| Message ID | 20240515160246.5181-2-larysa.zaremba@intel.com (mailing list archive) |
|---|---|
| State | Awaiting Upstream |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | Fix AF_XDP problems after changing queue number | expand |
On Wed, May 15, 2024 at 06:02:14PM +0200, Larysa Zaremba wrote: > Referenced commit has introduced a bitmap to distinguish between ZC and > copy-mode AF_XDP queues, because xsk_get_pool_from_qid() does not do this > for us. > > The bitmap would be especially useful when restoring previous state after > rebuild, if only it was not reallocated in the process. This leads to e.g. > xdpsock dying after changing number of queues. > > Instead of preserving the bitmap during the rebuild, remove it completely > and distinguish between ZC and copy-mode queues based on the presence of > a device associated with the pool. > > Fixes: e102db780e1c ("ice: track AF_XDP ZC enabled queues in bitmap") > Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> > Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com> Reviewed-by: Simon Horman <horms@kernel.org>
>-----Original Message----- >From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of >Zaremba, Larysa >Sent: Wednesday, May 15, 2024 9:32 PM >To: intel-wired-lan@lists.osuosl.org; Keller, Jacob E <jacob.e.keller@intel.com> >Cc: Fijalkowski, Maciej <maciej.fijalkowski@intel.com>; Jesper Dangaard Brouer ><hawk@kernel.org>; Daniel Borkmann <daniel@iogearbox.net>; Zaremba, >Larysa <larysa.zaremba@intel.com>; Kitszel, Przemyslaw ><przemyslaw.kitszel@intel.com>; John Fastabend ><john.fastabend@gmail.com>; Alexei Starovoitov <ast@kernel.org>; David S. >Miller <davem@davemloft.net>; Eric Dumazet <edumazet@google.com>; >netdev@vger.kernel.org; Jakub Kicinski <kuba@kernel.org>; >bpf@vger.kernel.org; Paolo Abeni <pabeni@redhat.com>; Magnus Karlsson ><magnus.karlsson@gmail.com>; Bagnucki, Igor <igor.bagnucki@intel.com>; >linux-kernel@vger.kernel.org >Subject: [Intel-wired-lan] [PATCH iwl-net 1/3] ice: remove af_xdp_zc_qps >bitmap > >Referenced commit has introduced a bitmap to distinguish between ZC and >copy-mode AF_XDP queues, because xsk_get_pool_from_qid() does not do >this for us. > >The bitmap would be especially useful when restoring previous state after >rebuild, if only it was not reallocated in the process. This leads to e.g. >xdpsock dying after changing number of queues. > >Instead of preserving the bitmap during the rebuild, remove it completely and >distinguish between ZC and copy-mode queues based on the presence of a >device associated with the pool. > >Fixes: e102db780e1c ("ice: track AF_XDP ZC enabled queues in bitmap") >Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> >Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com> >--- > drivers/net/ethernet/intel/ice/ice.h | 32 ++++++++++++++++-------- > drivers/net/ethernet/intel/ice/ice_lib.c | 8 ------ >drivers/net/ethernet/intel/ice/ice_xsk.c | 13 +++++----- > 3 files changed, 27 insertions(+), 26 deletions(-) > Tested-by: Chandan Kumar Rout <chandanx.rout@intel.com> (A Contingent Worker at Intel)
diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h index 6ad8002b22e1..d4d840729bda 100644 --- a/drivers/net/ethernet/intel/ice/ice.h +++ b/drivers/net/ethernet/intel/ice/ice.h @@ -409,7 +409,6 @@ struct ice_vsi { struct ice_tc_cfg tc_cfg; struct bpf_prog *xdp_prog; struct ice_tx_ring **xdp_rings; /* XDP ring array */ - unsigned long *af_xdp_zc_qps; /* tracks AF_XDP ZC enabled qps */ u16 num_xdp_txq; /* Used XDP queues */ u8 xdp_mapping_mode; /* ICE_MAP_MODE_[CONTIG|SCATTER] */ @@ -746,6 +745,25 @@ static inline void ice_set_ring_xdp(struct ice_tx_ring *ring) ring->flags |= ICE_TX_FLAGS_RING_XDP; } +/** + * ice_get_xp_from_qid - get ZC XSK buffer pool bound to a queue ID + * @vsi: pointer to VSI + * @qid: index of a queue to look at XSK buff pool presence + * + * Returns a pointer to xsk_buff_pool