diff options
author | K. Y. Srinivasan <kys@microsoft.com> | 2012-02-09 14:14:48 -0800 |
---|---|---|
committer | Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2012-02-13 11:35:20 -0800 |
commit | 89ae7d709357df3a60e81c57613859d91d093ace (patch) | |
tree | 1bb00c35612ea48371380a9e9348a80c81c86dd7 /drivers/scsi | |
parent | 0cbb613fa82fb09b6ef914ab0aaa92091f367bb6 (diff) |
Staging: hv: storvsc: Move the storage driver out of the staging area
The storage driver (storvsc_drv.c) handles all block storage devices
assigned to Linux guests hosted on Hyper-V. This driver has been in the
staging tree for a while and this patch moves it out of the staging area.
James was willing to apply this patch during the 3.3-rc phase and a decision
was taken to defer this to 3.4 since Greg had queued up a bunch of storvsc
patches for 3.4. Now that Greg has applied all of the pending storvsc patches,
I am sending this patch to move this driver out of staging. Based on James'
recommendation, this patch gets rid of the unneeded files in the staging/hv
directory.
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Acked-by: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Diffstat (limited to 'drivers/scsi')
-rw-r--r-- | drivers/scsi/Kconfig | 7 | ||||
-rw-r--r-- | drivers/scsi/Makefile | 3 | ||||
-rw-r--r-- | drivers/scsi/storvsc_drv.c | 1548 |
3 files changed, 1558 insertions, 0 deletions
diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig index 16570aa84aac..d3d18e89cb57 100644 --- a/drivers/scsi/Kconfig +++ b/drivers/scsi/Kconfig @@ -662,6 +662,13 @@ config VMWARE_PVSCSI To compile this driver as a module, choose M here: the module will be called vmw_pvscsi. +config HYPERV_STORAGE + tristate "Microsoft Hyper-V virtual storage driver" + depends on SCSI && HYPERV + default HYPERV + help + Select this option to enable the Hyper-V virtual storage driver. + config LIBFC tristate "LibFC module" select SCSI_FC_ATTRS diff --git a/drivers/scsi/Makefile b/drivers/scsi/Makefile index 2b887498be50..e4c1a69f8fab 100644 --- a/drivers/scsi/Makefile +++ b/drivers/scsi/Makefile @@ -142,6 +142,7 @@ obj-$(CONFIG_SCSI_BNX2_ISCSI) += libiscsi.o bnx2i/ obj-$(CONFIG_BE2ISCSI) += libiscsi.o be2iscsi/ obj-$(CONFIG_SCSI_PMCRAID) += pmcraid.o obj-$(CONFIG_VMWARE_PVSCSI) += vmw_pvscsi.o +obj-$(CONFIG_HYPERV_STORAGE) += hv_storvsc.o obj-$(CONFIG_ARM) += arm/ @@ -170,6 +171,8 @@ scsi_mod-$(CONFIG_SCSI_PROC_FS) += scsi_proc.o scsi_mod-y += scsi_trace.o scsi_mod-$(CONFIG_PM) += scsi_pm.o +hv_storvsc-y := storvsc_drv.o + scsi_tgt-y += scsi_tgt_lib.o scsi_tgt_if.o sd_mod-objs := sd.o diff --git a/drivers/scsi/storvsc_drv.c b/drivers/scsi/storvsc_drv.c new file mode 100644 index 000000000000..695ffc36e02d --- /dev/null +++ b/drivers/scsi/storvsc_drv.c @@ -0,0 +1,1548 @@ +/* + * Copyright (c) 2009, Microsoft Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., 59 Temple + * Place - Suite 330, Boston, MA 02111-1307 USA. + * + * Authors: + * Haiyang Zhang <haiyangz@microsoft.com> + * Hank Janssen <hjanssen@microsoft.com> + * K. Y. Srinivasan <kys@microsoft.com> + */ + +#include <linux/kernel.h> +#include <linux/wait.h> +#include <linux/sched.h> +#include <linux/completion.h> +#include <linux/string.h> +#include <linux/mm.h> +#include <linux/delay.h> +#include <linux/init.h> +#include <linux/slab.h> +#include <linux/module.h> +#include <linux/device.h> +#include <linux/hyperv.h> +#include <linux/mempool.h> +#include <scsi/scsi.h> +#include <scsi/scsi_cmnd.h> +#include <scsi/scsi_host.h> +#include <scsi/scsi_device.h> +#include <scsi/scsi_tcq.h> +#include <scsi/scsi_eh.h> +#include <scsi/scsi_devinfo.h> +#include <scsi/scsi_dbg.h> + +/* + * All wire protocol details (storage protocol between the guest and the host) + * are consolidated here. + * + * Begin protocol definitions. + */ + +/* + * Version history: + * V1 Beta: 0.1 + * V1 RC < 2008/1/31: 1.0 + * V1 RC > 2008/1/31: 2.0 + * Win7: 4.2 + */ + +#define VMSTOR_CURRENT_MAJOR 4 +#define VMSTOR_CURRENT_MINOR 2 + + +/* Packet structure describing virtual storage requests. */ +enum vstor_packet_operation { + VSTOR_OPERATION_COMPLETE_IO = 1, + VSTOR_OPERATION_REMOVE_DEVICE = 2, + VSTOR_OPERATION_EXECUTE_SRB = 3, + VSTOR_OPERATION_RESET_LUN = 4, + VSTOR_OPERATION_RESET_ADAPTER = 5, + VSTOR_OPERATION_RESET_BUS = 6, + VSTOR_OPERATION_BEGIN_INITIALIZATION = 7, + VSTOR_OPERATION_END_INITIALIZATION = 8, + VSTOR_OPERATION_QUERY_PROTOCOL_VERSION = 9, + VSTOR_OPERATION_QUERY_PROPERTIES = 10, + VSTOR_OPERATION_ENUMERATE_BUS = 11, + VSTOR_OPERATION_MAXIMUM = 11 +}; + +/* + * Platform neutral description of a scsi request - + * this remains the same across the write regardless of 32/64 bit + * note: it's patterned off the SCSI_PASS_THROUGH structure + */ +#define STORVSC_MAX_CMD_LEN 0x10 +#define STORVSC_SENSE_BUFFER_SIZE 0x12 +#define STORVSC_MAX_BUF_LEN_WITH_PADDING 0x14 + +struct vmscsi_request { + u16 length; + u8 srb_status; + u8 scsi_status; + + u8 port_number; + u8 path_id; + u8 target_id; + u8 lun; + + u8 cdb_length; + u8 sense_info_length; + u8 data_in; + u8 reserved; + + u32 data_transfer_length; + + union { + u8 cdb[STORVSC_MAX_CMD_LEN]; + u8 sense_data[STORVSC_SENSE_BUFFER_SIZE]; + u8 reserved_array[STORVSC_MAX_BUF_LEN_WITH_PADDING]; + }; +} __attribute((packed)); + + +/* + * This structure is sent during the intialization phase to get the different + * properties of the channel. + */ +struct vmstorage_channel_properties { + u16 protocol_version; + u8 path_id; + u8 target_id; + + /* Note: port number is only really known on the client side */ + u32 port_number; + u32 flags; + u32 max_transfer_bytes; + + /* + * This id is unique for each channel and will correspond with + * vendor specific data in the inquiry data. + */ + + u64 unique_id; +} __packed; + +/* This structure is sent during the storage protocol negotiations. */ +struct vmstorage_protocol_version { + /* Major (MSW) and minor (LSW) version numbers. */ + u16 major_minor; + + /* + * Revision number is auto-incremented whenever this file is changed + * (See FILL_VMSTOR_REVISION macro above). Mismatch does not + * definitely indicate incompatibility--but it does indicate mismatched + * builds. + * This is only used on the windows side. Just set it to 0. + */ + u16 revision; +} __packed; + +/* Channel Property Flags */ +#define STORAGE_CHANNEL_REMOVABLE_FLAG 0x1 +#define STORAGE_CHANNEL_EMULATED_IDE_FLAG 0x2 + +struct vstor_packet { + /* Requested operation type */ + enum vstor_packet_operation operation; + + /* Flags - see below for values */ + u32 flags; + + /* Status of the request returned from the server side. */ + u32 status; + + /* Data payload area */ + union { + /* + * Structure used to forward SCSI commands from the + * client to the server. + */ + struct vmscsi_request vm_srb; + + /* Structure used to query channel properties. */ + struct vmstorage_channel_properties storage_channel_properties; + + /* Used during version negotiations. */ + struct vmstorage_protocol_version version; + }; +} __packed; + +/* + * Packet Flags: + * + * This flag indicates that the server should send back a completion for this + * packet. + */ + +#define REQUEST_COMPLETION_FLAG 0x1 + +/* Matches Windows-end */ +enum storvsc_request_type { + WRITE_TYPE = 0, + READ_TYPE, + UNKNOWN_TYPE, +}; + +/* + * SRB status codes and masks; a subset of the codes used here. + */ + +#define SRB_STATUS_AUTOSENSE_VALID 0x80 +#define SRB_STATUS_INVALID_LUN 0x20 +#define SRB_STATUS_SUCCESS 0x01 +#define SRB_STATUS_ERROR 0x04 + +/* + * This is the end of Protocol specific defines. + */ + + +/* + * We setup a mempool to allocate request structures