From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id E955D45F5F; Sat, 28 Dec 2024 04:19:00 +0100 (CET) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id 5AA9C402F1; Sat, 28 Dec 2024 04:18:46 +0100 (CET) Received: from linux.microsoft.com (linux.microsoft.com [13.77.154.182]) by mails.dpdk.org (Postfix) with ESMTP id C806240280 for ; Sat, 28 Dec 2024 04:18:41 +0100 (CET) Received: by linux.microsoft.com (Postfix, from userid 1213) id D39CA206ADCA; Fri, 27 Dec 2024 19:18:40 -0800 (PST) DKIM-Filter: OpenDKIM Filter v2.11.0 linux.microsoft.com D39CA206ADCA DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.microsoft.com; s=default; t=1735355920; bh=MCQ5ogvmY6j6NwG65blbcXyCdkBIqdvR5Z5PI0c2dzo=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=fGIoEOSFIu5Oss8H6sDT0sVwyujn8nBTBt6bwZfbCCvybyuD0DWR2xHFycr58q/yD BAGWflVrCm5Y9hBPLTn3W61reRSgRQlO9d+54848IkHVyaNQKi8gs1gp8kGXd4nu7J gBZwKWf0ni+/BY6sfhz/dL0ufabyp8c/Gnl11+L8= From: Andre Muezerie To: andremue@linux.microsoft.com Cc: dev@dpdk.org, stephen@networkplumber.org Subject: [PATCH v3 3/3] drivers/net: add diagnostics macros to make code portable Date: Fri, 27 Dec 2024 19:18:30 -0800 Message-Id: <1735355910-1175-4-git-send-email-andremue@linux.microsoft.com> X-Mailer: git-send-email 1.8.3.1 In-Reply-To: <1735355910-1175-1-git-send-email-andremue@linux.microsoft.com> References: <1735263196-2809-1-git-send-email-andremue@linux.microsoft.com> <1735355910-1175-1-git-send-email-andremue@linux.microsoft.com> X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org It was a common pattern to have "GCC diagnostic ignored" pragmas sprinkled over the code and only activate these pragmas for certain compilers (gcc and clang). Clang supports GCC's pragma for compatibility with existing source code, so #pragma GCC diagnostic and #pragma clang diagnostic are synonyms for Clang (https://clang.llvm.org/docs/UsersManual.html). Now that effort is being made to make the code compatible with MSVC these expressions would become more complex. It makes sense to hide this complexity behind macros. This makes maintenance easier as these macros are defined in a single place. As a plus the code becomes more readable as well. Signed-off-by: Andre Muezerie --- drivers/net/axgbe/axgbe_rxtx.h | 9 ------ drivers/net/cpfl/cpfl_rxtx_vec_common.h | 4 --- drivers/net/dpaa2/dpaa2_rxtx.c | 16 ++-------- drivers/net/fm10k/fm10k_rxtx_vec.c | 19 +++++++++--- drivers/net/hns3/hns3_rxtx_vec_neon.h | 5 ++-- .../net/i40e/i40e_recycle_mbufs_vec_common.c | 2 -- drivers/net/i40e/i40e_rxtx_common_avx.h | 16 +++++++--- drivers/net/i40e/i40e_rxtx_vec_altivec.c | 11 +++++-- drivers/net/i40e/i40e_rxtx_vec_avx2.c | 16 +++++++--- drivers/net/i40e/i40e_rxtx_vec_avx512.c | 16 +++++++--- drivers/net/i40e/i40e_rxtx_vec_common.h | 4 --- drivers/net/i40e/i40e_rxtx_vec_neon.c | 9 ++++-- drivers/net/i40e/i40e_rxtx_vec_sse.c | 22 +++++++++++--- drivers/net/iavf/iavf_rxtx_vec_avx2.c | 21 +++++++++---- drivers/net/iavf/iavf_rxtx_vec_avx512.c | 27 +++++++++++++---- drivers/net/iavf/iavf_rxtx_vec_common.h | 10 ++++--- drivers/net/iavf/iavf_rxtx_vec_sse.c | 30 +++++++++++++++---- drivers/net/ice/ice_rxtx_common_avx.h | 10 ++++--- drivers/net/ice/ice_rxtx_vec_avx2.c | 16 +++++++--- drivers/net/ice/ice_rxtx_vec_avx512.c | 16 +++++++--- drivers/net/ice/ice_rxtx_vec_common.h | 4 --- drivers/net/ice/ice_rxtx_vec_sse.c | 22 +++++++++++--- drivers/net/idpf/idpf_rxtx_vec_common.h | 4 --- .../ixgbe/ixgbe_recycle_mbufs_vec_common.c | 2 -- drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c | 2 -- drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c | 19 +++++++++--- drivers/net/mlx5/mlx5_flow.c | 6 ++-- drivers/net/mlx5/mlx5_rxtx_vec_altivec.h | 5 ---- drivers/net/mlx5/mlx5_rxtx_vec_neon.h | 2 -- drivers/net/mlx5/mlx5_rxtx_vec_sse.h | 26 +++++++++++----- drivers/net/tap/tap_flow.c | 6 ++-- drivers/net/virtio/virtio_rxtx_simple.c | 4 --- 32 files changed, 247 insertions(+), 134 deletions(-) diff --git a/drivers/net/axgbe/axgbe_rxtx.h b/drivers/net/axgbe/axgbe_rxtx.h index a326ba9ac8..f5f74a0a39 100644 --- a/drivers/net/axgbe/axgbe_rxtx.h +++ b/drivers/net/axgbe/axgbe_rxtx.h @@ -6,15 +6,6 @@ #ifndef _AXGBE_RXTX_H_ #define _AXGBE_RXTX_H_ -/* to suppress gcc warnings related to descriptor casting*/ -#ifdef RTE_TOOLCHAIN_GCC -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - -#ifdef RTE_TOOLCHAIN_CLANG -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - /* Descriptor related defines */ #define AXGBE_MAX_RING_DESC 4096 /*should be power of 2*/ #define AXGBE_TX_DESC_MIN_FREE (AXGBE_MAX_RING_DESC >> 3) diff --git a/drivers/net/cpfl/cpfl_rxtx_vec_common.h b/drivers/net/cpfl/cpfl_rxtx_vec_common.h index 479e1ddcb9..5b98f86932 100644 --- a/drivers/net/cpfl/cpfl_rxtx_vec_common.h +++ b/drivers/net/cpfl/cpfl_rxtx_vec_common.h @@ -11,10 +11,6 @@ #include "cpfl_ethdev.h" #include "cpfl_rxtx.h" -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - #define CPFL_SCALAR_PATH 0 #define CPFL_VECTOR_PATH 1 #define CPFL_RX_NO_VECTOR_FLAGS ( \ diff --git a/drivers/net/dpaa2/dpaa2_rxtx.c b/drivers/net/dpaa2/dpaa2_rxtx.c index e3b6c7e460..21c306fd94 100644 --- a/drivers/net/dpaa2/dpaa2_rxtx.c +++ b/drivers/net/dpaa2/dpaa2_rxtx.c @@ -1962,14 +1962,6 @@ dpaa2_dev_tx_ordered(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) return num_tx; } -#if defined(RTE_TOOLCHAIN_GCC) -#pragma GCC diagnostic push -#pragma GCC diagnostic ignored "-Wcast-qual" -#elif defined(RTE_TOOLCHAIN_CLANG) -#pragma clang diagnostic push -#pragma clang diagnostic ignored "-Wcast-qual" -#endif - /* This function loopbacks all the received packets.*/ uint16_t dpaa2_dev_loopback_rx(void *queue, @@ -2083,7 +2075,10 @@ dpaa2_dev_loopback_rx(void *queue, if (unlikely((status & QBMAN_DQ_STAT_VALIDFRAME) == 0)) continue; } +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual fd[num_rx] = (struct qbman_fd *)qbman_result_DQ_fd(dq_storage); +__rte_diagnostic_pop dq_storage++; num_rx++; @@ -2118,8 +2113,3 @@ dpaa2_dev_loopback_rx(void *queue, return 0; } -#if defined(RTE_TOOLCHAIN_GCC) -#pragma GCC diagnostic pop -#elif defined(RTE_TOOLCHAIN_CLANG) -#pragma clang diagnostic pop -#endif diff --git a/drivers/net/fm10k/fm10k_rxtx_vec.c b/drivers/net/fm10k/fm10k_rxtx_vec.c index 68acaca75b..3d534a91ac 100644 --- a/drivers/net/fm10k/fm10k_rxtx_vec.c +++ b/drivers/net/fm10k/fm10k_rxtx_vec.c @@ -11,10 +11,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static void fm10k_reset_tx_queue(struct fm10k_tx_queue *txq); @@ -270,8 +266,11 @@ fm10k_rxq_rearm(struct fm10k_rx_queue *rxq) /* Clean up all the HW/SW ring content */ for (i = 0; i < RTE_FM10K_RXQ_REARM_THRESH; i++) { mb_alloc[i] = &rxq->fake_mbuf; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp[i].q, dma_addr0); +__rte_diagnostic_pop } rte_eth_devices[rxq->port_id].data->rx_mbuf_alloc_failed += @@ -316,8 +315,11 @@ fm10k_rxq_rearm(struct fm10k_rx_queue *rxq) dma_addr1 = _mm_and_si128(dma_addr1, hba_msk); /* flush desc with pa dma_addr */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp++->q, dma_addr0); _mm_store_si128((__m128i *)&rxdp++->q, dma_addr1); +__rte_diagnostic_pop /* enforce 512B alignment on default Rx virtual addresses */ mb0->data_off = (uint16_t)(RTE_PTR_ALIGN((char *)mb0->buf_addr @@ -465,7 +467,10 @@ fm10k_recv_raw_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts, /* Read desc statuses backwards to avoid race condition */ /* A.1 load desc[3] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs0[3] = _mm_loadu_si128((__m128i *)(rxdp + 3)); +__rte_diagnostic_pop rte_compiler_barrier(); /* B.2 copy 2 64 bit or 4 32 bit mbuf point into rx_pkts */ @@ -477,11 +482,14 @@ fm10k_recv_raw_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts, #endif /* A.1 load desc[2-0] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs0[2] = _mm_loadu_si128((__m128i *)(rxdp + 2)); rte_compiler_barrier(); descs0[1] = _mm_loadu_si128((__m128i *)(rxdp + 1)); rte_compiler_barrier(); descs0[0] = _mm_loadu_si128((__m128i *)(rxdp)); +__rte_diagnostic_pop #if defined(RTE_ARCH_X86_64) /* B.2 copy 2 mbuf point into rx_pkts */ @@ -736,7 +744,10 @@ vtx1(volatile struct fm10k_tx_desc *txdp, __m128i descriptor = _mm_set_epi64x(flags << 56 | (uint64_t)pkt->vlan_tci << 16 | (uint64_t)pkt->data_len, MBUF_DMA_ADDR(pkt)); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)txdp, descriptor); +__rte_diagnostic_pop } static inline void diff --git a/drivers/net/hns3/hns3_rxtx_vec_neon.h b/drivers/net/hns3/hns3_rxtx_vec_neon.h index bbb5478015..99f080f3e8 100644 --- a/drivers/net/hns3/hns3_rxtx_vec_neon.h +++ b/drivers/net/hns3/hns3_rxtx_vec_neon.h @@ -9,8 +9,6 @@ #include -#pragma GCC diagnostic ignored "-Wcast-qual" - static inline void hns3_vec_tx(volatile struct hns3_desc *desc, struct rte_mbuf *pkt) { @@ -22,8 +20,11 @@ hns3_vec_tx(volatile struct hns3_desc *desc, struct rte_mbuf *pkt) 0, ((uint64_t)HNS3_TXD_DEFAULT_VLD_FE_BDTYPE) << HNS3_UINT32_BIT }; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual vst1q_u64((uint64_t *)&desc->addr, val1); vst1q_u64((uint64_t *)&desc->tx.outer_vlan_tag, val2); +__rte_diagnostic_pop } static uint16_t diff --git a/drivers/net/i40e/i40e_recycle_mbufs_vec_common.c b/drivers/net/i40e/i40e_recycle_mbufs_vec_common.c index 14424c9921..6eafe51e3d 100644 --- a/drivers/net/i40e/i40e_recycle_mbufs_vec_common.c +++ b/drivers/net/i40e/i40e_recycle_mbufs_vec_common.c @@ -10,8 +10,6 @@ #include "i40e_ethdev.h" #include "i40e_rxtx.h" -#pragma GCC diagnostic ignored "-Wcast-qual" - void i40e_recycle_rx_descriptors_refill_vec(void *rx_queue, uint16_t nb_mbufs) { diff --git a/drivers/net/i40e/i40e_rxtx_common_avx.h b/drivers/net/i40e/i40e_rxtx_common_avx.h index 85958d6c81..bcabacf689 100644 --- a/drivers/net/i40e/i40e_rxtx_common_avx.h +++ b/drivers/net/i40e/i40e_rxtx_common_avx.h @@ -11,10 +11,6 @@ #include "i40e_ethdev.h" #include "i40e_rxtx.h" -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - #ifdef __AVX2__ static __rte_always_inline void i40e_rxq_rearm_common(struct i40e_rx_queue *rxq, __rte_unused bool avx512) @@ -36,8 +32,11 @@ i40e_rxq_rearm_common(struct i40e_rx_queue *rxq, __rte_unused bool avx512) dma_addr0 = _mm_setzero_si128(); for (i = 0; i < RTE_I40E_DESCS_PER_LOOP; i++) { rxep[i].mbuf = &rxq->fake_mbuf; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp[i].read, dma_addr0); +__rte_diagnostic_pop } } rte_eth_devices[rxq->port_id].data->rx_mbuf_alloc_failed += @@ -72,8 +71,11 @@ i40e_rxq_rearm_common(struct i40e_rx_queue *rxq, __rte_unused bool avx512) dma_addr1 = _mm_add_epi64(dma_addr1, hdr_room); /* flush desc with pa dma_addr */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp++->read, dma_addr0); _mm_store_si128((__m128i *)&rxdp++->read, dma_addr1); +__rte_diagnostic_pop } #else #ifdef __AVX512VL__ @@ -144,8 +146,11 @@ i40e_rxq_rearm_common(struct i40e_rx_queue *rxq, __rte_unused bool avx512) dma_addr4_7 = _mm512_add_epi64(dma_addr4_7, hdr_room); /* flush desc with pa dma_addr */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm512_store_si512((__m512i *)&rxdp->read, dma_addr0_3); _mm512_store_si512((__m512i *)&(rxdp + 4)->read, dma_addr4_7); +__rte_diagnostic_pop } } else #endif /* __AVX512VL__*/ @@ -190,8 +195,11 @@ i40e_rxq_rearm_common(struct i40e_rx_queue *rxq, __rte_unused bool avx512) dma_addr2_3 = _mm256_add_epi64(dma_addr2_3, hdr_room); /* flush desc with pa dma_addr */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm256_store_si256((__m256i *)&rxdp->read, dma_addr0_1); _mm256_store_si256((__m256i *)&(rxdp + 2)->read, dma_addr2_3); +__rte_diagnostic_pop } } diff --git a/drivers/net/i40e/i40e_rxtx_vec_altivec.c b/drivers/net/i40e/i40e_rxtx_vec_altivec.c index b6b0d38ec1..6771cc7928 100644 --- a/drivers/net/i40e/i40e_rxtx_vec_altivec.c +++ b/drivers/net/i40e/i40e_rxtx_vec_altivec.c @@ -15,8 +15,6 @@ #include -#pragma GCC diagnostic ignored "-Wcast-qual" - static inline void i40e_rxq_rearm(struct i40e_rx_queue *rxq) { @@ -286,7 +284,10 @@ _recv_raw_pkts_vec(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts, mbp1 = *(__vector unsigned long *)&sw_ring[pos]; /* Read desc statuses backwards to avoid race condition */ /* A.1 load desc[3] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[3] = *(__vector unsigned long *)(rxdp + 3); +__rte_diagnostic_pop rte_compiler_barrier(); /* B.2 copy 2 mbuf point into rx_pkts */ @@ -296,11 +297,14 @@ _recv_raw_pkts_vec(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts, mbp2 = *(__vector unsigned long *)&sw_ring[pos + 2]; /* A.1 load desc[2-0] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[2] = *(__vector unsigned long *)(rxdp + 2); rte_compiler_barrier(); descs[1] = *(__vector unsigned long *)(rxdp + 1); rte_compiler_barrier(); descs[0] = *(__vector unsigned long *)(rxdp); +__rte_diagnostic_pop /* B.2 copy 2 mbuf point into rx_pkts */ *(__vector unsigned long *)&rx_pkts[pos + 2] = mbp2; @@ -534,7 +538,10 @@ vtx1(volatile struct i40e_tx_desc *txdp, __vector unsigned long descriptor = (__vector unsigned long){ pkt->buf_iova + pkt->data_off, high_qw}; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual *(__vector unsigned long *)txdp = descriptor; +__rte_diagnostic_pop } static inline void diff --git a/drivers/net/i40e/i40e_rxtx_vec_avx2.c b/drivers/net/i40e/i40e_rxtx_vec_avx2.c index 