On X722, we can control whether or not the hardware performs ATR
eviction. Define the correct bit so we can twiddle it.
Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
When polling the bonded ports for RX packets the old driver would
always start with the first slave in the list. If the requested
number of packets is filled on the first port in a two port config
then the second port could be starved or have larger number of
missed packet errors.
The code attempts to start with a different slave each time RX poll
is done to help eliminate starvation of slave ports. The effect of
the previous code was much lower performance for two slaves in the
bond then just the one slave.
The performance drop was detected when the application can not poll
the rings of Rx packets fast enough and the packets per second for
two or more ports was at the threshold throughput of the application.
At this threshold the slaves would see very little or no drops in
the case of one slave. Then enable the second slave you would see
a large drop rate on the two slave bond and reduction in throughput.
Signed-off-by: Keith Wiles <keith.wiles@intel.com>
Acked-by: Declan Doherty <declan.doherty@intel.com>
Current code enables RX interrupts even if this it not
requested.
Fixes: ea121b2831 ("net/nfp: add Rx interrupts")
Cc: stable@dpdk.org
Signed-off-by: Alejandro Lucero <alejandro.lucero@netronome.com>
Chained mbufs hold data_len as the length of that particular mbuf
and pkt_len as the full packet length including all the chained
mbufs. It is not clear from the mbuf definition if pkt_len should
be set for all the mbufs in a chain, but code there for handling
mbufs suggests just the first mbuf requires to have pkt_len set.
NFP PMD was assuming pkt_len is set in all the chained mbufs and
unit tests for gather dma were building mbufs with pkt_len always
set. This patch gets rid of that assumption.
Fixes: b812daadad ("nfp: add Rx and Tx")
Cc: stable@dpdk.org
Signed-off-by: Alejandro Lucero <alejandro.lucero@netronome.com>
When LSO, not doing this can led to firmware disruption. It does
not show as error because TCP ends up sending data again later on.
Fixes: 9ba3d0ae20 ("net/nfp: add TSO support")
Cc: stable@dpdk.org
Signed-off-by: Alejandro Lucero <alejandro.lucero@netronome.com>
Typo in ecore_sriov.c; Ending line with , instead of ;
Fixes: 379cbb2c44 ("net/qede/base: semantic change")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
When a server doesn't support ARI, VF offsets begin at a much higher
number. As a result, ecore miscalculates the first_vf_in_pf and
initialization fails since base driver incorrectly learns there are
no SBs for its VF [as its VFs are out of range].
Fixes: 22d07d939c ("net/qede/base: update")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
Remove the unneeded conversion to LE when writing to the 32-bit
XSDM_REG_OPERATION_GEN register
Fixes: ec94dbc573 ("qede: add base driver")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
VFs are seeing the number of MACs available to them as '0',
and as a result configure themselves as PROMISC. This fix is to
prevent that.
Fixes: 86a2265e59 ("qede: add SRIOV support")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
Fix the logic for identifying which bit amongst the Multi-bit
attention sources is set.
Fixes: e6051bd6b0 ("qede: add interrupt handling support")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
There are some constellations where Due to lack of resource allocation
in MFW, There would be an insufficient number of L2 queues for all the
VFs.
This introduces a new feature ECORE_VF_L2_QUE which correctly numbers
the number of VF queues. Notice it might be larger than the actual
number of VFs in configuration space, in which case its the ecore
client responsibility not to try activating that many.
As part of the fix, also correct the nubmering of the VF queues. As
their numbering is dependent on the SBs of the PF, which might only be
partially used by L2 [as half would be assigned for RDMA which doesn't
require L2 queues], we make the numbering consecutive with that of the
L2 queues only.
Fixes: ec94dbc573 ("qede: add base driver")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
Prints in ecore_get_dev_info showed only chip revision,
and did that as number instead of letter.
I.e., BB A0 --> BB0, BB B0 --> BB1, AH A0 --> AH0, AH A1 --> AH0.
Correct the printing scheme into
{AH, BB} {A, B}{0, 1}
Fixes: ec94dbc573 ("qede: add base driver")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
Fix Timer(TM) block Internal Lookup Table(or ILT for logical to
physical address translation) initialization for SRIOV's coexistence
with other protocols.
Fixes: ec94dbc573 ("qede: add base driver")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
Remove the forcing of the driver's default resource allocation.
Fixes: 77f7222124 ("net/qede: add PCI ids for new chip variant")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
Set pointers to NULL after freeing the allocations. Change OSAL_FREE
macro to take care of this and cleanup relevant code.
Fixes: 26ae839d06 ("qede: add DCBX support")
Fixes: ec94dbc573 ("qede: add base driver")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
dcbx-update-flag is incorrectly converted to boolean before assigining
it to ramrod data, fix this typecasting. Also, added more debug
messages in the dcbx code paths.
Fixes: 26ae839d06 ("qede: add DCBX support")
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
Split control and datapath to make datapath substitutable and
possibly reusable with alternative control path.
libefx-based Tx datapath is bound to libefx control path, but
it should be possible to use other datapaths with alternative
control path(s).
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
If Rx refill threshold guarantees that refill happens for one or
more bulks, less checks may be done on refill.
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Split control and datapath to make datapath substitutable and
possibly reusable with alternative control path.
libefx-based Rx datapath is bound to libefx control path, but
other datapaths should be possible to use with alternative
control path(s).
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Rx queue flags should keep the information required on datapath.
It is a preparation to split control and data paths.
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Use different sets of libefx EvQ callbacks for management,
transmit and receive event queue. It makes event handling
more robust against unexpected events.
Also it is required for alternative datapath support.
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Adding support for the next items: eth, vlan, ipv4, udp, tcp and for the
next actions: queue, drop
Signed-off-by: Vasily Philipov <vasilyf@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Make priv_lock/priv_unlock functions and some other structs/defines visible
from different source files by placing them into mlx4.h header.
Signed-off-by: Vasily Philipov <vasilyf@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
If LSC interrupts are enabled, the application expects the link_update
ops to be executed by the PMD itself.
No link status change event is received upon probing, therefore the link
status update must be forced.
Fixes: c4da6caa42 ("mlx4: handle link status interrupts")
Cc: stable@dpdk.org
Signed-off-by: Gaetan Rivet <gaetan.rivet@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Enable Thunderx nicvf PMD driver in the common
config as it does not have build dependency
with any external library and/or architecture.
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
SIMPLEQ_* operations are not available in FreeBSD. Replacing
with equivalent STAILQ_* operations.
Fixes: f2546f8e51 ("net/thunderx/base: add functions to store qsets")
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
ConnectX-5 supports enhanced version of multi-packet send (MPS). An MPS Tx
descriptor can carry multiple packets either by including pointers of
packets or by inlining packets. Inlining packet data can be helpful to
better utilize PCIe bandwidth. In addition, Enhanced MPS supports hybrid
mode - mixing inlined packets and pointers in a descriptor. This feature is
enabled by default if supported by HW.
Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Some libefx-based drivers might need this functionality to
indicate DPCPU FW IDs as part of FW version info to assist
experienced users.
Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Andy Moreton <amoreton@solarflare.com>
