[dpdk-dev] Number of memory channels per processor socket?

[Abhishek Verma]

Hi,

I am new to DPDK and i tried searching the archives in case this had been
discussed but couldnt find any references and hence this email:

What is the significance of " -n NUM: Number of memory channels per
processor socket" which is passed as an EAL option? I have a virtual
machine (VM) spawned using VirtualBox and i am trying to use DPDK to get
faster access to packets there. I did lscpu, but that didnt give me
anything interesting that i could use here:

abhishekV at VirtualBox:~/dpdk/dpdk-2.0.0/x86_64-native-linuxapp-gcc/app$ sudo
lscpu
[sudo] password for abhishekV:

Architecture:  x86_64
CPU op-mode(s):32-bit, 64-bit
Byte Order:Little Endian
CPU(s):2
On-line CPU(s) list:   0,1
Thread(s) per core:1
Core(s) per socket:2
Socket(s): 1
NUMA node(s):  1
Vendor ID: GenuineIntel
CPU family:6
Model: 58
Stepping:  9
CPU MHz:   2594.017
BogoMIPS:  5188.03
L1d cache: 32K
L1d cache: 32K
L2d cache: 6144K
NUMA node0 CPU(s): 0,1
abhishekV at VirtualBox:~/dpdk/dpdk-2.0.0/x86_64-native-linuxapp-gcc/app$

I understand that i have two cores and hence should use -c 1 (since i have
cores 0 and 1). However, i dont understand what to give as the "-n" option.
Is it always 4?

Thanks, Abhishek

[dpdk-dev] Could not achieve wire speed for 40GE with any DPDK version on XL710 NIC's

[Pavel Odintsov]

Hello, folks!

We have execute bunch of tests for receive data with Intel XL710 40GE
NIC. We want to achieve wire speed on this platform for traffic
capture.

But we definitely can't do it. We tried with different versions of
DPDK: 1.4, 1.6, 1.8, 2.0. And have not success.

We achieved only 40Mpps and could do more.

Could anybody help us with this issue? Looks like this NIC's could not
work on wire speed :(

Platform: Intel Xeon E5 e5 2670 + XL 710.

-- 
Sincerely yours, Pavel Odintsov

[dpdk-dev] [PATCH 1/2] eal:Introduce rte_dma_wmb/rte_dma_rmb.

[WangDong]

These macro can be used to replace current PMD's compiler memory barrier 
(volatile varible) and rte_wmb.
In x86, they implement to compiler memory barrier.
In power, they implement to processor memory barrier.

---
 .../common/include/arch/ppc_64/rte_atomic.h|  4 
 .../common/include/arch/x86/rte_atomic.h   |  4 
 lib/librte_eal/common/include/generic/rte_atomic.h | 25 ++
 3 files changed, 33 insertions(+)

diff --git a/lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h 
b/lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h
index fb7af2b..8f4129d 100644
--- a/lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h
+++ b/lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h
@@ -72,6 +72,10 @@ extern "C" {
  */
 #definerte_rmb() {asm volatile("sync" : : : "memory"); }

+#define rte_dma_wmb() {asm volatile("sync" : : : "memory"); }
+
+#define rte_dma_rmb() {asm volatile("sync" : : : "memory"); 
+
 /*- 16 bit atomic operations 
-*/
 /* To be compatible with Power7, use GCC built-in functions for 16 bit
  * operations */
diff --git a/lib/librte_eal/common/include/arch/x86/rte_atomic.h 
b/lib/librte_eal/common/include/arch/x86/rte_atomic.h
index e93e8ee..7cfbe8f 100644
--- a/lib/librte_eal/common/include/arch/x86/rte_atomic.h
+++ b/lib/librte_eal/common/include/arch/x86/rte_atomic.h
@@ -53,6 +53,10 @@ extern "C" {

 #definerte_rmb() _mm_lfence()

+#define rte_dma_wmb() rte_compiler_barrier()
+
+#define rte_dma_rmb() rte_compiler_barrier()
+
 /*- 16 bit atomic operations 
-*/

 #ifndef RTE_FORCE_INTRINSICS
diff --git a/lib/librte_eal/common/include/generic/rte_atomic.h 
b/lib/librte_eal/common/include/generic/rte_atomic.h
index 6c7581a..a51 100644
--- a/lib/librte_eal/common/include/generic/rte_atomic.h
+++ b/lib/librte_eal/common/include/generic/rte_atomic.h
@@ -72,6 +72,31 @@ static inline void rte_wmb(void);
  */
 static inline void rte_rmb(void);

+/**
+ * Write memory barrier for DMA.
+ *
+ * Be used in PMD, unlike rte_wmb() which use processor memory barrier,
+ * this memory barrier focus on performance, if compiler memory barrier
+ * is sufficient for guarantee memory ordering, this function will
+ * use compiler memory barrier.
+ *
+ * This function is architecture dependent.
+ */
+static inline void rte_dma_wmb(void);
+
+/**
+ * Read memory barrier for DMA.
+ *
+ * Be used in PMD, unlike rte_rmb() which use processor memory barrier,
+ * this memory barrier focus on performance, if compiler memory barrier
+ * is sufficient for guarantee memory ordering, this function will
+ * use compiler memory barrier.
+ *
+ * This function is architecture dependent.
+ */
+static inline void rte_dma_rmb(void);
+
+
 #endif /* __DOXYGEN__ */

 /**
-- 
2.1.0

[dpdk-dev] [PATCH 1/2] eal:Introduce rte_dma_wmb/rte_dma_rmb.

[WangDong]

These macro can be used to replace current PMD's compiler memory barrier 
(volatile varible) and rte_wmb.
In x86, they implement to compiler memory barrier.
In power, they implement to processor memory barrier.

---
 .../common/include/arch/ppc_64/rte_atomic.h|  4 
 .../common/include/arch/x86/rte_atomic.h   |  4 
 lib/librte_eal/common/include/generic/rte_atomic.h | 25 ++
 3 files changed, 33 insertions(+)

diff --git a/lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h 
b/lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h
index fb7af2b..8f4129d 100644
--- a/lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h
+++ b/lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h
@@ -72,6 +72,10 @@ extern "C" {
  */
 #definerte_rmb() {asm volatile("sync" : : : "memory"); }

+#define rte_dma_wmb() {asm volatile("sync" : : : "memory"); }
+
+#define rte_dma_rmb() {asm volatile("sync" : : : "memory"); 
+
 /*- 16 bit atomic operations 
-*/
 /* To be compatible with Power7, use GCC built-in functions for 16 bit
  * operations */
diff --git a/lib/librte_eal/common/include/arch/x86/rte_atomic.h 
b/lib/librte_eal/common/include/arch/x86/rte_atomic.h
index e93e8ee..7cfbe8f 100644
--- a/lib/librte_eal/common/include/arch/x86/rte_atomic.h
+++ b/lib/librte_eal/common/include/arch/x86/rte_atomic.h
@@ -53,6 +53,10 @@ extern "C" {

 #definerte_rmb() _mm_lfence()

+#define rte_dma_wmb() rte_compiler_barrier()
+
+#define rte_dma_rmb() rte_compiler_barrier()
+
 /*- 16 bit atomic operations 
-*/

 #ifndef RTE_FORCE_INTRINSICS
diff --git a/lib/librte_eal/common/include/generic/rte_atomic.h 
b/lib/librte_eal/common/include/generic/rte_atomic.h
index 6c7581a..a51 100644
--- a/lib/librte_eal/common/include/generic/rte_atomic.h
+++ b/lib/librte_eal/common/include/generic/rte_atomic.h
@@ -72,6 +72,31 @@ static inline void rte_wmb(void);
  */
 static inline void rte_rmb(void);

+/**
+ * Write memory barrier for DMA.
+ *
+ * Be used in PMD, unlike rte_wmb() which use processor memory barrier,
+ * this memory barrier focus on performance, if compiler memory barrier
+ * is sufficient for guarantee memory ordering, this function will
+ * use compiler memory barrier.
+ *
+ * This function is architecture dependent.
+ */
+static inline void rte_dma_wmb(void);
+
+/**
+ * Read memory barrier for DMA.
+ *
+ * Be used in PMD, unlike rte_rmb() which use processor memory barrier,
+ * this memory barrier focus on performance, if compiler memory barrier
+ * is sufficient for guarantee memory ordering, this function will
+ * use compiler memory barrier.
+ *
+ * This function is architecture dependent.
+ */
+static inline void rte_dma_rmb(void);
+
+
 #endif /* __DOXYGEN__ */

 /**
-- 
2.1.0

[dpdk-dev] [PATCH 2/2] ixgbe:replace compiler memory barrier and rte_wmb with rte_dma_rmb and rte_dma_wmb.

