提交 c3e0ce9f 编写于 作者: A Alexander Shishkin 提交者: Xie XiuQi

perf: Copy parent's address filter offsets on clone

mainline inclusion
from mainline-5.1-rc1
commit 18736eef12137c59f60cc9f56dc5bea05c92e0eb
category: bugfix
bugzilla: 11595
CVE: NA

-------------------------------------------------
When a child event is allocated in the inherit_event() path, the VMA
based filter offsets are not copied from the parent, even though the
address space mapping of the new task remains the same, which leads to
no trace for the new task until exec.
Reported-by: NMansour Alharthi <malharthi9@gatech.edu>
Signed-off-by: NAlexander Shishkin <alexander.shishkin@linux.intel.com>
Tested-by: NMathieu Poirier <mathieu.poirier@linaro.org>
Acked-by: NPeter Zijlstra <peterz@infradead.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Fixes: 375637bc ("perf/core: Introduce address range filtering")
Link: http://lkml.kernel.org/r/20190215115655.63469-2-alexander.shishkin@linux.intel.comSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
(cherry picked from commit 18736eef12137c59f60cc9f56dc5bea05c92e0eb)
Signed-off-by: NZhen Lei <thunder.leizhen@huawei.com>
Reviewed-by: NCheng Jian <cj.chengjian@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
上级 27b25782
...@@ -1254,6 +1254,7 @@ static void put_ctx(struct perf_event_context *ctx) ...@@ -1254,6 +1254,7 @@ static void put_ctx(struct perf_event_context *ctx)
* perf_event_context::lock * perf_event_context::lock
* perf_event::mmap_mutex * perf_event::mmap_mutex
* mmap_sem * mmap_sem
* perf_addr_filters_head::lock
* *
* cpu_hotplug_lock * cpu_hotplug_lock
* pmus_lock * pmus_lock
...@@ -10099,6 +10100,20 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, ...@@ -10099,6 +10100,20 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu,
goto err_per_task; goto err_per_task;
} }
/*
* Clone the parent's vma offsets: they are valid until exec()
* even if the mm is not shared with the parent.
*/
if (event->parent) {
struct perf_addr_filters_head *ifh = perf_event_addr_filters(event);
raw_spin_lock_irq(&ifh->lock);
memcpy(event->addr_filters_offs,
event->parent->addr_filters_offs,
pmu->nr_addr_filters * sizeof(unsigned long));
raw_spin_unlock_irq(&ifh->lock);
}
/* force hw sync on the address filters */ /* force hw sync on the address filters */
event->addr_filters_gen = 1; event->addr_filters_gen = 1;
} }
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册