Hello Yesterday I was working on a network device driver problem. This is a custom built network device driver written by a friend of mine. The driver would insmod ok, but when it came time to ifconfig the network device associated with this driver, I got a bus error and the following oops message: Oops: kernel access of bad area, sig: 11 <stuff deleted> Call backtrace: FFFF050F C902AD0C C902F878 C00CDBCC C00C3B1C C00C9B84 C00DA880 C00DB6C4 C00FBF84 C0104314 C00BB49C C00BC740 C00BD02C C0003C50 <stuff deleted> So, the way I approached this problem was to examine the call backtrace address's. The FFFF050F address was most likely bogus so I looked at the next address, C902AD0C. I examined /proc/ksyms for the closet match to this address and noticed it was for a routine in the device driver. That was expected. However, the routine was for a function call which didn't look right. I thought that when you ifconfig a network device, the kernel calls the do_ioctl registered routine. The routine corresponding to C902AD0C was for a non-ioctal routine. A routine which isn't even called from the ioctl routine. Anyway, I put printk's and a early return ( return(0); ) in the ioctl routine and tried it again. The code still oop'd and the printk was not printed out. I was royally confused. I finally resorted to diffing the last set of changes and discovered that the code had disabled interrupts prior to requesting interrupts and then never reenabled interrupts. I fixed that and the problem went away. I am worried that a problem still exists since that doesn't seem like a good reason for the problem. Your comments are welcome. JD -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
