patches for DPDK stable branches
 help / color / mirror / Atom feed
From: Yongseok Koh <yskoh@mellanox.com>
To: Ferruh Yigit <ferruh.yigit@intel.com>
Cc: Thomas Monjalon <thomas@monjalon.net>, dpdk stable <stable@dpdk.org>
Subject: [dpdk-stable] patch 'mk: fix scope of disabling AVX512F support' has been queued to LTS release 17.11.6
Date: Fri,  8 Mar 2019 09:47:10 -0800	[thread overview]
Message-ID: <20190308174749.30771-32-yskoh@mellanox.com> (raw)
In-Reply-To: <20190308174749.30771-1-yskoh@mellanox.com>

Hi,

FYI, your patch has been queued to LTS release 17.11.6

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objection by 03/13/19. So please
shout if anyone has objection.

Also note that after the patch there's a diff of the upstream commit vs the patch applied
to the branch. If the code is different (ie: not only metadata diffs), due for example to
a change in context or macro names, please double check it.

Thanks.

Yongseok

---
>From e2e5744253733ce137e780699da00a627a25fc6a Mon Sep 17 00:00:00 2001
From: Ferruh Yigit <ferruh.yigit@intel.com>
Date: Mon, 14 Jan 2019 15:49:35 +0000
Subject: [PATCH] mk: fix scope of disabling AVX512F support

[ backported from upstream commit a32ca9a4ebc1350ce48df0222decef95a72b742b ]

AVX512 was disabled for GCC because of Bugzilla issue 97 [1],
the GCC defect submitted for the issue [2] highlighted that this is
a known binutils version 2.30 issue.

Narrowed the scope of no-avx512 to the this specific binutils version.

[1]
https://bugs.dpdk.org/show_bug.cgi?id=97

[2]
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88096

Fixes: 1bc8541000ff ("mk: disable gcc AVX512F support")

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
---
 doc/guides/rel_notes/known_issues.rst    | 35 ++++++++++++++++++++++++++++++++
 mk/rte.cpuflags.mk                       |  4 ++--
 mk/toolchain/gcc/rte.toolchain-compat.mk | 10 +++++++++
 3 files changed, 47 insertions(+), 2 deletions(-)

diff --git a/doc/guides/rel_notes/known_issues.rst b/doc/guides/rel_notes/known_issues.rst
index 3f6d8cb5f..fa0082137 100644
--- a/doc/guides/rel_notes/known_issues.rst
+++ b/doc/guides/rel_notes/known_issues.rst
@@ -742,3 +742,38 @@ igb_uio can not be used when running l3fwd-power
 
 **Driver/Module**:
    ``igb_uio`` module.
+
+
+AVX-512 support disabled
+------------------------
+
+**Description**:
+   ``AVX-512`` support has been disabled on some conditions.
+   This shouldn't be confused with ``CONFIG_RTE_ENABLE_AVX512`` config option which is already
+   disabled by default. This config option defines if ``AVX-512`` specific implementations of
+   some file to be used or not. What has been disabled is compiler feature to produce ``AVX-512``
+   instructions from any source code.
+
+   On DPDK v18.11 ``AVX-512`` is disabled for all ``GCC`` builds which reported to cause a performance
+   drop.
+
+   On DPDK v19.02 ``AVX-512`` disable scope is reduced to ``GCC`` and ``binutils version 2.30`` based
+   on information accured from the GCC community defect.
+
+**Reason**:
+   Generated ``AVX-512`` code cause crash:
+   https://bugs.dpdk.org/show_bug.cgi?id=97
+   https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88096
+
+**Resolution/Workaround**:
+   * Update ``binutils`` to newer version than ``2.30``.
+
+   OR
+
+   * Use different compiler, like ``clang`` for this case.
+
+**Affected Environment/Platform**:
+    ``GCC`` and ``binutils version 2.30``.
+
+**Driver/Module**:
+    ALL.
diff --git a/mk/rte.cpuflags.mk b/mk/rte.cpuflags.mk
index 2129510c2..eee4868b2 100644
--- a/mk/rte.cpuflags.mk
+++ b/mk/rte.cpuflags.mk
@@ -97,8 +97,8 @@ ifneq ($(filter $(AUTO_CPUFLAGS),__AVX512F__),)
 ifeq ($(CONFIG_RTE_ENABLE_AVX512),y)
 CPUFLAGS += AVX512F
 else
-# disable AVX512F support of gcc as a workaround for Bug 97
-ifeq ($(CONFIG_RTE_TOOLCHAIN_GCC),y)
+# disable AVX512F support for GCC & binutils 2.30 as a workaround for Bug 97
+ifeq ($(FORCE_DISABLE_AVX512),y)
 MACHINE_CFLAGS += -mno-avx512f
 endif
 endif
diff --git a/mk/toolchain/gcc/rte.toolchain-compat.mk b/mk/toolchain/gcc/rte.toolchain-compat.mk
index 8dbd4417e..ad3b68e63 100644
--- a/mk/toolchain/gcc/rte.toolchain-compat.mk
+++ b/mk/toolchain/gcc/rte.toolchain-compat.mk
@@ -47,6 +47,16 @@ HOST_GCC_MINOR = $(shell echo __GNUC_MINOR__ | $(HOSTCC) -E -x c - | tail -n 1)
 HOST_GCC_PATCHLEVEL = $(shell echo __GNUC_PATCHLEVEL__ | $(HOSTCC) -E -x c - | tail -n 1)
 HOST_GCC_VERSION = $(HOST_GCC_MAJOR)$(HOST_GCC_MINOR)
 
