Daniel Mrzyglod [Tue, 7 Jul 2015 08:09:33 +0000 (10:09 +0200)]
examples/ip_pipeline: rework firewall pipeline
Firewall pipeline implementation is split to two files.
pipeline_firewall.c file handles front-end functions (cli commands
parsing) pipeline_firewall_ops.c contains implementation of functions
done by pipeline (back-end).
Signed-off-by: Daniel Mrzyglod <danielx.t.mrzyglod@intel.com> Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Passthrough pipeline implementation is split to two files.
pipeline_passthrough.c file handles front-end functions (cli commands
parsing) pipeline_passthrough_ops.c contains implementation of functions
done by pipeline (back-end).
Master pipeline is responsible for command line handling and
communicationg with all other pipelines via message queues. Removed
cmdline.c file, as its functionality will be split over multiple
pipeline files.
New syntax of config files is needed for ip_pipeline example
enhancements. Some old files are temporarily disabled in the Makefile.
It is part of a bigger change.
David Marchand [Thu, 9 Jul 2015 09:19:26 +0000 (11:19 +0200)]
eal/linux: fix out of bound access in hugepage init
Using IBM advance toolchain on Ubuntu 14.04 (package 8.0-3), gcc is complaining
about out of bound accesses.
CC eal_hugepage_info.o
lib/librte_eal/linuxapp/eal/eal_hugepage_info.c:
In function ‘eal_hugepage_info_init’:
lib/librte_eal/linuxapp/eal/eal_hugepage_info.c:350:35:
error: array subscript is above array bounds [-Werror=array-bounds]
internal_config.hugepage_info[j].hugepage_sz)
^
lib/librte_eal/linuxapp/eal/eal_hugepage_info.c:350:35:
error: array subscript is above array bounds [-Werror=array-bounds]
lib/librte_eal/linuxapp/eal/eal_hugepage_info.c:349:37:
error: array subscript is above array bounds [-Werror=array-bounds]
if (internal_config.hugepage_info[j-1].hugepage_sz <
^
lib/librte_eal/linuxapp/eal/eal_hugepage_info.c:350:35:
error: array subscript is above array bounds [-Werror=array-bounds]
internal_config.hugepage_info[j].hugepage_sz)
Looking at the code, these warnings are invalid from my pov and they disappeared
when upgrading the toolchain to new version (8.0-4).
However, the code was buggy (sorting code is wrong), so fix this by using qsort
and adding a check on num_sizes to avoid potential out of bound accesses.
Signed-off-by: David Marchand <david.marchand@6wind.com> Acked-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
David Marchand [Thu, 9 Jul 2015 09:19:22 +0000 (11:19 +0200)]
eal/linux: remove useless casts in hugepage init
Rather than cast the huge pages number returned by get_num_hugepages, rework
this function so that it returns 0 when something goes wrong.
And no need for casts in log.
Signed-off-by: David Marchand <david.marchand@6wind.com> Acked-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
Bruce Liu [Wed, 8 Jul 2015 15:15:47 +0000 (23:15 +0800)]
hash: fix build on non-x86 arch
In the no-x86 branch, a parentheses is missing.
And need to include <limits.h> header to use CHAR_BIT.
Fixes: 49361c3f3cfa ("hash: remove duplicated code") Signed-off-by: Bruce Liu <2103458176@qq.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Thomas Monjalon [Thu, 2 Jul 2015 21:03:40 +0000 (23:03 +0200)]
mk: enable next abi preview
When a change makes really hard to keep ABI compatibility,
instead of waiting next release to break the ABI, it is smoother
to introduce the new code as a preview and disable it when packaging.
The flag RTE_NEXT_ABI must be used to "ifdef" the new code.
When the release is out, a dynamically linked application can use
the new shared libraries with the old ABI while developpers can prepare
their application for the next ABI by reading the deprecation notice
and easily testing the new code.
