stable.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Saeed Mahameed <saeedm@mellanox.com>
Subject: [PATCH 5.0 23/36] net/mlx5e: Disable rxhash when CQE compress is enabled
Date: Mon,  3 Jun 2019 11:09:11 +0200	[thread overview]
Message-ID: <20190603090522.514888806@linuxfoundation.org> (raw)
In-Reply-To: <20190603090520.998342694@linuxfoundation.org>

From: Saeed Mahameed <saeedm@mellanox.com>

[ Upstream commit c0194e2d0ef0e5ce5e21a35640d23a706827ae28 ]

When CQE compression is enabled (Multi-host systems), compressed CQEs
might arrive to the driver rx, compressed CQEs don't have a valid hash
offload and the driver already reports a hash value of 0 and invalid hash
type on the skb for compressed CQEs, but this is not good enough.

On a congested PCIe, where CQE compression will kick in aggressively,
gro will deliver lots of out of order packets due to the invalid hash
and this might cause a serious performance drop.

The only valid solution, is to disable rxhash offload at all when CQE
compression is favorable (Multi-host systems).

Fixes: 7219ab34f184 ("net/mlx5e: CQE compression")
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c |   13 +++++++++++++
 1 file changed, 13 insertions(+)

--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -3789,6 +3789,12 @@ static netdev_features_t mlx5e_fix_featu
 			netdev_warn(netdev, "Disabling LRO, not supported in legacy RQ\n");
 	}
 
+	if (MLX5E_GET_PFLAG(params, MLX5E_PFLAG_RX_CQE_COMPRESS)) {
+		features &= ~NETIF_F_RXHASH;
+		if (netdev->features & NETIF_F_RXHASH)
+			netdev_warn(netdev, "Disabling rxhash, not supported when CQE compress is active\n");
+	}
+
 	mutex_unlock(&priv->state_lock);
 
 	return features;
@@ -3915,6 +3921,9 @@ int mlx5e_hwstamp_set(struct mlx5e_priv
 	memcpy(&priv->tstamp, &config, sizeof(config));
 	mutex_unlock(&priv->state_lock);
 
+	/* might need to fix some features */
+	netdev_update_features(priv->netdev);
+
 	return copy_to_user(ifr->ifr_data, &config,
 			    sizeof(config)) ? -EFAULT : 0;
 }
@@ -4744,6 +4753,10 @@ static void mlx5e_build_nic_netdev(struc
 	if (!priv->channels.params.scatter_fcs_en)
 		netdev->features  &= ~NETIF_F_RXFCS;
 
+	/* prefere CQE compression over rxhash */
+	if (MLX5E_GET_PFLAG(&priv->channels.params, MLX5E_PFLAG_RX_CQE_COMPRESS))
+		netdev->features &= ~NETIF_F_RXHASH;
+
 #define FT_CAP(f) MLX5_CAP_FLOWTABLE(mdev, flow_table_properties_nic_receive.f)
 	if (FT_CAP(flow_modify_en) &&
 	    FT_CAP(modify_root) &&



  parent reply	other threads:[~2019-06-03  9:12 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-06-03  9:08 [PATCH 5.0 00/36] 5.0.21-stable review Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 01/36] bonding/802.3ad: fix slave link initialization transition states Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 02/36] cxgb4: offload VLAN flows regardless of VLAN ethtype Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 03/36] inet: switch IP ID generator to siphash Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 04/36] ipv4/igmp: fix another memory leak in igmpv3_del_delrec() Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 05/36] ipv4/igmp: fix build error if !CONFIG_IP_MULTICAST Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 06/36] ipv6: Consider sk_bound_dev_if when binding a raw socket to an address Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 07/36] ipv6: Fix redirect with VRF Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 08/36] llc: fix skb leak in llc_build_and_send_ui_pkt() Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 09/36] mlxsw: spectrum_acl: Avoid warning after identical rules insertion Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 10/36] net: dsa: mv88e6xxx: fix handling of upper half of STATS_TYPE_PORT Greg Kroah-Hartman
2019-06-03  9:08 ` [PATCH 5.0 11/36] net: fec: fix the clk mismatch in failed_reset path Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 12/36] net-gro: fix use-after-free read in napi_gro_frags() Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 13/36] net: mvneta: Fix err code path of probe Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 14/36] net: mvpp2: fix bad MVPP2_TXQ_SCHED_TOKEN_CNTR_REG queue value Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 15/36] net: phy: marvell10g: report if the PHY fails to boot firmware Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 16/36] net: sched: dont use tc_action->order during action dump Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 17/36] net: stmmac: fix reset gpio free missing Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 18/36] r8169: fix MAC address being lost in PCI D3 Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 19/36] usbnet: fix kernel crash after disconnect Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 20/36] net/mlx5: Avoid double free in fs init error unwinding path Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 21/36] tipc: Avoid copying bytes beyond the supplied data Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 22/36] net/mlx5: Allocate root ns memory using kzalloc to match kfree Greg Kroah-Hartman
2019-06-03  9:09 ` Greg Kroah-Hartman [this message]
2019-06-03  9:09 ` [PATCH 5.0 25/36] net: stmmac: dma channel control register need to be init first Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 26/36] bnxt_en: Fix aggregation buffer leak under OOM condition Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 27/36] bnxt_en: Fix possible BUG() condition when calling pci_disable_msix() Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 28/36] bnxt_en: Reduce memory usage when running in kdump kernel Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 29/36] net/tls: fix state removal with feature flags off Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 30/36] net/tls: dont ignore netdev notifications if no TLS features Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 31/36] cxgb4: Revert "cxgb4: Remove SGE_HOST_PAGE_SIZE dependency on page size" Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 32/36] net: correct zerocopy refcnt with udp MSG_MORE Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 33/36] crypto: vmx - ghash: do nosimd fallback manually Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 34/36] xen/pciback: Dont disable PCI_COMMAND on PCI device reset Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 35/36] Revert "tipc: fix modprobe tipc failed after switch order of device registration" Greg Kroah-Hartman
2019-06-03  9:09 ` [PATCH 5.0 36/36] tipc: fix modprobe tipc failed after switch order of device registration Greg Kroah-Hartman
2019-06-03 15:09 ` [PATCH 5.0 00/36] 5.0.21-stable review kernelci.org bot
2019-06-03 17:17 ` Guenter Roeck
2019-06-03 17:59 ` Naresh Kamboju
2019-06-03 18:33 ` Jon Hunter
2019-06-03 23:32 ` shuah

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190603090522.514888806@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=saeedm@mellanox.com \
    --cc=stable@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).