Wei Dai [Mon, 13 Mar 2017 08:59:26 +0000 (16:59 +0800)]
examples/performance-thread: remove useless include
There is no function to refer any part of execinfo.h, so remove the
reference to it.
And there is no this file in musl. So need to remove it to support musl.
Thomas Monjalon [Sun, 30 Apr 2017 14:11:40 +0000 (16:11 +0200)]
usertools: fix CPU layout with python 3
These differences in Python 3 were causing errors:
- xrange is replaced by range
- dict values are a view (instead of list)
- has_key is removed
Fixes: deb87e6777c0 ("usertools: use sysfs for CPU layout") Fixes: 63985c5f104e ("usertools: fix CPU layout for more than 2 threads") Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Reviewed-by: Gowrishankar Muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>
The existing code used to search for module files via modinfo has
several corner cases which can result in it failing where it should be
successful.
The call to lower() would cause results returned by 'modinfo' to be
forced to lowercase, results which were subsequently passed to
exists() which is case sensitive. This was most likely done to capture
all variants of failure strings modinfo might return
(ie. ERROR/Error/error/...) without thought negative effect to the
later call to exists(). For many this is a nonissue but if the module
path included non-lowercase alpha characters, something which is
easily possible with a non-lowercase kernel-extraversion string, this
would cause an issue.
We could move the call to lower() to the check for "error" but this
still leaves possible corner cases, for modules or module paths with
'error' in them.
Instead we will prevent modinfo's stderr from being used as a "good
value" for path, meaning we either get a valid path from modinfo, or
nothing at all. This removes all corner cases.
Ultimately these preliminary checks are unnecessary as exists() will
only return True if it is passed a valid path, passing it modinfo's
stderr would fail. In keeping with the original code, however, we do
some preliminary checks, but we are now free of corner cases.
Signed-off-by: Mark Asselstine <mark.asselstine@windriver.com>
Kuba Kozak [Thu, 27 Apr 2017 14:42:40 +0000 (16:42 +0200)]
net/ixgbe: support xstats by ID
To achieve functionality of retrieving only specific statistics
given by application there are two new functions added:
ixgbe_dev_xstats_get_by_ids() which retrieve
values of statistics specified by ids array
and ixgbe_dev_xstats_get_names_by_ids() which retrieve
names of statistics specified by ids array.
Kuba Kozak [Thu, 27 Apr 2017 14:42:39 +0000 (16:42 +0200)]
net/igb: support xstats by ID
To achieve functionality of retrieving only specific statistics
given by application there are two new functions added:
eth_igb_xstats_get_by_id() which retrieve
values of statistics specified by ids array
and eth_igb_xstats_get_names_by_id() which retrieve
names of statistics specified by ids array.
Kuba Kozak [Thu, 27 Apr 2017 14:42:38 +0000 (16:42 +0200)]
app/proc-info: support xstats by ID and by name
There are new arguments --xstats-ids and --xstats-name
in proc_info command line to retrieve statistics given by ids
and by name.
E.g. --xstats-ids="1,3,5,7,8"
E.g. --xstats-name rx_errors
Kuba Kozak [Thu, 27 Apr 2017 14:42:36 +0000 (16:42 +0200)]
ethdev: retrieve xstats by ID
Extended xstats API in ethdev library to allow grouping of stats
logically so they can be retrieved per logical grouping managed
by the application.
Added new functions rte_eth_xstats_get_names_by_id and
rte_eth_xstats_get_by_id using additional arguments (in compare
to rte_eth_xstats_get_names and rte_eth_xstats_get) - array of ids
and array of values.
doc: add description for modified xstats API
Documentation change for new extended statistics API functions.
The old API only allows retrieval of *all* of the NIC statistics
at once. Given this requires a MMIO read PCI transaction per statistic
it is an inefficient way of retrieving just a few key statistics.
Often a monitoring agent only has an interest in a few key statistics,
and the old API forces wasting CPU time and PCIe bandwidth in retrieving
*all* statistics; even those that the application didn't explicitly
show an interest in.
The new, more flexible API allow retrieval of statistics per ID.
