[dpdk-dev] [PATCH v2 4/4] net/ixgbe: delete HW rings when releasing queues

[Wang, Haiyue]

> -----Original Message-----
> From: wangyunjian <wangyunjian at huawei.com>
> Sent: Wednesday, September 22, 2021 14:58
> To: Wang, Haiyue <haiyue.wang at intel.com>; dev at dpdk.org
> Cc: Xing, Beilei <beilei.xing at intel.com>; Yang, Qiming <qiming.yang at intel.com>; Zhang, Qi Z
> <qi.z.zhang at intel.com>; dingxiaoxiong <dingxiaoxiong at huawei.com>
> Subject: RE: [dpdk-dev] [PATCH v2 4/4] net/ixgbe: delete HW rings when releasing queues
> 
> > -----Original Message-----
> > From: Wang, Haiyue [mailto:haiyue.wang at intel.com]
> > Sent: Wednesday, September 22, 2021 2:23 PM
> > To: wangyunjian <wangyunjian at huawei.com>; dev at dpdk.org
> > Cc: Xing, Beilei <beilei.xing at intel.com>; Yang, Qiming
> > <qiming.yang at intel.com>; Zhang, Qi Z <qi.z.zhang at intel.com>; dingxiaoxiong
> > <dingxiaoxiong at huawei.com>
> > Subject: RE: [dpdk-dev] [PATCH v2 4/4] net/ixgbe: delete HW rings when
> > releasing queues
> >
> > > -----Original Message-----
> > > From: Yunjian Wang <wangyunjian at huawei.com>
> > > Sent: Saturday, September 18, 2021 16:42
> > > To: dev at dpdk.org
> > > Cc: Wang, Haiyue <haiyue.wang at intel.com>; Xing, Beilei
> > > <beilei.xing at intel.com>; Yang, Qiming <qiming.yang at intel.com>; Zhang,
> > > Qi Z <qi.z.zhang at intel.com>; dingxiaoxiong at huawei.com; Yunjian Wang
> > > <wangyunjian at huawei.com>
> > > Subject: [dpdk-dev] [PATCH v2 4/4] net/ixgbe: delete HW rings when
> > > releasing queues
> > >
> > > Normally when closing the device the queue memzone should be freed.
> > > But the memzone will be not freed, when device setup ops like:
> > >
> > > rte_eth_bond_slave_remove
> > > -->__eth_bond_slave_remove_lock_free
> > > ---->slave_remove
> > > ------>rte_eth_dev_internal_reset
> > > -------->rte_eth_dev_rx_queue_config
> > > ---------->eth_dev_rx_queue_config
> > > ------------>ixgbe_dev_rx_queue_release
> > > rte_eth_dev_close
> > > -->ixgbe_dev_close
> > > ---->ixgbe_dev_free_queues
> > > ------>ixgbe_dev_rx_queue_release
> > >       (not been called due to nb_rx_queues and nb_tx_queues are 0)
> > >
> > > In order to free the memzone, we can release the memzone when
> > > releasing queues.
> > >
> >
> > After re-check the eth dev API, I think we can simplify the commit message to
> > such as:
> >
> > Fix memzone leak when re-configure the RX/TX queues.
> >
> > Please see 'rte_eth_dev_configure', when queue number is changed to small
> > size, the BIG memzone queue index will be lost. This will make it is a MUST
> > fix. ;-)
> >
> > And add the Fixes tag and CC to stable.
> >
> > What do you think ?
> 
> OK, thanks for your suggestion, will include them in next version.
> 

Also, please fix the patch format warnings. You can get this
from patchwork CI.

