Hey Michael,

Michael wrote:
> 
> Hello,
> 
> > sd0 at scsibus1 target 0 lun 0: <QUANTUM, FIREBALL SE4.3S, PJ0A> disk
> > fixed
> > sd0: 4110 MB, 7637 cyl, 4 head, 275 sec, 512 bytes/sect x 8418816
> > sectors
> > probe(mesh0:0:1:0): max sync rate 10.00Mb/s
> > sd1 at scsibus1 target 1 lun 0: <IBM OEM, DFHSS4F, 4141> disk fixed
> > sd1: drive offline
> ...
> > mesh0: timeout state 3
> > mesh0: intr/msk 00/07, exc 00, err 00, st0/1 21/40
> > mesh: resetting DMA
> 
> > note the last messages, the typical scsi/mesh/sync troubles...
> Yeah, looks like the mesh patch didn't make it into 2.0.2 for whatever reason.

well, more than 2.0.2, this is really just "cvs" of the same day. As you
see the kernel identifies itself as 3.99. Make pressure so that your
patches get commited ;)


> > I don't remember which magic michael used last time. I think he forced
> > sync on for some devices. This means in any case that negotiation still
> > has troubles.
> No, I forced sync off first, then I hacked mesh to cope with sync. The problem is that the unpatched mesh driver accepts sync negotiation from other SCSI devices, the bits in flags only keep it from starting the negotiation itself.

but if the driver can DO mesh sync transfer if you force it, why can't
it do when it is asked for after a negotiation ?

In any case I wanted to clean up my box. Since I tried to test kernel
stuff, your X stuff, I end up having a farily messed up computer
- userland is a pre 2.0 thing
- kernel is either 3.99 series or your only working 2.99 kernel
- X server is your patched version to work with my Matrox Millennium

anything I could do in my curernt situation? get userland to 2.0.2 for
example? and what about X?
I also have a TwinTurbo card now. I wonder how support for that one is ?
But since we have no DRI I wonder if there would a gain over the matrox
at all...

Since I no longer have a "test" and a "work" netbsd/macppc setup, I
wanted a lss "testing" box. Although I have to admit, I never had
particular problems with my current setup. Give the working kernel and X
server everything is quite fine. Of course fixing the keyboard under X
in a better way than my half-baked xmodmap would be more professional.

cheers,

  Riccardo