Wei Dai [Mon, 19 Jun 2017 04:14:38 +0000 (12:14 +0800)]
lpm: fix index of tbl8
From v20 to v1604, number of tbl8 can be up to 1<<24,
(uint8_t) or (uint16_t) may truncate the number of
index of tlb8 in v1604 and cause wrong number.
Fixes: dc81ebbacaeb ("lpm: extend IPv4 next hop field") Cc: stable@dpdk.org Signed-off-by: Wei Dai <wei.dai@intel.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Since vhost_user_set_features failure is not handled in any way, a
single error log has been added to at least to let the user know that
something has gone wrong.
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com> Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Yuanhan Liu [Fri, 2 Jun 2017 00:14:46 +0000 (08:14 +0800)]
vhost: fix crash on NUMA
The queue allocation was changed, from allocating one queue-pair at a
time to one queue at a time. Most of the changes have been done, but
just with one being missed: the size of copying the old queue is still
based on queue-pair at numa_realloc(), which leads to overwritten issue.
As a result, crash may happen.
Fix it by specifying the right copy size. Also, the net queue macros
are not used any more. Remove them.
Daniel Verkamp [Fri, 26 May 2017 11:59:15 +0000 (13:59 +0200)]
vhost: access VhostUsrMsg via packed struct
Accessing fields of a packed struct through unaligned pointers is
undefined behavior. Instead of passing pointers to particular fields,
a pointer to the root struct should be used. This patch does exactly
that.
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com> Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com> Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Thomas Monjalon [Thu, 15 Jun 2017 17:37:00 +0000 (19:37 +0200)]
mem: support page locking on FreeBSD
The function rte_mem_lock_page() was added for Linux only.
The file eal_common_memory.c is a better place to make it
available in FreeBSD also.
The issue is seen when trying to compile bnxt on FreeBSD:
bnxt_hwrm.c: undefined reference to `rte_mem_lock_page'
Fixes: 3097de6e6bfb ("mem: get physical address of any pointer") Reported-by: Fangfang Wei <fangfangx.wei@intel.com> Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Adrien Mazarguil [Thu, 15 Jun 2017 15:48:59 +0000 (17:48 +0200)]
ethdev: tidy up endianness handling in flow API
The flow API defines several structures whose fields must be specified in
network order. This commit documents them using explicit type names and
related endianness conversion macros.
Adrien Mazarguil [Thu, 15 Jun 2017 15:48:58 +0000 (17:48 +0200)]
eal: add static endianness conversion macros
These macros resolve to constant expressions that allow developers to
perform endianness conversion on static/const objects, even outside of
function scope as they do not translate to function calls.
This is most useful for static initializers and constant values (whenever
it has to be performed at compilation time). Run-time endianness conversion
of variable values should keep using rte_*_to_*() calls for best
performance.
Nelio Laranjeiro [Thu, 15 Jun 2017 15:48:57 +0000 (17:48 +0200)]
eal: introduce big and little endian types
This commit introduces new rte_{le,be}{16,32,64}_t types and updates
rte_{le,be,cpu}_to_{le,be,cpu}_*() accordingly.
These types are added for documentation purposes, mainly to clarify the
byte ordering to use for storage when not CPU order. Doing so eliminates
uncertainty and conversion mistakes.
Thomas Monjalon [Thu, 15 Jun 2017 09:25:06 +0000 (11:25 +0200)]
app/testpmd: fix build with bypass without ixgbe
When ixgbe bypass is not explicitly disabled while ixgbe is disabled:
app/test-pmd/testpmd.c:304:27: error:
‘RTE_PMD_IXGBE_BYPASS_TMT_OFF’ undeclared here
The ixgbe bypass feature is meaningful only if ixgbe is enabled.
So we need to check both.
A best fix will be to enable bypass always and remove this option.
Fixes: e261265e42a1 ("ethdev: move bypass functions to ixgbe PMD") Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Thomas Monjalon [Thu, 15 Jun 2017 09:34:16 +0000 (11:34 +0200)]
app/testpmd: fix build without ixgbe
cmd_set_vf_rxmode_parsed() was defined only in the build context
of RTE_LIBRTE_IXGBE_PMD:
app/test-pmd/cmdline.c:13817:27: error: ‘cmd_set_vf_rxmode’ undeclared here
Fixes: 4cfe399f6550 ("net/bnxt: support to set VF rxmode") Reported-by: Yongseok Koh <yskoh@mellanox.com> Reported-by: Jan Viktorin <viktorin@rehivetech.com> Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Adrien Mazarguil [Wed, 14 Jun 2017 14:48:51 +0000 (16:48 +0200)]
ethdev: add isolated mode to flow API
Isolated mode can be requested by applications on individual ports to avoid
ingress traffic outside of the flow rules they define.
