This patch fixes the incorrect session private mempool passing
to cryptodev.
Fixes: 41d561cbdd24 ("examples/fips_validation: add power on self test") Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com> Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Previous commit sets mtu to the same value as max_rx_pkt_len.
Though PMDs (at least Intel ones) consider MTU as
max_rx_pkt_len minus ether header, crc bytes, vlan tags.
Fixes: 73d2c1d3f33c ("examples/ip_fragmentation: support big packets") Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Ferruh Yigit [Tue, 15 Jan 2019 17:28:02 +0000 (17:28 +0000)]
examples/kni: fix crash while handling userspace request
When KNI interface receives RTE_KNI_REQ_CFG_NETWORK_IF request, it
stap/start the physical device which as a result of stop() can free all
the mbufs in its queue.
Meanwhile sample application continues to read from KNI interface queues
and push into device queues. This simultaneous access may cause a crash,
crash log can be found at defect description.
As a solution KNI sample application can do the proper synchronization,
and stop transfer between KNI interface and physical interface while
physical device stop/started.
Bugzilla ID: 116 Fixes: 3fc5ca2f6352 ("kni: initial import") Cc: stable@dpdk.org Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com> Acked-by: Phil Yang <phil.yang@arm.com>
Jiayu Hu [Wed, 16 Jan 2019 00:45:33 +0000 (08:45 +0800)]
gro: check invalid TCP header length
When the TCP header length of input packets is invalid (i.e., less
than 20 bytes or greater than 60 bytes), check_seq_option() will
access illegal memory area when compare TCP Options, which may
cause a segmentation fault.
This patch adds missing invalid TCP header length check to avoid
illegal memory accesses.
Fixes: 0d2cbe59b719 ("lib/gro: support TCP/IPv4") Fixes: 9e0b9d2ec0f4 ("gro: support VxLAN GRO") Cc: stable@dpdk.org Signed-off-by: Jiayu Hu <jiayu.hu@intel.com> Tested-by: Yinan Wang <yinan.wang@intel.com> Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Reshma Pattan [Tue, 15 Jan 2019 13:45:04 +0000 (13:45 +0000)]
app/pdump: fix vdev cleanup
Virtual devices added in pdump application
using rte_eal_hotplug_add should be removed explicitly
while exiting the pdump application, otherwise the
subsequent run of the pdump application will fail with the reason
that virtual devices with the same name already exists in primary.
Fixes: 6362f362a2 ("app/pdump: use EAL hotplug instead of ethdev attach") Cc: stable@dpdk.org Signed-off-by: Reshma Pattan <reshma.pattan@intel.com> Tested-by: Ferruh Yigit <ferruh.yigit@intel.com>
Liang Ma [Tue, 15 Jan 2019 10:01:37 +0000 (10:01 +0000)]
power: fix file descriptor leak
Coverity issue: 328528 Fixes: e6c6dc0f96c8 ("power: add p-state driver compatibility") Signed-off-by: Liang Ma <liang.j.ma@intel.com> Reviewed-by: Lei Yao <lei.a.yao@intel.com> Tested-by: Lei Yao <lei.a.yao@intel.com> Signed-off-by: David Hunt <david.hunt@intel.com>
Harry van Haaren [Thu, 17 Jan 2019 17:17:44 +0000 (17:17 +0000)]
build: fix meson check for binutils 2.30
This commit reworks the checks for binutils 2.30 and how
the flags to disable AVX512F are passed to the compiler.
Previously the #define for including AVX512 code was set,
while the -mno-avx512f argument was given to the compiler.
This would cause gcc to correctly refuse to emit AVX512
instructions, but the rte_memcpy code that includes AVX512
optimizations was being added to the build.
The check for binutils check is now moved to x86 as it is
irrelevant for other architectures, and the -mno-avx512f
flag is passed to a march_opts array in meson. As the
-mno-avx512 flag is added earlier in the build, the code
in rte_memcpy is no longer attempted to be compiled.
This commit also adds a message print in the meson configure
stage to alert the user of the workaround being employed.