FYI:

https://doc.dpdk.org/guides/contributing/patches.html

6.9. Checking the Patches

> >
> > > Signed-off-by: Yunjian Wang <wangyunjian at huawei.com>
> > > ---
> > >  drivers/net/ixgbe/ixgbe_rxtx.c | 6 ++++--
> > > drivers/net/ixgbe/ixgbe_rxtx.h | 2 ++
> > >  2 files changed, 6 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/drivers/net/ixgbe/ixgbe_rxtx.c
> > > b/drivers/net/ixgbe/ixgbe_rxtx.c index bfdfd5e755..1b6e0489f4 100644
> > > --- a/drivers/net/ixgbe/ixgbe_rxtx.c
> > > +++ b/drivers/net/ixgbe/ixgbe_rxtx.c
> > > @@ -2482,6 +2482,7 @@ ixgbe_tx_queue_release(struct ixgbe_tx_queue
> > *txq)
> > >  	if (txq != NULL && txq->ops != NULL) {
> > >  		txq->ops->release_mbufs(txq);
> > >  		txq->ops->free_swring(txq);
> > > +		rte_memzone_free(txq->mz);
> > >  		rte_free(txq);
> > >  	}
> > >  }
> > > @@ -2763,6 +2764,7 @@ ixgbe_dev_tx_queue_setup(struct rte_eth_dev
> > *dev,
> > >  		return -ENOMEM;
> > >  	}
> > >
> > > +	txq->mz = tz;
> > >  	txq->nb_tx_desc = nb_desc;
> > >  	txq->tx_rs_thresh = tx_rs_thresh;
> > >  	txq->tx_free_thresh = tx_free_thresh; @@ -2887,6 +2889,7 @@
> > > ixgbe_rx_queue_release(struct ixgbe_rx_queue *rxq)
> > >  		ixgbe_rx_queue_release_mbufs(rxq);
> > >  		rte_free(rxq->sw_ring);
> > >  		rte_free(rxq->sw_sc_ring);
> > > +		rte_memzone_free(rxq->mz);
> > >  		rte_free(rxq);
> > >  	}
> > >  }
> > > @@ -3162,6 +3165,7 @@ ixgbe_dev_rx_queue_setup(struct rte_eth_dev
> > *dev,
> > >  		return -ENOMEM;
> > >  	}
> > >
> > > +	rxq->mz = rz;
> > >  	/*
> > >  	 * Zero init all the descriptors in the ring.
> > >  	 */
> > > @@ -3433,14 +3437,12 @@ ixgbe_dev_free_queues(struct rte_eth_dev *dev)
> > >  	for (i = 0; i < dev->data->nb_rx_queues; i++) {
> > >  		ixgbe_dev_rx_queue_release(dev->data->rx_queues[i]);
> > >  		dev->data->rx_queues[i] = NULL;
> > > -		rte_eth_dma_zone_free(dev, "rx_ring", i);
> > >  	}
> > >  	dev->data->nb_rx_queues = 0;
> > >
> > >  	for (i = 0; i < dev->data->nb_tx_queues; i++) {
> > >  		ixgbe_dev_tx_queue_release(dev->data->tx_queues[i]);
> > >  		dev->data->tx_queues[i] = NULL;
> > > -		rte_eth_dma_zone_free(dev, "tx_ring", i);
> > >  	}
> > >  	dev->data->nb_tx_queues = 0;
> > >  }
> > > diff --git a/drivers/net/ixgbe/ixgbe_rxtx.h
> > > b/drivers/net/ixgbe/ixgbe_rxtx.h index 476ef62cfd..a1764f2b08 100644
> > > --- a/drivers/net/ixgbe/ixgbe_rxtx.h
> > > +++ b/drivers/net/ixgbe/ixgbe_rxtx.h
> > > @@ -138,6 +138,7 @@ struct ixgbe_rx_queue {
> > >  	struct rte_mbuf fake_mbuf;
> > >  	/** hold packets to return to application */
> > >  	struct rte_mbuf *rx_stage[RTE_PMD_IXGBE_RX_MAX_BURST*2];
> > > +	const struct rte_memzone *mz;
> > >  };
> > >
> > >  /**
> > > @@ -236,6 +237,7 @@ struct ixgbe_tx_queue {
> > >  	uint8_t		    using_ipsec;
> > >  	/**< indicates that IPsec TX feature is in use */  #endif
> > > +	const struct rte_memzone *mz;
> > >  };
> > >
> > >  struct ixgbe_txq_ops {
> > > --
> > > 2.23.0