Jerin Jacob [Tue, 6 Apr 2021 14:40:55 +0000 (20:10 +0530)]
common/cnxk: add model init and IO handling API
Add routines for SoC model identification and HW IO handling
routines specific to CN9K and CN10K Marvell SoC's.
These are based on arm64 ISA and behaviour specific to
Marvell SoC's.
Juraj Linkeš [Tue, 30 Mar 2021 06:40:19 +0000 (08:40 +0200)]
build: alias default build as generic
The current machine='default' build name is not descriptive. The actual
default build is machine='native'. Add an alternative string which does
the same build and better describes what we're building:
machine='generic'. Leave machine='default' for backwards compatibility.
Signed-off-by: Juraj Linkeš <juraj.linkes@pantheon.tech> Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Bruce Richardson [Thu, 25 Feb 2021 15:29:02 +0000 (15:29 +0000)]
build: hide debug messages in non-developer mode
The messages about what components have what dependency names, and
information about function versioning not being supported on windows are
only of interest to developers, so hide them when building in
non-developer mode.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Bruce Richardson [Thu, 25 Feb 2021 15:29:01 +0000 (15:29 +0000)]
build: enable a developer mode setting
To allow support for additional build checks and tests only really
relevant for developers, we add support for a developer mode option to
DPDK. The default, "auto", value for this enables developer mode if a
".git" folder is found at the root of the source tree - as was the case
with the previous "make" build system. There is also support for
explicitly enabling or disabling this option using "meson configure" if
so desired.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
The affinity of a control thread is set after it has been launched. If
setting the affinity fails, pthread_cancel is called followed by a call
to pthread_join, which can hang forever if the thread's start routine
doesn't call a pthread cancellation point.
This patch modifies the logic so that the control thread exits
gracefully if the affinity cannot be set successfully and removes the
call to pthread_cancel.
Fixes: 6383d2642b62 ("eal: set name when creating a control thread") Cc: stable@dpdk.org Signed-off-by: Luc Pelletier <lucp.at.work@gmail.com> Acked-by: Olivier Matz <olivier.matz@6wind.com> Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
The creation of control threads uses a pthread barrier for
synchronization. This patch fixes a race condition where the pthread
barrier could get destroyed while one of the threads has not yet
returned from the pthread_barrier_wait function, which could result in
undefined behaviour.
Thomas Monjalon [Mon, 8 Mar 2021 22:24:10 +0000 (23:24 +0100)]
vfio: reformat logs
The log messages had various issues:
- split on 2 lines, making search (grep) difficult
- long lines (can be split after the string)
- indented for no good reason (parent message may have higher log level)
- inconsistent use of __func__, not meaningful context for user
- lack of context (general message not mentioning VFIO)
- log level too high (more below)
Message having its level decreased from WARNING to NOTICE:
"not managed by VFIO driver, skipping"
Message having its level decreased from INFO to DEBUG:
"Probing VFIO support..."
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Thomas Monjalon [Mon, 5 Apr 2021 19:33:25 +0000 (21:33 +0200)]
app/testpmd: fix usage text
The options help text was including an incomplete and redundant
summary of the options before explaining each. The summary is dropped.
The details of the option --hairpin-mode had an extra space,
breaking the alignment with the next line.
There were some mismatches between options in the usage text
sed -rn 's/.*\(" *--([a-z-]*)[=: ].*/\1/p' app/test-pmd/parameters.c
and the options declared in lgopts array
sed -rn 's/.*\{.*"(.*)",.*,.*,.*},.*/\1/p' app/test-pmd/parameters.c
The misses were:
--no-numa
--enable-scatter
--tx-ip
--tx-udp
--noisy-lkup-num-reads-writes
The option --ports was not implemented.
Fixes: 01817b10d27c ("app/testpmd: change hairpin queues setup") Fixes: 3c156061b938 ("app/testpmd: add noisy neighbour forwarding mode") Fixes: bf5b2126bf44 ("app/testpmd: add ability to set Tx IP and UDP parameters") Fixes: 0499793854f5 ("app/testpmd: add scatter enabling option") Fixes: 999b2ee0fe45 ("app/testpmd: enable NUMA support by default") Fixes: af75078fece3 ("first public release") Cc: stable@dpdk.org Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru> Acked-by: Xiaoyun Li <xiaoyun.li@intel.com> Reviewed-by: Jens Freimann <jfreimann@redhat.com> Reviewed-by: Bing Zhao <bingz@nvidia.com> Acked-by: David Marchand <david.marchand@redhat.com>
Thomas Monjalon [Mon, 5 Apr 2021 19:33:24 +0000 (21:33 +0200)]
app/regex: fix usage text
The usage syntax help includes the program name which was fake.
It is replaced with the real name from argv.
Fixes: de06137cb295 ("app/regex: add RegEx test application") Cc: stable@dpdk.org Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Ori Kam <orika@nvidia.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru> Acked-by: David Marchand <david.marchand@redhat.com>
Thomas Monjalon [Mon, 5 Apr 2021 19:33:23 +0000 (21:33 +0200)]
app: fix exit messages
Some applications were printing useless messages with rte_exit()
after showing the help. Using exit() is enough in this case.
