> I have a bit of free space at the beginning of my hard drive so I wanted > to reinstall a native plan9 there, however installing from the iso (dled > on 23/12/2008) fails for me. I get some I/O errors at the fmtfossil stage, > such as those: > sdE0: i/o error d0 @0 > sdE0: LLBA 312,581,808 sectors > > I suppose this happens because of some changes with the sata driver since > I was able to install with the iso something like one year ago. I guess I > could try to create an alternate iso with Erik's sdiahci but I figured > I could try to install from 9vx first, if that's at all possible.
i just reviewed the changes on sources. i don't see anything that jumps out at me as a regression. mainly u32int's were changed to ulong and there are some locking fixes. not likely your problem. perhaps you were never using a driver from sources? there are a number of postitive changes in the current driver. it's been in production for a year or so and we've gotten a few fixes for esoteric cases from that. (there's nothing like a bad drive to ferret these things out.) lucio also found that i had absolutely no idea what i was doing with hbas that support Hsss (staggered spinup), at least for ich9m-based thinkpads. and i know there were some other folks who helped me kick disks out of power-mgmt funks. my current versions are: /n/sources/contrib/quanstro/src/9/pc/^(ahci.h sdiahci.c). /n/sources/contrib/quanstro/src/9loadaoe/^(ahci.h sdiahci.c). by the way, since in the plan 9 world int ≡ 32 bits, why is it customary to use ulong instead of uint for 32 bit values? the reason for doing this probablly passed before kenc was written. - erik
