diff options
Diffstat (limited to 'drivers/hv')
| -rw-r--r-- | drivers/hv/Makefile | 2 | ||||
| -rw-r--r-- | drivers/hv/channel.c | 61 | ||||
| -rw-r--r-- | drivers/hv/channel_mgmt.c | 52 | ||||
| -rw-r--r-- | drivers/hv/connection.c | 63 | ||||
| -rw-r--r-- | drivers/hv/hv.c | 2 | ||||
| -rw-r--r-- | drivers/hv/hv_balloon.c | 32 | ||||
| -rw-r--r-- | drivers/hv/hv_fcopy.c | 414 | ||||
| -rw-r--r-- | drivers/hv/hv_kvp.c | 21 | ||||
| -rw-r--r-- | drivers/hv/hv_snapshot.c | 2 | ||||
| -rw-r--r-- | drivers/hv/hv_util.c | 13 | ||||
| -rw-r--r-- | drivers/hv/hyperv_vmbus.h | 19 | ||||
| -rw-r--r-- | drivers/hv/ring_buffer.c | 17 | ||||
| -rw-r--r-- | drivers/hv/vmbus_drv.c | 118 |
13 files changed, 668 insertions, 148 deletions
diff --git a/drivers/hv/Makefile b/drivers/hv/Makefile index 0a74b566118..5e4dfa4cfe2 100644 --- a/drivers/hv/Makefile +++ b/drivers/hv/Makefile @@ -5,4 +5,4 @@ obj-$(CONFIG_HYPERV_BALLOON) += hv_balloon.o hv_vmbus-y := vmbus_drv.o \ hv.o connection.o channel.o \ channel_mgmt.o ring_buffer.o -hv_utils-y := hv_util.o hv_kvp.o hv_snapshot.o +hv_utils-y := hv_util.o hv_kvp.o hv_snapshot.o hv_fcopy.o diff --git a/drivers/hv/channel.c b/drivers/hv/channel.c index 69ea36f07b4..284cf66489f 100644 --- a/drivers/hv/channel.c +++ b/drivers/hv/channel.c @@ -27,6 +27,7 @@ #include <linux/slab.h> #include <linux/module.h> #include <linux/hyperv.h> +#include <linux/uio.h> #include "hyperv_vmbus.h" @@ -470,18 +471,26 @@ int vmbus_teardown_gpadl(struct vmbus_channel *channel, u32 gpadl_handle) } EXPORT_SYMBOL_GPL(vmbus_teardown_gpadl); +static void reset_channel_cb(void *arg) +{ + struct vmbus_channel *channel = arg; + + channel->onchannel_callback = NULL; +} + static void vmbus_close_internal(struct vmbus_channel *channel) { struct vmbus_channel_close_channel *msg; int ret; - unsigned long flags; channel->state = CHANNEL_OPEN_STATE; channel->sc_creation_callback = NULL; /* Stop callback and cancel the timer asap */ - spin_lock_irqsave(&channel->inbound_lock, flags); - channel->onchannel_callback = NULL; - spin_unlock_irqrestore(&channel->inbound_lock, flags); + if (channel->target_cpu != smp_processor_id()) + smp_call_function_single(channel->target_cpu, reset_channel_cb, + channel, true); + else + reset_channel_cb(channel); /* Send a closing message */ @@ -554,14 +563,14 @@ EXPORT_SYMBOL_GPL(vmbus_close); * * Mainly used by Hyper-V drivers. */ -int vmbus_sendpacket(struct vmbus_channel *channel, const void *buffer, +int vmbus_sendpacket(struct vmbus_channel *channel, void *buffer, u32 bufferlen, u64 requestid, enum vmbus_packet_type type, u32 flags) { struct vmpacket_descriptor desc; u32 packetlen = sizeof(struct vmpacket_descriptor) + bufferlen; u32 packetlen_aligned = ALIGN(packetlen, sizeof(u64)); - struct scatterlist bufferlist[3]; + struct kvec bufferlist[3]; u64 aligned_data = 0; int ret; bool signal = false; @@ -575,11 +584,12 @@ int vmbus_sendpacket(struct vmbus_channel *channel, const void *buffer, desc.len8 = (u16)(packetlen_aligned >> 3); desc.trans_id = requestid; - sg_init_table(bufferlist, 3); - sg_set_buf(&bufferlist[0], &desc, sizeof(struct vmpacket_descriptor)); - sg_set_buf(&bufferlist[1], buffer, bufferlen); - sg_set_buf(&bufferlist[2], &aligned_data, - packetlen_aligned - packetlen); + bufferlist[0].iov_base = &desc; + bufferlist[0].iov_len = sizeof(struct vmpacket_descriptor); + bufferlist[1].iov_base = buffer; + bufferlist[1].iov_len = bufferlen; + bufferlist[2].iov_base = &aligned_data; + bufferlist[2].iov_len = (packetlen_aligned - packetlen); ret = hv_ringbuffer_write(&channel->outbound, bufferlist, 3, &signal); @@ -605,7 +615,7 @@ int vmbus_sendpacket_pagebuffer(struct vmbus_channel *channel, u32 descsize; u32 packetlen; u32 packetlen_aligned; - struct scatterlist bufferlist[3]; + struct kvec bufferlist[3]; u64 aligned_data = 0; bool signal = false; @@ -637,11 +647,12 @@ int vmbus_sendpacket_pagebuffer(struct vmbus_channel *channel, desc.range[i].pfn = pagebuffers[i].pfn; } - sg_init_table(bufferlist, 3); - sg_set_buf(&bufferlist[0], &desc, descsize); - sg_set_buf(&bufferlist[1], buffer, bufferlen); - sg_set_buf(&bufferlist[2], &aligned_data, - packetlen_aligned - packetlen); + bufferlist[0].iov_base = &desc; + bufferlist[0].iov_len = descsize; + bufferlist[1].iov_base = buffer; + bufferlist[1].iov_len = bufferlen; + bufferlist[2].iov_base = &aligned_data; + bufferlist[2].iov_len = (packetlen_aligned - packetlen); ret = hv_ringbuffer_write(&channel->outbound, bufferlist, 3, &signal); @@ -665,14 +676,13 @@ int vmbus_sendpacket_multipagebuffer(struct vmbus_channel *channel, u32 descsize; u32 packetlen; u32 packetlen_aligned; - struct scatterlist bufferlist[3]; + struct kvec bufferlist[3]; u64 aligned_data = 0; bool signal = false; u32 pfncount = NUM_PAGES_SPANNED(multi_pagebuffer->offset, multi_pagebuffer->len); - - if ((pfncount < 0) || (pfncount > MAX_MULTIPAGE_BUFFER_COUNT)) + if (pfncount > MAX_MULTIPAGE_BUFFER_COUNT) return -EINVAL; /* @@ -700,11 +710,12 @@ int vmbus_sendpacket_multipagebuffer(struct