for this driver + * on a per-lun basis. The following define specifies the number of + * elements in the pool. + */ + +#define STORVSC_MIN_BUF_NR 64 +static int storvsc_ringbuffer_size = (20 * PAGE_SIZE); + +module_param(storvsc_ringbuffer_size, int, S_IRUGO); +MODULE_PARM_DESC(storvsc_ringbuffer_size, "Ring buffer size (bytes)"); + +#define STORVSC_MAX_IO_REQUESTS 128 + +/* + * In Hyper-V, each port/path/target maps to 1 scsi host adapter. In + * reality, the path/target is not used (ie always set to 0) so our + * scsi host adapter essentially has 1 bus with 1 target that contains + * up to 256 luns. + */ +#define STORVSC_MAX_LUNS_PER_TARGET 64 +#define STORVSC_MAX_TARGETS 1 +#define STORVSC_MAX_CHANNELS 1 + + + +struct storvsc_cmd_request { + struct list_head entry; + struct scsi_cmnd *cmd; + + unsigned int bounce_sgl_count; + struct scatterlist *bounce_sgl; + + struct hv_device *device; + + /* Synchronize the request/response if needed */ + struct completion wait_event; + + unsigned char *sense_buffer; + struct hv_multipage_buffer data_buffer; + struct vstor_packet vstor_packet; +}; + + +/* A storvsc device is a device object that contains a vmbus channel */ +struct storvsc_device { + struct hv_device *device; + + bool destroy; + bool drain_notify; + atomic_t num_outstanding_req; + struct Scsi_Host *host; + + wait_queue_head_t waiting_to_drain; + + /* + * Each unique Port/Path/Target represents 1 channel ie scsi + * controller. In reality, the pathid, targetid is always 0 + * and the port is set by us + */ + unsigned int port_number; + unsigned char path_id; + unsigned char target_id; + + /* Used for vsc/vsp channel reset process */ + struct storvsc_cmd_request init_request; + struct storvsc_cmd_request reset_request; +}; + +struct stor_mem_pools { + struct kmem_cache *request_pool; + mempool_t *request_mempool; +}; + +struct hv_host_device { + struct hv_device *dev; + unsigned int port; + unsigned char path; + unsigned char target; +}; + +struct storvsc_scan_work { + struct work_struct work; + struct Scsi_Host *host; + uint lun; +}; + +static void storvsc_bus_scan(struct work_struct *work) +{ + struct storvsc_scan_work *wrk; + int id, order_id; + + wrk = container_of(work, struct storvsc_scan_work, work); + for (id = 0; id < wrk->host->max_id; ++id) { + if (wrk->host->reverse_ordering) + order_id = wrk->host->max_id - id - 1; + else + order_id = id; + + scsi_scan_target(&wrk->host->shost_gendev, 0, + order_id, SCAN_WILD_CARD, 1); + } + kfree(wrk); +} + +static void storvsc_remove_lun(struct work_struct *work) +{ + struct storvsc_scan_work *wrk; + struct scsi_device *sdev; + + wrk = container_of(work, struct storvsc_scan_work, work); + if (!scsi_host_get(wrk->host)) + goto done; + + sdev = scsi_device_lookup(wrk->host, 0, 0, wrk->lun); + + if (sdev) { + scsi_remove_device(sdev); + scsi_device_put(sdev); + } + scsi_host_put(wrk->host); + +done: + kfree(wrk); +} + +/* + * Major/minor macros. Minor version is in LSB, meaning that earlier flat + * version numbers will be interpreted as "0.x" (i.e., 1 becomes 0.1). + */ + +static inline u16 storvsc_get_version(u8 major, u8 minor) +{ + u16 version; + + version = ((major << 8) | minor); + return version; +} + +/* + * We can get incoming messages from the host that are not in response to + * messages that we have sent out. An example of this would be messages + * received by the guest to notify dynamic addition/removal of LUNs. To + * deal with potential race conditions where the driver may be in the + * midst of being unloaded when we might receive an unsolicited message + * from the host, we have implemented a mechanism to gurantee sequential + * consistency: + * + * 1) Once the device is marked as being destroyed, we will fail all + * outgoing messages. + * 2) We permit incoming messages