Skip to contentSkip to conversation

bnxt_en: Bug fixes

PATCH Applied Michael Chan 3 participants 8 messages Jul 31 – Aug 5

Subsystems (3)
lore
Share & tools
lore

Michael Chan

author maintainer
#0 lore raw

cover bnxt_en: Bug fixes

This series include 3 bug fixes:

1. queue start bug fix on the VNIC's default ring. 2 refactoring patches preceed the actual bug fix.
2. Bug fix for TPA data corruption seen on some ARM systems.
3. PTP PPS setting bug fix.

Keegan Freyhof (1):
bnxt_en: Fix PTP PPS setting bug

Michael Chan (1):
bnxt_en: Disable EOP for TPA on all chips to prevent data corruption

Shravya KN (3):
bnxt_en: Move RSS table fill outside __bnxt_hwrm_vnic_set_rss()
bnxt_en: Determine and store default RX ring in vnic structure
bnxt_en: Refresh VNIC default ring on queue restart if needed

 drivers/net/ethernet/broadcom/bnxt/bnxt.c     | 61 ++++++++++++-------
 drivers/net/ethernet/broadcom/bnxt/bnxt.h     |  1 +
 drivers/net/ethernet/broadcom/bnxt/bnxt_ptp.c |  7 ++-
 3 files changed, 46 insertions(+), 23 deletions(-)
Signature · 1 line
2.51.0
author maintainer
#1 lore raw

PATCH 1/5 bnxt_en: Move RSS table fill outside __bnxt_hwrm_vnic_set_rss()

From: Shravya KN <[email protected]>

This is a refactor patch with no change in behavior. The caller will now fill the RSS table before calling __bnxt_hwrm_vnic_set_rss(). In the next patch, we'll add code to determine the default ring for the VNIC when we fill the RSS table.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +7 -7
index 7513618793da..48f4e38ad78d 100644
@@ -6706,6 +6706,9 @@ 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;
 
+	if (!vnic->rss_table)
+		return;
+
 	/* Fill the RSS indirection table with ring group ids */
 	for (i = 0, j = 0; i < HW_HASH_INDEX_SIZE; i++) {
 		if (!no_rss)
@@ -6745,13 +6748,8 @@ static void
 __bnxt_hwrm_vnic_set_rss(struct bnxt *bp, struct hwrm_vnic_rss_cfg_input *req,
 			 struct bnxt_vnic_info *vnic)
 {
-	if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
-		bnxt_fill_hw_rss_tbl_p5(bp, vnic);
-		if (bp->flags & BNXT_FLAG_CHIP_P7)
-			req->flags |= VNIC_RSS_CFG_REQ_FLAGS_IPSEC_HASH_TYPE_CFG_SUPPORT;
-	} else {
-		bnxt_fill_hw_rss_tbl(bp, vnic);
-	}
+	if (bp->flags & BNXT_FLAG_CHIP_P7)
+		req->flags |= VNIC_RSS_CFG_REQ_FLAGS_IPSEC_HASH_TYPE_CFG_SUPPORT;
 
 	if (bp->rss_hash_delta) {
 		req->hash_type = cpu_to_le32(bp->rss_hash_delta);
@@ -6803,6 +6801,7 @@ static int bnxt_hwrm_vnic_set_rss_p5(struct bnxt *bp,
 	if (!set_rss)
 		return hwrm_req_send(bp, req);
 
+	bnxt_fill_hw_rss_tbl_p5(bp, vnic);
 	__bnxt_hwrm_vnic_set_rss(bp, req, vnic);
 	ring_tbl_map = vnic->rss_table_dma_addr;
 	nr_ctxs = bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings);
@@ -10866,6 +10865,7 @@ static int __bnxt_setup_vnic(struct bnxt *bp, struct bnxt_vnic_info *vnic)
 	}
 
 skip_rss_ctx:
+	bnxt_fill_hw_rss_tbl(bp, vnic);
 	/* configure default vnic, ring grp */
 	rc = bnxt_hwrm_vnic_cfg(bp, vnic);
 	if (rc) {
Signature · 1 line
2.51.0
author maintainer
#2 lore raw

PATCH 2/5 bnxt_en: Determine and store default RX ring in vnic structure

In reply to Michael Chan

From: Shravya KN <[email protected]>

Each VNIC has a default RX ring. The purpose of the default RX ring is to provide a destination for any packets that cannot be parsed by the RSS logic. Up until now, the default RX ring is always Ring 0.

