While creating 1000 containers, perf is showing lot of time spent in
snmp_fold_field on a large cpu system.
The current patch tries to improve by reordering the statistics gathering.
Please note that similar overhead was also reported while creating
veth pairs https://lkml.org/lkml/2013/3/19/556
Changes in V3:
- use memset to initialize temp buffer in leaf function. (David)
- use memcpy to copy the buffer data to stat instead of unalign_pu (Joe)
- Move buffer definition to leaf function __snmp6_fill_stats64() (Eric)
-
Changes in V2:
- Allocate the stat calculation buffer in stack. (Eric)
Setup:
160 cpu (20 core) baremetal powerpc system with 1TB memory
1000 docker containers was created with command
docker run -itd ubuntu:15.04 /bin/bash in loop
observation:
Docker container creation linearly increased from around 1.6 sec to 7.5 sec
(at 1000 containers) perf data showed, creating veth interfaces resulting in
the below code path was taking more time.
rtnl_fill_ifinfo
-> inet6_fill_link_af
-> inet6_fill_ifla6_attrs
-> snmp_fold_field
proposed idea:
currently __snmp6_fill_stats64 calls snmp_fold_field that walks
through per cpu data to of an item (iteratively for around 36 items).
The patch tries to aggregate the statistics by going through
all the items of each cpu sequentially which is reducing cache
misses.
Performance of docker creation improved by around more than 2x
after the patch.
before the patch:
================
#time docker run -itd ubuntu:15.04 /bin/bash
3f45ba571a42e925c4ec4aaee0e48d7610a9ed82a4c931f83324d41822cf6617
real 0m6.836s
user 0m0.095s
sys 0m0.011s
perf record -a docker run -itd ubuntu:15.04 /bin/bash
=======================================================
# Samples: 32K of event 'cycles'
# Event count (approx.): 24688700190
# Overhead Command Shared Object Symbol
# ........ ............... ...................... ........................
50.73% docker [kernel.kallsyms] [k] snmp_fold_field
9.07% swapper [kernel.kallsyms] [k] snooze_loop
3.49% docker [kernel.kallsyms] [k] veth_stats_one
2.85% swapper [kernel.kallsyms] [k] _raw_spin_lock
1.37% docker docker [.] backtrace_qsort
1.31% docker docker [.] strings.FieldsFunc
cache-misses: 2.7%
after the patch:
=============
#time docker run -itd ubuntu:15.04 /bin/bash
9178273e9df399c8290b6c196e4aef9273be2876225f63b14a60cf97eacfafb5
real 0m3.249s
user 0m0.088s
sys 0m0.020s
perf record -a docker run -itd ubuntu:15.04 /bin/bash
=======================================================
# Samples: 18K of event 'cycles'
# Event count (approx.): 13958958797
# Overhead Command Shared Object Symbol
# ........ ............... .................... ..........................
10.57% docker docker [.] scanblock
8.37% swapper [kernel.kallsyms] [k] snooze_loop
6.91% docker [kernel.kallsyms] [k] snmp_get_cpu_field
6.67% docker [kernel.kallsyms] [k] veth_stats_one
3.96% docker docker [.] runtime_MSpan_Sweep
2.47% docker docker [.] strings.FieldsFunc
cache-misses: 1.41 %
Please let me know if you have suggestions/comments.
Thanks Eric, Joe and David for comments on V1 and V2.
Raghavendra K T (2):
net: Introduce helper functions to get the per cpu data
net: Optimize snmp stat aggregation by walking all the percpu data at
once
include/net/ip.h | 10 ++++++++++
net/ipv4/af_inet.c | 41 +++++++++++++++++++++++++++--------------
net/ipv6/addrconf.c | 22 +++++++++++++---------
3 files changed, 50 insertions(+), 23 deletions(-)
--
1.7.11.7
Signed-off-by: Raghavendra K T <[email protected]>
---
include/net/ip.h | 10 ++++++++++
net/ipv4/af_inet.c | 41 +++++++++++++++++++++++++++--------------
2 files changed, 37 insertions(+), 14 deletions(-)
diff --git a/include/net/ip.h b/include/net/ip.h
index d5fe9f2..93bf12e 100644
--- a/include/net/ip.h
+++ b/include/net/ip.h
@@ -202,10 +202,20 @@ void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
#define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
#define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
+u64 snmp_get_cpu_field(void __percpu *mib, int cpu, int offct);
unsigned long snmp_fold_field(void __percpu *mib, int offt);
#if BITS_PER_LONG==32
+u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
+ size_t syncp_offset);
u64 snmp_fold_field64(void __percpu *mib, int offt, size_t sync_off);
#else
+static inline u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
+ size_t syncp_offset)
+{
+ return snmp_get_cpu_field(mib, cpu, offct);
+
+}
+
static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_off)
{
return snmp_fold_field(mib, offt);
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 9532ee8..302e36b 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -1448,38 +1448,51 @@ int inet_ctl_sock_create(struct sock **sk, unsigned short family,
}
EXPORT_SYMBOL_GPL(inet_ctl_sock_create);
+u64 snmp_get_cpu_field(void __percpu *mib, int cpu, int offt)
+{
