Age | Commit message (Collapse) | Author |
|
Random packet check maybe failed for misc packets received.
When the received-packet not in the sent-packets, will continue to check
next packet for workaround this kind of issue.
Signed-off-by: Lijuan Tu <lijuanx.a.tu@intel.com>
|
|
This patch fixes a typo in vf_pf_reset_test_plan.rst.
Signed-off-by: Rami Rosen <rami.rosen@intel.com>
|
|
This patch fixes two trivial typos in userspace_ethtool_test_plan.rst.
Signed-off-by: Rami Rosen <rami.rosen@intel.com>
|
|
Send out html format report for single core performance case.
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
|
|
Support html format report for NIC performance and support to send it out
by email.
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
|
|
testpmd removed "crc-strip" parameter for it is enabled by default.
If needed to disable crc-strip, must use new parameter "disable-crc-strip".
Signed-off-by: xu,huilong <huilongx.xu@intel.com>
|
|
dpdk app output log update in dpdk17.05, so must change corroborate text in
ipfrag/ip_ressembly/kni/quota_watermak.
Signed-off-by: xu,huilong <huilongx.xu@intel.com>
|
|
Adapt to 17.05 source code tree.
Signed-off-by: xu,huilong <huilongx.xu@intel.com>
|
|
Support fortville 25G NIC in tso test suite.
Signed-off-by: xu,huilong <huilongx.xu@intel.com>
|
|
Fortville nic will transmit LLDP packets when using dpdk driver.
So add LLDP filter in nvgre and vxlan suites.
Signed-off-by: xu,huilong <huilongx.xu@intel.com>
|
|
|
|
i40e VF HW crc-strip default enabled after commit 60da774e.
|
|
Signed-off-by: Peng Yuan <yuan.peng@intel.com>
|
|
|
|
This patch fixes trivial typos in DDP and MPLS test suite.
Signed-off-by: Rami Rosen <rami.rosen@intel.com>
|
|
|
|
1. Adapt cases to support i350
2. VF device enable crc-strip as default
3. Port should bind to default driver before generating VF device
4. Cleanup vm resouces after case or suite end
Signed-off-by: Lijuan Tu <lijuanx.a.tu@intel.com>
|
|
Signed-off-by: Peng Yuan <yuan.peng@intel.com>
|
|
|
|
|
|
Signed-off-by: Peng Yuan <yuan.peng@intel.com>
|
|
This test used to modify config/common_linuxapp or config/common_bsdapp
to compile DPDK with PCAP, however, this config may lie as well in
config/common_base. If the PCAP variable is not found in OS specific
files, then it is assumed that config/common_base should be modified.
Scapy may contain utils itself, therefore importing dts utils is moved
after importing from scapy (so as to override utils).
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
for Cavium NICs
Following tests are skipped for Cavium NICs, since this functionality
is not implemented:
- test_add_remove_mac_address
- test_invalid_addresses
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
test_link_status_interrupt_port_available for Cavium NICs
The test is skipped right now as link_status_interrupt assumes CPU
frequency which makes the test fail (for Cavium NICs and thunderx
platform)
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
entering main loop" instead of "memory mapped"
Previously integrity test assumed that there should be "memory mapped"
in the output of l2fwd after start up, even though it does not appear.
Test can also be run when 2 same NICs are used now.
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
frames for Cavium NICs
When configuring a Cavium device the max_rx_pkt_len parameter is
ignored (and the device uses its maximum MTU). Therefore, test cases
which assume that bigger than allowed frames should be dropped fail.
Modified conf/dpdk_test_case_checklist.xls to skip
jumboframes_jumbo_nojumbo and jumboframes_bigger_jumbo tests.
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Cavium NICs do not filter PTP packets, so ieee1588 enable test case is
not applicable, modified conf/dpdk_test_case_checklist.xls
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Cavium NICs do not support dlow director, therefore they are added to
following tests in conf/dpdk_test_case_checklist.xls:
-fdir_flexword_drop_ipv4
-fdir_flexword_drop_ipv6
-fdir_flexword_fwd_ipv4
-fdir_flexword_fwd_ipv6
-fdir_noflexword_drop_ipv4
-fdir_noflexword_drop_ipv6
-fdir_noflexword_fwd_ipv4
-fdir_noflexword_fwd_ipv6
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Testpmd must be run with hardware VLAN filter disabled as thunder
driver does not support this feature.
