Thomas Monjalon [Tue, 1 Dec 2015 14:04:01 +0000 (15:04 +0100)]
mk: introduce new install syntax
The old install command was:
make install T=x86_64-native-linuxapp-gcc DESTDIR=install
It still works and can be replaced by these more standard commands:
make config T=x86_64-native-linuxapp-gcc 0=x86_64-native-linuxapp-gcc
make O=x86_64-native-linuxapp-gcc
make install O=x86_64-native-linuxapp-gcc prefix= DESTDIR=install
It means the "make install" do not perform any compilation anymore when T
is not used. It is done only in pre_install to keep compatibility with the
old syntax based on T= option.
The default prefix /usr/local is empty in the T= case which is
used only for a local install.
Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com> Acked-by: Panu Matilainen <pmatilai@redhat.com>
Thomas Monjalon [Tue, 1 Dec 2015 11:47:48 +0000 (12:47 +0100)]
mk: install a standard cutomizable tree
The rule "install" follows these conventions:
http://gnu.org/prep/standards/html_node/Directory-Variables.html
http://gnu.org/prep/standards/html_node/DESTDIR.html
The variable sdkdir has been added to the more standards ones,
to configure the directory used with RTE_SDK when using the DPDK makefiles
to build an application.
It is still possible to build DPDK with the "install T=" rule without
specifying any DESTDIR. In such case there is no install, as before.
The old usage of an installed SDK is:
make -C examples/helloworld RTE_SDK=$(readlink -m $DESTDIR) \
RTE_TARGET=x86_64-native-linuxapp-gcc
RTE_TARGET can be specified but is useless now with an installed SDK.
The RTE_SDK directory must now point to a different path depending of
the installation.
Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com> Acked-by: Panu Matilainen <pmatilai@redhat.com>
Thomas Monjalon [Tue, 1 Dec 2015 08:27:37 +0000 (09:27 +0100)]
mk: move installation procedure in install rule
The real installation was called "binary install" and was done
after the build when DESTDIR was specified.
Remove this limitation and move the code in install rule only.
Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com> Acked-by: Panu Matilainen <pmatilai@redhat.com>
Thomas Monjalon [Tue, 1 Dec 2015 07:36:00 +0000 (08:36 +0100)]
mk: remove multi-target install
The multi-target install create some subdirectories with the target name
which is not standard for a "make install" procedure.
The uninstall procedure cannot be applied properly (without removing
all files in a directory). It would need to pre-compute paths.
As it is a packaging issue, it is removed from the build system capabilities.
The variable BUILD_DIR is also renamed to RTE_OUTPUT used in other files.
Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com> Acked-by: Panu Matilainen <pmatilai@redhat.com>
Ferruh Yigit [Thu, 3 Dec 2015 13:51:08 +0000 (13:51 +0000)]
mk: fix combined lib build with ABI versioning
Fixes following error (observed when versioning macros used):
LD libdpdk.so
/usr/bin/ld: /root/dpdk/build/lib/libdpdk.so: version node not found
for symbol <function>@DPDK_x.y
Also resulting combined library contains symbol version information:
$ readelf -a build/lib/libdpdk.so | grep rte_eal_ | grep @ | head
<...> GLOBAL DEFAULT 12 rte_eal_alarm_set@@DPDK_2.0
<...> GLOBAL DEFAULT 12 rte_eal_pci_write_config@@DPDK_2.1
<...> GLOBAL DEFAULT 12 rte_eal_remote_launch@@DPDK_2.0
...
Versioning fixed by merging all version scripts into one automatically and
feeding it to final library.
Mike Sowka [Sat, 5 Dec 2015 22:27:56 +0000 (17:27 -0500)]
sched: fix build without SSE4
Irrelevant of the target, the preprocessor #ifdef SSE2 for the
grinder_pipe_exists function is inadequate since the __mm_testz_si128
function requires SSE4.1, PTEST instruction described in
https://en.wikipedia.org/wiki/SSE4#SSE4.1 (I do no have better spec
reference). I have bumped the preprocessor #ifdef to require SSE4.
