der Mouse writes:
> I'm working with a 2.0 i386 machine.  It's got a "hardware" (presumably
> really firmware) RAID card:
> 
> twe0 at pci3 dev 1 function 0: 3ware Escalade
> twe0: interrupting at irq 10
> twe0: 12 ports, Firmware FE7S 1.05.00.065, BIOS BE7X 1.08.00.048
> twe0: Monitor ME7X 1.01.00.038, PCB Rev5    , Achip 3.20    , Pchip 1.30-66 
> 
> It's got 12 disks on it, each 286168 MB (according to the list in
> dmesg - they're sold as 300G, but tha's disk-manufacturer gigabytes).
> Because it has a hard limit of 2T on a RAID 5, I've configured it with
> two RAID 5s, each six disks, and was planning to paste them together in
> software.  And the two arrays show up fine at boot time:
> 
> ld0 at twe0 unit 0: 64K stripe RAID5, status: Normal
> ld0: 1397 GB, 182405 cyl, 255 head, 63 sec, 512 bytes/sect x 2930351360 secto
> rs
> ld1 at twe0 unit 6: 64K stripe RAID5, status: Normal
> ld1: 1397 GB, 182405 cyl, 255 head, 63 sec, 512 bytes/sect x 2930351360 secto
> rs
> 
> But I've tried ccd and I've tried RAIDframe RAID0, and in each case,
> two components each 2930351232 sectors adds up to a result having only
> 1565734912 sectors (that's the RAIDframe value, ccd may differ slightly
> but is about the same) - both appear to be using 32-bit sector numbers.

How did you get this value from RAIDframe?

> Is there any way to get what I want here (a 3T filesystem)?  Would it
> be enough to go through the ccd driver and replace 32-bit types with
> daddr_t wherever suitable, or is there something more fundamental that
> would break?  I feel competent to do that to ccd, though I wouldn't
> feel competent to do it to RAIDframe.

RAIDframe already supports 64-bits worth of sectors in an array.  
Unfortunatly, I don't think NetBSD's disklabel does. :(

(hmm... I also see that raidctl probably only prints 32-bit 
quantities, even though it gets 64-bit ones..)

Later...

Greg Oster