From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bill Davidsen Subject: Re: 2.6.17-1.2174_FC5: : Disabling IRQ #7 / nobody cared / usb_hcd_irq Date: Wed, 06 Sep 2006 11:36:56 -0400 Message-ID: <44FEEB18.105@tmr.com> References: <790873504@web.de> <44FD8932.60906@tmr.com> <44FED3A4.2010204@web.de> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <44FED3A4.2010204@web.de> Sender: linux-laptop-owner@vger.kernel.org List-Id: Content-Type: text/plain; charset="us-ascii"; format="flowed" To: Simeon Penev Cc: linux-laptop@vger.kernel.org Did you all see that the Croc Hunter Simeon Penev wrote: >Hi Bill, > >thank you for the fast reply! > >Bill Davidsen wrote: > > >>Although IRQ7 is traditionally LP, it seems to get identified for almost >>any unexpected interrupt. I knew why once, but I haven't seen this for >>some time. The real problem is that the kernel disables the IRQ, making >>the system unusable. You can try any of several things: >>- try the last 2.6.16 kernel >>- remove the IRQ disable code, or add a counter. It happens only once on >>many machines, >> as something comes ready. >> >> > >I'll try that, but actually i wanted to help fix the problem if it's a >bug with the kernel. > > Matter of approach. I would not disable an IRQ until there had been some number of uncaught interrupts in a row, others feel the number is one. But there really is an interrupt, so it's probably not a "bug" in that sense. > > >>- disable any CPU speed stuff you have running (just for information) >> >> > >Yes, i'm using cpufreq, but it's a feature i'm not willing to give away >because the laptop is more silent. The fan works less. > > Noted, but "just for information" would let you see if the speedstep was the cause, or at least contributed. That would help run it down. > > >>Yes, that's a shotgun approach, I'm still running 2.6.16.2111 on FC4 >>because later kernels don't init the video or wireless correctly. >> >> >> > >Thank you once again. > I have a machine which does zero or one of these, so I'm familiar with the problem. I am running an older kernel which whines but doesn't disable, and since I don't need IRQ7 I wouldn't be too unhappy if it did disable. But I think there's an underlying hardware issue, or at least hardware to driver mismatch. Since you said irqpoll causes problems there's definitely something not working right. -- bill davidsen CTO TMR Associates, Inc Doing interesting things with small computers since 1979