[PATCH] net: sched: Add support for packet bursting.

Subsystems: netem network emulator, networking [general], tc subsystem, the rest

STALE1869d

5 messages, 3 authors, 2021-06-27 · open the first message on its own page

[PATCH] net: sched: Add support for packet bursting.

From: Niclas Hedam <hidden>
Date: 2021-06-25 12:03:14

Hello,

This patch implements packet bursting in the NetEm scheduler.
This allows system administrators to hold back outgoing
packets and release them at a multiple of a time quantum.
This feature can be used to prevent timing attacks caused
by network latency.

I'm currently publishing a paper on this, which is currently not
publicly available, but the idea is based on Predictive Black-Box
Mitigation of Timing Channels
(https://dl.acm.org/doi/pdf/10.1145/1866307.1866341).

Signed-off-by: Niclas Hedam <redacted>
---
 include/uapi/linux/pkt_sched.h |  2 ++
 net/sched/sch_netem.c          | 24 +++++++++++++++++++++---
 2 files changed, 23 insertions(+), 3 deletions(-)
diff --git a/include/uapi/linux/pkt_sched.h b/include/uapi/linux/pkt_sched.h
index 79a699f106b1..826d1dee6601 100644
--- a/include/uapi/linux/pkt_sched.h
+++ b/include/uapi/linux/pkt_sched.h
@@ -594,6 +594,7 @@ enum {
 	TCA_NETEM_DELAY_DIST,
 	TCA_NETEM_REORDER,
 	TCA_NETEM_CORRUPT,
+	TCA_NETEM_BURSTING,
 	TCA_NETEM_LOSS,
 	TCA_NETEM_RATE,
 	TCA_NETEM_ECN,
@@ -615,6 +616,7 @@ struct tc_netem_qopt {
 	__u32	gap;		/* re-ordering gap (0 for none) */
 	__u32   duplicate;	/* random packet dup  (0=none ~0=100%) */
 	__u32	jitter;		/* random jitter in latency (us) */
+	__u32	bursting;	/* send packets in bursts (us) */
 };

 struct tc_netem_corr {
diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c
index 0c345e43a09a..52d796287b86 100644
--- a/net/sched/sch_netem.c
+++ b/net/sched/sch_netem.c
@@ -85,6 +85,7 @@ struct netem_sched_data {
 	s64 latency;
 	s64 jitter;

+	u32 bursting;
 	u32 loss;
 	u32 ecn;
 	u32 limit;
@@ -467,7 +468,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch,
 	/* If a delay is expected, orphan the skb. (orphaning usually takes
 	 * place at TX completion time, so _before_ the link transit delay)
 	 */
-	if (q->latency || q->jitter || q->rate)
+	if (q->latency || q->jitter || q->rate || q->bursting)
 		skb_orphan_partial(skb);

 	/*
@@ -527,8 +528,17 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch,
 	qdisc_qstats_backlog_inc(sch, skb);

 	cb = netem_skb_cb(skb);
-	if (q->gap == 0 ||		/* not doing reordering */
-	    q->counter < q->gap - 1 ||	/* inside last reordering gap */
+	if (q->bursting > 0) {
+		u64 now;
+
+		now = ktime_get_ns();
+
+		cb->time_to_send = now - (now % q->bursting) + q->bursting;
+
+		++q->counter;
+		tfifo_enqueue(skb, sch);
+	} else if (q->gap == 0 ||		/* not doing reordering */
+	    q->counter < q->gap - 1 ||		/* inside last reordering gap */
 	    q->reorder < get_crandom(&q->reorder_cor)) {
 		u64 now;
 		s64 delay;
@@ -927,6 +937,7 @@ static const struct nla_policy netem_policy[TCA_NETEM_MAX + 1] = {
 	[TCA_NETEM_ECN]		= { .type = NLA_U32 },
 	[TCA_NETEM_RATE64]	= { .type = NLA_U64 },
 	[TCA_NETEM_LATENCY64]	= { .type = NLA_S64 },
+	[TCA_NETEM_BURSTING]	= { .type = NLA_U64 },
 	[TCA_NETEM_JITTER64]	= { .type = NLA_S64 },
 	[TCA_NETEM_SLOT]	= { .len = sizeof(struct tc_netem_slot) },
 };
@@ -1001,6 +1012,7 @@ static int netem_change(struct Qdisc *sch, struct nlattr *opt,