structure if there is a buffer pool + * attached and configured as zero-copy, NULL otherwise. + */ +static inline struct xsk_buff_pool *ice_get_xp_from_qid(struct ice_vsi *vsi, + u16 qid) +{ + struct xsk_buff_pool *pool = xsk_get_pool_from_qid(vsi->netdev, qid); + + if (!ice_is_xdp_ena_vsi(vsi)) + return NULL; + + return (pool && pool->dev) ? pool : NULL; +} + /** * ice_xsk_pool - get XSK buffer pool bound to a ring * @ring: Rx ring to use @@ -758,10 +776,7 @@ static inline struct xsk_buff_pool *ice_xsk_pool(struct ice_rx_ring *ring) struct ice_vsi *vsi = ring->vsi; u16 qid = ring->q_index; - if (!ice_is_xdp_ena_vsi(vsi) || !test_bit(qid, vsi->af_xdp_zc_qps)) - return NULL; - - return xsk_get_pool_from_qid(vsi->netdev, qid); + return ice_get_xp_from_qid(vsi, qid); } /** @@ -786,12 +801,7 @@ static inline void ice_tx_xsk_pool(struct ice_vsi *vsi, u16 qid) if (!ring) return; - if (!ice_is_xdp_ena_vsi(vsi) || !test_bit(qid, vsi->af_xdp_zc_qps)) { - ring->xsk_pool = NULL; - return; - } - - ring->xsk_pool = xsk_get_pool_from_qid(vsi->netdev, qid); + ring->xsk_pool = ice_get_xp_from_qid(vsi, qid); } /** diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c index 5371e91f6bbb..c0a7ff6c7e87 100644 --- a/drivers/net/ethernet/intel/ice/ice_lib.c +++ b/drivers/net/ethernet/intel/ice/ice_lib.c @@ -114,14 +114,8 @@ static int ice_vsi_alloc_arrays(struct ice_vsi *vsi) if (!vsi->q_vectors) goto err_vectors; - vsi->af_xdp_zc_qps = bitmap_zalloc(max_t(int, vsi->alloc_txq, vsi->alloc_rxq), GFP_KERNEL); - if (!vsi->af_xdp_zc_qps) - goto err_zc_qps; - return 0; -err_zc_qps: - devm_kfree(dev, vsi->q_vectors); err_vectors: devm_kfree(dev, vsi->rxq_map); err_rxq_map: @@ -309,8 +303,6 @@ static void ice_vsi_free_arrays(struct ice_vsi *vsi) dev = ice_pf_to_dev(pf); - bitmap_free(vsi->af_xdp_zc_qps); - vsi->af_xdp_zc_qps = NULL; /* free the ring and vector containers */ devm_kfree(dev, vsi->q_vectors); vsi->q_vectors = NULL; diff --git a/drivers/net/ethernet/intel/ice/ice_xsk.c b/drivers/net/ethernet/intel/ice/ice_xsk.c index aa81d1162b81..2015f66b0cf9 100644 --- a/drivers/net/ethernet/intel/ice/ice_xsk.c +++ b/drivers/net/ethernet/intel/ice/ice_xsk.c @@ -269,7 +269,6 @@ static int ice_xsk_pool_disable(struct ice_vsi *vsi, u16 qid) if (!pool) return -EINVAL; - clear_bit(qid, vsi->af_xdp_zc_qps); xsk_pool_dma_unmap(pool, ICE_RX_DMA_ATTR); return 0; @@ -300,8 +299,6 @@ ice_xsk_pool_enable(struct ice_vsi *vsi, struct xsk_buff_pool *pool, u16 qid) if (err) return err; - set_bit(qid, vsi->af_xdp_zc_qps); - return 0; } @@ -349,11 +346,13 @@ ice_realloc_rx_xdp_bufs(struct ice_rx_ring *rx_ring, bool pool_present) int ice_realloc_zc_buf(struct ice_vsi *vsi, bool zc) { struct ice_rx_ring *rx_ring; - unsigned long q; + uint i; + + ice_for_each_rxq(vsi, i) { + rx_ring = vsi->rx_rings[i]; + if (!rx_ring->xsk_pool) + continue; - for_each_set_bit(q, vsi->af_xdp_zc_qps, - max_t(int, vsi->alloc_txq, vsi->alloc_rxq)) { - rx_ring = vsi->rx_rings[q]; if (ice_realloc_rx_xdp_bufs(rx_ring, zc)) return -ENOMEM; }
Referenced commit has introduced a bitmap to distinguish between ZC and copy-mode AF_XDP queues, because xsk_get_pool_from_qid() does not do this for us. The bitmap would be especially useful when restoring previous state after rebuild, if only it was not reallocated in the process. This leads to e.g. xdpsock dying after changing number of queues. Instead of preserving the bitmap during the rebuild, remove it completely and distinguish between ZC and copy-mode queues based on the presence of a device associated with the pool. Fixes: e102db780e1c ("ice: track AF_XDP ZC enabled queues in bitmap") Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com> --- drivers/net/ethernet/intel/ice/ice.h | 32 ++++++++++++++++-------- drivers/net/ethernet/intel/ice/ice_lib.c | 8 ------ drivers/net/ethernet/intel/ice/ice_xsk.c | 13 +++++----- 3 files changed, 27 insertions(+), 26 deletions(-)