USB MSD access freezes computer - OS2

This is a discussion on USB MSD access freezes computer - OS2 ; Hello, I'm trying to track down a nasty problem: accessing USB mass storage devices randomly freezes the computer (ThinkPad R51, USB 2.0). The USB-HD access light stays on. No mouse movement, no Ctrl-Alt-Del, I have to power off of the ...

+ Reply to Thread
Results 1 to 12 of 12

Thread: USB MSD access freezes computer

  1. USB MSD access freezes computer

    Hello,

    I'm trying to track down a nasty problem: accessing USB mass storage
    devices randomly freezes the computer (ThinkPad R51, USB 2.0). The
    USB-HD access light stays on. No mouse movement, no Ctrl-Alt-Del, I have
    to power off of the computer.

    So far I've found the following:
    * It's not a problem of a particular device. I have a memory stick, an
    enternal HD and a card reader. Freezes happen with all three devices.
    * It's not a problem of the computer hardware. All devices work
    flawlessly with Windows.
    * It's not dependent on the file system. It happens with FAT and HPFS.
    * It even happens when reading the sectors of a physikal disks by means
    of IOCtl API. Sometimes after reading just a few sectors, sometimes
    after thousands.
    * It also happens when no other device is connected to the USB.
    * Non-MSD devices (Mouse, Printer, Scanner) do not cause freezes.
    * It happens with USBMSD.ADD and CWUSBMSD.ADD (the later one is old but
    it detects the card reader, the IBM version does not)

    I assume the problem must be somewhere in the MSD device driver chain.
    Question: which drivers are involved?
    USBUHCD.SYS (here: BldLevel 10.158)
    USBEHCD.SYS (10.84)
    USBD.SYS (10.158)
    USBMSD.ADD (10.145) or CWUSBMSD.ADD (9.23)
    DANIDASD.DMD (Revision: 1.04, File Version: 1.4.4)

    Anything else?

    I don't have access to newer versions because I don't have eCS or PA.
    However, I wouldn't mind if someone would send me his/her reliable set
    of drivers ;-) E-mail address is valid.

    Another idea: could a IRQ conflict be the cause of the freezes? "rmview
    /irq" shows USB and WLAN on IRQ 11. Switching off WLAN did not make any
    difference.

    Kind regards
    Stefan




  2. Re: USB MSD access freezes computer

    > I'm trying to track down a nasty problem: accessing USB mass storage
    > devices randomly freezes the computer (ThinkPad R51, USB 2.0). The
    > USB-HD access light stays on. No mouse movement, no Ctrl-Alt-Del, I have
    > to power off of the computer.

    ....
    >
    > Another idea: could a IRQ conflict be the cause of the freezes? "rmview
    > /irq" shows USB and WLAN on IRQ 11. Switching off WLAN did not make any
    > difference.


    Yes, I suspect an IRQ issue. I had the same problem at the beginning.
    When plugging in any USB device, the system froze.

    BIOS switches did not help, but shuffling around the relevant cards.
    Well, this is not an option for you.

    rmview usually does not show all hardware, just the ones resource manager
    compatible drivers are loaded for. Run pci.exe (look on hobbes for instance)
    instead. It reports all PCI devices and the IRQs they use.

    In my case it was a conflict between the SCSI controller and USB controllers
    which where not reported by rmview. The SCSI controller was not happy with
    IRQ sharing.

