summaryrefslogtreecommitdiffstats
path: root/drivers/hv
diff options
context:
space:
mode:
authorVitaly Kuznetsov <vkuznets@redhat.com>2015-02-28 11:38:58 -0800
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>2015-03-01 19:31:47 -0800
commitb05d8d9ef5ef21d1b18440430f950304836e1aaa (patch)
tree347dde1952b3f4a26c6a6c8c5871ee63a1b6b3de /drivers/hv
parent37f492ce8148d84261cdb9fe1e925199f47531ae (diff)
downloadop-kernel-dev-b05d8d9ef5ef21d1b18440430f950304836e1aaa.zip
op-kernel-dev-b05d8d9ef5ef21d1b18440430f950304836e1aaa.tar.gz
Drivers: hv: hv_balloon: eliminate the trylock path in acquire/release_region_mutex
When many memory regions are being added and automatically onlined the following lockup is sometimes observed: INFO: task udevd:1872 blocked for more than 120 seconds. ... Call Trace: [<ffffffff816ec0bc>] schedule_timeout+0x22c/0x350 [<ffffffff816eb98f>] wait_for_common+0x10f/0x160 [<ffffffff81067650>] ? default_wake_function+0x0/0x20 [<ffffffff816eb9fd>] wait_for_completion+0x1d/0x20 [<ffffffff8144cb9c>] hv_memory_notifier+0xdc/0x120 [<ffffffff816f298c>] notifier_call_chain+0x4c/0x70 ... When several memory blocks are going online simultaneously we got several hv_memory_notifier() trying to acquire the ha_region_mutex. When this mutex is being held by hot_add_req() all these competing acquire_region_mutex() do mutex_trylock, fail, and queue themselves into wait_for_completion(..). However when we do complete() from release_region_mutex() only one of them wakes up. This could be solved by changing complete() -> complete_all() memory onlining can be delayed as well, in that case we can still get several hv_memory_notifier() runners at the same time trying to grab the mutex. Only one of them will succeed and the others will hang for forever as complete() is not being called. We don't see this issue often because we have 5sec onlining timeout in hv_mem_hot_add() and usually all udev events arrive in this time frame. Get rid of the trylock path, waiting on the mutex is supposed to provide the required serialization. Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> Signed-off-by: K. Y. Srinivasan <kys@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Diffstat (limited to 'drivers/hv')
-rw-r--r--drivers/hv/hv_balloon.c35
1 files changed, 6 insertions, 29 deletions
diff --git a/drivers/hv/hv_balloon.c b/drivers/hv/hv_balloon.c
index 965c37a..ea3d33c 100644
--- a/drivers/hv/hv_balloon.c
+++ b/drivers/hv/hv_balloon.c
@@ -534,7 +534,6 @@ struct hv_dynmem_device {
struct task_struct *thread;
struct mutex ha_region_mutex;
- struct completion waiter_event;
/*
* A list of hot-add regions.
@@ -554,38 +553,17 @@ static struct hv_dynmem_device dm_device;
static void post_status(struct hv_dynmem_device *dm);
#ifdef CONFIG_MEMORY_HOTPLUG
-static void acquire_region_mutex(bool trylock)
-{
- if (trylock) {
- reinit_completion(&dm_device.waiter_event);
- while (!mutex_trylock(&dm_device.ha_region_mutex))
- wait_for_completion(&dm_device.waiter_event);
- } else {
- mutex_lock(&dm_device.ha_region_mutex);
- }
-}
-
-static void release_region_mutex(bool trylock)
-{
- if (trylock) {
- mutex_unlock(&dm_device.ha_region_mutex);
- } else {
- mutex_unlock(&dm_device.ha_region_mutex);
- complete(&dm_device.waiter_event);
- }
-}
-
static int hv_memory_notifier(struct notifier_block *nb, unsigned long val,
void *v)
{
switch (val) {
case MEM_GOING_ONLINE:
- acquire_region_mutex(true);
+ mutex_lock(&dm_device.ha_region_mutex);
break;
case MEM_ONLINE:
case MEM_CANCEL_ONLINE:
- release_region_mutex(true);
+ mutex_unlock(&dm_device.ha_region_mutex);
if (dm_device.ha_waiting) {
dm_device.ha_waiting = false;
complete(&dm_device.ol_waitevent);
@@ -646,7 +624,7 @@ static void hv_mem_hot_add(unsigned long start, unsigned long size,
init_completion(&dm_device.ol_waitevent);
dm_device.ha_waiting = true;
- release_region_mutex(false);
+ mutex_unlock(&dm_device.ha_region_mutex);
nid = memory_add_physaddr_to_nid(PFN_PHYS(start_pfn));
ret = add_memory(nid, PFN_PHYS((start_pfn)),
(HA_CHUNK << PAGE_SHIFT));
@@ -675,7 +653,7 @@ 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);
- acquire_region_mutex(false);
+ mutex_lock(&dm_device.ha_region_mutex);
post_status(&dm_device);
}
@@ -886,7 +864,7 @@ static void hot_add_req(struct work_struct *dummy)
resp.hdr.size = sizeof(struct dm_hot_add_response);
#ifdef CONFIG_MEMORY_HOTPLUG
- acquire_region_mutex(false);
+ mutex_lock(&dm_device.ha_region_mutex);
pg_start = dm->ha_wrk.ha_page_range.finfo.start_page;
pfn_cnt = dm->ha_wrk.ha_page_range.finfo.page_cnt;
@@ -918,7 +896,7 @@ static void hot_add_req(struct work_struct *dummy)
if (do_hot_add)
resp.page_count = process_hot_add(pg_start, pfn_cnt,
rg_start, rg_sz);
- release_region_mutex(false);
+ mutex_unlock(&dm_device.ha_region_mutex);
#endif
/*
* The result field of the response structure has the
@@ -1440,7 +1418,6 @@ 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_completion(&dm_device.waiter_event);
INIT_LIST_HEAD(&dm_device.ha_region_list);
mutex_init(&dm_device.ha_region_mutex);
INIT_WORK(&dm_device.balloon_wrk.wrk, balloon_up);
OpenPOWER on IntegriCloud