vmbus_channel *channel, memcpy(desc.range.pfn_array, multi_pagebuffer->pfn_array, pfncount * sizeof(u64)); - sg_init_table(bufferlist, 3); - sg_set_buf(&bufferlist[0], &desc, descsize); - sg_set_buf(&bufferlist[1], buffer, bufferlen); - sg_set_buf(&bufferlist[2], &aligned_data, - packetlen_aligned - packetlen); + bufferlist[0].iov_base = &desc; + bufferlist[0].iov_len = descsize; + bufferlist[1].iov_base = buffer; + bufferlist[1].iov_len = bufferlen; + bufferlist[2].iov_base = &aligned_data; + bufferlist[2].iov_len = (packetlen_aligned - packetlen); ret = hv_ringbuffer_write(&channel->outbound, bufferlist, 3, &signal); diff --git a/drivers/hv/channel_mgmt.c b/drivers/hv/channel_mgmt.c index fa920469bf1..ed9350d4276 100644 --- a/drivers/hv/channel_mgmt.c +++ b/drivers/hv/channel_mgmt.c @@ -149,6 +149,7 @@ static struct vmbus_channel *alloc_channel(void) spin_lock_init(&channel->sc_lock); INIT_LIST_HEAD(&channel->sc_list); + INIT_LIST_HEAD(&channel->percpu_list); channel->controlwq = create_workqueue("hv_vmbus_ctl"); if (!channel->controlwq) { @@ -188,7 +189,20 @@ static void free_channel(struct vmbus_channel *channel) queue_work(vmbus_connection.work_queue, &channel->work); } +static void percpu_channel_enq(void *arg) +{ + struct vmbus_channel *channel = arg; + int cpu = smp_processor_id(); + + list_add_tail(&channel->percpu_list, &hv_context.percpu_list[cpu]); +} +static void percpu_channel_deq(void *arg) +{ + struct vmbus_channel *channel = arg; + + list_del(&channel->percpu_list); +} /* * vmbus_process_rescind_offer - @@ -210,6 +224,12 @@ static void vmbus_process_rescind_offer(struct work_struct *work) msg.header.msgtype = CHANNELMSG_RELID_RELEASED; vmbus_post_msg(&msg, sizeof(struct vmbus_channel_relid_released)); + if (channel->target_cpu != smp_processor_id()) + smp_call_function_single(channel->target_cpu, + percpu_channel_deq, channel, true); + else + percpu_channel_deq(channel); + if (channel->primary_channel == NULL) { spin_lock_irqsave(&vmbus_connection.channel_lock, flags); list_del(&channel->listentry); @@ -245,6 +265,7 @@ static void vmbus_process_offer(struct work_struct *work) work); struct vmbus_channel *channel; bool fnew = true; + bool enq = false; int ret; unsigned long flags; @@ -264,12 +285,22 @@ static void vmbus_process_offer(struct work_struct *work) } } - if (fnew) + if (fnew) { list_add_tail(&newchannel->listentry, &vmbus_connection.chn_list); + enq = true; + } spin_unlock_irqrestore(&vmbus_connection.channel_lock, flags); + if (enq) { + if (newchannel->target_cpu != smp_processor_id()) + smp_call_function_single(newchannel->target_cpu, + percpu_channel_enq, + newchannel, true); + else + percpu_channel_enq(newchannel); + } if (!fnew) { /* * Check to see if this is a sub-channel. @@ -282,6 +313,14 @@ static void vmbus_process_offer(struct work_struct *work) spin_lock_irqsave(&channel->sc_lock, flags); list_add_tail(&newchannel->sc_list, &channel->sc_list); spin_unlock_irqrestore(&channel->sc_lock, flags); + + if (newchannel->target_cpu != smp_processor_id()) + smp_call_function_single(newchannel->target_cpu, + percpu_channel_enq, + newchannel, true); + else + percpu_channel_enq(newchannel); + newchannel->state = CHANNEL_OPEN_STATE; if (channel->sc_creation_callback != NULL) channel->sc_creation_callback(newchannel); @@ -365,7 +404,7 @@ static u32 next_vp; * performance critical channels (IDE, SCSI and Network) will be uniformly * distributed across all available CPUs. */ -static u32 get_vp_index(uuid_le *type_guid) +static void init_vp_index(struct vmbus_channel *channel, const uuid_le *type_guid) { u32 cur_cpu; int i; @@ -387,10 +426,13 @@ static u32 get_vp_index(uuid_le *type_guid) * Also if the channel is not a performance critical * channel, bind it to cpu 0. */ - return 0; + channel->target_cpu = 0; + channel->target_vp = 0; + return; } cur_cpu = (++next_vp % max_cpus); - return hv_context.vp_index[cur_cpu]; + channel->target_cpu = cur_cpu; + channel->target_vp = hv_context.vp_index[cur_cpu]; } /* @@ -438,7 +480,7 @@ static void vmbus_onoffer(struct vmbus_channel_message_header *hdr) offer->connection_id; } - newchannel->target_vp = get_vp_index(&offer->offer.if_type); + init_vp_index(newchannel, &offer->offer.if_type); memcpy(&newchannel->offermsg, offer, sizeof(struct vmbus_channel_offer_channel)); diff --git a/drivers/hv/connection.c b/drivers/hv/connection.c index af6edf9b193..ae22e3c1fc4 100644 --- a/drivers/hv/connection.c +++ b/drivers/hv/connection.c @@ -55,6 +55,9 @@ static __u32 vmbus_get_next_version(__u32 current_version) case (VERSION_WIN8): return VERSION_WIN7; + case (VERSION_WIN8_1): + return VERSION_WIN8; + case (VERSION_WS2008): default: return VERSION_INVAL; @@ -67,7 +70,6 @@ static int vmbus_negotiate_version(struct vmbus_channel_msginfo *msginfo, int ret = 0; struct vmbus_channel_initiate_contact *msg; unsigned long flags; - int t; init_completion(&msginfo->waitevent); @@ -78,6 +80,8 @@ static int vmbus_negotiate_version(struct vmbus_channel_msginfo *msginfo, msg->interrupt_page = virt_to_phys(vmbus_connection.int_page); msg->monitor_page1 = virt_to_phys(vmbus_connection.monitor_pages[0]); msg->monitor_page2 = virt_to_phys(vmbus_connection.monitor_pages[1]); + if (version == VERSION_WIN8_1) + msg->target_vcpu = hv_context.vp_index[smp_processor_id()]; /* * Add to list before we send the request since we may @@ -100,15 +104,7 @@ static int vmbus_negotiate_version(struct vmbus_channel_msginfo *msginfo, } /* Wait for the connection response */ - t = wait_for_completion_timeout(&msginfo->waitevent, 5*HZ); - if (t == 0) { - spin_lock_irqsave(&vmbus_connection.channelmsg_lock, - flags); - list_del(&msginfo->msglistentry); - spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, - flags); - return -ETIMEDOUT; - } + wait_for_completion(&msginfo->waitevent); spin_lock_irqsave(&vmbus_connection.channelmsg_lock, flags); list_del(&msginfo->msglistentry); @@ -228,8 +224,8 @@ cleanup: vmbus_connection.int_page = NULL; } - free_pages((unsigned long)vmbus_connection.monitor_pages[0], 1); - free_pages((unsigned long)vmbus_connection.monitor_pages[1], 1); + free_pages((unsigned long)vmbus_connection.monitor_pages[0], 0); + free_pages((unsigned long)vmbus_connection.monitor_pages[1], 0); vmbus_connection.monitor_pages[0] = NULL; vmbus_connection.monitor_pages[1] = NULL; @@ -238,6 +234,28 @@ cleanup: return ret; } +/* + * Map the given relid to the corresponding channel based on the + * per-cpu list of channels that have been affinitized to this CPU. + * This will be used in the channel callback path as we can do this + * mapping in a lock-free fashion. + */ +static struct vmbus_channel *pcpu_relid2channel(u32 relid) +{ + struct vmbus_channel *channel; + struct vmbus_channel *found_channel = NULL; + int cpu = smp_processor_id(); + struct list_head *pcpu_head = &hv_context.percpu_list[cpu]; + + list_for_each_entry(channel, pcpu_head, percpu_list) { + if (channel->offermsg.child_relid == relid) { + found_channel = channel; + break; + } + } + + return found_channel; +} /* * relid2channel - Get the channel object given its @@ -281,7 +299,6 @@ struct vmbus_channel *relid2channel(u32 relid) static void process_chn_event(u32 relid) { struct vmbus_channel *channel; - unsigned long flags; void *arg; bool read_state; u32 bytes_to_read; @@ -290,7 +307,7 @@ static void process_chn_event(u32 relid) * Find the channel based on this relid and invokes the * channel callback to process the event */ - channel = relid2channel(relid); + channel = pcpu_relid2channel(relid); if (!channel) { pr_err("channel not found for relid - %u\n", relid); @@ -300,13 +317,12 @@ static void process_chn_event(u32 relid) /* * A channel once created is persistent even when there * is no driver handling the device. An unloading driver - * sets the onchannel_callback to NULL under the - * protection of the channel inbound_lock. Thus, checking - * and invoking the driver specific callback takes care of - * orderly unloading of the driver. + * sets the onchannel_callback to NULL on the same CPU + * as where this interrupt is handled (in an interrupt context). + * Thus, checking and invoking the driver specific callback takes + * care of orderly unloading of the driver. */ - spin_lock_irqsave(&channel->inbound_lock, flags); if (channel->onchannel_callback != NULL) { arg = channel->channel_callback_context; read_state = channel->batched_reading; @@ -323,15 +339,18 @@ static void process_chn_event(u32 relid) */ do { - hv_begin_read(&channel->inbound); + if (read_state) + hv_begin_read(&channel->inbound); channel->onchannel_callback(arg); - bytes_to_read = hv_end_read(&channel->inbound); + if (read_state) + bytes_to_read = hv_end_read(&channel->inbound); + else + bytes_to_read = 0; } while (read_state && (bytes_to_read != 0)); } else { pr_err("no channel callback for relid - %u\n", relid); } - spin_unlock_irqrestore(&channel->inbound_lock, flags); } /* diff --git a/drivers/hv/hv.c b/drivers/hv/hv.c index bcb49502c3b..edfc8488cb0 100644 --- a/drivers/hv/hv.c +++ b/drivers/hv/hv.c @@ -383,6 +383,8 @@ void hv_synic_init(void *arg) */ rdmsrl(HV_X64_MSR_VP_INDEX, vp_index); hv_context.vp_index[cpu] = (u32)vp_index; + + INIT_LIST_HEAD(&hv_context.percpu_list[cpu]); return; } diff --git a/drivers/hv/hv_balloon.c b/drivers/hv/hv_balloon.c index 7e17a5495e0..5e90c5d771a 100644 --- a/drivers/hv/hv_balloon.c +++ b/drivers/hv/hv_balloon.c @@ -19,6 +19,7 @@ #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt #include <linux/kernel.h> +#include <linux/jiffies.h> #include <linux/mman.h> #include <linux/delay.h> #include <linux/init.h> @@ -459,6 +460,11 @@ static bool do_hot_add; */ static uint pressure_report_delay = 45; +/* + * The last time we posted a pressure report to host. + */ +static unsigned long last_post_time; + module_param(hot_add, bool, (S_IRUGO | S_IWUSR)); MODULE_PARM_DESC(hot_add, "If set attempt memory hot_add"); @@ -542,6 +548,7 @@ struct hv_dynmem_device { static struct hv_dynmem_device dm_device; +static void post_status(struct hv_dynmem_device *dm); #ifdef CONFIG_MEMORY_HOTPLUG