Dmitri wrote: >Quoting Ravi Kumar B S <[EMAIL PROTECTED]>: > >>i am using the usb bluetooth driver. i have configured the usb bluetooth >>driver as a module. i have two versions of the driver. one with the >>comments, extensive debug statements(printks) etc. the other one is the >>one which is cleaned does not have that many debug statements. >>now the problem is when i insmod the uncleaned one and connect the >>device, and do lsmod, the driver is listed as being used. (means, 1 in >>"used by" column of the lsmod result). >>but when i insmod the cleaned one, it is not listed as being used and >>open call to the device fails. >> >>any idea what is going wrong? >> > >Probably you made a mistake somewhere :-) > >In best case, you "cleaned out" too much, maybe overlook the scope of some >operators (like 'if') and introduced a bug. Diff the sources with emacs or >tkdiff, and review the changes. > >In worst case, presence of printk's moves the memory around, and you have >a stray pointer somewhere that reads or writes past the allocated >memory. With printks in place, this pointer happens to point to a safe >place; without them all fails. > >It helps sometimes to *slowly* move changes from one file to another, one >change at a time, and testing the driver. As soon as the change >breaks/fixes the code you know where the problem is. But sometimes *any* >change, *anywhere* affects the code (when you return something allocated >on the stack, for example). Then you have to stare at the code until you >find the bug. > >Dmitri > thanks. but the thing is that the cleaned up code works on another of my machine. and is not working only in a particular machine. i see the configuration as same in both the machines.
any idea? /Ravi kumar b-s _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
