The ARMv8.1 CASP instruction works with even register pairs and since
there no register constraint in older versions of GCC/Clang, use
explicit register allocation to satisfy CASP requirements.
Remove function level optimization specification as Clang doesn't have
support for it and explicit register allocation removes the need for it.
Fixes build issue with arm64-armv8a-linux-clang.
Fixes: ee338015e7a9 ("mempool/octeontx2: add optimized dequeue operation for arm64") Reported-by: Gavin Hu <gavin.hu@arm.com> Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com> Acked-by: Jerin Jacob <jerinj@marvell.com>
test/hash: init parameters in the correct function
Each test case initializes its hash parameters in the test case
function. To be consistent, generate keys function should initialize
hash parameters similarly.
Added unit test cases for EAL flags --proc-type=auto and
--create-uio-dev in order to cover the below functions
eal_proc_type_detect()
rte_eal_create_uio_dev()
On some systems when dpdk test is executed with make test command
autotest_runner crashes in first_cpu_on_node. This happens when list
of available cpus contains something that is not a cpu as first element.
Fixed by removing all non-cpu values from list of available cpus.
Unit test table_autotest results in segmentation fault.
Crash occurs in test_table_lpm_ipv6_combined().
Variable 'nht_pos0' used as array subscript is not initialized
in rte_table_lpm_ipv6_entry_add(). It will not be assigned,
if a rule does not exist.
In such case a junk number or invalid array index might result in
segmentation fault due to array out of bounds when
lpm->nht_users is used with such invalid array index.
Fix is to initialize the variables used for array subscript.
Bugzilla ID: 285 Fixes: d89a5bce1d ("lpm6: extend next hop field") Cc: stable@dpdk.org Signed-off-by: Jananee Parthasarathy <jananeex.m.parthasarathy@intel.com> Tested-by: David Marchand <david.marchand@redhat.com> Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Gage Eads [Thu, 20 Jun 2019 22:07:35 +0000 (17:07 -0500)]
doc: add a note for multi-process in mempool guide
The mempool library assigns handler ops indexes based on the dynamic load
order of mempool handlers. Indexes are used so a mempool can be used by
multiple processes, but this only works if all processes agree on the
mapping from index to mempool handler.
When using the '-d' argument, it's possible for different processes to load
mempool handlers in different orders, and thus have different
index->handler mappings. Using a mempool in multiple of such processes will
result in undefined behavior.
This commit adds a note to the mempool library programmer's guide warning
users against this.
Fixes: 449c49b93a6b ("mempool: support handler operations") Cc: stable@dpdk.org Signed-off-by: Gage Eads <gage.eads@intel.com> Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
Suppress the unaligned packed member address warnings by extending
the telemetry library build flags with -Wno-address-of-packed-member
option, through the WERROR_FLAGS makefile variable.
With this change additional warnings are turned on to be treated as errors,
which causes the following build issues to be seen:
- no previous prototype [-Werror=missing-prototypes]
- initialization discards ‘const’ qualifier from pointer target type
[-Werror=discarded-qualifiers]
- old-style function definition [-Werror=old-style-definition]
- variable may be used before its value is set (when using icc compiler).
Fixes: 0fe3a37924d4 ("telemetry: format json response when sending stats") Fixes: ee5ff0d3297e ("telemetry: add client feature and sockets") Fixes: 8877ac688b52 ("telemetry: introduce infrastructure") Fixes: 1b756087db93 ("telemetry: add parser for client socket messages") Fixes: fff6df7bf58e ("telemetry: fix using ports of different types") Fixes: 4080e46c8078 ("telemetry: support global metrics") Cc: stable@dpdk.org Signed-off-by: Reshma Pattan <reshma.pattan@intel.com> Signed-off-by: Flavia Musatescu <flavia.musatescu@intel.com> Acked-by: Kevin Laatz <kevin.laatz@intel.com>
event/octeontx2: add libatomic dependency for 32-bit clang
When compiling with clang on 32-bit platforms, we are missing
copies of 64-bit atomic functions. We can solve this by linking
against libatomic for the drivers and libs which need those
atomic ops.
