PCMCIA for LC2420 - Portable

This is a discussion on PCMCIA for LC2420 - Portable ; I have a linux certified LC 2420 laptop running linux from scratch (LFS) 5.0 and I recently moved to 2.6.0 kernel. I wanted to try a Linksys WPC54G on linux to enable wireless networking but I ran into a problem. ...

+ Reply to Thread
Results 1 to 13 of 13

Thread: PCMCIA for LC2420

  1. PCMCIA for LC2420

    I have a linux certified LC 2420 laptop running linux from scratch
    (LFS) 5.0 and I recently moved to 2.6.0 kernel. I wanted to try a
    Linksys WPC54G on linux to enable wireless networking but I ran into a
    problem. PCMCIA does not seem to work. I did not need PCMCIA before
    this so I did not enable it in the previous kernels (linux-2.4.*). I
    am using pcmcia-cs-3.2.7. When I insert any card (I also tried Adaptec
    1480A slimscsi) I hear no beeps, the output of /var/log/sys.log does
    not change and
    cardctl status
    says
    Socket 0:
    no card

    I know that there is no hardware problem because the laptop is dual
    boot and on Windoze 2000 side, it says new hardware found when I
    insert any card. Also the power light comes on for WPC54G (does not
    happen for linux).

    If I look at dmesg, it looks like it detected the Cardbus bridge:
    Linux Kernel Card Services
    options: [pci] [cardbus] [pm]
    Intel PCIC probe: not found.
    Yenta: CardBus bridge found at 0000:00:08.0 [1584:3000]
    Yenta: ISA IRQ list 0000, PCI irq17
    Socket status: 4d41b401
    cs: IO port probe 0x0c00-0x0cff: excluding 0xc00-0xc1f
    cs: IO port probe 0x0800-0x08ff: excluding 0x800-0x837 0x840-0x87f
    0x898-0x89f
    cs: IO port probe 0x0100-0x04ff: excluding 0x3c0-0x3df 0x480-0x48f
    0x4d0-0x4d7
    cs: IO port probe 0x0a00-0x0aff: clean.

    The relevant lspci output is:
    00:08.0 CardBus bridge: O2 Micro, Inc. OZ6912 Cardbus Controller
    Subsystem: Uniwill Computer Corp: Unknown device 3000
    Flags: bus master, stepping, slow devsel, latency 168, IRQ 17
    Memory at 3fff0000 (32-bit, non-prefetchable) [size=4K]
    Bus: primary=00, secondary=02, subordinate=05, sec-latency=176
    Memory window 0: 40000000-403ff000 (prefetchable)
    Memory window 1: 40400000-407ff000
    I/O window 0: 00004000-000040ff
    I/O window 1: 00004400-000044ff
    16-bit legacy interface ports at 0001

    Please help!

    Amit

  2. Re: PCMCIA for LC2420

    On Fri, 02 Jan 2004 09:09:49 -0800, Amit typed:

    > I have a linux certified LC 2420 laptop running linux from scratch (LFS)
    > 5.0 and I recently moved to 2.6.0 kernel. I wanted to try a Linksys
    > WPC54G on linux to enable wireless networking but I ran into a problem.
    > PCMCIA does not seem to work.


    Did you compile PCMCIA support into the kernel?

    Did you compile/add wireless support?

    Did you install the modules-init-tools as needed for the new kernel?

    ftp://ftp.kernel.org/pub/linux/kerne...rusty/modules/

    Get the latest version and read the README, I ran(from non 2.6 kernel);

    ../configure --prefix=/
    make moveold
    make
    make install

    Now translate your old /etc/modules.conf into /etc/modprobe.conf with the
    ../generate-modprobe.conf script that comes with module-init-tools:

    ../generate-modprobe.conf /etc/modprobe.conf

    I also had to edit the /etc/modprobe.conf file USB aliases to;

    alias usb-controller ehci-hcd
    alias usb-controller1 uhci-hcd


    --
    SCO + RICO Act = Justice

    Hi! I'm a .sig virus! Copy me to your .sig!