+LD_VERSION = $(shell $(LD) -v)
+# disable AVX512F support for GCC & binutils 2.30 as a workaround for Bug 97
+ifneq ($(filter 2.30%,$(LD_VERSION)),)
+FORCE_DISABLE_AVX512 := y
+# print warning only once for librte_eal
+ifneq ($(filter %librte_eal,$(CURDIR)),)
+$(warning AVX512 support disabled because of ld 2.30. See Bug 97)
+endif
+endif
+
 # if GCC is older than 4.x
 ifeq ($(shell test $(GCC_VERSION) -lt 40 && echo 1), 1)
 	MACHINE_CFLAGS =
-- 
2.11.0

---
  Diff of the applied patch vs upstream commit (please double-check if non-empty:
---
--- -	2019-03-08 09:46:41.889660902 -0800
+++ 0032-mk-fix-scope-of-disabling-AVX512F-support.patch	2019-03-08 09:46:40.143408000 -0800
@@ -1,8 +1,10 @@
-From a32ca9a4ebc1350ce48df0222decef95a72b742b Mon Sep 17 00:00:00 2001
+From e2e5744253733ce137e780699da00a627a25fc6a Mon Sep 17 00:00:00 2001
 From: Ferruh Yigit <ferruh.yigit@intel.com>
 Date: Mon, 14 Jan 2019 15:49:35 +0000
 Subject: [PATCH] mk: fix scope of disabling AVX512F support
 
+[ backported from upstream commit a32ca9a4ebc1350ce48df0222decef95a72b742b ]
+
 AVX512 was disabled for GCC because of Bugzilla issue 97 [1],
 the GCC defect submitted for the issue [2] highlighted that this is
 a known binutils version 2.30 issue.
@@ -15,43 +17,21 @@
 [2]
 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88096
 
-Fixes: 8d07c82b239f ("mk: disable gcc AVX512F support")
-Cc: stable@dpdk.org
+Fixes: 1bc8541000ff ("mk: disable gcc AVX512F support")
 
 Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
 Acked-by: Thomas Monjalon <thomas@monjalon.net>
 ---
- config/meson.build                       |  8 ++++++++
  doc/guides/rel_notes/known_issues.rst    | 35 ++++++++++++++++++++++++++++++++
- doc/guides/rel_notes/release_19_02.rst   | 15 ++++++++++++++
  mk/rte.cpuflags.mk                       |  4 ++--
  mk/toolchain/gcc/rte.toolchain-compat.mk | 10 +++++++++
- 5 files changed, 70 insertions(+), 2 deletions(-)
+ 3 files changed, 47 insertions(+), 2 deletions(-)
 