[WangDong]

---
 drivers/net/ixgbe/ixgbe_rxtx.c | 30 +-
 drivers/net/ixgbe/ixgbe_rxtx_vec.c |  3 +++
 2 files changed, 12 insertions(+), 21 deletions(-)

diff --git a/drivers/net/ixgbe/ixgbe_rxtx.c b/drivers/net/ixgbe/ixgbe_rxtx.c
index 3ace8a8..3316488 100644
--- a/drivers/net/ixgbe/ixgbe_rxtx.c
+++ b/drivers/net/ixgbe/ixgbe_rxtx.c
@@ -130,6 +130,7 @@ ixgbe_tx_free_bufs(struct ixgbe_tx_queue *txq)

/* check DD bit on threshold descriptor */
status = txq->tx_ring[txq->tx_next_dd].wb.status;
+   rte_dma_rmb();
if (! (status & IXGBE_ADVTXD_STAT_DD))
return 0;

@@ -320,7 +321,7 @@ tx_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
txq->tx_tail = 0;

/* update tail pointer */
-   rte_wmb();
+   rte_dma_wmb();
IXGBE_PCI_REG_WRITE(txq->tdt_reg_addr, txq->tx_tail);

return nb_pkts;
@@ -841,7 +842,6 @@ ixgbe_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
txd->read.cmd_type_len |= rte_cpu_to_le_32(cmd_type_len);
}
 end_of_tx:
-   rte_wmb();

/*
 * Set the Transmit Descriptor Tail (TDT)
@@ -849,6 +849,7 @@ end_of_tx:
PMD_TX_LOG(DEBUG, "port_id=%u queue_id=%u tx_tail=%u nb_tx=%u",
   (unsigned) txq->port_id, (unsigned) txq->queue_id,
   (unsigned) tx_id, (unsigned) nb_tx);
+   rte_dma_wmb();
IXGBE_PCI_REG_WRITE(txq->tdt_reg_addr, tx_id);
txq->tx_tail = tx_id;

@@ -975,6 +976,7 @@ ixgbe_rx_scan_hw_ring(struct ixgbe_rx_queue *rxq)

/* Compute how many status bits were set */
nb_dd = 0;
+   rte_dma_rmb();
for (j = 0; j < LOOK_AHEAD; ++j)
nb_dd += s[j] & IXGBE_RXDADV_STAT_DD;

@@ -1138,7 +1140,7 @@ rx_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
}

/* update tail pointer */
-   rte_wmb();
+   rte_dma_wmb();
IXGBE_PCI_REG_WRITE(rxq->rdt_reg_addr, cur_free_trigger);
}

@@ -1229,13 +1231,10 @@ ixgbe_recv_pkts(void *rx_queue, struct rte_mbuf 
**rx_pkts,
/*
 * The order of operations here is important as the DD status
 * bit must not be read after any other descriptor fields.
-* rx_ring and rxdp are pointing to volatile data so the order
-* of accesses cannot be reordered by the compiler. If they were
-* not volatile, they could be reordered which could lead to
-* using invalid descriptor fields when read from rxd.
 */
rxdp = &rx_ring[rx_id];
staterr = rxdp->wb.upper.status_error;
+   rte_dma_rmb();
if (! (staterr & rte_cpu_to_le_32(IXGBE_RXDADV_STAT_DD)))
break;
rxd = *rxdp;
@@ -1373,6 +1372,7 @@ ixgbe_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
   (unsigned) nb_rx);
rx_id = (uint16_t) ((rx_id == 0) ?
 (rxq->nb_rx_desc - 1) : (rx_id - 1));
+   rte_dma_wmb();
IXGBE_PCI_REG_WRITE(rxq->rdt_reg_addr, rx_id);
nb_hold = 0;
}
@@ -1494,17 +1494,6 @@ ixgbe_recv_pkts_lro(void *rx_queue, struct rte_mbuf 
**rx_pkts, uint16_t nb_pkts,

 next_desc:
/*
-* The code in this whole file uses the volatile pointer to
-* ensure the read ordering of the status and the rest of the
-* descriptor fields (on the compiler level only!!!). This is so
-* UGLY - why not to just use the compiler barrier instead? DPDK
-* even has the rte_compiler_barrier() for that.
-*
-* But most importantly this is just wrong because this doesn't
-* ensure memory ordering in a general case at all. For
-* instance, DPDK is supposed to work on Power CPUs where
-* compiler barrier may just not be enough!
-*
 * I tried to write only this function properly to have a
 * starting point (as a part of an LRO/RSC series) but the
 * compiler cursed at me when I tried to cast away the
@@ -1519,12 +1508,11 @@ next_desc:
 * TODO:
 *- Get rid of "volatile" crap and let the compiler do its
 *  job.
-*- Use the proper memory barrier (rte_rmb()) to ensure the
-*  memory ordering below.
 */
rxdp = &rx_ring[rx_id];
staterr = rte_le_to_cpu_32(rxdp->wb.upper.status_error);

+   rte_dma_rmb();
if (!(staterr & IXGBE_RXDADV_STAT_DD))
break;

@@ -1704,7 +1692,7 @@ next_desc:
   "nb_hold=%u nb_rx=%u

[dpdk-dev] RTM instruction compile failure for XABORT when AVX is active

[Matthew Hall]

Hi all,

I am getting a strange error compiling some RTM instructions when I upgraded my 
VM environment from VirtualBox 4 to VirtualBox 5 and AVX instructions from the 
host CPU became available. However when I am reading the opcode description for 
XABORT it supposedly allows an immediate operand for the argument so I can't 
understand why this code would not compile. Any advice?

  CC test_hash_scaling.o
In file included from /vagrant/external/dpdk/app/test/test_hash_scaling.c:37:
In file included from /vagrant/external/dpdk/build/include/rte_spinlock.h:42:
/vagrant/external/dpdk/build/include/rte_rtm.h:56:15: error: invalid operand 
for inline asm constraint 'i'
asm volatile(".byte 0xc6,0xf8,%P0" :: "i" (status) : "memory");

Matthew.

[dpdk-dev] RTM instruction compile failure for XABORT when AVX is active

[Matthew Hall]

It would appear there is some bug in the new lock elision patches that is 
preventing it from compiling with clang. Any suggestions?

On Jun 28, 2015, at 10:16 AM, Matthew Hall  wrote:

> Hi all,
> 
> I am getting a strange error compiling some RTM instructions when I upgraded 
> my VM environment from VirtualBox 4 to VirtualBox 5 and AVX instructions from 
> the host CPU became available. However when I am reading the opcode 
> description for XABORT it supposedly allows an immediate operand for the 
> argument so I can't understand why this code would not compile. Any advice?
> 
>  CC test_hash_scaling.o
> In file included from /vagrant/external/dpdk/app/test/test_hash_scaling.c:37:
> In file included from /vagrant/external/dpdk/build/include/rte_spinlock.h:42:
> /vagrant/external/dpdk/build/include/rte_rtm.h:56:15: error: invalid operand 
> for inline asm constraint 'i'
>asm volatile(".byte 0xc6,0xf8,%P0" :: "i" (status) : "memory");
> 
> Matthew.

[dpdk-dev] [PATCH v3 2/9] cxgbe: add cxgbe poll mode driver.

[Thomas Monjalon]

2015-06-18 17:47, Rahul Lakkireddy:
> +Chelsio cxgbe
> +M: Rahul Lakkireddy 
> +F: drivers/net/cxgbe/
> +F: doc/guides/nics/cxgbe.rst

Just a detail: the doc file is added in a later patch.
For consistency, this line should be added later.

[...]
> --- a/config/common_linuxapp
> +++ b/config/common_linuxapp
> @@ -208,6 +208,16 @@ CONFIG_RTE_LIBRTE_MLX4_TX_MP_CACHE=8
>  CONFIG_RTE_LIBRTE_MLX4_SOFT_COUNTERS=1
>  
>  #
> +# Compile burst-oriented Chelsio Terminator 10GbE/40GbE (CXGBE) PMD
> +#
> +CONFIG_RTE_LIBRTE_CXGBE_PMD=y
> +CONFIG_RTE_LIBRTE_CXGBE_DEBUG=n
> +CONFIG_RTE_LIBRTE_CXGBE_DEBUG_REG=n
> +CONFIG_RTE_LIBRTE_CXGBE_DEBUG_MBOX=n
> +CONFIG_RTE_LIBRTE_CXGBE_DEBUG_TX=n
> +CONFIG_RTE_LIBRTE_CXGBE_DEBUG_RX=n

What is the status of the driver with FreeBSD?
Could we, at least, add the options in common_bsdapp even disabled?

[dpdk-dev] [PATCH v3 3/9] cxgbe: add device configuration and RX support for cxgbe PMD.

[Thomas Monjalon]

2015-06-18 17:47, Rahul Lakkireddy:
> Adds RX support for the cxgbe poll mode driver.  This patch:
> 
> 1. Adds rx queue related eth_dev_ops.
> 2. Adds RSS support.
> 3. Adds dev_configure() and dev_infos_get() eth_dev_ops.
> 4. Adds rx_pkt_burst for receiving packets.
> 
> Signed-off-by: Rahul Lakkireddy 
> Signed-off-by: Kumar Sanghvi 

This patch doesn't build with 32-bit GCC because of some printf args (%lu).

[dpdk-dev] RTM instruction compile failure for XABORT when AVX is active

[Thomas Monjalon]

2015-06-28 10:59, Matthew Hall:
> It would appear there is some bug in the new lock elision patches that is 
> preventing it from compiling with clang. Any suggestions?