  3. Re: PCMCIA for LC2420

    Amit wrote:
    > I have a linux certified LC 2420 laptop running linux from scratch
    > (LFS) 5.0 and I recently moved to 2.6.0 kernel. I wanted to try a
    > Linksys WPC54G on linux to enable wireless networking but I ran into a
    > problem. PCMCIA does not seem to work. I did not need PCMCIA before
    > this so I did not enable it in the previous kernels (linux-2.4.*).


    It might be worth trying with a 2.4.23 kernel to see if it behaves the
    same. PCMCIA in 2.6.0 is somewhat unstable.

    > Yenta: CardBus bridge found at 0000:00:08.0 [1584:3000]
    > Yenta: ISA IRQ list 0000, PCI irq17
    > Socket status: 4d41b401


    This status value is bogus.

    > The relevant lspci output is:
    > 00:08.0 CardBus bridge: O2 Micro, Inc. OZ6912 Cardbus Controller
    > Subsystem: Uniwill Computer Corp: Unknown device 3000
    > Flags: bus master, stepping, slow devsel, latency 168, IRQ 17
    > Memory at 3fff0000 (32-bit, non-prefetchable) [size=4K]
    > Bus: primary=00, secondary=02, subordinate=05, sec-latency=176
    > Memory window 0: 40000000-403ff000 (prefetchable)
    > Memory window 1: 40400000-407ff000
    > I/O window 0: 00004000-000040ff
    > I/O window 1: 00004400-000044ff
    > 16-bit legacy interface ports at 0001


    How much memory is in this laptop?

    -- Dave

  4. Re: PCMCIA for LC2420

    Lenard wrote in message news:...
    > On Fri, 02 Jan 2004 09:09:49 -0800, Amit typed:
    >
    > > I have a linux certified LC 2420 laptop running linux from scratch (LFS)
    > > 5.0 and I recently moved to 2.6.0 kernel. I wanted to try a Linksys
    > > WPC54G on linux to enable wireless networking but I ran into a problem.
    > > PCMCIA does not seem to work.

    >
    > Did you compile PCMCIA support into the kernel?


    yes
    >
    > Did you compile/add wireless support?
    >

    yes even though it is not needed since wpc54g work using windoze
    drivers with ndiswrapper only

    > Did you install the modules-init-tools as needed for the new kernel?
    >

    absolutely!
    > ftp://ftp.kernel.org/pub/linux/kerne...rusty/modules/
    >
    > Get the latest version and read the README, I ran(from non 2.6 kernel);
    >
    > ./configure --prefix=/
    > make moveold
    > make
    > make install
    >
    > Now translate your old /etc/modules.conf into /etc/modprobe.conf with the
    > ./generate-modprobe.conf script that comes with module-init-tools:
    >
    > ./generate-modprobe.conf /etc/modprobe.conf


    yes!
    >
    > I also had to edit the /etc/modprobe.conf file USB aliases to;
    >
    > alias usb-controller ehci-hcd
    > alias usb-controller1 uhci-hcd


  5. Re: PCMCIA for LC2420

    dhinds@sonic.net wrote in message news:<_hiJb.5895$XF6.137174@typhoon.sonic.net>...
    >
    > It might be worth trying with a 2.4.23 kernel to see if it behaves the
    > same. PCMCIA in 2.6.0 is somewhat unstable.


    will do tomorrow and let you know if 2.6.1 is not out by that time
    >
    > > Yenta: CardBus bridge found at 0000:00:08.0 [1584:3000]
    > > Yenta: ISA IRQ list 0000, PCI irq17
    > > Socket status: 4d41b401

    >
    > This status value is bogus.


    I suspected so. BTW what is the relevance of the PCIC error?

    >
    > > The relevant lspci output is:
    > > 00:08.0 CardBus bridge: O2 Micro, Inc. OZ6912 Cardbus Controller
    > > Subsystem: Uniwill Computer Corp: Unknown device 3000
    > > Flags: bus master, stepping, slow devsel, latency 168, IRQ 17
    > > Memory at 3fff0000 (32-bit, non-prefetchable) [size=4K]
    > > Bus: primary=00, secondary=02, subordinate=05, sec-latency=176
    > > Memory window 0: 40000000-403ff000 (prefetchable)
    > > Memory window 1: 40400000-407ff000
    > > I/O window 0: 00004000-000040ff
    > > I/O window 1: 00004400-000044ff
    > > 16-bit legacy interface ports at 0001

    >
    > How much memory is in this laptop?
    >


    1GB. I have highmem of 4G enabled in the kernel.

