When enic's interrupt handler is called indicating an error, it scans
through the receive queues (RQs) on the adapter looking for errors.
But since the inclusion of rx scatter, some of the RQs may not be in
use, and you shouldn't check them for errors.
Fixes: 856d7ba7ed ("net/enic: support scattered Rx")
Signed-off-by: Nelson Escobar <neescoba@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>
The calculation of truncated packets didn't take into account packet
errors due to the adapter not having buffers, causing both the
ipackets, and imissed counts to be wrong if such errors occurred. In
order to properly calculate the number of packets truncated, we need
to subtract the count of errors due to no buffers.
Fixes: c44d9f01ad ("net/enic: count truncated packets")
Signed-off-by: Nelson Escobar <neescoba@cisco.com>
The exported device start and stop functions where not setting the queue
states to RTE_ETH_QUEUE_STATE_STARTED and RTE_ETH_QUEUE_STATE_STOPPED.
After starting the device, the RTE queue stop function would not call
the enic queue stop function since queue was already marked as stopped.
Put queue state updates in the lower level queue start/stop functions
which are called by both device and queue start/stop functions.
Fixes: fefed3d1e6 ("enic: new driver")
Reviewed-by: Nelson Escobar <neescoba@cisco.com>
Signed-off-by: John Daley <johndale@cisco.com>
The check that all Tx and Rx queues were set up was not
adequate when reconfiguring with a different number of
queues. Only the number of completion queues (CQs) was
being used to make the determination, but the CQ array
is shared between the underlying Rx and Tx queues.
Check that the internal Rx, Tx and CQs are all set up
before completing port configuration.
Fixes: fefed3d1e6 ("enic: new driver")
Signed-off-by: John Daley <johndale@cisco.com>
Reviewed-by: Nelson Escobar <neescoba@cisco.com>
One instance of a filter add failure was not incrementing the
the fail counter.
Fixes: 4c2c7bf41f ("net/enic: fix negative array index write")
Signed-off-by: John Daley <johndale@cisco.com>
In the burst Tx cleanup function, the reference count in mbufs
returned to the pool should to be decremented before they are
returned. Decrementing is not done by rte_mempool_put_bulk()
so it must be done separately using __rte_pktmbuf_prefree_seg().
Also when returning unsent buffers when the device is stopped
use rte_mbuf_free_seg() instead of rte_mempool_put() so that
reference counts are properly decremented.
Fixes: 36935afbc5 ("net/enic: refactor Tx mbuf recycling")
Reviewed-by: Nelson Escobar <neescoba@cisco.com>
Signed-off-by: John Daley <johndale@cisco.com>
The enic PMD code has diverged from code that was once
shared with the enic kernel mode driver for performance
reasons. It is confusing and misleading to print the
internal version number. Remove it.
Fixes: fefed3d1e6 ("enic: new driver")
Signed-off-by: John Daley <johndale@cisco.com>
RHEL 7.1's GCC for POWER8 reports the following error in one rte_memcpy()
macro call by mlx5:
error: array subscript is above array bounds [-Werror=array-bounds]
It appears to be a GCC bug which can be worked around by making parentheses
more explicit.
Signed-off-by: Olga Shern <olgas@mellanox.com>
Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
In multi-queue + FTAG use case, we need to turn tx_ftag_en on for all
the Tx queues.
Fixes: 7958b1310d ("fm10k: enable FTAG based forwarding")
Reported-by: Ricky Li <ricky.li@intel.com>
Signed-off-by: Xiao Wang <xiao.w.wang@intel.com>
This reverts commit 4761f57d58.
Introducing VLAN table by adding VLAN adminq command will cause NIC's
throughput drop obviously. It's a hardware issue.
With this revert, VLAN filtering can only work when promiscuous mode
is disabled.
Reverts: 4761f57d58 ("net/i40e: fix VLAN filtering in promiscuous mode")
Reported-by: Jeff Shaw <jeffrey.b.shaw@intel.com>
Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
Acked-by : Jing Chen <jing.d.chen@intel.com>
The error is reported using test build script:
$ scripts/test-build.sh x86_64-native-linuxapp-gcc
...
drivers/net/virtio/virtio_user_ethdev.c:345:2: error:
this ‘if’ clause does not guard... [-Werror=misleading-indentation]
if (rte_kvargs_count(kvlist, VIRTIO_USER_ARG_PATH) == 1)
^~
Fixes: 404bd6bfe3 ("net/virtio-user: fix return value not checked")
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Moved defines since the driver had no such information.
Used RTE_PCI_DEVICE in place of RTE_PCI_DEV_ID_DECL* stuff.
