The DevX counter management triggers an asynchronous event to get back
the new counters values from the HW.
The counter management doesn't trigger 2 parallel events for the same
pool, hence, the pool cannot be updated again in the event waiting time.
When the port is stopped, the DevX event mechanism wrongly was
destroyed what remained all the waiting pools in waiting state forever.
As a result, the counters of the stuck pools were never updated again.
Separate the DevX interrupt installation from the dev installation and
remove the DevX interrupt unregistration\registration from the
stop\start operations.
Now, the DevX interrupt should be installed in probe and uninstalled in
close.
Cc: stable@dpdk.org
Fixes: f15db67df0 ("net/mlx5: accelerate DV flow counter query")
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
When calculate queue index for queue action in
switch filter, vsi base queue should be included.
Fixes: 57c4f26935 ("net/ice: enable switch filter")
Cc: stable@dpdk.org
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
This patch enables the unaligned chunks feature for AF_XDP which allows
chunks to be placed at arbitrary places in the umem, as opposed to them
being required to be aligned to 2k. This allows for DPDK application
mempools to be mapped directly into the umem and in turn enable zero copy
transfer between umem and the PMD.
This patch replaces the zero copy via external mbuf mechanism introduced
in commit e9ff8bb719 ("net/af_xdp: enable zero copy by external mbuf").
The pmd_zero copy vdev argument is also removed as now the PMD will
auto-detect presence of the unaligned chunks feature and enable it if so
and otherwise fall back to copy mode if not detected.
Signed-off-by: Ciara Loftus <ciara.loftus@intel.com>
Signed-off-by: Kevin Laatz <kevin.laatz@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
The promiscuous and multicast fields are now initialised as enabled for
some virtual PMDs. This allows the devices to be used when running
applications that attempt to enable promiscuous or multicast mode.
Signed-off-by: Ciara Power <ciara.power@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
When sendto call fails and ENOBUFS/EAGAIN error is being set
some of the packets are actually successfully transmitted.
There is no available count of those packets, so in order to
make the statistics more accurate, all the previously enqueued
packets will be considered successful, even though this is not
entirely correct.
Statistics numbers before this update:
Pktgen:
Total Rx Pkts: 1360084
Tx Pkts: 2000000
testpmd:
RX-packets: 1408346 RX-missed: 0 RX-bytes: 84503418
TX-packets: 526486 TX-errors: 881851 TX-bytes: 31589724
Statistics numbers after this update:
Pktgen:
Total Rx Pkts: 1329872
Tx Pkts: 2000000
testpmd:
RX-packets: 1389156 RX-missed: 0 RX-bytes: 83349360
TX-packets: 1389156 TX-errors: 0 TX-bytes: 83349360
Fixes: 74b7fc0a0f ("net/af_packet: fix packet bytes counting")
Cc: stable@dpdk.org
Signed-off-by: Flavia Musatescu <flavia.musatescu@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
When to create hrxq for the drop, it could fail on creating qp and goto
the error handle which will release created ind_table by calling drop
release function, which takes rte_ethdev as the only parameter and uses
the priv->drop_queue.hrxq as input to release.
Unfortunately, at this point, the hrxq is not allocated and
priv->drop_queue.hrxq is still NULL, which leads to a segfault.
This patch fixes the above by allocating the hrxq at first place and
when the error happens, hrxq is released as the last one.
This patch also release other allocated resources by the correct order,
which is missing previously.
Fixes: 78be885295 ("net/mlx5: handle drop queues as regular queues")
Cc: stable@dpdk.org
Reported-by: Zengmo Gao <gaozengmo@jd.com>
Signed-off-by: Xiaoyu Min <jackmin@mellanox.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
This patch adds rss support for gtpu with input set teid and
pppoe/pppod with input set source mac and session id.
Signed-off-by: Simei Su <simei.su@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
This commit adds FDIR ID support to the AVX2 based receive
path routine. Support for both 16B and 32B descriptors is
implemented.
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Tested-by: Mesut Ali Ergin <mesut.a.ergin@intel.com>
This commit adds an implementation to the SSE vector implementation of
RX routine and moves some common defines from a c file to the header
file.
I40e can have 16 and 32 byte descriptors, and the Flow
Director ID data and indication-bit are in different locations
for each size descriptor. The support is implemented in two
separate functions as they require vastly different operations.
The 16B descriptor re-purposes the "filter-status" u32 field
to indicate FDIR ID when the FLM bit is set. No extra loads
are required, however we do have to store to mbuf->fdir.hi,
which is not stored to in the RX path before this patch.
The 32B descriptor requires loading the 2nd 16 bytes of each
descriptor, to get the FLEXBH_STAT and FD Filter ID from
qword3. The resulting data must also be stored to mbuf->fdir.hi,
same as the 16B code path.
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Tested-by: Mesut Ali Ergin <mesut.a.ergin@intel.com>
This commit adds a fdir_enable flag in a uint8_t sized hole
the rx queue structure The flag enables the rx code path to
easily identify if fdir is active. This can be used to skip
fdir id processing when it is not required.