19cf0ac718..90a1d4661a 100644 --- a/drivers/net/i40e/i40e_rxtx_vec_avx2.c +++ b/drivers/net/i40e/i40e_rxtx_vec_avx2.c @@ -15,10 +15,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static __rte_always_inline void i40e_rxq_rearm(struct i40e_rx_queue *rxq) { @@ -39,8 +35,11 @@ desc_fdir_processing_32b(volatile union i40e_rx_desc *rxdp, const uint32_t desc_idx) { /* 32B desc path: load rxdp.wb.qword2 for EXT_STATUS and FLEXBH_STAT */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual __m128i *rxdp_desc_0 = (void *)(&rxdp[desc_idx + 0].wb.qword2); __m128i *rxdp_desc_1 = (void *)(&rxdp[desc_idx + 1].wb.qword2); +__rte_diagnostic_pop const __m128i desc_qw2_0 = _mm_load_si128(rxdp_desc_0); const __m128i desc_qw2_1 = _mm_load_si128(rxdp_desc_1); @@ -276,6 +275,8 @@ _recv_raw_pkts_vec_avx2(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts, _mm256_loadu_si256((void *)&sw_ring[i + 4])); #endif +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc7 = _mm_load_si128((void *)(rxdp + 7)); rte_compiler_barrier(); const __m128i raw_desc6 = _mm_load_si128((void *)(rxdp + 6)); @@ -291,6 +292,7 @@ _recv_raw_pkts_vec_avx2(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts, const __m128i raw_desc1 = _mm_load_si128((void *)(rxdp + 1)); rte_compiler_barrier(); const __m128i raw_desc0 = _mm_load_si128((void *)(rxdp + 0)); +__rte_diagnostic_pop const __m256i raw_desc6_7 = _mm256_inserti128_si256( _mm256_castsi128_si256(raw_desc6), raw_desc7, 1); @@ -695,7 +697,10 @@ vtx1(volatile struct i40e_tx_desc *txdp, __m128i descriptor = _mm_set_epi64x(high_qw, pkt->buf_iova + pkt->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)txdp, descriptor); +__rte_diagnostic_pop } static inline void @@ -728,8 +733,11 @@ vtx(volatile struct i40e_tx_desc *txdp, __m256i desc0_1 = _mm256_set_epi64x( hi_qw1, pkt[1]->buf_iova + pkt[1]->data_off, hi_qw0, pkt[0]->buf_iova + pkt[0]->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm256_store_si256((void *)(txdp + 2), desc2_3); _mm256_store_si256((void *)txdp, desc0_1); +__rte_diagnostic_pop } /* do any last ones */ diff --git a/drivers/net/i40e/i40e_rxtx_vec_avx512.c b/drivers/net/i40e/i40e_rxtx_vec_avx512.c index 3b2750221b..b77989074f 100644 --- a/drivers/net/i40e/i40e_rxtx_vec_avx512.c +++ b/drivers/net/i40e/i40e_rxtx_vec_avx512.c @@ -15,10 +15,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - #define RTE_I40E_DESCS_PER_LOOP_AVX 8 static __rte_always_inline void @@ -41,8 +37,11 @@ desc_fdir_processing_32b(volatile union i40e_rx_desc *rxdp, const uint32_t desc_idx) { /* 32B desc path: load rxdp.wb.qword2 for EXT_STATUS and FLEXBH_STAT */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual __m128i *rxdp_desc_0 = (void *)(&rxdp[desc_idx + 0].wb.qword2); __m128i *rxdp_desc_1 = (void *)(&rxdp[desc_idx + 1].wb.qword2); +__rte_diagnostic_pop const __m128i desc_qw2_0 = _mm_load_si128(rxdp_desc_0); const __m128i desc_qw2_1 = _mm_load_si128(rxdp_desc_1); @@ -263,6 +262,8 @@ _recv_raw_pkts_vec_avx512(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts, __m256i raw_desc0_1, raw_desc2_3, raw_desc4_5, raw_desc6_7; /* load in descriptors, in reverse order */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc7 = _mm_load_si128((void *)(rxdp + 7)); rte_compiler_barrier(); @@ -286,6 +287,7 @@ _recv_raw_pkts_vec_avx512(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts, rte_compiler_barrier(); const __m128i raw_desc0 = _mm_load_si128((void *)(rxdp + 0)); +__rte_diagnostic_pop raw_desc6_7 = _mm256_inserti128_si256 @@ -875,7 +877,10 @@ vtx1(volatile struct i40e_tx_desc *txdp, struct rte_mbuf *pkt, uint64_t flags) __m128i descriptor = _mm_set_epi64x(high_qw, pkt->buf_iova + pkt->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)txdp, descriptor); +__rte_diagnostic_pop } static inline void @@ -909,7 +914,10 @@ vtx(volatile struct i40e_tx_desc *txdp, hi_qw2, pkt[2]->buf_iova + pkt[2]->data_off, hi_qw1, pkt[1]->buf_iova + pkt[1]->data_off, hi_qw0, pkt[0]->buf_iova + pkt[0]->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm512_storeu_si512((void *)txdp, desc0_3); +__rte_diagnostic_pop } /* do any last ones */ diff --git a/drivers/net/i40e/i40e_rxtx_vec_common.h b/drivers/net/i40e/i40e_rxtx_vec_common.h index 8b745630e4..ec59a68f9d 100644 --- a/drivers/net/i40e/i40e_rxtx_vec_common.h +++ b/drivers/net/i40e/i40e_rxtx_vec_common.h @@ -11,10 +11,6 @@ #include "i40e_ethdev.h" #include "i40e_rxtx.h" -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static inline uint16_t reassemble_packets(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_bufs, uint16_t nb_bufs, uint8_t *split_flags) diff --git a/drivers/net/i40e/i40e_rxtx_vec_neon.c b/drivers/net/i40e/i40e_rxtx_vec_neon.c index e1c5c7041b..f35f3e1e20 100644 --- a/drivers/net/i40e/i40e_rxtx_vec_neon.c +++ b/drivers/net/i40e/i40e_rxtx_vec_neon.c @@ -16,9 +16,6 @@ #include "i40e_rxtx.h" #include "i40e_rxtx_vec_common.h" - -#pragma GCC diagnostic ignored "-Wcast-qual" - static inline void i40e_rxq_rearm(struct i40e_rx_queue *rxq) { @@ -421,6 +418,8 @@ _recv_raw_pkts_vec(struct i40e_rx_queue *__rte_restrict rxq, int32x4_t len_shl = {0, 0, 0, PKTLEN_SHIFT}; /* A.1 load desc[3-0] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[3] = vld1q_u64((uint64_t *)(rxdp + 3)); descs[2] = vld1q_u64((uint64_t *)(rxdp + 2)); descs[1] = vld1q_u64((uint64_t *)(rxdp + 1)); @@ -433,6 +432,7 @@ _recv_raw_pkts_vec(struct i40e_rx_queue *__rte_restrict rxq, descs[2] = vld1q_lane_u64((uint64_t *)(rxdp + 2), descs[2], 0); descs[1] = vld1q_lane_u64((uint64_t *)(rxdp + 1), descs[1], 0); descs[0] = vld1q_lane_u64((uint64_t *)(rxdp), descs[0], 0); +__rte_diagnostic_pop /* B.1 load 4 mbuf point */ mbp1 = vld1q_u64((uint64_t *)&sw_ring[pos]); @@ -662,7 +662,10 @@ vtx1(volatile struct i40e_tx_desc *txdp, ((uint64_t)pkt->data_len << I40E_TXD_QW1_TX_BUF_SZ_SHIFT)); uint64x2_t descriptor = {pkt->buf_iova + pkt->data_off, high_qw}; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual vst1q_u64((uint64_t *)txdp, descriptor); +__rte_diagnostic_pop } static inline void diff --git a/drivers/net/i40e/i40e_rxtx_vec_sse.c b/drivers/net/i40e/i40e_rxtx_vec_sse.c index ad560d2b6b..8a5537bcf2 100644 --- a/drivers/net/i40e/i40e_rxtx_vec_sse.c +++ b/drivers/net/i40e/i40e_rxtx_vec_sse.c @@ -14,10 +14,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static inline void i40e_rxq_rearm(struct i40e_rx_queue *rxq) { @@ -41,8 +37,11 @@ i40e_rxq_rearm(struct i40e_rx_queue *rxq) dma_addr0 = _mm_setzero_si128(); for (i = 0; i < RTE_I40E_DESCS_PER_LOOP; i++) { rxep[i].mbuf = &rxq->fake_mbuf; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp[i].read, dma_addr0); +__rte_diagnostic_pop } } rte_eth_devices[rxq->port_id].data->rx_mbuf_alloc_failed += @@ -72,8 +71,11 @@ i40e_rxq_rearm(struct i40e_rx_queue *rxq) dma_addr1 = _mm_add_epi64(dma_addr1, hdr_room); /* flush desc with pa dma_addr */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp++->read, dma_addr0); _mm_store_si128((__m128i *)&rxdp++->read, dma_addr1); +__rte_diagnostic_pop } rxq->rxrearm_start += RTE_I40E_RXQ_REARM_THRESH; @@ -97,10 +99,13 @@ descs_to_fdir_32b(volatile union i40e_rx_desc *rxdp, struct rte_mbuf **rx_pkt) { /* 32B descriptors: Load 2nd half of descriptors for FDIR ID data */ __m128i desc0_qw23, desc1_qw23, desc2_qw23, desc3_qw23; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual desc0_qw23 = _mm_loadu_si128((__m128i *)&(rxdp + 0)->wb.qword2); desc1_qw23 = _mm_loadu_si128((__m128i *)&(rxdp + 1)->wb.qword2); desc2_qw23 = _mm_loadu_si128((__m128i *)&(rxdp + 2)->wb.qword2); desc3_qw23 = _mm_loadu_si128((__m128i *)&(rxdp + 3)->wb.qword2); +__rte_diagnostic_pop /* FDIR ID data: move last u32 of each desc to 4 u32 lanes */ __m128i v_unpack_01, v_unpack_23; @@ -462,7 +467,10 @@ _recv_raw_pkts_vec(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts, mbp1 = _mm_loadu_si128((__m128i *)&sw_ring[pos]); /* Read desc statuses backwards to avoid race condition */ /* A.1 load desc[3] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[3] = _mm_loadu_si128((__m128i *)(rxdp + 3)); +__rte_diagnostic_pop rte_compiler_barrier(); /* B.2 copy 2 64 bit or 4 32 bit mbuf point into rx_pkts */ @@ -474,11 +482,14 @@ _recv_raw_pkts_vec(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts, #endif /* A.1 load desc[2-0] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[2] = _mm_loadu_si128((__m128i *)(rxdp + 2)); rte_compiler_barrier(); descs[1] = _mm_loadu_si128((__m128i *)(rxdp + 1)); rte_compiler_barrier(); descs[0] = _mm_loadu_si128((__m128i *)(rxdp)); +__rte_diagnostic_pop #if defined(RTE_ARCH_X86_64) /* B.2 copy 2 mbuf point into rx_pkts */ @@ -681,7 +692,10 @@ vtx1(volatile struct i40e_tx_desc *txdp, __m128i descriptor = _mm_set_epi64x(high_qw, pkt->buf_iova + pkt->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)txdp, descriptor); +__rte_diagnostic_pop } static inline void diff --git a/drivers/net/iavf/iavf_rxtx_vec_avx2.c b/drivers/net/iavf/iavf_rxtx_vec_avx2.c index 49d41af953..f5503d9dae 100644 --- a/drivers/net/iavf/iavf_rxtx_vec_avx2.c +++ b/drivers/net/iavf/iavf_rxtx_vec_avx2.c @@ -6,10 +6,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static __rte_always_inline void iavf_rxq_rearm(struct iavf_rx_queue *rxq) { @@ -193,6 +189,8 @@ _iavf_recv_raw_pkts_vec_avx2(struct iavf_rx_queue *rxq, _mm256_loadu_si256((void *)&sw_ring[i + 4])); #endif +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc7 = _mm_load_si128((void *)(rxdp + 7)); rte_compiler_barrier(); const __m128i raw_desc6 = _mm_load_si128((void *)(rxdp + 6)); @@ -208,6 +206,7 @@ _iavf_recv_raw_pkts_vec_avx2(struct iavf_rx_queue *rxq, const __m128i raw_desc1 = _mm_load_si128((void *)(rxdp + 1)); rte_compiler_barrier(); const __m128i raw_desc0 = _mm_load_si128((void *)(rxdp + 0)); +__rte_diagnostic_pop const __m256i raw_desc6_7 = _mm256_inserti128_si256(_mm256_castsi128_si256(raw_desc6), raw_desc7, 1); @@ -509,7 +508,7 @@ _iavf_recv_raw_pkts_vec_avx2_flex_rxd(struct iavf_rx_queue *rxq, 0, rxq->mbuf_initializer); struct rte_mbuf **sw_ring = &rxq->sw_ring[rxq->rx_tail]; volatile union iavf_rx_flex_desc *rxdp = - (union iavf_rx_flex_desc *)rxq->rx_ring + rxq->rx_tail; + (volatile union iavf_rx_flex_desc *)rxq->rx_ring + rxq->rx_tail; rte_prefetch0(rxdp); @@ -742,6 +741,8 @@ _iavf_recv_raw_pkts_vec_avx2_flex_rxd(struct iavf_rx_queue *rxq, __m256i raw_desc0_1, raw_desc2_3, raw_desc4_5, raw_desc6_7; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc7 = _mm_load_si128((void *)(rxdp + 7)); rte_compiler_barrier(); @@ -765,6 +766,7 @@ _iavf_recv_raw_pkts_vec_avx2_flex_rxd(struct iavf_rx_queue *rxq, rte_compiler_barrier(); const __m128i raw_desc0 = _mm_load_si128((void *)(rxdp + 0)); +__rte_diagnostic_pop raw_desc6_7 = _mm256_inserti128_si256 @@ -959,6 +961,8 @@ _iavf_recv_raw_pkts_vec_avx2_flex_rxd(struct iavf_rx_queue *rxq, offloads & RTE_ETH_RX_OFFLOAD_TIMESTAMP || rxq->rx_flags & IAVF_RX_FLAGS_VLAN_TAG_LOC_L2TAG2_2) { /* load bottom half of every 32B desc */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc_bh7 = _mm_load_si128 ((void *)(&rxdp[7].wb.status_error1)); @@ -990,6 +994,7 @@ _iavf_recv_raw_pkts_vec_avx2_flex_rxd(struct iavf_rx_queue *rxq, const __m128i raw_desc_bh0 = _mm_load_si128 ((void *)(&rxdp[0].wb.status_error1)); +__rte_diagnostic_pop __m256i raw_desc_bh6_7 = _mm256_inserti128_si256 @@ -1664,7 +1669,10 @@ iavf_vtx1(volatile struct iavf_tx_desc *txdp, __m128i descriptor = _mm_set_epi64x(high_qw, pkt->buf_iova + pkt->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)txdp, descriptor); +__rte_diagnostic_pop } static __rte_always_inline void @@ -1719,8 +1727,11 @@ iavf_vtx(volatile struct iavf_tx_desc *txdp, pkt[1]->buf_iova + pkt[1]->data_off, hi_qw0, pkt[0]->buf_iova + pkt[0]->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm256_store_si256((void *)(txdp + 2), desc2_3); _mm256_store_si256((void *)txdp, desc0_1); +__rte_diagnostic_pop } /* do any last ones */ diff --git a/drivers/net/iavf/iavf_rxtx_vec_avx512.c b/drivers/net/iavf/iavf_rxtx_vec_avx512.c index d6a861bf80..04894debc8 100644 --- a/drivers/net/iavf/iavf_rxtx_vec_avx512.c +++ b/drivers/net/iavf/iavf_rxtx_vec_avx512.c @@ -6,10 +6,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - #define IAVF_DESCS_PER_LOOP_AVX 8 #define PKTLEN_SHIFT 10 @@ -164,6 +160,8 @@ _iavf_recv_raw_pkts_vec_avx512(struct iavf_rx_queue *rxq, #endif __m512i raw_desc0_3, raw_desc4_7; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc7 = _mm_load_si128((void *)(rxdp + 7)); rte_compiler_barrier(); @@ -187,6 +185,7 @@ _iavf_recv_raw_pkts_vec_avx512(struct iavf_rx_queue *rxq, rte_compiler_barrier(); const __m128i raw_desc0 = _mm_load_si128((void *)(rxdp + 0)); +__rte_diagnostic_pop raw_desc4_7 = _mm512_broadcast_i32x4(raw_desc4); raw_desc4_7 = _mm512_inserti32x4(raw_desc4_7, raw_desc5, 1); @@ -600,7 +599,7 @@ _iavf_recv_raw_pkts_vec_avx512_flex_rxd(struct iavf_rx_queue *rxq, rxq->mbuf_initializer); struct rte_mbuf **sw_ring = &rxq->sw_ring[rxq->rx_tail]; volatile union iavf_rx_flex_desc *rxdp = - (union iavf_rx_flex_desc *)rxq->rx_ring + rxq->rx_tail; + (volatile union iavf_rx_flex_desc *)rxq->rx_ring + rxq->rx_tail; rte_prefetch0(rxdp); @@ -733,6 +732,8 @@ _iavf_recv_raw_pkts_vec_avx512_flex_rxd(struct iavf_rx_queue *rxq, __m512i raw_desc0_3, raw_desc4_7; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc7 = _mm_load_si128((void *)(rxdp + 7)); rte_compiler_barrier(); @@ -756,6 +757,7 @@ _iavf_recv_raw_pkts_vec_avx512_flex_rxd(struct iavf_rx_queue *rxq, rte_compiler_barrier(); const __m128i raw_desc0 = _mm_load_si128((void *)(rxdp + 0)); +__rte_diagnostic_pop raw_desc4_7 = _mm512_broadcast_i32x4(raw_desc4); raw_desc4_7 = _mm512_inserti32x4(raw_desc4_7, raw_desc5, 1); @@ -1111,6 +1113,8 @@ _iavf_recv_raw_pkts_vec_avx512_flex_rxd(struct iavf_rx_queue *rxq, offloads & RTE_ETH_RX_OFFLOAD_TIMESTAMP || rxq->rx_flags & IAVF_RX_FLAGS_VLAN_TAG_LOC_L2TAG2_2) { /* load bottom half of every 32B desc */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc_bh7 = _mm_load_si128 ((void *)(&rxdp[7].wb.status_error1)); @@ -1142,6 +1146,7 @@ _iavf_recv_raw_pkts_vec_avx512_flex_rxd(struct iavf_rx_queue *rxq, const __m128i raw_desc_bh0 = _mm_load_si128 ((void *)(&rxdp[0].wb.status_error1)); +__rte_diagnostic_pop __m256i raw_desc_bh6_7 = _mm256_inserti128_si256 @@ -1983,7 +1988,10 @@ iavf_vtx1(volatile struct iavf_tx_desc *txdp, __m128i descriptor = _mm_set_epi64x(high_qw, pkt->buf_iova + pkt->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_storeu_si128((__m128i *)txdp, descriptor); +__rte_diagnostic_pop } #define IAVF_TX_LEN_MASK 0xAA @@ -2037,7 +2045,10 @@ iavf_vtx(volatile struct iavf_tx_desc *txdp, pkt[1]->buf_iova + pkt[1]->data_off, hi_qw0, pkt[0]->buf_iova + pkt[0]->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm512_storeu_si512((void *)txdp, desc0_3); +__rte_diagnostic_pop } /* do any last ones */ @@ -2225,7 +2236,10 @@ ctx_vtx1(volatile struct iavf_tx_desc *txdp, struct rte_mbuf *pkt, __m256i ctx_data_desc = _mm256_set_epi64x(high_data_qw, pkt->buf_iova + pkt->data_off, high_ctx_qw, low_ctx_qw); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm256_storeu_si256((__m256i *)txdp, ctx_data_desc); +__rte_diagnostic_pop } static __rte_always_inline void @@ -2300,7 +2314,10 @@ ctx_vtx(volatile struct iavf_tx_desc *txdp, hi_ctx_qw1, low_ctx_qw1, hi_data_qw0, pkt[0]->buf_iova + pkt[0]->data_off, hi_ctx_qw0, low_ctx_qw0); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm512_storeu_si512((void *)txdp, desc0_3); +__rte_diagnostic_pop } if (nb_pkts) diff --git a/drivers/net/iavf/iavf_rxtx_vec_common.h b/drivers/net/iavf/iavf_rxtx_vec_common.h index 5c5220048d..690e0749e4 100644 --- a/drivers/net/iavf/iavf_rxtx_vec_common.h +++ b/drivers/net/iavf/iavf_rxtx_vec_common.h @@ -11,10 +11,6 @@ #include "iavf.h" #include "iavf_rxtx.h" -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static __rte_always_inline uint16_t reassemble_packets(struct iavf_rx_queue *rxq, struct rte_mbuf **rx_bufs, uint16_t nb_bufs, uint8_t *split_flags) @@ -422,8 +418,11 @@ iavf_rxq_rearm_common(struct iavf_rx_queue *rxq, __rte_unused bool avx512) dma_addr0 = _mm_setzero_si128(); for (i = 0; i < IAVF_VPMD_DESCS_PER_LOOP; i++) { rxp[i] = &rxq->fake_mbuf; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp[i].read, dma_addr0); +__rte_diagnostic_pop } } rte_eth_devices[rxq->port_id].data->rx_mbuf_alloc_failed += @@ -458,8 +457,11 @@ iavf_rxq_rearm_common(struct iavf_rx_queue *rxq, __rte_unused bool avx512) dma_addr1 = _mm_add_epi64(dma_addr1, hdr_room); /* flush desc with pa dma_addr */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp++->read, dma_addr0); _mm_store_si128((__m128i *)&rxdp++->read, dma_addr1); +__rte_diagnostic_pop } #else #ifdef CC_AVX512_SUPPORT diff --git a/drivers/net/iavf/iavf_rxtx_vec_sse.c b/drivers/net/iavf/iavf_rxtx_vec_sse.c index 0db6fa8bd4..373df0c935 100644 --- a/drivers/net/iavf/iavf_rxtx_vec_sse.c +++ b/drivers/net/iavf/iavf_rxtx_vec_sse.c @@ -12,10 +12,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static inline void iavf_rxq_rearm(struct iavf_rx_queue *rxq) { @@ -38,8 +34,11 @@ iavf_rxq_rearm(struct iavf_rx_queue *rxq) dma_addr0 = _mm_setzero_si128(); for (i = 0; i < IAVF_VPMD_DESCS_PER_LOOP; i++) { rxp[i] = &rxq->fake_mbuf; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp[i].read, dma_addr0); +__rte_diagnostic_pop } } rte_eth_devices[rxq->port_id].data->rx_mbuf_alloc_failed += @@ -69,8 +68,11 @@ iavf_rxq_rearm(struct iavf_rx_queue *rxq) dma_addr1 = _mm_add_epi64(dma_addr1, hdr_room); /* flush desc with pa dma_addr */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp++->read, dma_addr0); _mm_store_si128((__m128i *)&rxdp++->read, dma_addr1); +__rte_diagnostic_pop } rxq->rxrearm_start += rxq->rx_free_thresh; @@ -578,7 +580,10 @@ _recv_raw_pkts_vec(struct iavf_rx_queue *rxq, struct rte_mbuf **rx_pkts, mbp1 = _mm_loadu_si128((__m128i *)&sw_ring[pos]); /* Read desc statuses backwards to avoid race condition */ /* A.1 load desc[3] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[3] = _mm_loadu_si128((__m128i *)(rxdp + 3)); +__rte_diagnostic_pop rte_compiler_barrier(); /* B.2 copy 2 64 bit or 4 32 bit mbuf point into rx_pkts */ @@ -590,11 +595,14 @@ _recv_raw_pkts_vec(struct iavf_rx_queue *rxq, struct rte_mbuf **rx_pkts, #endif /* A.1 load desc[2-0] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[2] = _mm_loadu_si128((__m128i *)(rxdp + 2)); rte_compiler_barrier(); descs[1] = _mm_loadu_si128((__m128i *)(rxdp + 1)); rte_compiler_barrier(); descs[0] = _mm_loadu_si128((__m128i *)(rxdp)); +__rte_diagnostic_pop #if defined(RTE_ARCH_X86_64) /* B.2 copy 2 mbuf point into rx_pkts */ @@ -783,7 +791,7 @@ _recv_raw_pkts_vec_flex_rxd(struct iavf_rx_queue *rxq, /* Just the act of getting into the function from the application is * going to cost about 7 cycles */ - rxdp = (union iavf_rx_flex_desc *)rxq->rx_ring + rxq->rx_tail; + rxdp = (volatile union iavf_rx_flex_desc *)rxq->rx_ring + rxq->rx_tail; rte_prefetch0(rxdp); @@ -864,7 +872,10 @@ _recv_raw_pkts_vec_flex_rxd(struct iavf_rx_queue *rxq, mbp1 = _mm_loadu_si128((__m128i *)&sw_ring[pos]); /* Read desc statuses backwards to avoid race condition */ /* A.1 load desc[3] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[3] = _mm_loadu_si128((__m128i *)(rxdp + 3)); +__rte_diagnostic_pop rte_compiler_barrier(); /* B.2 copy 2 64 bit or 4 32 bit mbuf point into rx_pkts */ @@ -876,11 +887,14 @@ _recv_raw_pkts_vec_flex_rxd(struct iavf_rx_queue *rxq, #endif /* A.1 load desc[2-0] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[2] = _mm_loadu_si128((__m128i *)(rxdp + 2)); rte_compiler_barrier(); descs[1] = _mm_loadu_si128((__m128i *)(rxdp + 1)); rte_compiler_barrier(); descs[0] = _mm_loadu_si128((__m128i *)(rxdp)); +__rte_diagnostic_pop #if defined(RTE_ARCH_X86_64) /* B.2 copy 2 mbuf point into rx_pkts */ @@ -927,6 +941,8 @@ _recv_raw_pkts_vec_flex_rxd(struct iavf_rx_queue *rxq, offloads & RTE_ETH_RX_OFFLOAD_TIMESTAMP || rxq->rx_flags & IAVF_RX_FLAGS_VLAN_TAG_LOC_L2TAG2_2) { /* load bottom half of every 32B desc */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs_bh[3] = _mm_load_si128 ((void *)(&rxdp[3].wb.status_error1)); rte_compiler_barrier(); @@ -938,6 +954,7 @@ _recv_raw_pkts_vec_flex_rxd(struct iavf_rx_queue *rxq, rte_compiler_barrier(); descs_bh[0] = _mm_load_si128 ((void *)(&rxdp[0].wb.status_error1)); +__rte_diagnostic_pop } if (offloads & RTE_ETH_RX_OFFLOAD_RSS_HASH) { @@ -1349,7 +1366,10 @@ vtx1(volatile struct iavf_tx_desc *txdp, struct rte_mbuf *pkt, uint64_t flags) __m128i descriptor = _mm_set_epi64x(high_qw, pkt->buf_iova + pkt->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)txdp, descriptor); +__rte_diagnostic_pop } static inline void diff --git a/drivers/net/ice/ice_rxtx_common_avx.h b/drivers/net/ice/ice_rxtx_common_avx.h index dacb87dcb0..b9cb1f5279 100644 --- a/drivers/net/ice/ice_rxtx_common_avx.h +++ b/drivers/net/ice/ice_rxtx_common_avx.h @@ -7,10 +7,6 @@ #include "ice_rxtx.h" -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - #ifdef __AVX2__ static __rte_always_inline void ice_rxq_rearm_common(struct ice_rx_queue *rxq, __rte_unused bool avx512) @@ -33,8 +29,11 @@ ice_rxq_rearm_common(struct ice_rx_queue *rxq, __rte_unused bool avx512) dma_addr0 = _mm_setzero_si128(); for (i = 0; i < ICE_DESCS_PER_LOOP; i++) { rxep[i].mbuf = &rxq->fake_mbuf; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp[i].read, dma_addr0); +__rte_diagnostic_pop } } rte_eth_devices[rxq->port_id].data->rx_mbuf_alloc_failed += @@ -77,8 +76,11 @@ ice_rxq_rearm_common(struct ice_rx_queue *rxq, __rte_unused bool avx512) dma_addr1 = _mm_add_epi64(dma_addr1, hdr_room); /* flush desc with pa dma_addr */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp++->read, dma_addr0); _mm_store_si128((__m128i *)&rxdp++->read, dma_addr1); +__rte_diagnostic_pop } #else #ifdef __AVX512VL__ diff --git a/drivers/net/ice/ice_rxtx_vec_avx2.c b/drivers/net/ice/ice_rxtx_vec_avx2.c index d6e88dbb29..6c6a810a15 100644 --- a/drivers/net/ice/ice_rxtx_vec_avx2.c +++ b/drivers/net/ice/ice_rxtx_vec_avx2.c @@ -7,10 +7,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static __rte_always_inline void ice_rxq_rearm(struct ice_rx_queue *rxq) { @@ -254,6 +250,8 @@ _ice_recv_raw_pkts_vec_avx2(struct ice_rx_queue *rxq, struct rte_mbuf **rx_pkts, _mm256_loadu_si256((void *)&sw_ring[i + 4])); #endif +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc7 = _mm_load_si128((void *)(rxdp + 7)); rte_compiler_barrier(); const __m128i raw_desc6 = _mm_load_si128((void *)(rxdp + 6)); @@ -269,6 +267,7 @@ _ice_recv_raw_pkts_vec_avx2(struct ice_rx_queue *rxq, struct rte_mbuf **rx_pkts, const __m128i raw_desc1 = _mm_load_si128((void *)(rxdp + 1)); rte_compiler_barrier(); const __m128i raw_desc0 = _mm_load_si128((void *)(rxdp + 0)); +__rte_diagnostic_pop const __m256i raw_desc6_7 = _mm256_inserti128_si256(_mm256_castsi128_si256(raw_desc6), raw_desc7, 1); @@ -444,6 +443,8 @@ _ice_recv_raw_pkts_vec_avx2(struct ice_rx_queue *rxq, struct rte_mbuf **rx_pkts, if (rxq->vsi->adapter->pf.dev_data->dev_conf.rxmode.offloads & RTE_ETH_RX_OFFLOAD_RSS_HASH) { /* load bottom half of every 32B desc */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc_bh7 = _mm_load_si128 ((void *)(&rxdp[7].wb.status_error1)); @@ -475,6 +476,7 @@ _ice_recv_raw_pkts_vec_avx2(struct ice_rx_queue *rxq, struct rte_mbuf **rx_pkts, const __m128i raw_desc_bh0 = _mm_load_si128 ((void *)(&rxdp[0].wb.status_error1)); +__rte_diagnostic_pop __m256i raw_desc_bh6_7 = _mm256_inserti128_si256 @@ -790,7 +792,10 @@ ice_vtx1(volatile struct ice_tx_desc *txdp, ice_txd_enable_offload(pkt, &high_qw); __m128i descriptor = _mm_set_epi64x(high_qw, rte_pktmbuf_iova(pkt)); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)txdp, descriptor); +__rte_diagnostic_pop } static __rte_always_inline void @@ -841,8 +846,11 @@ ice_vtx(volatile struct ice_tx_desc *txdp, _mm256_set_epi64x (hi_qw1, rte_pktmbuf_iova(pkt[1]), hi_qw0, rte_pktmbuf_iova(pkt[0])); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm256_store_si256((void *)(txdp + 2), desc2_3); _mm256_store_si256((void *)txdp, desc0_1); +__rte_diagnostic_pop } /* do any last ones */ diff --git a/drivers/net/ice/ice_rxtx_vec_avx512.c b/drivers/net/ice/ice_rxtx_vec_avx512.c index add095ef06..5a774550dc 100644 --- a/drivers/net/ice/ice_rxtx_vec_avx512.c +++ b/drivers/net/ice/ice_rxtx_vec_avx512.c @@ -7,10 +7,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - #define ICE_DESCS_PER_LOOP_AVX 8 static __rte_always_inline void @@ -243,6 +239,8 @@ _ice_recv_raw_pkts_vec_avx512(struct ice_rx_queue *rxq, __m256i raw_desc0_1, raw_desc2_3, raw_desc4_5, raw_desc6_7; /* load in descriptors, in reverse order */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc7 = _mm_load_si128((void *)(rxdp + 7)); rte_compiler_barrier(); @@ -266,6 +264,7 @@ _ice_recv_raw_pkts_vec_avx512(struct ice_rx_queue *rxq, rte_compiler_barrier(); const __m128i raw_desc0 = _mm_load_si128((void *)(rxdp + 0)); +__rte_diagnostic_pop raw_desc6_7 = _mm256_inserti128_si256 @@ -474,6 +473,8 @@ _ice_recv_raw_pkts_vec_avx512(struct ice_rx_queue *rxq, if (rxq->vsi->adapter->pf.dev_data->dev_conf.rxmode.offloads & RTE_ETH_RX_OFFLOAD_RSS_HASH) { /* load bottom half of every 32B desc */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc_bh7 = _mm_load_si128 ((void *)(&rxdp[7].wb.status_error1)); @@ -505,6 +506,7 @@ _ice_recv_raw_pkts_vec_avx512(struct ice_rx_queue *rxq, const __m128i raw_desc_bh0 = _mm_load_si128 ((void *)(&rxdp[0].wb.status_error1)); +__rte_diagnostic_pop __m256i raw_desc_bh6_7 = _mm256_inserti128_si256 @@ -987,7 +989,10 @@ ice_vtx1(volatile struct ice_tx_desc *txdp, ice_txd_enable_offload(pkt, &high_qw); __m128i descriptor = _mm_set_epi64x(high_qw, rte_pktmbuf_iova(pkt)); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)txdp, descriptor); +__rte_diagnostic_pop } static __rte_always_inline void @@ -1029,7 +1034,10 @@ ice_vtx(volatile struct ice_tx_desc *txdp, struct rte_mbuf **pkt, hi_qw2, rte_pktmbuf_iova(pkt[2]), hi_qw1, rte_pktmbuf_iova(pkt[1]), hi_qw0, rte_pktmbuf_iova(pkt[0])); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm512_storeu_si512((void *)txdp, desc0_3); +__rte_diagnostic_pop } /* do any last ones */ diff --git a/drivers/net/ice/ice_rxtx_vec_common.h b/drivers/net/ice/ice_rxtx_vec_common.h index 4b73465af5..45147decff 100644 --- a/drivers/net/ice/ice_rxtx_vec_common.h +++ b/drivers/net/ice/ice_rxtx_vec_common.h @@ -7,10 +7,6 @@ #include "ice_rxtx.h" -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static inline uint16_t ice_rx_reassemble_packets(struct ice_rx_queue *rxq, struct rte_mbuf **rx_bufs, uint16_t nb_bufs, uint8_t *split_flags) diff --git a/drivers/net/ice/ice_rxtx_vec_sse.c b/drivers/net/ice/ice_rxtx_vec_sse.c index c01d8ede29..f2991cee1a 100644 --- a/drivers/net/ice/ice_rxtx_vec_sse.c +++ b/drivers/net/ice/ice_rxtx_vec_sse.c @@ -6,10 +6,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static inline __m128i ice_flex_rxd_to_fdir_flags_vec(const __m128i fdir_id0_3) { @@ -52,8 +48,11 @@ ice_rxq_rearm(struct ice_rx_queue *rxq) dma_addr0 = _mm_setzero_si128(); for (i = 0; i < ICE_DESCS_PER_LOOP; i++) { rxep[i].mbuf = &rxq->fake_mbuf; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp[i].read, dma_addr0); +__rte_diagnostic_pop } } rte_eth_devices[rxq->port_id].data->rx_mbuf_alloc_failed += @@ -91,8 +90,11 @@ ice_rxq_rearm(struct ice_rx_queue *rxq) dma_addr1 = _mm_add_epi64(dma_addr1, hdr_room); /* flush desc with pa dma_addr */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp++->read, dma_addr0); _mm_store_si128((__m128i *)&rxdp++->read, dma_addr1); +__rte_diagnostic_pop } rxq->rxrearm_start += ICE_RXQ_REARM_THRESH; @@ -425,7 +427,10 @@ _ice_recv_raw_pkts_vec(struct ice_rx_queue *rxq, struct rte_mbuf **rx_pkts, mbp1 = _mm_loadu_si128((__m128i *)&sw_ring[pos]); /* Read desc statuses backwards to avoid race condition */ /* A.1 load desc[3] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[3] = _mm_loadu_si128((__m128i *)(rxdp + 3)); +__rte_diagnostic_pop rte_compiler_barrier(); /* B.2 copy 2 64 bit or 4 32 bit mbuf point into rx_pkts */ @@ -437,11 +442,14 @@ _ice_recv_raw_pkts_vec(struct ice_rx_queue *rxq, struct rte_mbuf **rx_pkts, #endif /* A.1 load desc[2-0] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[2] = _mm_loadu_si128((__m128i *)(rxdp + 2)); rte_compiler_barrier(); descs[1] = _mm_loadu_si128((__m128i *)(rxdp + 1)); rte_compiler_barrier(); descs[0] = _mm_loadu_si128((__m128i *)(rxdp)); +__rte_diagnostic_pop #if defined(RTE_ARCH_X86_64) /* B.2 copy 2 mbuf point into rx_pkts */ @@ -489,6 +497,8 @@ _ice_recv_raw_pkts_vec(struct ice_rx_queue *rxq, struct rte_mbuf **rx_pkts, if (rxq->vsi->adapter->pf.dev_data->dev_conf.rxmode.offloads & RTE_ETH_RX_OFFLOAD_RSS_HASH) { /* load bottom half of every 32B desc */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual const __m128i raw_desc_bh3 = _mm_load_si128 ((void *)(&rxdp[3].wb.status_error1)); @@ -504,6 +514,7 @@ _ice_recv_raw_pkts_vec(struct ice_rx_queue *rxq, struct rte_mbuf **rx_pkts, const __m128i raw_desc_bh0 = _mm_load_si128 ((void *)(&rxdp[0].wb.status_error1)); +__rte_diagnostic_pop /** * to shift the 32b RSS hash value to the @@ -680,7 +691,10 @@ ice_vtx1(volatile struct ice_tx_desc *txdp, struct rte_mbuf *pkt, ((uint64_t)pkt->data_len << ICE_TXD_QW1_TX_BUF_SZ_S)); __m128i descriptor = _mm_set_epi64x(high_qw, rte_pktmbuf_iova(pkt)); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)txdp, descriptor); +__rte_diagnostic_pop } static inline void diff --git a/drivers/net/idpf/idpf_rxtx_vec_common.h b/drivers/net/idpf/idpf_rxtx_vec_common.h index 2787d27616..002c1e6948 100644 --- a/drivers/net/idpf/idpf_rxtx_vec_common.h +++ b/drivers/net/idpf/idpf_rxtx_vec_common.h @@ -11,10 +11,6 @@ #include "idpf_ethdev.h" #include "idpf_rxtx.h" -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - #define IDPF_SCALAR_PATH 0 #define IDPF_VECTOR_PATH 1 #define IDPF_RX_NO_VECTOR_FLAGS ( \ diff --git a/drivers/net/ixgbe/ixgbe_recycle_mbufs_vec_common.c b/drivers/net/ixgbe/ixgbe_recycle_mbufs_vec_common.c index d451562269..92a89f8def 100644 --- a/drivers/net/ixgbe/ixgbe_recycle_mbufs_vec_common.c +++ b/drivers/net/ixgbe/ixgbe_recycle_mbufs_vec_common.c @@ -8,8 +8,6 @@ #include "ixgbe_ethdev.h" #include "ixgbe_rxtx.h" -#pragma GCC diagnostic ignored "-Wcast-qual" - void ixgbe_recycle_rx_descriptors_refill_vec(void *rx_queue, uint16_t nb_mbufs) { diff --git a/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c b/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c index 952b032eb6..4e7a64b39f 100644 --- a/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c +++ b/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c @@ -11,8 +11,6 @@ #include "ixgbe_rxtx.h" #include "ixgbe_rxtx_vec_common.h" -#pragma GCC diagnostic ignored "-Wcast-qual" - static inline void ixgbe_rxq_rearm(struct ixgbe_rx_queue *rxq) { diff --git a/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c b/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c index a77370cdb7..8e4048a32f 100644 --- a/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c +++ b/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c @@ -12,10 +12,6 @@ #include -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - static inline void ixgbe_rxq_rearm(struct ixgbe_rx_queue *rxq) { @@ -41,8 +37,11 @@ ixgbe_rxq_rearm(struct ixgbe_rx_queue *rxq) dma_addr0 = _mm_setzero_si128(); for (i = 0; i < RTE_IXGBE_DESCS_PER_LOOP; i++) { rxep[i].mbuf = &rxq->fake_mbuf; +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp[i].read, dma_addr0); +__rte_diagnostic_pop } } rte_eth_devices[rxq->port_id].data->rx_mbuf_alloc_failed += @@ -76,8 +75,11 @@ ixgbe_rxq_rearm(struct ixgbe_rx_queue *rxq) dma_addr1 = _mm_and_si128(dma_addr1, hba_msk); /* flush desc with pa dma_addr */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&rxdp++->read, dma_addr0); _mm_store_si128((__m128i *)&rxdp++->read, dma_addr1); +__rte_diagnostic_pop } rxq->rxrearm_start += RTE_IXGBE_RXQ_REARM_THRESH; @@ -466,7 +468,10 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts, /* Read desc statuses backwards to avoid race condition */ /* A.1 load desc[3] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[3] = _mm_loadu_si128((__m128i *)(rxdp + 3)); +__rte_diagnostic_pop rte_compiler_barrier(); /* B.2 copy 2 64 bit or 4 32 bit mbuf point into rx_pkts */ @@ -478,11 +483,14 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts, #endif /* A.1 load desc[2-0] */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual descs[2] = _mm_loadu_si128((__m128i *)(rxdp + 2)); rte_compiler_barrier(); descs[1] = _mm_loadu_si128((__m128i *)(rxdp + 1)); rte_compiler_barrier(); descs[0] = _mm_loadu_si128((__m128i *)(rxdp)); +__rte_diagnostic_pop #if defined(RTE_ARCH_X86_64) /* B.2 copy 2 mbuf point into rx_pkts */ @@ -676,7 +684,10 @@ vtx1(volatile union ixgbe_adv_tx_desc *txdp, __m128i descriptor = _mm_set_epi64x((uint64_t)pkt->pkt_len << 46 | flags | pkt->data_len, pkt->buf_iova + pkt->data_off); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual _mm_store_si128((__m128i *)&txdp->read, descriptor); +__rte_diagnostic_pop } static inline void diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c index 16ddd05448..0a89d2c414 100644 --- a/drivers/net/mlx5/mlx5_flow.c +++ b/drivers/net/mlx5/mlx5_flow.c @@ -7287,10 +7287,10 @@ flow_tunnel_from_rule(const struct mlx5_flow *flow) { struct mlx5_flow_tunnel *tunnel; -#pragma GCC diagnostic push -#pragma GCC diagnostic ignored "-Wcast-qual" +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual tunnel = (typeof(tunnel))flow->tunnel; -#pragma GCC diagnostic pop +__rte_diagnostic_pop return tunnel; } diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_altivec.h b/drivers/net/mlx5/mlx5_rxtx_vec_altivec.h index 240987d03d..b37483bcca 100644 --- a/drivers/net/mlx5/mlx5_rxtx_vec_altivec.h +++ b/drivers/net/mlx5/mlx5_rxtx_vec_altivec.h @@ -25,11 +25,6 @@ #include "mlx5_rxtx_vec.h" #include "mlx5_autoconf.h" -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#pragma GCC diagnostic ignored "-Wstrict-aliasing" -#endif - /** * Store free buffers to RX SW ring. * diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h index dc1d30753d..e7deacc1fb 100644 --- a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h +++ b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h @@ -25,8 +25,6 @@ #include "mlx5_rxtx_vec.h" #include "mlx5_autoconf.h" -#pragma GCC diagnostic ignored "-Wcast-qual" - /** * Store free buffers to RX SW ring. * diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_sse.h b/drivers/net/mlx5/mlx5_rxtx_vec_sse.h index 81a177fce7..f4d08d5b30 100644 --- a/drivers/net/mlx5/mlx5_rxtx_vec_sse.h +++ b/drivers/net/mlx5/mlx5_rxtx_vec_sse.h @@ -24,10 +24,6 @@ #include "mlx5_rxtx_vec.h" #include "mlx5_autoconf.h" -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - /** * Store free buffers to RX SW ring. * @@ -75,7 +71,8 @@ static inline uint16_t rxq_cq_decompress_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, struct rte_mbuf **elts, bool keep) { - volatile struct mlx5_mini_cqe8 *mcq = (void *)(cq + !rxq->cqe_comp_layout); + volatile struct mlx5_mini_cqe8 *mcq = + (volatile void *)(cq + !rxq->cqe_comp_layout); /* Title packet is pre-built. */ struct rte_mbuf *t_pkt = rxq->cqe_comp_layout ? &rxq->title_pkt : elts[0]; unsigned int pos; @@ -129,8 +126,11 @@ rxq_cq_decompress_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, * E. store flow tag (rte_flow mark). */ cycle: +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual if (rxq->cqe_comp_layout) rte_prefetch0((void *)(cq + mcqe_n)); +__rte_diagnostic_pop for (pos = 0; pos < mcqe_n; ) { __m128i mcqe1, mcqe2; __m128i rxdf1, rxdf2; @@ -138,6 +138,8 @@ rxq_cq_decompress_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, __m128i byte_cnt, invalid_mask; #endif +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual if (!rxq->cqe_comp_layout) for (i = 0; i < MLX5_VPMD_DESCS_PER_LOOP; ++i) if (likely(pos + i < mcqe_n)) @@ -145,6 +147,7 @@ rxq_cq_decompress_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, /* A.1 load mCQEs into a 128bit register. */ mcqe1 = _mm_loadu_si128((__m128i *)&mcq[pos % 8]); mcqe2 = _mm_loadu_si128((__m128i *)&mcq[pos % 8 + 2]); +__rte_diagnostic_pop /* B.1 store rearm data to mbuf. */ _mm_storeu_si128((__m128i *)&elts[pos]->rearm_data, rearm); _mm_storeu_si128((__m128i *)&elts[pos + 1]->rearm_data, rearm); @@ -354,9 +357,12 @@ rxq_cq_decompress_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, /* Move to next CQE and invalidate consumed CQEs. */ if (!rxq->cqe_comp_layout) { if (!(pos & 0x7) && pos < mcqe_n) { +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual if (pos + 8 < mcqe_n) rte_prefetch0((void *)(cq + pos + 8)); - mcq = (void *)(cq + pos); +__rte_diagnostic_pop + mcq = (volatile void *)(cq + pos); for (i = 0; i < 8; ++i) cq[inv++].op_own = MLX5_CQE_INVALIDATE; } @@ -371,7 +377,7 @@ rxq_cq_decompress_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, MLX5_CQE_FORMAT(cq->op_own) == MLX5_COMPRESSED) { pos = 0; elts = &elts[mcqe_n]; - mcq = (void *)cq; + mcq = (volatile void *)cq; mcqe_n = MLX5_CQE_NUM_MINIS(cq->op_own) + 1; pkts_n += mcqe_n; goto cycle; @@ -651,6 +657,8 @@ rxq_cq_process_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, p = _mm_andnot_si128(mask, p); /* A.1 load cqes. */ p3 = _mm_extract_epi16(p, 3); +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual cqes[3] = _mm_loadl_epi64((__m128i *) &cq[pos + p3].sop_drop_qpn); rte_compiler_barrier(); @@ -683,6 +691,7 @@ rxq_cq_process_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, cqes[2] = _mm_blend_epi16(cqes[2], cqe_tmp1, 0x30); cqe_tmp2 = _mm_loadl_epi64((__m128i *)&cq[pos + p3].rsvd4[2]); cqe_tmp1 = _mm_loadl_epi64((__m128i *)&cq[pos + p2].rsvd4[2]); +__rte_diagnostic_pop cqes[3] = _mm_blend_epi16(cqes[3], cqe_tmp2, 0x04); cqes[2] = _mm_blend_epi16(cqes[2], cqe_tmp1, 0x04); /* C.2 generate final structure for mbuf with swapping bytes. */ @@ -700,6 +709,8 @@ rxq_cq_process_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, /* E.1 extract op_own field. */ op_own_tmp2 = _mm_unpacklo_epi32(cqes[2], cqes[3]); /* C.1 load remained CQE data and extract necessary fields. */ +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual cqe_tmp2 = _mm_load_si128((__m128i *)&cq[pos + p1]); cqe_tmp1 = _mm_load_si128((__m128i *)&cq[pos]); cqes[1] = _mm_blendv_epi8(cqes[1], cqe_tmp2, blend_mask); @@ -710,6 +721,7 @@ rxq_cq_process_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, cqes[0] = _mm_blend_epi16(cqes[0], cqe_tmp1, 0x30); cqe_tmp2 = _mm_loadl_epi64((__m128i *)&cq[pos + p1].rsvd4[2]); cqe_tmp1 = _mm_loadl_epi64((__m128i *)&cq[pos].rsvd4[2]); +__rte_diagnostic_pop cqes[1] = _mm_blend_epi16(cqes[1], cqe_tmp2, 0x04); cqes[0] = _mm_blend_epi16(cqes[0], cqe_tmp1, 0x04); /* C.2 generate final structure for mbuf with swapping bytes. */ diff --git a/drivers/net/tap/tap_flow.c b/drivers/net/tap/tap_flow.c index c0e44bb1a7..373b773e2d 100644 --- a/drivers/net/tap/tap_flow.c +++ b/drivers/net/tap/tap_flow.c @@ -23,10 +23,10 @@ #ifdef HAVE_BPF_RSS /* Workaround for warning in bpftool generated skeleton code */ -#pragma GCC diagnostic push -#pragma GCC diagnostic ignored "-Wcast-qual" +__rte_diagnostic_push +__rte_diagnostic_ignored_wcast_qual #include "tap_rss.skel.h" -#pragma GCC diagnostic pop +__rte_diagnostic_pop #endif #define ISOLATE_HANDLE 1 diff --git a/drivers/net/virtio/virtio_rxtx_simple.c b/drivers/net/virtio/virtio_rxtx_simple.c index 438256970d..439e00a7e1 100644 --- a/drivers/net/virtio/virtio_rxtx_simple.c +++ b/drivers/net/virtio/virtio_rxtx_simple.c @@ -23,10 +23,6 @@ #include "virtio_rxtx_simple.h" -#ifndef __INTEL_COMPILER -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - int __rte_cold virtio_rxq_vec_setup(struct virtnet_rx *rxq) { -- 2.47.0.vfs.0.3