patches for DPDK stable branches
 help / color / mirror / Atom feed
From: Kevin Traynor <ktraynor@redhat.com>
To: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Cc: Phil Yang <phil.yang@arm.com>,
	Harry van Haaren <harry.van.haaren@intel.com>,
	dpdk stable <stable@dpdk.org>
Subject: [dpdk-stable] patch 'service: fix race condition for MT unsafe service' has been queued to LTS release 18.11.9
Date: Fri,  5 Jun 2020 19:24:16 +0100
Message-ID: <20200605182525.22483-19-ktraynor@redhat.com> (raw)
In-Reply-To: <20200605182525.22483-1-ktraynor@redhat.com>

Hi,

FYI, your patch has been queued to LTS release 18.11.9

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 06/10/20. So please
shout if anyone has objections.

Also note that after the patch there's a diff of the upstream commit vs the
patch applied to the branch. This will indicate if there was any rebasing
needed to apply to the stable branch. If there were code changes for rebasing
(ie: not only metadata diffs), please double check that the rebase was
correctly done.

Queued patches are on a temporary branch at:
https://github.com/kevintraynor/dpdk-stable-queue

This queued commit can be viewed at:
https://github.com/kevintraynor/dpdk-stable-queue/commit/817639c78342003d5d3d675f6bc0eca05ee2c6be

Thanks.

Kevin.

---
From 817639c78342003d5d3d675f6bc0eca05ee2c6be Mon Sep 17 00:00:00 2001
From: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Date: Wed, 6 May 2020 23:27:59 +0800
Subject: [PATCH] service: fix race condition for MT unsafe service

[ upstream commit 18cae99cb9096abe0f33b43e51eb8b781e8a5ff9 ]

A MT unsafe service might get configured to run on another core
while the service is running currently. This might result in the
MT unsafe service running on multiple cores simultaneously. Use
'execute_lock' always when the service is MT unsafe.

If the service is known to be mapped on a single lcore,
setting the service capability to MT safe will avoid taking
the lock and improve the performance.

Fixes: e9139a32f6e8 ("service: add function to run on app lcore")

Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Phil Yang <phil.yang@arm.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
---
 lib/librte_eal/common/include/rte_service.h           |  8 ++++++--
 lib/librte_eal/common/include/rte_service_component.h |  6 +++++-
 lib/librte_eal/common/rte_service.c                   | 11 +++++------
 3 files changed, 16 insertions(+), 9 deletions(-)

diff --git a/lib/librte_eal/common/include/rte_service.h b/lib/librte_eal/common/include/rte_service.h
index 11f673503b..21002209bc 100644
--- a/lib/librte_eal/common/include/rte_service.h
+++ b/lib/librte_eal/common/include/rte_service.h
@@ -105,10 +105,14 @@ int32_t rte_service_probe_capability(uint32_t id, uint32_t capability);
  * function enables or disables *lcore* to run *service_id*.
  *
- * If multiple cores are enabled on a service, an atomic is used to ensure that
- * only one cores runs the service at a time. The exception to this is when
+ * If multiple cores are enabled on a service, a lock is used to ensure that
+ * only one core runs the service at a time. The exception to this is when
  * a service indicates that it is multi-thread safe by setting the capability
  * called RTE_SERVICE_CAP_MT_SAFE. With the multi-thread safe capability set,
  * the service function can be run on multiple threads at the same time.
  *
+ * If the service is known to be mapped to a single lcore, setting the
+ * capability of the service to RTE_SERVICE_CAP_MT_SAFE can achieve
+ * better performance by avoiding the use of lock.
+ *
  * @param service_id the service to apply the lcore to
  * @param lcore The lcore that will be mapped to service
diff --git a/lib/librte_eal/common/include/rte_service_component.h b/lib/librte_eal/common/include/rte_service_component.h
index c12adbc256..259e3182d8 100644
--- a/lib/librte_eal/common/include/rte_service_component.h
+++ b/lib/librte_eal/common/include/rte_service_component.h
@@ -44,5 +44,5 @@ struct rte_service_spec {
  * Register a new service.
  *
- * A service represents a component that the requires CPU time periodically to
+ * A service represents a component that requires CPU time periodically to
  * achieve its purpose.
  *
@@ -57,4 +57,8 @@ struct rte_service_spec {
  * component is ready to be executed.
  *
+ * If the service is known to be mapped to a single lcore, setting the
+ * capability of the service to RTE_SERVICE_CAP_MT_SAFE can achieve
+ * better performance.
+ *
  * @param spec The specification of the service to register
  * @param[out] service_id A pointer to a uint32_t, which will be filled in
diff --git a/lib/librte_eal/common/rte_service.c b/lib/librte_eal/common/rte_service.c
index 5e757fbd19..e71e1efaa8 100644
--- a/lib/librte_eal/common/rte_service.c
+++ b/lib/librte_eal/common/rte_service.c
@@ -50,4 +50,8 @@ struct rte_service_spec_impl {
 
