Re: [dpdk-dev] [PATCH v1 06/11] config: add build enablement for IPN3KE
Hi Ferruh, It seems I have missed some comments, so just reply this again. > -Original Message- > From: Yigit, Ferruh [mailto:ferruh.yi...@linux.intel.com] > Sent: Wednesday, March 06, 2019 20:45 > To: Xu, Rosen ; dev@dpdk.org > Cc: Yigit, Ferruh ; Zhang, Tianfei > ; Wei, Dan ; Pei, Andy > ; Yang, Qiming ; Wang, > Haiyue ; Chen, Santos ; > Zhang, Zhang > Subject: Re: [dpdk-dev] [PATCH v1 06/11] config: add build enablement for > IPN3KE > > On 2/28/2019 7:13 AM, Rosen Xu wrote: > > Add build enablement for Intel FPGA Acceleration NIC IPN3KE. > > No need to make this a separate patch, can squash into previous one. I will merge it in patch v2. > > > > Signed-off-by: Rosen Xu > > --- > > config/common_base | 6 ++ > > 1 file changed, 6 insertions(+) > > > > diff --git a/config/common_base b/config/common_base index > > 7c6da51..4fac8ba 100644 > > --- a/config/common_base > > +++ b/config/common_base > > @@ -316,6 +316,12 @@ CONFIG_RTE_LIBRTE_AVF_DEBUG_RX=n > > CONFIG_RTE_LIBRTE_AVF_16BYTE_RX_DESC=n > > > > # > > +# Compile burst-oriented IPN3KE PMD driver # > > +CONFIG_RTE_LIBRTE_IPN3KE_PMD=n > > Why PMD is disabled by default? It should not be disabled by default. I will change it to y in patch v2. > > +CONFIG_RTE_LIBRTE_IPN3KE_DEBUG=n > > _DEBUG seems not used at all, if so can you please remove it? Currently it's not used, I will remove it in patch v2. > > + > > +# > > # Compile burst-oriented Mellanox ConnectX-3 (MLX4) PMD # > > CONFIG_RTE_LIBRTE_MLX4_PMD=n > >
Re: [dpdk-dev] [PATCH v2] ring: enforce reading the tails before ring operations
> -Original Message- > From: Honnappa Nagarahalli > Sent: Saturday, March 9, 2019 7:48 AM > To: tho...@monjalon.net; Ananyev, Konstantin > ; Gavin Hu (Arm Technology China) > > Cc: Ilya Maximets ; dev@dpdk.org; nd > ; jer...@marvell.com; hemant.agra...@nxp.com; > nipun.gu...@nxp.com; olivier.m...@6wind.com; Richardson, Bruce > ; chao...@linux.vnet.ibm.com; nd > > Subject: RE: [PATCH v2] ring: enforce reading the tails before ring > operations > > > 08/03/2019 16:50, Ananyev, Konstantin: > > > 08/03/2019 16:05, Gavin Hu (Arm Technology China): > > > > Anyway, on x86, smp_rmb, as a compiler barrier, applies to > load/store, not > > only load/load. > > > > > > Yes, that's true, but I think that's happened by coincidence, not > > > intentionally. > > > > > > > This is the case also for arm, arm64, ppc32, ppc64. > > > > I will submit a patch to expand the definition of this API. > > > > > > I understand your intention, but that does mean we would also need > to > > > change not only rte_smp_rmb() but rte_rmb() too (to keep things > consistent)? > > > That sounds worring. > > > Might be better to keep smp_rmb() definition as it is, and introduce > > > new function that fits your purposes (smp_rwmb or > smp_load_store_barrier)? > Looking at rte_rmb, rte_io_rmb, rte_cio_rmb implementations for Arm, > they all provide load/store barrier as well. If other architectures also > provide load/store barrier with rte_xxx_rmb, then we could extend the > meaning of the existing APIs. Further looking at rte_rmb, rte_io_rmb, rte_cio_rmb implementations for PPC64 and x86, They also provide load/store barrier. It is safe to extend the meaning of the existing rte_XXX_rmb API. > > Even if a new API is provided, we need to do provide the same APIs for IO > and CIO variants. Since rte_XXX_rmbs API for all architectures already provide the desired load/store ordering, a new API is redundant and not needed. > > > > How is it managed in other projects? > In my experience, I usually have been changing the algorithms to use C11 > memory model. So, I have not come across this issue yet. Others can > comment. > > >
Re: [dpdk-dev] [PATCH 4/4] .gitignore: ignore build directories from test-meson-build
09/03/2019 04:03, Stephen Hemminger: > On Sat, 09 Mar 2019 01:47:32 +0100 > Thomas Monjalon wrote: > > > 06/03/2019 18:05, Bruce Richardson: > > > test-meson-build.sh generates multiple build directories for various > > > targets. As these follow a few known patterns of reasonable lengths, > > > and since they don't need to be tracked in git, we can add them to > > > the gitignore file. > > > > > > Signed-off-by: Bruce Richardson > > > --- > > > --- a/.gitignore > > > +++ b/.gitignore > > > +build-arm64-* > > > +build-clang-* > > > +build-gcc-* > > > +build-x86-default > > > > We could even ignore build-* ? > > Why not? > build* > > Since the default build directory in build/ Yes, but "git ls build*" is not empty because of buildtools/ Could be "build" and "build-*".
[dpdk-dev] [Bug 205] make test fails with DPDK 19.02 (KeyError: 'Command')
https://bugs.dpdk.org/show_bug.cgi?id=205 Rami Rosen (ramir...@gmail.com) changed: What|Removed |Added Status|CONFIRMED |RESOLVED Resolution|--- |FIXED --- Comment #2 from Rami Rosen (ramir...@gmail.com) --- Following patch fixes this: https://mails.dpdk.org/archives/dev/2019-February/124648.html https://mails.dpdk.org/archives/dev/2019-February/124624.html so closing this BUG. -- You are receiving this mail because: You are the assignee for the bug.
