summaryrefslogtreecommitdiffstats
path: root/drivers/hv
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/hv')
-rw-r--r--drivers/hv/channel.c19
-rw-r--r--drivers/hv/channel_mgmt.c52
-rw-r--r--drivers/hv/connection.c39
-rw-r--r--drivers/hv/hv.c2
-rw-r--r--drivers/hv/hv_balloon.c29
-rw-r--r--drivers/hv/hyperv_vmbus.h11
-rw-r--r--drivers/hv/vmbus_drv.c10
7 files changed, 130 insertions, 32 deletions
diff --git a/drivers/hv/channel.c b/drivers/hv/channel.c
index 602ca86a648..284cf66489f 100644
--- a/drivers/hv/channel.c
+++ b/drivers/hv/channel.c
@@ -471,18 +471,26 @@ int vmbus_teardown_gpadl(struct vmbus_channel *channel, u32 gpadl_handle)
}
EXPORT_SYMBOL_GPL(vmbus_teardown_gpadl);
+static void reset_channel_cb(void *arg)
+{
+ struct vmbus_channel *channel = arg;
+
+ channel->onchannel_callback = NULL;
+}
+
static void vmbus_close_internal(struct vmbus_channel *channel)
{
struct vmbus_channel_close_channel *msg;
int ret;
- unsigned long flags;
channel->state = CHANNEL_OPEN_STATE;
channel->sc_creation_callback = NULL;
/* Stop callback and cancel the timer asap */
- spin_lock_irqsave(&channel->inbound_lock, flags);
- channel->onchannel_callback = NULL;
- spin_unlock_irqrestore(&channel->inbound_lock, flags);
+ if (channel->target_cpu != smp_processor_id())
+ smp_call_function_single(channel->target_cpu, reset_channel_cb,
+ channel, true);
+ else
+ reset_channel_cb(channel);
/* Send a closing message */
@@ -674,8 +682,7 @@ int vmbus_sendpacket_multipagebuffer(struct vmbus_channel *channel,
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;
/*
diff --git a/drivers/hv/channel_mgmt.c b/drivers/hv/channel_mgmt.c
index fa920469bf1..ed9350d4276 100644
--- a/drivers/hv/channel_mgmt.c
+++ b/drivers/hv/channel_mgmt.c
@@ -149,6 +149,7 @@ static struct vmbus_channel *alloc_channel(void)
spin_lock_init(&channel->sc_lock);
INIT_LIST_HEAD(&channel->sc_list);
+ INIT_LIST_HEAD(&channel->percpu_list);
channel->controlwq = create_workqueue("hv_vmbus_ctl");
if (!channel->controlwq) {
@@ -188,7 +189,20 @@ static void free_channel(struct vmbus_channel *channel)
queue_work(vmbus_connection.work_queue, &channel->work);
}
+static void percpu_channel_enq(void *arg)
+{
+ struct vmbus_channel *channel = arg;
+ int cpu = smp_processor_id();
+
+ list_add_tail(&channel->percpu_list, &hv_context.percpu_list[cpu]);
+}
+static void percpu_channel_deq(void *arg)
+{
+ struct vmbus_channel *channel = arg;
+
+ list_del(&channel->percpu_list);
+}
/*
* vmbus_process_rescind_offer -
@@ -210,6 +224,12 @@ static void vmbus_process_rescind_offer(struct work_struct *work)
msg.header.msgtype = CHANNELMSG_RELID_RELEASED;
vmbus_post_msg(&msg, sizeof(struct vmbus_channel_relid_released));
+ if (channel->target_cpu != smp_processor_id())
+ smp_call_function_single(channel->target_cpu,
+ percpu_channel_deq, channel, true);
+ else
+ percpu_channel_deq(channel);
+
if (channel->primary_channel == NULL) {
spin_lock_irqsave(&vmbus_connection.channel_lock, flags);
list_del(&channel->listentry);
@@ -245,6 +265,7 @@ static void vmbus_process_offer(struct work_struct *work)
work);
struct vmbus_channel *channel;
bool fnew = true;
+ bool enq = false;
int ret;
unsigned long flags;
@@ -264,12 +285,22 @@ static void vmbus_process_offer(struct work_struct *work)
}
}
- if (fnew)
+ if (fnew) {
list_add_tail(&newchannel->listentry,
&vmbus_connection.chn_list);
+ enq = true;
+ }
spin_unlock_irqrestore(&vmbus_connection.channel_lock, flags);
+ if (enq) {
+ if (newchannel->target_cpu != smp_processor_id())
+ smp_call_function_single(newchannel->target_cpu,
+ percpu_channel_enq,
+ newchannel, true);
+ else
+ percpu_channel_enq(newchannel);
+ }
if (!fnew) {
/*
* Check to see if this is a sub-channel.
