[PATCH 2/2] perf: validate cpu early in perf_event_alloc()

From: Oleg Nesterov
Date: Tue Jan 18 2011 - 11:18:20 EST


Starting from perf_event_alloc()->perf_init_event(), the kernel
assumes that event->cpu is either -1 or the valid CPU number.

Change perf_event_alloc() to validate this argument early. This
also means we can remove the similar check in find_get_context().

Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx>
---

kernel/perf_event.c | 10 ++++++----
1 file changed, 6 insertions(+), 4 deletions(-)

--- git/kernel/perf_event.c~2_perf_event_alloc 2011-01-18 16:56:40.000000000 +0100
+++ git/kernel/perf_event.c 2011-01-18 16:57:08.000000000 +0100
@@ -2233,9 +2233,6 @@ find_get_context(struct pmu *pmu, struct
if (perf_paranoid_cpu() && !capable(CAP_SYS_ADMIN))
return ERR_PTR(-EACCES);

- if (cpu < 0 || cpu >= nr_cpumask_bits)
- return ERR_PTR(-EINVAL);
-
/*
* We could be clever and allow to attach a event to an
* offline CPU and activate it when the CPU comes up, but
@@ -5541,6 +5538,11 @@ perf_event_alloc(struct perf_event_attr
struct hw_perf_event *hwc;
long err;

+ if ((unsigned)cpu >= nr_cpu_ids) {
+ if (!task || cpu != -1)
+ return ERR_PTR(-EINVAL);
+ }
+
event = kzalloc(sizeof(*event), GFP_KERNEL);
if (!event)
return ERR_PTR(-ENOMEM);
@@ -5589,7 +5591,7 @@ perf_event_alloc(struct perf_event_attr

if (!overflow_handler && parent_event)
overflow_handler = parent_event->overflow_handler;
-
+
event->overflow_handler = overflow_handler;

if (attr->disabled)

--
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/