The Atom N2600 does not have SSE4, http://ark.intel.com/products/58916,
and so I had trouble building rte_sched with optimized version of
grinder_pipe_exists, with following:
error: inlining failed in call to always_inline _mm_testz_si128’:
target specific option mismatch
GCC 4.9 correctly identifies my target as not having SSE4, and with
provided patch builds the non-optimized version of grinder_pipe_exists.
config: disable vector optimizations in sched library
As it causes issues when building with RTE_MACHINE=default due to SSE4.x
requirements and in other discussions was so far rated "lightly tested and
doesn't provide really significant performance improvement" let us disable
that in the default config.
(=> http://dpdk.org/ml/archives/dev/2015-November/029067.html)
Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com> Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Thomas Monjalon [Thu, 3 Dec 2015 04:16:55 +0000 (05:16 +0100)]
eal: fix build with Xen dom0 enabled
There is a new function in the EAL API for internal use.
It has neither a proper prefix nor a .map export:
libethdev.so: undefined reference to `is_xen_dom0_supported'
Fixes: 719dbebceb81 ("xen: allow determining DOM0 at runtime") Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com>
Building RTE_LIBRTE_PMD_XENVIRT was broken when RTE_BUILD_COMBINE_LIBS was
enabled (http://dpdk.org/ml/archives/dev/2015-November/028660.html).
Now the underlying issue is rather simple, the xen code needs libxenstore.
But rte.app.mk so far only considered that when RTE_BUILD_COMBINE_LIBS was
disabled.
While it is correct to create the DPDK sublib linking only in the
RTE_BUILD_COMBINE_LIBS=n case, the libxenstore should be added to the linked
libs in any case if RTE_LIBRTE_PMD_XENVIRT is enabled.
Reported-by: Thiago Martins <thiagocmartinsc@gmail.com> Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
Simon Kagstrom [Tue, 24 Nov 2015 07:50:01 +0000 (08:50 +0100)]
mk: influence CPU flags with user input
We have encountered a CPU where the AES-NI instruction set is disabled
due to export restrictions. Since the build machine and target machine
is different, using -native configs doesn't work, and on this CPU, the
application refuses to run due to the AES CPU flags being amiss.
The patch passes EXTRA_CFLAGS to the figure-out-cpu-flags helper,
which allows us to add -mno-aes to the compile flags and resolve this
problem.
Signed-off-by: Simon Kagstrom <simon.kagstrom@netinsight.net> Acked-by: Olivier Matz <olivier.matz@6wind.com>
After some strace, the problem comes from the --as-needed option passed to the
linker.
It is safer to specify libraries we depend on after the objects we are linking
into a shared library, especially when the linker is invoked with options like
--as-needed.
Fixes: bef06a8a0655 ("mk: set library dependencies in shared object file") Signed-off-by: David Marchand <david.marchand@6wind.com> Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
The validator attempts to disable all kernel modules but since
commit 36080ff96b0eb37a6da8c4fec1a2f8a57dfadf5b fails to do so
for KNI, causing the build stage to fail if kernel headers are missing.
With the introduction of CONFIG_RTE_KNI_KMOD, CONFIG_RTE_LIBRTE_KNI=n
can eventually be dropped but leaving it around for now as its
needed with pre-2.1 versions.
Thomas Monjalon [Thu, 22 Oct 2015 18:27:05 +0000 (20:27 +0200)]
scripts: add build tests
This script helps to build a list of target with some custom options.
It tries to enable most of the options.
The examples and documentation are also built.
It uses some configuration from exported variables.