When starting the next release cycle, the "ifdefs" will be removed
and the ABI break will be marked by incrementing LIBABIVER. The map
files will also be updated.
The default value is enabled to be developer compliant.
The packagers must disable it as done in pkg/dpdk.spec.
When enabled, all shared library numbers are incremented by appending
a minor .1 to the old ABI number. In the next release, only impacted
libraries will have a major +1 increment.
The impacted libraries must provide an alternative map file to use
with this option.
The ABI policy is updated.
Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com> Acked-by: Neil Horman <nhorman@tuxdriver.com>
Thomas Monjalon [Wed, 8 Jul 2015 10:34:26 +0000 (12:34 +0200)]
mk: remove variables identical to config ones
CONFIG_RTE_BUILD_SHARED_LIB and CONFIG_RTE_BUILD_COMBINE_LIBS does not
have quotes in their values (only y or n). That's why the variables
RTE_BUILD_SHARED_LIB and RTE_BUILD_COMBINE_LIBS are always identical to
their CONFIG_ counterpart, and are useless.
In order to have consistent naming of config options in the makefiles,
these options are removed and the "CONFIG_ prefixed" variables are used.
Fixes: e25e4d7ef16b ("mk: shared libraries") Fixes: 4d3d79e7a5c6 ("mk: combined library") Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com> Acked-by: Neil Horman <nhorman@tuxdriver.com>
Raz Amir [Thu, 5 Mar 2015 16:24:23 +0000 (18:24 +0200)]
nic_uio: probe and attach devices after unload
Added code that saves the pointers to the detached devices, during
driver loading, and during driver unloading, go over the list,
and re-attach them by calling device_probe_and_attach
on each device.
Signed-off-by: Raz Amir <razamir22@gmail.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Tetsuya Mukawa [Tue, 30 Jun 2015 08:26:51 +0000 (17:26 +0900)]
eal: enable port hotplug as default for linux and bsd
This patch removes CONFIG_RTE_LIBRTE_EAL_HOTPLUG option, and enables it
as default in both Linux and BSD.
Also, to support port hotplug, rte_eal_pci_scan() and below missing
symbols should be exported to ethdev library.
- rte_eal_parse_devargs_str()
- rte_eal_pci_close_one()
- rte_eal_pci_probe_one()
- rte_eal_pci_scan()
- rte_eal_vdev_init()
- rte_eal_vdev_uninit()
The patch consolidates below functions, and implement these
in eal_common_pci_uio.c.
- pci_uio_map_secondary()
- pci_uio_map_resource()
- pci_uio_unmap()
- pci_uio_find_resource()
- pci_uio_unmap_resource()
Signed-off-by: Tetsuya Mukawa <mukawa@igel.co.jp> Acked-by: David Marchand <david.marchand@6wind.com>
pci: introduce function to map uio resource by index
This patch adds a new function called pci_uio_map_resource_by_index().
The function hides how to map uio resource in linuxapp and bsdapp.
With the function, pci_uio_map_resource() will be more abstracted.
Signed-off-by: Tetsuya Mukawa <mukawa@igel.co.jp> Acked-by: David Marchand <david.marchand@6wind.com>
pci: introduce functions to alloc and free uio resource
This patch adds new functions called pci_uio_alloc_resource() and
pci_uio_free_resource().
The functions hides how to prepare or free uio resource in linuxapp
and bsdapp. With the function, pci_uio_map_resource() will be more
abstracted.
Signed-off-by: Tetsuya Mukawa <mukawa@igel.co.jp> Acked-by: David Marchand <david.marchand@6wind.com>
pci: fix uio mapping differences between linux and bsd
This patch fixes below.
- bsdapp
- Use map_id in pci_uio_map_resource().
- Fix interface of pci_map_resource().
- Move path variable of mapped_pci_resource structure to pci_map.
- linuxapp
- Remove redundant error message of linuxapp.