static void hv_bring_pgs_online(unsigned long start_pfn, unsigned long size) @@ -612,7 +619,7 @@ static void hv_mem_hot_add(unsigned long start, unsigned long size, * have not been "onlined" within the allowed time. */ wait_for_completion_timeout(&dm_device.ol_waitevent, 5*HZ); - + post_status(&dm_device); } return; @@ -951,11 +958,17 @@ static void post_status(struct hv_dynmem_device *dm) { struct dm_status status; struct sysinfo val; + unsigned long now = jiffies; + unsigned long last_post = last_post_time; if (pressure_report_delay > 0) { --pressure_report_delay; return; } + + if (!time_after(now, (last_post_time + HZ))) + return; + si_meminfo(&val); memset(&status, 0, sizeof(struct dm_status)); status.hdr.type = DM_STATUS_REPORT; @@ -983,6 +996,14 @@ static void post_status(struct hv_dynmem_device *dm) if (status.hdr.trans_id != atomic_read(&trans_id)) return; + /* + * If the last post time that we sampled has changed, + * we have raced, don't post the status. + */ + if (last_post != last_post_time) + return; + + last_post_time = jiffies; vmbus_sendpacket(dm->dev->channel, &status, sizeof(struct dm_status), (unsigned long)NULL, @@ -1117,7 +1138,7 @@ static void balloon_up(struct work_struct *dummy) if (ret == -EAGAIN) msleep(20); - + post_status(&dm_device); } while (ret == -EAGAIN); if (ret) { @@ -1144,8 +1165,10 @@ static void balloon_down(struct hv_dynmem_device *dm, struct dm_unballoon_response resp; int i; - for (i = 0; i < range_count; i++) + for (i = 0; i < range_count; i++) { free_balloon_pages(dm, &range_array[i]); + post_status(&dm_device); + } if (req->more_pages == 1) return; @@ -1171,7 +1194,8 @@ static int dm_thread_func(void *dm_dev) int t; while (!kthread_should_stop()) { - t = wait_for_completion_timeout(&dm_device.config_event, 1*HZ); + t = wait_for_completion_interruptible_timeout( + &dm_device.config_event, 1*HZ); /* * The host expects us to post information on the memory * pressure every second. diff --git a/drivers/hv/hv_fcopy.c b/drivers/hv/hv_fcopy.c new file mode 100644 index 00000000000..23b2ce294c4 --- /dev/null +++ b/drivers/hv/hv_fcopy.c @@ -0,0 +1,414 @@ +/* + * An implementation of file copy service. + * + * Copyright (C) 2014, Microsoft, Inc. + * + * Author : K. Y. Srinivasan <ksrinivasan@novell.com> + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License version 2 as published + * by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + */ + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#include <linux/semaphore.h> +#include <linux/fs.h> +#include <linux/nls.h> +#include <linux/workqueue.h> +#include <linux/cdev.h> +#include <linux/hyperv.h> +#include <linux/sched.h> +#include <linux/uaccess.h> +#include <linux/miscdevice.h> + +#include "hyperv_vmbus.h" + +#define WIN8_SRV_MAJOR 1 +#define WIN8_SRV_MINOR 1 +#define WIN8_SRV_VERSION (WIN8_SRV_MAJOR << 16 | WIN8_SRV_MINOR) + +/* + * Global state maintained for transaction that is being processed. + * For a class of integration services, including the "file copy service", + * the specified protocol is a "request/response" protocol which means that + * there can only be single outstanding transaction from the host at any + * given point in time. We use this to simplify memory management in this + * driver - we cache and process only one message at a time. + * + * While the request/response protocol is guaranteed by the host, we further + * ensure this by serializing packet processing in this driver - we do not + * read additional packets from the VMBUs until the current packet is fully + * handled. + * + * The transaction "active" state is set when we receive a request from the + * host and we cleanup this state when the transaction is completed - when we + * respond to the host with our response. When the transaction active state is + * set, we defer handling incoming packets. + */ + +static struct { + bool active; /* transaction status - active or not */ + int recv_len; /* number of bytes received. */ + struct hv_fcopy_hdr *fcopy_msg; /* current message */ + struct hv_start_fcopy message; /* sent to daemon */ + struct vmbus_channel *recv_channel; /* chn we got the request */ + u64 recv_req_id; /* request ID. */ + void *fcopy_context; /* for the channel callback */ + struct semaphore read_sema; +} fcopy_transaction; + +static bool opened; /* currently device opened */ + +/* + * Before we can accept copy messages from the host, we need + * to handshake with the user level daemon. This state tracks + * if we are in the handshake phase. + */ +static bool in_hand_shake = true; +static void fcopy_send_data(void); +static void fcopy_respond_to_host(int error); +static void fcopy_work_func(struct work_struct *dummy); +static DECLARE_DELAYED_WORK(fcopy_work, fcopy_work_func); +static u8 *recv_buffer; + +static void fcopy_work_func(struct work_struct *dummy) +{ + /* + * If the timer fires, the user-mode component has not responded; + * process the pending transaction. + */ + fcopy_respond_to_host(HV_E_FAIL); +} + +static int fcopy_handle_handshake(u32 