This config works on my machine:
export DPDK_DEP_PCAP=y
export DPDK_DEP_MOFED=y
mlxdep=/opt/mofed-3.0
export DPDK_DEP_CFLAGS=-I$mlxdep/include
export DPDK_DEP_LDFLAGS=-L$mlxdep/lib
export DPDK_BUILD_TEST_CONFIGS='x86_64-native-linuxapp-gcc+shared+next
x86_64-native-linuxapp-clang+shared+combined
i686-native-linuxapp-gcc+combined'
export DPDK_MAKE_JOBS=8
export DPDK_NOTIFY=notify-send
Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com>
v2:
- conditionally enable szedata2
- add quotes for CFLAGS and LDFLAGS
Thomas Monjalon [Thu, 22 Oct 2015 18:26:28 +0000 (20:26 +0200)]
scripts: add checkpatch wrapper
This script can be used to call checkpatch.pl from Linux with some
custom DPDK options.
The path to the original Linux script must be set in an environment
variable. A script is added to load any configuration variables
required by development tools from a file .develconfig, or
~/.config/dpdk/devel.config or /etc/dpdk/devel.config.
Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
v2:
- do not ignore COMPLEX_MACRO
- use option --no-tree to avoid silent failure
- add -q and -v options
Michael Qiu [Fri, 27 Nov 2015 03:36:05 +0000 (11:36 +0800)]
drivers: fix constant suffix for 32-bit platforms
There is a compilation issue with some compilers.
In i686 platform, long is 32bit, so XXX_CYCLECOUNTER_MASK
need define as 'ULL'
Fixes: 9c857bf6be87 ("igb: support ieee1588 functions for device time") Fixes: 1c4445e1f28e ("ixgbe: support ieee1588 functions for device time") Fixes: f3a4e40eca0c ("i40e: support ieee1588 functions for device time") Signed-off-by: Michael Qiu <michael.qiu@intel.com>
After introducing vPMD feature, fm10k driver will select best
Rx/Tx in running time. Original implementation selects Rx/Tx
silently without notifications.
This patch adds debug info to notify user what actual Rx/Tx
func are used.
Steps to reproduce the bug:
1. All Rx offloading is disabled and start the device, then
Vector Rx is used.
2. Stop the device. Re-configure to enable hw_ip_checksum = 1,
start the device again.
3. In this case, assume regular Rx should be used since Vector
Rx doesn't support ip checksum offload. But actually Vector
Rx is used and cause checksum won't be done by hardware.
The reason is after re-configuring, driver misses an "else" in
func fm10k_set_rx_function(). Then Rx func in last round are
used.
Fixes: 77a8ab47eb38 ("fm10k: select best Rx function") Reported-by: Xiao Wang <xiao.w.wang@intel.com> Signed-off-by: Chen Jing D(Mark) <jing.d.chen@intel.com> Acked-by: Xiao Wang <xiao.w.wang@intel.com>
Wenzhuo Lu [Fri, 27 Nov 2015 06:07:35 +0000 (14:07 +0800)]
igbvf: fix MAC address if none assigned by PF
If use DPDK PF + DPDK VF on igb NICs, we find the MAC address of VF
port is always 0. Because we forget to give it a value if this MAC
address is not assigned by PF. This patch'll assign a random MAC
address to igb VF port as ixgbe does.
Fan Zhang [Wed, 28 Oct 2015 17:11:21 +0000 (17:11 +0000)]
examples/ip_pipeline: add hash key mask parameter
This patch updates the flow_classification pipeline for added key_mask
parameter in 8/16-byte key hash parameters. The update provides user
optional key_mask configuration item applying to the packets.
Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com> Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Fan Zhang [Wed, 28 Oct 2015 17:11:16 +0000 (17:11 +0000)]
table: add key mask to 8 and 16-byte hash parameters
This patch relates to ABI change proposed for librte_table.
The key_mask parameter is added for 8-byte and 16-byte
key extendible bucket and LRU tables.The release notes
is updated and the deprecation notice is removed.
Marcin Kerlin [Tue, 20 Oct 2015 13:01:26 +0000 (15:01 +0200)]
examples/ip_pipeline: add bulk update of firewall rules
Added two new cli commands to firewall pipeline. Commands bulk add and
bulk delete takes as argument a file with rules to add/delete. The file
is parsed, and then rules are passed to backend functions which
add/delete records from pipeline tables.