Besides making ingress more deterministic, it allows PMDs to safely reuse
resources otherwise assigned to handle the remaining traffic, such as
global RSS configuration settings, VLAN filters, MAC address entries,
legacy filter API rules and so on in order to expand the set of possible
flow rule types.
To minimize code complexity, PMDs implementing this mode may provide
partial (or even no) support for flow rules when not enabled (e.g. no
priorities, no RSS action). Applications written to use the flow API are
therefore encouraged to enable it.
Once effective, leaving isolated mode may not be possible depending on PMD
implementation.
Ferruh Yigit [Tue, 13 Jun 2017 16:42:12 +0000 (17:42 +0100)]
hash: fix icc build
build error with icc version 17.0.4 (gcc version 7.0.0 compatibility):
In file included from .../dpdk/lib/librte_hash/rte_fbk_hash.h(59),
from .../dpdk/lib/librte_hash/rte_fbk_hash.c(54):
.../dpdk/x86_64-native-linuxapp-icc/include/rte_hash_crc.h(480):
error #1292: unknown attribute "fallthrough"
__attribute__ ((fallthrough));
^
In file included from .../dpdk/lib/librte_hash/rte_fbk_hash.h(59),
from .../dpdk/lib/librte_hash/rte_fbk_hash.c(54):
.../dpdk/x86_64-native-linuxapp-icc/include/rte_hash_crc.h(486):
error #1292: unknown attribute "fallthrough"
__attribute__ ((fallthrough));
^
This code patch hit when gcc > 7 installed and ICC doesn't recognize
fallthrough attribute.
Ferruh Yigit [Tue, 13 Jun 2017 16:42:11 +0000 (17:42 +0100)]
net/mlx5: fix build with gcc 7.1
build error:
.../dpdk/drivers/net/mlx5/mlx5_fdir.c:
In function ‘fdir_filter_to_flow_desc’:
.../dpdk/drivers/net/mlx5/mlx5_fdir.c:146:18:
error: this statement may fall through [-Werror=implicit-fallthrough=]
desc->dst_port = fdir_filter->input.flow.udp4_flow.dst_port;
~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
.../dpdk/drivers/net/mlx5/mlx5_fdir.c:147:2: note: here
case RTE_ETH_FLOW_NONFRAG_IPV4_OTHER:
^~~~
Ferruh Yigit [Tue, 13 Jun 2017 16:42:10 +0000 (17:42 +0100)]
net/enic: fix build with gcc 7.1
build error:
.../dpdk/drivers/net/enic/base/vnic_dev.c:
In function ‘vnic_dev_get_mac_addr’:
.../dpdk/drivers/net/enic/base/vnic_dev.c:470:12:
error: ‘a0’ is used uninitialized in this function
[-Werror=uninitialized]
args[0] = *a0;
^~~
...dpdk/drivers/net/enic/base/vnic_dev.c:
In function ‘vnic_dev_classifier’:
...dpdk/drivers/net/enic/base/vnic_dev.c:471:12:
error: ‘a1’ may be used uninitialized in this function
[-Werror=maybe-uninitialized]
args[1] = *a1;
^~~
Fixed by providing initial values.
Ferruh Yigit [Tue, 13 Jun 2017 16:42:09 +0000 (17:42 +0100)]
net/i40e: fix memset size
This causes build error with gcc 7.1.1 :
...dpdk/drivers/net/i40e/i40e_flow.c:2357:2:
error: ‘memset’ used with length equal to number of elements without
multiplication by element size [-Werror=memset-elt-size]
memset(off_arr, 0, I40E_MAX_FLXPLD_FIED);
^~~~~~
...dpdk/drivers/net/i40e/i40e_flow.c:2358:2:
error: ‘memset’ used with length equal to number of elements without
multiplication by element size [-Werror=memset-elt-size]
memset(len_arr, 0, I40E_MAX_FLXPLD_FIED);
^~~~~~
Fixed by providing correct size to memset.
