Re: [ovs-dev] [PATCH v2 3/3] system-dpdk: Test with mlx5 devices.
On 24 Nov 2021, at 17:14, Eelco Chaudron wrote: > On 23 Nov 2021, at 15:15, David Marchand wrote: > >> The DPDK unit test only runs if vfio or igb_uio kernel module is loaded: >> on systems with only mlx5, this test is always skipped. >> >> Besides, the test tries to grab the first device listed by dpdk-devbind.py, >> regardless of the PCI device status regarding kmod binding. >> >> Remove dependency on this DPDK script and use a minimal script that >> reads PCI sysfs. >> >> This script is not perfect, as one can imagine PCI devices bound to >> vfio-pci for virtual machines. >> Add a new environment variable DPDK_PCI_ADDR for testers to select the >> PCI device of their liking. >> For consistency and grep, the temporary file PCI_ADDR is renamed >> to DPDK_PCI_ADDR. >> >> Note: with mlx5 devices, there is now more OVS/DPDK warnings to waive. > > Uncovered some build issues with my scripts as now my MLX5 card in the system > is used :) > > Changes look good to me, one small nit. > > Acked-by: Eelco Chaudron > >> Signed-off-by: David Marchand >> --- >> Documentation/topics/testing.rst | 1 - >> tests/automake.mk| 1 + >> tests/system-dpdk-find-device.py | 44 >> tests/system-dpdk-macros.at | 10 ++-- >> tests/system-dpdk.at | 4 ++- >> 5 files changed, 50 insertions(+), 10 deletions(-) >> create mode 100755 tests/system-dpdk-find-device.py >> >> diff --git a/Documentation/topics/testing.rst >> b/Documentation/topics/testing.rst >> index 0ddcbd58ab..00f734a77a 100644 >> --- a/Documentation/topics/testing.rst >> +++ b/Documentation/topics/testing.rst >> @@ -343,7 +343,6 @@ To see a list of all the available tests, run:: >> >> These tests support a `DPDK supported NIC`_. The tests operate on a wider >> set of >> environments, for instance, when a virtual port is used. >> -They do require proper DPDK variables (``DPDK_DIR`` and ``DPDK_BUILD``). >> Moreover you need to have root privileges to load the required modules and >> to bind >> the NIC to the DPDK-compatible driver. >> >> diff --git a/tests/automake.mk b/tests/automake.mk >> index c519a5cb36..ead92ace0f 100644 >> --- a/tests/automake.mk >> +++ b/tests/automake.mk >> @@ -188,6 +188,7 @@ SYSTEM_OFFLOADS_TESTSUITE_AT = \ >> >> SYSTEM_DPDK_TESTSUITE_AT = \ >> tests/system-common-macros.at \ >> +tests/system-dpdk-find-device.py \ >> tests/system-dpdk-macros.at \ >> tests/system-dpdk-testsuite.at \ >> tests/system-dpdk.at >> diff --git a/tests/system-dpdk-find-device.py >> b/tests/system-dpdk-find-device.py >> new file mode 100755 >> index 00..4253326e75 >> --- /dev/null >> +++ b/tests/system-dpdk-find-device.py >> @@ -0,0 +1,44 @@ >> +#!/usr/bin/env python3 Forgot to complain about the this, i.e. “#!