  3. Re: USB MSD access freezes computer

    Hi Stefan

    Stefan Pelz wrote:
    > Hello,
    >
    > I'm trying to track down a nasty problem: accessing USB mass storage
    > devices randomly freezes the computer (ThinkPad R51, USB 2.0). The
    > USB-HD access light stays on. No mouse movement, no Ctrl-Alt-Del, I have
    > to power off of the computer.
    >
    > So far I've found the following:
    > * It's not a problem of a particular device. I have a memory stick, an
    > enternal HD and a card reader. Freezes happen with all three devices.
    > * It's not a problem of the computer hardware. All devices work
    > flawlessly with Windows.
    > * It's not dependent on the file system. It happens with FAT and HPFS.
    > * It even happens when reading the sectors of a physikal disks by means
    > of IOCtl API. Sometimes after reading just a few sectors, sometimes
    > after thousands.
    > * It also happens when no other device is connected to the USB.
    > * Non-MSD devices (Mouse, Printer, Scanner) do not cause freezes.
    > * It happens with USBMSD.ADD and CWUSBMSD.ADD (the later one is old but
    > it detects the card reader, the IBM version does not)
    >
    > I assume the problem must be somewhere in the MSD device driver chain.
    > Question: which drivers are involved?
    > USBUHCD.SYS (here: BldLevel 10.158)
    > USBEHCD.SYS (10.84)
    > USBD.SYS (10.158)
    > USBMSD.ADD (10.145) or CWUSBMSD.ADD (9.23)
    > DANIDASD.DMD (Revision: 1.04, File Version: 1.4.4)
    >
    > Anything else?
    >
    > I don't have access to newer versions because I don't have eCS or PA.
    > However, I wouldn't mind if someone would send me his/her reliable set
    > of drivers ;-) E-mail address is valid.
    >
    > Another idea: could a IRQ conflict be the cause of the freezes? "rmview
    > /irq" shows USB and WLAN on IRQ 11. Switching off WLAN did not make any
    > difference.
    >
    > Kind regards
    > Stefan
    >
    >
    >



    I suspect your problems stem from the older usb drivers in use,
    danisdasd and possibly irq conflict(s).

    Ideally you need the current IBM USB drivers and os2dasd.dmd which may
    help with the irq problem - if there is one.

    Maybe some nice, kind person will send you the most recent USB drivers
    along with os2dasd.dmd - check your email tomorrow :-)

    Something else that may be of help is USBcfg; there is a USB Removables
    FAQ included in the Help file that could be of interest. USBcfg is here
    http://hobbes.nmsu.edu/pub/os2/system/usbcfgb7a.zip

    Regards

    Pete

  4. Re: USB MSD access freezes computer

    On Wed, 21 Mar 2007 22:23:35 +0100, Heiko Nitzsche wrote:

    >rmview usually does not show all hardware, just the ones resource manager
    >compatible drivers are loaded for. Run pci.exe (look on hobbes for instance)
    >instead. It reports all PCI devices and the IRQs they use.


    Hallo Heiko,

    thanks for the suggestion, please find below the output of pci.exe. For
    better readability I've shortened the output and removed all devices
    without any IRQ (bridges).

    Regarding USB, pci.exe reports 3 UHCI and 1 EHCI controllers. How does
    that fit to the 2 USB ports the machine has? Perhaps the 3rd controller
    ist only used together with the docking station?

    I'm a bit surprised about the reported connections of system IRQs and
    PCI INTs. PCI has four INT lines A to D. I used to think that each line
    gets connected to one system IRQ. The output for the USB devices shows
    that all four lines are connected to IRQ 11. However, the SMBus, audio
    and modem controllers are on INT B but connected to IRQ 10? Something
    wrong here?

    Assuming that the output of pci.exe is ok, the following devices are
    using IRQ 11:
    * all USB controllers
    * Graphics controller
    * CardBus controller
    * FireWire controller
    * Ethernet controller
    * WLAN controller

    FireWire has no drivers and AFAIK modern graphic controllers do not use
    the IRQ. Therefore the USB controllers seem to share their IRQ with
    CardBus and the two network controllers. I'll create a test environment
    without CardBus and MAC drivers ...

    If I recall right, the BIOS allows to configure IRQ / INT settings.
    Maybe it's worth playing with these settings. However, if all devices
    which use the same INT will also get the same IRQ, then USB, CardBus and
    MACs will stay on the same IRQ anyway.

    BTW: The USB controllers are using different INTs. Does that mean that
    depending on the INT/IRQ configuration the USB ports could use different
    IRQs? Simply plugging a device into the other port could help resolving
    a IRQ conflict? And how does the EHCI controller come into the scenario?
    A USB 2.0 device will use two controllers and two IRQs or what?

    Kind Regards
    Stefan


    ------------------------------------------
    PCI BIOS Version 2.10 found!
    Number of PCI Busses : 6
    PCI Characteristics : Config Mechanism 1
    Searching for PCI Devices using the OEMHLP$ driver

    Bus 0 (PCI):
    Intel 82801DB/DBL USB UHCI Controller #1 (ICH4/ICH4-L A1 step)
    System IRQ 11, INT# A

    Intel 82801DB/DBL USB UHCI Controller #2 (ICH4/ICH4-L A1 step)
    System IRQ 11, INT# B

    Intel 82801DB/DBL USB UHCI Controller #3 (ICH4/ICH4-L A1 step)
    System IRQ 11, INT# C

