summaryrefslogtreecommitdiff
path: root/include/linux/bpf.h
diff options
context:
space:
mode:
authorAlexei Starovoitov <ast@fb.com>2016-02-02 09:39:54 +0300
committerDavid S. Miller <davem@davemloft.net>2016-02-06 11:34:36 +0300
commita10423b87a7eae75da79ce80a8d9475047a674ee (patch)
tree2a8296c94fc90056e7a8d2b2dd3b31d4d3d73d5b /include/linux/bpf.h
parent824bd0ce6c7c43a9e1e210abf124958e54d88342 (diff)
downloadlinux-a10423b87a7eae75da79ce80a8d9475047a674ee.tar.xz
bpf: introduce BPF_MAP_TYPE_PERCPU_ARRAY map
Primary use case is a histogram array of latency where bpf program computes the latency of block requests or other events and stores histogram of latency into array of 64 elements. All cpus are constantly running, so normal increment is not accurate, bpf_xadd causes cache ping-pong and this per-cpu approach allows fastest collision-free counters. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'include/linux/bpf.h')
-rw-r--r--include/linux/bpf.h1
1 files changed, 1 insertions, 0 deletions
diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 83d1926c61e4..141fb0d45731 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -151,6 +151,7 @@ struct bpf_array {
union {
char value[0] __aligned(8);
void *ptrs[0] __aligned(8);
+ void __percpu *pptrs[0] __aligned(8);
};
};
#define MAX_TAIL_CALL_CNT 32