rmmod unloads loadable modules from the running kernel.
rmmod tries to unload a set of modules from the kernel, with the restriction that they are not in use and that they are not referred to by other modules.
If more than one module is named on the command line, the modules will be removed in the given order. This supports unloading of stacked modules.
With the option '-r', a recursive removal of modules will be attempted. This means that if a top module in a stack is named on the command line, all modules that are used by this module will be removed as well, if possible.
Do autoclean: tag unused modules as
Save persistent data for the named modules, without unloading any modules. If no module names are specified then data is saved for all modules that have persistent data. Data is only saved if both the kernel and modutils support persistent data and /proc/ksyms contains an entry __insmod_modulename_Ppersistent_filename
Display a summary of options and immediately exit.
Remove a module stack.
Output everything to syslog(3) instead of the terminal.
If a module contains persistent data (see insmod(8) and modules.conf(5)) then removing the module always writes the persistent data to the filename in the insmod _P symbol entry. You can also save the persistent data at any time by rmmod -e__, this will not unload any modules.
Generated comment lines start with '#%', all generated comments are stripped from the existing file, other comments are preserved. The saved data values are written to the file, preserving the existing order of comments and assignments. New values are added at the end of the file. If the file contains values that do not exist in the module then these values are preserved but are preceded by a generated comment warning that they are not being used. The latter operation allows a user to switch between kernels without losing persistent data and without getting any error messages.