RioCar 20GB 2.00-beta7.

Has anyone else experience this. I pulled the unit out before shutting down the car. When I came back, I plugged the thing in and I got a message saying Hard drive not found, contact support. I started driving and let the thing sit for 2 minutes, then plugged it into the sled, and it worked fine. This happened again on the same day. I did the same thing, and it worked. Then, I started shutting the car off, waiting for the unit to shut down completely (I set it to 10 sec), then removing the unit. Now it works perfectly everytime I plug it back in when I do that.

Sounds like you need a new HD cable. I needed the same, still havn't recieved mine yet ; ( Call tech support and they'll send you a new one.

-Greg

Had the same thing yesterday with my new Mk2a. In my case the poor thing seems to have been treated roughly during shipping. On one side the shock protection was completely loose and the harddrive cable wasn't firmly attached to the mainboard anymore. So I pushed it back down (gently) and refitted the shock protection. All is well now.

I got this the first time yesterday (I've had this empeg for about 6 months now).
The message I got from the serial port when failing was:

...

empeg single channel IDE

Probing primary interface...

hdb: TOSHIBA MK1814GAV, ATA DISK drive

ide0: unexpected interrupt, status=0x80, count=1

ide0 at 0x000-0x007,0x038 on irq 6

hdb: TOSHIBA MK1814GAV, 17301MB w/0kB Cache, CHS=35152/16/63

empeg-flash driver initialized

smc chip id/revision 0x3349

smc9194.c:v0.12 03/06/96 by Erik Stahlman (erik@vt.edu)



SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP MEM:6144b MAC 00:02:d7:22:08:3

6

Partition check:

 hdb: hdb4

RAMDISK: ext2 filesystem found at block 0

RAMDISK: Loading 320 blocks [1 disk] into ram disk... done.

VFS: Mounted root (ext2 filesystem).

empeg-pump v0.03 (19980601)

Press Ctrl-A to enter pump...VFS: Cannot open root device 03:05

Kernel panic: VFS: Unable to mount root fs on 03:05

03:05 is I believe /dev/hda5

Opened it up and tested various configurations of single/dual drive, which contact on the cable that was used etc, but couldn't draw a conclusion from that.
It would get the "Disk not found" about 2/3rds of the time with both disks connected, but it worked with only one disk (so probably not the same problem as the original poster in this thread, as a 20GB (from factory) only has one drive). No amount of pressing on contacts, flexing the cable etc changed this.

Being pretty confused at this time, I tried reinstalling 2.0b7 developer (one never knows...) Lo and behold - everything works again! Can't provoke it to not find the disks... OK, then I reinstall HiJack - Boom, problem is back! Tried it with v115, v117 and v118 (as I saw Mark has released another couple of versions, I'll try it with the latest&greatest in a few minutes). I could consistently solve the problem by uploading 2.0b7 and consistently get it back by adding HiJack...

Will be back after finishing catching up on the board and testing the lastest HiJack.

/Michael

You've probably checked this but make sure that you enable "One or two drives (slower)" under "Hard disk detection" in Hijack, since you have it only with hijack.

No, it isn't HiJack and single disk speedup (at least not explicitly).

OK, this is from starting from a freshly applied 2.0b7.upgrade
* check that there's no mention of HiJack in config.ini
* apply HiJack v120 for MkII (file size: 497 264 byte,
MD5Sum: aa6fa76ce96f1de13a698c70f4c4f30a) via LogoEditor -
player starts OK directly after application, I can get in an play tunes
and use the HiJack functions
* turn off (press/hold top button, then remove power)
* apply power - HiJack v120 splash is shown, waving Tux and custom logo -
then I get "No hard..." about 2/3 to 3/4 of the time...
* apply 2.0b7 again and can't make it (about 20 tries) not boot properly

Lookin at the boot logs, here's the first difference:
When it can't find the disk:
...
empeg single channel IDE
Probing primary interface...
hdb: TOSHIBA MK1814GAV, ATA DISK drive
ide0: unexpected interrupt, status=0x80, count=1
ide0 at 0x000-0x007,0x038 on irq 6
hdb: TOSHIBA MK1814GAV, 17301MB w/0kB Cache, CHS=35152/16/63
empeg-flash driver initialized
smc chip id/revision 0x3349
smc9194.c:v0.12 03/06/96 by Erik Stahlman (erik@vt.edu)

SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP MEM:6144b MAC 00:02:d7:22:08:3
6
Partition check:
hdb: hdb4
RAMDISK: ext2 filesystem found at block 0
RAMDISK: Loading 320 blocks [1 disk] into ram disk... done.
VFS: Mounted root (ext2 filesystem).
empeg-pump v0.03 (19980601)
Press Ctrl-A to enter pump...VFS: Cannot open root device 03:05
Kernel panic: VFS: Unable to mount root fs on 03:05


Isn't 0x80 associated with the primary disk? (vague memories from LILO
remapping to please Windows...) Looks like it just doesn't see the primary disk...

Successful boot:
...
empeg single channel IDE
Probing primary interface...
hda: TOSHIBA MK3017GAP, ATA DISK drive
hdb: TOSHIBA MK1814GAV, ATA DISK drive
ide0 at 0x000-0x007,0x038 on irq 6
hda: TOSHIBA MK3017GAP, 28615MB w/0kB Cache, CHS=58140/16/63
hdb: TOSHIBA MK1814GAV, 17301MB w/0kB Cache, CHS=35152/16/63
empeg-flash driver initialized
smc chip id/revision 0x3349
smc9194.c:v0.12 03/06/96 by Erik Stahlman (erik@vt.edu)

SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP MEM:6144b MAC 00:02:d7:22:08:3
6
Partition check:
hda: hda1 < hda5 hda6 > hda2 hda3 hda4
hdb: hdb4
RAMDISK: ext2 filesystem found at block 0
RAMDISK: Loading 320 blocks [1 disk] into ram disk... done.
VFS: Mounted root (ext2 filesystem).
....



No touching the IDE-cable or player itself between attempts - power on/off by
pulling the cable from the power supply, cable between supply and empeg is
secured. And no amount of flexing the IDE-cable makes 2.0b7 w/o HiJack bahave
this way...

/Michael - seriously confused

>No, it isn't HiJack and single disk speedup (at least not explicitly).

Yes, it probably is that feature, since nothing else can cause this behaviour. The feature is not a config.ini feature, but rather a bit stored in flash memory. If that bit happens to have the wrong setting, then you're basically screwed as far as hijack is concerned.

I'll see if I can't fix this. In your logs, I notice the slave drive coming up before the master.. I suppose this is "standard" required drive behaviour (master waits for slave on reset), so maybe it will always happen this way.. In any case, this is detectable, and if a slave appears, the code can just wait for the master regardless of the setting in Hijack.

If the problem happens the other way around, master comes up and slave is ignored, then this is easy to fix by just going into the hijack menu and changing the 1/2 drive setting -- so nothing new needed for that scenario. The reason the first scenario is more difficult is because the stupid filesystem code panic's and shuts down, rather than just sitting there managing life as it's supposed to do..

I'll put the fix out in v122 later today.

-ml

As I said in a private reply to Mark, this was indeed it - by getting into HiJack real early, just as the display lights up the first time, I could get to the HiJack functions.
Vitals showed 0+18GB and 1/2 showed 1 disk only. Changing the 1/2 option didn't save that boot - still didn't find the disk - but subsequent reboots have all been successful.

The fix you describe definately sounds like it would solve any future cases of this bit flipping (one wonders what caused it to flip though...).

Again, thanks Mark!

/Michael

Okay, v122 is out, and has code to avoid this problem.

The cause was probably an upgrade from an earlier version of hijack that had different bit assignments in flash.. so the new version could have misinterpreted settings.

I *usually* try to avoid changing bit meanings, but sometimes it is very useful to rearrange the hijack flash area..

Cheers

-ml

an upgrade from an earlier version of hijack that had different bit assignments in flash

Aha...

I went from somewhere in the 80s to v115. Must try to keep up better

/Michael

I didn't have Hijack installed when this error occured. But I do have it now (after flashing the image first) and everything is going great.