diff options
Diffstat (limited to 'drivers/scsi/megaraid/megaraid_sas_fusion.c')
-rw-r--r-- | drivers/scsi/megaraid/megaraid_sas_fusion.c | 1334 |
1 files changed, 1080 insertions, 254 deletions
diff --git a/drivers/scsi/megaraid/megaraid_sas_fusion.c b/drivers/scsi/megaraid/megaraid_sas_fusion.c index 24778ba4b6e8..29650ba669da 100644 --- a/drivers/scsi/megaraid/megaraid_sas_fusion.c +++ b/drivers/scsi/megaraid/megaraid_sas_fusion.c @@ -47,6 +47,7 @@ #include <linux/blkdev.h> #include <linux/mutex.h> #include <linux/poll.h> +#include <linux/vmalloc.h> #include <scsi/scsi.h> #include <scsi/scsi_cmnd.h> @@ -181,32 +182,44 @@ inline void megasas_return_cmd_fusion(struct megasas_instance *instance, struct megasas_cmd_fusion *cmd) { cmd->scmd = NULL; - memset(cmd->io_request, 0, sizeof(struct MPI2_RAID_SCSI_IO_REQUEST)); + memset(cmd->io_request, 0, MEGA_MPI2_RAID_DEFAULT_IO_FRAME_SIZE); + cmd->r1_alt_dev_handle = MR_DEVHANDLE_INVALID; + cmd->cmd_completed = false; } /** * megasas_fire_cmd_fusion - Sends command to the FW + * @instance: Adapter soft state + * @req_desc: 32bit or 64bit Request descriptor + * + * Perform PCI Write. Ventura supports 32 bit Descriptor. + * Prior to Ventura (12G) MR controller supports 64 bit Descriptor. */ + static void megasas_fire_cmd_fusion(struct megasas_instance *instance, union MEGASAS_REQUEST_DESCRIPTOR_UNION *req_desc) { + if (instance->is_ventura) + writel(le32_to_cpu(req_desc->u.low), + &instance->reg_set->inbound_single_queue_port); + else { #if defined(writeq) && defined(CONFIG_64BIT) - u64 req_data = (((u64)le32_to_cpu(req_desc->u.high) << 32) | - le32_to_cpu(req_desc->u.low)); + u64 req_data = (((u64)le32_to_cpu(req_desc->u.high) << 32) | + le32_to_cpu(req_desc->u.low)); - writeq(req_data, &instance->reg_set->inbound_low_queue_port); + writeq(req_data, &instance->reg_set->inbound_low_queue_port); #else - unsigned long flags; - - spin_lock_irqsave(&instance->hba_lock, flags); - writel(le32_to_cpu(req_desc->u.low), - &instance->reg_set->inbound_low_queue_port); - writel(le32_to_cpu(req_desc->u.high), - &instance->reg_set->inbound_high_queue_port); - mmiowb(); - spin_unlock_irqrestore(&instance->hba_lock, flags); + unsigned long flags; + spin_lock_irqsave(&instance->hba_lock, flags); + writel(le32_to_cpu(req_desc->u.low), + &instance->reg_set->inbound_low_queue_port); + writel(le32_to_cpu(req_desc->u.high), + &instance->reg_set->inbound_high_queue_port); + mmiowb(); + spin_unlock_irqrestore(&instance->hba_lock, flags); #endif + } } /** @@ -229,7 +242,10 @@ megasas_fusion_update_can_queue(struct megasas_instance *instance, int fw_boot_c reg_set = instance->reg_set; - cur_max_fw_cmds = readl(&instance->reg_set->outbound_scratch_pad_3) & 0x00FFFF; + /* ventura FW does not fill outbound_scratch_pad_3 with queue depth */ + if (!instance->is_ventura) + cur_max_fw_cmds = + readl(&instance->reg_set->outbound_scratch_pad_3) & 0x00FFFF; if (dual_qdepth_disable || !cur_max_fw_cmds) cur_max_fw_cmds = instance->instancet->read_fw_status_reg(reg_set) & 0x00FFFF; @@ -243,7 +259,7 @@ megasas_fusion_update_can_queue(struct megasas_instance *instance, int fw_boot_c if (fw_boot_context == OCR_CONTEXT) { cur_max_fw_cmds = cur_max_fw_cmds - 1; - if (cur_max_fw_cmds <= instance->max_fw_cmds) { + if (cur_max_fw_cmds < instance->max_fw_cmds) { instance->cur_can_queue = cur_max_fw_cmds - (MEGASAS_FUSION_INTERNAL_CMDS + MEGASAS_FUSION_IOCTL_CMDS); @@ -255,7 +271,8 @@ megasas_fusion_update_can_queue(struct megasas_instance *instance, int fw_boot_c instance->ldio_threshold = ldio_threshold; if (!instance->is_rdpq) - instance->max_fw_cmds = min_t(u16, instance->max_fw_cmds, 1024); + instance->max_fw_cmds = + min_t(u16, instance->max_fw_cmds, 1024); if (reset_devices) instance->max_fw_cmds = min(instance->max_fw_cmds, @@ -271,7 +288,14 @@ megasas_fusion_update_can_queue(struct megasas_instance *instance, int fw_boot_c (MEGASAS_FUSION_INTERNAL_CMDS + MEGASAS_FUSION_IOCTL_CMDS); instance->cur_can_queue = instance->max_scsi_cmds; + instance->host->can_queue = instance->cur_can_queue; } + + if (instance->is_ventura) + instance->max_mpt_cmds = + instance->max_fw_cmds * RAID_1_PEER_CMDS; + else + instance->max_mpt_cmds = instance->max_fw_cmds; } /** * megasas_free_cmds_fusion - Free all the cmds in the free cmd pool @@ -285,7 +309,7 @@ megasas_free_cmds_fusion(struct megasas_instance *instance) struct megasas_cmd_fusion *cmd; /* SG, Sense */ - for (i = 0; i < instance->max_fw_cmds; i++) { + for (i = 0; i < instance->max_mpt_cmds; i++) { cmd = fusion->cmd_list[i]; if (cmd) { if (cmd->sg_frame) @@ -329,7 +353,7 @@ megasas_free_cmds_fusion(struct megasas_instance *instance) /* cmd_list */ - for (i = 0; i < instance->max_fw_cmds; i++) + for (i = 0; i < instance->max_mpt_cmds; i++) kfree(fusion->cmd_list[i]); kfree(fusion->cmd_list); @@ -343,7 +367,7 @@ megasas_free_cmds_fusion(struct megasas_instance *instance) static int megasas_create_sg_sense_fusion(struct megasas_instance *instance) { int i; - u32 max_cmd; + u16 max_cmd; struct fusion_context *fusion; struct megasas_cmd_fusion *cmd; @@ -353,7 +377,8 @@ static int megasas_create_sg_sense_fusion(struct megasas_instance *instance) fusion->sg_dma_pool = pci_pool_create("mr_sg", instance->pdev, - instance->max_chain_frame_sz, 4, 0); + instance->max_chain_frame_sz, + MR_DEFAULT_NVME_PAGE_SIZE, 0); /* SCSI_SENSE_BUFFERSIZE = 96 bytes */ fusion->sense_dma_pool = pci_pool_create("mr_sense", instance->pdev, @@ -381,33 +406,47 @@ static int megasas_create_sg_sense_fusion(struct megasas_instance *instance) return -ENOMEM; } } + + /* create sense buffer for the raid 1/10 fp */ + for (i = max_cmd; i < instance->max_mpt_cmds; i++) { + cmd = fusion->cmd_list[i]; + cmd->sense = pci_pool_alloc(fusion->sense_dma_pool, + GFP_KERNEL, &cmd->sense_phys_addr); + if (!cmd->sense) { + dev_err(&instance->pdev->dev, + "Failed from %s %d\n", __func__, __LINE__); + return -ENOMEM; + } + } + return 0; } int megasas_alloc_cmdlist_fusion(struct megasas_instance *instance) { - u32 max_cmd, i; + u32 max_mpt_cmd, i; struct fusion_context *fusion; fusion = instance->ctrl_context; - max_cmd = instance->max_fw_cmds; + max_mpt_cmd = instance->max_mpt_cmds; /* * fusion->cmd_list is an array of struct megasas_cmd_fusion pointers. * Allocate the dynamic array first and then allocate individual * commands. */ - fusion->cmd_list = kzalloc(sizeof(struct megasas_cmd_fusion *) * max_cmd, - GFP_KERNEL); + fusion->cmd_list = + kzalloc(sizeof(struct megasas_cmd_fusion *) * max_mpt_cmd, + GFP_KERNEL); if (!fusion->cmd_list) { dev_err(&instance->pdev->dev, "Failed from %s %d\n", __func__, __LINE__); return -ENOMEM; } - for (i = 0; i < max_cmd; i++) { + for (i = 0; i < max_mpt_cmd; i++) { fusion->cmd_list[i] = kzalloc(sizeof(struct megasas_cmd_fusion), GFP_KERNEL); if (!fusion->cmd_list[i]) { @@ -539,7 +578,7 @@ megasas_alloc_rdpq_fusion(struct megasas_instance *instance) } fusion->rdpq_virt[i].RDPQBaseAddress = - fusion->reply_frames_desc_phys[i]; + cpu_to_le64(fusion->reply_frames_desc_phys[i]); reply_desc = fusion->reply_frames_desc[i]; for (j = 0; j < fusion->reply_q_depth; j++, reply_desc++) @@ -642,13 +681,14 @@ megasas_alloc_cmds_fusion(struct megasas_instance *instance) */ /* SMID 0 is reserved. Set SMID/index from 1 */ - for (i = 0; i < instance->max_fw_cmds; i++) { + for (i = 0; i < instance->max_mpt_cmds; i++) { cmd = fusion->cmd_list[i]; offset = MEGA_MPI2_RAID_DEFAULT_IO_FRAME_SIZE * i; memset(cmd, 0, sizeof(struct megasas_cmd_fusion)); cmd->index = i + 1; cmd->scmd = NULL; - cmd->sync_cmd_idx = (i >= instance->max_scsi_cmds) ? + cmd->sync_cmd_idx = + (i >= instance->max_scsi_cmds && i < instance->max_fw_cmds) ? (i - instance->max_scsi_cmds) : (u32)ULONG_MAX; /* Set to Invalid */ cmd->instance = instance; @@ -658,6 +698,7 @@ megasas_alloc_cmds_fusion(struct megasas_instance *instance) memset(cmd->io_request, 0, sizeof(struct MPI2_RAID_SCSI_IO_REQUEST)); cmd->io_request_phys_addr = io_req_base_phys + offset; + cmd->r1_alt_dev_handle = MR_DEVHANDLE_INVALID; } if (megasas_create_sg_sense_fusion(instance)) @@ -725,6 +766,7 @@ megasas_ioc_init_fusion(struct megasas_instance *instance) const char *sys_info; MFI_CAPABILITIES *drv_ops; u32 scratch_pad_2; + unsigned long flags; fusion = instance->ctrl_context; @@ -781,6 +823,7 @@ megasas_ioc_init_fusion(struct megasas_instance *instance) MPI2_IOCINIT_MSGFLAG_RDPQ_ARRAY_MODE : 0; IOCInitMessage->SystemRequestFrameBaseAddress = cpu_to_le64(fusion->io_request_frames_phys); IOCInitMessage->HostMSIxVectors = instance->msix_vectors; + IOCInitMessage->HostPageSize = MR_DEFAULT_NVME_PAGE_SHIFT; init_frame = (struct megasas_init_frame *)cmd->frame; memset(init_frame, 0, MEGAMFI_FRAME_SIZE); @@ -796,7 +839,7 @@ megasas_ioc_init_fusion(struct megasas_instance *instance) drv_ops = (MFI_CAPABILITIES *) &(init_frame->driver_operations); /* driver support Extended MSIX */ - if (fusion->adapter_type == INVADER_SERIES) + if (fusion->adapter_type >= INVADER_SERIES) drv_ops->mfi_capabilities.support_additional_msix = 1; /* driver supports HA / Remote LUN over Fast Path interface */ drv_ops->mfi_capabilities.support_fp_remote_lun = 1; @@ -813,6 +856,7 @@ megasas_ioc_init_fusion(struct megasas_instance *instance) drv_ops->mfi_capabilities.support_ext_queue_depth = 1; drv_ops->mfi_capabilities.support_qd_throttling = 1; + drv_ops->mfi_capabilities.support_pd_map_target_id = 1; /* Convert capability to LE32 */ cpu_to_le32s((u32 *)&init_frame->driver_operations.mfi_capabilities); @@ -850,7 +894,14 @@ megasas_ioc_init_fusion(struct megasas_instance *instance) break; } - megasas_fire_cmd_fusion(instance, &req_desc); + /* For Ventura also IOC INIT required 64 bit Descriptor write. */ + spin_lock_irqsave(&instance->hba_lock, flags); + writel(le32_to_cpu(req_desc.u.low), + &instance->reg_set->inbound_low_queue_port); + writel(le32_to_cpu(req_desc.u.high), + &instance->reg_set->inbound_high_queue_port); + mmiowb(); + spin_unlock_irqrestore(&instance->hba_lock, flags); wait_and_poll(instance, cmd, MFI_POLL_TIMEOUT_SECS); @@ -1009,11 +1060,6 @@ megasas_get_ld_map_info(struct megasas_instance *instance) memset(ci, 0, fusion->max_map_sz); memset(dcmd->mbox.b, 0, MFI_MBOX_SIZE); -#if VD_EXT_DEBUG - dev_dbg(&instance->pdev->dev, - "%s sending MR_DCMD_LD_MAP_GET_INFO with size %d\n", - __func__, cpu_to_le32(size_map_info)); -#endif dcmd->cmd = MFI_CMD_DCMD; dcmd->cmd_status = 0xFF; dcmd->sge_count = 1; @@ -1065,10 +1111,11 @@ megasas_get_map_info(struct megasas_instance *instance) int megasas_sync_map_info(struct megasas_instance *instance) { - int ret = 0, i; + int i; struct megasas_cmd *cmd; struct megasas_dcmd_frame *dcmd; - u32 size_sync_info, num_lds; + u16 num_lds; + u32 size_sync_info; struct fusion_context *fusion; struct MR_LD_TARGET_SYNC *ci = NULL; struct MR_DRV_RAID_MAP_ALL *map; @@ -1134,7 +1181,7 @@ megasas_sync_map_info(struct megasas_instance *instance) instance->instancet->issue_dcmd(instance, cmd); - return ret; + return 0; } /* @@ -1220,7 +1267,8 @@ megasas_init_adapter_fusion(struct megasas_instance *instance) { struct megasas_register_set __iomem *reg_set; struct fusion_context *fusion; - u32 max_cmd, scratch_pad_2; + u16 max_cmd; + u32 scratch_pad_2; int i = 0, count; fusion = instance->ctrl_context; @@ -1230,13 +1278,6 @@ megasas_init_adapter_fusion(struct megasas_instance *instance) megasas_fusion_update_can_queue(instance, PROBE_CONTEXT); /* - * Reduce the max supported cmds by 1. This is to ensure that the - * reply_q_sz (1 more than the max cmd that driver may send) - * does not exceed max cmds that the FW can support - */ - instance->max_fw_cmds = instance->max_fw_cmds-1; - - /* * Only Driver's internal DCMDs and IOCTL DCMDs needs to have MFI frames */ instance->max_mfi_cmds = @@ -1247,12 +1288,12 @@ megasas_init_adapter_fusion(struct megasas_instance *instance) fusion->reply_q_depth = 2 * (((max_cmd + 1 + 15)/16)*16); fusion->request_alloc_sz = - sizeof(union MEGASAS_REQUEST_DESCRIPTOR_UNION) *max_cmd; + sizeof(union MEGASAS_REQUEST_DESCRIPTOR_UNION) * instance->max_mpt_cmds; fusion->reply_alloc_sz = sizeof(union MPI2_REPLY_DESCRIPTORS_UNION) *(fusion->reply_q_depth); fusion->io_frames_alloc_sz = MEGA_MPI2_RAID_DEFAULT_IO_FRAME_SIZE + - (MEGA_MPI2_RAID_DEFAULT_IO_FRAME_SIZE * - (max_cmd + 1)); /* Extra 1 for SMID 0 */ + (MEGA_MPI2_RAID_DEFAULT_IO_FRAME_SIZE + * (instance->max_mpt_cmds + 1)); /* Extra 1 for SMID 0 */ scratch_pad_2 = readl(&instance->reg_set->outbound_scratch_pad_2); /* If scratch_pad_2 & MEGASAS_MAX_CHAIN_SIZE_UNITS_MASK is set, @@ -1302,7 +1343,7 @@ megasas_init_adapter_fusion(struct megasas_instance *instance) fusion->last_reply_idx[i] = 0; /* - * For fusion adapters, 3 commands for IOCTL and 5 commands + * For fusion adapters, 3 commands for IOCTL and 8 commands * for driver's internal DCMDs. */ instance->max_scsi_cmds = instance->max_fw_cmds - @@ -1331,6 +1372,7 @@ megasas_init_adapter_fusion(struct megasas_instance *instance) } instance->flag_ieee = 1; + instance->r1_ldio_hint_default = MR_R1_LDIO_PIGGYBACK_DEFAULT; fusion->fast_path_io = 0; fusion->drv_map_pages = get_order(fusion->drv_map_sz); @@ -1388,96 +1430,348 @@ fail_alloc_mfi_cmds: */ void -map_cmd_status(struct megasas_cmd_fusion *cmd, u8 status, u8 ext_status) +map_cmd_status(struct fusion_context *fusion, + struct scsi_cmnd *scmd, u8 status, u8 ext_status, + u32 data_length, u8 *sense) { + u8 cmd_type; + int resid; + cmd_type = megasas_cmd_type(scmd); switch (status) { case MFI_STAT_OK: - cmd->scmd->result = DID_OK << 16; + scmd->result = DID_OK << 16; break; case MFI_STAT_SCSI_IO_FAILED: case MFI_STAT_LD_INIT_IN_PROGRESS: - cmd->scmd->result = (DID_ERROR << 16) | ext_status; + scmd->result = (DID_ERROR << 16) | ext_status; break; case MFI_STAT_SCSI_DONE_WITH_ERROR: - cmd->scmd->result = (DID_OK << 16) | ext_status; + scmd->result = (DID_OK << 16) | ext_status; if (ext_status == SAM_STAT_CHECK_CONDITION) { - memset(cmd->scmd->sense_buffer, 0, + memset(scmd->sense_buffer, 0, SCSI_SENSE_BUFFERSIZE); - memcpy(cmd->scmd->sense_buffer, cmd->sense, + memcpy(scmd->sense_buffer, sense, SCSI_SENSE_BUFFERSIZE); - cmd->scmd->result |= DRIVER_SENSE << 24; + scmd->result |= DRIVER_SENSE << 24; } + + /* + * If the IO request is partially completed, then MR FW will + * update "io_request->DataLength" field with actual number of + * bytes transferred.Driver will set residual bytes count in + * SCSI command structure. + */ + resid = (scsi_bufflen(scmd) - data_length); + scsi_set_resid(scmd, resid); + + if (resid && + ((cmd_type == READ_WRITE_LDIO) || + (cmd_type == READ_WRITE_SYSPDIO))) + scmd_printk(KERN_INFO, scmd, "BRCM Debug mfi stat 0x%x, data len" + " requested/completed 0x%x/0x%x\n", + status, scsi_bufflen(scmd), data_length); break; case MFI_STAT_LD_OFFLINE: case MFI_STAT_DEVICE_NOT_FOUND: - cmd->scmd->result = DID_BAD_TARGET << 16; + scmd->result = DID_BAD_TARGET << 16; break; case MFI_STAT_CONFIG_SEQ_MISMATCH: - cmd->scmd->result = DID_IMM_RETRY << 16; + scmd->result = DID_IMM_RETRY << 16; break; default: - dev_printk(KERN_DEBUG, &cmd->instance->pdev->dev, "FW status %#x\n", status); - cmd->scmd->result = DID_ERROR << 16; + scmd->result = DID_ERROR << 16; break; } } /** + * megasas_is_prp_possible - + * Checks if native NVMe PRPs can be built for the IO + * + * @instance: Adapter soft state + * @scmd: SCSI command from the mid-layer + * @sge_count: scatter gather element count. + * + * Returns: true: PRPs can be built + * false: IEEE SGLs needs to be built + */ +static bool +megasas_is_prp_possible(struct megasas_instance *instance, + struct scsi_cmnd *scmd, int sge_count) +{ + struct fusion_context *fusion; + int i; + u32 data_length = 0; + struct scatterlist *sg_scmd; + bool build_prp = false; + u32 mr_nvme_pg_size; + + mr_nvme_pg_size = max_t(u32, instance->nvme_page_size, + MR_DEFAULT_NVME_PAGE_SIZE); + fusion = instance->ctrl_context; + data_length = scsi_bufflen(scmd); + sg_scmd = scsi_sglist(scmd); + + /* + * NVMe uses one PRP for each page (or part of a page) + * look at the data length - if 4 pages or less then IEEE is OK + * if > 5 pages then we need to build a native SGL + * if > 4 and <= 5 pages, then check physical address of 1st SG entry + * if this first size in the page is >= the residual beyond 4 pages + * then use IEEE, otherwise use native SGL + */ + + if (data_length > (mr_nvme_pg_size * 5)) { + build_prp = true; + } else if ((data_length > (mr_nvme_pg_size * 4)) && + (data_length <= (mr_nvme_pg_size * 5))) { + /* check if 1st SG entry size is < residual beyond 4 pages */ + if (sg_dma_len(sg_scmd) < (data_length - (mr_nvme_pg_size * 4))) + build_prp = true; + } + +/* + * Below code detects gaps/holes in IO data buffers. + * What does holes/gaps mean? + * Any SGE except first one in a SGL starts at non NVME page size + * aligned address OR Any SGE except last one in a SGL ends at + * non NVME page size boundary. + * + * Driver has already informed block layer by setting boundary rules for + * bio merging done at NVME page size boundary calling kernel API + * blk_queue_virt_boundary inside slave_config. + * Still there is possibility of IO coming with holes to driver because of + * IO merging done by IO scheduler. + * + * With SCSI BLK MQ enabled, there will be no IO with holes as there is no + * IO scheduling so no IO merging. + * + * With SCSI BLK MQ disabled, IO scheduler may attempt to merge IOs and + * then sending IOs with holes. + * + * Though driver can request block layer to disable IO merging by calling- + * queue_flag_set_unlocked(QUEUE_FLAG_NOMERGES, sdev->request_queue) but + * user may tune sysfs parameter- nomerges again to 0 or 1. + * + * If in future IO scheduling is enabled with SCSI BLK MQ, + * this algorithm to detect holes will be required in driver + * for SCSI BLK MQ enabled case as well. + * + * + */ + scsi_for_each_sg(scmd, sg_scmd, sge_count, i) { + if ((i != 0) && (i != (sge_count - 1))) { + if (mega_mod64(sg_dma_len(sg_scmd), mr_nvme_pg_size) || + mega_mod64(sg_dma_address(sg_scmd), + mr_nvme_pg_size)) { + build_prp = false; + atomic_inc(&instance->sge_holes_type1); + break; + } + } + + if ((sge_count > 1) && (i == 0)) { + if ((mega_mod64((sg_dma_address(sg_scmd) + + sg_dma_len(sg_scmd)), + mr_nvme_pg_size))) { + build_prp = false; + atomic_inc(&instance->sge_holes_type2); + break; + } + } + + if ((sge_count > 1) && (i == (sge_count - 1))) { + if (mega_mod64(sg_dma_address(sg_scmd), + mr_nvme_pg_size)) { + build_prp = false; + atomic_inc(&instance->sge_holes_type3); + break; + } + } + } + + return build_prp; +} + +/** + * megasas_make_prp_nvme - + * Prepare PRPs(Physical Region Page)- SGLs specific to NVMe drives only + * + * @instance: Adapter soft state + * @scmd: SCSI command from the mid-layer + * @sgl_ptr: SGL to be filled in + * @cmd: Fusion command frame + * @sge_count: scatter gather element count. + * + * Returns: true: PRPs are built + * false: IEEE SGLs needs to be built + */ +static bool +megasas_make_prp_nvme(struct megasas_instance *instance, struct scsi_cmnd *scmd, + struct MPI25_IEEE_SGE_CHAIN64 *sgl_ptr, + struct megasas_cmd_fusion *cmd, int sge_count) +{ + int sge_len, offset, num_prp_in_chain = 0; + struct MPI25_IEEE_SGE_CHAIN64 *main_chain_element, *ptr_first_sgl; + u64 *ptr_sgl; + dma_addr_t ptr_sgl_phys; + u64 sge_addr; + u32 page_mask, page_mask_result; + struct scatterlist *sg_scmd; + u32 first_prp_len; + bool build_prp = false; + int data_len = scsi_bufflen(scmd); + struct fusion_context *fusion; + u32 mr_nvme_pg_size = max_t(u32, instance->nvme_page_size, + MR_DEFAULT_NVME_PAGE_SIZE); + + fusion = instance->ctrl_context; + + build_prp = megasas_is_prp_possible(instance, scmd, sge_count); + + if (!build_prp) + return false; + + /* + * Nvme has a very convoluted prp format. One prp is required + * for each page or partial page. Driver need to split up OS sg_list + * entries if it is longer than one page or cross a page + * boundary. Driver also have to insert a PRP list pointer entry as + * the last entry in each physical page of the PRP list. + * + * NOTE: The first PRP "entry" is actually placed in the first + * SGL entry in the main message as IEEE 64 format. The 2nd + * entry in the main message is the chain element, and the rest + * of the PRP entries are built in the contiguous pcie buffer. + */ + page_mask = mr_nvme_pg_size - 1; + ptr_sgl = (u64 *)cmd->sg_frame; + ptr_sgl_phys = cmd->sg_frame_phys_addr; + memset(ptr_sgl, 0, instance->max_chain_frame_sz); + + /* Build chain frame element which holds all prps except first*/ + main_chain_element = (struct MPI25_IEEE_SGE_CHAIN64 *) + ((u8 *)sgl_ptr + sizeof(struct MPI25_IEEE_SGE_CHAIN64)); + + main_chain_element->Address = cpu_to_le64(ptr_sgl_phys); + main_chain_element->NextChainOffset = 0; + main_chain_element->Flags = IEEE_SGE_FLAGS_CHAIN_ELEMENT | + IEEE_SGE_FLAGS_SYSTEM_ADDR | + MPI26_IEEE_SGE_FLAGS_NSF_NVME_PRP; + + /* Build first prp, sge need not to be page aligned*/ + ptr_first_sgl = sgl_ptr; + sg_scmd = scsi_sglist(scmd); + sge_addr = sg_dma_address(sg_scmd); + sge_len = sg_dma_len(sg_scmd); + + offset = (u32)(sge_addr & page_mask); + first_prp_len = mr_nvme_pg_size - offset; + + ptr_first_sgl->Address = cpu_to_le64(sge_addr); + ptr_first_sgl->Length = cpu_to_le32(first_prp_len); + + data_len -= first_prp_len; + + if (sge_len > first_prp_len) { + sge_addr += first_prp_len; + sge_len -= first_prp_len; + } else if (sge_len == first_prp_len) { + sg_scmd = sg_next(sg_scmd); + sge_addr = sg_dma_address(sg_scmd); + sge_len = sg_dma_len(sg_scmd); + } + + for (;;) { + offset = (u32)(sge_addr & page_mask); + + /* Put PRP pointer due to page boundary*/ + page_mask_result = (uintptr_t)(ptr_sgl + 1) & page_mask; + if (unlikely(!page_mask_result)) { + scmd_printk(KERN_NOTICE, + scmd, "page boundary ptr_sgl: 0x%p\n", + ptr_sgl); + ptr_sgl_phys += 8; + *ptr_sgl = cpu_to_le64(ptr_sgl_phys); + ptr_sgl++; + num_prp_in_chain++; + } + + *ptr_sgl = cpu_to_le64(sge_addr); + ptr_sgl++; + ptr_sgl_phys += 8; + num_prp_in_chain++; + + sge_addr += mr_nvme_pg_size; + sge_len -= mr_nvme_pg_size; + data_len -= mr_nvme_pg_size; + + if (data_len <= 0) + break; + + if (sge_len > 0) + continue; + + sg_scmd = sg_next(sg_scmd); + sge_addr = sg_dma_address(sg_scmd); + sge_len = sg_dma_len(sg_scmd); + } + + main_chain_element->Length = + cpu_to_le32(num_prp_in_chain * sizeof(u64)); + + atomic_inc(&instance->prp_sgl); + return build_prp; +} + +/** * megasas_make_sgl_fusion - Prepares 32-bit SGL * @instance: Adapter soft state * @scp: SCSI command from the mid-layer * @sgl_ptr: SGL to be filled in * @cmd: cmd we are working on + * @sge_count sge count * - * If successful, this function returns the number of SG elements. */ -static int +static void megasas_make_sgl_fusion(struct megasas_instance *instance, struct scsi_cmnd *scp, struct MPI25_IEEE_SGE_CHAIN64 *sgl_ptr, - struct megasas_cmd_fusion *cmd) + struct megasas_cmd_fusion *cmd, int sge_count) { - int i, sg_processed, sge_count; + int i, sg_processed; struct scatterlist *os_sgl; struct fusion_context *fusion; fusion = instance->ctrl_context; - if (fusion->adapter_type == INVADER_SERIES) { + if (fusion->adapter_type >= INVADER_SERIES) { struct MPI25_IEEE_SGE_CHAIN64 *sgl_ptr_end = sgl_ptr; sgl_ptr_end += fusion->max_sge_in_main_msg - 1; sgl_ptr_end->Flags = 0; } - sge_count = scsi_dma_map(scp); - - BUG_ON(sge_count < 0); - - if (sge_count > instance->max_num_sge || !sge_count) - return sge_count; - scsi_for_each_sg(scp, os_sgl, sge_count, i) { sgl_ptr->Length = cpu_to_le32(sg_dma_len(os_sgl)); sgl_ptr->Address = cpu_to_le64(sg_dma_address(os_sgl)); sgl_ptr->Flags = 0; - if (fusion->adapter_type == INVADER_SERIES) + if (fusion->adapter_type >= INVADER_SERIES) if (i == sge_count - 1) sgl_ptr->Flags = IEEE_SGE_FLAGS_END_OF_LIST; sgl_ptr++; - sg_processed = i + 1; if ((sg_processed == (fusion->max_sge_in_main_msg - 1)) && (sge_count > fusion->max_sge_in_main_msg)) { struct MPI25_IEEE_SGE_CHAIN64 *sg_chain; - if (fusion->adapter_type == INVADER_SERIES) { + if (fusion->adapter_type >= INVADER_SERIES) { if ((le16_to_cpu(cmd->io_request->IoFlags) & MPI25_SAS_DEVICE0_FLAGS_ENABLED_FAST_PATH) != MPI25_SAS_DEVICE0_FLAGS_ENABLED_FAST_PATH) @@ -1493,7 +1787,7 @@ megasas_make_sgl_fusion(struct megasas_instance *instance, sg_chain = sgl_ptr; /* Prepare chain element */ sg_chain->NextChainOffset = 0; - if (fusion->adapter_type == INVADER_SERIES) + if (fusion->adapter_type >= INVADER_SERIES) sg_chain->Flags = IEEE_SGE_FLAGS_CHAIN_ELEMENT; else sg_chain->Flags = @@ -1507,6 +1801,45 @@ megasas_make_sgl_fusion(struct megasas_instance *instance, memset(sgl_ptr, 0, instance->max_chain_frame_sz); } } + atomic_inc(&instance->ieee_sgl); +} + +/** + * megasas_make_sgl - Build Scatter Gather List(SGLs) + * @scp: SCSI command pointer + * @instance: Soft instance of controller + * @cmd: Fusion command pointer + * + * This function will build sgls based on device type. + * For nvme drives, there is different way of building sgls in nvme native + * format- PRPs(Physical Region Page). + * + * Returns the number of sg lists actually used, zero if the sg lists + * is NULL, or -ENOMEM if the mapping failed + */ +static +int megasas_make_sgl(struct megasas_instance *instance, struct scsi_cmnd *scp, + struct megasas_cmd_fusion *cmd) +{ + int sge_count; + bool build_prp = false; + struct MPI25_IEEE_SGE_CHAIN64 *sgl_chain64; + + sge_count = scsi_dma_map(scp); + + if ((sge_count > instance->max_num_sge) || (sge_count <= 0)) + return sge_count; + + sgl_chain64 = (struct MPI25_IEEE_SGE_CHAIN64 *)&cmd->io_request->SGL; + if ((le16_to_cpu(cmd->io_request->IoFlags) & + MPI25_SAS_DEVICE0_FLAGS_ENABLED_FAST_PATH) && + (cmd->pd_interface == NVME_PD)) + build_prp = megasas_make_prp_nvme(instance, scp, sgl_chain64, + cmd, sge_count); + + if (!build_prp) + megasas_make_sgl_fusion(instance, scp, sgl_chain64, + cmd, sge_count); return sge_count; } @@ -1525,7 +1858,7 @@ megasas_set_pd_lba(struct MPI2_RAID_SCSI_IO_REQUEST *io_request, u8 cdb_len, struct MR_DRV_RAID_MAP_ALL *local_map_ptr, u32 ref_tag) { struct MR_LD_RAID *raid; - u32 ld; + u16 ld; u64 start_blk = io_info->pdBlock; u8 *cdb = io_request->CDB.CDB32; u32 num_blocks = io_info->numBlocks; @@ -1574,6 +1907,7 @@ megasas_set_pd_lba(struct MPI2_RAID_SCSI_IO_REQUEST *io_request, u8 cdb_len, MPI2_SCSIIO_EEDPFLAGS_CHECK_REFTAG | MPI2_SCSIIO_EEDPFLAGS_CHECK_REMOVE_OP | MPI2_SCSIIO_EEDPFLAGS_CHECK_APPTAG | + MPI25_SCSIIO_EEDPFLAGS_DO_NOT_DISABLE_MODE | MPI2_SCSIIO_EEDPFLAGS_CHECK_GUARD); } else { io_request->EEDPFlags = cpu_to_le16( @@ -1688,6 +2022,166 @@ megasas_set_pd_lba(struct MPI2_RAID_SCSI_IO_REQUEST *io_request, u8 cdb_len, } /** + * megasas_stream_detect - stream detection on read and and write IOs + * @instance: Adapter soft state + * @cmd: Command to be prepared + * @io_info: IO Request info + * + */ + +/** stream detection on read and and write IOs */ +static void megasas_stream_detect(struct megasas_instance *instance, + struct megasas_cmd_fusion *cmd, + struct IO_REQUEST_INFO *io_info) +{ + struct fusion_context *fusion = instance->ctrl_context; + u32 device_id = io_info->ldTgtId; + struct LD_STREAM_DETECT *current_ld_sd + = fusion->stream_detect_by_ld[device_id]; + u32 *track_stream = ¤t_ld_sd->mru_bit_map, stream_num; + u32 shifted_values, unshifted_values; + u32 index_value_mask, shifted_values_mask; + int i; + bool is_read_ahead = false; + struct STREAM_DETECT *current_sd; + /* find possible stream */ + for (i = 0; i < MAX_STREAMS_TRACKED; ++i) { + stream_num = (*track_stream >> + (i * BITS_PER_INDEX_STREAM)) & + STREAM_MASK; + current_sd = ¤t_ld_sd->stream_track[stream_num]; + /* if we found a stream, update the raid + * context and also update the mruBitMap + */ + /* boundary condition */ + if ((current_sd->next_seq_lba) && + (io_info->ldStartBlock >= current_sd->next_seq_lba) && + (io_info->ldStartBlock <= (current_sd->next_seq_lba + 32)) && + (current_sd->is_read == io_info->isRead)) { + + if ((io_info->ldStartBlock != current_sd->next_seq_lba) && + ((!io_info->isRead) || (!is_read_ahead))) + /* + * Once the API availible we need to change this. + * At this point we are not allowing any gap + */ + continue; + + SET_STREAM_DETECTED(cmd->io_request->RaidContext.raid_context_g35); + current_sd->next_seq_lba = + io_info->ldStartBlock + io_info->numBlocks; + /* + * update the mruBitMap LRU + */ + shifted_values_mask = + (1 << i * BITS_PER_INDEX_STREAM) - 1; + shifted_values = ((*track_stream & shifted_values_mask) + << BITS_PER_INDEX_STREAM); + index_value_mask = + STREAM_MASK << i * BITS_PER_INDEX_STREAM; + unshifted_values = + *track_stream & ~(shifted_values_mask | + index_value_mask); + *track_stream = + unshifted_values | shifted_values | stream_num; + return; + } + } + /* + * if we did not find any stream, create a new one + * from the least recently used + */ + stream_num = (*track_stream >> + ((MAX_STREAMS_TRACKED - 1) * BITS_PER_INDEX_STREAM)) & + STREAM_MASK; + current_sd = ¤t_ld_sd->stream_track[stream_num]; + current_sd->is_read = io_info->isRead; + current_sd->next_seq_lba = io_info->ldStartBlock + io_info->numBlocks; + *track_stream = (((*track_stream & ZERO_LAST_STREAM) << 4) | stream_num); + return; +} + +/** + * megasas_set_raidflag_cpu_affinity - This function sets the cpu + * affinity (cpu of the controller) and raid_flags in the raid context + * based on IO type. + * + * @praid_context: IO RAID context + * @raid: LD raid map + * @fp_possible: Is fast path possible? + * @is_read: Is read IO? + * + */ +static void +megasas_set_raidflag_cpu_affinity(union RAID_CONTEXT_UNION *praid_context, + struct MR_LD_RAID *raid, bool fp_possible, + u8 is_read, u32 scsi_buff_len) +{ + u8 cpu_sel = MR_RAID_CTX_CPUSEL_0; + struct RAID_CONTEXT_G35 *rctx_g35; + + rctx_g35 = &praid_context->raid_context_g35; + if (fp_possible) { + if (is_read) { + if ((raid->cpuAffinity.pdRead.cpu0) && + (raid->cpuAffinity.pdRead.cpu1)) + cpu_sel = MR_RAID_CTX_CPUSEL_FCFS; + else if (raid->cpuAffinity.pdRead.cpu1) + cpu_sel = MR_RAID_CTX_CPUSEL_1; + } else { + if ((raid->cpuAffinity.pdWrite.cpu0) && + (raid->cpuAffinity.pdWrite.cpu1)) + cpu_sel = MR_RAID_CTX_CPUSEL_FCFS; + else if (raid->cpuAffinity.pdWrite.cpu1) + cpu_sel = MR_RAID_CTX_CPUSEL_1; + /* Fast path cache by pass capable R0/R1 VD */ + if ((raid->level <= 1) && + (raid->capability.fp_cache_bypass_capable)) { + rctx_g35->routing_flags |= + (1 << MR_RAID_CTX_ROUTINGFLAGS_SLD_SHIFT); + rctx_g35->raid_flags = + (MR_RAID_FLAGS_IO_SUB_TYPE_CACHE_BYPASS + << MR_RAID_CTX_RAID_FLAGS_IO_SUB_TYPE_SHIFT); + } + } + } else { + if (is_read) { + if ((raid->cpuAffinity.ldRead.cpu0) && + (raid->cpuAffinity.ldRead.cpu1)) + cpu_sel = MR_RAID_CTX_CPUSEL_FCFS; + else if (raid->cpuAffinity.ldRead.cpu1) + cpu_sel = MR_RAID_CTX_CPUSEL_1; + } else { + if ((raid->cpuAffinity.ldWrite.cpu0) && + (raid->cpuAffinity.ldWrite.cpu1)) + cpu_sel = MR_RAID_CTX_CPUSEL_FCFS; + else if (raid->cpuAffinity.ldWrite.cpu1) + cpu_sel = MR_RAID_CTX_CPUSEL_1; + + if (is_stream_detected(rctx_g35) && + (raid->level == 5) && + (raid->writeMode == MR_RL_WRITE_THROUGH_MODE) && + (cpu_sel == MR_RAID_CTX_CPUSEL_FCFS)) + cpu_sel = MR_RAID_CTX_CPUSEL_0; + } + } + + rctx_g35->routing_flags |= + (cpu_sel << MR_RAID_CTX_ROUTINGFLAGS_CPUSEL_SHIFT); + + /* Always give priority to MR_RAID_FLAGS_IO_SUB_TYPE_LDIO_BW_LIMIT + * vs MR_RAID_FLAGS_IO_SUB_TYPE_CACHE_BYPASS. + * IO Subtype is not bitmap. + */ + if ((raid->level == 1) && (!is_read)) { + if (scsi_buff_len > MR_LARGE_IO_MIN_SIZE) + praid_context->raid_context_g35.raid_flags = + (MR_RAID_FLAGS_IO_SUB_TYPE_LDIO_BW_LIMIT + << MR_RAID_CTX_RAID_FLAGS_IO_SUB_TYPE_SHIFT); + } +} + +/** * megasas_build_ldio_fusion - Prepares IOs to devices * @instance: Adapter soft state * @scp: SCSI command @@ -1701,29 +2195,36 @@ megasas_build_ldio_fusion(struct megasas_instance *instance, struct scsi_cmnd *scp, struct megasas_cmd_fusion *cmd) { - u8 fp_possible; + bool fp_possible; + u16 ld; u32 start_lba_lo, start_lba_hi, device_id, datalength = 0; + u32 scsi_buff_len; struct MPI2_RAID_SCSI_IO_REQUEST *io_request; union MEGASAS_REQUEST_DESCRIPTOR_UNION *req_desc; struct IO_REQUEST_INFO io_info; struct fusion_context *fusion; struct MR_DRV_RAID_MAP_ALL *local_map_ptr; u8 *raidLUN; + unsigned long spinlock_flags; + union RAID_CONTEXT_UNION *praid_context; + struct MR_LD_RAID *raid = NULL; + struct MR_PRIV_DEVICE *mrdev_priv; device_id = MEGASAS_DEV_INDEX(scp); fusion = instance->ctrl_context; io_request = cmd->io_request; - io_request->RaidContext.VirtualDiskTgtId = cpu_to_le16(device_id); - io_request->RaidContext.status = 0; - io_request->RaidContext.exStatus = 0; + io_request->RaidContext.raid_context.virtual_disk_tgt_id = + cpu_to_le16(device_id); + io_request->RaidContext.raid_context.status = 0; + io_request->RaidContext.raid_context.ex_status = 0; req_desc = (union MEGASAS_REQUEST_DESCRIPTOR_UNION *)cmd->request_desc; start_lba_lo = 0; start_lba_hi = 0; - fp_possible = 0; + fp_possible = false; /* * 6-byte READ(0x08) or WRITE(0x0A) cdb @@ -1779,22 +2280,27 @@ megasas_build_ldio_fusion(struct megasas_instance *instance, io_info.ldStartBlock = ((u64)start_lba_hi << 32) | start_lba_lo; io_info.numBlocks = datalength; io_info.ldTgtId = device_id; - io_request->DataLength = cpu_to_le32(scsi_bufflen(scp)); + io_info.r1_alt_dev_handle = MR_DEVHANDLE_INVALID; + scsi_buff_len = scsi_bufflen(scp); + io_request->DataLength = cpu_to_le32(scsi_buff_len); if (scp->sc_data_direction == PCI_DMA_FROMDEVICE) io_info.isRead = 1; local_map_ptr = fusion->ld_drv_map[(instance->map_id & 1)]; + ld = MR_TargetIdToLdGet(device_id, local_map_ptr); - if ((MR_TargetIdToLdGet(device_id, local_map_ptr) >= - instance->fw_supported_vd_count) || (!fusion->fast_path_io)) { - io_request->RaidContext.regLockFlags = 0; - fp_possible = 0; + if (ld < instance->fw_supported_vd_count) + raid = MR_LdRaidGet(ld, local_map_ptr); + + if (!raid || (!fusion->fast_path_io)) { + io_request->RaidContext.raid_context.reg_lock_flags = 0; + fp_possible = false; } else { if (MR_BuildRaidContext(instance, &io_info, - &io_request->RaidContext, + &io_request->RaidContext.raid_context, local_map_ptr, &raidLUN)) - fp_possible = io_info.fpOkForIo; + fp_possible = (io_info.fpOkForIo > 0) ? true : false; } /* Use raw_smp_processor_id() for now until cmd->request->cpu is CPU @@ -1803,6 +2309,54 @@ megasas_build_ldio_fusion(struct megasas_instance *instance, cmd->request_desc->SCSIIO.MSIxIndex = instance->msix_vectors ? raw_smp_processor_id() % instance->msix_vectors : 0; + praid_context = &io_request->RaidContext; + + if (instance->is_ventura) { + spin_lock_irqsave(&instance->stream_lock, spinlock_flags); + megasas_stream_detect(instance, cmd, &io_info); + spin_unlock_irqrestore(&instance->stream_lock, spinlock_flags); + /* In ventura if stream detected for a read and it is read ahead + * capable make this IO as LDIO + */ + if (is_stream_detected(&io_request->RaidContext.raid_context_g35) && + io_info.isRead && io_info.ra_capable) + fp_possible = false; + + /* FP for Optimal raid level 1. + * All large RAID-1 writes (> 32 KiB, both WT and WB modes) + * are built by the driver as LD I/Os. + * All small RAID-1 WT writes (<= 32 KiB) are built as FP I/Os + * (there is never a reason to process these as buffered writes) + * All small RAID-1 WB writes (<= 32 KiB) are built as FP I/Os + * with the SLD bit asserted. + */ + if (io_info.r1_alt_dev_handle != MR_DEVHANDLE_INVALID) { + mrdev_priv = scp->device->hostdata; + + if (atomic_inc_return(&instance->fw_outstanding) > + (instance->host->can_queue)) { + fp_possible = false; + atomic_dec(&instance->fw_outstanding); + } else if ((scsi_buff_len > MR_LARGE_IO_MIN_SIZE) || + atomic_dec_if_positive(&mrdev_priv->r1_ldio_hint)) { + fp_possible = false; + atomic_dec(&instance->fw_outstanding); + if (scsi_buff_len > MR_LARGE_IO_MIN_SIZE) + atomic_set(&mrdev_priv->r1_ldio_hint, + instance->r1_ldio_hint_default); + } + } + + /* If raid is NULL, set CPU affinity to default CPU0 */ + if (raid) + megasas_set_raidflag_cpu_affinity(praid_context, + raid, fp_possible, io_info.isRead, + scsi_buff_len); + else + praid_context->raid_context_g35.routing_flags |= + (MR_RAID_CTX_CPUSEL_0 << MR_RAID_CTX_ROUTINGFLAGS_CPUSEL_SHIFT); + } + if (fp_possible) { megasas_set_pd_lba(io_request, scp->cmd_len, &io_info, scp, local_map_ptr, start_lba_lo); @@ -1811,29 +2365,52 @@ megasas_build_ldio_fusion(struct megasas_instance *instance, (MPI2_REQ_DESCRIPT_FLAGS_FP_IO << MEGASAS_REQ_DESCRIPT_FLAGS_TYPE_SHIFT); if (fusion->adapter_type == INVADER_SERIES) { - if (io_request->RaidContext.regLockFlags == + if (io_request->RaidContext.raid_context.reg_lock_flags == REGION_TYPE_UNUSED) cmd->request_desc->SCSIIO.RequestFlags = (MEGASAS_REQ_DESCRIPT_FLAGS_NO_LOCK << MEGASAS_REQ_DESCRIPT_FLAGS_TYPE_SHIFT); - io_request->RaidContext.Type = MPI2_TYPE_CUDA; - io_request->RaidContext.nseg = 0x1; + io_request->RaidContext.raid_context.type + = MPI2_TYPE_CUDA; + io_request->RaidContext.raid_context.nseg = 0x1; io_request->IoFlags |= cpu_to_le16(MPI25_SAS_DEVICE0_FLAGS_ENABLED_FAST_PATH); - io_request->RaidContext.regLockFlags |= + io_request->RaidContext.raid_context.reg_lock_flags |= (MR_RL_FLAGS_GRANT_DESTINATION_CUDA | MR_RL_FLAGS_SEQ_NUM_ENABLE); + } else if (instance->is_ventura) { + io_request->RaidContext.raid_context_g35.nseg_type |= + (1 << RAID_CONTEXT_NSEG_SHIFT); + io_request->RaidContext.raid_context_g35.nseg_type |= + (MPI2_TYPE_CUDA << RAID_CONTEXT_TYPE_SHIFT); + io_request->RaidContext.raid_context_g35.routing_flags |= + (1 << MR_RAID_CTX_ROUTINGFLAGS_SQN_SHIFT); + io_request->IoFlags |= + cpu_to_le16(MPI25_SAS_DEVICE0_FLAGS_ENABLED_FAST_PATH); } - if ((fusion->load_balance_info[device_id].loadBalanceFlag) && - (io_info.isRead)) { + if (fusion->load_balance_info && + (fusion->load_balance_info[device_id].loadBalanceFlag) && + (io_info.isRead)) { io_info.devHandle = get_updated_dev_handle(instance, &fusion->load_balance_info[device_id], - &io_info); + &io_info, local_map_ptr); scp->SCp.Status |= MEGASAS_LOAD_BALANCE_FLAG; cmd->pd_r1_lb = io_info.pd_after_lb; + if (instance->is_ventura) + io_request->RaidContext.raid_context_g35.span_arm + = io_info.span_arm; + else + io_request->RaidContext.raid_context.span_arm + = io_info.span_arm; + } else scp->SCp.Status &= ~MEGASAS_LOAD_BALANCE_FLAG; + if (instance->is_ventura) + cmd->r1_alt_dev_handle = io_info.r1_alt_dev_handle; + else + cmd->r1_alt_dev_handle = MR_DEVHANDLE_INVALID; + if ((raidLUN[0] == 1) && (local_map_ptr->raidMap.devHndlInfo[io_info.pd_after_lb].validHandles > 1)) { instance->dev_handle = !(instance->dev_handle); @@ -1843,28 +2420,39 @@ megasas_build_ldio_fusion(struct megasas_instance *instance, cmd->request_desc->SCSIIO.DevHandle = io_info.devHandle; io_request->DevHandle = io_info.devHandle; + cmd->pd_interface = io_info.pd_interface; /* populate the LUN field */ memcpy(io_request->LUN, raidLUN, 8); } else { - io_request->RaidContext.timeoutValue = + io_request->RaidContext.raid_context.timeout_value = cpu_to_le16(local_map_ptr->raidMap.fpPdIoTimeoutSec); cmd->request_desc->SCSIIO.RequestFlags = (MEGASAS_REQ_DESCRIPT_FLAGS_LD_IO << MEGASAS_REQ_DESCRIPT_FLAGS_TYPE_SHIFT); if (fusion->adapter_type == INVADER_SERIES) { if (io_info.do_fp_rlbypass || - (io_request->RaidContext.regLockFlags == REGION_TYPE_UNUSED)) + (io_request->RaidContext.raid_context.reg_lock_flags + == REGION_TYPE_UNUSED)) cmd->request_desc->SCSIIO.RequestFlags = (MEGASAS_REQ_DESCRIPT_FLAGS_NO_LOCK << MEGASAS_REQ_DESCRIPT_FLAGS_TYPE_SHIFT); - io_request->RaidContext.Type = MPI2_TYPE_CUDA; - io_request->RaidContext.regLockFlags |= + io_request->RaidContext.raid_context.type + = MPI2_TYPE_CUDA; + io_request->RaidContext.raid_context.reg_lock_flags |= (MR_RL_FLAGS_GRANT_DESTINATION_CPU0 | MR_RL_FLAGS_SEQ_NUM_ENABLE); - io_request->RaidContext.nseg = 0x1; + io_request->RaidContext.raid_context.nseg = 0x1; + } else if (instance->is_ventura) { + io_request->RaidContext.raid_context_g35.routing_flags |= + (1 << MR_RAID_CTX_ROUTINGFLAGS_SQN_SHIFT); + io_request->RaidContext.raid_context_g35.nseg_type |= + (1 << RAID_CONTEXT_NSEG_SHIFT); + io_request->RaidContext.raid_context_g35.nseg_type |= + (MPI2_TYPE_CUDA << RAID_CONTEXT_TYPE_SHIFT); } io_request->Function = MEGASAS_MPI2_FUNCTION_LD_IO_REQUEST; io_request->DevHandle = cpu_to_le16(device_id); + } /* Not FP */ } @@ -1881,27 +2469,26 @@ static void megasas_build_ld_nonrw_fusion(struct megasas_instance *instance, { u32 device_id; struct MPI2_RAID_SCSI_IO_REQUEST *io_request; - u16 pd_index = 0; + u16 ld; struct MR_DRV_RAID_MAP_ALL *local_map_ptr; struct fusion_context *fusion = instance->ctrl_context; u8 span, physArm; __le16 devHandle; - u32 ld, arRef, pd; + u32 arRef, pd; struct MR_LD_RAID *raid; struct RAID_CONTEXT *pRAID_Context; u8 fp_possible = 1; io_request = cmd->io_request; device_id = MEGASAS_DEV_INDEX(scmd); - pd_index = MEGASAS_PD_INDEX(scmd); local_map_ptr = fusion->ld_drv_map[(instance->map_id & 1)]; io_request->DataLength = cpu_to_le32(scsi_bufflen(scmd)); /* get RAID_Context pointer */ - pRAID_Context = &io_request->RaidContext; + pRAID_Context = &io_request->RaidContext.raid_context; /* Check with FW team */ - pRAID_Context->VirtualDiskTgtId = cpu_to_le16(device_id); - pRAID_Context->regLockRowLBA = 0; - pRAID_Context->regLockLength = 0; + pRAID_Context->virtual_disk_tgt_id = cpu_to_le16(device_id); + pRAID_Context->reg_lock_row_lba = 0; + pRAID_Context->reg_lock_length = 0; if (fusion->fast_path_io && ( device_id < instance->fw_supported_vd_count)) { @@ -1909,10 +2496,11 @@ static void megasas_build_ld_nonrw_fusion(struct megasas_instance *instance, ld = MR_TargetIdToLdGet(device_id, local_map_ptr); if (ld >= instance->fw_supported_vd_count) fp_possible = 0; - - raid = MR_LdRaidGet(ld, local_map_ptr); - if (!(raid->capability.fpNonRWCapable)) - fp_possible = 0; + else { + raid = MR_LdRaidGet(ld, local_map_ptr); + if (!(raid->capability.fpNonRWCapable)) + fp_possible = 0; + } } else fp_possible = 0; @@ -1920,7 +2508,7 @@ static void megasas_build_ld_nonrw_fusion(struct megasas_instance *instance, io_request->Function = MEGASAS_MPI2_FUNCTION_LD_IO_REQUEST; io_request->DevHandle = cpu_to_le16(device_id); io_request->LUN[1] = scmd->device->lun; - pRAID_Context->timeoutValue = + pRAID_Context->timeout_value = cpu_to_le16 (scmd->request->timeout / HZ); cmd->request_desc->SCSIIO.RequestFlags = (MPI2_REQ_DESCRIPT_FLAGS_SCSI_IO << @@ -1928,9 +2516,11 @@ static void megasas_build_ld_nonrw_fusion(struct megasas_instance *instance, } else { /* set RAID context values */ - pRAID_Context->configSeqNum = raid->seqNum; - pRAID_Context->regLockFlags = REGION_TYPE_SHARED_READ; - pRAID_Context->timeoutValue = cpu_to_le16(raid->fpIoTimeoutForLd); + pRAID_Context->config_seq_num = raid->seqNum; + if (!instance->is_ventura) + pRAID_Context->reg_lock_flags = REGION_TYPE_SHARED_READ; + pRAID_Context->timeout_value = + cpu_to_le16(raid->fpIoTimeoutForLd); /* get the DevHandle for the PD (since this is fpNonRWCapable, this is a single disk RAID0) */ @@ -1965,7 +2555,8 @@ static void megasas_build_ld_nonrw_fusion(struct megasas_instance *instance, */ static void megasas_build_syspd_fusion(struct megasas_instance *instance, - struct scsi_cmnd *scmd, struct megasas_cmd_fusion *cmd, u8 fp_possible) + struct scsi_cmnd *scmd, struct megasas_cmd_fusion *cmd, + bool fp_possible) { u32 device_id; struct MPI2_RAID_SCSI_IO_REQUEST *io_request; @@ -1975,22 +2566,25 @@ megasas_build_syspd_fusion(struct megasas_instance *instance, struct MR_DRV_RAID_MAP_ALL *local_map_ptr; struct RAID_CONTEXT *pRAID_Context; struct MR_PD_CFG_SEQ_NUM_SYNC *pd_sync; + struct MR_PRIV_DEVICE *mr_device_priv_data; struct fusion_context *fusion = instance->ctrl_context; pd_sync = (void *)fusion->pd_seq_sync[(instance->pd_seq_map_id - 1) & 1]; device_id = MEGASAS_DEV_INDEX(scmd); pd_index = MEGASAS_PD_INDEX(scmd); os_timeout_value = scmd->request->timeout / HZ; + mr_device_priv_data = scmd->device->hostdata; + cmd->pd_interface = mr_device_priv_data->interface_type; io_request = cmd->io_request; /* get RAID_Context pointer */ - pRAID_Context = &io_request->RaidContext; - pRAID_Context->regLockFlags = 0; - pRAID_Context->regLockRowLBA = 0; - pRAID_Context->regLockLength = 0; + pRAID_Context = &io_request->RaidContext.raid_context; + pRAID_Context->reg_lock_flags = 0; + pRAID_Context->reg_lock_row_lba = 0; + pRAID_Context->reg_lock_length = 0; io_request->DataLength = cpu_to_le32(scsi_bufflen(scmd)); io_request->LUN[1] = scmd->device->lun; - pRAID_Context->RAIDFlags = MR_RAID_FLAGS_IO_SUB_TYPE_SYSTEM_PD + pRAID_Context->raid_flags = MR_RAID_FLAGS_IO_SUB_TYPE_SYSTEM_PD << MR_RAID_CTX_RAID_FLAGS_IO_SUB_TYPE_SHIFT; /* If FW supports PD sequence number */ @@ -1999,24 +2593,38 @@ megasas_build_syspd_fusion(struct megasas_instance *instance, /* TgtId must be incremented by 255 as jbod seq number is index * below raid map */ - pRAID_Context->VirtualDiskTgtId = - cpu_to_le16(device_id + (MAX_PHYSICAL_DEVICES - 1)); - pRAID_Context->configSeqNum = pd_sync->seq[pd_index].seqNum; + /* More than 256 PD/JBOD support for Ventura */ + if (instance->support_morethan256jbod) + pRAID_Context->virtual_disk_tgt_id = + pd_sync->seq[pd_index].pd_target_id; + else + pRAID_Context->virtual_disk_tgt_id = + cpu_to_le16(device_id + (MAX_PHYSICAL_DEVICES - 1)); + pRAID_Context->config_seq_num = pd_sync->seq[pd_index].seqNum; io_request->DevHandle = pd_sync->seq[pd_index].devHandle; - pRAID_Context->regLockFlags |= - (MR_RL_FLAGS_SEQ_NUM_ENABLE|MR_RL_FLAGS_GRANT_DESTINATION_CUDA); - pRAID_Context->Type = MPI2_TYPE_CUDA; - pRAID_Context->nseg = 0x1; + if (instance->is_ventura) { + io_request->RaidContext.raid_context_g35.routing_flags |= + (1 << MR_RAID_CTX_ROUTINGFLAGS_SQN_SHIFT); + io_request->RaidContext.raid_context_g35.nseg_type |= + (1 << RAID_CONTEXT_NSEG_SHIFT); + io_request->RaidContext.raid_context_g35.nseg_type |= + (MPI2_TYPE_CUDA << RAID_CONTEXT_TYPE_SHIFT); + } else { + pRAID_Context->type = MPI2_TYPE_CUDA; + pRAID_Context->nseg = 0x1; + pRAID_Context->reg_lock_flags |= + (MR_RL_FLAGS_SEQ_NUM_ENABLE|MR_RL_FLAGS_GRANT_DESTINATION_CUDA); + } } else if (fusion->fast_path_io) { - pRAID_Context->VirtualDiskTgtId = cpu_to_le16(device_id); - pRAID_Context->configSeqNum = 0; + pRAID_Context->virtual_disk_tgt_id = cpu_to_le16(device_id); + pRAID_Context->config_seq_num = 0; local_map_ptr = fusion->ld_drv_map[(instance->map_id & 1)]; io_request->DevHandle = local_map_ptr->raidMap.devHndlInfo[device_id].curDevHdl; } else { /* Want to send all IO via FW path */ - pRAID_Context->VirtualDiskTgtId = cpu_to_le16(device_id); - pRAID_Context->configSeqNum = 0; + pRAID_Context->virtual_disk_tgt_id = cpu_to_le16(device_id); + pRAID_Context->config_seq_num = 0; io_request->DevHandle = cpu_to_le16(0xFFFF); } @@ -2032,17 +2640,17 @@ megasas_build_syspd_fusion(struct megasas_instance *instance, cmd->request_desc->SCSIIO.RequestFlags = (MPI2_REQ_DESCRIPT_FLAGS_SCSI_IO << MEGASAS_REQ_DESCRIPT_FLAGS_TYPE_SHIFT); - pRAID_Context->timeoutValue = cpu_to_le16(os_timeout_value); - pRAID_Context->VirtualDiskTgtId = cpu_to_le16(device_id); + pRAID_Context->timeout_value = cpu_to_le16(os_timeout_value); + pRAID_Context->virtual_disk_tgt_id = cpu_to_le16(device_id); } else { /* system pd Fast Path */ io_request->Function = MPI2_FUNCTION_SCSI_IO_REQUEST; timeout_limit = (scmd->device->type == TYPE_DISK) ? 255 : 0xFFFF; - pRAID_Context->timeoutValue = + pRAID_Context->timeout_value = cpu_to_le16((os_timeout_value > timeout_limit) ? timeout_limit : os_timeout_value); - if (fusion->adapter_type == INVADER_SERIES) + if (fusion->adapter_type >= INVADER_SERIES) io_request->IoFlags |= cpu_to_le16(MPI25_SAS_DEVICE0_FLAGS_ENABLED_FAST_PATH); @@ -2066,9 +2674,11 @@ megasas_build_io_fusion(struct megasas_instance *instance, struct scsi_cmnd *scp, struct megasas_cmd_fusion *cmd) { - u16 sge_count; + int sge_count; u8 cmd_type; struct MPI2_RAID_SCSI_IO_REQUEST *io_request = cmd->io_request; + struct MR_PRIV_DEVICE *mr_device_priv_data; + mr_device_priv_data = scp->device->hostdata; /* Zero out some fields so they don't get reused */ memset(io_request->LUN, 0x0, 8); @@ -2078,9 +2688,9 @@ megasas_build_io_fusion(struct megasas_instance *instance, io_request->Control = 0; io_request->EEDPBlockSize = 0; io_request->ChainOffset = 0; - io_request->RaidContext.RAIDFlags = 0; - io_request->RaidContext.Type = 0; - io_request->RaidContext.nseg = 0; + io_request->RaidContext.raid_context.raid_flags = 0; + io_request->RaidContext.raid_context.type = 0; + io_request->RaidContext.raid_context.nseg = 0; memcpy(io_request->CDB.CDB32, scp->cmnd, scp->cmd_len); /* @@ -2097,12 +2707,14 @@ megasas_build_io_fusion(struct megasas_instance *instance, megasas_build_ld_nonrw_fusion(instance, scp, cmd); break; case READ_WRITE_SYSPDIO: + megasas_build_syspd_fusion(instance, scp, cmd, true); + break; case NON_READ_WRITE_SYSPDIO: - if (instance->secure_jbod_support && - (cmd_type == NON_READ_WRITE_SYSPDIO)) - megasas_build_syspd_fusion(instance, scp, cmd, 0); + if (instance->secure_jbod_support || + mr_device_priv_data->is_tm_capable) + megasas_build_syspd_fusion(instance, scp, cmd, false); else - megasas_build_syspd_fusion(instance, scp, cmd, 1); + megasas_build_syspd_fusion(instance, scp, cmd, true); break; default: break; @@ -2112,23 +2724,27 @@ megasas_build_io_fusion(struct megasas_instance *instance, * Construct SGL */ - sge_count = - megasas_make_sgl_fusion(instance, scp, - (struct MPI25_IEEE_SGE_CHAIN64 *) - &io_request->SGL, cmd); + sge_count = megasas_make_sgl(instance, scp, cmd); - if (sge_count > instance->max_num_sge) { - dev_err(&instance->pdev->dev, "Error. sge_count (0x%x) exceeds " - "max (0x%x) allowed\n", sge_count, - instance->max_num_sge); + if (sge_count > instance->max_num_sge || (sge_count < 0)) { + dev_err(&instance->pdev->dev, + "%s %d sge_count (%d) is out of range. Range is: 0-%d\n", + __func__, __LINE__, sge_count, instance->max_num_sge); return 1; } - /* numSGE store lower 8 bit of sge_count. - * numSGEExt store higher 8 bit of sge_count - */ - io_request->RaidContext.numSGE = sge_count; - io_request->RaidContext.numSGEExt = (u8)(sge_count >> 8); + if (instance->is_ventura) { + set_num_sge(&io_request->RaidContext.raid_context_g35, sge_count); + cpu_to_le16s(&io_request->RaidContext.raid_context_g35.routing_flags); + cpu_to_le16s(&io_request->RaidContext.raid_context_g35.nseg_type); + } else { + /* numSGE store lower 8 bit of sge_count. + * numSGEExt store higher 8 bit of sge_count + */ + io_request->RaidContext.raid_context.num_sge = sge_count; + io_request->RaidContext.raid_context.num_sge_ext = + (u8)(sge_count >> 8); + } io_request->SGLFlags = cpu_to_le16(MPI2_SGE_FLAGS_64_BIT_ADDRESSING); @@ -2149,25 +2765,61 @@ megasas_build_io_fusion(struct megasas_instance *instance, return 0; } -union MEGASAS_REQUEST_DESCRIPTOR_UNION * +static union MEGASAS_REQUEST_DESCRIPTOR_UNION * megasas_get_request_descriptor(struct megasas_instance *instance, u16 index) { u8 *p; struct fusion_context *fusion; - if (index >= instance->max_fw_cmds) { - dev_err(&instance->pdev->dev, "Invalid SMID (0x%x)request for " - "descriptor for scsi%d\n", index, - instance->host->host_no); - return NULL; - } fusion = instance->ctrl_context; - p = fusion->req_frames_desc - +sizeof(union MEGASAS_REQUEST_DESCRIPTOR_UNION) *index; + p = fusion->req_frames_desc + + sizeof(union MEGASAS_REQUEST_DESCRIPTOR_UNION) * index; return (union MEGASAS_REQUEST_DESCRIPTOR_UNION *)p; } + +/* megasas_prepate_secondRaid1_IO + * It prepares the raid 1 second IO + */ +void megasas_prepare_secondRaid1_IO(struct megasas_instance *instance, + struct megasas_cmd_fusion *cmd, + struct megasas_cmd_fusion *r1_cmd) +{ + union MEGASAS_REQUEST_DESCRIPTOR_UNION *req_desc, *req_desc2 = NULL; + struct fusion_context *fusion; + fusion = instance->ctrl_context; + req_desc = cmd->request_desc; + /* copy the io request frame as well as 8 SGEs data for r1 command*/ + memcpy(r1_cmd->io_request, cmd->io_request, + (sizeof(struct MPI2_RAID_SCSI_IO_REQUEST))); + memcpy(&r1_cmd->io_request->SGL, &cmd->io_request->SGL, + (fusion->max_sge_in_main_msg * sizeof(union MPI2_SGE_IO_UNION))); + /*sense buffer is different for r1 command*/ + r1_cmd->io_request->SenseBufferLowAddress = + cpu_to_le32(r1_cmd->sense_phys_addr); + r1_cmd->scmd = cmd->scmd; + req_desc2 = megasas_get_request_descriptor(instance, + (r1_cmd->index - 1)); + req_desc2->Words = 0; + r1_cmd->request_desc = req_desc2; + req_desc2->SCSIIO.SMID = cpu_to_le16(r1_cmd->index); + req_desc2->SCSIIO.RequestFlags = req_desc->SCSIIO.RequestFlags; + r1_cmd->request_desc->SCSIIO.DevHandle = cmd->r1_alt_dev_handle; + r1_cmd->io_request->DevHandle = cmd->r1_alt_dev_handle; + r1_cmd->r1_alt_dev_handle = cmd->io_request->DevHandle; + cmd->io_request->RaidContext.raid_context_g35.smid.peer_smid = + cpu_to_le16(r1_cmd->index); + r1_cmd->io_request->RaidContext.raid_context_g35.smid.peer_smid = + cpu_to_le16(cmd->index); + /*MSIxIndex of both commands request descriptors should be same*/ + r1_cmd->request_desc->SCSIIO.MSIxIndex = + cmd->request_desc->SCSIIO.MSIxIndex; + /*span arm is different for r1 cmd*/ + r1_cmd->io_request->RaidContext.raid_context_g35.span_arm = + cmd->io_request->RaidContext.raid_context_g35.span_arm + 1; +} + /** * megasas_build_and_issue_cmd_fusion -Main routine for building and * issuing non IOCTL cmd @@ -2178,7 +2830,7 @@ static u32 megasas_build_and_issue_cmd_fusion(struct megasas_instance *instance, struct scsi_cmnd *scmd) { - struct megasas_cmd_fusion *cmd; + struct megasas_cmd_fusion *cmd, *r1_cmd = NULL; union MEGASAS_REQUEST_DESCRIPTOR_UNION *req_desc; u32 index; struct fusion_context *fusion; @@ -2193,13 +2845,22 @@ megasas_build_and_issue_cmd_fusion(struct megasas_instance *instance, return SCSI_MLQUEUE_DEVICE_BUSY; } + if (atomic_inc_return(&instance->fw_outstanding) > + instance->host->can_queue) { + atomic_dec(&instance->fw_outstanding); + return SCSI_MLQUEUE_HOST_BUSY; + } + cmd = megasas_get_cmd_fusion(instance, scmd->request->tag); + if (!cmd) { + atomic_dec(&instance->fw_outstanding); + return SCSI_MLQUEUE_HOST_BUSY; + } + index = cmd->index; req_desc = megasas_get_request_descriptor(instance, index-1); - if (!req_desc) - return SCSI_MLQUEUE_HOST_BUSY; req_desc->Words = 0; cmd->request_desc = req_desc; @@ -2208,6 +2869,7 @@ megasas_build_and_issue_cmd_fusion(struct megasas_instance *instance, megasas_return_cmd_fusion(instance, cmd); dev_err(&instance->pdev->dev, "Error building command\n"); cmd->request_desc = NULL; + atomic_dec(&instance->fw_outstanding); return SCSI_MLQUEUE_HOST_BUSY; } @@ -2218,18 +2880,92 @@ megasas_build_and_issue_cmd_fusion(struct megasas_instance *instance, cmd->io_request->ChainOffset != 0xF) dev_err(&instance->pdev->dev, "The chain offset value is not " "correct : %x\n", cmd->io_request->ChainOffset); + /* + * if it is raid 1/10 fp write capable. + * try to get second command from pool and construct it. + * From FW, it has confirmed that lba values of two PDs + * corresponds to single R1/10 LD are always same + * + */ + /* driver side count always should be less than max_fw_cmds + * to get new command + */ + if (cmd->r1_alt_dev_handle != MR_DEVHANDLE_INVALID) { + r1_cmd = megasas_get_cmd_fusion(instance, + (scmd->request->tag + instance->max_fw_cmds)); + megasas_prepare_secondRaid1_IO(instance, cmd, r1_cmd); + } + /* * Issue the command to the FW */ - atomic_inc(&instance->fw_outstanding); megasas_fire_cmd_fusion(instance, req_desc); + if (r1_cmd) + megasas_fire_cmd_fusion(instance, r1_cmd->request_desc); + + return 0; } /** + * megasas_complete_r1_command - + * completes R1 FP write commands which has valid peer smid + * @instance: Adapter soft state + * @cmd_fusion: MPT command frame + * + */ +static inline void +megasas_complete_r1_command(struct megasas_instance *instance, + struct megasas_cmd_fusion *cmd) +{ + u8 *sense, status, ex_status; + u32 data_length; + u16 peer_smid; + struct fusion_context *fusion; + struct megasas_cmd_fusion *r1_cmd = NULL; + struct scsi_cmnd *scmd_local = NULL; + struct RAID_CONTEXT_G35 *rctx_g35; + + rctx_g35 = &cmd->io_request->RaidContext.raid_context_g35; + fusion = instance->ctrl_context; + peer_smid = le16_to_cpu(rctx_g35->smid.peer_smid); + + r1_cmd = fusion->cmd_list[peer_smid - 1]; + scmd_local = cmd->scmd; + status = rctx_g35->status; + ex_status = rctx_g35->ex_status; + data_length = cmd->io_request->DataLength; + sense = cmd->sense; + + cmd->cmd_completed = true; + + /* Check if peer command is completed or not*/ + if (r1_cmd->cmd_completed) { + rctx_g35 = &r1_cmd->io_request->RaidContext.raid_context_g35; + if (rctx_g35->status != MFI_STAT_OK) { + status = rctx_g35->status; + ex_status = rctx_g35->ex_status; + data_length = r1_cmd->io_request->DataLength; + sense = r1_cmd->sense; + } + + megasas_return_cmd_fusion(instance, r1_cmd); + map_cmd_status(fusion, scmd_local, status, ex_status, + le32_to_cpu(data_length), sense); + if (instance->ldio_threshold && + megasas_cmd_type(scmd_local) == READ_WRITE_LDIO) + atomic_dec(&instance->ldio_outstanding); + scmd_local->SCp.ptr = NULL; + megasas_return_cmd_fusion(instance, cmd); + scsi_dma_unmap(scmd_local); + scmd_local->scsi_done(scmd_local); + } +} + +/** * complete_cmd_fusion - Completes command * @instance: Adapter soft state * Completes all commands that is in reply descriptor queue @@ -2244,8 +2980,8 @@ complete_cmd_fusion(struct megasas_instance *instance, u32 MSIxIndex) struct megasas_cmd *cmd_mfi; struct megasas_cmd_fusion *cmd_fusion; u16 smid, num_completed; - u8 reply_descript_type; - u32 status, extStatus, device_id; + u8 reply_descript_type, *sense, status, extStatus; + u32 device_id, data_length; union desc_value d_val; struct LD_LOAD_BALANCE_INFO *lbinfo; int threshold_reply_count = 0; @@ -2275,20 +3011,17 @@ complete_cmd_fusion(struct megasas_instance *instance, u32 MSIxIndex) while (d_val.u.low != cpu_to_le32(UINT_MAX) && d_val.u.high != cpu_to_le32(UINT_MAX)) { - smid = le16_to_cpu(reply_desc->SMID); + smid = le16_to_cpu(reply_desc->SMID); cmd_fusion = fusion->cmd_list[smid - 1]; - - scsi_io_req = - (struct MPI2_RAID_SCSI_IO_REQUEST *) - cmd_fusion->io_request; - - if (cmd_fusion->scmd) - cmd_fusion->scmd->SCp.ptr = NULL; + scsi_io_req = (struct MPI2_RAID_SCSI_IO_REQUEST *) + cmd_fusion->io_request; scmd_local = cmd_fusion->scmd; - status = scsi_io_req->RaidContext.status; - extStatus = scsi_io_req->RaidContext.exStatus; + status = scsi_io_req->RaidContext.raid_context.status; + extStatus = scsi_io_req->RaidContext.raid_context.ex_status; + sense = cmd_fusion->sense; + data_length = scsi_io_req->DataLength; switch (scsi_io_req->Function) { case MPI2_FUNCTION_SCSI_TASK_MGMT: @@ -2303,37 +3036,33 @@ complete_cmd_fusion(struct megasas_instance *instance, u32 MSIxIndex) break; case MPI2_FUNCTION_SCSI_IO_REQUEST: /*Fast Path IO.*/ /* Update load balancing info */ - device_id = MEGASAS_DEV_INDEX(scmd_local); - lbinfo = &fusion->load_balance_info[device_id]; - if (cmd_fusion->scmd->SCp.Status & - MEGASAS_LOAD_BALANCE_FLAG) { + if (fusion->load_balance_info && + (cmd_fusion->scmd->SCp.Status & + MEGASAS_LOAD_BALANCE_FLAG)) { + device_id = MEGASAS_DEV_INDEX(scmd_local); + lbinfo = &fusion->load_balance_info[device_id]; atomic_dec(&lbinfo->scsi_pending_cmds[cmd_fusion->pd_r1_lb]); - cmd_fusion->scmd->SCp.Status &= - ~MEGASAS_LOAD_BALANCE_FLAG; + cmd_fusion->scmd->SCp.Status &= ~MEGASAS_LOAD_BALANCE_FLAG; } - if (reply_descript_type == - MPI2_RPY_DESCRIPT_FLAGS_SCSI_IO_SUCCESS) { - if (megasas_dbg_lvl == 5) - dev_err(&instance->pdev->dev, "\nFAST Path " - "IO Success\n"); - } - /* Fall thru and complete IO */ + //Fall thru and complete IO case MEGASAS_MPI2_FUNCTION_LD_IO_REQUEST: /* LD-IO Path */ - /* Map the FW Cmd Status */ - map_cmd_status(cmd_fusion, status, extStatus); - scsi_io_req->RaidContext.status = 0; - scsi_io_req->RaidContext.exStatus = 0; - if (megasas_cmd_type(scmd_local) == READ_WRITE_LDIO) - atomic_dec(&instance->ldio_outstanding); - megasas_return_cmd_fusion(instance, cmd_fusion); - scsi_dma_unmap(scmd_local); - scmd_local->scsi_done(scmd_local); atomic_dec(&instance->fw_outstanding); - + if (cmd_fusion->r1_alt_dev_handle == MR_DEVHANDLE_INVALID) { + map_cmd_status(fusion, scmd_local, status, + extStatus, le32_to_cpu(data_length), + sense); + if (instance->ldio_threshold && + (megasas_cmd_type(scmd_local) == READ_WRITE_LDIO)) + atomic_dec(&instance->ldio_outstanding); + scmd_local->SCp.ptr = NULL; + megasas_return_cmd_fusion(instance, cmd_fusion); + scsi_dma_unmap(scmd_local); + scmd_local->scsi_done(scmd_local); + } else /* Optimal VD - R1 FP command completion. */ + megasas_complete_r1_command(instance, cmd_fusion); break; case MEGASAS_MPI2_FUNCTION_PASSTHRU_IO_REQUEST: /*MFI command */ cmd_mfi = instance->cmd_list[cmd_fusion->sync_cmd_idx]; - /* Poll mode. Dummy free. * In case of Interrupt mode, caller has reverse check. */ @@ -2376,7 +3105,7 @@ complete_cmd_fusion(struct megasas_instance *instance, u32 MSIxIndex) * pending to be completed */ if (threshold_reply_count >= THRESHOLD_REPLY_COUNT) { - if (fusion->adapter_type == INVADER_SERIES) + if (instance->msix_combined) writel(((MSIxIndex & 0x7) << 24) | fusion->last_reply_idx[MSIxIndex], instance->reply_post_host_index_addr[MSIxIndex/8]); @@ -2392,7 +3121,7 @@ complete_cmd_fusion(struct megasas_instance *instance, u32 MSIxIndex) return IRQ_NONE; wmb(); - if (fusion->adapter_type == INVADER_SERIES) + if (instance->msix_combined) writel(((MSIxIndex & 0x7) << 24) | fusion->last_reply_idx[MSIxIndex], instance->reply_post_host_index_addr[MSIxIndex/8]); @@ -2405,6 +3134,22 @@ complete_cmd_fusion(struct megasas_instance *instance, u32 MSIxIndex) } /** + * megasas_sync_irqs - Synchronizes all IRQs owned by adapter + * @instance: Adapter soft state + */ +void megasas_sync_irqs(unsigned long instance_addr) +{ + u32 count, i; + struct megasas_instance *instance = + (struct megasas_instance *)instance_addr; + + count = instance->msix_vectors > 0 ? instance->msix_vectors : 1; + + for (i = 0; i < count; i++) + synchronize_irq(pci_irq_vector(instance->pdev, i)); +} + +/** * megasas_complete_cmd_dpc_fusion - Completes command * @instance: Adapter soft state * @@ -2489,7 +3234,7 @@ irqreturn_t megasas_isr_fusion(int irq, void *devp) * mfi_cmd: megasas_cmd pointer * */ -u8 +void build_mpt_mfi_pass_thru(struct megasas_instance *instance, struct megasas_cmd *mfi_cmd) { @@ -2518,7 +3263,7 @@ build_mpt_mfi_pass_thru(struct megasas_instance *instance, io_req = cmd->io_request; - if (fusion->adapter_type == INVADER_SERIES) { + if (fusion->adapter_type >= INVADER_SERIES) { struct MPI25_IEEE_SGE_CHAIN64 *sgl_ptr_end = (struct MPI25_IEEE_SGE_CHAIN64 *)&io_req->SGL; sgl_ptr_end += fusion->max_sge_in_main_msg - 1; @@ -2539,8 +3284,6 @@ build_mpt_mfi_pass_thru(struct megasas_instance *instance, MPI2_IEEE_SGE_FLAGS_IOCPLBNTA_ADDR; mpi25_ieee_chain->Length = cpu_to_le32(instance->max_chain_frame_sz); - - return 0; } /** @@ -2552,21 +3295,14 @@ build_mpt_mfi_pass_thru(struct megasas_instance *instance, union MEGASAS_REQUEST_DESCRIPTOR_UNION * build_mpt_cmd(struct megasas_instance *instance, struct megasas_cmd *cmd) { - union MEGASAS_REQUEST_DESCRIPTOR_UNION *req_desc; + union MEGASAS_REQUEST_DESCRIPTOR_UNION *req_desc = NULL; u16 index; - if (build_mpt_mfi_pass_thru(instance, cmd)) { - dev_err(&instance->pdev->dev, "Couldn't build MFI pass thru cmd\n"); - return NULL; - } - + build_mpt_mfi_pass_thru(instance, cmd); index = cmd->context.smid; req_desc = megasas_get_request_descriptor(instance, index - 1); - if (!req_desc) - return NULL; - req_desc->Words = 0; req_desc->SCSIIO.RequestFlags = (MPI2_REQ_DESCRIPT_FLAGS_SCSI_IO << MEGASAS_REQ_DESCRIPT_FLAGS_TYPE_SHIFT); @@ -2582,21 +3318,16 @@ build_mpt_cmd(struct megasas_instance *instance, struct megasas_cmd *cmd) * @cmd: mfi cmd pointer * */ -int +void megasas_issue_dcmd_fusion(struct megasas_instance *instance, struct megasas_cmd *cmd) { union MEGASAS_REQUEST_DESCRIPTOR_UNION *req_desc; req_desc = build_mpt_cmd(instance, cmd); - if (!req_desc) { - dev_info(&instance->pdev->dev, "Failed from %s %d\n", - __func__, __LINE__); - return DCMD_NOT_FIRED; - } megasas_fire_cmd_fusion(instance, req_desc); - return DCMD_SUCCESS; + return; } /** @@ -2771,6 +3502,14 @@ int megasas_wait_for_outstanding_fusion(struct megasas_instance *instance, " will reset adapter scsi%d.