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, pavan.chebbi@broadcom.com,
andrew.gospodarek@broadcom.com, richardcochran@gmail.com
Subject: [PATCH net-next 03/10] bnxt_en: Allow some TX packets to be unprocessed in NAPI
Date: Wed, 26 Jun 2024 09:43:00 -0700 [thread overview]
Message-ID: <20240626164307.219568-4-michael.chan@broadcom.com> (raw)
In-Reply-To: <20240626164307.219568-1-michael.chan@broadcom.com>
[-- Attachment #1: Type: text/plain, Size: 3872 bytes --]
The driver's current logic will always free all the TX SKBs up to
txr->tx_hw_cons within NAPI. In the next patches, we'll be adding
logic to handle TX timestamp completion and we may need to hold
some remaining TX SKBs if we don't have the timestamp completions
yet.
Modify __bnxt_poll_work_done() to clear each event bit separately to
allow bnapi->tx_int() to decide whether to clear BNXT_TX_CMP_EVENT or
not. bnapi->tx_int() will not clear BNXT_TX_CMP_EVENT if some TX
SKBs are held waiting for TX timestamps. Note that legacy chips will
never hold any SKBs this way. The SKB is always deferred to the PTP
worker slow path to retrieve the timestamp from firmware. On the new
P7 chips, the timestamp is returned by the hardware directly and we
can retrieve it directly from NAPI.
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 21 +++++++++++++++------
1 file changed, 15 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 996d9fe80495..784787a09dba 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -772,7 +772,8 @@ static netdev_tx_t bnxt_start_xmit(struct sk_buff *skb, struct net_device *dev)
return NETDEV_TX_OK;
}
-static void __bnxt_tx_int(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
+/* Returns true if some remaining TX packets not processed. */
+static bool __bnxt_tx_int(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
int budget)
{
struct netdev_queue *txq = netdev_get_tx_queue(bp->dev, txr->txq_index);
@@ -795,7 +796,7 @@ static void __bnxt_tx_int(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
if (unlikely(!skb)) {
bnxt_sched_reset_txr(bp, txr, cons);
- return;
+ return false;
}
tx_pkts++;
@@ -844,18 +845,22 @@ static void __bnxt_tx_int(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
__netif_txq_completed_wake(txq, tx_pkts, tx_bytes,
bnxt_tx_avail(bp, txr), bp->tx_wake_thresh,
READ_ONCE(txr->dev_state) == BNXT_DEV_STATE_CLOSING);
+
+ return false;
}
static void bnxt_tx_int(struct bnxt *bp, struct bnxt_napi *bnapi, int budget)
{
struct bnxt_tx_ring_info *txr;
+ bool more = false;
int i;
bnxt_for_each_napi_tx(i, bnapi, txr) {
if (txr->tx_hw_cons != RING_TX(bp, txr->tx_cons))
- __bnxt_tx_int(bp, txr, budget);
+ more |= __bnxt_tx_int(bp, txr, budget);
}
- bnapi->events &= ~BNXT_TX_CMP_EVENT;
+ if (!more)
+ bnapi->events &= ~BNXT_TX_CMP_EVENT;
}
static struct page *__bnxt_alloc_rx_page(struct bnxt *bp, dma_addr_t *mapping,
@@ -2952,8 +2957,10 @@ static int __bnxt_poll_work(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
}
}
- if (event & BNXT_REDIRECT_EVENT)
+ if (event & BNXT_REDIRECT_EVENT) {
xdp_do_flush();
+ event &= ~BNXT_REDIRECT_EVENT;
+ }
if (event & BNXT_TX_EVENT) {
struct bnxt_tx_ring_info *txr = bnapi->tx_ring[0];
@@ -2963,6 +2970,7 @@ static int __bnxt_poll_work(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
wmb();
bnxt_db_write_relaxed(bp, &txr->tx_db, prod);
+ event &= ~BNXT_TX_EVENT;
}
cpr->cp_raw_cons = raw_cons;
@@ -2980,13 +2988,14 @@ static void __bnxt_poll_work_done(struct bnxt *bp, struct bnxt_napi *bnapi,
struct bnxt_rx_ring_info *rxr = bnapi->rx_ring;
bnxt_db_write(bp, &rxr->rx_db, rxr->rx_prod);
+ bnapi->events &= ~BNXT_RX_EVENT;
}
if (bnapi->events & BNXT_AGG_EVENT) {
struct bnxt_rx_ring_info *rxr = bnapi->rx_ring;
bnxt_db_write(bp, &rxr->rx_agg_db, rxr->rx_agg_prod);
+ bnapi->events &= ~BNXT_AGG_EVENT;
}
- bnapi->events &= BNXT_TX_CMP_EVENT;
}
static int bnxt_poll_work(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
--
2.30.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
next prev parent reply other threads:[~2024-06-26 16:43 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-06-26 16:42 [PATCH net-next 00/10] bnxt_en: PTP updates for net-next Michael Chan
2024-06-26 16:42 ` [PATCH net-next 01/10] bnxt_en: Add new TX timestamp completion definitions Michael Chan
2024-06-27 9:00 ` Przemek Kitszel
2024-06-27 15:54 ` Michael Chan
2024-06-26 16:42 ` [PATCH net-next 02/10] bnxt_en: Add is_ts_pkt field to struct bnxt_sw_tx_bd Michael Chan
2024-06-28 0:08 ` Jakub Kicinski
2024-06-28 0:39 ` Michael Chan
2024-06-26 16:43 ` Michael Chan [this message]
2024-06-26 16:43 ` [PATCH net-next 04/10] bnxt_en: Add TX timestamp completion logic Michael Chan
2024-06-26 16:43 ` [PATCH net-next 05/10] bnxt_en: Add BCM5760X specific PHC registers mapping Michael Chan
2024-06-26 16:43 ` [PATCH net-next 06/10] bnxt_en: Refactor all PTP TX timestamp fields into a struct Michael Chan
2024-06-26 16:43 ` [PATCH net-next 07/10] bnxt_en: Remove an impossible condition check for PTP TX pending SKB Michael Chan
2024-06-26 16:43 ` [PATCH net-next 08/10] bnxt_en: Let bnxt_stamp_tx_skb() return error code Michael Chan
2024-06-26 16:43 ` [PATCH net-next 09/10] bnxt_en: Increase the max total outstanding PTP TX packets to 4 Michael Chan
2024-06-28 17:03 ` Simon Horman
2024-06-28 17:05 ` Simon Horman
2024-06-28 17:37 ` Michael Chan
2024-06-28 18:13 ` Simon Horman
2024-06-26 16:43 ` [PATCH net-next 10/10] bnxt_en: Remove atomic operations on ptp->tx_avail Michael Chan
2024-06-27 9:40 ` Przemek Kitszel
2024-06-27 15:34 ` Pavan Chebbi
2024-06-27 15:43 ` Michael Chan
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=20240626164307.219568-4-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 \
--cc=richardcochran@gmail.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.