+ return *(((unsigned long *)per_cpu_ptr(mib, cpu)) + offt);
+}
+EXPORT_SYMBOL_GPL(snmp_get_cpu_field);
+
unsigned long snmp_fold_field(void __percpu *mib, int offt)
{
unsigned long res = 0;
int i;
for_each_possible_cpu(i)
- res += *(((unsigned long *) per_cpu_ptr(mib, i)) + offt);
+ res += snmp_get_cpu_field(mib, i, offt);
return res;
}
EXPORT_SYMBOL_GPL(snmp_fold_field);
#if BITS_PER_LONG==32
+u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
+ size_t syncp_offset)
+{
+ void *bhptr;
+ struct u64_stats_sync *syncp;
+ u64 v;
+ unsigned int start;
+
+ bhptr = per_cpu_ptr(mib, cpu);
+ syncp = (struct u64_stats_sync *)(bhptr + syncp_offset);
+ do {
+ start = u64_stats_fetch_begin_irq(syncp);
+ v = *(((u64 *)bhptr) + offt);
+ } while (u64_stats_fetch_retry_irq(syncp, start));
+
+ return v;
+}
+EXPORT_SYMBOL_GPL(snmp_get_cpu_field64);
+
u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_offset)
{
u64 res = 0;
int cpu;
for_each_possible_cpu(cpu) {
- void *bhptr;
- struct u64_stats_sync *syncp;
- u64 v;
- unsigned int start;
-
- bhptr = per_cpu_ptr(mib, cpu);
- syncp = (struct u64_stats_sync *)(bhptr + syncp_offset);
- do {
- start = u64_stats_fetch_begin_irq(syncp);
- v = *(((u64 *) bhptr) + offt);
- } while (u64_stats_fetch_retry_irq(syncp, start));
-
- res += v;
+ res += snmp_get_cpu_field(mib, cpu, offct, syncp_offset);
}
return res;
}
--
1.7.11.7
Docker container creation linearly increased from around 1.6 sec to 7.5 sec
(at 1000 containers) and perf data showed 50% ovehead in snmp_fold_field.
reason: currently __snmp6_fill_stats64 calls snmp_fold_field that walks
through per cpu data of an item (iteratively for around 36 items).
idea: This patch tries to aggregate the statistics by going through
all the items of each cpu sequentially which is reducing cache
misses.
Docker creation got faster by more than 2x after the patch.
Result:
Before After
Docker creation time 6.836s 3.25s
cache miss 2.7% 1.41%
perf before:
50.73% docker [kernel.kallsyms] [k] snmp_fold_field
9.07% swapper [kernel.kallsyms] [k] snooze_loop
3.49% docker [kernel.kallsyms] [k] veth_stats_one
2.85% swapper [kernel.kallsyms] [k] _raw_spin_lock
perf after:
10.57% docker docker [.] scanblock
8.37% swapper [kernel.kallsyms] [k] snooze_loop
6.91% docker [kernel.kallsyms] [k] snmp_get_cpu_field
6.67% docker [kernel.kallsyms] [k] veth_stats_one
changes/ideas suggested:
Using buffer in stack (Eric), Usage of memset (David), Using memcpy in
place of unaligned_put (Joe).
Signed-off-by: Raghavendra K T <[email protected]>
---
net/ipv6/addrconf.c | 22 +++++++++++++---------
1 file changed, 13 insertions(+), 9 deletions(-)
Changes in V3:
- use memset to initialize temp buffer in leaf function. (David)
- use memcpy to copy the buffer data to stat instead of unalign_pu (Joe)
- Move buffer definition to leaf function __snmp6_fill_stats64() (Eric)
-
Changes in V2:
- Allocate the stat calculation buffer in stack. (Eric)
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index 21c2c81..379619a 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -4624,18 +4624,22 @@ static inline void __snmp6_fill_statsdev(u64 *stats, atomic_long_t *mib,
}
static inline void __snmp6_fill_stats64(u64 *stats, void __percpu *mib,
- int items, int bytes, size_t syncpoff)
+ int items, int bytes, size_t syncpoff)
{
- int i;
+ int i, c;
int pad = bytes - sizeof(u64) * items;
+ u64 buff[items];
+
BUG_ON(pad < 0);
- /* Use put_unaligned() because stats may not be aligned for u64. */
- put_unaligned(items, &stats[0]);
- for (i = 1; i < items; i++)
- put_unaligned(snmp_fold_field64(mib, i, syncpoff), &stats[i]);
+ memset(buff, 0, sizeof(buff));
+ buff[0] = items;
- memset(&stats[items], 0, pad);
+ for_each_possible_cpu(c)
+ for (i = 1; i < items; i++)
+ buff[i] += snmp_get_cpu_field64(mib, c, i, syncpoff);
+
+ memcpy(stats, buff, items * sizeof(u64));
}
static void snmp6_fill_stats(u64 *stats, struct inet6_dev *idev, int attrtype,
@@ -4643,8 +4647,8 @@ static void snmp6_fill_stats(u64 *stats, struct inet6_dev *idev, int attrtype,
{
switch (attrtype) {
case IFLA_INET6_STATS:
- __snmp6_fill_stats64(stats, idev->stats.ipv6,
- IPSTATS_MIB_MAX, bytes, offsetof(struct ipstats_mib, syncp));
+ __snmp6_fill_stats64(stats, idev->stats.ipv6, IPSTATS_MIB_MAX,
+ bytes, offsetof(struct ipstats_mib, syncp));
break;
case IFLA_INET6_ICMP6STATS:
__snmp6_fill_statsdev(stats, idev->stats.icmpv6dev->mibs, ICMP6_MIB_MAX, bytes);
--
1.7.11.7
On Sat, 2015-08-29 at 14:37 +0530, Raghavendra K T wrote:
>
> static inline void __snmp6_fill_stats64(u64 *stats, void __percpu *mib,
> - int items, int bytes, size_t syncpoff)
> + int items, int bytes, size_t syncpoff)
> {
> - int i;
> + int i, c;
> int pad = bytes - sizeof(u64) * items;
> + u64 buff[items];
> +
One last comment : using variable length arrays is confusing for the
reader, and sparse as well.