Re: [dpdk-dev] [PATCH 4/4] .gitignore: ignore build directories from test-meson-build
> Thomas Monjalon wrote: >Yes, but "git ls build*" is not empty because of buildtools/ >Could be "build" and "build-*". +1 > > > > >
[dpdk-dev] [PATCH v2 01/10] net/atlantic: fix negative error codes
These are just convention breakage on rte_errno,
no real harm from that.
Cc: sta...@dpdk.org
Signed-off-by: Igor Russkikh
Fixes: 2b1472d7150c ("net/atlantic: implement Tx path")
---
drivers/net/atlantic/atl_rxtx.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/atlantic/atl_rxtx.c b/drivers/net/atlantic/atl_rxtx.c
index 40c913795c79..aea58c0d5b83 100644
--- a/drivers/net/atlantic/atl_rxtx.c
+++ b/drivers/net/atlantic/atl_rxtx.c
@@ -812,12 +812,12 @@ atl_prep_pkts(__rte_unused void *tx_queue, struct
rte_mbuf **tx_pkts,
ol_flags = m->ol_flags;
if (m->nb_segs > AQ_HW_MAX_SEGS_SIZE) {
- rte_errno = -EINVAL;
+ rte_errno = EINVAL;
return i;
}
if (ol_flags & ATL_TX_OFFLOAD_NOTSUP_MASK) {
- rte_errno = -ENOTSUP;
+ rte_errno = ENOTSUP;
return i;
}
--
2.17.1
[dpdk-dev] [PATCH v2 00/10] net/atlantic: bugfixes and code cleanup
The patchset mainly fixes some Coverity issues, couple of minor bugs and overflows. It also improves eeprom get/set logic with ability of custom device address and dump length. v2: better git log, line lengths, Fixes tags. Igor Russkikh (7): net/atlantic: fix negative error codes net/atlantic: remove unused variable net/atlantic: extra checks for error codes net/atlantic: fix link configuration net/atlantic: eliminate excessive log levels on Rx/Tx net/atlantic: fix missing VLAN filter offload net/atlantic: fix xstats to return correct number of items Pavel Belous (3): net/atlantic: fix buffer overflow net/atlantic: use EEPROM magic as a device address net/atlantic: fix EEPROM get for small and uneven lengths drivers/net/atlantic/atl_ethdev.c | 75 ++- drivers/net/atlantic/atl_rxtx.c | 20 ++--- drivers/net/atlantic/atl_types.h | 7 +- drivers/net/atlantic/hw_atl/hw_atl_b0.c | 6 +- drivers/net/atlantic/hw_atl/hw_atl_utils.c| 9 ++- drivers/net/atlantic/hw_atl/hw_atl_utils.h| 23 +++--- .../net/atlantic/hw_atl/hw_atl_utils_fw2x.c | 64 +++- 7 files changed, 116 insertions(+), 88 deletions(-) -- 2.17.1
[dpdk-dev] [PATCH v2 02/10] net/atlantic: remove unused variable
Found by coverity scan. Signed-off-by: Igor Russkikh --- drivers/net/atlantic/hw_atl/hw_atl_b0.c | 6 +- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/drivers/net/atlantic/hw_atl/hw_atl_b0.c b/drivers/net/atlantic/hw_atl/hw_atl_b0.c index 9400e0edb999..53fd8e9f0237 100644 --- a/drivers/net/atlantic/hw_atl/hw_atl_b0.c +++ b/drivers/net/atlantic/hw_atl/hw_atl_b0.c @@ -31,7 +31,6 @@ static int hw_atl_b0_hw_qos_set(struct aq_hw_s *self) u32 tc = 0U; u32 buff_size = 0U; unsigned int i_priority = 0U; - bool is_rx_flow_control = false; /* TPS Descriptor rate init */ hw_atl_tps_tx_pkt_shed_desc_rate_curr_time_res_set(self, 0x0U); @@ -64,7 +63,6 @@ static int hw_atl_b0_hw_qos_set(struct aq_hw_s *self) /* QoS Rx buf size per TC */ tc = 0; - is_rx_flow_control = 0; buff_size = HW_ATL_B0_RXBUF_MAX; hw_atl_rpb_rx_pkt_buff_size_per_tc_set(self, buff_size, tc); @@ -76,9 +74,7 @@ static int hw_atl_b0_hw_qos_set(struct aq_hw_s *self) (buff_size * (1024U / 32U) * 50U) / 100U, tc); - hw_atl_rpb_rx_xoff_en_per_tc_set(self, -is_rx_flow_control ? 1U : 0U, -tc); + hw_atl_rpb_rx_xoff_en_per_tc_set(self, 0U, tc); /* QoS 802.1p priority -> TC mapping */ for (i_priority = 8U; i_priority--;) -- 2.17.1
[dpdk-dev] [PATCH v2 03/10] net/atlantic: extra checks for error codes
Found by Coverity scan. Checks are useless
because at these code places err is always zero.