Fixes: a32ca9a4ebc1 ("mk: fix scope of disabling AVX512F support") Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com> Tested-by: Ferruh Yigit <ferruh.yigit@intel.com>
Ilya Maximets [Tue, 15 Jan 2019 11:29:39 +0000 (14:29 +0300)]
eal: fix clang build with intrinsics forced
This fixes x86_64-native-linuxapp-clang build with
CONFIG_RTE_FORCE_INTRINSICS=y:
include/generic/rte_atomic.h:218:9: error:
implicit declaration of function '__atomic_exchange_2'
is invalid in C99 [-Werror,-Wimplicit-function-declaration]
include/generic/rte_atomic.h:501:9: error:
implicit declaration of function '__atomic_exchange_4'
is invalid in C99 [-Werror,-Wimplicit-function-declaration]
include/generic/rte_atomic.h:783:9: error:
implicit declaration of function '__atomic_exchange_8'
is invalid in C99 [-Werror,-Wimplicit-function-declaration]
We didn't caught this issue previously on other platforms because
CONFIG_RTE_FORCE_INTRINSICS enabled by default only for armv8.
Fixes: 7bdccb93078e ("eal: fix ARM build with clang") Cc: stable@dpdk.org Signed-off-by: Ilya Maximets <i.maximets@samsung.com>
Anatoly Burakov [Wed, 16 Jan 2019 12:12:53 +0000 (12:12 +0000)]
eal: check string parameter lengths
When specifying parameters such as hugefile prefix from the
command-line, it is possibly to supply an empty string. This may
lead to various problems: for example, if hugefile prefix is
empty, the runtime config path construction may end up
looking like "/var/run/dpdk//_config", which will technically
work, but is wrong and places files in the wrong place.
To fix it, check lengths of such user-specified parameters for
hugefile prefix, as well as hugepage dir and user-specified
mbuf pool ops string.
David Marchand [Thu, 17 Jan 2019 13:12:57 +0000 (14:12 +0100)]
eal: fix out of bound access when no CPU available
In the unlikely case when the dpdk application is started with no cpu
available in the [0, RTE_MAX_LCORE - 1] range, the master_lcore is
automatically chosen as RTE_MAX_LCORE which triggers an out of bound
access.
Either you have a crash then, or the initialisation fails later when
trying to pin the master thread on it.
In my test, with RTE_MAX_LCORE == 2:
Noa Ezra [Sun, 13 Jan 2019 11:37:06 +0000 (11:37 +0000)]
examples/ip_fragmentation: support big packets
In some vendors the RX and TX configuration must be the same, therefore
the MTU size need to be equal to max_rx_pkt_len.
The MTU is the largest size packet in bytes that can be sent on the
network, therefore before changing this parameter, the NIC could not
receive packets larger than 1500 bytes, which is the default MTU size.
In addition, scatter-gather need to be enabled in order to receive
frames bigger than mbuf size.
Signed-off-by: Noa Ezra <noae@mellanox.com> Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
David Hunt [Mon, 7 Jan 2019 11:39:20 +0000 (11:39 +0000)]
examples/power: fix core id with JSON commands
This patch fixes a bug introduced in the 64-core limitation
enhancement where the core_id is inadvertently converted from
virtual to physical even though it may already be a physical
core_id.
We should be using the core_type field, and only converting via
hypervisor when core_type is set to CORE_TYPE_VIRTUAL
Fixes: 5776b7a371d1 ("examples/power: allow VM to use lcores over 63") Signed-off-by: David Hunt <david.hunt@intel.com> Reviewed-by: Lei Yao <lei.a.yao@intel.com> Tested-by: Lei Yao <lei.a.yao@intel.com>
David Hunt [Tue, 8 Jan 2019 15:25:23 +0000 (15:25 +0000)]
examples/power: fix for p-state
The vm_power_manager starts by setting the environment to acpi
using rte_power_set_env(PM_ENV_ACPI_CPUFREQ). This causes a problem
starting vm_power_manager when the system is using the intel_pstate
driver. The env should be set to none, or not called at all, because
the library now auto-detects the environment to be either acpi or
intel_pstate. This patch sets the environment to none so that the
library can successfully auto-detect.
