Tru64 hits the PCI 0x480 an awful lot. It's documented as some kind of TLB invalidate. Does anyone know what that might mean? Could it explain why I see SCSI disk corruption?? I think it's supposed to go through state.pchip and invalidate a whole bunch of translations..
--- System.cpp 8 Feb 2008 20:08:13 -0000 1.62
+++ System.cpp 20 Feb 2008 15:43:12 -0000
@@ -1383,6 +1383,13 @@
return state.pchip[num].perr;
case 0x400:
return state.pchip[num].perrmask;
+ case 0x480:
+ /* tlbiv */
+ return 0;
+ case 0x4c0:
+ /* tlbia */
+ return 0;
+
case 0x800: // PCI reset
return 0;
default:
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
It seems that the 04C0h register is documented write-only, and the warnings are because Tru64 reads it. (I don't know why it does). So I guess it's just something to ignore.
(Still struggling as to why newfs/mount results in scsi disk corruption..)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Tru64 hits the PCI 0x480 an awful lot. It's documented as some kind of TLB invalidate. Does anyone know what that might mean? Could it explain why I see SCSI disk corruption?? I think it's supposed to go through state.pchip and invalidate a whole bunch of translations..
--- System.cpp 8 Feb 2008 20:08:13 -0000 1.62
+++ System.cpp 20 Feb 2008 15:43:12 -0000
@@ -1383,6 +1383,13 @@
return state.pchip[num].perr;
case 0x400:
return state.pchip[num].perrmask;
+ case 0x480:
+ /* tlbiv */
+ return 0;
+ case 0x4c0:
+ /* tlbia */
+ return 0;
+
case 0x800: // PCI reset
return 0;
default:
It seems that the 04C0h register is documented write-only, and the warnings are because Tru64 reads it. (I don't know why it does). So I guess it's just something to ignore.
(Still struggling as to why newfs/mount results in scsi disk corruption..)