I got it working now. Looks like the wrap system simulates some kind of C/H/S in lba mode. OpenBSD is still telling me that I'm in C/H/S mode:
Using drive 0, partition 3; Loading;..... But more important is that: 01F0 Master 848A SAMSUNG CF/ATA Phys C/H/S 1010/16/63 Log C/H/S 505/32/63 The log values seems to be identical on every CF card (except Cylinder). My two CF cards are totally different: 128MB - C/H/S 498/16/32 512MB - C/H/S 1010/16/63 I'm able to boot both cards with the sme image (created with the flashdist wrapper script - gzip image - written with phydiskwrite under windows). I set cylinders to 60 to get an 60MB image and everything is working fine now. Btw, why do I not need to change the bios setting for the m0n0wall image? Any idea? Regards Hagen Volpers >> I understand this is a problem of target systems translating C/H/S >> values differently. There is no problem in dynamicly using OpenBSD's >> idea of C/H/S values at build time. However, OpenBSD on two different >> machines can provide completely different C/H/S values on the exact >> same card. Correct me if im wrong. > [...] > > Just because flashdist asks for C/H/S doesn't mean that the image be applied > to a card with that exact C/H/S. This was the case before OpenBSD switched > to the LBA based MBR. Now, as long as the CF image fits on the card, it should > boot. It should boot, but it doesn't. I'm using a WRAP system and....: [...] Using drive 0, partition 3; Loading;..... [...] For some reason I cannot use LBA (even if I switch in WRAP bios). I wasn't able to figure out how. If I use your script everything is working... What I don't understand is, why other systems work (m0n0wall for example). Any idea? Regards Hagen Volpers
