Re: MULTIPROCESSOR enabled by default on all Alpha kernels

[Jason Thorpe <thorpej%me.com@localhost>]

> On Sep 20, 2020, at 12:51 PM, Dustin Marquess <dmarquess%gmail.com@localhost> wrote:
> 
> My only SMP Alpha is a CS20 running DS20L firmware. I know there were issues at one time with the built-in NIC and SCSI controllers. Did those ever get resolved? If so, I’ll get it racked and powered up.
> 
> I had swapped the SCSI controller out with one for a DS20L to run Tru64 on it at one time, but I still have the original (if I remember correctly, NetBSD didn’t like the DS20L one). I know one was Adaptec, I forget what the other one was.

The system (a DS25) I've been testing on has:

[     1.000000] ahc0 at pci1 dev 1 function 0: Adaptec aic7899 Ultra160 SCSI adapter  
[     1.000000] ahc0: interrupting at dec 6600 irq 0
[     1.000000] ahc0: aic7899: Ultra160 Wide Channel A, SCSI Id=7, 32/253 SCBs
[     1.000000] scsibus0 at ahc0: 16 targets, 8 luns per target
[     1.000000] ahc1 at pci1 dev 1 function 1: Adaptec aic7899 Ultra160 SCSI adapter  
[     1.000000] ahc1: interrupting at dec 6600 irq 1
[     1.000000] ahc1: aic7899: Ultra160 Wide Channel B, SCSI Id=7, 32/253 SCBs  
[     1.000000] scsibus1 at ahc1: 16 targets, 8 luns per target

...one of which hosts the boot drive:

[     6.656384] sd0 at scsibus0 target 0 lun 0: <COMPAQ, BD0726536C, 3B02> disk fixed
[     6.748180] sd0: 69464 MB, 28004 cyl, 20 head, 254 sec, 512 bytes/sect x 142264000 sectors

As far as network interfaces go, this system contains:

[     1.000000] fxp0 at pci0 dev 8 function 0: Intel i82559ER Ethernet (rev. 0x09)    
[     1.000000] fxp0: interrupting at dec 6600 irq 20
[     1.000000] fxp0: May need receiver lock-up workaround
[     1.000000] fxp0: Ethernet address 40:00:84:a5:1f:00
[     1.000000] inphy0 at fxp0 phy 1: i82555 10/100 media interface, rev. 4
[     1.000000] inphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto 

[     1.000000] bge0 at pci1 dev 5 function 0: Broadcom BCM5703X Gigabit Ethernet     
[     1.000000] bge0: interrupting at dec 6600 irq 4
[     1.000000] bge0: HW config 00000014, 00000000, 00000000, 00000000 00000000
[     1.000000] bge0: ASIC BCM5702/5703 A2 (0x1002), Ethernet address 00:02:a5:c6:2a:44
[     1.000000] brgphy0 at bge0 phy 1: BCM5703 1000BASE-T media interface, rev.2     
[     1.000000] brgphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto
[     1.000000] tsp2 at tsc0
[     1.000000] pci2 at tsp2 bus 0
[     1.000000] pci2: i/o space, memory space enabled, rd/line, rd/mult, wr/invok    
[     1.000000] bge1 at pci2 dev 2 function 0: Broadcom BCM5703 Gigabit Ethernet
[     1.000000] bge1: interrupting at dec 6600 irq 32
[     1.000000] bge1: HW config 00010055, 00000015, 00000000, 00000000 00000000
[     1.000000] bge1: ASIC BCM5702/5703 B0 (0x1100), Ethernet address 00:10:18:10:65:a6
[     1.000000] brgphy1 at bge1 phy 1: BCM5703 1000BASE-T media interface, rev.3     
[     1.000000] brgphy1: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto

On this system, the Intel fast Ethernet is not in use, but the two Broadcom gigabit interfaces are.

I have heard reports out in retrocomputing Twitter that > 1GB RAM is problematic, but I have not reproduced that issue -- this system has 12G of RAM.

The only ongoing issue I've been having with this test system is an occasional panic due to a memory management fault in the "siisata" driver (this system also has several SATA spinning rust / SSDs attached) in an error handling path, and it seems to correlate with "smartd" starting (from the "smartmontools" package).  It doesn't happen every time, only rarely, and gdb crashes when I try to load the kernel core, so I've enabled DDB so I can take a look if I happen to catch it in the act.

-- thorpej