$ make C=2 net/ipv6/addrconf.o
...
CHECK net/ipv6/addrconf.c
net/ipv6/addrconf.c:4733:18: warning: Variable length array is used.
net/ipv6/addrconf.c:4737:25: error: cannot size expression
I suggest you remove 'items' parameter and replace it by
IPSTATS_MIB_MAX, as __snmp6_fill_stats64() is called exactly once.
On Sat, 2015-08-29 at 07:32 -0700, Eric Dumazet wrote:
> On Sat, 2015-08-29 at 14:37 +0530, Raghavendra K T wrote:
>
> >
> > static inline void __snmp6_fill_stats64(u64 *stats, void __percpu *mib,
> > - int items, int bytes, size_t syncpoff)
> > + int items, int bytes, size_t syncpoff)
> > {
> > - int i;
> > + int i, c;
> > int pad = bytes - sizeof(u64) * items;
> > + u64 buff[items];
> > +
>
> One last comment : using variable length arrays is confusing for the
> reader, and sparse as well.
>
> $ make C=2 net/ipv6/addrconf.o
> ...
> CHECK net/ipv6/addrconf.c
> net/ipv6/addrconf.c:4733:18: warning: Variable length array is used.
> net/ipv6/addrconf.c:4737:25: error: cannot size expression
>
>
> I suggest you remove 'items' parameter and replace it by
> IPSTATS_MIB_MAX, as __snmp6_fill_stats64() is called exactly once.
If you respin, I suggest:
o remove "items" from the __snmp6_fill_stats64 arguments
and use IPSTATS_MIB_MAX in the function instead
o add braces around the for_each_possible_cpu loop
for_each_possible_cpu(c) {
for (i = 1; i < items; i++)
buff[i] += snmp_get_cpu_field64(mib, c, i, syncpoff);
}
On 08/29/2015 08:51 PM, Joe Perches wrote:
> On Sat, 2015-08-29 at 07:32 -0700, Eric Dumazet wrote:
>> On Sat, 2015-08-29 at 14:37 +0530, Raghavendra K T wrote:
>>
>>>
>>> static inline void __snmp6_fill_stats64(u64 *stats, void __percpu *mib,
>>> - int items, int bytes, size_t syncpoff)
>>> + int items, int bytes, size_t syncpoff)
>>> {
>>> - int i;
>>> + int i, c;
>>> int pad = bytes - sizeof(u64) * items;
>>> + u64 buff[items];
>>> +
>>
>> One last comment : using variable length arrays is confusing for the
>> reader, and sparse as well.
>>
>> $ make C=2 net/ipv6/addrconf.o
>> ...
>> CHECK net/ipv6/addrconf.c
>> net/ipv6/addrconf.c:4733:18: warning: Variable length array is used.
>> net/ipv6/addrconf.c:4737:25: error: cannot size expression
>>
>>
>> I suggest you remove 'items' parameter and replace it by
>> IPSTATS_MIB_MAX, as __snmp6_fill_stats64() is called exactly once.
>
> If you respin, I suggest:
>
> o remove "items" from the __snmp6_fill_stats64 arguments
> and use IPSTATS_MIB_MAX in the function instead
Yes, as also suggested by Eric.
> o add braces around the for_each_possible_cpu loop
>
> for_each_possible_cpu(c) {
> for (i = 1; i < items; i++)
> buff[i] += snmp_get_cpu_field64(mib, c, i, syncpoff);
> }
>
Sure. It makes it more readable.
will respin V4 with these changes (+ memset 0 for pad which I realized).