  6. Re: PCMCIA for LC2420

    Amit wrote:
    >>
    >> > Yenta: CardBus bridge found at 0000:00:08.0 [1584:3000]
    >> > Yenta: ISA IRQ list 0000, PCI irq17
    >> > Socket status: 4d41b401

    >>
    >> This status value is bogus.


    > I suspected so. BTW what is the relevance of the PCIC error?


    None whatsoever: it means you do not have an ISA-PCMCIA bridge.

    >> > 00:08.0 CardBus bridge: O2 Micro, Inc. OZ6912 Cardbus Controller
    >> > Subsystem: Uniwill Computer Corp: Unknown device 3000
    >> > Flags: bus master, stepping, slow devsel, latency 168, IRQ 17
    >> > Memory at 3fff0000 (32-bit, non-prefetchable) [size=4K]


    >> How much memory is in this laptop?


    > 1GB. I have highmem of 4G enabled in the kernel.


    Bingo. Note the memory address here of 0x3fff000: for some reason,
    your bridge's registers were mapped to overlay the top 4K of system
    memory.

    I remember this sort of problem coming up before and I'd thought it
    had been fixed. Aparently not. It is sometimes caused by the BIOS
    erroneously reporting a memory size a bit too small; you might check
    the memory map information reported in your boot logs.

    I don't know enough about the appropriate parts of the kernel to
    suggest a fix, unfortunately. You should report the bug to the
    linux-kernel mailing list. It is a bug in PCI resource allocation.

    -- Dave

  7. Re: PCMCIA for LC2420

    dhinds@sonic.net wrote:

    > I remember this sort of problem coming up before and I'd thought it
    > had been fixed. Aparently not. It is sometimes caused by the BIOS
    > erroneously reporting a memory size a bit too small; you might check
    > the memory map information reported in your boot logs.


    I checked on this and both 2.4.23 and 2.6.0 contain code to prevent
    this from happening. Can you double check that you are actually
    running the kernel you think you're running (use 'uname -a')?

    -- Dave

  8. Re: PCMCIA for LC2420

    dhinds@sonic.net wrote in message news:<87vJb.6249$XF6.140496@typhoon.sonic.net>...
    >
    > Bingo. Note the memory address here of 0x3fff000: for some reason,
    > your bridge's registers were mapped to overlay the top 4K of system
    > memory.
    >
    > I remember this sort of problem coming up before and I'd thought it
    > had been fixed. Aparently not. It is sometimes caused by the BIOS
    > erroneously reporting a memory size a bit too small; you might check
    > the memory map information reported in your boot logs.
    >
    > I don't know enough about the appropriate parts of the kernel to
    > suggest a fix, unfortunately. You should report the bug to the
    > linux-kernel mailing list. It is a bug in PCI resource allocation.
    >
    > -- Dave


    Just did it. I tried 2.4.23 and it also did not work. BTW the output
    of cat /proc/iomem is below. According to it system RAM ends at
    3ffeffff and yenta_socket begins at 3fff0000, i.e., no overlap?

    00000000-0009fbff : System RAM
    0009fc00-0009ffff : reserved
    000a0000-000bffff : Video RAM area
    000c0000-000c7fff : Video ROM
    000f0000-000fffff : System ROM
    00100000-3ffeffff : System RAM
    00100000-002c7c00 : Kernel code
    002c7c01-00389dff : Kernel data
    3fff0000-3fff0fff : 0000:00:08.0
    3fff0000-3fff0fff : yenta_socket
    40000000-403fffff : PCI CardBus #02
    40400000-407fffff : PCI CardBus #02
    c5b00000-d5cfffff : PCI Bus #01
    c8000000-cfffffff : 0000:01:00.0
    d5c80000-d5cfffff : 0000:01:00.0
    d5e00000-d7efffff : PCI Bus #01
    d6000000-d6ffffff : 0000:01:00.0
    dbffa000-dbffafff : 0000:00:04.0
    dbffa000-dbffafff : sis900
    dbffb000-dbffbfff : 0000:00:03.0
    dbffc000-dbffcfff : 0000:00:03.1
    dbffd000-dbffdfff : 0000:00:03.2
    dbffe000-dbffefff : 0000:00:03.3
    dbfff000-dbffffff : 0000:00:02.3
    dc000000-dfffffff : 0000:00:00.0

