2010-06-16 04:37:05

by KOSAKI Motohiro

[permalink] [raw]
Subject: [PATCH] mempolicy: reduce stack size of migrate_pages()


Now, migrate_pages() are using >500 bytes stack. This patch reduce it.

mm/mempolicy.c: In function 'sys_migrate_pages':
mm/mempolicy.c:1344: warning: the frame size of 528 bytes is larger than
512 bytes

Signed-off-by: KOSAKI Motohiro <[email protected]>
Cc: Christoph Lameter <[email protected]>
---
mm/mempolicy.c | 35 ++++++++++++++++++++++-------------
1 files changed, 22 insertions(+), 13 deletions(-)

diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index 13b09bd..1116427 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -1275,33 +1275,39 @@ SYSCALL_DEFINE4(migrate_pages, pid_t, pid, unsigned long, maxnode,
const unsigned long __user *, new_nodes)
{
const struct cred *cred = current_cred(), *tcred;
- struct mm_struct *mm;
+ struct mm_struct *mm = NULL;
struct task_struct *task;
- nodemask_t old;
- nodemask_t new;
nodemask_t task_nodes;
int err;
+ NODEMASK_SCRATCH(scratch);
+ nodemask_t *old = &scratch->mask1;
+ nodemask_t *new = &scratch->mask2;
+
+ if (!scratch)
+ return -ENOMEM;

- err = get_nodes(&old, old_nodes, maxnode);
+ err = get_nodes(old, old_nodes, maxnode);
if (err)
- return err;
+ goto out;

- err = get_nodes(&new, new_nodes, maxnode);
+ err = get_nodes(new, new_nodes, maxnode);
if (err)
- return err;
+ goto out;

/* Find the mm_struct */
read_lock(&tasklist_lock);
task = pid ? find_task_by_vpid(pid) : current;
if (!task) {
read_unlock(&tasklist_lock);
- return -ESRCH;
+ err = -ESRCH;
+ goto out;
}
mm = get_task_mm(task);
read_unlock(&tasklist_lock);

+ err = -EINVAL;
if (!mm)
- return -EINVAL;
+ goto out;

/*
* Check if this process has the right to modify the specified
@@ -1322,12 +1328,12 @@ SYSCALL_DEFINE4(migrate_pages, pid_t, pid, unsigned long, maxnode,

task_nodes = cpuset_mems_allowed(task);
/* Is the user allowed to access the target nodes? */
- if (!nodes_subset(new, task_nodes) && !capable(CAP_SYS_NICE)) {
+ if (!nodes_subset(*new, task_nodes) && !capable(CAP_SYS_NICE)) {
err = -EPERM;
goto out;
}

- if (!nodes_subset(new, node_states[N_HIGH_MEMORY])) {
+ if (!nodes_subset(*new, node_states[N_HIGH_MEMORY])) {
err = -EINVAL;
goto out;
}
@@ -1336,10 +1342,13 @@ SYSCALL_DEFINE4(migrate_pages, pid_t, pid, unsigned long, maxnode,
if (err)
goto out;

- err = do_migrate_pages(mm, &old, &new,
+ err = do_migrate_pages(mm, old, new,
capable(CAP_SYS_NICE) ? MPOL_MF_MOVE_ALL : MPOL_MF_MOVE);
out:
- mmput(mm);
+ if (mm)
+ mmput(mm);
+ NODEMASK_SCRATCH_FREE(scratch);
+
return err;
}

--
1.6.5.2



2010-06-16 05:00:48

by Kamezawa Hiroyuki

[permalink] [raw]
Subject: Re: [PATCH] mempolicy: reduce stack size of migrate_pages()

On Wed, 16 Jun 2010 13:36:57 +0900 (JST)
KOSAKI Motohiro <[email protected]> wrote:

>
> Now, migrate_pages() are using >500 bytes stack. This patch reduce it.
>
> mm/mempolicy.c: In function 'sys_migrate_pages':
> mm/mempolicy.c:1344: warning: the frame size of 528 bytes is larger than
> 512 bytes
>
> Signed-off-by: KOSAKI Motohiro <[email protected]>
> Cc: Christoph Lameter <[email protected]>
Reviewed-by: KAMEZAWA Hiroyuki <[email protected]>