Fixes: 6ced3dd72f5f ("net/i40e: support flexible payload parsing for FDIR") Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com> Acked-by: Beilei Xing <beilei.xing@intel.com>
Ferruh Yigit [Tue, 13 Jun 2017 16:42:08 +0000 (17:42 +0100)]
kni: fix build with gcc 7.1
build error:
.../dpdk/build/build/lib/librte_eal/linuxapp/kni/igb_main.c:
In function ‘igb_kni_probe’:
.../dpdk/build/build/lib/librte_eal/linuxapp/kni/igb_main.c:2483:30:
error: ‘%d’ directive output may be truncated writing between 1 and 5
bytes into a region of size between 0 and 11
[-Werror=format-truncation=]
"%d.%d, 0x%08x, %d.%d.%d",
^~
.../dpdk/build/build/lib/librte_eal/linuxapp/kni/igb_main.c:2483:8:
note: directive argument in the range [0, 65535]
"%d.%d, 0x%08x, %d.%d.%d",
^~~~~~~~~~~~~~~~~~~~~~~~~
.../dpdk/build/build/lib/librte_eal/linuxapp/kni/igb_main.c:2481:4:
note: ‘snprintf’ output between 23 and 43 bytes into a destination of
size 32
snprintf(adapter->fw_version,
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
sizeof(adapter->fw_version),
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
"%d.%d, 0x%08x, %d.%d.%d",
~~~~~~~~~~~~~~~~~~~~~~~~~~
fw.eep_major, fw.eep_minor, fw.etrack_id,
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
fw.or_major, fw.or_build, fw.or_patch);
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Fixed by increasing buffer size to 43 as suggested in compiler log.
Jerin Jacob [Fri, 9 Jun 2017 10:27:46 +0000 (15:57 +0530)]
net/thunderx: manage PCI device mapping for SQS VFs
Since the commit e84ad157b7bc ("pci: unmap resources if probe fails"),
EAL unmaps the PCI device if ethdev probe returns positive or
negative value.
nicvf thunderx PMD needs special treatment for Secondary queue set(SQS)
PCIe VF devices, where, it expects to not unmap or free the memory
without registering the ethdev subsystem.
Enable the same behavior by using RTE_PCI_DRV_KEEP_MAPPED_RES
PCI driver flag.
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com> Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Jerin Jacob [Fri, 9 Jun 2017 10:27:45 +0000 (15:57 +0530)]
eal/pci: introduce a PCI driver flag
Some ethdev devices like nicvf thunderx PMD need special treatment for
Secondary queue set(SQS) PCIe VF devices, where, it expects to not unmap
or free the memory without registering the ethdev subsystem.
Introducing a new RTE_PCI_DRV_KEEP_MAPPED_RES
PCI driver flag to request PCI subsystem to not unmap the mapped PCI
resources(PCI BAR address) if unsupported device detected.
Ferruh Yigit [Mon, 12 Jun 2017 15:25:12 +0000 (16:25 +0100)]
ethdev: remove driver name from device private data
rte_driver->name has the driver name and all physical and virtual
devices has access to it.
Previously it was not possible for virtual ethernet devices to access
rte_driver->name field (because eth_dev used to keep only pci_dev),
and it was required to save driver name in the device private struct.
After re-works on bus and vdev, it is possible for all bus types to
access rte_driver.
It is able to remove the driver name from ethdev device private data and
use eth_dev->device->driver->name.
Matej Vido [Mon, 12 Jun 2017 12:03:20 +0000 (14:03 +0200)]
net/szedata2: refactor ibuf and obuf address definition
This is to prepare for firmwares with multiple ibufs and obufs.
Ibufs and obufs are the modules in FPGA firmware implementing
the Ethernet port.
There is one ibuf+obuf per Ethernet port.
The cards and firmwares allow one physical port to be one Ethernet
port or split into more Ethernet ports, e.g. one 100GE physical
port can be one Ethernet port of 100GE or split into ten Ethernet
ports of 10GE.
All DMA queues in the device are shared between all Ethernet ports.
Offsets of ibufs and obufs are defined in array.
Functions which operate on ibufs and obufs iterate over this array.
Matej Vido [Mon, 12 Jun 2017 12:03:18 +0000 (14:03 +0200)]
net/szedata2: refactor ibuf and obuf names
Prefix "cgmii" is removed because it is too specific.
There are different ibuf/obuf modules in different firmwares
but the address space definition is the same.
This patch makes the name general.
Beilei Xing [Fri, 9 Jun 2017 08:21:23 +0000 (16:21 +0800)]
net/i40e: support ether pattern for FDIR
Previously, i40e PMD will select ethertype filter
parser when adding ether pattern rules. In fact,
FDIR also supports ether pattern.
This patch adds ether pattern support for FDIR.
Signed-off-by: Beilei Xing <beilei.xing@intel.com> Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
Harish Patil [Wed, 7 Jun 2017 07:42:22 +0000 (00:42 -0700)]
net/qede: refactor Tx routine
Refactor TX routine such that TX BD updates can all be grouped together.
