From: Michael Chan <michael.chan@broadcom.com>
To: davem@davemloft.net
Cc: netdev@vger.kernel.org, edumazet@google.com, kuba@kernel.org,
pabeni@redhat.com, andrew.gospodarek@broadcom.com,
Pavan Chebbi <pavan.chebbi@broadcom.com>
Subject: [PATCH net 3/5] bnxt_en: Fix RSS table entries calculation for P5_PLUS chips
Date: Wed, 17 Jan 2024 15:45:13 -0800 [thread overview]
Message-ID: <20240117234515.226944-4-michael.chan@broadcom.com> (raw)
In-Reply-To: <20240117234515.226944-1-michael.chan@broadcom.com>
[-- Attachment #1: Type: text/plain, Size: 3640 bytes --]
The existing formula used in the driver to calculate the number of RSS
table entries is to round up the number of RX rings to the next integer
multiples of 64 (e.g. 64, 128, 192, ..). This is incorrect. The valid
values supported by the chip are 64, 128, 256, 512 only (power of 2
starting from 64). When the number of RX rings is greater than 128, the
entry size will likely be wrong. Firmware will round down the invalid
value (e.g. 192 rounded down to 128) provided by the driver, causing some
RSS rings to not receive any packets.
We already have an existing function bnxt_calc_nr_ring_pages() to
do this calculation. Use it in bnxt_get_nr_rss_ctxs() to calculate the
number of RSS contexts correctly for P5_PLUS chips.
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Fixes: 7b3af4f75b81 ("bnxt_en: Add RSS support for 57500 chips.")
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 17 ++++++++++++-----
.../net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 3 ++-
2 files changed, 14 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 9fdc90bfce38..3d090d4403df 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -5935,8 +5935,12 @@ static u16 bnxt_get_max_rss_ring(struct bnxt *bp)
int bnxt_get_nr_rss_ctxs(struct bnxt *bp, int rx_rings)
{
- if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
- return DIV_ROUND_UP(rx_rings, BNXT_RSS_TABLE_ENTRIES_P5);
+ if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
+ if (!rx_rings)
+ return 0;
+ return bnxt_calc_nr_ring_pages(rx_rings - 1,
+ BNXT_RSS_TABLE_ENTRIES_P5);
+ }
if (BNXT_CHIP_TYPE_NITRO_A0(bp))
return 2;
return 1;
@@ -7001,10 +7005,11 @@ __bnxt_hwrm_reserve_pf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
req->num_rx_rings = cpu_to_le16(rx_rings);
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
+ u16 rss_ctx = bnxt_get_nr_rss_ctxs(bp, ring_grps);
+
req->num_cmpl_rings = cpu_to_le16(tx_rings + ring_grps);
req->num_msix = cpu_to_le16(cp_rings);
- req->num_rsscos_ctxs =
- cpu_to_le16(DIV_ROUND_UP(ring_grps, 64));
+ 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);
@@ -7051,8 +7056,10 @@ __bnxt_hwrm_reserve_vf_rings(struct bnxt *bp, int tx_rings, int rx_rings,
req->num_tx_rings = cpu_to_le16(tx_rings);
req->num_rx_rings = cpu_to_le16(rx_rings);
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS) {
+ u16 rss_ctx = bnxt_get_nr_rss_ctxs(bp, ring_grps);
+
req->num_cmpl_rings = cpu_to_le16(tx_rings + ring_grps);
- req->num_rsscos_ctxs = cpu_to_le16(DIV_ROUND_UP(ring_grps, 64));
+ 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);
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
index 27b983c0a8a9..1f6e0cd84f2e 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
@@ -1574,7 +1574,8 @@ u32 bnxt_get_rxfh_indir_size(struct net_device *dev)
struct bnxt *bp = netdev_priv(dev);
if (bp->flags & BNXT_FLAG_CHIP_P5_PLUS)
- return ALIGN(bp->rx_nr_rings, BNXT_RSS_TABLE_ENTRIES_P5);
+ return bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings) *
+ BNXT_RSS_TABLE_ENTRIES_P5;
return HW_HASH_INDEX_SIZE;
}
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
next prev parent reply other threads:[~2024-01-17 23:45 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-17 23:45 [PATCH net 0/5] bnxt_en: Bug fixes Michael Chan
2024-01-17 23:45 ` [PATCH net 1/5] bnxt_en: Wait for FLR to complete during probe Michael Chan
2024-01-17 23:45 ` [PATCH net 2/5] bnxt_en: Fix memory leak in bnxt_hwrm_get_rings() Michael Chan
2024-01-17 23:45 ` Michael Chan [this message]
2024-01-17 23:45 ` [PATCH net 4/5] bnxt_en: Prevent kernel warning when running offline self test Michael Chan
2024-01-17 23:45 ` [PATCH net 5/5] bnxt_en: Fix possible crash after creating sw mqprio TCs Michael Chan
2024-01-19 2:10 ` [PATCH net 0/5] bnxt_en: Bug fixes patchwork-bot+netdevbpf
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240117234515.226944-4-michael.chan@broadcom.com \
--to=michael.chan@broadcom.com \
--cc=andrew.gospodarek@broadcom.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=kuba@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=pavan.chebbi@broadcom.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is 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.