Re: [PATCH v5 00/11] remove component-specific logic for AVX builds
On Tue, Mar 25, 2025 at 5:03 PM Bruce Richardson wrote: > > A number of libs and drivers had special optimized AVX2 and AVX512 code > paths for performance reasons, and these tended to have copy-pasted > logic to build those files. Centralise that logic in the main > drivers/ and lib/ meson.build files to avoid duplication. > > v5: fix RTE_ARCH_X86 macro, which broke crc library > v4: rebase on latest main branch > minor fixes following feedback > limit use of -march=skylake-avx512 to when we don't already have a > -march flag supporting AVX512. > v3: add patch for event/dlb2 AVX512 handling. > add common code for libraries as well as drivers. > v2: add patch 4 to remove use of unnecessary CC_AVX2_SUPPORT flag > > > Bruce Richardson (11): > drivers: add generalized AVX build handling > net/intel: use common AVX build code > drivers/net: build use common AVX handling > drivers/net: remove AVX2 build-time define > event/dlb2: build using common AVX handling > lib: add generalized AVX build handling > acl: use common AVX build handling > fib: use common AVX build handling > net: simplify build-time logic for x86 > net: use common AVX512 build code > member: use common AVX512 build support > > drivers/event/dlb2/dlb2_sse.c | 4 ++ > drivers/event/dlb2/meson.build| 16 +--- > drivers/meson.build | 30 ++ > drivers/net/bnxt/bnxt_ethdev.c| 2 - > drivers/net/bnxt/meson.build | 10 + > drivers/net/enic/meson.build | 10 + > drivers/net/intel/i40e/meson.build| 26 +--- > drivers/net/intel/iavf/meson.build| 25 +--- > drivers/net/intel/ice/meson.build | 25 +--- > drivers/net/intel/idpf/meson.build| 25 +--- > drivers/net/nfp/meson.build | 10 + > drivers/net/octeon_ep/meson.build | 13 +- > drivers/net/octeon_ep/otx_ep_ethdev.c | 4 -- > drivers/net/virtio/meson.build| 9 + > lib/acl/meson.build | 54 ++--- > lib/fib/dir24_8.c | 6 +-- > lib/fib/meson.build | 18 + > lib/fib/trie.c| 6 +-- > lib/member/meson.build| 46 - > lib/meson.build | 34 +++- > lib/net/meson.build | 58 +++ > lib/net/rte_net_crc.c | 16 > 22 files changed, 114 insertions(+), 333 deletions(-) Thank you Bruce, this series lgtm. -- David Marchand
[PATCH] eventdev: fix dereferencing null atomic locks pointer in test-eventdev
Update atomic_init_locks to immediately return if memory allocation
fails. Atomic queue and atq tests updated to handle atomic locks being
null.
Coverity issue: 457876
Fixes: 9d619f82321b ("app/eventdev: introduce atomic tests")
Signed-off-by: Luka Jankovic
---
app/test-eventdev/test_atomic_atq.c| 2 ++
app/test-eventdev/test_atomic_common.h | 4 +++-
app/test-eventdev/test_atomic_queue.c | 2 ++
3 files changed, 7 insertions(+), 1 deletion(-)
diff --git a/app/test-eventdev/test_atomic_atq.c
b/app/test-eventdev/test_atomic_atq.c
index 4810d2eaae..73e2a53ae4 100644
--- a/app/test-eventdev/test_atomic_atq.c
+++ b/app/test-eventdev/test_atomic_atq.c
@@ -175,6 +175,8 @@ atomic_atq_eventdev_setup(struct evt_test *test, struct
evt_options *opt)
}
atomic_locks = atomic_init_locks(NB_STAGES, opt->nb_flows);
+ if (atomic_locks == NULL)
+ return -1;
return 0;
}
diff --git a/app/test-eventdev/test_atomic_common.h
b/app/test-eventdev/test_atomic_common.h
index a3cec4791d..0ee81328a4 100644
--- a/app/test-eventdev/test_atomic_common.h
+++ b/app/test-eventdev/test_atomic_common.h
@@ -66,8 +66,10 @@ atomic_init_locks(uint32_t nb_stages, uint32_t nb_flows)
rte_spinlock_t *atomic_locks = rte_calloc(NULL, num_locks,
sizeof(rte_spinlock_t), 0);
- if (atomic_locks == NULL)
+ if (atomic_locks == NULL) {
evt_err("Unable to allocate memory for spinlocks.");
+ return NULL;
+ }
for (uint32_t i = 0; i < num_locks; i++)
rte_spinlock_init(&atomic_locks[i]);
diff --git a/app/test-eventdev/test_atomic_queue.c
b/app/test-eventdev/test_atomic_queue.c
index c1a447bbac..8ce0849664 100644
--- a/app/test-eventdev/test_atomic_queue.c
+++ b/app/test-eventdev/test_atomic_queue.c
@@ -189,6 +189,8 @@ atomic_queue_eventdev_setup(struct evt_test *test, struct
evt_options *opt)
}
atomic_locks = atomic_init_locks(NB_STAGES, opt->nb_flows);
+ if (atomic_locks == NULL)
+ return -1;
return 0;
}
--
2.34.1
22.11.8 patches review and test
Hi all, Here is a list of patches targeted for stable release 22.11.8. The planned date for the final release is 2025/04/10. Please help with testing and validation of your use cases and report any issues/results with reply-all to this mail. For the final release the fixes and reported validations will be added to the release notes. A release candidate tarball can be found at: https://dpdk.org/browse/dpdk-stable/tag/?id=v22.11.8-rc1 These patches are located at branch 22.11 of dpdk-stable repo: https://dpdk.org/browse/dpdk-stable/ Thanks. Luca Boccassi --- Aleksandr Loktionov (2): net/igc/base: fix MAC address hash bit shift net/e1000/base: fix MAC address hash bit shift Amir Avivi (2): net/igc/base: fix iterator type net/e1000/base: fix iterator type Anatoly Burakov (3): net/e1000/base: correct mPHY access logic net/e1000: fix crashes in secondary processes net/ixgbe: fix crashes in secondary processes Andre Muezerie (5): eal/x86: fix some intrinsics header include for Windows net/bnxt: fix indication of allocation common/idpf: fix void function returning a value net/intel: fix void functions returning a value stack: fix pop in C11 implementation Andrew Boyer (1): doc: update ionic driver guide Ariel Otilibili (5): net/sfc: remove unnecessary assignment net/enetfec: remove useless assignment use Python raw string notation eal/linux: remove useless assignments mempool: fix errno in empty create Arkadiusz Kusztal (1): test/crypto: fix check for OOP header data Barbara Skobiej (3): net/igc/base: fix data type in MAC hash net/e1000/base: fix data type in MAC hash net/e1000/base: fix reset for 82580 Bruce Richardson (6): net/iavf: remove reset of Tx prepare function pointer net/intel: fix build with icx test/dma: fix pointers in IOVA as PA mode net/iavf: check interrupt registration failure net/iavf: fix crash on app exit on FreeBSD eal: fix undetected NUMA nodes Carolyn Wyborny (1): net/e1000/base: skip management check for 82575 Chengwen Feng (2): app/testpmd: show all DCB priority TC map app/testpmd: avoid crash in DCB config David Marchand (2): net/mlx5: fix leak in HWS flow counter action ci: point at GitHub mirror Dengdui Huang (3): net/hns3: fix mbuf freeing in simple Tx path net/hns3: fix copper port initialization net/hns3: fix reset timeout Dima Ruinskiy (5): net/igc/base: fix deadlock when writing i225 register net/igc/base: fix infinite loop net/igc/base: fix typo in LTR calculation net/igc/base: fix unused value net/e1000/base: fix unchecked return Ferruh Yigit (1): ethdev: convert string initialization Gavin Hu (1): net/mlx5: fix polling CQEs Gowrishankar Muthukrishnan (3): crypto/virtio: fix data queues iteration examples/vhost_crypto: fix user callbacks crypto/openssl: validate incorrect RSA signature Gregory Etelson (1): net/mlx5: fix IPIP tunnel verification Huisong Li (1): ethdev: fix functions available in new device event Jakub Buchocki (1): net/e1000/base: fix uninitialized variable Jianping Zhao (1): vhost: clear ring addresses when getting vring base Jie Hai (2): net/hns3: remove PVID info dump for VF net/hns3: rename RAS module Junfeng Guo (1): net/mlx5: fix actions translation error overwrite Kevin Traynor (1): vhost: add null callback checks Konstantin Ananyev (1): test/ring: fix init with custom number of lcores Long Li (3): net/netvsc: scan all net devices under the PCI device net/netvsc: remove device if its net devices removed net/mana: fix multi-process tracking Long Wu (1): net/bonding: fix dedicated queue setup Maayan Kashani (7): net/mlx5: adjust actions per rule limitation net/mlx5: fix flush of non-template flow rules net/mlx5: fix GRE flow match with SWS net/mlx5: fix LACP packet handling in isolated mode net/mlx5/hws: fix crash using represented port without ID net/mlx5/hws: fix GTP flags matching net/mlx5: fix GRE matching on root table Matthew Smith (1): net/ice: fix dropped packets when using VRRP Minggang Li (Gavin) (1): net/mlx5: fix Netlink socket leak Mingjin Ye (1): net/ixgbe: fix minimum Rx/Tx descriptors Nawal Kishor (1): common/cnxk: fix atomic load in batch ops Nicolas Chautru (1): test/bbdev: update FFT test vectors Nithin Dabilpuram (1): common/cnxk: fix inbound IPsec SA setup Nithinsen Kaithakadan (1): examples/ipsec-secgw: fix IV length in CTR 192/256 Pawel Malinowski (2): net/igc/base: fix semaphore timeout value net/e1000/base: fix semaphore timeout value Przemyslaw Ciesielski (4): net/igc/base: fix bitwise operation type net/igc/base: fix NVM data
[DPDK/eventdev Bug 1616] libtmp_rte_event_dlb2 sse fails to compile
https://bugs.dpdk.org/show_bug.cgi?id=1616 Thomas Monjalon (tho...@monjalon.net) changed: What|Removed |Added Resolution|--- |FIXED Status|IN_PROGRESS |RESOLVED --- Comment #5 from Thomas Monjalon (tho...@monjalon.net) --- Resolved in http://git.dpdk.org/dpdk/commit/?id=96357d5afe -- You are receiving this mail because: You are the assignee for the bug.
[DPDK/ethdev Bug 1586] Drivers are doing useless check for pthread_mutex_init return value
https://bugs.dpdk.org/show_bug.cgi?id=1586 Thomas Monjalon (tho...@monjalon.net) changed: What|Removed |Added Resolution|--- |FIXED Status|UNCONFIRMED |RESOLVED --- Comment #1 from Thomas Monjalon (tho...@monjalon.net) --- Resolved in http://git.dpdk.org/dpdk/commit/?id=4d2aa15076 -- You are receiving this mail because: You are the assignee for the bug.
[DPDK/core Bug 1559] rte_pktmbuf_pool_create doesn't set the correct rte_errno
https://bugs.dpdk.org/show_bug.cgi?id=1559 Thomas Monjalon (tho...@monjalon.net) changed: What|Removed |Added Resolution|--- |FIXED Status|UNCONFIRMED |RESOLVED --- Comment #1 from Thomas Monjalon (tho...@monjalon.net) --- Resolved in http://git.dpdk.org/dpdk/commit/?id=a81d8ceff4 -- You are receiving this mail because: You are the assignee for the bug.
[DPDK/ethdev Bug 1531] net/txgbe initialisation failed: EEPROM checksum is not valid.
https://bugs.dpdk.org/show_bug.cgi?id=1531 Thomas Monjalon (tho...@monjalon.net) changed: What|Removed |Added Resolution|--- |FIXED Status|IN_PROGRESS |RESOLVED --- Comment #5 from Thomas Monjalon (tho...@monjalon.net) --- Resolved in http://git.dpdk.org/dpdk/commit/?id=e389504ed4 -- You are receiving this mail because: You are the assignee for the bug.
[DPDK/core Bug 898] Memory leak during interprocess communication.
https://bugs.dpdk.org/show_bug.cgi?id=898 Thomas Monjalon (tho...@monjalon.net) changed: What|Removed |Added Resolution|--- |FIXED Status|UNCONFIRMED |RESOLVED --- Comment #3 from Thomas Monjalon (tho...@monjalon.net) --- Resolved in http://git.dpdk.org/dpdk/commit/?id=8a171e52ed -- You are receiving this mail because: You are the assignee for the bug.
[PATCH 2/7] net/ixgbe/base: correct definition of macro
The definition of IXGBE_LE32_TO_CPUS macro is meant to modify the value
in place - similar to the le32_to_cpus() macro in kernel. Fixing the
definition allows us to remove some warning flags, and removes the need
for the uintptr_t typecasts.
Fixes: aa4fc14d2cee ("ixgbe: update base driver")
Cc: sta...@dpdk.org
Signed-off-by: Bruce Richardson
---
drivers/net/intel/ixgbe/base/ixgbe_common.c | 4 ++--
drivers/net/intel/ixgbe/base/ixgbe_osdep.h | 2 +-
drivers/net/intel/ixgbe/base/meson.build| 2 +-
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/intel/ixgbe/base/ixgbe_common.c
b/drivers/net/intel/ixgbe/base/ixgbe_common.c
index d6425c5b78..fbc9605e4d 100644
--- a/drivers/net/intel/ixgbe/base/ixgbe_common.c
+++ b/drivers/net/intel/ixgbe/base/ixgbe_common.c
@@ -4610,7 +4610,7 @@ s32 ixgbe_host_interface_command(struct ixgbe_hw *hw, u32
*buffer,
/* first pull in the header so we know the buffer length */
for (bi = 0; bi < dword_len; bi++) {
buffer[bi] = IXGBE_READ_REG_ARRAY(hw, IXGBE_FLEX_MNG, bi);
- IXGBE_LE32_TO_CPUS((uintptr_t)&buffer[bi]);
+ IXGBE_LE32_TO_CPUS(&buffer[bi]);
}
/*
@@ -4646,7 +4646,7 @@ s32 ixgbe_host_interface_command(struct ixgbe_hw *hw, u32
*buffer,
/* Pull in the rest of the buffer (bi is where we left off) */
for (; bi <= dword_len; bi++) {
buffer[bi] = IXGBE_READ_REG_ARRAY(hw, IXGBE_FLEX_MNG, bi);
- IXGBE_LE32_TO_CPUS((uintptr_t)&buffer[bi]);
+ IXGBE_LE32_TO_CPUS(&buffer[bi]);
}
rel_out:
diff --git a/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
b/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
index cffc6a4ce8..6e5f7b4ae8 100644
--- a/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
+++ b/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
@@ -83,7 +83,7 @@ enum {
#define IXGBE_LE16_TO_CPU(_i) rte_le_to_cpu_16(_i)
#define IXGBE_LE32_TO_CPU(_i) rte_le_to_cpu_32(_i)
#define IXGBE_LE64_TO_CPU(_i) rte_le_to_cpu_64(_i)
-#define IXGBE_LE32_TO_CPUS(_i) rte_le_to_cpu_32(_i)
+#define IXGBE_LE32_TO_CPUS(_i) do { *_i = rte_le_to_cpu_32(*_i); } while(0)
#define IXGBE_CPU_TO_BE16(_i) rte_cpu_to_be_16(_i)
#define IXGBE_CPU_TO_BE32(_i) rte_cpu_to_be_32(_i)
#define IXGBE_BE32_TO_CPU(_i) rte_be_to_cpu_32(_i)
diff --git a/drivers/net/intel/ixgbe/base/meson.build
b/drivers/net/intel/ixgbe/base/meson.build
index 7e4fbdfa0f..f8b2ee6341 100644
--- a/drivers/net/intel/ixgbe/base/meson.build
+++ b/drivers/net/intel/ixgbe/base/meson.build
@@ -19,7 +19,7 @@ sources = [
'ixgbe_x550.c',
]
-error_cflags = ['-Wno-unused-value',
+error_cflags = [
'-Wno-unused-but-set-variable',
'-Wno-unused-parameter',
]
--
2.45.2
[PATCH 0/7] net/intel: clean up base code build
Simplify the build of the various intel base code directories, by clearing compiler warnings so the files can be compiled directly along with the regular driver files. Bruce Richardson (7): net/iavf/base: remove unused meson.build file net/ixgbe/base: correct definition of macro net/ixgbe/base: fix compilation warnings net/i40e/base: fix unused value warnings net/i40e/base: fix compiler warnings net/ice/base: reduce warnings for unused variables net/intel: simplify base code builds drivers/net/intel/e1000/base/meson.build| 9 ++-- drivers/net/intel/e1000/meson.build | 4 ++-- drivers/net/intel/fm10k/base/meson.build| 9 ++-- drivers/net/intel/fm10k/meson.build | 4 ++-- drivers/net/intel/i40e/base/i40e_diag.c | 2 +- drivers/net/intel/i40e/base/i40e_nvm.c | 2 +- drivers/net/intel/i40e/base/i40e_osdep.h| 4 ++-- drivers/net/intel/i40e/base/i40e_type.h | 14 - drivers/net/intel/i40e/base/meson.build | 23 ++--- drivers/net/intel/i40e/i40e_ethdev.c| 1 + drivers/net/intel/i40e/meson.build | 4 ++-- drivers/net/intel/iavf/base/meson.build | 10 - drivers/net/intel/ice/base/ice_osdep.h | 6 +++--- drivers/net/intel/ice/base/ice_switch.c | 2 -- drivers/net/intel/ice/base/ice_type.h | 2 +- drivers/net/intel/ice/base/meson.build | 1 - drivers/net/intel/ice/ice_fdir_filter.c | 2 +- drivers/net/intel/idpf/base/meson.build | 2 +- drivers/net/intel/idpf/meson.build | 7 --- drivers/net/intel/ixgbe/base/ixgbe_common.c | 4 ++-- drivers/net/intel/ixgbe/base/ixgbe_e610.c | 2 ++ drivers/net/intel/ixgbe/base/ixgbe_osdep.h | 21 +++ drivers/net/intel/ixgbe/base/meson.build| 20 ++ drivers/net/intel/ixgbe/meson.build | 4 ++-- 24 files changed, 56 insertions(+), 103 deletions(-) delete mode 100644 drivers/net/intel/iavf/base/meson.build -- 2.45.2
[PATCH 6/7] net/ice/base: reduce warnings for unused variables
Improve base code macros to reduce the number of issues with unused
variables in the code. Issues still remain with unused-but-set
variables, but completely unused variable warnings are eliminated.
Signed-off-by: Bruce Richardson
---
drivers/net/intel/ice/base/ice_osdep.h | 6 +++---
drivers/net/intel/ice/base/ice_switch.c | 2 --
drivers/net/intel/ice/base/ice_type.h | 2 +-
drivers/net/intel/ice/base/meson.build | 1 -
drivers/net/intel/ice/ice_fdir_filter.c | 2 +-
5 files changed, 5 insertions(+), 8 deletions(-)
diff --git a/drivers/net/intel/ice/base/ice_osdep.h
b/drivers/net/intel/ice/base/ice_osdep.h
index 7b96fcde03..ad6cde9896 100644
--- a/drivers/net/intel/ice/base/ice_osdep.h
+++ b/drivers/net/intel/ice/base/ice_osdep.h
@@ -196,9 +196,9 @@ struct __rte_packed_begin ice_virt_mem {
u32 size;
} __rte_packed_end;
-#define ice_malloc(h, s)rte_zmalloc(NULL, s, 0)
-#define ice_calloc(h, c, s) rte_calloc(NULL, c, s, 0)
-#define ice_free(h, m) rte_free(m)
+#define ice_malloc(h, s)((void)h, rte_zmalloc(NULL, s, 0))
+#define ice_calloc(h, c, s) ((void)h, rte_calloc(NULL, c, s, 0))
+#define ice_free(h, m) ((void)h, rte_free(m))
#define ice_memset(a, b, c, d) memset((a), (b), (c))
#define ice_memcpy(a, b, c, d) rte_memcpy((a), (b), (c))
diff --git a/drivers/net/intel/ice/base/ice_switch.c
b/drivers/net/intel/ice/base/ice_switch.c
index a3786961e6..468a9f055d 100644
--- a/drivers/net/intel/ice/base/ice_switch.c
+++ b/drivers/net/intel/ice/base/ice_switch.c
@@ -8190,7 +8190,6 @@ ice_add_adv_recipe(struct ice_hw *hw, struct
ice_adv_lkup_elem *lkups,
struct ice_sw_recipe *rm;
u8 i;
int status = ICE_SUCCESS;
- u16 cnt;
if (!ice_is_prof_rule(rinfo->tun_type) && !lkups_cnt)
return ICE_ERR_PARAM;
@@ -9795,7 +9794,6 @@ ice_rem_adv_rule(struct ice_hw *hw, struct
ice_adv_lkup_elem *lkups,
bool remove_rule = false;
struct ice_lock *rule_lock; /* Lock to protect filter rule list */
u16 i, rid, vsi_handle;
- bool is_add = false;
int status = ICE_SUCCESS;
ice_memset(&lkup_exts, 0, sizeof(lkup_exts), ICE_NONDMA_MEM);
diff --git a/drivers/net/intel/ice/base/ice_type.h
b/drivers/net/intel/ice/base/ice_type.h
index 35f832eb9f..297a5ea890 100644
--- a/drivers/net/intel/ice/base/ice_type.h
+++ b/drivers/net/intel/ice/base/ice_type.h
@@ -153,7 +153,7 @@ static inline u32 ice_round_to_num(u32 N, u32 R)
#define ICE_DBG_USER BIT_ULL(31)
#define ICE_DBG_ALL0xULL
-#define __ALWAYS_UNUSED
+#define __ALWAYS_UNUSED __rte_unused
#define IS_ETHER_ADDR_EQUAL(addr1, addr2) \
(((bool)u16 *)(addr1))[0] == ((u16 *)(addr2))[0]))) && \
diff --git a/drivers/net/intel/ice/base/meson.build
b/drivers/net/intel/ice/base/meson.build
index 38d092c370..e7ba9c34bc 100644
--- a/drivers/net/intel/ice/base/meson.build
+++ b/drivers/net/intel/ice/base/meson.build
@@ -39,7 +39,6 @@ if is_ms_compiler
else
error_cflags = [
'-Wno-unused-but-set-variable',
-'-Wno-unused-variable',
'-Wno-unused-parameter',
]
endif
diff --git a/drivers/net/intel/ice/ice_fdir_filter.c
b/drivers/net/intel/ice/ice_fdir_filter.c
index 940fe171b6..2ff0090aca 100644
--- a/drivers/net/intel/ice/ice_fdir_filter.c
+++ b/drivers/net/intel/ice/ice_fdir_filter.c
@@ -1090,7 +1090,7 @@ ice_fdir_input_set_conf(struct ice_pf *pf, enum
ice_fltr_ptype flow,
return -EINVAL;
seg_tun = (struct ice_flow_seg_info *)
- ice_malloc(hw, sizeof(*seg_tun) * ICE_FD_HW_SEG_MAX);
+ ice_malloc(pf->adapter->hw, sizeof(*seg_tun) *
ICE_FD_HW_SEG_MAX);
if (!seg_tun) {
PMD_DRV_LOG(ERR, "No memory can be allocated");
return -ENOMEM;
--
2.45.2
[PATCH 7/7] net/intel: simplify base code builds
Now that base-code warning flags have been removed from a number of
drivers, we can simplify their build configuration by just including the
base code files in the regular list of driver files passed back. There
is no need to use pre-compiled objects.