@@ -282,6 +313,14 @@ static void vmbus_process_offer(struct work_struct *work)
spin_lock_irqsave(&channel->sc_lock, flags);
list_add_tail(&newchannel->sc_list, &channel->sc_list);
spin_unlock_irqrestore(&channel->sc_lock, flags);
+
+ if (newchannel->target_cpu != smp_processor_id())
+ smp_call_function_single(newchannel->target_cpu,
+ percpu_channel_enq,
+ newchannel, true);
+ else
+ percpu_channel_enq(newchannel);
+
newchannel->state = CHANNEL_OPEN_STATE;
if (channel->sc_creation_callback != NULL)
channel->sc_creation_callback(newchannel);
@@ -365,7 +404,7 @@ static u32 next_vp;
* performance critical channels (IDE, SCSI and Network) will be uniformly
* distributed across all available CPUs.
*/
-static u32 get_vp_index(uuid_le *type_guid)
+static void init_vp_index(struct vmbus_channel *channel, const uuid_le *type_guid)
{
u32 cur_cpu;
int i;
@@ -387,10 +426,13 @@ static u32 get_vp_index(uuid_le *type_guid)
* Also if the channel is not a performance critical
* channel, bind it to cpu 0.
*/
- return 0;
+ channel->target_cpu = 0;
+ channel->target_vp = 0;
+ return;
}
cur_cpu = (++next_vp % max_cpus);
- return hv_context.vp_index[cur_cpu];
+ channel->target_cpu = cur_cpu;
+ channel->target_vp = hv_context.vp_index[cur_cpu];
}
/*
@@ -438,7 +480,7 @@ static void vmbus_onoffer(struct vmbus_channel_message_header *hdr)
offer->connection_id;
}
- newchannel->target_vp = get_vp_index(&offer->offer.if_type);
+ init_vp_index(newchannel, &offer->offer.if_type);
memcpy(&newchannel->offermsg, offer,
sizeof(struct vmbus_channel_offer_channel));
diff --git a/drivers/hv/connection.c b/drivers/hv/connection.c
index 2e7801af466..e84f4526eb3 100644
--- a/drivers/hv/connection.c
+++ b/drivers/hv/connection.c
@@ -224,8 +224,8 @@ cleanup:
vmbus_connection.int_page = NULL;
}
- free_pages((unsigned long)vmbus_connection.monitor_pages[0], 1);
- free_pages((unsigned long)vmbus_connection.monitor_pages[1], 1);
+ free_pages((unsigned long)vmbus_connection.monitor_pages[0], 0);
+ free_pages((unsigned long)vmbus_connection.monitor_pages[1], 0);
vmbus_connection.monitor_pages[0] = NULL;
vmbus_connection.monitor_pages[1] = NULL;
@@ -234,6 +234,28 @@ cleanup:
return ret;
}
+/*
+ * Map the given relid to the corresponding channel based on the
+ * per-cpu list of channels that have been affinitized to this CPU.
+ * This will be used in the channel callback path as we can do this
+ * mapping in a lock-free fashion.
+ */
+static struct vmbus_channel *pcpu_relid2channel(u32 relid)
+{
+ struct vmbus_channel *channel;
+ struct vmbus_channel *found_channel = NULL;
+ int cpu = smp_processor_id();
+ struct list_head *pcpu_head = &hv_context.percpu_list[cpu];
+
+ list_for_each_entry(channel, pcpu_head, percpu_list) {
+ if (channel->offermsg.child_relid == relid) {
+ found_channel = channel;
+ break;
+ }
+ }
+
+ return found_channel;
+}
/*
* relid2channel - Get the channel object given its
@@ -277,7 +299,6 @@ struct vmbus_channel *relid2channel(u32 relid)
static void process_chn_event(u32 relid)
{
struct vmbus_channel *channel;
- unsigned long flags;
void *arg;
bool read_state;
u32 bytes_to_read;
@@ -286,7 +307,7 @@ static void process_chn_event(u32 relid)
* Find the channel based on this relid and invokes the
* channel callback to process the event
*/
- channel = relid2channel(relid);
+ channel = pcpu_relid2channel(relid);
if (!channel) {
pr_err("channel not found for relid - %u\n", relid);
@@ -296,13 +317,12 @@ static void process_chn_event(u32 relid)
/*
* A channel once created is persistent even when there
* is no driver handling the device. An unloading driver
- * sets the onchannel_callback to NULL under the
- * protection of the channel inbound_lock. Thus, checking
- * and invoking the driver specific callback takes care of
- * orderly unloading of the driver.
+ * sets the onchannel_callback to NULL on the same CPU
+ * as where this interrupt is handled (in an interrupt context).
+ * Thus, checking and invoking the driver specific callback takes
+ * care of orderly unloading of the driver.
