Skip to contentSkip to conversation

bnxt_en: Driver update for net-next.

PATCH Superseded Michael Chan 3 participants 14 messages Jul 3, 2020 – Jul 5, 2020

Subsystems (3)

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Michael Chan

author maintainer
#0 lore raw

v2 cover bnxt_en: Driver update for net-next.

This patchset implements ethtool -X to setup user-defined RSS indirection table. The new infrastructure also allows the proper logical ring index to be used to populate the RSS indirection when queried by ethtool -x. Prior to these patches, we were incorrectly populating the output of ethtool -x with internal ring IDs which would make no sense to the user.

The last 2 patches add some cleanups to the VLAN acceleration logic and check the firmware capabilities before allowing VLAN acceleration offloads.

v2: Some RSS indirection table changes requested by Jakub Kicinski.

Edwin Peer (2):
bnxt_en: clean up VLAN feature bit handling
bnxt_en: allow firmware to disable VLAN offloads

Michael Chan (6):
bnxt_en: Set up the chip specific RSS table size.
bnxt_en: Add logical RSS indirection table structure.
bnxt_en: Add helper function to return the number of RSS contexts.
bnxt_en: Fill HW RSS table from the RSS logical indirection table.
bnxt_en: Return correct RSS indirection table entries to ethtool -x.
bnxt_en: Implement ethtool -X to set indirection table.

 drivers/net/ethernet/broadcom/bnxt/bnxt.c         | 221 +++++++++++++++-------
 drivers/net/ethernet/broadcom/bnxt/bnxt.h         |  20 ++
 drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c |  52 ++++-
 3 files changed, 225 insertions(+), 68 deletions(-)
Signature · 1 line
1.8.3.1
author maintainer
#1 lore raw

PATCH v2 1/8 bnxt_en: Set up the chip specific RSS table size.

Currently, we allocate one page for the hardware DMA RSS indirection table. While the size is currently big enough for all chips, future chip variations may support bigger sizes, so it is better to calculate and store the chip specific size and allocate accordingly.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 12 ++++++++----
 drivers/net/ethernet/broadcom/bnxt/bnxt.h |  7 +++++++
 2 files changed, 15 insertions(+), 4 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +8 -4
index 6a884df..4afc1df 100644
@@ -3538,7 +3538,7 @@ static void bnxt_free_vnic_attributes(struct bnxt *bp)
 		}
 
 		if (vnic->rss_table) {
-			dma_free_coherent(&pdev->dev, PAGE_SIZE,
+			dma_free_coherent(&pdev->dev, vnic->rss_table_size,
 					  vnic->rss_table,
 					  vnic->rss_table_dma_addr);
 			vnic->rss_table = NULL;
@@ -3603,7 +3603,13 @@ static int bnxt_alloc_vnic_attributes(struct bnxt *bp)
 			continue;
 
 		/* Allocate rss table and hash key */
-		vnic->rss_table = dma_alloc_coherent(&pdev->dev, PAGE_SIZE,
+		size = L1_CACHE_ALIGN(HW_HASH_INDEX_SIZE * sizeof(u16));
+		if (bp->flags & BNXT_FLAG_CHIP_P5)
+			size = L1_CACHE_ALIGN(BNXT_MAX_RSS_TABLE_SIZE_P5);
+
+		vnic->rss_table_size = size + HW_HASH_KEY_SIZE;
+		vnic->rss_table = dma_alloc_coherent(&pdev->dev,
+						     vnic->rss_table_size,
 						     &vnic->rss_table_dma_addr,
 						     GFP_KERNEL);
 		if (!vnic->rss_table) {
@@ -3611,8 +3617,6 @@ static int bnxt_alloc_vnic_attributes(struct bnxt *bp)
 			goto out;
 		}
 
-		size = L1_CACHE_ALIGN(HW_HASH_INDEX_SIZE * sizeof(u16));
-
 		vnic->rss_hash_key = ((void *)vnic->rss_table) + size;
 		vnic->rss_hash_key_dma_addr = vnic->rss_table_dma_addr + size;
 	}
drivers/net/ethernet/broadcom/bnxt/bnxt.h +7 -0
index 78e2fd6..5883b246 100644
@@ -1017,6 +1017,13 @@ struct bnxt_vnic_info {
 	__le16		*rss_table;
 	dma_addr_t	rss_hash_key_dma_addr;
 	u64		*rss_hash_key;
+	int		rss_table_size;
+#define BNXT_RSS_TABLE_ENTRIES_P5	64
+#define BNXT_RSS_TABLE_SIZE_P5		(BNXT_RSS_TABLE_ENTRIES_P5 * 4)
+#define BNXT_RSS_TABLE_MAX_TBL_P5	8
+#define BNXT_MAX_RSS_TABLE_SIZE_P5				\
+	(BNXT_RSS_TABLE_SIZE_P5 * BNXT_RSS_TABLE_MAX_TBL_P5)
+
 	u32		rx_mask;
 
 	u8		*mc_list;
Signature · 1 line
1.8.3.1
author maintainer
#2 lore raw

PATCH v2 2/8 bnxt_en: Add logical RSS indirection table structure.

In reply to Michael Chan

The driver currently does not keep track of the logical RSS indirection table. The hardware RSS table is set up with standard default ring distribution when initializing the chip. This makes it difficult to support user sepcified indirection table entries. As a first step, add the logical table in the main bnxt structure and allocate it according to chip specific table size. Add a function that sets up default RSS distribution based on the number of RX rings.

v2: Use kmalloc_array() since we init. all entries afterwards.
Use ALIGN() to roundup the RSS table size.
Use ethtool_rxfh_indir_default() to init. the default entries.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 51 +++++++++++++++++++++++++++++++
 drivers/net/ethernet/broadcom/bnxt/bnxt.h |  4 +++
 2 files changed, 55 insertions(+)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +51 -0
index 4afc1df..70f8302 100644
@@ -4830,6 +4830,48 @@ static u16 bnxt_cp_ring_for_tx(struct bnxt *bp, struct bnxt_tx_ring_info *txr)
 	}
 }
 