Signed-off-by: Bruce Richardson
---
drivers/net/intel/e1000/base/meson.build | 9 ++---
drivers/net/intel/e1000/meson.build | 4 ++--
drivers/net/intel/fm10k/base/meson.build | 9 ++---
drivers/net/intel/fm10k/meson.build | 4 ++--
drivers/net/intel/i40e/base/meson.build | 9 ++---
drivers/net/intel/i40e/meson.build | 4 ++--
drivers/net/intel/idpf/base/meson.build | 2 +-
drivers/net/intel/idpf/meson.build | 7 ---
drivers/net/intel/ixgbe/base/meson.build | 9 ++---
drivers/net/intel/ixgbe/meson.build | 4 ++--
10 files changed, 21 insertions(+), 40 deletions(-)
diff --git a/drivers/net/intel/e1000/base/meson.build
b/drivers/net/intel/e1000/base/meson.build
index 8cbd9f62e6..4fe86dc6df 100644
--- a/drivers/net/intel/e1000/base/meson.build
+++ b/drivers/net/intel/e1000/base/meson.build
@@ -1,7 +1,7 @@
# SPDX-License-Identifier: BSD-3-Clause
# Copyright(c) 2017 Intel Corporation
-sources = [
+base_sources = files(
'e1000_base.c',
'e1000_80003es2lan.c',
'e1000_82540.c',
@@ -21,9 +21,4 @@ sources = [
'e1000_osdep.c',
'e1000_phy.c',
'e1000_vf.c',
-]
-
-base_lib = static_library('e1000_base', sources,
-dependencies: static_rte_eal,
-c_args: cflags)
-base_objs = base_lib.extract_all_objects(recursive: true)
+)
diff --git a/drivers/net/intel/e1000/meson.build
b/drivers/net/intel/e1000/meson.build
index cd42c0042a..b52a843228 100644
--- a/drivers/net/intel/e1000/meson.build
+++ b/drivers/net/intel/e1000/meson.build
@@ -2,9 +2,9 @@
# Copyright(c) 2017 Intel Corporation
subdir('base')
-objs = [base_objs]
-sources = files(
+sources += base_sources
+sources += files(
'e1000_logs.c',
'em_ethdev.c',
'em_rxtx.c',
diff --git a/drivers/net/intel/fm10k/base/meson.build
b/drivers/net/intel/fm10k/base/meson.build
index a2640d1ee8..c45b31d6b2 100644
--- a/drivers/net/intel/fm10k/base/meson.build
+++ b/drivers/net/intel/fm10k/base/meson.build
@@ -1,16 +1,11 @@
# SPDX-License-Identifier: BSD-3-Clause
# Copyright(c) 2017 Intel Corporation
-sources = [
+base_sources = files(
'fm10k_api.c',
'fm10k_common.c',
'fm10k_mbx.c',
'fm10k_pf.c',
'fm10k_tlv.c',
'fm10k_vf.c',
-]
-
-base_lib = static_library('fm10k_base', sources,
-dependencies: static_rte_eal,
-c_args: cflags)
-base_objs = base_lib.extract_all_objects(recursive: true)
+)
diff --git a/drivers/net/intel/fm10k/meson.build
b/drivers/net/intel/fm10k/meson.build
index 69566add96..e08a00cb49 100644
--- a/drivers/net/intel/fm10k/meson.build
+++ b/drivers/net/intel/fm10k/meson.build
@@ -8,9 +8,9 @@ if is_windows
endif
subdir('base')
-objs = [base_objs]
-sources = files(
+sources += base_sources
+sources += files(
'fm10k_ethdev.c',
'fm10k_rxtx.c',
)
diff --git a/drivers/net/intel/i40e/base/meson.build
b/drivers/net/intel/i40e/base/meson.build
index 766383101b..7b4882ad1d 100644
--- a/drivers/net/intel/i40e/base/meson.build
+++ b/drivers/net/intel/i40e/base/meson.build
@@ -1,7 +1,7 @@
# SPDX-License-Identifier: BSD-3-Clause
# Copyright(c) 2017-2020 Intel Corporation
-sources = [
+base_sources = files(
'i40e_adminq.c',
'i40e_common.c',
'i40e_dcb.c',
@@ -9,9 +9,4 @@ sources = [
'i40e_hmc.c',
'i40e_lan_hmc.c',
'i40e_nvm.c',
-]
-
-base_lib = static_library('i40e_base', sources,
-dependencies: static_rte_eal,
-c_args: cflags)
-base_objs = base_lib.extract_all_objects(recursive: true)
+)
diff --git a/drivers/net/intel/i40e/meson.build
b/drivers/net/intel/i40e/meson.build
index 15993393fb..17b6715cd9 100644
--- a/drivers/net/intel/i40e/meson.build
+++ b/drivers/net/intel/i40e/meson.build
@@ -18,9 +18,9 @@ if dpdk_conf.get('RTE_IOVA_IN_MBUF') == 0
endif
subdir('base')
-objs = [base_objs]
-sources = files(
+sources += base_sources
+sources += files(
'i40e_ethdev.c',
'i40e_rxtx.c',
'i40e_pf.c',
diff --git a/drivers/net/intel/idpf/base/meson.build
b/drivers/net/intel/idpf/base/meson.build
index 7316e0a805..78782e463e 100644
--- a/drivers/net/intel/idpf/base/meson.build
+++ b/drivers/net/intel/idpf/base/meson.build
@@ -1,7 +1,7 @@
# SPDX-License-Identifier: BSD-3-Clause
# Copyright(c) 2023 Intel Corporation
-sources += files(
+base_sources = files(
'idpf_controlq.c',
'idpf_controlq_setup.c',
)
diff --git a/drivers/net/intel/idpf/meson.build
b/drivers/net/intel/idpf/meson.build
index 4b272d02b1..a8690da87b 100644
--- a/drivers/net/intel/idpf/meson.build
+++ b/drivers/net/intel/idpf/meson.build
@@ -7,9 +7,12 @@ if is_windows
subdir_done()
endif
+subdir('base')
+
includes += incl
[PATCH 4/7] net/i40e/base: fix unused value warnings
Fix warnings about unused values - parameters, variables, etc., and
remove the warning disable flags for them. Although modifying the
base-code files is not ideal, the changes required are minor, and only
affect two files from the imported base code.
Fixes: 8db9e2a1b232 ("i40e: base driver")
Cc: sta...@dpdk.org
Signed-off-by: Bruce Richardson
---
drivers/net/intel/i40e/base/i40e_nvm.c | 2 +-
drivers/net/intel/i40e/base/i40e_osdep.h | 4 ++--
drivers/net/intel/i40e/base/i40e_type.h | 14 +-
drivers/net/intel/i40e/base/meson.build | 3 ---
drivers/net/intel/i40e/i40e_ethdev.c | 1 +
5 files changed, 13 insertions(+), 11 deletions(-)
diff --git a/drivers/net/intel/i40e/base/i40e_nvm.c
b/drivers/net/intel/i40e/base/i40e_nvm.c
index 3e16a0d997..56dc4d9279 100644
--- a/drivers/net/intel/i40e/base/i40e_nvm.c
+++ b/drivers/net/intel/i40e/base/i40e_nvm.c
@@ -1743,7 +1743,7 @@ STATIC enum i40e_status_code
i40e_nvmupd_get_aq_result(struct i40e_hw *hw,
**/
STATIC enum i40e_status_code i40e_nvmupd_get_aq_event(struct i40e_hw *hw,
struct i40e_nvm_access *cmd,
- u8 *bytes, int *perrno)
+ u8 *bytes, __rte_unused int
*perrno)
{
u32 aq_total_len;
u32 aq_desc_len;
diff --git a/drivers/net/intel/i40e/base/i40e_osdep.h
b/drivers/net/intel/i40e/base/i40e_osdep.h
index c04f94732a..197f4678bf 100644
--- a/drivers/net/intel/i40e/base/i40e_osdep.h
+++ b/drivers/net/intel/i40e/base/i40e_osdep.h
@@ -184,8 +184,8 @@ struct __rte_packed_begin i40e_dma_mem {
const void *zone;
} __rte_packed_end;
-#define i40e_allocate_dma_mem(h, m, unused, s, a) \
- i40e_allocate_dma_mem_d(h, m, s, a)
+#define i40e_allocate_dma_mem(h, m, mt, s, a) \
+ i40e_allocate_dma_mem_d(h, m, mt, s, a)
#define i40e_free_dma_mem(h, m) i40e_free_dma_mem_d(h, m)
struct __rte_packed_begin i40e_virt_mem {
diff --git a/drivers/net/intel/i40e/base/i40e_type.h
b/drivers/net/intel/i40e/base/i40e_type.h
index 7cc746f82f..968e1982a6 100644
--- a/drivers/net/intel/i40e/base/i40e_type.h
+++ b/drivers/net/intel/i40e/base/i40e_type.h
@@ -14,11 +14,15 @@
#include "i40e_devids.h"
#define UNREFERENCED_XPARAMETER
-#define UNREFERENCED_1PARAMETER(_p) (_p);
-#define UNREFERENCED_2PARAMETER(_p, _q) (_p); (_q);
-#define UNREFERENCED_3PARAMETER(_p, _q, _r) (_p); (_q); (_r);
-#define UNREFERENCED_4PARAMETER(_p, _q, _r, _s) (_p); (_q); (_r); (_s);
-#define UNREFERENCED_5PARAMETER(_p, _q, _r, _s, _t) (_p); (_q); (_r); (_s);
(_t);
+#define UNREFERENCED_1PARAMETER(_p) (void)(_p)
+#define UNREFERENCED_2PARAMETER(_p, _q) \
+ do { (void)(_p); (void)(_q); } while (0)
+#define UNREFERENCED_3PARAMETER(_p, _q, _r) \
+ do { (void)(_p); (void)(_q); (void)(_r); } while (0)
+#define UNREFERENCED_4PARAMETER(_p, _q, _r, _s) \
+ do { (void)(_p); (void)(_q); (void)(_r); (void)(_s); } while (0)
+#define UNREFERENCED_5PARAMETER(_p, _q, _r, _s, _t) \
+ do { (void)(_p); (void)(_q); (void)(_r); (void)(_s); (void)(_t); }
while (0)
#define BIT(a) (1UL << (a))
#define BIT_ULL(a) (1ULL << (a))
diff --git a/drivers/net/intel/i40e/base/meson.build
b/drivers/net/intel/i40e/base/meson.build
index a0912b1788..2648e5d0c4 100644
--- a/drivers/net/intel/i40e/base/meson.build
+++ b/drivers/net/intel/i40e/base/meson.build
@@ -13,10 +13,7 @@ sources = [
error_cflags = [
'-Wno-sign-compare',
-'-Wno-unused-value',
'-Wno-strict-aliasing',
-'-Wno-unused-but-set-variable',
-'-Wno-unused-parameter',
]
c_args = cflags
foreach flag: error_cflags
diff --git a/drivers/net/intel/i40e/i40e_ethdev.c
b/drivers/net/intel/i40e/i40e_ethdev.c
index 1c5ab35a8b..90eba3419f 100644
--- a/drivers/net/intel/i40e/i40e_ethdev.c
+++ b/drivers/net/intel/i40e/i40e_ethdev.c
@@ -4694,6 +4694,7 @@ i40e_dev_rss_reta_query(struct rte_eth_dev *dev,
enum i40e_status_code
i40e_allocate_dma_mem_d(__rte_unused struct i40e_hw *hw,
struct i40e_dma_mem *mem,
+ __rte_unused enum i40e_memory_type mtype,
u64 size,
u32 alignment)
{
--
2.45.2
[PATCH 5/7] net/i40e/base: fix compiler warnings
Add a single-line fix to the base code, and then the remaining two
compiler warning disable flags can be removed from the driver base code
build file.
Fixes: 8db9e2a1b232 ("i40e: base driver")
Cc: sta...@dpdk.org
Signed-off-by: Bruce Richardson
---
drivers/net/intel/i40e/base/i40e_diag.c | 2 +-
drivers/net/intel/i40e/base/meson.build | 13 +
2 files changed, 2 insertions(+), 13 deletions(-)
diff --git a/drivers/net/intel/i40e/base/i40e_diag.c
b/drivers/net/intel/i40e/base/i40e_diag.c
index 4ca102cdd5..71b2e53e85 100644
--- a/drivers/net/intel/i40e/base/i40e_diag.c
+++ b/drivers/net/intel/i40e/base/i40e_diag.c
@@ -34,7 +34,7 @@ static enum i40e_status_code
i40e_diag_reg_pattern_test(struct i40e_hw *hw,
{
const u32 patterns[] = {0x5A5A5A5A, 0xA5A5A5A5, 0x, 0x};
u32 pat, val, orig_val;
- int i;
+ unsigned int i;
orig_val = rd32(hw, reg);
for (i = 0; i < ARRAY_SIZE(patterns); i++) {
diff --git a/drivers/net/intel/i40e/base/meson.build
b/drivers/net/intel/i40e/base/meson.build
index 2648e5d0c4..766383101b 100644
--- a/drivers/net/intel/i40e/base/meson.build
+++ b/drivers/net/intel/i40e/base/meson.build
@@ -11,18 +11,7 @@ sources = [
'i40e_nvm.c',
]
-error_cflags = [
-'-Wno-sign-compare',
-'-Wno-strict-aliasing',
-]
-c_args = cflags
-foreach flag: error_cflags
-if cc.has_argument(flag)
-c_args += flag
-endif
-endforeach
-
base_lib = static_library('i40e_base', sources,
dependencies: static_rte_eal,
-c_args: c_args)
+c_args: cflags)
base_objs = base_lib.extract_all_objects(recursive: true)
--
2.45.2
Re: [RFC PATCH] build: reduce use of AVX compiler flags
On Tue, Mar 25, 2025 at 05:22:15PM +, Bruce Richardson wrote:
> When doing a build for a target that already has the instruction sets
> for AVX2/AVX512 enabled, skip emitting the AVX compiler flags, or the
> skylake-avx512 '-march' flags, as they are unnecessary. Instead, when
> the default flags produce the desired output, just use them unmodified.
>
> Depends-on: series-34915 ("remove component-specific logic for AVX builds")
>
> Signed-off-by: Bruce Richardson
> ---
>
> This patchset depends on the previous AVX rework. However, sending it
> separately as a new RFC because it effectively increases the minimum
> compiler versions needed for x86 builds - from GCC 5 to 6, and
> Clang 3.6 to 3.9.
>
> For now, I've just documented that as an additional note in the GSG that
> these versions are recommended, but it would be simpler if we could just
> set them as the required minimum baseline (at least in the docs).
>
> Feedback on these compiler version requirements welcome.
>
+techboard
Ping for a little bit of feedback for this. Are we ok to bump the minimum
compiler versions as described above, or will I continue with the approach
in this RFC of keeping the minimum and just recommending the higher
versions for x86 platforms?
For reference GCC 6.1 was released April 2016[1], and, Clang 3.9 was
released Sept 2016[2]
/Bruce
[1] https://gcc.gnu.org/gcc-6/
[2] https://releases.llvm.org/
[PATCH v2 0/8] net/intel: clean up base code build
Simplify the build of the various intel base code directories, by clearing compiler warnings so the files can be compiled directly along with the regular driver files. v2: added missing patch for fm10k Bruce Richardson (8): net/fm10k/base: fix compilation warnings net/iavf/base: remove unused meson.build file net/ixgbe/base: correct definition of macro net/ixgbe/base: fix compilation warnings net/i40e/base: fix unused value warnings net/i40e/base: fix compiler warnings net/ice/base: reduce warnings for unused variables net/intel: simplify base code builds drivers/net/intel/e1000/base/meson.build| 9 ++-- drivers/net/intel/e1000/meson.build | 4 ++-- drivers/net/intel/fm10k/base/fm10k_mbx.c| 2 +- drivers/net/intel/fm10k/base/fm10k_osdep.h | 2 +- drivers/net/intel/fm10k/base/fm10k_pf.c | 8 +++ drivers/net/intel/fm10k/base/fm10k_type.h | 6 +++--- drivers/net/intel/fm10k/base/meson.build| 21 ++- drivers/net/intel/fm10k/meson.build | 4 ++-- drivers/net/intel/i40e/base/i40e_diag.c | 2 +- drivers/net/intel/i40e/base/i40e_nvm.c | 2 +- drivers/net/intel/i40e/base/i40e_osdep.h| 4 ++-- drivers/net/intel/i40e/base/i40e_type.h | 14 - drivers/net/intel/i40e/base/meson.build | 23 ++--- drivers/net/intel/i40e/i40e_ethdev.c| 1 + drivers/net/intel/i40e/meson.build | 4 ++-- drivers/net/intel/iavf/base/meson.build | 10 - drivers/net/intel/ice/base/ice_osdep.h | 6 +++--- drivers/net/intel/ice/base/ice_switch.c | 2 -- drivers/net/intel/ice/base/ice_type.h | 2 +- drivers/net/intel/ice/base/meson.build | 1 - drivers/net/intel/ice/ice_fdir_filter.c | 2 +- drivers/net/intel/idpf/base/meson.build | 2 +- drivers/net/intel/idpf/meson.build | 7 --- drivers/net/intel/ixgbe/base/ixgbe_common.c | 4 ++-- drivers/net/intel/ixgbe/base/ixgbe_e610.c | 2 ++ drivers/net/intel/ixgbe/base/ixgbe_osdep.h | 21 +++ drivers/net/intel/ixgbe/base/meson.build| 20 ++ drivers/net/intel/ixgbe/meson.build | 4 ++-- 28 files changed, 65 insertions(+), 124 deletions(-) delete mode 100644 drivers/net/intel/iavf/base/meson.build -- 2.45.2
[PATCH] acl: fix build with GCC 15 on aarch64
Caught in OBS for Fedora Rawhide on aarch64:
[ 198s] In file included from ../lib/acl/acl_run_neon.h:7,
[ 198s] from ../lib/acl/acl_run_neon.c:5:
[ 198s] In function ‘alloc_completion’,
[ 198s] inlined from ‘acl_start_next_trie’ at
../lib/acl/acl_run.h:140:24,
[ 198s] inlined from ‘search_neon_4.isra’ at
../lib/acl/acl_run_neon.h:239:20:
[ 198s] ../lib/acl/acl_run.h:93:25: error: ‘cmplt’ may be used
uninitialized [-Werror=maybe-uninitialized]
[ 198s]93 | if (p[n].count == 0) {
[ 198s] | ^~
[ 198s] ../lib/acl/acl_run_neon.h: In function ‘search_neon_4.isra’:
[ 198s] ../lib/acl/acl_run_neon.h:230:27: note: ‘cmplt’ declared here
[ 198s] 230 | struct completion cmplt[4];
[ 198s] | ^
The code was resetting sequentially cmpl[].count at the exact index that
later call to alloc_completion uses.
While this code seems correct, GCC 15 does not understand this (probably
when applying some optimisations).
Instead, reset cmpl[].count all at once in acl_set_flow, and cleanup the
various vectorized implementations accordingly.
Bugzilla ID: 1678
Cc: sta...@dpdk.org
Signed-off-by: David Marchand
---
lib/acl/acl_run.h | 5 +
lib/acl/acl_run_altivec.h | 8 ++--
lib/acl/acl_run_avx2.h| 4 +---
lib/acl/acl_run_neon.h| 8 ++--
lib/acl/acl_run_scalar.c | 4 +---
lib/acl/acl_run_sse.h | 8 ++--
6 files changed, 13 insertions(+), 24 deletions(-)
diff --git a/lib/acl/acl_run.h b/lib/acl/acl_run.h
index 7f092413cd..9fd3e60021 100644
--- a/lib/acl/acl_run.h
+++ b/lib/acl/acl_run.h
@@ -176,6 +176,8 @@ acl_set_flow(struct acl_flow_data *flows, struct completion
*cmplt,
uint32_t cmplt_size, const uint8_t **data, uint32_t *results,
uint32_t data_num, uint32_t categories, const uint64_t *trans)
{
+ unsigned int i;
+
flows->num_packets = 0;
flows->started = 0;
flows->trie = 0;
@@ -187,6 +189,9 @@ acl_set_flow(struct acl_flow_data *flows, struct completion
*cmplt,
flows->data = data;
flows->results = results;
flows->trans = trans;
+
+ for (i = 0; i < cmplt_size; i++)
+ cmplt[i].count = 0;
}
typedef void (*resolve_priority_t)
diff --git a/lib/acl/acl_run_altivec.h b/lib/acl/acl_run_altivec.h
index 2d398ffded..d5ccdb94f0 100644
--- a/lib/acl/acl_run_altivec.h
+++ b/lib/acl/acl_run_altivec.h
@@ -199,10 +199,8 @@ search_altivec_8(const struct rte_acl_ctx *ctx, const
uint8_t **data,
acl_set_flow(&flows, cmplt, RTE_DIM(cmplt), data, results,
total_packets, categories, ctx->trans_table);
- for (n = 0; n < MAX_SEARCHES_ALTIVEC8; n++) {
- cmplt[n].count = 0;
+ for (n = 0; n < MAX_SEARCHES_ALTIVEC8; n++)
index_array[n] = acl_start_next_trie(&flows, parms, n, ctx);
- }
/* Check for any matches. */
acl_match_check_x4(0, ctx, parms, &flows, (uint64_t *)&index_array[0]);
@@ -270,10 +268,8 @@ search_altivec_4(const struct rte_acl_ctx *ctx, const
uint8_t **data,
acl_set_flow(&flows, cmplt, RTE_DIM(cmplt), data, results,
total_packets, categories, ctx->trans_table);
- for (n = 0; n < MAX_SEARCHES_ALTIVEC4; n++) {
- cmplt[n].count = 0;
+ for (n = 0; n < MAX_SEARCHES_ALTIVEC4; n++)
index_array[n] = acl_start_next_trie(&flows, parms, n, ctx);
- }
/* Check for any matches. */
acl_match_check_x4(0, ctx, parms, &flows, index_array);
diff --git a/lib/acl/acl_run_avx2.h b/lib/acl/acl_run_avx2.h
index 0b8967f22e..e069fb85b2 100644
--- a/lib/acl/acl_run_avx2.h
+++ b/lib/acl/acl_run_avx2.h
@@ -171,10 +171,8 @@ search_avx2x16(const struct rte_acl_ctx *ctx, const
uint8_t **data,
acl_set_flow(&flows, cmplt, RTE_DIM(cmplt), data, results,
total_packets, categories, ctx->trans_table);
- for (n = 0; n < RTE_DIM(cmplt); n++) {
- cmplt[n].count = 0;
+ for (n = 0; n < RTE_DIM(cmplt); n++)
index_array[n] = acl_start_next_trie(&flows, parms, n, ctx);
- }
t0 = _mm256_set_epi64x(index_array[5], index_array[4],
index_array[1], index_array[0]);
diff --git a/lib/acl/acl_run_neon.h b/lib/acl/acl_run_neon.h
index 63074f871d..3b9bd0cc39 100644
--- a/lib/acl/acl_run_neon.h
+++ b/lib/acl/acl_run_neon.h
@@ -172,10 +172,8 @@ search_neon_8(const struct rte_acl_ctx *ctx, const uint8_t
**data,
acl_set_flow(&flows, cmplt, RTE_DIM(cmplt), data, results,
total_packets, categories, ctx->trans_table);
- for (n = 0; n < 8; n++) {
- cmplt[n].count = 0;
+ for (n = 0; n < 8; n++)
index_array[n] = acl_start_next_trie(&flows, parms, n, ctx);
- }
/* Check for any matches. */
acl_match_check_x4(0, ctx, parms, &flows, &index_array[
[RFC 1/2] eventdev: introduce event vector adapter
From: Pavan Nikhilesh
The event vector adapter supports offloading creation of
event vectors by vectorizing objects (mbufs/ptrs/u64s).
Applications can create a vector adapter associated with
an event queue and enqueue objects to be vectorized.
When the vector reaches the configured size or when the timeout
is reached, the vector adapter will enqueue the vector to the
event queue.
Signed-off-by: Pavan Nikhilesh
---
config/rte_config.h | 1 +
lib/eventdev/event_vector_adapter_pmd.h | 87 +
lib/eventdev/eventdev_pmd.h | 36 ++
lib/eventdev/meson.build| 3 +
lib/eventdev/rte_event_vector_adapter.c | 444 ++
lib/eventdev/rte_event_vector_adapter.h | 469
lib/eventdev/rte_eventdev.c | 21 ++
lib/eventdev/rte_eventdev.h | 8 +
lib/eventdev/version.map| 13 +
9 files changed, 1082 insertions(+)
create mode 100644 lib/eventdev/event_vector_adapter_pmd.h
create mode 100644 lib/eventdev/rte_event_vector_adapter.c
create mode 100644 lib/eventdev/rte_event_vector_adapter.h
diff --git a/config/rte_config.h b/config/rte_config.h
index 86897de75e..9535c48d81 100644
--- a/config/rte_config.h
+++ b/config/rte_config.h
@@ -92,6 +92,7 @@
#define RTE_EVENT_CRYPTO_ADAPTER_MAX_INSTANCE 32
#define RTE_EVENT_ETH_TX_ADAPTER_MAX_INSTANCE 32
#define RTE_EVENT_DMA_ADAPTER_MAX_INSTANCE 32
+#define RTE_EVENT_VECTOR_ADAPTER_MAX_INSTANCE_PER_QUEUE 32
/* rawdev defines */
#define RTE_RAWDEV_MAX_DEVS 64
diff --git a/lib/eventdev/event_vector_adapter_pmd.h
b/lib/eventdev/event_vector_adapter_pmd.h
new file mode 100644
index 00..dab0350564
--- /dev/null
+++ b/lib/eventdev/event_vector_adapter_pmd.h
@@ -0,0 +1,87 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2025 Marvell International Ltd.
+ * All rights reserved.
+ */
+#ifndef __EVENT_VECTOR_ADAPTER_PMD_H__
+#define __EVENT_VECTOR_ADAPTER_PMD_H__
+/**
+ * @file
+ * RTE Event Vector Adapter API (PMD Side)
+ *
+ * @note
+ * This file provides implementation helpers for internal use by PMDs. They
+ * are not intended to be exposed to applications and are not subject to ABI
+ * versioning.
+ */
+#include "eventdev_pmd.h"
+#include "rte_event_vector_adapter.h"
+
+typedef int (*rte_event_vector_adapter_create_t)(struct
rte_event_vector_adapter *adapter);
+/**< @internal Event vector adapter implementation setup */
+typedef int (*rte_event_vector_adapter_destroy_t)(struct
rte_event_vector_adapter *adapter);
+/**< @internal Event vector adapter implementation teardown */
+typedef int (*rte_event_vector_adapter_caps_get_t)(struct rte_eventdev *dev);
+/**< @internal Get capabilities for event vector adapter */
+typedef int (*rte_event_vector_adapter_stats_get_t)(const struct
rte_event_vector_adapter *adapter,
+ struct
rte_event_vector_adapter_stats *stats);
+/**< @internal Get statistics for event vector adapter */
+typedef int (*rte_event_vector_adapter_stats_reset_t)(
+ const struct rte_event_vector_adapter *adapter);
+/**< @internal Reset statistics for event vector adapter */
+
+/**
+ * @internal Structure containing the functions exported by an event vector
+ * adapter implementation.