It builds with clang.
My suggestion is to add the patch author in recipient's list ;)


> On Jun 28, 2015, at 10:16 AM, Matthew Hall  wrote:
> 
> > Hi all,
> > 
> > I am getting a strange error compiling some RTM instructions when I 
> > upgraded my VM environment from VirtualBox 4 to VirtualBox 5 and AVX 
> > instructions from the host CPU became available. However when I am reading 
> > the opcode description for XABORT it supposedly allows an immediate operand 
> > for the argument so I can't understand why this code would not compile. Any 
> > advice?
> > 
> >  CC test_hash_scaling.o
> > In file included from 
> > /vagrant/external/dpdk/app/test/test_hash_scaling.c:37:
> > In file included from 
> > /vagrant/external/dpdk/build/include/rte_spinlock.h:42:
> > /vagrant/external/dpdk/build/include/rte_rtm.h:56:15: error: invalid 
> > operand for inline asm constraint 'i'
> >asm volatile(".byte 0xc6,0xf8,%P0" :: "i" (status) : "memory");
> > 
> > Matthew.
> 

[dpdk-dev] [PATCHv3 2/3] rte_compat: Add MAP_STATIC_SYMBOL macro

[Thomas Monjalon]

2015-06-26 10:30, Neil Horman:
> On Fri, Jun 26, 2015 at 02:52:50PM +0200, Thomas Monjalon wrote:
> > 2015-06-25 10:35, Neil Horman:
> > > +/*
> > > + * MAP_STATIC_SYMBOL
> > > + * If a function has been bifurcated into multiple versions, none of 
> > > which
> > > + * are defined as the exported symbol name in the map file, this macro 
> > > can be
> > > + * used to alias a specific version of the symbol to its exported name.  
> > > For
> > > + * example, if you have 2 versions of a function foo_v1 and foo_v2, 
> > > where the
> > > + * former is mapped to foo at DPDK_1 and the latter is mapped to foo at 
> > > DPDK_2 when
> > > + * building a shared library, this macro can be used to map either 
> > > foo_v1 or
> > > + * foo_v2 to the symbol foo when building a static library, e.g.:
> > > + * MAP_STATIC_SYMBOL(void foo(), foo_v2);
> > > + */
> > > +#define MAP_STATIC_SYMBOL(f, p)
> > > +
> > >  #else
> > >  /*
> > >   * No symbol versioning in use
> > > @@ -104,7 +105,7 @@
> > >  #define __vsym
> > >  #define BASE_SYMBOL(b, n)
> > >  #define BIND_DEFAULT_SYMBOL(b, e, n)
> > > -
> > > +#define MAP_STATIC_SYMBOL(f, p) f  __attribute__((alias( RTE_STR(p
> > 
> > Is it working with clang and icc?
> No idea.  It should work with clang (though I don't have it installed at the
> moment), as the docs say the .symver directive is supported
> 
> as for icc, thats out of my control completely, as I don't have any access to
> it.
> 
> > Why not just define foo as foo_v2?
> I'm not sure what you mean here.  Are you suggesting that we just change the 
> abi
> so applications have to call foo_v2 rather than foo when we change the
> prototype.  I suppose we could do that, but that seems like it would be an 
> awful
> irritant to users.  They would rather have a single symbol to call if it does
> the same function.
> 
> > As this is the equivalent of BIND_DEFAULT_SYMBOL for the static case,
> > it would be easier to mix them in only one macro.
> > 
> Because of where its used.  If you use BIND_DEFAULT_SYMBOL to do the work of
> MAP_STATIC_SYMBOL, every compilation unit will define its own alias and you'll
> get symbol conflicts.

Oh, you mean it shouldn't be used in a .h file?
If so, this limitation should be noted in the description above.

OK for that solution, that's the best we have at the moment.

[dpdk-dev] [PATCHv3 3/3] ABI: Add some documentation

[Thomas Monjalon]

2015-06-26 10:54, Neil Horman:
> On Fri, Jun 26, 2015 at 03:00:17PM +0200, Thomas Monjalon wrote:
> > 2015-06-25 10:35, Neil Horman:
> > > +* ``VERSION_SYMBOL(b, e, n)``: Creates a symbol version table entry 
> > > binding
> > > +  unversioned symbol ``b`` to the internal function ``b_e``.
> > 
> > Should a versioned symbol @DPDK_
> > 
> Sure.

When fixed, this series can be applied.

> > > +* ``BASE_SYMBOL(b, e)``: Creates a symbol version table entry binding
> > > +  unversioned symbol ``b`` to the internal function ``b_e``.
> > 
> > Please give a use case of BASE_SYMBOL.
> > 
> No, I'd rather remove it if you really insist.  As noted before the way we set
> up the version map files means we currently have no need for this particular
> directive.  I only included it for completeness.  I think an example use in
> light of that fact would only confuse people.  If you're going to draw a line 
> in
> the sand around it, I'll just remove it.

No line in the sand, but it seems better to remove it to avoid confusing people.
ABI compat is already enough difficult to understand ;)

> > [...]
> > > +The addition of the new block tells the linker that a new version node is
> > > +available (DPDK_2.1), which contains the symbol rte_acl_create, and 
> > > inherits the
> > > +symbols from the DPDK_2.0 node.
> > 
> > which contains the old version of the symbol rte_acl_create
> > 
> I don't understand, is it not obvious that the DPDK_2.0 node contains the 2.0
> version of the symbol and the DPDK_2.1 node contains the 2.1 version of the
> symbol?

Yes it is.
I thought it was needed to insist on the existence of a new symbol while the
old one still exists. Just trying to be didactic.

Hope this documentation will help to unlock the patch series currently
breaking the ABI.

[dpdk-dev] [PATCH v3 2/7] mbuf: use the reserved 16 bits for double vlan

[Thomas Monjalon]

Neil, Olivier,
Your opinions are requested here.
Thanks

2015-06-25 08:31, Zhang, Helin:
> Hi Neil
[...]
> > -279,7 +285,7 @@ struct rte_mbuf {
> > uint16_t data_len;/**< Amount of data in segment buffer. */
> > uint32_t pkt_len; /**< Total pkt len: sum of all segments. */
> > uint16_t vlan_tci;/**< VLAN Tag Control Identifier (CPU order) 
> > */
> > -   uint16_t reserved;
> > +   uint16_t vlan_tci_outer;  /**< Outer VLAN Tag Control Identifier (CPU
> > +order) */
> Do you think here is a ABI break or not? Just using the reserved 16 bits, 
> which was
> intended for the second_vlan_tag. Thanks in advance!
> I did not see any "Incompatible" reported by validate_abi.sh.
> 
> Regards,
> Helin

[dpdk-dev] [PATCH v2 0/8] Dynamic RSS Configuration for Bonding

[Thomas Monjalon]

2015-06-26 07:33, Doherty, Declan:
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Tomasz Kulasek
> 
> Acked-by : Declan Doherty 

Please use checkpatch to fix formatting.

[dpdk-dev] [PATCH v2 08/11] hash: add new functionality to store data in hash table

[De Lara Guarch, Pablo]

Hi Stephen,

> -Original Message-
> From: Stephen Hemminger [mailto:stephen at networkplumber.org]
> Sent: Friday, June 26, 2015 5:50 PM
> To: De Lara Guarch, Pablo
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2 08/11] hash: add new functionality to
> store data in hash table
> 
> We did same thing with a slightly different method.
> 
> Subject: rte_hash: split key and bucket size
> 
> It is useful to store more data in the has bucket than just the key size.
> For example, storing an addresss and additional data.
> 
> Signed-off-by: Stephen Hemminger 

Did you send this patch? I did not see it in the mailing list...
Anyway, I like the idea of allowing the user to store variable size data (I was 
storing 8-byte data).
So I have sent a v3 with a more similar method, although I still use the new 
functions,
and use the parameter data_len for knowing the amount of bytes of data, and I 
keep constant the input key,
and return the data in another parameter.

Thanks for it!
Pablo

[dpdk-dev] [PATCH v3 00/11] Cuckoo hash

[Pablo de Lara]

This patchset is to replace the existing hash library with
a more efficient and functional approach, using the Cuckoo hash
method to deal with collisions. This method is based on using
two different hash functions to have two possible locations
in the hash table where an entry can be.
So, if a bucket is full, a new entry can push one of the items
in that bucket to its alternative location, making space for itself.

Advantages
~~
- Offers the option to store more entries when the target bucket is full
  (unlike the previous implementation)
- Memory efficient: for storing those entries, it is not necessary to
  request new memory, as the entries will be stored in the same table
- Constant worst lookup time: in worst case scenario, it always takes
  the same time to look up an entry, as there are only two possible locations
  where an entry can be.
- Storing data: user can store data in the hash table, unlike the
  previous implementation, but he can still use the old API

This implementation tipically offers over 90% utilization.
Notice that API has been extended, but old API remains. The main
change in ABI is that rte_hash structure is now private and the
deprecation of two macros.