David Hunt [Mon, 7 Jan 2019 14:40:51 +0000 (14:40 +0000)]
test/power: handle p-state mode
The cpufreq test breakes when the system is using the intel_pstate
driver for frequency management. The power library has recentyly been
updated to allow use of the intel_pstate driver, this patch fixes the
cpufreq test so that it can now use either acpi or pstate modes.
The library will auto-detect, and set the environment appropriately.
Fixes: ed7c51a6a680 ("app/test: vm power management") Cc: stable@dpdk.org Signed-off-by: David Hunt <david.hunt@intel.com>
David Hunt [Mon, 7 Jan 2019 14:40:23 +0000 (14:40 +0000)]
power: fix frequency list return code
The power_pstate_cpufreq_freqs() function was returning -1 in an
unsigned int, causing buffer over-runs when the results were being
processed. This function should be returning zero for all error
conditions, similar to it's acpi relation, power_acpi_cpufreq_freqs().
David Hunt [Mon, 7 Jan 2019 14:39:34 +0000 (14:39 +0000)]
power: fix frequency list to handle null buffer
This patch fixes a segfault in the case where a null buffer is passed
to the following functions:
power_acpi_cpufreq_freqs()
power_pstate_cpufreq_freqs()
Fixes: 445c6528b55f ("power: common interface for guest and host") Signed-off-by: David Hunt <david.hunt@intel.com>
David Hunt [Tue, 8 Jan 2019 14:59:26 +0000 (14:59 +0000)]
power: fix error handling on setting governor
In the power_set_governor_*() functions, we using fputs() on /sys
filesystem. However, we also need to call fflush() to ensure that
the write completes successfully. Otherwise the attempt to set the
power governor fails and the function returns as if it has
succeeded. This patch adds an fflush to ensure that the
write succeeds, otherwise returns an error.
gcc 6.2 and 6.4 fails to compile lib/librte_ipsec/sa.c
with the following errors:
lib/librte_ipsec/sa.c:
In function inline_outb_tun_pkt_process:
x86_64-native-linuxapp-gcc/include/rte_memcpy.h:337:2:
error: array subscript is above array bounds [-Werror=array-bounds]
rte_mov32((uint8_t *)dst + 1 * 32, (const uint8_t *)src + 1 * 32);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
...
It complains about the following lines of code:
esp_outb_tun_pkt_prepare(struct rte_ipsec_sa *sa, rte_be64_t
....
/* update spi, seqn and iv */
esph = (struct esp_hdr *)(ph + sa->hdr_len);
iv = (uint64_t *)(esph + 1);
rte_memcpy(iv, ivp, sa->iv_len);
While I believe it is a false positive,
it is too excessive to use rte_memcpy() here,
as IV length could be only 0/8/16 bytes.
So introduce small helper function to copy IV and use it
instead of rte_memcpy().
Fixes: 4d7ea3e1459b ("ipsec: implement SA data-path API") Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Yongseok Koh [Mon, 14 Jan 2019 21:16:22 +0000 (13:16 -0800)]
net/mlx5: fix instruction hotspot on replenishing Rx buffer
On replenishing Rx buffers for vectorized Rx, mbuf->buf_addr isn't needed
to be accessed as it is static and easily calculated from the mbuf address.
Accessing the mbuf content causes unnecessary load stall and it is worsened
on ARM.
Alejandro Lucero [Fri, 11 Jan 2019 13:25:53 +0000 (13:25 +0000)]
net/nfp: add CPP bridge as service
The Netronome's Network Flow Processor chip is highly programmable
with the goal of processing packets at high speed. Processing units
and other chip components are available from the host through the
PCIe CPP(Command Push Pull bus) interface. The NFP PF PMD configures
a CPP handler for setting up and working with vNICs, perform actions
like link up or down, or accessing extended stats from the MAC component.
There exist NFP host tools which access the NFP components for
programming and debugging but they require the CPP interface. When the
PMD is bound to the PF, the DPDK app owns the CPP interface, so these
host tools can not access the NFP through other means like NFP kernel
drivers.
