From patchwork Tue Nov 21 02:51:01 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Cao, Yahui" X-Patchwork-Id: 13462480 X-Patchwork-Delegate: kuba@kernel.org Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b="ll4ySK7p" Received: from mgamail.intel.com (mgamail.intel.com [134.134.136.100]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 35A4BCB; Mon, 20 Nov 2023 18:49:55 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1700534995; x=1732070995; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=Vsbg5WYpgIdaFM7kidb5HJYGh+XfqpqCp0zWMK3Lung=; b=ll4ySK7p3tEGIehaIYtQ+FRVV67Gbf46sLjEcsfI3nfMTX0k9j0kGPOl gYp29Hb/6DWN9+tUb61rfUW0VkkOR1HwCSAsOog/3/s5RZFtjdyYTHS2e MzXjOgf2YsQ1WqrHr8tupgSn+H/yBiPQ7ez3ElW6tGxJhGmREnSdMX2a1 wBi/cCNW8LiPSN6FXKt+fspWhCG3XlGLzcUm46Me+mQJ3s+eHLsVXOxK9 tSw3k3bckGOslW+2lRnN9SlLN1nVWqBDZBeooxWXU8uhGDDrCQuwIzbjY IWN3L6moSb3W4MMwecnpK6fW11m3L75S4u58xfyd7nhBjAx8/iYZnAqrY Q==; X-IronPort-AV: E=McAfee;i="6600,9927,10900"; a="458245861" X-IronPort-AV: E=Sophos;i="6.04,215,1695711600"; d="scan'208";a="458245861" Received: from fmsmga008.fm.intel.com ([10.253.24.58]) by orsmga105.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 20 Nov 2023 18:49:54 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=McAfee;i="6600,9927,10900"; a="832488229" X-IronPort-AV: E=Sophos;i="6.04,215,1695711600"; d="scan'208";a="832488229" Received: from dpdk-yahui-icx1.sh.intel.com ([10.67.111.85]) by fmsmga008.fm.intel.com with ESMTP; 20 Nov 2023 18:49:49 -0800 From: Yahui Cao To: intel-wired-lan@lists.osuosl.org Cc: kvm@vger.kernel.org, netdev@vger.kernel.org, lingyu.liu@intel.com, kevin.tian@intel.com, madhu.chittim@intel.com, sridhar.samudrala@intel.com, alex.williamson@redhat.com, jgg@nvidia.com, yishaih@nvidia.com, shameerali.kolothum.thodi@huawei.com, brett.creeley@amd.com, davem@davemloft.net, edumazet@google.com, kuba@kernel.org, pabeni@redhat.com Subject: [PATCH iwl-next v4 02/12] ice: Add function to get and set TX queue context Date: Tue, 21 Nov 2023 02:51:01 +0000 Message-Id: <20231121025111.257597-3-yahui.cao@intel.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20231121025111.257597-1-yahui.cao@intel.com> References: <20231121025111.257597-1-yahui.cao@intel.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Patchwork-Delegate: kuba@kernel.org Export TX queue context get and set function which is consumed by linux live migration driver to save and load device state. TX queue context contains static fields which does not change during TX traffic and dynamic fields which may change during TX traffic. Signed-off-by: Yahui Cao --- drivers/net/ethernet/intel/ice/ice_common.c | 216 +++++++++++++++++- drivers/net/ethernet/intel/ice/ice_common.h | 6 + .../net/ethernet/intel/ice/ice_hw_autogen.h | 15 ++ .../net/ethernet/intel/ice/ice_lan_tx_rx.h | 3 + 4 files changed, 239 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c index d0a3bed00921..8577a5ef423e 100644 --- a/drivers/net/ethernet/intel/ice/ice_common.c +++ b/drivers/net/ethernet/intel/ice/ice_common.c @@ -1645,7 +1645,10 @@ ice_read_rxq_ctx(struct ice_hw *hw, struct ice_rlan_ctx *rlan_ctx, return ice_get_ctx(ctx_buf, (u8 *)rlan_ctx, ice_rlan_ctx_info); } -/* LAN Tx Queue Context */ +/* LAN Tx Queue Context used for set Tx config by ice_aqc_opc_add_txqs, + * Bit[0-175] is valid + */ + const struct ice_ctx_ele ice_tlan_ctx_info[] = { /* Field Width LSB */ ICE_CTX_STORE(ice_tlan_ctx, base, 57, 0), @@ -1679,6 +1682,217 @@ const struct ice_ctx_ele ice_tlan_ctx_info[] = { { 0 } }; +/* LAN Tx Queue Context used for get Tx config from QTXCOMM_CNTX data, + * Bit[0-292] is valid, including internal queue state. Since internal + * queue state is dynamic field, its value will be cleared once queue + * is disabled + */ +static const struct ice_ctx_ele ice_tlan_ctx_data_info[] = { + /* Field Width LSB */ + ICE_CTX_STORE(ice_tlan_ctx, base, 57, 0), + ICE_CTX_STORE(ice_tlan_ctx, port_num, 3, 57), + ICE_CTX_STORE(ice_tlan_ctx, cgd_num, 5, 60), + ICE_CTX_STORE(ice_tlan_ctx, pf_num, 3, 65), + ICE_CTX_STORE(ice_tlan_ctx, vmvf_num, 10, 68), + ICE_CTX_STORE(ice_tlan_ctx, vmvf_type, 2, 78), + ICE_CTX_STORE(ice_tlan_ctx, src_vsi, 10, 80), + ICE_CTX_STORE(ice_tlan_ctx, tsyn_ena, 1, 90), + ICE_CTX_STORE(ice_tlan_ctx, internal_usage_flag, 1, 91), + ICE_CTX_STORE(ice_tlan_ctx, alt_vlan, 1, 92), + ICE_CTX_STORE(ice_tlan_ctx, cpuid, 8, 93), + ICE_CTX_STORE(ice_tlan_ctx, wb_mode, 1, 101), + ICE_CTX_STORE(ice_tlan_ctx, tphrd_desc, 1, 102), + ICE_CTX_STORE(ice_tlan_ctx, tphrd, 1, 103), + ICE_CTX_STORE(ice_tlan_ctx, tphwr_desc, 1, 104), + ICE_CTX_STORE(ice_tlan_ctx, cmpq_id, 9, 105), + ICE_CTX_STORE(ice_tlan_ctx, qnum_in_func, 14, 114), + ICE_CTX_STORE(ice_tlan_ctx, itr_notification_mode, 1, 128), + ICE_CTX_STORE(ice_tlan_ctx, adjust_prof_id, 6, 129), + ICE_CTX_STORE(ice_tlan_ctx, qlen, 13, 135), + ICE_CTX_STORE(ice_tlan_ctx, quanta_prof_idx, 4, 148), + ICE_CTX_STORE(ice_tlan_ctx, tso_ena, 1, 152), + ICE_CTX_STORE(ice_tlan_ctx, tso_qnum, 11, 153), + ICE_CTX_STORE(ice_tlan_ctx, legacy_int, 1, 164), + ICE_CTX_STORE(ice_tlan_ctx, drop_ena, 1, 165), + ICE_CTX_STORE(ice_tlan_ctx, cache_prof_idx, 2, 166), + ICE_CTX_STORE(ice_tlan_ctx, pkt_shaper_prof_idx, 3, 168), + ICE_CTX_STORE(ice_tlan_ctx, tail, 13, 184), + { 0 } +}; + +/** + * ice_copy_txq_ctx_from_hw - Copy txq context register from HW + * @hw: pointer to the hardware structure + * @ice_txq_ctx: pointer to the txq context + * + * Copy txq context from HW register space to dense structure + */ +static int +ice_copy_txq_ctx_from_hw(struct ice_hw *hw, u8 *ice_txq_ctx) +{ + u8 i; + + if (!ice_txq_ctx) + return -EINVAL; + + /* Copy each dword separately from HW */ + for (i = 0; i < ICE_TXQ_CTX_SIZE_DWORDS; i++) { + u32 *ctx = (u32 *)(ice_txq_ctx + (i * sizeof(u32))); + + *ctx = rd32(hw, GLCOMM_QTX_CNTX_DATA(i)); + + ice_debug(hw, ICE_DBG_QCTX, "qtxdata[%d]: %08X\n", i, *ctx); + } + + return 0; +} + +/** + * ice_copy_txq_ctx_to_hw - Copy txq context register into HW + * @hw: pointer to the hardware structure + * @ice_txq_ctx: pointer to the txq context + * + * Copy txq context from dense structure to HW register space + */ +static int +ice_copy_txq_ctx_to_hw(struct ice_hw *hw, u8 *ice_txq_ctx) +{ + u8 i; + + if (!ice_txq_ctx) + return -EINVAL; + + /* Copy each dword separately to HW */ + for (i = 0; i < ICE_TXQ_CTX_SIZE_DWORDS; i++) { + u32 *ctx = (u32 *)(ice_txq_ctx + (i * sizeof(u32))); + + wr32(hw, GLCOMM_QTX_CNTX_DATA(i), *ctx); + + ice_debug(hw, ICE_DBG_QCTX, "qtxdata[%d]: %08X\n", i, *ctx); + } + + return 0; +} + +/* Configuration access to tx ring context(from PF) is done via indirect + * interface, GLCOMM_QTX_CNTX_CTL/DATA registers. However, there registers + * are shared by all the PFs with single PCI card. Hence multiplied PF may + * access there registers simultaneously, causing access conflicts. Then + * card-level grained locking is required to protect these registers from + * being competed by PF devices within the same card. However, there is no + * such kind of card-level locking supported. Introduce a coarse grained + * global lock which is shared by all the PF driver. + * + * The overall flow is to acquire the lock, read/write TXQ context through + * GLCOMM_QTX_CNTX_CTL/DATA indirect interface and release the lock once + * access is completed. In this way, only one PF can have access to TXQ + * context safely. + */ +static DEFINE_MUTEX(ice_global_txq_ctx_lock); + +/** + * ice_read_txq_ctx - Read txq context from HW + * @hw: pointer to the hardware structure + * @tlan_ctx: pointer to the txq context + * @txq_index: the index of the Tx queue + * + * Read txq context from HW register space and then convert it from dense + * structure to sparse + */ +int +ice_read_txq_ctx(struct ice_hw *hw, struct ice_tlan_ctx *tlan_ctx, + u32 txq_index) +{ + u8 ctx_buf[ICE_TXQ_CTX_SZ] = { 0 }; + int status; + u32 txq_base; + u32 cmd, reg; + + if (!tlan_ctx) + return -EINVAL; + + if (txq_index > QTX_COMM_HEAD_MAX_INDEX) + return -EINVAL; + + /* Get TXQ base within card space */ + txq_base = rd32(hw, PFLAN_TX_QALLOC(hw->pf_id)); + txq_base = (txq_base & PFLAN_TX_QALLOC_FIRSTQ_M) >> + PFLAN_TX_QALLOC_FIRSTQ_S; + + cmd = (GLCOMM_QTX_CNTX_CTL_CMD_READ + << GLCOMM_QTX_CNTX_CTL_CMD_S) & GLCOMM_QTX_CNTX_CTL_CMD_M; + reg = cmd | GLCOMM_QTX_CNTX_CTL_CMD_EXEC_M | + (((txq_base + txq_index) << GLCOMM_QTX_CNTX_CTL_QUEUE_ID_S) & + GLCOMM_QTX_CNTX_CTL_QUEUE_ID_M); + + mutex_lock(&ice_global_txq_ctx_lock); + + wr32(hw, GLCOMM_QTX_CNTX_CTL, reg); + ice_flush(hw); + + status = ice_copy_txq_ctx_from_hw(hw, ctx_buf); + if (status) { + mutex_unlock(&ice_global_txq_ctx_lock); + return status; + } + + mutex_unlock(&ice_global_txq_ctx_lock); + + return ice_get_ctx(ctx_buf, (u8 *)tlan_ctx, ice_tlan_ctx_data_info); +} + +/** + * ice_write_txq_ctx - Write txq context from HW + * @hw: pointer to the hardware structure + * @tlan_ctx: pointer to the txq context + * @txq_index: the index of the Tx queue + * + * Convert txq context from sparse to dense structure and then write + * it to HW register space + */ +int +ice_write_txq_ctx(struct ice_hw *hw, struct ice_tlan_ctx *tlan_ctx, + u32 txq_index) +{ + u8 ctx_buf[ICE_TXQ_CTX_SZ] = { 0 }; + int status; + u32 txq_base; + u32 cmd, reg; + + if (!tlan_ctx) + return -EINVAL; + + if (txq_index > QTX_COMM_HEAD_MAX_INDEX) + return -EINVAL; + + ice_set_ctx(hw, (u8 *)tlan_ctx, ctx_buf, ice_tlan_ctx_info); + + /* Get TXQ base within card space */ + txq_base = rd32(hw, PFLAN_TX_QALLOC(hw->pf_id)); + txq_base = (txq_base & PFLAN_TX_QALLOC_FIRSTQ_M) >> + PFLAN_TX_QALLOC_FIRSTQ_S; + + cmd = (GLCOMM_QTX_CNTX_CTL_CMD_WRITE_NO_DYN + << GLCOMM_QTX_CNTX_CTL_CMD_S) & GLCOMM_QTX_CNTX_CTL_CMD_M; + reg = cmd | GLCOMM_QTX_CNTX_CTL_CMD_EXEC_M | + (((txq_base + txq_index) << GLCOMM_QTX_CNTX_CTL_QUEUE_ID_S) & + GLCOMM_QTX_CNTX_CTL_QUEUE_ID_M); + + mutex_lock(&ice_global_txq_ctx_lock); + + status = ice_copy_txq_ctx_to_hw(hw, ctx_buf); + if (status) { + mutex_lock(&ice_global_txq_ctx_lock); + return status; + } + + wr32(hw, GLCOMM_QTX_CNTX_CTL, reg); + ice_flush(hw); + + mutex_unlock(&ice_global_txq_ctx_lock); + + return 0; +} /* Sideband Queue command wrappers */ /** diff --git a/drivers/net/ethernet/intel/ice/ice_common.h b/drivers/net/ethernet/intel/ice/ice_common.h index df9c7f30592a..40fbb9088475 100644 --- a/drivers/net/ethernet/intel/ice/ice_common.h +++ b/drivers/net/ethernet/intel/ice/ice_common.h @@ -58,6 +58,12 @@ ice_write_rxq_ctx(struct ice_hw *hw, struct ice_rlan_ctx *rlan_ctx, int ice_read_rxq_ctx(struct ice_hw *hw, struct ice_rlan_ctx *rlan_ctx, u32 rxq_index); +int +ice_read_txq_ctx(struct ice_hw *hw, struct ice_tlan_ctx *tlan_ctx, + u32 txq_index); +int +ice_write_txq_ctx(struct ice_hw *hw, struct ice_tlan_ctx *tlan_ctx, + u32 txq_index); int ice_aq_get_rss_lut(struct