If a PMD wishes, it can be implemented to read just the required
NIC registers. As a result, the monitoring application no longer wastes
PCIe bandwidth and CPU time.
Kuba Kozak [Thu, 27 Apr 2017 14:42:35 +0000 (16:42 +0200)]
ethdev: revert xstats by ID
Revert patches to provide clear view for
upcoming changes. Reverted patches are listed below:
commit ea85e7d711b6 ("ethdev: retrieve xstats by ID")
commit a954495245c4 ("ethdev: get xstats ID by name")
commit 1223608adb9b ("app/proc-info: support xstats by ID")
commit 25e38f09af9c ("net/e1000: support xstats by ID")
commit 923419333f5a ("net/ixgbe: support xstats by ID")
Stop port before enabling QinQ.
Add commands to set inner and outer TPID's and start port.
Remove TPID's from flow validate and and flow create commands.
Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com> Acked-by: John McNamara <john.mcnamara@intel.com>
Rx queues configured with more than 1023 descriptors cause readv() calls to
fail due to more iovec entries than permitted by the kernel. As a result,
no packets can be received.
Quietly limit internal Rx queue size to the maximum number of iovec entries
to fix this issue.
Fixes: 0781f5762cfe ("net/tap: support segmented mbufs") Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com>
Jerin Jacob [Mon, 1 May 2017 18:41:55 +0000 (00:11 +0530)]
net/thunderx: fix deadlock in Rx path
RBDR buffers are refilled when SW consumes the buffers from CQ.
This creates deadlock case when CQ buffers exhausted due to lack
of RBDR buffers. Fix is to refill the RBDR when rx_free_thresh
meet, irrespective of the number of CQ buffers consumed.
Fixes: e2d7fc9f0a24 ("net/thunderx: add single and multi-segment Rx") Cc: stable@dpdk.org Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Markos Chandras [Thu, 16 Feb 2017 16:17:31 +0000 (16:17 +0000)]
examples/ethtool: fix link with ixgbe shared lib
When RTE_DEVEL_BUILD is unset, -rpath is unset.
So the ethtool app cannot link with ixgbe shared library
which is required by ethtool lib:
warning: librte_pmd_ixgbe.so.1, needed by
examples/ethtool/lib/x86_64-native-linuxapp-gcc/lib/librte_ethtool.so,
not found (try using -rpath or -rpath-link)
It is fixed by adding the library in the application link.
The library link is also improved to specify that this explicit link
to ixgbe is needed only in the shared lib mode.
Fixes: 077d223e25c3 ("examples/ethtool: use ixgbe public function") Signed-off-by: Markos Chandras <mchandras@suse.de> Acked-by: Remy Horton <remy.horton@intel.com> Acked-by: Timothy Redaelli <tredaelli@redhat.com> Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Pablo de Lara [Wed, 26 Apr 2017 11:29:52 +0000 (12:29 +0100)]
examples/l3fwd-power: fix Rx descriptor size
L3fwd power app monitors the RX queues to see if the polling frequency
should be adjusted (the busier the queue, the higher the frequency).
The app uses several thresholds in the ring to determine the frequency,
being 96 the highest one, when frequency should be highest.
The problem is that the difference between this value and the ring size
is not big enough (128 - 96 = 32 descriptors), which means that
if the descriptors are not replenished quick enough, queue might
not be busy, but the app would think that it is, because 96th descriptor
is set.
Therefore, by increasing this gap (increasing the RX ring size),
we make sure that this false measurement will not happen.
Fixes: b451aa39db31 ("examples/l3fwd-power: use DD bit rather than RX queue count") Cc: stable@dpdk.org Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com> Tested-by: Lei Yao <lei.a.yao@intel.com>
examples/l2fwd-keepalive: clean up shared mem on exit
This patch adds the unlinking/unmapping of shared host memory
on termination of l2fwd-keepalive. Previously it was only
cleaned on re-running of the example application.
The l2fwd-keepalive example has infinite processing loops and as a
result the only way to exit it is via SIGINT/SIGTERM (e.g. Control-C).