The vlan_insert() is buggy when it tries to handle the shared mbufs,
instead don't support inserting VLAN tag into shared mbufs and return
an error for that case.
Andy Pei [Wed, 22 May 2019 06:36:34 +0000 (14:36 +0800)]
raw/ifpga: fix unintentional integer overflow
cast unsigned int ports_per_retimer, unsigned int nums_retimer,
unsigned int nums_fvl and unsigned int ports_per_fvl to uint64_t
before multiply operation, to avoid Unintentional integer overflow.
Coverity issue: 337924, 337926 Fixes: d1cd4eb2d48e ("raw/ifpga: support ipn3ke") Cc: stable@dpdk.org Signed-off-by: Andy Pei <andy.pei@intel.com> Acked-by: Rosen Xu <rosen.xu@intel.com>
Tianfei Zhang [Fri, 21 Jun 2019 08:40:17 +0000 (16:40 +0800)]
raw/ifpga/base: fix retimer link status
Fix the readout retimer link status incorrectly when we
remove the linux intel-fpga-driver and run the DPDK application.
The linux driver will stop the retimer when remove the kernel
modules.
eval_call() blindly calls eval_max_bound() for external function
return value for all return types.
That causes wrong estimation for returned pointer min and max boundaries.
So any attempt to dereference that pointer value causes verifier to fail
with error message: "memory boundary violation at pc: ...".
To fix - estimate min/max boundaries based on the return value type.
Bugzilla ID: 298
Fixes: 8021917293d0 ("bpf: add extra validation for input BPF program") Cc: stable@dpdk.org Reported-by: Michel Machado <michel@digirati.com.br> Suggested-by: Michel Machado <michel@digirati.com.br> Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Add local APIs to trigger data copies, and retrieve handle values once
those copies are completed. Included are unit tests to validate the data
is copies correctly.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com> Acked-by: Jiayu Hu <jiayu.hu@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
Add stats functions to track what is happening in the driver, and put
unit tests to check those.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com> Acked-by: Jiayu Hu <jiayu.hu@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
Allow initializing a driver instance. Include selftest to validate these
functions.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com> Acked-by: Jiayu Hu <jiayu.hu@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
Add in the "info_get" function to the driver, to allow us to query the
device. This allows us to have the unit test pick up the presence of
supported hardware or not.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com> Acked-by: Jiayu Hu <jiayu.hu@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
raw/ioat: create device on probe and destroy on release
Add the create/destroy driver functions so that we can actually allocate
a rawdev and destroy it when done. No rawdev API functions are actually
implemented at this point.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com> Acked-by: Jiayu Hu <jiayu.hu@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
In order to allow binding/unbinding of devices for use by the
ioat_rawdev, we need to update the devbind script to add a new class
of device, and add device ids for the specific HW instances.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com> Acked-by: Jiayu Hu <jiayu.hu@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
Add stubs for ioat rawdev driver support in DPDK, specifically:
* makefile and meson build hooks
* initial public header file
* rawdev main C file, with probe and release functions
* release note update announcing the driver
* initial documentation for the new section in the rawdev doc
* unit test stubs for device unit tests
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com> Acked-by: Jiayu Hu <jiayu.hu@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
rawdev: allow devices to skip extra memory allocation
Some device drivers want to allocate their own private memory, and should
be allowed to do so. Therefore skip memory allocation and associated error
checks if zero-length private memory is requested.
While adjusting the code for new indent level, fix incorrect error
message.
With Debian and Ubuntu, the default installation path for the 64-bit
libraries is set to e.g. /usr/local/lib/x86_64-linux-gnu/, compared to
/usr/local/lib64 on Fedora and Redhat distributions. This causes issues
when using "pkg-config --define-prefix" since pkg-config assumes the prefix
to be the grandparent of where the .pc file is. On Ubuntu we then get the
cflags include path as being "/path/to/install-root/usr/local/lib/include"
i.e. with an extra "lib" in the path.
