diff options
Diffstat (limited to 'include/linux/hyperv.h')
| -rw-r--r-- | include/linux/hyperv.h | 147 |
1 files changed, 116 insertions, 31 deletions
diff --git a/include/linux/hyperv.h b/include/linux/hyperv.h index b10954a66939..42fe43fb0c80 100644 --- a/include/linux/hyperv.h +++ b/include/linux/hyperv.h @@ -674,6 +674,11 @@ enum hv_signal_policy { HV_SIGNAL_POLICY_EXPLICIT, }; +enum hv_numa_policy { + HV_BALANCED = 0, + HV_LOCALIZED, +}; + enum vmbus_device_type { HV_IDE = 0, HV_SCSI, @@ -691,7 +696,7 @@ enum vmbus_device_type { HV_FCOPY, HV_BACKUP, HV_DM, - HV_UNKOWN, + HV_UNKNOWN, }; struct vmbus_device { @@ -701,9 +706,6 @@ struct vmbus_device { }; struct vmbus_channel { - /* Unique channel id */ - int id; - struct list_head listentry; struct hv_device *device_obj; @@ -850,6 +852,43 @@ struct vmbus_channel { * ring lock to preserve the current behavior. */ bool acquire_ring_lock; + /* + * For performance critical channels (storage, networking + * etc,), Hyper-V has a mechanism to enhance the throughput + * at the expense of latency: + * When the host is to be signaled, we just set a bit in a shared page + * and this bit will be inspected by the hypervisor within a certain + * window and if the bit is set, the host will be signaled. The window + * of time is the monitor latency - currently around 100 usecs. This + * mechanism improves throughput by: + * + * A) Making the host more efficient - each time it wakes up, + * potentially it will process morev number of packets. The + * monitor latency allows a batch to build up. + * B) By deferring the hypercall to signal, we will also minimize + * the interrupts. + * + * Clearly, these optimizations improve throughput at the expense of + * latency. Furthermore, since the channel is shared for both + * control and data messages, control messages currently suffer + * unnecessary latency adversley impacting performance and boot + * time. To fix this issue, permit tagging the channel as being + * in "low latency" mode. In this mode, we will bypass the monitor + * mechanism. + */ + bool low_latency; + + /* + * NUMA distribution policy: + * We support teo policies: + * 1) Balanced: Here all performance critical channels are + * distributed evenly amongst all the NUMA nodes. + * This policy will be the default policy. + * 2) Localized: All channels of a given instance of a + * performance critical service will be assigned CPUs + * within a selected NUMA node. + */ + enum hv_numa_policy affinity_policy; }; @@ -870,6 +909,12 @@ static inline void set_channel_signal_state(struct vmbus_channel *c, c->signal_policy = policy; } +static inline void set_channel_affinity_state(struct vmbus_channel *c, + enum hv_numa_policy policy) +{ + c->affinity_policy = policy; +} + static inline void set_channel_read_state(struct vmbus_channel *c, bool state) { c->batched_reading = state; @@ -891,6 +936,16 @@ static inline void set_channel_pending_send_size(struct vmbus_channel *c, c->outbound.ring_buffer->pending_send_sz = size; } +static inline void set_low_latency_mode(struct vmbus_channel *c) +{ + c->low_latency = true; +} + +static inline void clear_low_latency_mode(struct vmbus_channel *c) +{ + c->low_latency = false; +} + void vmbus_onmessage(void *context); int vmbus_request_offers(void); @@ -1064,6 +1119,12 @@ struct hv_driver { struct device_driver driver; + /* dynamic device GUID's */ + struct { + spinlock_t lock; + struct list_head list; + } dynids; + int (*probe)(struct hv_device *, const struct hv_vmbus_device_id *); int (*remove)(struct hv_device *); void (*shutdown)(struct hv_device *); @@ -1257,6 +1318,27 @@ u64 hv_do_hypercall(u64 control, void *input, void *output); 0x80, 0x2e, 0x27, 0xed, 0xe1, 0x9f) /* + * Linux doesn't support the 3 devices: the first two are for + * Automatic Virtual Machine Activation, and the third is for + * Remote Desktop Virtualization. + * {f8e65716-3cb3-4a06-9a60-1889c5cccab5} + * {3375baf4-9e15-4b30-b765-67acb10d607b} + * {276aacf4-ac15-426c-98dd-7521ad3f01fe} + */ + +#define HV_AVMA1_GUID \ + .guid = UUID_LE(0xf8e65716, 