Author | Tokens | Token Proportion | Commits | Commit Proportion |
---|---|---|---|---|
Marco Chiappero | 1280 | 95.03% | 13 | 76.47% |
Tadeusz Struk | 60 | 4.45% | 2 | 11.76% |
Giovanni Cabiddu | 7 | 0.52% | 2 | 11.76% |
Total | 1347 | 17 |
// SPDX-License-Identifier: (BSD-3-Clause OR GPL-2.0-only) /* Copyright(c) 2015 - 2021 Intel Corporation */ #include <linux/bitfield.h> #include <linux/completion.h> #include <linux/minmax.h> #include <linux/types.h> #include "adf_accel_devices.h" #include "adf_common_drv.h" #include "adf_pfvf_msg.h" #include "adf_pfvf_utils.h" #include "adf_pfvf_vf_msg.h" #include "adf_pfvf_vf_proto.h" #define ADF_PFVF_MSG_COLLISION_DETECT_DELAY 10 #define ADF_PFVF_MSG_ACK_DELAY 2 #define ADF_PFVF_MSG_ACK_MAX_RETRY 100 /* How often to retry if there is no response */ #define ADF_PFVF_MSG_RESP_RETRIES 5 #define ADF_PFVF_MSG_RESP_TIMEOUT (ADF_PFVF_MSG_ACK_DELAY * \ ADF_PFVF_MSG_ACK_MAX_RETRY + \ ADF_PFVF_MSG_COLLISION_DETECT_DELAY) /** * adf_send_vf2pf_msg() - send VF to PF message * @accel_dev: Pointer to acceleration device * @msg: Message to send * * This function allows the VF to send a message to the PF. * * Return: 0 on success, error code otherwise. */ int adf_send_vf2pf_msg(struct adf_accel_dev *accel_dev, struct pfvf_message msg) { struct adf_pfvf_ops *pfvf_ops = GET_PFVF_OPS(accel_dev); u32 pfvf_offset = pfvf_ops->get_vf2pf_offset(0); return pfvf_ops->send_msg(accel_dev, msg, pfvf_offset, &accel_dev->vf.vf2pf_lock); } /** * adf_recv_pf2vf_msg() - receive a PF to VF message * @accel_dev: Pointer to acceleration device * * This function allows the VF to receive a message from the PF. * * Return: a valid message on success, zero otherwise. */ static struct pfvf_message adf_recv_pf2vf_msg(struct adf_accel_dev *accel_dev) { struct adf_pfvf_ops *pfvf_ops = GET_PFVF_OPS(accel_dev); u32 pfvf_offset = pfvf_ops->get_pf2vf_offset(0); return pfvf_ops->recv_msg(accel_dev, pfvf_offset, accel_dev->vf.pf_compat_ver); } /** * adf_send_vf2pf_req() - send VF2PF request message * @accel_dev: Pointer to acceleration device. * @msg: Request message to send * @resp: Returned PF response * * This function sends a message that requires a response from the VF to the PF * and waits for a reply. * * Return: 0 on success, error code otherwise. */ int adf_send_vf2pf_req(struct adf_accel_dev *accel_dev, struct pfvf_message msg, struct pfvf_message *resp) { unsigned long timeout = msecs_to_jiffies(ADF_PFVF_MSG_RESP_TIMEOUT); unsigned int retries = ADF_PFVF_MSG_RESP_RETRIES; int ret; reinit_completion(&accel_dev->vf.msg_received); /* Send request from VF to PF */ do { ret = adf_send_vf2pf_msg(accel_dev, msg); if (ret) { dev_err(&GET_DEV(accel_dev), "Failed to send request msg to PF\n"); return ret; } /* Wait for response, if it times out retry */ ret = wait_for_completion_timeout(&accel_dev->vf.msg_received, timeout); if (ret) { if (likely(resp)) *resp = accel_dev->vf.response; /* Once copied, set to an invalid value */ accel_dev->vf.response.type = 0; return 0; } dev_err(&GET_DEV(accel_dev), "PFVF response message timeout\n"); } while (--retries); return -EIO; } static int adf_vf2pf_blkmsg_data_req(struct adf_accel_dev *accel_dev, bool crc, u8 *type, u8 *data) { struct pfvf_message req = { 0 }; struct pfvf_message resp = { 0 }; u8 blk_type; u8 blk_byte; u8 msg_type; u8 max_data; int err; /* Convert the block type to {small, medium, large} size category */ if (*type <= ADF_VF2PF_SMALL_BLOCK_TYPE_MAX) { msg_type = ADF_VF2PF_MSGTYPE_SMALL_BLOCK_REQ; blk_type = FIELD_PREP(ADF_VF2PF_SMALL_BLOCK_TYPE_MASK, *type); blk_byte = FIELD_PREP(ADF_VF2PF_SMALL_BLOCK_BYTE_MASK, *data); max_data = ADF_VF2PF_SMALL_BLOCK_BYTE_MAX; } else if (*type <= ADF_VF2PF_MEDIUM_BLOCK_TYPE_MAX) { msg_type = ADF_VF2PF_MSGTYPE_MEDIUM_BLOCK_REQ; blk_type = FIELD_PREP(ADF_VF2PF_MEDIUM_BLOCK_TYPE_MASK, *type - ADF_VF2PF_SMALL_BLOCK_TYPE_MAX); blk_byte = FIELD_PREP(ADF_VF2PF_MEDIUM_BLOCK_BYTE_MASK, *data); max_data = ADF_VF2PF_MEDIUM_BLOCK_BYTE_MAX; } else if (*type <= ADF_VF2PF_LARGE_BLOCK_TYPE_MAX) { msg_type = ADF_VF2PF_MSGTYPE_LARGE_BLOCK_REQ; blk_type = FIELD_PREP(ADF_VF2PF_LARGE_BLOCK_TYPE_MASK, *type - ADF_VF2PF_MEDIUM_BLOCK_TYPE_MAX); blk_byte = FIELD_PREP(ADF_VF2PF_LARGE_BLOCK_BYTE_MASK, *data); max_data = ADF_VF2PF_LARGE_BLOCK_BYTE_MAX; } else { dev_err(&GET_DEV(accel_dev), "Invalid message type %u\n", *type); return -EINVAL; } /* Sanity check */ if (*data > max_data) { dev_err(&GET_DEV(accel_dev), "Invalid byte %s %u for message type %u\n", crc ? "count" : "index", *data, *type); return -EINVAL; } /* Build the block message */ req.type = msg_type; req.data = blk_type | blk_byte | FIELD_PREP(ADF_VF2PF_BLOCK_CRC_REQ_MASK, crc); err = adf_send_vf2pf_req(accel_dev, req, &resp); if (err) return err; *type = FIELD_GET(ADF_PF2VF_BLKMSG_RESP_TYPE_MASK, resp.data); *data = FIELD_GET(ADF_PF2VF_BLKMSG_RESP_DATA_MASK, resp.data); return 0; } static int adf_vf2pf_blkmsg_get_byte(struct adf_accel_dev *accel_dev, u8 type, u8 index, u8 *data) { int ret; ret = adf_vf2pf_blkmsg_data_req(accel_dev, false, &type, &index); if (ret < 0) return ret; if (unlikely(type != ADF_PF2VF_BLKMSG_RESP_TYPE_DATA)) { dev_err(&GET_DEV(accel_dev), "Unexpected BLKMSG response type %u, byte 0x%x\n", type, index); return -EFAULT; } *data = index; return 0; } static int adf_vf2pf_blkmsg_get_crc(struct adf_accel_dev *accel_dev, u8 type, u8 bytes, u8 *crc) { int ret; /* The count of bytes refers to a length, however shift it to a 0-based * count to avoid overflows. Thus, a request for 0 bytes is technically * valid. */ --bytes; ret = adf_vf2pf_blkmsg_data_req(accel_dev, true, &type, &bytes); if (ret < 0) return ret; if (unlikely(type != ADF_PF2VF_BLKMSG_RESP_TYPE_CRC)) { dev_err(&GET_DEV(accel_dev), "Unexpected CRC BLKMSG response type %u, crc 0x%x\n", type, bytes); return -EFAULT; } *crc = bytes; return 0; } /** * adf_send_vf2pf_blkmsg_req() - retrieve block message * @accel_dev: Pointer to acceleration VF device. * @type: The block message type, see adf_pfvf_msg.h for allowed values * @buffer: input buffer where to place the received data * @buffer_len: buffer length as input, the amount of written bytes on output * * Request a message of type 'type' over the block message transport. * This function will send the required amount block message requests and * return the overall content back to the caller through the provided buffer. * The buffer should be large enough to contain the requested message type, * otherwise the response will be truncated. * * Return: 0 on success, error code otherwise. */ int adf_send_vf2pf_blkmsg_req(struct adf_accel_dev *accel_dev, u8 type, u8 *buffer, unsigned int *buffer_len) { unsigned int index; unsigned int msg_len; int ret; u8 remote_crc; u8 local_crc; if (unlikely(type > ADF_VF2PF_LARGE_BLOCK_TYPE_MAX)) { dev_err(&GET_DEV(accel_dev), "Invalid block message type %d\n", type); return -EINVAL; } if (unlikely(*buffer_len < ADF_PFVF_BLKMSG_HEADER_SIZE)) { dev_err(&GET_DEV(accel_dev), "Buffer size too small for a block message\n"); return -EINVAL; } ret = adf_vf2pf_blkmsg_get_byte(accel_dev, type, ADF_PFVF_BLKMSG_VER_BYTE, &buffer[ADF_PFVF_BLKMSG_VER_BYTE]); if (unlikely(ret)) return ret; if (unlikely(!buffer[ADF_PFVF_BLKMSG_VER_BYTE])) { dev_err(&GET_DEV(accel_dev), "Invalid version 0 received for block request %u", type); return -EFAULT; } ret = adf_vf2pf_blkmsg_get_byte(accel_dev, type, ADF_PFVF_BLKMSG_LEN_BYTE, &buffer[ADF_PFVF_BLKMSG_LEN_BYTE]); if (unlikely(ret)) return ret; if (unlikely(!buffer[ADF_PFVF_BLKMSG_LEN_BYTE])) { dev_err(&GET_DEV(accel_dev), "Invalid size 0 received for block request %u", type); return -EFAULT; } /* We need to pick the minimum since there is no way to request a * specific version. As a consequence any scenario is possible: * - PF has a newer (longer) version which doesn't fit in the buffer * - VF expects a newer (longer) version, so we must not ask for * bytes in excess * - PF and VF share the same version, no problem */ msg_len = ADF_PFVF_BLKMSG_HEADER_SIZE + buffer[ADF_PFVF_BLKMSG_LEN_BYTE]; msg_len = min(*buffer_len, msg_len); /* Get the payload */ for (index = ADF_PFVF_BLKMSG_HEADER_SIZE; index < msg_len; index++) { ret = adf_vf2pf_blkmsg_get_byte(accel_dev, type, index, &buffer[index]); if (unlikely(ret)) return ret; } ret = adf_vf2pf_blkmsg_get_crc(accel_dev, type, msg_len, &remote_crc); if (unlikely(ret)) return ret; local_crc = adf_pfvf_calc_blkmsg_crc(buffer, msg_len); if (unlikely(local_crc != remote_crc)) { dev_err(&GET_DEV(accel_dev), "CRC error on msg type %d. Local %02X, remote %02X\n", type, local_crc, remote_crc); return -EIO; } *buffer_len = msg_len; return 0; } static bool adf_handle_pf2vf_msg(struct adf_accel_dev *accel_dev, struct pfvf_message msg) { switch (msg.type) { case ADF_PF2VF_MSGTYPE_RESTARTING: dev_dbg(&GET_DEV(accel_dev), "Restarting message received from PF\n"); adf_pf2vf_handle_pf_restarting(accel_dev); return false; case ADF_PF2VF_MSGTYPE_VERSION_RESP: case ADF_PF2VF_MSGTYPE_BLKMSG_RESP: case ADF_PF2VF_MSGTYPE_RP_RESET_RESP: dev_dbg(&GET_DEV(accel_dev), "Response Message received from PF (type 0x%.4x, data 0x%.4x)\n", msg.type, msg.data); accel_dev->vf.response = msg; complete(&accel_dev->vf.msg_received); return true; default: dev_err(&GET_DEV(accel_dev), "Unknown message from PF (type 0x%.4x, data: 0x%.4x)\n", msg.type, msg.data); } return false; } bool adf_recv_and_handle_pf2vf_msg(struct adf_accel_dev *accel_dev) { struct pfvf_message msg; msg = adf_recv_pf2vf_msg(accel_dev); if (msg.type) /* Invalid or no message */ return adf_handle_pf2vf_msg(accel_dev, msg); /* No replies for PF->VF messages at present */ return true; } /** * adf_enable_vf2pf_comms() - Function enables communication from vf to pf * * @accel_dev: Pointer to acceleration device virtual function. * * Return: 0 on success, error code otherwise. */ int adf_enable_vf2pf_comms(struct adf_accel_dev *accel_dev) { int ret; adf_pfvf_crc_init(); adf_enable_pf2vf_interrupts(accel_dev); ret = adf_vf2pf_request_version(accel_dev); if (ret) return ret; ret = adf_vf2pf_get_capabilities(accel_dev); if (ret) return ret; ret = adf_vf2pf_get_ring_to_svc(accel_dev); return ret; } EXPORT_SYMBOL_GPL(adf_enable_vf2pf_comms);
Information contained on this website is for historical information purposes only and does not indicate or represent copyright ownership.
Created with Cregit http://github.com/cregit/cregit
Version 2.0-RC1