 	/* per service statistics */
+	/* Indicates how many cores the service is mapped to run on.
+	 * It does not indicate the number of cores the service is running
+	 * on currently.
+	 */
 	rte_atomic32_t num_mapped_cores;
 	uint64_t calls;
@@ -370,10 +374,5 @@ service_run(uint32_t i, struct core_state *cs, uint64_t service_mask,
 	cs->service_active_on_lcore[i] = 1;
 
-	/* check do we need cmpset, if MT safe or <= 1 core
-	 * mapped, atomic ops are not required.
-	 */
-	const int use_atomics = (service_mt_safe(s) == 0) &&
-				(rte_atomic32_read(&s->num_mapped_cores) > 1);
-	if (use_atomics) {
+	if (service_mt_safe(s) == 0) {
 		if (!rte_atomic32_cmpset((uint32_t *)&s->execute_lock, 0, 1))
 			return -EBUSY;
-- 
2.21.3

---
  Diff of the applied patch vs upstream commit (please double-check if non-empty:
---
--- -	2020-06-05 19:20:51.965166867 +0100
+++ 0019-service-fix-race-condition-for-MT-unsafe-service.patch	2020-06-05 19:20:50.740042404 +0100
@@ -1 +1 @@
-From 18cae99cb9096abe0f33b43e51eb8b781e8a5ff9 Mon Sep 17 00:00:00 2001
+From 817639c78342003d5d3d675f6bc0eca05ee2c6be Mon Sep 17 00:00:00 2001
@@ -5,0 +6,2 @@
+[ upstream commit 18cae99cb9096abe0f33b43e51eb8b781e8a5ff9 ]
+
@@ -16 +17,0 @@
-Cc: stable@dpdk.org
@@ -22,3 +23,3 @@
- lib/librte_eal/common/rte_service.c            | 11 +++++------
- lib/librte_eal/include/rte_service.h           |  8 ++++++--
- lib/librte_eal/include/rte_service_component.h |  6 +++++-
+ lib/librte_eal/common/include/rte_service.h           |  8 ++++++--
+ lib/librte_eal/common/include/rte_service_component.h |  6 +++++-
+ lib/librte_eal/common/rte_service.c                   | 11 +++++------
@@ -27,29 +28,4 @@
-diff --git a/lib/librte_eal/common/rte_service.c b/lib/librte_eal/common/rte_service.c
-index 70d17a5d79..b8c465eb96 100644
---- a/lib/librte_eal/common/rte_service.c
-+++ b/lib/librte_eal/common/rte_service.c
-@@ -51,4 +51,8 @@ struct rte_service_spec_impl {
- 
- 	/* per service statistics */
-+	/* Indicates how many cores the service is mapped to run on.
-+	 * It does not indicate the number of cores the service is running
-+	 * on currently.
-+	 */
- 	rte_atomic32_t num_mapped_cores;
- 	uint64_t calls;
-@@ -371,10 +375,5 @@ service_run(uint32_t i, struct core_state *cs, uint64_t service_mask,
- 	cs->service_active_on_lcore[i] = 1;
- 
--	/* check do we need cmpset, if MT safe or <= 1 core
--	 * mapped, atomic ops are not required.
--	 */
--	const int use_atomics = (service_mt_safe(s) == 0) &&
--				(rte_atomic32_read(&s->num_mapped_cores) > 1);
--	if (use_atomics) {
-+	if (service_mt_safe(s) == 0) {
- 		if (!rte_atomic32_cmpset((uint32_t *)&s->execute_lock, 0, 1))
- 			return -EBUSY;
-diff --git a/lib/librte_eal/include/rte_service.h b/lib/librte_eal/include/rte_service.h
-index d8701dd4cf..3a1c735c58 100644
---- a/lib/librte_eal/include/rte_service.h
-+++ b/lib/librte_eal/include/rte_service.h
+diff --git a/lib/librte_eal/common/include/rte_service.h b/lib/librte_eal/common/include/rte_service.h
+index 11f673503b..21002209bc 100644
+--- a/lib/librte_eal/common/include/rte_service.h
++++ b/lib/librte_eal/common/include/rte_service.h
@@ -73,4 +49,4 @@
-diff --git a/lib/librte_eal/include/rte_service_component.h b/lib/librte_eal/include/rte_service_component.h
-index 16eab79eea..b75aba11b9 100644
---- a/lib/librte_eal/include/rte_service_component.h
-+++ b/lib/librte_eal/include/rte_service_component.h
+diff --git a/lib/librte_eal/common/include/rte_service_component.h b/lib/librte_eal/common/include/rte_service_component.h
+index c12adbc256..259e3182d8 100644
+--- a/lib/librte_eal/common/include/rte_service_component.h
++++ b/lib/librte_eal/common/include/rte_service_component.h
@@ -92,0 +69,25 @@
+diff --git a/lib/librte_eal/common/rte_service.c b/lib/librte_eal/common/rte_service.c
+index 5e757fbd19..e71e1efaa8 100644
+--- a/lib/librte_eal/common/rte_service.c
++++ b/lib/librte_eal/common/rte_service.c
+@@ -50,4 +50,8 @@ struct rte_service_spec_impl {
+ 
+ 	/* per service statistics */
++	/* Indicates how many cores the service is mapped to run on.
++	 * It does not indicate the number of cores the service is running
++	 * on currently.
++	 */
+ 	rte_atomic32_t num_mapped_cores;
+ 	uint64_t calls;
+@@ -370,10 +374,5 @@ service_run(uint32_t i, struct core_state *cs, uint64_t service_mask,
+ 	cs->service_active_on_lcore[i] = 1;
+ 
+-	/* check do we need cmpset, if MT safe or <= 1 core
+-	 * mapped, atomic ops are not required.
+-	 */
+-	const int use_atomics = (service_mt_safe(s) == 0) &&
+-				(rte_atomic32_read(&s->num_mapped_cores) > 1);
+-	if (use_atomics) {
++	if (service_mt_safe(s) == 0) {
+ 		if (!rte_atomic32_cmpset((uint32_t *)&s->execute_lock, 0, 1))
+ 			return -EBUSY;


  parent reply	other threads:[~2020-06-05 18:26 UTC|newest]

