My thanks to the several people who responded to my plea a few days ago
for help with DittoMax/parallel under Linux 2.2.12. I've made progress,
but not quite enough. Here's what I've learned (and what I still need to
learn):
- Because the parallel ports in the Compaq DeskPro do not support a true
EPP mode, I needed to purchase a separate card. I got a Siig PCI-based
parallel port. I had to use a DOS utility from Siig to put it into EPP
mode.
- After the port was installed and appropriate IO/IRQ values set at driver
installation, the bpck-fdc driver was unable to find the tape drive... in
fact it was unable to use the port at all. Attempts to look for the drive
resulted in an "unable to get hardware to register bpck fdc" message. Turns
out that the driver doesn't like to share interrupts, whereas normal behavior
for PCI devices is to all share a single interrupt.
- Using the /sbin/setpci utility, I changed the IRQ for the card to a
unique value. Now I'm starting to see hints of success. When I install
the drivers, log messages from the bpck-fdc driver report look promising;
I see messages like this:
bpck-fdc.c (bpck_fdc_probe) - Bpck parallel port tape drive interface for ftape
v4.03-pre-4 06/30/99.
bpck-fdc.c (bpck_fdc_log_adapter) - bpck floppy tape at 0x2050, mode 4 (EPP-32),
delay 0.
So far, so good. Time to use the drive.
Nothing works. Every open attempt fails, after some delay. I have log messages
like this:
Feb 3 17:13:35 morse kernel: [056] 0 ftape-calibr.c (ftape_calibrate) -
calibr_count: 1947, calibr_time: 9990 us.
Feb 3 17:13:35 morse kernel: [057] 0 ftape-calibr.c (ftape_calibrate) - TC for
`fdc_wait()' = 5130 nsec (at 1947 counts).
Feb 3 17:13:36 morse kernel: [058] 0 fdc-io.c (fdc_reset) - no drive polling
interrupt!.
Feb 3 17:13:36 morse kernel: [059] 0 fdc-io.c (fdc_probe) - Type i82078-1 or
5Volt i82078SL FDC found.
Feb 3 17:13:37 morse kernel: [060] 0 fdc-io.c (fdc_reset) - no drive polling
interrupt!.
Feb 3 17:14:08 morse kernel: [061] 0 ftape-ctl.c (ftape_activate_drive) - No tape
drive found.
Feb 3 17:14:10 morse kernel: [062] 0 fdc-io.c (fdc_reset) - no drive polling
interrupt!.
Feb 3 17:14:10 morse kernel: [063] 0 ftape-ctl.c (ftape_enable_R7ec0d5d4) -
ftape_activate_drive(ftape, &ftape->drive_type) failed: -19.
Feb 3 17:14:10 morse kernel: [064] 0 zftape-ctl.c (_zft_open) -
ftape_enable_R7ec0d5d4(sel) failed: -19.
Feb 3 17:14:10 morse kernel: [065] 0 zftape-init.c (zft_open) - _zft_open failed.
So I think I'm making progress, but I'm still facing a non-working tape
drive. Does anyone recognize any familiar (and solvable) problems here?
Nathan Meyers
[EMAIL PROTECTED]
On Wed, Feb 02, 2000 at 10:12:59AM -0800, Nathan Meyers wrote:
> Hi,
>
> I'm trying to use ftape with an external DittoMax drive on a Compaq DeskPro
> running the 2.2.12 kernel... having numerous problems, the worst of which is
> a kernel hang when I try to use the device. I'm wondering if anyone is
> having success with such a configuration. Here are the particulars:
>
> - I'm using the "unstable" ftape-4.x-1999_09_11 which, unlike the 4.03
> pre-release, can be compiled against the 2.2 kernel.
>
> - I'm inserting the modules with reasonable values:
>
> /sbin/modprobe ftape ft_fdc_driver=bpck-fdc
> /sbin/modprobe zftape
> /sbin/modprobe bpck-fdc
>
> - The address and IRQ values reported for the parallel device are correct,
> and printing works just fine (when the bpck-fdc driver isn't loaded, of
> course).
>
> - I have the device configured more or less like an EPP (more detail below).
>
> - As soon as I try to use the device (for example, ftmt -f /dev/qft0
> status), the system locks up and will respond only to a power cycle.
>
> Now, I know that the driver is particular about using EPP and not ECP
> (according to info at LinuxTapeCert.org)... would that possibly account for
> the lockup?
>
> Configuring the parallel port on the DeskPro is an interesting exercise. The
> BIOS configuration menu doesn't offer choices like "SPP, EPP, ECP", it
> offers choices that look like this:
>
> <IO Address>
> <IO Address> <IRQ>
> <IO Address> <IRQ> <DMA>
>
> In other words, you can choose a configuration that specifies an IO address
> but nothing else (like SPP?), one that specifies an IO address and an IRQ
> (like EPP?), or one that specifies an IO address, IRQ, and DMA (like ECP?).
> I've chosen one that specifies an IO address and IRQ, which are correctly
> recognized by the parallel driver. /proc/parport/0/hardware reports the
> following:
>
> base: 0x378
> irq: 7
> dma: none
> modes: SPP, ECP, ECPEPP, ECPPS2
>
> These modes seem a little strange for a device without DMA enabled, but the
> IO and IRQ values match the BIOS settings. Might these mode choices explain
> the driver lockup?
>
>
> I observe something else strange when I insert the drivers: in the
> /sbin/lsmod output: zftape reports a "Used by" value of -1.
>
>
> That's my life with ftape in a nutshell. Does anyone recognize the problems
> and/or have suggestions on what steps might get this stuff to work?
>
>
> Nathan Meyers
> [EMAIL PROTECTED]
>
