Re: [RFC PATCH] kernel/module: add a safer implementation of try_module_get()

On 2024-01-30 21:47, Luis Chamberlain wrote: > On Tue, Jan 30, 2024 at 08:36:14PM +0100, Marco Pagani wrote: >> The current implementation of try_module_get() requires the module to >> exist and be live as a precondition. While this may seem intuitive at >> first

[RFC PATCH] kernel/module: add a safer implementation of try_module_get()

efore attempting to take it. The tree lookup and refcount increment are executed after taking the module_mutex to prevent the module from being unloaded after looking up the tree. Signed-off-by: Marco Pagani --- include/linux/module.h | 15 +++ kernel/module/main.c

[PATCH] kernel/module: improve documentation for try_module_get()

The sentence "this call will fail if the module is already being removed" is potentially confusing and may contradict the rest of the documentation. If one tries to get a module that has already been removed using a stale pointer, the kernel will crash. Signed-off-by: Marco Pagani --