diff options
| author | Yunfeng Ye <yeyunfeng@huawei.com> | 2019-10-14 10:15:57 +0200 |
|---|---|---|
| committer | Peter Zijlstra <peterz@infradead.org> | 2019-10-17 21:31:55 +0200 |
| commit | d7e78706e43107fa269fe34b1a69e653f5ec9f2c (patch) | |
| tree | 123ce021f2ad3c6a16bc359bd5f98296d94be3db /kernel/events | |
| parent | perf/ring_buffer: Modify the parameter type of perf_mmap_free_page() (diff) | |
| download | linux-d7e78706e43107fa269fe34b1a69e653f5ec9f2c.tar.xz linux-d7e78706e43107fa269fe34b1a69e653f5ec9f2c.zip | |
perf/ring_buffer: Matching the memory allocate and free, in rb_alloc()
Currently perf_mmap_alloc_page() is used to allocate memory in
rb_alloc(), but using free_page() to free memory in the failure path.
It's better to use perf_mmap_free_page() instead.
Signed-off-by: Yunfeng Ye <yeyunfeng@huawei.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: <jolsa@redhat.co>
Cc: <acme@kernel.org>
Cc: <mingo@redhat.com>
Cc: <mark.rutland@arm.com>
Cc: <namhyung@kernel.org>
Cc: <alexander.shishkin@linux.intel.com>
Link: https://lkml.kernel.org/r/575c7e8c-90c7-4e3a-b41d-f894d8cdbd7f@huawei.com
Diffstat (limited to 'kernel/events')
| -rw-r--r-- | kernel/events/ring_buffer.c | 20 |
1 files changed, 10 insertions, 10 deletions
diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c index abc145cbfedf..246c83ac5643 100644 --- a/kernel/events/ring_buffer.c +++ b/kernel/events/ring_buffer.c @@ -754,6 +754,14 @@ static void *perf_mmap_alloc_page(int cpu) return page_address(page); } +static void perf_mmap_free_page(void *addr) +{ + struct page *page = virt_to_page(addr); + + page->mapping = NULL; + __free_page(page); +} + struct ring_buffer *rb_alloc(int nr_pages, long watermark, int cpu, int flags) { struct ring_buffer *rb; @@ -788,9 +796,9 @@ struct ring_buffer *rb_alloc(int nr_pages, long watermark, int cpu, int flags) fail_data_pages: for (i--; i >= 0; i--) - free_page((unsigned long)rb->data_pages[i]); + perf_mmap_free_page(rb->data_pages[i]); - free_page((unsigned long)rb->user_page); + perf_mmap_free_page(rb->user_page); fail_user_page: kfree(rb); @@ -799,14 +807,6 @@ fail: return NULL; } -static void perf_mmap_free_page(void *addr) -{ - struct page *page = virt_to_page(addr); - - page->mapping = NULL; - __free_page(page); -} - void rb_free(struct ring_buffer *rb) { int i; |