ron@mlfarm.uucp (Ronald Florence) (09/26/90)
I recently compiled and tried to install FAS 2.07 on our ps2/80 running Xenix 2.3.3. We have two IBM dual-asynch boards (four ports) in addition to the motherboard port. By defining NEED_PUT_GETCHAR, I got the new kernel to boot, with the following startup messages from FAS: %fas 0x03F8-0x03FF 04 - unit=0 type=* release=2.07.0 %fas 0x02F8-0x02FF 03 - unit=1 type=* release=2.07.0 %fas 0x3220-0x3227 03 - unit=2 type=* release=2.07.0 %fas 0x3228-0x322F 03 - unit=3 type=F release=2.07.0 %fas 0x4220-0x4227 03 - unit=4 type=F release=2.07.0 I am not sure why FAS thinks the on-board port and the first dual-asych board are not capable of FIFOs. So far, the question is academic, because I cannot get gettys running on the serial ports, and FAS seems to lock up the console until the only relief is the red switch. I used the following declarations in space.c: #define NUM_PHYSICAL_UNITS 5 uint fas_port [NUM_PHYSICAL_UNITS] = { 0x3f8, 0x2f8, 0x3220, 0x3228, 0x4220 }; uint fas_vec [NUM_PHYSICAL_UNITS] = { 0x4, 0x3, 0x3, 0x3, 0x3 }; uint fas_mcb [NUM_PHYSICAL_UNITS] = { 0, 0, 0, 0, 0 }; uint fas_modem [NUM_PHYSICAL_UNITS] = { EN_DTR | CA_DCD | UB_RING, EN_DTR | CA_DCD | UB_RING, EN_DTR | CA_DCD | UB_RING, EN_DTR | CA_DCD | UB_RING, EN_DTR | CA_DCD | UB_RING }; uint fas_flow [NUM_PHYSICAL_UNITS] = { HI_RTS | HO_CTS_ON_DSR, HI_RTS | HO_CTS_ON_DSR, HI_RTS | HO_CTS_ON_DSR, HI_RTS | HO_CTS_ON_DSR, HI_RTS | HO_CTS_ON_DSR }; and added the following devices to /dev, /etc/ttys, and /etc/ttytype: device major minor ttyF00 5 32 ttyF01 5 33 ttyF02 5 34 ttyF03 5 35 ttyF04 5 36 ttyFM00 5 192 Any ideas what is awry? Has anyone succeeded in getting FAS to work on a similar setup? Thanks. -- Ronald Florence {hsi,rayssd}!mlfarm!ron
max@lgc.com (Max Heffler) (09/27/90)
In article <RON.90Sep26124111@mlfarm.uucp> ron@mlfarm.uucp (Ronald Florence) writes: >I recently compiled and tried to install FAS 2.07 on our ps2/80 I am repeating my request for FAS 2.07 Part 2 of 3 or an FTP site with this. Thanx... -- Max Heffler internet: max@lgc.com Landmark Graphics Corp. uucp: ..!uunet!lgc!max 333 Cypress Run, Suite 100 phone: (713) 579-4751 Houston, Texas 77094
gemini@geminix.in-berlin.de (Uwe Doering) (09/29/90)
ron@mlfarm.uucp (Ronald Florence) writes: >I recently compiled and tried to install FAS 2.07 on our ps2/80 >running Xenix 2.3.3. We have two IBM dual-asynch boards (four ports) >in addition to the motherboard port. > >By defining NEED_PUT_GETCHAR, I got the new kernel to boot, with the >following startup messages from FAS: > > %fas 0x03F8-0x03FF 04 - unit=0 type=* release=2.07.0 > %fas 0x02F8-0x02FF 03 - unit=1 type=* release=2.07.0 > %fas 0x3220-0x3227 03 - unit=2 type=* release=2.07.0 > %fas 0x3228-0x322F 03 - unit=3 type=F release=2.07.0 > %fas 0x4220-0x4227 03 - unit=4 type=F release=2.07.0 > >I am not sure why FAS thinks the on-board port and the first >dual-asych board are not capable of FIFOs. This means that there are no NS16550A chips on the main board and the first dual-asynch board. Check out what type of chips they used instead. Or maybe there are versions of the NS16550 that aren't compatible with the `A' type. Don't know. All I can say is that I got the infos on how to auto-detect the FIFO type from the original National Semiconductor data sheet for the NS16550A, and obviously it works for the majority of FAS users. >So far, the question is >academic, because I cannot get gettys running on the serial ports, >and FAS seems to lock up the console until the only relief is the >red switch. > >I used the following declarations in space.c: > [contents of space.c deleted] Your space.c seems to be o.k. Did you enter the following line in /usr/sys/conf/master? sio 4 0537 104 fas 0 0 5 1 7 3 4 33 34 Someone told me that you have to run /usr/sys/conf/configure after you made the changes to the necessary config files. If this doesn't work the only reason for your problems I can think of is the different handling of interrupts. On the AT bus interrupts are edge triggered while on the micro channel they are level triggered. Maybe this breaks FAS' interrupt handler. Has anyone been able to run FAS on an IBM micro channel system? And don't forget to tell us under what operating system. Uwe -- Uwe Doering | Domain : gemini@geminix.in-berlin.de Berlin |---------------------------------------------------------------- Germany | Bangpath : ...!unido!fub!tmpmbx!geminix!gemini
ron@mlfarm.uucp (Ronald Florence) (09/30/90)
gemini@geminix.in-berlin.de (Uwe Doering) writes: > This means that there are no NS16550A chips on the main board and the > first dual-asynch board. The UART on the motherboard and the UARTs on the first dual-asynch board are NS16550s. The first dual-asynch board is identical to the the one which does show up as FIFO-capable. > Your space.c seems to be o.k. Did you enter the following line in > /usr/sys/conf/master? > > sio 4 0537 104 fas 0 0 5 1 7 3 4 33 34 Yes. And I ran configure after making the changes. > If this doesn't work the only reason for your problems I can think of is > the different handling of interrupts. On the AT bus interrupts are edge > triggered while on the micro channel they are level triggered. Maybe this > breaks FAS' interrupt handler. Thanks to Uwe for his prompt response, and to all for any further suggestions on getting FAS to run on a microchannel system. -- Ronald Florence {hsi,rayssd}!mlfarm!ron