+ */
+struct event_vector_adapter_ops {
+ rte_event_vector_adapter_create_t create;
+ /**< Set up adapter */
+ rte_event_vector_adapter_destroy_t destroy;
+ /**< Tear down adapter */
+ rte_event_vector_adapter_caps_get_t caps_get;
+ /**< Get capabilities from driver */
+ rte_event_vector_adapter_stats_get_t stats_get;
+ /**< Get adapter statistics */
+ rte_event_vector_adapter_stats_reset_t stats_reset;
+ /**< Reset adapter statistics */
+
+ rte_event_vector_adapter_enqueue_t enqueue;
+ /**< Enqueue ptrs into the event vector adapter */
+};
+/**
+ * @internal Adapter data; structure to be placed in shared memory to be
+ * accessible by various processes in a multi-process configuration.
+ */
+struct __rte_cache_aligned rte_event_vector_adapter_data {
+ uint32_t id;
+ /**< Event vector adapter ID */
+ uint8_t event_dev_id;
+ /**< Event device ID */
+ uint32_t socket_id;
+ /**< Socket ID where memory is allocated */
+ uint8_t event_port_id;
+ /**< Optional: event port ID used when the inbuilt port is absent */
+ const struct rte_memzone *mz;
+ /**< Event vector adapter memzone pointer */
+ struct rte_event_vector_adapter_conf conf;
+ /**< Configuration used to configure the adapter. */
+ uint32_t caps;
+ /**< Adapter capabilities */
+ void *adapter_priv;
+ /**< Vector adapter private data*/
+ uint32_t unified_service_id;
+ /**< Unified Service ID*/
+};
+
+static int
+dummy_vector_adapter_enqueue(struct rte_event_vector_adapter *adapter,
uintptr_t ptrs[],
+uint16_t num_
Re: [RFC v4 0/8] Symbol versioning and export rework
On Wed, Mar 26, 2025 at 01:02:32PM +0100, David Marchand wrote: > On Mon, Mar 17, 2025 at 4:43 PM David Marchand > wrote: > > > > So far, each DPDK library (or driver) exposing symbols in an ABI had to > > maintain a version.map and use some macros for symbol versioning, > > specially crafted with the GNU linker in mind. > > > > This series proposes to rework the whole principle, and instead rely on > > marking the symbol exports in the source code itself, then let it to the > > build framework to produce a version script adapted to the linker in use > > (think GNU linker vs MSVC linker). > > > > This greatly simplifies versioning symbols: a developer does not need to > > know anything about version.map, or that a versioned symbol must be > > renamed with _v26, annotated with __vsym, exported in a header etc... > > > > Checking symbol maps becomes unnecessary since generated by the build > > framework. > > > > Updating to a new ABI is just a matter of bumping the value in > > ABI_VERSION. > > > > > > Comments please. > > - I am considering making rte_function_versioning.h a non exported > header (precisely, moving it to buildtools/ and maybe renaming it). > +1 for not exporting it. -1 for moving to buildtools. I don't see the need to introduce yet another header path in DPDK. Let's just keep it where it is, or moved slightly in the EAL folder, and then not export it. > This header contains macros not prefixed with RTE_. > Using it requires some build trick (see use_function_versioning). > And I don't see symbol versioning as a MUST infrastructure that DPDK > needs to provide to datapath applications. > > Yet technically, this change would be an API breakage if some > applications indeed relied on it. > > Cc: techboard for info. > > > - On a similar note, this RFC series adds the rte_exports.h header > (defining RTE_EXPORT*_SYMBOL()) in config/, though its job is for > extracting a symbol list during the build. > So a better location is probably buildtools/. > Again, don't particularly like buildtools as a path, as it's not really a tool, just a header file. I'd rather keep the tools folders for scripts and the like. /Bruce
[RFC 2/2] eventdev: add default software vector adapter
From: Pavan Nikhilesh
When event device PMD doesn't support vector adapter,
the library will fallback to software implementation
which relies on service core to check for timeouts
and vectorizes the objects on enqueue.
Signed-off-by: Pavan Nikhilesh
---
lib/eventdev/eventdev_pmd.h | 2 +
lib/eventdev/rte_event_vector_adapter.c | 318
lib/eventdev/rte_eventdev.c | 2 +
3 files changed, 322 insertions(+)
diff --git a/lib/eventdev/eventdev_pmd.h b/lib/eventdev/eventdev_pmd.h
index d03461316b..dda8ad82c9 100644
--- a/lib/eventdev/eventdev_pmd.h
+++ b/lib/eventdev/eventdev_pmd.h
@@ -87,6 +87,8 @@ extern int rte_event_logtype;
#define RTE_EVENT_TIMER_ADAPTER_SW_CAP \
RTE_EVENT_TIMER_ADAPTER_CAP_PERIODIC
+#define RTE_EVENT_VECTOR_ADAPTER_SW_CAP RTE_EVENT_VECTOR_ADAPTER_CAP_SOV_EOV
+
#define RTE_EVENTDEV_DETACHED (0)
#define RTE_EVENTDEV_ATTACHED (1)
diff --git a/lib/eventdev/rte_event_vector_adapter.c
b/lib/eventdev/rte_event_vector_adapter.c
index 5f38a9a40b..c1d29530be 100644
--- a/lib/eventdev/rte_event_vector_adapter.c
+++ b/lib/eventdev/rte_event_vector_adapter.c
@@ -21,6 +21,10 @@
#define MZ_NAME_MAX_LEN64
#define DATA_MZ_NAME_FORMAT "rte_event_vector_adapter_data_%d_%d_%d"
+#define MAX_VECTOR_SIZE1024
+#define MIN_VECTOR_SIZE1
+#define MAX_VECTOR_NS 1E9
+#define MIN_VECTOR_NS 1E5
RTE_LOG_REGISTER_SUFFIX(ev_vector_logtype, adapter.vector, NOTICE);
#define RTE_LOGTYPE_EVVEC ev_vector_logtype
@@ -46,6 +50,9 @@ struct rte_event_vector_adapter
*adapters[RTE_EVENT_MAX_DEVS][RTE_EVENT_MAX_QUEU
}
\
} while (0)
+static const struct event_vector_adapter_ops sw_ops;
+static const struct rte_event_vector_adapter_info sw_info;
+
static int
validate_conf(const struct rte_event_vector_adapter_conf *conf,
struct rte_event_vector_adapter_info *info)
@@ -229,6 +236,11 @@ rte_event_vector_adapter_create_ext(const struct
rte_event_vector_adapter_conf *
}
}
+ if (adapter->ops == NULL) {
+ adapter->ops = &sw_ops;
+ info = sw_info;
+ }
+
rc = validate_conf(conf, &info);
if (rc < 0) {
adapter->ops = NULL;
@@ -338,6 +350,8 @@ rte_event_vector_adapter_lookup(uint32_t adapter_id)
return NULL;
}
}
+ if (adapter->ops == NULL)
+ adapter->ops = &sw_ops;
adapter->enqueue = adapter->ops->enqueue;
adapter->adapter_id = adapter_id;
@@ -384,6 +398,7 @@ rte_event_vector_adapter_info_get(uint8_t event_dev_id,
struct rte_event_vector_
if (dev->dev_ops->vector_adapter_info_get != NULL)
return dev->dev_ops->vector_adapter_info_get(dev, info);
+ *info = sw_info;
return 0;
}
@@ -442,3 +457,306 @@ rte_event_vector_adapter_stats_reset(struct
rte_event_vector_adapter *adapter)
return 0;
}
+
+/* Software vector adapter implementation. */
+
+struct sw_vector_adapter_service_data;
+struct sw_vector_adapter_data {
+ uint8_t dev_id;
+ uint8_t port_id;
+ uint16_t vector_sz;
+ uint64_t timestamp;
+ uint64_t event_meta;
+ uint64_t vector_tmo_ticks;
+ uint64_t fallback_event_meta;
+ struct rte_mempool *vector_mp;
+ struct rte_event_vector *vector;
+ RTE_ATOMIC(rte_mcslock_t *) lock;
+ struct rte_event_vector_adapter *adapter;
+ struct rte_event_vector_adapter_stats stats;
+ struct sw_vector_adapter_service_data *service_data;
+ RTE_TAILQ_ENTRY(sw_vector_adapter_data) next;
+};
+
+struct sw_vector_adapter_service_data {
+ uint32_t service_id;
+ RTE_ATOMIC(rte_mcslock_t *) lock;
+ RTE_TAILQ_HEAD(, sw_vector_adapter_data) adapter_list;
+};
+
+static inline struct sw_vector_adapter_data *
+sw_vector_adapter_priv(const struct rte_event_vector_adapter *adapter)
+{
+ return adapter->data->adapter_priv;
+}
+
+static int
+sw_vector_adapter_flush(struct sw_vector_adapter_data *sw)
+{
+ struct rte_event ev;
+
+ if (sw->vector == NULL)
+ return -ENOBUFS;
+
+ ev.event = sw->event_meta;
+ ev.vec = sw->vector;
+ if (rte_event_enqueue_burst(sw->dev_id, sw->port_id, &ev, 1) != 1)
+ return -ENOSPC;
+
+ sw->vector = NULL;
+ sw->timestamp = 0;
+ return 0;
+}
+
+static int
+sw_vector_adapter_service_func(void *arg)
+{
+ struct sw_vector_adapter_service_data *service_data = arg;
+ struct sw_vector_adapter_data *sw, *nextsw;
+ rte_mcslock_t me, me_adptr;
+ int ret;
+
+ rte_mcslock_lock(&service_data->lock, &me);
+ RTE_TAILQ_FOREACH_SAFE(sw, &service_data->adapter_list, next, nextsw)
+ {
+ if (!rte_mcslock_trylock(&sw->lock, &me_adptr))
+
[RFC 0/2] introduce event vector adapter
From: Pavan Nikhilesh The event vector adapter supports offloading the creation of event vectors by vectorizing objects (mbufs/ptrs/u64s). An event vector adapter has the following working model: ┌──┐ │ Vector ├─┐ │ adapter0 │ │ └──┘ │ ┌──┐ │ ┌──┐ │ Vector ├─┼──►│ Event │ │ adapter1 │ │ │ Queue0 │ └──┘ │ └──┘ ┌──┐ │ │ Vector ├─┘ │ adapter2 │ └──┘ ┌──┐ │ Vector ├─┐ │ adapter0 │ │ ┌──┐ └──┘ ├──►│ Event │ ┌──┐ │ │ Queue1 │ │ Vector ├─┘ └──┘ │ adapter1 │ └──┘ - A vector adapter can be seen as an extension to event queue. It helps in aggregating objects and generating a vector event which is enqueued to the event queue. - Multiple vector adapters can be created on an event queue, each with its own unique properties such as event properties, vector size, and timeout. Note: If the target event queue doesn't support RTE_EVENT_QUEUE_CFG_ALL_TYPES, then the vector adapter should use the same schedule type as the event queue. - Each vector adapter aggregates objects, generates a vector event and enqueues it to the event queue with the event properties mentioned in rte_event_vector_adapter_conf::ev. - After configuring the vector adapter, the Application needs to use the rte_event_vector_adapter_enqueue() function to enqueue objects i.e., mbufs/ptrs/u64s to the vector adapter. On reaching the configured vector size or timeout, the vector adapter enqueues the event vector to the event queue. Note: Application should use the event_type and sub_event_type properly identifying the contents of vector event on dequeue. - If the vector adapter advertises the RTE_EVENT_VECTOR_ADAPTER_CAP_SOV_EOV capability, application can use the RTE_EVENT_VECTOR_ENQ_[S|E]OV flags to indicate the start and end of a vector event. * When RTE_EVENT_VECTOR_ENQ_SOV is set, the vector adapter will flush any aggregation in progress as a vector event and start aggregating a new vector event with the enqueued ptr. * When RTE_EVENT_VECTOR_ENQ_EOV is set, the vector adapter will add the current ptr enqueued to the aggregated event and enqueue the vector event to the event queue. * If both flags are set, the vector adapter will flush the current aggregation as a vector event and enqueue the current ptr as a single event to the event queue. - If the vector adapter reaches the configured vector size, it will enqueue the aggregated vector event to the event queue. - If the vector adapter reaches the configured vector timeout, it will flush the current aggregation as a vector event if the minimum vector size is reached, if not it will enqueue the objects as single events to the event queue. - If the vector adapter is unable to aggregate the objects into a vector event, it will enqueue the objects as single events to the event queue with the event properties mentioned in rte_event_vector_adapter_conf::ev_fallback. Before using the vector adapter, the application has to create and configure an event device and based on the event device capability it might require creating an additional event port. When the application creates the vector adapter using the ``rte_event_vector_adapter_create()`` function, the event device driver capabilities are checked. If an in-built port is absent, the application uses the default function to create a new event port. For finer control over event port creation, the application should use the ``rte_event_vector_adapter_create_ext()`` function. The application can enqueue one or more objects to the vector adapter using the ``rte_event_vector_adapter_enqueue()`` function and control the aggregation using the flags. Vector adapters report stats using the ``rte_event_vector_adapter_stats_get()`` function and reset the stats using the ``rte_event_vector_adapter_stats_reset()``. The application can destroy the vector adapter using the ``rte_event_vector_adapter_destroy()`` function. Pavan Nikhilesh (2): eventdev: introduce event vector adapter eventdev: add default software vector adapter config/rte_config.h | 1 + lib/eventdev/event_vector_adapter_pmd.h | 87 +++ lib/eventdev/eventdev_pmd.h | 38 ++ lib/eventdev/meson.build| 3 + lib/eventdev/rte_event_vector_adapter.c | 762 lib/eventdev/rte_event_vector_adapter.h | 469 +++ lib/eventdev/rte_eventdev.c | 23 + lib/eventdev/rte_eventdev.h | 8 + lib/eventdev/version.map| 13 + 9 files changed, 1404 insertions(+) create mode 100644 lib/eventdev/event_vector_adapter_pmd.h
[PATCH v1] maintainers: update for iavf and i40e
Removing myself as maintainer for iavf and i40e. Signed-off-by: Ian Stokes --- MAINTAINERS | 2 -- 1 file changed, 2 deletions(-) diff --git a/MAINTAINERS b/MAINTAINERS index 4b01103f8e..5f2f413da8 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -802,7 +802,6 @@ F: doc/guides/nics/intel_vf.rst F: doc/guides/nics/features/ixgbe*.ini Intel i40e -M: Ian Stokes M: Bruce Richardson T: git://dpdk.org/next/dpdk-next-net-intel F: drivers/net/intel/i40e/ @@ -818,7 +817,6 @@ F: doc/guides/nics/features/fm10k*.ini Intel iavf M: Vladimir Medvedkin -M: Ian Stokes T: git://dpdk.org/next/dpdk-next-net-intel F: drivers/net/intel/iavf/ F: doc/guides/nics/features/iavf*.ini -- 2.34.1
Re: [RFC v4 0/8] Symbol versioning and export rework
On Wed, Mar 26, 2025 at 01:02:32PM +0100, David Marchand wrote: > On Mon, Mar 17, 2025 at 4:43 PM David Marchand > wrote: > > > > So far, each DPDK library (or driver) exposing symbols in an ABI had to > > maintain a version.map and use some macros for symbol versioning, > > specially crafted with the GNU linker in mind. > > > > This series proposes to rework the whole principle, and instead rely on > > marking the symbol exports in the source code itself, then let it to the > > build framework to produce a version script adapted to the linker in use > > (think GNU linker vs MSVC linker). > > > > This greatly simplifies versioning symbols: a developer does not need to > > know anything about version.map, or that a versioned symbol must be > > renamed with _v26, annotated with __vsym, exported in a header etc... > > > > Checking symbol maps becomes unnecessary since generated by the build > > framework. > > > > Updating to a new ABI is just a matter of bumping the value in > > ABI_VERSION. > > > > > > Comments please. > > - I am considering making rte_function_versioning.h a non exported > header (precisely, moving it to buildtools/ and maybe renaming it). > > This header contains macros not prefixed with RTE_. > Using it requires some build trick (see use_function_versioning). > And I don't see symbol versioning as a MUST infrastructure that DPDK > needs to provide to datapath applications. > > Yet technically, this change would be an API breakage if some > applications indeed relied on it. > Is it not needed for exporting if an exported library header had versioned symbols is in? /Bruce
Re: [PATCH v1] maintainers: update for iavf and i40e
On Wed, Mar 26, 2025 at 01:30:10PM +, Ian Stokes wrote: > Removing myself as maintainer for iavf and i40e. > > Signed-off-by: Ian Stokes > --- Acked-by: Bruce Richardson
[DPDK/core Bug 1683] use after on interrupt thread during EAL cleanup
https://bugs.dpdk.org/show_bug.cgi?id=1683 Bug ID: 1683 Summary: use after on interrupt thread during EAL cleanup Product: DPDK Version: unspecified Hardware: All OS: All Status: UNCONFIRMED Severity: normal Priority: Normal Component: core Assignee: dev@dpdk.org Reporter: david.march...@redhat.com Target Milestone: --- This was caught with ASan in a CI run in my GHA: + devtools/test-null.sh EAL: Detected CPU lcores: 4 EAL: Detected NUMA nodes: 1 EAL: Detected static linkage of DPDK EAL: Multi-process socket /run/user/1001/dpdk/rte/mp_socket EAL: Selected IOVA mode 'VA' testpmd: create a new mbuf pool : n=2048, size=2176, socket=0 testpmd: preferred mempool ops selected: ring_mp_mc Interactive-mode selected Auto-start selected Configuring Port 0 (socket 0) ... Shutting down port 1... = ==46768==ERROR: AddressSanitizer: heap-use-after-free on address 0x60608e44 at pc 0x5613ec543091 bp 0x7f87065fd1d0 sp 0x7f87065fd1c8 READ of size 4 at 0x60608e44 thread T1 #0 0x5613ec543090 in rte_intr_fd_get /home/runner/work/dpdk/dpdk/build/../lib/eal/common/eal_common_interrupts.c:210:22 #1 0x5613ec5a350b in eal_alarm_callback /home/runner/work/dpdk/dpdk/build/../lib/eal/linux/eal_alarm.c:125:19 #2 0x5613ec5acef1 in eal_intr_process_interrupts /home/runner/work/dpdk/dpdk/build/../lib/eal/linux/eal_interrupts.c:1026:5 #3 0x5613ec5acef1 in eal_intr_handle_interrupts /home/runner/work/dpdk/dpdk/build/../lib/eal/linux/eal_interrupts.c:1100:7 #4 0x5613ec5aba06 in eal_intr_thread_main /home/runner/work/dpdk/dpdk/build/../lib/eal/linux/eal_interrupts.c:1172:3 #5 0x7f870b294ac2 in start_thread nptl/./nptl/pthread_create.c:442:8 #6 0x7f870b32684f misc/../sysdeps/unix/sysv/linux/x86_64/clone3.S:81 0x60608e44 is located 4 bytes inside of 64-byte region [0x60608e40,0x60608e80) freed by thread T0 here: #0 0x5613eb24ba32 in free (/home/runner/work/dpdk/dpdk/build/app/dpdk-testpmd+0xb8ca32) (BuildId: 1c1d93dcb4fec8b525f3ed4b97885f75ba17fcb1) #1 0x5613ec5a1b7f in rte_eal_cleanup /home/runner/work/dpdk/dpdk/build/../lib/eal/linux/eal.c:1333:2 #2 0x5613eb3bf7bc in main /home/runner/work/dpdk/dpdk/build/../app/test-pmd/testpmd.c:4583:8 #3 0x7f870b229d8f in __libc_start_call_main csu/../sysdeps/nptl/libc_start_call_main.h:58:16 previously allocated by thread T0 here: #0 0x5613eb24bec8 in __interceptor_calloc (/home/runner/work/dpdk/dpdk/build/app/dpdk-testpmd+0xb8cec8) (BuildId: 1c1d93dcb4fec8b525f3ed4b97885f75ba17fcb1) #1 0x5613ec542b56 in rte_intr_instance_alloc /home/runner/work/dpdk/dpdk/build/../lib/eal/common/eal_common_interrupts.c:51:17 #2 0x5613ec5a26ed in rte_eal_alarm_init /home/runner/work/dpdk/dpdk/build/../lib/eal/linux/eal_alarm.c:66:16 #3 0x5613ec59f5a3 in rte_eal_init /home/runner/work/dpdk/dpdk/build/../lib/eal/linux/eal.c:1011:6 #4 0x5613eb3be5b3 in main /home/runner/work/dpdk/dpdk/build/../app/test-pmd/testpmd.c:4369:9 #5 0x7f870b229d8f in __libc_start_call_main csu/../sysdeps/nptl/libc_start_call_main.h:58:16 Thread T1 created by T0 here: #0 0x5613eb23515c in __interceptor_pthread_create (/home/runner/work/dpdk/dpdk/build/app/dpdk-testpmd+0xb7615c) (BuildId: 1c1d93dcb4fec8b525f3ed4b97885f75ba17fcb1) #1 0x5613ec59d50c in rte_thread_create /home/runner/work/dpdk/dpdk/build/../lib/eal/unix/rte_thread.c:199:8 #2 0x5613ec56011b in rte_thread_create_control /home/runner/work/dpdk/dpdk/build/../lib/eal/common/eal_common_thread.c:308:8 #3 0x5613ec56096c in rte_thread_create_internal_control /home/runner/work/dpdk/dpdk/build/../lib/eal/common/eal_common_thread.c:358:9 #4 0x5613ec5ab811 in rte_eal_intr_init /home/runner/work/dpdk/dpdk/build/../lib/eal/linux/eal_interrupts.c:1200:8 #5 0x5613ec59f58a in rte_eal_init /home/runner/work/dpdk/dpdk/build/../lib/eal/linux/eal.c:1006:6 #6 0x5613eb3be5b3 in main /home/runner/work/dpdk/dpdk/build/../app/test-pmd/testpmd.c:4369:9 #7 0x7f870b229d8f in __libc_start_call_main csu/../sysdeps/nptl/libc_start_call_main.h:58:16 SUMMARY: AddressSanitizer: heap-use-after-free /home/runner/work/dpdk/dpdk/build/../lib/eal/common/eal_common_interrupts.c:210:22 in rte_intr_fd_get Shadow bytes around the buggy address: 0x0c0c7fff9170: fa fa fa fa fd fd fd fd fd fd fd fa fa fa fa fa 0x0c0c7fff9180: fd fd fd fd fd fd fd fa fa fa fa fa 00 00 00 00 0x0c0c7fff9190: 00 00 06 fa fa fa fa fa fd fd fd fd fd fd fd fa 0x0c0c7fff91a0: fa fa fa fa fd fd fd fd fd fd fd fa fa fa fa fa 0x0c0c7fff91b0: fd fd fd fd fd fd fd fd fa fa fa fa fd fd fd fd =>0x0c0c7fff91c0: fd fd fd fa fa fa fa fa[fd]fd fd fd fd fd fd fd 0x0c0c7fff91d0: fa fa fa fa 00 00 00 00 00 00 00 00 fa fa fa fa 0x0c0c7fff91e0: fd fd fd fd fd fd fd fa fa fa fa fa fd fd fd fd 0x0c0c7fff91f0: fd fd fd fd fa fa fa fa fd fd f
Re: [RFC v4 0/8] Symbol versioning and export rework
On Mon, Mar 17, 2025 at 4:43 PM David Marchand wrote: > > So far, each DPDK library (or driver) exposing symbols in an ABI had to > maintain a version.map and use some macros for symbol versioning, > specially crafted with the GNU linker in mind. > > This series proposes to rework the whole principle, and instead rely on > marking the symbol exports in the source code itself, then let it to the > build framework to produce a version script adapted to the linker in use > (think GNU linker vs MSVC linker). > > This greatly simplifies versioning symbols: a developer does not need to > know anything about version.map, or that a versioned symbol must be > renamed with _v26, annotated with __vsym, exported in a header etc... > > Checking symbol maps becomes unnecessary since generated by the build > framework. > > Updating to a new ABI is just a matter of bumping the value in > ABI_VERSION. > > > Comments please. - I am considering making rte_function_versioning.h a non exported header (precisely, moving it to buildtools/ and maybe renaming it). This header contains macros not prefixed with RTE_. Using it requires some build trick (see use_function_versioning). And I don't see symbol versioning as a MUST infrastructure that DPDK needs to provide to datapath applications. Yet technically, this change would be an API breakage if some applications indeed relied on it. Cc: techboard for info. - On a similar note, this RFC series adds the rte_exports.h header (defining RTE_EXPORT*_SYMBOL()) in config/, though its job is for extracting a symbol list during the build. So a better location is probably buildtools/. -- David Marchand
[PATCH 3/7] net/ixgbe/base: fix compilation warnings
We can remove almost all of the "unused parameter" and "unused variable"
warnings by just improving the macro definitions in the osdep.h header.
Remaining two instances can be fixed by just one-line additions to the
code, so add those to give us a clean build with the warnings enabled.
