From: "Morten Brørup" <mb@smartsharesystems.com>
To: dev@dpdk.org
Cc: "Morten Brørup" <mb@smartsharesystems.com>
Subject: [TEST v10] ethdev: support single queue per port
Date: Wed, 6 Nov 2024 22:02:16 +0000 [thread overview]
Message-ID: <20241106220216.1505231-1-mb@smartsharesystems.com> (raw)
In-Reply-To: <20241024102253.785037-1-mb@smartsharesystems.com>
Configuring one queue per port (#define RTE_MAX_QUEUES_PER_PORT 1) fails
compilation with e.g.:
../drivers/net/bnxt/bnxt_rxq.c: In function 'bnxt_rx_queue_stop':
../drivers/net/bnxt/bnxt_rxq.c:587:34: error: array subscript 1 is above array bounds of 'uint8_t[1]' {aka 'unsigned char[1]'} [-Werror=array-bounds=]
587 | dev->data->rx_queue_state[q_id] = RTE_ETH_QUEUE_STATE_STOPPED;
| ~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~
In file included from ../drivers/net/bnxt/bnxt.h:16,
from ../drivers/net/bnxt/bnxt_rxq.c:10:
../lib/ethdev/ethdev_driver.h:168:17: note: while referencing 'rx_queue_state'
168 | uint8_t rx_queue_state[RTE_MAX_QUEUES_PER_PORT];
| ^~~~~~~~~~~~~~
Use the CI to test my ideas to fix this.
Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
---
v10:
* Rebased to main.
v9:
* Moved __rte_assume() in mana driver, to test if it works outside loop.
v8:
* Added __rte_unreachable() macro.
* GCC version 13 is required for __attribute__((assume())).
Provide alternative implementation, using __rte_unreachable(), for
older GCC versions.
* Added MSVC implementations of the two new macros.
v7:
* Introduce __rte_assume() in rte_common.h, and use this instead.
v6:
* Use __attribute__((assume(i < RTE_MAX_QUEUES_PER_PORT))) where the
indexing warning occurs, instead of pragma GCC optimize("no-peel-loops")
for the entire function.
v5:
* Wrap GCC optimizer pragmas in if defined(RTE_TOOLCHAIN_GCC).
v4:
* Workaound GCC optimizer incorrectly throwing a warning in these network
drivers:
* bnxt
* e1000
* failsafe
* hns3
v3:
* Fix net/ixgbe driver.
v2:
* Fix net/vmxnet3 driver.
---
config/rte_config.h | 4 ++--
drivers/net/bnxt/bnxt_ethdev.c | 2 ++
drivers/net/bnxt/bnxt_rxq.c | 1 +
drivers/net/e1000/igb_rxtx.c | 2 ++
drivers/net/failsafe/failsafe_ops.c | 10 ++++++--
drivers/net/hns3/hns3_rxtx.c | 2 ++
drivers/net/mana/tx.c | 1 +
drivers/net/vmxnet3/vmxnet3_ethdev.c | 34 +++++++++++++++++-----------
drivers/net/vmxnet3/vmxnet3_ethdev.h | 4 ++--
lib/eal/include/rte_common.h | 27 ++++++++++++++++++++++
10 files changed, 68 insertions(+), 19 deletions(-)
diff --git a/config/rte_config.h b/config/rte_config.h
index fd6f8a2f1a..924192c71c 100644
--- a/config/rte_config.h
+++ b/config/rte_config.h
@@ -65,8 +65,8 @@
#define RTE_MBUF_DEFAULT_MEMPOOL_OPS "ring_mp_mc"
/* ether defines */
-#define RTE_MAX_QUEUES_PER_PORT 1024
-#define RTE_ETHDEV_QUEUE_STAT_CNTRS 16 /* max 256 */
+#define RTE_MAX_QUEUES_PER_PORT 1 /* default 1024 */
+#define RTE_ETHDEV_QUEUE_STAT_CNTRS 1 /* max 256, default 16 */
#define RTE_ETHDEV_RXTX_CALLBACKS 1
#define RTE_MAX_MULTI_HOST_CTRLS 4
diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c
index 1f7c0d77d5..136e308437 100644
--- a/drivers/net/bnxt/bnxt_ethdev.c
+++ b/drivers/net/bnxt/bnxt_ethdev.c
@@ -910,6 +910,7 @@ static int bnxt_start_nic(struct bnxt *bp)
struct bnxt_rx_queue *rxq = bp->rx_queues[j];
if (!rxq->rx_deferred_start) {