Signed-off-by: Maciej Gajdzica <maciejx.t.gajdzica@intel.com> Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Marcin Kerlin [Tue, 20 Oct 2015 13:01:23 +0000 (15:01 +0200)]
table: add bulk adding and deleting
New functions prototypes for bulk add/delete added to table API. New
functions allows adding/deleting multiple records with single function
call. For now those functions are implemented only for ACL table. For
other tables these function pointers are set to NULL.
Signed-off-by: Maciej Gajdzica <maciejx.t.gajdzica@intel.com> Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Piotr Azarewicz [Wed, 28 Oct 2015 13:30:44 +0000 (14:30 +0100)]
port: fix reassembly and fragmentation
Bug fixes for ring ports with IPv4/IPv6 reassembly support.
Previous implementation can't work properly due to incorrect choosing
process function.
Also, assuming that, when processing ip packet, ip header is know we can
set l3_len parameter here.
Fix usage RTE_MBUF_METADATA_* macros due to redefinition the macros.
Jerin Jacob [Mon, 23 Nov 2015 18:45:38 +0000 (00:15 +0530)]
mk: add xgene1 machine target based on armv8-a
created the new xgene1 machine target to address the difference
in optional armv8-a CRC extension availability compared to
default armv8-a machine target(enabled CRC extension by default)
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Jerin Jacob [Mon, 23 Nov 2015 18:45:36 +0000 (00:15 +0530)]
hash: use armv8-a CRC32 instructions
armv8-a has optional CRC32 extension, march=armv8-a+crc enables code
generation for the ARMv8-A architecture together with
the optional CRC32 extensions.
added RTE_MACHINE_CPUFLAG_CRC32 to detect the availability of
CRC32 extension in compile time. At run-time, The RTE_CPUFLAG_CRC32
can be used to find the availability.
armv8-a+crc target support added in GCC 4.9,
Used inline assembly and emulated __ARM_FEATURE_CRC32 to work
with tool-chain < 4.9
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Daniel Mrzyglod [Tue, 24 Nov 2015 16:31:18 +0000 (17:31 +0100)]
net: fix build with gcc 4.4.7 and strict aliasing
This fix is for IPv6 checksum offload error on RHEL65.
Any optimalisation above -O0 provide error in IPv6 checksum
flag "-fstrict-aliasing" is default for optimalisation above -O0.
Step 1: testpmd -c 0x6 -n 4 -- -i --portmask=0x3 --disable-hw-vlan
--enable-rx-cksum --crc-strip --txqflags=0
Step 2: settings and start
set verbose 1
set fwd csum
start
Step 3: send scapy with bad checksum IPv6/TCP packet
Ether(src="52:00:00:00:00:00",
dst="90:e2:ba:4a:33:5d")/IPv6(src="::1")/TCP(chksum=0xf)/("X"*46)
Step 4: Received packets:
RESULTS: IPv6/TCP': ['0xd41'] or other unexpected.
EXPECTED RESULTS: IPv6/TCP': ['0x9f5e']
Fixes: 2b039d5f20a3 ("net: fix build with gcc 4.4.7 and strict aliasing") Signed-off-by: Daniel Mrzyglod <danielx.t.mrzyglod@intel.com> Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Thomas Monjalon [Wed, 25 Nov 2015 17:27:21 +0000 (18:27 +0100)]
cryptodev: mark experimental state
The crypto API is in an early state.
It requires more discussions and experiments to declare it stable,
as discussed in http://dpdk.org/ml/archives/dev/2015-November/028634.html
A documentation section will be required in the guides.
Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com>
Declan Doherty [Wed, 25 Nov 2015 13:25:17 +0000 (13:25 +0000)]
examples/l2fwd-crypto: add sample application
This patch creates a new sample applicaiton based off the l2fwd
application which performs specified crypto operations on IP packet
payloads which are forwarding.
Declan Doherty [Wed, 25 Nov 2015 13:25:15 +0000 (13:25 +0000)]
aesni_mb: add driver for multi buffer based crypto
This patch provides the initial implementation of the AES-NI multi-buffer
based crypto poll mode driver using DPDK's new cryptodev framework.
