Subject: Re: kernel link errors on sup Jun 12 05:06:48 1997
To: Bill Harris <billh@airmail.net>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: port-pmax
Date: 06/20/1997 13:25:23
>Need I worry about this?


>text    data    bss     dec     hex     filename
>684104  38288   70812   793204  c1a74   netbsd
>Warning: 4104 byte intersegment gap.
>^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

hi Bill,

No, as Ted Lemon says, this isn't worrisome.  The ELF-to-a.out
converter left slightly more than a page of free space between
segments, and it's letting you know it did that.

But -current is still getting changes to support mips3 (r4000,r4400)
CPUs.  To do that, we've had to re-work a lot of the internal kernel
(locore) interfaces: the low-level `glue' code that helps the kernel
handle interrupts, system calls, page-fault traps, etc; and also how
C kernel code calls into assembly code, to do things like cache-flushes.

I'm running a -current kernel, and I don't think it's as stable as it
has been and should be (and will be in a week or so.).  If this is a
`production' machine, the safest thing is to stick with a snapshot
kernel, like the June 7 snapshot.  Anyone running a kernel that's
older than the SII fixes should definitely upgrade to a newer
snapshot.

Of course, if you *can* tolerate a few bugs -- maybe the clock will
run fast, but I don't think the kernel will crash -- it helps us out
if you run -current, and send in PRs for any bugs you encounter.