diff options
Diffstat (limited to 'drivers/hv')
| -rw-r--r-- | drivers/hv/Kconfig | 2 | ||||
| -rw-r--r-- | drivers/hv/Makefile | 2 | ||||
| -rw-r--r-- | drivers/hv/channel.c | 208 | ||||
| -rw-r--r-- | drivers/hv/channel_mgmt.c | 349 | ||||
| -rw-r--r-- | drivers/hv/connection.c | 295 | ||||
| -rw-r--r-- | drivers/hv/hv.c | 143 | ||||
| -rw-r--r-- | drivers/hv/hv_balloon.c | 639 | ||||
| -rw-r--r-- | drivers/hv/hv_fcopy.c | 414 | ||||
| -rw-r--r-- | drivers/hv/hv_kvp.c | 61 | ||||
| -rw-r--r-- | drivers/hv/hv_snapshot.c | 281 | ||||
| -rw-r--r-- | drivers/hv/hv_util.c | 134 | ||||
| -rw-r--r-- | drivers/hv/hyperv_vmbus.h | 95 | ||||
| -rw-r--r-- | drivers/hv/ring_buffer.c | 150 | ||||
| -rw-r--r-- | drivers/hv/vmbus_drv.c | 640 |
14 files changed, 2764 insertions, 649 deletions
diff --git a/drivers/hv/Kconfig b/drivers/hv/Kconfig index b38ef6d8d04..0403b51d20b 100644 --- a/drivers/hv/Kconfig +++ b/drivers/hv/Kconfig @@ -2,7 +2,7 @@ menu "Microsoft Hyper-V guest support" config HYPERV tristate "Microsoft Hyper-V client drivers" - depends on X86 && ACPI && PCI + depends on X86 && ACPI && PCI && X86_LOCAL_APIC && HYPERVISOR_GUEST help Select this option to run Linux as a Hyper-V client operating system. diff --git a/drivers/hv/Makefile b/drivers/hv/Makefile index e6abfa02d8b..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_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 773a2f25a8f..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" @@ -47,63 +48,19 @@ static void vmbus_setevent(struct vmbus_channel *channel) (unsigned long *) vmbus_connection.send_int_page + (channel->offermsg.child_relid >> 5)); - monitorpage = vmbus_connection.monitor_pages; - monitorpage++; /* Get the child to parent monitor page */ + /* Get the child to parent monitor page */ + monitorpage = vmbus_connection.monitor_pages[1]; sync_set_bit(channel->monitor_bit, (unsigned long *)&monitorpage->trigger_group [channel->monitor_grp].pending); } else { - vmbus_set_event(channel->offermsg.child_relid); + vmbus_set_event(channel); } } /* - * vmbus_get_debug_info -Retrieve various channel debug info - */ -void vmbus_get_debug_info(struct vmbus_channel *channel, - struct vmbus_channel_debug_info *debuginfo) -{ - struct hv_monitor_page *monitorpage; - u8 monitor_group = (u8)channel->offermsg.monitorid / 32; - u8 monitor_offset = (u8)channel->offermsg.monitorid % 32; - - debuginfo->relid = channel->offermsg.child_relid; - debuginfo->state = channel->state; - memcpy(&debuginfo->interfacetype, - &channel->offermsg.offer.if_type, sizeof(uuid_le)); - memcpy(&debuginfo->interface_instance, - &channel->offermsg.offer.if_instance, - sizeof(uuid_le)); - - monitorpage = (struct hv_monitor_page *)vmbus_connection.monitor_pages; - - debuginfo->monitorid = channel->offermsg.monitorid; - - debuginfo->servermonitor_pending = - monitorpage->trigger_group[monitor_group].pending; - debuginfo->servermonitor_latency = - monitorpage->latency[monitor_group][monitor_offset]; - debuginfo->servermonitor_connectionid = - monitorpage->parameter[monitor_group] - [monitor_offset].connectionid.u.id; - - monitorpage++; - - debuginfo->clientmonitor_pending = - monitorpage->trigger_group[monitor_group].pending; - debuginfo->clientmonitor_latency = - monitorpage->latency[monitor_group][monitor_offset]; - debuginfo->clientmonitor_connectionid = - monitorpage->parameter[monitor_group] - [monitor_offset].connectionid.u.id; - - hv_ringbuffer_get_debuginfo(&channel->inbound, &debuginfo->inbound); - hv_ringbuffer_get_debuginfo(&channel->outbound, &debuginfo->outbound); -} - -/* * vmbus_open - Open the specified channel. */ int vmbus_open(struct vmbus_channel *newchannel, u32 send_ringbuffer_size, @@ -116,6 +73,15 @@ int vmbus_open(struct vmbus_channel *newchannel, u32 send_ringbuffer_size, unsigned long flags; int ret, t, err = 0; + spin_lock_irqsave(&newchannel->sc_lock, flags); + if (newchannel->state == CHANNEL_OPEN_STATE) { + newchannel->state = CHANNEL_OPENING_STATE; + } else { + spin_unlock_irqrestore(&newchannel->sc_lock, flags); + return -EINVAL; + } + spin_unlock_irqrestore(&newchannel->sc_lock, flags); + newchannel->onchannel_callback = onchannelcallback; newchannel->channel_callback_context = context; @@ -181,7 +147,7 @@ int vmbus_open(struct vmbus_channel *newchannel, u32 send_ringbuffer_size, open_msg->ringbuffer_gpadlhandle = newchannel->ringbuffer_gpadlhandle; open_msg->downstream_ringbuffer_pageoffset = send_ringbuffer_size >> PAGE_SHIFT; - open_msg->server_contextarea_gpadlhandle = 0; + open_msg->target_vp = newchannel->target_vp; if (userdatalen > MAX_USER_DEFINED_BYTES) { err = -EINVAL; @@ -216,6 +182,9 @@ int vmbus_open(struct vmbus_channel *newchannel, u32 send_ringbuffer_size, list_del(&open_info->msglistentry); spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, flags); + if (err == 0) + newchannel->state = CHANNEL_OPENED_STATE; + kfree(open_info); return err; @@ -241,7 +210,6 @@ static int create_gpadl_header(void *kbuffer, u32 size, { int i; int pagecount; - unsigned long long pfn; struct vmbus_channel_gpadl_header *gpadl_header; struct vmbus_channel_gpadl_body *gpadl_body; struct vmbus_channel_msginfo *msgheader; @@ -251,7 +219,6 @@ static int create_gpadl_header(void *kbuffer, u32 size, int pfnsum, pfncount, pfnleft, pfncurr, pfnsize; pagecount = size >> PAGE_SHIFT; - pfn = virt_to_phys(kbuffer) >> PAGE_SHIFT; /* do we need a gpadl body msg */ pfnsize = MAX_SIZE_CHANNEL_MESSAGE - @@ -280,7 +247,8 @@ static int create_gpadl_header(void *kbuffer, u32 size, gpadl_header->range[0].byte_offset = 0; gpadl_header->range[0].byte_count = size; for (i = 0; i < pfncount; i++) - gpadl_header->range[0].pfn_array[i] = pfn+i; + gpadl_header->range[0].pfn_array[i] = slow_virt_to_phys( + kbuffer + PAGE_SIZE * i) >> PAGE_SHIFT; *msginfo = msgheader; *messagecount = 1; @@ -333,7 +301,9 @@ static int create_gpadl_header(void *kbuffer, u32 size, * so the hypervisor gurantees that this is ok. */ for (i = 0; i < pfncurr; i++) - gpadl_body->pfn[i] = pfn + pfnsum + i; + gpadl_body->pfn[i] = slow_virt_to_phys( + kbuffer + PAGE_SIZE * (pfnsum + i)) >> + PAGE_SHIFT; /* add to msg header */ list_add_tail(&msgbody->msglistentry, @@ -359,7 +329,8 @@ static int create_gpadl_header(void *kbuffer, u32 size, gpadl_header->range[0].byte_offset = 0; gpadl_header->range[0].byte_count = size; for (i = 0; i < pagecount; i++) - gpadl_header->range[0].pfn_array[i] = pfn+i; + gpadl_header->range[0].pfn_array[i] = slow_virt_to_phys( + kbuffer + PAGE_SIZE * i) >> PAGE_SHIFT; *msginfo = msgheader; *messagecount = 1; @@ -376,7 +347,7 @@ nomem: * vmbus_establish_gpadl - Estabish a GPADL for the specified buffer * * @channel: a channel - * @kbuffer: from kmalloc + * @kbuffer: from kmalloc or vmalloc * @size: page-size multiple * @gpadl_handle: some funky thing */ @@ -500,19 +471,26 @@ int vmbus_teardown_gpadl(struct vmbus_channel *channel, u32 gpadl_handle) } EXPORT_SYMBOL_GPL(vmbus_teardown_gpadl); -/* - * vmbus_close - Close the specified channel - */ -void vmbus_close(struct vmbus_channel *channel) +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 */ @@ -538,6 +516,37 @@ void vmbus_close(struct vmbus_channel *channel) } + +/* + * vmbus_close - Close the specified channel + */ +void vmbus_close(struct vmbus_channel *channel) +{ + struct list_head *cur, *tmp; + struct vmbus_channel *cur_channel; + + if (channel->primary_channel != NULL) { + /* + * We will only close sub-channels when + * the primary is closed. + */ + return; + } + /* + * Close all the sub-channels first and then close the + * primary channel. + */ + list_for_each_safe(cur, tmp, &channel->sc_list) { + cur_channel = list_entry(cur, struct vmbus_channel, sc_list); + if (cur_channel->state != CHANNEL_OPENED_STATE) + continue; + vmbus_close_internal(cur_channel); + } + /* + * Now close the primary. + */ + vmbus_close_internal(channel); +} EXPORT_SYMBOL_GPL(vmbus_close); /** @@ -554,16 +563,17 @@ 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; /* Setup the descriptor */ @@ -574,15 +584,16 @@ 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); + ret = hv_ringbuffer_write(&channel->outbound, bufferlist, 3, &signal); - if (ret == 0 && !hv_get_ringbuffer_interrupt_mask(&channel->outbound)) + if (ret == 0 && signal) vmbus_setevent(channel); return ret; @@ -604,8 +615,9 @@ 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; if (pagecount > MAX_PAGE_BUFFER_COUNT) return -EINVAL; @@ -635,15 +647,16 @@ 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); + ret = hv_ringbuffer_write(&channel->outbound, bufferlist, 3, &signal); - if (ret == 0 && !hv_get_ringbuffer_interrupt_mask(&channel->outbound)) + if (ret == 0 && signal) vmbus_setevent(channel); return ret; @@ -663,13 +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; /* @@ -697,15 +710,16 @@ 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); + ret = hv_ringbuffer_write(&channel->outbound, bufferlist, 3, &signal); - if (ret == 0 && !hv_get_ringbuffer_interrupt_mask(&channel->outbound)) + if (ret == 0 && signal) vmbus_setevent(channel); return ret; @@ -732,6 +746,7 @@ int vmbus_recvpacket(struct vmbus_channel *channel, void *buffer, u32 packetlen; u32 userlen; int ret; + bool signal = false; *buffer_actual_len = 0; *requestid = 0; @@ -758,8 +773,10 @@ int vmbus_recvpacket(struct vmbus_channel *channel, void *buffer, /* Copy over the packet to the user buffer */ ret = hv_ringbuffer_read(&channel->inbound, buffer, userlen, - (desc.offset8 << 3)); + (desc.offset8 << 3), &signal); + if (signal) + vmbus_setevent(channel); return 0; } @@ -774,8 +791,8 @@ int vmbus_recvpacket_raw(struct vmbus_channel *channel, void *buffer, { struct vmpacket_descriptor desc; u32 packetlen; - u32 userlen; int ret; + bool signal = false; *buffer_actual_len = 0; *requestid = 0; @@ -788,7 +805,6 @@ int vmbus_recvpacket_raw(struct vmbus_channel *channel, void *buffer, packetlen = desc.len8 << 3; - userlen = packetlen - (desc.offset8 << 3); *buffer_actual_len = packetlen; @@ -802,8 +818,12 @@ int vmbus_recvpacket_raw(struct vmbus_channel *channel, void *buffer, *requestid = desc.trans_id; /* Copy over the entire packet to the user buffer */ - ret = hv_ringbuffer_read(&channel->inbound, buffer, packetlen, 0); + ret = hv_ringbuffer_read(&channel->inbound, buffer, packetlen, 0, + &signal); - return 0; + if (signal) + vmbus_setevent(channel); + + return ret; } EXPORT_SYMBOL_GPL(vmbus_recvpacket_raw); diff --git a/drivers/hv/channel_mgmt.c b/drivers/hv/channel_mgmt.c index 2f84c5cff8d..ed9350d4276 100644 --- a/drivers/hv/channel_mgmt.c +++ b/drivers/hv/channel_mgmt.c @@ -48,30 +48,39 @@ struct vmbus_channel_message_table_entry { * @negop is of type &struct icmsg_negotiate. * Set up and fill in default negotiate response message. * - * The max_fw_version specifies the maximum framework version that - * we can support and max _srv_version specifies the maximum service - * version we can support. A special value MAX_SRV_VER can be - * specified to indicate that we can handle the maximum version - * exposed by the host. + * The fw_version specifies the framework version that + * we can support and srv_version specifies the service + * version we can support. * * Mainly used by Hyper-V drivers. */ -void vmbus_prep_negotiate_resp(struct icmsg_hdr *icmsghdrp, +bool vmbus_prep_negotiate_resp(struct icmsg_hdr *icmsghdrp, struct icmsg_negotiate *negop, u8 *buf, - int max_fw_version, int max_srv_version) + int fw_version, int srv_version) { - int icframe_vercnt; - int icmsg_vercnt; + int icframe_major, icframe_minor; + int icmsg_major, icmsg_minor; + int fw_major, fw_minor; + int srv_major, srv_minor; int i; + bool found_match = false; icmsghdrp->icmsgsize = 0x10; + fw_major = (fw_version >> 16); + fw_minor = (fw_version & 0xFFFF); + + srv_major = (srv_version >> 16); + srv_minor = (srv_version & 0xFFFF); negop = (struct icmsg_negotiate *)&buf[ sizeof(struct vmbuspipe_hdr) + sizeof(struct icmsg_hdr)]; - icframe_vercnt = negop->icframe_vercnt; - icmsg_vercnt = negop->icmsg_vercnt; + icframe_major = negop->icframe_vercnt; + icframe_minor = 0; + + icmsg_major = negop->icmsg_vercnt; + icmsg_minor = 0; /* * Select the framework version number we will @@ -79,26 +88,48 @@ void vmbus_prep_negotiate_resp(struct icmsg_hdr *icmsghdrp, */ for (i = 0; i < negop->icframe_vercnt; i++) { - if (negop->icversion_data[i].major <= max_fw_version) - icframe_vercnt = negop->icversion_data[i].major; + if ((negop->icversion_data[i].major == fw_major) && + (negop->icversion_data[i].minor == fw_minor)) { + icframe_major = negop->icversion_data[i].major; + icframe_minor = negop->icversion_data[i].minor; + found_match = true; + } } + if (!found_match) + goto fw_error; + + found_match = false; + for (i = negop->icframe_vercnt; (i < negop->icframe_vercnt + negop->icmsg_vercnt); i++) { - if (negop->icversion_data[i].major <= max_srv_version) - icmsg_vercnt = negop->icversion_data[i].major; + if ((negop->icversion_data[i].major == srv_major) && + (negop->icversion_data[i].minor == srv_minor)) { + icmsg_major = negop->icversion_data[i].major; + icmsg_minor = negop->icversion_data[i].minor; + found_match = true; + } } /* - * Respond with the maximum framework and service + * Respond with the framework and service * version numbers we can support. */ - negop->icframe_vercnt = 1; - negop->icmsg_vercnt = 1; - negop->icversion_data[0].major = icframe_vercnt; - negop->icversion_data[0].minor = 0; - negop->icversion_data[1].major = icmsg_vercnt; - negop->icversion_data[1].minor = 0; + +fw_error: + if (!found_match) { + negop->icframe_vercnt = 0; + negop->icmsg_vercnt = 0; + } else { + negop->icframe_vercnt = 1; + negop->icmsg_vercnt = 1; + } + + negop->icversion_data[0].major = icframe_major; + negop->icversion_data[0].minor = icframe_minor; + negop->icversion_data[1].major = icmsg_major; + negop->icversion_data[1].minor = icmsg_minor; + return found_match; } EXPORT_SYMBOL_GPL(vmbus_prep_negotiate_resp); @@ -115,6 +146,10 @@ static struct vmbus_channel *alloc_channel(void) return NULL; spin_lock_init(&channel->inbound_lock); + 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) { @@ -154,7 +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 - @@ -165,8 +213,34 @@ static void vmbus_process_rescind_offer(struct work_struct *work) struct vmbus_channel *channel = container_of(work, struct vmbus_channel, work); + unsigned long flags; + struct vmbus_channel *primary_channel; + struct vmbus_channel_relid_released msg; + + if (channel->device_obj) + vmbus_device_unregister(channel->device_obj); + memset(&msg, 0, sizeof(struct vmbus_channel_relid_released)); + msg.child_relid = channel->offermsg.child_relid; + 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); - vmbus_device_unregister(channel->device_obj); + if (channel->primary_channel == NULL) { + spin_lock_irqsave(&vmbus_connection.channel_lock, flags); + list_del(&channel->listentry); + spin_unlock_irqrestore(&vmbus_connection.channel_lock, flags); + } else { + primary_channel = channel->primary_channel; + spin_lock_irqsave(&primary_channel->sc_lock, flags); + list_del(&channel->sc_list); + spin_unlock_irqrestore(&primary_channel->sc_lock, flags); + } + free_channel(channel); } void vmbus_free_channels(void) @@ -191,6 +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; @@ -210,18 +285,61 @@ 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. + */ + if (newchannel->offermsg.offer.sub_channel_index != 0) { + /* + * Process the sub-channel. + */ + newchannel->primary_channel = channel; + 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); + + return; + } + free_channel(newchannel); return; } /* + * This state is used to indicate a successful open + * so that when we do close the channel normally, we + * can cleanup properly + */ + newchannel->state = CHANNEL_OPEN_STATE; + + /* * Start the process of binding this offer to the driver * We need to set the DeviceObject field before calling * vmbus_child_dev_add() @@ -247,14 +365,74 @@ static void vmbus_process_offer(struct work_struct *work) kfree(newchannel->device_obj); free_channel(newchannel); - } else { + } +} + +enum { + IDE = 0, + SCSI, + NIC, + MAX_PERF_CHN, +}; + +/* + * This is an array of device_ids (device types) that are performance critical. + * We attempt to distribute the interrupt load for these devices across + * all available CPUs. + */ +static const struct hv_vmbus_device_id hp_devs[] = { + /* IDE */ + { HV_IDE_GUID, }, + /* Storage - SCSI */ + { HV_SCSI_GUID, }, + /* Network */ + { HV_NIC_GUID, }, +}; + + +/* + * We use this state to statically distribute the channel interrupt load. + */ +static u32 next_vp; + +/* + * Starting with Win8, we can statically distribute the incoming + * channel interrupt load by binding a channel to VCPU. We + * implement here a simple round robin scheme for distributing + * the interrupt load. + * We will bind channels that are not performance critical to cpu 0 and + * performance critical channels (IDE, SCSI and Network) will be uniformly + * distributed across all available CPUs. + */ +static void init_vp_index(struct vmbus_channel *channel, const uuid_le *type_guid) +{ + u32 cur_cpu; + int i; + bool perf_chn = false; + u32 max_cpus = num_online_cpus(); + + for (i = IDE; i < MAX_PERF_CHN; i++) { + if (!memcmp(type_guid->b, hp_devs[i].guid, + sizeof(uuid_le))) { + perf_chn = true; + break; + } + } + if ((vmbus_proto_version == VERSION_WS2008) || + (vmbus_proto_version == VERSION_WIN7) || (!perf_chn)) { /* - * This state is used to indicate a successful open - * so that when we do close the channel normally, we - * can cleanup properly + * Prior to win8, all channel interrupts are + * delivered on cpu 0. + * Also if the channel is not a performance critical + * channel, bind it to cpu 0. */ - newchannel->state = CHANNEL_OPEN_STATE; + channel->target_cpu = 0; + channel->target_vp = 0; + return; } + cur_cpu = (++next_vp % max_cpus); + channel->target_cpu = cur_cpu; + channel->target_vp = hv_context.vp_index[cur_cpu]; } /* @@ -275,6 +453,35 @@ static void vmbus_onoffer(struct vmbus_channel_message_header *hdr) return; } + /* + * By default we setup state to enable batched + * reading. A specific service can choose to + * disable this prior to opening the channel. + */ + newchannel->batched_reading = true; + + /* + * Setup state for signalling the host. + */ + newchannel->sig_event = (struct hv_input_signal_event *) + (ALIGN((unsigned long) + &newchannel->sig_buf, + HV_HYPERCALL_PARAM_ALIGN)); + + newchannel->sig_event->connectionid.asu32 = 0; + newchannel->sig_event->connectionid.u.id = VMBUS_EVENT_CONNECTION_ID; + newchannel->sig_event->flag_number = 0; + newchannel->sig_event->rsvdz = 0; + + if (vmbus_proto_version != VERSION_WS2008) { + newchannel->is_dedicated_interrupt = + (offer->is_dedicated_interrupt != 0); + newchannel->sig_event->connectionid.u.id = + offer->connection_id; + } + + init_vp_index(newchannel, &offer->offer.if_type); + memcpy(&newchannel->offermsg, offer, sizeof(struct vmbus_channel_offer_channel)); newchannel->monitor_grp = (u8)offer->monitorid / 32; @@ -581,4 +788,86 @@ cleanup: return ret; } -/* eof */ +/* + * Retrieve the (sub) channel on which to send an outgoing request. + * When a primary channel has multiple sub-channels, we choose a + * channel whose VCPU binding is closest to the VCPU on which + * this call is being made. + */ +struct vmbus_channel *vmbus_get_outgoing_channel(struct vmbus_channel *primary) +{ + struct list_head *cur, *tmp; + int cur_cpu = hv_context.vp_index[smp_processor_id()]; + struct vmbus_channel *cur_channel; + struct vmbus_channel *outgoing_channel = primary; + int cpu_distance, new_cpu_distance; + + if (list_empty(&primary->sc_list)) + return outgoing_channel; + + list_for_each_safe(cur, tmp, &primary->sc_list) { + cur_channel = list_entry(cur, struct vmbus_channel, sc_list); + if (cur_channel->state != CHANNEL_OPENED_STATE) + continue; + + if (cur_channel->target_vp == cur_cpu) + return cur_channel; + + cpu_distance = ((outgoing_channel->target_vp > cur_cpu) ? + (outgoing_channel->target_vp - cur_cpu) : + (cur_cpu - outgoing_channel->target_vp)); + + new_cpu_distance = ((cur_channel->target_vp > cur_cpu) ? + (cur_channel->target_vp - cur_cpu) : + (cur_cpu - cur_channel->target_vp)); + + if (cpu_distance < new_cpu_distance) + continue; + + outgoing_channel = cur_channel; + } + + return outgoing_channel; +} +EXPORT_SYMBOL_GPL(vmbus_get_outgoing_channel); + +static void invoke_sc_cb(struct vmbus_channel *primary_channel) +{ + struct list_head *cur, *tmp; + struct vmbus_channel *cur_channel; + + if (primary_channel->sc_creation_callback == NULL) + return; + + list_for_each_safe(cur, tmp, &primary_channel->sc_list) { + cur_channel = list_entry(cur, struct vmbus_channel, sc_list); + + primary_channel->sc_creation_callback(cur_channel); + } +} + +void vmbus_set_sc_create_callback(struct vmbus_channel *primary_channel, + void (*sc_cr_cb)(struct vmbus_channel *new_sc)) +{ + primary_channel->sc_creation_callback = sc_cr_cb; +} +EXPORT_SYMBOL_GPL(vmbus_set_sc_create_callback); + +bool vmbus_are_subchannels_present(struct vmbus_channel *primary) +{ + bool ret; + + ret = !list_empty(&primary->sc_list); + + if (ret) { + /* + * Invoke the callback on sub-channel creation. + * This will present a uniform interface to the + * clients. + */ + invoke_sc_cb(primary); + } + + return ret; +} +EXPORT_SYMBOL_GPL(vmbus_are_subchannels_present); diff --git a/drivers/hv/connection.c b/drivers/hv/connection.c index 650c9f0b664..ae22e3c1fc4 100644 --- a/drivers/hv/connection.c +++ b/drivers/hv/connection.c @@ -30,6 +30,7 @@ #include <linux/slab.h> #include <linux/vmalloc.h> #include <linux/hyperv.h> +#include <linux/export.h> #include <asm/hyperv.h> #include "hyperv_vmbus.h" @@ -40,15 +41,93 @@ struct vmbus_connection vmbus_connection = { }; /* + * Negotiated protocol version with the host. + */ +__u32 vmbus_proto_version; +EXPORT_SYMBOL_GPL(vmbus_proto_version); + +static __u32 vmbus_get_next_version(__u32 current_version) +{ + switch (current_version) { + case (VERSION_WIN7): + return VERSION_WS2008; + + case (VERSION_WIN8): + return VERSION_WIN7; + + case (VERSION_WIN8_1): + return VERSION_WIN8; + + case (VERSION_WS2008): + default: + return VERSION_INVAL; + } +} + +static int vmbus_negotiate_version(struct vmbus_channel_msginfo *msginfo, + __u32 version) +{ + int ret = 0; + struct vmbus_channel_initiate_contact *msg; + unsigned long flags; + + init_completion(&msginfo->waitevent); + + msg = (struct vmbus_channel_initiate_contact *)msginfo->msg; + + msg->header.msgtype = CHANNELMSG_INITIATE_CONTACT; + msg->vmbus_version_requested = version; + 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 + * receive the response before returning from this routine + */ + spin_lock_irqsave(&vmbus_connection.channelmsg_lock, flags); + list_add_tail(&msginfo->msglistentry, + &vmbus_connection.chn_msg_list); + + spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, flags); + + ret = vmbus_post_msg(msg, + sizeof(struct vmbus_channel_initiate_contact)); + if (ret != 0) { + spin_lock_irqsave(&vmbus_connection.channelmsg_lock, flags); + list_del(&msginfo->msglistentry); + spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, + flags); + return ret; + } + + /* Wait for the connection response */ + wait_for_completion(&msginfo->waitevent); + + spin_lock_irqsave(&vmbus_connection.channelmsg_lock, flags); + list_del(&msginfo->msglistentry); + spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, flags); + + /* Check if successful */ + if (msginfo->response.version_response.version_supported) { + vmbus_connection.conn_state = CONNECTED; + } else { + return -ECONNREFUSED; + } + + return ret; +} + +/* * vmbus_connect - Sends a connect request on the partition service connection */ int vmbus_connect(void) { int ret = 0; - int t; struct vmbus_channel_msginfo *msginfo = NULL; - struct vmbus_channel_initiate_contact *msg; - unsigned long flags; + __u32 version; /* Initialize the vmbus connection */ vmbus_connection.conn_state = CONNECTING; @@ -84,9 +163,10 @@ int vmbus_connect(void) * Setup the monitor notification facility. The 1st page for * parent->child and the 2nd page for child->parent */ - vmbus_connection.monitor_pages = - (void *)__get_free_pages((GFP_KERNEL|__GFP_ZERO), 1); - if (vmbus_connection.monitor_pages == NULL) { + vmbus_connection.monitor_pages[0] = (void *)__get_free_pages((GFP_KERNEL|__GFP_ZERO), 0); + vmbus_connection.monitor_pages[1] = (void *)__get_free_pages((GFP_KERNEL|__GFP_ZERO), 0); + if ((vmbus_connection.monitor_pages[0] == NULL) || + (vmbus_connection.monitor_pages[1] == NULL)) { ret = -ENOMEM; goto cleanup; } @@ -99,69 +179,41 @@ int vmbus_connect(void) goto cleanup; } - init_completion(&msginfo->waitevent); - - msg = (struct vmbus_channel_initiate_contact *)msginfo->msg; - - msg->header.msgtype = CHANNELMSG_INITIATE_CONTACT; - msg->vmbus_version_requested = VMBUS_REVISION_NUMBER; - msg->interrupt_page = virt_to_phys(vmbus_connection.int_page); - msg->monitor_page1 = virt_to_phys(vmbus_connection.monitor_pages); - msg->monitor_page2 = virt_to_phys( - (void *)((unsigned long)vmbus_connection.monitor_pages + - PAGE_SIZE)); - /* - * Add to list before we send the request since we may - * receive the response before returning from this routine + * Negotiate a compatible VMBUS version number with the + * host. We start with the highest number we can support + * and work our way down until we negotiate a compatible + * version. */ - spin_lock_irqsave(&vmbus_connection.channelmsg_lock, flags); - list_add_tail(&msginfo->msglistentry, - &vmbus_connection.chn_msg_list); - spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, flags); + version = VERSION_CURRENT; - ret = vmbus_post_msg(msg, - sizeof(struct vmbus_channel_initiate_contact)); - if (ret != 0) { - spin_lock_irqsave(&vmbus_connection.channelmsg_lock, flags); - list_del(&msginfo->msglistentry); - spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, - flags); - goto cleanup; - } + do { + ret = vmbus_negotiate_version(msginfo, version); + if (ret == -ETIMEDOUT) + goto cleanup; - /* 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); - ret = -ETIMEDOUT; - goto cleanup; - } + if (vmbus_connection.conn_state == CONNECTED) + break; - spin_lock_irqsave(&vmbus_connection.channelmsg_lock, flags); - list_del(&msginfo->msglistentry); - spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, flags); + version = vmbus_get_next_version(version); + } while (version != VERSION_INVAL); - /* Check if successful */ - if (msginfo->response.version_response.version_supported) { - vmbus_connection.conn_state = CONNECTED; - } else { - pr_err("Unable to connect, " - "Version %d not supported by Hyper-V\n", - VMBUS_REVISION_NUMBER); - ret = -ECONNREFUSED; + if (version == VERSION_INVAL) goto cleanup; - } + + vmbus_proto_version = version; + pr_info("Hyper-V Host Build:%d-%d.%d-%d-%d.%d; Vmbus version:%d.