This patch adds a CPP bridge using the rte_service API which can be
enabled by a DPDK app. Interestingly, DPDK clients like OVS will not
enable specific service cores, but this can be performed with a
secondary process specifically enabling this CPP bridge service and
therefore giving access to the NFP to those host tools.
This patch adds the type-of-service and time-to-live IP header
fields validation on E-Switch, both for match pattern and
VXLAN encapsulation action IP header itesm. The E-Switch flows
will use the common mlx5_flow_validate_item_ipv4/6 routines
with added extra parameter, specifying the supported fields
mask.
This patch adds the type-of-service and time-to-live IP header
fields support on E-Switch. There match pattern for both fields
with masking is added. Also these fields can be set for VXLAN
tunnel encapsulation header.
This issue is critical for some Open VSwitch configuration
on overlayed (tunneled) networks, where the tos field can be
inherited from outer header to inner header.
net/mlx5: add TOS and TTL flower match and tunnel keys
This patch is a preparation for adding the type-of-service and
time-to-live IP header fields support on E-Switch. There are
two types of keys added - one for match pattern, other for
tunnel encapsulation header.
This issue is critical for some Open VSwitch configuration
on overlayed (tunneled) networks, where the tos field can be
inherited from outer header to inner header.
RH7.2 with kernel 3.10.0-327 does not support VXLAN
devices metadata and IFLA_VXLAN_COLLECT_METADATA
key is neither defined nor supported. We must specify
VNI parameter, which will be actually ignored by kernel,
applied rules will be processed by mlx5 kernel driver
and the actual VNI from rules will be used.
net/mlx5: switch to detached VXLAN network devices
Current design uses the VXLAN virtual devices attached
to outer network interface for decapsulation. Kernel
allows to use non-attached devices, so now we can create
not attached device and use it both for encapsulation
and decapsulation. Devices management becomes simpler,
less VXLAN devices are created and used.
The VXLAN encapsulation neigh/local rules will use
the new introduced structure, which keeps the
rules lists, related to specified outer interface,
instead of attached VTEP structure. It allows us to
unbind VTEP structure from keeping the rules for
interface.
Currently the VXLAN encapsulation neigh/local rules
are stored in the list contained in the VTEP device
structure. Encapsulation VTEP device is attached to
outer interface and stored rules are related to this
underlying interface. We are going to use unattached
VXLAN devices for encapsulation (kernel does not use
attached interface to find egress one), so we should
introduce the structure to keep interface related
neigh/local rules instead of VTEP structure. This
patch introduces internal tcf_irule structure, and
its create/delete methods.
net/mlx5: support ethernet type for tunnels on E-Switch
This patch add support for inner and outer ethernet types for the
E-Switch Flows with tunnels. Inner and outer ethernet type match
can be specified with ethernet items, vlan items, or implicitly
deduced from IP address items. The tcm_info field in Netlink message
tcm structure is filled always with outer protocol.
This patch updates the validation routine for the E-Switch Flows.
The ethernet type field can be specified within inner and outer
tunnel ethernet items, by vlan item or implicitly deduced from
IP address items. The validation routine checks all these items
and their combinations for mutual compatibility issues and possible
conflicts.
This patch updates the translation routine for the E-Switch Flows.
Inner tunnel pattern items are translated into Netlink message,
support for tunnel inner IP addresses (v4 or v6), IP protocol,
and TCP and UDP ports is added.
We are going to support Flows matching with outer tunnel items
and not containing the explicit tunnel decap action (this one
might be drop, redirect or table jump, for exapmle).
So we can not rely on presence of tunnel decap action in the
list to decide whether the Flow is for tunnel, instead we will
use the presence of tunnel item. Item translation is rebound
to presence of tunnel items, instead of relying on decap action.
There is no way to tell kernel driver the outer address type
(IPv4 or IPv6) but specify the address flower key. The outer
address key is put on Netlink with zero mask if there is no
RTE item is specified in the list.