+static int bnxt_alloc_rss_indir_tbl(struct bnxt *bp)
+{
+	int entries;
+
+	if (bp->flags & BNXT_FLAG_CHIP_P5)
+		entries = BNXT_MAX_RSS_TABLE_ENTRIES_P5;
+	else
+		entries = HW_HASH_INDEX_SIZE;
+
+	bp->rss_indir_tbl_entries = entries;
+	bp->rss_indir_tbl = kmalloc_array(entries, sizeof(*bp->rss_indir_tbl),
+					  GFP_KERNEL);
+	if (!bp->rss_indir_tbl)
+		return -ENOMEM;
+	return 0;
+}
+
+static void bnxt_set_dflt_rss_indir_tbl(struct bnxt *bp)
+{
+	u16 max_rings, max_entries, pad, i;
+
+	if (!bp->rx_nr_rings)
+		return;
+
+	if (BNXT_CHIP_TYPE_NITRO_A0(bp))
+		max_rings = bp->rx_nr_rings - 1;
+	else
+		max_rings = bp->rx_nr_rings;
+
+	if (bp->flags & BNXT_FLAG_CHIP_P5)
+		max_entries = ALIGN(max_rings, BNXT_RSS_TABLE_ENTRIES_P5);
+	else
+		max_entries = HW_HASH_INDEX_SIZE;
+
+	for (i = 0; i < max_entries; i++)
+		bp->rss_indir_tbl[i] = ethtool_rxfh_indir_default(i, max_rings);
+
+	pad = bp->rss_indir_tbl_entries - max_entries;
+	if (pad)
+		memset(&bp->rss_indir_tbl[i], 0, pad * sizeof(u16));
+}
+
 static int bnxt_hwrm_vnic_set_rss(struct bnxt *bp, u16 vnic_id, bool set_rss)
 {
 	u32 i, j, max_rings;
@@ -11514,6 +11556,8 @@ static void bnxt_remove_one(struct pci_dev *pdev)
 	bnxt_free_ctx_mem(bp);
 	kfree(bp->ctx);
 	bp->ctx = NULL;
+	kfree(bp->rss_indir_tbl);
+	bp->rss_indir_tbl = NULL;
 	bnxt_free_port_stats(bp);
 	free_netdev(dev);
 }
@@ -12034,6 +12078,11 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	 */
 	bp->tx_nr_rings_per_tc = bp->tx_nr_rings;
 
+	rc = bnxt_alloc_rss_indir_tbl(bp);
+	if (rc)
+		goto init_err_pci_clean;
+	bnxt_set_dflt_rss_indir_tbl(bp);
+
 	if (BNXT_PF(bp)) {
 		if (!bnxt_pf_wq) {
 			bnxt_pf_wq =
@@ -12078,6 +12127,8 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	bnxt_free_ctx_mem(bp);
 	kfree(bp->ctx);
 	bp->ctx = NULL;
+	kfree(bp->rss_indir_tbl);
+	bp->rss_indir_tbl = NULL;
 
 init_err_free:
 	free_netdev(dev);
drivers/net/ethernet/broadcom/bnxt/bnxt.h +4 -0
index 5883b246..6de2813 100644
@@ -1023,6 +1023,8 @@ struct bnxt_vnic_info {
 #define BNXT_RSS_TABLE_MAX_TBL_P5	8
 #define BNXT_MAX_RSS_TABLE_SIZE_P5				\
 	(BNXT_RSS_TABLE_SIZE_P5 * BNXT_RSS_TABLE_MAX_TBL_P5)
+#define BNXT_MAX_RSS_TABLE_ENTRIES_P5				\
+	(BNXT_RSS_TABLE_ENTRIES_P5 * BNXT_RSS_TABLE_MAX_TBL_P5)
 
 	u32		rx_mask;
 
@@ -1655,6 +1657,8 @@ struct bnxt {
 	struct bnxt_ring_grp_info	*grp_info;
 	struct bnxt_vnic_info	*vnic_info;
 	int			nr_vnics;
+	u16			*rss_indir_tbl;
+	u16			rss_indir_tbl_entries;
 	u32			rss_hash_cfg;
 
 	u16			max_mtu;
Signature · 1 line
1.8.3.1
author maintainer
#3 lore raw

PATCH v2 3/8 bnxt_en: Add helper function to return the number of RSS contexts.

In reply to Michael Chan

On some chips, this varies based on the number of RX rings. Add this helper function and refactor the existing code to use it.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 13 +++++++++++--
 drivers/net/ethernet/broadcom/bnxt/bnxt.h |  1 +
 2 files changed, 12 insertions(+), 2 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +11 -2
index 70f8302..96e678b 100644
@@ -4872,6 +4872,15 @@ static void bnxt_set_dflt_rss_indir_tbl(struct bnxt *bp)
 		memset(&bp->rss_indir_tbl[i], 0, pad * sizeof(u16));
 }
 
+int bnxt_get_nr_rss_ctxs(struct bnxt *bp, int rx_rings)
+{
+	if (bp->flags & BNXT_FLAG_CHIP_P5)
+		return DIV_ROUND_UP(rx_rings, BNXT_RSS_TABLE_ENTRIES_P5);
+	if (BNXT_CHIP_TYPE_NITRO_A0(bp))
+		return 2;
+	return 1;
+}
+
 static int bnxt_hwrm_vnic_set_rss(struct bnxt *bp, u16 vnic_id, bool set_rss)
 {
 	u32 i, j, max_rings;
@@ -4927,7 +4936,7 @@ static int bnxt_hwrm_vnic_set_rss_p5(struct bnxt *bp, u16 vnic_id, bool set_rss)
 	req.hash_mode_flags = VNIC_RSS_CFG_REQ_HASH_MODE_FLAGS_DEFAULT;
 	req.ring_grp_tbl_addr = cpu_to_le64(vnic->rss_table_dma_addr);
 	req.hash_key_tbl_addr = cpu_to_le64(vnic->rss_hash_key_dma_addr);
-	nr_ctxs = DIV_ROUND_UP(bp->rx_nr_rings, 64);
+	nr_ctxs = bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings);
 	for (i = 0, k = 0; i < nr_ctxs; i++) {
 		__le16 *ring_tbl = vnic->rss_table;
 		int rc;
@@ -7680,7 +7689,7 @@ static int __bnxt_setup_vnic_p5(struct bnxt *bp, u16 vnic_id)
 {
 	int rc, i, nr_ctxs;
 
-	nr_ctxs = DIV_ROUND_UP(bp->rx_nr_rings, 64);
+	nr_ctxs = bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings);
 	for (i = 0; i < nr_ctxs; i++) {
 		rc = bnxt_hwrm_vnic_ctx_alloc(bp, vnic_id, i);
 		if (rc) {
drivers/net/ethernet/broadcom/bnxt/bnxt.h +1 -0
index 6de2813..5890913 100644
@@ -2039,6 +2039,7 @@ int bnxt_alloc_rx_data(struct bnxt *bp, struct bnxt_rx_ring_info *rxr,
 int hwrm_send_message_silent(struct bnxt *, void *, u32, int);
 int bnxt_hwrm_func_drv_rgtr(struct bnxt *bp, unsigned long *bmap,
 			    int bmap_size, bool async_only);
+int bnxt_get_nr_rss_ctxs(struct bnxt *bp, int rx_rings);
 int bnxt_hwrm_vnic_cfg(struct bnxt *bp, u16 vnic_id);
 int __bnxt_hwrm_get_tx_rings(struct bnxt *bp, u16 fid, int *tx_rings);
 int bnxt_nq_rings_in_use(struct bnxt *bp);
Signature · 1 line
1.8.3.1
author maintainer
#4 lore raw

PATCH v2 4/8 bnxt_en: Fill HW RSS table from the RSS logical indirection table.