Based on the TX offloads requested the TX bitfields are calculated in
a temporary variable and TX BDs are updated at the end. This will minimize
the if checks also. This change is done to easily accommodate newer TX
offload operations in the future.
Harish Patil [Wed, 7 Jun 2017 07:42:19 +0000 (00:42 -0700)]
net/qede: refactoring multi-queue implementation
This patch does the following refactoring and cleanup:
- As part of multi-queue support a struct member called 'type' was added
in struct qede_fastpath in order to identify whether a queue is RX or
TX and take actions based on that. This was unnecessary in the first
place since pointers to RX and TX queues are already available in
rte_eth_dev->data. So all usage of fp->type is removed.
- Remove remaining additional layer of internal callbacks for RX/TX
queues and fastpath related operations from the qed_eth_ops_pass.
With this change the files qede_eth_if.[c,h] are no longer needed.
- Add new per-queue start/stop APIs instead of clubbing it all together.
- Remove multiple TXQs references (num_tc and fp->txqs) since CoS is not
supported.
- Enable sharing of the status block for each queue pair.
- Remove enum qede_dev_state and instead make use of existing port
states RTE_ETH_QUEUE_STATE_STOPPED/RTE_ETH_QUEUE_STATE_STARTED.
- Move qede_dev_start() and qede_dev_stop() to qede_ethdev.c from
qede_rxtc.c.
Harish Patil [Wed, 7 Jun 2017 07:42:18 +0000 (00:42 -0700)]
net/qede: refactoring vport handling code
The refactoring is mainly for two reasons:
- To remove an additional layer of internal callbacks for all vport
related operations from the struct qed_eth_ops_pass. Instead, we
can invoke base APIs directly.
- Splitting a single large vport-update configuration into multiple and
independent vport-update operations. Each configuration would touch
only the required config bits that needs an update.
Ajit Khaparde [Thu, 1 Jun 2017 17:07:21 +0000 (12:07 -0500)]
net/bnxt: configure a default VF VLAN
This patch adds code to insert a default VF VLAN.
Also track the current default VLAN per vnic for the VF.
When setting the default VLAN, avoid setting it to the current value.
Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Ajit Khaparde [Thu, 1 Jun 2017 17:07:19 +0000 (12:07 -0500)]
net/bnxt: determine the Rx status of VF
This patch adds code to determine the Rx status of a VF.
It adds the rte_pmd_bnxt_get_vf_rx_status call, which calculates
the VNIC count of the function to get the Rx status.
Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Ajit Khaparde [Thu, 1 Jun 2017 17:07:16 +0000 (12:07 -0500)]
net/bnxt: support Tx loopback, set VF MAC and queues drop
Add functions rte_pmd_bnxt_set_tx_loopback,
rte_pmd_bnxt_set_all_queues_drop_en and
rte_pmd_bnxt_set_vf_mac_addr to configure tx_loopback,
queue_drop and VF MAC address setting in the hardware.
It also adds the necessary functions to send the HWRM commands
to the firmware.
Signed-off-by: Steeven Li <steeven.li@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Ajit Khaparde [Thu, 1 Jun 2017 17:07:13 +0000 (12:07 -0500)]
net/bnxt: refactor the query stats
1) Use hwrm_stat_ctx_query command to query statistics
Using hwrm_stat_ctx_query command will allow polling
the statistics from hardware instead of using the current push
model from the hardware which does a DMA of the stats to the host
at fixed intervals.
2) Use the rx_mbuf_alloc_fail to track mbuf alloc failures.
3) We were wrongly incrementing hwrm_cmd_seq in bnxt_hwrm_stat_clear
and bnxt_hwrm_stat_ctx_alloc functions. This patch fixes that.
Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Ajit Khaparde [Thu, 1 Jun 2017 17:07:10 +0000 (12:07 -0500)]
net/bnxt: support LRO
This patch adds support to enable and disable LRO
To support this feature, the driver creates an aggregator ring.
When the hardware starts doing LRO, it sends a tpa_start completion.
When the driver receives a tpa_end completion, it indicates that the
LRO chaining is complete.
Signed-off-by: Steeven Li <steeven.li@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Ajit Khaparde [Thu, 1 Jun 2017 17:07:09 +0000 (12:07 -0500)]
net/bnxt: support set MTU
This patch adds support to modify MTU using the set_mtu dev_op.
To support frames > 2k, the PMD creates an aggregator ring.
When a frame greater than 2k is received, it is fragmented
and the resulting fragments are DMA'ed to the aggregator ring.
Now the driver can support jumbo frames upto 9500 bytes.