Changes in v3:

- Now user can store variable size data, instead of 32 or 64-bit size data,
  using the new parameter "data_len" in rte_hash_parameters
- Add lookup_bulk_with_hash function in performance  unit tests
- Add new functions that handle data in performance unit tests
- Remove duplicates in performance unit tests
- Fix rte_hash_reset, which was not reseting the last entry

Changes in v2:

- Fixed issue where table could not store maximum number of entries
- Fixed issue where lookup burst could not be more than 32 (instead of 64)
- Remove unnecessary macros and add other useful ones
- Added missing library dependencies
- Used directly rte_hash_secondary instead of rte_hash_alt
- Renamed rte_hash.c to rte_cuckoo_hash.c to ease the view of the new library
- Renamed test_hash_perf.c temporarily to ease the view of the improved unit 
test
- Moved rte_hash, rte_bucket and rte_hash_key structures to rte_cuckoo_hash.c to
  make them private
- Corrected copyright dates
- Added an optimized function to compare keys that are multiple of 16 bytes
- Improved the way to use primary/secondary signatures. Now both are stored in
  the bucket, so there is no need to calculate them if required.
  Also, there is no need to use the MSB of a signature to differenciate between
  an empty entry and signature 0, since we are storing both signatures,
  which cannot be both 0.
- Removed rte_hash_rehash, as it was a very expensive operation.
  Therefore, the add function returns now -ENOSPC if key cannot be added
  because of a loop.
- Prefetched new slot for new key in add function to improve performance.
- Made doxygen comments more clear.
- Removed unnecessary rte_hash_del_key_data and rte_hash_del_key_with_data,
  as we can use the lookup functions if we want to get the data before deleting.
- Removed some unnecessary includes in rte_hash.h
- Removed some unnecessary variables in rte_cuckoo_hash.c
- Removed some unnecessary checks before creating a new hash table 
- Added documentation (in release notes and programmers guide)
- Added new unit tests and replaced the performance one for hash tables

Pablo de Lara (11):
  eal: add const in prefetch functions
  hash: move rte_hash structure to C file and make it internal
  test/hash: enhance hash unit tests
  test/hash: rename new hash perf unit test back to original name
  hash: replace existing hash library with cuckoo hash implementation
  hash: add new lookup_bulk_with_hash function
  hash: add new function rte_hash_reset
  hash: add new functionality to store data in hash table
  MAINTAINERS: claim responsability for hash library
  doc: announce ABI change of librte_hash
  doc: update hash documentation

 MAINTAINERS|1 +
 app/test/Makefile  |1 +
 app/test/test_hash.c   |  189 +--
 app/test/test_hash_perf.c  |  965 ++
 doc/guides/prog_guide/hash_lib.rst |   77 +-
 doc/guides/rel_notes/abi.rst   |2 +
 .../common/include/arch/ppc_64/rte_prefetch.h  |6 +-
 .../common/include/arch/x86/rte_prefetch.h |   14 +-
 .../common/include/generic/rte_prefetch.h  |8 +-
 lib/librte_hash/Makefile   |8 +-
 lib/librte_hash/rte_cuckoo_hash.c  | 1393 
 lib/librte_hash/rte_hash.c |  471 ---
 lib/librte_hash/rte_hash.h |  275 +++-
 lib/librte_hash/rte_hash_version.map   |   15 +
 14 files changed, 2250 insertions(+), 1175 deletions(-)
 create mode 100644 lib/librte_hash/rte_cuckoo_hash.c
 delete mode 100644 lib/librte_hash/rte_hash.c

-- 
2.4.2

[dpdk-dev] [PATCH v3 07/11] hash: add new function rte_hash_reset

[Pablo de Lara]

Added reset function to be able to empty the table,
without having to destroy and create it again.

Signed-off-by: Pablo de Lara 
---
 app/test/test_hash.c |  3 +--
 app/test/test_hash_perf.c| 11 +++
 lib/librte_hash/rte_cuckoo_hash.c| 21 +
 lib/librte_hash/rte_hash.h   | 11 +--
 lib/librte_hash/rte_hash_version.map |  1 +
 5 files changed, 35 insertions(+), 12 deletions(-)

diff --git a/app/test/test_hash.c b/app/test/test_hash.c
index 52de1bd..b1ca939 100644
--- a/app/test/test_hash.c
+++ b/app/test/test_hash.c
@@ -1156,8 +1156,7 @@ static int test_average_table_utilization(void)
no_space = 0;

/* Reset the table */
-   rte_hash_free(handle);
-   rte_hash_create(&ut_params);
+   rte_hash_reset(handle);
}

const unsigned average_keys_added = added_keys_until_no_space / 
ITERATIONS;
diff --git a/app/test/test_hash_perf.c b/app/test/test_hash_perf.c
index 86295b8..6e8b255 100644
--- a/app/test/test_hash_perf.c
+++ b/app/test/test_hash_perf.c
@@ -368,14 +368,10 @@ free_table(unsigned table_index)
rte_hash_free(h[table_index]);
 }

-static int
+static void
 reset_table(unsigned table_index)
 {
-   free_table(table_index);
-   if (create_table(table_index) != 0)
-   return -1;
-
-   return 0;
+   rte_hash_reset(h[table_index]);
 }

 static int
@@ -417,8 +413,7 @@ run_all_tbl_perf_tests(void)
if (timed_deletes(1, i) < 0)
return -1;

-   if (reset_table(i) < 0)
-   return -1;
+   reset_table(i);

printf("\n - WITH JUST KEYS -\n\n");

diff --git a/lib/librte_hash/rte_cuckoo_hash.c 
b/lib/librte_hash/rte_cuckoo_hash.c
index 1e70069..44f87fb 100644
--- a/lib/librte_hash/rte_cuckoo_hash.c
+++ b/lib/librte_hash/rte_cuckoo_hash.c
@@ -358,6 +358,27 @@ rte_hash_secondary_hash(const hash_sig_t primary_hash)
return (primary_hash ^ ((tag + 1) * alt_bits_xor));
 }

+void
+rte_hash_reset(struct rte_hash *h)
+{
+   void *ptr;
+   unsigned i;
+
+   if (h == NULL)
+   return;
+
+   memset(h->buckets, 0, h->num_buckets * sizeof(struct rte_hash_bucket));
+   memset(h->key_store, 0, h->key_entry_size * (h->entries + 1));
+
+   /* clear the free ring */
+   while (rte_ring_dequeue(h->free_slots, &ptr) == 0)
+   rte_pause();
+
+   /* Repopulate the free slots ring. Entry zero is reserved for key 
misses */
+   for (i = 1; i < h->entries + 1; i++)
+   rte_ring_sp_enqueue(h->free_slots, (void *)((uintptr_t) i));
+}
+
 /*
  * Try to insert a new entry. If bucket has space, hash value and key index
  * to the key table are copied.
diff --git a/lib/librte_hash/rte_hash.h b/lib/librte_hash/rte_hash.h
index 7f7e75f..fa327c2 100644
--- a/lib/librte_hash/rte_hash.h
+++ b/lib/librte_hash/rte_hash.h
@@ -132,8 +132,15 @@ void
 rte_hash_free(struct rte_hash *h);

 /**
- * Add a key to an existing hash table.
- * This operation is not multi-thread safe
+ * Reset all hash structure, by zeroing all entries
+ * @param h
+ *   Hash table to reset
+ */
+void
+rte_hash_reset(struct rte_hash *h);
+
+/**
+ * Add a key to an existing hash table. This operation is not multi-thread safe
  * and should only be called from one thread.
  *
  * @param h
diff --git a/lib/librte_hash/rte_hash_version.map 
b/lib/librte_hash/rte_hash_version.map
index fd92def..f011054 100644
--- a/lib/librte_hash/rte_hash_version.map
+++ b/lib/librte_hash/rte_hash_version.map
@@ -22,6 +22,7 @@ DPDK_2.1 {
global:

rte_hash_lookup_bulk_with_hash;
+   rte_hash_reset;

local: *;
 } DPDK_2.0;
-- 
2.4.2

[dpdk-dev] [PATCH v3 02/11] hash: move rte_hash structure to C file and make it internal

[Pablo de Lara]

rte_hash structure should not be a public structure,
and therefore it should be moved to the C file and be declared
as internal.

This patch also removes part of a unit test that was checking
a field of the structure.

Signed-off-by: Pablo de Lara 
---
 app/test/test_hash.c   |  6 +-
 lib/librte_hash/rte_hash.c | 30 +-
 lib/librte_hash/rte_hash.h | 37 +
 3 files changed, 35 insertions(+), 38 deletions(-)

diff --git a/app/test/test_hash.c b/app/test/test_hash.c
index 4ecb11b..4300de9 100644
--- a/app/test/test_hash.c
+++ b/app/test/test_hash.c
@@ -1,7 +1,7 @@
 /*-
  *   BSD LICENSE
  *
- *   Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
  *   All rights reserved.
  *
  *   Redistribution and use in source and binary forms, with or without
@@ -1110,10 +1110,6 @@ test_hash_creation_with_good_parameters(void)
printf("Creating hash with null hash_func failed\n");
return -1;
}
-   if (handle->hash_func == NULL) {
-   printf("Hash function should have been DEFAULT_HASH_FUNC\n");
-   return -1;
-   }

/* this test is trying to create a hash with the same name as previous 
one.
 * this should return a pointer to the hash we previously created.
diff --git a/lib/librte_hash/rte_hash.c b/lib/librte_hash/rte_hash.c
index 67dff5b..5100a75 100644
--- a/lib/librte_hash/rte_hash.c
+++ b/lib/librte_hash/rte_hash.c
@@ -1,7 +1,7 @@
 /*-
  *   BSD LICENSE
  *
- *   Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
  *   All rights reserved.
  *
  *   Redistribution and use in source and binary forms, with or without
@@ -92,6 +92,27 @@ EAL_REGISTER_TAILQ(rte_hash_tailq)
 /* The high bit is always set in real signatures */
 #define NULL_SIGNATURE  0