+ __rte_assume(j < RTE_MAX_QUEUES_PER_PORT);
bp->eth_dev->data->rx_queue_state[j] =
RTE_ETH_QUEUE_STATE_STARTED;
rxq->rx_started = true;
@@ -930,6 +931,7 @@ static int bnxt_start_nic(struct bnxt *bp)
struct bnxt_tx_queue *txq = bp->tx_queues[j];
if (!txq->tx_deferred_start) {
+ __rte_assume(j < RTE_MAX_QUEUES_PER_PORT);
bp->eth_dev->data->tx_queue_state[j] =
RTE_ETH_QUEUE_STATE_STARTED;
txq->tx_started = true;
diff --git a/drivers/net/bnxt/bnxt_rxq.c b/drivers/net/bnxt/bnxt_rxq.c
index 1c25c57ca6..1651c26545 100644
--- a/drivers/net/bnxt/bnxt_rxq.c
+++ b/drivers/net/bnxt/bnxt_rxq.c
@@ -584,6 +584,7 @@ int bnxt_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id)
return -EINVAL;
}
+ __rte_assume(q_id < RTE_MAX_QUEUES_PER_PORT);
dev->data->rx_queue_state[q_id] = RTE_ETH_QUEUE_STATE_STOPPED;
rxq->rx_started = false;
PMD_DRV_LOG_LINE(DEBUG, "Rx queue stopped");
diff --git a/drivers/net/e1000/igb_rxtx.c b/drivers/net/e1000/igb_rxtx.c
index d61eaad2de..4276bb6d31 100644
--- a/drivers/net/e1000/igb_rxtx.c
+++ b/drivers/net/e1000/igb_rxtx.c
@@ -1868,6 +1868,7 @@ igb_dev_clear_queues(struct rte_eth_dev *dev)
struct igb_rx_queue *rxq;
for (i = 0; i < dev->data->nb_tx_queues; i++) {
+ __rte_assume(i < RTE_MAX_QUEUES_PER_PORT);
txq = dev->data->tx_queues[i];
if (txq != NULL) {
igb_tx_queue_release_mbufs(txq);
@@ -1877,6 +1878,7 @@ igb_dev_clear_queues(struct rte_eth_dev *dev)
}
for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ __rte_assume(i < RTE_MAX_QUEUES_PER_PORT);
rxq = dev->data->rx_queues[i];
if (rxq != NULL) {
igb_rx_queue_release_mbufs(rxq);
diff --git a/drivers/net/failsafe/failsafe_ops.c b/drivers/net/failsafe/failsafe_ops.c
index 9c013e0419..5321c3385c 100644
--- a/drivers/net/failsafe/failsafe_ops.c
+++ b/drivers/net/failsafe/failsafe_ops.c
@@ -111,12 +111,14 @@ fs_set_queues_state_start(struct rte_eth_dev *dev)
uint16_t i;
for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ __rte_assume(i < RTE_MAX_QUEUES_PER_PORT);
rxq = dev->data->rx_queues[i];
if (rxq != NULL && !rxq->info.conf.rx_deferred_start)
dev->data->rx_queue_state[i] =
RTE_ETH_QUEUE_STATE_STARTED;
}
for (i = 0; i < dev->data->nb_tx_queues; i++) {
+ __rte_assume(i < RTE_MAX_QUEUES_PER_PORT);
txq = dev->data->tx_queues[i];
if (txq != NULL && !txq->info.conf.tx_deferred_start)
dev->data->tx_queue_state[i] =
@@ -176,14 +178,18 @@ fs_set_queues_state_stop(struct rte_eth_dev *dev)
{
uint16_t i;
- for (i = 0; i < dev->data->nb_rx_queues; i++)
+ for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ __rte_assume(i < RTE_MAX_QUEUES_PER_PORT);
if (dev->data->rx_queues[i] != NULL)
dev->data->rx_queue_state[i] =
RTE_ETH_QUEUE_STATE_STOPPED;
- for (i = 0; i < dev->data->nb_tx_queues; i++)
+ }
+ for (i = 0; i < dev->data->nb_tx_queues; i++) {
+ __rte_assume(i < RTE_MAX_QUEUES_PER_PORT);
if (dev->data->tx_queues[i] != NULL)
dev->data->tx_queue_state[i] =
RTE_ETH_QUEUE_STATE_STOPPED;
+ }
}
static int
diff --git a/drivers/net/hns3/hns3_rxtx.c b/drivers/net/hns3/hns3_rxtx.c
index 5941b966e0..03bbbc435f 100644
--- a/drivers/net/hns3/hns3_rxtx.c
+++ b/drivers/net/hns3/hns3_rxtx.c
@@ -1309,6 +1309,7 @@ hns3_start_tqps(struct hns3_hw *hw)
hns3_enable_all_queues(hw, true);
for (i = 0; i < hw->data->nb_tx_queues; i++) {