Some applications were using a redundant printf or fprintf() before
calling rte_exit(). The messages are unified in a single rte_exit().
Some rte_exit() calls were missing a line feed or returning a wrong code.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Wisam Jaddo <wisamm@nvidia.com> Acked-by: Ori Kam <orika@nvidia.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru> Acked-by: David Marchand <david.marchand@redhat.com>
EAL: lib.eal log level changed from info to debug
EAL: lib.ethdev log level changed from info to debug
EAL: lib.ethdev log level changed from debug to info
EAL: lib.ethdev log level changed from info to debug
EAL: lib.ethdev log level changed from debug to info
EAL: lib.telemetry log level changed from disabled to warning
EAL: lib.eal log level changed from info to debug
EAL: lib.ethdev log level changed from info to debug
EAL: lib.ethdev log level changed from debug to info
EAL: lib.telemetry log level changed from disabled to warning
Fixes: 6c7216eefd63 ("eal: fix log level of early messages") Fixes: 1c806ae5c3ac ("eal/windows: support command line options parsing") Cc: stable@dpdk.org Signed-off-by: David Marchand <david.marchand@redhat.com> Acked-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Lukasz Wojciechowski <l.wojciechow@partner.samsung.com> Tested-by: Lukasz Wojciechowski <l.wojciechow@partner.samsung.com>
David Marchand [Fri, 9 Apr 2021 11:04:52 +0000 (13:04 +0200)]
log: track log level changes
Add a log message when registering log types and changing log levels.
__rte_log_register previous handled both legacy and dynamic logtypes.
To simplify the code, __rte_log_register is reworked to only handle
dynamic logtypes and takes a log level.
Example:
$ DPDK_TEST=logs_autotest ./build/app/test/dpdk-test --no-huge -m 512 \
--log-level=lib.eal:debug
...
RTE>>logs_autotest
== dynamic log types
EAL: logtype1 log level changed from disabled to info
EAL: logtype2 log level changed from disabled to info
EAL: logtype1 log level changed from info to error
EAL: logtype3 log level changed from error to emergency
EAL: logtype2 log level changed from info to emergency
EAL: logtype3 log level changed from emergency to debug
EAL: logtype1 log level changed from error to debug
EAL: logtype2 log level changed from emergency to debug
error message
critical message
critical message
error message
== static log types
TESTAPP1: error message
TESTAPP1: critical message
TESTAPP2: critical message
TESTAPP1: error message
Test OK
Suggested-by: Ilya Maximets <i.maximets@ovn.org> Signed-off-by: David Marchand <david.marchand@redhat.com> Acked-by: Andrew Rybchenko <arybchenko@solarflare.com> Acked-by: Lukasz Wojciechowski <l.wojciechow@partner.samsung.com> Acked-by: Thomas Monjalon <thomas@monjalon.net>
Thomas Monjalon [Thu, 8 Apr 2021 16:47:13 +0000 (18:47 +0200)]
log: add option argument help
The option --log-level was not completely described in the usage text,
and it was difficult to guess the names of the log types and levels.
A new value "help" is accepted after --log-level to give more details
about the syntax and listing the log types and levels.
The array "levels" used for level name parsing is replaced with
a (modified) existing function which was used in rte_log_dump().
The new function rte_log_list_types() is exported in the API
for allowing an application to give this info to the user
if not exposing the EAL option --log-level.
The list of log types cannot include all drivers if not linked in the
application (shared object plugin case).
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru> Reviewed-by: David Marchand <david.marchand@redhat.com>
Thomas Monjalon [Thu, 8 Apr 2021 16:47:12 +0000 (18:47 +0200)]
log: catch invalid level option number
The parsing check for invalid log level was not trying to catch
irrelevant numeric values.
A log level 0 becomes a failure in parsing so it can be caught early.
A log level higher than the max (8) is accepted with a warning message.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru> Reviewed-by: David Marchand <david.marchand@redhat.com>
Thomas Monjalon [Thu, 8 Apr 2021 16:47:11 +0000 (18:47 +0200)]
log: introduce macro for maximum level
RTE_DIM(...) and RTE_LOG_DEBUG were used to get the highest log level.
For better clarity a new constant RTE_LOG_MAX is introduced
and mapped to RTE_LOG_DEBUG.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru> Reviewed-by: David Marchand <david.marchand@redhat.com>
Thomas Monjalon [Thu, 8 Apr 2021 16:47:10 +0000 (18:47 +0200)]
log: move private functions
Some private log functions had a wrong "rte_" prefix.
All private log functions are moved from eal_private.h
to the new file eal_log.h:
rte_eal_log_init -> eal_log_init
rte_log_save_regexp -> eal_log_save_regexp
rte_log_save_pattern -> eal_log_save_pattern
eal_log_set_default
The static functions in the file eal_common_log.c are renamed:
rte_log_save_level -> log_save_level
rte_log_lookup -> log_lookup
rte_log_init -> log_init
__rte_log_register -> log_register
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru> Reviewed-by: David Marchand <david.marchand@redhat.com>
test: proceed if timer subsystem already initialized
rte_timer_subsystem_init() may return -EALREADY if the timer subsystem
was already initialized. This can happen i.e. in PMD code (see
eth_ena_dev_init). This is not an error, rather a notification as the
initialization function simply returns without any action taken.