%d\n", + host_info_eax, host_info_ebx >> 16, + host_info_ebx & 0xFFFF, host_info_ecx, + host_info_edx >> 24, host_info_edx & 0xFFFFFF, + version >> 16, version & 0xFFFF); kfree(msginfo); return 0; cleanup: + pr_err("Unable to connect to host\n"); vmbus_connection.conn_state = DISCONNECTED; if (vmbus_connection.work_queue) @@ -172,16 +224,38 @@ cleanup: vmbus_connection.int_page = NULL; } - if (vmbus_connection.monitor_pages) { - free_pages((unsigned long)vmbus_connection.monitor_pages, 1); - vmbus_connection.monitor_pages = NULL; - } + 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; kfree(msginfo); 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 @@ -192,12 +266,26 @@ struct vmbus_channel *relid2channel(u32 relid) struct vmbus_channel *channel; struct vmbus_channel *found_channel = NULL; unsigned long flags; + struct list_head *cur, *tmp; + struct vmbus_channel *cur_sc; spin_lock_irqsave(&vmbus_connection.channel_lock, flags); list_for_each_entry(channel, &vmbus_connection.chn_list, listentry) { if (channel->offermsg.child_relid == relid) { found_channel = channel; break; + } else if (!list_empty(&channel->sc_list)) { + /* + * Deal with sub-channels. + */ + list_for_each_safe(cur, tmp, &channel->sc_list) { + cur_sc = list_entry(cur, struct vmbus_channel, + sc_list); + if (cur_sc->offermsg.child_relid == relid) { + found_channel = cur_sc; + break; + } + } } } spin_unlock_irqrestore(&vmbus_connection.channel_lock, flags); @@ -211,13 +299,15 @@ 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; /* * 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); @@ -227,19 +317,40 @@ 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) - channel->onchannel_callback(channel->channel_callback_context); - else + if (channel->onchannel_callback != NULL) { + arg = channel->channel_callback_context; + read_state = channel->batched_reading; + /* + * This callback reads the messages sent by the host. + * We can optimize host to guest signaling by ensuring: + * 1. While reading the channel, we disable interrupts from + * host. + * 2. Ensure that we process all posted messages from the host + * before returning from this callback. + * 3. Once we return, enable signaling from the host. Once this + * state is set we check to see if additional packets are + * available to read. In this case we repeat the process. + */ + + do { + if (read_state) + hv_begin_read(&channel->inbound); + channel->onchannel_callback(arg); + 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); } /* @@ -248,10 +359,32 @@ static void process_chn_event(u32 relid) void vmbus_on_event(unsigned long data) { u32 dword; - u32 maxdword = MAX_NUM_CHANNELS_SUPPORTED >> 5; + u32 maxdword; int bit; u32 relid; - u32 *recv_int_page = vmbus_connection.recv_int_page; + u32 *recv_int_page = NULL; + void *page_addr; + int cpu = smp_processor_id(); + union hv_synic_event_flags *event; + + if ((vmbus_proto_version == VERSION_WS2008) || + (vmbus_proto_version == VERSION_WIN7)) { + maxdword = MAX_NUM_CHANNELS_SUPPORTED >> 5; + recv_int_page = vmbus_connection.recv_int_page; + } else { + /* + * When the host is win8 and beyond, the event page + * can be directly checked to get the id of the channel + * that has the interrupt pending. + */ + maxdword = HV_EVENT_FLAGS_DWORD_COUNT; + page_addr = hv_context.synic_event_page[cpu]; + event = (union hv_synic_event_flags *)page_addr + + VMBUS_MESSAGE_SINT; + recv_int_page = event->flags32; + } + + /* Check events */ if (!recv_int_page) @@ -307,12 +440,16 @@ int vmbus_post_msg(void *buffer, size_t buflen) /* * vmbus_set_event - Send an event notification to the parent */ -int vmbus_set_event(u32 child_relid) +int vmbus_set_event(struct vmbus_channel *channel) { - /* Each u32 represents 32 channels */ - sync_set_bit(child_relid & 31, - (unsigned long *)vmbus_connection.send_int_page + - (child_relid >> 5)); + u32 child_relid = channel->offermsg.child_relid; + + if (!channel->is_dedicated_interrupt) { + /* Each u32 represents 32 channels */ + sync_set_bit(child_relid & 31, + (unsigned long *)vmbus_connection.send_int_page + + (child_relid >> 5)); + } - return hv_signal_event(); + return hv_signal_event(channel->sig_event); } diff --git a/drivers/hv/hv.c b/drivers/hv/hv.c index 3648f8f0f36..edfc8488cb0 100644 --- a/drivers/hv/hv.c +++ b/drivers/hv/hv.c @@ -27,6 +27,7 @@ #include <linux/vmalloc.h> #include <linux/hyperv.h> #include <linux/version.h> +#include <linux/interrupt.h> #include <asm/hyperv.h> #include "hyperv_vmbus.h" @@ -34,13 +35,16 @@ struct hv_context hv_context = { .synic_initialized = false, .hypercall_page = NULL, - .signal_event_param = NULL, - .signal_event_buffer = NULL, }; /* * query_hypervisor_info - Get version info of the windows hypervisor */ +unsigned int host_info_eax; +unsigned int host_info_ebx; +unsigned int host_info_ecx; +unsigned int host_info_edx; + static int query_hypervisor_info(void) { unsigned int eax; @@ -70,13 +74,10 @@ static int query_hypervisor_info(void) edx = 0; op = HVCPUID_VERSION; cpuid(op, &eax, &ebx, &ecx, &edx); - pr_info("Hyper-V Host OS Build:%d-%d.%d-%d-%d.%d\n", - eax, - ebx >> 16, - ebx & 0xFFFF, - ecx, - edx >> 24, - edx & 0xFFFFFF); + host_info_eax = eax; + host_info_ebx = ebx; + host_info_ecx = ecx; + host_info_edx = edx; } return max_leaf; } @@ -137,6 +138,10 @@ int hv_init(void) memset(hv_context.synic_event_page, 0, sizeof(void *) * NR_CPUS); memset(hv_context.synic_message_page, 0, sizeof(void *) * NR_CPUS); + memset(hv_context.vp_index, 0, + sizeof(int) * NR_CPUS); + memset(hv_context.event_dpc, 0, + sizeof(void *) * NR_CPUS); max_leaf = query_hypervisor_info(); @@ -168,24 +173,6 @@ int hv_init(void) hv_context.hypercall_page = virtaddr; - /* Setup the global signal event param for the signal event hypercall */ - hv_context.signal_event_buffer = - kmalloc(sizeof(struct hv_input_signal_event_buffer), - GFP_KERNEL); - if (!hv_context.signal_event_buffer) - goto cleanup; - - hv_context.signal_event_param = - (struct hv_input_signal_event *) - (ALIGN((unsigned long) - hv_context.signal_event_buffer, - HV_HYPERCALL_PARAM_ALIGN)); - hv_context.signal_event_param->connectionid.asu32 = 0; - hv_context.signal_event_param->connectionid.u.id = - VMBUS_EVENT_CONNECTION_ID; - hv_context.signal_event_param->flag_number = 0; - hv_context.signal_event_param->rsvdz = 0; - return 0; cleanup: @@ -213,10 +200,6 @@ void hv_cleanup(void) /* Reset our OS id */ wrmsrl(HV_X64_MSR_GUEST_OS_ID, 0); - kfree(hv_context.signal_event_buffer); - hv_context.signal_event_buffer = NULL; - hv_context.signal_event_param = NULL; - if (hv_context.hypercall_page) { hypercall_msr.as_uint64 = 0; wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64); @@ -273,16 +256,68 @@ int hv_post_message(union hv_connection_id connection_id, * * This involves a hypercall. */ -u16 hv_signal_event(void) +u16 hv_signal_event(void *con_id) { u16 status; - status = do_hypercall(HVCALL_SIGNAL_EVENT, - hv_context.signal_event_param, - NULL) & 0xFFFF; + status = (do_hypercall(HVCALL_SIGNAL_EVENT, con_id, NULL) & 0xFFFF); + return status; } + +int hv_synic_alloc(void) +{ + size_t size = sizeof(struct tasklet_struct); + int cpu; + + for_each_online_cpu(cpu) { + hv_context.event_dpc[cpu] = kmalloc(size, GFP_ATOMIC); + if (hv_context.event_dpc[cpu] == NULL) { + pr_err("Unable to allocate event dpc\n"); + goto err; + } + tasklet_init(hv_context.event_dpc[cpu], vmbus_on_event, cpu); + + hv_context.synic_message_page[cpu] = + (void *)get_zeroed_page(GFP_ATOMIC); + + if (hv_context.synic_message_page[cpu] == NULL) { + pr_err("Unable to allocate SYNIC message page\n"); + goto err; + } + + hv_context.synic_event_page[cpu] = + (void *)get_zeroed_page(GFP_ATOMIC); + + if (hv_context.synic_event_page[cpu] == NULL) { + pr_err("Unable to allocate SYNIC event page\n"); + goto err; + } + } + + return 0; +err: + return -ENOMEM; +} + +static void hv_synic_free_cpu(int cpu) +{ + kfree(hv_context.event_dpc[cpu]); + if (hv_context.synic_event_page[cpu]) + free_page((unsigned long)hv_context.synic_event_page[cpu]); + if (hv_context.synic_message_page[cpu]) + free_page((unsigned long)hv_context.synic_message_page[cpu]); +} + +void hv_synic_free(void) +{ + int cpu; + + for_each_online_cpu(cpu) + hv_synic_free_cpu(cpu); +} + /* * hv_synic_init - Initialize the Synthethic Interrupt Controller. * @@ -290,15 +325,15 @@ u16 hv_signal_event(void) * retrieve the initialized message and event pages. Otherwise, we create and * initialize the message and event pages. */ -void hv_synic_init(void *irqarg) +void hv_synic_init(void *arg) { u64 version; union hv_synic_simp simp; union hv_synic_siefp siefp; union hv_synic_sint shared_sint; union hv_synic_scontrol sctrl; + u64 vp_index; - u32 irq_vector = *((u32 *)(irqarg)); int cpu = smp_processor_id(); if (!hv_context.hypercall_page) @@ -307,22 +342,6 @@ void hv_synic_init(void *irqarg) /* Check the version */ rdmsrl(HV_X64_MSR_SVERSION, version); - hv_context.synic_message_page[cpu] = - (void *)get_zeroed_page(GFP_ATOMIC); - - if (hv_context.synic_message_page[cpu] == NULL) { - pr_err("Unable to allocate SYNIC message page\n"); - goto cleanup; - } - - hv_context.synic_event_page[cpu] = - (void *)get_zeroed_page(GFP_ATOMIC); - - if (hv_context.synic_event_page[cpu] == NULL) { - pr_err("Unable to allocate SYNIC event page\n"); - goto cleanup; - } - /* Setup the Synic's message page */ rdmsrl(HV_X64_MSR_SIMP, simp.as_uint64); simp.simp_enabled = 1; @@ -343,9 +362,9 @@ void hv_synic_init(void *irqarg) rdmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64); shared_sint.as_uint64 = 0; - shared_sint.vector = irq_vector; /* HV_SHARED_SINT_IDT_VECTOR + 0x20; */ + shared_sint.vector = HYPERVISOR_CALLBACK_VECTOR; shared_sint.masked = false; - shared_sint.auto_eoi = false; + shared_sint.auto_eoi = true; wrmsrl(HV_X64_MSR_SINT0 + VMBUS_MESSAGE_SINT, shared_sint.as_uint64); @@ -356,14 +375,16 @@ void hv_synic_init(void *irqarg) wrmsrl(HV_X64_MSR_SCONTROL, sctrl.as_uint64); hv_context.synic_initialized = true; - return; -cleanup: - if (hv_context.synic_event_page[cpu]) - free_page((unsigned long)hv_context.synic_event_page[cpu]); + /* + * Setup the mapping between Hyper-V's notion + * of cpuid and Linux' notion of cpuid. + * This array will be indexed using Linux cpuid. + */ + rdmsrl(HV_X64_MSR_VP_INDEX, vp_index); + hv_context.vp_index[cpu] = (u32)vp_index; - if (hv_context.synic_message_page[cpu]) - free_page((unsigned long)hv_context.synic_message_page[cpu]); + INIT_LIST_HEAD(&hv_context.percpu_list[cpu]); return; } diff --git a/drivers/hv/hv_balloon.c b/drivers/hv/hv_balloon.c index dd289fd179c..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> @@ -29,7 +30,6 @@ #include <linux/memory_hotplug.h> #include <linux/memory.h> #include <linux/notifier.h> -#include <linux/mman.h> #include <linux/percpu_counter.h> #include <linux/hyperv.h> @@ -118,7 +118,14 @@ union dm_caps { struct { __u64 balloon:1; __u64 hot_add:1; - __u64 reservedz:62; + /* + * To support guests that may have alignment + * limitations on hot-add, the guest can specify + * its alignment requirements; a value of n + * represents an alignment of 2^n in mega bytes. + */ + __u64 hot_add_alignment:4; + __u64 reservedz:58; } cap_bits; __u64 caps; } __packed; @@ -413,12 +420,56 @@ struct dm_info_msg { * End protocol definitions. */ -static bool hot_add; +/* + * State to manage hot adding memory into the guest. + * The range start_pfn : end_pfn specifies the range + * that the host has asked us to hot add. The range + * start_pfn : ha_end_pfn specifies the range that we have + * currently hot added. We hot add in multiples of 128M + * chunks; it is possible that we may not be able to bring + * online all the pages in the region. The range + * covered_start_pfn : covered_end_pfn defines the pages that can + * be brough online. + */ + +struct hv_hotadd_state { + struct list_head list; + unsigned long start_pfn; + unsigned long covered_start_pfn; + unsigned long covered_end_pfn; + unsigned long ha_end_pfn; + unsigned long end_pfn; +}; + +struct balloon_state { + __u32 num_pages; + struct work_struct wrk; +}; + +struct hot_add_wrk { + union dm_mem_page_range ha_page_range; + union dm_mem_page_range ha_region_range; + struct work_struct wrk; +}; + +static bool hot_add = true; static bool do_hot_add; +/* + * Delay reporting memory pressure by + * the specified number of seconds. + */ +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"); +module_param(pressure_report_delay, uint, (S_IRUGO | S_IWUSR)); +MODULE_PARM_DESC(pressure_report_delay, "Delay in secs in reporting pressure"); static atomic_t trans_id = ATOMIC_INIT(0); static int dm_ring_size = (5 * PAGE_SIZE); @@ -440,6 +491,7 @@ enum hv_dm_state { static __u8 recv_buffer[PAGE_SIZE]; static __u8 *send_buffer; #define PAGES_IN_2M 512 +#define HA_CHUNK (32 * 1024) struct hv_dynmem_device { struct hv_device *dev; @@ -453,7 +505,28 @@ struct hv_dynmem_device { unsigned int num_pages_ballooned; /* - * This thread handles both balloon/hot-add + * State to manage the ballooning (up) operation. + */ + struct balloon_state balloon_wrk; + + /* + * State to execute the "hot-add" operation. + */ + struct hot_add_wrk ha_wrk; + + /* + * This state tracks if the host has specified a hot-add + * region. + */ + bool host_specified_ha_region; + + /* + * State to synchronize hot-add. + */ + struct completion ol_waitevent; + bool ha_waiting; + /* + * This thread handles hot-add * requests from the host as well as notifying * the host with regards to memory pressure in * the guest. @@ -461,6 +534,11 @@ struct hv_dynmem_device { struct task_struct *thread; /* + * A list of hot-add regions. + */ + struct list_head ha_region_list; + + /* * We start with the highest version we can support * and downgrade based on the host; we save here the * next version to try. @@ -470,35 +548,356 @@ struct hv_dynmem_device { static struct hv_dynmem_device dm_device; -static void hot_add_req(struct hv_dynmem_device *dm, struct dm_hot_add *msg) +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) { + int i; - struct dm_hot_add_response resp; + for (i = 0; i < size; i++) { + struct page *pg; + pg = pfn_to_page(start_pfn + i); + __online_page_set_limits(pg); + __online_page_increment_counters(pg); + __online_page_free(pg); + } +} + +static void hv_mem_hot_add(unsigned long start, unsigned long size, + unsigned long pfn_count, + struct hv_hotadd_state *has) +{ + int ret = 0; + int i, nid; + unsigned long start_pfn; + unsigned long processed_pfn; + unsigned long total_pfn = pfn_count; + + for (i = 0; i < (size/HA_CHUNK); i++) { + start_pfn = start + (i * HA_CHUNK); + has->ha_end_pfn += HA_CHUNK; + + if (total_pfn > HA_CHUNK) { + processed_pfn = HA_CHUNK; + total_pfn -= HA_CHUNK; + } else { + processed_pfn = total_pfn; + total_pfn = 0; + } + + has->covered_end_pfn += processed_pfn; + + init_completion(&dm_device.ol_waitevent); + dm_device.ha_waiting = true; + + nid = memory_add_physaddr_to_nid(PFN_PHYS(start_pfn)); + ret = add_memory(nid, PFN_PHYS((start_pfn)), + (HA_CHUNK << PAGE_SHIFT)); + + if (ret) { + pr_info("hot_add memory failed error is %d\n", ret); + if (ret == -EEXIST) { + /* + * This error indicates that the error + * is not a transient failure. This is the + * case where the guest's physical address map + * precludes hot adding memory. Stop all further + * memory hot-add. + */ + do_hot_add = false; + } + has->ha_end_pfn -= HA_CHUNK; + has->covered_end_pfn -= processed_pfn; + break; + } + + /* + * Wait for the memory block to be onlined. + * Since the hot add has succeeded, it is ok to + * proceed even if the pages in the hot added region + * have not been "onlined" within the allowed time. + */ + wait_for_completion_timeout(&dm_device.ol_waitevent, 5*HZ); + post_status(&dm_device); + } + + return; +} + +static void hv_online_page(struct page *pg) +{ + struct list_head *cur; + struct hv_hotadd_state *has; + unsigned long cur_start_pgp; + unsigned long cur_end_pgp; + + if (dm_device.ha_waiting) { + dm_device.ha_waiting = false; + complete(&dm_device.ol_waitevent); + } + + list_for_each(cur, &dm_device.ha_region_list) { + has = list_entry(cur, struct hv_hotadd_state, list); + cur_start_pgp = (unsigned long) + pfn_to_page(has->covered_start_pfn); + cur_end_pgp = (unsigned long)pfn_to_page(has->covered_end_pfn); + + if (((unsigned long)pg >= cur_start_pgp) && + ((unsigned long)pg < cur_end_pgp)) { + /* + * This frame is currently backed; online the + * page. + */ + __online_page_set_limits(pg); + __online_page_increment_counters(pg); + __online_page_free(pg); + has->covered_start_pfn++; + } + } +} + +static bool pfn_covered(unsigned long start_pfn, unsigned long pfn_cnt) +{ + struct list_head *cur; + struct hv_hotadd_state *has; + unsigned long residual, new_inc; + + if (list_empty(&dm_device.ha_region_list)) + return false; + + list_for_each(cur, &dm_device.ha_region_list) { + has = list_entry(cur, struct hv_hotadd_state, list); + + /* + * If the pfn range we are dealing with is not in the current + * "hot add block", move on. + */ + if ((start_pfn >= has->end_pfn)) + continue; + /* + * If the current hot add-request extends beyond + * our current limit; extend it. + */ + if ((start_pfn + pfn_cnt) > has->end_pfn) { + residual = (start_pfn + pfn_cnt - has->end_pfn); + /* + * Extend the region by multiples of HA_CHUNK. + */ + new_inc = (residual / HA_CHUNK) * HA_CHUNK; + if (residual % HA_CHUNK) + new_inc += HA_CHUNK; + + has->end_pfn += new_inc; + } + + /* + * If the current start pfn is not where the covered_end + * is, update it. + */ + + if (has->covered_end_pfn != start_pfn) { + has->covered_end_pfn = start_pfn; + has->covered_start_pfn = start_pfn; + } + return true; + + } + + return false; +} + +static unsigned long handle_pg_range(unsigned long pg_start, + unsigned long pg_count) +{ + unsigned long start_pfn = pg_start; + unsigned long pfn_cnt = pg_count; + unsigned long size; + struct list_head *cur; + struct hv_hotadd_state *has; + unsigned long pgs_ol = 0; + unsigned long old_covered_state; + + if (list_empty(&dm_device.ha_region_list)) + return 0; + + list_for_each(cur, &dm_device.ha_region_list) { + has = list_entry(cur, struct hv_hotadd_state, list); + + /* + * If the pfn range we are dealing with is not in the current + * "hot add block", move on. + */ + if ((start_pfn >= has->end_pfn)) + continue; - if (do_hot_add) { + old_covered_state = has->covered_end_pfn; - pr_info("Memory hot add not supported\n"); + if (start_pfn < has->ha_end_pfn) { + /* + * This is the case where we are backing pages + * in an already hot added region. Bring + * these pages online first. + */ + pgs_ol = has->ha_end_pfn - start_pfn; + if (pgs_ol > pfn_cnt) + pgs_ol = pfn_cnt; + hv_bring_pgs_online(start_pfn, pgs_ol); + has->covered_end_pfn += pgs_ol; + has->covered_start_pfn += pgs_ol; + pfn_cnt -= pgs_ol; + } + if ((has->ha_end_pfn < has->end_pfn) && (pfn_cnt > 0)) { + /* + * We have some residual hot add range + * that needs to be hot added; hot add + * it now. Hot add a multiple of + * of HA_CHUNK that fully covers the pages + * we have. + */ + size = (has->end_pfn - has->ha_end_pfn); + if (pfn_cnt <= size) { + size = ((pfn_cnt / HA_CHUNK) * HA_CHUNK); + if (pfn_cnt % HA_CHUNK) + size += HA_CHUNK; + } else { + pfn_cnt = size; + } + hv_mem_hot_add(has->ha_end_pfn, size, pfn_cnt, has); + } /* - * Currently we do not support hot add. - * Just fail the request. + * If we managed to online any pages that were given to us, + * we declare success. */ + return has->covered_end_pfn - old_covered_state; + + } + + return 0; +} + +static unsigned long process_hot_add(unsigned long pg_start, + unsigned long pfn_cnt, + unsigned long rg_start, + unsigned long rg_size) +{ + struct hv_hotadd_state *ha_region = NULL; + + if (pfn_cnt == 0) + return 0; + + if (!dm_device.host_specified_ha_region) + if (pfn_covered(pg_start, pfn_cnt)) + goto do_pg_range; + + /* + * If the host has specified a hot-add range; deal with it first. + */ + + if (rg_size != 0) { + ha_region = kzalloc(sizeof(struct hv_hotadd_state), GFP_KERNEL); + if (!ha_region) + return 0; + + INIT_LIST_HEAD(&ha_region->list); + + list_add_tail(&ha_region->list, &dm_device.ha_region_list); + ha_region->start_pfn = rg_start; + ha_region->ha_end_pfn = rg_start; + ha_region->covered_start_pfn = pg_start; + ha_region->covered_end_pfn = pg_start; + ha_region->end_pfn = rg_start + rg_size; } +do_pg_range: + /* + * Process the page range specified; bringing them + * online if possible. + */ + return handle_pg_range(pg_start, pfn_cnt); +} + +#endif + +static void hot_add_req(struct work_struct *dummy) +{ + struct dm_hot_add_response resp; +#ifdef CONFIG_MEMORY_HOTPLUG + unsigned long pg_start, pfn_cnt; + unsigned long rg_start, rg_sz; +#endif + struct hv_dynmem_device *dm = &dm_device; + memset(&resp, 0, sizeof(struct dm_hot_add_response)); resp.hdr.type = DM_MEM_HOT_ADD_RESPONSE; resp.hdr.size = sizeof(struct dm_hot_add_response); - resp.hdr.trans_id = atomic_inc_return(&trans_id); - resp.page_count = 0; - resp.result = 0; +#ifdef CONFIG_MEMORY_HOTPLUG + pg_start = dm->ha_wrk.ha_page_range.finfo.start_page; + pfn_cnt = dm->ha_wrk.ha_page_range.finfo.page_cnt; + + rg_start = dm->ha_wrk.ha_region_range.finfo.start_page; + rg_sz = dm->ha_wrk.ha_region_range.finfo.page_cnt; + + if ((rg_start == 0) && (!dm->host_specified_ha_region)) { + unsigned long region_size; + unsigned long region_start; + + /* + * The host has not specified the hot-add region. + * Based on the hot-add page range being specified, + * compute a hot-add region that can cover the pages + * that need to be hot-added while ensuring the alignment + * and size requirements of Linux as it relates to hot-add. + */ + region_start = pg_start; + region_size = (pfn_cnt / HA_CHUNK) * HA_CHUNK; + if (pfn_cnt % HA_CHUNK) + region_size += HA_CHUNK; + + region_start = (pg_start / HA_CHUNK) * HA_CHUNK; + + rg_start = region_start; + rg_sz = region_size; + } + + if (do_hot_add) + resp.page_count = process_hot_add(pg_start, pfn_cnt, + rg_start, rg_sz); +#endif + /* + * The result field of the response structure has the + * following semantics: + * + * 1. If all or some pages hot-added: Guest should return success. + * + * 2. If no pages could be hot-added: + * + * If the guest returns success, then the host + * will not attempt any further hot-add operations. This + * signifies a permanent failure. + * + * If the guest returns failure, then this failure will be + * treated as a transient failure and the host may retry the + * hot-add operation after some delay. + */ + if (resp.page_count > 0) + resp.result = 1; + else if (!do_hot_add) + resp.result = 1; + else + resp.result = 0; + + if (!do_hot_add || (resp.page_count == 0)) + pr_info("Memory hot add failed\n"); dm->state = DM_INITIALIZED; + resp.hdr.trans_id = atomic_inc_return(&trans_id); vmbus_sendpacket(dm->dev->channel, &resp, sizeof(struct dm_hot_add_response), (unsigned long)NULL, VM_PKT_DATA_INBAND, 0); - } static void process_info(struct hv_dynmem_device *dm, struct dm_info_msg *msg) @@ -517,6 +916,34 @@ static void process_info(struct hv_dynmem_device *dm, struct dm_info_msg *msg) } } +static unsigned long compute_balloon_floor(void) +{ + unsigned long min_pages; +#define MB2PAGES(mb) ((mb) << (20 - PAGE_SHIFT)) + /* Simple continuous piecewiese linear function: + * max MiB -> min MiB gradient + * 0 0 + * 16 16 + * 32 24 + * 128 72 (1/2) + * 512 168 (1/4) + * 2048 360 (1/8) + * 8192 552 (1/32) + * 32768 1320 + * 131072 4392 + */ + if (totalram_pages < MB2PAGES(128)) + min_pages = MB2PAGES(8) + (totalram_pages >> 1); + else if (totalram_pages < MB2PAGES(512)) + min_pages = MB2PAGES(40) + (totalram_pages >> 2); + else if (totalram_pages < MB2PAGES(2048)) + min_pages = MB2PAGES(104) + (totalram_pages >> 3); + else + min_pages = MB2PAGES(296) + (totalram_pages >> 5); +#undef MB2PAGES + return min_pages; +} + /* * Post our status as it relates memory pressure to the * host. Host expects the guests to post this status @@ -530,16 +957,53 @@ static void process_info(struct hv_dynmem_device *dm, struct dm_info_msg *msg) 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; status.hdr.size = sizeof(struct dm_status); status.hdr.trans_id = atomic_inc_return(&trans_id); + /* + * The host expects the guest to report free memory. + * Further, the host expects the pressure information to + * include the ballooned out pages. + * For a given amount of memory that we are managing, we + * need to compute a floor below which we should not balloon. + * Compute this and add it to the pressure report. + */ + status.num_avail = val.freeram; + status.num_committed = vm_memory_committed() + + dm->num_pages_ballooned + + compute_balloon_floor(); - status.num_committed = vm_memory_committed(); + /* + * If our transaction ID is no longer current, just don't + * send the status. This can happen if we were interrupted + * after we picked our transaction ID. + */ + 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, @@ -547,8 +1011,6 @@ static void post_status(struct hv_dynmem_device *dm) } - - static void free_balloon_pages(struct hv_dynmem_device *dm, union dm_mem_page_range *range_array) { @@ -597,6 +1059,14 @@ static int alloc_balloon_pages(struct hv_dynmem_device *dm, int num_pages, dm->num_pages_ballooned += alloc_unit; + /* + * If we allocatted 2M pages; split them so we + * can free them in any order we get. + */ + + if (alloc_unit != 1) + split_page(pg, get_order(alloc_unit << PAGE_SHIFT)); + bl_resp->range_count++; bl_resp->range_array[i].finfo.start_page = page_to_pfn(pg); @@ -610,9 +1080,9 @@ static int alloc_balloon_pages(struct hv_dynmem_device *dm, int num_pages, -static void balloon_up(struct hv_dynmem_device *dm, struct dm_balloon *req) +static void balloon_up(struct work_struct *dummy) { - int num_pages = req->num_pages; + int num_pages = dm_device.balloon_wrk.num_pages; int num_ballooned = 0; struct dm_balloon_response *bl_resp; int alloc_unit; @@ -623,28 +1093,33 @@ static void balloon_up(struct hv_dynmem_device *dm, struct dm_balloon *req) /* - * Currently, we only support 4k allocations. + * We will attempt 2M allocations. However, if we fail to + * allocate 2M chunks, we will go back to 4k allocations. */ - alloc_unit = 1; + alloc_unit = 512; while (!done) { bl_resp = (struct dm_balloon_response *)send_buffer; memset(send_buffer, 0, PAGE_SIZE); bl_resp->hdr.type = DM_BALLOON_RESPONSE; - bl_resp->hdr.trans_id = atomic_inc_return(&trans_id); bl_resp->hdr.size = sizeof(struct dm_balloon_response); bl_resp->more_pages = 1; num_pages -= num_ballooned; - num_ballooned = alloc_balloon_pages(dm, num_pages, + num_ballooned = alloc_balloon_pages(&dm_device, num_pages, bl_resp, alloc_unit, &alloc_error); + if ((alloc_error) && (alloc_unit != 1)) { + alloc_unit = 1; + continue; + } + if ((alloc_error) || (num_ballooned == num_pages)) { bl_resp->more_pages = 0; done = true; - dm->state = DM_INITIALIZED; + dm_device.state = DM_INITIALIZED; } /* @@ -654,6 +1129,7 @@ static void balloon_up(struct hv_dynmem_device *dm, struct dm_balloon *req) */ do { + bl_resp->hdr.trans_id = atomic_inc_return(&trans_id); ret = vmbus_sendpacket(dm_device.dev->channel, bl_resp, bl_resp->hdr.size, @@ -662,7 +1138,7 @@ static void balloon_up(struct hv_dynmem_device *dm, struct dm_balloon *req) if (ret == -EAGAIN) msleep(20); - + post_status(&dm_device); } while (ret == -EAGAIN); if (ret) { @@ -672,7 +1148,7 @@ static void balloon_up(struct hv_dynmem_device *dm, struct dm_balloon *req) pr_info("Balloon response failed\n"); for (i = 0; i < bl_resp->range_count; i++) - free_balloon_pages(dm, + free_balloon_pages(&dm_device, &bl_resp->range_array[i]); done = true; @@ -689,8 +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; @@ -714,10 +1192,10 @@ static int dm_thread_func(void *dm_dev) { struct hv_dynmem_device *dm = dm_dev; int t; - unsigned long scan_start; 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. @@ -726,22 +1204,6 @@ static int dm_thread_func(void *dm_dev) if (t == 0) post_status(dm); - scan_start = jiffies; - switch (dm->state) { - case DM_BALLOON_UP: - balloon_up(dm, (struct dm_balloon *)recv_buffer); - break; - - case DM_HOT_ADD: - hot_add_req(dm, (struct dm_hot_add *)recv_buffer); - break; - default: - break; - } - - if (!time_in_range(jiffies, scan_start, scan_start + HZ)) - post_status(dm); - } return 0; @@ -814,6 +1276,10 @@ static void balloon_onchannelcallback(void *context) struct dm_message *dm_msg; struct dm_header *dm_hdr; struct hv_dynmem_device *dm = hv_get_drvdata(dev); + struct dm_balloon *bal_msg; + struct dm_hot_add *ha_msg; + union dm_mem_page_range *ha_pg_range; + union dm_mem_page_range *ha_region; memset(recv_buffer, 0, sizeof(recv_buffer)); vmbus_recvpacket(dev->channel, recv_buffer, @@ -835,8 +1301,12 @@ static void balloon_onchannelcallback(void *context) break; case DM_BALLOON_REQUEST: + if (dm->state == DM_BALLOON_UP) + pr_warn("Currently ballooning\n"); + bal_msg = (struct dm_balloon *)recv_buffer; dm->state = DM_BALLOON_UP; - complete(&dm->config_event); + dm_device.balloon_wrk.num_pages = bal_msg->num_pages; + schedule_work(&dm_device.balloon_wrk.wrk); break; case DM_UNBALLOON_REQUEST: @@ -846,8 +1316,31 @@ static void balloon_onchannelcallback(void *context) break; case DM_MEM_HOT_ADD_REQUEST: + if (dm->state == DM_HOT_ADD) + pr_warn("Currently hot-adding\n"); dm->state = DM_HOT_ADD; - complete(&dm->config_event); + ha_msg = (struct dm_hot_add *)recv_buffer; + if (ha_msg->hdr.size == sizeof(struct dm_hot_add)) { + /* + * This is a normal hot-add request specifying + * hot-add memory. + */ + ha_pg_range = &ha_msg->range; + dm->ha_wrk.ha_page_range = *ha_pg_range; + dm->ha_wrk.ha_region_range.page_range = 0; + } else { + /* + * Host is specifying that we first hot-add + * a region and then partially populate this + * region. + */ + dm->host_specified_ha_region = true; + ha_pg_range = &ha_msg->range; + ha_region = &ha_pg_range[1]; + dm->ha_wrk.ha_page_range = *ha_pg_range; + dm->ha_wrk.ha_region_range = *ha_region; + } + schedule_work(&dm_device.ha_wrk.wrk); break; case DM_INFO_MESSAGE: @@ -890,6 +1383,10 @@ static int balloon_probe(struct hv_device *dev, dm_device.next_version = DYNMEM_PROTOCOL_VERSION_WIN7; init_completion(&dm_device.host_event); init_completion(&dm_device.config_event); + INIT_LIST_HEAD(&dm_device.ha_region_list); + INIT_WORK(&dm_device.balloon_wrk.wrk, balloon_up); + INIT_WORK(&dm_device.ha_wrk.wrk, hot_add_req); + dm_device.host_specified_ha_region = false; dm_device.thread = kthread_run(dm_thread_func, &dm_device, "hv_balloon"); @@ -898,6 +1395,10 @@ static int balloon_probe(struct hv_device *dev, goto probe_error1; } +#ifdef CONFIG_MEMORY_HOTPLUG + set_online_page_callback(&hv_online_page); +#endif + hv_set_drvdata(dev, &dm_device); /* * Initiate the hand shake with the host and negotiate @@ -915,8 +1416,7 @@ static int balloon_probe(struct hv_device *dev, ret = vmbus_sendpacket(dev->channel, &version_req, sizeof(struct dm_version_request), (unsigned long)NULL, - VM_PKT_DATA_INBAND, - VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + VM_PKT_DATA_INBAND, 0); if (ret) goto probe_error2; @@ -943,13 +1443,13 @@ static int balloon_probe(struct hv_device *dev, cap_msg.hdr.trans_id = atomic_inc_return(&trans_id); cap_msg.caps.cap_bits.balloon = 1; + cap_msg.caps.cap_bits.hot_add = 1; + /* - * While we currently don't support hot-add, - * we still advertise this capability since the - * host requires that guests partcipating in the - * dynamic memory protocol support hot add. + * Specify our alignment requirements as it relates + * memory hot-add. Specify 128MB alignment. */ - cap_msg.caps.cap_bits.hot_add = 1; + cap_msg.caps.cap_bits.hot_add_alignment = 7; /* * Currently the host does not use these @@ -962,8 +1462,7 @@ static int balloon_probe(struct hv_device *dev, ret = vmbus_sendpacket(dev->channel, &cap_msg, sizeof(struct dm_capabilities), (unsigned long)NULL, - VM_PKT_DATA_INBAND, - VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + VM_PKT_DATA_INBAND, 0); if (ret) goto probe_error2; @@ -987,6 +1486,9 @@ static int balloon_probe(struct hv_device *dev, return 0; probe_error2: +#ifdef CONFIG_MEMORY_HOTPLUG + restore_online_page_callback(&hv_online_page); +#endif kthread_stop(dm_device.thread); probe_error1: @@ -999,13 +1501,26 @@ probe_error0: static int balloon_remove(struct hv_device *dev) { struct hv_dynmem_device *dm = hv_get_drvdata(dev); + struct list_head *cur, *tmp; + struct hv_hotadd_state *has; if (dm->num_pages_ballooned != 0) pr_warn("Ballooned pages: %d\n", dm->num_pages_ballooned); + cancel_work_sync(&dm->balloon_wrk.wrk); + cancel_work_sync(&dm->ha_wrk.wrk); + vmbus_close(dev->channel); kthread_stop(dm->thread); kfree(send_buffer); +#ifdef CONFIG_MEMORY_HOTPLUG + restore_online_page_callback(&hv_online_page); +#endif + list_for_each_safe(cur, tmp, &dm->ha_region_list) { + has = list_entry(cur, struct hv_hotadd_state, list); + list_del(&has->list); + kfree(has); + } return 0; } @@ -1013,9 +1528,7 @@ static int balloon_remove(struct hv_device *dev) static const struct hv_vmbus_device_id id_table[] = { /* Dynamic Memory Class ID */ /* 525074DC-8985-46e2-8057-A307DC18A502 */ - { VMBUS_DEVICE(0xdc, 0x74, 0x50, 0X52, 0x85, 0x89, 0xe2, 0x46, - 0x80, 0x57, 0xa3, 0x07, 0xdc, 0x18, 0xa5, 0x02) - }, + { HV_DM_GUID, }, { }, }; @@ -1034,15 +1547,7 @@ static int __init init_balloon_drv(void) return vmbus_driver_register(&balloon_drv); } -static void exit_balloon_drv(void) -{ - - vmbus_driver_unregister(&balloon_drv); -} - module_init(init_balloon_drv); -module_exit(exit_balloon_drv); MODULE_DESCRIPTION("Hyper-V Balloon"); -MODULE_VERSION(HV_DRV_VERSION); MODULE_LICENSE("GPL"); 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 ed50e9e83c6..521c14625b3 100644 --- a/drivers/hv/hv_kvp.c +++ b/drivers/hv/hv_kvp.c @@ -29,6 +29,20 @@ #include <linux/hyperv.h> +/* + * Pre win8 version numbers used in ws2008 and ws 2008 r2 (win7) + */ +#define WS2008_SRV_MAJOR 1 +#define WS2008_SRV_MINOR 0 +#define WS2008_SRV_VERSION (WS2008_SRV_MAJOR << 16 | WS2008_SRV_MINOR) + +#define WIN7_SRV_MAJOR 3 +#define WIN7_SRV_MINOR 0 +#define WIN7_SRV_VERSION (WIN7_SRV_MAJOR << 16 | WIN7_SRV_MINOR) + +#define WIN8_SRV_MAJOR 4 +#define WIN8_SRV_MINOR 0 +#define WIN8_SRV_VERSION (WIN8_SRV_MAJOR << 16 | WIN8_SRV_MINOR) /* * Global state maintained for transaction that is being processed. @@ -76,7 +90,9 @@ static u8 *recv_buffer; /* * Register the kernel component with the user-level daemon. * As part of this registration, pass the LIC version number. + * This number has no meaning, it satisfies the registration protocol. */ +#define HV_DRV_VERSION "3.1" static void kvp_register(int reg_value) @@ -97,7 +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); } } @@ -111,6 +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; @@ -139,7 +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; } @@ -419,7 +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; @@ -552,7 +579,7 @@ response_done: vmbus_sendpacket(channel, recv_buffer, buf_len, req_id, VM_PKT_DATA_INBAND, 0); - + poll_channel(channel); } /* @@ -575,6 +602,8 @@ void hv_kvp_onchannelcallback(void *context) struct icmsg_hdr *icmsghdrp; struct icmsg_negotiate *negop = NULL; + int util_fw_version; + int kvp_srv_version; if (kvp_transaction.active) { /* @@ -585,7 +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) { @@ -593,8 +622,28 @@ void hv_kvp_onchannelcallback(void *context) sizeof(struct vmbuspipe_hdr)]; if (icmsghdrp->icmsgtype == ICMSGTYPE_NEGOTIATE) { + /* + * Based on the host, select appropriate + * framework and service versions we will + * negotiate. + */ + switch (vmbus_proto_version) { + case (VERSION_WS2008): + util_fw_version = UTIL_WS2K8_FW_VERSION; + kvp_srv_version = WS2008_SRV_VERSION; + break; + case (VERSION_WIN7): + util_fw_version = UTIL_FW_VERSION; + kvp_srv_version = WIN7_SRV_VERSION; + break; + default: + util_fw_version = UTIL_FW_VERSION; + kvp_srv_version = WIN8_SRV_VERSION; + } vmbus_prep_negotiate_resp(icmsghdrp, negop, - recv_buffer, MAX_SRV_VER, MAX_SRV_VER); + recv_buffer, util_fw_version, + kvp_srv_version); + } else { kvp_msg = (struct hv_kvp_msg *)&recv_buffer[ sizeof(struct vmbuspipe_hdr) + diff --git a/drivers/hv/hv_snapshot.c b/drivers/hv/hv_snapshot.c new file mode 100644 index 00000000000..34f14fddb66 --- /dev/null +++ b/drivers/hv/hv_snapshot.c @@ -0,0 +1,281 @@ +/* + * An implementation of host initiated guest snapshot. + * + * + * Copyright (C) 2013, Microsoft, Inc. + * Author : K. Y. Srinivasan <kys@microsoft.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/net.h> +#include <linux/nls.h> +#include <linux/connector.h> +#include <linux/workqueue.h> +#include <linux/hyperv.h> + +#define VSS_MAJOR 5 +#define VSS_MINOR 0 +#define VSS_VERSION (VSS_MAJOR << 16 | VSS_MINOR) + + + +/* + * Global state maintained for transaction that is being processed. + * Note that only one transaction can be active at any point in time. + * + * This state is set when we receive a request from the host; we + * cleanup this state when the transaction is completed - when we respond + * to the host with the key value. + */ + +static struct { + bool active; /* transaction status - active or not */ + int recv_len; /* number of bytes received. */ + struct vmbus_channel *recv_channel; /* chn we got the request */ + u64 recv_req_id; /* request ID. */ + struct hv_vss_msg *msg; /* current message */ +} vss_transaction; + + +static void vss_respond_to_host(int error); + +static struct cb_id vss_id = { CN_VSS_IDX, CN_VSS_VAL }; +static const char vss_name[] = "vss_kernel_module"; +static __u8 *recv_buffer; + +static void vss_send_op(struct work_struct *dummy); +static DECLARE_WORK(vss_send_op_work, vss_send_op); + +/* + * Callback when data is received from user mode. + */ + +static void +vss_cn_callback(struct cn_msg *msg, struct netlink_skb_parms *nsp) +{ + struct hv_vss_msg *vss_msg; + + vss_msg = (struct hv_vss_msg *)msg->data; + + if (vss_msg->vss_hdr.operation == VSS_OP_REGISTER) { + pr_info("VSS daemon registered\n"); + vss_transaction.active = false; + if (vss_transaction.recv_channel != NULL) + hv_vss_onchannelcallback(vss_transaction.recv_channel); + return; + + } + vss_respond_to_host(vss_msg->error); +} + + +static void vss_send_op(struct work_struct *dummy) +{ + int op = vss_transaction.msg->vss_hdr.operation; + struct cn_msg *msg; + struct hv_vss_msg *vss_msg; + + msg = kzalloc(sizeof(*msg) + sizeof(*vss_msg), GFP_ATOMIC); + if (!msg) + return; + + vss_msg = (struct hv_vss_msg *)msg->data; + + msg->id.idx = CN_VSS_IDX; + msg->id.val = CN_VSS_VAL; + + vss_msg->vss_hdr.operation = op; + msg->len = sizeof(struct hv_vss_msg); + + cn_netlink_send(msg, 0, 0, GFP_ATOMIC); + kfree(msg); + + return; +} + +/* + * Send a response back to the host. + */ + +static void +vss_respond_to_host(int error) +{ + struct icmsg_hdr *icmsghdrp; + u32 buf_len; + struct vmbus_channel *channel; + u64 req_id; + + /* + * If a transaction is not active; log and return. + */ + + if (!vss_transaction.active) { + /* + * This is a spurious call! + */ + pr_warn("VSS: Transaction not active\n"); + return; + } + /* + * Copy the global state for completing the transaction. Note that + * only one transaction can be active at a time. + */ + + buf_len = vss_transaction.recv_len; + channel = vss_transaction.recv_channel; + req_id = vss_transaction.recv_req_id; + vss_transaction.active = false; + + icmsghdrp = (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; + + icmsghdrp->status = error; + + icmsghdrp->icflags = ICMSGHDRFLAG_TRANSACTION | ICMSGHDRFLAG_RESPONSE; + + vmbus_sendpacket(channel, recv_buffer, buf_len, req_id, + VM_PKT_DATA_INBAND, 0); + +} + +/* + * This callback is invoked when we get a VSS message from the host. + * The host ensures that only one VSS transaction can be active at a time. + */ + +void hv_vss_onchannelcallback(void *context) +{ + struct vmbus_channel *channel = context; + u32 recvlen; + u64 requestid; + struct hv_vss_msg *vss_msg; + + + struct icmsg_hdr *icmsghdrp; + struct icmsg_negotiate *negop = NULL; + + if (vss_transaction.active) { + /* + * We will defer processing this callback once + * the current transaction is complete. + */ + vss_transaction.recv_channel = channel; + return; + } + + vmbus_recvpacket(channel, recv_buffer, PAGE_SIZE * 2, &recvlen, + &requestid); + + if (recvlen > 0) { + icmsghdrp = (struct icmsg_hdr *)&recv_buffer[ + sizeof(struct vmbuspipe_hdr)]; + + if (icmsghdrp->icmsgtype == ICMSGTYPE_NEGOTIATE) { + vmbus_prep_negotiate_resp(icmsghdrp, negop, + recv_buffer, UTIL_FW_VERSION, + VSS_VERSION); + } else { + vss_msg = (struct hv_vss_msg *)&recv_buffer[ + sizeof(struct vmbuspipe_hdr) + + sizeof(struct icmsg_hdr)]; + + /* + * Stash away this global state for completing the + * transaction; note transactions are serialized. + */ + + vss_transaction.recv_len = recvlen; + vss_transaction.recv_channel = channel; + vss_transaction.recv_req_id = requestid; + vss_transaction.active = true; + vss_transaction.msg = (struct hv_vss_msg *)vss_msg; + + switch (vss_msg->vss_hdr.operation) { + /* + * Initiate a "freeze/thaw" + * operation in the guest. + * We respond to the host once + * the operation is complete. + * + * We send the message to the + * user space daemon and the + * operation is performed in + * the daemon. + */ + case VSS_OP_FREEZE: + case VSS_OP_THAW: + schedule_work(&vss_send_op_work); + return; + + case VSS_OP_HOT_BACKUP: + vss_msg->vss_cf.flags = + VSS_HBU_NO_AUTO_RECOVERY; + vss_respond_to_host(0); + return; + + case VSS_OP_GET_DM_INFO: + vss_msg->dm_info.flags = 0; + vss_respond_to_host(0); + return; + + default: + vss_respond_to_host(0); + return; + + } + + } + + icmsghdrp->icflags = ICMSGHDRFLAG_TRANSACTION + | ICMSGHDRFLAG_RESPONSE; + + vmbus_sendpacket(channel, recv_buffer, + recvlen, requestid, + VM_PKT_DATA_INBAND, 0); + } + +} + +int +hv_vss_init(struct hv_util_service *srv) +{ + int err; + + err = cn_add_callback(&vss_id, vss_name, vss_cn_callback); + if (err) + return err; + 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. + */ + vss_transaction.active = true; + return 0; +} + +void hv_vss_deinit(void) +{ + cn_del_callback(&vss_id); + cancel_work_sync(&vss_send_op_work); +} diff --git a/drivers/hv/hv_util.c b/drivers/hv/hv_util.c index a0667de7a04..3b9c9ef0deb 100644 --- a/drivers/hv/hv_util.c +++ b/drivers/hv/hv_util.c @@ -28,6 +28,34 @@ #include <linux/reboot.h> #include <linux/hyperv.h> +#include "hyperv_vmbus.h" + +#define SD_MAJOR 3 +#define SD_MINOR 0 +#define SD_VERSION (SD_MAJOR << 16 | SD_MINOR) + +#define SD_WS2008_MAJOR 1 +#define SD_WS2008_VERSION (SD_WS2008_MAJOR << 16 | SD_MINOR) + +#define TS_MAJOR 3 +#define TS_MINOR 0 +#define TS_VERSION (TS_MAJOR << 16 | TS_MINOR) + +#define TS_WS2008_MAJOR 1 +#define TS_WS2008_VERSION (TS_WS2008_MAJOR << 16 | TS_MINOR) + +#define HB_MAJOR 3 +#define HB_MINOR 0 +#define HB_VERSION (HB_MAJOR << 16 | HB_MINOR) + +#define HB_WS2008_MAJOR 1 +#define HB_WS2008_VERSION (HB_WS2008_MAJOR << 16 | HB_MINOR) + +static int sd_srv_version; +static int ts_srv_version; +static int hb_srv_version; +static int util_fw_version; + static void shutdown_onchannelcallback(void *context); static struct hv_util_service util_shutdown = { .util_cb = shutdown_onchannelcallback, @@ -49,12 +77,34 @@ static struct hv_util_service util_kvp = { .util_deinit = hv_kvp_deinit, }; +static struct hv_util_service util_vss = { + .util_cb = hv_vss_onchannelcallback, + .util_init = hv_vss_init, + .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); +} + +/* + * Perform the shutdown operation in a thread context. + */ +static DECLARE_WORK(shutdown_work, perform_shutdown); + static void shutdown_onchannelcallback(void *context) { struct vmbus_channel *channel = context; u32 recvlen; u64 requestid; - u8 execute_shutdown = false; + bool execute_shutdown = false; u8 *shut_txf_buf = util_shutdown.recv_buffer; struct shutdown_msg_data *shutdown_msg; @@ -71,7 +121,8 @@ static void shutdown_onchannelcallback(void *context) if (icmsghdrp->icmsgtype == ICMSGTYPE_NEGOTIATE) { vmbus_prep_negotiate_resp(icmsghdrp, negop, - shut_txf_buf, MAX_SRV_VER, MAX_SRV_VER); + shut_txf_buf, util_fw_version, + sd_srv_version); } else { shutdown_msg = (struct shutdown_msg_data *)&shut_txf_buf[ @@ -106,7 +157,7 @@ static void shutdown_onchannelcallback(void *context) } if (execute_shutdown == true) - orderly_poweroff(true); + schedule_work(&shutdown_work); } /* @@ -187,6 +238,7 @@ static void timesync_onchannelcallback(void *context) struct icmsg_hdr *icmsghdrp; struct ictimesync_data *timedatap; u8 *time_txf_buf = util_timesynch.recv_buffer; + struct icmsg_negotiate *negop = NULL; vmbus_recvpacket(channel, time_txf_buf, PAGE_SIZE, &recvlen, &requestid); @@ -196,8 +248,10 @@ static void timesync_onchannelcallback(void *context) sizeof(struct vmbuspipe_hdr)]; if (icmsghdrp->icmsgtype == ICMSGTYPE_NEGOTIATE) { - vmbus_prep_negotiate_resp(icmsghdrp, NULL, time_txf_buf, - MAX_SRV_VER, MAX_SRV_VER); + vmbus_prep_negotiate_resp(icmsghdrp, negop, + time_txf_buf, + util_fw_version, + ts_srv_version); } else { timedatap = (struct ictimesync_data *)&time_txf_buf[ sizeof(struct vmbuspipe_hdr) + @@ -227,6 +281,7 @@ static void heartbeat_onchannelcallback(void *context) struct icmsg_hdr *icmsghdrp; struct heartbeat_msg_data *heartbeat_msg; u8 *hbeat_txf_buf = util_heartbeat.recv_buffer; + struct icmsg_negotiate *negop = NULL; vmbus_recvpacket(channel, hbeat_txf_buf, PAGE_SIZE, &recvlen, &requestid); @@ -236,8 +291,9 @@ static void heartbeat_onchannelcallback(void *context) sizeof(struct vmbuspipe_hdr)]; if (icmsghdrp->icmsgtype == ICMSGTYPE_NEGOTIATE) { - vmbus_prep_negotiate_resp(icmsghdrp, NULL, - hbeat_txf_buf, MAX_SRV_VER, MAX_SRV_VER); + vmbus_prep_negotiate_resp(icmsghdrp, negop, + hbeat_txf_buf, util_fw_version, + hb_srv_version); } else { heartbeat_msg = (struct heartbeat_msg_data *)&hbeat_txf_buf[ @@ -263,7 +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) { @@ -274,12 +330,41 @@ static int util_probe(struct hv_device *dev, } } + /* + * The set of services managed by the util driver are not performance + * critical and do not need batched reading. Furthermore, some services + * such as KVP can only handle one message from the host at a time. + * Turn off batched reading for all util drivers before we open the + * channel. + */ + + set_channel_read_state(dev->channel, false); + ret = vmbus_open(dev->channel, 4 * PAGE_SIZE, 4 * PAGE_SIZE, NULL, 0, srv->util_cb, dev->channel); if (ret) goto error; hv_set_drvdata(dev, srv); + /* + * Based on the host; initialize the framework and + * service version numbers we will negotiate. + */ + switch (vmbus_proto_version) { + case (VERSION_WS2008): + util_fw_version = UTIL_WS2K8_FW_VERSION; + sd_srv_version = SD_WS2008_VERSION; + ts_srv_version = TS_WS2008_VERSION; + hb_srv_version = HB_WS2008_VERSION; + break; + + default: + util_fw_version = UTIL_FW_VERSION; + sd_srv_version = SD_VERSION; + ts_srv_version = TS_VERSION; + hb_srv_version = HB_VERSION; + } + return 0; error: @@ -304,21 +389,29 @@ static int util_remove(struct hv_device *dev) static const struct hv_vmbus_device_id id_table[] = { /* Shutdown guid */ - { VMBUS_DEVICE(0x31, 0x60, 0x0B, 0X0E, 0x13, 0x52, 0x34, 0x49, - 0x81, 0x8B, 0x38, 0XD9, 0x0C, 0xED, 0x39, 0xDB) - .driver_data = (unsigned long)&util_shutdown }, + { HV_SHUTDOWN_GUID, + .driver_data = (unsigned long)&util_shutdown + }, /* Time synch guid */ - { VMBUS_DEVICE(0x30, 0xe6, 0x27, 0x95, 0xae, 0xd0, 0x7b, 0x49, - 0xad, 0xce, 0xe8, 0x0a, 0xb0, 0x17, 0x5c, 0xaf) - .driver_data = (unsigned long)&util_timesynch }, + { HV_TS_GUID, + .driver_data = (unsigned long)&util_timesynch + }, /* Heartbeat guid */ - { VMBUS_DEVICE(0x39, 0x4f, 0x16, 0x57, 0x15, 0x91, 0x78, 0x4e, - 0xab, 0x55, 0x38, 0x2f, 0x3b, 0xd5, 0x42, 0x2d) - .driver_data = (unsigned long)&util_heartbeat }, + { HV_HEART_BEAT_GUID, + .driver_data = (unsigned long)&util_heartbeat + }, /* KVP guid */ - { VMBUS_DEVICE(0xe7, 0xf4, 0xa0, 0xa9, 0x45, 0x5a, 0x96, 0x4d, - 0xb8, 0x27, 0x8a, 0x84, 0x1e, 0x8c, 0x3, 0xe6) - .driver_data = (unsigned long)&util_kvp }, + { HV_KVP_GUID, + .driver_data = (unsigned long)&util_kvp + }, + /* VSS GUID */ + { HV_VSS_GUID, + .driver_data = (unsigned long)&util_vss + }, + /* File copy GUID */ + { HV_FCOPY_GUID, + .driver_data = (unsigned long)&util_fcopy + }, { }, }; @@ -350,5 +443,4 @@ module_init(init_hyperv_utils); module_exit(exit_hyperv_utils); MODULE_DESCRIPTION("Hyper-V Utilities"); -MODULE_VERSION(HV_DRV_VERSION); MODULE_LICENSE("GPL"); diff --git a/drivers/hv/hyperv_vmbus.h b/drivers/hv/hyperv_vmbus.h index d8d1fadb398..22b750749a3 100644 --- a/drivers/hv/hyperv_vmbus.h +++ b/drivers/hv/hyperv_vmbus.h @@ -101,15 +101,6 @@ enum hv_message_type { /* Define invalid partition identifier. */ #define HV_PARTITION_ID_INVALID ((u64)0x0) -/* Define connection identifier type. */ -union hv_connection_id { - u32 asu32; - struct { - u32 id:24; - u32 reserved:8; - } u; -}; - /* Define port identifier type. */ union hv_port_id { u32 asu32; @@ -338,13 +329,6 @@ struct hv_input_post_message { u64 payload[HV_MESSAGE_PAYLOAD_QWORD_COUNT]; }; -/* Definition of the hv_signal_event hypercall input structure. */ -struct hv_input_signal_event { - union hv_connection_id connectionid; - u16 flag_number; - u16 rsvdz; -}; - /* * Versioning definitions used for guests reporting themselves to the * hypervisor, and visa versa. @@ -498,11 +482,6 @@ static const uuid_le VMBUS_SERVICE_ID = { -struct hv_input_signal_event_buffer { - u64 align8; - struct hv_input_signal_event event; -}; - struct hv_context { /* We only support running on top of Hyper-V * So at this point this really can only contain the Hyper-V ID @@ -513,20 +492,40 @@ struct hv_context { bool synic_initialized; - /* - * This is used as an input param to HvCallSignalEvent hypercall. The - * input param is immutable in our usage and must be dynamic mem (vs - * stack or global). */ - struct hv_input_signal_event_buffer *signal_event_buffer; - /* 8-bytes aligned of the buffer above */ - struct hv_input_signal_event *signal_event_param; - void *synic_message_page[NR_CPUS]; void *synic_event_page[NR_CPUS]; + /* + * Hypervisor's notion of virtual processor ID is different from + * Linux' notion of CPU ID. This information can only be retrieved + * in the context of the calling CPU. Setup a map for easy access + * to this information: + * + * vp_index[a] is the Hyper-V's processor ID corresponding to + * Linux cpuid 'a'. + */ + u32 vp_index[NR_CPUS]; + /* + * Starting with win8, we can take channel interrupts on any CPU; + * we will manage the tasklet that handles events on a per CPU + * 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; +struct hv_ring_buffer_debug_info { + u32 current_interrupt_mask; + u32 current_read_index; + u32 current_write_index; + u32 bytes_avail_toread; + u32 bytes_avail_towrite; +}; /* Hv Interface */ @@ -538,12 +537,23 @@ extern int hv_post_message(union hv_connection_id connection_id, enum hv_message_type message_type, void *payload, size_t payload_size); -extern u16 hv_signal_event(void); +extern u16 hv_signal_event(void *con_id); + +extern int hv_synic_alloc(void); + +extern void hv_synic_free(void); extern void hv_synic_init(void *irqarg); extern void hv_synic_cleanup(void *arg); +/* + * Host version information. + */ +extern unsigned int host_info_eax; +extern unsigned int host_info_ebx; +extern unsigned int host_info_ecx; +extern unsigned int host_info_edx; /* Interface */ @@ -554,8 +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); + struct kvec *kv_list, + u32 kv_count, bool *signal); int hv_ringbuffer_peek(struct hv_ring_buffer_info *ring_info, void *buffer, u32 buflen); @@ -563,13 +573,16 @@ int hv_ringbuffer_peek(struct hv_ring_buffer_info *ring_info, void *buffer, int hv_ringbuffer_read(struct hv_ring_buffer_info *ring_info, void *buffer, u32 buflen, - u32 offset); + u32 offset, bool *signal); -u32 hv_get_ringbuffer_interrupt_mask(struct hv_ring_buffer_info *ring_info); void hv_ringbuffer_get_debuginfo(struct hv_ring_buffer_info *ring_info, struct hv_ring_buffer_debug_info *debug_info); +void hv_begin_read(struct hv_ring_buffer_info *rbi); + +u32 hv_end_read(struct hv_ring_buffer_info *rbi); + /* * Maximum channels is determined by the size of the interrupt page * which is PAGE_SIZE. 1/2 of PAGE_SIZE is for send endpoint interrupt @@ -611,7 +624,7 @@ struct vmbus_connection { * 2 pages - 1st page for parent->child notification and 2nd * is child->parent notification */ - void *monitor_pages; + struct hv_monitor_page *monitor_pages[2]; struct list_head chn_msg_list; spinlock_t channelmsg_lock; @@ -636,9 +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); @@ -657,9 +670,13 @@ int vmbus_connect(void); int vmbus_post_msg(void *buffer, size_t buflen); -int vmbus_set_event(u32 child_relid); +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 7233c88f01b..15db66b7414 100644 --- a/drivers/hv/ring_buffer.c +++ b/drivers/hv/ring_buffer.c @@ -26,9 +26,112 @@ #include <linux/kernel.h> #include <linux/mm.h> #include <linux/hyperv.h> +#include <linux/uio.h> #include "hyperv_vmbus.h" +void hv_begin_read(struct hv_ring_buffer_info *rbi) +{ + rbi->ring_buffer->interrupt_mask = 1; + mb(); +} + +u32 hv_end_read(struct hv_ring_buffer_info *rbi) +{ + u32 read; + u32 write; + + rbi->ring_buffer->interrupt_mask = 0; + mb(); + + /* + * Now check to see if the ring buffer is still empty. + * If it is not, we raced and we need to process new + * incoming messages. + */ + hv_get_ringbuffer_availbytes(rbi, &read, &write); + + return read; +} + +/* + * When we write to the ring buffer, check if the host needs to + * be signaled. Here is the details of this protocol: + * + * 1. The host guarantees that while it is draining the + * ring buffer, it will set the interrupt_mask to + * indicate it does not need to be interrupted when + * new data is placed. + * + * 2. The host guarantees that it will completely drain + * the