Netdev List
 help / color / mirror / Atom feed
From: Huazhong Tan <tanhuazhong@huawei.com>
To: <davem@davemloft.net>
Cc: <netdev@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
	<huangdaode@hisilicon.com>, <yisen.zhuang@huawei.com>,
	<salil.mehta@huawei.com>, <linuxarm@huawei.com>,
	Jian Shen <shenjian15@huawei.com>, Peng Li <lipeng321@huawei.com>,
	Huazhong Tan <tanhuazhong@huawei.com>
Subject: [PATCH net-next 08/12] net: hns3: fix for rss result nonuniform
Date: Thu, 31 Jan 2019 04:55:48 +0800	[thread overview]
Message-ID: <20190130205552.8512-9-tanhuazhong@huawei.com> (raw)
In-Reply-To: <20190130205552.8512-1-tanhuazhong@huawei.com>

From: Jian Shen <shenjian15@huawei.com>

The rss result is more uniform when use recommended hash key from
microsoft, instead of the one generated by netdev_rss_key_fill().
Also using hash algorithm "xor" is better than "toeplitz".

This patch modifies the default hash key and hash algorithm.

Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
 .../hisilicon/hns3/hns3pf/hclge_main.c         | 18 +++++++++++++++---
 .../hisilicon/hns3/hns3vf/hclgevf_main.c       | 14 +++++++++++---
 2 files changed, 26 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index 795ebedde284..42f0f8824b47 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -295,6 +295,14 @@ static const struct hclge_mac_mgr_tbl_entry_cmd hclge_mgr_table[] = {
 	},
 };
 
+static const u8 hclge_hash_key[] = {
+	0x6D, 0x5A, 0x56, 0xDA, 0x25, 0x5B, 0x0E, 0xC2,
+	0x41, 0x67, 0x25, 0x3D, 0x43, 0xA3, 0x8F, 0xB0,
+	0xD0, 0xCA, 0x2B, 0xCB, 0xAE, 0x7B, 0x30, 0xB4,
+	0x77, 0xCB, 0x2D, 0xA3, 0x80, 0x30, 0xF2, 0x0C,
+	0x6A, 0x42, 0xB7, 0x3B, 0xBE, 0xAC, 0x01, 0xFA
+};
+
 static int hclge_mac_update_stats_defective(struct hclge_dev *hdev)
 {
 #define HCLGE_MAC_CMD_NUM 21
@@ -3652,8 +3660,11 @@ void hclge_rss_indir_init_cfg(struct hclge_dev *hdev)
 
 static void hclge_rss_init_cfg(struct hclge_dev *hdev)
 {
+	int i, rss_algo = HCLGE_RSS_HASH_ALGO_TOEPLITZ;
 	struct hclge_vport *vport = hdev->vport;
-	int i;
+
+	if (hdev->pdev->revision >= 0x21)
+		rss_algo = HCLGE_RSS_HASH_ALGO_SIMPLE;
 
 	for (i = 0; i < hdev->num_vmdq_vport + 1; i++) {
 		vport[i].rss_tuple_sets.ipv4_tcp_en =
@@ -3673,9 +3684,10 @@ static void hclge_rss_init_cfg(struct hclge_dev *hdev)
 		vport[i].rss_tuple_sets.ipv6_fragment_en =
 			HCLGE_RSS_INPUT_TUPLE_OTHER;
 
-		vport[i].rss_algo = HCLGE_RSS_HASH_ALGO_TOEPLITZ;
+		vport[i].rss_algo = rss_algo;
 
-		netdev_rss_key_fill(vport[i].rss_hash_key, HCLGE_RSS_KEY_SIZE);
+		memcpy(vport[i].rss_hash_key, hclge_hash_key,
+		       HCLGE_RSS_KEY_SIZE);
 	}
 
 	hclge_rss_indir_init_cfg(hdev);
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
index fc99a0c70164..93f306f42cf4 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
@@ -21,6 +21,14 @@ static const struct pci_device_id ae_algovf_pci_tbl[] = {
 	{0, }
 };
 
+static const u8 hclgevf_hash_key[] = {
+	0x6D, 0x5A, 0x56, 0xDA, 0x25, 0x5B, 0x0E, 0xC2,
+	0x41, 0x67, 0x25, 0x3D, 0x43, 0xA3, 0x8F, 0xB0,
+	0xD0, 0xCA, 0x2B, 0xCB, 0xAE, 0x7B, 0x30, 0xB4,
+	0x77, 0xCB, 0x2D, 0xA3, 0x80, 0x30, 0xF2, 0x0C,
+	0x6A, 0x42, 0xB7, 0x3B, 0xBE, 0xAC, 0x01, 0xFA
+};
+
 MODULE_DEVICE_TABLE(pci, ae_algovf_pci_tbl);
 
 static const u32 cmdq_reg_addr_list[] = {HCLGEVF_CMDQ_TX_ADDR_L_REG,
@@ -1789,9 +1797,9 @@ static int hclgevf_rss_init_hw(struct hclgevf_dev *hdev)
 	rss_cfg->rss_size = hdev->rss_size_max;
 
 	if (hdev->pdev->revision >= 0x21) {
-		rss_cfg->hash_algo = HCLGEVF_RSS_HASH_ALGO_TOEPLITZ;
-		netdev_rss_key_fill(rss_cfg->rss_hash_key,
-				    HCLGEVF_RSS_KEY_SIZE);
+		rss_cfg->hash_algo = HCLGEVF_RSS_HASH_ALGO_SIMPLE;
+		memcpy(rss_cfg->rss_hash_key, hclgevf_hash_key,
+		       HCLGEVF_RSS_KEY_SIZE);
 
 		ret = hclgevf_set_rss_algo_key(hdev, rss_cfg->hash_algo,
 					       rss_cfg->rss_hash_key);
-- 
2.20.1



  parent reply	other threads:[~2019-01-30 20:58 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-01-30 20:55 [PATCH net-next 00/12] code optimizations & bugfixes for HNS3 driver Huazhong Tan
2019-01-30 20:55 ` [PATCH net-next 01/12] net: hns3: reuse the definition of l3 and l4 header info union Huazhong Tan
2019-01-30 20:55 ` [PATCH net-next 02/12] net: hns3: fix VF dump register issue Huazhong Tan
2019-01-30 20:55 ` [PATCH net-next 03/12] net: hns3: use the correct interface to stop|open port Huazhong Tan
2019-01-30 20:55 ` [PATCH net-next 04/12] net: hns3: change hnae3_register_ae_dev() to int Huazhong Tan
2019-01-30 20:55 ` [PATCH net-next 05/12] net: hns3: only support tc 0 for VF Huazhong Tan
2019-01-30 20:55 ` [PATCH net-next 06/12] net: hns3: Fix NULL deref when unloading driver Huazhong Tan
2019-01-30 20:55 ` [PATCH net-next 07/12] net: hns3: fix netif_napi_del() not do problem when unloading Huazhong Tan
2019-01-30 20:55 ` Huazhong Tan [this message]
2019-01-30 20:55 ` [PATCH net-next 09/12] net: hns3: fix improper error handling in the hclge_init_ae_dev() Huazhong Tan
2019-01-30 20:55 ` [PATCH net-next 10/12] net: hns3: fix an issue for hclgevf_ae_get_hdev Huazhong Tan
2019-01-30 20:55 ` [PATCH net-next 11/12] net: hns3: stop sending keep alive msg to PF when VF is resetting Huazhong Tan
2019-01-30 20:55 ` [PATCH net-next 12/12] net: hns3: keep flow director state unchanged when reset Huazhong Tan
2019-01-30 22:50 ` [PATCH net-next 00/12] code optimizations & bugfixes for HNS3 driver David Miller

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=20190130205552.8512-9-tanhuazhong@huawei.com \
    --to=tanhuazhong@huawei.com \
    --cc=davem@davemloft.net \
    --cc=huangdaode@hisilicon.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linuxarm@huawei.com \
    --cc=lipeng321@huawei.com \
    --cc=netdev@vger.kernel.org \
    --cc=salil.mehta@huawei.com \
    --cc=shenjian15@huawei.com \
    --cc=yisen.zhuang@huawei.com \
    /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