-diff --git a/config/meson.build b/config/meson.build
-index db32499b3..40802fc88 100644
---- a/config/meson.build
-+++ b/config/meson.build
-@@ -43,6 +43,14 @@ toolchain = cc.get_id()
- dpdk_conf.set_quoted('RTE_TOOLCHAIN', toolchain)
- dpdk_conf.set('RTE_TOOLCHAIN_' + toolchain.to_upper(), 1)
- 
-+# get binutils version for the workaround of Bug 97
-+ldver = run_command('ld', '-v').stdout().strip()
-+if ldver.contains('2.30')
-+	if cc.has_argument('-mno-avx512f')
-+		machine_args += '-mno-avx512f'
-+	endif
-+endif
-+
- add_project_link_arguments('-Wl,--no-as-needed', language: 'c')
- dpdk_extra_ldflags += '-Wl,--no-as-needed'
- 
 diff --git a/doc/guides/rel_notes/known_issues.rst b/doc/guides/rel_notes/known_issues.rst
-index a1face9c1..358dfa321 100644
+index 3f6d8cb5f..fa0082137 100644
 --- a/doc/guides/rel_notes/known_issues.rst
 +++ b/doc/guides/rel_notes/known_issues.rst
-@@ -826,3 +826,38 @@ Kernel crash when hot-unplug igb_uio device while DPDK application is running
+@@ -742,3 +742,38 @@ igb_uio can not be used when running l3fwd-power
  
  **Driver/Module**:
     ``igb_uio`` module.
@@ -90,37 +70,11 @@
 +
 +**Driver/Module**:
 +    ALL.
-diff --git a/doc/guides/rel_notes/release_19_02.rst b/doc/guides/rel_notes/release_19_02.rst
-index 832ac8c2a..5a2f62ca9 100644
---- a/doc/guides/rel_notes/release_19_02.rst
-+++ b/doc/guides/rel_notes/release_19_02.rst
-@@ -334,6 +334,21 @@ Known Issues
-    Also, make sure to start the actual text at the margin.
-    =========================================================
- 
-+* ``AVX-512`` support has been disabled for ``GCC`` builds when ``binutils 2.30``
-+  is detected [1] because of a crash [2]. This can affect ``native`` machine type
-+  build targets on the platforms that support ``AVX512F`` like ``Intel Skylake``
-+  processors, and can cause a possible performance drop. The immediate workaround
-+  is to use ``clang`` compiler on these platforms.
-+  Initial workaround in DPDK v18.11 was to disable ``AVX-512`` support for ``GCC``
-+  completely, but based on information on defect submitted to GCC community [3],
-+  issue has been identified as ``binutils 2.30`` issue. Since currently only GCC
-+  generates ``AVX-512`` instructions, the scope is limited to ``GCC`` and
-+  ``binutils 2.30``
-+
-+  - [1]: Commit ("mk: fix scope of disabling AVX512F support")
-+  - [2]: https://bugs.dpdk.org/show_bug.cgi?id=97
-+  - [3]: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88096
-+
- 
- Tested Platforms
- ----------------
 diff --git a/mk/rte.cpuflags.mk b/mk/rte.cpuflags.mk
-index c3291b17a..541211c61 100644
+index 2129510c2..eee4868b2 100644
 --- a/mk/rte.cpuflags.mk
 +++ b/mk/rte.cpuflags.mk
-@@ -69,8 +69,8 @@ ifneq ($(filter $(AUTO_CPUFLAGS),__AVX512F__),)
+@@ -97,8 +97,8 @@ ifneq ($(filter $(AUTO_CPUFLAGS),__AVX512F__),)
  ifeq ($(CONFIG_RTE_ENABLE_AVX512),y)
  CPUFLAGS += AVX512F
  else
@@ -132,10 +86,10 @@
  endif
  endif
 diff --git a/mk/toolchain/gcc/rte.toolchain-compat.mk b/mk/toolchain/gcc/rte.toolchain-compat.mk
-index 44904295c..dbddc986e 100644
+index 8dbd4417e..ad3b68e63 100644
 --- a/mk/toolchain/gcc/rte.toolchain-compat.mk
 +++ b/mk/toolchain/gcc/rte.toolchain-compat.mk
-@@ -20,6 +20,16 @@ HOST_GCC_MINOR = $(shell echo __GNUC_MINOR__ | $(HOSTCC) -E -x c - | tail -n 1)
+@@ -47,6 +47,16 @@ HOST_GCC_MINOR = $(shell echo __GNUC_MINOR__ | $(HOSTCC) -E -x c - | tail -n 1)
  HOST_GCC_PATCHLEVEL = $(shell echo __GNUC_PATCHLEVEL__ | $(HOSTCC) -E -x c - | tail -n 1)
  HOST_GCC_VERSION = $(HOST_GCC_MAJOR)$(HOST_GCC_MINOR)
  

  parent reply	other threads:[~2019-03-08 17:48 UTC|newest]