In reply to Michael Chan

Now that we have the logical table, we can fill the HW RSS table using the logical table's entries and converting them to the HW specific format. Re-initialize the logical table to standard distribution if the number of RX rings changes during ring reservation.

v2: Use ALIGN() to roundup the RSS table size.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 88 ++++++++++++++++++-------------
 1 file changed, 52 insertions(+), 36 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +52 -36
index 96e678b..6c90a94 100644
@@ -4881,9 +4881,51 @@ int bnxt_get_nr_rss_ctxs(struct bnxt *bp, int rx_rings)
 	return 1;
 }
 
+static void __bnxt_fill_hw_rss_tbl(struct bnxt *bp, struct bnxt_vnic_info *vnic)
+{
+	bool no_rss = !(vnic->flags & BNXT_VNIC_RSS_FLAG);
+	u16 i, j;
+
+	/* Fill the RSS indirection table with ring group ids */
+	for (i = 0, j = 0; i < HW_HASH_INDEX_SIZE; i++) {
+		if (!no_rss)
+			j = bp->rss_indir_tbl[i];
+		vnic->rss_table[i] = cpu_to_le16(vnic->fw_grp_ids[j]);
+	}
+}
+
+static void __bnxt_fill_hw_rss_tbl_p5(struct bnxt *bp,
+				      struct bnxt_vnic_info *vnic)
+{
+	__le16 *ring_tbl = vnic->rss_table;
+	struct bnxt_rx_ring_info *rxr;
+	u16 tbl_size, i;
+
+	tbl_size = ALIGN(bp->rx_nr_rings, BNXT_RSS_TABLE_ENTRIES_P5);
+
+	for (i = 0; i < tbl_size; i++) {
+		u16 ring_id, j;
+
+		j = bp->rss_indir_tbl[i];
+		rxr = &bp->rx_ring[j];
+
+		ring_id = rxr->rx_ring_struct.fw_ring_id;
+		*ring_tbl++ = cpu_to_le16(ring_id);
+		ring_id = bnxt_cp_ring_for_rx(bp, rxr);
+		*ring_tbl++ = cpu_to_le16(ring_id);
+	}
+}
+
+static void bnxt_fill_hw_rss_tbl(struct bnxt *bp, struct bnxt_vnic_info *vnic)
+{
+	if (bp->flags & BNXT_FLAG_CHIP_P5)
+		__bnxt_fill_hw_rss_tbl_p5(bp, vnic);
+	else
+		__bnxt_fill_hw_rss_tbl(bp, vnic);
+}
+
 static int bnxt_hwrm_vnic_set_rss(struct bnxt *bp, u16 vnic_id, bool set_rss)
 {
-	u32 i, j, max_rings;
 	struct bnxt_vnic_info *vnic = &bp->vnic_info[vnic_id];
 	struct hwrm_vnic_rss_cfg_input req = {0};
 
@@ -4893,24 +4935,9 @@ static int bnxt_hwrm_vnic_set_rss(struct bnxt *bp, u16 vnic_id, bool set_rss)
 
 	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_VNIC_RSS_CFG, -1, -1);
 	if (set_rss) {
+		bnxt_fill_hw_rss_tbl(bp, vnic);
 		req.hash_type = cpu_to_le32(bp->rss_hash_cfg);
 		req.hash_mode_flags = VNIC_RSS_CFG_REQ_HASH_MODE_FLAGS_DEFAULT;
-		if (vnic->flags & BNXT_VNIC_RSS_FLAG) {
-			if (BNXT_CHIP_TYPE_NITRO_A0(bp))
-				max_rings = bp->rx_nr_rings - 1;
-			else
-				max_rings = bp->rx_nr_rings;
-		} else {
-			max_rings = 1;
-		}
-
-		/* Fill the RSS indirection table with ring group ids */
-		for (i = 0, j = 0; i < HW_HASH_INDEX_SIZE; i++, j++) {
-			if (j == max_rings)
-				j = 0;
-			vnic->rss_table[i] = cpu_to_le16(vnic->fw_grp_ids[j]);
-		}
-
 		req.ring_grp_tbl_addr = cpu_to_le64(vnic->rss_table_dma_addr);
 		req.hash_key_tbl_addr =
 			cpu_to_le64(vnic->rss_hash_key_dma_addr);
@@ -4922,9 +4949,9 @@ static int bnxt_hwrm_vnic_set_rss(struct bnxt *bp, u16 vnic_id, bool set_rss)
 static int bnxt_hwrm_vnic_set_rss_p5(struct bnxt *bp, u16 vnic_id, bool set_rss)
 {
 	struct bnxt_vnic_info *vnic = &bp->vnic_info[vnic_id];
-	u32 i, j, k, nr_ctxs, max_rings = bp->rx_nr_rings;
-	struct bnxt_rx_ring_info *rxr = &bp->rx_ring[0];
 	struct hwrm_vnic_rss_cfg_input req = {0};
+	dma_addr_t ring_tbl_map;
+	u32 i, nr_ctxs;
 
 	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_VNIC_RSS_CFG, -1, -1);
 	req.vnic_id = cpu_to_le16(vnic->fw_vnic_id);
@@ -4932,31 +4959,18 @@ static int bnxt_hwrm_vnic_set_rss_p5(struct bnxt *bp, u16 vnic_id, bool set_rss)
 		hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
 		return 0;
 	}
+	bnxt_fill_hw_rss_tbl(bp, vnic);
 	req.hash_type = cpu_to_le32(bp->rss_hash_cfg);
 	req.hash_mode_flags = VNIC_RSS_CFG_REQ_HASH_MODE_FLAGS_DEFAULT;
-	req.ring_grp_tbl_addr = cpu_to_le64(vnic->rss_table_dma_addr);
 	req.hash_key_tbl_addr = cpu_to_le64(vnic->rss_hash_key_dma_addr);
+	ring_tbl_map = vnic->rss_table_dma_addr;
 	nr_ctxs = bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings);
-	for (i = 0, k = 0; i < nr_ctxs; i++) {
-		__le16 *ring_tbl = vnic->rss_table;
+	for (i = 0; i < nr_ctxs; ring_tbl_map += BNXT_RSS_TABLE_SIZE_P5, i++) {
 		int rc;
 
