> I think it was the merger of the 64-bit PowerPC with the 32-bit ppc that led > to things like the phys_addr_t type, etc. There must be some other aspect > new to the powerpc tree which requires adjustment by the driver.
I'm pretty sure that when I had things working, it was with the powerpc (not ppc) tree, with the embedded device tree, etc. I think one difference might be that my platform had less memory or perhaps different firmware, so that the kernel addresses for PCI resources were below 4GB. I seem to remember maybe there was some automagic in ioremap() that expanded the IO addresses to the right place. It does seem that the driver is loading properly and then simply not getting receive completions for the packets sent by the subnet manager. Since other events are occurring, I would guess that something is going wrong with posting the receives, though I don't know exactly what it might be. One thing to try would be to load the ib_mthca module with fw_cmd_doorbell=1 That won't fix anything, but if it fails in a different way that might give a clue as to what's wrong. - R. -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html