> this is another post on this error message that I get on my
> sparcstation 2:

> netbsd: ms0: input error (0x3447)

UTSL.

This - at least in the source I have handy - comes from
ms_zs_softint(), in sys/dev/sun/ms_zs.c:

        get = ms->ms_rbget;
        while (get != ms->ms_rbput) {
                ring_data = ms->ms_rbuf[get];
                get = (get + 1) & MS_RX_RING_MASK;

                /* low byte of ring_data is rr1 */
                c = (ring_data >> 8) & 0xff;

                if (ring_data & ZSRR1_DO)
                        intr_flags |= INTR_RX_OVERRUN;
                if (ring_data & (ZSRR1_FE | ZSRR1_PE)) {
                        log(LOG_ERR, "%s: input error (0x%x)\n",
                                ms->ms_dev.dv_xname, ring_data);
                        c = -1; /* signal input error */
                }

                /* Pass this up to the "middle" layer. */
                ms_input(ms, c);
        }

0x47 has ZSRR1_FE but not ZSRR1_PE set; the host is seeing a framing
error from the mouse.  Normally I'd guess this means that the sending
and receiving baudrates are mismatched; this seems unlikely with a
mouse.  It can also mean there's noise on the line; this too seems
unlikely, unless perhaps a wire has flexed too much and is making and
breaking contact as you move the mouse.

> [A]fter the machine has been up for some time the error disappears.

That is very peculiar.  It seems to point to something thermal, but I
have trouble imagining what.

Have you tried a different mouse?  If you can, that's one of the first
things I'd try.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B