We neglected to take care of this default RX ring when adding the queue restart feature. If ring 0 (default ring) is re-started, it may now have a new FW ring ID after freeing the old one and allocating a new one. The VNIC now may have a stale default ring and it may generate an internal exception. This exception may appear in dmesg:

FW reported unknown error type 10

The best way to resolve this issue is to use a more appropriate ring for the default ring instead of always ring 0. Ring 0 may not even be in the RSS table, especially on a new RSS context.

This patch adds the logic to determine and store the proper default RX ring for a VNIC. For an RSS VNIC, the default ring is the lowest ring number in the RSS table. The next patch will add proper logic to update the VNIC if the default ring changes after queue restart.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 33 +++++++++++++++--------
 drivers/net/ethernet/broadcom/bnxt/bnxt.h |  1 +
 2 files changed, 23 insertions(+), 11 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +22 -11
index 48f4e38ad78d..ff098d959741 100644
@@ -6704,25 +6704,36 @@ int bnxt_get_nr_rss_ctxs(struct bnxt *bp, int rx_rings)
 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;
+	u16 i, j, min_j = bp->rx_nr_rings - 1;
 
 	if (!vnic->rss_table)
-		return;
+		goto skip_rss_tbl;
 
 	/* 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];
+		min_j = min(j, min_j);
 		vnic->rss_table[i] = cpu_to_le16(vnic->fw_grp_ids[j]);
 	}
+
+skip_rss_tbl:
+	if (vnic->rss_table && !no_rss)
+		vnic->default_rx_ring = min_j;
+	else if (vnic->flags & BNXT_VNIC_RFS_FLAG)
+		vnic->default_rx_ring = vnic->vnic_id - 1;
+	else if ((vnic->vnic_id == 1) && BNXT_CHIP_TYPE_NITRO_A0(bp))
+		vnic->default_rx_ring = bp->rx_nr_rings - 1;
+	else
+		vnic->default_rx_ring = 0;
 }
 
 static void bnxt_fill_hw_rss_tbl_p5(struct bnxt *bp,
 				    struct bnxt_vnic_info *vnic)
 {
+	u16 tbl_size, i, min_j = bp->rx_nr_rings - 1;
 	__le16 *ring_tbl = vnic->rss_table;
 	struct bnxt_rx_ring_info *rxr;
-	u16 tbl_size, i;
 
 	tbl_size = bnxt_get_rxfh_indir_size(bp->dev);
 
@@ -6735,6 +6746,7 @@ static void bnxt_fill_hw_rss_tbl_p5(struct bnxt *bp,
 			j = ethtool_rxfh_context_indir(vnic->rss_ctx)[i];
 		else
 			j = bp->rss_indir_tbl[i];
+		min_j = min(j, min_j);
 		rxr = &bp->rx_ring[j];
 
 		ring_id = rxr->rx_ring_struct.fw_ring_id;
@@ -6742,6 +6754,7 @@ static void bnxt_fill_hw_rss_tbl_p5(struct bnxt *bp,
 		ring_id = bnxt_cp_ring_for_rx(bp, rxr);
 		*ring_tbl++ = cpu_to_le16(ring_id);
 	}
+	vnic->default_rx_ring = min_j;
 }
 
 static void
@@ -6938,8 +6951,9 @@ int bnxt_hwrm_vnic_cfg(struct bnxt *bp, struct bnxt_vnic_info *vnic)
 		return rc;
 
 	if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
-		struct bnxt_rx_ring_info *rxr = &bp->rx_ring[0];
+		struct bnxt_rx_ring_info *rxr;
 
+		rxr = &bp->rx_ring[vnic->default_rx_ring];
 		req->default_rx_ring_id =
 			cpu_to_le16(rxr->rx_ring_struct.fw_ring_id);
 		req->default_cmpl_ring_id =
@@ -6972,13 +6986,7 @@ int bnxt_hwrm_vnic_cfg(struct bnxt *bp, struct bnxt_vnic_info *vnic)
 		req->cos_rule = cpu_to_le16(0xffff);
 	}
 