  9. Re: PCMCIA for LC2420

    Amit wrote:

    > Just did it. I tried 2.4.23 and it also did not work. BTW the output
    > of cat /proc/iomem is below. According to it system RAM ends at
    > 3ffeffff and yenta_socket begins at 3fff0000, i.e., no overlap?


    Well right, that is where the kernel thinks your system RAM ends. Do
    you think you actually bought a "1 GB minus 4K" system?

    I'm not sure why the kernel is not bumping up the yenta driver memory
    to the next 1 MB boundary like it is supposed to: that's the way this
    problem was supposed to have been fixed.

    You should probably report this on the linux-kernel mailing list.
    Maybe someone else will have a good idea. Post your /proc/iomem map
    as well as the memory map shown at the start of your boot messages; it
    will look something like this:

    > BIOS-provided physical RAM map:
    > BIOS-e820: 0000000000000000 - 000000000009f800 (usable)
    > BIOS-e820: 000000000009f800 - 00000000000a0000 (reserved)
    > BIOS-e820: 00000000000e0000 - 0000000000100000 (reserved)
    > BIOS-e820: 0000000000100000 - 000000000bff0000 (usable)
    > BIOS-e820: 000000000bff0000 - 000000000bffec00 (ACPI data)
    > BIOS-e820: 000000000bffec00 - 000000000c000000 (ACPI NVS)
    > BIOS-e820: 00000000fff80000 - 0000000100000000 (reserved)


    I suppose it is possible that your BIOS is actually misconfiguring the
    device; in that case the kernel won't notice the problem and will
    assume the BIOS did something reasonable.

    -- Dave

  10. Re: PCMCIA for LC2420

    dhinds@sonic.net wrote in message news:<5GLJb.6445$XF6.143691@typhoon.sonic.net>...
    > Amit wrote:
    >
    >
    > Well right, that is where the kernel thinks your system RAM ends. Do
    > you think you actually bought a "1 GB minus 4K" system?
    >
    > I'm not sure why the kernel is not bumping up the yenta driver memory
    > to the next 1 MB boundary like it is supposed to: that's the way this
    > problem was supposed to have been fixed.
    >
    > You should probably report this on the linux-kernel mailing list.
    > Maybe someone else will have a good idea. Post your /proc/iomem map
    > as well as the memory map shown at the start of your boot messages; it
    > will look something like this:
    >


    I just now posted this to link-kernel mailing list. It has the
    complete dmesg output. The relevant portions are:
    BIOS-provided physical RAM map:
    BIOS-e820: 0000000000000000 - 000000000009fc00 (usable)
    BIOS-e820: 000000000009fc00 - 00000000000a0000 (reserved)
    BIOS-e820: 00000000000f0000 - 0000000000100000 (reserved)
    BIOS-e820: 0000000000100000 - 000000003fff0000 (usable)
    BIOS-e820: 000000003fff0000 - 000000003fff8000 (ACPI data)
    BIOS-e820: 000000003fff8000 - 0000000040000000 (ACPI NVS)
    BIOS-e820: 00000000fec00000 - 00000000fec01000 (reserved)
    BIOS-e820: 00000000fee00000 - 00000000fee01000 (reserved)
    BIOS-e820: 00000000ffee0000 - 00000000fff0ffff (reserved)
    BIOS-e820: 00000000fff80000 - 0000000100000000 (reserved)
    user-defined physical RAM map:
    user: 0000000000000000 - 000000000009fc00 (usable)
    user: 000000000009fc00 - 00000000000a0000 (reserved)
    user: 00000000000f0000 - 0000000000100000 (reserved)
    user: 0000000000100000 - 000000003fff0000 (usable)
    127MB HIGHMEM available.
    896MB LOWMEM available.
    found SMP MP-table at 000fba70
    hm, page 000fb000 reserved twice.
    hm, page 000fc000 reserved twice.
    hm, page 000f6000 reserved twice.
    hm, page 000f7000 reserved twice.
    On node 0 totalpages: 262128
    DMA zone: 4096 pages, LIFO batch:1
    Normal zone: 225280 pages, LIFO batch:16
    HighMem zone: 32752 pages, LIFO batch:7
    DMI 2.3 present.

