2003-08-31 03:06:46

by Matthew Wilcox

Subject: [PATCH] vmalloc might sleep

So let's whack people upside the head when they misuse it.

vmalloc-might-sleep.diff:

Index: mm/vmalloc.c
===================================================================
RCS file: /var/cvs/linux-2.6/mm/vmalloc.c,v
retrieving revision 1.2
diff -u -p -r1.2 vmalloc.c
--- a/mm/vmalloc.c 23 Aug 2003 02:47:26 -0000 1.2
+++ b/mm/vmalloc.c 31 Aug 2003 03:04:25 -0000
@@ -438,7 +438,8 @@ fail:
*/
void *vmalloc(unsigned long size)
{
- return __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM, PAGE_KERNEL);
+ might_sleep();
+ return __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM, PAGE_KERNEL);
}

/**
@@ -451,6 +452,7 @@ void *vmalloc(unsigned long size)
*/
void *vmalloc_32(unsigned long size)
{
+ might_sleep();
return __vmalloc(size, GFP_KERNEL, PAGE_KERNEL);
}


--
"It's not Hollywood. War is real, war is primarily not about defeat or
victory, it is about death. I've seen thousands and thousands of dead bodies.
Do you think I want to have an academic debate on this subject?" -- Robert Fisk

2003-08-31 03:20:11

by Al Viro

Subject: Re: [PATCH] vmalloc might sleep

On Sun, Aug 31, 2003 at 04:06:43AM +0100, Matthew Wilcox wrote:
>
> So let's whack people upside the head when they misuse it.
>
> vmalloc-might-sleep.diff:
>
> Index: mm/vmalloc.c
> ===================================================================
> RCS file: /var/cvs/linux-2.6/mm/vmalloc.c,v
> retrieving revision 1.2
> diff -u -p -r1.2 vmalloc.c
> --- a/mm/vmalloc.c 23 Aug 2003 02:47:26 -0000 1.2
> +++ b/mm/vmalloc.c 31 Aug 2003 03:04:25 -0000
> @@ -438,7 +438,8 @@ fail:
> */
> void *vmalloc(unsigned long size)
> {
> - return __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM, PAGE_KERNEL);
> + might_sleep();
> + return __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM, PAGE_KERNEL);

__vmalloc() goes kmalloc() to get an array of struct page pointers. And
both for vmalloc() and vmalloc_32() kmalloc() will be called with GFP_KERNEL
as gfp_mask. Which already has might_sleep().