From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <4C2DBC81.705@domain.hid> Date: Fri, 02 Jul 2010 12:16:33 +0200 From: Gilles Chanteperdrix MIME-Version: 1.0 References: <4C289319.4050807@domain.hid> <4C2A1B0A.3000204@domain.hid> <4C2BB5C0.3020407@domain.hid> In-Reply-To: Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Subject: Re: [Xenomai-help] IRQ conflict issues with and ESD PMC card. List-Id: Help regarding installation and common use of Xenomai List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Nape Ibrahim Lentsoane Cc: xenomai@xenomai.org Nape Ibrahim Lentsoane wrote: > > > On Wed, Jun 30, 2010 at 11:23 PM, Gilles Chanteperdrix > > wrote: > > Nape Ibrahim Lentsoane wrote: > > Okay, we updated the BIOS with a custom made version from > Advantech. At > > least we are able too boot now but > > now my usb mouse/port does not work. By the way I'm running on an > > industrial SBC from Advantech model MIC-3390. > > The contents of /proc/interrupts and /proc/xenomai/irq are in > > "pro_interrupts.txt" and "xenomai_irq_2.txt" respectively > > as attached. > > > > Something strange which I do not understand is that ran the '# > lsmod | grep > > xeno_can' and look at the results in the > > attached file "lsmod_grep_xeno_can.txt". What is puzzling is that I > > configured the Xenomai CAN stuff to load as modules > > but here they are automatically loaded, how's that? > > > > I'll try to enable the APIC and local APIC and see what happens. > > Ok. Now we can see things more clearly. First: the APIC is already > enabled in your kernel configuration, and seems not to be used, the ACPI > is broken, I do not know if this is the problem, but it seems your APIC > is not used. This is bad news. If that is possible, you should try and > use a BIOS with ACPI. I mean, come on, this has been the standard for > over 10 years now. See: > http://en.wikipedia.org/wiki/ACPI > > Second, your kernel is not really well configured: > - you have enabled multi-processor support whereas you have a single > CPU, disable it > - you probably use the generic IDE driver, whereas you should use the > one for the hardware you really have, use lspci to find out what the > hardware you use is. > - you enabled Xenomai SMI workaround, at this stage, this is a bad idea, > you should follow the guidelines in the "TROUBLESHOOTING" guide in order > to enable it. But this is too early, please keep it disabled for now. > > And now, the really bad news: you have the IRQ shared between Xenomai > driver and Linux driver issue. IRQ 12 is used both by the Xenomai CAN > driver and by the Linux USB controller. This is probably the reason why > the USB is not working. See the FAQ for suggestions on how you may solve > this issue. > > Okay, I can work without the USB for now but I don't understand > something here. > We have have similar other SBCs with Arcnet cards installed and running > RTNet and > some RTAI and they seem work well even if an IRQ is shared between RT > and non-RT. > For an example our real time Arcnet driver is sharing an IRQ with the > SMBus and there no issues there. > Just out of curiosity, why is this an problem with Xenomai? It is as much a problem with RTAI as with Xenomai. You probably have no issues with RTAI because you only use the real-time or non real-time driver, or you have high IRQ latencies due to the masking until the non-RT driver handles the interrupt. In other words, you may have issues but do not see them. -- Gilles.