diff options
Diffstat (limited to 'drivers/crypto/marvell')
19 files changed, 383 insertions, 176 deletions
diff --git a/drivers/crypto/marvell/cesa/cesa.c b/drivers/crypto/marvell/cesa/cesa.c index fa08f10e6f3f..9c21f5d835d2 100644 --- a/drivers/crypto/marvell/cesa/cesa.c +++ b/drivers/crypto/marvell/cesa/cesa.c @@ -94,7 +94,7 @@ static int mv_cesa_std_process(struct mv_cesa_engine *engine, u32 status) static int mv_cesa_int_process(struct mv_cesa_engine *engine, u32 status) { - if (engine->chain.first && engine->chain.last) + if (engine->chain_hw.first && engine->chain_hw.last) return mv_cesa_tdma_process(engine, status); return mv_cesa_std_process(engine, status); diff --git a/drivers/crypto/marvell/cesa/cesa.h b/drivers/crypto/marvell/cesa/cesa.h index d215a6bed6bc..50ca1039fdaa 100644 --- a/drivers/crypto/marvell/cesa/cesa.h +++ b/drivers/crypto/marvell/cesa/cesa.h @@ -440,8 +440,10 @@ struct mv_cesa_dev { * SRAM * @queue: fifo of the pending crypto requests * @load: engine load counter, useful for load balancing - * @chain: list of the current tdma descriptors being processed - * by this engine. + * @chain_hw: list of the current tdma descriptors being processed + * by the hardware. + * @chain_sw: list of the current tdma descriptors that will be + * submitted to the hardware. * @complete_queue: fifo of the processed requests by the engine * * Structure storing CESA engine information. @@ -463,7 +465,8 @@ struct mv_cesa_engine { struct gen_pool *pool; struct crypto_queue queue; atomic_t load; - struct mv_cesa_tdma_chain chain; + struct mv_cesa_tdma_chain chain_hw; + struct mv_cesa_tdma_chain chain_sw; struct list_head complete_queue; int irq; }; diff --git a/drivers/crypto/marvell/cesa/cipher.c b/drivers/crypto/marvell/cesa/cipher.c index cf62db50f958..3fe0fd9226cf 100644 --- a/drivers/crypto/marvell/cesa/cipher.c +++ b/drivers/crypto/marvell/cesa/cipher.c @@ -75,9 +75,12 @@ mv_cesa_skcipher_dma_cleanup(struct skcipher_request *req) static inline void mv_cesa_skcipher_cleanup(struct skcipher_request *req) { struct mv_cesa_skcipher_req *creq = skcipher_request_ctx(req); + struct mv_cesa_engine *engine = creq->base.engine; if (mv_cesa_req_get_type(&creq->base) == CESA_DMA_REQ) mv_cesa_skcipher_dma_cleanup(req); + + atomic_sub(req->cryptlen, &engine->load); } static void mv_cesa_skcipher_std_step(struct skcipher_request *req) @@ -212,7 +215,6 @@ mv_cesa_skcipher_complete(struct crypto_async_request *req) struct mv_cesa_engine *engine = creq->base.engine; unsigned int ivsize; - atomic_sub(skreq->cryptlen, &engine->load); ivsize = crypto_skcipher_ivsize(crypto_skcipher_reqtfm(skreq)); if (mv_cesa_req_get_type(&creq->base) == CESA_DMA_REQ) { @@ -459,6 +461,9 @@ static int mv_cesa_skcipher_queue_req(struct skcipher_request *req, struct mv_cesa_skcipher_req *creq = skcipher_request_ctx(req); struct mv_cesa_engine *engine; + if (!req->cryptlen) + return 0; + ret = mv_cesa_skcipher_req_init(req, tmpl); if (ret) return ret; diff --git a/drivers/crypto/marvell/cesa/hash.c b/drivers/crypto/marvell/cesa/hash.c index f150861ceaf6..5103d36cdfdb 100644 --- a/drivers/crypto/marvell/cesa/hash.c +++ b/drivers/crypto/marvell/cesa/hash.c @@ -110,9 +110,12 @@ static inline void mv_cesa_ahash_dma_cleanup(struct ahash_request *req) static inline void mv_cesa_ahash_cleanup(struct ahash_request *req) { struct mv_cesa_ahash_req *creq = ahash_request_ctx(req); + struct mv_cesa_engine *engine = creq->base.engine; if (mv_cesa_req_get_type(&creq->base) == CESA_DMA_REQ) mv_cesa_ahash_dma_cleanup(req); + + atomic_sub(req->nbytes, &engine->load); } static void mv_cesa_ahash_last_cleanup(struct ahash_request *req) @@ -362,16 +365,13 @@ static void mv_cesa_ahash_complete(struct crypto_async_request *req) if (mv_cesa_req_get_type(&creq->base) == CESA_DMA_REQ && (creq->base.chain.last->flags & CESA_TDMA_TYPE_MSK) == CESA_TDMA_RESULT) { - __le32 *data = NULL; + const void *data; /* * Result is already in the correct endianness when the SA is * used */ data = creq->base.chain.last->op->ctx.hash.hash; - for (i = 0; i < digsize / 4; i++) - creq->state[i] = le32_to_cpu(data[i]); - memcpy(ahashreq->result, data, digsize); } else { for (i = 0; i < digsize / 4; i++) @@ -395,8 +395,6 @@ static void mv_cesa_ahash_complete(struct crypto_async_request *req) } } } - - atomic_sub(ahashreq->nbytes, &engine->load); } static void mv_cesa_ahash_prepare(struct crypto_async_request *req, @@ -663,7 +661,7 @@ static int mv_cesa_ahash_dma_req_init(struct ahash_request *req) if (ret) goto err_free_tdma; - if (iter.src.sg) { + if (iter.base.len > iter.src.op_offset) { /* * Add all the new data, inserting an operation block and * launch command between each full SRAM block-worth of diff --git a/drivers/crypto/marvell/cesa/tdma.c b/drivers/crypto/marvell/cesa/tdma.c index 388a06e180d6..243305354420 100644 --- a/drivers/crypto/marvell/cesa/tdma.c +++ b/drivers/crypto/marvell/cesa/tdma.c @@ -38,6 +38,15 @@ void mv_cesa_dma_step(struct mv_cesa_req *dreq) { struct mv_cesa_engine *engine = dreq->engine; + spin_lock_bh(&engine->lock); + if (engine->chain_sw.first == dreq->chain.first) { + engine->chain_sw.first = NULL; + engine->chain_sw.last = NULL; + } + engine->chain_hw.first = dreq->chain.first; + engine->chain_hw.last = dreq->chain.last; + spin_unlock_bh(&engine->lock); + writel_relaxed(0, engine->regs + CESA_SA_CFG); mv_cesa_set_int_mask(engine, CESA_SA_INT_ACC0_IDMA_DONE); @@ -96,25 +105,27 @@ void mv_cesa_dma_prepare(struct mv_cesa_req *dreq, void mv_cesa_tdma_chain(struct mv_cesa_engine *engine, struct mv_cesa_req *dreq) { - if (engine->chain.first == NULL && engine->chain.last == NULL) { - engine->chain.first = dreq->chain.first; - engine->chain.last = dreq->chain.last; - } else { - struct mv_cesa_tdma_desc *last; + struct mv_cesa_tdma_desc *last = engine->chain_sw.last; - last = engine->chain.last; + /* + * Break the DMA chain if the request being queued needs the IV + * regs to be set before lauching the request. + */ + if (!last || dreq->chain.first->flags & CESA_TDMA_SET_STATE) + engine->chain_sw.first = dreq->chain.first; + else { last->next = dreq->chain.first; - engine->chain.last = dreq->chain.last; - - /* - * Break the DMA chain if the CESA_TDMA_BREAK_CHAIN is set on - * the last element of the current chain, or if the request - * being queued needs the IV regs to be set before lauching - * the request. - */ - if (!(last->flags & CESA_TDMA_BREAK_CHAIN) && - !(dreq->chain.first->flags & CESA_TDMA_SET_STATE)) - last->next_dma = cpu_to_le32(dreq->chain.first->cur_dma); + last->next_dma = cpu_to_le32(dreq->chain.first->cur_dma); + } + last = dreq->chain.last; + engine->chain_sw.last = last; + /* + * Break the DMA chain if the CESA_TDMA_BREAK_CHAIN is set on + * the last element of the current chain. + */ + if (last->flags & CESA_TDMA_BREAK_CHAIN) { + engine->chain_sw.first = NULL; + engine->chain_sw.last = NULL; } } @@ -127,7 +138,7 @@ int mv_cesa_tdma_process(struct mv_cesa_engine *engine, u32 status) tdma_cur = readl(engine->regs + CESA_TDMA_CUR); - for (tdma = engine->chain.first; tdma; tdma = next) { + for (tdma = engine->chain_hw.first; tdma; tdma = next) { spin_lock_bh(&engine->lock); next = tdma->next; spin_unlock_bh(&engine->lock); @@ -149,12 +160,12 @@ int mv_cesa_tdma_process(struct mv_cesa_engine *engine, u32 status) &backlog); /* Re-chaining to the next request */ - engine->chain.first = tdma->next; + engine->chain_hw.first = tdma->next; tdma->next = NULL; /* If this is the last request, clear the chain */ - if (engine->chain.first == NULL) - engine->chain.last = NULL; + if (engine->chain_hw.first == NULL) + engine->chain_hw.last = NULL; spin_unlock_bh(&engine->lock); ctx = crypto_tfm_ctx(req->tfm); diff --git a/drivers/crypto/marvell/octeontx2/cn10k_cpt.c b/drivers/crypto/marvell/octeontx2/cn10k_cpt.c index 5cae8fafa151..d4aab9e20f2a 100644 --- a/drivers/crypto/marvell/octeontx2/cn10k_cpt.c +++ b/drivers/crypto/marvell/octeontx2/cn10k_cpt.c @@ -6,6 +6,7 @@ #include "otx2_cptvf.h" #include "otx2_cptlf.h" #include "cn10k_cpt.h" +#include "otx2_cpt_common.h" static void cn10k_cpt_send_cmd(union otx2_cpt_inst_s *cptinst, u32 insts_num, struct otx2_cptlf_info *lf); @@ -27,7 +28,7 @@ static struct cpt_hw_ops cn10k_hw_ops = { static void cn10k_cpt_send_cmd(union otx2_cpt_inst_s *cptinst, u32 insts_num, struct otx2_cptlf_info *lf) { - void __iomem *lmtline = lf->lmtline; + void *lmtline = lf->lfs->lmt_info.base + (lf->slot * LMTLINE_SIZE); u64 val = (lf->slot & 0x7FF); u64 tar_addr = 0; @@ -41,15 +42,49 @@ static void cn10k_cpt_send_cmd(union otx2_cpt_inst_s *cptinst, u32 insts_num, dma_wmb(); /* Copy CPT command to LMTLINE */ - memcpy_toio(lmtline, cptinst, insts_num * OTX2_CPT_INST_SIZE); + memcpy(lmtline, cptinst, insts_num * OTX2_CPT_INST_SIZE); cn10k_lmt_flush(val, tar_addr); } +void cn10k_cpt_lmtst_free(struct pci_dev *pdev, struct otx2_cptlfs_info *lfs) +{ + struct otx2_lmt_info *lmt_info = &lfs->lmt_info; + + if (!lmt_info->base) + return; + + dma_free_attrs(&pdev->dev, lmt_info->size, + lmt_info->base - lmt_info->align, + lmt_info->iova - lmt_info->align, + DMA_ATTR_FORCE_CONTIGUOUS); +} +EXPORT_SYMBOL_NS_GPL(cn10k_cpt_lmtst_free, "CRYPTO_DEV_OCTEONTX2_CPT"); + +static int cn10k_cpt_lmtst_alloc(struct pci_dev *pdev, + struct otx2_cptlfs_info *lfs, u32 size) +{ + struct otx2_lmt_info *lmt_info = &lfs->lmt_info; + dma_addr_t align_iova; + dma_addr_t iova; + + lmt_info->base = dma_alloc_attrs(&pdev->dev, size, &iova, GFP_KERNEL, + DMA_ATTR_FORCE_CONTIGUOUS); + if (!lmt_info->base) + return -ENOMEM; + + align_iova = ALIGN((u64)iova, LMTLINE_ALIGN); + lmt_info->iova = align_iova; + lmt_info->align = align_iova - iova; + lmt_info->size = size; + lmt_info->base += lmt_info->align; + return 0; +} + int cn10k_cptpf_lmtst_init(struct otx2_cptpf_dev *cptpf) { struct pci_dev *pdev = cptpf->pdev; - resource_size_t size; - u64 lmt_base; + u32 size; + int ret; if (!test_bit(CN10K_LMTST, &cptpf->cap_flag)) { cptpf->lfs.ops = &otx2_hw_ops; @@ -57,18 +92,19 @@ int cn10k_cptpf_lmtst_init(struct otx2_cptpf_dev *cptpf) } cptpf->lfs.ops = &cn10k_hw_ops; - lmt_base = readq(cptpf->reg_base + RVU_PF_LMTLINE_ADDR); - if (!lmt_base) { - dev_err(&pdev->dev, "PF LMTLINE address not configured\n"); - return -ENOMEM; + size = OTX2_CPT_MAX_VFS_NUM * LMTLINE_SIZE + LMTLINE_ALIGN; + ret = cn10k_cpt_lmtst_alloc(pdev, &cptpf->lfs, size); + if (ret) { + dev_err(&pdev->dev, "PF-%d LMTLINE memory allocation failed\n", + cptpf->pf_id); + return ret; } - size = pci_resource_len(pdev, PCI_MBOX_BAR_NUM); - size -= ((1 + cptpf->max_vfs) * MBOX_SIZE); - cptpf->lfs.lmt_base = devm_ioremap_wc(&pdev->dev, lmt_base, size); - if (!cptpf->lfs.lmt_base) { - dev_err(&pdev->dev, - "Mapping of PF LMTLINE address failed\n"); - return -ENOMEM; + + ret = otx2_cpt_lmtst_tbl_setup_msg(&cptpf->lfs); + if (ret) { + dev_err(&pdev->dev, "PF-%d: LMTST Table setup failed\n", + cptpf->pf_id); + cn10k_cpt_lmtst_free(pdev, &cptpf->lfs); } return 0; @@ -78,18 +114,25 @@ EXPORT_SYMBOL_NS_GPL(cn10k_cptpf_lmtst_init, "CRYPTO_DEV_OCTEONTX2_CPT"); int cn10k_cptvf_lmtst_init(struct otx2_cptvf_dev *cptvf) { struct pci_dev *pdev = cptvf->pdev; - resource_size_t offset, size; + u32 size; + int ret; if (!test_bit(CN10K_LMTST, &cptvf->cap_flag)) return 0; - offset = pci_resource_start(pdev, PCI_MBOX_BAR_NUM); - size = pci_resource_len(pdev, PCI_MBOX_BAR_NUM); - /* Map VF LMILINE region */ - cptvf->lfs.lmt_base = devm_ioremap_wc(&pdev->dev, offset, size); - if (!cptvf->lfs.lmt_base) { - dev_err(&pdev->dev, "Unable to map BAR4\n"); - return -ENOMEM; + size = cptvf->lfs.lfs_num * LMTLINE_SIZE + LMTLINE_ALIGN; + ret = cn10k_cpt_lmtst_alloc(pdev, &cptvf->lfs, size); + if (ret) { + dev_err(&pdev->dev, "VF-%d LMTLINE memory allocation failed\n", + cptvf->vf_id); + return ret; + } + + ret = otx2_cpt_lmtst_tbl_setup_msg(&cptvf->lfs); + if (ret) { + dev_err(&pdev->dev, "VF-%d: LMTST Table setup failed\n", + cptvf->vf_id); + cn10k_cpt_lmtst_free(pdev, &cptvf->lfs); } return 0; diff --git a/drivers/crypto/marvell/octeontx2/cn10k_cpt.h b/drivers/crypto/marvell/octeontx2/cn10k_cpt.h index 92be3ecf570f..ea5990048c21 100644 --- a/drivers/crypto/marvell/octeontx2/cn10k_cpt.h +++ b/drivers/crypto/marvell/octeontx2/cn10k_cpt.h @@ -50,6 +50,7 @@ static inline u8 otx2_cpt_get_uc_compcode(union otx2_cpt_res_s *result) int cn10k_cptpf_lmtst_init(struct otx2_cptpf_dev *cptpf); int cn10k_cptvf_lmtst_init(struct otx2_cptvf_dev *cptvf); +void cn10k_cpt_lmtst_free(struct pci_dev *pdev, struct otx2_cptlfs_info *lfs); void cn10k_cpt_ctx_flush(struct pci_dev *pdev, u64 cptr, bool inval); int cn10k_cpt_hw_ctx_init(struct pci_dev *pdev, struct cn10k_cpt_errata_ctx *er_ctx); diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h b/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h index c5b7c57574ef..062def303dce 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h +++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h @@ -18,9 +18,8 @@ #define OTX2_CPT_MAX_VFS_NUM 128 #define OTX2_CPT_RVU_FUNC_ADDR_S(blk, slot, offs) \ (((blk) << 20) | ((slot) << 12) | (offs)) -#define OTX2_CPT_RVU_PFFUNC(pf, func) \ - ((((pf) & RVU_PFVF_PF_MASK) << RVU_PFVF_PF_SHIFT) | \ - (((func) & RVU_PFVF_FUNC_MASK) << RVU_PFVF_FUNC_SHIFT)) + +#define OTX2_CPT_RVU_PFFUNC(pdev, pf, func) rvu_make_pcifunc(pdev, pf, func) #define OTX2_CPT_INVALID_CRYPTO_ENG_GRP 0xFF #define OTX2_CPT_NAME_LENGTH 64 @@ -145,11 +144,8 @@ static inline u64 otx2_cpt_read64(void __iomem *reg_base, u64 blk, u64 slot, static inline bool is_dev_otx2(struct pci_dev *pdev) { - if (pdev->device == OTX2_CPT_PCI_PF_DEVICE_ID || - pdev->device == OTX2_CPT_PCI_VF_DEVICE_ID) - return true; - - return false; + return pdev->device == OTX2_CPT_PCI_PF_DEVICE_ID || + pdev->device == OTX2_CPT_PCI_VF_DEVICE_ID; } static inline bool is_dev_cn10ka(struct pci_dev *pdev) @@ -159,12 +155,10 @@ static inline bool is_dev_cn10ka(struct pci_dev *pdev) static inline bool is_dev_cn10ka_ax(struct pci_dev *pdev) { - if (pdev->subsystem_device == CPT_PCI_SUBSYS_DEVID_CN10K_A && - ((pdev->revision & 0xFF) == 4 || (pdev->revision & 0xFF) == 0x50 || - (pdev->revision & 0xff) == 0x51)) - return true; - - return false; + return pdev->subsystem_device == CPT_PCI_SUBSYS_DEVID_CN10K_A && + ((pdev->revision & 0xFF) == 4 || + (pdev->revision & 0xFF) == 0x50 || + (pdev->revision & 0xFF) == 0x51); } static inline bool is_dev_cn10kb(struct pci_dev *pdev) @@ -174,11 +168,8 @@ static inline bool is_dev_cn10kb(struct pci_dev *pdev) static inline bool is_dev_cn10ka_b0(struct pci_dev *pdev) { - if (pdev->subsystem_device == CPT_PCI_SUBSYS_DEVID_CN10K_A && - (pdev->revision & 0xFF) == 0x54) - return true; - - return false; + return pdev->subsystem_device == CPT_PCI_SUBSYS_DEVID_CN10K_A && + (pdev->revision & 0xFF) == 0x54; } static inline void otx2_cpt_set_hw_caps(struct pci_dev *pdev, @@ -192,18 +183,12 @@ static inline void otx2_cpt_set_hw_caps(struct pci_dev *pdev, static inline bool cpt_is_errata_38550_exists(struct pci_dev *pdev) { - if (is_dev_otx2(pdev) || is_dev_cn10ka_ax(pdev)) - return true; - - return false; + return is_dev_otx2(pdev) || is_dev_cn10ka_ax(pdev); } static inline bool cpt_feature_sgv2(struct pci_dev *pdev) { - if (!is_dev_otx2(pdev) && !is_dev_cn10ka_ax(pdev)) - return true; - - return false; + return !is_dev_otx2(pdev) && !is_dev_cn10ka_ax(pdev); } int otx2_cpt_send_ready_msg(struct otx2_mbox *mbox, struct pci_dev *pdev); @@ -223,5 +208,6 @@ int otx2_cpt_detach_rsrcs_msg(struct otx2_cptlfs_info *lfs); int otx2_cpt_msix_offset_msg(struct otx2_cptlfs_info *lfs); int otx2_cpt_sync_mbox_msg(struct otx2_mbox *mbox); int otx2_cpt_lf_reset_msg(struct otx2_cptlfs_info *lfs, int slot); +int otx2_cpt_lmtst_tbl_setup_msg(struct otx2_cptlfs_info *lfs); #endif /* __OTX2_CPT_COMMON_H */ diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c b/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c index b8b7c8a3c0ca..95f3de3a34eb 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c +++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c @@ -255,3 +255,28 @@ int otx2_cpt_lf_reset_msg(struct otx2_cptlfs_info *lfs, int slot) return ret; } EXPORT_SYMBOL_NS_GPL(otx2_cpt_lf_reset_msg, "CRYPTO_DEV_OCTEONTX2_CPT"); + +int otx2_cpt_lmtst_tbl_setup_msg(struct otx2_cptlfs_info *lfs) +{ + struct otx2_mbox *mbox = lfs->mbox; + struct pci_dev *pdev = lfs->pdev; + struct lmtst_tbl_setup_req *req; + + req = (struct lmtst_tbl_setup_req *) + otx2_mbox_alloc_msg_rsp(mbox, 0, sizeof(*req), + sizeof(struct msg_rsp)); + if (!req) { + dev_err(&pdev->dev, "RVU MBOX failed to alloc message.