    Intel 82801DB/DBL USB 2.0 EHCI Controller (ICH4/ICH4-L A1 step)
    System IRQ 11, INT# D

    Intel 82801DBM (ICH4-M) UltraATA/100 EIDE Controller
    System IRQ (disabled), INT# A

    Intel 82801DB/DBL SMBus Controller (ICH4/ICH4-L A1 step)
    System IRQ 10, INT# B

    Intel 82801DB/DBL AC'97 Audio Controller (ICH4/ICH4-L A1 step)
    System IRQ 10, INT# B

    Intel 82801DB/DBL AC'97 Modem Controller (ICH4/ICH4-L A1 step)
    System IRQ 10, INT# B

    Bus 1 (AGP):
    ATI Technologies Inc, Mobility Radeon 9000
    System IRQ 11, INT# A

    Bus 2 (PCI):
    Texas Instruments, PCI4520 PC Card CardBus Controller
    System IRQ 11, INT# A

    Texas Instruments, Device 802Ah Unknown (-> FireWire)
    System IRQ 11, INT# C

    Intel Corporation, 82540EP Gigabit Ethernet Controller (Mobile)
    System IRQ 11, INT# A

    Intel Corporation, PRO/Wireless 2200BG Network Connection
    System IRQ 11, INT# A


    ROM PCI IRQ routing table Tests....
    ROM IRQ routing table found at F000hEA0h
    Table Version 1.0 - OK
    Table size 272 bytes - OK
    Table Checksum 5Dh - OK
    PCI Interrupt Router: 82801DBM (ICH4-M) LPC Interface Bridge
    IRQ's dedicated to PCI : None
    The ROM PCI IRQ routing table appears to be OK.

    IRQ Summary: IRQs 10,11,14,15 are used by PCI devices
    Shared IRQs: IRQ 10 is shared by 3 PCI Devices
    IRQ 11 is shared by 9 PCI Devices




  5. Re: USB MSD access freezes computer

    Hi Stefan

    Stefan Pelz wrote:
    > On Wed, 21 Mar 2007 22:23:35 +0100, Heiko Nitzsche wrote:
    >
    >> rmview usually does not show all hardware, just the ones resource manager
    >> compatible drivers are loaded for. Run pci.exe (look on hobbes for instance)
    >> instead. It reports all PCI devices and the IRQs they use.

    >
    > Hallo Heiko,
    >
    > thanks for the suggestion, please find below the output of pci.exe. For
    > better readability I've shortened the output and removed all devices
    > without any IRQ (bridges).
    >
    > Regarding USB, pci.exe reports 3 UHCI and 1 EHCI controllers. How does
    > that fit to the 2 USB ports the machine has? Perhaps the 3rd controller
    > ist only used together with the docking station?
    >



    And what would you guess the 4th controller is for? :-)


    The number of USB Ports is not really related to the number of
    Controllers except for the fact that if there were no Controllers the
    ports would not work...

    EHCI is USB2
    UHCI is USB1.x

    If you need more than 2 USB ports get a 4 port USB Hub and attach it to
    a Port. You will then have 5 free USB ports - but still only 4 Controllers.


    At the risk of further confusion: My system has 1 EHCI and 1 OHCI, total
    2 USB Controllers. I currently have 7 USB ports available on the system
    case and am awaiting delivery of a monitor that includes a 4 port USB
    hub giving a potential for 10 available ports when the monitor is
    installed. Still only the 2 Controllers though.



    > I'm a bit surprised about the reported connections of system IRQs and
    > PCI INTs. PCI has four INT lines A to D. I used to think that each line
    > gets connected to one system IRQ. The output for the USB devices shows
    > that all four lines are connected to IRQ 11. However, the SMBus, audio
    > and modem controllers are on INT B but connected to IRQ 10? Something
    > wrong here?
    >
    > Assuming that the output of pci.exe is ok, the following devices are
    > using IRQ 11:
    > * all USB controllers
    > * Graphics controller
    > * CardBus controller
    > * FireWire controller
    > * Ethernet controller
    > * WLAN controller
    >
    > FireWire has no drivers and AFAIK modern graphic controllers do not use
    > the IRQ. Therefore the USB controllers seem to share their IRQ with
    > CardBus and the two network controllers. I'll create a test environment
    > without CardBus and MAC drivers ...
    >
    > If I recall right, the BIOS allows to configure IRQ / INT settings.
    > Maybe it's worth playing with these settings. However, if all devices
    > which use the same INT will also get the same IRQ, then USB, CardBus and
    > MACs will stay on the same IRQ anyway.
    >



    I think you just need to ensure that all drivers sharing an IRQ can
    share successfully. I think some earlier drivers - nic, audio and usb -
    have problems with IRQ sharing.