Signed-off-by: Igor Russkikh
---
drivers/net/atlantic/hw_atl/hw_atl_utils.c | 5 +
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/drivers/net/atlantic/hw_atl/hw_atl_utils.c
b/drivers/net/atlantic/hw_atl/hw_atl_utils.c
index f11093a50404..13f02b9f99c5 100644
--- a/drivers/net/atlantic/hw_atl/hw_atl_utils.c
+++ b/drivers/net/atlantic/hw_atl/hw_atl_utils.c
@@ -462,8 +462,6 @@ int hw_atl_utils_fw_rpc_wait(struct aq_hw_s *self,
goto err_exit;
}
} while (sw.tid != fw.tid || 0xU == fw.len);
- if (err < 0)
- goto err_exit;
if (rpc) {
if (fw.len) {
@@ -875,8 +873,7 @@ static int aq_fw1x_set_wol(struct aq_hw_s *self, bool
wol_enabled, u8 *mac)
}
err = hw_atl_utils_fw_rpc_call(self, rpc_size);
- if (err < 0)
- goto err_exit;
+
err_exit:
return err;
}
--
2.17.1
[dpdk-dev] [PATCH v2 05/10] net/atlantic: use EEPROM magic as a device address
From: Pavel Belous
Default dev addr is replaced with magic field from the request.
Length is allowed to be less than maximum.
SMBUS access bit definitions also better organised now.
Signed-off-by: Igor Russkikh
Signed-off-by: Pavel Belous
---
drivers/net/atlantic/atl_ethdev.c | 24 +++-
drivers/net/atlantic/atl_types.h | 7 +++--
drivers/net/atlantic/hw_atl/hw_atl_utils.c| 4 +++
drivers/net/atlantic/hw_atl/hw_atl_utils.h| 23 ---
.../net/atlantic/hw_atl/hw_atl_utils_fw2x.c | 28 +++
5 files changed, 56 insertions(+), 30 deletions(-)
diff --git a/drivers/net/atlantic/atl_ethdev.c
b/drivers/net/atlantic/atl_ethdev.c
index 5bc04f55cc21..925f9722ec66 100644
--- a/drivers/net/atlantic/atl_ethdev.c
+++ b/drivers/net/atlantic/atl_ethdev.c
@@ -1102,24 +1102,30 @@ atl_dev_get_eeprom_length(struct rte_eth_dev *dev
__rte_unused)
return SFP_EEPROM_SIZE;
}
-static int
-atl_dev_get_eeprom(struct rte_eth_dev *dev, struct rte_dev_eeprom_info *eeprom)
+int atl_dev_get_eeprom(struct rte_eth_dev *dev,
+ struct rte_dev_eeprom_info *eeprom)
{
struct aq_hw_s *hw = ATL_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ uint32_t dev_addr = SMBUS_DEVICE_ID;
if (hw->aq_fw_ops->get_eeprom == NULL)
return -ENOTSUP;
- if (eeprom->length != SFP_EEPROM_SIZE || eeprom->data == NULL)
+ if (eeprom->length > SFP_EEPROM_SIZE || eeprom->data == NULL)
return -EINVAL;
- return hw->aq_fw_ops->get_eeprom(hw, eeprom->data, eeprom->length);
+ if (eeprom->magic)
+ dev_addr = eeprom->magic;
+
+ return hw->aq_fw_ops->get_eeprom(hw, dev_addr, eeprom->data,
+eeprom->length);
}
-static int
-atl_dev_set_eeprom(struct rte_eth_dev *dev, struct rte_dev_eeprom_info *eeprom)
+int atl_dev_set_eeprom(struct rte_eth_dev *dev,
+ struct rte_dev_eeprom_info *eeprom)
{
struct aq_hw_s *hw = ATL_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ uint32_t dev_addr = SMBUS_DEVICE_ID;
if (hw->aq_fw_ops->set_eeprom == NULL)
return -ENOTSUP;
@@ -1127,7 +1133,11 @@ atl_dev_set_eeprom(struct rte_eth_dev *dev, struct
rte_dev_eeprom_info *eeprom)
if (eeprom->length != SFP_EEPROM_SIZE || eeprom->data == NULL)
return -EINVAL;
- return hw->aq_fw_ops->set_eeprom(hw, eeprom->data, eeprom->length);
+ if (eeprom->magic)
+ dev_addr = eeprom->magic;
+
+ return hw->aq_fw_ops->set_eeprom(hw, dev_addr,
+eeprom->data, eeprom->length);
}
static int
diff --git a/drivers/net/atlantic/atl_types.h b/drivers/net/atlantic/atl_types.h
index 3d90f6caefc2..46628dc0464d 100644
--- a/drivers/net/atlantic/atl_types.h
+++ b/drivers/net/atlantic/atl_types.h
@@ -137,9 +137,12 @@ struct aq_fw_ops {
int (*led_control)(struct aq_hw_s *self, u32 mode);
- int (*get_eeprom)(struct aq_hw_s *self, u32 *data, u32 len);
+ int (*get_eeprom)(struct aq_hw_s *self, int dev_addr,
+ u32 *data, u32 len);
+
+ int (*set_eeprom)(struct aq_hw_s *self, int dev_addr,
+ u32 *data, u32 len);
- int (*set_eeprom)(struct aq_hw_s *self, u32 *data, u32 len);
};
struct atl_sw_stats {
diff --git a/drivers/net/atlantic/hw_atl/hw_atl_utils.c
b/drivers/net/atlantic/hw_atl/hw_atl_utils.c
index 13f02b9f99c5..4299b7016e2f 100644
--- a/drivers/net/atlantic/hw_atl/hw_atl_utils.c
+++ b/drivers/net/atlantic/hw_atl/hw_atl_utils.c
@@ -305,6 +305,10 @@ int hw_atl_utils_fw_downld_dwords(struct aq_hw_s *self,
u32 a,
AQ_HW_WAIT_FOR(!(0x100 & aq_hw_read_reg(self,
HW_ATL_MIF_CMD)),
1, 1000U);
+ if (err) {
+ err = -ETIMEDOUT;
+ goto err_exit;
+ }
*(p++) = aq_hw_read_reg(self, HW_ATL_MIF_VAL);
a += 4;
diff --git a/drivers/net/atlantic/hw_atl/hw_atl_utils.h
b/drivers/net/atlantic/hw_atl/hw_atl_utils.h
index 5f3f70847310..f2a87826c0d1 100644
--- a/drivers/net/atlantic/hw_atl/hw_atl_utils.h
+++ b/drivers/net/atlantic/hw_atl/hw_atl_utils.h
@@ -8,6 +8,7 @@
#ifndef HW_ATL_UTILS_H
#define HW_ATL_UTILS_H
+#define BIT(x) (1UL << (x))
#define HW_ATL_FLUSH() { (void)aq_hw_read_reg(self, 0x10); }
/* Hardware tx descriptor */
@@ -389,18 +390,8 @@ enum hal_atl_utils_fw_state_e {
#define HAL_ATLANTIC_UTILS_FW_MSG_OFFLOAD_DEL 10U
#define HAL_ATLANTIC_UTILS_FW_MSG_CABLE_DIAG 13U // 0xd
-#define SMBUS_READ_REQUEST BIT(13)
-#define SMBUS_WRITE_REQUEST BIT(14)
#define SMBUS_DEVICE_ID 0x50
-enum hw_atl_fw2x_rate {
- FW2X_RATE_100M= 0x20,
- FW2X_RATE_1G = 0x100,
- FW2X_RATE_2
[dpdk-dev] [PATCH v2 04/10] net/atlantic: fix buffer overflow
From: Pavel Belous
Found by Coverity scan. This is a real memory corruption.