+struct rte_hash {
+   char name[RTE_HASH_NAMESIZE];   /**< Name of the hash. */
+   uint32_t entries;   /**< Total table entries. */
+   uint32_t bucket_entries;/**< Bucket entries. */
+   uint32_t key_len;   /**< Length of hash key. */
+   rte_hash_function hash_func;/**< Function used to calculate hash. */
+   uint32_t hash_func_init_val;/**< Init value used by hash_func. */
+   uint32_t num_buckets;   /**< Number of buckets in table. */
+   uint32_t bucket_bitmask;/**< Bitmask for getting bucket index
+   from hash signature. */
+   hash_sig_t sig_msb; /**< MSB is always set in valid signatures. */
+   uint8_t *sig_tbl;   /**< Flat array of hash signature buckets. */
+   uint32_t sig_tbl_bucket_size;   /**< Signature buckets may be padded for
+  alignment reasons, and this is the
+  bucket size used by sig_tbl. */
+   uint8_t *key_tbl;   /**< Flat array of key value buckets. */
+   uint32_t key_tbl_key_size;  /**< Keys may be padded for alignment
+  reasons, and this is the key size
+  used by key_tbl. */
+};
+
 /* Returns a pointer to the first signature in specified bucket. */
 static inline hash_sig_t *
 get_sig_tbl_bucket(const struct rte_hash *h, uint32_t bucket_index)
@@ -291,6 +312,13 @@ rte_hash_free(struct rte_hash *h)
rte_free(te);
 }

+hash_sig_t
+rte_hash_hash(const struct rte_hash *h, const void *key)
+{
+   /* calc hash result by key */
+   return h->hash_func(key, h->key_len, h->hash_func_init_val);
+}
+
 static inline int32_t
 __rte_hash_add_key_with_hash(const struct rte_hash *h,
const void *key, hash_sig_t sig)
diff --git a/lib/librte_hash/rte_hash.h b/lib/librte_hash/rte_hash.h
index 821a9d4..79827a6 100644
--- a/lib/librte_hash/rte_hash.h
+++ b/lib/librte_hash/rte_hash.h
@@ -1,7 +1,7 @@
 /*-
  *   BSD LICENSE
  *
- *   Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
  *   All rights reserved.
  *
  *   Redistribution and use in source and binary forms, with or without
@@ -40,9 +40,6 @@
  * RTE Hash Table
  */

-#include 
-#include 
-
 #ifdef __cplusplus
 extern "C" {
 #endif
@@ -84,27 +81,8 @@ struct rte_hash_parameters {
int socket_id;  /**< NUMA Socket ID for memory. */
 };

-/** A hash table structure. */
-struct rte_hash {
-   char name[RTE_HASH_NAMESIZE];   /**< Name of the hash. */
-   uint32_t entries;   /**< Total table entries. */
-   uint32_t bucket_entries;/**< Bucket entries. */
-   uint32_t key_len;   /**< Length of hash key. */
-   rte_hash_function hash_func;/**< Function used

[dpdk-dev] [PATCH v3 06/11] hash: add new lookup_bulk_with_hash function

[Pablo de Lara]

Previous implementation was lacking a function
to look up a burst of entries, given precalculated hash values.
This patch implements such function, quite useful for
looking up keys from packets that have precalculated hash values
from a 5-tuple key.

Added the function in the hash unit tests as well.

Signed-off-by: Pablo de Lara 
---
 app/test/test_hash.c |  19 ++-
 app/test/test_hash_perf.c|  24 +++-
 lib/librte_hash/rte_cuckoo_hash.c| 222 ++-
 lib/librte_hash/rte_hash.h   |  27 +
 lib/librte_hash/rte_hash_version.map |   8 ++
 5 files changed, 291 insertions(+), 9 deletions(-)

diff --git a/app/test/test_hash.c b/app/test/test_hash.c
index 0176219..52de1bd 100644
--- a/app/test/test_hash.c
+++ b/app/test/test_hash.c
@@ -456,6 +456,7 @@ static int test_five_keys(void)
 {
struct rte_hash *handle;
const void *key_array[5] = {0};
+   hash_sig_t hashes[5];
int pos[5];
int expected_pos[5];
unsigned i;
@@ -475,12 +476,24 @@ static int test_five_keys(void)
}

/* Lookup */
-   for(i = 0; i < 5; i++)
+   for (i = 0; i < 5; i++) {
key_array[i] = &keys[i];
+   hashes[i] = rte_hash_hash(handle, &keys[i]);
+   }

ret = rte_hash_lookup_multi(handle, &key_array[0], 5, (int32_t *)pos);
-   if(ret == 0)
-   for(i = 0; i < 5; i++) {
+   if (ret == 0)
+   for (i = 0; i < 5; i++) {
+   print_key_info("Lkp", key_array[i], pos[i]);
+   RETURN_IF_ERROR(pos[i] != expected_pos[i],
+   "failed to find key (pos[%u]=%d)", i, 
pos[i]);
+   }
+
+   /* Lookup with precalculated hashes */
+   ret = rte_hash_lookup_multi_with_hash(handle, &key_array[0], hashes,
+   5, (int32_t *)pos);
+   if (ret == 0)
+   for (i = 0; i < 5; i++) {
print_key_info("Lkp", key_array[i], pos[i]);
RETURN_IF_ERROR(pos[i] != expected_pos[i],
"failed to find key (pos[%u]=%d)", i, 
pos[i]);
diff --git a/app/test/test_hash_perf.c b/app/test/test_hash_perf.c
index 978731c..86295b8 100644
--- a/app/test/test_hash_perf.c
+++ b/app/test/test_hash_perf.c
@@ -289,18 +289,27 @@ timed_lookups(unsigned with_hash, unsigned table_index)
 }

 static int
-timed_lookups_multi(unsigned table_index)
+timed_lookups_multi(unsigned with_hash, unsigned table_index)
 {
unsigned i, j, k;
int32_t positions_burst[BURST_SIZE];
const void *keys_burst[BURST_SIZE];
const uint64_t start_tsc = rte_rdtsc();
+   hash_sig_t *hash_array;

for (i = 0; i < NUM_LOOKUPS/KEYS_TO_ADD; i++) {
for (j = 0; j < KEYS_TO_ADD/BURST_SIZE; j++) {
for (k = 0; k < BURST_SIZE; k++)
keys_burst[k] = keys[j * BURST_SIZE + k];
-   rte_hash_lookup_bulk(h[table_index],
+   if (with_hash) {
+   hash_array = &signatures[j * BURST_SIZE];
+   rte_hash_lookup_bulk_with_hash(h[table_index],
+   (const void **) keys_burst,
+   (const hash_sig_t *) hash_array,
+   BURST_SIZE,
+   positions_burst);
+   } else
+   rte_hash_lookup_bulk(h[table_index],
(const void **) keys_burst,
BURST_SIZE,
positions_burst);
@@ -311,12 +320,12 @@ timed_lookups_multi(unsigned table_index)
const uint64_t time_taken = end_tsc - start_tsc;
const float seconds_taken = (float)time_taken/rte_get_tsc_hz();

-   cycles[table_index][LOOKUP_MULTI][0] = time_taken/NUM_LOOKUPS;
+   cycles[table_index][LOOKUP_MULTI][with_hash] = time_taken/NUM_LOOKUPS;

printf("%"PRIu64" lookups in %f seconds\n", (uint64_t)NUM_LOOKUPS,
seconds_taken);
printf("Average %"PRIu64" tsc ticks per lookup\n",
-   cycles[table_index][LOOKUP_MULTI][0]);
+   cycles[table_index][LOOKUP_MULTI][with_hash]);
printf("Average %"PRIu64" lookups per second\n",
(NUM_LOOKUPS * rte_get_tsc_hz())/time_taken);
return 0;
@@ -398,6 +407,11 @@ run_all_tbl_perf_tests(void)
if (timed_lookups(1, i) < 0)
return -1;

+   printf("\nTimed lookups multi\n");
+   printf("--\n");
+   if (timed_lookups_multi(1, i) < 0)
+   return -1;
+

[dpdk-dev] [PATCH v3 01/11] eal: add const in prefetch functions

[Pablo de Lara]

rte_prefetchX functions included volatile void *p as parameter,
but the function does not modify it, so it should include the const keyword.

Signed-off-by: Pablo de Lara 
Acked-by: Bruce Richardson 
---
 lib/librte_eal/common/include/arch/ppc_64/rte_prefetch.h |  6 +++---
 lib/librte_eal/common/include/arch/x86/rte_prefetch.h| 14 +++---
 lib/librte_eal/common/include/generic/rte_prefetch.h |  8 
 3 files changed, 14 insertions(+), 14 deletions(-)

diff --git a/lib/librte_eal/common/include/arch/ppc_64/rte_prefetch.h 
b/lib/librte_eal/common/include/arch/ppc_64/rte_prefetch.h
index 9df0d13..fea3be1 100644
--- a/lib/librte_eal/common/include/arch/ppc_64/rte_prefetch.h
+++ b/lib/librte_eal/common/include/arch/ppc_64/rte_prefetch.h
@@ -39,17 +39,17 @@ extern "C" {

 #include "generic/rte_prefetch.h"

-static inline void rte_prefetch0(volatile void *p)
+static inline void rte_prefetch0(const volatile void *p)
 {
asm volatile ("dcbt 0,%[p],1" : : [p] "r" (p));
 }

-static inline void rte_prefetch1(volatile void *p)
+static inline void rte_prefetch1(const volatile void *p)
 {
asm volatile ("dcbt 0,%[p],1" : : [p] "r" (p));
 }