    > BTW: The USB controllers are using different INTs. Does that mean that
    > depending on the INT/IRQ configuration the USB ports could use different
    > IRQs? Simply plugging a device into the other port could help resolving
    > a IRQ conflict? And how does the EHCI controller come into the scenario?
    > A USB 2.0 device will use two controllers and two IRQs or what?
    >



    No, USB2 does not use 2 Controllers (you only have 1 fitted) or 2 IRQs.

    If you plug in a USB2 device the USB2 Controller will be responsible for
    control. If the device is USB1.x then the USB1 Controller connected to
    the hub with that port attached will take control of the device.


    Regards

    Pete




    > Kind Regards
    > Stefan
    >
    >
    > ------------------------------------------
    > PCI BIOS Version 2.10 found!
    > Number of PCI Busses : 6
    > PCI Characteristics : Config Mechanism 1
    > Searching for PCI Devices using the OEMHLP$ driver
    >
    > Bus 0 (PCI):
    > Intel 82801DB/DBL USB UHCI Controller #1 (ICH4/ICH4-L A1 step)
    > System IRQ 11, INT# A
    >
    > Intel 82801DB/DBL USB UHCI Controller #2 (ICH4/ICH4-L A1 step)
    > System IRQ 11, INT# B
    >
    > Intel 82801DB/DBL USB UHCI Controller #3 (ICH4/ICH4-L A1 step)
    > System IRQ 11, INT# C
    >
    > Intel 82801DB/DBL USB 2.0 EHCI Controller (ICH4/ICH4-L A1 step)
    > System IRQ 11, INT# D
    >
    > Intel 82801DBM (ICH4-M) UltraATA/100 EIDE Controller
    > System IRQ (disabled), INT# A
    >
    > Intel 82801DB/DBL SMBus Controller (ICH4/ICH4-L A1 step)
    > System IRQ 10, INT# B
    >
    > Intel 82801DB/DBL AC'97 Audio Controller (ICH4/ICH4-L A1 step)
    > System IRQ 10, INT# B
    >
    > Intel 82801DB/DBL AC'97 Modem Controller (ICH4/ICH4-L A1 step)
    > System IRQ 10, INT# B
    >
    > Bus 1 (AGP):
    > ATI Technologies Inc, Mobility Radeon 9000
    > System IRQ 11, INT# A
    >
    > Bus 2 (PCI):
    > Texas Instruments, PCI4520 PC Card CardBus Controller
    > System IRQ 11, INT# A
    >
    > Texas Instruments, Device 802Ah Unknown (-> FireWire)
    > System IRQ 11, INT# C
    >
    > Intel Corporation, 82540EP Gigabit Ethernet Controller (Mobile)
    > System IRQ 11, INT# A
    >
    > Intel Corporation, PRO/Wireless 2200BG Network Connection
    > System IRQ 11, INT# A
    >
    >
    > ROM PCI IRQ routing table Tests....
    > ROM IRQ routing table found at F000hEA0h
    > Table Version 1.0 - OK
    > Table size 272 bytes - OK
    > Table Checksum 5Dh - OK
    > PCI Interrupt Router: 82801DBM (ICH4-M) LPC Interface Bridge
    > IRQ's dedicated to PCI : None
    > The ROM PCI IRQ routing table appears to be OK.
    >
    > IRQ Summary: IRQs 10,11,14,15 are used by PCI devices
    > Shared IRQs: IRQ 10 is shared by 3 PCI Devices
    > IRQ 11 is shared by 9 PCI Devices
    >
    >
    >


  6. Re: USB MSD access freezes computer

    >> thanks for the suggestion, please find below the output of pci.exe. For
    >> better readability I've shortened the output and removed all devices
    >> without any IRQ (bridges).
    >>
    >> Regarding USB, pci.exe reports 3 UHCI and 1 EHCI controllers. How does
    >> that fit to the 2 USB ports the machine has? Perhaps the 3rd controller
    >> ist only used together with the docking station?