'pci_uio_map_resource()' is implemented in both linuxapp and bsdapp,
but interface is different. The patch fixes the function of bsdapp
to do same as linuxapp. After applying it, file descriptor should be
opened and closed out of pci_map_resource().
Signed-off-by: Tetsuya Mukawa <mukawa@igel.co.jp> Acked-by: Bernard Iremonger <bernard.iremonger@intel.com> Acked-by: David Marchand <david.marchand@6wind.com>
To merge pci code of linuxapp and bsdapp, this patch changes names
like below.
- uio_map to pci_map
- uio_resource to mapped_pci_resource
- uio_res_list to mapped_pci_res_list
Signed-off-by: Tetsuya Mukawa <mukawa@igel.co.jp> Acked-by: Bernard Iremonger <bernard.iremonger@intel.com> Acked-by: David Marchand <david.marchand@6wind.com>
pci: fix memory leaks and needless increment of map address
This patch fixes following memory leaks.
- When open() is failed, uio_res and fds won't be freed in
pci_uio_map_resource().
- When pci_map_resource() is failed but path is allocated correctly,
path and fds won't be freed in pci_uio_map_recource().
Also, some mapped resources should be freed.
- When pci_uio_unmap() is called, path should be freed.
Also, fixes below.
- When pci_map_resource() is failed, mapaddr will be MAP_FAILED.
In this case, pci_map_addr should not be incremented in
pci_uio_map_resource().
- To shrink code, move close().
- Remove fail variable.
Signed-off-by: Tetsuya Mukawa <mukawa@igel.co.jp> Acked-by: Bernard Iremonger <bernard.iremonger@intel.com> Acked-by: David Marchand <david.marchand@6wind.com>
This patch fixes coding style of below files in linuxapp and bsdapp.
- eal_pci.c
- eal_pci_uio.c
Signed-off-by: Tetsuya Mukawa <mukawa@igel.co.jp> Acked-by: Stephen Hemminger <stephen@networkplumber.org> Acked-by: David Marchand <david.marchand@6wind.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Jingjing Wu [Wed, 10 Jun 2015 06:24:31 +0000 (14:24 +0800)]
ethdev: rename and extend the mirror type
This path renames the mirror type in rte_eth_mirror_conf and macros,
and rework the mirror set in ixgbe drivers by using new definition.
It also fixes some coding style.
Test report: http://dpdk.org/ml/archives/dev/2015-June/019118.html
Helin Zhang [Thu, 11 Jun 2015 07:03:59 +0000 (15:03 +0800)]
examples/ipv4_multicast: support double vlan stripping and insertion
The outer vlan should be copied from source packet buffer to
support double vlan stripping and insertion, as double vlan can be
stripped or inserted by some of NIC hardware.
Helin Zhang [Thu, 11 Jun 2015 07:03:58 +0000 (15:03 +0800)]
app/testpmd: add qinq stripping and insertion
If double vlan is detected, its stripped flag and vlan tags can be
printed on rxonly mode. Test command of 'tx_vlan set' is expanded
to set both single and double vlan tags on TX side for each packets
to be sent out.
Helin Zhang [Thu, 11 Jun 2015 07:03:56 +0000 (15:03 +0800)]
i40e: support double vlan stripping and insertion
It configures specific registers to enable double vlan stripping
on RX side and insertion on TX side.
The RX descriptors will be parsed, the vlan tags and flags will be
saved to corresponding mbuf fields if vlan tag is detected.
The TX descriptors will be configured according to the
configurations in mbufs, to trigger the hardware insertion of
double vlan tags for each packets sent out.
Helin Zhang [Thu, 11 Jun 2015 07:03:54 +0000 (15:03 +0800)]
ixgbe: remove a discarded source line
Little endian to CPU order conversion had been added for reading
vlan tag from RX descriptor, while its original source line was
forgotten to delete. That's a discarded source line and should be
deleted.