-static inline void rte_prefetch2(volatile void *p)
+static inline void rte_prefetch2(const volatile void *p)
 {
asm volatile ("dcbt 0,%[p],1" : : [p] "r" (p));
 }
diff --git a/lib/librte_eal/common/include/arch/x86/rte_prefetch.h 
b/lib/librte_eal/common/include/arch/x86/rte_prefetch.h
index ec2454d..8e6e02c 100644
--- a/lib/librte_eal/common/include/arch/x86/rte_prefetch.h
+++ b/lib/librte_eal/common/include/arch/x86/rte_prefetch.h
@@ -1,7 +1,7 @@
 /*-
  *   BSD LICENSE
  *
- *   Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
  *   All rights reserved.
  *
  *   Redistribution and use in source and binary forms, with or without
@@ -40,19 +40,19 @@ extern "C" {

 #include "generic/rte_prefetch.h"

-static inline void rte_prefetch0(volatile void *p)
+static inline void rte_prefetch0(const volatile void *p)
 {
-   asm volatile ("prefetcht0 %[p]" : [p] "+m" (*(volatile char *)p));
+   asm volatile ("prefetcht0 %[p]" : : [p] "m" (*(const volatile char 
*)p));
 }

-static inline void rte_prefetch1(volatile void *p)
+static inline void rte_prefetch1(const volatile void *p)
 {
-   asm volatile ("prefetcht1 %[p]" : [p] "+m" (*(volatile char *)p));
+   asm volatile ("prefetcht1 %[p]" : : [p] "m" (*(const volatile char 
*)p));
 }

-static inline void rte_prefetch2(volatile void *p)
+static inline void rte_prefetch2(const volatile void *p)
 {
-   asm volatile ("prefetcht2 %[p]" : [p] "+m" (*(volatile char *)p));
+   asm volatile ("prefetcht2 %[p]" : : [p] "m" (*(const volatile char 
*)p));
 }

 #ifdef __cplusplus
diff --git a/lib/librte_eal/common/include/generic/rte_prefetch.h 
b/lib/librte_eal/common/include/generic/rte_prefetch.h
index 217f319..725715f 100644
--- a/lib/librte_eal/common/include/generic/rte_prefetch.h
+++ b/lib/librte_eal/common/include/generic/rte_prefetch.h
@@ -1,7 +1,7 @@
 /*-
  *   BSD LICENSE
  *
- *   Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
  *   All rights reserved.
  *
  *   Redistribution and use in source and binary forms, with or without
@@ -51,14 +51,14 @@
  * @param p
  *   Address to prefetch
  */
-static inline void rte_prefetch0(volatile void *p);
+static inline void rte_prefetch0(const volatile void *p);

 /**
  * Prefetch a cache line into all cache levels except the 0th cache level.
  * @param p
  *   Address to prefetch
  */
-static inline void rte_prefetch1(volatile void *p);
+static inline void rte_prefetch1(const volatile void *p);

 /**
  * Prefetch a cache line into all cache levels except the 0th and 1th cache
@@ -66,6 +66,6 @@ static inline void rte_prefetch1(volatile void *p);
  * @param p
  *   Address to prefetch
  */
-static inline void rte_prefetch2(volatile void *p);
+static inline void rte_prefetch2(const volatile void *p);

 #endif /* _RTE_PREFETCH_H_ */
-- 
2.4.2

[dpdk-dev] [PATCH v3 09/11] MAINTAINERS: claim responsability for hash library

[Pablo de Lara]

Signed-off-by: Pablo de Lara 
---
 MAINTAINERS | 1 +
 1 file changed, 1 insertion(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index 54f0973..a536992 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -339,6 +339,7 @@ F: doc/guides/sample_app_ug/l3_forward_access_ctrl.rst

 Hashes
 M: Bruce Richardson 
+M: Pablo de Lara 
 F: lib/librte_hash/
 F: doc/guides/prog_guide/hash_lib.rst
 F: app/test/test_hash*
-- 
2.4.2

[dpdk-dev] [PATCH v3 03/11] test/hash: enhance hash unit tests

[Pablo de Lara]

Add new unit test for calculating the average table utilization,
using random keys, based on number of entries that can be added
until we encounter one that cannot be added (bucket if full)

Also, replace current hash_perf unit test to see performance more clear.
The current hash_perf unit test takes too long and add keys that
may or may not fit in the table and look up/delete that may not be
in the table. This new unit test gets a set of keys that we know
that fits in the table, and then measure the time to add/look up/delete
them.

Signed-off-by: Pablo de Lara 
---
 app/test/Makefile |   2 +-
 app/test/test_hash.c  |  59 
 app/test/test_hash_perf.c | 702 --
 app/test/test_hash_perf_new.c | 560 +
 4 files changed, 620 insertions(+), 703 deletions(-)
 delete mode 100644 app/test/test_hash_perf.c
 create mode 100644 app/test/test_hash_perf_new.c

diff --git a/app/test/Makefile b/app/test/Makefile
index 2e2758c..8624e95 100644
--- a/app/test/Makefile
+++ b/app/test/Makefile
@@ -82,7 +82,7 @@ SRCS-y += test_memcpy.c
 SRCS-y += test_memcpy_perf.c

 SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash.c
-SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_perf.c
+SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_perf_new.c
 SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_functions.c
 SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_scaling.c

diff --git a/app/test/test_hash.c b/app/test/test_hash.c
index 4300de9..46174db 100644
--- a/app/test/test_hash.c
+++ b/app/test/test_hash.c
@@ -1147,6 +1147,63 @@ test_hash_creation_with_good_parameters(void)
return 0;
 }

+#define ITERATIONS 50
+/*
+ * Test to see the average table utilization (entries added/max entries)
+ * before hitting a random entry that cannot be added
+ */
+static int test_average_table_utilization(void)
+{
+   struct rte_hash *handle;
+   void *simple_key;
+   unsigned i, j, no_space = 0;
+   double added_keys_until_no_space = 0;
+   int ret;
+
+   ut_params.entries = 1 << 20;
+   ut_params.name = "test_average_utilization";
+   ut_params.hash_func = rte_hash_crc;
+   handle = rte_hash_create(&ut_params);
+   RETURN_IF_ERROR(handle == NULL, "hash creation failed");
+
+   simple_key = rte_zmalloc(NULL, ut_params.key_len, 0);
+
+   for (j = 0; j < ITERATIONS; j++) {
+   while (!no_space) {
+   for (i = 0; i < ut_params.key_len; i++)
+   ((uint8_t *) simple_key)[i] = rte_rand() % 255;
+
+   ret = rte_hash_add_key(handle, simple_key);
+   print_key_info("Add", simple_key, ret);
+
+   if (ret == -ENOSPC) {
+   if (-ENOENT != rte_hash_lookup(handle, 
simple_key))
+   printf("Found key that should not be 
present\n");
+   no_space = 1;
+   } else {
+   if (ret < 0)
+   rte_free(simple_key);
+   RETURN_IF_ERROR(ret < 0, "failed to add key 
(ret=%d)", ret);
+   added_keys_until_no_space++;
+   }
+   }
+   no_space = 0;
+
+   /* Reset the table */
+   rte_hash_free(handle);
+   rte_hash_create(&ut_params);
+   }
+
+   const unsigned average_keys_added = added_keys_until_no_space / 
ITERATIONS;
+
+   printf("Average table utilization = %.2f%% (%u/%u)\n",
+   ((double) average_keys_added / ut_params.entries * 100),
+   average_keys_added, ut_params.entries);
+   rte_hash_free(handle);
+
+   return 0;
+}
+
 static uint8_t key[16] = {0x00, 0x01, 0x02, 0x03,
0x04, 0x05, 0x06, 0x07,
0x08, 0x09, 0x0a, 0x0b,
@@ -1405,6 +1462,8 @@ test_hash(void)
return -1;
if (test_hash_creation_with_good_parameters() < 0)
return -1;
+   if (test_average_table_utilization() < 0)
+   return -1;

run_hash_func_tests();

diff --git a/app/test/test_hash_perf.c b/app/test/test_hash_perf.c
deleted file mode 100644
index d0e5ce0..000
--- a/app/test/test_hash_perf.c
+++ /dev/null
@@ -1,702 +0,0 @@
-/*-
- *   BSD LICENSE
- *
- *   Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
- *   All rights reserved.
- *
- *   Redistribution and use in source and binary forms, with or without
- *   modification, are permitted provided that the following conditions
- *   are met:
- *
- * * Redistributions of source code must retain the above copyright
- *   notice, this list of conditions and the following disclaimer.
- * * Redistributions in binary form must reproduce the above copyright
- *   notice, this list of conditions and the following disclaimer in
- *   the

[dpdk-dev] [PATCH v3 10/11] doc: announce ABI change of librte_hash

[Pablo de Lara]

rte_hash structure is now private for version 2.1, and two
of the macros in rte_hash.h are now deprecated, so this patch
adds notice of these changes.

Signed-off-by: Pablo de Lara 
---
 doc/guides/rel_notes/abi.rst | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/doc/guides/rel_notes/abi.rst b/doc/guides/rel_notes/abi.rst
index f00a6ee..fae09fd 100644
--- a/doc/guides/rel_notes/abi.rst
+++ b/doc/guides/rel_notes/abi.rst
@@ -38,3 +38,5 @@ Examples of Deprecation Notices

 Deprecation Notices
 ---
+* Structure rte_hash in librte_hash library has been changed and has been made 
private in relese 2.1, as applications should have never accessed to its 
internal data (library should have been marked as internal).
+* The Macros #RTE_HASH_BUCKET_ENTRIES_MAX and #RTE_HASH_KEY_LENGTH_MAX are 
deprecated and will be removed with version 2.2.
-- 
2.4.2

[dpdk-dev] [PATCH v3 04/11] test/hash: rename new hash perf unit test back to original name

[Pablo de Lara]

To be able to see the diff more clear, new performance unit test was
named differently from the old unit test. This patch renames the
new unit test as the old one.