    1 USB controller controls at least 1 USB port (can be more).
    If you have just 2 ports routed outside, the 3rd controller
    is probably only available via the docking station. This is
    also common for PC mainboards where you often have to install
    additional slot ports via cable (usually shipped with the board)
    to use the additional USB controllers.
    Mine has for instance 3 USB ports controlled by 2 USB controllers
    directly on the ATX panel. 4 additional ports controlled by the
    other 2 controllers can be accessed by a slot panel with 4 ports.

    The EHCI controller implements the USB 2.0 interface to the
    USB controllers (those reported as UHCI). It is an add-on controller
    that controls ALL other USB controllers (speed/protocol selection).
    Thus it only needs one IRQ. If the EHCI driver is not loaded, only
    USB 1.x speed can be used.

    In your case the EHCI driver is quite old while the UHCI driver
    are relatively new. Usually these drivers must match in version.
    So possible the problem arrises due to this. You can REM the EHCI
    driver and test again. The device should then work in USB 1.x mode.

    For the newer USB base drivers there is a newer kernel loader (os2ldr)
    required. I think this had to do with IRQ sharing, but I'm not sure.

    Under Windows 2000/XP you can have a look at the USB controllers
    and the currently attached devices in the Hardware Manager. Go
    to "USB Controller" and expand the list. Each controller is reported
    and also all USB hubs are shown (named as USB-Root-Hub). When you
    open the properties of the root hubs, you can see on the power
    management page the connected USB devices. This helps finding the
    best distribution of devices performance and power wise. A root hub
    does not necessarily relate to one USB port. One root hub may have
    multiple USB ports (I have 5 root hubs but 8 USB ports and 4 controllers).

    > At the risk of further confusion: My system has 1 EHCI and 1 OHCI, total
    > 2 USB Controllers. I currently have 7 USB ports available on the system
    > case and am awaiting delivery of a monitor that includes a 4 port USB
    > hub giving a potential for 10 available ports when the monitor is
    > installed. Still only the 2 Controllers though.


    Yes, but be aware that if you run multiple devices in parallel on
    the same controller via a hub, they will share the bandwidth. It's
    the same scenario as with a network hub.

    >> Assuming that the output of pci.exe is ok, the following devices are
    >> using IRQ 11:
    >> * all USB controllers
    >> * Graphics controller
    >> * CardBus controller
    >> * FireWire controller
    >> * Ethernet controller
    >> * WLAN controller
    >>
    >> FireWire has no drivers and AFAIK modern graphic controllers do not use
    >> the IRQ. Therefore the USB controllers seem to share their IRQ with
    >> CardBus and the two network controllers. I'll create a test environment
    >> without CardBus and MAC drivers ...
    >>
    >> If I recall right, the BIOS allows to configure IRQ / INT settings.
    >> Maybe it's worth playing with these settings. However, if all devices
    >> which use the same INT will also get the same IRQ, then USB, CardBus and
    >> MACs will stay on the same IRQ anyway.


    Usually the mainboard manual contains a table showing the wiring
    dependencies of INT A-D and their mapping to IRQs. I guess the USB
    controllers are all integrated in the Intel southbridge and thus
    probably always use the same IRQ. The USB drivers usually work fine
    with this (at least the newer ones).

    >
    > I think you just need to ensure that all drivers sharing an IRQ can
    > share successfully. I think some earlier drivers - nic, audio and usb -
    > have problems with IRQ sharing.


    Yeah, this is a good starting point.

    You can also try to disable all other devices in BIOS and enable
    one after the other to find the black sheep.

