On Thursday 19 November 2009 06:21:45 William Bourque wrote: > > Michael Buesch wrote: > > Enforce all device constraints on the descriptor memory region. > > > > There are several constraints on the descriptor memory, as documented > > in the specification. The current code does not enforce them and/or > > incorrectly enforces them. > > > > Those constraints are: > > - The address limitations on 30/32bit engines, that also apply to > > the skbs. > > - The 4k alignment requirement on 30/32bit engines. > > - The 8k alignment requirement on 64bit engines. > > > > No gain on my bcm4312 [14e4:4315] on HP Mini 1116R (Atom architecture). > The DMA errors are still present.
Well, this patch doesn't fix known breakages. It just _ensures_ that the device constraints are met, so we can rule that out and can go on looking for other bugs. Another thing is that it removes the extremely ugly always-do-GFP_DMA thing. > Also, just saying, but it seems Larry's "pm_qos_update_requirement" > patch had some good effects; I can hardly get any connectivity without > it. With the patch, the wireless seems to be stable for a few minutes > before generating DMA errors... without the patch the error start piling > up as soon the interface get up. I think we do need the PM_QOS stuff. It clearly fixes part of the problem. I think we have multiple problems here. -- Greetings, Michael. _______________________________________________ Bcm43xx-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