+		req.ring_grp_tbl_addr = cpu_to_le64(ring_tbl_map);
 		req.ring_table_pair_index = i;
 		req.rss_ctx_idx = cpu_to_le16(vnic->fw_rss_cos_lb_ctx[i]);
-		for (j = 0; j < 64; j++) {
-			u16 ring_id;
-
-			ring_id = rxr->rx_ring_struct.fw_ring_id;
-			*ring_tbl++ = cpu_to_le16(ring_id);
-			ring_id = bnxt_cp_ring_for_rx(bp, rxr);
-			*ring_tbl++ = cpu_to_le16(ring_id);
-			rxr++;
-			k++;
-			if (k == max_rings) {
-				k = 0;
-				rxr = &bp->rx_ring[0];
-			}
-		}
 		rc = hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
 		if (rc)
 			return rc;
@@ -8251,6 +8265,8 @@ int bnxt_reserve_rings(struct bnxt *bp, bool irq_re_init)
 			rc = bnxt_init_int_mode(bp);
 		bnxt_ulp_irq_restart(bp, rc);
 	}
+	bnxt_set_dflt_rss_indir_tbl(bp);
+
 	if (rc) {
 		netdev_err(bp->dev, "ring reservation/IRQ init failure rc: %d\n", rc);
 		return rc;
Signature · 1 line
1.8.3.1
author maintainer
#5 lore raw

PATCH v2 5/8 bnxt_en: Return correct RSS indirection table entries to ethtool -x.

In reply to Michael Chan

Now that we have the logical indirection table, we can return these proper logical indices directly to ethtool -x instead of the physical IDs.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +11 -4
index 6b88143..46f3978 100644
@@ -1275,6 +1275,12 @@ static int bnxt_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd)
 
 static u32 bnxt_get_rxfh_indir_size(struct net_device *dev)
 {
+	struct bnxt *bp = netdev_priv(dev);
+
+	if (bp->flags & BNXT_FLAG_CHIP_P5) {
+		return (bp->rx_nr_rings + BNXT_RSS_TABLE_ENTRIES_P5 - 1) &
+		       ~(BNXT_RSS_TABLE_ENTRIES_P5 - 1);
+	}
 	return HW_HASH_INDEX_SIZE;
 }
 
@@ -1288,7 +1294,7 @@ static int bnxt_get_rxfh(struct net_device *dev, u32 *indir, u8 *key,
 {
 	struct bnxt *bp = netdev_priv(dev);
 	struct bnxt_vnic_info *vnic;
-	int i = 0;
+	u32 i, tbl_size;
 
 	if (hfunc)
 		*hfunc = ETH_RSS_HASH_TOP;
@@ -1297,9 +1303,10 @@ static int bnxt_get_rxfh(struct net_device *dev, u32 *indir, u8 *key,
 		return 0;
 
 	vnic = &bp->vnic_info[0];
-	if (indir && vnic->rss_table) {
-		for (i = 0; i < HW_HASH_INDEX_SIZE; i++)
-			indir[i] = le16_to_cpu(vnic->rss_table[i]);
+	if (indir && bp->rss_indir_tbl) {
+		tbl_size = bnxt_get_rxfh_indir_size(dev);
+		for (i = 0; i < tbl_size; i++)
+			indir[i] = bp->rss_indir_tbl[i];
 	}
 
 	if (key && vnic->rss_hash_key)
Signature · 1 line
1.8.3.1
maintainer
#6 lore raw

Re: [PATCH net-next v2 5/8] bnxt_en: Return correct RSS indirection table entries to ethtool -x.

drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c static int bnxt_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd) · line 1283

Quoted 4 lines · On Fri, 3 Jul 2020 03:19:44 -0400 Michael Chan wrote:
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +4 -0
+	if (bp->flags & BNXT_FLAG_CHIP_P5) {
+		return (bp->rx_nr_rings + BNXT_RSS_TABLE_ENTRIES_P5 - 1) &
+		       ~(BNXT_RSS_TABLE_ENTRIES_P5 - 1);
+	}

ALIGN() here as well?

author maintainer
#7 lore raw

PATCH v2 6/8 bnxt_en: Implement ethtool -X to set indirection table.

In reply to Michael Chan

With the new infrastructure in place, we can now support the setting of the indirection table from ethtool.

When changing channels, in a rare case that firmware cannot reserve the rings that were promised, the RSS map will revert to default.

v2: When changing channels, if the RSS table size changes and RSS map
is non-default, return error.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c         |  7 ++++-
 drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 37 +++++++++++++++++++++++
 2 files changed, 43 insertions(+), 1 deletion(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +6 -1
index 6c90a94..0edb692 100644
@@ -6061,6 +6061,10 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
 		rx = rx_rings << 1;
 	cp = sh ? max_t(int, tx, rx_rings) : tx + rx_rings;
 	bp->tx_nr_rings = tx;
+
+	/* Reset the RSS indirection if we cannot reserve all the RX rings */
+	if (rx_rings != bp->rx_nr_rings)
+		bp->dev->priv_flags &= ~IFF_RXFH_CONFIGURED;
 	bp->rx_nr_rings = rx_rings;
 	bp->cp_nr_rings = cp;
 
@@ -8265,7 +8269,8 @@ int bnxt_reserve_rings(struct bnxt *bp, bool irq_re_init)
 			rc = bnxt_init_int_mode(bp);
 		bnxt_ulp_irq_restart(bp, rc);
 	}
-	bnxt_set_dflt_rss_indir_tbl(bp);
+	if (!netif_is_rxfh_configured(bp->dev))
+		bnxt_set_dflt_rss_indir_tbl(bp);
 
 	if (rc) {
 		netdev_err(bp->dev, "ring reservation/IRQ init failure rc: %d\n", rc);
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +37 -0
index 46f3978..9098818 100644
@@ -926,6 +926,13 @@ static int bnxt_set_channels(struct net_device *dev,
 		return rc;
 	}
 
+	if (bnxt_get_nr_rss_ctxs(bp, req_rx_rings) !=
+	    bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings) &&
+	    (dev->priv_flags & IFF_RXFH_CONFIGURED)) {
+		netdev_warn(dev, "RSS table size change required, RSS table entries must be default to proceed\n");
+		return -EINVAL;
+	}
+
 	if (netif_running(dev)) {
 		if (BNXT_PF(bp)) {
 			/* TODO CHIMP_FW: Send message to all VF's
@@ -1315,6 +1322,35 @@ static int bnxt_get_rxfh(struct net_device *dev, u32 *indir, u8 *key,
 	return 0;
 }
 