This patch updates the validation routine for the E-Switch Flows.
The inner/outer item flags are added and set correctly, the
validation routine will accept and check the inner items
which follow the tunnel item (like VNI).
net/mlx5: remove checks for outer tunnel items on E-Switch
This patch removes unnecessary outer tunnel parameters check in the
validation routine for the E-Switch Flows. IPv4/IPv6 may have any
spec and mask, and transferred to tc without changes, all checks
are performed by kernel.
We are going to support Flows matching with outer tunnel items
and not containing the explicit tunnel decap action (this one
might be drop, redirect or table jump, for exapmle). So we can
not rely on presence of tunnel decap action in the list to decide
whether the Flow is for tunnel, instead we will use the presence
of tunnel item (like RTE_FLOW_ITEM_TYPE_VXLAN) in the item list.
The tunnel pattern checks within Flow validation routine are
rebound to presence of tunnel item. VXLAN decap action checks
for presence of VXLAN VNI item.
The tunnel UDP item is checked at the point of processing the tunnel
item (i.e. VXLAN). We can not perform UDP item check as tunnel once
UDP item encountered in the list, because it is not known yet whether
the tunnel item follows. The pointer to UDP item is saved and
checked as outer ones if tunnel item found.
Haiyue Wang [Thu, 10 Jan 2019 12:07:31 +0000 (20:07 +0800)]
net/i40e: perform basic validation on VF messages
Do the VF message basic validation such as OPCODE message length check,
some special OPCODE message format check, to protect the i40e PMD from
malicious VF message attack.
Fixes: 4861cde46116 ("i40e: new poll mode driver") Cc: stable@dpdk.org Signed-off-by: Haiyue Wang <haiyue.wang@intel.com> Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Yongseok Koh [Mon, 14 Jan 2019 10:00:22 +0000 (02:00 -0800)]
app/testpmd: fix missing prefetch in macswap mode
Prefetching packet was missing when do_macswap() was optimized.
Fixes: 62b52877adbe ("app/testpmd: batch MAC swap for performance on x86") Signed-off-by: Yongseok Koh <yskoh@mellanox.com> Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Yongseok Koh [Mon, 14 Jan 2019 10:00:21 +0000 (02:00 -0800)]
app/testpmd: fix pointer reference in macswap
The pointer is misused and could set wrong MAC address. As a result, some
of packets can be dropped in receiver side due to MAC address mismatch.
This can be shown as performance degradation.
Bugzilla ID: 188 Fixes: 62b52877adbe ("app/testpmd: batch MAC swap for performance on x86") Signed-off-by: Yongseok Koh <yskoh@mellanox.com> Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Printing pointer in log is uninformative (unless in a debugger),
instead print the assigned kernel device name which correlates
well with what TAP is doing.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org> Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Having a global variable which is set to "TUN" or "TAP" during
probe is a potential bug if probing is ever done in different
processes or contexts. Let's fix it now by using existing enum
that has type of connection.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by Keith Wiles <keith.wiles@intel.com>
Assigning tun and tap index in DPDK tap device driver is racy
and fails if used with primary/secondary. Instead use the kernel
feature of device wildcarding where if a name with %d is used
the kernel will fill in the next available device.
Fixes: 02f96a0a82d1 ("net/tap: add TUN/TAP device PMD") Cc: stable@dpdk.org Reported-by: Haifeng Li <hfli@netitest.com> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by Keith Wiles <keith.wiles@intel.com>
If interface name is passed to remote or iface then check
the length and for invalid characters. This avoids problems where
name gets truncated or rejected by kernel.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by Keith Wiles <keith.wiles@intel.com>
Thomas Monjalon [Wed, 9 Jan 2019 14:23:19 +0000 (15:23 +0100)]
config: add static linkage of mlx dependency
The libraries provided by rdma-core may be statically linked
if enabling CONFIG_RTE_IBVERBS_LINK_STATIC in the make-based build.
If CONFIG_RTE_BUILD_SHARED_LIB is disabled, the applications
will embed the mlx PMDs with ibverbs and the mlx libraries.