Fixes: 23fcffe8ffac ("ixgbe: fix id and hash with flow director") Signed-off-by: Helin Zhang <helin.zhang@intel.com> Acked-by: Jingjing Wu <jingjing.wu@intel.com>
Zoltan Kiss [Tue, 23 Jun 2015 18:43:12 +0000 (19:43 +0100)]
ethdev: fix check of threshold for Tx freeing
The parameter tx_free_thresh is not consistent between the drivers:
some use it as rte_eth_tx_burst() requires, some release buffers when
the number of free descriptors drop below this value.
Let's use it as most fast-path code does, which is the latter, and update
comments throughout the code to reflect that.
Signed-off-by: Zoltan Kiss <zoltan.kiss@linaro.org> Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
The function to clear the TX ring when a port was being closed, e.g. on
exit in testpmd, was not checking the mbuf refcnt before freeing it.
Since the function in the vector driver to clear the ring after TX does
not set the pointer to NULL post-free, this caused crashes if mbuf
debugging was turned on.
To reproduce the issue, ensure the follow config variables are set:
RTE_IXGBE_INC_VECTOR
RTE_LIBRTE_MBUF_DEBUG
Then compile up and run testpmd using 10G ports with the vector driver.
Start traffic and let some flow through, then type "stop" and "quit" at
the testpmd prompt, and crash will occur. Output below:
testpmd> quit
Stopping port 0...done
Stopping port 1...PANIC in rte_mbuf_sanity_check():
bad ref cnt
[New Thread 0x7fffabfff700 (LWP 145312)]
[New Thread 0x7fffb47fe700 (LWP 145311)]
[New Thread 0x7fffb4fff700 (LWP 145310)]
[New Thread 0x7ffff6cd5700 (LWP 145309)]
18: [/home/bruce/dpdk.org/x86_64-native-linuxapp-gcc/app/testpmd(_start+0x29)
<....snip for brevity...>
Program received signal SIGABRT, Aborted.
0x00007ffff7120a98 in raise () from /lib64/libc.so.6
A similar error occurs when clearing the RX ring, which is also fixed by
this patch.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
As well as the fast-path functions in the rxtx code, there are also
functions which set up and tear down the descriptor rings. Since these
are not performance critical functions, there is no need to have them
extensively optimized, so we add __attribute__((cold)) to their
definitions. This has the side-effect of making debugging them easier as
the compiler does not optimize them as heavily, so more variables are
accessible by default in gdb.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
When a target is finished building, it reports just "Build complete".
When building multiple targets simultaneously, e.g.
make install T=x86_64-native-linuxapp-*
and one target fails, it's not always obvious which of the builds
failed. To help this, we add the actual target that is completed to the
"Build complete" message.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
This fixes lots of complaints from checkpatch about whitespace
and other non-material things. Important to fix the old code now
befor more code does copy/paste of same junk.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Since rte_eth_devices is in BSS, the address of rte_eth_devices
plus port_id can not ever be NULL for a valid port_id number.
Remove this unnecessary check and fix the code style.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org> Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
Daniel Mrzyglod [Mon, 29 Jun 2015 15:06:03 +0000 (17:06 +0200)]
cfgfile: fix unitialised buffer
Nature of the problem was not initialised buffer[256], there were probability
that operation system will provide previously used memory and on special condition
there were probability that string operations will work on random data that
could provide unexpected program behaviour.
Signed-off-by: Daniel Mrzyglod <danielx.t.mrzyglod@intel.com> Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Software implementation of the Toeplitz hash function used by RSS.
Can be used either for packet distribution on single queue NIC
or for simulating of RSS computation on specific NIC (for example
after GRE header decapsulating).
Signed-off-by: Vladimir Medvedkin <medvedkinv@gmail.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
When compiling the cxgbe driver with icc, multiple errors about using
enums as integers appear across a number of files, including in the base
code and in the DPDK-specific driver code.
