[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
R_WAITSTATES not set properly by kernel
I am experiencing a strange thing; the value of
R_WAITSTATES which I put into my kernelconfig, does not
seem to be written properly.
I'm using devboard_82 (2.4.20 kernel).
I can see the only piece of code in the kernel that
writes a value to this register in dram_init.S, and
I assume that this will be called very early, when
the kernel starts. (Or maybe not, since I'm using SDRAM?)
We use a device that needs two readstrobes for a 32bit
read, so we try to configure early waitstates for it.
I've set the config value of R_WAITSTATES to 0xffff95a6,
but our pheripheral device only gets one readpulse when
we do a 32bit read.
Now I've added a few lines to my driver, on init it now
printk's the value of CONFIG_ETRAX_DEF_R_WAITSTATES, used
by dram_init.S. It still has value 0xffff95a6.
After that, the driver writes this value to R_WAITSTATES again.
Now, when we do a 32bit read, our device gets two readpulses.
Why doesn't the pheripheral waitstate config work like it should?