\n"); + return -EFAULT; + } + + req->hdr.id = MBOX_MSG_LMTST_TBL_SETUP; + req->hdr.sig = OTX2_MBOX_REQ_SIG; + req->hdr.pcifunc = 0; + + req->use_local_lmt_region = true; + req->lmt_iova = lfs->lmt_info.iova; + + return otx2_cpt_send_mbox_msg(mbox, pdev); +} +EXPORT_SYMBOL_NS_GPL(otx2_cpt_lmtst_tbl_setup_msg, "CRYPTO_DEV_OCTEONTX2_CPT"); diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_reqmgr.h b/drivers/crypto/marvell/octeontx2/otx2_cpt_reqmgr.h index e27e849b01df..e64ca30335de 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cpt_reqmgr.h +++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_reqmgr.h @@ -34,6 +34,9 @@ #define SG_COMP_2 2 #define SG_COMP_1 1 +#define OTX2_CPT_DPTR_RPTR_ALIGN 8 +#define OTX2_CPT_RES_ADDR_ALIGN 32 + union otx2_cpt_opcode { u16 flags; struct { @@ -347,22 +350,48 @@ static inline struct otx2_cpt_inst_info * cn10k_sgv2_info_create(struct pci_dev *pdev, struct otx2_cpt_req_info *req, gfp_t gfp) { - u32 dlen = 0, g_len, sg_len, info_len; - int align = OTX2_CPT_DMA_MINALIGN; + u32 dlen = 0, g_len, s_len, sg_len, info_len; struct otx2_cpt_inst_info *info; - u16 g_sz_bytes, s_sz_bytes; u32 total_mem_len; int i; - g_sz_bytes = ((req->in_cnt + 2) / 3) * - sizeof(struct cn10kb_cpt_sglist_component); - s_sz_bytes = ((req->out_cnt + 2) / 3) * - sizeof(struct cn10kb_cpt_sglist_component); + /* Allocate memory to meet below alignment requirement: + * ------------------------------------ + * | struct otx2_cpt_inst_info | + * | (No alignment required) | + * | --------------------------------| + * | | padding for ARCH_DMA_MINALIGN | + * | | alignment | + * |------------------------------------| + * | SG List Gather/Input memory | + * | Length = multiple of 32Bytes | + * | Alignment = 8Byte | + * |---------------------------------- | + * | SG List Scatter/Output memory | + * | Length = multiple of 32Bytes | + * | Alignment = 8Byte | + * | -------------------------------| + * | | padding for 32B alignment | + * |------------------------------------| + * | Result response memory | + * | Alignment = 32Byte | + * ------------------------------------ + */ + + info_len = sizeof(*info); + + g_len = ((req->in_cnt + 2) / 3) * + sizeof(struct cn10kb_cpt_sglist_component); + s_len = ((req->out_cnt + 2) / 3) * + sizeof(struct cn10kb_cpt_sglist_component); + sg_len = g_len + s_len; - g_len = ALIGN(g_sz_bytes, align); - sg_len = ALIGN(g_len + s_sz_bytes, align); - info_len = ALIGN(sizeof(*info), align); - total_mem_len = sg_len + info_len + sizeof(union otx2_cpt_res_s); + /* Allocate extra memory for SG and response address alignment */ + total_mem_len = ALIGN(info_len, OTX2_CPT_DPTR_RPTR_ALIGN); + total_mem_len += (ARCH_DMA_MINALIGN - 1) & + ~(OTX2_CPT_DPTR_RPTR_ALIGN - 1); + total_mem_len += ALIGN(sg_len, OTX2_CPT_RES_ADDR_ALIGN); + total_mem_len += sizeof(union otx2_cpt_res_s); info = kzalloc(total_mem_len, gfp); if (unlikely(!info)) @@ -372,7 +401,8 @@ cn10k_sgv2_info_create(struct pci_dev *pdev, struct otx2_cpt_req_info *req, dlen += req->in[i].size; info->dlen = dlen; - info->in_buffer = (u8 *)info + info_len; + info->in_buffer = PTR_ALIGN((u8 *)info + info_len, ARCH_DMA_MINALIGN); + info->out_buffer = info->in_buffer + g_len; info->gthr_sz = req->in_cnt; info->sctr_sz = req->out_cnt; @@ -384,7 +414,7 @@ cn10k_sgv2_info_create(struct pci_dev *pdev, struct otx2_cpt_req_info *req, } if (sgv2io_components_setup(pdev, req->out, req->out_cnt, - &info->in_buffer[g_len])) { + info->out_buffer)) { dev_err(&pdev->dev, "Failed to setup scatter list\n"); goto destroy_info; } @@ -401,8 +431,10 @@ cn10k_sgv2_info_create(struct pci_dev *pdev, struct otx2_cpt_req_info *req, * Get buffer for union otx2_cpt_res_s response * structure and its physical address */ - info->completion_addr = info->in_buffer + sg_len; - info->comp_baddr = info->dptr_baddr + sg_len; + info->completion_addr = PTR_ALIGN((info->in_buffer + sg_len), + OTX2_CPT_RES_ADDR_ALIGN); + info->comp_baddr = ALIGN((info->dptr_baddr + sg_len), + OTX2_CPT_RES_ADDR_ALIGN); return info; @@ -417,10 +449,9 @@ static inline struct otx2_cpt_inst_info * otx2_sg_info_create(struct pci_dev *pdev, struct otx2_cpt_req_info *req, gfp_t gfp) { - int align = OTX2_CPT_DMA_MINALIGN; struct otx2_cpt_inst_info *info; - u32 dlen, align_dlen, info_len; - u16 g_sz_bytes, s_sz_bytes; + u32 dlen, info_len; + u16 g_len, s_len; u32 total_mem_len; if (unlikely(req->in_cnt > OTX2_CPT_MAX_SG_IN_CNT || @@ -429,22 +460,54 @@ otx2_sg_info_create(struct pci_dev *pdev, struct otx2_cpt_req_info *req, return NULL; } - g_sz_bytes = ((req->in_cnt + 3) / 4) * - sizeof(struct otx2_cpt_sglist_component); - s_sz_bytes = ((req->out_cnt + 3) / 4) * - sizeof(struct otx2_cpt_sglist_component); + /* Allocate memory to meet below alignment requirement: + * ------------------------------------ + * | struct otx2_cpt_inst_info | + * | (No alignment required) | + * | --------------------------------| + * | | padding for ARCH_DMA_MINALIGN | + * | | alignment | + * |------------------------------------| + * | SG List Header of 8 Byte | + * |------------------------------------| + * | SG List Gather/Input memory | + * | Length = multiple of 32Bytes | + * | Alignment = 8Byte | + * |---------------------------------- | + * | SG List Scatter/Output memory | + * | Length = multiple of 32Bytes | + * | Alignment = 8Byte | + * | -------------------------------| + * | | padding for 32B alignment | + * |------------------------------------| + * | Result response memory | + * | Alignment = 32Byte | + * ------------------------------------ + */ + + info_len = sizeof(*info); + + g_len = ((req->in_cnt + 3) / 4) * + sizeof(struct otx2_cpt_sglist_component); + s_len = ((req->out_cnt + 3) / 4) * + sizeof(struct otx2_cpt_sglist_component); + + dlen = g_len + s_len + SG_LIST_HDR_SIZE; - dlen = g_sz_bytes + s_sz_bytes + SG_LIST_HDR_SIZE; - align_dlen = ALIGN(dlen, align); - info_len = ALIGN(sizeof(*info), align); - total_mem_len = align_dlen + info_len + sizeof(union otx2_cpt_res_s); + /* Allocate extra memory for SG and response address alignment */ + total_mem_len = ALIGN(info_len, OTX2_CPT_DPTR_RPTR_ALIGN); + total_mem_len += (ARCH_DMA_MINALIGN - 1) & + ~(OTX2_CPT_DPTR_RPTR_ALIGN - 1); + total_mem_len += ALIGN(dlen, OTX2_CPT_RES_ADDR_ALIGN); + total_mem_len += sizeof(union otx2_cpt_res_s); info = kzalloc(total_mem_len, gfp); if (unlikely(!info)) return NULL; info->dlen = dlen; - info->in_buffer = (u8 *)info + info_len; + info->in_buffer = PTR_ALIGN((u8 *)info + info_len, ARCH_DMA_MINALIGN); + info->out_buffer = info->in_buffer + SG_LIST_HDR_SIZE + g_len; ((u16 *)info->in_buffer)[0] = req->out_cnt; ((u16 *)info->in_buffer)[1] = req->in_cnt; @@ -460,7 +523,7 @@ otx2_sg_info_create(struct pci_dev *pdev, struct otx2_cpt_req_info *req, } if (setup_sgio_components(pdev, req->out, req->out_cnt, - &info->in_buffer[8 + g_sz_bytes])) { + info->out_buffer)) { dev_err(&pdev->dev, "Failed to setup scatter list\n"); goto destroy_info; } @@ -476,8 +539,10 @@ otx2_sg_info_create(struct pci_dev *pdev, struct otx2_cpt_req_info *req, * Get buffer for union otx2_cpt_res_s response * structure and its physical address */ - info->completion_addr = info->in_buffer + align_dlen; - info->comp_baddr = info->dptr_baddr + align_dlen; + info->completion_addr = PTR_ALIGN((info->in_buffer + dlen), + OTX2_CPT_RES_ADDR_ALIGN); + info->comp_baddr = ALIGN((info->dptr_baddr + dlen), + OTX2_CPT_RES_ADDR_ALIGN); return info; @@ -490,6 +555,7 @@ struct otx2_cptlf_wqe; int otx2_cpt_do_request(struct pci_dev *pdev, struct otx2_cpt_req_info *req, int cpu_num); void otx2_cpt_post_process(struct otx2_cptlf_wqe *wqe); -int otx2_cpt_get_kcrypto_eng_grp_num(struct pci_dev *pdev); +int otx2_cpt_get_eng_grp_num(struct pci_dev *pdev, + enum otx2_cpt_eng_type); #endif /* __OTX2_CPT_REQMGR_H */ diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptlf.c b/drivers/crypto/marvell/octeontx2/otx2_cptlf.c index b5d66afcc030..dc7c7a2650a5 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cptlf.c +++ b/drivers/crypto/marvell/octeontx2/otx2_cptlf.c @@ -433,10 +433,7 @@ int otx2_cptlf_init(struct otx2_cptlfs_info *lfs, u8 eng_grp_mask, int pri, for (slot = 0; slot < lfs->lfs_num; slot++) { lfs->lf[slot].lfs = lfs; lfs->lf[slot].slot = slot; - if (lfs->lmt_base) - lfs->lf[slot].lmtline = lfs->lmt_base + - (slot * LMTLINE_SIZE); - else + if (!lfs->lmt_info.base) lfs->lf[slot].lmtline = lfs->reg_base + OTX2_CPT_RVU_FUNC_ADDR_S(BLKADDR_LMT, slot, OTX2_CPT_LMT_LF_LMTLINEX(0)); diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptlf.h b/drivers/crypto/marvell/octeontx2/otx2_cptlf.h index bd8604be2952..1b9f75214d18 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cptlf.h +++ b/drivers/crypto/marvell/octeontx2/otx2_cptlf.h @@ -105,18 +105,27 @@ struct cpt_hw_ops { gfp_t gfp); }; +#define LMTLINE_SIZE 128 +#define LMTLINE_ALIGN 128 +struct otx2_lmt_info { + void *base; + dma_addr_t iova; + u32 size; + u8 align; +}; + struct otx2_cptlfs_info { /* Registers start address of VF/PF LFs are attached to */ void __iomem *reg_base; -#define LMTLINE_SIZE 128 - void __iomem *lmt_base; + struct otx2_lmt_info lmt_info; struct pci_dev *pdev; /* Device LFs are attached to */ struct otx2_cptlf_info lf[OTX2_CPT_MAX_LFS_NUM]; struct otx2_mbox *mbox; struct cpt_hw_ops *ops; u8 are_lfs_attached; /* Whether CPT LFs are attached */ u8 lfs_num; /* Number of CPT LFs */ - u8 kcrypto_eng_grp_num; /* Kernel crypto engine group number */ + u8 kcrypto_se_eng_grp_num; /* Crypto symmetric engine group number */ + u8 kcrypto_ae_eng_grp_num; /* Crypto asymmetric engine group number */ u8 kvf_limits; /* Kernel crypto limits */ atomic_t state; /* LF's state. started/reset */ int blkaddr; /* CPT blkaddr: BLKADDR_CPT0/BLKADDR_CPT1 */ diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c b/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c