This PMD is dependent on Intel's multibuffer library, see the whitepaper
"Fast Multi-buffer IPsec Implementations on Intel® Architecture
Processors", see ref 1 for details on the library's design and ref 2 to
download the library itself. This initial implementation is limited to
supporting the chained operations of "hash then cipher" or "cipher then
hash" for the following cipher and hash algorithms:
Important Note:
Due to the fact that the multi-buffer library is designed for
accelerating IPsec crypto operation, the digest's generated for the HMAC
functions are truncated to lengths specified by IPsec RFC's, ie RFC2404
for using HMAC-SHA-1 with IPsec specifies that the digest is truncate
from 20 to 12 bytes.
Build instructions:
To build DPDK with the AESNI_MB_PMD the user is required to download
(ref 2) and compile the multi-buffer library on there system before
building DPDK. The environmental variable AESNI_MULTI_BUFFER_LIB_PATH
must be exported with the path where you extracted and built the multi
buffer library and finally set CONFIG_RTE_LIBRTE_PMD_AESNI_MB=y in
config/common_linuxapp.
Current status: It's doesn't support crypto operation
across chained mbufs, or cipher only or hash only operations.
Declan Doherty [Wed, 25 Nov 2015 13:25:14 +0000 (13:25 +0000)]
qat: add driver for QuickAssist devices
This patch adds a PMD for the Intel Quick Assist Technology DH895xxC
hardware accelerator.
This patch depends on a QAT PF driver for device initialization. See
the file docs/guides/cryptodevs/qat.rst for configuration details
This patch supports a limited subset of QAT device functionality,
currently supporting chaining of cipher and hash operations for the
following algorithmsd:
Some limitation on this patchset which shall be contributed in a
subsequent release:
- Chained mbufs are not supported.
- Hash only is not supported.
- Cipher only is not supported.
- Only in-place is currently supported (destination address is
the same as source address).
- Only supports session-oriented API implementation (session-less
APIs are not supported).
Signed-off-by: Declan Doherty <declan.doherty@intel.com> Signed-off-by: John Griffin <john.griffin@intel.com> Signed-off-by: Des O Dea <des.j.o.dea@intel.com> Signed-off-by: Fiona Trahe <fiona.trahe@intel.com> Acked-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
Declan Doherty [Wed, 25 Nov 2015 13:25:13 +0000 (13:25 +0000)]
mbuf_offload: introduce library to attach offloads to mbuf
This library add support for adding a chain of offload operations to a
mbuf. It contains the definition of the rte_mbuf_offload structure as
well as helper functions for attaching offloads to mbufs and a mempool
management functions.
This initial implementation supports attaching multiple offload
operations to a single mbuf, but only a single offload operation of a
specific type can be attach to that mbuf.
Signed-off-by: Declan Doherty <declan.doherty@intel.com> Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Declan Doherty [Wed, 25 Nov 2015 13:25:12 +0000 (13:25 +0000)]
cryptodev: introduce API and framework for crypto devices
This patch contains the initial proposed APIs and device framework for
integrating crypto packet processing into DPDK.
features include:
- Crypto device configuration / management APIs
- Definitions of supported cipher algorithms and operations.
- Definitions of supported hash/authentication algorithms and
operations.
- Crypto session management APIs
- Crypto operation data structures and APIs allocation of crypto
operation structure used to specify the crypto operations to
be performed on a particular mbuf.
- Extension of mbuf to contain crypto operation data pointer and
extra flags.
- Burst enqueue / dequeue APIs for processing of crypto operations.