Fixes: af75078fece3 ("first public release")
Fixes: c6cb313da739 ("net/ixgbe/base: add link management for E610")
Cc: sta...@dpdk.org
Signed-off-by: Bruce Richardson
---
drivers/net/intel/ixgbe/base/ixgbe_e610.c | 2 ++
drivers/net/intel/ixgbe/base/ixgbe_osdep.h | 19 +++
drivers/net/intel/ixgbe/base/meson.build | 11 ---
3 files changed, 13 insertions(+), 19 deletions(-)
diff --git a/drivers/net/intel/ixgbe/base/ixgbe_e610.c
b/drivers/net/intel/ixgbe/base/ixgbe_e610.c
index 5474c3012a..7420c78d07 100644
--- a/drivers/net/intel/ixgbe/base/ixgbe_e610.c
+++ b/drivers/net/intel/ixgbe/base/ixgbe_e610.c
@@ -1054,6 +1054,7 @@ static void ixgbe_parse_vsi_func_caps(struct ixgbe_hw *hw,
struct ixgbe_hw_func_caps *func_p,
struct ixgbe_aci_cmd_list_caps_elem *cap)
{
+ UNREFERENCED_PARAMETER(cap);
func_p->guar_num_vsi = ixgbe_get_num_per_func(hw, IXGBE_MAX_VSI);
}
@@ -1770,6 +1771,7 @@ s32 ixgbe_aci_set_event_mask(struct ixgbe_hw *hw, u8
port_num, u16 mask)
struct ixgbe_aci_cmd_set_event_mask *cmd;
struct ixgbe_aci_desc desc;
+ UNREFERENCED_PARAMETER(port_num);
cmd = &desc.params.set_event_mask;
ixgbe_fill_dflt_direct_cmd_desc(&desc, ixgbe_aci_opc_set_event_mask);
diff --git a/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
b/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
index 6e5f7b4ae8..398c38bffd 100644
--- a/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
+++ b/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
@@ -57,13 +57,16 @@
/* Bunch of defines for shared code bogosity */
#ifndef UNREFERENCED_PARAMETER
-#define UNREFERENCED_PARAMETER(_p)
+#define UNREFERENCED_PARAMETER(_p) (void)(_p)
#endif
-#define UNREFERENCED_1PARAMETER(_p)
-#define UNREFERENCED_2PARAMETER(_p, _q)
-#define UNREFERENCED_3PARAMETER(_p, _q, _r)
-#define UNREFERENCED_4PARAMETER(_p, _q, _r, _s)
-#define UNREFERENCED_5PARAMETER(_p, _q, _r, _s, _t)
+#define UNREFERENCED_1PARAMETER(_p) (void)(_p)
+#define UNREFERENCED_2PARAMETER(_p, _q) do { (void)(_p); (void)(_q); } while(0)
+#define UNREFERENCED_3PARAMETER(_p, _q, _r) \
+ do { (void)(_p); (void)(_q); (void)(_r); } while(0)
+#define UNREFERENCED_4PARAMETER(_p, _q, _r, _s) \
+ do { (void)(_p); (void)(_q); (void)(_r); (void)(_s); } while(0)
+#define UNREFERENCED_5PARAMETER(_p, _q, _r, _s, _t) \
+ do { (void)(_p); (void)(_q); (void)(_r); (void)(_s); (void)(_t); }
while(0)
/* Shared code error reporting */
enum {
@@ -130,8 +133,8 @@ static inline uint32_t ixgbe_read_addr(volatile void* addr)
IXGBE_PCI_REG_ADDR((hw), (reg) + ((index) << 2))
/* Not implemented !! */
-#define IXGBE_READ_PCIE_WORD(hw, reg) 0
-#define IXGBE_WRITE_PCIE_WORD(hw, reg, value) do { } while(0)
+#define IXGBE_READ_PCIE_WORD(hw, reg) ((void)hw, (void)(reg), 0)
+#define IXGBE_WRITE_PCIE_WORD(hw, reg, value) do { (void)hw; (void)reg;
(void)value; } while(0)
#define IXGBE_WRITE_FLUSH(a) IXGBE_READ_REG(a, IXGBE_STATUS)
diff --git a/drivers/net/intel/ixgbe/base/meson.build
b/drivers/net/intel/ixgbe/base/meson.build
index f8b2ee6341..64e0bfd7be 100644
--- a/drivers/net/intel/ixgbe/base/meson.build
+++ b/drivers/net/intel/ixgbe/base/meson.build
@@ -19,17 +19,6 @@ sources = [
'ixgbe_x550.c',
]
-error_cflags = [
-'-Wno-unused-but-set-variable',
-'-Wno-unused-parameter',
-]
-c_args = cflags
-foreach flag: error_cflags
-if cc.has_argument(flag)
-c_args += flag
-endif
-endforeach
-
base_lib = static_library('ixgbe_base', sources,
dependencies: [static_rte_eal, static_rte_net],
c_args: c_args)
--
2.45.2
[PATCH v2 1/8] net/fm10k/base: fix compilation warnings
The fixes required to re-enable warnings in the fm10k base code are
trivial, so let's make the changes and get a clean compile without any
warning disable flags.
* provide definitions for the UNREFERENCED_PARAMETER macros
* fix the spelling of the work "fallthrough" in comments
* provide a definition of FM10K_READ_PCI_WORD in os_dep.h that marks the
parameters as used.
Fixes: 7223d200c227 ("fm10k: add base driver")
Cc: sta...@dpdk.org
Signed-off-by: Bruce Richardson
---
drivers/net/intel/fm10k/base/fm10k_mbx.c | 2 +-
drivers/net/intel/fm10k/base/fm10k_osdep.h | 2 +-
drivers/net/intel/fm10k/base/fm10k_pf.c| 8
drivers/net/intel/fm10k/base/fm10k_type.h | 6 +++---
drivers/net/intel/fm10k/base/meson.build | 14 +-
5 files changed, 10 insertions(+), 22 deletions(-)
diff --git a/drivers/net/intel/fm10k/base/fm10k_mbx.c
b/drivers/net/intel/fm10k/base/fm10k_mbx.c
index 2bb0d82efe..9028403757 100644
--- a/drivers/net/intel/fm10k/base/fm10k_mbx.c
+++ b/drivers/net/intel/fm10k/base/fm10k_mbx.c
@@ -1602,7 +1602,7 @@ s32 fm10k_pfvf_mbx_init(struct fm10k_hw *hw, struct
fm10k_mbx_info *mbx,
mbx->mbmem_reg = FM10K_MBMEM_VF(id, 0);
break;
}
- /* fallthough */
+ /* fallthrough */
default:
return FM10K_MBX_ERR_NO_MBX;
}
diff --git a/drivers/net/intel/fm10k/base/fm10k_osdep.h
b/drivers/net/intel/fm10k/base/fm10k_osdep.h
index a727a57481..5f8ff10474 100644
--- a/drivers/net/intel/fm10k/base/fm10k_osdep.h
+++ b/drivers/net/intel/fm10k/base/fm10k_osdep.h
@@ -67,7 +67,7 @@ typedef uint64_t u64;
#define FM10K_PCI_REG_WRITE(reg, value) rte_write32((value), (reg))
/* not implemented */
-#define FM10K_READ_PCI_WORD(hw, reg) 0
+#define FM10K_READ_PCI_WORD(hw, reg) ((void)hw, (void)reg, 0)
#define FM10K_WRITE_MBX(hw, reg, value) FM10K_WRITE_REG(hw, reg, value)
#define FM10K_READ_MBX(hw, reg) FM10K_READ_REG(hw, reg)
diff --git a/drivers/net/intel/fm10k/base/fm10k_pf.c
b/drivers/net/intel/fm10k/base/fm10k_pf.c
index 439dd224de..b54116a4b5 100644
--- a/drivers/net/intel/fm10k/base/fm10k_pf.c
+++ b/drivers/net/intel/fm10k/base/fm10k_pf.c
@@ -1362,19 +1362,19 @@ STATIC u8 fm10k_iov_supported_xcast_mode_pf(struct
fm10k_vf_info *vf_info,
case FM10K_XCAST_MODE_PROMISC:
if (vf_flags & FM10K_VF_FLAG_PROMISC_CAPABLE)
return FM10K_XCAST_MODE_PROMISC;
- /* fallthough */
+ /* fallthrough */
case FM10K_XCAST_MODE_ALLMULTI:
if (vf_flags & FM10K_VF_FLAG_ALLMULTI_CAPABLE)
return FM10K_XCAST_MODE_ALLMULTI;
- /* fallthough */
+ /* fallthrough */
case FM10K_XCAST_MODE_MULTI:
if (vf_flags & FM10K_VF_FLAG_MULTI_CAPABLE)
return FM10K_XCAST_MODE_MULTI;
- /* fallthough */
+ /* fallthrough */
case FM10K_XCAST_MODE_NONE:
if (vf_flags & FM10K_VF_FLAG_NONE_CAPABLE)
return FM10K_XCAST_MODE_NONE;
- /* fallthough */
+ /* fallthrough */
default:
break;
}
diff --git a/drivers/net/intel/fm10k/base/fm10k_type.h
b/drivers/net/intel/fm10k/base/fm10k_type.h
index 84781ba9b2..437fb1c55e 100644
--- a/drivers/net/intel/fm10k/base/fm10k_type.h
+++ b/drivers/net/intel/fm10k/base/fm10k_type.h
@@ -83,9 +83,9 @@ struct fm10k_hw;
#define FM10K_NOT_IMPLEMENTED 0x7FFF
#define UNREFERENCED_XPARAMETER
-#define UNREFERENCED_1PARAMETER(_p) (_p)
-#define UNREFERENCED_2PARAMETER(_p, _q)do { (_p); (_q); } while (0)
-#define UNREFERENCED_3PARAMETER(_p, _q, _r) do { (_p); (_q); (_r); } while (0)
+#define UNREFERENCED_1PARAMETER(_p) (void)(_p)
+#define UNREFERENCED_2PARAMETER(_p, _q)do { (void)(_p);
(void)(_q); } while (0)
+#define UNREFERENCED_3PARAMETER(_p, _q, _r) do { (void)(_p); (void)(_q);
(void)(_r); } while (0)
/* Start of PF registers */
#define FM10K_CTRL 0x
diff --git a/drivers/net/intel/fm10k/base/meson.build
b/drivers/net/intel/fm10k/base/meson.build
index f24e453fd0..a2640d1ee8 100644
--- a/drivers/net/intel/fm10k/base/meson.build
+++ b/drivers/net/intel/fm10k/base/meson.build
@@ -10,19 +10,7 @@ sources = [
'fm10k_vf.c',
]
-error_cflags = [
-'-Wno-unused-parameter',
-'-Wno-unused-value',
-'-Wno-implicit-fallthrough',
-]
-c_args = cflags
-foreach flag: error_cflags
-if cc.has_argument(flag)
-c_args += flag
-endif
-endforeach
-
base_lib = static_library('fm10k_base', sources,
dependencies: static_rte_eal,
-c_args: c_args)
+c_args: cflags)
base_objs = base_lib.extract_all_objects(recursive: true)
--
2.45.2
Re: [RFC v4 0/8] Symbol versioning and export rework
On Wed, Mar 26, 2025 2:37 PM Bruce Richardson wrote: > > - I am considering making rte_function_versioning.h a non exported > > header (precisely, moving it to buildtools/ and maybe renaming it). > > > > This header contains macros not prefixed with RTE_. > > Using it requires some build trick (see use_function_versioning). > > And I don't see symbol versioning as a MUST infrastructure that DPDK > > needs to provide to datapath applications. > > > > Yet technically, this change would be an API breakage if some > > applications indeed relied on it. > > > > Is it not needed for exporting if an exported library header had versioned > symbols is in? That could be, though we don't expose such versionned symbols atm (and I don't think we ever did in DPDK). The only public symbol is always the symbol implemented for the latest ABI for newly compiled applications. Wrt to your other comment on the header(s) location, I don't mind moving to lib/eal/common. -- David Marchand
RE: [RFC 0/2] introduce event vector adapter
++ > -Original Message- > From: pbhagavat...@marvell.com > Sent: Wednesday, March 26, 2025 6:45 PM > To: Jerin Jacob > Cc: dev@dpdk.org; Pavan Nikhilesh Bhagavatula > > Subject: [RFC 0/2] introduce event vector adapter > > From: Pavan Nikhilesh > > The event vector adapter supports offloading the creation of event vectors > by vectorizing objects (mbufs/ptrs/u64s). > > An event vector adapter has the following working model: > > ┌──┐ > │ Vector ├─┐ > │ adapter0 │ │ > └──┘ │ > ┌──┐ │ ┌──┐ > │ Vector ├─┼──►│ Event │ > │ adapter1 │ │ │ Queue0 │ > └──┘ │ └──┘ > ┌──┐ │ > │ Vector ├─┘ > │ adapter2 │ > └──┘ > > ┌──┐ > │ Vector ├─┐ > │ adapter0 │ │ ┌──┐ > └──┘ ├──►│ Event │ > ┌──┐ │ │ Queue1 │ > │ Vector ├─┘ └──┘ > │ adapter1 │ > └──┘ > > - A vector adapter can be seen as an extension to event queue. It helps in >aggregating objects and generating a vector event which is enqueued to the >event queue. > > - Multiple vector adapters can be created on an event queue, each with its >own unique properties such as event properties, vector size, and timeout. >Note: If the target event queue doesn't support > RTE_EVENT_QUEUE_CFG_ALL_TYPES, > then the vector adapter should use the same schedule type as the > event > queue. > > - Each vector adapter aggregates objects, generates a vector event and >enqueues it to the event queue with the event properties mentioned in >rte_event_vector_adapter_conf::ev. > > - After configuring the vector adapter, the Application needs to use the >rte_event_vector_adapter_enqueue() function to enqueue objects i.e., >mbufs/ptrs/u64s to the vector adapter. >On reaching the configured vector size or timeout, the vector adapter >enqueues the event vector to the event queue. >Note: Application should use the event_type and sub_event_type properly > identifying the contents of vector event on dequeue. > > - If the vector adapter advertises the > RTE_EVENT_VECTOR_ADAPTER_CAP_SOV_EOV > capability, application can use the RTE_EVENT_VECTOR_ENQ_[S|E]OV flags > to indicate the start and end of a vector event. > * When RTE_EVENT_VECTOR_ENQ_SOV is set, the vector adapter will flush > any > aggregation in progress as a vector event and start aggregating a new > vector event with the enqueued ptr. > * When RTE_EVENT_VECTOR_ENQ_EOV is set, the vector adapter will add the > current ptr enqueued to the aggregated event and enqueue the vector event > to the event queue. > * If both flags are set, the vector adapter will flush the current > aggregation > as a vector event and enqueue the current ptr as a single event to the > event > queue. > > - If the vector adapter reaches the configured vector size, it will enqueue >the aggregated vector event to the event queue. > > - If the vector adapter reaches the configured vector timeout, it will flush >the current aggregation as a vector event if the minimum vector size is >reached, if not it will enqueue the objects as single events to the event >queue. > > - If the vector adapter is unable to aggregate the objects into a vector > event, >it will enqueue the objects as single events to the event queue with the > event >properties mentioned in rte_event_vector_adapter_conf::ev_fallback. > > Before using the vector adapter, the application has to create and configure > an event device and based on the event device capability it might require > creating an additional event port. > > When the application creates the vector adapter using the > ``rte_event_vector_adapter_create()`` function, the event device driver > capabilities are checked. If an in-built port is absent, the application > uses the default function to create a new event port. > For finer control over event port creation, the application should use > the ``rte_event_vector_adapter_create_ext()`` function. > > The application can enqueue one or more objects to the vector adapter using > the > ``rte_event_vector_adapter_enqueue()`` function and control the aggregation > using the flags. > > Vector adapters report stats using the > ``rte_event_vector_adapter_stats_get()`` > function and reset the stats using the > ``rte_event_vector_adapter_stats_reset()``. > > The application can destroy the vector adapter using the > ``rte_event_vector_adapter_destroy()`` function. > > Pavan Nikhilesh (2): > eventdev: introduce event vector adapter > eventdev: add default software vector adapter > > config/rte_config.h | 1 + > lib/eventdev/event_vector_adapter_pmd.h | 87 +++ > lib/eventdev
[PATCH 1/7] net/iavf/base: remove unused meson.build file
The meson.build file in the base folder was never used, as the base
files were just included in the main sources list for iavf.
Fixes: f1fdc9ddba5e ("drivers: move iavf common folder to iavf net")
Signed-off-by: Bruce Richardson
---
drivers/net/intel/iavf/base/meson.build | 10 --
1 file changed, 10 deletions(-)
delete mode 100644 drivers/net/intel/iavf/base/meson.build
diff --git a/drivers/net/intel/iavf/base/meson.build
b/drivers/net/intel/iavf/base/meson.build
deleted file mode 100644
index 273e88e921..00
--- a/drivers/net/intel/iavf/base/meson.build
+++ /dev/null
@@ -1,10 +0,0 @@
-# SPDX-License-Identifier: BSD-3-Clause
-# Copyright(c) 2019-2021 Intel Corporation
-
-sources = files('iavf_adminq.c', 'iavf_common.c', 'iavf_impl.c')
-
-if cc.has_argument('-Wno-pointer-to-int-cast')
-cflags += '-Wno-pointer-to-int-cast'
-endif
-
-require_iova_in_mbuf = false
--
2.45.2
[PATCH v2 3/8] net/ixgbe/base: correct definition of macro
The definition of IXGBE_LE32_TO_CPUS macro is meant to modify the value
in place - similar to the le32_to_cpus() macro in kernel. Fixing the
definition allows us to remove some warning flags, and removes the need
for the uintptr_t typecasts.
Fixes: aa4fc14d2cee ("ixgbe: update base driver")
Cc: sta...@dpdk.org
Signed-off-by: Bruce Richardson
---
drivers/net/intel/ixgbe/base/ixgbe_common.c | 4 ++--
drivers/net/intel/ixgbe/base/ixgbe_osdep.h | 2 +-
drivers/net/intel/ixgbe/base/meson.build| 2 +-
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/intel/ixgbe/base/ixgbe_common.c
b/drivers/net/intel/ixgbe/base/ixgbe_common.c
index d6425c5b78..fbc9605e4d 100644
--- a/drivers/net/intel/ixgbe/base/ixgbe_common.c
+++ b/drivers/net/intel/ixgbe/base/ixgbe_common.c
@@ -4610,7 +4610,7 @@ s32 ixgbe_host_interface_command(struct ixgbe_hw *hw, u32
*buffer,
/* first pull in the header so we know the buffer length */
for (bi = 0; bi < dword_len; bi++) {
buffer[bi] = IXGBE_READ_REG_ARRAY(hw, IXGBE_FLEX_MNG, bi);
- IXGBE_LE32_TO_CPUS((uintptr_t)&buffer[bi]);
+ IXGBE_LE32_TO_CPUS(&buffer[bi]);
}
/*
@@ -4646,7 +4646,7 @@ s32 ixgbe_host_interface_command(struct ixgbe_hw *hw, u32
*buffer,
/* Pull in the rest of the buffer (bi is where we left off) */
for (; bi <= dword_len; bi++) {
buffer[bi] = IXGBE_READ_REG_ARRAY(hw, IXGBE_FLEX_MNG, bi);
- IXGBE_LE32_TO_CPUS((uintptr_t)&buffer[bi]);
+ IXGBE_LE32_TO_CPUS(&buffer[bi]);
}
rel_out:
diff --git a/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
b/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
index cffc6a4ce8..6e5f7b4ae8 100644
--- a/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
+++ b/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
@@ -83,7 +83,7 @@ enum {
#define IXGBE_LE16_TO_CPU(_i) rte_le_to_cpu_16(_i)
#define IXGBE_LE32_TO_CPU(_i) rte_le_to_cpu_32(_i)
#define IXGBE_LE64_TO_CPU(_i) rte_le_to_cpu_64(_i)
-#define IXGBE_LE32_TO_CPUS(_i) rte_le_to_cpu_32(_i)
+#define IXGBE_LE32_TO_CPUS(_i) do { *_i = rte_le_to_cpu_32(*_i); } while(0)
#define IXGBE_CPU_TO_BE16(_i) rte_cpu_to_be_16(_i)
#define IXGBE_CPU_TO_BE32(_i) rte_cpu_to_be_32(_i)
#define IXGBE_BE32_TO_CPU(_i) rte_be_to_cpu_32(_i)
diff --git a/drivers/net/intel/ixgbe/base/meson.build
b/drivers/net/intel/ixgbe/base/meson.build
index 7e4fbdfa0f..f8b2ee6341 100644
--- a/drivers/net/intel/ixgbe/base/meson.build
+++ b/drivers/net/intel/ixgbe/base/meson.build
@@ -19,7 +19,7 @@ sources = [
'ixgbe_x550.c',
]
-error_cflags = ['-Wno-unused-value',
+error_cflags = [
'-Wno-unused-but-set-variable',
'-Wno-unused-parameter',
]
--
2.45.2
[PATCH v2 4/8] net/ixgbe/base: fix compilation warnings
We can remove almost all of the "unused parameter" and "unused variable"
warnings by just improving the macro definitions in the osdep.h header.
Remaining two instances can be fixed by just one-line additions to the
code, so add those to give us a clean build with the warnings enabled.
Fixes: af75078fece3 ("first public release")
Fixes: c6cb313da739 ("net/ixgbe/base: add link management for E610")
Cc: sta...@dpdk.org
Signed-off-by: Bruce Richardson
---
drivers/net/intel/ixgbe/base/ixgbe_e610.c | 2 ++
drivers/net/intel/ixgbe/base/ixgbe_osdep.h | 19 +++
drivers/net/intel/ixgbe/base/meson.build | 11 ---
3 files changed, 13 insertions(+), 19 deletions(-)
diff --git a/drivers/net/intel/ixgbe/base/ixgbe_e610.c
b/drivers/net/intel/ixgbe/base/ixgbe_e610.c
index 5474c3012a..7420c78d07 100644
--- a/drivers/net/intel/ixgbe/base/ixgbe_e610.c
+++ b/drivers/net/intel/ixgbe/base/ixgbe_e610.c
@@ -1054,6 +1054,7 @@ static void ixgbe_parse_vsi_func_caps(struct ixgbe_hw *hw,
struct ixgbe_hw_func_caps *func_p,
struct ixgbe_aci_cmd_list_caps_elem *cap)
{
+ UNREFERENCED_PARAMETER(cap);
func_p->guar_num_vsi = ixgbe_get_num_per_func(hw, IXGBE_MAX_VSI);
}
@@ -1770,6 +1771,7 @@ s32 ixgbe_aci_set_event_mask(struct ixgbe_hw *hw, u8
port_num, u16 mask)
struct ixgbe_aci_cmd_set_event_mask *cmd;
struct ixgbe_aci_desc desc;
+ UNREFERENCED_PARAMETER(port_num);
cmd = &desc.params.set_event_mask;
ixgbe_fill_dflt_direct_cmd_desc(&desc, ixgbe_aci_opc_set_event_mask);
diff --git a/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
b/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
index 6e5f7b4ae8..398c38bffd 100644
--- a/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
+++ b/drivers/net/intel/ixgbe/base/ixgbe_osdep.h
@@ -57,13 +57,16 @@
/* Bunch of defines for shared code bogosity */
#ifndef UNREFERENCED_PARAMETER
-#define UNREFERENCED_PARAMETER(_p)
+#define UNREFERENCED_PARAMETER(_p) (void)(_p)
#endif
-#define UNREFERENCED_1PARAMETER(_p)
-#define UNREFERENCED_2PARAMETER(_p, _q)
-#define UNREFERENCED_3PARAMETER(_p, _q, _r)
-#define UNREFERENCED_4PARAMETER(_p, _q, _r, _s)
-#define UNREFERENCED_5PARAMETER(_p, _q, _r, _s, _t)
+#define UNREFERENCED_1PARAMETER(_p) (void)(_p)
+#define UNREFERENCED_2PARAMETER(_p, _q) do { (void)(_p); (void)(_q); } while(0)
+#define UNREFERENCED_3PARAMETER(_p, _q, _r) \
+ do { (void)(_p); (void)(_q); (void)(_r); } while(0)
+#define UNREFERENCED_4PARAMETER(_p, _q, _r, _s) \
+ do { (void)(_p); (void)(_q); (void)(_r); (void)(_s); } while(0)
+#define UNREFERENCED_5PARAMETER(_p, _q, _r, _s, _t) \
+ do { (void)(_p); (void)(_q); (void)(_r); (void)(_s); (void)(_t); }
while(0)
/* Shared code error reporting */
enum {
@@ -130,8 +133,8 @@ static inline uint32_t ixgbe_read_addr(volatile void* addr)
IXGBE_PCI_REG_ADDR((hw), (reg) + ((index) << 2))
/* Not implemented !! */
-#define IXGBE_READ_PCIE_WORD(hw, reg) 0
-#define IXGBE_WRITE_PCIE_WORD(hw, reg, value) do { } while(0)
+#define IXGBE_READ_PCIE_WORD(hw, reg) ((void)hw, (void)(reg), 0)
+#define IXGBE_WRITE_PCIE_WORD(hw, reg, value) do { (void)hw; (void)reg;
(void)value; } while(0)
#define IXGBE_WRITE_FLUSH(a) IXGBE_READ_REG(a, IXGBE_STATUS)
diff --git a/drivers/net/intel/ixgbe/base/meson.build
b/drivers/net/intel/ixgbe/base/meson.build
index f8b2ee6341..64e0bfd7be 100644
--- a/drivers/net/intel/ixgbe/base/meson.build
+++ b/drivers/net/intel/ixgbe/base/meson.build
@@ -19,17 +19,6 @@ sources = [
'ixgbe_x550.c',
]
-error_cflags = [
-'-Wno-unused-but-set-variable',
-'-Wno-unused-parameter',
-]
-c_args = cflags
-foreach flag: error_cflags
-if cc.has_argument(flag)
-c_args += flag
-endif
-endforeach
-
base_lib = static_library('ixgbe_base', sources,
dependencies: [static_rte_eal, static_rte_net],
c_args: c_args)
--
2.45.2
[PATCH v2 5/8] net/i40e/base: fix unused value warnings
Fix warnings about unused values - parameters, variables, etc., and
remove the warning disable flags for them. Although modifying the
base-code files is not ideal, the changes required are minor, and only
affect two files from the imported base code.