/usr/bin/env python3” ;) >> +# Copyright (c) 2021 Red Hat, Inc. >> +# >> +# Licensed under the Apache License, Version 2.0 (the "License"); >> +# you may not use this file except in compliance with the License. >> +# You may obtain a copy of the License at: >> +# >> +# http://www.apache.org/licenses/LICENSE-2.0 >> +# >> +# Unless required by applicable law or agreed to in writing, software >> +# distributed under the License is distributed on an "AS IS" BASIS, >> +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. >> +# See the License for the specific language governing permissions and >> +# limitations under the License. >> + >> + >> +from pathlib import Path >> +import os >> +import sys >> + >> +# The tester might want to select a PCI device, if so, trust it. >> +if 'DPDK_PCI_ADDR' in os.environ: >> +print(os.environ['DPDK_PCI_ADDR']) >> +sys.exit(0) >> + >> +mlx5_ib_available = Path('/sys/module/mlx5_ib').exists() >> + >> +for device in sorted(Path('/sys/bus/pci/devices').iterdir()): >> +class_path = device / 'class' >> +# Only consider Network class devices >> +if class_path.read_text().strip() != '0x02': >> +continue >> +kmod_path = device / 'driver' / 'module' >> +kmod_name = kmod_path.resolve().name >> +# Devices of interest: >> +# - device is bound to vfio_pci or igb_uio, >> +# - device is bound to mlx5_core and mlx5 is loaded, >> +if (kmod_name not in ['vfio_pci', 'igb_uio'] and >> +(kmod_name not in ['mlx5_core'] or not mlx5_ib_available)): > > Flake8 is complaining about this. But for some reason, I like this better ;) > Quick online search reviews unclear interpretations, so I would leave it as > is… > > system-dpdk-find-device.py:39:9: E129 visually indented line with same indent > as next logical line > >> +continue >> +print(device.resolve().name) >> +sys.exit(0) >> + >> +sys.exit(1) >> diff --git a/tests/system-dpdk-macros.at b/tests/system-dpdk-macros.at >> index b3614f0565..066fae0ba5 100644 >> --- a/tests/system-dpdk-macros.at >> +++ b/tests/system-dpdk-macros.at >> @@ -22,14 +22,8 @@
Re: [ovs-dev] [PATCH v2 3/3] system-dpdk: Test with mlx5 devices.
On 23 Nov 2021, at 15:15, David Marchand wrote: > The DPDK unit test only runs if vfio or igb_uio kernel module is loaded: > on systems with only mlx5, this test is always skipped. > > Besides, the test tries to grab the first device listed by dpdk-devbind.py, > regardless of the PCI device status regarding kmod binding. > > Remove dependency on this DPDK script and use a minimal script that > reads PCI sysfs. > > This script is not perfect, as one can imagine PCI devices bound to > vfio-pci for virtual machines. > Add a new environment variable DPDK_PCI_ADDR for testers to select the > PCI device of their liking. > For consistency and grep, the temporary file PCI_ADDR is renamed > to DPDK_PCI_ADDR. > > Note: with mlx5 devices, there is now more OVS/DPDK warnings to waive. Uncovered some build issues with my scripts as now my MLX5 card in the system is used :) Changes look good to me, one small nit. Acked-by: Eelco Chaudron > Signed-off-by: David Marchand > --- > Documentation/topics/testing.rst | 1 - > tests/automake.mk| 1 + > tests/system-dpdk-find-device.py | 44 > tests/system-dpdk-macros.at | 10 ++-- > tests/system-dpdk.at | 4 ++- > 5 files changed, 50 insertions(+), 10 deletions(-) > create mode 100755 tests/system-dpdk-find-device.py > > diff --git a/Documentation/topics/testing.rst > b/Documentation/topics/testing.rst > index 0ddcbd58ab..00f734a77a 100644 > --- a/Documentation/topics/testing.rst > +++ b/Documentation/topics/testing.rst > @@ -343,7 +343,6 @@ To see a list of all the available tests, run:: > > These tests support a `DPDK supported