+static int bnxt_set_rxfh(struct net_device *dev, const u32 *indir,
+			 const u8 *key, const u8 hfunc)
+{
+	struct bnxt *bp = netdev_priv(dev);
+	int rc = 0;
+
+	if (hfunc && hfunc != ETH_RSS_HASH_TOP)
+		return -EOPNOTSUPP;
+
+	if (key)
+		return -EOPNOTSUPP;
+
+	if (indir) {
+		u32 i, pad, tbl_size = bnxt_get_rxfh_indir_size(dev);
+
+		for (i = 0; i < tbl_size; i++)
+			bp->rss_indir_tbl[i] = indir[i];
+		pad = bp->rss_indir_tbl_entries - tbl_size;
+		if (pad)
+			memset(&bp->rss_indir_tbl[i], 0, pad * sizeof(u16));
+	}
+
+	if (netif_running(bp->dev)) {
+		bnxt_close_nic(bp, false, false);
+		rc = bnxt_open_nic(bp, false, false);
+	}
+	return rc;
+}
+
 static void bnxt_get_drvinfo(struct net_device *dev,
 			     struct ethtool_drvinfo *info)
 {
@@ -3621,6 +3657,7 @@ void bnxt_ethtool_free(struct bnxt *bp)
 	.get_rxfh_indir_size    = bnxt_get_rxfh_indir_size,
 	.get_rxfh_key_size      = bnxt_get_rxfh_key_size,
 	.get_rxfh               = bnxt_get_rxfh,
+	.set_rxfh		= bnxt_set_rxfh,
 	.flash_device		= bnxt_flash_device,
 	.get_eeprom_len         = bnxt_get_eeprom_len,
 	.get_eeprom             = bnxt_get_eeprom,
Signature · 1 line
1.8.3.1
maintainer
#8 lore raw

Re: [PATCH net-next v2 6/8] bnxt_en: Implement ethtool -X to set indirection table.

Quoted 8 lines · On Fri, 3 Jul 2020 03:19:45 -0400 Michael Chan wrote:

With the new infrastructure in place, we can now support the setting of the indirection table from ethtool.

When changing channels, in a rare case that firmware cannot reserve the rings that were promised, the RSS map will revert to default.

v2: When changing channels, if the RSS table size changes and RSS map
is non-default, return error.

Sorry, still not what I had in mind :(

Quoted 7 lines
drivers/net/ethernet/broadcom/bnxt/bnxt.c +0 -0
index 6c90a94..0edb692 100644

drivers/net/ethernet/broadcom/bnxt/bnxt.c static int __bnxt_reserve_rings(struct bnxt *bp) · line 6067

Quoted 7 lines
drivers/net/ethernet/broadcom/bnxt/bnxt.c +4 -0
@@ -6061,6 +6061,10 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
 		rx = rx_rings << 1;
 	cp = sh ? max_t(int, tx, rx_rings) : tx + rx_rings;
 	bp->tx_nr_rings = tx;
+
+	/* Reset the RSS indirection if we cannot reserve all the RX rings */
+	if (rx_rings != bp->rx_nr_rings)
+		bp->dev->priv_flags &= ~IFF_RXFH_CONFIGURED;

Ethtool core will check that if user set RSS table explicitly and now changes ring count - the RSS table will not point to rings which will no longer exist. IOW if RSS table is set we're likely increasing the number of rings, and I'd venture to guess that the FW is not gonna give us less rings than we previously had. So it seems like we may be clearing the flag, and the RSS table unnecessarily here.

What I was suggesting, was that it perhaps be better to modulo the number or rings in __bnxt_fill_hw_rss_tbl* by the actual table size, but leave the user-set RSS table in place. That'd be the lowest LoC.

Also I still think that there should be a warning printed when FW gave us less rings than expected.

Quoted 18 lines
drivers/net/ethernet/broadcom/bnxt/bnxt.c +2 -1
 	bp->rx_nr_rings = rx_rings;
 	bp->cp_nr_rings = cp;
 
@@ -8265,7 +8269,8 @@ int bnxt_reserve_rings(struct bnxt *bp, bool irq_re_init)
 			rc = bnxt_init_int_mode(bp);
 		bnxt_ulp_irq_restart(bp, rc);
 	}
-	bnxt_set_dflt_rss_indir_tbl(bp);
+	if (!netif_is_rxfh_configured(bp->dev))
+		bnxt_set_dflt_rss_indir_tbl(bp);
 
 	if (rc) {
 		netdev_err(bp->dev, "ring reservation/IRQ init failure rc: %d\n", rc);
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +0 -0
index 46f3978..9098818 100644

drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c static int bnxt_set_channels(struct net_device *dev, · line 931

Quoted 6 lines
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +3 -0
@@ -926,6 +926,13 @@ static int bnxt_set_channels(struct net_device *dev,
 		return rc;
 	}
 
+	if (bnxt_get_nr_rss_ctxs(bp, req_rx_rings) !=
+	    bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings) &&
+	    (dev->priv_flags & IFF_RXFH_CONFIGURED)) {

In this case copy the old values over and zero-fill the new rings.

Quoted 22 lines
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +15 -0
+		netdev_warn(dev, "RSS table size change required, RSS table entries must be default to proceed\n");
+		return -EINVAL;
+	}
+
 	if (netif_running(dev)) {
 		if (BNXT_PF(bp)) {
 			/* TODO CHIMP_FW: Send message to all VF's
@@ -1315,6 +1322,35 @@ static int bnxt_get_rxfh(struct net_device *dev, u32 *indir, u8 *key,
 	return 0;
 }
 
+static int bnxt_set_rxfh(struct net_device *dev, const u32 *indir,
+			 const u8 *key, const u8 hfunc)
+{
+	struct bnxt *bp = netdev_priv(dev);
+	int rc = 0;
+
+	if (hfunc && hfunc != ETH_RSS_HASH_TOP)
+		return -EOPNOTSUPP;
+
+	if (key)
+		return -EOPNOTSUPP;

drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c static int bnxt_get_rxfh(struct net_device *dev, u32 *indir, u8 *key, · line 1345

Quoted 10 lines
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +10 -0
+
+	if (indir) {
+		u32 i, pad, tbl_size = bnxt_get_rxfh_indir_size(dev);
+
+		for (i = 0; i < tbl_size; i++)
+			bp->rss_indir_tbl[i] = indir[i];
+		pad = bp->rss_indir_tbl_entries - tbl_size;
+		if (pad)
+			memset(&bp->rss_indir_tbl[i], 0, pad * sizeof(u16));
+	}

I see in patch 4 there is a:

	bool no_rss = !(vnic->flags & BNXT_VNIC_RSS_FLAG);

Should there be some check here to only allow users to change the indirection table if RSS is supported / allowed?

author maintainer
#9 lore raw

Re: [PATCH net-next v2 6/8] bnxt_en: Implement ethtool -X to set indirection table.