Fixes: 50247fe03fe0 ("test/timer: exercise new APIs in secondary process") Cc: stable@dpdk.org Signed-off-by: Stanislaw Kardach <kda@semihalf.com> Reviewed-by: Michal Krawczyk <mk@semihalf.com>
timer: clarify error if subsystem already initialized
rte_timer_subsystem_init() may return -EALREADY if it has been already
initialized. Therefore put explicitly into doxygen that this is not a
failure for the application.
Signed-off-by: Stanislaw Kardach <kda@semihalf.com> Reviewed-by: Michal Krawczyk <mk@semihalf.com> Acked-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com>
UMR (User-Mode Memory Registration) WQE can present data buffers
scattered within multiple mbufs with single indirect mkey. Take
advantage of the UMR WQE, scattered mbuf in one operation can be
presented to an indirect mkey. The RegEx which only accepts one
mkey can now process the whole scattered mbuf in one operation.
The maximum scattered mbuf can be supported in one UMR WQE is now
defined as 64. The mbufs from multiple operations can be combined
into one UMR WQE as well if there is enough space in the KLM array,
since the operations can address their own mbuf's content by the
mkey's address and length. However, one operation's scattered mbuf's
can't be placed in two different UMR WQE's KLM array, if the UMR
WQE's KLM does not has enough free space for one operation, the
extra UMR WQE will be engaged.
In case the UMR WQE's indirect mkey will be over wrapped by the SQ's
WQE move, the mkey's index used by the UMR WQE should be the index
of last the RegEX WQE in the operations. As one operation consumes
one WQE set, build the RegEx WQE by reverse helps address the mkey
more efficiently. Once the operations in one burst consumes multiple
mkeys, when the mkey KLM array is full, the reverse WQE set index
will always be the last of the new mkey's for the new UMR WQE.
In GGA mode, the SQ WQE's memory layout becomes UMR/NOP and RegEx
WQE by interleave. The UMR and RegEx WQE can be called as WQE set.
The SQ's pi and ci will also be increased as WQE set not as WQE.
For operations don't have scattered mbuf, uses the mbuf's mkey directly,
the WQE set combination is NOP + RegEx.
For operations have scattered mbuf but share the UMR WQE with others,
the WQE set combination is NOP + RegEx.
For operations complete the UMR WQE, the WQE set combination is UMR +
RegEx.
Signed-off-by: John Hurley <jhurley@nvidia.com> Signed-off-by: Suanming Mou <suanmingm@nvidia.com> Acked-by: Ori Kam <orika@nvidia.com>
Thomas Monjalon [Tue, 6 Apr 2021 13:22:04 +0000 (15:22 +0200)]
drivers: align log names
The log levels are configured by using the name of the logs.
Some drivers are aligned to follow a common log name standard:
pmd.class.driver[.sub]
Some "common" drivers skip the "class" part:
pmd.driver.sub
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Rosen Xu <rosen.xu@intel.com> Acked-by: Xiao Wang <xiao.w.wang@intel.com> Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com> Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Acked-by: Min Hu (Connor) <humin29@huawei.com>
Thomas Monjalon [Tue, 6 Apr 2021 13:22:03 +0000 (15:22 +0200)]
drivers: fix log level after loading
When compiled as a shared object, and loaded at runtime as a plugin,
the drivers should get the log level set earlier at EAL init
by the user through --log-level option.
The function for applying the log level setting is
rte_log_register_type_and_pick_level().
It is called by most drivers via RTE_LOG_REGISTER().
The drivers common/mlx5, bcmfs and e1000 were missing,
so the user-specified log level was not applied when
those drivers were loaded as plugins.
The macro RTE_LOG_REGISTER() is used for those drivers.
The unnecessary protection for double registration
is removed from e1000.
David Marchand [Tue, 6 Apr 2021 09:25:45 +0000 (11:25 +0200)]
eal: fix telemetry log type on registration failure
rte_log_register_type_and_pick_level() returns an int.
Casting to a uin32_t will make us miss the -1 passed in case of failure.
Fallback to EAL log type like RTE_LOG_REGISTER.
Fixes: 37b881a96194 ("telemetry: use log function from pointer") Signed-off-by: David Marchand <david.marchand@redhat.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Thomas Monjalon [Tue, 6 Apr 2021 13:22:02 +0000 (15:22 +0200)]
log: choose EAL log type on registration failure
In the unlikely case where something goes wrong
while registering a log type,
the fallback is to use the EAL log type.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Reviewed-by: David Marchand <david.marchand@redhat.com> Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
David Marchand [Tue, 6 Apr 2021 17:59:10 +0000 (19:59 +0200)]
build: remove Windows export symbol list
Rather than have two files that keeps getting out of sync, let's
annotate the version.map to generate the Windows export file.
Some mlx5 symbols (haswell_broadwell_cpu, mlx5_glue, mlx5_os_*) were
only exported for Windows.
All of them are available and used by Linux too, so this patch adds
them in version.map.
