From: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
To: <dev@dpdk.org>
Cc: <matan@nvidia.com>, <rasland@nvidia.com>, <orika@nvidia.com>,
<dsosnowski@nvidia.com>, Tim Martin <timothym@nvidia.com>,
<stable@dpdk.org>
Subject: [PATCH v2 3/4] net/mlx5: fix Tx tracing to use single clock source
Date: Mon, 14 Oct 2024 11:04:33 +0300 [thread overview]
Message-ID: <20241014080434.1211629-4-viacheslavo@nvidia.com> (raw)
In-Reply-To: <20241014080434.1211629-1-viacheslavo@nvidia.com>
From: Tim Martin <timothym@nvidia.com>
The prior commit introduced tracing for mlx5, but there is a mixture of
two unrelated clocks used: the TSC for host work submission timestamps
and the NIC HW clock for CQE completion times. It is necessary to have
timestamps from a single common clock, and the NIC HW clock is the
better choice since it can be used with externally synchronized clocks.
This patch adds the NIC HW clock as an additional logged parameter for
trace_tx_entry, trace_tx_exit, and trace_tx_wqe. The included trace
analysis python script is also updated to use the new clock when
it is available.
Fixes: a1e910f5b8d4 ("net/mlx5: introduce tracepoints")
Fixes: 9725191a7e14 ("net/mlx5: add Tx datapath trace analyzing script")
Cc: stable@dpdk.org
Signed-off-by: Tim Martin <timothym@nvidia.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
---
drivers/net/mlx5/mlx5_trace.h | 9 ++++++---
drivers/net/mlx5/mlx5_tx.h | 21 +++++++++++++++++----
drivers/net/mlx5/tools/mlx5_trace.py | 12 +++++++++---
3 files changed, 32 insertions(+), 10 deletions(-)
diff --git a/drivers/net/mlx5/mlx5_trace.h b/drivers/net/mlx5/mlx5_trace.h
index a8f0b372c8..4fc3584acc 100644
--- a/drivers/net/mlx5/mlx5_trace.h
+++ b/drivers/net/mlx5/mlx5_trace.h
@@ -22,21 +22,24 @@ extern "C" {
/* TX burst subroutines trace points. */
RTE_TRACE_POINT_FP(
rte_pmd_mlx5_trace_tx_entry,
- RTE_TRACE_POINT_ARGS(uint16_t port_id, uint16_t queue_id),
+ RTE_TRACE_POINT_ARGS(uint64_t real_time, uint16_t port_id, uint16_t queue_id),
+ rte_trace_point_emit_u64(real_time);
rte_trace_point_emit_u16(port_id);
rte_trace_point_emit_u16(queue_id);
)
RTE_TRACE_POINT_FP(
rte_pmd_mlx5_trace_tx_exit,
- RTE_TRACE_POINT_ARGS(uint16_t nb_sent, uint16_t nb_req),
+ RTE_TRACE_POINT_ARGS(uint64_t real_time, uint16_t nb_sent, uint16_t nb_req),
+ rte_trace_point_emit_u64(real_time);
rte_trace_point_emit_u16(nb_sent);
rte_trace_point_emit_u16(nb_req);
)
RTE_TRACE_POINT_FP(
rte_pmd_mlx5_trace_tx_wqe,
- RTE_TRACE_POINT_ARGS(uint32_t opcode),
+ RTE_TRACE_POINT_ARGS(uint64_t real_time, uint32_t opcode),
+ rte_trace_point_emit_u64(real_time);
rte_trace_point_emit_u32(opcode);
)
diff --git a/drivers/net/mlx5/mlx5_tx.h b/drivers/net/mlx5/mlx5_tx.h
index 587e6a9f7d..55568c41b1 100644
--- a/drivers/net/mlx5/mlx5_tx.h
+++ b/drivers/net/mlx5/mlx5_tx.h
@@ -404,6 +404,14 @@ static __rte_always_inline uint64_t mlx5_read_pcibar_clock(struct rte_eth_dev *d
return 0;
}
+static __rte_always_inline uint64_t mlx5_read_pcibar_clock_from_txq(struct mlx5_txq_data *txq)
+{
+ struct mlx5_txq_ctrl *txq_ctrl = container_of(txq, struct mlx5_txq_ctrl, txq);
+ struct rte_eth_dev *dev = ETH_DEV(txq_ctrl->priv);
+
+ return mlx5_read_pcibar_clock(dev);
+}
+
/**
* Set Software Parser flags and offsets in Ethernet Segment of WQE.
* Flags must be preliminary initialized to zero.
