Reb53c0a2+1797340/253010041] 
Aug 28 18:18:11 grisey kernel:    [ide-mod:ide_devfs_handle_Reb53c0a2+2583456/252223925] [sys_init_module+1285/1448] [ide-mod:ide_devfs_handle_Reb53c0a2+1783680/253023701] [system_call+51/56] 
Aug 28 18:18:11 grisey kernel: 
Aug 28 18:18:11 grisey kernel: Code: 89 50 40 8b 83 c4 00 00 00 85 c0 7c 0a 53 50 e8 c5 0d 6e ef 

-------------------------------------------------------
This sf.net email is sponsored by: Jabber - The world's fastest growing real-time communications platform! Don't just IM. Build it in!
http://www.jabber.com/osdn/xim

#7 lore raw

Hi, I'm running woody w/ pcmcia-cs_3.1.33-6 and my own custom build, but unpatched kernel 2.4.18, and a recent ALSA CVS checkout. I haven't tried my brand spanking new hdsp very much (been working), but it seems to load up nicely. I'll try to compile up 2.4.19 tonight to see if that breaks anything, and if it doesn't I'll wager that the problem lies with cardservices. BTW, I have a fujitsu laptop I'm using it with, which appears to have very similar internals to yours, lots of ali devices all over the place.

784 - Michael C. Piantedosi - [email protected]

-------------------------------------------------------
This sf.net email is sponsored by: Jabber - The world's fastest growing real-time communications platform! Don't just IM. Build it in!
http://www.jabber.com/osdn/xim

#8 lore raw
In reply to Orm Finnendahl
Quoted 2 lines

Now the snd-hammerfall-mem install o.k. but snd-hdsp segfaults (and I have to reboot in order to get rid of the module).

well, pcmci-cs may be at fault, but this is still a bug in the driver. it shouldn't segfault. i guess i've assumed that certain things are always true and that this is broken by pcmcia devices.

Quoted 4 lines

I would like to know whether the problem is on the side of the cardservices (e.g. by not providing the necessary irq and io information for the hdsp driver) or how the whole mechanism works. I couldn't find the relvant information anywhere.

it sounds fairly likely that its pcmcia-cs that is the problem. it seems to be the main focal point for the difference between people with working pcmcia H-DSP's and those who do not.

i don't know what this code does that makes so much difference. i say so much, because if you notice, its not the hdsp driver that segfaults, but some part of the IDE code.

--p

-------------------------------------------------------
This sf.net email is sponsored by: Jabber - The world's fastest growing real-time communications platform! Don't just IM. Build it in!
http://www.jabber.com/osdn/xim

#9 lore raw
In reply to Orm Finnendahl

On 28/08/02 Orm Finnendahl wrote...

Quoted 11 lines

Am Mittwoch, den 28. August 2002 um 08:30:21 Uhr (-0400) schrieb Paul Davis:

Yes and no. This is what happened:

I compiled a new debian kernel package with a 2.4.19 kernel. My working kernel was 2.4.18-686 and in addition I had the original 2.4.18-bf2.4 (both available with debian). After I had compiled and installed the kernel, it didn't work until I had deinstalled and reinstalled the old 2.4.18-686 kernel and the appropriate pcmcia-cs and pcmcia-modules. After that, I recompiled the whole ALSA.

Quoted 3 lines

Now the snd-hammerfall-mem install o.k. but snd-hdsp segfaults (and I have to reboot in order to get rid of the module).

Hi, i'm trying to convince the PCMCIA HDSP to work, and I have more or less the same problem (I believe that's is pcmcia-cs side but i dont have other cards to test, apart from the built in wirless interface that is correctly recognized).

I have a custom 2.4.18 (ACPI and low latency patched) and latest CVS alsa driver on Debian Woody and pcmcia 3.1.33-6 (i82365). The laptop is Toshiba Tecra 9100.

Loading snd-hdsp provokes seg fault with basically the same log as Orm Finnendahl's (see below).