Fixes: 8db9e2a1b232 ("i40e: base driver")
Cc: sta...@dpdk.org
Signed-off-by: Bruce Richardson
---
drivers/net/intel/i40e/base/i40e_nvm.c | 2 +-
drivers/net/intel/i40e/base/i40e_osdep.h | 4 ++--
drivers/net/intel/i40e/base/i40e_type.h | 14 +-
drivers/net/intel/i40e/base/meson.build | 3 ---
drivers/net/intel/i40e/i40e_ethdev.c | 1 +
5 files changed, 13 insertions(+), 11 deletions(-)
diff --git a/drivers/net/intel/i40e/base/i40e_nvm.c
b/drivers/net/intel/i40e/base/i40e_nvm.c
index 3e16a0d997..56dc4d9279 100644
--- a/drivers/net/intel/i40e/base/i40e_nvm.c
+++ b/drivers/net/intel/i40e/base/i40e_nvm.c
@@ -1743,7 +1743,7 @@ STATIC enum i40e_status_code
i40e_nvmupd_get_aq_result(struct i40e_hw *hw,
**/
STATIC enum i40e_status_code i40e_nvmupd_get_aq_event(struct i40e_hw *hw,
struct i40e_nvm_access *cmd,
- u8 *bytes, int *perrno)
+ u8 *bytes, __rte_unused int
*perrno)
{
u32 aq_total_len;
u32 aq_desc_len;
diff --git a/drivers/net/intel/i40e/base/i40e_osdep.h
b/drivers/net/intel/i40e/base/i40e_osdep.h
index c04f94732a..197f4678bf 100644
--- a/drivers/net/intel/i40e/base/i40e_osdep.h
+++ b/drivers/net/intel/i40e/base/i40e_osdep.h
@@ -184,8 +184,8 @@ struct __rte_packed_begin i40e_dma_mem {
const void *zone;
} __rte_packed_end;
-#define i40e_allocate_dma_mem(h, m, unused, s, a) \
- i40e_allocate_dma_mem_d(h, m, s, a)
+#define i40e_allocate_dma_mem(h, m, mt, s, a) \
+ i40e_allocate_dma_mem_d(h, m, mt, s, a)
#define i40e_free_dma_mem(h, m) i40e_free_dma_mem_d(h, m)
struct __rte_packed_begin i40e_virt_mem {
diff --git a/drivers/net/intel/i40e/base/i40e_type.h
b/drivers/net/intel/i40e/base/i40e_type.h
index 7cc746f82f..968e1982a6 100644
--- a/drivers/net/intel/i40e/base/i40e_type.h
+++ b/drivers/net/intel/i40e/base/i40e_type.h
@@ -14,11 +14,15 @@
#include "i40e_devids.h"
#define UNREFERENCED_XPARAMETER
-#define UNREFERENCED_1PARAMETER(_p) (_p);
-#define UNREFERENCED_2PARAMETER(_p, _q) (_p); (_q);
-#define UNREFERENCED_3PARAMETER(_p, _q, _r) (_p); (_q); (_r);
-#define UNREFERENCED_4PARAMETER(_p, _q, _r, _s) (_p); (_q); (_r); (_s);
-#define UNREFERENCED_5PARAMETER(_p, _q, _r, _s, _t) (_p); (_q); (_r); (_s);
(_t);
+#define UNREFERENCED_1PARAMETER(_p) (void)(_p)
+#define UNREFERENCED_2PARAMETER(_p, _q) \
+ do { (void)(_p); (void)(_q); } while (0)
+#define UNREFERENCED_3PARAMETER(_p, _q, _r) \
+ do { (void)(_p); (void)(_q); (void)(_r); } while (0)
+#define UNREFERENCED_4PARAMETER(_p, _q, _r, _s) \
+ do { (void)(_p); (void)(_q); (void)(_r); (void)(_s); } while (0)
+#define UNREFERENCED_5PARAMETER(_p, _q, _r, _s, _t) \
+ do { (void)(_p); (void)(_q); (void)(_r); (void)(_s); (void)(_t); }
while (0)
#define BIT(a) (1UL << (a))
#define BIT_ULL(a) (1ULL << (a))
diff --git a/drivers/net/intel/i40e/base/meson.build
b/drivers/net/intel/i40e/base/meson.build
index a0912b1788..2648e5d0c4 100644
--- a/drivers/net/intel/i40e/base/meson.build
+++ b/drivers/net/intel/i40e/base/meson.build
@@ -13,10 +13,7 @@ sources = [
error_cflags = [
'-Wno-sign-compare',
-'-Wno-unused-value',
'-Wno-strict-aliasing',
-'-Wno-unused-but-set-variable',
-'-Wno-unused-parameter',
]
c_args = cflags
foreach flag: error_cflags
diff --git a/drivers/net/intel/i40e/i40e_ethdev.c
b/drivers/net/intel/i40e/i40e_ethdev.c
index 1c5ab35a8b..90eba3419f 100644
--- a/drivers/net/intel/i40e/i40e_ethdev.c
+++ b/drivers/net/intel/i40e/i40e_ethdev.c
@@ -4694,6 +4694,7 @@ i40e_dev_rss_reta_query(struct rte_eth_dev *dev,
enum i40e_status_code
i40e_allocate_dma_mem_d(__rte_unused struct i40e_hw *hw,
struct i40e_dma_mem *mem,
+ __rte_unused enum i40e_memory_type mtype,
u64 size,
u32 alignment)
{
--
2.45.2
[PATCH v2 2/8] net/iavf/base: remove unused meson.build file
The meson.build file in the base folder was never used, as the base
files were just included in the main sources list for iavf.
Fixes: f1fdc9ddba5e ("drivers: move iavf common folder to iavf net")
Signed-off-by: Bruce Richardson
---
drivers/net/intel/iavf/base/meson.build | 10 --
1 file changed, 10 deletions(-)
delete mode 100644 drivers/net/intel/iavf/base/meson.build
diff --git a/drivers/net/intel/iavf/base/meson.build
b/drivers/net/intel/iavf/base/meson.build
deleted file mode 100644
index 273e88e921..00
--- a/drivers/net/intel/iavf/base/meson.build
+++ /dev/null
@@ -1,10 +0,0 @@
-# SPDX-License-Identifier: BSD-3-Clause
-# Copyright(c) 2019-2021 Intel Corporation
-
-sources = files('iavf_adminq.c', 'iavf_common.c', 'iavf_impl.c')
-
-if cc.has_argument('-Wno-pointer-to-int-cast')
-cflags += '-Wno-pointer-to-int-cast'
-endif
-
-require_iova_in_mbuf = false
--
2.45.2
Re: Clarification on Minimum Supported Kernel Version for DPDK
I’ll set up the CI testing environment on kernel 4.19 as per the current minimum requirement then. Thanks, Cody Cheng On Sat, Mar 22, 2025 at 9:02 AM Morten Brørup wrote: > > > From: Stephen Hemminger [mailto:step...@networkplumber.org] > > Sent: Friday, 21 March 2025 16.53 > > > > On Fri, 21 Mar 2025 07:28:45 +0100 > > Morten Brørup wrote: > > > > > @Kevin, @Stephen, @Bruce, > > > > > > I cannot reliably answer Cody's question, and it may need further > > discussion. > > > > > > What is your opinion on minimum Linux kernel version requirements? > > > > > > @Thomas: In the future, the DPDK release notes should mention the > > minimum Linux kernel requirements. > > > > > > > From: Cody Cheng [mailto:cch...@iol.unh.edu] > > > > Sent: Thursday, 20 March 2025 21.28 > > > > > > > > Hi Morten, > > > > > > > > I am in the process of setting up a test environment at the UNH > > DPDK > > > > Community Test Lab that follows the minimum supported kernel > > version > > > > for DPDK. According to the DPDK documentation, the minimum > > supported > > > > kernel version is 4.19. However, the oldest long term stable kernel > > > > version listed on kernel.org is 5.4.291. > > > > > > > > Should the test environment be set up on kernel version 4.19 or > > > > 5.4.291? > > > > > > The kernel 4.19 support stems from still supporting RHEL/CentOS 7. > > > I wonder if this exception mentioned in the documentation [1] is > > still valid, or if we should bump it to RHEL/CentOS 8, which ships with > > kernel 4.18 [1]. > > > > > > RHEL/CentOS 7 support was discussed at by tech board long ago [2], > > but I cannot find a conclusion about the kernel version; the discussion > > was mostly about compiler support. > > > > > > [1]: https://doc.dpdk.org/guides/linux_gsg/sys_reqs.html#system- > > software > > > [2]: > > https://docs.redhat.com/en/documentation/red_hat_enterprise_linux/8/htm > > l-single/8.0_release_notes/index#overview > > > [3]: https://mails.dpdk.org/archives/dev/2023-February/263516.html > > > > My opinion has always been that DPDK only offers certain guarantees > > about testing: > > - oldest current LTS > > - oldest supported version of Redhat/Ubuntu/SUSE enterprise kernels > > > > after that in the embedded space, the user is likely to be ok but any > > kernel > > related issues are their problem not the communities to deal with. > > Generally, if some new DPDK feature requires a new kernel (or new kernel > feature), the details should be mentioned in the release notes. > And preferably, that feature should degrade gracefully when the feature is > not present. > > For the embedded space, we could support the oldest current version available > as Super LTS [4], which is 4.4. And for now, we could stick with the second > oldest, 4.19, which is what we currently have. > > [4]: > https://wiki.linuxfoundation.org/civilinfrastructureplatform/start#kernel_maintainership > > Some old kernel version might not be officially supported by the Kernel > community, but an embedded vendor might have tested the relevant features > extensively and thus trust it more than some new and officially supported > version. > So let's not require a newer version than we absolutely must, on technical > grounds. > It seems that kernel 4.19 is the current minimum requirement, so let's stick > with that, until there are valid technical reasons for requiring a newer > version. > > Anyway, it seems we need to clarify the policy for kernel version > requirements. > It's easy regarding the distros; DPDK running on those require their shipped > kernel version, at minimum. > It's for everything else clarification is needed. > > And it's not just embedded. Virtual appliances can be tricky too... with our > SmartShare VM we had to add support for running as a guest under an ancient > QEMU host version, because that is the hypervisor used by one of the big > system providers in our most important target market. > > In non-cloud market segments, a lot of really old stuff is still being used > in production, working perfectly fine. > > > > > The two parts most likely to cause issues are vfio-pci and vhost > > related stuff. > > There is also small chance of issues with the memory handling in EAL. > And maybe handling of many CPU cores, and most likely something related to > the new cache steering feature. >
[PATCH v2 6/8] net/i40e/base: fix compiler warnings
Add a single-line fix to the base code, and then the remaining two
compiler warning disable flags can be removed from the driver base code
build file.
Fixes: 8db9e2a1b232 ("i40e: base driver")
Cc: sta...@dpdk.org
Signed-off-by: Bruce Richardson
---
drivers/net/intel/i40e/base/i40e_diag.c | 2 +-
drivers/net/intel/i40e/base/meson.build | 13 +
2 files changed, 2 insertions(+), 13 deletions(-)
diff --git a/drivers/net/intel/i40e/base/i40e_diag.c
b/drivers/net/intel/i40e/base/i40e_diag.c
index 4ca102cdd5..71b2e53e85 100644
--- a/drivers/net/intel/i40e/base/i40e_diag.c
+++ b/drivers/net/intel/i40e/base/i40e_diag.c
@@ -34,7 +34,7 @@ static enum i40e_status_code
i40e_diag_reg_pattern_test(struct i40e_hw *hw,
{
const u32 patterns[] = {0x5A5A5A5A, 0xA5A5A5A5, 0x, 0x};
u32 pat, val, orig_val;
- int i;
+ unsigned int i;
orig_val = rd32(hw, reg);
for (i = 0; i < ARRAY_SIZE(patterns); i++) {
diff --git a/drivers/net/intel/i40e/base/meson.build
b/drivers/net/intel/i40e/base/meson.build
index 2648e5d0c4..766383101b 100644
--- a/drivers/net/intel/i40e/base/meson.build
+++ b/drivers/net/intel/i40e/base/meson.build
@@ -11,18 +11,7 @@ sources = [
'i40e_nvm.c',
]
-error_cflags = [
-'-Wno-sign-compare',
-'-Wno-strict-aliasing',
-]
-c_args = cflags
-foreach flag: error_cflags
-if cc.has_argument(flag)
-c_args += flag
-endif
-endforeach
-
base_lib = static_library('i40e_base', sources,
dependencies: static_rte_eal,
-c_args: c_args)
+c_args: cflags)
base_objs = base_lib.extract_all_objects(recursive: true)
--
2.45.2
[PATCH v2 8/8] net/intel: simplify base code builds
Now that base-code warning flags have been removed from a number of
drivers, we can simplify their build configuration by just including the
base code files in the regular list of driver files passed back. There
is no need to use pre-compiled objects.
Signed-off-by: Bruce Richardson
---
drivers/net/intel/e1000/base/meson.build | 9 ++---
drivers/net/intel/e1000/meson.build | 4 ++--
drivers/net/intel/fm10k/base/meson.build | 9 ++---
drivers/net/intel/fm10k/meson.build | 4 ++--
drivers/net/intel/i40e/base/meson.build | 9 ++---
drivers/net/intel/i40e/meson.build | 4 ++--
drivers/net/intel/idpf/base/meson.build | 2 +-
drivers/net/intel/idpf/meson.build | 7 ---
drivers/net/intel/ixgbe/base/meson.build | 9 ++---
drivers/net/intel/ixgbe/meson.build | 4 ++--
10 files changed, 21 insertions(+), 40 deletions(-)
diff --git a/drivers/net/intel/e1000/base/meson.build
b/drivers/net/intel/e1000/base/meson.build
index 8cbd9f62e6..4fe86dc6df 100644
--- a/drivers/net/intel/e1000/base/meson.build
+++ b/drivers/net/intel/e1000/base/meson.build
@@ -1,7 +1,7 @@
# SPDX-License-Identifier: BSD-3-Clause
# Copyright(c) 2017 Intel Corporation
-sources = [
+base_sources = files(
'e1000_base.c',
'e1000_80003es2lan.c',
'e1000_82540.c',
@@ -21,9 +21,4 @@ sources = [
'e1000_osdep.c',
'e1000_phy.c',
'e1000_vf.c',
-]
-
-base_lib = static_library('e1000_base', sources,
-dependencies: static_rte_eal,
-c_args: cflags)
-base_objs = base_lib.extract_all_objects(recursive: true)
+)
diff --git a/drivers/net/intel/e1000/meson.build
b/drivers/net/intel/e1000/meson.build
index cd42c0042a..b52a843228 100644
--- a/drivers/net/intel/e1000/meson.build
+++ b/drivers/net/intel/e1000/meson.build
@@ -2,9 +2,9 @@
# Copyright(c) 2017 Intel Corporation
subdir('base')
-objs = [base_objs]
-sources = files(
+sources += base_sources
+sources += files(
'e1000_logs.c',
'em_ethdev.c',
'em_rxtx.c',
diff --git a/drivers/net/intel/fm10k/base/meson.build
b/drivers/net/intel/fm10k/base/meson.build
index a2640d1ee8..c45b31d6b2 100644
--- a/drivers/net/intel/fm10k/base/meson.build
+++ b/drivers/net/intel/fm10k/base/meson.build
@@ -1,16 +1,11 @@
# SPDX-License-Identifier: BSD-3-Clause
# Copyright(c) 2017 Intel Corporation
-sources = [
+base_sources = files(
'fm10k_api.c',
'fm10k_common.c',
'fm10k_mbx.c',
'fm10k_pf.c',
'fm10k_tlv.c',
'fm10k_vf.c',
-]
-
-base_lib = static_library('fm10k_base', sources,
-dependencies: static_rte_eal,
-c_args: cflags)
-base_objs = base_lib.extract_all_objects(recursive: true)
+)
diff --git a/drivers/net/intel/fm10k/meson.build
b/drivers/net/intel/fm10k/meson.build
index 69566add96..e08a00cb49 100644
--- a/drivers/net/intel/fm10k/meson.build
+++ b/drivers/net/intel/fm10k/meson.build
@@ -8,9 +8,9 @@ if is_windows
endif
subdir('base')
-objs = [base_objs]
-sources = files(
+sources += base_sources
+sources += files(
'fm10k_ethdev.c',
'fm10k_rxtx.c',
)
diff --git a/drivers/net/intel/i40e/base/meson.build
b/drivers/net/intel/i40e/base/meson.build
index 766383101b..7b4882ad1d 100644
--- a/drivers/net/intel/i40e/base/meson.build
+++ b/drivers/net/intel/i40e/base/meson.build
@@ -1,7 +1,7 @@
# SPDX-License-Identifier: BSD-3-Clause
# Copyright(c) 2017-2020 Intel Corporation
-sources = [
+base_sources = files(
'i40e_adminq.c',
'i40e_common.c',
'i40e_dcb.c',
@@ -9,9 +9,4 @@ sources = [
'i40e_hmc.c',
'i40e_lan_hmc.c',
'i40e_nvm.c',
-]
-
-base_lib = static_library('i40e_base', sources,
-dependencies: static_rte_eal,
-c_args: cflags)
-base_objs = base_lib.extract_all_objects(recursive: true)
+)
diff --git a/drivers/net/intel/i40e/meson.build
b/drivers/net/intel/i40e/meson.build
index 15993393fb..17b6715cd9 100644
--- a/drivers/net/intel/i40e/meson.build
+++ b/drivers/net/intel/i40e/meson.build
@@ -18,9 +18,9 @@ if dpdk_conf.get('RTE_IOVA_IN_MBUF') == 0
endif
subdir('base')
-objs = [base_objs]
-sources = files(
+sources += base_sources
+sources += files(
'i40e_ethdev.c',
'i40e_rxtx.c',
'i40e_pf.c',
diff --git a/drivers/net/intel/idpf/base/meson.build
b/drivers/net/intel/idpf/base/meson.build
index 7316e0a805..78782e463e 100644
--- a/drivers/net/intel/idpf/base/meson.build
+++ b/drivers/net/intel/idpf/base/meson.build
@@ -1,7 +1,7 @@
# SPDX-License-Identifier: BSD-3-Clause
# Copyright(c) 2023 Intel Corporation
-sources += files(
+base_sources = files(
'idpf_controlq.c',
'idpf_controlq_setup.c',
)
diff --git a/drivers/net/intel/idpf/meson.build
b/drivers/net/intel/idpf/meson.build
index 4b272d02b1..a8690da87b 100644
--- a/drivers/net/intel/idpf/meson.build
+++ b/drivers/net/intel/idpf/meson.build
@@ -7,9 +7,12 @@ if is_windows
subdir_done()
endif
+subdir('base')
+
includes += incl
[PATCH v2 7/8] net/ice/base: reduce warnings for unused variables
Improve base code macros to reduce the number of issues with unused
variables in the code. Issues still remain with unused-but-set
variables, but completely unused variable warnings are eliminated.
Signed-off-by: Bruce Richardson
---
drivers/net/intel/ice/base/ice_osdep.h | 6 +++---
drivers/net/intel/ice/base/ice_switch.c | 2 --
drivers/net/intel/ice/base/ice_type.h | 2 +-
drivers/net/intel/ice/base/meson.build | 1 -
drivers/net/intel/ice/ice_fdir_filter.c | 2 +-
5 files changed, 5 insertions(+), 8 deletions(-)
diff --git a/drivers/net/intel/ice/base/ice_osdep.h
b/drivers/net/intel/ice/base/ice_osdep.h
index 7b96fcde03..ad6cde9896 100644
--- a/drivers/net/intel/ice/base/ice_osdep.h
+++ b/drivers/net/intel/ice/base/ice_osdep.h
@@ -196,9 +196,9 @@ struct __rte_packed_begin ice_virt_mem {
u32 size;
} __rte_packed_end;
-#define ice_malloc(h, s)rte_zmalloc(NULL, s, 0)
-#define ice_calloc(h, c, s) rte_calloc(NULL, c, s, 0)
-#define ice_free(h, m) rte_free(m)
+#define ice_malloc(h, s)((void)h, rte_zmalloc(NULL, s, 0))
+#define ice_calloc(h, c, s) ((void)h, rte_calloc(NULL, c, s, 0))
+#define ice_free(h, m) ((void)h, rte_free(m))
#define ice_memset(a, b, c, d) memset((a), (b), (c))
#define ice_memcpy(a, b, c, d) rte_memcpy((a), (b), (c))
diff --git a/drivers/net/intel/ice/base/ice_switch.c
b/drivers/net/intel/ice/base/ice_switch.c
index a3786961e6..468a9f055d 100644
--- a/drivers/net/intel/ice/base/ice_switch.c
+++ b/drivers/net/intel/ice/base/ice_switch.c
@@ -8190,7 +8190,6 @@ ice_add_adv_recipe(struct ice_hw *hw, struct
ice_adv_lkup_elem *lkups,
struct ice_sw_recipe *rm;
u8 i;
int status = ICE_SUCCESS;
- u16 cnt;
if (!ice_is_prof_rule(rinfo->tun_type) && !lkups_cnt)
return ICE_ERR_PARAM;
@@ -9795,7 +9794,6 @@ ice_rem_adv_rule(struct ice_hw *hw, struct
ice_adv_lkup_elem *lkups,
bool remove_rule = false;
struct ice_lock *rule_lock; /* Lock to protect filter rule list */
u16 i, rid, vsi_handle;
- bool is_add = false;
int status = ICE_SUCCESS;
ice_memset(&lkup_exts, 0, sizeof(lkup_exts), ICE_NONDMA_MEM);
diff --git a/drivers/net/intel/ice/base/ice_type.h
b/drivers/net/intel/ice/base/ice_type.h
index 35f832eb9f..297a5ea890 100644
--- a/drivers/net/intel/ice/base/ice_type.h
+++ b/drivers/net/intel/ice/base/ice_type.h
@@ -153,7 +153,7 @@ static inline u32 ice_round_to_num(u32 N, u32 R)
#define ICE_DBG_USER BIT_ULL(31)
#define ICE_DBG_ALL0xULL
-#define __ALWAYS_UNUSED
+#define __ALWAYS_UNUSED __rte_unused
#define IS_ETHER_ADDR_EQUAL(addr1, addr2) \
(((bool)u16 *)(addr1))[0] == ((u16 *)(addr2))[0]))) && \
diff --git a/drivers/net/intel/ice/base/meson.build
b/drivers/net/intel/ice/base/meson.build
index 38d092c370..e7ba9c34bc 100644
--- a/drivers/net/intel/ice/base/meson.build
+++ b/drivers/net/intel/ice/base/meson.build
@@ -39,7 +39,6 @@ if is_ms_compiler
else
error_cflags = [
'-Wno-unused-but-set-variable',
-'-Wno-unused-variable',
'-Wno-unused-parameter',
]
endif
diff --git a/drivers/net/intel/ice/ice_fdir_filter.c
b/drivers/net/intel/ice/ice_fdir_filter.c
index 940fe171b6..2ff0090aca 100644
--- a/drivers/net/intel/ice/ice_fdir_filter.c
+++ b/drivers/net/intel/ice/ice_fdir_filter.c
@@ -1090,7 +1090,7 @@ ice_fdir_input_set_conf(struct ice_pf *pf, enum
ice_fltr_ptype flow,
return -EINVAL;
seg_tun = (struct ice_flow_seg_info *)
- ice_malloc(hw, sizeof(*seg_tun) * ICE_FD_HW_SEG_MAX);
+ ice_malloc(pf->adapter->hw, sizeof(*seg_tun) *
ICE_FD_HW_SEG_MAX);
if (!seg_tun) {
PMD_DRV_LOG(ERR, "No memory can be allocated");
return -ENOMEM;
--
2.45.2
[PATCH v5] test: improve resiliency of malloc autotest
The test case "test_multi_alloc_statistics" was brittle in that it did
some allocations and frees and then checked statistics without
considering the initial state of the malloc heaps. This meant that,
depending on what allocations/frees were done beforehand, the test can
sometimes fail.
We can improve resiliency by running the test using a new malloc heap,
which means it is unaffected by any previous allocations.
Bugzilla ID: 1579
Fixes: a40a1f8231b4 ("app: various tests update")
Cc: sta...@dpdk.org
Signed-off-by: Bruce Richardson
Acked-by: Chengwen Feng
---
v5: resubmit for 25.07 release, and to see if CI errors are cleared.
v4:
* added cleanup support in case of failure, to allow multiple runs of
the test in a single session.
v3:
* switched allocation from mmap to malloc allowing it work on windows
* use explicit alignment of the malloc return value to ensure memory
added to heap is page-aligned.
v2:
* removed unnecessary extra include
* only added new code for non-windows, since using mmap for allocation.
