Use this page to discuss devices that do not currently have any Linux PCMCIA support. Before posting, be sure you've checked any relevant sections of the PCMCIA-HOWTO, such as:

Messages Inline: 0 1

Question New Media Corp. BASICS Joysitck Adapter

Date: 1999, Nov 27
From: R. Timothy Stewart roadkill

Hello all,
  Just curious if anyone has any information on the card mentioned in the title, or how to get information. I would like to attempt a driver for it, but when I emailed New Media they refused to supply any information. Any help would be appreciated. Thanks!

-TimS

Ok Sedlbauer card works fine

Date: 1999, Dec 09
From: Valentijn Sessink Valentijn

As subject says: sedlbauer Speed Star works fine for me. Out of the box, with 3.1.5., Slackware 7, Vaio F304 laptop.

V.

Question PCMCIA ALAS sound driver

Date: 2000, Feb 23
From: Steve Feil stevef

I have seen some high quality PCMCIA sound cards advertised
on the web, however I am disappointed that PCMCIA sound cards
are not supported in Linux.

I have one company in mind that may be willing to supply the
specifications for a PCMCIA sound card they produce.

I would be willing to learn how to write a driver, however
right now I'm just gathering background information.

According to what I have read, the reason that PCMCIA sound
cards are not supported in Linux, is because they use DMA. 
Going on this assumption I have four questions, which I'm 
posting to the PCMCIA mailing list as well as the one for ALSA.

1. Is DMA used on PCMCIA cards other than sound cards?

2. Have any PCMCIA DMA drivers been written for Linux?

3. Can a DMA driver written for a PCMCIA card (other 
   than sound) be easily adapted for a sound card?

4. If a PCMCIA DMA driver module were written, how easy would
   it be for a ALSA module to interface with the DMA module?

None It is a difficult issue

Re: Question PCMCIA ALAS sound driver (Steve Feil)
Date: 2000, Feb 23
From: David Hinds <dhinds@pcmcia.sourceforge.org>

> 1. Is DMA used on PCMCIA cards other than sound cards?

Generally no; I am not aware of any.  Some floppy disk interface cards
can optionally use DMA, but Linux doesn't support that mode.

> 2. Have any PCMCIA DMA drivers been written for Linux?

No.

> 3. Can a DMA driver written for a PCMCIA card (other 
>    than sound) be easily adapted for a sound card?

I should give some background about DMA and PCMCIA.  Early PCMCIA
bridges, and the original PCMCIA specification, did not support DMA,
period.  Some vendors later started adding DMA support, but because
there was no standard, each made up their own proprietary scheme for
doing it.  This was all pre-PCI.  Then with PCI and CardBus bridges, a
new issue came up: the host ISA DMA controller doesn't have anything
to do with PCI, and can't talk to an ISA device like a PCMCIA sound
card sitting behind a CardBus bridge on the PCI bus.  So vendors came
up with yet more DMA kludges.  One relatively common scheme is for the
CardBus bridge to have a set of extra registers that look like an ISA
DMA controller, and the bridge emulates the DMA engine using PCI bus
master operations.  Unfortunately, if you look at the Linux ISA DMA
driver, there is no simple way to support an extra DMA controller.

So, to support PCMCIA DMA devices, you need to:

(1) provide support in the CardBus bridge driver for configuring the
    DMA controller, for bridges that have one.
(2) rewrite Linux DMA support to support additional DMA controllers
    that are "bound" to particular devices.

In short, it is messy and non-trivial.

-- Dave

Question howto motivate people writing drivers?

Date: 2000, May 12
From: chris navaryder

Hi all,

Sorry for asking this sort of dumb question. I've been checking the hardware-howto several times in the last years and haven't seen any progress on the pcmcia-sound front. I'd really *love* to see a card supported as this would mean sound for my wearable computer.