This issue only applies for test installs on Ubuntu and similar distros,
and is not a problem for regular installs since the --define-prefix
parameter would not be passed to pkg-config in those cases.
The workaround for this in our test build script is to explicitly make
"lib" the "libdir" setting for the install, overriding the distro-provided
default.
Fixes: 7f80a2102bbb ("devtools: test pkg-config file") Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Luca Boccassi <bluca@debian.org>
Wei Zhao [Mon, 10 Jun 2019 06:45:10 +0000 (14:45 +0800)]
app/testpmd: fix offloads config
There is no need to use default offloads configuration
if offloads configuration has been pass down from upper layer.
The default offloads are overwritten if not zero.
Fixes: 5e91aeef218c ("app/testpmd: fix offload flags after port config") Cc: stable@dpdk.org Signed-off-by: Wei Zhao <wei.zhao1@intel.com> Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
For tx checksum offload it is not mandatory to provide
a valid packet_type in addition to a valid ol_flags.
This patch only use ol_flags information for this feature
as a performance improvement.
The patch enables the backend of rte_flow. It transfers
rte_flow_xxx to device specific data structure and
configures packet process engine's binary classifier
(switch) properly.
Leyi Rong [Mon, 1 Jul 2019 05:04:49 +0000 (13:04 +0800)]
net/ice/base: support ethertype switch filters
Add protocol definitions for ethertype.
Add ice_ethtype_hdr structure definition into the ice_prot_hdr union.
Add ethtype offsets into the training packet maps.
Signed-off-by: Dan Nowlin <dan.nowlin@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Leyi Rong <leyi.rong@intel.com> Tested-by: Wei Zhao <wei.zhao1@intel.com> Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Leyi Rong [Mon, 1 Jul 2019 05:04:48 +0000 (13:04 +0800)]
net/ice/base: support IPv6 based switch filters
- Add IPv6 switch rule support.
- Add IPv6 training packet.
- Correct name of IPv6 header variable.
- Fix enum values so that they point to the proper
ice_prot_ext_tbl_entry field.
Signed-off-by: Kevin Scott <kevin.c.scott@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Leyi Rong <leyi.rong@intel.com> Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Dan Wei [Sat, 22 Jun 2019 14:25:21 +0000 (10:25 -0400)]
net/ipn3ke: modify AFU configurations
Modify AFU configurations for new BBS (Blue Bitstream) of A10 on N3000
card:
- AFU register access: RTL changes the UPL (User Programmable Logic
which is the container of vBNG IP) base address and the read/write
commands of register indirect access.
- Poll the INIT_STS register to wait for the vBNG IP and DDR reset
completion.
- Refine log for debug: print UPL_version not only for vBNG bit stream,
but also for other bit streams.
Fixes: c01c748e4ae6 ("net/ipn3ke: add new driver") Cc: stable@dpdk.org Signed-off-by: Dan Wei <dan.wei@intel.com> Acked-by: Rosen Xu <rosen.xu@intel.com>
Pavan Nikhilesh [Fri, 28 Jun 2019 18:23:53 +0000 (23:53 +0530)]
event/octeontx2: add devargs to control timer adapters
Add devargs to control each event timer adapter i.e. TIM rings internal
parameters uniquely. The following dict format is expected
[ring-chnk_slots-disable_npa-stats_ena]. 0 represents default values.
Pavan Nikhilesh [Fri, 28 Jun 2019 18:23:52 +0000 (23:53 +0530)]
event/octeontx2: add devargs to limit timer adapters
Add devargs to limit the max number of TIM rings reserved on probe.
Since, TIM rings are HW resources we can avoid starving other
applications by not grabbing all the rings.