Another strange thing is that, after inserting the card, lspci doesnt work anymore:

~ # lspci
pcilib: Cannot open /proc/bus/pci/04/00.0
lspci: Unable to read 64 bytes of configuration space. ~ #

I dont think that the snd-hdsp modules itself crashes, after trying to load it and getting the mentioned log, lsmod gives:

snd-hdsp              214596   1  (initializing)
snd-hammerfall-mem      1824   0  [snd-hdsp]

and hdsp keeps initializing forever (moreover it cant be removedn and it hangs the system when I try to shutdown)

Another thing detail is that cardctl ident gives more information than cardinfo:

~ # cardctl ident
Socket 0:
product info: "TOSHIBA", "Wireless LAN Card", "Version 01.01", ""
manfid: 0x0156, 0x0002
function: 6 (network)
Socket 1:
product info: "Martin Kirst", "Hammerfall DSP", "CardBus Audio
Card 1.00"
PCI id: 0x10ee, 0x3fc5
Socket 2:
no product info available

BUT if I use cardinfo, it detects that the Socket 1 is in use but doesnt print any information about it.

I guess the problem is somewhat related in the IRQ routing or assignment ..

Quoted 10 lines
Aug 28 18:18:11 grisey kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000040
Aug 28 18:18:11 grisey kernel:  printing eip:
Aug 28 18:18:11 grisey kernel: d0a2741b
Aug 28 18:18:11 grisey kernel: *pde = 00000000
Aug 28 18:18:11 grisey kernel: Oops: 0002
Aug 28 18:18:11 grisey kernel: CPU:    0
Aug 28 18:18:11 grisey kernel: EIP:    0010:[ide-mod:ide_devfs_handle_Reb53c0a2+1796923/253010458]    Not tainted
Aug 28 18:18:11 grisey kernel: EFLAGS: 00010246
Aug 28 18:18:11 grisey kernel: eax: 00000000   ebx: cba581a8   ecx: cba58140   edx: 00000000
Aug 28 18:18:11 grisey kernel: esi: fffffff0   edi: d0ae7480   ebp: cba581a8   esp: cba5debc
Quoted 8 lines
Aug 28 18:18:11 grisey kernel: ds: 0018   es: 0018   ss: 0018
Aug 28 18:18:11 grisey kernel: Process modprobe (pid: 547, stackpage=cba5d000)
Aug 28 18:18:11 grisey kernel: Stack: cba58000 d0a2748c cba581a8 d0a05aa3 cba58000 cba58000 d0a27578 cba58000
Aug 28 18:18:11 grisey kernel:        d0ae6f90 cf7aa41c d0ae7480 00000000 c017a12e cf7aa41c d0ae6f90 cf7aa41c
Aug 28 18:18:11 grisey kernel:        d0ae7480 00000000 000c34ac c017a194 d0ae7480 cf7aa41c d0a24000 00000004
Aug 28 18:18:11 grisey kernel: Call Trace: [ide-mod:ide_devfs_handle_Reb53c0a2+1797036/253010345] [ide-mod:ide_devfs_handle_Reb53c0a2+1659331/253148050] [ide-mod:ide_devfs_handle_Reb53c0a2+1797272/253010109] [ide-mod:ide_devfs_handle_Reb53c0a2+2582192/252225189] [ide-mod:ide_devfs_handle_Reb53c0a2+2583456/252223925]
Aug 28 18:18:11 grisey kernel:    [pci_announce_device+54/84] [ide-mod:ide_devfs_handle_Reb53c0a2+2582192/252225189] [ide-mod:ide_devfs_handle_Reb53c0a2+2583456/252223925] [pci_register_driver+72/96] [ide-mod:ide_devfs_handle_Reb53c0a2+2583456/252223925] [ide-mod:ide_devfs_handle_Reb53c0a2+1797340/253010041]
Aug 28 18:18:11 grisey kernel:    [ide-mod:ide_devfs_handle_Reb53c0a2+2583456/252223925] [sys_init_module+1285/1448] [ide-mod:ide_devfs_handle_Reb53c0a2+1783680/253023701] [system_call+51/56]

-------------------------------------------------------
This sf.net email is sponsored by: Jabber - The world's fastest growing real-time communications platform! Don't just IM. Build it in!
http://www.jabber.com/osdn/xim

#2 lore raw
In reply to Paul Davis

I have a somewhat working pcmcia hdsp configuration.
I would be very interested in removing the somewhat from the previous statement... and can certainly provide any details from my configuration that may be of assistance.

I'm using the planet ccrma packages, which means 2.4.19-1 kernel with low latency (and other) patches alsa 0.9.0-36
on top of red hat 7.3

For some reason, the hotplug script can't successfully load the snd-hdsp driver but a simple
"modprobe snd-hdsp" does work. This is true at boot time (coldplug) and upon insertion (hotplug). Can anyone offer any insight into this?