-	if (vnic->flags & BNXT_VNIC_RSS_FLAG)
-		ring = 0;
-	else if (vnic->flags & BNXT_VNIC_RFS_FLAG)
-		ring = vnic->vnic_id - 1;
-	else if ((vnic->vnic_id == 1) && BNXT_CHIP_TYPE_NITRO_A0(bp))
-		ring = bp->rx_nr_rings - 1;
-
+	ring = vnic->default_rx_ring;
 	grp_idx = bp->rx_ring[ring].bnapi->index;
 	req->dflt_ring_grp = cpu_to_le16(bp->grp_info[grp_idx].fw_grp_id);
 	req->lb_rule = cpu_to_le16(0xffff);
@@ -11171,6 +11179,9 @@ static int bnxt_setup_nitroa0_vnic(struct bnxt *bp)
 		return rc;
 	}
 
+	/* Setup the proper default RX ring */
+	bnxt_fill_hw_rss_tbl(bp, vnic);
+
 	rc = bnxt_hwrm_vnic_cfg(bp, vnic);
 	if (rc) {
 		netdev_err(bp->dev, "Cannot allocate special vnic for NS2 A0: %x\n",
drivers/net/ethernet/broadcom/bnxt/bnxt.h +1 -0
index 6335dfc14c98..44ea8fc7b766 100644
@@ -1334,6 +1334,7 @@ struct bnxt_vnic_info {
 #define BNXT_VNIC_RSSCTX_FLAG		0x40
 	struct ethtool_rxfh_context *rss_ctx;
 	u32		vnic_id;
+	u16		default_rx_ring;
 };
 
 struct bnxt_rss_ctx {
Signature · 1 line
2.51.0
author maintainer
#3 lore raw

PATCH 3/5 bnxt_en: Refresh VNIC default ring on queue restart if needed

In reply to Michael Chan

From: Shravya KN <[email protected]>

When a queue is restarted, refresh VNIC_CFG for all VNICs whose default RX ring is the restarted ring. This will eliminate this possible FW warning caused by a stale default ring in the VNIC:

FW reported unknown error type 10

Fixes 5ac066b7b062 ("bnxt_en: Fix queue start to update vnic RSS table")
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +5 -0
index ff098d959741..97cae19643c3 100644
@@ -11098,6 +11098,11 @@ static int bnxt_set_vnic_mru_p5(struct bnxt *bp, struct bnxt_vnic_info *vnic,
 				   vnic->vnic_id, rc);
 			return rc;
 		}
+		if (rxr_id == vnic->default_rx_ring) {
+			rc = bnxt_hwrm_vnic_cfg(bp, vnic);
+			if (rc)
+				return rc;
+		}
 	}
 	vnic->mru = mru;
 	bnxt_hwrm_vnic_update(bp, vnic,
Signature · 1 line
2.51.0
author maintainer
#4 lore raw

PATCH 4/5 bnxt_en: Disable EOP for TPA on all chips to prevent data corruption

In reply to Michael Chan

EOP (End of frame padding) on the AGG ring may cause overlapping of zero padding at the end of one segment with the next segment's data. If Relaxed Ordering (RO) is enabled, the zero padding may overwrite valid data in the next segment and corrupt the data. Older chips (P5 and older) do not automatically disable RO when EOP is enabled. On some ARM systems, data corruption was reported on 57508 (P5) chips with RO enabled.

Always disable EOP on all chips on the AGG rings when TPA is enabled to fix the data corruption.

Fixes bfcd8d791ec1 ("bnxt_en: Add fast path logic for TPA on 57500 chips.")
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +7 -4
index 97cae19643c3..a0ea8c994b32 100644
@@ -4611,11 +4611,14 @@ static void bnxt_init_one_rx_agg_ring_rxbd(struct bnxt *bp,
 		type = ((u32)rxr->rx_page_size << RX_BD_LEN_SHIFT) |
 			RX_BD_TYPE_RX_AGG_BD;
 