    >> ------------------------------------------
    >> PCI BIOS Version 2.10 found!
    >> Number of PCI Busses : 6
    >> PCI Characteristics : Config Mechanism 1 Searching for PCI Devices
    >> using the OEMHLP$ driver
    >>
    >> Bus 0 (PCI):
    >> Intel 82801DB/DBL USB UHCI Controller #1 (ICH4/ICH4-L A1 step)
    >> System IRQ 11, INT# A
    >>
    >> Intel 82801DB/DBL USB UHCI Controller #2 (ICH4/ICH4-L A1 step)
    >> System IRQ 11, INT# B
    >>
    >> Intel 82801DB/DBL USB UHCI Controller #3 (ICH4/ICH4-L A1 step)
    >> System IRQ 11, INT# C
    >>
    >> Intel 82801DB/DBL USB 2.0 EHCI Controller (ICH4/ICH4-L A1 step)
    >> System IRQ 11, INT# D
    >>
    >> Intel 82801DBM (ICH4-M) UltraATA/100 EIDE Controller
    >> System IRQ (disabled), INT# A
    >>
    >> Intel 82801DB/DBL SMBus Controller (ICH4/ICH4-L A1 step)
    >> System IRQ 10, INT# B
    >>
    >> Intel 82801DB/DBL AC'97 Audio Controller (ICH4/ICH4-L A1 step)
    >> System IRQ 10, INT# B
    >>
    >> Intel 82801DB/DBL AC'97 Modem Controller (ICH4/ICH4-L A1 step)
    >> System IRQ 10, INT# B
    >>
    >> Bus 1 (AGP):
    >> ATI Technologies Inc, Mobility Radeon 9000
    >> System IRQ 11, INT# A
    >>
    >> Bus 2 (PCI):
    >> Texas Instruments, PCI4520 PC Card CardBus Controller
    >> System IRQ 11, INT# A
    >>
    >> Texas Instruments, Device 802Ah Unknown (-> FireWire)
    >> System IRQ 11, INT# C
    >>
    >> Intel Corporation, 82540EP Gigabit Ethernet Controller (Mobile)
    >> System IRQ 11, INT# A
    >>
    >> Intel Corporation, PRO/Wireless 2200BG Network Connection
    >> System IRQ 11, INT# A
    >>
    >>
    >> ROM PCI IRQ routing table Tests....
    >> ROM IRQ routing table found at F000hEA0h
    >> Table Version 1.0 - OK
    >> Table size 272 bytes - OK
    >> Table Checksum 5Dh - OK
    >> PCI Interrupt Router: 82801DBM (ICH4-M) LPC Interface Bridge
    >> IRQ's dedicated to PCI : None
    >> The ROM PCI IRQ routing table appears to be OK.
    >>
    >> IRQ Summary: IRQs 10,11,14,15 are used by PCI devices
    >> Shared IRQs: IRQ 10 is shared by 3 PCI Devices
    >> IRQ 11 is shared by 9 PCI Devices


  7. Re: USB MSD access freezes computer

    Peter Brown wrote:
    > Hi Stefan
    >
    > Stefan Pelz wrote:
    >> On Wed, 21 Mar 2007 22:23:35 +0100, Heiko Nitzsche wrote:
    >>
    >>> rmview usually does not show all hardware, just the ones resource
    >>> manager
    >>> compatible drivers are loaded for. Run pci.exe (look on hobbes for
    >>> instance)
    >>> instead. It reports all PCI devices and the IRQs they use.

    >>
    >> Hallo Heiko,
    >>
    >> thanks for the suggestion, please find below the output of pci.exe. For
    >> better readability I've shortened the output and removed all devices
    >> without any IRQ (bridges).
    >>
    >> Regarding USB, pci.exe reports 3 UHCI and 1 EHCI controllers. How does
    >> that fit to the 2 USB ports the machine has? Perhaps the 3rd controller
    >> ist only used together with the docking station?
    >>

    >
    >
    > And what would you guess the 4th controller is for? :-)
    >
    >
    > The number of USB Ports is not really related to the number of
    > Controllers except for the fact that if there were no Controllers the
    > ports would not work...
    >
    > EHCI is USB2
    > UHCI is USB1.x
    >
    > If you need more than 2 USB ports get a 4 port USB Hub and attach it to
    > a Port. You will then have 5 free USB ports - but still only 4 Controllers.
    >
    >
    > At the risk of further confusion: My system has 1 EHCI and 1 OHCI, total
    > 2 USB Controllers. I currently have 7 USB ports available on the system
    > case and am awaiting delivery of a monitor that includes a 4 port USB
    > hub giving a potential for 10 available ports when the monitor is
    > installed. Still only the 2 Controllers though.


    And at the risk of even more confusion: on my JVC741 laptop there are 3
    UHCI controllers + 1 EHCI, yet the machine has only two USB slots (and
    no, there's no docking station option). So what purpose could the third
    UHCI controller serve...?

    >> I'm a bit surprised about the reported connections of system IRQs and
    >> PCI INTs. PCI has four INT lines A to D. I used to think that each line
    >> gets connected to one system IRQ. The output for the USB devices shows
    >> that all four lines are connected to IRQ 11. However, the SMBus, audio
    >> and modem controllers are on INT B but connected to IRQ 10? Something
    >> wrong here?