version) +{ + switch (version) { + case FCOPY_CURRENT_VERSION: + break; + default: + /* + * For now we will fail the registration. + * If and when we have multiple versions to + * deal with, we will be backward compatible. + * We will add this code when needed. + */ + return -EINVAL; + } + pr_info("FCP: user-mode registering done. Daemon version: %d\n", + version); + fcopy_transaction.active = false; + if (fcopy_transaction.fcopy_context) + hv_fcopy_onchannelcallback(fcopy_transaction.fcopy_context); + in_hand_shake = false; + return 0; +} + +static void fcopy_send_data(void) +{ + struct hv_start_fcopy *smsg_out = &fcopy_transaction.message; + int operation = fcopy_transaction.fcopy_msg->operation; + struct hv_start_fcopy *smsg_in; + + /* + * The strings sent from the host are encoded in + * in utf16; convert it to utf8 strings. + * The host assures us that the utf16 strings will not exceed + * the max lengths specified. We will however, reserve room + * for the string terminating character - in the utf16s_utf8s() + * function we limit the size of the buffer where the converted + * string is placed to W_MAX_PATH -1 to guarantee + * that the strings can be properly terminated! + */ + + switch (operation) { + case START_FILE_COPY: + memset(smsg_out, 0, sizeof(struct hv_start_fcopy)); + smsg_out->hdr.operation = operation; + smsg_in = (struct hv_start_fcopy *)fcopy_transaction.fcopy_msg; + + utf16s_to_utf8s((wchar_t *)smsg_in->file_name, W_MAX_PATH, + UTF16_LITTLE_ENDIAN, + (__u8 *)smsg_out->file_name, W_MAX_PATH - 1); + + utf16s_to_utf8s((wchar_t *)smsg_in->path_name, W_MAX_PATH, + UTF16_LITTLE_ENDIAN, + (__u8 *)smsg_out->path_name, W_MAX_PATH - 1); + + smsg_out->copy_flags = smsg_in->copy_flags; + smsg_out->file_size = smsg_in->file_size; + break; + + default: + break; + } + up(&fcopy_transaction.read_sema); + return; +} + +/* + * Send a response back to the host. + */ + +static void +fcopy_respond_to_host(int error) +{ + struct icmsg_hdr *icmsghdr; + u32 buf_len; + struct vmbus_channel *channel; + u64 req_id; + + /* + * Copy the global state for completing the transaction. Note that + * only one transaction can be active at a time. This is guaranteed + * by the file copy protocol implemented by the host. Furthermore, + * the "transaction active" state we maintain ensures that there can + * only be one active transaction at a time. + */ + + buf_len = fcopy_transaction.recv_len; + channel = fcopy_transaction.recv_channel; + req_id = fcopy_transaction.recv_req_id; + + fcopy_transaction.active = false; + + icmsghdr = (struct icmsg_hdr *) + &recv_buffer[sizeof(struct vmbuspipe_hdr)]; + + if (channel->onchannel_callback == NULL) + /* + * We have raced with util driver being unloaded; + * silently return. + */ + return; + + icmsghdr->status = error; + icmsghdr->icflags = ICMSGHDRFLAG_TRANSACTION | ICMSGHDRFLAG_RESPONSE; + vmbus_sendpacket(channel, recv_buffer, buf_len, req_id, + VM_PKT_DATA_INBAND, 0); +} + +void hv_fcopy_onchannelcallback(void *context) +{ + struct vmbus_channel *channel = context; + u32 recvlen; + u64 requestid; + struct hv_fcopy_hdr *fcopy_msg; + struct icmsg_hdr *icmsghdr; + struct icmsg_negotiate *negop = NULL; + int util_fw_version; + int fcopy_srv_version; + + if (fcopy_transaction.active) { + /* + * We will defer processing this callback once + * the current transaction is complete. + */ + fcopy_transaction.fcopy_context = context; + return; + } + + vmbus_recvpacket(channel, recv_buffer, PAGE_SIZE * 2, &recvlen, + &requestid); + if (recvlen <= 0) + return; + + icmsghdr = (struct icmsg_hdr *)&recv_buffer[ + sizeof(struct vmbuspipe_hdr)]; + if (icmsghdr->icmsgtype == ICMSGTYPE_NEGOTIATE) { + util_fw_version = UTIL_FW_VERSION; + fcopy_srv_version = WIN8_SRV_VERSION; + vmbus_prep_negotiate_resp(icmsghdr, negop, recv_buffer, + util_fw_version, fcopy_srv_version); + } else { + fcopy_msg = (struct hv_fcopy_hdr *)&recv_buffer[ + sizeof(struct vmbuspipe_hdr) + + sizeof(struct icmsg_hdr)]; + + /* + * Stash away this global state for completing the + * transaction; note transactions are serialized. + */ + + fcopy_transaction.active = true; + fcopy_transaction.recv_len = recvlen; + fcopy_transaction.recv_channel = channel; + fcopy_transaction.recv_req_id = requestid; + fcopy_transaction.fcopy_msg = fcopy_msg; + + /* + * Send the information to the user-level daemon. + */ + schedule_delayed_work(&fcopy_work, 5*HZ); + fcopy_send_data(); + return; + } + icmsghdr->icflags = ICMSGHDRFLAG_TRANSACTION | ICMSGHDRFLAG_RESPONSE; + vmbus_sendpacket(channel, recv_buffer, recvlen, requestid, + VM_PKT_DATA_INBAND, 0); +} + +/* + * Create a char device that can support read/write for passing + * the payload. + */ + +static ssize_t fcopy_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + void *src; + size_t copy_size; + int operation; + + /* + * Wait until there is something to be read. + */ + if (down_interruptible(&fcopy_transaction.read_sema)) + return -EINTR; + + /* + * The channel may be rescinded and in this case, we will wakeup