Thread overview: 90+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-05 18:23 [dpdk-stable] patch 'app/testpmd: add parsing for QinQ VLAN headers' " Kevin Traynor
2020-06-05 18:23 ` [dpdk-stable] patch 'doc: fix log level example in Linux guide' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'eal: fix typo in endian conversion macros' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/avp: fix gcc 10 maybe-uninitialized warning' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'eal/x86: ignore gcc 10 stringop-overflow warnings' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'kvargs: fix invalid token parsing on FreeBSD' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'eal/ppc: fix build with gcc 9.3' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/i40e: fix flow director for ARP packets' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'doc: add i40e limitation for flow director' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/i40e: fix flush of flow director filter' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'vhost: fix peer close check' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'vhost: prevent zero-copy with incompatible client mode' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/softnic: fix memory leak for thread' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/softnic: fix resource leak for pipeline' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/octeontx: fix dangling pointer on init failure' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/mlx5: fix RSS enablement' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'eventdev: fix probe and remove for secondary process' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'event/dsw: avoid reusing previously recorded events' " Kevin Traynor
2020-06-05 18:24 ` Kevin Traynor [this message]
2020-06-05 18:24 ` [dpdk-stable] patch 'service: fix identification of service running on other lcore' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'service: remove rte prefix from static functions' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'remove references to private PCI probe function' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'examples/l2fwd-keepalive: fix mbuf pool size' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'mem: fix overflow on allocation' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'examples/eventdev: fix crash on exit' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'test/flow_classify: enable multi-sockets system' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'bbdev: fix doxygen comments' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'crypto/ccp: fix fd leak on probe failure' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'app/crypto-perf: fix display of sample test vector' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'crypto/qat: fix cipher descriptor for ZUC and SNOW' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'crypto/kasumi: fix extern declaration' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/failsafe: fix fd leak' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'app/testpmd: fix statistics after reset' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/netvsc: fix comment spelling' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'bus/vmbus: " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/ixgbe: fix link status synchronization on BSD' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/tap: fix crash in flow destroy' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/bnxt: fix FW version query' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'app/testpmd: fix memory failure handling for i40e DDP' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'bus/fslmc: fix dereferencing null pointer' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/dpaa2: fix 10G port negotiation' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'bus/fslmc: fix size of qman fq descriptor' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/ring: fix device pointer on allocation' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/mlx5: fix matching for UDP tunnels with Verbs' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/mlx4: fix drop queue error handling' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/mlx5: fix Tx queue release debug log timing' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'app: remove extra new line after link duplex' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'examples: " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'event/dsw: fix enqueue burst return value' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'app/eventdev: check Tx adapter service ID' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'crypto/caam_jr: fix check of file descriptors' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'crypto/caam_jr: fix IRQ functions return type' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'build: disable gcc 10 zero-length-bounds warning' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'eal: fix C++17 compilation' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/e1000: fix port hotplug for multi-process' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'doc: fix multicast filter feature announcement' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/ixgbe: fix statistics in flow control mode' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/qede: fix link state configuration' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/vmxnet3: handle bad host framing' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/qede: fix port reconfiguration' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/bnxt: fix error log for command timeout' " Kevin Traynor
2020-06-05 18:24 ` [dpdk-stable] patch 'net/bnxt: fix using RSS config struct' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'app/testpmd: fix DCB set' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'net/ixgbe/base: update copyright' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'net/i40e/base: " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'pci: accept 32-bit domain numbers' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'pci: reject negative values in PCI id' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'kvargs: fix strcmp helper documentation' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'mempool/dpaa2: install missing header with meson' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'fix same typo in multiple places' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'examples/kni: fix MTU change to setup Tx queue' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'net/i40e: fix wild pointer' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'net/i40e: fix queue related exception handling' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'vhost: fix zero-copy server mode' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'net/mvpp2: fix build with gcc 10' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'examples/vm_power: drop Unix path limit redefinition' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'doc: fix build with doxygen 1.8.18' " Kevin Traynor
2020-06-05 22:46   ` Thomas Monjalon
2020-06-10 14:48     ` Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'cryptodev: fix SHA-1 digest enum comment' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'net/sfc/base: fix manual filter delete in EF10' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'net/i40e: fix setting L2TAG' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'net/iavf: " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'net/ixgbe: check driver type in MACsec API' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'examples/kni: fix crash during MTU set' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'examples/ip_pipeline: remove check of null response' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'doc: add NASM installation steps' " Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'doc: fix typo in contributors guide' " Kevin Traynor
2020-06-05 18:25 ` Kevin Traynor
2020-06-05 18:25 ` [dpdk-stable] patch 'doc: prefer https when pointing to dpdk.org' " Kevin Traynor

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200605182525.22483-19-ktraynor@redhat.com \
    --to=ktraynor@redhat.com \
    --cc=harry.van.haaren@intel.com \
    --cc=honnappa.nagarahalli@arm.com \
    --cc=phil.yang@arm.com \
    --cc=stable@dpdk.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

patches for DPDK stable branches

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://inbox.dpdk.org/stable/0 stable/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 stable stable/ https://inbox.dpdk.org/stable \
		stable@dpdk.org
	public-inbox-index stable

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.stable


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git