+ __rte_assume(i < RTE_MAX_QUEUES_PER_PORT);
txq = hw->data->tx_queues[i];
if (txq->enabled)
hw->data->tx_queue_state[i] =
@@ -1316,6 +1317,7 @@ hns3_start_tqps(struct hns3_hw *hw)
}
for (i = 0; i < hw->data->nb_rx_queues; i++) {
+ __rte_assume(i < RTE_MAX_QUEUES_PER_PORT);
rxq = hw->data->rx_queues[i];
if (rxq->enabled)
hw->data->rx_queue_state[i] =
diff --git a/drivers/net/mana/tx.c b/drivers/net/mana/tx.c
index 272a28bcba..e2352e6300 100644
--- a/drivers/net/mana/tx.c
+++ b/drivers/net/mana/tx.c
@@ -74,6 +74,7 @@ mana_start_tx_queues(struct rte_eth_dev *dev)
if (dev->data->tx_queue_state[i] == RTE_ETH_QUEUE_STATE_STARTED)
return -EINVAL;
+ __rte_assume(priv->num_queues < RTE_MAX_QUEUES_PER_PORT);
for (i = 0; i < priv->num_queues; i++) {
struct mana_txq *txq;
struct ibv_qp_init_attr qp_attr = { 0 };
diff --git a/drivers/net/vmxnet3/vmxnet3_ethdev.c b/drivers/net/vmxnet3/vmxnet3_ethdev.c
index 78fac63ab6..1752c58069 100644
--- a/drivers/net/vmxnet3/vmxnet3_ethdev.c
+++ b/drivers/net/vmxnet3/vmxnet3_ethdev.c
@@ -1470,42 +1470,52 @@ vmxnet3_dev_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats)
struct vmxnet3_hw *hw = dev->data->dev_private;
struct UPT1_TxStats txStats;
struct UPT1_RxStats rxStats;
+ uint64_t packets, bytes;
VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_CMD, VMXNET3_CMD_GET_STATS);
for (i = 0; i < hw->num_tx_queues; i++) {
vmxnet3_tx_stats_get(hw, i, &txStats);
- stats->q_opackets[i] = txStats.ucastPktsTxOK +
+ packets = txStats.ucastPktsTxOK +
txStats.mcastPktsTxOK +
txStats.bcastPktsTxOK;
- stats->q_obytes[i] = txStats.ucastBytesTxOK +
+ bytes = txStats.ucastBytesTxOK +
txStats.mcastBytesTxOK +
txStats.bcastBytesTxOK;
- stats->opackets += stats->q_opackets[i];
- stats->obytes += stats->q_obytes[i];
+ stats->opackets += packets;
+ stats->obytes += bytes;
stats->oerrors += txStats.pktsTxError + txStats.pktsTxDiscard;
+
+ if (i < RTE_ETHDEV_QUEUE_STAT_CNTRS) {
+ stats->q_opackets[i] = packets;
+ stats->q_obytes[i] = bytes;
+ }
}
for (i = 0; i < hw->num_rx_queues; i++) {
vmxnet3_rx_stats_get(hw, i, &rxStats);
- stats->q_ipackets[i] = rxStats.ucastPktsRxOK +
+ packets = rxStats.ucastPktsRxOK +
rxStats.mcastPktsRxOK +
rxStats.bcastPktsRxOK;
- stats->q_ibytes[i] = rxStats.ucastBytesRxOK +
+ bytes = rxStats.ucastBytesRxOK +
rxStats.mcastBytesRxOK +
rxStats.bcastBytesRxOK;
- stats->ipackets += stats->q_ipackets[i];
- stats->ibytes += stats->q_ibytes[i];
-
- stats->q_errors[i] = rxStats.pktsRxError;
+ stats->ipackets += packets;
+ stats->ibytes += bytes;
stats->ierrors += rxStats.pktsRxError;
stats->imissed += rxStats.pktsRxOutOfBuf;
+
+ if (i < RTE_ETHDEV_QUEUE_STAT_CNTRS) {
+ stats->q_ipackets[i] = packets;
+ stats->q_ibytes[i] = bytes;
+ stats->q_errors[i] = rxStats.pktsRxError;
+ }
}
return 0;
@@ -1521,8 +1531,6 @@ vmxnet3_dev_stats_reset(struct rte_eth_dev *dev)
VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_CMD, VMXNET3_CMD_GET_STATS);
- RTE_BUILD_BUG_ON(RTE_ETHDEV_QUEUE_STAT_CNTRS < VMXNET3_MAX_TX_QUEUES);
-
for (i = 0; i < hw->num_tx_queues; i++) {
vmxnet3_hw_tx_stats_get(hw, i, &txStats);
memcpy(&hw->snapshot_tx_stats[i], &txStats,
@@ -1566,7 +1574,7 @@ vmxnet3_dev_info_get(struct rte_eth_dev *dev,
dev_info->min_rx_bufsize = 1518 + RTE_PKTMBUF_HEADROOM;
