Thanks for the help,
[email protected] wrote:
Have you tried reading other registers in this space? Are all giving
you 0xffffffff?
Yes, I have accessed 1k of contiguous addresses in this space (with a
for loop around a ddi_get32). On my PCI Bus Analyzer, the device
returns 0x00000000 for empty registers -- but I still see 0xffffffff
returned in my driver.
I have also written a scratch register and read the results. When I
write my scratch registers with 0xdeaddead, I see 0xdeaddead written to
my scratch register with by PCI Bus Analyzer; and I see it read out
(with my subsequent ddi_get32) as 0xdeaddead, but my function still
returns 0xffffffff.
So this appears to be only a get issue. My puts seems to work fine.
***
As an aside, I tried to read the register space with:
int32_t val=0xabcd;
status=ddi_peek32(dip,addr,&val);
It never completes (I don't see a read on my bus analyzer), and instead
returns a status=DDI_FAILURE while corrupting val with invalid data.
I don't know how to diagnose what ddi_peek32 doesn't like.
Does your code compile without any warnings?
No warnings. I am using the default sun compiler:
% cc -V
cc: Sun Ceres C 5.10 SunOS_1386 2008/10/22
Since the space is supposed to be memory mapped, have you tried
looking at the memory using mdb?
No I have not. I will start learning how to use that tool today.
Thanks for pointing me in that direction.
-Adam
_______________________________________________
driver-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/driver-discuss
