In my ongoing soap opera - trying to use a DittoMax with a parallel port
card from Siig - I've finally reached some sort of conclusion: memory is
being corrupted during driver activity. A lot of other strange behavior
turns out to be manifestations of this memory corruption.
The corruption appears to be happening when bpck_fdc_read_reg_value()
is in use, possibly when it is reading the FT_BPCK_REG_STAT register.
But it's hard to be sure, the corruption also mangles the logging output,
making it difficult to identify exactly when things go wrong. The
corruption does not not happen immediately - the drive can be found
when the driver is loaded - but it happens shortly after I try any
drive operation.
So that's where my efforts sit with ftape: corruption is clearly happening
during bpck-fdc driver use. Race condition? Failure to disable IRQ at
some critical time? Cosmic ray? Hardware weirdness?
Any insights would be most welcome.
Nathan Meyers
[EMAIL PROTECTED]