index 12971300296d..1c5c262af48d 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c +++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c @@ -639,6 +639,12 @@ static int cptpf_device_init(struct otx2_cptpf_dev *cptpf) /* Disable all cores */ ret = otx2_cpt_disable_all_cores(cptpf); + otx2_cptlf_set_dev_info(&cptpf->lfs, cptpf->pdev, cptpf->reg_base, + &cptpf->afpf_mbox, BLKADDR_CPT0); + if (cptpf->has_cpt1) + otx2_cptlf_set_dev_info(&cptpf->cpt1_lfs, cptpf->pdev, + cptpf->reg_base, &cptpf->afpf_mbox, + BLKADDR_CPT1); return ret; } @@ -786,19 +792,19 @@ static int otx2_cptpf_probe(struct pci_dev *pdev, cptpf->max_vfs = pci_sriov_get_totalvfs(pdev); cptpf->kvf_limits = 1; - err = cn10k_cptpf_lmtst_init(cptpf); + /* Initialize CPT PF device */ + err = cptpf_device_init(cptpf); if (err) goto unregister_intr; - /* Initialize CPT PF device */ - err = cptpf_device_init(cptpf); + err = cn10k_cptpf_lmtst_init(cptpf); if (err) goto unregister_intr; /* Initialize engine groups */ err = otx2_cpt_init_eng_grps(pdev, &cptpf->eng_grps); if (err) - goto unregister_intr; + goto free_lmtst; err = sysfs_create_group(&dev->kobj, &cptpf_sysfs_group); if (err) @@ -814,6 +820,8 @@ sysfs_grp_del: sysfs_remove_group(&dev->kobj, &cptpf_sysfs_group); cleanup_eng_grps: otx2_cpt_cleanup_eng_grps(pdev, &cptpf->eng_grps); +free_lmtst: + cn10k_cpt_lmtst_free(pdev, &cptpf->lfs); unregister_intr: cptpf_disable_afpf_mbox_intr(cptpf); destroy_afpf_mbox: @@ -848,6 +856,8 @@ static void otx2_cptpf_remove(struct pci_dev *pdev) cptpf_disable_afpf_mbox_intr(cptpf); /* Destroy AF-PF mbox */ cptpf_afpf_mbox_destroy(cptpf); + /* Free LMTST memory */ + cn10k_cpt_lmtst_free(pdev, &cptpf->lfs); pci_set_drvdata(pdev, NULL); } diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c b/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c index ec1ac7e836a3..b4b2d3d1cbc2 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c +++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c @@ -142,7 +142,7 @@ static int send_inline_ipsec_inbound_msg(struct otx2_cptpf_dev *cptpf, memset(req, 0, sizeof(*req)); req->hdr.id = MBOX_MSG_CPT_INLINE_IPSEC_CFG; req->hdr.sig = OTX2_MBOX_REQ_SIG; - req->hdr.pcifunc = OTX2_CPT_RVU_PFFUNC(cptpf->pf_id, 0); + req->hdr.pcifunc = OTX2_CPT_RVU_PFFUNC(cptpf->pdev, cptpf->pf_id, 0); req->dir = CPT_INLINE_INBOUND; req->slot = slot; req->sso_pf_func_ovrd = cptpf->sso_pf_func_ovrd; @@ -184,7 +184,8 @@ static int rx_inline_ipsec_lf_cfg(struct otx2_cptpf_dev *cptpf, u8 egrp, nix_req->gen_cfg.opcode = cpt_inline_rx_opcode(pdev); nix_req->gen_cfg.param1 = req->param1; nix_req->gen_cfg.param2 = req->param2; - nix_req->inst_qsel.cpt_pf_func = OTX2_CPT_RVU_PFFUNC(cptpf->pf_id, 0); + nix_req->inst_qsel.cpt_pf_func = + OTX2_CPT_RVU_PFFUNC(cptpf->pdev, cptpf->pf_id, 0); nix_req->inst_qsel.cpt_slot = 0; ret = otx2_cpt_send_mbox_msg(&cptpf->afpf_mbox, pdev); if (ret) @@ -264,8 +265,6 @@ static int handle_msg_rx_inline_ipsec_lf_cfg(struct otx2_cptpf_dev *cptpf, return -ENOENT; } - otx2_cptlf_set_dev_info(&cptpf->lfs, cptpf->pdev, cptpf->reg_base, - &cptpf->afpf_mbox, BLKADDR_CPT0); cptpf->lfs.global_slot = 0; cptpf->lfs.ctx_ilen_ovrd = cfg_req->ctx_ilen_valid; cptpf->lfs.ctx_ilen = cfg_req->ctx_ilen; @@ -278,9 +277,6 @@ static int handle_msg_rx_inline_ipsec_lf_cfg(struct otx2_cptpf_dev *cptpf, if (cptpf->has_cpt1) { cptpf->rsrc_req_blkaddr = BLKADDR_CPT1; - otx2_cptlf_set_dev_info(&cptpf->cpt1_lfs, cptpf->pdev, - cptpf->reg_base, &cptpf->afpf_mbox, - BLKADDR_CPT1); cptpf->cpt1_lfs.global_slot = num_lfs; cptpf->cpt1_lfs.ctx_ilen_ovrd = cfg_req->ctx_ilen_valid; cptpf->cpt1_lfs.ctx_ilen = cfg_req->ctx_ilen; @@ -397,9 +393,8 @@ void otx2_cptpf_vfpf_mbox_handler(struct work_struct *work) msg = (struct mbox_msghdr *)(mdev->mbase + offset); /* Set which VF sent this message based on mbox IRQ */ - msg->pcifunc = ((u16)cptpf->pf_id << RVU_PFVF_PF_SHIFT) | - ((vf->vf_id + 1) & RVU_PFVF_FUNC_MASK); - + msg->pcifunc = rvu_make_pcifunc(cptpf->pdev, cptpf->pf_id, + (vf->vf_id + 1)); err = cptpf_handle_vf_req(cptpf, vf, msg, msg->next_msgoff - offset); /* @@ -474,8 +469,7 @@ static void process_afpf_mbox_msg(struct otx2_cptpf_dev *cptpf, switch (msg->id) { case MBOX_MSG_READY: - cptpf->pf_id = (msg->pcifunc >> RVU_PFVF_PF_SHIFT) & - RVU_PFVF_PF_MASK; + cptpf->pf_id = rvu_get_pf(cptpf->pdev, msg->pcifunc); break; case MBOX_MSG_MSIX_OFFSET: rsp_msix = (struct msix_offset_rsp *) msg; @@ -507,6 +501,7 @@ static void process_afpf_mbox_msg(struct otx2_cptpf_dev *cptpf, case MBOX_MSG_CPT_INLINE_IPSEC_CFG: case MBOX_MSG_NIX_INLINE_IPSEC_CFG: case MBOX_MSG_CPT_LF_RESET: + case MBOX_MSG_LMTST_TBL_SETUP: break; default: diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c b/drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c index 42c5484ce66a..cc47e361089a 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c +++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c @@ -176,7 +176,9 @@ static int cptx_set_ucode_base(struct otx2_cpt_eng_grp_info *eng_grp, /* Set PF number for microcode fetches */ ret = otx2_cpt_write_af_reg(&cptpf->afpf_mbox, cptpf->pdev, CPT_AF_PF_FUNC, - cptpf->pf_id << RVU_PFVF_PF_SHIFT, blkaddr); + rvu_make_pcifunc(cptpf->pdev, + cptpf->pf_id, 0), + blkaddr); if (ret) return ret; @@ -1491,11 +1493,13 @@ int otx2_cpt_discover_eng_capabilities(struct otx2_cptpf_dev *cptpf) union otx2_cpt_opcode opcode; union otx2_cpt_res_s *result; union otx2_cpt_inst_s inst; + dma_addr_t result_baddr; dma_addr_t rptr_baddr; struct pci_dev *pdev; - u32 len, compl_rlen; + int timeout = 10000; + void *base, *rptr; int ret, etype; - void *rptr; + u32 len; /* * We don't get capabilities if it was already done @@ -1513,29 +1517,33 @@ int otx2_cpt_discover_eng_capabilities(struct otx2_cptpf_dev *cptpf) if (ret) goto delete_grps; - otx2_cptlf_set_dev_info(lfs, cptpf->pdev, cptpf->reg_base, - &cptpf->afpf_mbox, BLKADDR_CPT0); ret = otx2_cptlf_init(lfs, OTX2_CPT_ALL_ENG_GRPS_MASK, OTX2_CPT_QUEUE_HI_PRIO, 1); if (ret) goto delete_grps; - compl_rlen = ALIGN(sizeof(union otx2_cpt_res_s), OTX2_CPT_DMA_MINALIGN); - len = compl_rlen + LOADFVC_RLEN; + /* Allocate extra memory for "rptr" and "result" pointer alignment */ + len = LOADFVC_RLEN + ARCH_DMA_MINALIGN + + sizeof(union otx2_cpt_res_s) + OTX2_CPT_RES_ADDR_ALIGN; - result = kzalloc(len, GFP_KERNEL); - if (!result) { + base = kzalloc(len, GFP_KERNEL); + if (!base) { ret = -ENOMEM; goto lf_cleanup; } - rptr_baddr = dma_map_single(&pdev->dev, (void *)result, len, - DMA_BIDIRECTIONAL); + + rptr = PTR_ALIGN(base, ARCH_DMA_MINALIGN); + rptr_baddr = dma_map_single(&pdev->dev, rptr, len, DMA_BIDIRECTIONAL); if (dma_mapping_error(&pdev->dev, rptr_baddr)) { dev_err(&pdev->dev, "DMA mapping failed\n"); ret = -EFAULT; - goto free_result; + goto free_rptr; } - rptr = (u8 *)result + compl_rlen; + + result = (union otx2_cpt_res_s *)PTR_ALIGN(rptr + LOADFVC_RLEN, + OTX2_CPT_RES_ADDR_ALIGN); + result_baddr = ALIGN(rptr_baddr + LOADFVC_RLEN, + OTX2_CPT_RES_ADDR_ALIGN); /* Fill in the command */ opcode.s.major = LOADFVC_MAJOR_OP; @@ -1547,27 +1555,38 @@ int otx2_cpt_discover_eng_capabilities(struct otx2_cptpf_dev *cptpf) /* 64-bit swap for microcode data reads, not needed for addresses */ cpu_to_be64s(&iq_cmd.cmd.u); iq_cmd.dptr = 0; - iq_cmd.rptr = rptr_baddr + compl_rlen; + iq_cmd.rptr = rptr_baddr; iq_cmd.cptr.u = 0; for (etype = 1; etype < OTX2_CPT_MAX_ENG_TYPES; etype++) { result->s.compcode = OTX2_CPT_COMPLETION_CODE_INIT; iq_cmd.cptr.s.grp = otx2_cpt_get_eng_grp(&cptpf->eng_grps, etype); - otx2_cpt_fill_inst(&inst, &iq_cmd, rptr_baddr); + otx2_cpt_fill_inst(&inst, &iq_cmd, result_baddr); lfs->ops->send_cmd(&inst, 1, &cptpf->lfs.lf[0]); + timeout = 10000; while (lfs->ops->cpt_get_compcode(result) == - OTX2_CPT_COMPLETION_CODE_INIT) + OTX2_CPT_COMPLETION_CODE_INIT) { cpu_relax(); + udelay(1); + timeout--; + if (!timeout) { + ret = -ENODEV; + cptpf->is_eng_caps_discovered = false; + dev_warn(&pdev->dev, "Timeout on CPT load_fvc completion poll\n"); + goto error_no_response; + } + } cptpf->eng_caps[etype].u = be64_to_cpup(rptr); } - dma_unmap_single(&pdev->dev, rptr_baddr, len, DMA_BIDIRECTIONAL); cptpf->is_eng_caps_discovered = true; -free_result: - kfree(result); +error_no_response: + dma_unmap_single(&pdev->dev, rptr_baddr, len, DMA_BIDIRECTIONAL); +free_rptr: + kfree(base); lf_cleanup: otx2_cptlf_shutdown(lfs); delete_grps: diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.c b/drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.c index 7eb0bc13994d..8d9f394d6b50 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.c +++ b/drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.c @@ -384,7 +384,8 @@ static inline int cpt_enc_dec(struct skcipher_request *req, u32 enc) req_info->req_type = OTX2_CPT_ENC_DEC_REQ; req_info->is_enc = enc; req_info->is_trunc_hmac = false; - req_info->ctrl.s.grp = otx2_cpt_get_kcrypto_eng_grp_num(pdev); + req_info->ctrl.s.grp = otx2_cpt_get_eng_grp_num(pdev, + OTX2_CPT_SE_TYPES); req_info->req.cptr = ctx->er_ctx.hw_ctx; req_info->req.cptr_dma = ctx->er_ctx.cptr_dma; @@ -1288,7 +1289,8 @@ static int cpt_aead_enc_dec(struct aead_request *req, u8 reg_type, u8 enc) if (status) return status; - req_info->ctrl.s.grp = otx2_cpt_get_kcrypto_eng_grp_num(pdev); + req_info->ctrl.s.grp = otx2_cpt_get_eng_grp_num(pdev, + OTX2_CPT_SE_TYPES); /* * We perform an asynchronous send and once diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c b/drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c index d84eebdf2fa8..c1c44a7b89fa 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c +++ b/drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c @@ -265,17 +265,33 @@ static int cptvf_lf_init(struct otx2_cptvf_dev *cptvf) u8 eng_grp_msk; /* Get engine group number for symmetric crypto */ - cptvf->lfs.kcrypto_eng_grp_num = OTX2_CPT_INVALID_CRYPTO_ENG_GRP; + cptvf->lfs.kcrypto_se_eng_grp_num = OTX2_CPT_INVALID_CRYPTO_ENG_GRP; ret = otx2_cptvf_send_eng_grp_num_msg(cptvf, OTX2_CPT_SE_TYPES); if (ret) return ret; - if (cptvf->lfs.kcrypto_eng_grp_num == OTX2_CPT_INVALID_CRYPTO_ENG_GRP) { - dev_err(dev, "Engine group for kernel crypto not available\n"); - ret = -ENOENT; + if (cptvf->lfs.kcrypto_se_eng_grp_num == + OTX2_CPT_INVALID_CRYPTO_ENG_GRP) { + dev_err(dev, + "Symmetric Engine group for crypto not available\n"); + return -ENOENT; + } + + /* Get engine group number for asymmetric crypto */ + cptvf->lfs.kcrypto_ae_eng_grp_num = OTX2_CPT_INVALID_CRYPTO_ENG_GRP; + ret = otx2_cptvf_send_eng_grp_num_msg(cptvf, OTX2_CPT_AE_TYPES); + if (ret) return