On Saturday 14 November 2009 23:29:38 Larry Finger wrote: > Michael, > > I now have MMIO traces after modprobe of wl and b43, both after a fresh cold > reboot. There are many similaries (of course), but some differences that could > be important in the DMA errors. > > In both dumps, I see the code advancing through the cores. I don't see the PCI > configuration read/writes, thus I don't see the core selection. Once the > SSB_IDHIGH register has the correct value for the 802.11 core, I see the code > read the SPROM. In Broadcom's driver, I see additional read/write cycles that > do > not happen in b43, but I cannot identify the target registers. The operations > in > question are: > > read16 base+0x0800 -> 0x2801 > write16 base+0x0800 <- 0x3801 > read16 base+0x280a -> 0x6dbe > write16 base+0x280a <- 0xedbe > > Any idea what is located at offsets 0x0800 and 0x280a? Nothing in the b43 dump > reads or writes either of those locations.
Are you sure this still is on the ssb? It could be another device in another region that just happens to be mapped after ssb. -- Greetings, Michael. _______________________________________________ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