Note: Existing version.map annotation achieved with:
$ for dir in lib/librte_eal drivers/common/mlx5; do
./buildtools/map-list-symbol.sh $dir/*.map |
while read file version sym; do
! git grep -qw $sym $dir/*.def || continue;
sed -i -e "s/$sym;/$sym; # WINDOWS_NO_EXPORT/" $dir/*.map;
done;
done
Signed-off-by: David Marchand <david.marchand@redhat.com> Tested-by: Tal Shnaiderman <talshn@nvidia.com> Acked-by: Thomas Monjalon <thomas@monjalon.net>
David Marchand [Wed, 7 Apr 2021 09:06:56 +0000 (11:06 +0200)]
service: clean references to removed symbol
rte_service_get_id() was removed in v17.11 but the API description
still referenced it and a version node was still present in EAL map.
Fixes: 8edc9aaaf217 ("service: use id in get by name function") Cc: stable@dpdk.org Signed-off-by: David Marchand <david.marchand@redhat.com> Reviewed-by: Harry van Haaren <harry.van.haaren@intel.com> Acked-by: Thomas Monjalon <thomas@monjalon.net>
Avoid race with unregister interrupt handler if interrupt
source has some active callbacks at the moment, use wrapper
around rte_intr_callback_unregister() to check for -EAGAIN
return value and to loop until rte_intr_callback_unregister()
succeeds.
Roy Shterman [Mon, 22 Feb 2021 10:41:31 +0000 (12:41 +0200)]
mem: fix freeing segments in --huge-unlink mode
When using huge_unlink we unlink the segment right
after allocation. Although we unlink the file we keep
the fd in fd_list so file still exist just the path deleted.
When freeing the hugepage we need to close the fd and assign
it with (-1) in fd_list for the page to be released.
The current flow fails rte_malloc in the following flow when working
with --huge-unlink option:
1. alloc_seg() for segment A -
We allocate a segment, unlink the path to the segment
and keep the file descriptor in fd_list.
2. free_seg() for segment A -
We clear the segment metadata and return - without closing fd
or assigning (-1) in fd list.
3. alloc_seg() for segment A again -
We find segment A as available, try to allocate it,
find the old fd in fd_list try to unlink it
as part of alloc_seg() but failed because path doesn't exist.
The impact of such error is falsely failing rte_malloc()
although we have hugepages available.
Fixes: d435aad37da7 ("mem: support --huge-unlink mode") Cc: stable@dpdk.org Signed-off-by: Roy Shterman <roy.shterman@vastdata.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Thomas Monjalon [Mon, 5 Apr 2021 09:15:05 +0000 (11:15 +0200)]
bus/pci: rename probe/remove operation types
The names of the prototypes pci_probe_t and pci_remove_t
are missing a prefix rte_.
These function types are simply renamed.
No compatibility break is expected for the applications
because it is considered as an internal name in the driver interface.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Reviewed-by: Xueming Li <xuemingl@nvidia.com> Reviewed-by: David Marchand <david.marchand@redhat.com>
power: do not skip saving original P-state governor
Currently, when we set the pstate governor to "performance", we check if
it is already set to this value, and if it is, we skip setting it.
However, we never save this value anywhere, so that next time we come
back and request the governor to be set to its original value, the
original value is empty.
Fix it by saving the original pstate governor first. While we're at it,
replace `strlcpy` with `rte_strscpy`.
Previous fix for base frequency handling in pstate mode introduced a
couple of issues:
- When base_frequency file does not exist, it simply bails out because
of what appears to be accidental addition of FOPEN_OR_ERR_RET. This is
incorrect, as absence of this file is not fatal and is in fact
expected on kernel versions earlier than 5.3
- When base_frequency file does exist, it gets opened, but never gets
closed, resulting in a resource leak
Both issues also manifest themselves as Coverity defects (dead code, and
a resource leak), so this fix addresses both.
David Marchand [Thu, 1 Apr 2021 19:58:42 +0000 (21:58 +0200)]
doc: fix sphinx rtd theme import in GHA
If the rtd theme is available, passing it by name is enough to select
it. Sphinx itself recognises the "sphinx_rtd_theme" name as a special
case and tries to find its path automatically.
On the other hand, passing a html_theme_path makes sphinx parse all
themes availables in this path, which in some environment (like GHA) is
/usr/share and makes sphinx error on the first zipfile it finds (in GHA,
some Azure CLI thingy) that has no sphinx theme in it.
Fixes: 46562be65094 ("doc: import sphinx rtd theme when available") Cc: stable@dpdk.org Signed-off-by: David Marchand <david.marchand@redhat.com> Acked-by: Aaron Conole <aconole@redhat.com>
Matan Azrad [Mon, 1 Mar 2021 10:41:31 +0000 (10:41 +0000)]
vdpa/mlx5: fix virtq cleaning
The HW virtq object can be destroyed either when the device is closed or
when the state of the virtq becomes disabled.
Some parameters of the virtq should continue to be managed when the
virtq state is changed but all of them must be initialized when the
device is closed.
Wrongly, the enable parameter stayed on when the device is closed what
might cause creation of invalid virtq in the next time a device is
assigned to the driver.
Clean all the virtqs memory when the device is closed.