Quoted 13 lines · Jakub Kicinski
Quoted 11 lines · On Fri, 3 Jul 2020 03:19:45 -0400 Michael Chan wrote:
drivers/net/ethernet/broadcom/bnxt/bnxt.c +3 -0
index 6c90a94..0edb692 100644
@@ -6061,6 +6061,10 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
              rx = rx_rings << 1;
      cp = sh ? max_t(int, tx, rx_rings) : tx + rx_rings;
      bp->tx_nr_rings = tx;
+
+     /* Reset the RSS indirection if we cannot reserve all the RX rings */
+     if (rx_rings != bp->rx_nr_rings)
Quoted 8 lines

drivers/net/ethernet/broadcom/bnxt/bnxt.c static int __bnxt_reserve_rings(struct bnxt *bp) · line 6067

Quoted 1 line
drivers/net/ethernet/broadcom/bnxt/bnxt.c +1 -0
+             bp->dev->priv_flags &= ~IFF_RXFH_CONFIGURED;

Ethtool core will check that if user set RSS table explicitly and now changes ring count - the RSS table will not point to rings which will no longer exist. IOW if RSS table is set we're likely increasing the number of rings, and I'd venture to guess that the FW is not gonna give us less rings than we previously had. So it seems like we may be clearing the flag, and the RSS table unnecessarily here.

Right, the user most likely wants to increase the rings with RSS map already set. The user can decrease rings too if the higher rings don't have any weight. We call bnxt_check_rings() to see if firmware has the requested rings before we proceed. Once we proceed, we're committed and if firmware cannot give us the rings it promised earlier, the RSS map may no longer be valid if the rings are even fewer than what we had before. This is rare, as I said earlier, but it can happen, especially on the VFs. The resources are less guaranteed on the VFs.

Quoted 7 lines

What I was suggesting, was that it perhaps be better to modulo the number or rings in __bnxt_fill_hw_rss_tbl* by the actual table size, but leave the user-set RSS table in place. That'd be the lowest LoC.

Also I still think that there should be a warning printed when FW gave us less rings than expected.

Sure, we can add some warnings if we don't warn already.

Quoted 22 lines
Quoted 22 lines
drivers/net/ethernet/broadcom/bnxt/bnxt.c +2 -1
      bp->rx_nr_rings = rx_rings;
      bp->cp_nr_rings = cp;
 
@@ -8265,7 +8269,8 @@ int bnxt_reserve_rings(struct bnxt *bp, bool irq_re_init)
                      rc = bnxt_init_int_mode(bp);
              bnxt_ulp_irq_restart(bp, rc);
      }
-     bnxt_set_dflt_rss_indir_tbl(bp);
+     if (!netif_is_rxfh_configured(bp->dev))
+             bnxt_set_dflt_rss_indir_tbl(bp);
 
      if (rc) {
              netdev_err(bp->dev, "ring reservation/IRQ init failure rc: %d\n", rc);
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +0 -0
index 46f3978..9098818 100644
@@ -926,6 +926,13 @@ static int bnxt_set_channels(struct net_device *dev,
              return rc;
      }
 
Quoted 5 lines

drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c static int bnxt_set_channels(struct net_device *dev, · line 931

Quoted 3 lines
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +3 -0
+     if (bnxt_get_nr_rss_ctxs(bp, req_rx_rings) !=
+         bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings) &&
+         (dev->priv_flags & IFF_RXFH_CONFIGURED)) {

In this case copy the old values over and zero-fill the new rings.

If the existing RSS table has 128 entries and a custom map. And now the table needs to expand to 192 entries with 64 new entries, we cannot just copy, right? The weights will all be messed up if we try. I think requiring the user to change back to default or force it back to default are the only sane options.

Quoted 36 lines
Quoted 23 lines
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +15 -0
+             netdev_warn(dev, "RSS table size change required, RSS table entries must be default to proceed\n");
+             return -EINVAL;
+     }
+
      if (netif_running(dev)) {
              if (BNXT_PF(bp)) {
                      /* TODO CHIMP_FW: Send message to all VF's
@@ -1315,6 +1322,35 @@ static int bnxt_get_rxfh(struct net_device *dev, u32 *indir, u8 *key,
      return 0;
 }
 
+static int bnxt_set_rxfh(struct net_device *dev, const u32 *indir,
+                      const u8 *key, const u8 hfunc)
+{
+     struct bnxt *bp = netdev_priv(dev);
+     int rc = 0;
+
+     if (hfunc && hfunc != ETH_RSS_HASH_TOP)
+             return -EOPNOTSUPP;
+
+     if (key)
+             return -EOPNOTSUPP;

drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c static int bnxt_get_rxfh(struct net_device *dev, u32 *indir, u8 *key, · line 1345

Quoted 10 lines
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +10 -0
+
+     if (indir) {
+             u32 i, pad, tbl_size = bnxt_get_rxfh_indir_size(dev);
+
+             for (i = 0; i < tbl_size; i++)
+                     bp->rss_indir_tbl[i] = indir[i];
+             pad = bp->rss_indir_tbl_entries - tbl_size;
+             if (pad)
+                     memset(&bp->rss_indir_tbl[i], 0, pad * sizeof(u16));
+     }

I see in patch 4 there is a:

Quoted 4 lines
        bool no_rss = !(vnic->flags & BNXT_VNIC_RSS_FLAG);

Should there be some check here to only allow users to change the indirection table if RSS is supported / allowed?

RSS is always supported on the function. The first VNIC hw structure that we allocate always supports RSS. We allocate additional VNICs for aRFS and those VNICs do not use RSS and the BNXT_VNIC_RSS_FLAG will not be set on those VNICs. That's what the code in patch 4 is doing. The bp->rss_indir_tbl always applies to the first VNIC that supports RSS.

maintainer
#10 lore raw

Re: [PATCH net-next v2 6/8] bnxt_en: Implement ethtool -X to set indirection table.

Quoted 27 lines · On Fri, 3 Jul 2020 17:31:59 -0700 Michael Chan wrote:
Quoted 21 lines
Quoted 21 lines
      bp->rx_nr_rings = rx_rings;
      bp->cp_nr_rings = cp;
drivers/net/ethernet/broadcom/bnxt/bnxt.c +2 -1
@@ -8265,7 +8269,8 @@ int bnxt_reserve_rings(struct bnxt *bp, bool irq_re_init)
                      rc = bnxt_init_int_mode(bp);
              bnxt_ulp_irq_restart(bp, rc);
      }
-     bnxt_set_dflt_rss_indir_tbl(bp);
+     if (!netif_is_rxfh_configured(bp->dev))
+             bnxt_set_dflt_rss_indir_tbl(bp);
 
      if (rc) {
              netdev_err(bp->dev, "ring reservation/IRQ init failure rc: %d\n", rc);
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +0 -0
index 46f3978..9098818 100644
@@ -926,6 +926,13 @@ static int bnxt_set_channels(struct net_device *dev,
              return rc;
      }
 
Quoted 5 lines

drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c static int bnxt_set_channels(struct net_device *dev, · line 931

Quoted 3 lines
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +3 -0
+     if (bnxt_get_nr_rss_ctxs(bp, req_rx_rings) !=
+         bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings) &&
+         (dev->priv_flags & IFF_RXFH_CONFIGURED)) {  

In this case copy the old values over and zero-fill the new rings.

