Re: [PATCH] vmalloc_32 should use GFP_KERNEL

[Andrew Morton]

On Wed, 18 Jul 2007 16:25:34 +1000 Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx> wrote:

> I've noticed lots of failures of vmalloc_32 on machines where it
> shouldn't have failed unless it was doing an atomic operation.
> 
> Looking closely, I noticed that:
> 
> #if defined(CONFIG_64BIT) && defined(CONFIG_ZONE_DMA32)
> #define GFP_VMALLOC32 GFP_DMA32
> #elif defined(CONFIG_64BIT) && defined(CONFIG_ZONE_DMA)
> #define GFP_VMALLOC32 GFP_DMA
> #else
> #define GFP_VMALLOC32 GFP_KERNEL
> #endif
> 
> Which seems to be incorrect, it should always -or- in the DMA flags
> on top of GFP_KERNEL, thus this patch.
> 
> Signed-off-by: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx>
> ---
> 
> This fixes frequent errors launchin X with the nouveau DRM for example.
> 
> Index: linux-work/mm/vmalloc.c
> ===================================================================
> --- linux-work.orig/mm/vmalloc.c	2007-07-18 16:22:00.000000000 +1000
> +++ linux-work/mm/vmalloc.c	2007-07-18 16:22:11.000000000 +1000
> @@ -578,9 +578,9 @@ void *vmalloc_exec(unsigned long size)
>  }
>  
>  #if defined(CONFIG_64BIT) && defined(CONFIG_ZONE_DMA32)
> -#define GFP_VMALLOC32 GFP_DMA32
> +#define GFP_VMALLOC32 GFP_DMA32 | GFP_KERNEL
>  #elif defined(CONFIG_64BIT) && defined(CONFIG_ZONE_DMA)
> -#define GFP_VMALLOC32 GFP_DMA
> +#define GFP_VMALLOC32 GFP_DMA | GFP_KERNEL
>  #else
>  #define GFP_VMALLOC32 GFP_KERNEL
>  #endif
> 

whoops, yes.

Are those errors serious and common enough for 2.6.22.x?  
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/