So, what is needed to convince one of you driver-coders to write something? (eg. free sample card, $$$, eternal gratitude...) I'm sure there's quite a bunch of other people waiting for this. Would any of them be willing to aid in 'motivating'? (we could set up a fund or something)

thanks, chris

Sad No easy answer for this either

Re: Question howto motivate people writing drivers? (chris)
Date: 2000, May 12
From: David Hinds <dhinds@pcmcia.sourceforge.org>

Traditionally, people write linux drivers for things because they have a specific need. You might be able to find someone who would do it for money, but you might not. The fact that there has been no progress on PCMCIA sound support reflects two things: implementing it is hard, and few people need it. Only a tiny minority of linux PCMCIA users (who are themselves a tiny minority of linux users) have any use for PCMCIA sound cards, because the vast majority of commercial laptops come with sound support superior to what could be done in a PCMCIA device.

I know someone at DEC did some work to develop a custom kernel for an ARM-based system to support a PCMCIA sound card. It was a very specific hack for a very specific hardware platform, and the trick is not applicable to other architectures.

-- Dave

Idea Lucent LT Win Modem Global (Pcmcia)

Date: 2000, Jul 10
From: William wic

Any one did try the card... its not working under Linux SuSE 6.4... Any solution??
Message is :
Socket 0:
  no product info available
Socket 1:
  product info: "LT WIN MODEM GLOBAL", "", "", ""

None Sorry...

Re: Idea Lucent LT Win Modem Global (Pcmcia) (William)
Date: 2000, Jul 10
From: David Hinds <dhinds@pcmcia.sourceforge.org>

On Mon, Jul 10, 2000 at 11:48:18AM -0700, William wrote:

>   product info: "LT WIN MODEM GLOBAL", "", "", ""

It is a WinModem.  Not supported.

-- Dave

Question Driver for DVD cards

Date: 2000, Aug 07
From: David Kaplan dmk

Hi,

A while back I tried to use a ZDVD PCMCIA card in my laptop and it didnt work at all. At the time, I was told that the problem was that it used a weird memory mode. Just recently I found out that the kernel now seems to support that video mode (I assume it is MTRR) and that people are beginning to play DVDs with linux.

I was wondering if anyone was working on a driver for this card? I havent a clue how to write one, but perhaps it wouldnt be too hard now that the kernel supports the memory mode.

Question AVM Fritz ISDN Card troubles.

Date: 2000, Aug 19
From: Andreas Fink afink

Hi Folks,

I got a strange effect. Setup is RedHat 6.2, Tosiba Portege 7200CT. Other PCMCIA devices work well (CD ROM, Xircom Network etc). I'm trying to get a Fritz Card PCMCIA ISDN Adapter (also known as AVM A1 PCMCIA) to work with my kernel. If you read the docs in the kernel and on www.isdn4linux.de, this card is fully supported. If I read the docs in pcmcia_cs, this card is not mentioned. The card is a passive ISDN card so it doesnt fit into the section "Modem cards which have their own UART". So there are differentiating statements. The german Suse distribution comes with the device listed in /etc/pcmcia/config. I copied that into my RedHat config file and adopted the modules accordingly. All modules load properly so that part works for shure.

Anyway, the situation I have now is if you insert the card you get a high beep followed by a low beep. The correct ISDN drivers are ready to be installed and get loaded properly at the insertion. However I get this message in syslog:

Aug 19 01:58:04 florida kernel: cs: IO port probe 0x0100-0x04ff: excluding 0x170-0x177 0x370-0x37f 0x4d0-0x4d7 
Aug 19 01:58:04 florida kernel: cs: IO port probe 0x0a00-0x0aff: clean. 
Aug 19 01:58:08 florida cardmgr[690]: initializing socket 0
Aug 19 01:58:08 florida kernel: cs: memory probe 0x0d0000-0x0dffff: clean. 
Aug 19 01:58:08 florida cardmgr[690]: socket 0: AVM ISDN-Controller A1
Aug 19 01:58:08 florida cardmgr[690]:   product info: "AVM", "ISDN A"
Aug 19 01:58:08 florida cardmgr[690]:   function: 6 (network)u
Aug 19 01:58:09 florida cardmgr[690]: get dev info on socket 0 failed: Resource temporarily unavailable

It seems to be that this last line is the killer. What I see off /proc is that the card has an IRQ but the io ports dont get assigned somehow. Under Windows the card gets either 0x1000 or 0x1400 (or maybe it was 0x100 and 0x140). Both areas are valid in the config. I tried various IRQ's with no difference. The BIOS of the Toshiba is set to CardBus 16. That works best and I dont have any problem with any other card.