 	q->latency = PSCHED_TICKS2NS(qopt->latency);
 	q->jitter = PSCHED_TICKS2NS(qopt->jitter);
+	q->bursting = PSCHED_TICKS2NS(qopt->bursting);
 	q->limit = qopt->limit;
 	q->gap = qopt->gap;
 	q->counter = 0;
@@ -1032,6 +1044,9 @@ static int netem_change(struct Qdisc *sch, struct nlattr *opt,
 	if (tb[TCA_NETEM_LATENCY64])
 		q->latency = nla_get_s64(tb[TCA_NETEM_LATENCY64]);

+	if (tb[TCA_NETEM_BURSTING])
+		q->bursting = nla_get_u64(tb[TCA_NETEM_BURSTING]);
+
 	if (tb[TCA_NETEM_JITTER64])
 		q->jitter = nla_get_s64(tb[TCA_NETEM_JITTER64]);
@@ -1150,6 +1165,9 @@ static int netem_dump(struct Qdisc *sch, struct sk_buff *skb)
 			     UINT_MAX);
 	qopt.jitter = min_t(psched_tdiff_t, PSCHED_NS2TICKS(q->jitter),
 			    UINT_MAX);
+	qopt.bursting = min_t(psched_tdiff_t, PSCHED_NS2TICKS(q->bursting),
+			    UINT_MAX);
+
 	qopt.limit = q->limit;
 	qopt.loss = q->loss;
 	qopt.gap = q->gap;
--
2.25.1

Re: [PATCH] net: sched: Add support for packet bursting.

From: kernel test robot <hidden>
Date: 2021-06-25 20:54:23

Hi Niclas,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on net/master]
[also build test ERROR on ipvs/master net-next/master linus/master v5.13-rc7 next-20210625]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Niclas-Hedam/net-sched-Add-support-for-packet-bursting/20210625-200450
base:   https://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git be7f62eebaff2f86c1467a2d33930a0a7a87675b
config: m68k-allyesconfig (attached as .config)
compiler: m68k-linux-gcc (GCC) 9.3.0
reproduce (this is a W=1 build):
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # https://github.com/0day-ci/linux/commit/dca330eed26fd2835927462dcca58379dd5599a5
        git remote add linux-review https://github.com/0day-ci/linux
        git fetch --no-tags linux-review Niclas-Hedam/net-sched-Add-support-for-packet-bursting/20210625-200450
        git checkout dca330eed26fd2835927462dcca58379dd5599a5
        # save the attached .config to linux build tree
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=m68k 

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <redacted>

All errors (new ones prefixed by >>):

   m68k-linux-ld: net/sched/sch_netem.o: in function `netem_enqueue':
quoted
sch_netem.c:(.text+0x10aa): undefined reference to `__umoddi3'
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org

Re: [PATCH] net: sched: Add support for packet bursting.

From: kernel test robot <hidden>
Date: 2021-06-25 22:35:20

Hi Niclas,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on net/master]
[also build test ERROR on ipvs/master net-next/master linus/master v5.13-rc7 next-20210625]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Niclas-Hedam/net-sched-Add-support-for-packet-bursting/20210625-200450
base:   https://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git be7f62eebaff2f86c1467a2d33930a0a7a87675b
config: sh-allmodconfig (attached as .config)
compiler: sh4-linux-gcc (GCC) 9.3.0
reproduce (this is a W=1 build):
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # https://github.com/0day-ci/linux/commit/dca330eed26fd2835927462dcca58379dd5599a5
        git remote add linux-review https://github.com/0day-ci/linux
        git fetch --no-tags linux-review Niclas-Hedam/net-sched-Add-support-for-packet-bursting/20210625-200450
        git checkout dca330eed26fd2835927462dcca58379dd5599a5
        # save the attached .config to linux build tree
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=sh 

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <redacted>

All errors (new ones prefixed by >>, old ones prefixed by <<):
quoted
ERROR: modpost: "__umoddi3" [net/sched/sch_netem.ko] undefined!
ERROR: modpost: "__delay" [drivers/net/mdio/mdio-cavium.ko] undefined!
ERROR: modpost: "__udivdi3" [fs/btrfs/btrfs.ko] undefined!
ERROR: modpost: "__umoddi3" [fs/btrfs/btrfs.ko] undefined!