-		/* On P7, setting EOP will cause the chip to disable
-		 * Relaxed Ordering (RO) for TPA data.  Disable EOP for
-		 * potentially higher performance with RO.
+		/* Disable EOP if TPA is enabled to prevent overlapping zero
+		 * padding with the next segment's data.  On P7_PLUS, EOP will
+		 * automatically disable Relaxed Ordering (RO) to prevent
+		 * potential data corruption (and may degrade performance).  On
+		 * older chips, RO will not be automatically disabled and may
+		 * cause corruption.
 		 */
-		if (BNXT_CHIP_P5_AND_MINUS(bp) || !(bp->flags & BNXT_FLAG_TPA))
+		if (!(bp->flags & BNXT_FLAG_TPA))
 			type |= RX_BD_FLAGS_AGG_EOP;
 
 		bnxt_init_rxbd_pages(ring, type);
Signature · 1 line
2.51.0
author maintainer
#5 lore raw

PATCH 5/5 bnxt_en: Fix PTP PPS setting bug

In reply to Michael Chan

From: Keegan Freyhof <[email protected]>

The existing driver logic is always turning on PTP_CLK_REQ_PPS regardless of the "on" parameter passed to bnxt_ptp_enable(). During shutdown, PTP_CLK_REQ_PPS may be turned off and this bug will do the opposite and may trigger a PCIe PTM request TLP. On some systems this can trigger a PCIe AER.

Fix it by properly configuring PTP_CLK_REQ_PPS based on the "on" parameter.

Fixes 9e518f25802c ("bnxt_en: 1PPS functions to configure TSIO pins")
---
 drivers/net/ethernet/broadcom/bnxt/bnxt_ptp.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt_ptp.c +5 -2
index 5d41dc1bc782..b89293dfd2d1 100644
@@ -495,12 +495,15 @@ static int bnxt_ptp_enable(struct ptp_clock_info *ptp_info,
 		return rc;
 	case PTP_CLK_REQ_PPS:
 		/* Configure PHC PPS IN */
-		rc = bnxt_ptp_cfg_pin(bp, 0, BNXT_PPS_PIN_PPS_IN);
+		pin_id = 0;
+		if (!on)
+			break;
+		rc = bnxt_ptp_cfg_pin(bp, pin_id, BNXT_PPS_PIN_PPS_IN);
 		if (rc)
 			return rc;
 		rc = bnxt_ptp_cfg_event(bp, BNXT_PPS_EVENT_INTERNAL);
 		if (!rc)
-			ptp->pps_info.pins[0].event = BNXT_PPS_EVENT_INTERNAL;
+			ptp->pps_info.pins[pin_id].event = BNXT_PPS_EVENT_INTERNAL;
 		return rc;
 	default:
 		netdev_err(ptp->bp->dev, "Unrecognized PIN function\n");
Signature · 1 line
2.51.0
#6 lore raw
In reply to Michael Chan
This series was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:

Hello:

This series was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:

Quoted 8 lines · On Fri, 31 Jul 2026 12:09:32 -0700 you wrote:

This series include 3 bug fixes:

1. queue start bug fix on the VNIC's default ring. 2 refactoring patches preceed the actual bug fix.
2. Bug fix for TPA data corruption seen on some ARM systems.
3. PTP PPS setting bug fix.

[...]

Here is the summary with links:
- [net,1/5] bnxt_en: Move RSS table fill outside __bnxt_hwrm_vnic_set_rss()

    https://git.kernel.org/netdev/net/c/1d0fc6c7ea49
  - [net,2/5] bnxt_en: Determine and store default RX ring in vnic structure
    https://git.kernel.org/netdev/net/c/53f01cd594e2
  - [net,3/5] bnxt_en: Refresh VNIC default ring on queue restart if needed
    https://git.kernel.org/netdev/net/c/0b137529a899
  - [net,4/5] bnxt_en: Disable EOP for TPA on all chips to prevent data corruption
    https://git.kernel.org/netdev/net/c/c3faf548a00f
  - [net,5/5] bnxt_en: Fix PTP PPS setting bug
    https://git.kernel.org/netdev/net/c/80eaf88efec3

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
maintainer
#7 lore raw
In reply to Michael Chan
Quoted 6 lines · On Fri, 31 Jul 2026 12:09:32 -0700 Michael Chan wrote:

This series include 3 bug fixes:

1. queue start bug fix on the VNIC's default ring. 2 refactoring patches preceed the actual bug fix.
2. Bug fix for TPA data corruption seen on some ARM systems.
3. PTP PPS setting bug fix.

Just to avoid any further misunderstandings - you can have 15 outstanding patches per tree, so 15 for net and 15 for net-next. There can be multiple series for each tree, as long as they don't conflict.