There is no need in extra RTE_ALIGN macros since the
request/result structures are 4-byte aligned by definition.
Cc: sta...@dpdk.org
Fixes: ce4e8d418097 ("net/atlantic: implement EEPROM get/set")
Signed-off-by: Igor Russkikh
Signed-off-by: Pavel Belous
---
drivers/net/atlantic/hw_atl/hw_atl_utils_fw2x.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/atlantic/hw_atl/hw_atl_utils_fw2x.c
b/drivers/net/atlantic/hw_atl/hw_atl_utils_fw2x.c
index 6841d9bce39c..f90ccfe9e010 100644
--- a/drivers/net/atlantic/hw_atl/hw_atl_utils_fw2x.c
+++ b/drivers/net/atlantic/hw_atl/hw_atl_utils_fw2x.c
@@ -501,7 +501,7 @@ static int aq_fw2x_get_eeprom(struct aq_hw_s *self, u32
*data, u32 len)
/* Write SMBUS request to cfg memory */
err = hw_atl_utils_fw_upload_dwords(self, self->rpc_addr,
(u32 *)(void *)&request,
- RTE_ALIGN(sizeof(request), sizeof(u32)));
+ sizeof(request) / sizeof(u32));
if (err < 0)
return err;
@@ -523,7 +523,7 @@ static int aq_fw2x_get_eeprom(struct aq_hw_s *self, u32
*data, u32 len)
err = hw_atl_utils_fw_downld_dwords(self, self->rpc_addr + sizeof(u32),
&result,
- RTE_ALIGN(sizeof(result), sizeof(u32)));
+ sizeof(result) / sizeof(u32));
if (err < 0)
return err;
@@ -558,7 +558,7 @@ static int aq_fw2x_set_eeprom(struct aq_hw_s *self, u32
*data, u32 len)
/* Write SMBUS request to cfg memory */
err = hw_atl_utils_fw_upload_dwords(self, self->rpc_addr,
(u32 *)(void *)&request,
- RTE_ALIGN(sizeof(request), sizeof(u32)));
+ sizeof(request) / sizeof(u32));
if (err < 0)
return err;
@@ -589,7 +589,7 @@ static int aq_fw2x_set_eeprom(struct aq_hw_s *self, u32
*data, u32 len)
/* Read status of write operation */
err = hw_atl_utils_fw_downld_dwords(self, self->rpc_addr + sizeof(u32),
&result,
- RTE_ALIGN(sizeof(result), sizeof(u32)));
+ sizeof(result) / sizeof(u32));
if (err < 0)
return err;
--
2.17.1
[dpdk-dev] [PATCH v2 07/10] net/atlantic: fix link configuration
In case link speed is re configured after port start, it does not
takes the requested speed value, but instead just sets full autoneg
mask.