---
app/test/test_malloc.c | 77 +++---
1 file changed, 64 insertions(+), 13 deletions(-)
diff --git a/app/test/test_malloc.c b/app/test/test_malloc.c
index 02a7d8ef20..d6f8da3b24 100644
--- a/app/test/test_malloc.c
+++ b/app/test/test_malloc.c
@@ -25,6 +25,7 @@
#include
#include
#include
+#include
#include
#define N 1
@@ -267,11 +268,48 @@ test_str_to_size(void)
static int
test_multi_alloc_statistics(void)
{
+ int ret = -1; /* default return is error, cleared at end on success */
int socket = 0;
struct rte_malloc_socket_stats pre_stats, post_stats ,first_stats,
second_stats;
size_t size = 2048;
int align = 1024;
int overhead = 0;
+ const size_t pgsz = rte_mem_page_size();
+ const size_t heap_size = (1 << 22);
+
+ if (pgsz > heap_size) {
+ printf("Page size (%zu) is bigger than heap size, skipping
alloc stats test\n",
+ pgsz);
+ return TEST_SKIPPED;
+ }
+ if (heap_size % pgsz != 0) {
+ printf("Heap size (%zu) is not a multiple of page size (%zu),
skipping alloc stats test\n",
+ heap_size, pgsz);
+ return TEST_SKIPPED;
+ }
+
+ if (rte_malloc_heap_create(__func__) != 0) {
+ printf("Failed to create test malloc heap\n");
+ goto end;
+ }
+
+ /* Allocate some memory using malloc and add it to our test heap. */
+ void *unaligned_memory = malloc(heap_size + pgsz);
+ if (unaligned_memory == NULL) {
+ printf("Failed to allocate memory\n");
+ goto cleanup_empty_heap;
+ }
+ void *memory = RTE_PTR_ALIGN(unaligned_memory, pgsz);
+ if (rte_malloc_heap_memory_add(__func__, memory, heap_size, NULL,
+ heap_size / pgsz, pgsz) != 0) {
+ printf("Failed to add memory to heap\n");
+ goto cleanup_allocated_memory;
+ }
+ socket = rte_malloc_heap_get_socket(__func__);
+ if (socket < 0) {
+ printf("Failed to get socket for test malloc heap.\n");
+ goto cleanup_all;
+ }
/* Dynamically calculate the overhead by allocating one cacheline and
* then comparing what was allocated from the heap.
@@ -280,7 +318,7 @@ test_multi_alloc_statistics(void)
void *dummy = rte_malloc_socket(NULL, RTE_CACHE_LINE_SIZE, 0, socket);
if (dummy == NULL)
- return -1;
+ goto cleanup_all;
rte_malloc_get_socket_stats(socket, &post_stats);
@@ -295,7 +333,8 @@ test_multi_alloc_statistics(void)
void *p1 = rte_malloc_socket("stats", size , align, socket);
if (!p1)
- return -1;
+ goto cleanup_all;
+
rte_free(p1);
rte_malloc_dump_stats(stdout, "stats");
@@ -308,7 +347,7 @@ test_multi_alloc_statistics(void)
(post_stats.alloc_count != pre_stats.alloc_count) ||
(post_stats.free_count != pre_stats.free_count)) {
printf("Malloc statistics are incorrect - freed alloc\n");
- return -1;
+ goto cleanup_all;
}
/* Check two consecutive allocations */
size = 1024;
@@ -316,12 +355,12 @@ test_multi_alloc_statistics(void)
rte_malloc_get_socket_stats(socket,&pre_stats);
void *p2 = rte_malloc_socket("add", size ,align, socket);
if (!p2)
- return -1;
+ goto cleanup_all;
rte_malloc_get_socket_stats(socket,&first_stats);
void *p3 = rte_malloc_socket("add2", size,align, socket);
if (!p3)
- return -1;
+ goto cleanup_all;
rte_malloc_get_socket_stats(socket,&second_stats);
@@ -333,34 +372,34 @@ test_multi_alloc_statistics(void)
if(second
RE: [EXTERNAL] Re: [RFC 2/2] eventdev: add default software vector adapter
> On Wed, 26 Mar 2025 18:44:36 +0530
> wrote:
>
> > +struct sw_vector_adapter_service_data {
> > + uint32_t service_id;
> > + RTE_ATOMIC(rte_mcslock_t *) lock;
> > + RTE_TAILQ_HEAD(, sw_vector_adapter_data) adapter_list;
> > +};
>
> Why the indirect pointer to the lock? rather than embedding it in
> the structure?
IIUC, the lock itself is declared and used as a pointer right?
I looked at examples from test_mcslock.c, and this seemed correct.
RE: [RFC PATCH] build: reduce use of AVX compiler flags
> From: Bruce Richardson [mailto:bruce.richard...@intel.com]
> Sent: Wednesday, 26 March 2025 17.22
>
> On Tue, Mar 25, 2025 at 05:22:15PM +, Bruce Richardson wrote:
> > When doing a build for a target that already has the instruction sets
> > for AVX2/AVX512 enabled, skip emitting the AVX compiler flags, or the
> > skylake-avx512 '-march' flags, as they are unnecessary. Instead, when
> > the default flags produce the desired output, just use them
> unmodified.
> >
> > Depends-on: series-34915 ("remove component-specific logic for AVX
> builds")
> >
> > Signed-off-by: Bruce Richardson
> > ---
> >
> > This patchset depends on the previous AVX rework. However, sending it
> > separately as a new RFC because it effectively increases the minimum
> > compiler versions needed for x86 builds - from GCC 5 to 6, and
> > Clang 3.6 to 3.9.
> >
> > For now, I've just documented that as an additional note in the GSG
> that
> > these versions are recommended, but it would be simpler if we could
> just
> > set them as the required minimum baseline (at least in the docs).
> >
> > Feedback on these compiler version requirements welcome.
> >
>
> +techboard
>
> Ping for a little bit of feedback for this. Are we ok to bump the
> minimum
> compiler versions as described above, or will I continue with the
> approach
> in this RFC of keeping the minimum and just recommending the higher
> versions for x86 platforms?
>
> For reference GCC 6.1 was released April 2016[1], and, Clang 3.9 was
> released Sept 2016[2]
>
> /Bruce
>
> [1] https://gcc.gnu.org/gcc-6/
> [2] https://releases.llvm.org/
Considering GCC versions shipped with RHEL [3]...
We kind of support RHEL 7, but we already require a newer compiler (GCC 5) than
shipped with RHEL 7 (GCC 4.8).
RHEL 8 ships with GCC 8, which was released in May 2018 [4]. Maybe we can jump
to GCC 8?
BTW, we should also apply the same principle I argued [5] should apply for
upgrading the Kernel requirements: There should be a need for specific feature
or similar - which there is with your patch - and the details should be
mentioned in the release notes.
[3]: https://access.redhat.com/solutions/19458
[4]: https://gcc.gnu.org/gcc-8/
[5]:
https://inbox.dpdk.org/dev/CAMEVEZutf4sJ=eqfonw_bjw0tgtwqtbf_tk_y38qzblccco...@mail.gmail.com/T/#me7c8f1dbe4331ccf232d43512d6ddb51458c568a
Re: [EXTERNAL] Re: [RFC 2/2] eventdev: add default software vector adapter
On Wed, 26 Mar 2025 17:25:32 +
Pavan Nikhilesh Bhagavatula wrote:
> > On Wed, 26 Mar 2025 18:44:36 +0530
> > wrote:
> >
> > > +struct sw_vector_adapter_service_data {
> > > + uint32_t service_id;
> > > + RTE_ATOMIC(rte_mcslock_t *) lock;
> > > + RTE_TAILQ_HEAD(, sw_vector_adapter_data) adapter_list;
> > > +};
> >
> > Why the indirect pointer to the lock? rather than embedding it in
> > the structure?
>
> IIUC, the lock itself is declared and used as a pointer right?
> I looked at examples from test_mcslock.c, and this seemed correct.
>
Forgot, these locks used linked list of waiters, and root is a pointer.
[DPDK/ethdev Bug 1582] virtio: reader/writer lock mismatch
https://bugs.dpdk.org/show_bug.cgi?id=1582 Thomas Monjalon (tho...@monjalon.net) changed: What|Removed |Added Status|UNCONFIRMED |RESOLVED Resolution|--- |FIXED --- Comment #2 from Thomas Monjalon (tho...@monjalon.net) --- Resolved in http://git.dpdk.org/dpdk/commit/?id=22aa9a9c70 -- You are receiving this mail because: You are the assignee for the bug.
[PATCH 0/4] net/null optimizations
While reviewing use of null device for testing, noticed several things this driver is doing which are unnecessary. Stephen Hemminger (4): net/null: Tx optimizations net/null: fix packet copy net/null: optimize Rx net/null: count all queues drivers/net/null/#rte_eth_null.c# | 754 ++ drivers/net/null/rte_eth_null.c | 95 ++-- 2 files changed, 788 insertions(+), 61 deletions(-) create mode 100644 drivers/net/null/#rte_eth_null.c# -- 2.47.2
RE: [RFC v4 0/8] Symbol versioning and export rework
> From: David Marchand [mailto:david.march...@redhat.com] > Sent: Wednesday, 26 March 2025 13.03 > [...] > And I don't see symbol versioning as a MUST infrastructure that DPDK > needs to provide to datapath applications. Agree. Getting rid of the exotic stuff like symbol versioning would make the DPDK API easier understandable for DPDK application developers. Also, it's my impression that symbol versioning does not really play a major role in keeping or breaking ABI compatibility. Conclusion: Not worth the complexity.
[RFC v2 3/3] eventdev: refactor DMA adapter ops
From: Pavan Nikhilesh
Migrate all invocations of rte_event_dma_adapter_op
API to rte_dma_op.
Signed-off-by: Pavan Nikhilesh
Acked-by: Amit Prakash Shukla
---
app/test-eventdev/test_perf_common.c | 6 +-
app/test-eventdev/test_perf_common.h | 4 +-
app/test/test_event_dma_adapter.c | 6 +-
.../prog_guide/eventdev/event_dma_adapter.rst | 6 +-
drivers/dma/cnxk/cnxk_dmadev.c| 2 +-
drivers/dma/cnxk/cnxk_dmadev_fp.c | 12 ++--
lib/eventdev/rte_event_dma_adapter.c | 18 +++---
lib/eventdev/rte_event_dma_adapter.h | 57 ---
8 files changed, 27 insertions(+), 84 deletions(-)
diff --git a/app/test-eventdev/test_perf_common.c
b/app/test-eventdev/test_perf_common.c
index 627f07caa1..4e0109db52 100644
--- a/app/test-eventdev/test_perf_common.c
+++ b/app/test-eventdev/test_perf_common.c
@@ -562,11 +562,11 @@ crypto_adapter_enq_op_fwd(struct prod_data *p)
static inline void
dma_adapter_enq_op_fwd(struct prod_data *p)
{
- struct rte_event_dma_adapter_op *ops[BURST_SIZE] = {NULL};
+ struct rte_dma_op *ops[BURST_SIZE] = {NULL};
struct test_perf *t = p->t;
const uint32_t nb_flows = t->nb_flows;
const uint64_t nb_pkts = t->nb_pkts;
- struct rte_event_dma_adapter_op op;
+ struct rte_dma_op op;
struct rte_event evts[BURST_SIZE];
const uint8_t dev_id = p->dev_id;
struct evt_options *opt = t->opt;
@@ -2114,7 +2114,7 @@ perf_mempool_setup(struct evt_test *test, struct
evt_options *opt)
} else if (opt->prod_type == EVT_PROD_TYPE_EVENT_DMA_ADPTR) {
t->pool = rte_mempool_create(test->name, /* mempool name */
opt->pool_sz, /* number of
elements*/
-sizeof(struct
rte_event_dma_adapter_op) +
+sizeof(struct rte_dma_op) +
(sizeof(struct
rte_dma_sge) * 2),
cache_sz, /* cache
size*/
0, NULL, NULL, NULL, /* obj
constructor */
diff --git a/app/test-eventdev/test_perf_common.h
b/app/test-eventdev/test_perf_common.h
index d7333ad390..63078b0ee2 100644
--- a/app/test-eventdev/test_perf_common.h
+++ b/app/test-eventdev/test_perf_common.h
@@ -139,7 +139,7 @@ perf_mark_fwd_latency(enum evt_prod_type prod_type, struct
rte_event *const ev)
}
pe->timestamp = rte_get_timer_cycles();
} else if (prod_type == EVT_PROD_TYPE_EVENT_DMA_ADPTR) {
- struct rte_event_dma_adapter_op *op = ev->event_ptr;
+ struct rte_dma_op *op = ev->event_ptr;
op->user_meta = rte_get_timer_cycles();
} else {
@@ -297,7 +297,7 @@ perf_process_last_stage_latency(struct rte_mempool *const
pool, enum evt_prod_ty
tstamp = pe->timestamp;
rte_crypto_op_free(op);
} else if (prod_type == EVT_PROD_TYPE_EVENT_DMA_ADPTR) {
- struct rte_event_dma_adapter_op *op = ev->event_ptr;
+ struct rte_dma_op *op = ev->event_ptr;
to_free_in_bulk = op;
tstamp = op->user_meta;
diff --git a/app/test/test_event_dma_adapter.c
b/app/test/test_event_dma_adapter.c
index 9988d4fc7b..7f72a4e81d 100644
--- a/app/test/test_event_dma_adapter.c
+++ b/app/test/test_event_dma_adapter.c
@@ -234,7 +234,7 @@ test_op_forward_mode(void)
{
struct rte_mbuf *src_mbuf[TEST_MAX_OP];
struct rte_mbuf *dst_mbuf[TEST_MAX_OP];
- struct rte_event_dma_adapter_op *op;
+ struct rte_dma_op *op;
struct rte_event ev[TEST_MAX_OP];
int ret, i;
@@ -266,7 +266,7 @@ test_op_forward_mode(void)
op->vchan = TEST_DMA_VCHAN_ID;
op->event_meta = dma_response_info.event;
- /* Fill in event info and update event_ptr with
rte_event_dma_adapter_op */
+ /* Fill in event info and update event_ptr with rte_dma_op */
memset(&ev[i], 0, sizeof(struct rte_event));
ev[i].event = 0;
ev[i].op = RTE_EVENT_OP_NEW;
@@ -396,7 +396,7 @@ configure_dmadev(void)
rte_socket_id());
RTE_TEST_ASSERT_NOT_NULL(params.dst_mbuf_pool, "Can't create
DMA_DST_MBUFPOOL\n");
- elt_size = sizeof(struct rte_event_dma_adapter_op) + (sizeof(struct
rte_dma_sge) * 2);
+ elt_size = sizeof(struct rte_dma_op) + (sizeof(struct rte_dma_sge) * 2);
params.op_mpool = rte_mempool_create("EVENT_DMA_OP_POOL",
DMA_OP_POOL_SIZE, elt_size, 0,
0, NULL, NULL, NULL, NULL,
rte_socket_id(), 0);
RTE_TEST_ASSERT_NOT_NULL(params.op_mpool, "Can't create DMA_OP_POOL\n");
diff --git a/doc/guides/prog_guide/eventdev/event_dma_ada
[RFC v2 1/3] dmadev: add enqueue dequeue operations
From: Pavan Nikhilesh
Add enqueue/dequeue operations that use struct rte_dma_op
to communicate with the dma device.
These operations need to be enabled at dma device configuration
time by setting the flag rte_dma_conf::enable_enq_deq if the
device supports RTE_DMA_CAPA_OPS_ENQ_DEQ capability.
The enqueue dequeue operations are not compatible with
rte_dma_copy, rte_dma_copy_sg, rte_dma_fill, rte_dma_submit,
rte_dma_completed, rte_dma_completed_status range of APIs.
Signed-off-by: Pavan Nikhilesh
---
lib/dmadev/rte_dmadev.c | 18
lib/dmadev/rte_dmadev.h | 145 +++
lib/dmadev/rte_dmadev_core.h | 10 ++
lib/dmadev/rte_dmadev_trace_fp.h | 20
lib/dmadev/rte_dmadev_trace_points.c | 6 ++
5 files changed, 199 insertions(+)
diff --git a/lib/dmadev/rte_dmadev.c b/lib/dmadev/rte_dmadev.c
index 4a075ee1c6..a4dc65ac44 100644
--- a/lib/dmadev/rte_dmadev.c
+++ b/lib/dmadev/rte_dmadev.c
@@ -917,6 +917,22 @@ dummy_burst_capacity(__rte_unused const void *dev_private,
return 0;
}
+static uint16_t
+dummy_enqueue(__rte_unused void *dev_private, __rte_unused uint16_t vchan,
+ __rte_unused struct rte_dma_op **ops, __rte_unused uint16_t
nb_ops)
+{
+ RTE_DMA_LOG(ERR, "Enqueue not configured or not supported.");
+ return 0;
+}
+
+static uint16_t
+dummy_dequeue(__rte_unused void *dev_private, __rte_unused uint16_t vchan,
+ __rte_unused struct rte_dma_op **ops, __rte_unused uint16_t
nb_ops)
+{
+ RTE_DMA_LOG(ERR, "Enqueue not configured or not supported.");
+ return 0;
+}
+
static void
dma_fp_object_dummy(struct rte_dma_fp_object *obj)
{
@@ -928,6 +944,8 @@ dma_fp_object_dummy(struct rte_dma_fp_object *obj)
obj->completed= dummy_completed;
obj->completed_status = dummy_completed_status;
obj->burst_capacity = dummy_burst_capacity;
+ obj->enqueue = dummy_enqueue;
+ obj->dequeue = dummy_dequeue;
}
static int
diff --git a/lib/dmadev/rte_dmadev.h b/lib/dmadev/rte_dmadev.h
index 550dbfbf75..55ea07dccc 100644
--- a/lib/dmadev/rte_dmadev.h
+++ b/lib/dmadev/rte_dmadev.h
@@ -265,6 +265,11 @@ int16_t rte_dma_next_dev(int16_t start_dev_id);
* known from 'nb_priorities' field in struct rte_dma_info.
*/
#define RTE_DMA_CAPA_PRI_POLICY_SP RTE_BIT64(8)
+/** Support enqueue and dequeue operations.
+ *
+ * @see struct rte_dma_op
+ */
+#define RTE_DMA_CAPA_OPS_ENQ_DEQRTE_BIT64(9)
/** Support copy operation.
* This capability start with index of 32, so that it could leave gap between
@@ -351,6 +356,15 @@ struct rte_dma_conf {
* Lowest value indicates higher priority and vice-versa.
*/
uint16_t priority;
+ /** Indicates whether to use enqueue dequeue operations using
rte_dma_op.
+* false-default mode, true-enqueue, dequeue mode.
+* This value can be set to true only when ENQ_DEQ_OPS capability is
+* supported. When enabled, only calls to `rte_dma_enqueue_ops` and
+* `rte_dma_dequeue_ops` are valid.
+*
+* @see RTE_DMA_CAPA_OPS_ENQ_DEQ
+*/
+ bool enable_enq_deq;
};
/**
@@ -794,6 +808,63 @@ struct rte_dma_sge {
uint32_t length; /**< The DMA operation length. */
};
+/**
+ * A structure used to hold event based DMA operation entry. All the
information
+ * required for a DMA transfer shall be populated in "struct rte_dma_op"
+ * instance.
+ */
+struct rte_dma_op {
+ uint64_t flags;
+ /**< Flags related to the operation.
+* @see RTE_DMA_OP_FLAG_*
+*/
+ struct rte_mempool *op_mp;
+ /**< Mempool from which op is allocated. */
+ enum rte_dma_status_code status;
+ /**< Status code for this operation. */
+ uint32_t rsvd;
+ /**< Reserved for future use. */
+ uint64_t impl_opaque[2];
+ /**< Implementation-specific opaque data.
+* An dma device implementation use this field to hold
+* implementation specific values to share between dequeue and enqueue
+* operations.
+* The application should not modify this field.
+*/
+ uint64_t user_meta;
+ /**< Memory to store user specific metadata.
+* The dma device implementation should not modify this area.
+*/
+ uint64_t event_meta;
+ /**< Event metadata of DMA completion event.
+* Used when RTE_EVENT_DMA_ADAPTER_CAP_INTERNAL_PORT_VCHAN_EV_BIND is
not
+* supported in OP_NEW mode.
+* @see rte_event_dma_adapter_mode::RTE_EVENT_DMA_ADAPTER_OP_NEW
+* @see RTE_EVENT_DMA_ADAPTER_CAP_INTERNAL_PORT_VCHAN_EV_BIND
+*
+* Used when RTE_EVENT_DMA_ADAPTER_CAP_INTERNAL_PORT_OP_FWD is not
+* supported in OP_FWD mode.
+* @see rte_event_dma_adapter_mode::RTE_EVENT_DMA_ADAPTER_OP_FORWARD
+* @see RTE_EVENT_DMA_ADAPTER_CAP_INTERNAL_PORT_OP_FWD
+*
+* @see struct rte_e
[RFC v2 2/3] dma/cnxk: implement enqueue dequeue ops
From: Pavan Nikhilesh
Implement DMA enqueue/dequeue operations when
application enables it via configuration.