Fixes: c47d6e83334e ("vdpa/mlx5: support queue update") Cc: stable@dpdk.org Signed-off-by: Matan Azrad <matan@nvidia.com> Acked-by: Xueming Li <xuemingl@nvidia.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
David Marchand [Mon, 1 Feb 2021 17:46:01 +0000 (18:46 +0100)]
net/virtio: remove duplicated port ID from virtio-user
The private virtio_user_dev structure embeds a virtio_hw which itself
contains the ethdev port_id.
Make use of it and remove the duplicate port_id field.
Signed-off-by: David Marchand <david.marchand@redhat.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Similar as single dequeue, the multiple accesses of descriptor length
will lead to potential risk. One-time access of descriptor length can
eliminate this risk.
Fixes: 75ed51697820 ("vhost: add packed ring batch dequeue") Cc: stable@dpdk.org Signed-off-by: Marvin Liu <yong.liu@intel.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Marvin Liu [Wed, 31 Mar 2021 06:49:38 +0000 (14:49 +0800)]
vhost: fix packed ring potential buffer overflow
Similar as split ring, the multiple accesses of descriptor length will
lead to potential risk. One-time access of descriptor length can
eliminate this risk.
Fixes: 2f3225a7d69b ("vhost: add vector filling support for packed ring") Cc: stable@dpdk.org Signed-off-by: Marvin Liu <yong.liu@intel.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Marvin Liu [Wed, 31 Mar 2021 06:49:37 +0000 (14:49 +0800)]
vhost: fix split ring potential buffer overflow
In vhost datapath, descriptor's length are mostly used in two coherent
operations. First step is used for address translation, second step is
used for memory transaction from guest to host. But the interval between
two steps will give a window for malicious guest, in which can change
descriptor length after vhost calculated buffer size. Thus may lead to
buffer overflow in vhost side. This potential risk can be eliminated by
accessing the descriptor length once.
Fixes: 1be4ebb1c464 ("vhost: support indirect descriptor in mergeable Rx") Cc: stable@dpdk.org Signed-off-by: Marvin Liu <yong.liu@intel.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Maxime Coquelin [Tue, 23 Mar 2021 09:02:18 +0000 (10:02 +0100)]
vhost: move dirty logging cache out of virtqueue
This patch moves the per-virtqueue's dirty logging cache
out of the virtqueue struct, by allocating it dynamically
only when live-migration is enabled.
It saves 8 cachelines in vhost_virtqueue struct.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com> Reviewed-by: Chenbo Xia <chenbo.xia@intel.com> Reviewed-by: David Marchand <david.marchand@redhat.com> Tested-by: Balazs Nemeth <bnemeth@redhat.com>
Maxime Coquelin [Tue, 16 Mar 2021 09:38:24 +0000 (10:38 +0100)]
net/virtio: allocate fake mbuf in Rx queue
While it is worth clarifying whether the fake mbuf
in virtnet_rx struct is really necessary, it is sure
that it heavily impacts cache usage by being part of
the struct. Indeed, it uses two cachelines, and
requires alignment on a cacheline.
Before this series, it means it took 120 bytes in
virtnet_rx struct:
Balazs Nemeth [Fri, 26 Mar 2021 11:01:30 +0000 (12:01 +0100)]
net/qede: remove unnecessary field in Rx entry and simplify
The member page_offset is always zero. Having this in the qede_rx_entry
makes it larger than it needs to be and this has cache performance
implications so remove that field. In addition, since qede_rx_entry only
has an rte_mbuf*, remove the definition of qede_rx_entry.
Signed-off-by: Balazs Nemeth <bnemeth@redhat.com> Reviewed-by: Igor Russkikh <irusskikh@marvell.com>
Balazs Nemeth [Fri, 26 Mar 2021 11:01:29 +0000 (12:01 +0100)]
net/qede: prefetch next packet to free
While handling the current mbuf, pull the next mbuf into the cache. Note
that the last four mbufs pulled into the cache are not handled, but that
doesn't matter.
Signed-off-by: Balazs Nemeth <bnemeth@redhat.com> Reviewed-by: Igor Russkikh <irusskikh@marvell.com>
Balazs Nemeth [Fri, 26 Mar 2021 11:01:27 +0000 (12:01 +0100)]
net/qede: free packets in bulk
rte_pktmbuf_free_bulk calls rte_mempool_put_bulk with the number of
pending packets to return to the mempool. In contrast, rte_pktmbuf_free
calls rte_mempool_put that calls rte_mempool_put_bulk with one object.
An important performance related downside of adding one packet at a time
to the mempool is that on each call, the per-core cache pointer needs to
be read from tls while a single rte_mempool_put_bulk only reads from the
tls once.
Signed-off-by: Balazs Nemeth <bnemeth@redhat.com> Reviewed-by: Igor Russkikh <irusskikh@marvell.com>
Balazs Nemeth [Fri, 26 Mar 2021 11:01:26 +0000 (12:01 +0100)]
net/qede: assume mbuf to free is never null
The ring txq->sw_tx_ring is managed with txq->sw_tx_cons. As long as
txq->sw_tx_cons is correct, there is no need to check if
txq->sw_tx_ring[idx] is null explicitly.
Signed-off-by: Balazs Nemeth <bnemeth@redhat.com> Reviewed-by: Igor Russkikh <irusskikh@marvell.com>
Leaving the unused flags member here has a few performance implications.
