[PATCH net v2 2/2] bnge: fix RSS table sizing for large RX ring counts
flat view
WARM3d
REVIEWED: 1 (1M)
From: Vikas Gupta <vikas.gupta@broadcom.com>
Date: 2026-10-15 12:19:11
Also in:
lkml
Subsystem:
broadcom bng_en 800 gigabit ethernet driver, networking drivers, the rest · Maintainers:
Vikas Gupta, Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds
Revision v2 of 2 in this series; 1 review trailer (1 from subsystem maintainers).
Revisions (2)
- v1 [diff vs current]
- v2 current
From: Dharmender Garg <redacted>
bnge_adjust_pow_two() adds one block before rounding up to a power of
two only when the quotient is 0 or 1. A larger quotient that is
already a power of two is returned unchanged, so 129-192 RX rings
get a 128-entry RSS indirection table and 257-320 rings a 256-entry
one. The rings above the table size never receive default RSS traffic.
bnxt_calc_nr_ring_pages() always adds one, which is correct.
Make bnge_adjust_pow_two() always add one block before rounding up.
The same helper also sized the ring pages; to keep the ring size
calculation unchanged, move that to a separate bnge_calc_nr_ring_pages()
helper that retains the original rounding.
Fixes: 627c67f038d2 ("bng_en: Add resource management support")
Signed-off-by: Dharmender Garg <redacted>
Reviewed-by: Vikas Gupta <vikas.gupta@broadcom.com>
---
drivers/net/ethernet/broadcom/bnge/bnge_netdev.c | 10 +++++-----
drivers/net/ethernet/broadcom/bnge/bnge_resc.h | 11 +++++++++++
2 files changed, 16 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
index a4288f0258f8..ecba8aff3052 100644
--- a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
+++ b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c@@ -3339,8 +3339,8 @@ void bnge_set_ring_params(struct bnge_dev *bd) } agg_ring_size = ring_size * agg_factor; - bn->rx_agg_nr_pages = bnge_adjust_pow_two(agg_ring_size, - RX_DESC_CNT); + bn->rx_agg_nr_pages = bnge_calc_nr_ring_pages(agg_ring_size, + RX_DESC_CNT); if (bn->rx_agg_nr_pages > MAX_RX_AGG_PAGES) { u32 tmp = agg_ring_size;
@@ -3363,11 +3363,11 @@ void bnge_set_ring_params(struct bnge_dev *bd) bn->rx_buf_use_size = rx_size; bn->rx_buf_size = rx_space; - bn->rx_nr_pages = bnge_adjust_pow_two(ring_size, RX_DESC_CNT); + bn->rx_nr_pages = bnge_calc_nr_ring_pages(ring_size, RX_DESC_CNT); bn->rx_ring_mask = (bn->rx_nr_pages * RX_DESC_CNT) - 1; ring_size = bn->tx_ring_size; - bn->tx_nr_pages = bnge_adjust_pow_two(ring_size, TX_DESC_CNT); + bn->tx_nr_pages = bnge_calc_nr_ring_pages(ring_size, TX_DESC_CNT); bn->tx_ring_mask = (bn->tx_nr_pages * TX_DESC_CNT) - 1; max_rx_cmpl = bn->rx_ring_size;
@@ -3377,7 +3377,7 @@ void bnge_set_ring_params(struct bnge_dev *bd) ring_size = max_rx_cmpl * 2 + agg_ring_size + bn->tx_ring_size; bn->cp_ring_size = ring_size; - bn->cp_nr_pages = bnge_adjust_pow_two(ring_size, CP_DESC_CNT); + bn->cp_nr_pages = bnge_calc_nr_ring_pages(ring_size, CP_DESC_CNT); if (bn->cp_nr_pages > MAX_CP_PAGES) { bn->cp_nr_pages = MAX_CP_PAGES; bn->cp_ring_size = MAX_CP_PAGES * CP_DESC_CNT - 1;
diff --git a/drivers/net/ethernet/broadcom/bnge/bnge_resc.h b/drivers/net/ethernet/broadcom/bnge/bnge_resc.h
index b62a634669f6..58462524939f 100644
--- a/drivers/net/ethernet/broadcom/bnge/bnge_resc.h
+++ b/drivers/net/ethernet/broadcom/bnge/bnge_resc.h@@ -81,6 +81,17 @@ bnge_adjust_pow_two(u32 total_ent, u16 ent_per_blk) { u32 blks = total_ent / ent_per_blk; + if (!blks) + return 1; + + return roundup_pow_of_two(blks + 1); +} + +static inline u32 +bnge_calc_nr_ring_pages(u32 total_ent, u16 ent_per_blk) +{ + u32 blks = total_ent / ent_per_blk; + if (blks == 0 || blks == 1) return ++blks;
--
2.52.0