Re: [PATCH] fix error handling in load_module()

[Rusty Russell <rusty@rustcorp.com.au>]

On Tue, 22 Sep 2009 12:11:21 am Tejun Heo wrote:
> Hello, Andrew.
> 
> Andrew Morton wrote:
> > My reverse engineering of the secret, undocumented percpu_modfree()
> > indicates that its mad inventor intended that percpu_modfree(NULL) be a
> > valid thing to do.
> > 
> > It calls free_percpu(), all implementations of which appear to secretly
> > support free_percpu(NULL).
> 
> Eh... unfortunately, the original percpu_modfree() implementation
> didn't seem to support it.

OK, I'll Andrew's fix for Tejun, and after his (spot-on!) comment about
percpu_modfree never taking NULL, I've fixed the one caller to match
the other two:

Subject: module: don't call percpu_modfree on NULL pointer.

The general one handles NULL, the static obsolescent
(CONFIG_HAVE_LEGACY_PER_CPU_AREA) one in module.c doesn't; Eric's
commit 720eba31 assumed it did, and various frobbings since then kept
that assumption.

All other callers in module.c all protect it with an if; this effectively
does the same as free_init is only goto if we fail percpu_modalloc().

Reported-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Eric Dumazet <dada1@cosmosbay.com>
Cc: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Américo Wang <xiyou.wangcong@gmail.com>

diff --git a/kernel/module.c b/kernel/module.c
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -2522,8 +2522,8 @@ static noinline struct module *load_modu
  free_unload:
 	module_unload_free(mod);
 #if defined(CONFIG_MODULE_UNLOAD) && defined(CONFIG_SMP)
+	percpu_modfree(mod->refptr);
  free_init:
-	percpu_modfree(mod->refptr);
 #endif
 	module_free(mod, mod->module_init);
  free_core: