* [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements
@ 2024-02-20 23:03 Michael Chan
2024-02-20 23:03 ` [PATCH net-next 01/10] bnxt_en: Refactor ring reservation functions Michael Chan
` (10 more replies)
0 siblings, 11 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem; +Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek
[-- Attachment #1: Type: text/plain, Size: 1838 bytes --]
The current Ntuple filter implementation has a limitation on 5750X (P5)
and newer chips. The destination ring of the ntuple filter must be
a valid ring in the RSS indirection table. Ntuple filters may not work
if the RSS indirection table is modified by the user to only contain a
subset of the rings. If an ntuple filter is set to a ring destination
that is not in the RSS indirection table, the packet matching that
filter will be placed in a random ring instead of the specified
destination ring.
This series of patches will fix the problem by using a separate VNIC
for ntuple filters. The default VNIC will be dedicated for RSS and
so the indirection table can be setup in any way and will not affect
ntuple filters using the separate VNIC.
Quite a bit of refactoring is needed to do the the VNIC and RSS
context accounting in the first few patches. This is technically a
bug fix, but I think the changes are too big for -net.
Michael Chan (3):
bnxt_en: Refactor ring reservation functions
bnxt_en: Explicitly specify P5 completion rings to reserve
bnxt_en: Check additional resources in bnxt_check_rings()
Pavan Chebbi (6):
bnxt_en: Improve RSS context reservation infrastructure
bnxt_en: Refactor bnxt_set_features()
bnxt_en: Define BNXT_VNIC_DEFAULT for the default vnic index
bnxt_en: Provision for an additional VNIC for ntuple filters
bnxt_en: Create and setup the additional VNIC for adding ntuple
filters
bnxt_en: Use the new VNIC to create ntuple filters
Venkat Duvvuru (1):
bnxt_en: Add bnxt_get_total_vnics() to calculate number of VNICs
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 487 +++++++++++-------
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 20 +
.../net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 4 +-
3 files changed, 311 insertions(+), 200 deletions(-)
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply [flat|nested] 12+ messages in thread
* [PATCH net-next 01/10] bnxt_en: Refactor ring reservation functions
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
@ 2024-02-20 23:03 ` Michael Chan
2024-02-20 23:03 ` [PATCH net-next 02/10] bnxt_en: Explicitly specify P5 completion rings to reserve Michael Chan
` (9 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem
Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek,
Ajit Khaparde
[-- Attachment #1: Type: text/plain, Size: 18114 bytes --]
The current functions to reserve hardware rings pass in 6 different ring
or resource types as parameters. Add a structure bnxt_hw_rings to
consolidate all these parameters and pass the structure pointer instead
to these functions. Add 2 related helper functions also. This makes
the code cleaner and makes it easier to add new resources to be
reserved.
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 253 +++++++++++-----------
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 9 +
2 files changed, 132 insertions(+), 130 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 6f415425dc14..4a30eff0791b 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -7098,8 +7098,7 @@ int __bnxt_hwrm_get_tx_rings(struct bnxt *bp, u16 fid, int *tx_rings)
static bool bnxt_rfs_supported(struct bnxt *bp);
static struct hwrm_func_cfg_input *
-__bnxt_hwrm_reserve_pf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
- int ring_grps, int cp_rings, int stats, int vnics)
+__bnxt_hwrm_reserve_pf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
struct hwrm_func_cfg_input *req;
u32 enables = 0;
@@ -7108,52 +7107,51 @@ __bnxt_hwrm_reserve_pf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
return NULL;
req->fid = cpu_to_le16(0xffff);
- enables |= tx_rings ? FUNC_CFG_REQ_ENABLES_NUM_TX_RINGS : 0;
- req->num_tx_rings = cpu_to_le16(tx_rings);
+ enables |= hwr->tx ? FUNC_CFG_REQ_ENABLES_NUM_TX_RINGS : 0;
+ req->num_tx_rings = cpu_to_le16(hwr->tx);
if (BNXT_NEW_RM(bp)) {
- enables |= rx_rings ? FUNC_CFG_REQ_ENABLES_NUM_RX_RINGS : 0;
- enables |= stats ? FUNC_CFG_REQ_ENABLES_NUM_STAT_CTXS : 0;
+ enables |= hwr->rx ? FUNC_CFG_REQ_ENABLES_NUM_RX_RINGS : 0;
+ enables |= hwr->stat ? FUNC_CFG_REQ_ENABLES_NUM_STAT_CTXS : 0;
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
- enables |= cp_rings ? FUNC_CFG_REQ_ENABLES_NUM_MSIX : 0;
- enables |= tx_rings + ring_grps ?
+ enables |= hwr->cp ? FUNC_CFG_REQ_ENABLES_NUM_MSIX : 0;
+ enables |= hwr->tx + hwr->grp ?
FUNC_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
- enables |= rx_rings ?
- FUNC_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
+ enables |= hwr->rx ?
+ FUNC_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
} else {
- enables |= cp_rings ?
+ enables |= hwr->cp ?
FUNC_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
- enables |= ring_grps ?
+ enables |= hwr->grp ?
FUNC_CFG_REQ_ENABLES_NUM_HW_RING_GRPS |
FUNC_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
}
- enables |= vnics ? FUNC_CFG_REQ_ENABLES_NUM_VNICS : 0;
+ enables |= hwr->vnic ? FUNC_CFG_REQ_ENABLES_NUM_VNICS : 0;
- req->num_rx_rings = cpu_to_le16(rx_rings);
+ req->num_rx_rings = cpu_to_le16(hwr->rx);
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
- u16 rss_ctx = bnxt_get_nr_rss_ctxs(bp, ring_grps);
+ u16 rss_ctx = bnxt_get_nr_rss_ctxs(bp, hwr->grp);
- req->num_cmpl_rings = cpu_to_le16(tx_rings + ring_grps);
- req->num_msix = cpu_to_le16(cp_rings);
+ req->num_cmpl_rings = cpu_to_le16(hwr->tx + hwr->grp);
+ req->num_msix = cpu_to_le16(hwr->cp);
req->num_rsscos_ctxs = cpu_to_le16(rss_ctx);
} else {
- req->num_cmpl_rings = cpu_to_le16(cp_rings);
- req->num_hw_ring_grps = cpu_to_le16(ring_grps);
+ req->num_cmpl_rings = cpu_to_le16(hwr->cp);
+ req->num_hw_ring_grps = cpu_to_le16(hwr->grp);
req->num_rsscos_ctxs = cpu_to_le16(1);
if (!(bp->rss_cap & BNXT_RSS_CAP_NEW_RSS_CAP) &&
bnxt_rfs_supported(bp))
req->num_rsscos_ctxs =
- cpu_to_le16(ring_grps + 1);
+ cpu_to_le16(hwr->grp + 1);
}
- req->num_stat_ctxs = cpu_to_le16(stats);
- req->num_vnics = cpu_to_le16(vnics);
+ req->num_stat_ctxs = cpu_to_le16(hwr->stat);
+ req->num_vnics = cpu_to_le16(hwr->vnic);
}
req->enables = cpu_to_le32(enables);
return req;
}
static struct hwrm_func_vf_cfg_input *
-__bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
- int ring_grps, int cp_rings, int stats, int vnics)
+__bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
struct hwrm_func_vf_cfg_input *req;
u32 enables = 0;
@@ -7161,51 +7159,48 @@ __bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
if (hwrm_req_init(bp, req, HWRM_FUNC_VF_CFG))
return NULL;
- enables |= tx_rings ? FUNC_VF_CFG_REQ_ENABLES_NUM_TX_RINGS : 0;
- enables |= rx_rings ? FUNC_VF_CFG_REQ_ENABLES_NUM_RX_RINGS |
- FUNC_VF_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
- enables |= stats ? FUNC_VF_CFG_REQ_ENABLES_NUM_STAT_CTXS : 0;
+ enables |= hwr->tx ? FUNC_VF_CFG_REQ_ENABLES_NUM_TX_RINGS : 0;
+ enables |= hwr->rx ? FUNC_VF_CFG_REQ_ENABLES_NUM_RX_RINGS |
+ FUNC_VF_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
+ enables |= hwr->stat ? FUNC_VF_CFG_REQ_ENABLES_NUM_STAT_CTXS : 0;
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
- enables |= tx_rings + ring_grps ?
+ enables |= hwr->tx + hwr->grp ?
FUNC_VF_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
} else {
- enables |= cp_rings ?
- FUNC_VF_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
- enables |= ring_grps ?
+ enables |= hwr->cp ? FUNC_VF_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
+ enables |= hwr->grp ?
FUNC_VF_CFG_REQ_ENABLES_NUM_HW_RING_GRPS : 0;
}
- enables |= vnics ? FUNC_VF_CFG_REQ_ENABLES_NUM_VNICS : 0;
+ enables |= hwr->vnic ? FUNC_VF_CFG_REQ_ENABLES_NUM_VNICS : 0;
enables |= FUNC_VF_CFG_REQ_ENABLES_NUM_L2_CTXS;
req->num_l2_ctxs = cpu_to_le16(BNXT_VF_MAX_L2_CTX);
- req->num_tx_rings = cpu_to_le16(tx_rings);
- req->num_rx_rings = cpu_to_le16(rx_rings);
+ req->num_tx_rings = cpu_to_le16(hwr->tx);
+ req->num_rx_rings = cpu_to_le16(hwr->rx);
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
- u16 rss_ctx = bnxt_get_nr_rss_ctxs(bp, ring_grps);
+ u16 rss_ctx = bnxt_get_nr_rss_ctxs(bp, hwr->grp);
- req->num_cmpl_rings = cpu_to_le16(tx_rings + ring_grps);
+ req->num_cmpl_rings = cpu_to_le16(hwr->tx + hwr->grp);
req->num_rsscos_ctxs = cpu_to_le16(rss_ctx);
} else {
- req->num_cmpl_rings = cpu_to_le16(cp_rings);
- req->num_hw_ring_grps = cpu_to_le16(ring_grps);
+ req->num_cmpl_rings = cpu_to_le16(hwr->cp);
+ req->num_hw_ring_grps = cpu_to_le16(hwr->grp);
req->num_rsscos_ctxs = cpu_to_le16(BNXT_VF_MAX_RSS_CTX);
}
- req->num_stat_ctxs = cpu_to_le16(stats);
- req->num_vnics = cpu_to_le16(vnics);
+ req->num_stat_ctxs = cpu_to_le16(hwr->stat);
+ req->num_vnics = cpu_to_le16(hwr->vnic);
req->enables = cpu_to_le32(enables);
return req;
}
static int
-bnxt_hwrm_reserve_pf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
- int ring_grps, int cp_rings, int stats, int vnics)
+bnxt_hwrm_reserve_pf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
struct hwrm_func_cfg_input *req;
int rc;
- req = __bnxt_hwrm_reserve_pf_rings(bp, tx_rings, rx_rings, ring_grps,
- cp_rings, stats, vnics);
+ req = __bnxt_hwrm_reserve_pf_rings(bp, hwr);
if (!req)
return -ENOMEM;
@@ -7219,25 +7214,23 @@ bnxt_hwrm_reserve_pf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
return rc;
if (bp->hwrm_spec_code < 0x10601)
- bp->hw_resc.resv_tx_rings = tx_rings;
+ bp->hw_resc.resv_tx_rings = hwr->tx;
return bnxt_hwrm_get_rings(bp);
}
static int
-bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
- int ring_grps, int cp_rings, int stats, int vnics)
+bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
struct hwrm_func_vf_cfg_input *req;
int rc;
if (!BNXT_NEW_RM(bp)) {
- bp->hw_resc.resv_tx_rings = tx_rings;
+ bp->hw_resc.resv_tx_rings = hwr->tx;
return 0;
}
- req = __bnxt_hwrm_reserve_vf_rings(bp, tx_rings, rx_rings, ring_grps,
- cp_rings, stats, vnics);
+ req = __bnxt_hwrm_reserve_vf_rings(bp, hwr);
if (!req)
return -ENOMEM;
@@ -7248,15 +7241,12 @@ bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
return bnxt_hwrm_get_rings(bp);
}
-static int bnxt_hwrm_reserve_rings(struct bnxt *bp, int tx, int rx, int grp,
- int cp, int stat, int vnic)
+static int bnxt_hwrm_reserve_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
if (BNXT_PF(bp))
- return bnxt_hwrm_reserve_pf_rings(bp, tx, rx, grp, cp, stat,
- vnic);
+ return bnxt_hwrm_reserve_pf_rings(bp, hwr);
else
- return bnxt_hwrm_reserve_vf_rings(bp, tx, rx, grp, cp, stat,
- vnic);
+ return bnxt_hwrm_reserve_vf_rings(bp, hwr);
}
int bnxt_nq_rings_in_use(struct bnxt *bp)
@@ -7352,47 +7342,60 @@ static bool bnxt_need_reserve_rings(struct bnxt *bp)
return false;
}
-static int __bnxt_reserve_rings(struct bnxt *bp)
+static void bnxt_copy_reserved_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
struct bnxt_hw_resc *hw_resc = &bp->hw_resc;
- int cp = bnxt_nq_rings_in_use(bp);
- int tx = bp->tx_nr_rings;
- int rx = bp->rx_nr_rings;
- int grp, rx_rings, rc;
- int vnic = 1, stat;
+
+ hwr->tx = hw_resc->resv_tx_rings;
+ if (BNXT_NEW_RM(bp)) {
+ hwr->rx = hw_resc->resv_rx_rings;
+ hwr->cp = hw_resc->resv_irqs;
+ hwr->grp = hw_resc->resv_hw_ring_grps;
+ hwr->vnic = hw_resc->resv_vnics;
+ hwr->stat = hw_resc->resv_stat_ctxs;
+ }
+}
+
+static bool bnxt_rings_ok(struct bnxt *bp, struct bnxt_hw_rings *hwr)
+{
+ return hwr->tx && hwr->rx && hwr->cp && hwr->grp && hwr->vnic &&
+ hwr->stat;
+}
+
+static int __bnxt_reserve_rings(struct bnxt *bp)
+{
+ struct bnxt_hw_rings hwr = {0};
+ int rx_rings, rc;
bool sh = false;
int tx_cp;
if (!bnxt_need_reserve_rings(bp))
return 0;
+ hwr.cp = bnxt_nq_rings_in_use(bp);
+ hwr.tx = bp->tx_nr_rings;
+ hwr.rx = bp->rx_nr_rings;
+ hwr.vnic = 1;
if (bp->flags & BNXT_FLAG_SHARED_RINGS)
sh = true;
if ((bp->flags & BNXT_FLAG_RFS) &&
!(bp->flags & BNXT_FLAG_CHIP_P5_PLUS))
- vnic = rx + 1;
+ hwr.vnic = hwr.rx + 1;
if (bp->flags & BNXT_FLAG_AGG_RINGS)
- rx <<= 1;
- grp = bp->rx_nr_rings;
- stat = bnxt_get_func_stat_ctxs(bp);
+ hwr.rx <<= 1;
+ hwr.grp = bp->rx_nr_rings;
+ hwr.stat = bnxt_get_func_stat_ctxs(bp);
- rc = bnxt_hwrm_reserve_rings(bp, tx, rx, grp, cp, stat, vnic);
+ rc = bnxt_hwrm_reserve_rings(bp, &hwr);
if (rc)
return rc;
- tx = hw_resc->resv_tx_rings;
- if (BNXT_NEW_RM(bp)) {
- rx = hw_resc->resv_rx_rings;
- cp = hw_resc->resv_irqs;
- grp = hw_resc->resv_hw_ring_grps;
- vnic = hw_resc->resv_vnics;
- stat = hw_resc->resv_stat_ctxs;
- }
+ bnxt_copy_reserved_rings(bp, &hwr);
- rx_rings = rx;
+ rx_rings = hwr.rx;
if (bp->flags & BNXT_FLAG_AGG_RINGS) {
- if (rx >= 2) {
- rx_rings = rx >> 1;
+ if (hwr.rx >= 2) {
+ rx_rings = hwr.rx >> 1;
} else {
if (netif_running(bp->dev))
return -ENOMEM;
@@ -7404,17 +7407,17 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
bnxt_set_ring_params(bp);
}
}
- rx_rings = min_t(int, rx_rings, grp);
- cp = min_t(int, cp, bp->cp_nr_rings);
- if (stat > bnxt_get_ulp_stat_ctxs(bp))
- stat -= bnxt_get_ulp_stat_ctxs(bp);
- cp = min_t(int, cp, stat);
- rc = bnxt_trim_rings(bp, &rx_rings, &tx, cp, sh);
+ rx_rings = min_t(int, rx_rings, hwr.grp);
+ hwr.cp = min_t(int, hwr.cp, bp->cp_nr_rings);
+ if (hwr.stat > bnxt_get_ulp_stat_ctxs(bp))
+ hwr.stat -= bnxt_get_ulp_stat_ctxs(bp);
+ hwr.cp = min_t(int, hwr.cp, hwr.stat);
+ rc = bnxt_trim_rings(bp, &rx_rings, &hwr.tx, hwr.cp, sh);
if (bp->flags & BNXT_FLAG_AGG_RINGS)
- rx = rx_rings << 1;
- tx_cp = bnxt_num_tx_to_cp(bp, tx);
- cp = sh ? max_t(int, tx_cp, rx_rings) : tx_cp + rx_rings;
- bp->tx_nr_rings = tx;
+ hwr.rx = rx_rings << 1;
+ tx_cp = bnxt_num_tx_to_cp(bp, hwr.tx);
+ hwr.cp = sh ? max_t(int, tx_cp, rx_rings) : tx_cp + rx_rings;
+ bp->tx_nr_rings = hwr.tx;
/* If we cannot reserve all the RX rings, reset the RSS map only
* if absolutely necessary
@@ -7431,9 +7434,9 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
}
}
bp->rx_nr_rings = rx_rings;
- bp->cp_nr_rings = cp;
+ bp->cp_nr_rings = hwr.cp;
- if (!tx || !rx || !cp || !grp || !vnic || !stat)
+ if (!bnxt_rings_ok(bp, &hwr))
return -ENOMEM;
if (!netif_is_rxfh_configured(bp->dev))
@@ -7442,9 +7445,7 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
return rc;
}
-static int bnxt_hwrm_check_vf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
- int ring_grps, int cp_rings, int stats,
- int vnics)
+static int bnxt_hwrm_check_vf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
struct hwrm_func_vf_cfg_input *req;
u32 flags;
@@ -7452,8 +7453,7 @@ static int bnxt_hwrm_check_vf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
if (!BNXT_NEW_RM(bp))
return 0;
- req = __bnxt_hwrm_reserve_vf_rings(bp, tx_rings, rx_rings, ring_grps,
- cp_rings, stats, vnics);
+ req = __bnxt_hwrm_reserve_vf_rings(bp, hwr);
flags = FUNC_VF_CFG_REQ_FLAGS_TX_ASSETS_TEST |
FUNC_VF_CFG_REQ_FLAGS_RX_ASSETS_TEST |
FUNC_VF_CFG_REQ_FLAGS_CMPL_ASSETS_TEST |
@@ -7467,15 +7467,12 @@ static int bnxt_hwrm_check_vf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
return hwrm_req_send_silent(bp, req);
}
-static int bnxt_hwrm_check_pf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
- int ring_grps, int cp_rings, int stats,
- int vnics)
+static int bnxt_hwrm_check_pf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
struct hwrm_func_cfg_input *req;
u32 flags;
- req = __bnxt_hwrm_reserve_pf_rings(bp, tx_rings, rx_rings, ring_grps,
- cp_rings, stats, vnics);
+ req = __bnxt_hwrm_reserve_pf_rings(bp, hwr);
flags = FUNC_CFG_REQ_FLAGS_TX_ASSETS_TEST;
if (BNXT_NEW_RM(bp)) {
flags |= FUNC_CFG_REQ_FLAGS_RX_ASSETS_TEST |
@@ -7493,20 +7490,15 @@ static int bnxt_hwrm_check_pf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
return hwrm_req_send_silent(bp, req);
}
-static int bnxt_hwrm_check_rings(struct bnxt *bp, int tx_rings, int rx_rings,
- int ring_grps, int cp_rings, int stats,
- int vnics)
+static int bnxt_hwrm_check_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
if (bp->hwrm_spec_code < 0x10801)
return 0;
if (BNXT_PF(bp))
- return bnxt_hwrm_check_pf_rings(bp, tx_rings, rx_rings,
- ring_grps, cp_rings, stats,
- vnics);
+ return bnxt_hwrm_check_pf_rings(bp, hwr);
- return bnxt_hwrm_check_vf_rings(bp, tx_rings, rx_rings, ring_grps,
- cp_rings, stats, vnics);
+ return bnxt_hwrm_check_vf_rings(bp, hwr);
}
static void bnxt_hwrm_coal_params_qcaps(struct bnxt *bp)
@@ -12342,21 +12334,22 @@ static bool bnxt_rfs_supported(struct bnxt *bp)
/* If runtime conditions support RFS */
static bool bnxt_rfs_capable(struct bnxt *bp)
{
- int vnics, max_vnics, max_rss_ctxs;
+ struct bnxt_hw_rings hwr = {0};
+ int max_vnics, max_rss_ctxs;
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
return bnxt_rfs_supported(bp);
if (!(bp->flags & BNXT_FLAG_MSIX_CAP) || !bnxt_can_reserve_rings(bp) || !bp->rx_nr_rings)
return false;
- vnics = 1 + bp->rx_nr_rings;
+ hwr.vnic = 1 + bp->rx_nr_rings;
max_vnics = bnxt_get_max_func_vnics(bp);
max_rss_ctxs = bnxt_get_max_func_rss_ctxs(bp);
/* RSS contexts not a limiting factor */
if (bp->rss_cap & BNXT_RSS_CAP_NEW_RSS_CAP)
max_rss_ctxs = max_vnics;
- if (vnics > max_vnics || vnics > max_rss_ctxs) {
+ if (hwr.vnic > max_vnics || hwr.vnic > max_rss_ctxs) {
if (bp->rx_nr_rings > 1)
netdev_warn(bp->dev,
"Not enough resources to support NTUPLE filters, enough resources for up to %d rx rings\n",
@@ -12367,15 +12360,16 @@ static bool bnxt_rfs_capable(struct bnxt *bp)
if (!BNXT_NEW_RM(bp))
return true;
- if (vnics == bp->hw_resc.resv_vnics)
+ if (hwr.vnic == bp->hw_resc.resv_vnics)
return true;
- bnxt_hwrm_reserve_rings(bp, 0, 0, 0, 0, 0, vnics);
- if (vnics <= bp->hw_resc.resv_vnics)
+ bnxt_hwrm_reserve_rings(bp, &hwr);
+ if (hwr.vnic <= bp->hw_resc.resv_vnics)
return true;
netdev_warn(bp->dev, "Unable to reserve resources to support NTUPLE filters.\n");
- bnxt_hwrm_reserve_rings(bp, 0, 0, 0, 0, 0, 1);
+ hwr.vnic = 1;
+ bnxt_hwrm_reserve_rings(bp, &hwr);
return false;
}
@@ -13299,9 +13293,8 @@ int bnxt_check_rings(struct bnxt *bp, int tx, int rx, bool sh, int tcs,
int tx_xdp)
{
int max_rx, max_tx, max_cp, tx_sets = 1, tx_cp;
- int tx_rings_needed, stats;
+ struct bnxt_hw_rings hwr = {0};
int rx_rings = rx;
- int cp, vnics;
if (tcs)
tx_sets = tcs;
@@ -13314,26 +13307,26 @@ int bnxt_check_rings(struct bnxt *bp, int tx, int rx, bool sh, int tcs,
if (bp->flags & BNXT_FLAG_AGG_RINGS)
rx_rings <<= 1;
- tx_rings_needed = tx * tx_sets + tx_xdp;
- if (max_tx < tx_rings_needed)
+ hwr.rx = rx_rings;
+ hwr.tx = tx * tx_sets + tx_xdp;
+ if (max_tx < hwr.tx)
return -ENOMEM;
- vnics = 1;
+ hwr.vnic = 1;
if ((bp->flags & (BNXT_FLAG_RFS | BNXT_FLAG_CHIP_P5_PLUS)) ==
BNXT_FLAG_RFS)
- vnics += rx;
+ hwr.vnic += rx;
- tx_cp = __bnxt_num_tx_to_cp(bp, tx_rings_needed, tx_sets, tx_xdp);
- cp = sh ? max_t(int, tx_cp, rx) : tx_cp + rx;
- if (max_cp < cp)
+ tx_cp = __bnxt_num_tx_to_cp(bp, hwr.tx, tx_sets, tx_xdp);
+ hwr.cp = sh ? max_t(int, tx_cp, rx) : tx_cp + rx;
+ if (max_cp < hwr.cp)
return -ENOMEM;
- stats = cp;
+ hwr.stat = hwr.cp;
if (BNXT_NEW_RM(bp)) {
- cp += bnxt_get_ulp_msix_num(bp);
- stats += bnxt_get_ulp_stat_ctxs(bp);
+ hwr.cp += bnxt_get_ulp_msix_num(bp);
+ hwr.stat += bnxt_get_ulp_stat_ctxs(bp);
}
- return bnxt_hwrm_check_rings(bp, tx_rings_needed, rx_rings, rx, cp,
- stats, vnics);
+ return bnxt_hwrm_check_rings(bp, &hwr);
}
static void bnxt_unmap_bars(struct bnxt *bp, struct pci_dev *pdev)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index 60bdd0673ec8..bb3b31f8d02f 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -1254,6 +1254,15 @@ struct bnxt_vnic_info {
#define BNXT_VNIC_RFS_NEW_RSS_FLAG 0x10
};
+struct bnxt_hw_rings {
+ int tx;
+ int rx;
+ int grp;
+ int cp;
+ int stat;
+ int vnic;
+};
+
struct bnxt_hw_resc {
u16 min_rsscos_ctxs;
u16 max_rsscos_ctxs;
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH net-next 02/10] bnxt_en: Explicitly specify P5 completion rings to reserve
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
2024-02-20 23:03 ` [PATCH net-next 01/10] bnxt_en: Refactor ring reservation functions Michael Chan
@ 2024-02-20 23:03 ` Michael Chan
2024-02-20 23:03 ` [PATCH net-next 03/10] bnxt_en: Improve RSS context reservation infrastructure Michael Chan
` (8 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem
Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek,
Ajit Khaparde
[-- Attachment #1: Type: text/plain, Size: 4433 bytes --]
The current code assumes that every RX ring group and every TX ring
requires a completion ring on P5_PLUS chips. Now that we have the
bnxt_hw_rings structure, add the cp_p5 field so that it can
be explicitly specified. This makes the logic more clear.
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 19 ++++++++++++-------
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 1 +
2 files changed, 13 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 4a30eff0791b..4f3d2b1c9989 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -7114,7 +7114,7 @@ __bnxt_hwrm_reserve_pf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
enables |= hwr->stat ? FUNC_CFG_REQ_ENABLES_NUM_STAT_CTXS : 0;
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
enables |= hwr->cp ? FUNC_CFG_REQ_ENABLES_NUM_MSIX : 0;
- enables |= hwr->tx + hwr->grp ?
+ enables |= hwr->cp_p5 ?
FUNC_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
enables |= hwr->rx ?
FUNC_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
@@ -7131,7 +7131,7 @@ __bnxt_hwrm_reserve_pf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
u16 rss_ctx = bnxt_get_nr_rss_ctxs(bp, hwr->grp);
- req->num_cmpl_rings = cpu_to_le16(hwr->tx + hwr->grp);
+ req->num_cmpl_rings = cpu_to_le16(hwr->cp_p5);
req->num_msix = cpu_to_le16(hwr->cp);
req->num_rsscos_ctxs = cpu_to_le16(rss_ctx);
} else {
@@ -7164,7 +7164,7 @@ __bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
FUNC_VF_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
enables |= hwr->stat ? FUNC_VF_CFG_REQ_ENABLES_NUM_STAT_CTXS : 0;
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
- enables |= hwr->tx + hwr->grp ?
+ enables |= hwr->cp_p5 ?
FUNC_VF_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
} else {
enables |= hwr->cp ? FUNC_VF_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
@@ -7180,7 +7180,7 @@ __bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
u16 rss_ctx = bnxt_get_nr_rss_ctxs(bp, hwr->grp);
- req->num_cmpl_rings = cpu_to_le16(hwr->tx + hwr->grp);
+ req->num_cmpl_rings = cpu_to_le16(hwr->cp_p5);
req->num_rsscos_ctxs = cpu_to_le16(rss_ctx);
} else {
req->num_cmpl_rings = cpu_to_le16(hwr->cp);
@@ -7350,6 +7350,8 @@ static void bnxt_copy_reserved_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
if (BNXT_NEW_RM(bp)) {
hwr->rx = hw_resc->resv_rx_rings;
hwr->cp = hw_resc->resv_irqs;
+ if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
+ hwr->cp_p5 = hw_resc->resv_cp_rings;
hwr->grp = hw_resc->resv_hw_ring_grps;
hwr->vnic = hw_resc->resv_vnics;
hwr->stat = hw_resc->resv_stat_ctxs;
@@ -7359,7 +7361,7 @@ static void bnxt_copy_reserved_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
static bool bnxt_rings_ok(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
return hwr->tx && hwr->rx && hwr->cp && hwr->grp && hwr->vnic &&
- hwr->stat;
+ hwr->stat && (hwr->cp_p5 || !(bp->flags & BNXT_FLAG_CHIP_P5_PLUS));
}
static int __bnxt_reserve_rings(struct bnxt *bp)
@@ -7378,8 +7380,9 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
hwr.vnic = 1;
if (bp->flags & BNXT_FLAG_SHARED_RINGS)
sh = true;
- if ((bp->flags & BNXT_FLAG_RFS) &&
- !(bp->flags & BNXT_FLAG_CHIP_P5_PLUS))
+ if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
+ hwr.cp_p5 = hwr.rx + hwr.tx;
+ else if (bp->flags & BNXT_FLAG_RFS)
hwr.vnic = hwr.rx + 1;
if (bp->flags & BNXT_FLAG_AGG_RINGS)
hwr.rx <<= 1;
@@ -13326,6 +13329,8 @@ int bnxt_check_rings(struct bnxt *bp, int tx, int rx, bool sh, int tcs,
hwr.cp += bnxt_get_ulp_msix_num(bp);
hwr.stat += bnxt_get_ulp_stat_ctxs(bp);
}
+ if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
+ hwr.cp_p5 = hwr.tx + rx;
return bnxt_hwrm_check_rings(bp, &hwr);
}
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index bb3b31f8d02f..681a579afcc2 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -1259,6 +1259,7 @@ struct bnxt_hw_rings {
int rx;
int grp;
int cp;
+ int cp_p5;
int stat;
int vnic;
};
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH net-next 03/10] bnxt_en: Improve RSS context reservation infrastructure
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
2024-02-20 23:03 ` [PATCH net-next 01/10] bnxt_en: Refactor ring reservation functions Michael Chan
2024-02-20 23:03 ` [PATCH net-next 02/10] bnxt_en: Explicitly specify P5 completion rings to reserve Michael Chan
@ 2024-02-20 23:03 ` Michael Chan
2024-02-20 23:03 ` [PATCH net-next 04/10] bnxt_en: Check additional resources in bnxt_check_rings() Michael Chan
` (7 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem; +Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek
[-- Attachment #1: Type: text/plain, Size: 7635 bytes --]
From: Pavan Chebbi <pavan.chebbi@broadcom.com>
Add RSS context fields to struct bnxt_hw_rings and struct bnxt_hw_resc.
With these, we can now specific the exact number of RSS contexts to
reserve and store the reserved value. The original code relies on
other resources to infer the number of RSS contexts to reserve and the
reserved value is not stored. This improved infrastructure will make
the RSS context accounting more complete and is needed by later
patches.
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 57 ++++++++++++++---------
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 2 +
2 files changed, 36 insertions(+), 23 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 4f3d2b1c9989..e6fad2a8b817 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -7043,6 +7043,7 @@ static int bnxt_hwrm_get_rings(struct bnxt *bp)
hw_resc->resv_hw_ring_grps =
le32_to_cpu(resp->alloc_hw_ring_grps);
hw_resc->resv_vnics = le16_to_cpu(resp->alloc_vnics);
+ hw_resc->resv_rsscos_ctxs = le16_to_cpu(resp->alloc_rsscos_ctx);
cp = le16_to_cpu(resp->alloc_cmpl_rings);
stats = le16_to_cpu(resp->alloc_stat_ctx);
hw_resc->resv_irqs = cp;
@@ -7116,32 +7117,23 @@ __bnxt_hwrm_reserve_pf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
enables |= hwr->cp ? FUNC_CFG_REQ_ENABLES_NUM_MSIX : 0;
enables |= hwr->cp_p5 ?
FUNC_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
- enables |= hwr->rx ?
- FUNC_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
} else {
enables |= hwr->cp ?
FUNC_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
enables |= hwr->grp ?
- FUNC_CFG_REQ_ENABLES_NUM_HW_RING_GRPS |
- FUNC_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
+ FUNC_CFG_REQ_ENABLES_NUM_HW_RING_GRPS : 0;
}
enables |= hwr->vnic ? FUNC_CFG_REQ_ENABLES_NUM_VNICS : 0;
-
+ enables |= hwr->rss_ctx ? FUNC_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS :
+ 0;
req->num_rx_rings = cpu_to_le16(hwr->rx);
+ req->num_rsscos_ctxs = cpu_to_le16(hwr->rss_ctx);
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
- u16 rss_ctx = bnxt_get_nr_rss_ctxs(bp, hwr->grp);
-
req->num_cmpl_rings = cpu_to_le16(hwr->cp_p5);
req->num_msix = cpu_to_le16(hwr->cp);
- req->num_rsscos_ctxs = cpu_to_le16(rss_ctx);
} else {
req->num_cmpl_rings = cpu_to_le16(hwr->cp);
req->num_hw_ring_grps = cpu_to_le16(hwr->grp);
- req->num_rsscos_ctxs = cpu_to_le16(1);
- if (!(bp->rss_cap & BNXT_RSS_CAP_NEW_RSS_CAP) &&
- bnxt_rfs_supported(bp))
- req->num_rsscos_ctxs =
- cpu_to_le16(hwr->grp + 1);
}
req->num_stat_ctxs = cpu_to_le16(hwr->stat);
req->num_vnics = cpu_to_le16(hwr->vnic);
@@ -7163,6 +7155,7 @@ __bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
enables |= hwr->rx ? FUNC_VF_CFG_REQ_ENABLES_NUM_RX_RINGS |
FUNC_VF_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
enables |= hwr->stat ? FUNC_VF_CFG_REQ_ENABLES_NUM_STAT_CTXS : 0;
+ enables |= hwr->rss_ctx ? FUNC_VF_CFG_REQ_ENABLES_NUM_RSSCOS_CTXS : 0;
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
enables |= hwr->cp_p5 ?
FUNC_VF_CFG_REQ_ENABLES_NUM_CMPL_RINGS : 0;
@@ -7177,15 +7170,12 @@ __bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
req->num_l2_ctxs = cpu_to_le16(BNXT_VF_MAX_L2_CTX);
req->num_tx_rings = cpu_to_le16(hwr->tx);
req->num_rx_rings = cpu_to_le16(hwr->rx);
+ req->num_rsscos_ctxs = cpu_to_le16(hwr->rss_ctx);
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
- u16 rss_ctx = bnxt_get_nr_rss_ctxs(bp, hwr->grp);
-
req->num_cmpl_rings = cpu_to_le16(hwr->cp_p5);
- req->num_rsscos_ctxs = cpu_to_le16(rss_ctx);
} else {
req->num_cmpl_rings = cpu_to_le16(hwr->cp);
req->num_hw_ring_grps = cpu_to_le16(hwr->grp);
- req->num_rsscos_ctxs = cpu_to_le16(BNXT_VF_MAX_RSS_CTX);
}
req->num_stat_ctxs = cpu_to_le16(hwr->stat);
req->num_vnics = cpu_to_le16(hwr->vnic);
@@ -7289,6 +7279,20 @@ static int bnxt_get_func_stat_ctxs(struct bnxt *bp)
return cp + ulp_stat;
}
+static int bnxt_get_total_rss_ctxs(struct bnxt *bp, struct bnxt_hw_rings *hwr)
+{
+ if (!hwr->grp)
+ return 0;
+ if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
+ return bnxt_get_nr_rss_ctxs(bp, hwr->grp);
+
+ if (BNXT_VF(bp))
+ return BNXT_VF_MAX_RSS_CTX;
+ if (!(bp->rss_cap & BNXT_RSS_CAP_NEW_RSS_CAP) && bnxt_rfs_supported(bp))
+ return hwr->grp + 1;
+ return 1;
+}
+
/* Check if a default RSS map needs to be setup. This function is only
* used on older firmware that does not require reserving RX rings.
*/
@@ -7355,6 +7359,7 @@ static void bnxt_copy_reserved_rings(struct bnxt *bp, struct bnxt_hw_rings *hwr)
hwr->grp = hw_resc->resv_hw_ring_grps;
hwr->vnic = hw_resc->resv_vnics;
hwr->stat = hw_resc->resv_stat_ctxs;
+ hwr->rss_ctx = hw_resc->resv_rsscos_ctxs;
}
}
@@ -7387,6 +7392,7 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
if (bp->flags & BNXT_FLAG_AGG_RINGS)
hwr.rx <<= 1;
hwr.grp = bp->rx_nr_rings;
+ hwr.rss_ctx = bnxt_get_total_rss_ctxs(bp, &hwr);
hwr.stat = bnxt_get_func_stat_ctxs(bp);
rc = bnxt_hwrm_reserve_rings(bp, &hwr);
@@ -11210,6 +11216,7 @@ static void bnxt_clear_reservations(struct bnxt *bp, bool fw_reset)
hw_resc->resv_rx_rings = 0;
hw_resc->resv_hw_ring_grps = 0;
hw_resc->resv_vnics = 0;
+ hw_resc->resv_rsscos_ctxs = 0;
if (!fw_reset) {
bp->tx_nr_rings = 0;
bp->rx_nr_rings = 0;
@@ -12340,6 +12347,7 @@ static bool bnxt_rfs_capable(struct bnxt *bp)
struct bnxt_hw_rings hwr = {0};
int max_vnics, max_rss_ctxs;
+ hwr.rss_ctx = 1;
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
return bnxt_rfs_supported(bp);
if (!(bp->flags & BNXT_FLAG_MSIX_CAP) || !bnxt_can_reserve_rings(bp) || !bp->rx_nr_rings)
@@ -12349,10 +12357,10 @@ static bool bnxt_rfs_capable(struct bnxt *bp)
max_vnics = bnxt_get_max_func_vnics(bp);
max_rss_ctxs = bnxt_get_max_func_rss_ctxs(bp);
- /* RSS contexts not a limiting factor */
- if (bp->rss_cap & BNXT_RSS_CAP_NEW_RSS_CAP)
- max_rss_ctxs = max_vnics;
- if (hwr.vnic > max_vnics || hwr.vnic > max_rss_ctxs) {
+ if (!(bp->rss_cap & BNXT_RSS_CAP_NEW_RSS_CAP))
+ hwr.rss_ctx = hwr.vnic;
+
+ if (hwr.vnic > max_vnics || hwr.rss_ctx > max_rss_ctxs) {
if (bp->rx_nr_rings > 1)
netdev_warn(bp->dev,
"Not enough resources to support NTUPLE filters, enough resources for up to %d rx rings\n",
@@ -12363,15 +12371,18 @@ static bool bnxt_rfs_capable(struct bnxt *bp)
if (!BNXT_NEW_RM(bp))
return true;
- if (hwr.vnic == bp->hw_resc.resv_vnics)
+ if (hwr.vnic == bp->hw_resc.resv_vnics &&
+ hwr.rss_ctx <= bp->hw_resc.resv_rsscos_ctxs)
return true;
bnxt_hwrm_reserve_rings(bp, &hwr);
- if (hwr.vnic <= bp->hw_resc.resv_vnics)
+ if (hwr.vnic <= bp->hw_resc.resv_vnics &&
+ hwr.rss_ctx <= bp->hw_resc.resv_rsscos_ctxs)
return true;
netdev_warn(bp->dev, "Unable to reserve resources to support NTUPLE filters.\n");
hwr.vnic = 1;
+ hwr.rss_ctx = 0;
bnxt_hwrm_reserve_rings(bp, &hwr);
return false;
}
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index 681a579afcc2..4ff090981809 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -1262,11 +1262,13 @@ struct bnxt_hw_rings {
int cp_p5;
int stat;
int vnic;
+ int rss_ctx;
};
struct bnxt_hw_resc {
u16 min_rsscos_ctxs;
u16 max_rsscos_ctxs;
+ u16 resv_rsscos_ctxs;
u16 min_cp_rings;
u16 max_cp_rings;
u16 resv_cp_rings;
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH net-next 04/10] bnxt_en: Check additional resources in bnxt_check_rings()
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
` (2 preceding siblings ...)
2024-02-20 23:03 ` [PATCH net-next 03/10] bnxt_en: Improve RSS context reservation infrastructure Michael Chan
@ 2024-02-20 23:03 ` Michael Chan
2024-02-20 23:03 ` [PATCH net-next 05/10] bnxt_en: Add bnxt_get_total_vnics() to calculate number of VNICs Michael Chan
` (6 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem
Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek,
Somnath Kotur
[-- Attachment #1: Type: text/plain, Size: 1348 bytes --]
bnxt_check_rings() is called to check if we have enough resource
assets to satisfy the new number of ethtool channels. If the asset
test fails, the ethtool operation will fail gracefully. Otherwise
we will proceed and commit to use the new number of channels. If it
fails to allocate any resources, the chip will fail to come up.
For completeness, check all possible resources before committing to
the new settings. Add the missing ring group and RSS context asset
tests in bnxt_check_rings().
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index e6fad2a8b817..71ac165dfb52 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -13339,6 +13339,8 @@ int bnxt_check_rings(struct bnxt *bp, int tx, int rx, bool sh, int tcs,
if (BNXT_NEW_RM(bp)) {
hwr.cp += bnxt_get_ulp_msix_num(bp);
hwr.stat += bnxt_get_ulp_stat_ctxs(bp);
+ hwr.grp = rx;
+ hwr.rss_ctx = bnxt_get_total_rss_ctxs(bp, &hwr);
}
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
hwr.cp_p5 = hwr.tx + rx;
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH net-next 05/10] bnxt_en: Add bnxt_get_total_vnics() to calculate number of VNICs
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
` (3 preceding siblings ...)
2024-02-20 23:03 ` [PATCH net-next 04/10] bnxt_en: Check additional resources in bnxt_check_rings() Michael Chan
@ 2024-02-20 23:03 ` Michael Chan
2024-02-20 23:03 ` [PATCH net-next 06/10] bnxt_en: Refactor bnxt_set_features() Michael Chan
` (5 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem
Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek,
Venkat Duvvuru
[-- Attachment #1: Type: text/plain, Size: 2692 bytes --]
From: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com>
Refactor the code by adding a new function to calculate the number of
required VNICs. This is used in multiple places when reserving or
checking resources.
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 28 ++++++++++++++---------
1 file changed, 17 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 71ac165dfb52..95af3cd7dba0 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -7308,13 +7308,22 @@ static void bnxt_check_rss_tbl_no_rmgr(struct bnxt *bp)
}
}
+static int bnxt_get_total_vnics(struct bnxt *bp, int rx_rings)
+{
+ if (bp->flags & BNXT_FLAG_RFS) {
+ if (!(bp->flags & BNXT_FLAG_CHIP_P5_PLUS))
+ return rx_rings + 1;
+ }
+ return 1;
+}
+
static bool bnxt_need_reserve_rings(struct bnxt *bp)
{
struct bnxt_hw_resc *hw_resc = &bp->hw_resc;
int cp = bnxt_cp_rings_in_use(bp);
int nq = bnxt_nq_rings_in_use(bp);
int rx = bp->rx_nr_rings, stat;
- int vnic = 1, grp = rx;
+ int vnic, grp = rx;
if (hw_resc->resv_tx_rings != bp->tx_nr_rings &&
bp->hwrm_spec_code >= 0x10601)
@@ -7329,9 +7338,9 @@ static bool bnxt_need_reserve_rings(struct bnxt *bp)
bnxt_check_rss_tbl_no_rmgr(bp);
return false;
}
- if ((bp->flags & BNXT_FLAG_RFS) &&
- !(bp->flags & BNXT_FLAG_CHIP_P5_PLUS))
- vnic = rx + 1;
+
+ vnic = bnxt_get_total_vnics(bp, rx);
+
if (bp->flags & BNXT_FLAG_AGG_RINGS)
rx <<= 1;
stat = bnxt_get_func_stat_ctxs(bp);
@@ -7382,13 +7391,13 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
hwr.cp = bnxt_nq_rings_in_use(bp);
hwr.tx = bp->tx_nr_rings;
hwr.rx = bp->rx_nr_rings;
- hwr.vnic = 1;
if (bp->flags & BNXT_FLAG_SHARED_RINGS)
sh = true;
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
hwr.cp_p5 = hwr.rx + hwr.tx;
- else if (bp->flags & BNXT_FLAG_RFS)
- hwr.vnic = hwr.rx + 1;
+
+ hwr.vnic = bnxt_get_total_vnics(bp, hwr.rx);
+
if (bp->flags & BNXT_FLAG_AGG_RINGS)
hwr.rx <<= 1;
hwr.grp = bp->rx_nr_rings;
@@ -13326,10 +13335,7 @@ int bnxt_check_rings(struct bnxt *bp, int tx, int rx, bool sh, int tcs,
if (max_tx < hwr.tx)
return -ENOMEM;
- hwr.vnic = 1;
- if ((bp->flags & (BNXT_FLAG_RFS | BNXT_FLAG_CHIP_P5_PLUS)) ==
- BNXT_FLAG_RFS)
- hwr.vnic += rx;
+ hwr.vnic = bnxt_get_total_vnics(bp, rx);
tx_cp = __bnxt_num_tx_to_cp(bp, hwr.tx, tx_sets, tx_xdp);
hwr.cp = sh ? max_t(int, tx_cp, rx) : tx_cp + rx;
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH net-next 06/10] bnxt_en: Refactor bnxt_set_features()
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
` (4 preceding siblings ...)
2024-02-20 23:03 ` [PATCH net-next 05/10] bnxt_en: Add bnxt_get_total_vnics() to calculate number of VNICs Michael Chan
@ 2024-02-20 23:03 ` Michael Chan
2024-02-20 23:03 ` [PATCH net-next 07/10] bnxt_en: Define BNXT_VNIC_DEFAULT for the default vnic index Michael Chan
` (4 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem
Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek,
Kalesh AP
[-- Attachment #1: Type: text/plain, Size: 1802 bytes --]
From: Pavan Chebbi <pavan.chebbi@broadcom.com>
Refactor bnxt_set_features() function to have a common
function to re-init. We'll need this to reinitialize when
ntuple configuration changes.
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 19 ++++++++++++-------
1 file changed, 12 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 95af3cd7dba0..2fe5262f6c41 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -12431,6 +12431,16 @@ static netdev_features_t bnxt_fix_features(struct net_device *dev,
return features;
}
+static int bnxt_reinit_features(struct bnxt *bp, bool irq_re_init,
+ bool link_re_init, u32 flags, bool update_tpa)
+{
+ bnxt_close_nic(bp, irq_re_init, link_re_init);
+ bp->flags = flags;
+ if (update_tpa)
+ bnxt_set_ring_params(bp);
+ return bnxt_open_nic(bp, irq_re_init, link_re_init);
+}
+
static int bnxt_set_features(struct net_device *dev, netdev_features_t features)
{
struct bnxt *bp = netdev_priv(dev);
@@ -12479,14 +12489,9 @@ static int bnxt_set_features(struct net_device *dev, netdev_features_t features)
return rc;
}
- if (re_init) {
- bnxt_close_nic(bp, false, false);
- bp->flags = flags;
- if (update_tpa)
- bnxt_set_ring_params(bp);
+ if (re_init)
+ return bnxt_reinit_features(bp, false, false, flags, update_tpa);
- return bnxt_open_nic(bp, false, false);
- }
if (update_tpa) {
bp->flags = flags;
rc = bnxt_set_tpa(bp,
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH net-next 07/10] bnxt_en: Define BNXT_VNIC_DEFAULT for the default vnic index
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
` (5 preceding siblings ...)
2024-02-20 23:03 ` [PATCH net-next 06/10] bnxt_en: Refactor bnxt_set_features() Michael Chan
@ 2024-02-20 23:03 ` Michael Chan
2024-02-20 23:03 ` [PATCH net-next 08/10] bnxt_en: Provision for an additional VNIC for ntuple filters Michael Chan
` (3 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem
Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek,
Kalesh AP, Somnath Kotur
[-- Attachment #1: Type: text/plain, Size: 8615 bytes --]
From: Pavan Chebbi <pavan.chebbi@broadcom.com>
Replace hard coded 0 index with more meaningful BNXT_VNIC_DEFAULT.
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 39 ++++++++++---------
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 2 +
.../net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 4 +-
3 files changed, 24 insertions(+), 21 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 2fe5262f6c41..8877043febd2 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -4229,6 +4229,7 @@ static int bnxt_alloc_vnics(struct bnxt *bp)
static void bnxt_init_vnics(struct bnxt *bp)
{
+ struct bnxt_vnic_info *vnic0 = &bp->vnic_info[BNXT_VNIC_DEFAULT];
int i;
for (i = 0; i < bp->nr_vnics; i++) {
@@ -4242,7 +4243,7 @@ static void bnxt_init_vnics(struct bnxt *bp)
vnic->fw_l2_ctx_id = INVALID_HW_RING_ID;
if (bp->vnic_info[i].rss_hash_key) {
- if (!i) {
+ if (i == BNXT_VNIC_DEFAULT) {
u8 *key = (void *)vnic->rss_hash_key;
int k;
@@ -4268,8 +4269,7 @@ static void bnxt_init_vnics(struct bnxt *bp)
bp->toeplitz_prefix |= key[k];
}
} else {
- memcpy(vnic->rss_hash_key,
- bp->vnic_info[0].rss_hash_key,
+ memcpy(vnic->rss_hash_key, vnic0->rss_hash_key,
HW_HASH_KEY_SIZE);
}
}
@@ -5000,6 +5000,7 @@ static void bnxt_free_mem(struct bnxt *bp, bool irq_re_init)
static int bnxt_alloc_mem(struct bnxt *bp, bool irq_re_init)
{
+ struct bnxt_vnic_info *vnic0 = &bp->vnic_info[BNXT_VNIC_DEFAULT];
int i, j, rc, size, arr_size;
void *bnapi;
@@ -5128,8 +5129,8 @@ static int bnxt_alloc_mem(struct bnxt *bp, bool irq_re_init)
if (rc)
goto alloc_mem_err;
- bp->vnic_info[0].flags |= BNXT_VNIC_RSS_FLAG | BNXT_VNIC_MCAST_FLAG |
- BNXT_VNIC_UCAST_FLAG;
+ vnic0->flags |= BNXT_VNIC_RSS_FLAG | BNXT_VNIC_MCAST_FLAG |
+ BNXT_VNIC_UCAST_FLAG;
rc = bnxt_alloc_vnic_attributes(bp);
if (rc)
goto alloc_mem_err;
@@ -6178,7 +6179,7 @@ static int bnxt_hwrm_vnic_set_rss_p5(struct bnxt *bp, u16 vnic_id, bool set_rss)
static void bnxt_hwrm_update_rss_hash_cfg(struct bnxt *bp)
{
- struct bnxt_vnic_info *vnic = &bp->vnic_info[0];
+ struct bnxt_vnic_info *vnic = &bp->vnic_info[BNXT_VNIC_DEFAULT];
struct hwrm_vnic_rss_qcfg_output *resp;
struct hwrm_vnic_rss_qcfg_input *req;
@@ -6282,6 +6283,7 @@ static u32 bnxt_get_roce_vnic_mode(struct bnxt *bp)
int bnxt_hwrm_vnic_cfg(struct bnxt *bp, u16 vnic_id)
{
+ struct bnxt_vnic_info *vnic0 = &bp->vnic_info[BNXT_VNIC_DEFAULT];
struct bnxt_vnic_info *vnic = &bp->vnic_info[vnic_id];
struct hwrm_vnic_cfg_input *req;
unsigned int ring = 0, grp_idx;
@@ -6311,8 +6313,7 @@ int bnxt_hwrm_vnic_cfg(struct bnxt *bp, u16 vnic_id)
req->enables |= cpu_to_le32(VNIC_CFG_REQ_ENABLES_RSS_RULE |
VNIC_CFG_REQ_ENABLES_MRU);
} else if (vnic->flags & BNXT_VNIC_RFS_NEW_RSS_FLAG) {
- req->rss_rule =
- cpu_to_le16(bp->vnic_info[0].fw_rss_cos_lb_ctx[0]);
+ req->rss_rule = cpu_to_le16(vnic0->fw_rss_cos_lb_ctx[0]);
req->enables |= cpu_to_le32(VNIC_CFG_REQ_ENABLES_RSS_RULE |
VNIC_CFG_REQ_ENABLES_MRU);
req->flags |= cpu_to_le32(VNIC_CFG_REQ_FLAGS_RSS_DFLT_CR_MODE);
@@ -6409,7 +6410,7 @@ static int bnxt_hwrm_vnic_alloc(struct bnxt *bp, u16 vnic_id,
vnic_no_ring_grps:
for (i = 0; i < BNXT_MAX_CTX_PER_VNIC; i++)
vnic->fw_rss_cos_lb_ctx[i] = INVALID_HW_RING_ID;
- if (vnic_id == 0)
+ if (vnic_id == BNXT_VNIC_DEFAULT)
req->flags = cpu_to_le32(VNIC_ALLOC_REQ_FLAGS_DEFAULT);
resp = hwrm_req_hold(bp, req);
@@ -9896,7 +9897,7 @@ static bool bnxt_mc_list_updated(struct bnxt *, u32 *);
static int bnxt_init_chip(struct bnxt *bp, bool irq_re_init)
{
- struct bnxt_vnic_info *vnic = &bp->vnic_info[0];
+ struct bnxt_vnic_info *vnic = &bp->vnic_info[BNXT_VNIC_DEFAULT];
int rc = 0;
unsigned int rx_nr_rings = bp->rx_nr_rings;
@@ -9925,7 +9926,7 @@ static int bnxt_init_chip(struct bnxt *bp, bool irq_re_init)
rx_nr_rings--;
/* default vnic 0 */
- rc = bnxt_hwrm_vnic_alloc(bp, 0, 0, rx_nr_rings);
+ rc = bnxt_hwrm_vnic_alloc(bp, BNXT_VNIC_DEFAULT, 0, rx_nr_rings);
if (rc) {
netdev_err(bp->dev, "hwrm vnic alloc failure rc: %x\n", rc);
goto err_out;
@@ -9934,7 +9935,7 @@ static int bnxt_init_chip(struct bnxt *bp, bool irq_re_init)
if (BNXT_VF(bp))
bnxt_hwrm_func_qcfg(bp);
- rc = bnxt_setup_vnic(bp, 0);
+ rc = bnxt_setup_vnic(bp, BNXT_VNIC_DEFAULT);
if (rc)
goto err_out;
if (bp->rss_cap & BNXT_RSS_CAP_RSS_HASH_TYPE_DELTA)
@@ -11594,7 +11595,7 @@ static void bnxt_cfg_one_usr_fltr(struct bnxt *bp, struct bnxt_filter_base *fltr
if (fltr->type == BNXT_FLTR_TYPE_NTUPLE) {
ntp_fltr = container_of(fltr, struct bnxt_ntuple_filter, base);
- l2_fltr = bp->vnic_info[0].l2_filters[0];
+ l2_fltr = bp->vnic_info[BNXT_VNIC_DEFAULT].l2_filters[0];
atomic_inc(&l2_fltr->refcnt);
ntp_fltr->l2_fltr = l2_fltr;
if (bnxt_hwrm_cfa_ntuple_filter_alloc(bp, ntp_fltr)) {
@@ -12148,8 +12149,8 @@ void bnxt_get_ring_err_stats(struct bnxt *bp,
static bool bnxt_mc_list_updated(struct bnxt *bp, u32 *rx_mask)
{
+ struct bnxt_vnic_info *vnic = &bp->vnic_info[BNXT_VNIC_DEFAULT];
struct net_device *dev = bp->dev;
- struct bnxt_vnic_info *vnic = &bp->vnic_info[0];
struct netdev_hw_addr *ha;
u8 *haddr;
int mc_count = 0;
@@ -12183,7 +12184,7 @@ static bool bnxt_mc_list_updated(struct bnxt *bp, u32 *rx_mask)
static bool bnxt_uc_list_updated(struct bnxt *bp)
{
struct net_device *dev = bp->dev;
- struct bnxt_vnic_info *vnic = &bp->vnic_info[0];
+ struct bnxt_vnic_info *vnic = &bp->vnic_info[BNXT_VNIC_DEFAULT];
struct netdev_hw_addr *ha;
int off = 0;
@@ -12210,7 +12211,7 @@ static void bnxt_set_rx_mode(struct net_device *dev)
if (!test_bit(BNXT_STATE_OPEN, &bp->state))
return;
- vnic = &bp->vnic_info[0];
+ vnic = &bp->vnic_info[BNXT_VNIC_DEFAULT];
mask = vnic->rx_mask;
mask &= ~(CFA_L2_SET_RX_MASK_REQ_MASK_PROMISCUOUS |
CFA_L2_SET_RX_MASK_REQ_MASK_MCAST |
@@ -12241,7 +12242,7 @@ static void bnxt_set_rx_mode(struct net_device *dev)
static int bnxt_cfg_rx_mode(struct bnxt *bp)
{
struct net_device *dev = bp->dev;
- struct bnxt_vnic_info *vnic = &bp->vnic_info[0];
+ struct bnxt_vnic_info *vnic = &bp->vnic_info[BNXT_VNIC_DEFAULT];
struct netdev_hw_addr *ha;
int i, off = 0, rc;
bool uc_update;
@@ -14081,7 +14082,7 @@ u32 bnxt_get_ntp_filter_idx(struct bnxt *bp, struct flow_keys *fkeys,
if (skb)
return skb_get_hash_raw(skb) & BNXT_NTP_FLTR_HASH_MASK;
- vnic = &bp->vnic_info[0];
+ vnic = &bp->vnic_info[BNXT_VNIC_DEFAULT];
return bnxt_toeplitz(bp, fkeys, (void *)vnic->rss_hash_key);
}
@@ -14176,7 +14177,7 @@ static int bnxt_rx_flow_steer(struct net_device *dev, const struct sk_buff *skb,
u32 flags;
if (ether_addr_equal(dev->dev_addr, eth->h_dest)) {
- l2_fltr = bp->vnic_info[0].l2_filters[0];
+ l2_fltr = bp->vnic_info[BNXT_VNIC_DEFAULT].l2_filters[0];
atomic_inc(&l2_fltr->refcnt);
} else {
struct bnxt_l2_key key;
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index 4ff090981809..328bbf72acaf 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -1213,6 +1213,8 @@ struct bnxt_ring_grp_info {
u16 cp_fw_ring_id;
};
+#define BNXT_VNIC_DEFAULT 0
+
struct bnxt_vnic_info {
u16 fw_vnic_id; /* returned by Chimp during alloc */
#define BNXT_MAX_CTX_PER_VNIC 8
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
index dfb5944683f8..1d240a27455a 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
@@ -1314,7 +1314,7 @@ static int bnxt_add_ntuple_cls_rule(struct bnxt *bp,
if (!new_fltr)
return -ENOMEM;
- l2_fltr = bp->vnic_info[0].l2_filters[0];
+ l2_fltr = bp->vnic_info[BNXT_VNIC_DEFAULT].l2_filters[0];
atomic_inc(&l2_fltr->refcnt);
new_fltr->l2_fltr = l2_fltr;
fmasks = &new_fltr->fmasks;
@@ -1763,7 +1763,7 @@ static int bnxt_get_rxfh(struct net_device *dev,
if (!bp->vnic_info)
return 0;
- vnic = &bp->vnic_info[0];
+ vnic = &bp->vnic_info[BNXT_VNIC_DEFAULT];
if (rxfh->indir && bp->rss_indir_tbl) {
tbl_size = bnxt_get_rxfh_indir_size(dev);
for (i = 0; i < tbl_size; i++)
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH net-next 08/10] bnxt_en: Provision for an additional VNIC for ntuple filters
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
` (6 preceding siblings ...)
2024-02-20 23:03 ` [PATCH net-next 07/10] bnxt_en: Define BNXT_VNIC_DEFAULT for the default vnic index Michael Chan
@ 2024-02-20 23:03 ` Michael Chan
2024-02-20 23:03 ` [PATCH net-next 09/10] bnxt_en: Create and setup the additional VNIC for adding " Michael Chan
` (2 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem
Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek,
Somnath Kotur, Kalesh AP
[-- Attachment #1: Type: text/plain, Size: 4868 bytes --]
From: Pavan Chebbi <pavan.chebbi@broadcom.com>
On newer chips that support the ring table index method for
ntuple filters, the current scheme of using the same VNIC for
both RSS and ntuple filters will not work in all cases. An
ntuple filter can only be directed to a destination ring if
that destination ring is also in the RSS indirection table.
To support ntuple filters with any arbitratry RSS indirection
table that may only include a subset of the rings, we need to
use a separate VNIC for ntuple filters.
This patch provisions the additional VNIC. The next patch will
allocate additional VNIC from firmware and set it up.
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 33 +++++++++++++++++++----
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 4 +++
2 files changed, 32 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 8877043febd2..0ca9dd397622 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -4211,8 +4211,12 @@ static int bnxt_alloc_vnics(struct bnxt *bp)
int num_vnics = 1;
#ifdef CONFIG_RFS_ACCEL
- if ((bp->flags & (BNXT_FLAG_RFS | BNXT_FLAG_CHIP_P5_PLUS)) == BNXT_FLAG_RFS)
- num_vnics += bp->rx_nr_rings;
+ if (bp->flags & BNXT_FLAG_RFS) {
+ if (BNXT_SUPPORTS_NTUPLE_VNIC(bp))
+ num_vnics++;
+ else if (!(bp->flags & BNXT_FLAG_CHIP_P5_PLUS))
+ num_vnics += bp->rx_nr_rings;
+ }
#endif
if (BNXT_CHIP_TYPE_NITRO_A0(bp))
@@ -7284,9 +7288,13 @@ static int bnxt_get_total_rss_ctxs(struct bnxt *bp, struct bnxt_hw_rings *hwr)
{
if (!hwr->grp)
return 0;
- if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
- return bnxt_get_nr_rss_ctxs(bp, hwr->grp);
+ if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
+ int rss_ctx = bnxt_get_nr_rss_ctxs(bp, hwr->grp);
+ if (BNXT_SUPPORTS_NTUPLE_VNIC(bp))
+ rss_ctx *= hwr->vnic;
+ return rss_ctx;
+ }
if (BNXT_VF(bp))
return BNXT_VF_MAX_RSS_CTX;
if (!(bp->rss_cap & BNXT_RSS_CAP_NEW_RSS_CAP) && bnxt_rfs_supported(bp))
@@ -7312,6 +7320,8 @@ static void bnxt_check_rss_tbl_no_rmgr(struct bnxt *bp)
static int bnxt_get_total_vnics(struct bnxt *bp, int rx_rings)
{
if (bp->flags & BNXT_FLAG_RFS) {
+ if (BNXT_SUPPORTS_NTUPLE_VNIC(bp))
+ return 2;
if (!(bp->flags & BNXT_FLAG_CHIP_P5_PLUS))
return rx_rings + 1;
}
@@ -8962,6 +8972,10 @@ static int bnxt_hwrm_cfa_adv_flow_mgnt_qcaps(struct bnxt *bp)
CFA_ADV_FLOW_MGNT_QCAPS_RESP_FLAGS_RFS_RING_TBL_IDX_V2_SUPPORTED)
bp->fw_cap |= BNXT_FW_CAP_CFA_RFS_RING_TBL_IDX_V2;
+ if (flags &
+ CFA_ADV_FLOW_MGNT_QCAPS_RESP_FLAGS_RFS_RING_TBL_IDX_V3_SUPPORTED)
+ bp->fw_cap |= BNXT_FW_CAP_CFA_RFS_RING_TBL_IDX_V3;
+
if (flags &
CFA_ADV_FLOW_MGNT_QCAPS_RESP_FLAGS_NTUPLE_FLOW_RX_EXT_IP_PROTO_SUPPORTED)
bp->fw_cap |= BNXT_FW_CAP_CFA_NTUPLE_RX_EXT_IP_PROTO;
@@ -12358,16 +12372,25 @@ static bool bnxt_rfs_capable(struct bnxt *bp)
int max_vnics, max_rss_ctxs;
hwr.rss_ctx = 1;
+ if (BNXT_SUPPORTS_NTUPLE_VNIC(bp)) {
+ /* 2 VNICS: default + Ntuple */
+ hwr.vnic = 2;
+ hwr.rss_ctx = bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings) *
+ hwr.vnic;
+ goto check_reserve_vnic;
+ }
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
return bnxt_rfs_supported(bp);
if (!(bp->flags & BNXT_FLAG_MSIX_CAP) || !bnxt_can_reserve_rings(bp) || !bp->rx_nr_rings)
return false;
hwr.vnic = 1 + bp->rx_nr_rings;
+check_reserve_vnic:
max_vnics = bnxt_get_max_func_vnics(bp);
max_rss_ctxs = bnxt_get_max_func_rss_ctxs(bp);
- if (!(bp->rss_cap & BNXT_RSS_CAP_NEW_RSS_CAP))
+ if (!(bp->flags & BNXT_FLAG_CHIP_P5_PLUS) &&
+ !(bp->rss_cap & BNXT_RSS_CAP_NEW_RSS_CAP))
hwr.rss_ctx = hwr.vnic;
if (hwr.vnic > max_vnics || hwr.rss_ctx > max_rss_ctxs) {
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index 328bbf72acaf..e9158407b181 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -2328,12 +2328,16 @@ struct bnxt {
#define BNXT_FW_CAP_BACKING_STORE_V2 BIT_ULL(36)
#define BNXT_FW_CAP_VNIC_TUNNEL_TPA BIT_ULL(37)
#define BNXT_FW_CAP_CFA_NTUPLE_RX_EXT_IP_PROTO BIT_ULL(38)
+ #define BNXT_FW_CAP_CFA_RFS_RING_TBL_IDX_V3 BIT_ULL(39)
u32 fw_dbg_cap;
#define BNXT_NEW_RM(bp) ((bp)->fw_cap & BNXT_FW_CAP_NEW_RM)
#define BNXT_PTP_USE_RTC(bp) (!BNXT_MH(bp) && \
((bp)->fw_cap & BNXT_FW_CAP_PTP_RTC))
+#define BNXT_SUPPORTS_NTUPLE_VNIC(bp) \
+ (BNXT_PF(bp) && ((bp)->fw_cap & BNXT_FW_CAP_CFA_RFS_RING_TBL_IDX_V3))
+
u32 hwrm_spec_code;
u16 hwrm_cmd_seq;
u16 hwrm_cmd_kong_seq;
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH net-next 09/10] bnxt_en: Create and setup the additional VNIC for adding ntuple filters
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
` (7 preceding siblings ...)
2024-02-20 23:03 ` [PATCH net-next 08/10] bnxt_en: Provision for an additional VNIC for ntuple filters Michael Chan
@ 2024-02-20 23:03 ` Michael Chan
2024-02-20 23:03 ` [PATCH net-next 10/10] bnxt_en: Use the new VNIC to create " Michael Chan
2024-02-22 14:40 ` [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements patchwork-bot+netdevbpf
10 siblings, 0 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem
Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek,
Somnath Kotur
[-- Attachment #1: Type: text/plain, Size: 4737 bytes --]
From: Pavan Chebbi <pavan.chebbi@broadcom.com>
Allocate and setup the additional VNIC for ntuple filters if this
new method is supported by the firmware. Even though this VNIC is
only used for ntuple filters with direct ring destinations, we still
setup the RSS hash to be identical to the default VNIC so that each
RX packet will have the correct hash in the RX completion. This
VNIC is always at VNIC index BNXT_VNIC_NTUPLE.
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 44 +++++++++++++++++------
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 2 ++
2 files changed, 36 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 0ca9dd397622..b8ef6c2ba40c 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -5135,6 +5135,10 @@ static int bnxt_alloc_mem(struct bnxt *bp, bool irq_re_init)
vnic0->flags |= BNXT_VNIC_RSS_FLAG | BNXT_VNIC_MCAST_FLAG |
BNXT_VNIC_UCAST_FLAG;
+ if (BNXT_SUPPORTS_NTUPLE_VNIC(bp) && (bp->flags & BNXT_FLAG_RFS))
+ bp->vnic_info[BNXT_VNIC_NTUPLE].flags |=
+ BNXT_VNIC_RSS_FLAG | BNXT_VNIC_NTUPLE_FLAG;
+
rc = bnxt_alloc_vnic_attributes(bp);
if (rc)
goto alloc_mem_err;
@@ -6090,7 +6094,10 @@ static void bnxt_fill_hw_rss_tbl_p5(struct bnxt *bp,
for (i = 0; i < tbl_size; i++) {
u16 ring_id, j;
- j = bp->rss_indir_tbl[i];
+ if (vnic->flags & BNXT_VNIC_NTUPLE_FLAG)
+ j = ethtool_rxfh_indir_default(i, bp->rx_nr_rings);
+ else
+ j = bp->rss_indir_tbl[i];
rxr = &bp->rx_ring[j];
ring_id = rxr->rx_ring_struct.fw_ring_id;
@@ -9844,10 +9851,28 @@ static int bnxt_setup_vnic(struct bnxt *bp, u16 vnic_id)
return __bnxt_setup_vnic(bp, vnic_id);
}
+static int bnxt_alloc_and_setup_vnic(struct bnxt *bp, u16 vnic_id,
+ u16 start_rx_ring_idx, int rx_rings)
+{
+ int rc;
+
+ rc = bnxt_hwrm_vnic_alloc(bp, vnic_id, start_rx_ring_idx, rx_rings);
+ if (rc) {
+ netdev_err(bp->dev, "hwrm vnic %d alloc failure rc: %x\n",
+ vnic_id, rc);
+ return rc;
+ }
+ return bnxt_setup_vnic(bp, vnic_id);
+}
+
static int bnxt_alloc_rfs_vnics(struct bnxt *bp)
{
int i, rc = 0;
+ if (BNXT_SUPPORTS_NTUPLE_VNIC(bp))
+ return bnxt_alloc_and_setup_vnic(bp, BNXT_VNIC_NTUPLE, 0,
+ bp->rx_nr_rings);
+
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
return 0;
@@ -9863,14 +9888,7 @@ static int bnxt_alloc_rfs_vnics(struct bnxt *bp)
vnic->flags |= BNXT_VNIC_RFS_FLAG;
if (bp->rss_cap & BNXT_RSS_CAP_NEW_RSS_CAP)
vnic->flags |= BNXT_VNIC_RFS_NEW_RSS_FLAG;
- rc = bnxt_hwrm_vnic_alloc(bp, vnic_id, ring_id, 1);
- if (rc) {
- netdev_err(bp->dev, "hwrm vnic %d alloc failure rc: %x\n",
- vnic_id, rc);
- break;
- }
- rc = bnxt_setup_vnic(bp, vnic_id);
- if (rc)
+ if (bnxt_alloc_and_setup_vnic(bp, vnic_id, ring_id, 1))
break;
}
return rc;
@@ -12467,12 +12485,12 @@ static int bnxt_reinit_features(struct bnxt *bp, bool irq_re_init,
static int bnxt_set_features(struct net_device *dev, netdev_features_t features)
{
+ bool update_tpa = false, update_ntuple = false;
struct bnxt *bp = netdev_priv(dev);
u32 flags = bp->flags;
u32 changes;
int rc = 0;
bool re_init = false;
- bool update_tpa = false;
flags &= ~BNXT_FLAG_ALL_CONFIG_FEATS;
if (features & NETIF_F_GRO_HW)
@@ -12503,6 +12521,9 @@ static int bnxt_set_features(struct net_device *dev, netdev_features_t features)
if (changes & ~BNXT_FLAG_TPA)
re_init = true;
+ if (changes & BNXT_FLAG_RFS)
+ update_ntuple = true;
+
if (flags != bp->flags) {
u32 old_flags = bp->flags;
@@ -12513,6 +12534,9 @@ static int bnxt_set_features(struct net_device *dev, netdev_features_t features)
return rc;
}
+ if (update_ntuple)
+ return bnxt_reinit_features(bp, true, false, flags, update_tpa);
+
if (re_init)
return bnxt_reinit_features(bp, false, false, flags, update_tpa);
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index e9158407b181..dd849e715c9b 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -1214,6 +1214,7 @@ struct bnxt_ring_grp_info {
};
#define BNXT_VNIC_DEFAULT 0
+#define BNXT_VNIC_NTUPLE 1
struct bnxt_vnic_info {
u16 fw_vnic_id; /* returned by Chimp during alloc */
@@ -1254,6 +1255,7 @@ struct bnxt_vnic_info {
#define BNXT_VNIC_MCAST_FLAG 4
#define BNXT_VNIC_UCAST_FLAG 8
#define BNXT_VNIC_RFS_NEW_RSS_FLAG 0x10
+#define BNXT_VNIC_NTUPLE_FLAG 0x20
};
struct bnxt_hw_rings {
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH net-next 10/10] bnxt_en: Use the new VNIC to create ntuple filters
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
` (8 preceding siblings ...)
2024-02-20 23:03 ` [PATCH net-next 09/10] bnxt_en: Create and setup the additional VNIC for adding " Michael Chan
@ 2024-02-20 23:03 ` Michael Chan
2024-02-22 14:40 ` [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements patchwork-bot+netdevbpf
10 siblings, 0 replies; 12+ messages in thread
From: Michael Chan @ 2024-02-20 23:03 UTC (permalink / raw
To: davem
Cc: netdev, edumazet, kuba, pabeni, pavan.chebbi, andrew.gospodarek,
Somnath Kotur, Kalesh AP
[-- Attachment #1: Type: text/plain, Size: 2844 bytes --]
From: Pavan Chebbi <pavan.chebbi@broadcom.com>
The newly created vnic (BNXT_VNIC_NTUPLE) is ready to be used to create
ntuple filters when supported by firmware. All RX rings can be used
regardless of the RSS indirection setting on the default VNIC.
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 33 ++++++++++++++++++-----
1 file changed, 27 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index b8ef6c2ba40c..9968d67e6c77 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -5785,6 +5785,29 @@ void bnxt_fill_ipv6_mask(__be32 mask[4])
mask[i] = cpu_to_be32(~0);
}
+static void
+bnxt_cfg_rfs_ring_tbl_idx(struct bnxt *bp,
+ struct hwrm_cfa_ntuple_filter_alloc_input *req,
+ u16 rxq)
+{
+ if (BNXT_SUPPORTS_NTUPLE_VNIC(bp)) {
+ struct bnxt_vnic_info *vnic;
+ u32 enables;
+
+ vnic = &bp->vnic_info[BNXT_VNIC_NTUPLE];
+ req->dst_id = cpu_to_le16(vnic->fw_vnic_id);
+ enables = CFA_NTUPLE_FILTER_ALLOC_REQ_ENABLES_RFS_RING_TBL_IDX;
+ req->enables |= cpu_to_le32(enables);
+ req->rfs_ring_tbl_idx = cpu_to_le16(rxq);
+ } else {
+ u32 flags;
+
+ flags = CFA_NTUPLE_FILTER_ALLOC_REQ_FLAGS_DEST_RFS_RING_IDX;
+ req->flags |= cpu_to_le32(flags);
+ req->dst_id = cpu_to_le16(rxq);
+ }
+}
+
int bnxt_hwrm_cfa_ntuple_filter_alloc(struct bnxt *bp,
struct bnxt_ntuple_filter *fltr)
{
@@ -5794,7 +5817,6 @@ int bnxt_hwrm_cfa_ntuple_filter_alloc(struct bnxt *bp,
struct flow_keys *keys = &fltr->fkeys;
struct bnxt_l2_filter *l2_fltr;
struct bnxt_vnic_info *vnic;
- u32 flags = 0;
int rc;
rc = hwrm_req_init(bp, req, HWRM_CFA_NTUPLE_FILTER_ALLOC);
@@ -5805,16 +5827,15 @@ int bnxt_hwrm_cfa_ntuple_filter_alloc(struct bnxt *bp,
req->l2_filter_id = l2_fltr->base.filter_id;
if (fltr->base.flags & BNXT_ACT_DROP) {
- flags = CFA_NTUPLE_FILTER_ALLOC_REQ_FLAGS_DROP;
+ req->flags =
+ cpu_to_le32(CFA_NTUPLE_FILTER_ALLOC_REQ_FLAGS_DROP);
} else if (bp->fw_cap & BNXT_FW_CAP_CFA_RFS_RING_TBL_IDX_V2) {
- flags = CFA_NTUPLE_FILTER_ALLOC_REQ_FLAGS_DEST_RFS_RING_IDX;
- req->dst_id = cpu_to_le16(fltr->base.rxq);
+ bnxt_cfg_rfs_ring_tbl_idx(bp, req, fltr->base.rxq);
} else {
vnic = &bp->vnic_info[fltr->base.rxq + 1];
req->dst_id = cpu_to_le16(vnic->fw_vnic_id);
}
- req->flags = cpu_to_le32(flags);
- req->enables = cpu_to_le32(BNXT_NTP_FLTR_FLAGS);
+ req->enables |= cpu_to_le32(BNXT_NTP_FLTR_FLAGS);
req->ethertype = htons(ETH_P_IP);
req->ip_addr_type = CFA_NTUPLE_FILTER_ALLOC_REQ_IP_ADDR_TYPE_IPV4;
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
^ permalink raw reply related [flat|nested] 12+ messages in thread
* Re: [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
` (9 preceding siblings ...)
2024-02-20 23:03 ` [PATCH net-next 10/10] bnxt_en: Use the new VNIC to create " Michael Chan
@ 2024-02-22 14:40 ` patchwork-bot+netdevbpf
10 siblings, 0 replies; 12+ messages in thread
From: patchwork-bot+netdevbpf @ 2024-02-22 14:40 UTC (permalink / raw
To: Michael Chan
Cc: davem, netdev, edumazet, kuba, pabeni, pavan.chebbi,
andrew.gospodarek
Hello:
This series was applied to netdev/net-next.git (main)
by Paolo Abeni <pabeni@redhat.com>:
On Tue, 20 Feb 2024 15:03:07 -0800 you wrote:
> The current Ntuple filter implementation has a limitation on 5750X (P5)
> and newer chips. The destination ring of the ntuple filter must be
> a valid ring in the RSS indirection table. Ntuple filters may not work
> if the RSS indirection table is modified by the user to only contain a
> subset of the rings. If an ntuple filter is set to a ring destination
> that is not in the RSS indirection table, the packet matching that
> filter will be placed in a random ring instead of the specified
> destination ring.
>
> [...]
Here is the summary with links:
- [net-next,01/10] bnxt_en: Refactor ring reservation functions
https://git.kernel.org/netdev/net-next/c/257bbf45af81
- [net-next,02/10] bnxt_en: Explicitly specify P5 completion rings to reserve
https://git.kernel.org/netdev/net-next/c/ae8186b2d406
- [net-next,03/10] bnxt_en: Improve RSS context reservation infrastructure
https://git.kernel.org/netdev/net-next/c/438ba39b25fe
- [net-next,04/10] bnxt_en: Check additional resources in bnxt_check_rings()
https://git.kernel.org/netdev/net-next/c/929429986773
- [net-next,05/10] bnxt_en: Add bnxt_get_total_vnics() to calculate number of VNICs
https://git.kernel.org/netdev/net-next/c/8c81ae6c54c1
- [net-next,06/10] bnxt_en: Refactor bnxt_set_features()
https://git.kernel.org/netdev/net-next/c/5d5b90fb4e90
- [net-next,07/10] bnxt_en: Define BNXT_VNIC_DEFAULT for the default vnic index
https://git.kernel.org/netdev/net-next/c/ef4ee64e9990
- [net-next,08/10] bnxt_en: Provision for an additional VNIC for ntuple filters
https://git.kernel.org/netdev/net-next/c/532c034e4b2b
- [net-next,09/10] bnxt_en: Create and setup the additional VNIC for adding ntuple filters
https://git.kernel.org/netdev/net-next/c/93e90104bd12
- [net-next,10/10] bnxt_en: Use the new VNIC to create ntuple filters
https://git.kernel.org/netdev/net-next/c/f6eff053a60c
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2024-02-22 14:40 UTC | newest]
Thread overview: 12+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-02-20 23:03 [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements Michael Chan
2024-02-20 23:03 ` [PATCH net-next 01/10] bnxt_en: Refactor ring reservation functions Michael Chan
2024-02-20 23:03 ` [PATCH net-next 02/10] bnxt_en: Explicitly specify P5 completion rings to reserve Michael Chan
2024-02-20 23:03 ` [PATCH net-next 03/10] bnxt_en: Improve RSS context reservation infrastructure Michael Chan
2024-02-20 23:03 ` [PATCH net-next 04/10] bnxt_en: Check additional resources in bnxt_check_rings() Michael Chan
2024-02-20 23:03 ` [PATCH net-next 05/10] bnxt_en: Add bnxt_get_total_vnics() to calculate number of VNICs Michael Chan
2024-02-20 23:03 ` [PATCH net-next 06/10] bnxt_en: Refactor bnxt_set_features() Michael Chan
2024-02-20 23:03 ` [PATCH net-next 07/10] bnxt_en: Define BNXT_VNIC_DEFAULT for the default vnic index Michael Chan
2024-02-20 23:03 ` [PATCH net-next 08/10] bnxt_en: Provision for an additional VNIC for ntuple filters Michael Chan
2024-02-20 23:03 ` [PATCH net-next 09/10] bnxt_en: Create and setup the additional VNIC for adding " Michael Chan
2024-02-20 23:03 ` [PATCH net-next 10/10] bnxt_en: Use the new VNIC to create " Michael Chan
2024-02-22 14:40 ` [PATCH net-next 00/10] bnxt_en: Ntuple filter improvements patchwork-bot+netdevbpf
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.