Signed-off-by: Pavan Nikhilesh
---
drivers/dma/cnxk/cnxk_dmadev.c| 25 +-
drivers/dma/cnxk/cnxk_dmadev.h| 7 ++
drivers/dma/cnxk/cnxk_dmadev_fp.c | 140 ++
3 files changed, 171 insertions(+), 1 deletion(-)
diff --git a/drivers/dma/cnxk/cnxk_dmadev.c b/drivers/dma/cnxk/cnxk_dmadev.c
index 90bb69011f..1ce3563250 100644
--- a/drivers/dma/cnxk/cnxk_dmadev.c
+++ b/drivers/dma/cnxk/cnxk_dmadev.c
@@ -19,7 +19,7 @@ cnxk_dmadev_info_get(const struct rte_dma_dev *dev, struct
rte_dma_info *dev_inf
dev_info->dev_capa = RTE_DMA_CAPA_MEM_TO_MEM | RTE_DMA_CAPA_MEM_TO_DEV |
RTE_DMA_CAPA_DEV_TO_MEM | RTE_DMA_CAPA_DEV_TO_DEV |
RTE_DMA_CAPA_OPS_COPY | RTE_DMA_CAPA_OPS_COPY_SG |
-RTE_DMA_CAPA_M2D_AUTO_FREE;
+RTE_DMA_CAPA_M2D_AUTO_FREE |
RTE_DMA_CAPA_OPS_ENQ_DEQ;
if (roc_feature_dpi_has_priority()) {
dev_info->dev_capa |= RTE_DMA_CAPA_PRI_POLICY_SP;
dev_info->nb_priorities = CN10K_DPI_MAX_PRI;
@@ -114,6 +114,21 @@ cnxk_dmadev_configure(struct rte_dma_dev *dev, const
struct rte_dma_conf *conf,
if (roc_feature_dpi_has_priority())
dpivf->rdpi.priority = conf->priority;
+ if (conf->enable_enq_deq) {
+ dev->fp_obj->copy = NULL;
+ dev->fp_obj->fill = NULL;
+ dev->fp_obj->submit = NULL;
+ dev->fp_obj->copy_sg = NULL;
+ dev->fp_obj->completed = NULL;
+ dev->fp_obj->completed_status = NULL;
+
+ dev->fp_obj->enqueue = cnxk_dma_ops_enqueue;
+ dev->fp_obj->dequeue = cnxk_dma_ops_dequeue;
+
+ if (roc_model_is_cn10k())
+ dev->fp_obj->enqueue = cn10k_dma_ops_enqueue;
+ }
+
return 0;
}
@@ -270,6 +285,14 @@ cnxk_dmadev_vchan_setup(struct rte_dma_dev *dev, uint16_t
vchan,
return -ENOMEM;
}
+ size = (max_desc * sizeof(struct rte_dma_op *));
+ dpi_conf->c_desc.ops = rte_zmalloc(NULL, size, RTE_CACHE_LINE_SIZE);
+ if (dpi_conf->c_desc.ops == NULL) {
+ plt_err("Failed to allocate for ops array");
+ rte_free(dpi_conf->c_desc.compl_ptr);
+ return -ENOMEM;
+ }
+
for (i = 0; i < max_desc; i++)
dpi_conf->c_desc.compl_ptr[i * CNXK_DPI_COMPL_OFFSET] =
CNXK_DPI_REQ_CDATA;
diff --git a/drivers/dma/cnxk/cnxk_dmadev.h b/drivers/dma/cnxk/cnxk_dmadev.h
index 9a232a5464..18039e43fb 100644
--- a/drivers/dma/cnxk/cnxk_dmadev.h
+++ b/drivers/dma/cnxk/cnxk_dmadev.h
@@ -93,6 +93,7 @@ struct cnxk_dpi_cdesc_data_s {
uint16_t head;
uint16_t tail;
uint8_t *compl_ptr;
+ struct rte_dma_op **ops;
};
struct cnxk_dpi_conf {
@@ -131,5 +132,11 @@ int cn10k_dmadev_copy(void *dev_private, uint16_t vchan,
rte_iova_t src, rte_iov
int cn10k_dmadev_copy_sg(void *dev_private, uint16_t vchan, const struct
rte_dma_sge *src,
const struct rte_dma_sge *dst, uint16_t nb_src,
uint16_t nb_dst,
uint64_t flags);
+uint16_t cnxk_dma_ops_enqueue(void *dev_private, uint16_t vchan, struct
rte_dma_op **ops,
+ uint16_t nb_ops);
+uint16_t cn10k_dma_ops_enqueue(void *dev_private, uint16_t vchan, struct
rte_dma_op **ops,
+ uint16_t nb_ops);
+uint16_t cnxk_dma_ops_dequeue(void *dev_private, uint16_t vchan, struct
rte_dma_op **ops,
+ uint16_t nb_ops);
#endif
diff --git a/drivers/dma/cnxk/cnxk_dmadev_fp.c
b/drivers/dma/cnxk/cnxk_dmadev_fp.c
index 36fc40c7e0..419425c386 100644
--- a/drivers/dma/cnxk/cnxk_dmadev_fp.c
+++ b/drivers/dma/cnxk/cnxk_dmadev_fp.c
@@ -665,3 +665,143 @@ cnxk_dma_adapter_dequeue(uintptr_t get_work1)
return (uintptr_t)op;
}
+
+uint16_t
+cnxk_dma_ops_enqueue(void *dev_private, uint16_t vchan, struct rte_dma_op
**ops, uint16_t nb_ops)
+{
+ struct cnxk_dpi_vf_s *dpivf = dev_private;
+ struct cnxk_dpi_conf *dpi_conf = &dpivf->conf[vchan];
+ const struct rte_dma_sge *fptr, *lptr;
+ uint16_t src, dst, nwords = 0;
+ struct rte_dma_op *op;
+ uint16_t space, i;
+ uint8_t *comp_ptr;
+ uint64_t hdr[4];
+ int rc;
+
+ space = (dpi_conf->c_desc.max_cnt + 1) -
+ ((dpi_conf->c_desc.tail - dpi_conf->c_desc.head) &
dpi_conf->c_desc.max_cnt);
+ space = RTE_MIN(space, nb_ops);
+
+ for (i = 0; i < space; i++) {
+ op = ops[i];
+ comp_ptr =
+ &dpi_conf->c_desc.compl_ptr[dpi_conf->c_desc.tail *
CNXK_DPI_COMPL_OFFSET];
+ dpi_conf->c_desc.ops[dpi_conf->c_desc.tail] = op;
+ CNXK_DPI_STRM_INC(dpi_conf->c_desc, tail);
+
+
[RFC v2 0/3] Introduce DMA enqueue/dequeue operations
From: Pavan Nikhilesh Introduce DMA enqueue/dequeue operations to the DMA device library. The enqueue/dequeue operations allow applications to communicate with the DMA device using the rte_dma_op structure, providing a more flexible and efficient way to manage DMA operations. The programming model for the enqueue/dequeue operations is as follows: * Query DMA devices capability for RTE_DMA_CAPA_OPS_ENQ_DEQ through rte_dma_info::dev_capa. * Enable enqueue/dequeue operations on DMA device by enabling rte_dma_conf::enable_enq_deq during device configuration. * Allocate a mempool for rte_dma_ops with object size of rte_dma_op + (sizeof(struct rte_dma_sge) * rte_dma_info::max_sges * 2). * Configure vchans and start the dma device. * Get an rte_dma_op from the mempool, fill it with the necessary information and use rte_dma_enqueue() to enqueue the operation. * Call rte_dma_dequeue() operation to get the array of finished operations. * Free the rte_dma_op back to the mempool. The enqueue dequeue operations are not compatible with rte_dma_copy, rte_dma_copy_sg, rte_dma_fill, rte_dma_submit, rte_dma_completed and rte_dma_completed_status range of APIs. Pavan Nikhilesh (3): dmadev: add enqueue dequeue operations dma/cnxk: implement enqueue dequeue ops eventdev: refactor DMA adapter ops app/test-eventdev/test_perf_common.c | 6 +- app/test-eventdev/test_perf_common.h | 4 +- app/test/test_event_dma_adapter.c | 6 +- .../prog_guide/eventdev/event_dma_adapter.rst | 6 +- drivers/dma/cnxk/cnxk_dmadev.c| 27 +++- drivers/dma/cnxk/cnxk_dmadev.h| 7 + drivers/dma/cnxk/cnxk_dmadev_fp.c | 152 +- lib/dmadev/rte_dmadev.c | 18 +++ lib/dmadev/rte_dmadev.h | 145 + lib/dmadev/rte_dmadev_core.h | 10 ++ lib/dmadev/rte_dmadev_trace_fp.h | 20 +++ lib/dmadev/rte_dmadev_trace_points.c | 6 + lib/eventdev/rte_event_dma_adapter.c | 18 +-- lib/eventdev/rte_event_dma_adapter.h | 57 --- 14 files changed, 397 insertions(+), 85 deletions(-) -- 2.43.0
RE: [EXTERNAL] Re: [patch v2 0/6] Support VMBUS channels without monitoring enabled
> Subject: Re: [EXTERNAL] Re: [patch v2 0/6] Support VMBUS channels without > monitoring enabled > > On Wed, 12 Mar 2025 00:33:52 + > Long Li wrote: > > > > Subject: [EXTERNAL] Re: [patch v2 0/6] Support VMBUS channels > > > without monitoring enabled > > > > > > On Mon, 10 Mar 2025 14:42:51 -0700 > > > lon...@linuxonhyperv.com wrote: > > > > > > > From: Long Li > > > > > > > > Hyperv may expose VMBUS channels without monitoring enabled. In > > > > this case, it programs almost all the data traffic to VF. > > > > > > > > This patchset enabled vmbus/netvsc to use channels without > > > > monitoring enabled. > > > > > > > > > CI still reports a build issue > > > > There are ABI changes to rte_vmbus_* calls. This patch added > rte_vmbus_device* as the 1st parameter to those calls. > > > > This will be a breaking change, and it only affects hn_netvsc as it's the > > only PMD > using the vmbus. > > > > Reading ./doc/guides/contributing/abi_policy.rst, I think the best option > > is to > use RTE_NEXT_ABI. But I can't find its definition in the code base. > > > > Please advise on how to proceed with making those breaking ABI changes. > > > > Thanks, > > Long > > Can't take it as is, here are some options: > > 1. Version the API even though should only be used internally. Use API > versioning >as transistion until 25.11. > 2. Wait for 25.11 and just fix it now, and do deprecation notice now. > > 3. Mark the API's as internal (in 25.11) and do deprecation notice now. > > 4. Make new functions with different names, and mark old ones as deprecated, > then remove in 25.11 Hi Stephen, I have sent deprecation notice and it has been accepted: https://patchwork.dpdk.org/project/dpdk/patch/1742242184-19600-1-git-send-email-lon...@linuxonhyperv.com/ Can you take this patch series? Thanks, Long
[PATCH 2/4] net/null: fix packet copy
If doing copy on transmit, can potentially copy past the data
in the mbuf. Change to only copy data from that segment.
Fixes: c743e50c475f ("null: new poll mode driver")
Cc: muk...@igel.co.jp
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
drivers/net/null/rte_eth_null.c | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/net/null/rte_eth_null.c b/drivers/net/null/rte_eth_null.c
index 3a05af2e32..baae81c572 100644
--- a/drivers/net/null/rte_eth_null.c
+++ b/drivers/net/null/rte_eth_null.c
@@ -35,7 +35,7 @@ struct null_queue {
struct pmd_internals *internals;
struct rte_mempool *mb_pool;
- struct rte_mbuf *dummy_packet;
+ void *dummy_packet;
RTE_ATOMIC(uint64_t) rx_pkts;
RTE_ATOMIC(uint64_t) tx_pkts;
@@ -156,14 +156,14 @@ eth_null_tx(void *q, struct rte_mbuf **bufs, uint16_t
nb_bufs)
static uint16_t
eth_null_copy_tx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs)
{
- int i;
struct null_queue *h = q;
- unsigned int packet_size;
+ unsigned int i;
- packet_size = h->internals->packet_size;
for (i = 0; i < nb_bufs; i++) {
- rte_memcpy(h->dummy_packet, rte_pktmbuf_mtod(bufs[i], void *),
- packet_size);
+ struct rte_mbuf *m = bufs[i];
+ unsigned int len = RTE_MIN(h->internals->packet_size,
m->data_len);
+
+ rte_memcpy(h->dummy_packet, rte_pktmbuf_mtod(m, void *), len);
rte_pktmbuf_free(bufs[i]);
}
--
2.47.2
Re: [RFC PATCH] build: reduce use of AVX compiler flags
On Wed, 26 Mar 2025 19:06:58 +0100
Morten Brørup wrote:
> > From: Bruce Richardson [mailto:bruce.richard...@intel.com]
> > Sent: Wednesday, 26 March 2025 17.22
> >
> > On Tue, Mar 25, 2025 at 05:22:15PM +, Bruce Richardson wrote:
> > > When doing a build for a target that already has the instruction sets
> > > for AVX2/AVX512 enabled, skip emitting the AVX compiler flags, or the
> > > skylake-avx512 '-march' flags, as they are unnecessary. Instead, when
> > > the default flags produce the desired output, just use them
> > unmodified.
> > >
> > > Depends-on: series-34915 ("remove component-specific logic for AVX
> > builds")
> > >
> > > Signed-off-by: Bruce Richardson
> > > ---
> > >
> > > This patchset depends on the previous AVX rework. However, sending it
> > > separately as a new RFC because it effectively increases the minimum
> > > compiler versions needed for x86 builds - from GCC 5 to 6, and
> > > Clang 3.6 to 3.9.
> > >
> > > For now, I've just documented that as an additional note in the GSG
> > that
> > > these versions are recommended, but it would be simpler if we could
> > just
> > > set them as the required minimum baseline (at least in the docs).
> > >
> > > Feedback on these compiler version requirements welcome.
> > >
> >
> > +techboard
> >
> > Ping for a little bit of feedback for this. Are we ok to bump the
> > minimum
> > compiler versions as described above, or will I continue with the
> > approach
> > in this RFC of keeping the minimum and just recommending the higher
> > versions for x86 platforms?
> >
> > For reference GCC 6.1 was released April 2016[1], and, Clang 3.9 was
> > released Sept 2016[2]
> >
> > /Bruce
> >
> > [1] https://gcc.gnu.org/gcc-6/
> > [2] https://releases.llvm.org/
>
> Considering GCC versions shipped with RHEL [3]...
> We kind of support RHEL 7, but we already require a newer compiler (GCC 5)
> than shipped with RHEL 7 (GCC 4.8).
> RHEL 8 ships with GCC 8, which was released in May 2018 [4]. Maybe we can
> jump to GCC 8?
>
> BTW, we should also apply the same principle I argued [5] should apply for
> upgrading the Kernel requirements: There should be a need for specific
> feature or similar - which there is with your patch - and the details should
> be mentioned in the release notes.
>
> [3]: https://access.redhat.com/solutions/19458
> [4]: https://gcc.gnu.org/gcc-8/
> [5]:
> https://inbox.dpdk.org/dev/CAMEVEZutf4sJ=eqfonw_bjw0tgtwqtbf_tk_y38qzblccco...@mail.gmail.com/T/#me7c8f1dbe4331ccf232d43512d6ddb51458c568a
>
RHEL 7 reached end of life on June 30, 2024.
DPDK need no longer support it on future versions.
[PATCH v6 04/14] net/zxdh: modify dtb queue ops
Modify the implementation of the dtb queue
request and release interfaces,
and add the implementation of queue initialization.
Signed-off-by: Bingbin Chen
---
drivers/net/zxdh/zxdh_np.c | 491 +
drivers/net/zxdh/zxdh_np.h | 97
2 files changed, 482 insertions(+), 106 deletions(-)
diff --git a/drivers/net/zxdh/zxdh_np.c b/drivers/net/zxdh/zxdh_np.c
index 00e02cb36f..3f5d286ffe 100644
--- a/drivers/net/zxdh/zxdh_np.c
+++ b/drivers/net/zxdh/zxdh_np.c
@@ -17,16 +17,15 @@
static ZXDH_DEV_MGR_T g_dev_mgr;
static ZXDH_SDT_MGR_T g_sdt_mgr;
-static uint32_t g_dpp_dtb_int_enable;
static uint32_t g_table_type[ZXDH_DEV_CHANNEL_MAX][ZXDH_DEV_SDT_ID_MAX];
static ZXDH_PPU_CLS_BITMAP_T g_ppu_cls_bit_map[ZXDH_DEV_CHANNEL_MAX];
static ZXDH_DTB_MGR_T *p_dpp_dtb_mgr[ZXDH_DEV_CHANNEL_MAX];
-static ZXDH_RISCV_DTB_MGR *p_riscv_dtb_queue_mgr[ZXDH_DEV_CHANNEL_MAX];
static ZXDH_SDT_TBL_DATA_T
g_sdt_info[ZXDH_DEV_CHANNEL_MAX][ZXDH_DEV_SDT_ID_MAX];
static ZXDH_PPU_STAT_CFG_T g_ppu_stat_cfg[ZXDH_DEV_CHANNEL_MAX];
static uint64_t g_np_fw_compat_addr[ZXDH_DEV_CHANNEL_MAX];
static const ZXDH_VERSION_COMPATIBLE_REG_T g_np_sdk_version = {
ZXDH_NPSDK_COMPAT_ITEM_ID, 1, 0, 0, 0, {0} };
+static const uint32_t hardware_ep_id[5] = {5, 6, 7, 8, 9};
static const ZXDH_FIELD_T g_smmu0_smmu0_cpu_ind_cmd_reg[] = {
{"cpu_ind_rw", ZXDH_FIELD_FLAG_RW, 31, 1, 0x0, 0x0},
@@ -544,6 +543,25 @@ zxdh_np_dev_opr_spinlock_get(uint32_t dev_id, uint32_t
type, ZXDH_SPINLOCK_T **p
return ZXDH_OK;
}
+static uint32_t
+zxdh_np_dev_dtb_opr_spinlock_get(uint32_t dev_id, uint32_t type,
+ uint32_t index, ZXDH_SPINLOCK_T **p_spinlock_out)
+{
+ ZXDH_DEV_MGR_T *p_dev_mgr = &g_dev_mgr;
+ ZXDH_DEV_CFG_T *p_dev_info = p_dev_mgr->p_dev_array[dev_id];
+
+ switch (type) {
+ case ZXDH_DEV_SPINLOCK_T_DTB:
+ *p_spinlock_out = &p_dev_info->dtb_queue_spinlock[index];
+ break;
+ default:
+ PMD_DRV_LOG(ERR, "spinlock type is invalid!");
+ return ZXDH_ERR;
+ }
+
+ return ZXDH_OK;
+}
+
static uint32_t
zxdh_np_dev_read_channel(uint32_t dev_id, uint32_t addr, uint32_t size,
uint32_t *p_data)
{
@@ -938,6 +956,7 @@ zxdh_np_dev_add(uint32_t dev_id, ZXDH_DEV_TYPE_E dev_type,
{
ZXDH_DEV_CFG_T *p_dev_info = NULL;
ZXDH_DEV_MGR_T *p_dev_mgr = NULL;
+ uint32_t i = 0;
p_dev_mgr = &g_dev_mgr;
if (!p_dev_mgr->is_init) {
@@ -973,7 +992,9 @@ zxdh_np_dev_add(uint32_t dev_id, ZXDH_DEV_TYPE_E dev_type,
p_dev_info->p_pcie_read_fun = zxdh_np_dev_pcie_default_read;
rte_spinlock_init(&p_dev_info->dtb_spinlock.spinlock);
- rte_spinlock_init(&p_dev_info->smmu0_spinlock.spinlock);
+
+ for (i = 0; i < ZXDH_DTB_QUEUE_NUM_MAX; i++)
+ rte_spinlock_init(&p_dev_info->dtb_queue_spinlock[i].spinlock);
return ZXDH_OK;
}
@@ -1329,6 +1350,89 @@ zxdh_np_agent_channel_reg_write(uint32_t dev_id,
return ret;
}
+static uint32_t
+zxdh_np_agent_channel_dtb_sync_send(uint32_t dev_id,
+
ZXDH_AGENT_CHANNEL_DTB_MSG_T *p_msg,
+ uint32_t *p_data,
+ uint32_t rep_len)
+{
+ uint32_t ret = ZXDH_OK;
+
+ ZXDH_AGENT_CHANNEL_MSG_T agent_msg = {0};
+ agent_msg.msg = (void *)p_msg;
+ agent_msg.msg_len = sizeof(ZXDH_AGENT_CHANNEL_DTB_MSG_T);
+
+ ret = zxdh_np_agent_channel_sync_send(dev_id, &agent_msg, p_data,
rep_len);
+ if (ret != ZXDH_OK) {
+ PMD_DRV_LOG(ERR, "zxdh_np_agent_channel_sync_send failed");
+ return ZXDH_ERR;
+ }
+
+ return ZXDH_OK;
+}
+
+static uint32_t
+zxdh_np_agent_channel_dtb_queue_request(uint32_t dev_id,
+ char
p_name[32],
+
uint32_t vport_info,
+
uint32_t *p_queue_id)
+{
+ uint32_t rc = ZXDH_OK;
+
+ uint32_t rsp_buff[2] = {0};
+ uint32_t msg_result = 0;
+ uint32_t queue_id = 0;
+ ZXDH_AGENT_CHANNEL_DTB_MSG_T msgcfg = {
+ .dev_id = 0,
+ .type= ZXDH_DTB_MSG,
+ .oper= ZXDH_QUEUE_REQUEST,
+ .vport = vport_info,
+ };
+ memcpy(msgcfg.name, p_name, strnlen(p_name, ZXDH_PORT_NAME_MAX));
+
+ PMD_DRV_LOG(DEBUG, "msgcfg.name=%s", msgcfg.name);
+
+ rc = zxdh_np_agent_channel_dtb_sync_send(dev_id, &msgcfg, rsp_buff,
sizeof(rsp_buff));
+ ZXDH_COMM_CHECK_DEV_RC(dev_id, rc,
"zxdh_np_agent_channel_dtb_sync_send");
+
+ msg_result = rsp_buff[0];
+ queue_id = rsp_buff[1];
+
+ PMD_DRV_LOG(DEBUG, "dev_id: %u, msg_result: %u", dev_id, msg_r
[PATCH 3/4] net/null: optimize Rx
No other rx_burst function checks args, remove it.
Since rx_burst can only safely be called by a single thread
at a time, there is no need for atomic operations on statistics.
Signed-off-by: Stephen Hemminger
---
drivers/net/null/rte_eth_null.c | 22 +-
1 file changed, 5 insertions(+), 17 deletions(-)
diff --git a/drivers/net/null/rte_eth_null.c b/drivers/net/null/rte_eth_null.c
index baae81c572..7ac29b3f81 100644
--- a/drivers/net/null/rte_eth_null.c
+++ b/drivers/net/null/rte_eth_null.c
@@ -37,7 +37,7 @@ struct null_queue {
struct rte_mempool *mb_pool;
void *dummy_packet;
- RTE_ATOMIC(uint64_t) rx_pkts;
+ uint64_t rx_pkts;
RTE_ATOMIC(uint64_t) tx_pkts;
};
@@ -88,9 +88,6 @@ eth_null_rx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs)
struct null_queue *h = q;
unsigned int packet_size;
- if ((q == NULL) || (bufs == NULL))
- return 0;
-
packet_size = h->internals->packet_size;
if (rte_pktmbuf_alloc_bulk(h->mb_pool, bufs, nb_bufs) != 0)
return 0;
@@ -101,10 +98,8 @@ eth_null_rx(void *q, struct rte_mbuf **bufs, uint16_t
nb_bufs)
bufs[i]->port = h->internals->port_id;
}
- /* NOTE: review for potential ordering optimization */
- rte_atomic_fetch_add_explicit(&h->rx_pkts, i, rte_memory_order_seq_cst);
-
- return i;
+ h->rx_pkts += nb_bufs;
+ return nb_bufs;
}
static uint16_t
@@ -114,9 +109,6 @@ eth_null_copy_rx(void *q, struct rte_mbuf **bufs, uint16_t
nb_bufs)
struct null_queue *h = q;
unsigned int packet_size;
- if ((q == NULL) || (bufs == NULL))
- return 0;
-
packet_size = h->internals->packet_size;
if (rte_pktmbuf_alloc_bulk(h->mb_pool, bufs, nb_bufs) != 0)
return 0;
@@ -129,10 +121,8 @@ eth_null_copy_rx(void *q, struct rte_mbuf **bufs, uint16_t
nb_bufs)
bufs[i]->port = h->internals->port_id;
}
- /* NOTE: review for potential ordering optimization */
- rte_atomic_fetch_add_explicit(&h->rx_pkts, i, rte_memory_order_seq_cst);
-
- return i;
+ h->rx_pkts += nb_bufs;
+ return nb_bufs;
}
static uint16_t
@@ -326,7 +316,6 @@ eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats
*igb_stats)
RTE_MIN(dev->data->nb_rx_queues,
RTE_DIM(internal->rx_null_queues)));
for (i = 0; i < num_stats; i++) {
- /* NOTE: review for atomic access */
igb_stats->q_ipackets[i] =
internal->rx_null_queues[i].rx_pkts;
rx_total += igb_stats->q_ipackets[i];
@@ -360,7 +349,6 @@ eth_stats_reset(struct rte_eth_dev *dev)
internal = dev->data->dev_private;
for (i = 0; i < RTE_DIM(internal->rx_null_queues); i++)
- /* NOTE: review for atomic access */
internal->rx_null_queues[i].rx_pkts = 0;
for (i = 0; i < RTE_DIM(internal->tx_null_queues); i++)
internal->tx_null_queues[i].tx_pkts = 0;
--
2.47.2
[PATCH 4/4] net/null: count all queues
If RTE_ETHDEV_QUEUE_STAT_CNTRS is less than the number of queues
in a device, the device should still count packets for all queues.
Remove the igb_ prefix which was inherited from other driver.
Signed-off-by: Stephen Hemminger
---
drivers/net/null/rte_eth_null.c | 31 +--
1 file changed, 13 insertions(+), 18 deletions(-)
diff --git a/drivers/net/null/rte_eth_null.c b/drivers/net/null/rte_eth_null.c
index 7ac29b3f81..82b49d471d 100644
--- a/drivers/net/null/rte_eth_null.c
+++ b/drivers/net/null/rte_eth_null.c
@@ -304,31 +304,26 @@ eth_dev_info(struct rte_eth_dev *dev,
static int
eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *igb_stats)
{
- unsigned int i, num_stats;
+ const struct pmd_internals *internal = dev->data->dev_private;
+ unsigned int i;
unsigned long rx_total = 0, tx_total = 0;
- const struct pmd_internals *internal;
- if ((dev == NULL) || (igb_stats == NULL))
- return -EINVAL;
+ for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ uint64_t pkts = internal->rx_null_queues[i].rx_pkts;
- internal = dev->data->dev_private;
- num_stats = RTE_MIN((unsigned int)RTE_ETHDEV_QUEUE_STAT_CNTRS,
- RTE_MIN(dev->data->nb_rx_queues,
- RTE_DIM(internal->rx_null_queues)));
- for (i = 0; i < num_stats; i++) {
- igb_stats->q_ipackets[i] =
- internal->rx_null_queues[i].rx_pkts;
- rx_total += igb_stats->q_ipackets[i];
+ if (i < RTE_ETHDEV_QUEUE_STAT_CNTRS)
+ igb_stats->q_ipackets[i] = pkts;
+
+ rx_total += pkts;
}
- num_stats = RTE_MIN((unsigned int)RTE_ETHDEV_QUEUE_STAT_CNTRS,
- RTE_MIN(dev->data->nb_tx_queues,
- RTE_DIM(internal->tx_null_queues)));
- for (i = 0; i < num_stats; i++) {
+ for (i = 0; i < dev->data->nb_tx_queues; i++) {
uint64_t pkts =
rte_atomic_load_explicit(&internal->tx_null_queues[i].tx_pkts,
- rte_memory_order_relaxed);
+
rte_memory_order_relaxed);
+
+ if (i < RTE_ETHDEV_QUEUE_STAT_CNTRS)
+ igb_stats->q_opackets[i] = pkts;
- igb_stats->q_opackets[i] = pkts;
tx_total += pkts;
}
--
2.47.2
[PATCH 1/4] net/null: Tx optimizations
All the null device does is call pktmbuf_free, can be marked lockless.
The statitistics need to use atomic for this.
Use rte_pktmbuf_free_bulk instead of a loop.
And pktmbuf_free handles multi-segment packets without problems.
There is no reason for eth_null_tx to check arguments,
no other tx_burst function does.
Signed-off-by: Stephen Hemminger
---
drivers/net/null/#rte_eth_null.c# | 754 ++
drivers/net/null/rte_eth_null.c | 34 +-
2 files changed, 766 insertions(+), 22 deletions(-)
create mode 100644 drivers/net/null/#rte_eth_null.c#
diff --git a/drivers/net/null/#rte_eth_null.c#
b/drivers/net/null/#rte_eth_null.c#
new file mode 100644
index 00..9bac780e9e
--- /dev/null
+++ b/drivers/net/null/#rte_eth_null.c#
@@ -0,0 +1,754 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright (C) IGEL Co.,Ltd.
+ * All rights reserved.
