On Tue, Feb 17, 2009 at 09:16:41AM -0700, Bjorn Helgaas wrote: > On Saturday 14 February 2009 02:30:02 pm Henrique de Moraes Holschuh wrote: > > On Thu, 12 Feb 2009, Bjorn Helgaas wrote: > > > about why they recommend it be disabled? Presumably they'd like to > > > have the EDAC functionality (maybe on Windows?), so I would think that > > > they'd only recommend disabling the device if it were broken or there > > > were some other avenue for supporting EDAC. > > > > Or if whatever Windows does to PCI devices by default on suspend/resume is > > not appropriate for the overflow device... etc. > > > > All I know is that this idiotic hiding of the overflow device breaks EDAC > > functionality on all of my perfectly EDAC'able D875PBZ unless I manually > > patch the kernel to unhide it. > > > > And since I have been running these boards with EDAC active using local > > hacks to deal with the overflow device for *years* (well before EDAC was > > even merged in mainline, and was still called bluesmoke), I know for sure it > > is safe on this board. > > > > In fact, do we have any reports of misbehaviour when one unhides the device? > > Because I only know of sucess cases... > > I don't have any data in this case, so I'm not arguing against you. But > here's why I asked the question of whether we have any indication of why > Intel recommends disabling it: I work for a hardware manufacturer, so I > see some of the stuff that happens before shipping a product. Sometimes > there are hardware defects that show up only in extremely rare situations > that a user is very unlikely to encounter. But when they do happen, they > might cause a crash or hang or even worse, a silent corruption of data. > Sometimes it's impossible to fix the original hardware defect, so the best > resolution is to disable the relevant functionality by leaving it hidden, > e.g., completely undocumented, having the BIOS turn it off, leaving a > connector off the motherboard, etc. Combining the first two "hardware workarounds" is what got us here. My opinion is that "turning it off" is a special case of "using", so when that gets undocumented theres non-zero probability that someone will find that the device can be turned off and might want to know what else it can do. Documenting bugs is better than hiding them. Regarding original topic: this device *is* documented, at least in the part EDAC guys are using, but without any mention of the disable bit. Best Regards, Michal Miroslaw -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