Kconfig warnings: (for reference only)
   WARNING: unmet direct dependencies detected for SND_ATMEL_SOC_PDC
   Depends on SOUND && !UML && SND && SND_SOC && SND_ATMEL_SOC && HAS_DMA
   Selected by
   - SND_ATMEL_SOC_SSC && SOUND && !UML && SND && SND_SOC && SND_ATMEL_SOC
   - SND_ATMEL_SOC_SSC_PDC && SOUND && !UML && SND && SND_SOC && SND_ATMEL_SOC && ATMEL_SSC

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org

Re: [PATCH] net: sched: Add support for packet bursting.

From: Cong Wang <hidden>
Date: 2021-06-27 18:33:01

On Fri, Jun 25, 2021 at 5:03 AM Niclas Hedam [off-list ref] wrote:
quoted hunk
diff --git a/include/uapi/linux/pkt_sched.h b/include/uapi/linux/pkt_sched.h
index 79a699f106b1..826d1dee6601 100644
--- a/include/uapi/linux/pkt_sched.h
+++ b/include/uapi/linux/pkt_sched.h
@@ -594,6 +594,7 @@ enum {
        TCA_NETEM_DELAY_DIST,
        TCA_NETEM_REORDER,
        TCA_NETEM_CORRUPT,
+       TCA_NETEM_BURSTING,
        TCA_NETEM_LOSS,
        TCA_NETEM_RATE,
        TCA_NETEM_ECN,
You can't add a new enum in the middle, as it is UAPI.

Thanks.

Re: [PATCH] net: sched: Add support for packet bursting.

From: Niclas Hedam <hidden>
Date: 2021-06-27 19:12:18

Hi Cong,

Good point.
Here is a new patch.

From 71843907bdb9cdc4e24358f0c16a8778f2762dc7 Mon Sep 17 00:00:00 2001
From: Niclas Hedam <redacted>
Date: Fri, 25 Jun 2021 13:37:18 +0200
Subject: [PATCH] net: sched: Add support for packet bursting.

This commit implements packet bursting in the NetEm scheduler.
This allows system administrators to hold back outgoing
packets and release them at a multiple of a time quantum.
This feature can be used to prevent timing attacks caused
by network latency.

Signed-off-by: Niclas Hedam <redacted>
---
 include/uapi/linux/pkt_sched.h |  2 ++
 net/sched/sch_netem.c          | 24 +++++++++++++++++++++---
 2 files changed, 23 insertions(+), 3 deletions(-)
diff --git a/include/uapi/linux/pkt_sched.h b/include/uapi/linux/pkt_sched.h
index 79a699f106b1..1ba49f141dae 100644
--- a/include/uapi/linux/pkt_sched.h
+++ b/include/uapi/linux/pkt_sched.h
@@ -603,6 +603,7 @@ enum {
 	TCA_NETEM_JITTER64,
 	TCA_NETEM_SLOT,
 	TCA_NETEM_SLOT_DIST,
+        TCA_NETEM_BURSTING,
 	__TCA_NETEM_MAX,
 };
@@ -615,6 +616,7 @@ struct tc_netem_qopt {
 	__u32	gap;		/* re-ordering gap (0 for none) */
 	__u32   duplicate;	/* random packet dup  (0=none ~0=100%) */
 	__u32	jitter;		/* random jitter in latency (us) */
+	__u32	bursting;	/* send packets in bursts (us) */
 };

 struct tc_netem_corr {
diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c
index 0c345e43a09a..52d796287b86 100644
--- a/net/sched/sch_netem.c
+++ b/net/sched/sch_netem.c
@@ -85,6 +85,7 @@ struct netem_sched_data {
 	s64 latency;
 	s64 jitter;