The resulting shutdown is unclean, which is fixed by adding a signal
handler that causes the processing loops to break.
Bruce Richardson [Fri, 28 Apr 2017 10:18:15 +0000 (11:18 +0100)]
examples/performance-thread: fix build on FreeBSD 10.0
While later releases in the FreeBSD 10 series have a CPU_COUNT macro
defined, FreeBSD 10.0 and 10.1 do not have this macro. Therefore we provide
a basic fallback implementation of the macro for platforms where it is not
defined.
Fixes: 433ba6228f9a ("examples/performance-thread: add pthread_shim app") Cc: stable@dpdk.org Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Allain Legacy [Fri, 31 Mar 2017 19:13:20 +0000 (15:13 -0400)]
app/testpmd: load commands from file at runtime
Adds support to testpmd to load a set of cmdline CLI commands at runtime.
This can be helpful when needing to cut-n-paste many commands where
cut-n-paste may not be practical.
Allain Legacy [Fri, 31 Mar 2017 19:13:19 +0000 (15:13 -0400)]
app/testpmd: load commands from file at startup
Adds support to testpmd to load a set of cmdline CLI commands at startup.
This can be helpful when needing to cut-n-paste many commands each time
testpmd is restarted. This option will work in both interactive and
non-interactive modes.
Bit-rate collation should only be done by one core. This patch adds
an option to select which core performs the bit-rate calculation,
which is also disabled by default.
Fixes: 7e4441c8efb9 ("app/testpmd: add bitrate statistics calculation") Signed-off-by: Remy Horton <remy.horton@intel.com> Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Jiayu Hu [Fri, 28 Apr 2017 01:32:50 +0000 (09:32 +0800)]
app/testpmd: fix exit without freeing resources
When testpmd exits, it frees the acquired resources (e.g. stop ports).
However, when we terminate it by Ctrl-d, testpmd exits directly without
releasing the resources. In this patch, we fix this exit issue.
Fixes: af75078fece3 ("first public release") Cc: stable@dpdk.org Signed-off-by: Jiayu Hu <jiayu.hu@intel.com> Reviewed-by: Jianfeng Tan <jianfeng.tan@intel.com>
The number of mbufs in pools is not consistent depending on the
options passed by the user and the number of ports, especially
in numa mode, when the number of mbuf is specified by the user.
When the user specifies the number of mbuf (per pool), it should
overrides the default value.
./build/app/testpmd -- --total-num-mbufs=8000 -i
<mbuf_pool_socket_0>: n=256000, size=2176, socket=0
<mbuf_pool_socket_1>: n=256000, size=2176, socket=1
# BAD, should be n=8000 for each socket
Since
commit 999b2ee0fe45 ("app/testpmd: enable NUMA support by default"),
testpmd is started with numa enabled by default. This highlights a
floating point exception when started with --total-num-mbufs without any
port (division by 0). This bug was already triggered before this commit
if the --no-numa option was given.
This commit adds a check of the nb_ports value before doing the
division. By looking at this code, it appears that the creation of the
mbuf pool is not consistent for the number of mbufs depending on the
configuration. This is fixed in the next commit.
Fixes: b6ea6408fbc7 ("ethdev: store numa_node per device") Cc: stable@dpdk.org Signed-off-by: Olivier Matz <olivier.matz@6wind.com> Acked-by: Jingjing Wu <jingjing.wu@intel.com>
Pablo de Lara [Tue, 25 Apr 2017 15:32:36 +0000 (16:32 +0100)]
examples/l2fwd-crypto: fix packets array index
Fixes: c0f87eb5252b ("cryptodev: change burst API to be crypto op oriented") Cc: stable@dpdk.org Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Pablo de Lara [Wed, 26 Apr 2017 13:33:03 +0000 (14:33 +0100)]
crypto/openssl: fix AES-GCM capability
Crypto OpenSSL PMD supports 16, 24 and 32 byte keys,
for AES GCM.