First, each qede_tx_entry takes up more memory which has caching
implications as less entries fit in a cache line while multiple entries
are frequently handled in batches. Second, an array of qede_tx_entry
entries is incompatible with existing APIs that expect an array of
rte_mbuf pointers. Consequently, an extra array need to be allocated
before calling such APIs and each entry needs to be copied over.
This patch omits the flags field and replaces the qede_tx_entry entry
by a simple rte_mbuf pointer.
Signed-off-by: Balazs Nemeth <bnemeth@redhat.com> Reviewed-by: Igor Russkikh <irusskikh@marvell.com>
Modification of the 802.1Q Tag Identifier, VXLAN Network
Identifier or GENEVE Network Identifier is not supported.
Reject attempt to modify these fields via the MODIFY_FIELD
action and document this mlx5 driver limitation.
Fixes: 641dbe4fb053 ("net/mlx5: support modify field flow action") Cc: stable@dpdk.org Signed-off-by: Alexander Kozyrev <akozyrev@nvidia.com> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
There is a limitation about copying one header field to another for
the Flow group 0. Such copy action is not allowed there. But setting
a header field with an immediate value is perfectly fine.
Allow the MODIFY_FIELD action on group 0 in case the source field
is an immediate value or a pointer to it.
Signed-off-by: Alexander Kozyrev <akozyrev@nvidia.com> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
net/mlx5: check extended metadata for mark modification
The MODIFY_FIELD action requires the extended metadata support
in order to manipulate on MARK register. Check if it is supported
and reject the MODIFY_FIELD action if it is not.
Fixes: 641dbe4fb053 ("net/mlx5: support modify field flow action") Cc: stable@dpdk.org Signed-off-by: Alexander Kozyrev <akozyrev@nvidia.com> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
Xueming Li [Sun, 28 Mar 2021 13:48:13 +0000 (13:48 +0000)]
net/mlx5: fix setting VF default MAC through representor
With kernel bonding, there was an error when setting VF MAC address
through representor. The Netlink API requires ifindex of owner PF, not
bonding device ifindex.
Uses owner PF ifindex to modify VF default MAC in case of bonding
device.
Fixes: c21e5facf7d2 ("net/mlx5: use bond index for netdev operations") Cc: stable@dpdk.org Signed-off-by: Xueming Li <xuemingl@nvidia.com> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
Xueming Li [Sun, 28 Mar 2021 13:48:12 +0000 (13:48 +0000)]
net/mlx5: save bonding member ports information
Since kernel bonding netdev doesn't provide statistics counter that
reflects all member ports, PMD has to manually summarize counters from
each member ports.
As a preparation, this patch collects bonding member port information
and saves to shared context data.
Signed-off-by: Xueming Li <xuemingl@nvidia.com> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
Xueming Li [Sun, 28 Mar 2021 13:48:11 +0000 (13:48 +0000)]
net/mlx5: support list of representor PF
To probe representors from different kernel bonding PFs, had to specify
2 separate devargs like this:
-a 03:00.0,representor=pf0vf[0-3] -a 03:00.0,representor=pf1vf[0-3]
This patch supports range or list of PF section in devargs, so the
alternative short devargs of above is:
-a 03:00.0,representor=pf[0-1]vf[0-3]
Signed-off-by: Xueming Li <xuemingl@nvidia.com> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
Xueming Li [Sun, 28 Mar 2021 13:48:10 +0000 (13:48 +0000)]
net/mlx5: refactor bonding representor probing
To probe representor on 2nd PF of kernel bonding device, had to specify
PF1 BDF in devarg:
<PF1_BDF>,representor=0
When closing bonding device, all representors had to be closed together
and this implies all representors have to use primary PF of bonding
device. So after probing representor port on 2nd PF, when locating new
probed device using device argument, the filter used 2nd PF as PCI
address and failed to locate new device.
Conflict happened by using current representor devargs:
- Use PCI BDF to specify representor owner PF
- Use PCI BDF to locate probed representor device.
- PMD uses primary PCI BDF as PCI device.
To resolve such conflicts, new representor syntax is introduced here:
<primary BDF>,representor=pfXvfY
All representors must use primary PF as owner PCI device, PMD internally
locate owner PCI address by checking representor "pfX" part. To EAL, all
representors are registered to primary PCI device, the 2nd PF is hidden
to EAL, thus all search should be consistent.
Same to VF representor, HPF (host PF on BlueField) uses same syntax to
probe, example: representor=pf1vf[0-3,-1]
This patch also adds pf index into kernel bonding representor port name:
<BDF>_<ib_name>_representor_pf<X>vf<Y>
Signed-off-by: Xueming Li <xuemingl@nvidia.com> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
Xueming Li [Sun, 28 Mar 2021 13:48:09 +0000 (13:48 +0000)]
net/mlx5: revert setting bonding representor to first PF
With kernel bonding, representors on second PF are being probed by
devargs:
<primary_bdf>,representor=pf1vf<N>
No need to save primary PF port ID and lookup when probing sibling
ports, revert patch [1]
[1]:
commit e6818853c022 ("net/mlx5: set representor to first PF in bonding mode")
Signed-off-by: Xueming Li <xuemingl@nvidia.com> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
Xueming Li [Sun, 28 Mar 2021 13:48:07 +0000 (13:48 +0000)]
common/mlx5: support sub-function representor parsing
This patch supports representor name parsing for SF.
In sysfs, representor name stored under "phys_port_name" sysfs key,
similar to VF representor, switch port name of SF representor is
"pf<x>sf<y>".
For netlink message, net SF type is supported.
Examples:
pf0sf1
pf0sf[0-3]
Signed-off-by: Xueming Li <xuemingl@nvidia.com> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
Qi Zhang [Thu, 25 Mar 2021 12:42:41 +0000 (20:42 +0800)]
net/ice: refine RSS configure
The ICE_RSS_ANY_HEADERS will try to enable outer RSS for
non-tunnel case and inner RSS for tunnel case. This confuse
user.
As we already have ICE_RSS_INNER_HEADER for tunnel case,
So, replace ICE_RSS_ANY_HEADERS with ICE_RSS_OUTER_HEADERS
for all exist flow which only specified the outer pattern.
To enable inner RSS for any tunnel cases, a separated rule
should be enabled.
The patch also remove some unnecessary condition check for GTPU
in base code, as we already can support outer RSS for GTPU.
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Xuan Ding <xuan.ding@intel.com>
Murphy Yang [Mon, 29 Mar 2021 08:28:45 +0000 (08:28 +0000)]
net/ixgbe: fix RSS RETA being reset after port start
If one calls ‘rte_eth_dev_rss_reta_update’ with ixgbe before starting
the device (but after setting everything else), then RSS RETA
configuration will be zero after starting the device.
This patch gives a notification if the port not started.
Bugzilla ID: 664 Fixes: 249358424eab ("ixgbe: RSS RETA configuration") Cc: stable@dpdk.org Signed-off-by: Murphy Yang <murphyx.yang@intel.com> Acked-by: Haiyue Wang <haiyue.wang@intel.com>
Qi Zhang [Mon, 1 Mar 2021 07:57:14 +0000 (15:57 +0800)]
net/iavf: fix TSO max segment size
According to Intel® AVF spec
(https://www.intel.com/content/dam/
www/public/us/en/documents/product-specifications/
ethernet-adaptive-virtual-function-hardware-spec.pdf)
section 2.2.2.3:
The max segment size(MSS) of TSO should not be set lower than 88.
Robin Zhang [Fri, 12 Mar 2021 08:52:09 +0000 (08:52 +0000)]
net/i40e: announce request queue capability in PF
A new feature requesting additional queues from PF is added in iavf;
before sending VIRTCHNL_OP_REQUEST_QUEUES op code, the offload
capability flag VIRTCHNL_VF_OFFLOAD_REQ_QUEUES will be checked.
And due to DPDK PF is still used by some cases, add this offload
capability flag in i40e PF.
Fixes: cbdbd360f77f ("net/i40e: support AVF basic interface") Cc: stable@dpdk.org Signed-off-by: Robin Zhang <robinx.zhang@intel.com> Acked-by: Jeff Guo <jia.guo@intel.com>
David Hunt [Thu, 11 Mar 2021 11:55:10 +0000 (11:55 +0000)]
net/iavf: implement power management
Implement support for the power management API by implementing a
`get_monitor_addr` function that will return an address of an RX ring's
status bit.
This patch is basically a cut-and-paste of the changes already
committed in ixgbe, i40e and ice drivers in 21.02. This extends
the availability of the power-saving mechanism to the iavf driver,
which is needed for those use-cases using virtual functions.
Patchset where PMD Power Management added in 21.02:
http://patchwork.dpdk.org/project/dpdk/list/?series=14756
Signed-off-by: David Hunt <david.hunt@intel.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Feifei Wang [Wed, 10 Mar 2021 02:40:29 +0000 (10:40 +0800)]
net/i40e: fix parsing packet type for NEON
In i40e NEON vector Rx path, the packet descs processing is incorrect.
This caused wrong packet type been filled in mbuf.
To fix this, when shifting the pktlen field to be 16-bit aligned, it
only needs to process the high 16bit of the packet descs instead of
the high 32bit.
Test Results:
Architecture: arm64
NIC: XL710
Driver: i40e
Package: Ether()/IP()/
Without this patch:
desc_to_ptype_v: ptype = 7 (error)
With this patch:
desc_to_ptype_v: ptype = 23 (correct)
Fixes: ae0eb310f253 ("net/i40e: implement vector PMD for ARM") Cc: stable@dpdk.org Signed-off-by: Feifei Wang <feifei.wang2@arm.com> Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com> Tested-by: Kathleen Capella <kathleen.capella@arm.com>
Ivan Malov [Fri, 26 Mar 2021 09:39:27 +0000 (12:39 +0300)]
net/sfc: fix error path inconsistency
At the fail label, there's a statement to set general errno and
error message. However, before the label is reached, a custom
error message can be set by the code which parses actions.
This custom (action-specific) message, when present,
must not be replaced by the general one.
