On Wed, 4 Mar 2009 18:30:47 -0500 Denis Hainsworth <de...@alumni.brandeis.edu> wrote:
> So since machine1 and machine3 are running the same exact software it > would seem to indicate a hardware issue. Yet the same hardware > running openbsd 4.4 also seems ok which tends to indicate some odd > interaction between the hardware and openbsd 4.2. > > So if anyone has any thoughts they would be appreciated as I have been > going around in circles on this issue for a while now. I realize you said you're new to OpenBSD, but the rule to remember is, if you're not running a GENERIC kernel, then you're on your own. It's already difficult to replicate problems using similar hardware and the same GENERIC kernel, but trying to replicate problems found in an undefined custom kernel is just wasted effort. This is the reason why custom kernels are unsupported. You said the support for the controller was hacked into 4.2 by adding the PCI Device-ID, and you also said you are "unaware of any other changes" but that certainly doesn't mean other changes were not made. All of this means you cannot replicate the problem with the 4.2 GENERIC kernel because there is no default support for the device. Additionally, 4.2 is a year and a half old, and you're only noticing this issue now, so you obviously did not do much testing or monitoring of your custom support for the device. As your test with 4.4 has shown, proper support for the device has been added, and the self-inflicted performance problem you mentioned no longer exists. I realize you said there's some supposed "reason" why you cannot upgrade, but this self-imposed limitation leaves you only one choice: You can back-port all of the changes from 4.4-release, or 4.4-stable, or even better 4.5-beta to your customized 4.2 build. Even if you did succeed in back-porting all the changes to 4.2, you'd still be running your own custom FrankenSource monster, and once again, no one would realistically be able to help you with it. It's a whole lot easier to just upgrade. -- J.C. Roberts