Within the past three days, someone posted to the expert list on
a lost interrupt which stopped his installation of Mandrake 7.0.
I have managed to duplicate the error, at least I think so. I
hope the poster will respond with his equipment configuration.
OK VIA MVP4 Chipset, K6-2 standard clocking 100MHz bus and
500MHz processor--dropped to 80 and then to 75 and finally to 66
in an attempt to pass the error.
Settings for UDMA/PIO/Prefetch and HDD Block mode were progressed
truth-table style. None seemed to have any effect on the error.
an 80-pin UDMA cable was removed. Still no effect.
The drive caused an error as follows
Invalid code 0000
<register dump>
Kernel Panic
Attempting to kill inactive process
In swapping processes
hdc: lost interrupt (That was my Creative CDRW)
The drive was a seagate barracuda 7200 rpm 7.6ms nominally 10.2
G. It looked like the 6.0/6.1 problem with large UDMA Seagate
Drives and MVP super7 chipsets all over again. I salvaged one of
those Seagate drives by putting it as /dev/hdc and setting it to
NORMAL, so I tried that.
hda: lost interrupt (that was my Creative CDRW)
Uh huh. Well I put a 2.5G Maxtor drive at hda instead of the
CD-ROM
Passed the boot point where it was failing and signal 11ed after
loading second stage from CD-ROM
Now I wanted the exact wording for the error so I restored the
Seagate to drive /dev/hda and ran the CD boot again.
Signal 11ed on second stage.
I dropped the clock 5%
THe thing is installing. UDMA is auto, Prefetch is on, PIO is
auto, HDD Block is enabled mem is 8ns and cycle time is 2.
Installing at 95Khz/475MHz processor without any trouble.
I would guess it is an initial load of some NVRAM in the Seagate
which is "cured" by setting it temporarily to drive C.
IN the original failure position, I was able to install FreeBSD
and to remove FreeBSD and install Win98 and to remove win98 and
fail on VEnus and Helios with the typical large Seagate errors
previously reported. None of these installs had any effect on
the Seagate drive as far as its behavior toward an AIR install
was concerned.
We are obviously dealing with something very subtle at the
edge--the bleeding edge it would seem. What is different about
Seagate drives? I tried Fujitsu, Maxtor, IBM, and Quantum in the
same position and none showed the problem. It seems to take
three conditions
1. Large Seagate Drive (8.4G or bigger)
2. VIA MVP3 or MVP4 chipset Super7 motherboard with IDT, AMD,
INtel, or Cyrix processor
3. Linux-Mandrake 6.0 6.1(6.5MacMillan) or 7.0
And it seems to be cured by functioning as /dev/hdc in an
attempted install with a non-seagate /dev/hda
Once the "cure" is performed, it will install.
But will it boot?
LIL-
Using the boot floppy does bring it up. Obviously Seagate is
doing something very diferent from other manufacturers.
I have observed similar or worse problems on an Intel TX chipset
with a Seagate large Drive where the install seemed to go
perfectly well but the HDD was completely corrupt afterward. On
a TX chipset board, the position of the drive mattered not a
whit. It installed without error messages and was completely
corrupt.
On SiS super-7 chipsets on boards I have a lot of trouble making
hiccup during burn-in testing, I also have trouble with Seagate
large drives, though they work like champs as /dev/hdc and
NORMAL.
OK this system is running with the large Seagate in /dev/hda
(primary boot) but it will boot only from boot disks.
SO is that where you encountered the "Lost Interrupt"? Was it a
Seagate Drive? If it was, what was the Chipset and Processor?
And does anyone need a cheap Seagate barracuda, capable of
hyperfast accesses in Windows or FreeBSD?
Civileme
--
experimentation involving more than 500 trials with an
ordinary slice of bread and a tablespoon of peanut butter
has determined that the probability a random toss will
land sticky side down (SSD) is approximately .98