Fixes: d61f70b4c918 ("crypto/libcrypto: add driver for OpenSSL library") Cc: stable@dpdk.org Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Huanle Han [Mon, 20 Feb 2017 14:04:46 +0000 (22:04 +0800)]
net/virtio: fix crash when closing twice
This commit fixs segment fault when rte_eth_dev_close() is called on
a virtio dev more than once. Assigning zero after free to avoids
freed memory to be accessed again.
Fixes: 69c80d4ef89b ("net/virtio: allocate queue at init stage") Cc: stable@dpdk.org Signed-off-by: Huanle Han <hanxueluo@gmail.com> Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Jianfeng Tan [Fri, 21 Apr 2017 02:28:09 +0000 (02:28 +0000)]
net/virtio-user: fix device type check
Segfault happens when using virtio-user after commit 7f0a669e7b04
("ethdev: add allocation helper for virtual drivers").
It's due to we use ethdev->device to recognize physical devices,
but after above commit, this field is also filled for virtual
devices. Then we obtain the wrong pci_dev pointer and accessing
its field when copying pci info results in segfault.
To fix it, we use hw->virtio_user_dev to differentiate physical
devices from virtual devices.
Fixes: 6a7c0dfcdf40 ("net/virtio: do not depend on PCI device of ethdev") Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com> Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Jianfeng Tan [Thu, 27 Apr 2017 07:35:39 +0000 (07:35 +0000)]
net/virtio: fix link status always down
The virtio port link status will always be DOWN:
The commit aa9f06061765 ("net/virtio: fix link status always being up")
introduces a flag to help checking the status. If this flag is not set,
status will be always down. However, in dev start, this flag is set
after link status update, then we miss the chance to change the status
to UP in dev start.
To fix this bug, we simply move the link status update after the flag
setting so that the status can be correctly updated.
Fixes: aa9f06061765 ("net/virtio: fix link status always being up") Cc: stable@dpdk.org Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com> Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Jianfeng Tan [Thu, 27 Apr 2017 07:35:36 +0000 (07:35 +0000)]
net/virtio: fix MSI-X for modern devices
The field, use_msix, in struct virtio_hw is not updated for modern
device, and is always zero. And now we depend on the status feature
and MSI-X to report LSC support (which is also not a correct
behavior). As a result, LSC is always disabled for modern devices.
To fix this, we just recognize MSI-X capability when going through
capability list, and update the info in virtio.
Fixes: 6ba1f63b5ab0 ("virtio: support specification 1.0") Cc: stable@dpdk.org Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com> Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Zhiyong Yang [Thu, 27 Apr 2017 09:41:23 +0000 (17:41 +0800)]
vhost: workaround MQ fails to startup
vhost since dpdk17.02 + qemu2.7 and above will cause failures of
new connection when negotiating to set MQ. (one queue pair works
well).
Because there exist some bugs in qemu code when introducing
VHOST_USER_PROTOCOL_F_REPLY_ACK to qemu. when dealing with the vhost
message VHOST_USER_SET_MEM_TABLE for the second time, qemu indeed
doesn't send the messge (The message needs to be sent only once)but
still will be waiting for dpdk's reply ack, then, qemu is always
freezing. DPDK code indeed works in the right way.
The feature VHOST_USER_PROTOCOL_F_REPLY_ACK has to be disabled
by default at the dpdk side in order to avoid the feature support of
DPDK + qemu at the same time. if doing like that, MQ can works well.
Cc: stable@dpdk.org Reported-by: Ciara Loftus <ciara.loftus@intel.com> Signed-off-by: Zhiyong Yang <zhiyong.yang@intel.com> Tested-by: Ciara Loftus <ciara.loftus@intel.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com> Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
In case of fiber and link speed set to 1Gb at peer side (with autoneg
or with defined speed), link status could be not properly updated at
time cable is plugged-in.
Indeed if cable was not plugged when device has been configured and
started then link status will not be updated properly with new speed
as no link setup will be triggered.
To avoid this issue, IXGBE_FLAG_NEED_LINK_CONFIG is set to try a link
setup each time link_update() is triggered and current link status is
down. When cable is plugged-in, link setup will be performed via
ixgbe_setup_link().
