On Thu, May 18, 2006 at 06:42:02PM -0400, Brian Walsh wrote: > Brian Walsh wrote: > > Brian Walsh wrote: > >> Greg KH wrote: > >>> On Wed, May 17, 2006 at 01:26:33PM -0400, Brian Walsh wrote: > >>>> I am having a problem with a cardbus modem. The cardbus device has an > >>>> internal > >>>> OHCI host controller and 2 usbserial devices attached to that > >>>> controller. I am > >>>> running a 2.6.16 kernel on an Arm ixp425 Coyote based system. > >>>> > >>>> The device works fine. However to insert the device I must first unload > >>>> the > >>>> ohci_hcd module. Insert the card and then let hotplug take care of > >>>> reinitializing everything. All the ohci host devices are reattached > >>>> along with > >>>> all the attached usb devices. > >>>> > >>>> If I insert the card without unloading the ohci_hcd module hotplug will > >>>> see and > >>>> attach the ohci host controller but never attaches the usbserial devices. > >>> Can you post the kernel log messages (with CONFIG_USB_DEBUG enabled) for > >>> when you do this? > >>> > >>> Oh, and does this work properly on a i386 machine? > >>> > >> I posted the log files. They were a bit long to post to the list. > >> > >> http://www.walsh.ws/usb_hang.txt > >> http://www.walsh.ws/usb_reload.txt > >> > >> In usb_reload.txt I first unload the ohci_hcd module then insert the > >> cardbus card. > >> > >> In usb_hang.txt the ohci_hcd module is already loaded when I insert the > >> card. > >> > >> I can not say at this time if it works on an i386 machine, as soon as I > >> can set > >> one up I will test it out. > >> > >> Brian > >> > >> > > > > After looking through the logs again it appears this may be an Arm related > > problem. The dmabounce message is saying there is a device on the pci bus > > and > > then the drivers attach. When it fails the dmabounce message is coming > > after > > the usb drivers have attached to the device. Anyway I am posting this > > problem > > over to the arm-kernel list as well. > > > > > Ok I am still not sure where the problem exactly lies but here is the fix I > came > up with. > > I moved the call to platform_notify up in the call to device_add. This allows > the device to fully attach to the system before the driver starts to mess with > it. I don't know if it is the best solution but it works for my hardware. > > Comments?
What does your platform_notify() do? What arch is this? Pointers to code anywhere? And as it looks like only ppc and arm platforms use this callback, you should really check with the maintainers of those arches before I could accept a patch like this. thanks, greg k-h ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
