On 11/13/2009 05:16 AM, Michael Buesch wrote: > Ok, so my guess is that the DMA allocator simply returned high memory > that was unusable to the device. My new code explicitly checks for that (and a > few other things) and retries with GFP_DMA in case the address has illegal > bits set. > That's the same thing we do for the frame buffers, so I don't see anything > wrong with it. > > Yeah, the new code is big and scary, but I think it makes a whole lot more > sense than > what we have now. Especially if I add some comments and do more cleanups.
I agree that the new code makes a lot of sense. I'm also beginning to believe that the problem lies outside b43, and that it is merely triggered by some interaction with ACPI and/or the BIOS. From what I found in looking back through the DMA error reports, most (if not all) people with the problem have netbook computers with Intel ATOM processors. I am considering posting on LKML and the ACPI mailing lists to see if we can get any ideas from those experts. Please comment on the draft text below: ============================ A number of users are experiencing DMA descriptor or data errors using 64-bit DMA with the Broadcom BCM4312 wireless device. After careful review and a rewrite of the DMA code in the driver, we have not been able to fix the problem, but we have determined the following: (1) The problem is much more likely to occur on netbook systems. Several of the developers have this card in regular notebook systems. None of us have the problem, thus it may occur only on netbooks, but several brand/model combinations are affected including Dell Inspiron 910 and Acer Aspire One A150. (2) If CONFIG_ACPI_PROCESSOR is not set on affected systems, the error rate is much lower. (3) When a DMA descriptor error occurs, a dump of the descriptors does not reveal any obvious problems. I do not know enough about either the ACPI or DMA code to begin debugging in either of those regions. Any suggestions on debugging strategies, or links to similar problems would be appreciated. ============================ Larry _______________________________________________ Bcm43xx-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
