On Sat, Jan 19, 2019 at 2:52 AM Kurt Jaeger <pi_at_freebsd.org> wrote: > Hi! > > > uart is the new thing. sio info should be ignored. > > > > Chances are good that this device doesn't have the proper entries in the > > puc driver. Do you have any pci devices that show up as unclaimed? > > In a different box, I got this: > > none1_at_pci0:7:4:0: class=0x070002 card=0x000814a1 chip=0x000814a1 > rev=0xb0 hdr=0x00 > vendor = 'Systembase Co Ltd' > class = simple comms > subclass = UART > bar [10] = type I/O Port, range 32, base 0x1040, size 64, enabled > bar [14] = type I/O Port, range 32, base 0x1000, size 64, enabled > This is the one you want. You'll need to add vendor 14a1 device 8 to the puc tables. Do you need help with this? It will be a bit tricky because each of these defines several ports, I think. and: > > pcib7_at_pci0:6:0:0: class=0x060400 card=0x00000000 chip=0x10801b21 > rev=0x04 hdr=0x01 > vendor = 'ASMedia Technology Inc.' > device = 'ASM1083/1085 PCIe to PCI Bridge' > class = bridge > subclass = PCI-PCI > This is something else. > The chips on the card are: > > ASMedia asm1083 b0bk4911b3 1543 (?) > SystemBase SB16C1058PCI 1624 > > It only detects four (or six?) serials... > > So I think I found a 'somehow' working setup and have to add stuff to > sys/dev/puc/pucdata.c to match it. Thanks for the pointer! That's right. Ask me if you need help. There's several different ways that hardware vendors slice and dice the UARTs, and there's no standard. Clock rate may be an issue too, since newer cards have faster baud clocks to support higher rates, but this means to get the right right you have to use a different divisor than the older 16550A typically needed. Luckily this is well supported. WarnerReceived on Sat Jan 19 2019 - 15:28:43 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:20 UTC