@@ -841,6 +849,7 @@ mlx5_tx_cseg_init(struct mlx5_txq_data *__rte_restrict txq,
unsigned int olx)
{
struct mlx5_wqe_cseg *__rte_restrict cs = &wqe->cseg;
+ uint64_t real_time;
/* For legacy MPW replace the EMPW by TSO with modifier. */
if (MLX5_TXOFF_CONFIG(MPW) && opcode == MLX5_OPCODE_ENHANCED_MPSW)
@@ -854,9 +863,12 @@ mlx5_tx_cseg_init(struct mlx5_txq_data *__rte_restrict txq,
cs->flags = RTE_BE32(MLX5_COMP_ONLY_FIRST_ERR <<
MLX5_COMP_MODE_OFFSET);
cs->misc = RTE_BE32(0);
- if (__rte_trace_point_fp_is_enabled())
- rte_pmd_mlx5_trace_tx_entry(txq->port_id, txq->idx);
- rte_pmd_mlx5_trace_tx_wqe((txq->wqe_ci << 8) | opcode);
+ if (__rte_trace_point_fp_is_enabled()) {
+ real_time = mlx5_read_pcibar_clock_from_txq(txq);
+ if (!loc->pkts_sent)
+ rte_pmd_mlx5_trace_tx_entry(real_time, txq->port_id, txq->idx);
+ rte_pmd_mlx5_trace_tx_wqe(real_time, (txq->wqe_ci << 8) | opcode);
+ }
}
/**
@@ -3818,7 +3830,8 @@ mlx5_tx_burst_tmpl(struct mlx5_txq_data *__rte_restrict txq,
__mlx5_tx_free_mbuf(txq, pkts, loc.mbuf_free, olx);
/* Trace productive bursts only. */
if (__rte_trace_point_fp_is_enabled() && loc.pkts_sent)
- rte_pmd_mlx5_trace_tx_exit(loc.pkts_sent, pkts_n);
+ rte_pmd_mlx5_trace_tx_exit(mlx5_read_pcibar_clock_from_txq(txq),
+ loc.pkts_sent, pkts_n);
return loc.pkts_sent;
}
diff --git a/drivers/net/mlx5/tools/mlx5_trace.py b/drivers/net/mlx5/tools/mlx5_trace.py
index 67461520a9..5eb634a490 100755
--- a/drivers/net/mlx5/tools/mlx5_trace.py
+++ b/drivers/net/mlx5/tools/mlx5_trace.py
@@ -174,7 +174,9 @@ def do_tx_entry(msg, trace):
return
# allocate the new burst and append to the queue
burst = MlxBurst()
- burst.call_ts = msg.default_clock_snapshot.ns_from_origin
+ burst.call_ts = event["real_time"]
+ if burst.call_ts == 0:
+ burst.call_ts = msg.default_clock_snapshot.ns_from_origin
trace.tx_blst[cpu_id] = burst
pq_id = event["port_id"] << 16 | event["queue_id"]
queue = trace.tx_qlst.get(pq_id)
@@ -194,7 +196,9 @@ def do_tx_exit(msg, trace):
burst = trace.tx_blst.get(cpu_id)
if burst is None:
return
- burst.done_ts = msg.default_clock_snapshot.ns_from_origin
+ burst.done_ts = event["real_time"]
+ if burst.done_ts == 0:
+ burst.done_ts = msg.default_clock_snapshot.ns_from_origin
burst.req = event["nb_req"]
burst.done = event["nb_sent"]
trace.tx_blst.pop(cpu_id)
@@ -210,7 +214,9 @@ def do_tx_wqe(msg, trace):
wqe = MlxWqe()
wqe.wait_ts = trace.tx_wlst.get(cpu_id)
if wqe.wait_ts is None:
- wqe.wait_ts = msg.default_clock_snapshot.ns_from_origin
+ wqe.wait_ts = event["real_time"]
+ if wqe.wait_ts == 0:
+ wqe.wait_ts = msg.default_clock_snapshot.ns_from_origin
wqe.opcode = event["opcode"]
burst.wqes.append(wqe)
--
2.34.1
prev parent reply other threads:[~2024-10-14 8:05 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20241009114028.973284-1-viacheslavo@nvidia.com>
2024-10-09 11:40 ` [PATCH 1/4] net/mlx5/tools: fix trace dump multiple burst completions Viacheslav Ovsiienko
2024-10-09 13:08 ` Dariusz Sosnowski
2024-10-09 11:40 ` [PATCH 2/4] net/mlx5: fix real time counter reading from PCI BAR Viacheslav Ovsiienko
2024-10-09 11:40 ` [PATCH 3/4] net/mlx5: fix Tx tracing to use single clock source Viacheslav Ovsiienko
[not found] ` <20241014080434.1211629-1-viacheslavo@nvidia.com>
2024-10-14 8:04 ` [PATCH v2 1/4] net/mlx5/tools: fix trace dump multiple burst completions Viacheslav Ovsiienko
2024-10-14 8:04 ` [PATCH v2 2/4] net/mlx5: fix real time counter reading from PCI BAR Viacheslav Ovsiienko
2024-10-14 8:04 ` Viacheslav Ovsiienko [this message]
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=20241014080434.1211629-4-viacheslavo@nvidia.com \
--to=viacheslavo@nvidia.com \
--cc=dev@dpdk.org \
--cc=dsosnowski@nvidia.com \
--cc=matan@nvidia.com \
--cc=orika@nvidia.com \
--cc=rasland@nvidia.com \
--cc=stable@dpdk.org \
--cc=timothym@nvidia.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).