Signed-off-by: Des O Dea <des.j.o.dea@intel.com> Signed-off-by: John Griffin <john.griffin@intel.com> Signed-off-by: Fiona Trahe <fiona.trahe@intel.com> Signed-off-by: Declan Doherty <declan.doherty@intel.com> Acked-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
Bruce Richardson [Tue, 24 Nov 2015 17:37:57 +0000 (17:37 +0000)]
ethdev: add sanity checks to functions
The functions rte_eth_rx_queue_count and rte_eth_descriptor_done are
supported by very few PMDs. Therefore, it is best to check for support
for the functions in the ethdev library, so as to avoid run-time crashes
at run-time if the application goes to use those APIs. Similarly, the
port parameter should also be checked for validity.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Bruce Richardson [Tue, 24 Nov 2015 17:37:56 +0000 (17:37 +0000)]
ethdev: remove duplicated debug functions
The functions for rx/tx burst, for rx_queue_count and descriptor_done in
the ethdev library all had two copies of the code. One copy in
rte_ethdev.h was inlined for performance, while a second was in
rte_ethdev.c for debugging purposes only. We can eliminate the second
copy of the functions by moving the additional debug checks into the
copies of the functions in the header file. [Any compilation for
debugging at optimization level 0 will not inline the function so the
result should be same as when the function was in the .c file.]
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Declan Doherty [Wed, 25 Nov 2015 13:25:09 +0000 (13:25 +0000)]
ethdev: make error checking macros public
Move the function pointer and port id checking macros to rte_ethdev and
rte_dev header files, so that they can be used in the static inline
functions there. Also replace the RTE_LOG call within
RTE_PMD_DEBUG_TRACE so this macro can be built with the -pedantic flag
Declan Doherty [Wed, 25 Nov 2015 13:25:08 +0000 (13:25 +0000)]
ethdev: prefix internal error checking macros
The macros to check that the function pointers and port ids are valid
for an ethdev are potentially useful to have in a common headers for
use with all PMDs. However, since they would then become externally
visible, we apply the RTE_ & RTE_ETH_ prefix to them as approtiate.
Signed-off-by: Declan Doherty <declan.doherty@intel.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Ferruh Yigit [Wed, 25 Nov 2015 11:13:55 +0000 (11:13 +0000)]
eal/linux: fix build with glibc < 2.12
pthread_setname_np() function added in glibc 2.12, using this function
in older glibc versions cause compile error:
error: implicit declaration of function "pthread_setname_np"
This patch adds "rte_thread_setname" macro and set it according
glibc >= 2.12 check, thread naming disabled for older glibc versions,
glibc versions that support "pthread_setname_np" will keep using this
function.
Fixes: 67b6d3039e9e ("eal: set name to threads") Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Simon Kagstrom [Tue, 17 Nov 2015 08:14:48 +0000 (09:14 +0100)]
sched: release enqueued mbufs when freeing port
Otherwise mbufs will leak when the port is destroyed. The
rte_sched_port_qbase() and rte_sched_port_qsize() functions are used
in free now, so move them up.
Signed-off-by: Simon Kagstrom <simon.kagstrom@netinsight.net>
Increase the number of possible subports per port to allow up to 16 bits.
It is still possible that this will require excessive RAM.
Although mbuf structure is changed, it is ABI compatiable since it
just expands existing sched part of structure to overlap pre-existing hole
in the hash element of structure.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Pablo de Lara [Tue, 24 Nov 2015 16:59:41 +0000 (16:59 +0000)]
kni: fix build on Ubuntu 12.04 with kernel 3.13.0-30
Fixes following error when Ubuntu 12.04 uses kernel 3.13.0-30-generic,
since skb_set_hash() is implemented in the kernel from 3.13.0-30,
which is declared as UBUNTU_KERNEL_VERSION(3,13,0,30,0) and not
UBUNTU_KERNEL_VERSION(3,13,0,30,54)
In file included
from /usr/src/linux-headers-3.13.0-30-generic/include/linux/if_ether.h:23:0,
from /tmp/dpdk/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_osdep.h:39,
from /tmp/dpdk/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_hw.h:31,
from /tmp/dpdk/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_api.h:31,
from /tmp/dpdk/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_mbx.h:31,
from /tmp/dpdk/x86_64-native-linuxapp-gcc/build/lib/librte_eal/linuxapp/kni/e1000_mbx.c:28:
/usr/src/linux-headers-3.13.0-30-generic/include/linux/skbuff.h:740:1:
note: previous definition of ‘skb_set_hash’ was here
skb_set_hash(struct sk_buff *skb, __u32 hash, enum pkt_hash_types type)
^
Fixes: e88c3b0a ("kni: fix build on Ubuntu 12.04.5") Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Xiaobo Chi [Wed, 11 Nov 2015 06:57:23 +0000 (14:57 +0800)]
vhost: fix kernel module insertion
Problem:if I firstly insert my kmod_test.ko, then insert eventfd_link.ko,
error will happen with hint "Device or resource busy". This is because
the default minor device number, 0, has been occupied by my kmod_test.ko .