2010-06-18 21:39:27

by Andrew Morton

[permalink] [raw]
Subject: Re: [PATCH] mempolicy: reduce stack size of migrate_pages()

On Wed, 16 Jun 2010 13:36:57 +0900 (JST)
KOSAKI Motohiro <[email protected]> wrote:

>
> Now, migrate_pages() are using >500 bytes stack. This patch reduce it.
>
> mm/mempolicy.c: In function 'sys_migrate_pages':
> mm/mempolicy.c:1344: warning: the frame size of 528 bytes is larger than
> 512 bytes
>
> Signed-off-by: KOSAKI Motohiro <[email protected]>
> Cc: Christoph Lameter <[email protected]>
> ---
> mm/mempolicy.c | 35 ++++++++++++++++++++++-------------
> 1 files changed, 22 insertions(+), 13 deletions(-)
>
> diff --git a/mm/mempolicy.c b/mm/mempolicy.c
> index 13b09bd..1116427 100644
> --- a/mm/mempolicy.c
> +++ b/mm/mempolicy.c
> @@ -1275,33 +1275,39 @@ SYSCALL_DEFINE4(migrate_pages, pid_t, pid, unsigned long, maxnode,
> const unsigned long __user *, new_nodes)
> {
> const struct cred *cred = current_cred(), *tcred;
> - struct mm_struct *mm;
> + struct mm_struct *mm = NULL;
> struct task_struct *task;
> - nodemask_t old;
> - nodemask_t new;
> nodemask_t task_nodes;
> int err;
> + NODEMASK_SCRATCH(scratch);
> + nodemask_t *old = &scratch->mask1;
> + nodemask_t *new = &scratch->mask2;
>
> + if (!scratch)
> + return -ENOMEM;

It doesn't matter in practice, but it makes me all queazy to see code
which plays with pointers which might be NULL.

--- a/mm/mempolicy.c~mempolicy-reduce-stack-size-of-migrate_pages-fix
+++ a/mm/mempolicy.c
@@ -1279,13 +1279,16 @@ SYSCALL_DEFINE4(migrate_pages, pid_t, pi
struct task_struct *task;
nodemask_t task_nodes;
int err;
+ nodemask_t *old;
+ nodemask_t *new;
NODEMASK_SCRATCH(scratch);
- nodemask_t *old = &scratch->mask1;
- nodemask_t *new = &scratch->mask2;

if (!scratch)
return -ENOMEM;

+ old = &scratch->mask1;
+ new = &scratch->mask2;
+
err = get_nodes(old, old_nodes, maxnode);
if (err)
goto out;
_

2010-06-21 11:46:20

by KOSAKI Motohiro

[permalink] [raw]
Subject: Re: [PATCH] mempolicy: reduce stack size of migrate_pages()

> > const struct cred *cred = current_cred(), *tcred;
> > - struct mm_struct *mm;
> > + struct mm_struct *mm = NULL;
> > struct task_struct *task;
> > - nodemask_t old;
> > - nodemask_t new;
> > nodemask_t task_nodes;
> > int err;
> > + NODEMASK_SCRATCH(scratch);
> > + nodemask_t *old = &scratch->mask1;
> > + nodemask_t *new = &scratch->mask2;
> >
> > + if (!scratch)
> > + return -ENOMEM;
>
> It doesn't matter in practice, but it makes me all queazy to see code
> which plays with pointers which might be NULL.

I see. thanks.
Do we need to send you updated patch?


>
> --- a/mm/mempolicy.c~mempolicy-reduce-stack-size-of-migrate_pages-fix
> +++ a/mm/mempolicy.c
> @@ -1279,13 +1279,16 @@ SYSCALL_DEFINE4(migrate_pages, pid_t, pi
> struct task_struct *task;
> nodemask_t task_nodes;
> int err;
> + nodemask_t *old;
> + nodemask_t *new;
> NODEMASK_SCRATCH(scratch);
> - nodemask_t *old = &scratch->mask1;
> - nodemask_t *new = &scratch->mask2;
>
> if (!scratch)
> return -ENOMEM;
>
> + old = &scratch->mask1;
> + new = &scratch->mask2;
> +
> err = get_nodes(old, old_nodes, maxnode);
> if (err)
> goto out;
> _
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/