Cc: sta...@dpdk.org
Fixes: 7943ba05f67c ("net/atlantic: add link status and interrupt management")
Signed-off-by: Igor Russkikh
---
drivers/net/atlantic/atl_ethdev.c | 44 ++-
1 file changed, 20 insertions(+), 24 deletions(-)
diff --git a/drivers/net/atlantic/atl_ethdev.c
b/drivers/net/atlantic/atl_ethdev.c
index 925f9722ec66..c869ff321bf5 100644
--- a/drivers/net/atlantic/atl_ethdev.c
+++ b/drivers/net/atlantic/atl_ethdev.c
@@ -465,8 +465,6 @@ atl_dev_start(struct rte_eth_dev *dev)
struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev);
struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
uint32_t intr_vector = 0;
- uint32_t *link_speeds;
- uint32_t speed = 0;
int status;
int err;
@@ -543,6 +541,8 @@ atl_dev_start(struct rte_eth_dev *dev)
goto error;
}
+ err = atl_dev_set_link_up(dev);
+
err = hw->aq_fw_ops->update_link_status(hw);
if (err)
@@ -550,26 +550,6 @@ atl_dev_start(struct rte_eth_dev *dev)
dev->data->dev_link.link_status = hw->aq_link_status.mbps != 0;
- link_speeds = &dev->data->dev_conf.link_speeds;
-
- speed = 0x0;
-
- if (*link_speeds == ETH_LINK_SPEED_AUTONEG) {
- speed = hw->aq_nic_cfg->link_speed_msk;
- } else {
- if (*link_speeds & ETH_LINK_SPEED_10G)
- speed |= AQ_NIC_RATE_10G;
- if (*link_speeds & ETH_LINK_SPEED_5G)
- speed |= AQ_NIC_RATE_5G;
- if (*link_speeds & ETH_LINK_SPEED_1G)
- speed |= AQ_NIC_RATE_1G;
- if (*link_speeds & ETH_LINK_SPEED_2_5G)
- speed |= AQ_NIC_RATE_2G5;
- if (*link_speeds & ETH_LINK_SPEED_100M)
- speed |= AQ_NIC_RATE_100M;
- }
-
- err = hw->aq_fw_ops->set_link_speed(hw, speed);
if (err)
goto error;
@@ -657,9 +637,25 @@ static int
atl_dev_set_link_up(struct rte_eth_dev *dev)
{
struct aq_hw_s *hw = ATL_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ uint32_t link_speeds = dev->data->dev_conf.link_speeds;
+ uint32_t speed_mask = 0;
+
+ if (link_speeds == ETH_LINK_SPEED_AUTONEG) {
+ speed_mask = hw->aq_nic_cfg->link_speed_msk;
+ } else {
+ if (link_speeds & ETH_LINK_SPEED_10G)
+ speed_mask |= AQ_NIC_RATE_10G;
+ if (link_speeds & ETH_LINK_SPEED_5G)
+ speed_mask |= AQ_NIC_RATE_5G;
+ if (link_speeds & ETH_LINK_SPEED_1G)
+ speed_mask |= AQ_NIC_RATE_1G;
+ if (link_speeds & ETH_LINK_SPEED_2_5G)
+ speed_mask |= AQ_NIC_RATE_2G5;
+ if (link_speeds & ETH_LINK_SPEED_100M)
+ speed_mask |= AQ_NIC_RATE_100M;
+ }
- return hw->aq_fw_ops->set_link_speed(hw,
- hw->aq_nic_cfg->link_speed_msk);
+ return hw->aq_fw_ops->set_link_speed(hw, speed_mask);
}
/*
--
2.17.1
[dpdk-dev] [PATCH v2 08/10] net/atlantic: eliminate excessive log levels on Rx/Tx
Default rxtx logging used ERR level, that caused logger to always
trigger. That may cause perf degradation even if logger was not enabled but
compiled in.
Signed-off-by: Igor Russkikh
---
drivers/net/atlantic/atl_rxtx.c | 16
1 file changed, 4 insertions(+), 12 deletions(-)
diff --git a/drivers/net/atlantic/atl_rxtx.c b/drivers/net/atlantic/atl_rxtx.c
index aea58c0d5b83..f1519594adad 100644
--- a/drivers/net/atlantic/atl_rxtx.c
+++ b/drivers/net/atlantic/atl_rxtx.c
@@ -946,7 +946,7 @@ atl_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts)
break;
}
- PMD_RX_LOG(ERR, "port_id=%u queue_id=%u tail=%u "
+ PMD_RX_LOG(DEBUG, "port_id=%u queue_id=%u tail=%u "
"eop=0x%x pkt_len=%u hash=0x%x hash_type=0x%x",
(unsigned int)rxq->port_id,
(unsigned int)rxq->queue_id,
@@ -981,7 +981,7 @@ atl_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts)
while (true) {
new_mbuf = rte_mbuf_raw_alloc(rxq->mb_pool);
if (new_mbuf == NULL) {
- PMD_RX_LOG(ERR,
+ PMD_RX_LOG(DEBUG,
"RX mbuf alloc failed port_id=%u "
"queue_id=%u", (unsigned int)rxq->port_id,
(unsigned int)rxq->queue_id);
@@ -1084,7 +1084,7 @@ atl_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts)
adapter->sw_stats.q_ibytes[rxq->queue_id] +=
rx_mbuf_first->pkt_len;
- PMD_RX_LOG(ERR, "add mbuf segs=%d pkt_len=%d",
+ PMD_RX_LOG(DEBUG, "add mbuf segs=%d pkt_len=%d",
rx_mbuf_first->nb_segs,
rx_mbuf_first->pkt_len);
}
@@ -1104,7 +1104,7 @@ atl_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts)
*/
nb_hold = (uint16_t)(nb_hold + rxq->nb_rx_hold);
if (nb_hold > rxq->rx_free_thresh) {
- PMD_RX_LOG(ERR, "port_id=%u queue_id=%u rx_tail=%u "
+ PMD_RX_LOG(DEBUG, "port_id=%u queue_id=%u rx_tail=%u "
"nb_hold=%u nb_rx=%u",
(unsigned int)rxq->port_id, (unsigned int)rxq->queue_id,
(unsigned int)tail, (unsigned int)nb_hold,
@@ -1129,8 +1129,6 @@ atl_xmit_cleanup(struct atl_tx_queue *txq)
struct hw_atl_txd_s *txd;
int to_clean = 0;
- PMD_INIT_FUNC_TRACE();