+	u32 bursting;
 	u32 loss;
 	u32 ecn;
 	u32 limit;
@@ -467,7 +468,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch,
 	/* If a delay is expected, orphan the skb. (orphaning usually takes
 	 * place at TX completion time, so _before_ the link transit delay)
 	 */
-	if (q->latency || q->jitter || q->rate)
+	if (q->latency || q->jitter || q->rate || q->bursting)
 		skb_orphan_partial(skb);

 	/*
@@ -527,8 +528,17 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch,
 	qdisc_qstats_backlog_inc(sch, skb);

 	cb = netem_skb_cb(skb);
-	if (q->gap == 0 ||		/* not doing reordering */
-	    q->counter < q->gap - 1 ||	/* inside last reordering gap */
+	if (q->bursting > 0) {
+		u64 now;
+
+		now = ktime_get_ns();
+
+		cb->time_to_send = now - (now % q->bursting) + q->bursting;
+
+		++q->counter;
+		tfifo_enqueue(skb, sch);
+	} else if (q->gap == 0 ||		/* not doing reordering */
+	    q->counter < q->gap - 1 ||		/* inside last reordering gap */
 	    q->reorder < get_crandom(&q->reorder_cor)) {
 		u64 now;
 		s64 delay;
@@ -927,6 +937,7 @@ static const struct nla_policy netem_policy[TCA_NETEM_MAX + 1] = {
 	[TCA_NETEM_ECN]		= { .type = NLA_U32 },
 	[TCA_NETEM_RATE64]	= { .type = NLA_U64 },
 	[TCA_NETEM_LATENCY64]	= { .type = NLA_S64 },
+	[TCA_NETEM_BURSTING]	= { .type = NLA_U64 },
 	[TCA_NETEM_JITTER64]	= { .type = NLA_S64 },
 	[TCA_NETEM_SLOT]	= { .len = sizeof(struct tc_netem_slot) },
 };
@@ -1001,6 +1012,7 @@ static int netem_change(struct Qdisc *sch, struct nlattr *opt,

 	q->latency = PSCHED_TICKS2NS(qopt->latency);
 	q->jitter = PSCHED_TICKS2NS(qopt->jitter);
+	q->bursting = PSCHED_TICKS2NS(qopt->bursting);
 	q->limit = qopt->limit;
 	q->gap = qopt->gap;
 	q->counter = 0;
@@ -1032,6 +1044,9 @@ static int netem_change(struct Qdisc *sch, struct nlattr *opt,
 	if (tb[TCA_NETEM_LATENCY64])
 		q->latency = nla_get_s64(tb[TCA_NETEM_LATENCY64]);

+	if (tb[TCA_NETEM_BURSTING])
+		q->bursting = nla_get_u64(tb[TCA_NETEM_BURSTING]);
+
 	if (tb[TCA_NETEM_JITTER64])
 		q->jitter = nla_get_s64(tb[TCA_NETEM_JITTER64]);
@@ -1150,6 +1165,9 @@ static int netem_dump(struct Qdisc *sch, struct sk_buff *skb)
 			     UINT_MAX);
 	qopt.jitter = min_t(psched_tdiff_t, PSCHED_NS2TICKS(q->jitter),
 			    UINT_MAX);
+	qopt.bursting = min_t(psched_tdiff_t, PSCHED_NS2TICKS(q->bursting),
+			    UINT_MAX);
+
 	qopt.limit = q->limit;
 	qopt.loss = q->loss;
 	qopt.gap = q->gap;
--
2.25.1
On 27 Jun 2021, at 20:32, Cong Wang [off-list ref] wrote:

On Fri, Jun 25, 2021 at 5:03 AM Niclas Hedam [off-list ref] wrote:
quoted
diff --git a/include/uapi/linux/pkt_sched.h b/include/uapi/linux/pkt_sched.h
index 79a699f106b1..826d1dee6601 100644
--- a/include/uapi/linux/pkt_sched.h
+++ b/include/uapi/linux/pkt_sched.h
@@ -594,6 +594,7 @@ enum {
       TCA_NETEM_DELAY_DIST,
       TCA_NETEM_REORDER,
       TCA_NETEM_CORRUPT,
+       TCA_NETEM_BURSTING,
       TCA_NETEM_LOSS,
       TCA_NETEM_RATE,
       TCA_NETEM_ECN,
You can't add a new enum in the middle, as it is UAPI.

Thanks.

Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help