Signed-off-by: Steeven Li <steeven.li@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Ajit Khaparde [Thu, 1 Jun 2017 17:07:05 +0000 (12:07 -0500)]
net/bnxt: support VLAN filter and strip
This patch adds VLAN strip and offload callbacks.
To add a VLAN filter:
For each VNIC and each associated filter(s)
if VLAN exists:
if VLAN matches vlan_id
VLAN filter already exists, just skip and continue
else
add a new MAC+VLAN filter
else
Remove the old MAC only filter
Add a new MAC+VLAN filter
To remove a VLAN filter:
For each VNIC and each associated filter(s)
if VLAN exists && VLAN matches vlan_id
remove the MAC+VLAN filter
add a new MAC only filter
else
VLAN filter doesn't exist, just skip and continue
Ajit Khaparde [Thu, 1 Jun 2017 17:07:03 +0000 (12:07 -0500)]
net/bnxt: support tunneling
Add support for udp_tunnel_port_add/del dev_ops to configure a UDP port
for VXLAN and Geneve Tunnel protocols.
The HWRM supports only one global destination port for a tunnel type,
use a reference counter to keep track of its usage.
Cache the configured VXLAN/Geneve ports and use that value to check
if the right UDP port is being freed up.
Skip calling bnxt_hwrm_tunnel_dst_port_alloc if the same UDP port is
being programmed.
Skip calling bnxt_hwrm_tunnel_dst_port_free if no UDP port has been
configured.
1) For a VF, query the firmware to determine if a MAC address is
already configured. If not configure a random default MAC address.
2) Do not initialize the default completion ring in
bnxt_alloc_hwrm_rings().
3) While registering for async events with the firmware,
use func_vf_cfg for a VF and use func_cfg for a PF.
4) Query the VNIC plcmode config using the bnxt_hwrm_vnic_plcmodes_qcfg
before a VNIC is updated. Reconfigure the VNIC with the plcmode
configuration queried earlier. Not doing this could overwrite
the plcmodes in some cases.
5) Reorg the bnxt_handle_fwd_req to properly handle the forwarded
requests. The previous code did not handle it completely.
Ajit Khaparde [Thu, 1 Jun 2017 17:06:59 +0000 (12:06 -0500)]
net/bnxt: refactor to properly allocate resources for PF/VF
1) Move the function reset to bnxt_dev_init.
On the same lines, setup, enable and request interrupt to init path.
Memory allocation is also being done in the init path.
2) After a function reset, configure the VFs. Distribute resources
evenly between all functions (PF and VF) for now. In the future, this
should be controllable.
3) The bnxt_vf_info and bnxt_pf_info had lot of duplication. Move the
common items to struct bnxt. And only unique items specific to PF
remain in the struct bnxt_pf_info.
4) Program the firmware to allow certain commands sent by a VF.
Disallowing these will prevent clean VF driver cleanup.
5) Since PF/VF need to allocate resources from a pool in the hardware,
use func_qcaps and func_qcfg to appropriately query the capabilities
and available resources.
6) If a PF is being initialized and no VFs are allocated, explicitly
call func_cfg to allocate the resources.
7) Once resources are requested from the firmware, update local copy
of resource count in struct bnxt only after sending the func_qcfg to
make sure the allocation request in the firmware went through.
The changes in this patch will be used by the subsequent patches
to allow proper initialization of PF/VF instance.
Ajit Khaparde [Thu, 1 Jun 2017 17:06:58 +0000 (12:06 -0500)]
net/bnxt: update to HWRM version 1.7.7
This patch updates the Broadcom bnxt PMD to version 1.7.7
Most of the changes in the patch are in the hsi_struct_def_dpdk.h - an
autogenerated file. The changes in the *.c files are because of changes
in the macro names.
Yongseok Koh [Wed, 31 May 2017 17:48:45 +0000 (10:48 -0700)]
net/mlx5: fix redundant free of Tx buffer
SW completion ring of Tx (txq->elts) stores individual mbufs even if a
multi-segmented packet is sent. rte_pktmbuf_free_seg() must be used when
cleaning up the completion ring. Otherwise, chained mbufs are redundantly
freed and finally it would cause a crash.
Andrew Rybchenko [Fri, 26 May 2017 13:50:44 +0000 (14:50 +0100)]
net/sfc: rely on one desc is one mbuf in simple EF10 Tx reap
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com> Reviewed-by: Andy Moreton <amoreton@solarflare.com> Reviewed-by: David Riddoch <driddoch@solarflare.com>
git.droids-corp.org / dpdk.git / log