when the device is being destroyed, + * only to properly account for messages already sent out. + */ + +static inline struct storvsc_device *get_out_stor_device( + struct hv_device *device) +{ + struct storvsc_device *stor_device; + + stor_device = hv_get_drvdata(device); + + if (stor_device && stor_device->destroy) + stor_device = NULL; + + return stor_device; +} + + +static inline void storvsc_wait_to_drain(struct storvsc_device *dev) +{ + dev->drain_notify = true; + wait_event(dev->waiting_to_drain, + atomic_read(&dev->num_outstanding_req) == 0); + dev->drain_notify = false; +} + +static inline struct storvsc_device *get_in_stor_device( + struct hv_device *device) +{ + struct storvsc_device *stor_device; + + stor_device = hv_get_drvdata(device); + + if (!stor_device) + goto get_in_err; + + /* + * If the device is being destroyed; allow incoming + * traffic only to cleanup outstanding requests. + */ + + if (stor_device->destroy && + (atomic_read(&stor_device->num_outstanding_req) == 0)) + stor_device = NULL; + +get_in_err: + return stor_device; + +} + +static void destroy_bounce_buffer(struct scatterlist *sgl, + unsigned int sg_count) +{ + int i; + struct page *page_buf; + + for (i = 0; i < sg_count; i++) { + page_buf = sg_page((&sgl[i])); + if (page_buf != NULL) + __free_page(page_buf); + } + + kfree(sgl); +} + +static int do_bounce_buffer(struct scatterlist *sgl, unsigned int sg_count) +{ + int i; + + /* No need to check */ + if (sg_count < 2) + return -1; + + /* We have at least 2 sg entries */ + for (i = 0; i < sg_count; i++) { + if (i == 0) { + /* make sure 1st one does not have hole */ + if (sgl[i].offset + sgl[i].length != PAGE_SIZE) + return i; + } else if (i == sg_count - 1) { + /* make sure last one does not have hole */ + if (sgl[i].offset != 0) + return i; + } else { + /* make sure no hole in the middle */ + if (sgl[i].length != PAGE_SIZE || sgl[i].offset != 0) + return i; + } + } + return -1; +} + +static struct scatterlist *create_bounce_buffer(struct scatterlist *sgl, + unsigned int sg_count, + unsigned int len, + int write) +{ + int i; + int num_pages; + struct scatterlist *bounce_sgl; + struct page *page_buf; + unsigned int buf_len = ((write == WRITE_TYPE) ? 0 : PAGE_SIZE); + + num_pages = ALIGN(len, PAGE_SIZE) >> PAGE_SHIFT; + + bounce_sgl = kcalloc(num_pages, sizeof(struct scatterlist), GFP_ATOMIC); + if (!bounce_sgl) + return NULL; + + for (i = 0; i < num_pages; i++) { + page_buf = alloc_page(GFP_ATOMIC); + if (!page_buf) + goto cleanup; + sg_set_page(&bounce_sgl[i], page_buf, buf_len, 0); + } + + return bounce_sgl; + +cleanup: + destroy_bounce_buffer(bounce_sgl, num_pages); + return NULL; +} + +/* Assume the original sgl has enough room */ +static unsigned int copy_from_bounce_buffer(struct scatterlist *orig_sgl, + struct scatterlist *bounce_sgl, + unsigned int orig_sgl_count, + unsigned int bounce_sgl_count) +{ + int i; + int j = 0; + unsigned long src, dest; + unsigned int srclen, destlen, copylen; + unsigned int total_copied = 0; + unsigned long bounce_addr = 0; + unsigned long dest_addr = 0; + unsigned long flags; + + local_irq_save(flags); + + for (i = 0; i < orig_sgl_count; i++) { + dest_addr = (unsigned long)kmap_atomic(sg_page((&orig_sgl[i])), + KM_IRQ0) + orig_sgl[i].offset; + dest = dest_addr; + destlen = orig_sgl[i].length; + + if (bounce_addr == 0) + bounce_addr = + (unsigned long)kmap_atomic(sg_page((&bounce_sgl[j])), + KM_IRQ0); + + while (destlen) { + src = bounce_addr + bounce_sgl[j].offset; + srclen = bounce_sgl[j].length - bounce_sgl[j].offset; + + copylen = min(srclen, destlen); + memcpy((void *)dest, (void *)src, copylen); + + total_copied += copylen; + bounce_sgl[j].offset += copylen; + destlen -= copylen; + dest += copylen; + + if (bounce_sgl[j].offset == bounce_sgl[j].length) { + /* full */ + kunmap_atomic((void *)bounce_addr, KM_IRQ0); + j++; + + /* + * It is possible that the number of elements + * in the bounce buffer may not be equal to + * the number of elements in the original + * scatter list. Handle this correctly. + */ + + if (j == bounce_sgl_count) { + /* + * We are done; cleanup and return. + */ + kunmap_atomic((void *)(dest_addr - + orig_sgl[i].offset), + KM_IRQ0); + local_irq_restore(flags); + return total_copied; + } + + /* if we need to use another bounce buffer */ + if (destlen || i != orig_sgl_count - 1) + bounce_addr = + (unsigned long)kmap_atomic( + sg_page((&bounce_sgl[j])), KM_IRQ0); + } else if (destlen == 0 && i == orig_sgl_count - 1) { + /* unmap the last bounce that is < PAGE_SIZE */ + kunmap_atomic((void *)bounce_addr, KM_IRQ0); + } + } + + kunmap_atomic((void *)(dest_addr - orig_sgl[i].offset), + KM_IRQ0); + } + + local_irq_restore(flags); + + return total_copied; +} + +/* Assume the bounce_sgl has enough room ie using the create_bounce_buffer() */ +static unsigned int copy_to_bounce_buffer(struct scatterlist *orig_sgl, + struct scatterlist *bounce_sgl, + unsigned int orig_sgl_count) +{ + int i; + int j = 0; + unsigned long src, dest; + unsigned int srclen, destlen, copylen; + unsigned int total_copied = 0; + unsigned long bounce_addr = 0; + unsigned long src_addr = 0; + unsigned long flags; + + local_irq_save(flags); + + for (i = 0; i < orig_sgl_count; i++) { + src_addr = (unsigned long)kmap_atomic(sg_page((&orig_sgl[i])), + KM_IRQ0) + orig_sgl[i].offset; + src = src_addr; + srclen = orig_sgl[i].length; + + if (bounce_addr == 0) + bounce_addr = + (unsigned long)kmap_atomic(sg_page((&bounce_sgl[j])), + KM_IRQ0); + + while (srclen) { + /* assume bounce offset always == 0 */ + dest = bounce_addr + bounce_sgl[j].length; + destlen = PAGE_SIZE - bounce_sgl[j].length; + + copylen = min(srclen, destlen); + memcpy((void *)dest, (void *)src, copylen); + + total_copied += copylen; + bounce_sgl[j].length += copylen; + srclen -= copylen; + src += copylen; + + if (bounce_sgl[j].length == PAGE_SIZE) { + /* full..move to next entry */ + kunmap_atomic((void *)bounce_addr, KM_IRQ0); + j++; + + /* if we need to use another bounce buffer */ + if (srclen || i != orig_sgl_count - 1) + bounce_addr = + (unsigned long)kmap_atomic( + sg_page((&bounce_sgl[j])), KM_IRQ0); + + } else if (srclen == 0 && i == orig_sgl_count - 1) { + /* unmap the last bounce that is < PAGE_SIZE */ + kunmap_atomic((void *)bounce_addr, KM_IRQ0); + } + } + + kunmap_atomic((void *)(src_addr - orig_sgl[i].offset), KM_IRQ0); + } + + local_irq_restore(flags); + + return total_copied; +} + +static int storvsc_channel_init(struct hv_device *device) +{ + struct storvsc_device *stor_device; + struct storvsc_cmd_request *request; + struct vstor_packet *vstor_packet; + int ret, t; + + stor_device = get_out_stor_device(device); + if (!stor_device) + return -ENODEV; + + request = &stor_device->init_request; + vstor_packet = &request->vstor_packet; + + /* + * Now, initiate the vsc/vsp initialization protocol on the open + * channel + */ + memset(request, 0, sizeof(struct storvsc_cmd_request)); + init_completion(&request->wait_event); + vstor_packet->operation = VSTOR_OPERATION_BEGIN_INITIALIZATION; + vstor_packet->flags = REQUEST_COMPLETION_FLAG; + + ret = vmbus_sendpacket(device->channel, vstor_packet, + sizeof(struct vstor_packet), + (unsigned long)request, + VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + if (ret != 0) + goto cleanup; + + t = wait_for_completion_timeout(&request->wait_event, 5*HZ); + if (t == 0) { + ret = -ETIMEDOUT; + goto cleanup; + } + + if (vstor_packet->operation != VSTOR_OPERATION_COMPLETE_IO || + vstor_packet->status != 0) + goto cleanup; + + + /* reuse the packet for version range supported */ + memset(vstor_packet, 