ice_hw *hw, struct ice_aq_get_set_rss_lut_params *get_params); diff --git a/drivers/net/ethernet/intel/ice/ice_hw_autogen.h b/drivers/net/ethernet/intel/ice/ice_hw_autogen.h index 86936b758ade..7410da715ad4 100644 --- a/drivers/net/ethernet/intel/ice/ice_hw_autogen.h +++ b/drivers/net/ethernet/intel/ice/ice_hw_autogen.h @@ -8,6 +8,7 @@ #define QTX_COMM_DBELL(_DBQM) (0x002C0000 + ((_DBQM) * 4)) #define QTX_COMM_HEAD(_DBQM) (0x000E0000 + ((_DBQM) * 4)) +#define QTX_COMM_HEAD_MAX_INDEX 16383 #define QTX_COMM_HEAD_HEAD_S 0 #define QTX_COMM_HEAD_HEAD_M ICE_M(0x1FFF, 0) #define PF_FW_ARQBAH 0x00080180 @@ -258,6 +259,9 @@ #define VPINT_ALLOC_PCI_VALID_M BIT(31) #define VPINT_MBX_CTL(_VSI) (0x0016A000 + ((_VSI) * 4)) #define VPINT_MBX_CTL_CAUSE_ENA_M BIT(30) +#define PFLAN_TX_QALLOC(_PF) (0x001D2580 + ((_PF) * 4)) +#define PFLAN_TX_QALLOC_FIRSTQ_S 0 +#define PFLAN_TX_QALLOC_FIRSTQ_M ICE_M(0x3FFF, 0) #define GLLAN_RCTL_0 0x002941F8 #define QRX_CONTEXT(_i, _QRX) (0x00280000 + ((_i) * 8192 + (_QRX) * 4)) #define QRX_CTRL(_QRX) (0x00120000 + ((_QRX) * 4)) @@ -362,6 +366,17 @@ #define GLNVM_ULD_POR_DONE_1_M BIT(8) #define GLNVM_ULD_PCIER_DONE_2_M BIT(9) #define GLNVM_ULD_PE_DONE_M BIT(10) +#define GLCOMM_QTX_CNTX_CTL 0x002D2DC8 +#define GLCOMM_QTX_CNTX_CTL_QUEUE_ID_S 0 +#define GLCOMM_QTX_CNTX_CTL_QUEUE_ID_M ICE_M(0x3FFF, 0) +#define GLCOMM_QTX_CNTX_CTL_CMD_S 16 +#define GLCOMM_QTX_CNTX_CTL_CMD_M ICE_M(0x7, 16) +#define GLCOMM_QTX_CNTX_CTL_CMD_READ 0 +#define GLCOMM_QTX_CNTX_CTL_CMD_WRITE 1 +#define GLCOMM_QTX_CNTX_CTL_CMD_RESET 3 +#define GLCOMM_QTX_CNTX_CTL_CMD_WRITE_NO_DYN 4 +#define GLCOMM_QTX_CNTX_CTL_CMD_EXEC_M BIT(19) +#define GLCOMM_QTX_CNTX_DATA(_i) (0x002D2D40 + ((_i) * 4)) #define GLPCI_CNF2 0x000BE004 #define GLPCI_CNF2_CACHELINE_SIZE_M BIT(1) #define PF_FUNC_RID 0x0009E880 diff --git a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h index 89f986a75cc8..79e07c863ae0 100644 --- a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h +++ b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h @@ -431,6 +431,8 @@ enum ice_rx_flex_desc_status_error_1_bits { #define ICE_RXQ_CTX_SIZE_DWORDS 8 #define ICE_RXQ_CTX_SZ (ICE_RXQ_CTX_SIZE_DWORDS * sizeof(u32)) +#define ICE_TXQ_CTX_SIZE_DWORDS 10 +#define ICE_TXQ_CTX_SZ (ICE_TXQ_CTX_SIZE_DWORDS * sizeof(u32)) #define ICE_TX_CMPLTNQ_CTX_SIZE_DWORDS 22 #define ICE_TX_DRBELL_Q_CTX_SIZE_DWORDS 5 #define GLTCLAN_CQ_CNTX(i, CQ) (GLTCLAN_CQ_CNTX0(CQ) + ((i) * 0x0800)) @@ -649,6 +651,7 @@ struct ice_tlan_ctx { u8 cache_prof_idx; u8 pkt_shaper_prof_idx; u8 int_q_state; /* width not needed - internal - DO NOT WRITE!!! */ + u16 tail; }; /* The ice_ptype_lkup table is used to convert from the 10-bit ptype in the