I tried the latest kernel 2.4.0test4 with the lates PCMCIA_CS tools downloaded today and I get the same results whatsoever.

Anyone having any clue how to track this problem down to the source?

Andreas Fink

Agree I have the same problem! Opened up incident with manufacturer

Re: Question AVM Fritz ISDN Card troubles. (Andreas Fink)
Date: 2000, Aug 26
From: Marc Beuchat mbeuchat

Andreas,

I am pleased to see that I am not the only one having this problem :-(

I have also concluded that there is an IO port issue but don't know how to proceed. I opened up a request with the card manufacturer on Friday 25 Aug. Here is a transcript of the problem statement:

----------------------------------------------------------------------------

Product: FRITZ!Card PCMCIA

ISDN-Controller : FRITZ!Card PCMCIA

Operating system : linux 2.2.16 (redhat 6.2 distribution)

CAPI driver : Rev 1.12

I have a FRITZ!Card PCMCIA (v2.0) which I use with my Sony VAIO C1XN laptop. It worked well when I was running Windows 2000 on the laptop. I recently installed linux redhat 6.2 on the laptop. I compiled the kernel 2.2.16 and compiled the pcmcia cardservices package.

Everything works. I am able to use my other pcmcia cards (3com and Xircom) network and modem cards.

I manually added the FRITZ!Card drive to the pcmcia card services (fcpcmcia_cs.c) and recompiled. I configured /etc/pcmcia/config to recognize the card. I compiled the fcpcmcia.o module. I compiled capi and isdn modules into the kernel. When I insert the card, the fcpcmcia_cs.o module is unable to determine the IO port of the card. To help understand what is happening, here are the kernel messages when I start the pcmcia card services:

======================

Linux PCMCIA Card Services 3.1.19
  kernel build: 2.2.16 #4 Thu Aug 24 22:56:09 CEST 2000
  options:  [pci] [cardbus] [apm] [pnp]
PCI routing table version 1.0 at 0xfdf40
PnP: PNP BIOS installation structure at 0xc00f6d20
PnP: PNP BIOS version 1.0, entry at f0000:b334, dseg at 400
Intel PCIC probe:
  Ricoh RL5C475 rev 80 PCI-to-CardBus at slot 00:0c, mem 0x68000000
    host opts [0]: [serial irq] [io 3/6/1] [mem 3/6/1] [pci irq 9] [lat 168/176]
 [bus 32/34]
    ISA irqs (default) = 3,4,5,7,10,11,15 PCI status changes
cs: IO port probe 0x1000-0x17ff: clean.
cs: IO port probe 0x0100-0x04ff: excluding 0x170-0x177 0x220-0x22f 0x330-0x337 0x370-0x377 0x388-0x38f
cs: IO port probe 0x0a00-0x0aff: clean.

======================

and here are the kernel messages when I insert the FRITZ!Card:

======================

cs: memory probe 0xa0000000-0xa0ffffff: clean. CAPI-driver Rev 1.12: Loaded

fcpcmcia: Base adress: 0xC4040050
fcpcmcia: AVM FRITZ!Card PCMCIA driver, revision 0.0
fcpcmcia: Loading...
fcpcmcia: Driver 'fcpcmcia' attached to stack
kcapi: driver fcpcmcia attached fcpcmcia: Loaded.
fcpcmcia_cs: testing i/o 0x140-0x148
fcpcmcia_cs: testing i/o 0x300-0x308
fcpcmcia_cs: RequestIO: No more items

=======================

I was wondering if I need to manually configure the IO port in /etc/pcmcia/config.opts.

Do you have any suggestions?

Thank You.

Marc Beuchat

----------------------------------------------------------------------------

I'll keep you tuned if I hear back from them.

