h.OHCI error?
Dave Higton (281) 668 posts |
In OCHIDriver.h.OHCI, there should be 23 unused bits in the middle of the HcInterruptStatus register, according to the document I have (HCIR1_0A.PDF, Open Host Controller Interface Specification for USB, release 1.0a, 09/14/99 2:33 PM) on page 128, but our H file defines 22 bits. The consequence is that the OC and MIE bits are in the wrong position. I have no idea whether the discrepancy is of any consequence. |
Dave Higton (281) 668 posts |
Also, the fm_number register should have 16 bits allocated to FN, the Frame Number, but actually has 14 bits – and 2 bits unaccounted for. |
Jeffrey Lee (213) 6048 posts |
Well spotted – I’ve submitted a fix now. I think the only place that header gets used is for the keyboard scanning code in the Iyonix HAL. And even then, it makes mixed use of that h.OHCI and h.ohciregs, so it doesn’t look like those broken register definitions were actually being used anywhere. (And as you’ve probably worked out by now, the main OHCI driver sources are in the NetBSD folder, not Controllers.OHCIDriver, and the NetBSD code doesn’t make use of the headers that OHCIDriver exports) |