.../drivers/net/cxgbe/cxgbe_main.c(386): error #188: enumerated type mixed
with another type
t4_get_port_type_description(pi->port_type));
^
For the errors in the base driver code we use the CFLAGS_BASE_DRIVER
approach used by other drivers to disable warnings.
For errors in the DPDK-specific code, typecasts are used to fix the
errors in the code itself.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Shaopeng He [Fri, 15 May 2015 08:56:02 +0000 (16:56 +0800)]
fm10k: support Xen domain0
fm10k was failing to run in XEN domain0, as the physical
memory for DMA should be allocated and translated
in a different way for XEN domain0. So
rte_memzone_reserve_bounded() should be used for DMA
memory allocation, and rte_mem_phy2mch() should be used
for DMA memory address translation to support running
fm10k PMD in XEN domain0.
Signed-off-by: Shaopeng He <shaopeng.he@intel.com> Acked-by: Jijiang Liu <jijiang.liu@intel.com>
Adrien Mazarguil [Tue, 30 Jun 2015 09:28:09 +0000 (11:28 +0200)]
mlx4: update documentation for MOFED 3.0
- Add RX/TX L3/L4 checksum offloading and validation.
- Update kernel module parameters section.
- Update prerequisites for MOFED and firmware versions.
- Remove optimized external libraries section. MOFED now provides enhanced
support directly without having to install modified libraries.
Adrien Mazarguil [Tue, 30 Jun 2015 09:28:06 +0000 (11:28 +0200)]
mlx4: add VXLAN checksum offload
Depending on adapters features and VXLAN support in the kernel, VXLAN frames
can be automatically recognized, in which case checksum validation and
generation occurs on inner and outer L3 and L4.
Signed-off-by: Olga Shern <olgas@mellanox.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Gilad Berman [Tue, 30 Jun 2015 09:28:05 +0000 (11:28 +0200)]
mlx4: add L3/L4 checksum offload
Mellanox ConnectX-3 adapters can handle L3 (IPv4) and L4 (TCP, UDP, TCP6,
UDP6) RX checksums validation and TX checksums generation, with and without
802.1Q (VLAN) headers.
Alex Rosenbaum [Tue, 30 Jun 2015 09:28:03 +0000 (11:28 +0200)]
mlx4: shrink Tx queue elements for better performance
TX queue elements (struct txq_elt) contain WR and SGE structures required by
ibv_post_send(). This commit replaces them with a single pointer to the
related TX mbuf considering that:
- There is no need to keep these structures around forever since the
hardware doesn't access them after ibv_post_send() and send_pending*()
have returned.
- The TX queue index stored in the WR ID field is not used for completions
anymore since they use a separate counter (elts_comp_cd).
- The WR structure itself was only useful for ibv_post_send(), it is
currently only used to store the mbuf data address and an offset to the
mbuf structure in the WR ID field. send_pending*() callbacks only require
SGEs or buffer pointers.
Therefore for single segment mbufs, send_pending() or send_pending_inline()
can be used directly without involving SGEs. For scattered mbufs, SGEs are
allocated on the stack and passed to send_pending_sg_list().
Signed-off-by: Alex Rosenbaum <alexr@mellanox.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Adrien Mazarguil [Tue, 30 Jun 2015 09:28:01 +0000 (11:28 +0200)]
mlx4: use MOFED 3.0 fast verbs interface for Tx operations
The "raw" post send interface was experimental and has been deprecated. This
commit replaces it with a new low level interface that dissociates post and
flush (doorbell) operations for improved QP performance.
The CQ polling function is updated as well.
Signed-off-by: Alex Rosenbaum <alexr@mellanox.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Alex Rosenbaum [Tue, 30 Jun 2015 09:28:00 +0000 (11:28 +0200)]
mlx4: avoid requesting Tx completion events to improve performance
Instead of requesting a completion event for each TX burst, request it on a
fixed schedule once every MLX4_PMD_TX_PER_COMP_REQ (currently 64) packets to
improve performance.