Signed-off-by: Pablo de Lara 
---
 app/test/Makefile |   2 +-
 app/test/test_hash_perf.c | 560 ++
 app/test/test_hash_perf_new.c | 560 --
 3 files changed, 561 insertions(+), 561 deletions(-)
 create mode 100644 app/test/test_hash_perf.c
 delete mode 100644 app/test/test_hash_perf_new.c

diff --git a/app/test/Makefile b/app/test/Makefile
index 8624e95..2e2758c 100644
--- a/app/test/Makefile
+++ b/app/test/Makefile
@@ -82,7 +82,7 @@ SRCS-y += test_memcpy.c
 SRCS-y += test_memcpy_perf.c

 SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash.c
-SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_perf_new.c
+SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_perf.c
 SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_functions.c
 SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_scaling.c

diff --git a/app/test/test_hash_perf.c b/app/test/test_hash_perf.c
new file mode 100644
index 000..978731c
--- /dev/null
+++ b/app/test/test_hash_perf.c
@@ -0,0 +1,560 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
+ *   All rights reserved.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in
+ *   the documentation and/or other materials provided with the
+ *   distribution.
+ * * Neither the name of Intel Corporation nor the names of its
+ *   contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include 
+#include 
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include "test.h"
+
+#define KEYS_TO_ADD (1 << 18)
+#define MAX_ENTRIES (KEYS_TO_ADD * 4) /* 25% table utilization */
+#define NUM_LOOKUPS (KEYS_TO_ADD * 10) /* Loop among keys added, several times 
*/
+#define BUCKET_SIZE 4
+#define NUM_BUCKETS (MAX_ENTRIES / BUCKET_SIZE)
+#define MAX_KEYSIZE 64
+#define NUM_KEYSIZES 10
+#define NUM_SHUFFLES 10
+#define BURST_SIZE 16
+
+enum operations {
+   ADD = 0,
+   LOOKUP,
+   LOOKUP_MULTI,
+   DELETE,
+   NUM_OPERATIONS
+};
+
+static uint32_t hashtest_key_lens[] = {
+   4, 8, 16, 32, 48, 64, /* standard key sizes */
+   9,/* IPv4 SRC + DST + protocol, unpadded */
+   13,   /* IPv4 5-tuple, unpadded */
+   37,   /* IPv6 5-tuple, unpadded */
+   40/* IPv6 5-tuple, padded to 8-byte boundary */
+};
+struct rte_hash *h[NUM_KEYSIZES];
+/* Array that stores if a slot is full */
+uint8_t slot_taken[MAX_ENTRIES];
+/* Array to store number of cycles per operation */
+uint64_t cycles[NUM_KEYSIZES][NUM_OPERATIONS][2];
+/* Array to store all input keys */
+uint8_t keys[KEYS_TO_ADD][MAX_KEYSIZE];
+/* Array to store the precomputed hash for 'keys' */
+hash_sig_t signatures[KEYS_TO_ADD];
+/* Array to store how many busy entries have each bucket */
+uint8_t buckets[NUM_BUCKETS];
+
+/* Parameters used for hash table in unit test functions. */
+static struct rte_hash_parameters ut_params = {
+   .entries = MAX_ENTRIES,
+   .bucket_entries = BUCKET_SIZE,
+   .hash_func = rte_jhash,
+   .hash_func_init_val = 0,
+};
+
+static int
+create_table(unsigned table_index)
+{
+   char name[RTE_HASH_NAMESIZE];
+
+   sprintf(name, "test_hash%d", hashtest_key_lens[table_index]);
+   ut_params.name = name;
+   ut_params.key_len = hashtest_key_lens[table_index];
+   ut_params.socket_id = rte_socket_id();
+   h[table_index] = rte_hash_find_existing(name);
+  

[dpdk-dev] [PATCH v3 11/11] doc: update hash documentation

[Pablo de Lara]

Updates hash library documentation, reflecting
the new implementation changes.

Signed-off-by: Pablo de Lara 
---
 doc/guides/prog_guide/hash_lib.rst | 77 +++---
 1 file changed, 64 insertions(+), 13 deletions(-)

diff --git a/doc/guides/prog_guide/hash_lib.rst 
b/doc/guides/prog_guide/hash_lib.rst
index 9b83835..40e24bf 100644
--- a/doc/guides/prog_guide/hash_lib.rst
+++ b/doc/guides/prog_guide/hash_lib.rst
@@ -1,5 +1,5 @@
 ..  BSD LICENSE
-Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
 All rights reserved.

 Redistribution and use in source and binary forms, with or without
@@ -50,8 +50,6 @@ The hash also allows the configuration of some low-level 
implementation related

 *   Hash function to translate the key into a bucket index

-*   Number of entries per bucket
-
 The main methods exported by the hash are:

 *   Add entry with key: The key is provided as input. If a new entry is 
successfully added to the hash for the specified key,
@@ -65,10 +63,27 @@ The main methods exported by the hash are:
 *   Lookup for entry with key: The key is provided as input. If an entry with 
the specified key is found in the hash (lookup hit),
 then the position of the entry is returned, otherwise (lookup miss) a 
negative value is returned.

-The current hash implementation handles the key management only.
-The actual data associated with each key has to be managed by the user using a 
separate table that
+Apart from these method explained above, the API allows the user three more 
options:
+
+*   Add / lookup / delete with key and precomputed hash: Both the key and its 
precomputed hash are provided as input. This allows
+the user to perform these operations faster, as hash is already computed.
+
+*   Add / lookup / delete with key and data: A pair of key-value is provided 
as input. This allows the user to store
+not only the key, but also a user-defined number of bytes of data, which 
should perform better
+than storing the data in an external table
+
+*   Combination of the two options above: User can provide key, precomputed 
hash and data.
+
+Also, the API contains a method to allow the user to look up entries in 
bursts, achieving higher performance
+than looking up individual entries, as the function prefetches next entries at 
the time it is operating
+with the first ones, which reduces significantly the impact of the necessary 
memory accesses.
+Notice that this method uses a pipeline of 8 entries (4 stages of 2 entries), 
so it is highly recommended
+to use at least 8 entries per burst.
+
+The actual data associated with each key can be either managed by the user 
using a separate table that
 mirrors the hash in terms of number of entries and position of each entry,
-as shown in the Flow Classification use case describes in the following 
sections.
+as shown in the Flow Classification use case describes in the following 
sections,
+or stored in the hash table itself.

 The example hash tables in the L2/L3 Forwarding sample applications defines 
which port to forward a packet to based on a packet flow identified by the 
five-tuple lookup.
 However, this table could also be used for more sophisticated features and 
provide many other functions and actions that could be performed on the packets 
and flows.
@@ -76,17 +91,26 @@ However, this table could also be used for more 
sophisticated features and provi
 Implementation Details
 --

-The hash table is implemented as an array of entries which is further divided 
into buckets,
-with the same number of consecutive array entries in each bucket.
-For any input key, there is always a single bucket where that key can be 
stored in the hash,
-therefore only the entries within that bucket need to be examined when the key 
is looked up.
+The hash table has two main tables:
+
+* First table is an array of entries which is further divided into buckets,
+  with the same number of consecutive array entries in each bucket. Each entry 
contains the computed primary
+  and secondary hashes of a given key (explained below), and an index to the 
second table.
+
+* The second table is an array of all the keys stored in the hash table and 
its data associated to each key.
+
+The hash library uses the cuckoo hash method to resolve collisions.
+For any input key, there are two possible buckets (primary and 
secondary/alternative location)
+where that key can be stored in the hash, therefore only the entries within 
those bucket need to be examined
+when the key is looked up.
 The lookup speed is achieved by reducing the number of entries to be scanned 
from the total
-number of hash entries down to the number of entries in a hash bucket,
+number of hash entries down to the number of entries in the two hash buckets,
 as opposed to the basic method of linearly scanning all the entries in the 
array.
 The hash uses a hash

[dpdk-dev] [PATCH v3 08/11] hash: add new functionality to store data in hash table

[Pablo de Lara]

Usually hash tables not only store keys, but also data associated
to them. In order to maintain the existing API,
the key index will still be returned when
adding/looking up/deleting an entry, but user will be able
to store/look up data associated to a key.

Unit tests have been updated to use these new functions.

