Date: Thu, 26 Mar 1998 10:22:51 +1100 From: Bruce Evans <bde@zeta.org.au> To: ady@warpnet.ro, current@FreeBSD.ORG Subject: Re: silo overflows (Was Re: 3.0-RELEASE?) Message-ID: <199803252322.KAA29759@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
> Likewise, here it is my experience with a CYCLADES Cyclom 16YeP multiport >board & FreeBSD 3.0-971117-SNAP, both UP & SMP case (please nota that I >used the COM ports before installing the multiport board and did get zero >"silo overflow"s; also, setting speed to 115200 rather than 57600 makes >the no. of overflows grow faster): This problem is well known. The PCI version of the cy driver doesn't use fast interrupts, so at 115200 bps, it is likely to drop about N fifos full of input whenever another driver preempts of masks cy interrupts for N+0.5 msec. The keyboard driver usually masks (non-fast) cy interrupts for more than several msec to program the keyboard LEDs; PIO drivers like the PIO wd driver can preempt (non-fast) interrupts for even longer if the disk is fast and the controller is slow... >---------- >Mar 25 19:46:01 ady /kernel: cy0: 1 more silo overflow (total 177) >Mar 25 19:50:26 ady /kernel: cy0: 1 more silo overflow (total 178) >Mar 25 19:50:58 ady /kernel: cy2: 1 more silo overflow (total 13) >Mar 25 20:02:03 ady /kernel: cy0: 1 more silo overflow (total 179) >Mar 25 20:03:11 ady /kernel: cy0: 1 more silo overflow (total 180) It's doing well to get only one at a time :-). Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199803252322.KAA29759>