Signed-off-by: Laurent Hardy <laurent.hardy@6wind.com> Acked-by: Wei Dai <wei.dai@intel.com>
Wei Wang [Tue, 25 Apr 2017 06:30:49 +0000 (14:30 +0800)]
net/bonding: fix updating slave link status
We need to update dev->data->dev_link before handling LSC event.
Otherwise it will still have the initial value after the startup of
the program before interrupt callback was executed.
Fixes: 414b202343ce ("bonding: fix initial link status of slave") Cc: stable@dpdk.org Signed-off-by: Wei Wang <lnykww@gmail.com> Acked-by: Declan Doherty <declan.doherty@intel.com>
If segments are used, just mbufs previously linked to head descriptor
of a mbuf chain are released. Other Tx descriptor are used for the
mbuf chain but they keep their linked mbufs without releasing them.
It is not a fatal issue because sooner or later those descriptors will
be head descriptors or just used for a single mbuf packet, then those
linked mbufs will be released.
However, this leads to apps needing bigger mbufs pools and some
confusion about memory requirements. Indeed, because larger pools, some
performance impact could also be expected due to cache misses.
With this patch all Tx descriptors will release linked mbufs inside the
xmit function, and rte_pktmbuf_seg_free is used instead of
rte_pktmbuf_free.
Matt Peters [Tue, 25 Apr 2017 12:28:45 +0000 (08:28 -0400)]
net/i40e: fix mbuf alloc failed counter
When an mbuf alloc fails during the mempool get operation for the
i40e bulk alloc receive function, the rx_mbuf_alloc_failed counter
is not incremented to record the error.
This fix ensures consistency with the other i40e receive procedures and
other net drivers.
Changes included in this fix
- limit CFLAGS to base files
- fix to remove/mark unused members
- add checks for debug config option
- make qede_set_mtu() and qede_udp_dst_port_del() static and others
non-static as appropriate
- move local APIs qede_vlan_offload_set() and qede_rx_cqe_to_pkt_type()
- initialize variables as required
Firmware supports ntuple configuration which is always based on 4-tuples.
So remove RTE_ETH_FLOW_FRAG_IPV4 and RTE_ETH_FLOW_FRAG_IPV6 as valid flows.
Also merge the two switch statements into one.
Fixes: 622075356e8f ("net/qede: support ntuple and flow director filter") Signed-off-by: Harish Patil <harish.patil@cavium.com>
Use new mbuf flags PKT_RX_VLAN_STRIPPED and PKT_RX_QINQ_STRIPPED
introduced by the patch:
commit b37b528d957c ("mbuf: add new Rx flags for stripped VLAN")
This defect is a functional issue where the RX CQE pointer remains
uninitialized in the LRO code path which can cause null pointer exception
while accessing VLAN or RSS hash value from CQE.
- Add a common routine to handle ETH_RX_CQE_TYPE_TPA_CONT and
ETH_RX_CQE_TYPE_TPA_END.
- Remove enum qede_agg_state since there is no need to maintain
aggregation state.
- Modify the segment chaining logic by tracking head and tail
TPA segments of each aggregation.
- Add more debug and comments.
- Mark the packet as PKT_RX_LRO.
Fixes: 29540be7efce ("net/qede: support LRO/TSO offloads") Signed-off-by: Harish Patil <harish.patil@cavium.com>
- In qede_mac_addr_set(), in order to configure default MAC address we
first delete the existing MAC address before trying to add new one. During
init time, there is no MAC filter to begin with, so trying to remove a
non-existing MAC address causes a firmware exception. This can be prevented
by internally calling qede_mac_addr_add() which has the checks in place to
delete a MAC address only if it was added before.
- Remove setting of the default MAC address from within
qede_dev_configure() since rte_eth_dev_start() calls mac_addr_set() anyway.
Testpmd failed to start when CONFIG_RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC
is disabled, the root cause is the length of sw_ring and queue are
incorrect with the above configuration.
Andrew Rybchenko [Fri, 21 Apr 2017 12:16:42 +0000 (13:16 +0100)]
net/sfc: fix LSC interrupt support for UIO cases
Recently link status change interrupt was enabled by default in testpmd,
it has opened the driver bug with not working LSC interrupt if either
igb_uio or uio_pci_generic kernel driver is used.