Marc

More odd IO request: base 0300 align 0020

Re: Agree I have the same problem! Opened up incident with manufacturer (Marc Beuchat)
Date: 2000, Aug 26
From: Marc Beuchat mbeuchat

Here is some more information concerning the problem with the FRITZ!Card PCMCIA. I compiled the pcmcia package and the fcpcmcia_cs module with debugging turned on. After inserting the card, I get the following messages:

===========================================

fcpcmcia_cs: fcpcmcia_config(0xc3f7cc80)

i82365: SetMemMap(0, 0, 0x23, 300 ns, 0xa0000000-0xa0000fff, 0x00000)

i82365: SetMemMap(0, 0, 0x23, 300 ns, 0xa0000000-0xa0000fff, 0x00000)

fcpcmcia_cs: testing i/o 0x140-0x148

odd IO request: base 0140 align 0020

i82365: SetMemMap(0, 0, 0x23, 300 ns, 0xa0000000-0xa0000fff, 0x00000)

fcpcmcia_cs: testing i/o 0x300-0x308

odd IO request: base 0300 align 0020

fcpcmcia_cs: RequestIO: No more items

============================================

I haven't had the time to figure out what these "odd" messages indicate. However they may help narrow down the problem.

Marc

Idea Bug in fcpcmcia_cs.c for FRITZ!Card pcmcia from AVM

Re: More odd IO request: base 0300 align 0020 (Marc Beuchat)
Date: 2000, Aug 27
From: Marc Beuchat mbeuchat

I did some debugging of the source code and discovered that the following:

fcpcmcia_cs: sets up the request io structure for Port1.
fcpcmcia_cs: calls CardServices() with function RequestIO and request io structure. 
cs: function RequestIO calls request_io() with the request info.
cs: request_io() calls alloc_io_space() for Port1. This is when the "odd io request" message is generated.  However the function returns normally.
cs: request_io() calls alloc_io_space() for Port2. This calls ends with an error.

request_io() calls alloc_io_space() for Port2 since fcpcmcia_cs initializes the socket structures for both ports (see line 164 and 165 of the source code in function fcpcmcia_attach). However, when it calls CardServices() with function RequestIO, it does not specify the base address for Port 2, hence the error.

To fix the problem, I commented out lines 164 and 165 in fcpcmcia_cs.c:

/*
  link->io.NumPorts2 = 16
  link->io.Attributes2 = IO_DATA_PATH_WIDTH_16
*/

With this change, the card configures correctly. Unfortunately I do not have access to an ISDN line today to test if the card actually works!! I will check this tomorrow and post the results.

I will also ask AVM support to confirm if the use of Port2 is necessary or if this is a bug. Let's see!

Marc

None Re: Idea: Bug in fcpcmcia_cs.c for FRITZ!Card pcmcia from AVM

Re: Idea Bug in fcpcmcia_cs.c for FRITZ!Card pcmcia from AVM (Marc Beuchat)
Date: 2000, Aug 27
From: Andreas Fink afink

At 06:24 27-08-2000, you wrote:
>I did some debugging of the source code and discovered that the
>following:
>
>fcpcmcia_cs: sets up the request io structure for Port1.
>fcpcmcia_cs: calls CardServices() with function RequestIO and request io structure. 
>
>cs: function RequestIO calls request_io() with the request info.
>
>cs: request_io() calls alloc_io_space() for Port1. This is when the "odd io request" message is generated.  However the function returns normally.
>cs: request_io() calls alloc_io_space() for Port2. This calls ends with an error.
>
>request_io() calls alloc_io_space() for Port2 since fcpcmcia_cs
>initializes the socket structures for both ports (see line 164 and 165
>of the source code in function fcpcmcia_attach). However, when it calls
>CardServices() with function RequestIO, it does not specify the base
>address for Port 2, hence the error.
>
>To fix the problem, I commented out lines 164 and 165 in fcpcmcia_cs.c:
>
>/*
>  link->io.NumPorts2 = 16
>  link->io.Attributes2 = IO_DATA_PATH_WIDTH_16
>*/
>
>With this change, the card configures correctly. Unfortunately I do not
>have access to an ISDN line today to test if the card actually works!! I
>will check this tomorrow and post the results.
>
>I will also ask AVM support to confirm if the use of Port2 is necessary
>or if this is a bug. Let's see!
>
>Marc

Hey good work! I was guessing that this could be something like that but I was not sure if that part is in the pcmcia driver or the card driver.