root@distro:~/test$ lsmod
Module Size Used by
kmod_test 927 0
vboxsf 35930 4
vboxguest 222130 1 vboxsf
microcode 10315 0
autofs4 25051 0
root@distro:~/test$ insmod ./eventfd_link.ko
insmod: ERROR: could not insert module ./eventfd_link.ko: Device or
resource busy
Explanation: For miscdevices, the major device_no is same, so the minor
device_no should be set to ditinguish different misc devices; if not set
the minor, it may fail while insmod due to the default minor value, 0, has
been used by other miscdevice. MISC_DYNAMIC_MINOR means to let Linux
kernel dynamically assign one minor devide number while loading.
Signed-off-by: Xiaobo Chi <xiaobo.chi@nokia.com> Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Victor Kaplansky [Tue, 24 Nov 2015 07:25:35 +0000 (15:25 +0800)]
vhost: fix enabling vring per queue
The VHOST_USER_SET_VRING_ENABLE request was sent for each queue-pair.
However, it's changed to be sent per queue in the queue-pair by QEMU
commit dc3db6ad ("vhost-user: start/stop all rings"). The change
is reasonable, as we send all other request per queue, instead of
queue-pair.
Hence we should do proper changes to adapt to the QEMU change here.
Otherwise, a segfault will be triggered when last TX queue was enabled.
Signed-off-by: Victor Kaplansky <victork@redhat.com> Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Jianfeng Tan [Thu, 12 Nov 2015 06:06:01 +0000 (14:06 +0800)]
vhost: fix mmap not aligned with hugepage size
This patch fixes a bug under lower version linux kernel, mmap()
fails when length is not aligned with hugepage size. mmap()
without flag of MAP_ANONYMOUS, should be called with length
argument aligned with hugepagesz at older longterm version
Linux, like 2.6.32 and 3.2.72, or mmap() will fail with EINVAL.
This bug was fixed in Linux kernel by commit: dab2d3dc45ae7343216635d981d43637e1cb7d45
To avoid failure, make sure in caller to keep length aligned.
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com> Acked-by: Huawei Xie <huawei.xie@intel.com>
Or Ami [Mon, 23 Nov 2015 14:44:42 +0000 (15:44 +0100)]
mlx4: allow operation in secondary processes
Secondary processes are expected to use queues and other resources
allocated by the primary, however Verbs resources can only be shared
between processes when inherited through fork().
This limitation can be worked around for TX by configuring separate queues
from secondary processes.
Signed-off-by: Or Ami <ora@mellanox.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Nelio Laranjeiro [Mon, 23 Nov 2015 14:44:48 +0000 (15:44 +0100)]
mlx5: fix available entries in Tx rings
The number of available entries in TX rings is taken before performing
completion, effectively making rings smaller than they are and causing
TX performance issues under load.
Fixes: 2e22920b85d9 ("mlx5: support non-scattered Tx and Rx") Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com> Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Nelio Laranjeiro [Mon, 23 Nov 2015 14:44:41 +0000 (15:44 +0100)]
mlx4: fix available entries in Tx rings
The number of available entries in TX rings is taken before performing
completion, effectively making rings smaller than they are and causing
TX performance issues under load.
git.droids-corp.org / dpdk.git / log