The flag is zero by default (as rxq is zmalloc-ed at startup),
and the flag is set to 1 on configuration of a flow director rule.
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Tested-by: Mesut Ali Ergin <mesut.a.ergin@intel.com>
One issue caught by Coverity 340835
*unlock: axgbe_phy_set_mode unlocks pdata->phy_mutex
*double_unlock: axgbe_phy_sfp_detect unlocks pdata->phy_mutex
while it is unlocked.
In axgbe_phy_sfp_detect()/axgbe_phy_set_redrv_mode(),
axgbe_phy_get_comm_ownership() and axgbe_phy_put_comm_ownership()
are invoked subsequently.
Currently in axgbe_phy_get_comm_ownership(), during one of the case
'phy_data->comm_owned' is not protected and before returning 0, lock is
not called and unlock is called in axgbe_phy_put_comm_ownership()
directly which is incorrect.
Ideally, the variable 'phy_data->comm_owned' needs to be protected.
During success scenario, lock is called in
axgbe_phy_get_comm_ownership() followed by unlock in
axgbe_phy_put_comm_ownership(). In failure case, unlock is invoked in
axgbe_phy_get_comm_ownership() itself appropriately.
The fix is to protect 'phy_data->comm_owned' in the identified case
ensuring locks/unlocks properly exist.
Coverity issue: 340835
Fixes: a5c7273771 ("net/axgbe: add phy programming APIs")
Cc: stable@dpdk.org
Signed-off-by: Pallantla Poornima <pallantlax.poornima@intel.com>
Acked-by: Ravi Kumar <ravi1.kumar@amd.com>
This commit adds support for matching flows on Geneve headers.
Signed-off-by: Moti Haimovsky <motih@mellanox.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
This commit add querying the HCA which FLEX protocols are already
enabled.
Signed-off-by: Moti Haimovsky <motih@mellanox.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
mlx5_link_update immediately returns when called with no-wait parameter
and its call for retrieving the link status returns with EAGAIN error.
This is too harsh on busy systems where a first call fails with EAGAIN
from time to time.
This patch adds a (very limited) retry on such cases in order to allow
retrieving the link status.
Signed-off-by: Moti Haimovsky <motih@mellanox.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
This patch converts some of the casts to unaligned integer types.
The memcpy call is replaced with explicit copying because it
may require type qualifiers in some environments.
Fixes: 18a1c20044 ("net/mlx5: implement Tx burst template")
Cc: stable@dpdk.org
Reported-by: Jeremy Plsek <jplsek@iol.unh.edu>
Signed-off-by: Ali Alnubani <alialnu@mellanox.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
The transmitter packets counter was not updated
correctly in the loop inside the tx_burst routines.
Fixes: f32a3f5216 ("net/mlx5: fix completion queue overflow for large burst")
Cc: stable@dpdk.org
Signed-off-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
There is no dependency on this library for mlx4.
Fixes: 1dd7c7e38c ("net/mlx4: support meson build")
Cc: stable@dpdk.org
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Raslan Darawsheh <rasland@mellanox.com>
Sync mail box data structures to version 0x0002.
This patch checks for mismatch in mail box revision and
avoids initializing octeontx2 pci device if
there is a mismatch.
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>
When app restarts, RSS/FDIR profile ID will increase which may
run out of all the profile resources.
The patch cleanup RSS/FDIR profile resources when device init
to fix this issue.
Fixes: d7d150b930 ("net/ice: enable RSS when device init")
Cc: stable@dpdk.org
Signed-off-by: Ying A Wang <ying.a.wang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
This patch add FDIR statistical counter support
and it includes RTE_FLOW count actions support
and query support.
RTE_FLOW count actions support id and shared.
RTE_FLOW query record packet hits by default.
Signed-off-by: Yahui Cao <yahui.cao@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
The patch integrates the counter resource init/release into fdir's
init/release scenario
Signed-off-by: Yahui Cao <yahui.cao@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
FDIR can send packet to a group of queues and distruibte it by RSS.
Signed-off-by: Yahui Cao <yahui.cao@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
Add ice_create_fdir_filter to create a rule. If a flow is matched by
flow director filter, filter rule will be set to HW. For now common
pattern and queue/passthru/drop/mark actions are supported.
Signed-off-by: Yahui Cao <yahui.cao@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
This patch adds a HW FDIR rule to the FDIR HW table
without adding a FDIR filter.
Signed-off-by: Beilei Xing <beilei.xing@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
Enable flow director engine, including initialization and teardown.
- Control VSI create and release.
- Queue pair allocated, set up and release.
- Programming packet create and release.
- FDIR profile create and release.
Signed-off-by: Beilei Xing <beilei.xing@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
This patch fixes (dereference after null check) coverity issue.
The address of first segmented packets was not set correctly during
reassembling packets which led to this issue.
Coverity issue: 13245
Fixes: 8a44c15aa5 ("net/ixgbe: extract non-x86 specific code from vector driver")
Cc: stable@dpdk.org
Signed-off-by: Joyce Kong <joyce.kong@arm.com>
Reviewed-by: Xiao Zhang <xiao.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
This patch fixes (dereference after null check) coverity issue.