    Not necessarily.
    On my JVC laptop, many PCI devices seem to be connected to just one, two
    or three (rather than 4) INT lines. It's simply the wiring, nothing to
    be done about it.
    IIRC all PCI stuff (4 X USB, 2 X graphics, Firewire, sound, modem,
    Cardbus, LAN, WLAN & I surely forgot one or two) is connected to IRQ 10
    or 11 by default on OS/2.
    Under Windows, many more IRQs are available because that can use the
    APIC and ACPI, which OS/2 cannot (yet reliably).

    Try to use SPCIIRQ.SYS (hobbes) to change IRQ settings of PCI devices.
    (On my JVC 741 I found that SPCIIRQ.SYS can be used for changing the IRQ
    of just one PCI device. YMMV.)

    As other posters have suggested, I think upgrading USB drivers & os2ldr
    might help.

    Philip

  8. Re: USB MSD access freezes computer

    > And at the risk of even more confusion: on my JVC741 laptop there are 3
    > UHCI controllers + 1 EHCI, yet the machine has only two USB slots (and
    > no, there's no docking station option). So what purpose could the third
    > UHCI controller serve...?


    Do you have a built-in cardreader?

    You can check under Windows XP with Device Manager what is connected there.
    (see my other post)

  9. Re: USB MSD access freezes computer

    Heiko Nitzsche wrote:
    >> And at the risk of even more confusion: on my JVC741 laptop there are
    >> 3 UHCI controllers + 1 EHCI, yet the machine has only two USB slots
    >> (and no, there's no docking station option). So what purpose could the
    >> third UHCI controller serve...?

    >
    > Do you have a built-in cardreader?


    No.

    > You can check under Windows XP with Device Manager what is connected there.
    > (see my other post)


    That doesn't help much, it even adds to the confusion :-)

    My JVC got 4 root hubs. Two of them are connected to the physical USB
    ports (could be checked when I've had something plugged in the physical
    USB ports). The other two root hubs show no connection at all.
    So, I'm still puzzled (but not worried) why I have 3 UHCI controllers &
    4 root hubs, but only two physical USB ports.

    Other, similar JVC models do have built-in card readers etc. So it may
    well be possible that JVC (actually Asus) designers just swapped in some
    mobo they had lying around anyway. "Too bad for the superfluous mobo
    connections..."
    (Makes some sense as the BIOS reports two IDE controllers too, each with
    master + slave. Heaven knows how the second IDE ctlr (or the primary
    slave) can be connected to anything.)

    Thanks anyway,

    Philip

  10. Re: USB MSD access freezes computer

    Sir:

    Philip Nienhuis wrote:
    > Heiko Nitzsche wrote:
    >>> And at the risk of even more confusion: on my JVC741 laptop there are
    >>> 3 UHCI controllers + 1 EHCI, yet the machine has only two USB slots
    >>> (and no, there's no docking station option). So what purpose could
    >>> the third UHCI controller serve...?

    >>
    >> Do you have a built-in cardreader?

    >
    > No.
    >
    >> You can check under Windows XP with Device Manager what is connected
    >> there.
    >> (see my other post)

    >
    > That doesn't help much, it even adds to the confusion :-)
    >
    > My JVC got 4 root hubs. Two of them are connected to the physical USB
    > ports (could be checked when I've had something plugged in the physical
    > USB ports). The other two root hubs show no connection at all.
    > So, I'm still puzzled (but not worried) why I have 3 UHCI controllers &
    > 4 root hubs, but only two physical USB ports.
    >
    > Other, similar JVC models do have built-in card readers etc. So it may
    > well be possible that JVC (actually Asus) designers just swapped in some
    > mobo they had lying around anyway. "Too bad for the superfluous mobo
    > connections..."
    > (Makes some sense as the BIOS reports two IDE controllers too, each with
    > master + slave. Heaven knows how the second IDE ctlr (or the primary
    > slave) can be connected to anything.)


    If you were to get the manual for the main board, you'll find that there
    are USB ports (pin connectors) on the main board for additional external
    ports, you supply the wire & hardware. I've seen a pair for a front
    mounted panel and another pair for internal USB devices like memory card
    readers and IR ports, or for additional rear ports. If you know this,
    why ask the questions in this thread about them?