dev_info->max_rx_pktlen = 16384; /* includes CRC, cf MAXFRS register */
dev_info->min_mtu = VMXNET3_MIN_MTU;
- dev_info->max_mtu = VMXNET3_MAX_MTU;
+ dev_info->max_mtu = VMXNET3_VERSION_GE_6(hw) ? VMXNET3_V6_MAX_MTU : VMXNET3_MAX_MTU;
dev_info->speed_capa = RTE_ETH_LINK_SPEED_10G;
dev_info->max_mac_addrs = VMXNET3_MAX_MAC_ADDRS;
diff --git a/drivers/net/vmxnet3/vmxnet3_ethdev.h b/drivers/net/vmxnet3/vmxnet3_ethdev.h
index 2b3e2c4caa..e9ded6663d 100644
--- a/drivers/net/vmxnet3/vmxnet3_ethdev.h
+++ b/drivers/net/vmxnet3/vmxnet3_ethdev.h
@@ -121,8 +121,8 @@ struct vmxnet3_hw {
#define VMXNET3_VFT_TABLE_SIZE (VMXNET3_VFT_SIZE * sizeof(uint32_t))
UPT1_TxStats saved_tx_stats[VMXNET3_EXT_MAX_TX_QUEUES];
UPT1_RxStats saved_rx_stats[VMXNET3_EXT_MAX_RX_QUEUES];
- UPT1_TxStats snapshot_tx_stats[VMXNET3_MAX_TX_QUEUES];
- UPT1_RxStats snapshot_rx_stats[VMXNET3_MAX_RX_QUEUES];
+ UPT1_TxStats snapshot_tx_stats[VMXNET3_EXT_MAX_TX_QUEUES];
+ UPT1_RxStats snapshot_rx_stats[VMXNET3_EXT_MAX_RX_QUEUES];
uint16_t tx_prod_offset;
uint16_t rx_prod_offset[2];
/* device capability bit map */
diff --git a/lib/eal/include/rte_common.h b/lib/eal/include/rte_common.h
index c79f9ed319..2f143c87e4 100644
--- a/lib/eal/include/rte_common.h
+++ b/lib/eal/include/rte_common.h
@@ -366,6 +366,16 @@ static void __attribute__((destructor(RTE_PRIO(prio)), used)) func(void)
#define __rte_noreturn __attribute__((noreturn))
#endif
+/**
+ * Hint point in program never reached
+ */
+#if defined(RTE_TOOLCHAIN_GCC) || defined(RTE_TOOLCHAIN_CLANG)
+#define __rte_unreachable() __extension__(__builtin_unreachable())
+#else
+/* MSVC or ICC */
+#define __rte_unreachable() __assume(0)
+#endif
+
/**
* Issue a warning in case the function's return value is ignored.
*
@@ -423,6 +433,23 @@ static void __attribute__((destructor(RTE_PRIO(prio)), used)) func(void)
#define __rte_cold __attribute__((cold))
#endif
+/**
+ * Hint precondition
+ *
+ * @warning Depending on the compiler, any code in ``condition`` might be executed.
+ * This currently only occurs with GCC prior to version 13.
+ */
+#if defined(RTE_TOOLCHAIN_GCC) && (GCC_VERSION >= 130000)
+#define __rte_assume(condition) __attribute__((assume(condition)))
+#elif defined(RTE_TOOLCHAIN_GCC)
+#define __rte_assume(condition) do { if (!(condition)) __rte_unreachable(); } while (0)
+#elif defined(RTE_TOOLCHAIN_CLANG)
+#define __rte_assume(condition) __extension__(__builtin_assume(condition))
+#else
+/* MSVC or ICC */
+#define __rte_assume(condition) __assume(condition)
+#endif
+
/**
* Disable AddressSanitizer on some code
*/
--
2.43.0
next prev parent reply other threads:[~2024-11-06 22:02 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-24 10:22 [PATCH] ethdev: TEST " Morten Brørup
2024-10-24 21:11 ` [PATCH v7] " Morten Brørup
2024-10-25 8:23 ` [PATCH v8] " Morten Brørup
2024-11-06 21:16 ` [TEST v9] " Morten Brørup
2024-11-06 22:02 ` Morten Brørup [this message]
2024-11-06 23:25 ` [TEST v10] ethdev: " Stephen Hemminger
2024-11-07 9:35 ` Morten Brørup
2024-11-07 9:26 ` [TEST v11] ethdev: test " Morten Brørup
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=20241106220216.1505231-1-mb@smartsharesystems.com \
--to=mb@smartsharesystems.com \
--cc=dev@dpdk.org \
/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).