-
if (txq != NULL) {
sw_ring = txq->sw_ring;
int head = txq->tx_head;
@@ -1181,11 +1179,7 @@ atl_tso_setup(struct rte_mbuf *tx_pkt, union
hw_atl_txc_s *txc)
uint32_t tx_cmd = 0;
uint64_t ol_flags = tx_pkt->ol_flags;
- PMD_INIT_FUNC_TRACE();
-
if (ol_flags & PKT_TX_TCP_SEG) {
- PMD_DRV_LOG(DEBUG, "xmit TSO pkt");
-
tx_cmd |= tx_desc_cmd_lso | tx_desc_cmd_l4cs;
txc->cmd = 0x4;
@@ -1240,8 +1234,6 @@ atl_xmit_pkt(struct aq_hw_s *hw, struct atl_tx_queue *txq,
u32 tx_cmd = 0U;
int desc_count = 0;
- PMD_INIT_FUNC_TRACE();
-
tail = txq->tx_tail;
txc = (union hw_atl_txc_s *)&txq->hw_ring[tail];
--
2.17.1
[dpdk-dev] [PATCH v2 06/10] net/atlantic: fix EEPROM get for small and uneven lengths
From: Pavel Belous
Fixes: ce4e8d418097 ("net/atlantic: implement EEPROM get/set")
Cc: sta...@dpdk.org
Signed-off-by: Igor Russkikh
Signed-off-by: Pavel Belous
---
.../net/atlantic/hw_atl/hw_atl_utils_fw2x.c | 28 ---
1 file changed, 24 insertions(+), 4 deletions(-)
diff --git a/drivers/net/atlantic/hw_atl/hw_atl_utils_fw2x.c
b/drivers/net/atlantic/hw_atl/hw_atl_utils_fw2x.c
index 78ca1eb1faaa..dc619bfdd75e 100644
--- a/drivers/net/atlantic/hw_atl/hw_atl_utils_fw2x.c
+++ b/drivers/net/atlantic/hw_atl/hw_atl_utils_fw2x.c
@@ -534,13 +534,33 @@ static int aq_fw2x_get_eeprom(struct aq_hw_s *self, int
dev_addr,
return err;
if (result == 0) {
- err = hw_atl_utils_fw_downld_dwords(self,
+ u32 num_dwords = len / sizeof(u32);
+ u32 bytes_remains = len % sizeof(u32);
+
+ if (num_dwords) {
+ err = hw_atl_utils_fw_downld_dwords(self,
self->rpc_addr + sizeof(u32) * 2,
data,
- RTE_ALIGN(len, sizeof(u32)));
+ num_dwords);
- if (err < 0)
- return err;
+ if (err < 0)
+ return err;
+ }
+
+ if (bytes_remains) {
+ u32 val = 0;
+
+ err = hw_atl_utils_fw_downld_dwords(self,
+ self->rpc_addr + sizeof(u32) * 2 + num_dwords,
+ &val,
+ sizeof(u32));
+
+ if (err < 0)
+ return err;
+
+ rte_memcpy((u8 *)data + len - bytes_remains,
+ &val, bytes_remains);
+ }
}
return 0;
--
2.17.1
[dpdk-dev] [PATCH v2 10/10] net/atlantic: fix xstats to return correct number of items
Max number of xstats items was returned instead of actual number
of filled in records.
Cc: sta...@dpdk.org
Fixes: fbe059e87209 ("net/atlantic: implement device statistics")
Signed-off-by: Igor Russkikh
---
drivers/net/atlantic/atl_ethdev.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/atlantic/atl_ethdev.c
b/drivers/net/atlantic/atl_ethdev.c
index f2674f0cafe9..0f6edf878fe8 100644
--- a/drivers/net/atlantic/atl_ethdev.c
+++ b/drivers/net/atlantic/atl_ethdev.c
@@ -758,7 +758,7 @@ atl_dev_xstats_get_names(struct rte_eth_dev *dev
__rte_unused,
snprintf(xstats_names[i].name, RTE_ETH_XSTATS_NAME_SIZE, "%s",
atl_xstats_tbl[i].name);
- return size;
+ return i;
}
static int
@@ -778,7 +778,7 @@ atl_dev_xstats_get(struct rte_eth_dev *dev, struct
rte_eth_xstat *stats,
atl_xstats_tbl[i].offset);
}
- return n;
+ return i;
}
static int
--
2.17.1
[dpdk-dev] [PATCH v2 09/10] net/atlantic: fix missing VLAN filter offload
Original vlan offload code declared callbacks, but did not
enable the feature offload bit
Cc: sta...@dpdk.org
Fixes: f7c2c2c8c558 ("net/atlantic: implement VLAN filters and offloads")
Signed-off-by: Igor Russkikh
---
drivers/net/atlantic/atl_ethdev.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/net/atlantic/atl_ethdev.c
b/drivers/net/atlantic/atl_ethdev.c
index c869ff321bf5..f2674f0cafe9 100644
--- a/drivers/net/atlantic/atl_ethdev.c
+++ b/drivers/net/atlantic/atl_ethdev.c
@@ -165,7 +165,8 @@ static struct rte_pci_driver rte_atl_pmd = {
| DEV_RX_OFFLOAD_IPV4_CKSUM \
| DEV_RX_OFFLOAD_UDP_CKSUM \
| DEV_RX_OFFLOAD_TCP_CKSUM \
- | DEV_RX_OFFLOAD_JUMBO_FRAME)
+ | DEV_RX_OFFLOAD_JUMBO_FRAME \
+ | DEV_RX_OFFLOAD_VLAN_FILTER)
#define ATL_TX_OFFLOADS (DEV_TX_OFFLOAD_VLAN_INSERT \
| DEV_TX_OFFLOAD_IPV4_CKSUM \
--
2.17.1
Re: [dpdk-dev] [PATCH v2 04/10] net/atlantic: fix buffer overflow
On Sat, 9 Mar 2019 14:03:30 +
Igor Russkikh wrote:
> From: Pavel Belous
>
> Found by Coverity scan. This is a real memory corruption.
> There is no need in extra RTE_ALIGN macros since the
> request/result structures are 4-byte aligned by definition.
When fixing bugs found by Coverity it is best to mark these with
the Coverity number. The convention (in Linux kernel) is shown
by these examples:
Addresses-Coverity-ID: 1476095 ("Bad bit shift operation")
Detected by CoverityScan, CID#1476031 ("Dereference before null check")
Should this go in the style guide somewhere?