+ */
+
+#include
+
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+
+#define ETH_NULL_PACKET_SIZE_ARG "size"
+#define ETH_NULL_PACKET_COPY_ARG "copy"
+#define ETH_NULL_PACKET_NO_RX_ARG "no-rx"
+
+static unsigned int default_packet_size = 64;
+static unsigned int default_packet_copy;
+static unsigned int default_no_rx;
+
+static const char *valid_arguments[] = {
+ ETH_NULL_PACKET_SIZE_ARG,
+ ETH_NULL_PACKET_COPY_ARG,
+ ETH_NULL_PACKET_NO_RX_ARG,
+ NULL
+};
+
+struct pmd_internals;
+
+struct null_queue {
+ struct pmd_internals *internals;
+
+ struct rte_mempool *mb_pool;
+ void *dummy_packet;
+
+ uint64_t rx_pkts;
+ RTE_ATOMIC(uint64_t) tx_pkts;
+};
+
+struct pmd_options {
+ unsigned int packet_copy;
+ unsigned int packet_size;
+ unsigned int no_rx;
+};
+
+struct pmd_internals {
+ unsigned int packet_size;
+ unsigned int packet_copy;
+ unsigned int no_rx;
+ uint16_t port_id;
+
+ struct null_queue rx_null_queues[RTE_MAX_QUEUES_PER_PORT];
+ struct null_queue tx_null_queues[RTE_MAX_QUEUES_PER_PORT];
+
+ struct rte_ether_addr eth_addr;
+ /** Bit mask of RSS offloads, the bit offset also means flow type */
+ uint64_t flow_type_rss_offloads;
+
+ rte_spinlock_t rss_lock;
+
+ uint16_t reta_size;
+ struct rte_eth_rss_reta_entry64 reta_conf[RTE_ETH_RSS_RETA_SIZE_128 /
+ RTE_ETH_RETA_GROUP_SIZE];
+
+ uint8_t rss_key[40];/**< 40-byte hash key. */
+};
+static struct rte_eth_link pmd_link = {
+ .link_speed = RTE_ETH_SPEED_NUM_10G,
+ .link_duplex = RTE_ETH_LINK_FULL_DUPLEX,
+ .link_status = RTE_ETH_LINK_DOWN,
+ .link_autoneg = RTE_ETH_LINK_FIXED,
+};
+
+RTE_LOG_REGISTER_DEFAULT(eth_null_logtype, NOTICE);
+#define RTE_LOGTYPE_ETH_NULL eth_null_logtype
+
+#define PMD_LOG(level, ...) \
+ RTE_LOG_LINE_PREFIX(level, ETH_NULL, "%s(): ", __func__, __VA_ARGS__)
+
+static uint16_t
+eth_null_rx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs)
+{
+ int i;
+ struct null_queue *h = q;
+ unsigned int packet_size;
+
+ packet_size = h->internals->packet_size;
+ if (rte_pktmbuf_alloc_bulk(h->mb_pool, bufs, nb_bufs) != 0)
+ return 0;
+
+ for (i = 0; i < nb_bufs; i++) {
+ bufs[i]->data_len = (uint16_t)packet_size;
+ bufs[i]->pkt_len = packet_size;
+ bufs[i]->port = h->internals->port_id;
+ }
+
+ h->rx_pkts += nb_bufs;
+ return nb_bufs;
+}
+
+static uint16_t
+eth_null_copy_rx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs)
+{
+ int i;
+ struct null_queue *h = q;
+ unsigned int packet_size;
+
+ packet_size = h->internals->packet_size;
+ if (rte_pktmbuf_alloc_bulk(h->mb_pool, bufs, nb_bufs) != 0)
+ return 0;
+
+ for (i = 0; i < nb_bufs; i++) {
+ rte_memcpy(rte_pktmbuf_mtod(bufs[i], void *), h->dummy_packet,
+ packet_size);
+ bufs[i]->data_len = (uint16_t)packet_size;
+ bufs[i]->pkt_len = packet_size;
+ bufs[i]->port = h->internals->port_id;
+ }
+
+ h->rx_pkts += nb_bufs;
+ return nb_bufs;
+}
+
+static uint16_t
+eth_null_no_rx(void *q __rte_unused, struct rte_mbuf **bufs __rte_unused,
+ uint16_t nb_bufs __rte_unused)
+{
+ return 0;
+}
+
+static uint16_t
+eth_null_tx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs)
+{
+ struct null_queue *h = q;
+
+ rte_pktmbuf_free_bulk(bufs, nb_bufs);
+ rte_atomic_fetch_add_explicit(&h->tx_pkts, nb_bufs,
rte_memory_order_relaxed);
+
+ return nb_bufs;
+}
+
+static uint16_t
+eth_null_copy_tx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs)
+{
+ struct null_queue *h = q;
+ unsigned int i;
+
+ for (i = 0; i < nb_bufs; i++) {
+ struct rte_mbuf *m = bufs[i];
+ unsigned int len
+ = RTE_MIN(h
RE: [RFC v2 0/3] Introduce DMA enqueue/dequeue operations
++ > -Original Message- > From: pbhagavat...@marvell.com > Sent: Thursday, March 27, 2025 1:07 AM > To: Jerin Jacob > Cc: dev@dpdk.org; Pavan Nikhilesh Bhagavatula > > Subject: [RFC v2 0/3] Introduce DMA enqueue/dequeue operations > > From: Pavan Nikhilesh > > Introduce DMA enqueue/dequeue operations to the DMA device library. > > The enqueue/dequeue operations allow applications to communicate with the > DMA device using the rte_dma_op structure, providing a more flexible and > efficient way to manage DMA operations. > > The programming model for the enqueue/dequeue operations is as follows: > > * Query DMA devices capability for RTE_DMA_CAPA_OPS_ENQ_DEQ through > rte_dma_info::dev_capa. > * Enable enqueue/dequeue operations on DMA device by enabling > rte_dma_conf::enable_enq_deq during device configuration. > * Allocate a mempool for rte_dma_ops with object size of > rte_dma_op + (sizeof(struct rte_dma_sge) * rte_dma_info::max_sges * 2). > * Configure vchans and start the dma device. > * Get an rte_dma_op from the mempool, fill it with the necessary > information and use rte_dma_enqueue() to enqueue the operation. > * Call rte_dma_dequeue() operation to get the array of finished operations. > * Free the rte_dma_op back to the mempool. > > The enqueue dequeue operations are not compatible with rte_dma_copy, > rte_dma_copy_sg, rte_dma_fill, rte_dma_submit, rte_dma_completed and > rte_dma_completed_status range of APIs. > > Pavan Nikhilesh (3): > dmadev: add enqueue dequeue operations > dma/cnxk: implement enqueue dequeue ops > eventdev: refactor DMA adapter ops > > app/test-eventdev/test_perf_common.c | 6 +- > app/test-eventdev/test_perf_common.h | 4 +- > app/test/test_event_dma_adapter.c | 6 +- > .../prog_guide/eventdev/event_dma_adapter.rst | 6 +- > drivers/dma/cnxk/cnxk_dmadev.c| 27 +++- > drivers/dma/cnxk/cnxk_dmadev.h| 7 + > drivers/dma/cnxk/cnxk_dmadev_fp.c | 152 +- > lib/dmadev/rte_dmadev.c | 18 +++ > lib/dmadev/rte_dmadev.h | 145 + > lib/dmadev/rte_dmadev_core.h | 10 ++ > lib/dmadev/rte_dmadev_trace_fp.h | 20 +++ > lib/dmadev/rte_dmadev_trace_points.c | 6 + > lib/eventdev/rte_event_dma_adapter.c | 18 +-- > lib/eventdev/rte_event_dma_adapter.h | 57 --- > 14 files changed, 397 insertions(+), 85 deletions(-) > > -- > 2.43.0
[PATCH] version: 25.07-rc0
Start a new release cycle with empty release notes. Bump version and ABI minor. Signed-off-by: David Marchand --- .github/workflows/build.yml| 2 +- ABI_VERSION| 2 +- VERSION| 2 +- doc/guides/rel_notes/index.rst | 1 + doc/guides/rel_notes/release_25_07.rst | 138 + 5 files changed, 142 insertions(+), 3 deletions(-) create mode 100644 doc/guides/rel_notes/release_25_07.rst diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml index fba46b920f..0cc4d12b0b 100644 --- a/.github/workflows/build.yml +++ b/.github/workflows/build.yml @@ -12,7 +12,7 @@ defaults: env: REF_GIT_BRANCH: main REF_GIT_REPO: https://github.com/DPDK/dpdk - REF_GIT_TAG: v24.11 + REF_GIT_TAG: v25.03 jobs: checkpatch: diff --git a/ABI_VERSION b/ABI_VERSION index 8b9bee5b58..a5615e1308 100644 --- a/ABI_VERSION +++ b/ABI_VERSION @@ -1 +1 @@ -25.1 +25.2 diff --git a/VERSION b/VERSION index 16164a3867..fedf12952a 100644 --- a/VERSION +++ b/VERSION @@ -1 +1 @@ -25.03.0 +25.07.0-rc0 diff --git a/doc/guides/rel_notes/index.rst b/doc/guides/rel_notes/index.rst index fc0309113e..6462f01966 100644 --- a/doc/guides/rel_notes/index.rst +++ b/doc/guides/rel_notes/index.rst @@ -8,6 +8,7 @@ Release Notes :maxdepth: 1 :numbered: +release_25_07 release_25_03 release_24_11 release_24_07 diff --git a/doc/guides/rel_notes/release_25_07.rst b/doc/guides/rel_notes/release_25_07.rst new file mode 100644 index 00..cd1025aac0 --- /dev/null +++ b/doc/guides/rel_notes/release_25_07.rst @@ -0,0 +1,138 @@ +.. SPDX-License-Identifier: BSD-3-Clause + Copyright 2025 The DPDK contributors + +.. include:: + +DPDK Release 25.07 +== + +.. **Read this first.** + + The text in the sections below explains how to update the release notes. + + Use proper spelling, capitalization and punctuation in all sections. + + Variable and config names should be quoted as fixed width text: + ``LIKE_THIS``. + + Build the docs and view the output file to ensure the changes are correct:: + + ninja -C build doc + xdg-open build/doc/guides/html/rel_notes/release_25_07.html + + +New Features + + +.. This section should contain new features added in this release. + Sample format: + + * **Add a title in the past tense with a full stop.** + + Add a short 1-2 sentence description in the past tense. + The description should be enough to allow someone scanning + the release notes to understand the new feature. + + If the feature adds a lot of sub-features you can use a bullet list + like this: + + * Added feature foo to do something. + * Enhanced feature bar to do something else. + + Refer to the previous release notes for examples. + + Suggested order in release notes items: + * Core libs (EAL, mempool, ring, mbuf, buses) + * Device abstraction libs and PMDs (ordered alphabetically by vendor name) + - ethdev (lib, PMDs) + - cryptodev (lib, PMDs) + - eventdev (lib, PMDs) + - etc + * Other libs + * Apps, Examples, Tools (if significant) + + This section is a comment. Do not overwrite or remove it. + Also, make sure to start the actual text at the margin. + === + + +Removed Items +- + +.. This section should contain removed items in this release. Sample format: + + * Add a short 1-2 sentence description of the removed item + in the past tense. + + This section is a comment. Do not overwrite or remove it. + Also, make sure to start the actual text at the margin. + === + + +API Changes +--- + +.. This section should contain API changes. Sample format: + + * sample: Add a short 1-2 sentence description of the API change + which was announced in the previous releases and made in this release. + Start with a scope label like "ethdev:". + Use fixed width quotes for ``function_names`` or ``struct_names``. + Use the past tense. + + This section is a comment. Do not overwrite or remove it. + Also, make sure to start the actual text at the margin. + === + + +ABI Changes +--- + +.. This section should contain ABI changes. Sample format: + + * sample: Add a short 1-2 sentence description of the ABI change + which was announced in the previous releases and made in this release. + Start with a scope label like "ethdev:". + Use fixed width quotes for ``function_names`` or ``struct_names``. + Use the past tense. + + This section is a comment. Do not overwrite or remove it. + Also, make sure to start the actual text at the margin. + === + +* No ABI change that would break compatibility with 24.11. + + +Kn
Re: [PATCH] version: 25.07-rc0
On Wed, Mar 26, 2025 at 10:00:03AM +0100, David Marchand wrote: > Start a new release cycle with empty release notes. > Bump version and ABI minor. > > Signed-off-by: David Marchand > --- One query below, otherwise LGTM Acked-by: Bruce Richardson > .github/workflows/build.yml| 2 +- > ABI_VERSION| 2 +- > VERSION| 2 +- > doc/guides/rel_notes/index.rst | 1 + > doc/guides/rel_notes/release_25_07.rst | 138 + > 5 files changed, 142 insertions(+), 3 deletions(-) > create mode 100644 doc/guides/rel_notes/release_25_07.rst > > diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml > index fba46b920f..0cc4d12b0b 100644 > --- a/.github/workflows/build.yml > +++ b/.github/workflows/build.yml > @@ -12,7 +12,7 @@ defaults: > env: >REF_GIT_BRANCH: main >REF_GIT_REPO: https://github.com/DPDK/dpdk > - REF_GIT_TAG: v24.11 > + REF_GIT_TAG: v25.03 > Do we not keep comparing against 24.11 for the whole of the 25-ABI cycle?
RE: [PATCH v5 00/11] remove component-specific logic for AVX builds
> A number of libs and drivers had special optimized AVX2 and AVX512 code > paths for performance reasons, and these tended to have copy-pasted > logic to build those files. Centralise that logic in the main > drivers/ and lib/ meson.build files to avoid duplication. > > v5: fix RTE_ARCH_X86 macro, which broke crc library > v4: rebase on latest main branch > minor fixes following feedback > limit use of -march=skylake-avx512 to when we don't already have a > -march flag supporting AVX512. > v3: add patch for event/dlb2 AVX512 handling. > add common code for libraries as well as drivers. > v2: add patch 4 to remove use of unnecessary CC_AVX2_SUPPORT flag > > > Bruce Richardson (11): > drivers: add generalized AVX build handling > net/intel: use common AVX build code > drivers/net: build use common AVX handling > drivers/net: remove AVX2 build-time define > event/dlb2: build using common AVX handling > lib: add generalized AVX build handling > acl: use common AVX build handling > fib: use common AVX build handling > net: simplify build-time logic for x86 > net: use common AVX512 build code > member: use common AVX512 build support > > drivers/event/dlb2/dlb2_sse.c | 4 ++ > drivers/event/dlb2/meson.build| 16 +--- > drivers/meson.build | 30 ++ > drivers/net/bnxt/bnxt_ethdev.c| 2 - > drivers/net/bnxt/meson.build | 10 + > drivers/net/enic/meson.build | 10 + > drivers/net/intel/i40e/meson.build| 26 +--- > drivers/net/intel/iavf/meson.build| 25 +--- > drivers/net/intel/ice/meson.build | 25 +--- > drivers/net/intel/idpf/meson.build| 25 +--- > drivers/net/nfp/meson.build | 10 + > drivers/net/octeon_ep/meson.build | 13 +- > drivers/net/octeon_ep/otx_ep_ethdev.c | 4 -- > drivers/net/virtio/meson.build| 9 + > lib/acl/meson.build | 54 ++--- > lib/fib/dir24_8.c | 6 +-- > lib/fib/meson.build | 18 + > lib/fib/trie.c| 6 +-- > lib/member/meson.build| 46 - > lib/meson.build | 34 +++- > lib/net/meson.build | 58 +++ > lib/net/rte_net_crc.c | 16 > 22 files changed, 114 insertions(+), 333 deletions(-) > > -- Series-Acked-by: Konstantin Ananyev > 2.45.2
Re: [PATCH] version: 25.07-rc0
On Wed, Mar 26, 2025 at 10:11 AM Bruce Richardson
wrote:
> > diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
> > index fba46b920f..0cc4d12b0b 100644
> > --- a/.github/workflows/build.yml
> > +++ b/.github/workflows/build.yml
> > @@ -12,7 +12,7 @@ defaults:
> > env:
> >REF_GIT_BRANCH: main
> >REF_GIT_REPO: https://github.com/DPDK/dpdk
> > - REF_GIT_TAG: v24.11
> > + REF_GIT_TAG: v25.03
> >
> Do we not keep comparing against 24.11 for the whole of the 25-ABI cycle?
Comparing the new release against v24.11 would not detect breakage of
symbols introduced for the next ABI, during v25.03.
Like for example the changes in 52633e3a3fa8 ("net: add thread-safe CRC API").
--
David Marchand
Re: [PATCH] version: 25.07-rc0
On Wed, Mar 26, 2025 at 11:09:50AM +0100, David Marchand wrote:
> On Wed, Mar 26, 2025 at 10:11 AM Bruce Richardson
> wrote:
> > > diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
> > > index fba46b920f..0cc4d12b0b 100644
> > > --- a/.github/workflows/build.yml
> > > +++ b/.github/workflows/build.yml
> > > @@ -12,7 +12,7 @@ defaults:
> > > env:
> > >REF_GIT_BRANCH: main
> > >REF_GIT_REPO: https://github.com/DPDK/dpdk
> > > - REF_GIT_TAG: v24.11
> > > + REF_GIT_TAG: v25.03
> > >
> > Do we not keep comparing against 24.11 for the whole of the 25-ABI cycle?
>
> Comparing the new release against v24.11 would not detect breakage of
> symbols introduced for the next ABI, during v25.03.
> Like for example the changes in 52633e3a3fa8 ("net: add thread-safe CRC API").
>
Ah, yes, thanks for the explanation.
/Bruce
Re: [RFC v4 0/8] Symbol versioning and export rework
On Wed, Mar 26, 2025 at 02:54:32PM +0100, David Marchand wrote: > On Wed, Mar 26, 2025 2:37 PM Bruce Richardson > wrote: > > > - I am considering making rte_function_versioning.h a non exported > > > header (precisely, moving it to buildtools/ and maybe renaming it). > > > > > > This header contains macros not prefixed with RTE_. > > > Using it requires some build trick (see use_function_versioning). > > > And I don't see symbol versioning as a MUST infrastructure that DPDK > > > needs to provide to datapath applications. > > > > > > Yet technically, this change would be an API breakage if some > > > applications indeed relied on it. > > > > > > > Is it not needed for exporting if an exported library header had versioned > > symbols is in? > > That could be, though we don't expose such versionned symbols atm (and > I don't think we ever did in DPDK). > The only public symbol is always the symbol implemented for the latest > ABI for newly compiled applications. > Ok, that's good. Looking through code and code history I do indeed see that the use of function versioning in the past seems to have all been done through the C file. Therefore, no issues with making the header internal-only and not exported. /Bruce
Re: [RFC 2/2] eventdev: add default software vector adapter
On Wed, 26 Mar 2025 18:44:36 +0530
wrote:
> +
> +struct sw_vector_adapter_service_data {
> + uint32_t service_id;
> + RTE_ATOMIC(rte_mcslock_t *) lock;
> + RTE_TAILQ_HEAD(, sw_vector_adapter_data) adapter_list;
> +};
Do you really need mcslock here?
mcslock is for locks where there is large amount of contention and lots of
CPU's.
This doesn't seem like that.
[PATCH v2 1/1] net/af_packet: register supported param fanout mode
Although af_packet PMD supports setting the fanout algorithm, it is not
explicitly declared during registration. Add fanout parameter registration
to facilitate users in quickly obtaining supported fanout algorithms.
Additionally, add configuration parameter printing during initialization.
Fixes: d3bc77ab05a4 ("net/af_packet: allow changing fanout mode")
Cc: sta...@dpdk.org
Signed-off-by: Shaokai Zhang
Reviewed-by: Joey Xing
---
drivers/net/af_packet/rte_eth_af_packet.c | 8 +++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/drivers/net/af_packet/rte_eth_af_packet.c
b/drivers/net/af_packet/rte_eth_af_packet.c
index ab34285ed9..55171944cc 100644
--- a/drivers/net/af_packet/rte_eth_af_packet.c
+++ b/drivers/net/af_packet/rte_eth_af_packet.c
@@ -1172,6 +1172,11 @@ rte_eth_from_packet(struct rte_vdev_device *dev,
PMD_LOG(INFO, "%s:\tblock count %d", name, blockcount);
PMD_LOG(INFO, "%s:\tframe size %d", name, framesize);
PMD_LOG(INFO, "%s:\tframe count %d", name, framecount);
+ PMD_LOG(INFO, "%s:\tqdisc bypass %d", name, qdisc_bypass);
+ if (fanout_mode)
+ PMD_LOG(INFO, "%s:\tfanout mode %s", name, fanout_mode);
+ else
+ PMD_LOG(INFO, "%s:\tfanout mode %s", name, "default
PACKET_FANOUT_HASH");
if (rte_pmd_init_internals(dev, *sockfd, qpairs,
blocksize, blockcount,
@@ -1274,4 +1279,5 @@ RTE_PMD_REGISTER_PARAM_STRING(net_af_packet,
"blocksz= "
"framesz= "
"framecnt= "
- "qdisc_bypass=<0|1>");
+ "qdisc_bypass=<0|1> "
+ "fanout_mode=");
--
2.47.1.windows.2
RE: [EXTERNAL] [PATCH] app/crypto-perf: fix aad offset alignment
> -Original Message-
> From: Akhil Goyal
> Sent: Monday, 17 March 2025 12:23
> To: Shani Peretz ; dev@dpdk.org
> Cc: Suanming Mou ; sta...@dpdk.org; Brian
> Dooley ; Pablo de Lara
>
> Subject: RE: [EXTERNAL] [PATCH] app/crypto-perf: fix aad offset alignment
>
> External email: Use caution opening links or attachments
>
>
> Hi,
> > AAD offset in AES-GCM crypto test was calculated by adding 16-byte
> > alignment after the IV, which is only needed in AES-CCM.
>
> Agreed that CCM has a requirement for 16B alignment.
> But for GCM, does it break any protocol? Can we not align to byte boundary
> for performance?
> This is a performance application which mainly focus on getting the best
> throughput.
> Did you check if it is having some performance degradation?
>
> >
> > The patch correct the AAD offset calculation in AES-GCM algorithm tests.
> >
> > Fixes: 0b242422d385 ("app/crypto-perf: set AAD after the crypto
> > operation")
> > Cc: sta...@dpdk.org
> >
> > Signed-off-by: Shani Peretz
> > ---
> > app/test-crypto-perf/cperf_ops.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/app/test-crypto-perf/cperf_ops.c
> > b/app/test-crypto-perf/cperf_ops.c
> > index 6d5f510220..f9be51e17f 100644
> > --- a/app/test-crypto-perf/cperf_ops.c
> > +++ b/app/test-crypto-perf/cperf_ops.c
> > @@ -688,7 +688,9 @@ cperf_set_ops_aead(struct rte_crypto_op **ops,
> > uint16_t i;
> > /* AAD is placed after the IV */
> > uint16_t aad_offset = iv_offset +
> > - RTE_ALIGN_CEIL(test_vector->aead_iv.length, 16);
> > + ((options->aead_algo == RTE_CRYPTO_AEAD_AES_CCM)
> > ?
> > + RTE_ALIGN_CEIL(test_vector->aead_iv.length, 16) :
> > + test_vector->aead_iv.length);
> >
> > for (i = 0; i < nb_ops; i++) {
> > struct rte_crypto_sym_op *sym_op = ops[i]->sym;
> > --
> > 2.25.1
I checked the throughput test, and I haven't noticed any degradation compared
to upstream. I can share the results if needed.
Note that regardless of the performance it fixes several segmentation faults in
the test.
(The problem is that we allocate the crypto_op_private_size without alignment,
but we try to access as if it was 16 byte alignment)
RE: [PATCH v0 1/1] net/octeon_ep: fix buffer refill failure
> -Original Message-
> From: Vamsi Krishna
> Sent: Monday, March 24, 2025 4:39 PM
> To: dev@dpdk.org
> Cc: Jerin Jacob ; Vamsi Krishna Attunuru
> ; Vidya Sagar Velumuri
> Subject: [PATCH v0 1/1] net/octeon_ep: fix buffer refill failure
>
> From: Vidya Sagar Velumuri
>
> Fix the buffer refill for SDP output queue
> Fixes: ab09499e809c ("net/octeon_ep: add SSE Rx")
>
> Signed-off-by: Vidya Sagar Velumuri
> Acked-by: Vamsi Krishna Attunuru
Updated the git commit as follows and applied to dpdk-next-net-mrvl/for-main.
Thanks
commit 5b1a795452a09eb6a0b0a4d8eccd113c1c53f131 (HEAD -> for-main)
Author: Vidya Sagar Velumuri
Date: Mon Mar 24 16:38:50 2025 +0530
net/octeon_ep: fix buffer refill
Fix the buffer refill for SDP output queue
Fixes: ab09499e809c ("net/octeon_ep: add SSE Rx")
Cc: sta...@dpdk.org
Signed-off-by: Vidya Sagar Velumuri
Acked-by: Vamsi Attunuru
> ---
> drivers/net/octeon_ep/cnxk_ep_rx.h | 7 +++
> 1 file changed, 3 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/net/octeon_ep/cnxk_ep_rx.h
> b/drivers/net/octeon_ep/cnxk_ep_rx.h
> index 83677e1c6f..5db32c4448 100644
> --- a/drivers/net/octeon_ep/cnxk_ep_rx.h
> +++ b/drivers/net/octeon_ep/cnxk_ep_rx.h
> @@ -72,11 +72,10 @@ cnxk_ep_rx_refill(struct otx_ep_droq *droq)
> desc_refilled = count;
> count = droq->read_idx;
> rc = cnxk_ep_rx_refill_mbuf(droq, count);
> - if (unlikely(rc)) {
> + if (unlikely(rc))
> droq->stats.rx_alloc_failure++;
> - return;
> - }
> - desc_refilled += count;
> + else
> + desc_refilled += count;
> }
>
> /* Flush the droq descriptor data to memory to be sure
> --
> 2.34.1
Re: [PATCH v4] rust: support raw DPDK API
Hello Bruce, Hi Gregory, some high-level feedback: * I'd suggest for future revisions splitting this into two patches. The rust example should be separate from a patch adding the basic rust infrastructure. In progress. * For the example, I'd suggest that the helloworld rust example should correspond to the regular "C" helloworld example, i.e just init DPDK and print helloworld. If we want a basic packet handling example, it should be an l2fwd one - again to align with the "C" examples. The next patch update will not include that example. * However, in terms of the examples, if the idea is to provide higher level "rustic" APIs for DPDK, the examples should be written to use that. Any use of the low-level C APIs should be in a testing app only, rather than an example app. IMHO We want to encourage proper rust usage, not C-code just rewritten to use a different compiler. This project, in its current stage, is a bit of an experiment. I'm trying to figure out how much effort it will take to provide DPDK API for Rust application. Rust API is the final goal. Regards, Gregory