Now for the card driver, it should also be possible to use the hisax driver instead as the card is supported in the kernel. I must say that I don't like the  AVM supplied driver because its 100% based for the Suse distribution and SUSE does everything different. Just to make the developers life easier. Recompiling the drivers for my RedHat kernel was a small nightmare until I fixed the Makefile coorrectly. Install of course does put the files at the wrong place and so on. So I prefer to try to fix the hisax driver because that means its also going to work on other platforms like PowerPC or Alpha. I've heard that the card might not work with PowerPC because of the byte order. When I get home to my ISDN line I will try your fix and I will also see what happens on my PowerBook G3.





Andreas Fink

------------------------------------------------------------
Home Address: Albisstr. 48, 8932 Mettmenstetten, Switzerland
Tel: +41-1-7683534, Fax: +41-1-7683531, GSM: +41-79-2457333
------------------------------------------------------------

Sad Patch didnt work for me

Re: None Re: Idea: Bug in fcpcmcia_cs.c for FRITZ!Card pcmcia from AVM (Andreas Fink)
Date: 2000, Aug 30
From: Andreas Fink afink

The uncommenting of those lines did not work for me. 
I still get the same error messages like resource 
unavailable.

You have apparently traced through the code. Can you tell 
me which module calls which one in what sequence? 
What entries do you have in /etc/pcmcia/config?


Andreas Fink

Feedback Re: Patch didnt work for me

Re: Sad Patch didnt work for me (Andreas Fink)
Date: 2000, Aug 31
From: Marc Beuchat mbeuchat

Andreas,

Once I made the patch change, I noticed that instead of having an RequestIO error, I had a RequestIRQ error. I was using PCI interrupts (irq_mode=0). I turned this off so that ISA interrupt routing would be used. That solved the problem. Unfortunately, I still can't make an ISDN connection now that I have the card recognized.

Last night I was looking at the avm ftp site. I went into ftp://ftp.avm.de/pub/cardware/b1/linux and found all the latest stuff including patches for pcmcia-cs. I downloaded the whole think, recompiled pcmcia-cs with the avm patch. The card is now recognized without a problem. I need to compile isdn4linux and capi4linux tonight and see if it works!

I never thought to check in b1 for drivers, however it seems to be the most recent code and it supports the a1 pcmcia card. I recommend that you try this out. You will need to downgrade to pcmcia-cs 3.0.9, but I don't see that as a problem. I am sure they will soon have a patch for the more recent versions.

Marc

Question How to handle a CardBus card without manfid?

Date: 2000, Sep 01
From: Albrecht Dress Albrecht.Dress

I currently try to hack a driver for a IEEE1394/FireWire CardBus Card by Newer Technology. The card is based upon the TI PCI-Lynx chipset, and a "normal" driver is available from the ieee1394 project, which I try to use.

Unfortunately, the card is not recognised at all upon insertion. I hacked around in the cardmgr sources and added some debug info to find the cause. The problem seems to be that at the beginning of lookup_card within cardmgr.c the ioctl() call returns 0, but has_cis is also 0. The check for a funcid.func does also not work as its value is 0xff.

Maybe I should add that the card was designed for a Mac. I can see it using lspci, but all io and memory resources are disabled ("Control: I/O- Mem- BusMaster- etc.). Might this be the cause of the problem? Any other idea how I could it get being recognised?

Thanks for your help, Albrecht.

None Fixed in latest drivers

Re: Question How to handle a CardBus card without manfid? (Albrecht Dress)
Date: 2000, Sep 01
From: David Hinds <dhinds@pcmcia.sourceforge.org>

The 3.1.20 drivers will let you add something like:

  card "your card"
    pci 0x1234, 0x5678
    bind "..."

to /etc/pcmcia/config, to configure a card based on its PCI vendor and
device ID's.

-- Dave
Unsupported PCMCIA Devices


Add Message to: "Unsupported PCMCIA Devices"

Members Subscribe Admin Mode Show Frames Help for HyperNews at pcmcia-cs.sourceforge.net 1.10