[dpdk-dev] [Bug 222] When setting CONFIG_RTE_LIBRTE_PDUMP=n, build fails with "undefined reference to `test_pdump'"
https://bugs.dpdk.org/show_bug.cgi?id=222 Bug ID: 222 Summary: When setting CONFIG_RTE_LIBRTE_PDUMP=n, build fails with "undefined reference to `test_pdump'" Product: DPDK Version: 19.02 Hardware: x86 OS: Linux Status: CONFIRMED Severity: normal Priority: Normal Component: other Assignee: dev@dpdk.org Reporter: ramir...@gmail.com Target Milestone: --- Running the following steps: Cloning DPDK tree (git clone http://dpdk.org/git/dpdk) make config T=x86_64-native-linuxapp-gcc setting CONFIG_RTE_LIBRTE_PDUMP=n in build/.config running "make" ... ... CC test_ipsec.o LD test /usr/bin/ld: test.o: in function `main': test.c:(.text.startup+0x91): undefined reference to `test_pdump' /usr/bin/ld: test_eal_flags.o: in function `process_dup.constprop.7': test_eal_flags.c:(.text+0x4a4): undefined reference to `flag_for_send_pkts' /usr/bin/ld: test_eal_flags.c:(.text+0x4b3): undefined reference to `send_pkts' collect2: error: ld returned 1 exit status make[3]: *** [/work/down/temp/dpdk/mk/rte.app.mk:399: test] Error 1 make[2]: *** [/work/down/temp/dpdk/mk/rte.subdir.mk:37: test] Error 2 make[1]: *** [/work/down/temp/dpdk/mk/rte.sdkbuild.mk:48: app] Error 2 make: *** [/work/down/temp/dpdk/mk/rte.sdkroot.mk:99: all] Error 2 .. ... This happens on 4.19.4-300.fc29.x86_64. The error is reproducible and consistent. It did not happen on previous releases of DPDK (though I don't remember exactly which) -- You are receiving this mail because: You are the assignee for the bug.
Re: [dpdk-dev] [PATCH v1 05/11] drivers/net/ipn3ke: add IPN3KE PMD driver
On Thu, 2019-02-28 at 15:13 +0800, Rosen Xu wrote: > Add Intel FPGA Acceleration NIC IPN3KE PMD driver. > > Signed-off-by: Rosen Xu > Signed-off-by: Andy Pei > Signed-off-by: Dan Wei > --- > drivers/net/Makefile | 1 + > drivers/net/ipn3ke/Makefile | 33 + > drivers/net/ipn3ke/ipn3ke_ethdev.c | 814 + > drivers/net/ipn3ke/ipn3ke_ethdev.h | 742 + > drivers/net/ipn3ke/ipn3ke_flow.c | 1407 > > drivers/net/ipn3ke/ipn3ke_flow.h | 104 ++ > drivers/net/ipn3ke/ipn3ke_logs.h | 30 + > drivers/net/ipn3ke/ipn3ke_representor.c | 890 ++ > drivers/net/ipn3ke/ipn3ke_tm.c | 2217 > + > drivers/net/ipn3ke/ipn3ke_tm.h | 135 ++ > drivers/net/ipn3ke/meson.build | 9 + > drivers/net/ipn3ke/rte_pmd_ipn3ke_version.map | 4 + > 12 files changed, 6386 insertions(+) > create mode 100644 drivers/net/ipn3ke/Makefile > create mode 100644 drivers/net/ipn3ke/ipn3ke_ethdev.c > create mode 100644 drivers/net/ipn3ke/ipn3ke_ethdev.h > create mode 100644 drivers/net/ipn3ke/ipn3ke_flow.c > create mode 100644 drivers/net/ipn3ke/ipn3ke_flow.h > create mode 100644 drivers/net/ipn3ke/ipn3ke_logs.h > create mode 100644 drivers/net/ipn3ke/ipn3ke_representor.c > create mode 100644 drivers/net/ipn3ke/ipn3ke_tm.c > create mode 100644 drivers/net/ipn3ke/ipn3ke_tm.h > create mode 100644 drivers/net/ipn3ke/meson.build > create mode 100644 drivers/net/ipn3ke/rte_pmd_ipn3ke_version.map > > diff --git a/drivers/net/Makefile b/drivers/net/Makefile > index 670d7f7..f66263c 100644 > --- a/drivers/net/Makefile > +++ b/drivers/net/Makefile > @@ -32,6 +32,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_FM10K_PMD) += fm10k > DIRS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e > DIRS-$(CONFIG_RTE_LIBRTE_ICE_PMD) += ice > DIRS-$(CONFIG_RTE_LIBRTE_IXGBE_PMD) += ixgbe > +DIRS-$(CONFIG_RTE_LIBRTE_IPN3KE_PMD) += ipn3ke > DIRS-$(CONFIG_RTE_LIBRTE_LIO_PMD) += liquidio > DIRS-$(CONFIG_RTE_LIBRTE_MLX4_PMD) += mlx4 > DIRS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += mlx5 You need to do the equivalent in drivers/net/meson.build -- Kind regards, Luca Boccassi
Re: [dpdk-dev] mlx5 under FreeBSD
Thursday, March 7, 2019 5:28 PM, Mit Matelske:
> Subject: Re: [dpdk-dev] mlx5 under FreeBSD
>
>
> >
> > - Original Message -
> > From: "shahafs"
> > To: "Mit Matelske"
> >
> > What you print here is the verbs layer, and looks OK. I am more interested
> on the values before the command reach the device.
> > Can you print the values of: rx_hash_fn and rx_hash_toeplitz_key from the
> CREATE_TIR input mailbox command?
> >
> > Relevant function on mlx5_ib kernel module are: mlx5_core_create_tir or
> create_rss_raw_qp_tir.