Fixes: 06bc197796e2 ("net/sfc: interrupts support sufficient for event queue init") Cc: stable@dpdk.org Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Wei Dai [Thu, 20 Apr 2017 03:06:06 +0000 (11:06 +0800)]
net/ixgbe: align register setting when RSC is disabled
When Receive Side Coalescing (RSC) is not used, the RSC Disable
(RSC_DIS) filed of register Receive Filter Control Register (RFCTL)
should be set according to ixgbe datasheet.
Signed-off-by: Wei Dai <wei.dai@intel.com> Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
Wei Dai [Tue, 18 Apr 2017 06:57:02 +0000 (14:57 +0800)]
net/ixgbe/base: support MAC X550em/X557 LED on/off
This patch updates ixgbe_led_[on|off]_t_X550em for MAC or PHY connected
LEDs. To support both MAC or PHY connected LEDs, both MAC and PHY led
control registers are configured.
Signed-off-by: Wei Dai <wei.dai@intel.com> Tested-by: Yuan Peng <yuan.peng@intel.com>
Wei Dai [Tue, 18 Apr 2017 06:57:01 +0000 (14:57 +0800)]
net/ixgbe/base: support 2.5G KX physical layer
Add another define for the 2.5G KX physical layer. This requires all
variables and support functions that deal with "physical_layer" to be
bumped to u64 as this is the 33rd define.
Signed-off-by: Wei Dai <wei.dai@intel.com> Tested-by: Yuan Peng <yuan.peng@intel.com>
Wei Dai [Tue, 18 Apr 2017 06:57:00 +0000 (14:57 +0800)]
net/ixgbe/base: acquire PHY semaphore before device reset
A recent firmware change fixed an issue to acquire the PHY semaphore
before accessing PHY registers. This led to a case where SW can issue
a device reset clearing the MDIO registers. This patch makes SW acquire
the PHY semaphore before issuing a device reset.
Signed-off-by: Wei Dai <wei.dai@intel.com> Tested-by: Yuan Peng <yuan.peng@intel.com>
Wenzhuo Lu [Tue, 25 Apr 2017 03:11:25 +0000 (11:11 +0800)]
cmdline: fix parsing
When parsing a CLI, all the CLI instances are checked
one by one. Even if an instance already matches the CLI,
the parsing will not stop for ambiguous check.
The problem is that the following check may change the
parsing result of the previous one even if the following
instance doesn't match.
Use a temporary validate for the parsing result when
trying to match an instance and only store the result
when it matches, so the previous result has no chance
to be changed.
Fixes: af75078fece3 ("first public release") Cc: stable@dpdk.org Signed-off-by: Wenzhuo Lu <wenzhuo.lu@intel.com> Acked-by: Olivier Matz <olivier.matz@6wind.com>
In file included from build/include/rte_latencystats.h:43:0,
from /tmp/check-includes.sh.6580.c:1:
build/include/rte_metrics.h:91:2: error: unknown type name 'uint16_t'
[...]
In file included from /tmp/check-includes.sh.6580.c:1:0:
build/include/rte_latencystats.h:66:19: error: expected declaration
specifiers or '...' before '*' token
[...]
Fixes: 5cd3cac9ed22 ("latency: added new library for latency stats") Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
eventdev: fix headers for strict compilation flags
Exported headers must allow compilation with the strictest flags. This
commit addresses the following errors:
In file included from build/include/rte_eventdev_pmd.h:55:0,
from /tmp/check-includes.sh.25816.c:1:
build/include/rte_eventdev.h:908:8: error: struct has no named members
[-Werror=pedantic]
[...]
In file included from /tmp/check-includes.sh.25816.c:1:0:
build/include/rte_eventdev_pmd.h:65:35: error: ISO C does not permit named
variadic macros [-Werror=variadic-macros]
[...]