NIC`_. The tests operate on a wider > set of > environments, for instance, when a virtual port is used. > -They do require proper DPDK variables (``DPDK_DIR`` and ``DPDK_BUILD``). > Moreover you need to have root privileges to load the required modules and > to bind > the NIC to the DPDK-compatible driver. > > diff --git a/tests/automake.mk b/tests/automake.mk > index c519a5cb36..ead92ace0f 100644 > --- a/tests/automake.mk > +++ b/tests/automake.mk > @@ -188,6 +188,7 @@ SYSTEM_OFFLOADS_TESTSUITE_AT = \ > > SYSTEM_DPDK_TESTSUITE_AT = \ > tests/system-common-macros.at \ > + tests/system-dpdk-find-device.py \ > tests/system-dpdk-macros.at \ > tests/system-dpdk-testsuite.at \ > tests/system-dpdk.at > diff --git a/tests/system-dpdk-find-device.py > b/tests/system-dpdk-find-device.py > new file mode 100755 > index 00..4253326e75 > --- /dev/null > +++ b/tests/system-dpdk-find-device.py > @@ -0,0 +1,44 @@ > +#!/usr/bin/env python3 > +# Copyright (c) 2021 Red Hat, Inc. > +# > +# Licensed under the Apache License, Version 2.0 (the "License"); > +# you may not use this file except in compliance with the License. > +# You may obtain a copy of the License at: > +# > +# http://www.apache.org/licenses/LICENSE-2.0 > +# > +# Unless required by applicable law or agreed to in writing, software > +# distributed under the License is distributed on an "AS IS" BASIS, > +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. > +# See the License for the specific language governing permissions and > +# limitations under the License. > + > + > +from pathlib import Path > +import os > +import sys > + > +# The tester might want to select a PCI device, if so, trust it. > +if 'DPDK_PCI_ADDR' in os.environ: > +print(os.environ['DPDK_PCI_ADDR']) > +sys.exit(0) > + > +mlx5_ib_available = Path('/sys/module/mlx5_ib').exists() > + > +for device in sorted(Path('/sys/bus/pci/devices').iterdir()): > +class_path = device / 'class' > +# Only consider Network class devices > +if class_path.read_text().strip() != '0x02': > +continue > +kmod_path = device / 'driver' / 'module' > +kmod_name = kmod_path.resolve().name > +# Devices of interest: > +# - device is bound to vfio_pci or igb_uio, > +# - device is bound to mlx5_core and mlx5 is loaded, > +if (kmod_name not in ['vfio_pci', 'igb_uio'] and > +(kmod_name not in ['mlx5_core'] or not mlx5_ib_available)): Flake8 is complaining about this. But for some reason, I like this better ;) Quick online search reviews unclear interpretations, so I would leave it as is… system-dpdk-find-device.py:39:9: E129 visually indented line with same indent as next logical line > +continue > +print(device.resolve().name) > +sys.exit(0) > + > +sys.exit(1) > diff --git a/tests/system-dpdk-macros.at b/tests/system-dpdk-macros.at > index b3614f0565..066fae0ba5 100644 > --- a/tests/system-dpdk-macros.at > +++ b/tests/system-dpdk-macros.at > @@ -22,14 +22,8 @@ m4_define([OVS_DPDK_PRE_PHY_SKIP], >[dnl Perform the precheck > OVS_DPDK_PRE_CHECK() > > - dnl Check if VFIO or UIO driver is loaded > - AT_SKIP_IF([ ! (lsmod | grep -E "igb_uio|vfio") ], [], [stdout]) > - > - dnl Find PCI address candidate, skip if there is no
[ovs-dev] [PATCH v2 3/3] system-dpdk: Test with mlx5 devices.
