RCU init call bootmem allocation after slub us ready.

[Yinghai Lu]

[ 0.000000] SLUB: Genslabs=15, HWalign=64, Order=0-3, MinObjects=0,
CPUs=160, Nodes=8
[ 0.000000] Hierarchical RCU implementation.
[ 0.000000] RCU restricting CPUs from NR_CPUS=4096 to nr_cpu_ids=160.
[ 0.000000] ------------[ cut here ]------------
[ 0.000000] WARNING: at mm/nobootmem.c:234
___alloc_bootmem_nopanic+0x3a/0x7d()
[ 0.000000] Modules linked in:
[ 0.000000] CPU: 0 PID: 0 Comm: swapper/0 Not tainted
3.9.0-yh-10495-gfc7ffbc-dirty #1543
[ 0.000000] Hardware name: Oracle Corporation Sun Fire X4800 M2 /
, BIOS 15013200 04/19/2012
[ 0.000000] 0000000000000009 ffffffff82a01e58 ffffffff8209dc61
ffffffff82a01e98
[ 0.000000] ffffffff8108a9d0 0000000000000000 0000000000000000
0000000000000200
[ 0.000000] ffffffffffffffff 0000000000000040 0000000000000000
ffffffff82a01ea8
[ 0.000000] Call Trace:
[ 0.000000] [<ffffffff8209dc61>] dump_stack+0x19/0x1b
[ 0.000000] [<ffffffff8108a9d0>] warn_slowpath_common+0x70/0xa0
[ 0.000000] [<ffffffff8108aa1a>] warn_slowpath_null+0x1a/0x20
[ 0.000000] [<ffffffff82dd54cd>] ___alloc_bootmem_nopanic+0x3a/0x7d
[ 0.000000] [<ffffffff82dd5695>] __alloc_bootmem+0x12/0x38
[ 0.000000] [<ffffffff82de976a>] alloc_bootmem_cpumask_var+0x1a/0x21
[ 0.000000] [<ffffffff82dce4da>] rcu_init+0x66/0x283
[ 0.000000] [<ffffffff82dd15ed>] ? perf_event_init+0x136/0x15f
[ 0.000000] [<ffffffff82dabf95>] start_kernel+0x24a/0x423
[ 0.000000] [<ffffffff82dababe>] ? repair_env_string+0x5e/0x5e
[ 0.000000] [<ffffffff820936b6>] ? __memblock_reserve+0x81/0x89
[ 0.000000] [<ffffffff82dab5a8>] x86_64_start_reservations+0x2f/0x31
[ 0.000000] [<ffffffff82dab79e>] x86_64_start_kernel+0x1f4/0x1fc
[ 0.000000] ---[ end trace 53ae8bb075006d8f ]---
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majo...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

[Paul E. McKenney]

Do you have commit fa0f93cd6a (rcu: Don't allocate bootmem from rcu_init())?
If not, could you please try applying it?

Thanx, Paul

[Yinghai Lu]

On Mon, May 6, 2013 at 12:06 PM, Paul E. McKenney
<pau...@linux.vnet.ibm.com> wrote:

> Do you have commit fa0f93cd6a (rcu: Don't allocate bootmem from rcu_init())?
> If not, could you please try applying it?

oh, must be in tip/rcu branch...

i only test tip/x86 for a while.

sorry for the noise.

Yinghai

[Paul E. McKenney]

As Dave Jones pointed out, this one is not yet in mainline. Here is the
corresponding patch.

Thanx, Paul

------------------------------------------------------------------------

rcu: Don't allocate bootmem from rcu_init()

When rcu_init() is called we already have slab working, allocating
bootmem at that point results in warnings and an allocation from
slab. This commit therefore changes alloc_bootmem_cpumask_var() to
alloc_cpumask_var() in rcu_bootup_announce_oddness(), which is called
from rcu_init().

Signed-off-by: Sasha Levin <sasha...@oracle.com>
Signed-off-by: Paul E. McKenney <pau...@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <jo...@joshtriplett.org>

diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h
index ca6e39c..44b0998 100644
--- a/kernel/rcutree_plugin.h
+++ b/kernel/rcutree_plugin.h
@@ -86,7 +86,7 @@ static void __init rcu_bootup_announce_oddness(void)
#ifdef CONFIG_RCU_NOCB_CPU
#ifndef CONFIG_RCU_NOCB_CPU_NONE
if (!have_rcu_nocb_mask) {
- alloc_bootmem_cpumask_var(&rcu_nocb_mask);
+ alloc_cpumask_var(&rcu_nocb_mask, GFP_KERNEL);
have_rcu_nocb_mask = true;
}
#ifdef CONFIG_RCU_NOCB_CPU_ZERO

[Yinghai Lu]

Yes, found that in you reply to Dave Jones.

>
> Thanx, Paul
>
> ------------------------------------------------------------------------
>
> rcu: Don't allocate bootmem from rcu_init()
>
> When rcu_init() is called we already have slab working, allocating
> bootmem at that point results in warnings and an allocation from
> slab. This commit therefore changes alloc_bootmem_cpumask_var() to
> alloc_cpumask_var() in rcu_bootup_announce_oddness(), which is called
> from rcu_init().
>
> Signed-off-by: Sasha Levin <sasha...@oracle.com>
> Signed-off-by: Paul E. McKenney <pau...@linux.vnet.ibm.com>
> Reviewed-by: Josh Triplett <jo...@joshtriplett.org>
>
> diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h
> index ca6e39c..44b0998 100644
> --- a/kernel/rcutree_plugin.h
> +++ b/kernel/rcutree_plugin.h
> @@ -86,7 +86,7 @@ static void __init rcu_bootup_announce_oddness(void)
> #ifdef CONFIG_RCU_NOCB_CPU
> #ifndef CONFIG_RCU_NOCB_CPU_NONE
> if (!have_rcu_nocb_mask) {
> - alloc_bootmem_cpumask_var(&rcu_nocb_mask);
> + alloc_cpumask_var(&rcu_nocb_mask, GFP_KERNEL);

alloc_bootmem will return zeroed buffer.

so need to use zalloc_cpumask_var instead.