Mike Tancsa wrote: > At 09:00 AM 11/19/2009, Karl Denninger wrote: >> Well ok then the uart driver is BROKEN. >> >> It simply locks up on the port after some period of time, returning >> nothing. I have found no way to reset the port other than a reboot >> either. That's a "surprise" that people running fax servers and other >> similar things are going to be very unhappy about. > > Which serial card are you using ? I have a number of PCI cards (lava > for example) that are working quite well with the puc and uart driver > combo. Perhaps you could post some details about the hardware used > thats having issues. > > ---Mike puc0: <Oxford Semiconductor OX16PCI954 UARTs> port 0x4060-0x407f,0x4040-0x405f mem 0x94503000-0x94503fff,0x94502000-0x94502fff irq 16 at device 0.0 on pci3 puc0: [FILTER] uart2: <16550 or compatible> on puc0 uart2: [FILTER] uart3: <16550 or compatible> on puc0 uart3: [FILTER] uart4: <16550 or compatible> on puc0 uart4: [FILTER] uart5: <16550 or compatible> on puc0 uart5: [FILTER]
It's a generic board with four ports sitting on the PCI bus. Nothing special, no smarts, just four 16550 uarts. Port "0" is on the motherboard and I am using that to talk to my NMEA clock (using PPS, which is working). I also am using one of the "troubled" ports to talk to an APC UPS, again, without problems. Trivial and reasonably-trivial applications work fine on the uart/puc combination. Hylafax requires correct modem control that is FLAWLESS or it WILL blow up. This is an area where smart cards have in the past run into driver trouble (going back years; I have lots of experience with driver issues on so-called "smart" cards going back more than a decade in this sort of application!) and now, it appears that sickness has translated into the UART driver as well. The system in question sends hundreds of faxes a day and any sort of squirrelly driver problems in the serial subsystem show up almost instantly (I updated to RC2 on that machine last night and it took less than an hour for the modems to wedge this morning.) I tried to flip back to sio() as the driver in GENERIC and it appears that can't be done, or I'm missing some interdependency somewhere (doing so throws all sorts of errors related to the include files) -- Karl
_______________________________________________ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"