The address of first segmented packets was not set correctly during
reassembling packets which led to this issue.
Coverity issue: 343422, 343403
Fixes: ca74903b75 ("net/i40e: extract non-x86 specific code from vector driver")
Cc: stable@dpdk.org
Signed-off-by: Joyce Kong <joyce.kong@arm.com>
Reviewed-by: Xiao Zhang <xiao.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
Needs to call _rte_eth_dev_callback_process to run registered
callbacks when link status change.
Fixes: cf911d90e3 ("net/ice: support link update")
Cc: stable@dpdk.org
Signed-off-by: Leyi Rong <leyi.rong@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
The patch reworks packet process engine's binary classifier
(switch) for the new framework. It also adds support for new
packet type like PPPoE for switch filter.
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
The patch adds a manifest for all patterns which can be selected
by low level filter engine to compose their supported list.
Signed-off-by: Ying A Wang <ying.a.wang@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
The patch reworks the generic flow API (rte_flow) implementation.
It introduces an abstract layer which provides a unified interface
for low-level filter engine (switch, fdir, hash) to register supported
patterns and actions and implement flow validate/create/destroy/flush/
query activities.
The patch also removes the existing switch filter implementation to
avoid compile error. Switch filter implementation for the new framework
will be added in the following patch.
Signed-off-by: Ying A Wang <ying.a.wang@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
Added a devarg to control the mode in generic flow API.
We use none-pipeline mode by default.
Signed-off-by: Qiming Yang <qiming.yang@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
The patch removes redundant code and cleans up some wrong indentations.
Signed-off-by: Ying A Wang <ying.a.wang@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
Previous fix removes usage of rte_pktmbuf_prepend() to get pointer
to virtio net header which changes mbuf data_off and data_len.
Size of virtio net header is added to segment length when Tx descriptor
is composed, but segment address (calculated using data_off) is not
adjusted to take size of virtio net header into account.
Fixes: 1ae55ad38e ("net/virtio: fix mbuf data and packet length mismatch")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Tiwei Bie <tiwei.bie@intel.com>
The loop to read packets does not take all packets as the number of
available packets (nb_used) is decremented in the loop.
Fixes: 52b5a707e6 ("net/virtio: add Altivec Rx")
Signed-off-by: David Christensen <drc@linux.vnet.ibm.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
With vectorized functions, only the rx stats for number of packets is
incremented.
Update also the other statistics.
Fixes: 52b5a707e6 ("net/virtio: add Altivec Rx")
Signed-off-by: David Christensen <drc@linux.vnet.ibm.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
In case VIRTIO_F_ORDER_PLATFORM(36) is not negotiated, then the frontend
and backend are assumed to be implemented in software, that is they can
run on identical CPUs in an SMP configuration.
Thus a weak form of memory barriers like rte_smp_r/wmb, other than
rte_cio_r/wmb, is sufficient for this case(vq->hw->weak_barriers == 1)
and yields better performance.
For the above case, this patch helps yielding even better performance
by replacing the two-way barriers with C11 one-way barriers for used
flags in packed ring.
Signed-off-by: Joyce Kong <joyce.kong@arm.com>
Reviewed-by: Gavin Hu <gavin.hu@arm.com>
Reviewed-by: Phil Yang <phil.yang@arm.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
In case VIRTIO_F_ORDER_PLATFORM(36) is not negotiated, then the frontend
and backend are assumed to be implemented in software, that is they can
run on identical CPUs in an SMP configuration.
Thus a weak form of memory barriers like rte_smp_r/wmb, other than
rte_cio_r/wmb, is sufficient for this case(vq->hw->weak_barriers == 1)
and yields better performance.
For the above case, this patch helps yielding even better performance
by replacing the two-way barriers with C11 one-way barriers for avail
flags in packed ring.
Meanwhile, a read barrier is required to ensure ordering between
descriptor's flags and content reads [1]. With C11, load-acquire can
enforce the ordering instead of rmb barrier.
[1] https://patchwork.dpdk.org/patch/49109/
Signed-off-by: Joyce Kong <joyce.kong@arm.com>
Reviewed-by: Gavin Hu <gavin.hu@arm.com>
Reviewed-by: Phil Yang <phil.yang@arm.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
When configuring i40e rx queue, the temporary variable to store max
packet length is not big enough which leads to integer overflow issue.
This patch fixes the issue by removing the variable and using the
expression directly since the variable is only used once.
Fixes: 4861cde461 ("i40e: new poll mode driver")
Cc: stable@dpdk.org
Signed-off-by: Xiao Zhang <xiao.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
If the default dev args 'proto_xtr' is not in the first position, it
will overwrite the parsed queue map value, so use an new variable to
save the default.
And enhance the error message printing to show the right information.
Fixes: 7e124ff12c ("net/ice: support protocol extraction per Rx queue")
Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
The original protocol extraction memory free is removed by introducing
new design for releasing port upon close.
Fixes: bd513ece3c ("net/ice: release port upon close")
Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