Signed-off-by: David Marchand <david.marchand@6wind.com>
Reused defines from the driver and moved broadcom vendor id macro.
Used RTE_PCI_DEVICE in place of RTE_PCI_DEV_ID_DECL* stuff.
Signed-off-by: David Marchand <david.marchand@6wind.com>
Moved cisco vendor id since the driver had no such information.
Used RTE_PCI_DEVICE in place of RTE_PCI_DEV_ID_DECL* stuff.
Signed-off-by: David Marchand <david.marchand@6wind.com>
Moved vmware device ids macro since the driver had no such information.
Used RTE_PCI_DEVICE in place of RTE_PCI_DEV_ID_DECL* stuff.
Signed-off-by: David Marchand <david.marchand@6wind.com>
Reused defines from the driver.
Used RTE_PCI_DEVICE in place of RTE_PCI_DEV_ID_DECL* stuff.
Signed-off-by: David Marchand <david.marchand@6wind.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Reused defines from the driver and added a Intel vendor id macro for use by
igb later.
Used RTE_PCI_DEVICE in place of RTE_PCI_DEV_ID_DECL* stuff.
igb/igbvf is left as is, waiting for kni/ethtool cleanup.
Signed-off-by: David Marchand <david.marchand@6wind.com>
The older versions of rte_eth_bond_8023ad_conf_get and
rte_eth_bond_8023ad_setup were available in the old way since 2.0 - at
least according to the map file.
But versioning in the code was set to 16.04.
That breaks compatibility checks for 2.0 on that library.
For example with the dpdk abi checker:
http://people.canonical.com/~paelzer/compat_report.html
To fix, version the old symbols on the 2.0 version as they were
initially added to the map file.
See http://people.canonical.com/~paelzer/compat_report.html
Fixes: dc40f17a ("net/bonding: allow external state machine in mode 4")
Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
This is for target i686-native-linuxapp-gcc and gcc6,
Compilation error is:
In file included from
include/rte_mempool.h:77:0, from
drivers/net/virtio/virtio_rxtx_simple.c:
In function `virtio_xmit_pkts_simple':
include/rte_memcpy.h:551:2: error:
array subscript is above array bounds
rte_mov16((uint8_t *)dst + 1 * 16, (const uint8_t *)src + 1 * 16);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Call stack is as following:
virtio_xmit_pkts_simple
virtio_xmit_cleanup
rte_mempool_put_bulk
rte_mempool_generic_put
__mempool_generic_put
rte_memcpy
The array used as source buffer in virtio_xmit_cleanup (free) is a
pointer array with 32 elements, in 32bit this makes 128 bytes.
in rte_memcpy() implementation, there a code piece as following:
if (size > 256) {
rte_move128(...);
rte_move128(...); <--- [1]
....
}
The compiler traces the array all through the call stack and knows the
size of array is 128 and generates a warning on above [1] which tries to
access beyond byte 128.
But unfortunately it ignores the "(size > 256)" check.
Giving a hint to compiler that variable "size" is related to the size of
the source buffer fixes compiler warning.
Fixes: 863bfb4744 ("mempool: optimize copy in cache")
Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Fixes a bug where rte_eth_vhost_get_queue_event would not return enabled
queues after a guest application restart.
Fixes: ee584e9710 ("vhost: add driver on top of the library")
Signed-off-by: Rich Lane <rich.lane@bigswitch.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
There is a logic bug in this code, that could lead to null pointer
dereference when cvq is NULL. Fix this problem by changing logic
&& to logic ||.
>> CID 127480: Null pointer dereferences (FORWARD_NULL)
>> Dereferencing null pointer "cvq".
if (!cvq && !cvq->vq) {
...
}
Coverity issue: 127480
Fixes: 01ad44fd37 ("net/virtio: split Rx/Tx queue")
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
When use strcpy() to copy string with length exceeding the last
parameter of strcpy(), it may lead to the destination string
unterminated.
We replaced strncpy with snprintf to make sure it's NULL terminated.
Coverity issue: 127476
Fixes: ce2eabdd43 ("net/virtio-user: add virtual device")
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
The return value by rte_kvargs_parse is not free(d), which leads
to memory leak.
Coverity issue: 127482
Fixes: ce2eabdd43 ("net/virtio-user: add virtual device")
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
When parsing /proc/self/maps to get hugepage information, the string
was being copied with strcpy(), which could, theoretically but in fact
not possiblly, overflow the destination buffer. Anyway, to avoid the
false alarm, we replaced strncpy with snprintf for safely copying the
strings.