\n", instance->host->host_no); megasas_complete_cmd_dpc_fusion((unsigned long)instance); + if (instance->requestorId && reason) { + dev_warn(&instance->pdev->dev, "SR-IOV Found FW in FAULT" + " state while polling during" + " I/O timeout handling for %d\n", + instance->host->host_no); + *convert = 1; + } + retval = 1; goto out; } @@ -2790,7 +3529,7 @@ int megasas_wait_for_outstanding_fusion(struct megasas_instance *instance, } /* If SR-IOV VF mode & I/O timeout, check for HB timeout */ - if (instance->requestorId && reason) { + if (instance->requestorId && (reason == SCSIIO_TIMEOUT_OCR)) { if (instance->hb_host_mem->HB.fwCounter != instance->hb_host_mem->HB.driverCounter) { instance->hb_host_mem->HB.driverCounter = @@ -3030,12 +3769,6 @@ megasas_issue_tm(struct megasas_instance *instance, u16 device_handle, req_desc = megasas_get_request_descriptor(instance, (cmd_fusion->index - 1)); - if (!req_desc) { - dev_err(&instance->pdev->dev, "Failed from %s %d\n", - __func__, __LINE__); - megasas_return_cmd(instance, cmd_mfi); - return -ENOMEM; - } cmd_fusion->request_desc = req_desc; req_desc->Words = 0; @@ -3092,7 +3825,7 @@ megasas_issue_tm(struct megasas_instance *instance, u16 device_handle, break; else { instance->instancet->disable_intr(instance); - msleep(1000); + megasas_sync_irqs((unsigned long)instance); megasas_complete_cmd_dpc_fusion ((unsigned long)instance); instance->instancet->enable_intr(instance); @@ -3173,13 +3906,13 @@ static u16 megasas_get_tm_devhandle(struct scsi_device *sdev) instance = (struct megasas_instance *)sdev->host->hostdata; fusion = instance->ctrl_context; - if (sdev->channel < MEGASAS_MAX_PD_CHANNELS) { + if (!MEGASAS_IS_LOGICAL(sdev)) { if (instance->use_seqnum_jbod_fp) { - pd_index = (sdev->channel * MEGASAS_MAX_DEV_PER_CHANNEL) + - sdev->id; - pd_sync = (void *)fusion->pd_seq_sync - [(instance->pd_seq_map_id - 1) & 1]; - devhandle = pd_sync->seq[pd_index].devHandle; + pd_index = (sdev->channel * MEGASAS_MAX_DEV_PER_CHANNEL) + + sdev->id; + pd_sync = (void *)fusion->pd_seq_sync + [(instance->pd_seq_map_id - 1) & 1]; + devhandle = pd_sync->seq[pd_index].devHandle; } else sdev_printk(KERN_ERR, sdev, "Firmware expose tmCapable" " without JBOD MAP support from %s %d\n", __func__, __LINE__); @@ -3212,6 +3945,9 @@ int megasas_task_abort_fusion(struct scsi_cmnd *scmd) instance = (struct megasas_instance *)scmd->device->host->hostdata; fusion = instance->ctrl_context; + scmd_printk(KERN_INFO, scmd, "task abort called for scmd(%p)\n", scmd); + scsi_print_command(scmd); + if (atomic_read(&instance->adprecovery) != MEGASAS_HBA_OPERATIONAL) { dev_err(&instance->pdev->dev, "Controller is not OPERATIONAL," "SCSI host:%d\n", instance->host->host_no); @@ -3292,6 +4028,9 @@ int megasas_reset_target_fusion(struct scsi_cmnd *scmd) instance = (struct megasas_instance *)scmd->device->host->hostdata; fusion = instance->ctrl_context; + sdev_printk(KERN_INFO, scmd->device, + "target reset called for scmd(%p)\n", scmd); + if (atomic_read(&instance->adprecovery) != MEGASAS_HBA_OPERATIONAL) { dev_err(&instance->pdev->dev, "Controller is not OPERATIONAL," "SCSI host:%d\n", instance->host->host_no); @@ -3362,7 +4101,7 @@ int megasas_check_mpio_paths(struct megasas_instance *instance, struct scsi_cmnd *scmd) { struct megasas_instance *peer_instance = NULL; - int retval = (DID_RESET << 16); + int retval = (DID_REQUEUE << 16); if (instance->peerIsPresent) { peer_instance = megasas_get_peer_instance(instance); @@ -3377,9 +4116,9 @@ int megasas_check_mpio_paths(struct megasas_instance *instance, /* Core fusion reset function */ int megasas_reset_fusion(struct Scsi_Host *shost, int reason) { - int retval = SUCCESS, i, convert = 0; + int retval = SUCCESS, i, j, convert = 0; struct megasas_instance *instance; - struct megasas_cmd_fusion *cmd_fusion; + struct megasas_cmd_fusion *cmd_fusion, *r1_cmd; struct fusion_context *fusion; u32 abs_state, status_reg, reset_adapter; u32 io_timeout_in_crash_mode = 0; @@ -3440,7 +4179,7 @@ int megasas_reset_fusion(struct Scsi_Host *shost, int reason) set_bit(MEGASAS_FUSION_IN_RESET, &instance->reset_flags); atomic_set(&instance->adprecovery, MEGASAS_ADPRESET_SM_POLLING); instance->instancet->disable_intr(instance); - msleep(1000); + megasas_sync_irqs((unsigned long)instance); /* First try waiting for commands to complete */ if (megasas_wait_for_outstanding_fusion(instance, reason, @@ -3451,23 +4190,40 @@ int megasas_reset_fusion(struct Scsi_Host *shost, int reason) if (convert) reason = 0; + if (megasas_dbg_lvl & OCR_LOGS) + dev_info(&instance->pdev->dev, "\nPending SCSI commands:\n"); + /* Now return commands back to the OS */ for (i = 0 ; i < instance->max_scsi_cmds; i++) { cmd_fusion = fusion->cmd_list[i]; + /*check for extra commands issued by driver*/ + if (instance->is_ventura) { + r1_cmd = fusion->cmd_list[i + instance->max_fw_cmds]; + megasas_return_cmd_fusion(instance, r1_cmd); + } scmd_local = cmd_fusion->scmd; if (cmd_fusion->scmd) { + if (megasas_dbg_lvl & OCR_LOGS) { + sdev_printk(KERN_INFO, + cmd_fusion->scmd->device, "SMID: 0x%x\n", + cmd_fusion->index); + scsi_print_command(cmd_fusion->scmd); + } + scmd_local->result = megasas_check_mpio_paths(instance, scmd_local); - if (megasas_cmd_type(scmd_local) == READ_WRITE_LDIO) + if (instance->ldio_threshold && + megasas_cmd_type(scmd_local) == READ_WRITE_LDIO) atomic_dec(&instance->ldio_outstanding); megasas_return_cmd_fusion(instance, cmd_fusion); scsi_dma_unmap(scmd_local); scmd_local->scsi_done(scmd_local); - atomic_dec(&instance->fw_outstanding); } } + atomic_set(&instance->fw_outstanding, 0); + status_reg = instance->instancet->read_fw_status_reg( instance->reg_set); abs_state = status_reg & MFI_STATE_MASK; @@ -3528,11 +4284,13 @@ transition_to_ready: __func__, __LINE__); megaraid_sas_kill_hba(instance); retval = FAILED; + goto out; } /* Reset load balance info */ - memset(fusion->load_balance_info, 0, - sizeof(struct LD_LOAD_BALANCE_INFO) - *MAX_LOGICAL_DRIVES_EXT); + if (fusion->load_balance_info) + memset(fusion->load_balance_info, 0, + (sizeof(struct LD_LOAD_BALANCE_INFO) * + MAX_LOGICAL_DRIVES_EXT)); if (!megasas_get_map_info(instance)) megasas_sync_map_info(instance); @@ -3540,7 +4298,17 @@ transition_to_ready: megasas_setup_jbod_map(instance); shost_for_each_device(sdev, shost) - megasas_update_sdev_properties(sdev); + megasas_set_dynamic_target_properties(sdev); + + /* reset stream detection array */ + if (instance->is_ventura) { + for (j = 0; j < MAX_LOGICAL_DRIVES_EXT; ++j) { + memset(fusion->stream_detect_by_ld[j], + 0, sizeof(struct LD_STREAM_DETECT)); + fusion->stream_detect_by_ld[j]->mru_bit_map + = MR_STREAM_BITMAP; + } + } clear_bit(MEGASAS_FUSION_IN_RESET, &instance->reset_flags); @@ -3676,6 +4444,64 @@ void megasas_fusion_ocr_wq(struct work_struct *work) megasas_reset_fusion(instance->host, 0); } +/* Allocate fusion context */ +int +megasas_alloc_fusion_context(struct megasas_instance *instance) +{ + struct fusion_context *fusion; + + instance->ctrl_context_pages = get_order(sizeof(struct fusion_context)); + instance->ctrl_context = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, + instance->ctrl_context_pages); + if (!instance->ctrl_context) { + /* fall back to using vmalloc for fusion_context */ + instance->ctrl_context = vzalloc(sizeof(struct fusion_context)); + if (!instance->ctrl_context) { + dev_err(&instance->pdev->dev, "Failed from %s %d\n", __func__, __LINE__); + return -ENOMEM; + } + } + + fusion = instance->ctrl_context; + + fusion->load_balance_info_pages = get_order(MAX_LOGICAL_DRIVES_EXT * + sizeof(struct LD_LOAD_BALANCE_INFO)); + fusion->load_balance_info = + (struct LD_LOAD_BALANCE_INFO *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, + fusion->load_balance_info_pages); + if (!fusion->load_balance_info) { + fusion->load_balance_info = vzalloc(MAX_LOGICAL_DRIVES_EXT * + sizeof(struct LD_LOAD_BALANCE_INFO)); + if (!fusion->load_balance_info) + dev_err(&instance->pdev->dev, "Failed to allocate load_balance_info, " + "continuing without Load Balance support\n"); + } + + return 0; +} + +void +megasas_free_fusion_context(struct megasas_instance *instance) +{ + struct fusion_context *fusion = instance->ctrl_context; + + if (fusion) { + if (fusion->load_balance_info) { + if (is_vmalloc_addr(fusion->load_balance_info)) + vfree(fusion->load_balance_info); + else + free_pages((ulong)fusion->load_balance_info, + fusion->load_balance_info_pages); + } + + if (is_vmalloc_addr(fusion)) + vfree(fusion); + else + free_pages((ulong)fusion, + instance->ctrl_context_pages); + } +} + struct megasas_instance_template megasas_instance_template_fusion = { .enable_intr = megasas_enable_intr_fusion, .disable_intr = megasas_disable_intr_fusion, |