    Kernel command line: root=/dev/hda5 ro mem=1048512K
    Initializing CPU#0
    PID hash table entries: 4096 (order 12: 32768 bytes)
    Detected 2390.695 MHz processor.
    Console: colour VGA+ 80x25
    Memory: 1034492k/1048512k available (1589k kernel code, 13080k
    reserved, 717k data, 136k init, 131008k highmem)

  11. Re: PCMCIA for LC2420

    Amit wrote:

    > I just now posted this to link-kernel mailing list. It has the
    > complete dmesg output. The relevant portions are:
    > BIOS-provided physical RAM map:
    > BIOS-e820: 0000000000000000 - 000000000009fc00 (usable)
    > BIOS-e820: 000000000009fc00 - 00000000000a0000 (reserved)
    > BIOS-e820: 00000000000f0000 - 0000000000100000 (reserved)
    > BIOS-e820: 0000000000100000 - 000000003fff0000 (usable)
    > BIOS-e820: 000000003fff0000 - 000000003fff8000 (ACPI data)
    > BIOS-e820: 000000003fff8000 - 0000000040000000 (ACPI NVS)
    > BIOS-e820: 00000000fec00000 - 00000000fec01000 (reserved)
    > BIOS-e820: 00000000fee00000 - 00000000fee01000 (reserved)
    > BIOS-e820: 00000000ffee0000 - 00000000fff0ffff (reserved)
    > BIOS-e820: 00000000fff80000 - 0000000100000000 (reserved)
    > user-defined physical RAM map:
    > user: 0000000000000000 - 000000000009fc00 (usable)
    > user: 000000000009fc00 - 00000000000a0000 (reserved)
    > user: 00000000000f0000 - 0000000000100000 (reserved)
    > user: 0000000000100000 - 000000003fff0000 (usable)


    ?!?!?

    > Kernel command line: root=/dev/hda5 ro mem=1048512K


    Err. Can you tell me why you are booting with an option that says
    that you have 1 GB minus 4K of memory? This mem= option is the cause
    of your problem.

    -- Dave

  12. Re: PCMCIA for LC2420

    >
    > Err. Can you tell me why you are booting with an option that says
    > that you have 1 GB minus 4K of memory? This mem= option is the cause
    > of your problem.
    >
    > -- Dave


    This is what Russell King pointed to on the linux kernel mailing list.
    I was using grub as the boot loader and grub automatically determines
    the RAM size and sticks it in the boot option of the kernel. It so
    happens that it invariably gets the memory size wrong and one needs to
    compile grub with --disable-auto-linux-mem-opt to stop this
    imbecility. Now things work. I got Adaptec 1480A recognized and am
    trying to get the linksys wpc54g to work

    Thanks,

    Amit

  13. Re: PCMCIA for LC2420

    Amit wrote:
    >>
    >> Err. Can you tell me why you are booting with an option that says
    >> that you have 1 GB minus 4K of memory? This mem= option is the cause
    >> of your problem.


    > This is what Russell King pointed to on the linux kernel mailing list.
    > I was using grub as the boot loader and grub automatically determines
    > the RAM size and sticks it in the boot option of the kernel. It so
    > happens that it invariably gets the memory size wrong and one needs to
    > compile grub with --disable-auto-linux-mem-opt to stop this
    > imbecility.


    Argh!! Someone needs to tell them that it is unacceptable to
    automatically override the kernel unless you know what you're
    doing is correct!

    -- Dave

+ Reply to Thread