Coverity issue: 127484
Fixes: 6a84c37e39 ("net/virtio-user: add vhost-user adapter layer")
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
When return values of function calls are not checked, Coverity will
report errors like:
if (rte_kvargs_count(kvlist, VIRTIO_USER_ARG_PATH) == 1)
>>> CID 127477: (CHECKED_RETURN)
>>> Calling "rte_kvargs_process" without checking return value
(as is done elsewhere 25 out of 30 times).
rte_kvargs_process(kvlist, VIRTIO_USER_ARG_PATH,
&get_string_arg, &path);
Coverity issue: 127477, 127478
Fixes: ce2eabdd43 ("net/virtio-user: add virtual device")
Fixes: 6a84c37e39 ("net/virtio-user: add vhost-user adapter layer")
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
On some older systems, such as SUSE 11, the compiling error shows
as:
.../dpdk/drivers/net/virtio/virtio_user/virtio_user_dev.c:67:22:
error: ‘O_CLOEXEC’ undeclared (first use in this function)
The fix is to use EFD_CLOEXEC, which is defined in sys/eventfd.h,
instead of O_CLOEXEC which needs _GNU_SOURCE defined on some old
systems.
Fixes: 37a7eb2ae8 ("net/virtio-user: add device emulation layer")
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Avoid submitting doorbell when:
* no packets have been submitted to TX
* no free resources have been submitted while RX
Sending doorbell without actual work to be performed by device
violates ENA specification and can lead to unpredictable behavior.
Fixes: 1173fca25a ("ena: add polling-mode driver")
Signed-off-by: Alexander Matushevsky <matua@amazon.com>
Signed-off-by: Jan Medala <jan@semihalf.com>
enic_set_mac_address() meant to remove the old MAC address before
setting the new one, but accidentally tried removing the new MAC
address before setting the new MAC address.
Fixes: fefed3d1e6 ("enic: new driver")
Signed-off-by: Nelson Escobar <neescoba@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>
enic_disable() removed the MAC address when a port was shut down but
enic_enable() didn't add the MAC address back when the port was
started again. Move where we set the MAC address for the adapter from
enic_setup_finish() to a enic_enable() so that port restarting works
properly.
Fixes: fefed3d1e6 ("enic: new driver")
Signed-off-by: Nelson Escobar <neescoba@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>
This patch adds support for extended statistics for BNX2X PMD.
Signed-off-by: Rasesh Mody <rasesh.mody@qlogic.com>
Acked-by: Remy Horton <remy.horton@intel.com>
This patch adds support for extended statistics for QEDE PMD.
Signed-off-by: Rasesh Mody <rasesh.mody@qlogic.com>
Acked-by: Remy Horton <remy.horton@intel.com>
Fixed issue could occur when Mbuf starvation happens in a middle of
reception of a segmented packet. In such a situation, the PMD has to
release all segments of that packet. The end condition was wrong
causing it to free an Mbuf still handled by the NIC.
Fixes: 9964b965ad ("net/mlx5: re-add Rx scatter support")
Reported-by: Yongseok Koh <yskoh@mellanox.com>
Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
In mlx5 rx function, the packet_type and ol_flags mbuf fields are not
properly initialized when no rx offload feature is enabled (checksum, l2
tun checksum, vlan_strip, crc). Thus, these fields can have a value
different of 0 depending on their value when the mbuf was freed.
This can result in an incorrect application behavior if invalid
ol_flags/ptype are set, or memory corruptions if IND_ATTACHED_MBUF is
set in ol_flags.
Fixes: 081f7eae24 ("mlx5: process offload flags only when requested")
Signed-off-by: Maxime Leroy <maxime.leroy@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
The enic PMD was using the same variables in the enic structure to
track two different things. Initially rq_count, wq_count, cq_count,
and intr_count were set to the values obtained from the VIC adapters
as the maximum resources allocated on the VIC, then in
enic_set_vnic_res(), they were set to the counts of resources actually
used, discarding the initial values. The checks in enic_set_vnic_res()
were technically incorrect if it is called more than once on a port,
which happens when using bonding, but were harmless in practice as the
checks couldn't fail on the second call.
The enic rx-scatter patch misunderstood the subtleties of
enic_set_vnic_res(), and naively added a multiply by two to the
rq_count check. This resulted in the rq_count check failing when
enic_set_vnic_res() was called a second time, ie when using bonding.
This patch adds new variables to the enic structure to track the
maximum resources the VIC is configured to provide so that the
information isn't later lost and calls to enic_set_vnic_res() do
the expected thing.
Fixes: 856d7ba7ed ("net/enic: support scattered Rx")
Signed-off-by: Nelson Escobar <neescoba@cisco.com>
When calling i40e_flowtype_to_pctype in i40e_get_hash_filter_global_config
and i40e_set_hash_filter_global_config, function i40e_flowtype_to_pctype
will be possibly make an out-of-bounds access, because size of the array
is 15. So check the flow type is valid before calling
i40e_flowtype_to_pctype.