Signed-off-by: Alex Rosenbaum <alexr@mellanox.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Or Ami [Tue, 30 Jun 2015 09:27:58 +0000 (11:27 +0200)]
mlx4: query netdevice to get initial MAC address
Querying the netdevice instead of deriving the port's MAC address from its
GID is less prone to errors. There is no guarantee that the GID will always
contain it nor that the algorithm won't change.
Signed-off-by: Or Ami <ora@mellanox.com> Signed-off-by: Olga Shern <olgas@mellanox.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Olga Shern [Tue, 30 Jun 2015 09:27:57 +0000 (11:27 +0200)]
mlx4: fix support for multiple vlan filters
This commit fixes the "Multiple RX VLAN filters can be configured, but only
the first one works" bug. Since a single flow specification cannot contain
several VLAN definitions, the flows table is extended with MLX4_MAX_VLAN_IDS
possible specifications per configured MAC address.
Signed-off-by: Olga Shern <olgas@mellanox.com> Signed-off-by: Or Ami <ora@mellanox.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Or Ami [Tue, 30 Jun 2015 09:27:56 +0000 (11:27 +0200)]
mlx4: remove provision for flow creation failure in DMFS A0 mode
Starting from MLNX_OFED 3.0 FW 2.34.5000 when working with optimized
steering mode (-7) QPs can be attached to the port's MAC, therefore no need
for the check.
Signed-off-by: Or Ami <ora@mellanox.com> Signed-off-by: Olga Shern <olgas@mellanox.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Olga Shern [Tue, 30 Jun 2015 09:27:53 +0000 (11:27 +0200)]
mlx4: improve accuracy of link status information
Query interface properties using the ethtool API instead of Verbs
through ibv_query_port(). The returned information is more accurate for
Ethernet links since several link speeds cannot be mapped to Verbs
semantics.
Signed-off-by: Olga Shern <olgas@mellanox.com> Signed-off-by: Alex Rosenbaum <alexr@mellanox.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Olga Shern [Tue, 30 Jun 2015 09:27:52 +0000 (11:27 +0200)]
mlx4: allow application to partially fork
Although using the PMD from a forked process is still unsupported, this
commit makes Verbs safe enough for applications to call fork() for other
purposes.
Signed-off-by: Olga Shern <olgas@mellanox.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Adrien Mazarguil [Tue, 30 Jun 2015 09:27:47 +0000 (11:27 +0200)]
mlx4: fix possible crash on scattered mbuf allocation failure
When failing to allocate a segment, mlx4_rx_burst_sp() may call
rte_pktmbuf_free() on an incomplete scattered mbuf whose next pointer
in the last segment is not set.
Rahul Lakkireddy [Mon, 29 Jun 2015 23:28:42 +0000 (04:58 +0530)]
cxgbe: add documentation
Adds cxgbe poll mode driver documentation under the usual doc/guides/nics/
directory with the rest of the drivers. The documentation covers cxgbe
implementation details, features and limitations, prerequisites,
configuration, and a sample application usage.
Rahul Lakkireddy [Mon, 29 Jun 2015 23:28:35 +0000 (04:58 +0530)]
cxgbe: add pmd skeleton
Adds cxgbe poll mode driver for DPDK under drivers/net/cxgbe directory.
This patch:
1. Adds the Makefile to compile cxgbe pmd.
2. Registers and initializes the cxgbe pmd driver.
Enable cxgbe PMD for compilation and linking with changes to:
1. config/common_linuxapp to add macros for cxgbe pmd.
2. drivers/net/Makefile to add cxgbe pmd to the compile list.
3. mk/rte.app.mk to add cxgbe pmd to link.
Update MAINTAINERS file to claim responsibility for the cxgbe PMD.
git.droids-corp.org / dpdk.git / log