    BTW, USB hubs do not show up in reports, as they are nothing more than
    wire connectors, usually. Hubs can be root of their own chain of ports,
    but since each chain can have about 255 ports, what is the point of
    adding such a root device? Now know that Hubs are not controllers in
    the sense of UHCI, EHCI, OHCI devices are controllers. Controllers have
    the distinction of having a different bus on one side like PCI. Thus,
    your JVC board has four controllers, not four hubs. One of the
    controllers is an EHCI controller, whose job is to control other USB
    controllers. Then you have three UHCI controllers, each has up to two
    ports attached (out of the about 255 ports available to be attached).
    One UHCI controller has two ports currently attached and the others have
    pin connectors somewhere on the main board for additional ports. You
    can run from a command prompt :> PCI.exe > some.file and then open
    some.file with a reader like the system editor to see what I say is
    correct. If you don't have PCI.exe, you can get it here
    and an updated
    device list from the original author's web site
    (replace
    the one in the first package).

    PS. None of this fixes the original problem that started this thread,
    whose answer is purchase newer drivers by buying latest eCS.
    --
    Bill
    Thanks a Million!

  11. Re: USB MSD access freezes computer

    William L. Hartzell wrote:
    > Sir:
    >
    > Philip Nienhuis wrote:
    >> Heiko Nitzsche wrote:
    >>>> And at the risk of even more confusion: on my JVC741 laptop there
    >>>> are 3 UHCI controllers + 1 EHCI, yet the machine has only two USB
    >>>> slots (and no, there's no docking station option). So what purpose
    >>>> could the third UHCI controller serve...?
    >>>
    >>> Do you have a built-in cardreader?

    >>
    >> No.
    >>
    >>> You can check under Windows XP with Device Manager what is connected
    >>> there.
    >>> (see my other post)

    >>
    >> That doesn't help much, it even adds to the confusion :-)
    >>
    >> My JVC got 4 root hubs. Two of them are connected to the physical USB
    >> ports (could be checked when I've had something plugged in the
    >> physical USB ports). The other two root hubs show no connection at all.
    >> So, I'm still puzzled (but not worried) why I have 3 UHCI controllers
    >> & 4 root hubs, but only two physical USB ports.
    >>
    >> Other, similar JVC models do have built-in card readers etc. So it may
    >> well be possible that JVC (actually Asus) designers just swapped in
    >> some mobo they had lying around anyway. "Too bad for the superfluous
    >> mobo connections..."
    >> (Makes some sense as the BIOS reports two IDE controllers too, each
    >> with master + slave. Heaven knows how the second IDE ctlr (or the
    >> primary slave) can be connected to anything.)

    >
    > If you were to get the manual for the main board, you'll find that there
    > are USB ports (pin connectors) on the main board for additional external
    > ports, you supply the wire & hardware. I've seen a pair for a front
    > mounted panel and another pair for internal USB devices like memory card
    > readers and IR ports, or for additional rear ports. If you know this,
    > why ask the questions in this thread about them?


    Note that I was writing about a mini-laptop (see 2nd line of my previous
    posting or 6th line here), not a desktop mobo where you can easily
    attach any cable you like.
    Other than that, you've hit the nail on the head :-)
    Just to be sure I'll mention that I have had this laptop completely
    taken apart to be able to change the HD: no sign of extra USB pin
    connectors on the mobo.

    BTW if I only could get that "manual for the main board" .... (JVC
    MP-XP741 = ASUS S200, if you happen to stumble on it some day.)

    Oh, and I wrote in response to the OP who wondered why he had 3 UHCI
    controllers and just 2 physical USB ports on his laptop (like me).

    Philip

  12. Re: USB MSD access freezes computer - solved

    On Wed, 21 Mar 2007 23:45:37 GMT, Peter Brown wrote:

    >I suspect your problems stem from the older usb drivers in use,
    >danisdasd and possibly irq conflict(s).


    Meanwhile, I've upgraded the USB drivers and the problems are gone. I've transfered about 10GB by now, without any
    problems. However, the data transfer rate of the HD is a bit disapointing (Its a 2,5" HD in a very slim USB case). SysBench,
    DFSee and a selfmade program all deliver around 4.2 MB/s.

    >Ideally you need the current IBM USB drivers and os2dasd.dmd which may
    >help with the irq problem - if there is one.


    I'm still using danidasd.dmd. I'm hesitating to switch to os2dasd.dmd because - if I recall right - I tried this one or two years
    ago and couldn't boot anymore (the famous "cannot operate harddisk" message). Maybe I should give it another try?

    Perhaps this would also help with the device detection problem? See other post.

    Kind regards
    Stefan




+ Reply to Thread