Quoted 5 lines

If the existing RSS table has 128 entries and a custom map. And now the table needs to expand to 192 entries with 64 new entries, we cannot just copy, right? The weights will all be messed up if we try. I think requiring the user to change back to default or force it back to default are the only sane options.

I see it now, you're right.

author maintainer
#11 lore raw

PATCH v2 7/8 bnxt_en: clean up VLAN feature bit handling

In reply to Michael Chan

From: Edwin Peer <[email protected]>

The hardware VLAN offload feature on our NIC does not have separate knobs for handling customer and service tags on RX. Either offloading of both must be enabled or both must be disabled. Introduce definitions for the combined feature set in order to clean up the code and make this constraint more clear. Technically these features can be separately enabled on TX, however, since the default is to turn both on, the combined TX feature set is also introduced for code consistency.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 34 ++++++++++++-------------------
 drivers/net/ethernet/broadcom/bnxt/bnxt.h |  5 +++++
 2 files changed, 18 insertions(+), 21 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +13 -21
index 0edb692..c7a6a2a 100644
@@ -1614,7 +1614,7 @@ static inline struct sk_buff *bnxt_tpa_end(struct bnxt *bp,
 		skb_set_hash(skb, tpa_info->rss_hash, tpa_info->hash_type);
 
 	if ((tpa_info->flags2 & RX_CMP_FLAGS2_META_FORMAT_VLAN) &&
-	    (skb->dev->features & NETIF_F_HW_VLAN_CTAG_RX)) {
+	    (skb->dev->features & BNXT_HW_FEATURE_VLAN_ALL_RX)) {
 		u16 vlan_proto = tpa_info->metadata >>
 			RX_CMP_FLAGS2_METADATA_TPID_SFT;
 		u16 vtag = tpa_info->metadata & RX_CMP_FLAGS2_METADATA_TCI_MASK;
@@ -1832,7 +1832,7 @@ static int bnxt_rx_pkt(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
 
 	if ((rxcmp1->rx_cmp_flags2 &
 	     cpu_to_le32(RX_CMP_FLAGS2_META_FORMAT_VLAN)) &&
-	    (skb->dev->features & NETIF_F_HW_VLAN_CTAG_RX)) {
+	    (skb->dev->features & BNXT_HW_FEATURE_VLAN_ALL_RX)) {
 		u32 meta_data = le32_to_cpu(rxcmp1->rx_cmp_meta_data);
 		u16 vtag = meta_data & RX_CMP_FLAGS2_METADATA_TCI_MASK;
 		u16 vlan_proto = meta_data >> RX_CMP_FLAGS2_METADATA_TPID_SFT;
@@ -9911,24 +9911,16 @@ static netdev_features_t bnxt_fix_features(struct net_device *dev,
 	/* Both CTAG and STAG VLAN accelaration on the RX side have to be
 	 * turned on or off together.
 	 */
-	vlan_features = features & (NETIF_F_HW_VLAN_CTAG_RX |
-				    NETIF_F_HW_VLAN_STAG_RX);
-	if (vlan_features != (NETIF_F_HW_VLAN_CTAG_RX |
-			      NETIF_F_HW_VLAN_STAG_RX)) {
-		if (dev->features & NETIF_F_HW_VLAN_CTAG_RX)
-			features &= ~(NETIF_F_HW_VLAN_CTAG_RX |
-				      NETIF_F_HW_VLAN_STAG_RX);
+	vlan_features = features & BNXT_HW_FEATURE_VLAN_ALL_RX;
+	if (vlan_features != BNXT_HW_FEATURE_VLAN_ALL_RX) {
+		if (dev->features & BNXT_HW_FEATURE_VLAN_ALL_RX)
+			features &= ~BNXT_HW_FEATURE_VLAN_ALL_RX;
 		else if (vlan_features)
-			features |= NETIF_F_HW_VLAN_CTAG_RX |
-				    NETIF_F_HW_VLAN_STAG_RX;
+			features |= BNXT_HW_FEATURE_VLAN_ALL_RX;
 	}
 #ifdef CONFIG_BNXT_SRIOV
-	if (BNXT_VF(bp)) {
-		if (bp->vf.vlan) {
-			features &= ~(NETIF_F_HW_VLAN_CTAG_RX |
-				      NETIF_F_HW_VLAN_STAG_RX);
-		}
-	}
+	if (BNXT_VF(bp) && bp->vf.vlan)
+		features &= ~BNXT_HW_FEATURE_VLAN_ALL_RX;
 #endif
 	return features;
 }
@@ -9951,7 +9943,7 @@ static int bnxt_set_features(struct net_device *dev, netdev_features_t features)
 	if (bp->flags & BNXT_FLAG_NO_AGG_RINGS)
 		flags &= ~BNXT_FLAG_TPA;
 
-	if (features & NETIF_F_HW_VLAN_CTAG_RX)
+	if (features & BNXT_HW_FEATURE_VLAN_ALL_RX)
 		flags |= BNXT_FLAG_STRIP_VLAN;
 
 	if (features & NETIF_F_NTUPLE)
@@ -12039,8 +12031,8 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	dev->gso_partial_features = NETIF_F_GSO_UDP_TUNNEL_CSUM |
 				    NETIF_F_GSO_GRE_CSUM;
 	dev->vlan_features = dev->hw_features | NETIF_F_HIGHDMA;
-	dev->hw_features |= NETIF_F_HW_VLAN_CTAG_RX | NETIF_F_HW_VLAN_CTAG_TX |
-			    NETIF_F_HW_VLAN_STAG_RX | NETIF_F_HW_VLAN_STAG_TX;
+	dev->hw_features |= BNXT_HW_FEATURE_VLAN_ALL_RX |
+			    BNXT_HW_FEATURE_VLAN_ALL_TX;
 	if (BNXT_SUPPORTS_TPA(bp))
 		dev->hw_features |= NETIF_F_GRO_HW;
 	dev->features |= dev->hw_features | NETIF_F_HIGHDMA;
@@ -12096,7 +12088,7 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	bnxt_fw_init_one_p3(bp);
 
-	if (dev->hw_features & NETIF_F_HW_VLAN_CTAG_RX)
+	if (dev->hw_features & BNXT_HW_FEATURE_VLAN_ALL_RX)
 		bp->flags |= BNXT_FLAG_STRIP_VLAN;
 
 	rc = bnxt_init_int_mode(bp);
drivers/net/ethernet/broadcom/bnxt/bnxt.h +5 -0
index 5890913..13c4064 100644
@@ -1906,6 +1906,11 @@ struct bnxt {
 #define BNXT_PCIE_STATS_OFFSET(counter)			\
 	(offsetof(struct pcie_ctx_hw_stats, counter) / 8)
 
