[SunRescue] NetBSD 1.4 miniroot problem fixed
bsdbob at weedcon1.cropsci.ncsu.edu
Sun Dec 5 20:36:27 CST 1999
> I give up. I originally selected NetBSD because it has the best Sun-3
> support and I was hoping to have all the Suns run the same OS. But at this
> point, it's just not worth it. Support for older SPARCs is just too broken.
> I could drop back to 1.3.3, which seems to work OK modulo the occasional
> panic caused by some kind of SCSI-related timeout, but I don't want to be
> stuck with yet another dead OS. I just looked at OpenBSD's page for SPARCs
> and they claim to support everything I need, so OpenBSD, here I come!
Be aware that OpenBSD's build, or at least on my hardware still has that
scsi timeout thing. I have tracked it down to the point in the code where
si.c is reading and comparing values in the buffer (maybe to see if
the buffer is empty), and it will dump out to debugger occasionally
at that point. It is a timing thing, since the code does not actually
break, and the second pass around it has caught up with itself.
The routine is si_dma_start where the message is issued.
If you compare the sun3 code and the sun4 code for si.c from NetBSD
vs OpenBSD, it is missing a splhigh() call that may affect the timing.
IFF I can keep the machine stable enough for a kernel build, I will try
wacking it a bit for a fix. I was thinking of using an IPX as a build
box for the kernel and then ftping the kernel over to the sun4 box for
testing. The code works fine, apparently, on sun4c hardware.
Which exact hardware are you running?
More information about the rescue