0, sizeof(struct vstor_packet)); + vstor_packet->operation = VSTOR_OPERATION_QUERY_PROTOCOL_VERSION; + vstor_packet->flags = REQUEST_COMPLETION_FLAG; + + vstor_packet->version.major_minor = + storvsc_get_version(VMSTOR_CURRENT_MAJOR, VMSTOR_CURRENT_MINOR); + + /* + * The revision number is only used in Windows; set it to 0. + */ + vstor_packet->version.revision = 0; + + ret = vmbus_sendpacket(device->channel, vstor_packet, + sizeof(struct vstor_packet), + (unsigned long)request, + VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + if (ret != 0) + goto cleanup; + + t = wait_for_completion_timeout(&request->wait_event, 5*HZ); + if (t == 0) { + ret = -ETIMEDOUT; + goto cleanup; + } + + if (vstor_packet->operation != VSTOR_OPERATION_COMPLETE_IO || + vstor_packet->status != 0) + goto cleanup; + + + memset(vstor_packet, 0, sizeof(struct vstor_packet)); + vstor_packet->operation = VSTOR_OPERATION_QUERY_PROPERTIES; + vstor_packet->flags = REQUEST_COMPLETION_FLAG; + vstor_packet->storage_channel_properties.port_number = + stor_device->port_number; + + ret = vmbus_sendpacket(device->channel, vstor_packet, + sizeof(struct vstor_packet), + (unsigned long)request, + VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + + if (ret != 0) + goto cleanup; + + t = wait_for_completion_timeout(&request->wait_event, 5*HZ); + if (t == 0) { + ret = -ETIMEDOUT; + goto cleanup; + } + + if (vstor_packet->operation != VSTOR_OPERATION_COMPLETE_IO || + vstor_packet->status != 0) + goto cleanup; + + stor_device->path_id = vstor_packet->storage_channel_properties.path_id; + stor_device->target_id + = vstor_packet->storage_channel_properties.target_id; + + memset(vstor_packet, 0, sizeof(struct vstor_packet)); + vstor_packet->operation = VSTOR_OPERATION_END_INITIALIZATION; + vstor_packet->flags = REQUEST_COMPLETION_FLAG; + + ret = vmbus_sendpacket(device->channel, vstor_packet, + sizeof(struct vstor_packet), + (unsigned long)request, + VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + + if (ret != 0) + goto cleanup; + + t = wait_for_completion_timeout(&request->wait_event, 5*HZ); + if (t == 0) { + ret = -ETIMEDOUT; + goto cleanup; + } + + if (vstor_packet->operation != VSTOR_OPERATION_COMPLETE_IO || + vstor_packet->status != 0) + goto cleanup; + + +cleanup: + return ret; +} + + +static void storvsc_command_completion(struct storvsc_cmd_request *cmd_request) +{ + struct scsi_cmnd *scmnd = cmd_request->cmd; + struct hv_host_device *host_dev = shost_priv(scmnd->device->host); + void (*scsi_done_fn)(struct scsi_cmnd *); + struct scsi_sense_hdr sense_hdr; + struct vmscsi_request *vm_srb; + struct storvsc_scan_work *wrk; + struct stor_mem_pools *memp = scmnd->device->hostdata; + + vm_srb = &cmd_request->vstor_packet.vm_srb; + if (cmd_request->bounce_sgl_count) { + if (vm_srb->data_in == READ_TYPE) + copy_from_bounce_buffer(scsi_sglist(scmnd), + cmd_request->bounce_sgl, + scsi_sg_count(scmnd), + cmd_request->bounce_sgl_count); + destroy_bounce_buffer(cmd_request->bounce_sgl, + cmd_request->bounce_sgl_count); + } + + /* + * If there is an error; offline the device since all + * error recovery strategies would have already been + * deployed on the host side. + */ + if (vm_srb->srb_status == SRB_STATUS_ERROR) + scmnd->result = DID_TARGET_FAILURE << 16; + else + scmnd->result = vm_srb->scsi_status; + + /* + * If the LUN is invalid; remove the device. + */ + if (vm_srb->srb_status == SRB_STATUS_INVALID_LUN) { + struct storvsc_device *stor_dev; + struct hv_device *dev = host_dev->dev; + struct Scsi_Host *host; + + stor_dev = get_in_stor_device(dev); + host = stor_dev->host; + + wrk = kmalloc(sizeof(struct storvsc_scan_work), + GFP_ATOMIC); + if (!wrk) { + scmnd->result = DID_TARGET_FAILURE << 16; + } else { + wrk->host = host; + wrk->lun = vm_srb->lun; + INIT_WORK(&wrk->work, storvsc_remove_lun); + schedule_work(&wrk->work); + } + } + + if (scmnd->result) { + if (scsi_normalize_sense(scmnd->sense_buffer, + SCSI_SENSE_BUFFERSIZE, &sense_hdr)) + scsi_print_sense_hdr("storvsc", &sense_hdr); + } + + scsi_set_resid(scmnd, + cmd_request->data_buffer.len - + vm_srb->data_transfer_length); + + scsi_done_fn = scmnd->scsi_done; + + scmnd->host_scribble = NULL; + scmnd->scsi_done = NULL; + + scsi_done_fn(scmnd); + + mempool_free(cmd_request, memp->request_mempool); +} + +static void storvsc_on_io_completion(struct hv_device *device, + struct vstor_packet *vstor_packet, + struct storvsc_cmd_request *request) +{ + struct storvsc_device *stor_device; + struct vstor_packet *stor_pkt; + + stor_device = hv_get_drvdata(device); + stor_pkt = &request->vstor_packet; + + /* + * The current SCSI handling on the host side does + * not correctly handle: + * INQUIRY command with page code parameter set to 0x80 + * MODE_SENSE command with cmd[2] == 0x1c + * + * Setup srb and scsi status so this won't be fatal. + * We do this so we can distinguish truly fatal failues + * (srb status == 0x4) and off-line the device in that case. + */ + + if ((stor_pkt->vm_srb.cdb[0] == INQUIRY) || + (stor_pkt->vm_srb.cdb[0] == MODE_SENSE)) { + vstor_packet->vm_srb.scsi_status = 0; + vstor_packet->vm_srb.srb_status = SRB_STATUS_SUCCESS; + } + + + /* Copy over the status...etc */ + stor_pkt->vm_srb.scsi_status = vstor_packet->vm_srb.scsi_status; + stor_pkt->vm_srb.srb_status = vstor_packet->vm_srb.srb_status; + stor_pkt->vm_srb.sense_info_length = + vstor_packet->vm_srb.sense_info_length; + + if (vstor_packet->vm_srb.scsi_status != 0 || + vstor_packet->vm_srb.srb_status != SRB_STATUS_SUCCESS){ + dev_warn(&device->device, + "cmd 0x%x scsi status 0x%x srb status 0x%x\n", + stor_pkt->vm_srb.cdb[0], + vstor_packet->vm_srb.scsi_status, + vstor_packet->vm_srb.srb_status); + } + + if ((vstor_packet->vm_srb.scsi_status & 0xFF) == 0x02) { + /* CHECK_CONDITION */ + if (vstor_packet->vm_srb.srb_status & + SRB_STATUS_AUTOSENSE_VALID) { + /* autosense data available */ + dev_warn(&device->device, + "stor pkt %p autosense data valid - len %d\n", + request, + vstor_packet->vm_srb.sense_info_length); + + memcpy(request->sense_buffer, + vstor_packet->vm_srb.sense_data, + vstor_packet->vm_srb.sense_info_length); + + } + } + + stor_pkt->vm_srb.data_transfer_length = + vstor_packet->vm_srb.data_transfer_length; + + storvsc_command_completion(request); + + if (atomic_dec_and_test(&stor_device->num_outstanding_req) && + stor_device->drain_notify) + wake_up(&stor_device->waiting_to_drain); + + +} + +static void storvsc_on_receive(struct hv_device *device, + struct vstor_packet *vstor_packet, + struct storvsc_cmd_request *request) +{ + struct storvsc_scan_work *work; + struct storvsc_device *stor_device; + + switch (vstor_packet->operation) { + case VSTOR_OPERATION_COMPLETE_IO: + storvsc_on_io_completion(device, vstor_packet, request); + break; + + case VSTOR_OPERATION_REMOVE_DEVICE: + case VSTOR_OPERATION_ENUMERATE_BUS: + stor_device = get_in_stor_device(device); + work = kmalloc(sizeof(struct storvsc_scan_work), GFP_ATOMIC); + if (!work) + return; + + INIT_WORK(&work->work, storvsc_bus_scan); + work->host = stor_device->host; + schedule_work(&work->work); + break; + + default: + break; + } +} + +static void storvsc_on_channel_callback(void *context) +{ + struct hv_device *device = (struct hv_device *)context; + struct storvsc_device *stor_device; + u32 bytes_recvd; + u64 request_id; + unsigned char packet[ALIGN(sizeof(struct vstor_packet), 8)]; + struct storvsc_cmd_request *request; + int ret; + + + stor_device = get_in_stor_device(device); + if (!stor_device) + return; + + do { + ret = vmbus_recvpacket(device->channel, packet, + ALIGN(sizeof(struct vstor_packet), 