Thread overview: 75+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-03-08 17:46 [dpdk-stable] patch 'net/bnx2x: cleanup info logs' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/i40e: fix getting RSS configuration' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/i40e: fix using recovery mode firmware' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/ixgbe: fix overwriting RSS RETA' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/mlx5: fix validation of Rx queue number' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'ethdev: fix typo in queue setup error log' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'drivers/net: fix several Tx prepare functions' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'vhost: fix crash after mmap failure' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/i40e: revert fix offload not supported mask' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/i40e: remove redundant reset of queue number' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'doc: fix garbage text in generated HTML guides' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'eventdev: fix xstats documentation typo' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'crypto/qat: fix block size error handling' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'timer: fix race condition' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/i40e: fix statistics inconsistency' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'vhost: fix race condition when adding fd in the fdset' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/ena: add supported RSS offloads types' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/ena: update completion queue after cleanup' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net: fix underflow for checksum of invalid IPv4 packets' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/tap: add buffer overflow checks before checksum' " Yongseok Koh
2019-03-08 17:46 ` [dpdk-stable] patch 'net/af_packet: fix setting MTU decrements sockaddr twice' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/tap: fix possible uninitialized variable access' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'app/testpmd: expand RED queue thresholds to 64 bits' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/i40e: fix get RSS conf' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'devtools: fix wrong headline lowercase for arm' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'drivers/crypto: fix PMDs memory leak' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'test/crypto: fix misleading trace message' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'examples/ipsec-secgw: fix outbound codepath for single SA' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'examples/ipsec-secgw: make local variables static' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'kni: fix build on RHEL 8' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'kni: fix build on RHEL8 for arm and Power9' " Yongseok Koh
2019-03-08 17:47 ` Yongseok Koh [this message]
2019-03-08 17:47 ` [dpdk-stable] patch 'net/ixgbe: fix over using multicast table for VF' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'vhost: fix possible out of bound access in vector filling' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/fm10k: fix internal switch initial status' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/dpaa: fix secondary process' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'examples/flow_filtering: fix example documentation' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'doc: fix a parameter name in testpmd guide' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'app/testpmd: fix quit to stop all ports before close' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/bonding: fix possible null pointer reference' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/sfc: discard last seen VLAN TCI if Tx packet is dropped' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/dpaa2: fix device init for secondary process' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/sfc: fix typo in preprocessor check' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'drivers: fix sprintf with snprintf' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/mlx5: fix instruction hotspot on replenishing Rx buffer' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'examples/tep_term: remove unused constant' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'eal: fix core number validation' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'eal: fix out of bound access when no CPU available' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'eal: check string parameter lengths' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'gro: check invalid TCP header length' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/i40e: fix VF overwrite PF RSS LUT for X722' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/sfc: fix VF error/missed stats mapping' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/sfc: fix datapath name references in logs' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'vhost: fix memory leak on realloc failure' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'examples/vhost: fix path allocation failure handling' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/sfc: fix Rx packets counter' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'doc: add missing loopback option in testpmd guide' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'efd: fix tail queue leak' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/qede: fix performance bottleneck in Rx path' " Yongseok Koh
2019-03-12 17:04   ` [dpdk-stable] [EXT] " Shahed Shaikh
2019-03-12 22:03     ` Yongseok Koh
2019-03-19 19:12       ` Shahed Shaikh
2019-03-27 18:47         ` Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/qede: remove prefetch in Tx " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'doc: fix references in power management guide' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'ethdev: fix errno to have positive value' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'gso: fix VxLAN/GRE tunnel checks' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'kni: fix build for dev_open in Linux 5.0' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'kni: fix build for igb_ndo_bridge_setlink " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'vfio: fix error message' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/i40e: fix queue region DCB configure' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/virtio-user: fix used ring in cvq handling' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/ena: fix dev init with multi-process' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'net/ena: fix errno to positive value' " Yongseok Koh
2019-03-08 17:47 ` [dpdk-stable] patch 'doc: add dependency for PDF in contributing guide' " Yongseok Koh

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=20190308174749.30771-32-yskoh@mellanox.com \
    --to=yskoh@mellanox.com \
    --cc=ferruh.yigit@intel.com \
    --cc=stable@dpdk.org \
    --cc=thomas@monjalon.net \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).