Patchwork powerpc/mpic: fix WARN_ON from alloc_bootmem

login
register
mail settings
Submitter Jeremy Kerr
Date June 29, 2009, 6:36 a.m.
Message ID <1246257419.938592.49700983608.1.gpush@pingu>
Download mbox | patch
Permalink /patch/29262/
State Superseded
Headers show

Comments

Jeremy Kerr - June 29, 2009, 6:36 a.m.
Current powerpc/merge on a QS21 reports an oops on boot:

------------[ cut here ]------------
Badness at mm/bootmem.c:535
 [snip]
Call Trace:
[c000000000763a80] [c0000000005dd068] .alloc_arch_preferred_bootmem+0x28/0x9c (unreliable)
[c000000000763b10] [c0000000005ddaac] .___alloc_bootmem_nopanic+0x5c/0x130
[c000000000763bc0] [c0000000005ddbd8] .___alloc_bootmem+0x28/0x68
[c000000000763c50] [c0000000005dddf0] .__alloc_bootmem+0x1c/0x30
[c000000000763cd0] [c0000000005d1080] .mpic_alloc+0x20c/0x9f8
[c000000000763dc0] [c0000000005d3348] .cell_init_irq+0x70/0x114
[c000000000763e60] [c0000000005c41f8] .init_IRQ+0x44/0x64
[c000000000763ee0] [c0000000005c0914] .start_kernel+0x238/0x440
[c000000000763f90] [c000000000007368] .start_here_common+0x1c/0x34

Looks like we're using alloc_bootmem after the slab is up and running,
in mpic_init.

This change uses a kzalloc instead - in fact, we've already done a
kzalloc() earlier in this function.

Signed-off-by: Jeremy Kerr <jk@ozlabs.org>

---
 arch/powerpc/sysdev/mpic.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
Kumar Gala - June 29, 2009, 2:51 p.m.
On Jun 29, 2009, at 1:36 AM, Jeremy Kerr wrote:

> Current powerpc/merge on a QS21 reports an oops on boot:
>
> ------------[ cut here ]------------
> Badness at mm/bootmem.c:535
> [snip]
> Call Trace:
> [c000000000763a80] [c0000000005dd068] .alloc_arch_preferred_bootmem 
> +0x28/0x9c (unreliable)
> [c000000000763b10] [c0000000005ddaac] .___alloc_bootmem_nopanic+0x5c/ 
> 0x130
> [c000000000763bc0] [c0000000005ddbd8] .___alloc_bootmem+0x28/0x68
> [c000000000763c50] [c0000000005dddf0] .__alloc_bootmem+0x1c/0x30
> [c000000000763cd0] [c0000000005d1080] .mpic_alloc+0x20c/0x9f8
> [c000000000763dc0] [c0000000005d3348] .cell_init_irq+0x70/0x114
> [c000000000763e60] [c0000000005c41f8] .init_IRQ+0x44/0x64
> [c000000000763ee0] [c0000000005c0914] .start_kernel+0x238/0x440
> [c000000000763f90] [c000000000007368] .start_here_common+0x1c/0x34
>
> Looks like we're using alloc_bootmem after the slab is up and running,
> in mpic_init.
>
> This change uses a kzalloc instead - in fact, we've already done a
> kzalloc() earlier in this function.
>
> Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
>
> ---
> arch/powerpc/sysdev/mpic.c |    3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/arch/powerpc/sysdev/mpic.c b/arch/powerpc/sysdev/mpic.c
> index d46de1f..a5980e6 100644
> --- a/arch/powerpc/sysdev/mpic.c
> +++ b/arch/powerpc/sysdev/mpic.c
> @@ -1109,9 +1109,8 @@ struct mpic * __init mpic_alloc(struct  
> device_node *node,
> 			psize /= 4;
> 			bits = intvec_top + 1;
> 			mapsize = BITS_TO_LONGS(bits) * sizeof(unsigned long);
> -			mpic->protected = alloc_bootmem(mapsize);
> +			mpic->protected = kzalloc(mapsize, GFP_KERNEL);
> 			BUG_ON(mpic->protected == NULL);
> -			memset(mpic->protected, 0, mapsize);
> 			for (i = 0; i < psize; i++) {
> 				if (psrc[i] > intvec_top)
> 					continue;

I posted a similar patch that is in linus's tree:

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=85355bb272db31a3f2dd99d547eef794805e1319

- k

Patch

diff --git a/arch/powerpc/sysdev/mpic.c b/arch/powerpc/sysdev/mpic.c
index d46de1f..a5980e6 100644
--- a/arch/powerpc/sysdev/mpic.c
+++ b/arch/powerpc/sysdev/mpic.c
@@ -1109,9 +1109,8 @@  struct mpic * __init mpic_alloc(struct device_node *node,
 			psize /= 4;
 			bits = intvec_top + 1;
 			mapsize = BITS_TO_LONGS(bits) * sizeof(unsigned long);
-			mpic->protected = alloc_bootmem(mapsize);
+			mpic->protected = kzalloc(mapsize, GFP_KERNEL);
 			BUG_ON(mpic->protected == NULL);
-			memset(mpic->protected, 0, mapsize);
 			for (i = 0; i < psize; i++) {
 				if (psrc[i] > intvec_top)
 					continue;