On Tue, Feb 21, 2006 at 10:30:38AM -0500, Gabriel Somlo wrote: > On 2/20/06, Greg KH <greg@xxxxxxxxx> wrote: > > > > When I place the card into a 64-bit PCI-66 slot (not pci-x, 3.3v only), > > > the driver's probe function crashes the kernel when it attempts to call > > > pci_read_config_dword. > > > > It does? That's not good. > > Thanks for the reply ! > > > What exactly does the code look like that is causing this crash? > > So: foo_init() calls pci_module_init() and passes it a struct pci_driver > with foo_init_one() as the 'probe' callback. > > I traced foo_init_one() (using printk statements), and it does pretty much > the usual stuff: pci_enable_device(), pci_set_master(), pci_set_mwi(), > figure out IO memory using pci_resource_start() and pci_resource_len(), > set up the IRQ stuff with requst_irq()... > > This all seems to work fine, up until the point where pci_read_config_dword() > is called. There, the system just freezes up. No panic, no crash output, just > everything locks up (the caps lock light no the keyboard still toggles, but > that's pretty much about it). Hm, do you have a pointer to the source code itself? That's going to be the only way we can try to fix this. > > And, which kernel version is this happening on? > > It's the latest Fedora kernel -- 2.6.15-1.1831_FC4smp. > > > What exactly is the crash output? > > The single-processor version of the kernel (2.6.15-1.1831_FC4) prints some > extra info before the crash: > > ACPI: PCI Interrupt Link [LNKF] enabled at IRQ 11 > PCI: setting IRQ 11 as level-triggered > > Then, nothing... Then how do you know that this line causes the crash? :) thanks, greg k-h -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/