Pavan Nikhilesh [Fri, 28 Jun 2019 18:23:42 +0000 (23:53 +0530)]
event/octeontx2: add devargs to modify chunk slots
Add devargs support to modify number of chunk slots. Chunks are used to
store event timers, a chunk can be visualised as an array where the last
element points to the next chunk and rest of them are used to store
events. TIM traverses the list of chunks and enqueues the event timers
to SSO.
If no argument is passed then a default value of 255 is taken.
Pavan Nikhilesh [Fri, 28 Jun 2019 18:23:41 +0000 (23:53 +0530)]
event/octeontx2: add devargs to disable NPA
If the chunks are allocated from NPA then TIM can automatically free
them when traversing the list of chunks.
Add devargs to disable NPA and use software mempool to manage chunks.
Pavan Nikhilesh [Fri, 28 Jun 2019 18:23:39 +0000 (23:53 +0530)]
event/octeontx2: create and free timer adapter
When the application calls timer adapter create the following is used:
- Allocate a TIM lf based on number of lf's provisioned.
- Verify the config parameters supplied.
- Allocate memory required for
* Buckets based on min and max timeout supplied.
* Allocate the chunk pool based on the number of timers.
On Free:
- Free the allocated bucket and chunk memory.
- Free the TIM lf allocated.
Pavan Nikhilesh [Fri, 28 Jun 2019 18:23:34 +0000 (23:53 +0530)]
event/octeontx2: add devargs to control SSO GGRP QoS
SSO GGRPs i.e. queue uses DRAM & SRAM buffers to hold in-flight
events. By default the buffers are assigned to the SSO GGRPs to
satisfy minimum HW requirements. SSO is free to assign the remaining
buffers to GGRPs based on a preconfigured threshold.
We can control the QoS of SSO GGRP by modifying the above mentioned
thresholds. GGRPs that have higher importance can be assigned higher
thresholds than the rest.
Pavan Nikhilesh [Fri, 28 Jun 2019 18:23:32 +0000 (23:53 +0530)]
event/octeontx2: add devargs to force legacy mode
Octeontx2 SSO by default is set to use dual workslot mode.
Add devargs option to force legacy mode i.e. single workslot mode.
Example:
--dev "0002:0e:00.0,single_ws=1"
Pavan Nikhilesh [Fri, 28 Jun 2019 18:23:28 +0000 (23:53 +0530)]
event/octeontx2: add SSO dual workslot mode
OcteonTx2 AP core SSO cache contains two entries each entry caches
state of an single GWS aka event port.
AP core requests events from SSO by using following sequence :
1. Write to SSOW_LF_GWS_OP_GET_WORK
2. Wait for SSO to complete scheduling by polling on SSOW_LF_GWS_TAG[63]
3. SSO notifies core by clearing SSOW_LF_GWS_TAG[63] and if work is
valid SSOW_LF_GWS_WQP is non-zero.
The above sequence uses only one in-core cache entry.
In dual workslot mode we try to use both the in-core cache entries by
triggering GET_WORK on a second workslot as soon as the above sequence
completes. This effectively hides the schedule latency of SSO if there
are enough events with unique flow_tags in-flight.
This mode reserves two SSO GWS lf's for each event port effectively
doubling single core performance.
Dual workslot mode is the default mode of operation in octeontx2.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com> Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Pavan Nikhilesh [Fri, 28 Jun 2019 18:23:18 +0000 (23:53 +0530)]
event/octeontx2: add devargs for inflight buffer count
The number of events for a *open system* event device is specified
as -1 as per the eventdev specification.
Since, Octeontx2 SSO inflight events are only limited by DRAM size, the
xae_cnt devargs parameter is introduced to provide upper limit for
in-flight events.
Example:
--dev "0002:0e:00.0,xae_cnt=8192"
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com> Acked-by: Jerin Jacob <jerinj@marvell.com>
git.droids-corp.org / dpdk.git / log