In the process fix other occurances of the same problem
Coverity issue: 37793, 37794
Fixes: 782c8c92f1 ("i40e: add hash configuration")
Fixes: f2b2e2354b ("i40e: split function for hash and flow director input")
Fixes: 98f0557076 ("i40e: configure input fields for RSS or flow director")
Signed-off-by: Beilei Xing <beilei.xing@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Null-checking vsi suggests that it may be null, but it
has been dereferenced before null-checking. So move the
check to before the assignment statement using the pointer.
Coverity issue: 119265, 119266
Fixes: d0a349409b ("i40e: support AQ based RSS config")
Fixes: 647d1eaf75 ("i40evf: support AQ based RSS config")
Signed-off-by: Beilei Xing <beilei.xing@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
The condition, "(pf->flags | I40E_FLAG_VMDQ)" will always be true,
regardless of the value of the flags operand, because I40E_FLAG_VMDQ
is 4ULL - meaning at least one bit will always be set in the result.
That will cause log error when VMDq is disabled.
Since the original intent behind the condition is to check if VMDq
is enabled, fix the code by changing "|" to "&".
Coverity issue: 13219, 13221
Fixes: 4805ed59e9 ("i40e: enhance mac address operations")
Signed-off-by: Beilei Xing <beilei.xing@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Currently we are wrongly setting HWRM_CFA_L2_SET_RX_MASK_INPUT_MASK_MCAST
flag in bnxt_hwrm_cfa_l2_set_rx_mask() which is preventing promiscuous
and multicast promiscuous settings from working correctly.
This patch fixes it.
Fixes: 244bc98b0d ("net/bnxt: set L2 Rx mask")
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Memory zones should be freed using the proper memzone_free function not
rte_free which is for malloc calls.
After allocating memzone it's required to zeroize memory in it, so do so
before storing the handle for later freeing.
Fixes: 9ba7981ec9 ("ena: add communication layer for DPDK")
Signed-off-by: Alexander Matushevsky <matua@amazon.com>
Signed-off-by: Jakub Palider <jpa@semihalf.com>
Signed-off-by: Jan Medala <jan@semihalf.com>
While allocating (coherent) memory, get information about calling node Id
and pass that as parameter when reserving a memzone.
Signed-off-by: Alexander Matushevsky <matua@amazon.com>
Signed-off-by: Jakub Palider <jpa@semihalf.com>
Signed-off-by: Jan Medala <jan@semihalf.com>
Depending on HW revision readless communcation between host and device
may be unavailable. In that case prevent PMD from setting up readless
communication mechanism.
"readless" refers to ability to read ENA registers without actually
issuing read request from host (x86). Instead, host programs 2 registers
on the device that triggers a DMA from device to host and reports a
register value. However, this functionality is not going to be available
in all types of devices. The decision if this mode is supported or not,
is taken from revision_id in pci configuration space.
Signed-off-by: Alexander Matushevsky <matua@amazon.com>
Signed-off-by: Jakub Palider <jpa@semihalf.com>
Signed-off-by: Jan Medala <jan@semihalf.com>
Increase maintenance and debug potentiality with dedicated areas of memory
where additional information can be stored by the ENA device.
Signed-off-by: Alexander Matushevsky <matua@amazon.com>
Signed-off-by: Jakub Palider <jpa@semihalf.com>
Signed-off-by: Jan Medala <jan@semihalf.com>
Synchronize ENA communication layer with latest ENA FW version.
Signed-off-by: Alexander Matushevsky <matua@amazon.com>
Signed-off-by: Jakub Palider <jpa@semihalf.com>
Signed-off-by: Jan Medala <jan@semihalf.com>
If you stop, then start a port that had already received some packets,
the NIC could fetch discriptors from the wrong location. This could
effectivly reduce the size of the Rx queue by a random amount and
cause packet drop or reduced performance.
Reset the NIC fetch index to 0 when allocating and posting mbuf
addresses to the NIC.
Fixes: 947d860c82 ("enic: improve Rx performance")
Signed-off-by: John Daley <johndale@cisco.com>
Reviewed-by: Nelson Escobar <neescoba@cisco.com>
If you stop then start a port that had already sent some packets,
there was a segfault due to not resetting the number of completed
sends to zero.
Fixes: a3b1e9551c ("net/enic: streamline mbuf handling in Tx path")
Signed-off-by: Nelson Escobar <neescoba@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>
Fixes: 62072098b5 ("mlx5: support setting link up or down")
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