*/
- spin_lock_irqsave(&channel->inbound_lock, flags);
if (channel->onchannel_callback != NULL) {
arg = channel->channel_callback_context;
read_state = channel->batched_reading;
@@ -327,7 +347,6 @@ static void process_chn_event(u32 relid)
pr_err("no channel callback for relid - %u\n", relid);
}
- spin_unlock_irqrestore(&channel->inbound_lock, flags);
}
/*
diff --git a/drivers/hv/hv.c b/drivers/hv/hv.c
index bcb49502c3b..edfc8488cb0 100644
--- a/drivers/hv/hv.c
+++ b/drivers/hv/hv.c
@@ -383,6 +383,8 @@ void hv_synic_init(void *arg)
*/
rdmsrl(HV_X64_MSR_VP_INDEX, vp_index);
hv_context.vp_index[cpu] = (u32)vp_index;
+
+ INIT_LIST_HEAD(&hv_context.percpu_list[cpu]);
return;
}
diff --git a/drivers/hv/hv_balloon.c b/drivers/hv/hv_balloon.c
index 7e6d78dc943..5e90c5d771a 100644
--- a/drivers/hv/hv_balloon.c
+++ b/drivers/hv/hv_balloon.c
@@ -19,6 +19,7 @@
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/kernel.h>
+#include <linux/jiffies.h>
#include <linux/mman.h>
#include <linux/delay.h>
#include <linux/init.h>
@@ -459,6 +460,11 @@ static bool do_hot_add;
*/
static uint pressure_report_delay = 45;
+/*
+ * The last time we posted a pressure report to host.
+ */
+static unsigned long last_post_time;
+
module_param(hot_add, bool, (S_IRUGO | S_IWUSR));
MODULE_PARM_DESC(hot_add, "If set attempt memory hot_add");
@@ -542,6 +548,7 @@ struct hv_dynmem_device {
static struct hv_dynmem_device dm_device;
+static void post_status(struct hv_dynmem_device *dm);
#ifdef CONFIG_MEMORY_HOTPLUG
static void hv_bring_pgs_online(unsigned long start_pfn, unsigned long size)
@@ -612,7 +619,7 @@ static void hv_mem_hot_add(unsigned long start, unsigned long size,
* have not been "onlined" within the allowed time.
*/
wait_for_completion_timeout(&dm_device.ol_waitevent, 5*HZ);
-
+ post_status(&dm_device);
}
return;
@@ -951,11 +958,17 @@ static void post_status(struct hv_dynmem_device *dm)
{
struct dm_status status;
struct sysinfo val;
+ unsigned long now = jiffies;
+ unsigned long last_post = last_post_time;
if (pressure_report_delay > 0) {
--pressure_report_delay;
return;
}
+
+ if (!time_after(now, (last_post_time + HZ)))
+ return;
+
si_meminfo(&val);
memset(&status, 0, sizeof(struct dm_status));
status.hdr.type = DM_STATUS_REPORT;
@@ -983,6 +996,14 @@ static void post_status(struct hv_dynmem_device *dm)
if (status.hdr.trans_id != atomic_read(&trans_id))
return;
+ /*
+ * If the last post time that we sampled has changed,
+ * we have raced, don't post the status.
+ */
+ if (last_post != last_post_time)
+ return;
+
+ last_post_time = jiffies;
vmbus_sendpacket(dm->dev->channel, &status,
sizeof(struct dm_status),
(unsigned long)NULL,
@@ -1117,7 +1138,7 @@ static void balloon_up(struct work_struct *dummy)
if (ret == -EAGAIN)
msleep(20);
-
+ post_status(&dm_device);
} while (ret == -EAGAIN);
if (ret) {
@@ -1144,8 +1165,10 @@ static void balloon_down(struct hv_dynmem_device *dm,
struct dm_unballoon_response resp;
int i;
- for (i = 0; i < range_count; i++)
+ for (i = 0; i < range_count; i++) {
free_balloon_pages(dm, &range_array[i]);
+ post_status(&dm_device);
+ }
if (req->more_pages == 1)
return;
diff --git a/drivers/hv/hyperv_vmbus.h b/drivers/hv/hyperv_vmbus.h
index 860134da803..22b750749a3 100644
--- a/drivers/hv/hyperv_vmbus.h
+++ b/drivers/hv/hyperv_vmbus.h
@@ -510,6 +510,11 @@ struct hv_context {
* basis.
*/
struct tasklet_struct *event_dpc[NR_CPUS];
+ /*
+ * To optimize the mapping of relid to channel, maintain
+ * per-cpu list of the channels based on their CPU affinity.
+ */
+ struct list_head percpu_list[NR_CPUS];
};
extern struct hv_context hv_context;
@@ -644,9 +649,9 @@ extern struct vmbus_connection vmbus_connection;
/* General vmbus interface */
-struct hv_device *vmbus_device_create(uuid_le *type,
- uuid_le *instance,
- struct vmbus_channel *channel);
+struct hv_device *vmbus_device_create(const uuid_le *type,
+ const uuid_le *instance,
+ struct vmbus_channel *channel);
int vmbus_device_register(struct hv_device *child_device_obj);
void vmbus_device_unregister(struct hv_device *device_obj);
diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c
index 8e53a3c2607..4d6b26979fb 100644
--- a/drivers/hv/vmbus_drv.c
+++ b/drivers/hv/vmbus_drv.c
@@ -435,7 +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)
{
@@ -450,7 +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)))
@@ -779,9 +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;