the + * the thread blocked on the semaphore and we will use the opened + * state to correctly handle this case. + */ + if (!opened) + return -ENODEV; + + operation = fcopy_transaction.fcopy_msg->operation; + + if (operation == START_FILE_COPY) { + src = &fcopy_transaction.message; + copy_size = sizeof(struct hv_start_fcopy); + if (count < copy_size) + return 0; + } else { + src = fcopy_transaction.fcopy_msg; + copy_size = sizeof(struct hv_do_fcopy); + if (count < copy_size) + return 0; + } + if (copy_to_user(buf, src, copy_size)) + return -EFAULT; + + return copy_size; +} + +static ssize_t fcopy_write(struct file *file, const char __user *buf, + size_t count, loff_t *ppos) +{ + int response = 0; + + if (count != sizeof(int)) + return -EINVAL; + + if (copy_from_user(&response, buf, sizeof(int))) + return -EFAULT; + + if (in_hand_shake) { + if (fcopy_handle_handshake(response)) + return -EINVAL; + return sizeof(int); + } + + /* + * Complete the transaction by forwarding the result + * to the host. But first, cancel the timeout. + */ + if (cancel_delayed_work_sync(&fcopy_work)) + fcopy_respond_to_host(response); + + return sizeof(int); +} + +static int fcopy_open(struct inode *inode, struct file *f) +{ + /* + * The user level daemon that will open this device is + * really an extension of this driver. We can have only + * active open at a time. + */ + if (opened) + return -EBUSY; + + /* + * The daemon is alive; setup the state. + */ + opened = true; + return 0; +} + +static int fcopy_release(struct inode *inode, struct file *f) +{ + /* + * The daemon has exited; reset the state. + */ + in_hand_shake = true; + opened = false; + return 0; +} + + +static const struct file_operations fcopy_fops = { + .read = fcopy_read, + .write = fcopy_write, + .release = fcopy_release, + .open = fcopy_open, +}; + +static struct miscdevice fcopy_misc = { + .minor = MISC_DYNAMIC_MINOR, + .name = "vmbus/hv_fcopy", + .fops = &fcopy_fops, +}; + +static int fcopy_dev_init(void) +{ + return misc_register(&fcopy_misc); +} + +static void fcopy_dev_deinit(void) +{ + + /* + * The device is going away - perhaps because the + * host has rescinded the channel. Setup state so that + * user level daemon can gracefully exit if it is blocked + * on the read semaphore. + */ + opened = false; + /* + * Signal the semaphore as the device is + * going away. + */ + up(&fcopy_transaction.read_sema); + misc_deregister(&fcopy_misc); +} + +int hv_fcopy_init(struct hv_util_service *srv) +{ + recv_buffer = srv->recv_buffer; + + /* + * When this driver loads, the user level daemon that + * processes the host requests may not yet be running. + * Defer processing channel callbacks until the daemon + * has registered. + */ + fcopy_transaction.active = true; + sema_init(&fcopy_transaction.read_sema, 0); + + return fcopy_dev_init(); +} + +void hv_fcopy_deinit(void) +{ + cancel_delayed_work_sync(&fcopy_work); + fcopy_dev_deinit(); +} diff --git a/drivers/hv/hv_kvp.c b/drivers/hv/hv_kvp.c index 09988b28962..521c14625b3 100644 --- a/drivers/hv/hv_kvp.c +++ b/drivers/hv/hv_kvp.c @@ -113,7 +113,7 @@ kvp_register(int reg_value) kvp_msg->kvp_hdr.operation = reg_value; strcpy(version, HV_DRV_VERSION); msg->len = sizeof(struct hv_kvp_msg); - cn_netlink_send(msg, 0, GFP_ATOMIC); + cn_netlink_send(msg, 0, 0, GFP_ATOMIC); kfree(msg); } } @@ -127,6 +127,17 @@ kvp_work_func(struct work_struct *dummy) kvp_respond_to_host(NULL, HV_E_FAIL); } +static void poll_channel(struct vmbus_channel *channel) +{ + if (channel->target_cpu != smp_processor_id()) + smp_call_function_single(channel->target_cpu, + hv_kvp_onchannelcallback, + channel, true); + else + hv_kvp_onchannelcallback(channel); +} + + static int kvp_handle_handshake(struct hv_kvp_msg *msg) { int ret = 1; @@ -155,7 +166,7 @@ static int kvp_handle_handshake(struct hv_kvp_msg *msg) kvp_register(dm_reg_value); kvp_transaction.active = false; if (kvp_transaction.kvp_context) - hv_kvp_onchannelcallback(kvp_transaction.kvp_context); + poll_channel(kvp_transaction.kvp_context); } return ret; } @@ -435,7 +446,7 @@ kvp_send_key(struct work_struct *dummy) } msg->len = sizeof(struct hv_kvp_msg); - cn_netlink_send(msg, 0, GFP_ATOMIC); + cn_netlink_send(msg, 0, 0, GFP_ATOMIC); kfree(msg); return; @@ -568,7 +579,7 @@ response_done: vmbus_sendpacket(channel, recv_buffer, buf_len, req_id, VM_PKT_DATA_INBAND, 0); - + poll_channel(channel); } /* @@ -603,7 +614,7 @@ void hv_kvp_onchannelcallback(void *context) return; } - vmbus_recvpacket(channel, recv_buffer, PAGE_SIZE * 2, &recvlen, + vmbus_recvpacket(channel, recv_buffer, PAGE_SIZE * 4, &recvlen, &requestid); if (recvlen > 0) { diff --git a/drivers/hv/hv_snapshot.c b/drivers/hv/hv_snapshot.c index 0c354622437..34f14fddb66 100644 --- a/drivers/hv/hv_snapshot.c +++ b/drivers/hv/hv_snapshot.c @@ -98,7 +98,7 @@ static void vss_send_op(struct work_struct *dummy) vss_msg->vss_hdr.operation = op; msg->len = sizeof(struct hv_vss_msg); - cn_netlink_send(msg, 0, GFP_ATOMIC); + cn_netlink_send(msg, 0, 0, GFP_ATOMIC); kfree(msg); return; diff --git a/drivers/hv/hv_util.c b/drivers/hv/hv_util.c index 62dfd246b94..3b9c9ef0deb 100644 --- a/drivers/hv/hv_util.c +++ b/drivers/hv/hv_util.c @@ -28,6 +28,7 @@ #include <linux/reboot.h> #include <linux/hyperv.h> +#include "hyperv_vmbus.h" #define SD_MAJOR 3 #define SD_MINOR 0 @@ -82,6 +83,12 @@ static struct hv_util_service util_vss = { .util_deinit = hv_vss_deinit, }; +static struct hv_util_service util_fcopy = { + .util_cb = hv_fcopy_onchannelcallback, + .util_init = hv_fcopy_init, + .util_deinit = hv_fcopy_deinit, +}; + static void perform_shutdown(struct work_struct *dummy) { orderly_poweroff(true); @@ -312,7 +319,7 @@ static int util_probe(struct hv_device *dev, (struct hv_util_service *)dev_id->driver_data; int ret; - srv->recv_buffer = kmalloc(PAGE_SIZE * 2, GFP_KERNEL); + srv->recv_buffer = kmalloc(PAGE_SIZE * 4, GFP_KERNEL); if (!srv->recv_buffer) return -ENOMEM; if (srv->util_init) { @@ -401,6 +408,10 @@ static const struct hv_vmbus_device_id id_table[] = { { HV_VSS_GUID, .driver_data = (unsigned long)&util_vss }, + /* File copy GUID */ + { HV_FCOPY_GUID, + .driver_data = (unsigned long)&util_fcopy + }, { }, }; diff --git a/drivers/hv/hyperv_vmbus.h b/drivers/hv/hyperv_vmbus.h index e05517616a0..22b750749a3 100644 --- a/drivers/hv/hyperv_vmbus.h +++ b/drivers/hv/hyperv_vmbus.h @@ -510,6 +510,11 @@ struct hv_context { * basis. */ struct tasklet_struct *event_dpc[NR_CPUS]; + /* + * To optimize the mapping of relid to channel, maintain + * per-cpu list of the channels based on their CPU affinity. + */ + struct list_head percpu_list[NR_CPUS]; }; extern struct hv_context hv_context; @@ -559,8 +564,8 @@ int hv_ringbuffer_init(struct hv_ring_buffer_info *ring_info, void *buffer, void hv_ringbuffer_cleanup(struct hv_ring_buffer_info *ring_info); int hv_ringbuffer_write(struct hv_ring_buffer_info *ring_info, - struct scatterlist *sglist, - u32 sgcount, bool *signal); + struct kvec *kv_list, + u32 kv_count, bool *signal); int hv_ringbuffer_peek(struct hv_ring_buffer_info *ring_info, void *buffer, u32 buflen); @@ -644,9 +649,9 @@ extern struct vmbus_connection vmbus_connection; /* General vmbus interface */ -struct hv_device *vmbus_device_create(uuid_le *type, - uuid_le *instance, - struct vmbus_channel *channel); +struct hv_device *vmbus_device_create(const uuid_le *type, + const uuid_le *instance, + struct vmbus_channel *channel); int vmbus_device_register(struct hv_device *child_device_obj); void vmbus_device_unregister(struct hv_device *device_obj); @@ -669,5 +674,9 @@ int vmbus_set_event(struct vmbus_channel *channel); void vmbus_on_event(unsigned long data); +int hv_fcopy_init(struct hv_util_service *); +void hv_fcopy_deinit(void); +void hv_fcopy_onchannelcallback(void *); + #endif /* _HYPERV_VMBUS_H */ diff --git a/drivers/hv/ring_buffer.c b/drivers/hv/ring_buffer.c index 26c93cf9f6b..15db66b7414 100644 --- a/drivers/hv/ring_buffer.c +++ b/drivers/hv/ring_buffer.c @@ -26,6 +26,7 @@ #include <linux/kernel.h> #include <linux/mm.h> #include <linux/hyperv.h> +#include <linux/uio.h> #include "hyperv_vmbus.h" @@ -387,23 +388,20 @@ void hv_ringbuffer_cleanup(struct hv_ring_buffer_info *ring_info) * */ int hv_ringbuffer_write(struct hv_ring_buffer_info *outring_info, - struct scatterlist *sglist, u32 sgcount, bool *signal) + struct kvec *kv_list, u32 kv_count, bool *signal) { int i = 0; u32 bytes_avail_towrite; u32 bytes_avail_toread; u32 totalbytes_towrite = 0; - struct scatterlist *sg; u32 next_write_location; u32 old_write; u64 prev_indices = 0; unsigned long flags; - for_each_sg(sglist, sg, sgcount, i) - { - totalbytes_towrite += sg->length; - } + for (i = 0; i < kv_count; i++) + totalbytes_towrite += kv_list[i].iov_len; totalbytes_towrite += sizeof(u64); @@ -427,12 +425,11 @@ int hv_ringbuffer_write(struct hv_ring_buffer_info *outring_info, old_write = next_write_location; - for_each_sg(sglist, sg, sgcount, i) - { + for (i = 0; i < kv_count; i++) { next_write_location = hv_copyto_ringbuffer(outring_info, next_write_location, - sg_virt(sg), - sg->length); + kv_list[i].iov_base, + kv_list[i].iov_len); } /* Set previous packet start */ diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c index 077bb1bdac3..4d6b26979fb 100644 --- a/drivers/hv/vmbus_drv.c +++ b/drivers/hv/vmbus_drv.c @@ -25,7 +25,6 @@ #include <linux/init.h> #include <linux/module.h> #include <linux/device.h> -#include <linux/irq.h> #include <linux/interrupt.h> #include <linux/sysctl.h> #include <linux/slab.h> @@ -44,6 +43,12 @@ static struct tasklet_struct msg_dpc; static struct completion probe_event; static int irq; +struct resource hyperv_mmio = { + .name = "hyperv mmio", + .flags = IORESOURCE_MEM, +}; +EXPORT_SYMBOL_GPL(hyperv_mmio); + static int vmbus_exists(void) { if (hv_acpi_dev == NULL) @@ -430,7 +435,7 @@ static int vmbus_uevent(struct device *device, struct kobj_uevent_env *env) return ret; } -static uuid_le null_guid; +static const uuid_le null_guid; static inline bool is_null_guid(const __u8 *guid) { @@ -445,7 +450,7 @@ static inline bool