+#define BNXT_HW_FEATURE_VLAN_ALL_RX				\
+	(NETIF_F_HW_VLAN_CTAG_RX | NETIF_F_HW_VLAN_STAG_RX)
+#define BNXT_HW_FEATURE_VLAN_ALL_TX				\
+	(NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_STAG_TX)
+
 #define I2C_DEV_ADDR_A0				0xa0
 #define I2C_DEV_ADDR_A2				0xa2
 #define SFF_DIAG_SUPPORT_OFFSET			0x5c
Signature · 1 line
1.8.3.1
author maintainer
#12 lore raw

PATCH v2 8/8 bnxt_en: allow firmware to disable VLAN offloads

In reply to Michael Chan

From: Edwin Peer <[email protected]>

Bare-metal use cases require giving firmware and the embedded application processor control over VLAN offloads. The driver should not attempt to override or utilize this feature in such scenarios since it will not work as expected.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 22 +++++++++++++++++++---
 drivers/net/ethernet/broadcom/bnxt/bnxt.h |  3 +++
 2 files changed, 22 insertions(+), 3 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +19 -3
index c7a6a2a..5553bad 100644
@@ -5208,6 +5208,14 @@ static int bnxt_hwrm_vnic_qcaps(struct bnxt *bp)
 		if (flags &
 		    VNIC_QCAPS_RESP_FLAGS_ROCE_MIRRORING_CAPABLE_VNIC_CAP)
 			bp->flags |= BNXT_FLAG_ROCE_MIRROR_CAP;
+
+		/* Older P5 fw before EXT_HW_STATS support did not set
+		 * VLAN_STRIP_CAP properly.
+		 */
+		if ((flags & VNIC_QCAPS_RESP_FLAGS_VLAN_STRIP_CAP) ||
+		    ((bp->flags & BNXT_FLAG_CHIP_P5) &&
+		     !(bp->fw_cap & BNXT_FW_CAP_EXT_HW_STATS_SUPPORTED)))
+			bp->fw_cap |= BNXT_FW_CAP_VLAN_RX_STRIP;
 		bp->max_tpa_v2 = le16_to_cpu(resp->max_aggs_supported);
 		if (bp->max_tpa_v2)
 			bp->hw_ring_stats_size =
@@ -7028,7 +7036,7 @@ static int __bnxt_hwrm_func_qcaps(struct bnxt *bp)
 	struct hwrm_func_qcaps_input req = {0};
 	struct hwrm_func_qcaps_output *resp = bp->hwrm_cmd_resp_addr;
 	struct bnxt_hw_resc *hw_resc = &bp->hw_resc;
-	u32 flags;
+	u32 flags, flags_ext;
 
 	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_FUNC_QCAPS, -1, -1);
 	req.fid = cpu_to_le16(0xffff);
@@ -7053,6 +7061,12 @@ static int __bnxt_hwrm_func_qcaps(struct bnxt *bp)
 		bp->fw_cap |= BNXT_FW_CAP_ERROR_RECOVERY;
 	if (flags & FUNC_QCAPS_RESP_FLAGS_ERR_RECOVER_RELOAD)
 		bp->fw_cap |= BNXT_FW_CAP_ERR_RECOVER_RELOAD;
+	if (!(flags & FUNC_QCAPS_RESP_FLAGS_VLAN_ACCELERATION_TX_DISABLED))
+		bp->fw_cap |= BNXT_FW_CAP_VLAN_TX_INSERT;
+
+	flags_ext = le32_to_cpu(resp->flags_ext);
+	if (flags_ext & FUNC_QCAPS_RESP_FLAGS_EXT_EXT_HW_STATS_SUPPORTED)
+		bp->fw_cap |= BNXT_FW_CAP_EXT_HW_STATS_SUPPORTED;
 
 	bp->tx_push_thresh = 0;
 	if ((flags & FUNC_QCAPS_RESP_FLAGS_PUSH_MODE_SUPPORTED) &&
@@ -12031,8 +12045,10 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	dev->gso_partial_features = NETIF_F_GSO_UDP_TUNNEL_CSUM |
 				    NETIF_F_GSO_GRE_CSUM;
 	dev->vlan_features = dev->hw_features | NETIF_F_HIGHDMA;
-	dev->hw_features |= BNXT_HW_FEATURE_VLAN_ALL_RX |
-			    BNXT_HW_FEATURE_VLAN_ALL_TX;
+	if (bp->fw_cap & BNXT_FW_CAP_VLAN_RX_STRIP)
+		dev->hw_features |= BNXT_HW_FEATURE_VLAN_ALL_RX;
+	if (bp->fw_cap & BNXT_FW_CAP_VLAN_TX_INSERT)
+		dev->hw_features |= BNXT_HW_FEATURE_VLAN_ALL_TX;
 	if (BNXT_SUPPORTS_TPA(bp))
 		dev->hw_features |= NETIF_F_GRO_HW;
 	dev->features |= dev->hw_features | NETIF_F_HIGHDMA;
drivers/net/ethernet/broadcom/bnxt/bnxt.h +3 -0
index 13c4064..d556e56 100644
@@ -1716,6 +1716,9 @@ struct bnxt {
 	#define BNXT_FW_CAP_ERR_RECOVER_RELOAD		0x00100000
 	#define BNXT_FW_CAP_HOT_RESET			0x00200000
 	#define BNXT_FW_CAP_SHARED_PORT_CFG		0x00400000
+	#define BNXT_FW_CAP_VLAN_RX_STRIP		0x01000000
+	#define BNXT_FW_CAP_VLAN_TX_INSERT		0x02000000
+	#define BNXT_FW_CAP_EXT_HW_STATS_SUPPORTED	0x04000000
 
 #define BNXT_NEW_RM(bp)		((bp)->fw_cap & BNXT_FW_CAP_NEW_RM)
 	u32			hwrm_spec_code;
Signature · 1 line
1.8.3.1
maintainer
#13 lore raw
In reply to Michael Chan

From: Michael Chan <[email protected]>
Date: Fri, 3 Jul 2020 03:19:39 -0400

Quoted 11 lines

This patchset implements ethtool -X to setup user-defined RSS indirection table. The new infrastructure also allows the proper logical ring index to be used to populate the RSS indirection when queried by ethtool -x. Prior to these patches, we were incorrectly populating the output of ethtool -x with internal ring IDs which would make no sense to the user.

The last 2 patches add some cleanups to the VLAN acceleration logic and check the firmware capabilities before allowing VLAN acceleration offloads.

v2: Some RSS indirection table changes requested by Jakub Kicinski.

Jakub, please review.