8), + &bytes_recvd, &request_id); + if (ret == 0 && bytes_recvd > 0) { + + request = (struct storvsc_cmd_request *) + (unsigned long)request_id; + + if ((request == &stor_device->init_request) || + (request == &stor_device->reset_request)) { + + memcpy(&request->vstor_packet, packet, + sizeof(struct vstor_packet)); + complete(&request->wait_event); + } else { + storvsc_on_receive(device, + (struct vstor_packet *)packet, + request); + } + } else { + break; + } + } while (1); + + return; +} + +static int storvsc_connect_to_vsp(struct hv_device *device, u32 ring_size) +{ + struct vmstorage_channel_properties props; + int ret; + + memset(&props, 0, sizeof(struct vmstorage_channel_properties)); + + ret = vmbus_open(device->channel, + ring_size, + ring_size, + (void *)&props, + sizeof(struct vmstorage_channel_properties), + storvsc_on_channel_callback, device); + + if (ret != 0) + return ret; + + ret = storvsc_channel_init(device); + + return ret; +} + +static int storvsc_dev_remove(struct hv_device *device) +{ + struct storvsc_device *stor_device; + unsigned long flags; + + stor_device = hv_get_drvdata(device); + + spin_lock_irqsave(&device->channel->inbound_lock, flags); + stor_device->destroy = true; + spin_unlock_irqrestore(&device->channel->inbound_lock, flags); + + /* + * At this point, all outbound traffic should be disable. We + * only allow inbound traffic (responses) to proceed so that + * outstanding requests can be completed. + */ + + storvsc_wait_to_drain(stor_device); + + /* + * Since we have already drained, we don't need to busy wait + * as was done in final_release_stor_device() + * Note that we cannot set the ext pointer to NULL until + * we have drained - to drain the outgoing packets, we need to + * allow incoming packets. + */ + spin_lock_irqsave(&device->channel->inbound_lock, flags); + hv_set_drvdata(device, NULL); + spin_unlock_irqrestore(&device->channel->inbound_lock, flags); + + /* Close the channel */ + vmbus_close(device->channel); + + kfree(stor_device); + return 0; +} + +static int storvsc_do_io(struct hv_device *device, + struct storvsc_cmd_request *request) +{ + struct storvsc_device *stor_device; + struct vstor_packet *vstor_packet; + int ret = 0; + + vstor_packet = &request->vstor_packet; + stor_device = get_out_stor_device(device); + + if (!stor_device) + return -ENODEV; + + + request->device = device; + + + vstor_packet->flags |= REQUEST_COMPLETION_FLAG; + + vstor_packet->vm_srb.length = sizeof(struct vmscsi_request); + + + vstor_packet->vm_srb.sense_info_length = STORVSC_SENSE_BUFFER_SIZE; + + + vstor_packet->vm_srb.data_transfer_length = + request->data_buffer.len; + + vstor_packet->operation = VSTOR_OPERATION_EXECUTE_SRB; + + if (request->data_buffer.len) { + ret = vmbus_sendpacket_multipagebuffer(device->channel, + &request->data_buffer, + vstor_packet, + sizeof(struct vstor_packet), + (unsigned long)request); + } else { + ret = vmbus_sendpacket(device->channel, vstor_packet, + sizeof(struct vstor_packet), + (unsigned long)request, + VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + } + + if (ret != 0) + return ret; + + atomic_inc(&stor_device->num_outstanding_req); + + return ret; +} + +static int storvsc_device_alloc(struct scsi_device *sdevice) +{ + struct stor_mem_pools *memp; + int number = STORVSC_MIN_BUF_NR; + + memp = kzalloc(sizeof(struct stor_mem_pools), GFP_KERNEL); + if (!memp) + return -ENOMEM; + + memp->request_pool = + kmem_cache_create(dev_name(&sdevice->sdev_dev), + sizeof(struct storvsc_cmd_request), 0, + SLAB_HWCACHE_ALIGN, NULL); + + if (!memp->request_pool) + goto err0; + + memp->request_mempool = mempool_create(number, mempool_alloc_slab, + mempool_free_slab, + memp->request_pool); + + if (!memp->request_mempool) + goto err1; + + sdevice->hostdata = memp; + + return 0; + +err1: + kmem_cach |