Fixes: 662286ae61d2 ("net/sfc: add actions parsing stub to MAE backend") Cc: stable@dpdk.org Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru> Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru> Reviewed-by: Andy Moreton <amoreton@xilinx.com>
Ciara Loftus [Thu, 25 Mar 2021 08:22:09 +0000 (08:22 +0000)]
net/af_xdp: mark recvfrom return as ignored
Coverity complains that the return value of recvfrom() in the AF_XDP
datapath is not checked. We don't care about the return value because in
the case of an error we still return 0 from the receive function to
indicate no packets were received. So to make Coverity happy we cast the
return to 'void'.
Coverity issue: 369671 Fixes: 63e8989fe5a4 ("net/af_xdp: use recvfrom instead of poll syscall") Signed-off-by: Ciara Loftus <ciara.loftus@intel.com>
Guoyang Zhou [Tue, 23 Mar 2021 13:17:51 +0000 (21:17 +0800)]
net/hinic: fix crash in secondary process
Some apps, such as fstack, will use secondary process to access the
memory of eth_dev_ops, and they want to get the info of dev, but hinic
driver does not initialized it when in secondary process.
Tyler Retzlaff [Fri, 12 Mar 2021 22:20:06 +0000 (14:20 -0800)]
ethdev: install driver headers
Introduce a meson option 'enable_driver_sdk', when true installs internal
driver headers for ethdev. This allows drivers that do not depend on
stable api/abi to be built external to the dpdk source tree.
Chengchang Tang [Tue, 23 Mar 2021 13:45:56 +0000 (21:45 +0800)]
net/hns3: fix long task queue pairs reset time
Currently, the queue reset process needs to be performed one by one,
which is inefficient. However, the queues reset in the same function is
almost at the same stage. To optimize the queue reset process, a new
function has been added to the firmware command HNS3_OPC_CFG_RST_TRIGGER
to reset all queues in the same function at a time. And the related
queue reset MBX message is adjusted in the same way too.
Fixes: bba636698316 ("net/hns3: support Rx/Tx and related operations") Cc: stable@dpdk.org Signed-off-by: Chengchang Tang <tangchengchang@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
Huisong Li [Tue, 23 Mar 2021 13:45:55 +0000 (21:45 +0800)]
net/hns3: fix link update when failed to get link info
In the "hns3_dev_link_update" API, the link information of the port is
obtained first, and then 'dev_link' in dev->data is updated. When the
driver is resetting or fails to obtain link info, the current driver
still reports the previous link info to the user. This may cause that
the dev->data->dev_link may be inconsistent with the hw link status.
Therefore, the link status consistency between the hardware, driver,
and framework can be ensured in this interface regardless of whether
the driver is normal or abnormal.
Fixes: 109e4dd1bd7a ("net/hns3: get link state change through mailbox") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
Chengchang Tang [Tue, 23 Mar 2021 13:45:54 +0000 (21:45 +0800)]
net/hns3: fix Tx checksum for UDP packets with special port
For Kunpeng920 network engine, UDP packets with destination port 6081,
4789 or 4790 will be identified as tunnel packets. If the UDP CKSUM
offload is set in the mbuf, and the TX tunnel mask is not set, the
CKSUM of these packets will be wrong. In this case, the upper layer
user may not identify the packet as a tunnel packet, and processes it
as non-tunnel packet, and expect to offload the outer UDP CKSUM, so
they may not fill the outer L2/L3 length to mbuf. However, the HW
identifies these packet as tunnel packets and therefore offload the
inner UDP CKSUM. As a result, the inner and outer UDP CKSUM are
incorrect. And for non-tunnel UDP packets with preceding special
destination port will also exist similar checksum error.
For the new generation Kunpeng930 network engine, the above errata
have been fixed. Therefore, the concept of udp_cksum_mode is
introduced. There are two udp_cksum_mode for hns3 PMD,
HNS3_SPECIAL_PORT_HW_CKSUM_MODE means HW could solve the above
problem. And in HNS3_SPECIAL_PORT_SW_CKSUM_MODE, hns3 PMD will check
packets in the Tx prepare and perform the UDP CKSUM for such packets
to avoid a checksum error.
Fixes: bba636698316 ("net/hns3: support Rx/Tx and related operations") Cc: stable@dpdk.org Signed-off-by: Chengchang Tang <tangchengchang@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
Huisong Li [Tue, 23 Mar 2021 13:45:52 +0000 (21:45 +0800)]
net/hns3: fix build for SVE path
The 'queue_full_cnt' stats have been encapsulated in 'dfx_stats'.
However, the modification in the SVE algorithm is omitted.
As a result, the driver fails to be compiled when the SVE
algorithm is used.
Fixes: 9b77f1fe303f ("net/hns3: encapsulate DFX stats in datapath") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
Huisong Li [Tue, 23 Mar 2021 13:45:51 +0000 (21:45 +0800)]
net/hns3: fix reporting undefined speed
There may be a case in future that the speed obtained from firmware
is undefined (such as, 400G or other rate), and link status of device is
up. At this case, PMD driver will reports 100Mbps to the user in the
"hns3_dev_link_update" API, which is unreasonable. Besides, if the
speed from firmware is zero, driver should report zero instead of
100Mbps.
Fixes: 59fad0f32135 ("net/hns3: support link update operation") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
git.droids-corp.org / dpdk.git / log