If CONFIG_RTE_BUILD_SHARED_LIB is enabled,
the mlx PMDs will embed ibverbs and the mlx libraries.
Support with meson may be added later.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Shahaf Shuler <shahafs@mellanox.com>
Thomas Monjalon [Wed, 9 Jan 2019 14:23:18 +0000 (15:23 +0100)]
config: gather options for dlopen mlx dependency
Rename options CONFIG_RTE_LIBRTE_MLX4_DLOPEN_DEPS and
CONFIG_RTE_LIBRTE_MLX5_DLOPEN_DEPS to a single option
CONFIG_RTE_IBVERBS_LINK_DLOPEN.
Rename meson option enable_driver_mlx_glue to ibverbs_link.
There was no good reason for setting a different link option
for mlx4 and mlx5. Having a single common option makes it
easier to understand and unify make and meson systems.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Shahaf Shuler <shahafs@mellanox.com>
David Marchand [Fri, 11 Jan 2019 14:11:37 +0000 (15:11 +0100)]
ethdev: declare Tx prepare API as not experimental
The Tx prepare API was introduced at a time when the experimental API
tag mechanism did not exist yet and is missing the tag.
However, this API has been there since 17.02 and did not suffer any
change since its introduction.
Let's just remove the warning from the documentation.
Signed-off-by: David Marchand <david.marchand@redhat.com> Acked-by: Ferruh Yigit <ferruh.yigit@intel.com> Acked-by: Andrew Rybchenko <arybchenko@solarflare.com> Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Shreyansh Jain [Fri, 11 Jan 2019 12:25:08 +0000 (12:25 +0000)]
bus/fslmc: support secondary process
Previously FSLMC bus only supported blacklisting of DPNI (eth),
DPSECI (crypto) devices. With this patch, devices like DPIO,
DPMCP, and other DP* can also be blacklisted/whitelisted.
This is a required condition for secondary processes where the
secondary needs to be passed a mutually exclusive list of
resources as compared the primary and all other secondaries.
This patch also moves the DPIO memory from malloc to hugepage so
that in future in case the DPIO list can be shared, it can be
accessed in secondaries.
Once this patch is done, multi-process cases can be executed by
whitelisting/blacklisting devices in each instance.
Shreyansh Jain [Fri, 11 Jan 2019 12:25:05 +0000 (12:25 +0000)]
net/dpaa2: change reference to private device
The I/O threads for DPAA2 take their reference for bpool ID, the
port ID and other info like qdid, from the rte_eth_dev. Further,
to get this data during I/O operation, a reference of the RTE
device is kept in the queue structure (dpaa2_queue).
In case of secondary processes, rte_eth_dev is not same as the
primary process. Thus, the reference goes invalid.
This patch changes the implementation to use the dev_private
rather than the rte_eth_dev as that is shared area across
all the processes.
Shreyansh Jain [Fri, 11 Jan 2019 12:25:02 +0000 (12:25 +0000)]
mempool/dpaa2: support saving context of buffer pool
Initial design was to have the buffer pool per process where a
global static array stores the bpids. But, in case of secondary
processes, this would not allow the I/O threads to translate the
bpid in Rx'd packets.
This patch moves the array to a global area (rte_malloc) and in
case of Rx thread not containing a valid reference to the array,
reference is build using the handle avaialble in the dpaa2_queue.
Nipun Gupta [Fri, 11 Jan 2019 12:24:59 +0000 (12:24 +0000)]
net/dpaa2: support custom hash key
The DPAA2 hw can support a special offset based
configuration to program distribution on hash.
This is for all cases, which are not directly supported.
e.g. HASH based distribution on inner ip header
of a GRE tunnel.
Nipun Gupta [Fri, 11 Jan 2019 12:24:55 +0000 (12:24 +0000)]
net/dpaa2: add dpdmux initialization and configuration
This patch introduces an rte pmd API to configure dpdmux from
the application.
dpdmux can work in association with dpni as an additional
distribution capability on the NIC.