0x3cb3, 0x4a06, 0x9a, 0x60, \ + 0x18, 0x89, 0xc5, 0xcc, 0xca, 0xb5) + +#define HV_AVMA2_GUID \ + .guid = UUID_LE(0x3375baf4, 0x9e15, 0x4b30, 0xb7, 0x65, \ + 0x67, 0xac, 0xb1, 0x0d, 0x60, 0x7b) + +#define HV_RDV_GUID \ + .guid = UUID_LE(0x276aacf4, 0xac15, 0x426c, 0x98, 0xdd, \ + 0x75, 0x21, 0xad, 0x3f, 0x01, 0xfe) + +/* * Common header for Hyper-V ICs */ @@ -1344,6 +1426,15 @@ struct ictimesync_data { u8 flags; } __packed; +struct ictimesync_ref_data { + u64 parenttime; + u64 vmreferencetime; + u8 flags; + char leapflags; + char stratum; + u8 reserved[3]; +} __packed; + struct hyperv_service_callback { u8 msg_type; char *log_msg; @@ -1357,8 +1448,12 @@ extern bool vmbus_prep_negotiate_resp(struct icmsg_hdr *, struct icmsg_negotiate *, u8 *, int, int); +void hv_event_tasklet_disable(struct vmbus_channel *channel); +void hv_event_tasklet_enable(struct vmbus_channel *channel); + void hv_process_channel_removal(struct vmbus_channel *channel, u32 relid); +void vmbus_setevent(struct vmbus_channel *channel); /* * Negotiated version with the Host. */ @@ -1391,10 +1486,11 @@ hv_get_ring_buffer(struct hv_ring_buffer_info *ring_info) * there is room for the producer to send the pending packet. */ -static inline bool hv_need_to_signal_on_read(struct hv_ring_buffer_info *rbi) +static inline void hv_signal_on_read(struct vmbus_channel *channel) { u32 cur_write_sz; u32 pending_sz; + struct hv_ring_buffer_info *rbi = &channel->inbound; /* * Issue a full memory barrier before making the signaling decision. @@ -1412,14 +1508,14 @@ static inline bool hv_need_to_signal_on_read(struct hv_ring_buffer_info *rbi) pending_sz = READ_ONCE(rbi->ring_buffer->pending_send_sz); /* If the other end is not blocked on write don't bother. */ if (pending_sz == 0) - return false; + return; cur_write_sz = hv_get_bytes_to_write(rbi); if (cur_write_sz >= pending_sz) - return true; + vmbus_setevent(channel); - return false; + return; } /* @@ -1431,31 +1527,23 @@ static inline struct vmpacket_descriptor * get_next_pkt_raw(struct vmbus_channel *channel) { struct hv_ring_buffer_info *ring_info = &channel->inbound; - u32 read_loc = ring_info->priv_read_index; + u32 priv_read_loc = ring_info->priv_read_index; void *ring_buffer = hv_get_ring_buffer(ring_info); - struct vmpacket_descriptor *cur_desc; - u32 packetlen; u32 dsize = ring_info->ring_datasize; - u32 delta = read_loc - ring_info->ring_buffer->read_index; + /* + * delta is the difference between what is available to read and + * what was already consumed in place. We commit read index after + * the whole batch is processed. + */ + u32 delta = priv_read_loc >= ring_info->ring_buffer->read_index ? + priv_read_loc - ring_info->ring_buffer->read_index : + (dsize - ring_info->ring_buffer->read_index) + priv_read_loc; u32 bytes_avail_toread = (hv_get_bytes_to_read(ring_info) - delta); if (bytes_avail_toread < sizeof(struct vmpacket_descriptor)) return NULL; - if ((read_loc + sizeof(*cur_desc)) > dsize) - return NULL; - - cur_desc = ring_buffer + read_loc; - packetlen = cur_desc->len8 << 3; - - /* - * If the packet under consideration is wrapping around, - * return failure. - */ - if ((read_loc + packetlen + VMBUS_PKT_TRAILER) > (dsize - 1)) - return NULL; - - return cur_desc; + return ring_buffer + priv_read_loc; } /* @@ -1467,16 +1555,14 @@ static inline void put_pkt_raw(struct vmbus_channel *channel, struct vmpacket_descriptor *desc) { struct hv_ring_buffer_info *ring_info = &channel->inbound; - u32 read_loc = ring_info->priv_read_index; u32 packetlen = desc->len8 << 3; u32 dsize = ring_info->ring_datasize; - if ((read_loc + packetlen + VMBUS_PKT_TRAILER) > dsize) - BUG(); /* * Include the packet trailer. */ ring_info->priv_read_index += packetlen + VMBUS_PKT_TRAILER; + ring_info->priv_read_index %= dsize; } /* @@ -1501,8 +1587,7 @@ static inline void commit_rd_index(struct vmbus_channel *channel) virt_rmb(); ring_info->ring_buffer->read_index = ring_info->priv_read_index; - if (hv_need_to_signal_on_read(ring_info)) - vmbus_set_event(channel); + hv_signal_on_read(channel); } |