ring buffer before exiting the read loop. Further, + * once the ring buffer is empty, it will clear the + * interrupt_mask and re-check to see if new data has + * arrived. + */ + +static bool hv_need_to_signal(u32 old_write, struct hv_ring_buffer_info *rbi) +{ + mb(); + if (rbi->ring_buffer->interrupt_mask) + return false; + + /* check interrupt_mask before read_index */ + rmb(); + /* + * This is the only case we need to signal when the + * ring transitions from being empty to non-empty. + */ + if (old_write == rbi->ring_buffer->read_index) + return true; + + return false; +} + +/* + * To optimize the flow management on the send-side, + * when the sender is blocked because of lack of + * sufficient space in the ring buffer, potential the + * consumer of the ring buffer can signal the producer. + * This is controlled by the following parameters: + * + * 1. pending_send_sz: This is the size in bytes that the + * producer is trying to send. + * 2. The feature bit feat_pending_send_sz set to indicate if + * the consumer of the ring will signal when the ring + * state transitions from being full to a state where + * there is room for the producer to send the pending packet. + */ + +static bool hv_need_to_signal_on_read(u32 old_rd, + struct hv_ring_buffer_info *rbi) +{ + u32 prev_write_sz; + u32 cur_write_sz; + u32 r_size; + u32 write_loc = rbi->ring_buffer->write_index; + u32 read_loc = rbi->ring_buffer->read_index; + u32 pending_sz = rbi->ring_buffer->pending_send_sz; + + /* + * If the other end is not blocked on write don't bother. + */ + if (pending_sz == 0) + return false; + + r_size = rbi->ring_datasize; + cur_write_sz = write_loc >= read_loc ? r_size - (write_loc - read_loc) : + read_loc - write_loc; + + prev_write_sz = write_loc >= old_rd ? r_size - (write_loc - old_rd) : + old_rd - write_loc; + + + if ((prev_write_sz < pending_sz) && (cur_write_sz >= pending_sz)) + return true; + + return false; +} /* * hv_get_next_write_location() @@ -239,19 +342,6 @@ void hv_ringbuffer_get_debuginfo(struct hv_ring_buffer_info *ring_info, } } - -/* - * - * hv_get_ringbuffer_interrupt_mask() - * - * Get the interrupt mask for the specified ring buffer - * - */ -u32 hv_get_ringbuffer_interrupt_mask(struct hv_ring_buffer_info *rbi) -{ - return rbi->ring_buffer->interrupt_mask; -} - /* * * hv_ringbuffer_init() @@ -298,22 +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) + 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); @@ -335,12 +423,13 @@ int hv_ringbuffer_write(struct hv_ring_buffer_info *outring_info, /* Write to the ring buffer */ next_write_location = hv_get_next_write_location(outring_info); - for_each_sg(sglist, sg, sgcount, i) - { + old_write = next_write_location; + + 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 */ @@ -351,14 +440,16 @@ int hv_ringbuffer_write(struct hv_ring_buffer_info *outring_info, &prev_indices, sizeof(u64)); - /* Make sure we flush all writes before updating the writeIndex */ - smp_wmb(); + /* Issue a full memory barrier before updating the write index */ + mb(); /* Now, update the write location */ hv_set_next_write_location(outring_info, next_write_location); spin_unlock_irqrestore(&outring_info->ring_lock, flags); + + *signal = hv_need_to_signal(old_write, outring_info); return 0; } @@ -414,13 +505,14 @@ int hv_ringbuffer_peek(struct hv_ring_buffer_info *Inring_info, * */ int hv_ringbuffer_read(struct hv_ring_buffer_info *inring_info, void *buffer, - u32 buflen, u32 offset) + u32 buflen, u32 offset, bool *signal) { u32 bytes_avail_towrite; u32 bytes_avail_toread; u32 next_read_location = 0; u64 prev_indices = 0; unsigned long flags; + u32 old_read; if (buflen <= 0) return -EINVAL; @@ -431,6 +523,8 @@ int hv_ringbuffer_read(struct hv_ring_buffer_info *inring_info, void *buffer, &bytes_avail_toread, &bytes_avail_towrite); + old_read = bytes_avail_toread; + /* Make sure there is something to read */ if (bytes_avail_toread < buflen) { spin_unlock_irqrestore(&inring_info->ring_lock, flags); @@ -454,12 +548,14 @@ int hv_ringbuffer_read(struct hv_ring_buffer_info *inring_info, void *buffer, /* Make sure all reads are done before we update the read index since */ /* the writer may start writing to the read area once the read index */ /*is updated */ - smp_mb(); + mb(); /* Update the read index */ hv_set_next_read_location(inring_info, next_read_location); spin_unlock_irqrestore(&inring_info->ring_lock, flags); + *signal = hv_need_to_signal_on_read(old_read, inring_info); + return 0; } diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c index 8e1a9ec5300..4d6b26979fb 100644 --- a/drivers/hv/vmbus_drv.c +++ b/drivers/hv/vmbus_drv.c @@ -25,43 +25,29 @@ #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> #include <linux/acpi.h> -#include <acpi/acpi_bus.h> #include <linux/completion.h> #include <linux/hyperv.h> +#include <linux/kernel_stat.h> #include <asm/hyperv.h> #include <asm/hypervisor.h> +#include <asm/mshyperv.h> #include "hyperv_vmbus.h" - static struct acpi_device *hv_acpi_dev; static struct tasklet_struct msg_dpc; -static struct tasklet_struct event_dpc; static struct completion probe_event; static int irq; -struct hv_device_info { - u32 chn_id; - u32 chn_state; - uuid_le chn_type; - uuid_le chn_instance; - - u32 monitor_id; - u32 server_monitor_pending; - u32 server_monitor_latency; - u32 server_monitor_conn_id; - u32 client_monitor_pending; - u32 client_monitor_latency; - u32 client_monitor_conn_id; - - struct hv_dev_port_info inbound; - struct hv_dev_port_info outbound; +struct resource hyperv_mmio = { + .name = "hyperv mmio", + .flags = IORESOURCE_MEM, }; +EXPORT_SYMBOL_GPL(hyperv_mmio); static int vmbus_exists(void) { @@ -71,169 +57,361 @@ static int vmbus_exists(void) return 0; } +#define VMBUS_ALIAS_LEN ((sizeof((struct hv_vmbus_device_id *)0)->guid) * 2) +static void print_alias_name(struct hv_device *hv_dev, char *alias_name) +{ + int i; + for (i = 0; i < VMBUS_ALIAS_LEN; i += 2) + sprintf(&alias_name[i], "%02x", hv_dev->dev_type.b[i/2]); +} -static void get_channel_info(struct hv_device *device, - struct hv_device_info *info) +static u8 channel_monitor_group(struct vmbus_channel *channel) { - struct vmbus_channel_debug_info debug_info; + return (u8)channel->offermsg.monitorid / 32; +} - if (!device->channel) - return; +static u8 channel_monitor_offset(struct vmbus_channel *channel) +{ + return (u8)channel->offermsg.monitorid % 32; +} - vmbus_get_debug_info(device->channel, &debug_info); +static u32 channel_pending(struct vmbus_channel *channel, + struct hv_monitor_page *monitor_page) +{ + u8 monitor_group = channel_monitor_group(channel); + return monitor_page->trigger_group[monitor_group].pending; +} - info->chn_id = debug_info.relid; - info->chn_state = debug_info.state; - memcpy(&info->chn_type, &debug_info.interfacetype, - sizeof(uuid_le)); - memcpy(&info->chn_instance, &debug_info.interface_instance, - sizeof(uuid_le)); +static u32 channel_latency(struct vmbus_channel *channel, + struct hv_monitor_page *monitor_page) +{ + u8 monitor_group = channel_monitor_group(channel); + u8 monitor_offset = channel_monitor_offset(channel); + return monitor_page->latency[monitor_group][monitor_offset]; +} - info->monitor_id = debug_info.monitorid; +static u32 channel_conn_id(struct vmbus_channel *channel, + struct hv_monitor_page *monitor_page) +{ + u8 monitor_group = channel_monitor_group(channel); + u8 monitor_offset = channel_monitor_offset(channel); + return monitor_page->parameter[monitor_group][monitor_offset].connectionid.u.id; +} - info->server_monitor_pending = debug_info.servermonitor_pending; - info->server_monitor_latency = debug_info.servermonitor_latency; - info->server_monitor_conn_id = debug_info.servermonitor_connectionid; +static ssize_t id_show(struct device *dev, struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); - info->client_monitor_pending = debug_info.clientmonitor_pending; - info->client_monitor_latency = debug_info.clientmonitor_latency; - info->client_monitor_conn_id = debug_info.clientmonitor_connectionid; + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "%d\n", hv_dev->channel->offermsg.child_relid); +} +static DEVICE_ATTR_RO(id); - info->inbound.int_mask = debug_info.inbound.current_interrupt_mask; - info->inbound.read_idx = debug_info.inbound.current_read_index; - info->inbound.write_idx = debug_info.inbound.current_write_index; - info->inbound.bytes_avail_toread = - debug_info.inbound.bytes_avail_toread; - info->inbound.bytes_avail_towrite = - debug_info.inbound.bytes_avail_towrite; +static ssize_t state_show(struct device *dev, struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); - info->outbound.int_mask = - debug_info.outbound.current_interrupt_mask; - info->outbound.read_idx = debug_info.outbound.current_read_index; - info->outbound.write_idx = debug_info.outbound.current_write_index; - info->outbound.bytes_avail_toread = - debug_info.outbound.bytes_avail_toread; - info->outbound.bytes_avail_towrite = - debug_info.outbound.bytes_avail_towrite; + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "%d\n", hv_dev->channel->state); } +static DEVICE_ATTR_RO(state); -#define VMBUS_ALIAS_LEN ((sizeof((struct hv_vmbus_device_id *)0)->guid) * 2) -static void print_alias_name(struct hv_device *hv_dev, char *alias_name) +static ssize_t monitor_id_show(struct device *dev, + struct device_attribute *dev_attr, char *buf) { - int i; - for (i = 0; i < VMBUS_ALIAS_LEN; i += 2) - sprintf(&alias_name[i], "%02x", hv_dev->dev_type.b[i/2]); + struct hv_device *hv_dev = device_to_hv_device(dev); + + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "%d\n", hv_dev->channel->offermsg.monitorid); } +static DEVICE_ATTR_RO(monitor_id); -/* - * vmbus_show_device_attr - Show the device attribute in sysfs. - * - * This is invoked when user does a - * "cat /sys/bus/vmbus/devices/<busdevice>/<attr name>" - */ -static ssize_t vmbus_show_device_attr(struct device *dev, - struct device_attribute *dev_attr, - char *buf) +static ssize_t class_id_show(struct device *dev, + struct device_attribute *dev_attr, char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "{%pUl}\n", + hv_dev->channel->offermsg.offer.if_type.b); +} +static DEVICE_ATTR_RO(class_id); + +static ssize_t device_id_show(struct device *dev, + struct device_attribute *dev_attr, char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "{%pUl}\n", + hv_dev->channel->offermsg.offer.if_instance.b); +} +static DEVICE_ATTR_RO(device_id); + +static ssize_t modalias_show(struct device *dev, + struct device_attribute *dev_attr, char *buf) { struct hv_device *hv_dev = device_to_hv_device(dev); - struct hv_device_info *device_info; char alias_name[VMBUS_ALIAS_LEN + 1]; - int ret = 0; - device_info = kzalloc(sizeof(struct hv_device_info), GFP_KERNEL); - if (!device_info) - return ret; + print_alias_name(hv_dev, alias_name); + return sprintf(buf, "vmbus:%s\n", alias_name); +} +static DEVICE_ATTR_RO(modalias); - get_channel_info(hv_dev, device_info); - - if (!strcmp(dev_attr->attr.name, "class_id")) { - ret = sprintf(buf, "{%pUl}\n", device_info->chn_type.b); - } else if (!strcmp(dev_attr->attr.name, "device_id")) { - ret = sprintf(buf, "{%pUl}\n", device_info->chn_instance.b); - } else if (!strcmp(dev_attr->attr.name, "modalias")) { - print_alias_name(hv_dev, alias_name); - ret = sprintf(buf, "vmbus:%s\n", alias_name); - } else if (!strcmp(dev_attr->attr.name, "state")) { - ret = sprintf(buf, "%d\n", device_info->chn_state); - } else if (!strcmp(dev_attr->attr.name, "id")) { - ret = sprintf(buf, "%d\n", device_info->chn_id); - } else if (!strcmp(dev_attr->attr.name, "out_intr_mask")) { - ret = sprintf(buf, "%d\n", device_info->outbound.int_mask); - } else if (!strcmp(dev_attr->attr.name, "out_read_index")) { - ret = sprintf(buf, "%d\n", device_info->outbound.read_idx); - } else if (!strcmp(dev_attr->attr.name, "out_write_index")) { - ret = sprintf(buf, "%d\n", device_info->outbound.write_idx); - } else if (!strcmp(dev_attr->attr.name, "out_read_bytes_avail")) { - ret = sprintf(buf, "%d\n", - device_info->outbound.bytes_avail_toread); - } else if (!strcmp(dev_attr->attr.name, "out_write_bytes_avail")) { - ret = sprintf(buf, "%d\n", - device_info->outbound.bytes_avail_towrite); - } else if (!strcmp(dev_attr->attr.name, "in_intr_mask")) { - ret = sprintf(buf, "%d\n", device_info->inbound.int_mask); - } else if (!strcmp(dev_attr->attr.name, "in_read_index")) { - ret = sprintf(buf, "%d\n", device_info->inbound.read_idx); - } else if (!strcmp(dev_attr->attr.name, "in_write_index")) { - ret = sprintf(buf, "%d\n", device_info->inbound.write_idx); - } else if (!strcmp(dev_attr->attr.name, "in_read_bytes_avail")) { - ret = sprintf(buf, "%d\n", - device_info->inbound.bytes_avail_toread); - } else if (!strcmp(dev_attr->attr.name, "in_write_bytes_avail")) { - ret = sprintf(buf, "%d\n", - device_info->inbound.bytes_avail_towrite); - } else if (!strcmp(dev_attr->attr.name, "monitor_id")) { - ret = sprintf(buf, "%d\n", device_info->monitor_id); - } else if (!strcmp(dev_attr->attr.name, "server_monitor_pending")) { - ret = sprintf(buf, "%d\n", device_info->server_monitor_pending); - } else if (!strcmp(dev_attr->attr.name, "server_monitor_latency")) { - ret = sprintf(buf, "%d\n", device_info->server_monitor_latency); - } else if (!strcmp(dev_attr->attr.name, "server_monitor_conn_id")) { - ret = sprintf(buf, "%d\n", - device_info->server_monitor_conn_id); - } else if (!strcmp(dev_attr->attr.name, "client_monitor_pending")) { - ret = sprintf(buf, "%d\n", device_info->client_monitor_pending); - } else if (!strcmp(dev_attr->attr.name, "client_monitor_latency")) { - ret = sprintf(buf, "%d\n", device_info->client_monitor_latency); - } else if (!strcmp(dev_attr->attr.name, "client_monitor_conn_id")) { - ret = sprintf(buf, "%d\n", - device_info->client_monitor_conn_id); - } +static ssize_t server_monitor_pending_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); - kfree(device_info); - return ret; + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "%d\n", + channel_pending(hv_dev->channel, + vmbus_connection.monitor_pages[1])); +} +static DEVICE_ATTR_RO(server_monitor_pending); + +static ssize_t client_monitor_pending_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "%d\n", + channel_pending(hv_dev->channel, + vmbus_connection.monitor_pages[1])); +} +static DEVICE_ATTR_RO(client_monitor_pending); + +static ssize_t server_monitor_latency_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "%d\n", + channel_latency(hv_dev->channel, + vmbus_connection.monitor_pages[0])); +} +static DEVICE_ATTR_RO(server_monitor_latency); + +static ssize_t client_monitor_latency_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "%d\n", + channel_latency(hv_dev->channel, + vmbus_connection.monitor_pages[1])); +} +static DEVICE_ATTR_RO(client_monitor_latency); + +static ssize_t server_monitor_conn_id_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "%d\n", + channel_conn_id(hv_dev->channel, + vmbus_connection.monitor_pages[0])); +} +static DEVICE_ATTR_RO(server_monitor_conn_id); + +static ssize_t client_monitor_conn_id_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + + if (!hv_dev->channel) + return -ENODEV; + return sprintf(buf, "%d\n", + channel_conn_id(hv_dev->channel, + vmbus_connection.monitor_pages[1])); +} +static DEVICE_ATTR_RO(client_monitor_conn_id); + +static ssize_t out_intr_mask_show(struct device *dev, + struct device_attribute *dev_attr, char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + struct hv_ring_buffer_debug_info outbound; + + if (!hv_dev->channel) + return -ENODEV; + hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound); + return sprintf(buf, "%d\n", outbound.current_interrupt_mask); } +static DEVICE_ATTR_RO(out_intr_mask); + +static ssize_t out_read_index_show(struct device *dev, + struct device_attribute *dev_attr, char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + struct hv_ring_buffer_debug_info outbound; + + if (!hv_dev->channel) + return -ENODEV; + hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound); + return sprintf(buf, "%d\n", outbound.current_read_index); +} +static DEVICE_ATTR_RO(out_read_index); + +static ssize_t out_write_index_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + struct hv_ring_buffer_debug_info outbound; + + if (!hv_dev->channel) + return -ENODEV; + hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound); + return sprintf(buf, "%d\n", outbound.current_write_index); +} +static DEVICE_ATTR_RO(out_write_index); + +static ssize_t out_read_bytes_avail_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + struct hv_ring_buffer_debug_info outbound; + + if (!hv_dev->channel) + return -ENODEV; + hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound); + return sprintf(buf, "%d\n", outbound.bytes_avail_toread); +} +static DEVICE_ATTR_RO(out_read_bytes_avail); + +static ssize_t out_write_bytes_avail_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + struct hv_ring_buffer_debug_info outbound; + + if (!hv_dev->channel) + return -ENODEV; + hv_ringbuffer_get_debuginfo(&hv_dev->channel->outbound, &outbound); + return sprintf(buf, "%d\n", outbound.bytes_avail_towrite); +} +static DEVICE_ATTR_RO(out_write_bytes_avail); + +static ssize_t in_intr_mask_show(struct device *dev, + struct device_attribute *dev_attr, char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + struct hv_ring_buffer_debug_info inbound; + + if (!hv_dev->channel) + return -ENODEV; + hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound); + return sprintf(buf, "%d\n", inbound.current_interrupt_mask); +} +static DEVICE_ATTR_RO(in_intr_mask); + +static ssize_t in_read_index_show(struct device *dev, + struct device_attribute *dev_attr, char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + struct hv_ring_buffer_debug_info inbound; + + if (!hv_dev->channel) + return -ENODEV; + hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound); + return sprintf(buf, "%d\n", inbound.current_read_index); +} +static DEVICE_ATTR_RO(in_read_index); + +static ssize_t in_write_index_show(struct device *dev, + struct device_attribute *dev_attr, char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + struct hv_ring_buffer_debug_info inbound; + + if (!hv_dev->channel) + return -ENODEV; + hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound); + return sprintf(buf, "%d\n", inbound.current_write_index); +} +static DEVICE_ATTR_RO(in_write_index); + +static ssize_t in_read_bytes_avail_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + struct hv_ring_buffer_debug_info inbound; + + if (!hv_dev->channel) + return -ENODEV; + hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound); + return sprintf(buf, "%d\n", inbound.bytes_avail_toread); +} +static DEVICE_ATTR_RO(in_read_bytes_avail); + +static ssize_t in_write_bytes_avail_show(struct device *dev, + struct device_attribute *dev_attr, + char *buf) +{ + struct hv_device *hv_dev = device_to_hv_device(dev); + struct hv_ring_buffer_debug_info inbound; + + if (!hv_dev->channel) + return -ENODEV; + hv_ringbuffer_get_debuginfo(&hv_dev->channel->inbound, &inbound); + return sprintf(buf, "%d\n", inbound.bytes_avail_towrite); +} +static DEVICE_ATTR_RO(in_write_bytes_avail); /* Set up per device attributes in /sys/bus/vmbus/devices/<bus device> */ -static struct device_attribute vmbus_device_attrs[] = { - __ATTR(id, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(state, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(class_id, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(device_id, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(monitor_id, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(modalias, S_IRUGO, vmbus_show_device_attr, NULL), - - __ATTR(server_monitor_pending, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(server_monitor_latency, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(server_monitor_conn_id, S_IRUGO, vmbus_show_device_attr, NULL), - - __ATTR(client_monitor_pending, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(client_monitor_latency, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(client_monitor_conn_id, S_IRUGO, vmbus_show_device_attr, NULL), - - __ATTR(out_intr_mask, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(out_read_index, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(out_write_index, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(out_read_bytes_avail, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(out_write_bytes_avail, S_IRUGO, vmbus_show_device_attr, NULL), - - __ATTR(in_intr_mask, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(in_read_index, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(in_write_index, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(in_read_bytes_avail, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR(in_write_bytes_avail, S_IRUGO, vmbus_show_device_attr, NULL), - __ATTR_NULL +static struct attribute *vmbus_attrs[] = { + &dev_attr_id.attr, + &dev_attr_state.attr, + &dev_attr_monitor_id.attr, + &dev_attr_class_id.attr, + &dev_attr_device_id.attr, + &dev_attr_modalias.attr, + &dev_attr_server_monitor_pending.attr, + &dev_attr_client_monitor_pending.attr, + &dev_attr_server_monitor_latency.attr, + &dev_attr_client_monitor_latency.attr, + &dev_attr_server_monitor_conn_id.attr, + &dev_attr_client_monitor_conn_id.attr, + &dev_attr_out_intr_mask.attr, + &dev_attr_out_read_index.attr, + &dev_attr_out_write_index.attr, + &dev_attr_out_read_bytes_avail.attr, + &dev_attr_out_write_bytes_avail.attr, + &dev_attr_in_intr_mask.attr, + &dev_attr_in_read_index.attr, + &dev_attr_in_write_index.attr, + &dev_attr_in_read_bytes_avail.attr, + &dev_attr_in_write_bytes_avail.attr, + NULL, }; - +ATTRIBUTE_GROUPS(vmbus); /* * vmbus_uevent - add uevent for our device @@ -257,7 +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) { @@ -272,7 +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))) @@ -382,12 +560,9 @@ static struct bus_type hv_bus = { .remove = vmbus_remove, .probe = vmbus_probe, .uevent = vmbus_uevent, - .dev_attrs = vmbus_device_attrs, + .dev_groups = vmbus_groups, }; -static const char *driver_name = "hyperv"; - - struct onmessage_work_context { struct work_struct work; struct hv_message msg; @@ -433,7 +608,7 @@ static void vmbus_on_msg_dpc(unsigned long data) * will not deliver any more messages since there is * no empty slot */ - smp_mb(); + mb(); if (msg->header.message_flags.msg_pending) { /* @@ -446,7 +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; @@ -454,34 +629,46 @@ static irqreturn_t vmbus_isr(int irq, void *dev_id) union hv_synic_event_flags *event; bool handled = false; + page_addr = hv_context.synic_event_page[cpu]; + if (page_addr == NULL) + return; + + event = (union hv_synic_event_flags *)page_addr + + VMBUS_MESSAGE_SINT; /* * Check for events before checking for messages. This is the order * in which events and messages are checked in Windows guests on * Hyper-V, and the Windows team suggested we do the same. */ - page_addr = hv_context.synic_event_page[cpu]; - event = (union hv_synic_event_flags *)page_addr + VMBUS_MESSAGE_SINT; + if ((vmbus_proto_version == VERSION_WS2008) || + (vmbus_proto_version == VERSION_WIN7)) { - /* Since we are a child, we only need to check bit 0 */ - if (sync_test_and_clear_bit(0, (unsigned long *) &event->flags32[0])) { + /* Since we are a child, we only need to check bit 0 */ + if (sync_test_and_clear_bit(0, + (unsigned long *) &event->flags32[0])) { + handled = true; + } + } else { + /* + * Our host is win8 or above. The signaling mechanism + * has changed and we can directly look at the event page. + * If bit n is set then we have an interrup on the channel + * whose id is n. + */ handled = true; - tasklet_schedule(&event_dpc); } + if (handled) + tasklet_schedule(hv_context.event_dpc[cpu]); + + page_addr = hv_context.synic_message_page[cpu]; 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; } /* @@ -496,7 +683,6 @@ static irqreturn_t vmbus_isr(int irq, void *dev_id) static int vmbus_bus_init(int irq) { int ret; - unsigned int vector; /* Hypervisor initialization...setup hypercall page..etc */ ret = hv_init(); @@ -506,39 +692,33 @@ static int vmbus_bus_init(int irq) } tasklet_init(&msg_dpc, vmbus_on_msg_dpc, 0); - tasklet_init(&event_dpc, vmbus_on_event, 0); ret = bus_register(&hv_bus); 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; - } - - vector = IRQ0_VECTOR + irq; + hv_setup_vmbus_irq(vmbus_isr); + ret = hv_synic_alloc(); + if (ret) + goto err_alloc; /* - * Notify the hypervisor of our irq and + * Initialize the per-cpu interrupt state and * connect to the host. */ - on_each_cpu(hv_synic_init, (void *)&vector, 1); + on_each_cpu(hv_synic_init, NULL, 1); ret = vmbus_connect(); if (ret) - goto err_irq; + goto err_alloc; vmbus_request_offers(); return 0; -err_irq: - free_irq(irq, hv_acpi_dev); +err_alloc: + hv_synic_free(); + hv_remove_vmbus_irq(); -err_unregister: bus_unregister(&hv_bus); err_cleanup: @@ -575,8 +755,6 @@ int __vmbus_driver_register(struct hv_driver *hv_driver, struct module *owner, c ret = driver_register(&hv_driver->driver); - vmbus_request_offers(); - return ret; } EXPORT_SYMBOL_GPL(__vmbus_driver_register); @@ -601,9 +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; @@ -647,7 +825,7 @@ int vmbus_device_register(struct hv_device *child_device_obj) if (ret) pr_err("Unable to register child device\n"); else - pr_info("child device %s registered\n", + pr_debug("child device %s registered\n", dev_name(&child_device_obj->device)); return ret; @@ -659,30 +837,33 @@ int vmbus_device_register(struct hv_device *child_device_obj) */ void vmbus_device_unregister(struct hv_device *device_obj) { + pr_debug("child device %s unregistered\n", + dev_name(&device_obj->device)); + /* * Kick off the process of unregistering the device. * This will call vmbus_remove() and eventually vmbus_device_release() */ device_unregister(&device_obj->device); - - pr_info("child device %s unregistered\n", - dev_name(&device_obj->device)); } /* - * 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) { - - if (res->type == ACPI_RESOURCE_TYPE_IRQ) { - struct acpi_resource_irq *irqp; - irqp = &res->data.irq; - - *((unsigned int *)irq) = irqp->interrupts[0]; + switch (res->type) { + case ACPI_RESOURCE_TYPE_IRQ: + irq = res->data.irq.interrupts[0]; + break; + + case ACPI_RESOURCE_TYPE_ADDRESS64: + hyperv_mmio.start = res->data.address64.minimum; + hyperv_mmio.end = res->data.address64.maximum; + break; } return AE_OK; @@ -691,18 +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[] = { @@ -732,7 +929,6 @@ static int __init hv_acpi_init(void) /* * Get irq resources first. */ - ret = acpi_bus_register_driver(&vmbus_acpi_driver); if (ret) @@ -763,8 +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(); @@ -773,7 +968,6 @@ static void __exit vmbus_exit(void) MODULE_LICENSE("GPL"); -MODULE_VERSION(HV_DRV_VERSION); subsys_initcall(hv_acpi_init); module_exit(vmbus_exit); |