On Thu, 2006-01-12 at 11:25 -0500, Medeiros Edward M NPRI wrote: [...] > Thank you for your reply. I am not implementing either a > "block" or "char" driver but rather a PCI device driver I fear/think you must since you probably want to deliver that data in user-space (and since you want ot use copy_to_user(), the destination *is* the user-space). > based on "pci_skel.c" from the "Linux Device Drivers, > 3rd edition" book. "pci_skel.c" has "probe()", > "remove()", "init()", and "exit()" functions only > and no "ioctl" functionality. It is unclear to me > how I can use a function like "copy_to_user()" > that requires a user-space destination address > with this type of driver.....Any suggestions? Ah, at that level (that means: at the lower part of your driver - user-space is on top of the upper part). I never really programmed PCI stuff but from the struct pci_driver decalartion as in http://www.cs.fsu.edu/~baker/devices/lxr/http/source/linux/include/linux/pci.h#L663 you kave just some "basic" infrastructure to start and stop the PCI card which is controlled by your driver. The interface at this low-level is just used by a char/block driver to manage the PCI card with probing etc. If, how and how much data a given PCI card delivers or accepts and what must be done to access that data depends completely on the given PCI card (usually you have I/O ports or memory areas to read from and/or write to - either via DMA or directly or God knows what hardware developer are creating). So at that level you have at most kernel space pointers. The user-space pointers probably come from a char or block driver. Perhaps you want to reread the first chapters about his "skull" driver to find the stuff I'm talking about. Bernd -- Firmix Software GmbH http://www.firmix.at/ mobil: +43 664 4416156 fax: +43 1 7890849-55 Embedded Linux Development and Services -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
