[PATCH v2] mm: module_alloc: check if size is 0

[Veli-Pekka Peltola]

Hi,

On 03/07/2012 03:09 PM, Veli-Pekka Peltola wrote:
> After commit de7d2b567d040e3b67fe7121945982f14343213d (mm/vmalloc.c: report
> more vmalloc failures) users will get a warning if vmalloc_node_range() is
> called with size 0. This happens if module's init size equals to 0. This
> patch changes ARM, MIPS and x86 module_alloc() to return NULL before calling
> vmalloc_node_range() that would also return NULL and print a warning.
>
> Signed-off-by: Veli-Pekka Peltola<veli-pekka.peltola at bluegiga.com>
> Cc: Russell King<linux at arm.linux.org.uk>
> Cc: Thomas Gleixner<tglx at linutronix.de>
> Cc: Ingo Molnar<mingo at redhat.com>
> Cc: "H. Peter Anvin"<hpa at zytor.com>
> Cc: x86 at kernel.org
> ---
> I found this with ARM but after checking out various implementations of
> module_alloc() I thought it would be better to fix all at once.
>
> One way to replicate the warning:
> compile kernel with CONFIG_KALLSYMS=n
> insmod a module without init, I used usb-common.ko
>
> Changes since v1:
>   - changed style as hpa suggested
>
>   arch/arm/kernel/module.c  |    2 ++
>   arch/mips/kernel/module.c |    2 ++
>   arch/x86/kernel/module.c  |    2 +-
>   3 files changed, 5 insertions(+), 1 deletions(-)
>
> diff --git a/arch/arm/kernel/module.c b/arch/arm/kernel/module.c
> index 1e9be5d..17648e2 100644
> --- a/arch/arm/kernel/module.c
> +++ b/arch/arm/kernel/module.c
> @@ -39,6 +39,8 @@
>   #ifdef CONFIG_MMU
>   void *module_alloc(unsigned long size)
>   {
> +	if (!size)
> +		return NULL;
>   	return __vmalloc_node_range(size, 1, MODULES_VADDR, MODULES_END,
>   				GFP_KERNEL, PAGE_KERNEL_EXEC, -1,
>   				__builtin_return_address(0));
> diff --git a/arch/mips/kernel/module.c b/arch/mips/kernel/module.c
> index a5066b1..1a51de1 100644
> --- a/arch/mips/kernel/module.c
> +++ b/arch/mips/kernel/module.c
> @@ -47,6 +47,8 @@ static DEFINE_SPINLOCK(dbe_lock);
>   #ifdef MODULE_START
>   void *module_alloc(unsigned long size)
>   {
> +	if (!size)
> +		return NULL;
>   	return __vmalloc_node_range(size, 1, MODULE_START, MODULE_END,
>   				GFP_KERNEL, PAGE_KERNEL, -1,
>   				__builtin_return_address(0));
> diff --git a/arch/x86/kernel/module.c b/arch/x86/kernel/module.c
> index 925179f..fd44d69 100644
> --- a/arch/x86/kernel/module.c
> +++ b/arch/x86/kernel/module.c
> @@ -38,7 +38,7 @@
>
>   void *module_alloc(unsigned long size)
>   {
> -	if (PAGE_ALIGN(size)>  MODULES_LEN)
> +	if (!size || PAGE_ALIGN(size)>  MODULES_LEN)
>   		return NULL;
>   	return __vmalloc_node_range(size, 1, MODULES_VADDR, MODULES_END,
>   				GFP_KERNEL | __GFP_HIGHMEM, PAGE_KERNEL_EXEC,

Any comments on this? Should I split all architectures to separate patches?

I just tested 3.3 on ARM and x86, both printed a warning and call trace 
without this patch.

--
Veli-Pekka Peltola