Rogier Wolff wrote: > > Kurt Garloff wrote: > > > A simple test shows that the module count is working for > > > the sg driver. So it seems as if open() calls to char > > > devices do an auto module count increment. However I > > > was unable to find the code that does this (e.g. no sign > > > of it in fs/devices.c). > > > > > > Perhaps someone else could shed some light on this. > > > > No idea. > > Hi, > > FYI: The module counts are handled in a macro called something like > get_fops. Thanks Roger. So sg_open() is called from fs/devices.c chrdev_open() which also calls get_chrfops() which calls the macro fops_get() defined in include/linux/fs.h ! There is also a fops_put() for dealing with the decrement on close. It's a bit like a cryptic crossword. I haven't changed the sg driver in lk 2.4 in the last 6 months, but there have been about half a dozen changes by others. It would be nice if we were alerted, especially when code is removed, without comment. Block devices take another route via fs/block_dev.c which doesn't seem to involve calls to the macro fops_get() macro. So that means osst might be double incrementing the module count. It is not alone, Alan might like to do a 'grep MOD_INC drivers/char/*.c' if he is running out of things to do ... Doug Gilbert Dyslexia rules K.O. :-) - To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [EMAIL PROTECTED]
