[dpdk-dev] [PATCH] common: introduce an integer log2 function

[Jerin Jacob]

-----Original Message-----
> Date: Thu, 6 Apr 2017 16:23:55 +0200
> From: Olivier MATZ <olivier.matz at 6wind.com>
> To: dev at dpdk.org, Jerin Jacob <jerin.jacob at caviumnetworks.com>
> Cc: alejandro.lucero at netronome.com
> Subject: Re: [dpdk-dev] [PATCH] common: introduce an integer log2 function
> X-Mailer: Claws Mail 3.14.1 (GTK+ 2.24.31; x86_64-pc-linux-gnu)
> 
> I forgot to CC Alejandro for nfp.
> 
> Also CC Jerin: I think the same could be done in thunderx, but
> I did not want to do the patch since it changes the base/ driver.
> 
> 
> On Thu,  6 Apr 2017 16:15:36 +0200, Olivier Matz <olivier.matz at 6wind.com> wrote:
> > At some places, the log2() function is used despite this function
> > works on float. This introduces a dependency to the math lib but
> > most of the time it is not required because we want an integer log2.
> > 
> > Add a new helper to do this job and fix 2 drivers.
> > 
> > Signed-off-by: Olivier Matz <olivier.matz at 6wind.com>
> > ---
> >  drivers/net/nfp/nfp_net.c                  |  6 ++----
> >  lib/librte_eal/common/include/rte_common.h | 17 +++++++++++++++++
> >  mk/rte.app.mk                              |  2 +-
> >  3 files changed, 20 insertions(+), 5 deletions(-)
> > 
> > diff --git a/drivers/net/nfp/nfp_net.c b/drivers/net/nfp/nfp_net.c
> > index a1ad97ad5..0e6a729ef 100644
> > --- a/drivers/net/nfp/nfp_net.c
> > +++ b/drivers/net/nfp/nfp_net.c
> > @@ -39,8 +39,6 @@
> >   * Netronome vNIC DPDK Poll-Mode Driver: Main entry point
> >   */
> >  
> > -#include <math.h>
> > -
> >  #include <rte_byteorder.h>
> >  #include <rte_common.h>
> >  #include <rte_log.h>
> > @@ -1493,7 +1491,7 @@ nfp_net_rx_queue_setup(struct rte_eth_dev *dev,
> >  	 * of descriptors in log2 format
> >  	 */
> >  	nn_cfg_writeq(hw, NFP_NET_CFG_RXR_ADDR(queue_idx), rxq->dma);
> > -	nn_cfg_writeb(hw, NFP_NET_CFG_RXR_SZ(queue_idx), log2(nb_desc));
> > +	nn_cfg_writeb(hw, NFP_NET_CFG_RXR_SZ(queue_idx), rte_log2_u32(nb_desc));
> >  
> >  	return 0;
> >  }
> > @@ -1647,7 +1645,7 @@ nfp_net_tx_queue_setup(struct rte_eth_dev *dev, uint16_t queue_idx,
> >  	 * of descriptors in log2 format
> >  	 */
> >  	nn_cfg_writeq(hw, NFP_NET_CFG_TXR_ADDR(queue_idx), txq->dma);
> > -	nn_cfg_writeb(hw, NFP_NET_CFG_TXR_SZ(queue_idx), log2(nb_desc));
> > +	nn_cfg_writeb(hw, NFP_NET_CFG_TXR_SZ(queue_idx), rte_log2_u32(nb_desc));
> >  
> >  	return 0;
> >  }
> > diff --git a/lib/librte_eal/common/include/rte_common.h b/lib/librte_eal/common/include/rte_common.h
> > index e057f6e21..eea0054c1 100644
> > --- a/lib/librte_eal/common/include/rte_common.h
> > +++ b/lib/librte_eal/common/include/rte_common.h
> > @@ -326,6 +326,23 @@ rte_bsf32(uint32_t v)
> >  	return __builtin_ctz(v);
> >  }
> >  
> > +/**
> > + * Return the rounded-up log2 of a integer.
> > + *
> > + * @param v
> > + *     The input parameter.
> > + * @return
> > + *     The rounded-up log2 of the input, or 0 if the input is 0.
> > + */
> > +static inline uint32_t
> > +rte_log2_u32(uint32_t v)
> > +{
> > +	if (v == 0)
> > +		return 0;
> > +	v = rte_align32pow2(v);
> > +	return rte_bsf32(v);
> > +}

I think, It will good be if we add unit test case for the same.

> > +
> >  #ifndef offsetof
> >  /** Return the offset of a field in a structure. */
> >  #define offsetof(TYPE, MEMBER)  __builtin_offsetof (TYPE, MEMBER)
> > diff --git a/mk/rte.app.mk b/mk/rte.app.mk
> > index 4c659e971..2fcd388e6 100644
> > --- a/mk/rte.app.mk
> > +++ b/mk/rte.app.mk
> > @@ -125,7 +125,7 @@ endif
> >  _LDLIBS-$(CONFIG_RTE_LIBRTE_LIO_PMD)        += -lrte_pmd_lio
> >  _LDLIBS-$(CONFIG_RTE_LIBRTE_MLX4_PMD)       += -lrte_pmd_mlx4 -libverbs
> >  _LDLIBS-$(CONFIG_RTE_LIBRTE_MLX5_PMD)       += -lrte_pmd_mlx5 -libverbs
> > -_LDLIBS-$(CONFIG_RTE_LIBRTE_NFP_PMD)        += -lrte_pmd_nfp -lm
> > +_LDLIBS-$(CONFIG_RTE_LIBRTE_NFP_PMD)        += -lrte_pmd_nfp
> >  _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_NULL)       += -lrte_pmd_null
> >  _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_PCAP)       += -lrte_pmd_pcap -lpcap
> >  _LDLIBS-$(CONFIG_RTE_LIBRTE_QEDE_PMD)       += -lrte_pmd_qede
>