This commit addresses the following compilation errors:
In file included from build/include/rte_cryptodev_scheduler.h:37:0,
from /tmp/check-includes.sh.5355.c:1:
build/include/rte_cryptodev_scheduler_operations.h:43:30: error: unknown
type name 'uint8_t' struct rte_cryptodev *dev, uint8_t slave_id);
[...]
Exported headers must allow compilation with the strictest flags. This
commit addresses the following errors:
In file included from /tmp/check-includes.sh.20132.c:1:0:
build/include/rte_vhost.h:73:30: error: ISO C forbids zero-size array
'regions' [-Werror=pedantic]
[...]
Also:
- Add C++ awareness to rte_vhost.h for consistency with rte_eth_vhost.h.
- Move Linux includes into C++ block to prevent linking issues with
exported symbols.
- Update check-includes.sh following the removal of rte_virtio_net.h.
Finally, update check-includes.sh to ignore rte_vhost.h and rte_eth_vhost.h
from now on since the Linux headers they depend on are not clean enough:
In file included from /usr/include/linux/vhost.h:17:0,
from build/include/rte_vhost.h:43,
from build/include/rte_eth_vhost.h:44,
from /tmp/check-includes.sh.20132.c:1:
/usr/include/linux/virtio_ring.h: In function 'vring_init':
/usr/include/linux/virtio_ring.h:146:16: error: pointer of type 'void *'
used in arithmetic [-Werror=pointer-arith]
[...]
In file included from build/include/rte_vhost.h:43:0,
from build/include/rte_eth_vhost.h:44,
from /tmp/check-includes.sh.20132.c:1:
/usr/include/linux/vhost.h: At top level:
/usr/include/linux/vhost.h:73:3: error: ISO C99 doesn't support unnamed
structs/unions [-Werror=pedantic]
[...]
This commit addresses several errors related to missing includes such as:
In file included from /tmp/check-includes.sh.15315.c:1:0:
build/include/rte_avp_fifo.h:77:22: error: 'struct rte_avp_fifo' declared
inside parameter list [-Werror]
[...]
build/include/rte_avp_fifo.h: In function 'avp_fifo_init':
build/include/rte_avp_fifo.h:81:3: error: implicit declaration of function
'rte_panic' [-Werror=implicit-function-declaration]
[...]
build/include/rte_avp_fifo.h:83:6: error: dereferencing pointer to
incomplete type
[...]
build/include/rte_avp_fifo.h:109:2: error: implicit declaration of
function 'rte_wmb' [-Werror=implicit-function-declaration]
[...]
In file included from /tmp/check-includes.sh.15315.c:1:0:
build/include/rte_avp_common.h:104:2: error: unknown type name 'uint64_t'
[...]
build/include/rte_avp_common.h:386:15: error: 'ETHER_ADDR_LEN' undeclared
here (not in a function)
[...]
It addresses errors with strict compilation flags:
In file included from /tmp/check-includes.sh.15315.c:1:0:
build/include/rte_avp_common.h:122:3: error: ISO C99 doesn't support
unnamed structs/unions [-Werror=pedantic]
[...]
build/include/rte_avp_common.h:136:17: error: ISO C forbids zero-size
array 'buffer' [-Werror=pedantic]
[...]
This commit addresses the following compilation errors:
In file included from build/include/rte_flow_driver.h:50:0,
from /tmp/check-includes.sh.1397.cc:1:
build/include/rte_flow.h:428:2: error: expected primary-expression before
'.' token
[...]
build/include/rte_flow.h:469:1: sorry, unimplemented: non-trivial
designated initializers not supported
[...]
In file included from build/include/rte_flow_driver.h:50:0,
from /tmp/check-includes.sh.1397.cc:1:
build/include/rte_flow.h:631:1: error: C99 designator 'label_tc_s' outside
aggregate initializer
[...]
build/include/rte_flow.h:631:1: error: initializer-string for array of
chars is too long [-fpermissive]
[...]
build/include/rte_flow.h:650:1: sorry, unimplemented: non-trivial
designated initializers not supported
[...]
C++ does not support the C99-style designated initializers used in this
file for the default item masks. While the resulting symbols are primarily
useful to PMDs (written in C), they are exposed as part of the public API
for documentation purposes and to assist application writers.