I don't have the latest pcmcia-cs (I think I've got 3.1.22 but I'm not on that machine at the moment) and don't appear to be having any segfault trouble... however occasionally I do get unknown device errors when pcmcia starts up and sometimes the card isn't registered at all.

daniel

Quoted 21 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]]On Behalf Of Paul Davis
Sent: Wednesday, 28 August 2002 10:30 PM
To: Orm Finnendahl
Cc: [email protected]
Subject: Re: [Alsa-devel] PCMCIA and snd-hdsp

Quoted 4 lines

after recompiling a kernel on my laptop, my RME HDSP configuration is broken. Does anybody have a working setup with the RME PCMCIA card and would be willing to share some information about the process of initialization off list?

i don't have the working setup (with pcmcia), but i would obviously like to hear more about whatever has broken. do you still have the old kernel, and does it still work?

--p

Quoted 9 lines

-------------------------------------------------------
This sf.net email is sponsored by: Jabber - The world's fastest growing real-time communications platform! Don't just IM. Build it in!
http://www.jabber.com/osdn/xim
_______________________________________________
Alsa-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/alsa-devel

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

#3 lore raw
Quoted 4 lines

For some reason, the hotplug script can't successfully load the snd-hdsp driver but a simple
"modprobe snd-hdsp" does work. This is true at boot time (coldplug) and upon insertion (hotplug). Can anyone offer any insight into this?

i know nothing about the hotplug script of pcmcia-cs, but i bet its because it doesn't use the dependency tree that depmod builds. you would have the same problem with insmod(8) for example. modprobe doesn't just load the named module, it loads all the required modules as well.

--p

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

#11 lore raw
In reply to a message not in the archive

Hi all,

I got the RME working again. It even works with the new pcmcia-cs. I also posted a message on pcmcia-cs in order to get the config file right (its in /etc/pcmcia/config and I got part of it working, but it doesn't yet do it all the way; I'll post it if I succeed).

I had to do about 10 kernel compiles until I found the right combination of modules or compiled-in features. I don't exactly know, what made it work. Following are the specs of my working kernel-config file which might be important:

- Support for hot pluggable devices are compiled into the kernel
- pcmcia is compiled into the kernel
- Cardbus Support is compiled into the kernel
- i82365 compatible bridge is compiled into the kernel
- all other options concerning PCMCIA are disabled
- Hotplug support (PCI) is disabled
- Plug and Play support is disabled

After that I recompiled the Alsa-driver and AlsaLib and it worked right away.

Am Mittwoch, den 28. August 2002 um 14:59:32 Uhr (-0400) schrieb Paul Davis:

Quoted 3 lines

i don't know what this code does that makes so much difference. i say so much, because if you notice, its not the hdsp driver that segfaults, but some part of the IDE code.

I noticed that.

I guess, the best bet is not to use modules at all if pcmcia is concerned, but compile everything into the kernel (although it ran with the old card-services on my machine using modules). Also it seems that the card is not cardbus-compatible and therefore the compatible bridge is needed.

Thanks to all of you, who helped!!

Orm

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

#4 lore raw
In reply to Paul Davis
Quoted 11 lines
Quoted 5 lines

For some reason, the hotplug script can't successfully load the snd-hdsp driver but a simple
"modprobe snd-hdsp" does work. This is true at boot time (coldplug) and upon insertion (hotplug). Can anyone offer any insight into this?

i know nothing about the hotplug script of pcmcia-cs, but i bet its because it doesn't use the dependency tree that depmod builds. you would have the same problem with insmod(8) for example. modprobe doesn't just load the named module, it loads all the required modules as well.

No, the script does appear to do a modprobe. I've checked the scripts themselves and watched their behaviour. The appropriate required modules are loaded (including snd-hammerfall-mem if required). All of those modules load correctly. However, the snd-hdsp module complains that it can't find the module (not that there are unresolved symbols). Then all of the other loaded modules are unloaded (including snd-hammerfall-mem).

I'm not very experienced with shell scripts so it will take me quite a while to scour through and look at the process in more detail...

Does anyone have their hdsp hotplugging successfully?

Also, I can't get alsactl to restore mixer settings correctly. I can see that they've been written correctly in the asound.state file but they don't seem to be restored properly. What am I doing wrong?

many thanks,
daniel

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf