From: Rasesh Mody <rasesh.mody@cavium.com>
To: dev@dpdk.org, ferruh.yigit@intel.com
Cc: Rasesh Mody <rasesh.mody@cavium.com>, Dept-EngDPDKDev@cavium.com
Subject: [dpdk-dev] [PATCH 50/53] net/qede/base: disable aRFS for NPAR and 100G
Date: Mon, 18 Sep 2017 18:51:40 -0700 [thread overview]
Message-ID: <1505785903-1741-21-git-send-email-rasesh.mody@cavium.com> (raw)
In-Reply-To: <1505785903-1741-1-git-send-email-rasesh.mody@cavium.com>
Disable accelerated RFS for NPAR and 100G using ECORE_MF_DISABLE_ARFS
multi function mode bit.
Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
---
drivers/net/qede/base/ecore.h | 2 ++
drivers/net/qede/base/ecore_cxt.c | 10 +++++++++-
drivers/net/qede/base/ecore_dev.c | 6 +++++-
drivers/net/qede/base/ecore_l2.c | 3 +++
drivers/net/qede/qede_if.h | 1 +
drivers/net/qede/qede_main.c | 2 ++
6 files changed, 22 insertions(+), 2 deletions(-)
diff --git a/drivers/net/qede/base/ecore.h b/drivers/net/qede/base/ecore.h
index 3b51fc2..dc09847 100644
--- a/drivers/net/qede/base/ecore.h
+++ b/drivers/net/qede/base/ecore.h
@@ -528,6 +528,8 @@ enum ecore_mf_mode_bit {
/* TODO - if we ever re-utilize any of this logic, we can rename */
ECORE_MF_UFP_SPECIFIC,
+
+ ECORE_MF_DISABLE_ARFS,
};
enum ecore_ufp_mode {
diff --git a/drivers/net/qede/base/ecore_cxt.c b/drivers/net/qede/base/ecore_cxt.c
index 3ebeb12..fed7926 100644
--- a/drivers/net/qede/base/ecore_cxt.c
+++ b/drivers/net/qede/base/ecore_cxt.c
@@ -1991,6 +1991,8 @@ enum _ecore_status_t ecore_cxt_set_pf_params(struct ecore_hwfn *p_hwfn)
switch (p_hwfn->hw_info.personality) {
case ECORE_PCI_ETH:
{
+ u32 count = 0;
+
struct ecore_eth_pf_params *p_params =
&p_hwfn->pf_params.eth_pf_params;
@@ -1999,7 +2001,13 @@ enum _ecore_status_t ecore_cxt_set_pf_params(struct ecore_hwfn *p_hwfn)
ecore_cxt_set_proto_cid_count(p_hwfn, PROTOCOLID_ETH,
p_params->num_cons,
p_params->num_vf_cons);
- p_hwfn->p_cxt_mngr->arfs_count = p_params->num_arfs_filters;
+
+ count = p_params->num_arfs_filters;
+
+ if (!OSAL_TEST_BIT(ECORE_MF_DISABLE_ARFS,
+ &p_hwfn->p_dev->mf_bits))
+ p_hwfn->p_cxt_mngr->arfs_count = count;
+
break;
}
default:
diff --git a/drivers/net/qede/base/ecore_dev.c b/drivers/net/qede/base/ecore_dev.c
index 0568470..9511110 100644
--- a/drivers/net/qede/base/ecore_dev.c
+++ b/drivers/net/qede/base/ecore_dev.c
@@ -3501,7 +3501,8 @@ static enum _ecore_status_t ecore_hw_get_resc(struct ecore_hwfn *p_hwfn,
p_hwfn->p_dev->mf_bits = 1 << ECORE_MF_LLH_MAC_CLSS |
1 << ECORE_MF_LLH_PROTO_CLSS |
1 << ECORE_MF_LL2_NON_UNICAST |
- 1 << ECORE_MF_INTER_PF_SWITCH;
+ 1 << ECORE_MF_INTER_PF_SWITCH |
+ 1 << ECORE_MF_DISABLE_ARFS;
break;
case NVM_CFG1_GLOB_MF_MODE_DEFAULT:
p_hwfn->p_dev->mf_bits = 1 << ECORE_MF_LLH_MAC_CLSS |
@@ -3514,6 +3515,9 @@ static enum _ecore_status_t ecore_hw_get_resc(struct ecore_hwfn *p_hwfn,
DP_INFO(p_hwfn, "Multi function mode is 0x%lx\n",
p_hwfn->p_dev->mf_bits);
+ if (ECORE_IS_CMT(p_hwfn->p_dev))
+ p_hwfn->p_dev->mf_bits |= (1 << ECORE_MF_DISABLE_ARFS);
+
/* It's funny since we have another switch, but it's easier
* to throw this away in linux this way. Long term, it might be
* better to have have getters for needed ECORE_MF_* fields,
diff --git a/drivers/net/qede/base/ecore_l2.c b/drivers/net/qede/base/ecore_l2.c
index 01fe880..e3afc8a 100644
--- a/drivers/net/qede/base/ecore_l2.c
+++ b/drivers/net/qede/base/ecore_l2.c
@@ -2072,6 +2072,9 @@ void ecore_arfs_mode_configure(struct ecore_hwfn *p_hwfn,
struct ecore_ptt *p_ptt,
struct ecore_arfs_config_params *p_cfg_params)
{
+ if (OSAL_TEST_BIT(ECORE_MF_DISABLE_ARFS, &p_hwfn->p_dev->mf_bits))
+ return;
+
if (p_cfg_params->arfs_enable) {
ecore_gft_config(p_hwfn, p_ptt, p_hwfn->rel_pf_id,
p_cfg_params->tcp,
diff --git a/drivers/net/qede/qede_if.h b/drivers/net/qede/qede_if.h
index 1f97b59..bf80ccb 100644
--- a/drivers/net/qede/qede_if.h
+++ b/drivers/net/qede/qede_if.h
@@ -40,6 +40,7 @@ struct qed_dev_info {
#define QED_MFW_VERSION_3_OFFSET 24
uint32_t flash_size;
+ bool b_arfs_capable;
bool b_inter_pf_switch;
bool tx_switching;
u16 mtu;
diff --git a/drivers/net/qede/qede_main.c b/drivers/net/qede/qede_main.c
index 2f6a4dc..a0c9e03 100644
--- a/drivers/net/qede/qede_main.c
+++ b/drivers/net/qede/qede_main.c
@@ -378,6 +378,8 @@ static int qed_slowpath_start(struct ecore_dev *edev,
if (IS_PF(edev)) {
dev_info->b_inter_pf_switch =
OSAL_TEST_BIT(ECORE_MF_INTER_PF_SWITCH, &edev->mf_bits);
+ if (!OSAL_TEST_BIT(ECORE_MF_DISABLE_ARFS, &edev->mf_bits))
+ dev_info->b_arfs_capable = true;
dev_info->tx_switching = false;
dev_info->smart_an = ecore_mcp_is_smart_an_supported(p_hwfn);
--
1.7.10.3
next prev parent reply other threads:[~2017-09-19 1:52 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-09-19 1:51 [dpdk-dev] [PATCH 30/53] net/qede/base: read per queue coalescing from HW Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 31/53] net/qede/base: refactor device's number of ports logic Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 32/53] net/qede/base: use proper units for rate limiting Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 33/53] net/qede/base: use available macro Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 34/53] net/qede/base: use function pointers for spq async callback Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 35/53] net/qede/base: fix API return types Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 36/53] net/qede/base: semantic changes Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 37/53] net/qede/base: handle the error condition properly Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 38/53] net/qede/base: add new macro for CMT mode Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 39/53] net/qede/base: change verbosity Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 40/53] net/qede/base: fix number of app table entries Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 41/53] net/qede/base: update firmware to 8.30.12.0 Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 42/53] net/qede/base: add UFP support Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 43/53] net/qede/base: add support for mapped doorbell Bars for VFs Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 44/53] net/qede/base: add support for driver attribute repository Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 45/53] net/qede/base: move define to header file Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 46/53] net/qede/base: dcbx dscp related extensions Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 47/53] net/qede/base: add feature support for per-PF virtual link Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 48/53] net/qede/base: catch an init command write failure Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 49/53] net/qede/base: retain dcbx config till actually applied Rasesh Mody
2017-09-19 1:51 ` Rasesh Mody [this message]
2017-09-19 1:51 ` [dpdk-dev] [PATCH 51/53] net/qede/base: add support for WoL writes Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 52/53] net/qede/base: remove unused input parameter Rasesh Mody
2017-09-19 1:51 ` [dpdk-dev] [PATCH 53/53] net/qede/base: update PMD version to 2.6.0.1 Rasesh Mody
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=1505785903-1741-21-git-send-email-rasesh.mody@cavium.com \
--to=rasesh.mody@cavium.com \
--cc=Dept-EngDPDKDev@cavium.com \
--cc=dev@dpdk.org \
--cc=ferruh.yigit@intel.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;
as well as URLs for NNTP newsgroup(s).