Signed-off-by: Pablo de Lara 
---
 app/test/Makefile|   1 +
 app/test/test_hash_perf.c| 246 ---
 lib/librte_hash/rte_cuckoo_hash.c| 167 +++-
 lib/librte_hash/rte_hash.h   | 145 -
 lib/librte_hash/rte_hash_version.map |   6 +
 5 files changed, 424 insertions(+), 141 deletions(-)

diff --git a/app/test/Makefile b/app/test/Makefile
index 2e2758c..5481ca1 100644
--- a/app/test/Makefile
+++ b/app/test/Makefile
@@ -83,6 +83,7 @@ SRCS-y += test_memcpy_perf.c

 SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash.c
 SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_perf.c
+SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_cuckoo_hash_perf.c
 SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_functions.c
 SRCS-$(CONFIG_RTE_LIBRTE_HASH) += test_hash_scaling.c

diff --git a/app/test/test_hash_perf.c b/app/test/test_hash_perf.c
index 6e8b255..2fcb9a6 100644
--- a/app/test/test_hash_perf.c
+++ b/app/test/test_hash_perf.c
@@ -55,6 +55,7 @@
 #define NUM_KEYSIZES 10
 #define NUM_SHUFFLES 10
 #define BURST_SIZE 16
+#define DATA_MULTIPLIER 13 /* Used for storing data */

 enum operations {
ADD = 0,
@@ -75,7 +76,7 @@ struct rte_hash *h[NUM_KEYSIZES];
 /* Array that stores if a slot is full */
 uint8_t slot_taken[MAX_ENTRIES];
 /* Array to store number of cycles per operation */
-uint64_t cycles[NUM_KEYSIZES][NUM_OPERATIONS][2];
+uint64_t cycles[NUM_KEYSIZES][NUM_OPERATIONS][2][2];
 /* Array to store all input keys */
 uint8_t keys[KEYS_TO_ADD][MAX_KEYSIZE];
 /* Array to store the precomputed hash for 'keys' */
@@ -92,11 +93,19 @@ static struct rte_hash_parameters ut_params = {
 };

 static int
-create_table(unsigned table_index)
+create_table(unsigned with_data, unsigned table_index)
 {
char name[RTE_HASH_NAMESIZE];

-   sprintf(name, "test_hash%d", hashtest_key_lens[table_index]);
+   if (with_data) {
+   sprintf(name, "test_hash%d_data", 
hashtest_key_lens[table_index]);
+   /* Table will store 8-byte data */
+   ut_params.data_len = sizeof(uint64_t);
+   } else {
+   sprintf(name, "test_hash%d", hashtest_key_lens[table_index]);
+   ut_params.data_len = 0;
+   }
+
ut_params.name = name;
ut_params.key_len = hashtest_key_lens[table_index];
ut_params.socket_id = rte_socket_id();
@@ -228,15 +237,25 @@ get_input_keys(unsigned table_index)
 }

 static int
-timed_adds(unsigned with_hash, unsigned table_index) {
+timed_adds(unsigned with_hash, unsigned with_data, unsigned table_index) {
unsigned i;
const uint64_t start_tsc = rte_rdtsc();
+   uint64_t data;

for (i = 0; i < KEYS_TO_ADD; i++) {
-   if (with_hash)
+   data = signatures[i] * DATA_MULTIPLIER;
+   if (with_hash && with_data)
+   rte_hash_add_key_with_hash_data(h[table_index],
+   (const void *) keys[i],
+   signatures[i], &data);
+   else if (with_hash && !with_data)
rte_hash_add_key_with_hash(h[table_index],
(const void *) keys[i],
signatures[i]);
+   else if (!with_hash && with_data)
+   rte_hash_add_key_data(h[table_index],
+   (const void *) keys[i],
+   &data);
else
rte_hash_add_key(h[table_index], keys[i]);
}
@@ -245,29 +264,38 @@ timed_adds(unsigned with_hash, unsigned table_index) {
const uint64_t time_taken = end_tsc - start_tsc;
const float seconds_taken = (float)time_taken/rte_get_tsc_hz();

-   cycles[table_index][ADD][with_hash] = time_taken/KEYS_TO_ADD;
+   cycles[table_index][ADD][with_hash][with_data] = time_taken/KEYS_TO_ADD;

printf("\n%"PRIu64" adds in %f seconds\n", (uint64_t)KEYS_TO_ADD,
seconds_taken);
printf("Average %"PRIu64" tsc ticks per add\n",
-   cycles[table_index][ADD][with_hash]);
+   cycles[table_index][ADD][with_hash][with_data]);
printf("Average %"PRIu64" adds per second\n",
(KEYS_TO_ADD * rte_get_tsc_hz())/time_taken);
return 0;
 }

 static int
-timed_lookups(unsigned with_hash, unsigned table_index)
+timed_lookups(unsigned with_hash, unsigned with_data, unsigned table_index)
 {
unsigned i, j;
const uint64_t 

[dpdk-dev] [PATCH v3 05/11] hash: replace existing hash library with cuckoo hash implementation

[Pablo de Lara]

This patch replaces the existing hash library with another approach,
using the Cuckoo Hash method to resolve collisions (open addressing),
which pushes items from a full bucket when a new entry tries
to be added in it, storing the evicted entry in an alternative location,
using a secondary hash function.

This gives the user the ability to store more entries when a bucket
is full, in comparison with the previous implementation.
Therefore, the unit test has been updated, as some scenarios have changed
(such as the previous removed restriction).

Also note that the API has not been changed, although new fields
have been added in the rte_hash structure (structure is internal now).
The main change when creating a new table is that the number of entries
per bucket is fixed now, so its parameter is ignored now
(still there to maintain the same parameters structure).
The hash unit test has been updated to reflect these changes.

As a last note, the maximum burst size in lookup_burst function
hash been increased to 64, to improve performance.

Signed-off-by: Pablo de Lara 
---
 app/test/test_hash.c  |  106 +---
 lib/librte_hash/Makefile  |8 +-
 lib/librte_hash/rte_cuckoo_hash.c | 1056 +
 lib/librte_hash/rte_hash.c|  499 --
 lib/librte_hash/rte_hash.h|   67 ++-
 5 files changed, 1128 insertions(+), 608 deletions(-)
 create mode 100644 lib/librte_hash/rte_cuckoo_hash.c
 delete mode 100644 lib/librte_hash/rte_hash.c

diff --git a/app/test/test_hash.c b/app/test/test_hash.c
index 46174db..0176219 100644
--- a/app/test/test_hash.c
+++ b/app/test/test_hash.c
@@ -169,7 +169,6 @@ static struct flow_key keys[5] = { {
 /* Parameters used for hash table in unit test functions. Name set later. */
 static struct rte_hash_parameters ut_params = {
.entries = 64,
-   .bucket_entries = 4,
.key_len = sizeof(struct flow_key), /* 13 */
.hash_func = rte_jhash,
.hash_func_init_val = 0,
@@ -516,9 +515,18 @@ static int test_five_keys(void)
pos[i] = rte_hash_lookup(handle, &keys[i]);
print_key_info("Lkp", &keys[i], pos[i]);
RETURN_IF_ERROR(pos[i] != -ENOENT,
-   "failed to find key (pos[%u]=%d)", i, pos[i]);
+   "found non-existent key (pos[%u]=%d)", i, 
pos[i]);
}

+   /* Lookup multi */
+   ret = rte_hash_lookup_multi(handle, &key_array[0], 5, (int32_t *)pos);
+   if (ret == 0)
+   for (i = 0; i < 5; i++) {
+   print_key_info("Lkp", key_array[i], pos[i]);
+   RETURN_IF_ERROR(pos[i] != -ENOENT,
+   "found not-existent key (pos[%u]=%d)", 
i, pos[i]);
+   }
+
rte_hash_free(handle);

return 0;
@@ -527,21 +535,18 @@ static int test_five_keys(void)
 /*
  * Add keys to the same bucket until bucket full.
  * - add 5 keys to the same bucket (hash created with 4 keys per bucket):
- *   first 4 successful, 5th unsuccessful
- * - lookup the 5 keys: 4 hits, 1 miss
- * - add the 5 keys again: 4 OK, one error as bucket is full
- * - lookup the 5 keys: 4 hits (updated data), 1 miss
- * - delete the 5 keys: 5 OK (even if the 5th is not in the table)
+ *   first 4 successful, 5th successful, pushing existing item in bucket
+ * - lookup the 5 keys: 5 hits
+ * - add the 5 keys again: 5 OK
+ * - lookup the 5 keys: 5 hits (updated data)
+ * - delete the 5 keys: 5 OK
  * - lookup the 5 keys: 5 misses
- * - add the 5th key: OK
- * - lookup the 5th key: hit
  */
 static int test_full_bucket(void)
 {
struct rte_hash_parameters params_pseudo_hash = {
.name = "test4",
.entries = 64,
-   .bucket_entries = 4,
.key_len = sizeof(struct flow_key), /* 13 */
.hash_func = pseudo_hash,
.hash_func_init_val = 0,
@@ -555,7 +560,7 @@ static int test_full_bucket(void)
handle = rte_hash_create(¶ms_pseudo_hash);
RETURN_IF_ERROR(handle == NULL, "hash creation failed");

-   /* Fill bucket*/
+   /* Fill bucket */
for (i = 0; i < 4; i++) {
pos[i] = rte_hash_add_key(handle, &keys[i]);
print_key_info("Add", &keys[i], pos[i]);
@@ -563,47 +568,39 @@ static int test_full_bucket(void)
"failed to add key (pos[%u]=%d)", i, pos[i]);
expected_pos[i] = pos[i];
}
-   /* This shouldn't work because the bucket is full */
+   /*
+* This should work and will push one of the items
+* in the bucket because it is full
+*/
pos[4] = rte_hash_add_key(handle, &keys[4]);
print_key_info("Add", &keys[4], pos[4]);
-   RETURN_IF_ERROR(pos[4] != -ENOSPC,
-   "fail: added key to full bucket (pos[4]=%d)", pos[4]);
+