is_null_guid(const __u8 *guid) */ static const struct hv_vmbus_device_id *hv_vmbus_get_id( const struct hv_vmbus_device_id *id, - __u8 *guid) + const __u8 *guid) { for (; !is_null_guid(id->guid); id++) if (!memcmp(&id->guid, guid, sizeof(uuid_le))) @@ -558,9 +563,6 @@ static struct bus_type hv_bus = { .dev_groups = vmbus_groups, }; -static const char *driver_name = "hyperv"; - - struct onmessage_work_context { struct work_struct work; struct hv_message msg; @@ -619,7 +621,7 @@ static void vmbus_on_msg_dpc(unsigned long data) } } -static irqreturn_t vmbus_isr(int irq, void *dev_id) +static void vmbus_isr(void) { int cpu = smp_processor_id(); void *page_addr; @@ -629,7 +631,7 @@ static irqreturn_t vmbus_isr(int irq, void *dev_id) page_addr = hv_context.synic_event_page[cpu]; if (page_addr == NULL) - return IRQ_NONE; + return; event = (union hv_synic_event_flags *)page_addr + VMBUS_MESSAGE_SINT; @@ -665,28 +667,8 @@ static irqreturn_t vmbus_isr(int irq, void *dev_id) msg = (struct hv_message *)page_addr + VMBUS_MESSAGE_SINT; /* Check if there are actual msgs to be processed */ - if (msg->header.message_type != HVMSG_NONE) { - handled = true; + if (msg->header.message_type != HVMSG_NONE) tasklet_schedule(&msg_dpc); - } - - if (handled) - return IRQ_HANDLED; - else - return IRQ_NONE; -} - -/* - * vmbus interrupt flow handler: - * vmbus interrupts can concurrently occur on multiple CPUs and - * can be handled concurrently. - */ - -static void vmbus_flow_handler(unsigned int irq, struct irq_desc *desc) -{ - kstat_incr_irqs_this_cpu(irq, desc); - - desc->action->handler(irq, desc->action->dev_id); } /* @@ -715,25 +697,7 @@ static int vmbus_bus_init(int irq) if (ret) goto err_cleanup; - ret = request_irq(irq, vmbus_isr, 0, driver_name, hv_acpi_dev); - - if (ret != 0) { - pr_err("Unable to request IRQ %d\n", - irq); - goto err_unregister; - } - - /* - * Vmbus interrupts can be handled concurrently on - * different CPUs. Establish an appropriate interrupt flow - * handler that can support this model. - */ - irq_set_handler(irq, vmbus_flow_handler); - - /* - * Register our interrupt handler. - */ - hv_register_vmbus_handler(irq, vmbus_isr); + hv_setup_vmbus_irq(vmbus_isr); ret = hv_synic_alloc(); if (ret) @@ -753,9 +717,8 @@ static int vmbus_bus_init(int irq) err_alloc: hv_synic_free(); - free_irq(irq, hv_acpi_dev); + hv_remove_vmbus_irq(); -err_unregister: bus_unregister(&hv_bus); err_cleanup: @@ -816,9 +779,9 @@ EXPORT_SYMBOL_GPL(vmbus_driver_unregister); * vmbus_device_create - Creates and registers a new child device * on the vmbus. */ -struct hv_device *vmbus_device_create(uuid_le *type, - uuid_le *instance, - struct vmbus_channel *channel) +struct hv_device *vmbus_device_create(const uuid_le *type, + const uuid_le *instance, + struct vmbus_channel *channel) { struct hv_device *child_device_obj; @@ -886,18 +849,21 @@ void vmbus_device_unregister(struct hv_device *device_obj) /* - * VMBUS is an acpi enumerated device. Get the the IRQ information - * from DSDT. + * VMBUS is an acpi enumerated device. Get the the information we + * need from DSDT. */ -static acpi_status vmbus_walk_resources(struct acpi_resource *res, void *irq) +static acpi_status vmbus_walk_resources(struct acpi_resource *res, void *ctx) { + switch (res->type) { + case ACPI_RESOURCE_TYPE_IRQ: + irq = res->data.irq.interrupts[0]; + break; - if (res->type == ACPI_RESOURCE_TYPE_IRQ) { - struct acpi_resource_irq *irqp; - irqp = &res->data.irq; - - *((unsigned int *)irq) = irqp->interrupts[0]; + case ACPI_RESOURCE_TYPE_ADDRESS64: + hyperv_mmio.start = res->data.address64.minimum; + hyperv_mmio.end = res->data.address64.maximum; + break; } return AE_OK; @@ -906,18 +872,34 @@ static acpi_status vmbus_walk_resources(struct acpi_resource *res, void *irq) static int vmbus_acpi_add(struct acpi_device *device) { acpi_status result; + int ret_val = -ENODEV; hv_acpi_dev = device; result = acpi_walk_resources(device->handle, METHOD_NAME__CRS, - vmbus_walk_resources, &irq); + vmbus_walk_resources, NULL); - if (ACPI_FAILURE(result)) { - complete(&probe_event); - return -ENODEV; + if (ACPI_FAILURE(result)) + goto acpi_walk_err; + /* + * The parent of the vmbus acpi device (Gen2 firmware) is the VMOD that + * has the mmio ranges. Get that. + */ + if (device->parent) { + result = acpi_walk_resources(device->parent->handle, + METHOD_NAME__CRS, + vmbus_walk_resources, NULL); + + if (ACPI_FAILURE(result)) + goto acpi_walk_err; + if (hyperv_mmio.start && hyperv_mmio.end) + request_resource(&iomem_resource, &hyperv_mmio); } + ret_val = 0; + +acpi_walk_err: complete(&probe_event); - return 0; + return ret_val; } static const struct acpi_device_id vmbus_acpi_device_ids[] = { @@ -947,7 +929,6 @@ static int __init hv_acpi_init(void) /* * Get irq resources first. */ - ret = acpi_bus_register_driver(&vmbus_acpi_driver); if (ret) @@ -978,8 +959,7 @@ cleanup: static void __exit vmbus_exit(void) { - - free_irq(irq, hv_acpi_dev); + hv_remove_vmbus_irq(); vmbus_free_channels(); bus_unregister(&hv_bus); hv_cleanup(); |