>
> In create_rss_raw_qp_tir, starting at line 1374 I added the following printfs:
>
>switch (ucmd.rx_hash_function) {
>case MLX5_RX_HASH_FUNC_TOEPLITZ:
>{
> void *rss_key = MLX5_ADDR_OF(tirc, tirc, rx_hash_toeplitz_key);
> size_t len = MLX5_FLD_SZ_BYTES(tirc, rx_hash_toeplitz_key);
>
> if (len != ucmd.rx_key_len) {
> err = -EINVAL;
> goto err;
> }
>
> printf("setting MLX5_RX_HASH_FUNC_TOEPLITZ\n");
> MLX5_SET(tirc, tirc, rx_hash_fn, MLX5_RX_HASH_FUNC_TOEPLITZ);
> MLX5_SET(tirc, tirc, rx_hash_symmetric, 1);
> memcpy(rss_key, ucmd.rx_hash_key, len);
> printf("rx_hash_toeplitz_key(%ld):\n", len); for(int ix = 0; ix < len; ++ix) {
>printf("%02X ", ((u_char *)rss_key)[ix]);
>if (((ix + 1) % 10) == 0) printf("\n"); }
> break;
>}
>default:
> err = -EOPNOTSUPP;
> goto err;
>}
>
>
> Which resulted in:
>
> Mar 7 09:25:56 fb100g1 kernel: setting MLX5_RX_HASH_FUNC_TOEPLITZ
> Mar 7 09:25:56 fb100g1 kernel: rx_hash_toeplitz_key(40):
> Mar 7 09:25:56 fb100g1 kernel: 2C C6 81 D1 5B DB F4 F7 FC A2 Mar 7 09:25:56
> fb100g1 kernel: 83 19 DB 1A 3E 94 6B 9E 38 D9 Mar 7 09:25:56 fb100g1 kernel:
> 2C 9C 03 D1 AD 99 44 A7 D9 56 Mar 7 09:25:56 fb100g1 kernel: 3D 59 06 3C 25
> F3 FC 1F DC 2A
>
> So, to me, everything looks good?
Yes it looks good. Wired, because it doesn't match the syndrome you got from
the device.
I don't have any more good ideas here.
Have you tried to compare the input parameters for this command between a
working version of Mellanox PMD over Linux and this version?
>
> Thanks again!
>
> Mit Matelske
>
>
>
> >
> > All this is "stock" FB 12.0 and DPDK 18.08 code, FWIW.
> >
> > Any and all help is appreciated. I'm a little out of my depths with this.
> >
> > Thanks!
> >
> > Mit Matelske
> >
> > >
> > > Thanks!
> > >
> > > Mit Matelske
> > >
> > > - Original Message -
> > > From: "shahafs"
> > > To: "Mit Matelske" , "stephen"
> > >
> > > Cc: "Thomas Monjalon" , "dev"
> > ,
> > > "Yongseok Koh"
> > > Sent: Sunday, March 3, 2019 2:43:41 AM
> > > Subject: RE: [dpdk-dev] mlx5 under FreeBSD
> > >
> > > Friday, March 1, 2019 5:58 PM, Mit Matelske:
> > > > Subject: Re: [dpdk-dev] mlx5 under FreeBSD
> > > >
> > > > I was wondering if I could lean on the experts to help me out with
> > > > an
> > issue?
> > > > I'm finally testing this port but don't understand why the
> > > > following error is
> > > > happening:
> > > >
> > > > net_mlx5: mlx5_rxq.c:2151: mlx5_hrxq_drop_new(): port 0 cannot
> > > > allocate QP for drop queue
> > > >
> > > > The driver is giving the following error corresponding with that
> > > > DPDK
> > > > message:
> > > >
> > > > Feb 28 15:00:44 fb100g1 kernel: mlx5_core0: ERR:
> > > > mlx5_cmd_check:701:(pid
> > > > 6819): CREATE_TIR(0x900) op_mod(0x0) failed, status bad
> > > > parameter(0x3), syndrome (0x569ee6)
> > >
> > > Hard to say why. I tried to look for the syndrome above on FW but
> > > couldn't find. Are you sure this is correct one?
> > > Which FW version you use?
> > >
> > > >
> > > > The mlx5_core_create_tir function completes successfully for the
> > > > en driver, just not the ib driver when trying to build the hash tables.
> > > > Any
> > > ideas?
> > > >
> > > > I've instrumented libmlx5.so.1, mlx5.ko and mlx5ib.ko to debug
> > > > this and everything looks fine according to my limited knowledge.
> > > > Not sure why the card is balking on this action.
> > > >
> > > > Thanks!
> > > >
> > > > Mit Matelske
> > > >
> > > > - Original Message -
> > > > From: "stephen"
> > > > To: "Mit Matelske"
> > > > Cc: "shahafs" , "Thomas Monjalon"
> > > > , "dev" , "Yongseok Koh"
> > > >
> > > > Sent: Wednesday, November 21, 2018 2:12:04 PM
> > > > Subject: Re: [dpdk-dev] mlx5 under FreeBSD
> > > >
> > > > On Wed, 21 Nov 2018 11:11:51 -0600 (CST) Mit Matelske
> > > > wrote:
> > > >
> > > > > >
> > > > > > Rather than building a complex shim, why not just have an OS
> > > > > > dependent section of the driver. The bigger issue is that the
> > > > > > kernel side functionality in BSD is different or missing. For
> > > > > > example TAP
> > > > API is different and not multi-queue.
> > > > >
> > > > > Once I started working on it, I had to go this route. I haven't
> > > > > been able to but much time into it yet, but it is coming along:
> > > > >
> > > > > EAL: PCI device :b3:00.0 on NUMA socket 0
> > > > > EAL:
