N
Nju Njoroge
Guest
Hello,
I have been using the Riscwatch box w/ RISCwatch ver. 5.1 software. I
am running Linux on our embedded PPC405D. The trace captures the
instructions with their virtual addresses. However, when an interrupt
is detected, the processor uses the physical addresses, at which point
Riscwatch complains it is "not configured" to read those addresses. I
suspect that it thinks that the physical address is virtual, so when
it goes into the TLB and/or page table for the translation, it can't
find it since there isn't obviously a translation. In other words, I
think that Riscwatch should switch to real mode instead at that point.
How can configure RISCwatch to access those interrupt address
locations that need physical address? Anyone encountered this problem
and found a way around it? I e-mailed IBM about it, but haven't gotten
a response yet...
Thanks,
NN
I have been using the Riscwatch box w/ RISCwatch ver. 5.1 software. I
am running Linux on our embedded PPC405D. The trace captures the
instructions with their virtual addresses. However, when an interrupt
is detected, the processor uses the physical addresses, at which point
Riscwatch complains it is "not configured" to read those addresses. I
suspect that it thinks that the physical address is virtual, so when
it goes into the TLB and/or page table for the translation, it can't
find it since there isn't obviously a translation. In other words, I
think that Riscwatch should switch to real mode instead at that point.
How can configure RISCwatch to access those interrupt address
locations that need physical address? Anyone encountered this problem
and found a way around it? I e-mailed IBM about it, but haven't gotten
a response yet...
Thanks,
NN