Nipun Gupta [Fri, 11 Jan 2019 12:24:49 +0000 (12:24 +0000)]
net/dpaa2: add dpdmux mc flib
dpdmux object is added as a part of net driver as it is used to
de-multiplex packets to separate interfaces on basis of specific rules.
These rules can be configured from the software
Akhil Goyal [Fri, 11 Jan 2019 12:24:30 +0000 (12:24 +0000)]
net/dpaa2: enable optional timestamp in mbuf
This patch enables the population of timestamp field
in mbuf on packet receive.
It may give performance impact on LX2xxx platforms.
So, it has been made optional for Lx2xxx platform.
One shall call, rte_dpaa2_enable_ts() to enable it.
Shreyansh Jain [Fri, 11 Jan 2019 12:24:23 +0000 (12:24 +0000)]
bus/fslmc: fix parse method for bus devices
Current code expects that bus->parse() would get a string containing
the name of the bus. That is incorrect. bus->parse() is expected
to have strings like:
dpni.1,key=val
dpio.2,key=val
Jens Freimann [Fri, 11 Jan 2019 09:39:28 +0000 (10:39 +0100)]
net/virtio: check head desc with correct wrap counter
In virtio_pq_send_command() we check for a used descriptor
and wait in an idle loop until it becomes used. We can't use
vq->used_wrap_counter here to check for the first descriptor
we made available because the ring could have wrapped. Let's use
the used_wrap_counter that matches the state of the head descriptor.
Fixes: ec194c2f1895 ("net/virtio: support packed queue in send command") Signed-off-by: Jens Freimann <jfreimann@redhat.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Ilya Maximets [Wed, 9 Jan 2019 14:50:15 +0000 (17:50 +0300)]
net/virtio: support platform memory ordering
VIRTIO_F_ORDER_PLATFORM is required to use proper memory barriers
in case of HW vhost implementations like vDPA.
DMA barriers (rte_cio_*) are sufficent for that purpose.
Previously known as VIRTIO_F_IO_BARRIER.
Signed-off-by: Ilya Maximets <i.maximets@samsung.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com>
Ilya Maximets [Wed, 9 Jan 2019 14:50:14 +0000 (17:50 +0300)]
net/virtio: update memory ordering comment for VQ notify
We're not using IO ports in case of modern device even on IA.
Also, this comment useless for other architectures.
Signed-off-by: Ilya Maximets <i.maximets@samsung.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com>
Ilya Maximets [Wed, 9 Jan 2019 14:50:13 +0000 (17:50 +0300)]
net/virtio: add barrier before reading the flags
Reading the used->flags could be reordered with avail->idx update.
vhost in kernel disables notifications for the time of packets
receiving, like this:
1. disable notify
2. process packets
3. enable notify
4. has more packets ? goto 1
In case of reordering, virtio driver could read the flags on
step 2 while notifications disabled and update avail->idx after
the step 4, i.e. vhost will exit the loop on step 4 with
notifications enabled, but virtio will not notify.
Fixes: c1f86306a026 ("virtio: add new driver") Cc: stable@dpdk.org Reported-by: Shahaf Shuler <shahafs@mellanox.com> Signed-off-by: Ilya Maximets <i.maximets@samsung.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com>
Rami Rosen [Wed, 9 Jan 2019 14:57:30 +0000 (16:57 +0200)]
examples/vhost: remove unnecessary method and constant
This cleanup patch removes a method and a constant which
are now unnecessary in the VHOST sample application, namely
the validate_num_devices() method and the MAX_DEVICES
constant.
Signed-off-by: Rami Rosen <ramirose@gmail.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Ivan Malov [Wed, 9 Jan 2019 11:06:39 +0000 (11:06 +0000)]
net/sfc: discard last seen VLAN TCI if Tx packet is dropped
Early processing of a packet on transmit may change last seen
VLAN TCI in the queue context. If such a packet is eventually
dropped, last seen VLAN TCI must be set to its previous value.
Fixes: 7fd636815a43 ("net/sfc: support VLAN offload on transmit path") Cc: stable@dpdk.org Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru> Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
git.droids-corp.org / dpdk.git / log