Since disabling promiscuous and allmulticast is not supported for
thunder driver following test cases are set to be skipped:
dynamic_config_broadcast, dynamic_config_disable_promiscuous,
dynamic_config_allmulticast
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Sniffing packets with scapy seems to fail, switched to tcpdump.
Taken into account capabilities of Cavium nics, created
framework/test_capabilities.py to contain information useful when
skipping not applicable parts of tests.
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Testpmd should be run with hardware vlan filter disabled for cavium
nics as the driver does not offer this kind of capability and the
application stops at the beginning, so if cavium nic is used the
--disable-hw-vlan-filter flag is added.
Shutdown api test is also modified so as to skip disabling jumbo frames
mixed with vlan if hardware-assisted VLAN filtering not
configured (rte_eth_dev_set_vlan_offload return -ENOTSUP + there is
"failed diag" in the tespmd output).
Tests to be skipped for cavium nics are added to
conf/dpdk_test_case_checklist.xls
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Result from tcpdump_command may contain something more than only number,
added splitting and taking first element of this operation so that
casting to int proceeds
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Some kernels may not use IPv6 and hence there is no point using ping6.
If lack of link by IPv6 is detected then IPv4 pings are sent
(previously ports where skipped if they had no IPv6 connectivity).
DPDK binding script uses "--force" argument, since interfaces are
detected as active.
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
When collecting pci devices info in
pci_devices_information_uncached_linux devices are checked if they are
from cavium and if so linkspeed is detected to append only 10Gb
interfaces
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Changed default number of set hugepages from 1024 to 8 if architecture
is arm64 and huge pages size is 524288 kB.
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Below patch makes scripts look for correct DPDK binding script,
checking whether 'usertools' or 'tools' directory is present and which
script exist in 'tools' directory ('dpdk-devbind.py' or
'dpdk_nic_bind.py'). Looking for correct pathc is done in get_dpdk_bind_script method.
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Below patch adds preliminary settings to support Cavium NICs.
Signed-off-by: Angela Czubak <aczubak at caviumnetworks.com>
|
|
Signed-off-by: Peng Yuan <yuan.peng@intel.com>
|
|
Remove all the match network device object by copy list first.
Signed-off-by: Lijuan Tu <lijuanx.a.tu@intel.com>
|
|
Signed-off-by: xu,gang <gangx.xu@intel.com>
|
|
Previously, a single VLAN header is treated as inner VLAN,
but generally, a single VLAN header is treated as the outer
VLAN header.
Signed-off-by: Lijuan Tu <lijuanx.a.tu@intel.com>
|
|
Previously, a single VLAN header is treated as inner VLAN,
but generally, a single VLAN header is treated as the outer
VLAN header.
i40e refer to dpdk commit: 5b2d37858d32ec6aba78f15dc065f286108d6b4e
ixgbe refer to dpdk commit: 2b84092427b1800cba7d3204b95703b1718fcd28
Signed-off-by: Lijuan Tu <lijuanx.a.tu@intel.com>
|
|
- add QAT pmd performance auto test
- add AESNI_MB pmd performance auto test
- add AESNI_GCM pmd performance auto test
- add KASUMI pmd performance auto test
- add SNOW3G pmd performance auto test
- add ZUC pmd performance auto test
- add openssl pmd performance auto test
- add crypto performance test plan
- update license date
|
|
- Fix pep8 issues
- Add DOCSIS support
- Add AES GCM support
- Refine the test vectors
|
|
1. Change the packet generator from "scapy" to packet module
2. Ehanhce the cipher/hash output generator with cryptography module,
instead of openssl
3. Update comments and clean up code
|
|
Signed-off-by: Lu,peipei <peipeix.lu@intel.com>
|
|
|
|
v4:
- Move the killall process to tear_down, managed by dts framework
- Change the "print" to logger method
v3:
- Fix PEP8 issues
v2:
- Update copyright dates
- Update coryright icon
|
|
- Refactor Test App building and installation to a new function
- Add Test App building and installation in retry step
(which build dpdk without -j option), so that fix sometime
retry building successfully, but test app miss
|
|
Signed-off-by: xu,huilong <huilongx.xu@intel.com>
|