ret; + + if (cptvf->lfs.kcrypto_ae_eng_grp_num == + OTX2_CPT_INVALID_CRYPTO_ENG_GRP) { + dev_err(dev, + "Asymmetric Engine group for crypto not available\n"); + return -ENOENT; } - eng_grp_msk = 1 << cptvf->lfs.kcrypto_eng_grp_num; + + eng_grp_msk = BIT(cptvf->lfs.kcrypto_se_eng_grp_num) | + BIT(cptvf->lfs.kcrypto_ae_eng_grp_num); ret = otx2_cptvf_send_kvf_limits_msg(cptvf); if (ret) @@ -283,8 +299,6 @@ static int cptvf_lf_init(struct otx2_cptvf_dev *cptvf) lfs_num = cptvf->lfs.kvf_limits; - otx2_cptlf_set_dev_info(lfs, cptvf->pdev, cptvf->reg_base, - &cptvf->pfvf_mbox, cptvf->blkaddr); ret = otx2_cptlf_init(lfs, eng_grp_msk, OTX2_CPT_QUEUE_HI_PRIO, lfs_num); if (ret) @@ -378,10 +392,6 @@ static int otx2_cptvf_probe(struct pci_dev *pdev, otx2_cpt_set_hw_caps(pdev, &cptvf->cap_flag); - ret = cn10k_cptvf_lmtst_init(cptvf); - if (ret) - goto clear_drvdata; - /* Initialize PF<=>VF mailbox */ ret = cptvf_pfvf_mbox_init(cptvf); if (ret) @@ -396,6 +406,9 @@ static int otx2_cptvf_probe(struct pci_dev *pdev, cptvf_hw_ops_get(cptvf); + otx2_cptlf_set_dev_info(&cptvf->lfs, cptvf->pdev, cptvf->reg_base, + &cptvf->pfvf_mbox, cptvf->blkaddr); + ret = otx2_cptvf_send_caps_msg(cptvf); if (ret) { dev_err(&pdev->dev, "Couldn't get CPT engine capabilities.\n"); @@ -404,13 +417,19 @@ static int otx2_cptvf_probe(struct pci_dev *pdev, if (cptvf->eng_caps[OTX2_CPT_SE_TYPES] & BIT_ULL(35)) cptvf->lfs.ops->cpt_sg_info_create = cn10k_sgv2_info_create; + ret = cn10k_cptvf_lmtst_init(cptvf); + if (ret) + goto unregister_interrupts; + /* Initialize CPT LFs */ ret = cptvf_lf_init(cptvf); if (ret) - goto unregister_interrupts; + goto free_lmtst; return 0; +free_lmtst: + cn10k_cpt_lmtst_free(pdev, &cptvf->lfs); unregister_interrupts: cptvf_disable_pfvf_mbox_intrs(cptvf); destroy_pfvf_mbox: @@ -434,6 +453,8 @@ static void otx2_cptvf_remove(struct pci_dev *pdev) cptvf_disable_pfvf_mbox_intrs(cptvf); /* Destroy PF-VF mbox */ cptvf_pfvf_mbox_destroy(cptvf); + /* Free LMTST memory */ + cn10k_cpt_lmtst_free(pdev, &cptvf->lfs); pci_set_drvdata(pdev, NULL); } diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c b/drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c index d9fa5f6e204d..5277bcfa275e 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c +++ b/drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c @@ -75,6 +75,7 @@ static void process_pfvf_mbox_mbox_msg(struct otx2_cptvf_dev *cptvf, struct otx2_cpt_caps_rsp *eng_caps; struct cpt_rd_wr_reg_msg *rsp_reg; struct msix_offset_rsp *rsp_msix; + u8 grp_num; int i; if (msg->id >= MBOX_MSG_MAX) { @@ -122,7 +123,11 @@ static void process_pfvf_mbox_mbox_msg(struct otx2_cptvf_dev *cptvf, break; case MBOX_MSG_GET_ENG_GRP_NUM: rsp_grp = (struct otx2_cpt_egrp_num_rsp *) msg; - cptvf->lfs.kcrypto_eng_grp_num = rsp_grp->eng_grp_num; + grp_num = rsp_grp->eng_grp_num; + if (rsp_grp->eng_type == OTX2_CPT_SE_TYPES) + cptvf->lfs.kcrypto_se_eng_grp_num = grp_num; + else if (rsp_grp->eng_type == OTX2_CPT_AE_TYPES) + cptvf->lfs.kcrypto_ae_eng_grp_num = grp_num; break; case MBOX_MSG_GET_KVF_LIMITS: rsp_limits = (struct otx2_cpt_kvf_limits_rsp *) msg; @@ -134,6 +139,7 @@ static void process_pfvf_mbox_mbox_msg(struct otx2_cptvf_dev *cptvf, sizeof(cptvf->eng_caps)); break; case MBOX_MSG_CPT_LF_RESET: + case MBOX_MSG_LMTST_TBL_SETUP: break; default: dev_err(&cptvf->pdev->dev, "Unsupported msg %d received.\n", @@ -188,7 +194,7 @@ int otx2_cptvf_send_eng_grp_num_msg(struct otx2_cptvf_dev *cptvf, int eng_type) } req->hdr.id = MBOX_MSG_GET_ENG_GRP_NUM; req->hdr.sig = OTX2_MBOX_REQ_SIG; - req->hdr.pcifunc = OTX2_CPT_RVU_PFFUNC(cptvf->vf_id, 0); + req->hdr.pcifunc = OTX2_CPT_RVU_PFFUNC(cptvf->pdev, cptvf->vf_id, 0); req->eng_type = eng_type; return otx2_cpt_send_mbox_msg(mbox, pdev); @@ -209,7 +215,7 @@ int otx2_cptvf_send_kvf_limits_msg(struct otx2_cptvf_dev *cptvf) } req->id = MBOX_MSG_GET_KVF_LIMITS; req->sig = OTX2_MBOX_REQ_SIG; - req->pcifunc = OTX2_CPT_RVU_PFFUNC(cptvf->vf_id, 0); + req->pcifunc = OTX2_CPT_RVU_PFFUNC(cptvf->pdev, cptvf->vf_id, 0); return otx2_cpt_send_mbox_msg(mbox, pdev); } @@ -229,7 +235,7 @@ int otx2_cptvf_send_caps_msg(struct otx2_cptvf_dev *cptvf) } req->id = MBOX_MSG_GET_CAPS; req->sig = OTX2_MBOX_REQ_SIG; - req->pcifunc = OTX2_CPT_RVU_PFFUNC(cptvf->vf_id, 0); + req->pcifunc = OTX2_CPT_RVU_PFFUNC(cptvf->pdev, cptvf->vf_id, 0); return otx2_cpt_send_mbox_msg(mbox, pdev); } diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptvf_reqmgr.c b/drivers/crypto/marvell/octeontx2/otx2_cptvf_reqmgr.c index 426244107037..e71494486c64 100644 --- a/drivers/crypto/marvell/octeontx2/otx2_cptvf_reqmgr.c +++ b/drivers/crypto/marvell/octeontx2/otx2_cptvf_reqmgr.c @@ -391,9 +391,19 @@ void otx2_cpt_post_process(struct otx2_cptlf_wqe *wqe) &wqe->lfs->lf[wqe->lf_num].pqueue); } -int otx2_cpt_get_kcrypto_eng_grp_num(struct pci_dev *pdev) +int otx2_cpt_get_eng_grp_num(struct pci_dev *pdev, + enum otx2_cpt_eng_type eng_type) { struct otx2_cptvf_dev *cptvf = pci_get_drvdata(pdev); - return cptvf->lfs.kcrypto_eng_grp_num; + switch (eng_type) { + case OTX2_CPT_SE_TYPES: + return cptvf->lfs.kcrypto_se_eng_grp_num; + case OTX2_CPT_AE_TYPES: + return cptvf->lfs.kcrypto_ae_eng_grp_num; + default: + dev_err(&cptvf->pdev->dev, "Unsupported engine type"); + break; + } + return -ENXIO; } |