Considering that:
- using pre-C99 initialization style for compatibility with C++ would
render them difficult to understand (all struct members must be
initialized)
- using both initialization styles would be needlessly verbose
- not exposing them at all would defeat their purpose
- applications do not normally need these symbols at run time
This commit hides these symbols from C++ applications. Specific C++
initializers will be added later if necessary.
Fixes: 6de5c0f1302c ("ethdev: define default item masks in flow API") Fixes: 7cd048321d1d ("ethdev: add MPLS and GRE flow API items") Cc: stable@dpdk.org Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com> Acked-by: Beilei Xing <beilei.xing@intel.com> Acked-by: Shahaf Shuler <shahafs@mellanox.com>
In file included from /tmp/check-includes.sh.681.c:1:0:
build/include/rte_mbuf_ptype.h:587:35: error: unknown type name 'uint32_t'
[...]
build/include/rte_mbuf_ptype.h:662:51: error: unknown type name 'size_t'
[...]
The RTE_FUNC_*_RET() and RTE_PROC_*_RET() macro definitions in rte_dev.h
require RTE_PMD_DEBUG_TRACE(). This macro is defined as needed by users of
rte_dev.h since its value depends on their own debug settings.
It may be defined multiple times as a result when including files from
various components simultaneously. Worse, these redefinitions may be
inconsistent. This causes the following compilation errors:
In file included from /tmp/check-includes.sh.13890.c:27:0:
build/include/rte_eventdev_pmd.h:58:0: error: "RTE_PMD_DEBUG_TRACE"
redefined [-Werror]
[...]
In file included from build/include/rte_ethdev_pci.h:39:0,
from /tmp/check-includes.sh.13890.c:13:
build/include/rte_ethdev.h:1042:0: note: this is the location of the
previous definition
[...]
In file included from /tmp/check-includes.sh.13890.c:83:0:
build/include/rte_cryptodev_pmd.h:65:0: error: "RTE_PMD_DEBUG_TRACE"
redefined [-Werror]
[...]
In file included from /tmp/check-includes.sh.13890.c:27:0:
build/include/rte_eventdev_pmd.h:58:0: note: this is the location of
the previous definition
[...]
This commit moves the RTE_PMD_DEBUG_TRACE() definition to rte_dev.h where
it is enabled consistently depending on global configuration settings and
removes redundant definitions.
Also when disabled, RTE_PMD_DEBUG_TRACE() is now defined as (void)0 to
avoid empty statements warnings if used outside { } blocks.
Zhihong Wang [Thu, 27 Apr 2017 23:00:14 +0000 (19:00 -0400)]
config: make AVX and AVX512 configurable
Making AVX and AVX512 configurable is useful for performance and power
testing.
The similar kernel patch at https://patchwork.kernel.org/patch/9618883/.
AVX512 support like in rte_memcpy has been in DPDK since 16.04, but it's
still unproven in rich use cases in hardware. Therefore it's marked as
experimental for now, will enable it after enough field test and possible
optimization.
Signed-off-by: Zhihong Wang <zhihong.wang@intel.com> Reviewed-by: Zhiyong Yang <zhiyong.yang@intel.com> Reviewed-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Jerin Jacob [Wed, 26 Apr 2017 16:29:19 +0000 (21:59 +0530)]
config: set cache line as 128B for generic arm64
armv8 implementations may have 64B or 128B cache line.
Setting to the maximum available cache line size in generic config to
address minimum DMA alignment across all arm64 implementations.
Increasing the cacheline size has no negative impact to cache invalidation
on systems with a smaller cache line.
The need for the minimum DMA alignment has impact on functional aspects
of the platform so default config should cater the functional aspects.
There is an impact on memory usage with this scheme, but that's not too
important for the single image arm64 distribution use case.
The arm64 linux kernel followed the similar approach for single
arm64 image use case.
http://lxr.free-electrons.com/source/arch/arm64/include/asm/cache.h
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com> Acked-by: Jianbo Liu <jianbo.liu@linaro.org> Acked-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
git.droids-corp.org / dpdk.git / log