From owner-freebsd-current Tue May 7 23:42:32 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id XAA29105 for current-outgoing; Tue, 7 May 1996 23:42:32 -0700 (PDT) Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id XAA29097 for ; Tue, 7 May 1996 23:42:27 -0700 (PDT) Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id QAA27730; Wed, 8 May 1996 16:39:56 +1000 Date: Wed, 8 May 1996 16:39:56 +1000 From: Bruce Evans Message-Id: <199605080639.QAA27730@godzilla.zeta.org.au> To: current@freebsd.org, kmitch@phantasma.bevc.blacksburg.va.us Subject: Re: COM_MULTIPORT and -current Sender: owner-current@freebsd.org X-Loop: FreeBSD.org Precedence: bulk >I have a STB 4port card that I got from a friend running -stable, and >popped it into my -current system. I changed all of the appropriate >jumpers on the card. >He had his kernel config as (stable): >device sio0 at isa? port "IO_COM1" tty irq 4 vector siointr >device sio1 at isa? port 0x3e8 tty flags 0x105 irq 15 vector siointr >... >When I got it, I set mine up as (current): >options COM_MULTIPORT >device sio2 at isa? port 0x3e8 tty flags 0x505 >device sio3 at isa? port 0x2e8 tty flags 0x505 >device sio4 at isa? port 0x1e8 tty flags 0x505 >device sio5 at isa? port 0x1f8 tty flags 0x505 irq 5 vector siointr >This doesn't work. I get: >sio2 not found at 0x3e8 >... I can't help much with this. Perhaps there is a conflict with sio1. -current is a bit fussier than 2.1R. All sio devices should be configured, even if they are disabled, so that the irqs for unused devices can be disabled. 2.1R had disabled the irqs for the standard com1-4 but this broke some S3 boards. >If I change my config to: >options COM_MULTIPORT >device sio2 at isa? port 0x3e8 tty flags 0x205 >device sio3 at isa? port 0x2e8 tty flags 0x205 >device sio4 at isa? port 0x1e8 tty flags 0x205 >device sio5 at isa? port 0x1f8 tty flags 0x205 irq 5 vector siointr >everything appears to work fine (NOTE: now the master is not the same as the >port with the irq). This results in: This apparently "works" (up to about 16000 bps on a 16550) by configuring all the ports to use polled modee. One of the interrupt tests in the probe apparently fails for the normal configuration, but the interrupt tests in the probe are skipped for polled mode. (For non-multiport sio devices, polled mode is configured by leaving out the irq from the device line. For multiport devices, it is normal to leave out the irq from all lines except the one for the multiport master, so polled mode (for all devices) is configured by leaving out the irq from the multiport master line.) Bruce