The DPDK unit test only runs if vfio or igb_uio kernel module is loaded: on systems with only mlx5, this test is always skipped. Besides, the test tries to grab the first device listed by dpdk-devbind.py, regardless of the PCI device status regarding kmod binding. Remove dependency on this DPDK script and use a minimal script that reads PCI sysfs. This script is not perfect, as one can imagine PCI devices bound to vfio-pci for virtual machines. Add a new environment variable DPDK_PCI_ADDR for testers to select the PCI device of their liking. For consistency and grep, the temporary file PCI_ADDR is renamed to DPDK_PCI_ADDR. Note: with mlx5 devices, there is now more OVS/DPDK warnings to waive. Signed-off-by: David Marchand --- Documentation/topics/testing.rst | 1 - tests/automake.mk| 1 + tests/system-dpdk-find-device.py | 44 tests/system-dpdk-macros.at | 10 ++-- tests/system-dpdk.at | 4 ++- 5 files changed, 50 insertions(+), 10 deletions(-) create mode 100755 tests/system-dpdk-find-device.py diff --git a/Documentation/topics/testing.rst b/Documentation/topics/testing.rst index 0ddcbd58ab..00f734a77a 100644 --- a/Documentation/topics/testing.rst +++ b/Documentation/topics/testing.rst @@ -343,7 +343,6 @@ To see a list of all the available tests, run:: These tests support a `DPDK supported NIC`_. The tests operate on a wider set of environments, for instance, when a virtual port is used. -They do require proper DPDK variables (``DPDK_DIR`` and ``DPDK_BUILD``). Moreover you need to have root privileges to load the required modules and to bind the NIC to the DPDK-compatible driver. diff --git a/tests/automake.mk b/tests/automake.mk index c519a5cb36..ead92ace0f 100644 --- a/tests/automake.mk +++ b/tests/automake.mk @@ -188,6 +188,7 @@ SYSTEM_OFFLOADS_TESTSUITE_AT = \ SYSTEM_DPDK_TESTSUITE_AT = \ tests/system-common-macros.at \ + tests/system-dpdk-find-device.py \ tests/system-dpdk-macros.at \ tests/system-dpdk-testsuite.at \ tests/system-dpdk.at diff --git a/tests/system-dpdk-find-device.py b/tests/system-dpdk-find-device.py new file mode 100755 index 00..4253326e75 --- /dev/null +++ b/tests/system-dpdk-find-device.py @@ -0,0 +1,44 @@ +#!/usr/bin/env python3 +# Copyright (c) 2021 Red Hat, Inc. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at: +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +from pathlib import Path +import os +import sys + +# The tester might want to select a PCI device, if so, trust it. +if 'DPDK_PCI_ADDR' in os.environ: +print(os.environ['DPDK_PCI_ADDR']) +sys.exit(0) + +mlx5_ib_available = Path('/sys/module/mlx5_ib').exists() + +for device in sorted(Path('/sys/bus/pci/devices').iterdir()): +class_path = device / 'class' +# Only consider Network class devices +if class_path.read_text().strip() != '0x02': +continue +kmod_path = device / 'driver' / 'module' +kmod_name = kmod_path.resolve().name +# Devices of interest: +# - device is bound to vfio_pci or igb_uio, +# - device is bound to mlx5_core and mlx5 is loaded, +if (kmod_name not in ['vfio_pci', 'igb_uio'] and +(kmod_name not in ['mlx5_core'] or not mlx5_ib_available)): +continue +print(device.resolve().name) +sys.exit(0) + +sys.exit(1) diff --git a/tests/system-dpdk-macros.at b/tests/system-dpdk-macros.at index b3614f0565..066fae0ba5 100644 --- a/tests/system-dpdk-macros.at +++ b/tests/system-dpdk-macros.at @@ -22,14 +22,8 @@ m4_define([OVS_DPDK_PRE_PHY_SKIP], [dnl Perform the precheck OVS_DPDK_PRE_CHECK() - dnl Check if VFIO or UIO driver is loaded - AT_SKIP_IF([ ! (lsmod | grep -E "igb_uio|vfio") ], [], [stdout]) - - dnl Find PCI address candidate, skip if there is no DPDK-compatible NIC - AT_CHECK([$DPDK_DIR/usertools/dpdk-devbind.py -s | head -n +4 | tail -1], [], [stdout]) - AT_CHECK([cat stdout | cut -d" " -s -f1 > PCI_ADDR]) - AT_SKIP_IF([ ! test -s PCI_ADDR ]) - + dnl Check if a device is available for DPDK + AT_SKIP_IF([ ! $abs_top_srcdir/tests/system-dpdk-find-device.py > DPDK_PCI_ADDR]) ]) diff --git a/tests/system-dpdk.at b/tests/system-dpdk.at index 7b5d063f62..894e01316f 100644 --- a/tests/system-dpdk.at +++ b/tests/system-dpdk.at @@ -7,6 +7,8 @@ m4_define([SYSTEM_DPDK_ALLOWED_LOGS],[ \@does not exist. The Open vSwitch kernel module is probably not loaded.@d \@EAL: Invalid NUMA socket, default to 0@d \@EAL: No