@@ -385,11 +385,15 @@ static struct vas_window *vas_allocate_window(int vas_id, u64 flags,
* same fault IRQ is not freed by the OS before.
*/
mutex_lock(&vas_pseries_mutex);
- if (migration_in_progress)
+ if (migration_in_progress) {
rc = -EBUSY;
- else
+ } else {
rc = allocate_setup_window(txwin, (u64 *)&domain[0],
cop_feat_caps->win_type);
+ if (!rc)
+ caps->nr_open_wins_progress++;
+ }
+
mutex_unlock(&vas_pseries_mutex);
if (rc)
goto out;
@@ -404,8 +408,17 @@ static struct vas_window *vas_allocate_window(int vas_id, u64 flags,
goto out_free;
txwin->win_type = cop_feat_caps->win_type;
- mutex_lock(&vas_pseries_mutex);
+
/*
+ * The migration SUSPEND thread sets migration_in_progress and
+ * closes all open windows from the list. But the window is
+ * added to the list after open and modify HCALLs. So possible
+ * that migration_in_progress is set before modify HCALL which
+ * may cause some windows are still open when the hypervisor
+ * initiates the migration.
+ * So checks the migration_in_progress flag again and close all
+ * open windows.
+ *
* Possible to lose the acquired credit with DLPAR core
* removal after the window is opened. So if there are any
* closed windows (means with lost credits), do not give new
@@ -413,9 +426,11 @@ static struct vas_window *vas_allocate_window(int vas_id, u64 flags,
* after the existing windows are reopened when credits are
* available.
*/
- if (!caps->nr_close_wins) {
+ mutex_lock(&vas_pseries_mutex);
+ if (!caps->nr_close_wins && !migration_in_progress) {
list_add(&txwin->win_list, &caps->list);
caps->nr_open_windows++;
+ caps->nr_open_wins_progress--;
mutex_unlock(&vas_pseries_mutex);
vas_user_win_add_mm_context(&txwin->vas_win.task_ref);
return &txwin->vas_win;
@@ -433,6 +448,12 @@ static struct vas_window *vas_allocate_window(int vas_id, u64 flags,
*/
free_irq_setup(txwin);
h_deallocate_vas_window(txwin->vas_win.winid);
+ /*
+ * Hold mutex and reduce nr_open_wins_progress counter.
+ */
+ mutex_lock(&vas_pseries_mutex);
+ caps->nr_open_wins_progress--;
+ mutex_unlock(&vas_pseries_mutex);
out:
atomic_dec(&cop_feat_caps->nr_used_credits);
kfree(txwin);
@@ -937,14 +958,14 @@ int vas_migration_handler(int action)
struct vas_caps *vcaps;
int i, rc = 0;
+ pr_info("VAS migration event %d\n", action);
+
/*
* NX-GZIP is not enabled. Nothing to do for migration.
*/
if (!copypaste_feat)
return rc;
- mutex_lock(&vas_pseries_mutex);
-
if (action == VAS_SUSPEND)
migration_in_progress = true;
else
@@ -990,12 +1011,27 @@ int vas_migration_handler(int action)
switch (action) {
case VAS_SUSPEND:
+ mutex_lock(&vas_pseries_mutex);
rc = reconfig_close_windows(vcaps, vcaps->nr_open_windows,
true);
+ /*
+ * Windows are included in the list after successful
+ * open. So wait for closing these in-progress open
+ * windows in vas_allocate_window() which will be
+ * done if the migration_in_progress is set.
+ */
+ while (vcaps->nr_open_wins_progress) {
+ mutex_unlock(&vas_pseries_mutex);
+ msleep(10);
+ mutex_lock(&vas_pseries_mutex);
+ }
+ mutex_unlock(&vas_pseries_mutex);
break;
case VAS_RESUME:
+ mutex_lock(&vas_pseries_mutex);
atomic_set(&caps->nr_total_credits, new_nr_creds);
rc = reconfig_open_windows(vcaps, new_nr_creds, true);
+ mutex_unlock(&vas_pseries_mutex);
break;
default:
/* should not happen */
@@ -1011,8 +1047,9 @@ int vas_migration_handler(int action)
goto out;
}
+ pr_info("VAS migration event (%d) successful\n", action);
+
out:
- mutex_unlock(&vas_pseries_mutex);
return rc;
}
@@ -91,6 +91,8 @@ struct vas_cop_feat_caps {
struct vas_caps {
struct vas_cop_feat_caps caps;
struct list_head list; /* List of open windows */
+ int nr_open_wins_progress; /* Number of open windows in */
+ /* progress. Used in migration */
int nr_close_wins; /* closed windows in the hypervisor for DLPAR */
int nr_open_windows; /* Number of successful open windows */
u8 feat; /* Feature type */
The hypervisor returns migration failure if all VAS windows are not closed. During pre-migration stage, vas_migration_handler() sets migration_in_progress flag and closes all windows from the list. The allocate VAS window routine checks the migration flag, setup the window and then add it to the list. So there is possibility of the migration handler missing the window that is still in the process of setup. t1: Allocate and open VAS t2: Migration event window lock vas_pseries_mutex If migration_in_progress set unlock vas_pseries_mutex return open window HCALL unlock vas_pseries_mutex Modify window HCALL lock vas_pseries_mutex setup window migration_in_progress=true Closes all windows from the list // May miss windows that are // not in the list unlock vas_pseries_mutex lock vas_pseries_mutex return if nr_closed_windows == 0 // No DLPAR CPU or migration add window to the list // Window will be added to the // list after the setup is completed unlock vas_pseries_mutex return unlock vas_pseries_mutex Close VAS window // due to DLPAR CPU or migration return -EBUSY This patch resolves the issue with the following steps: - Set the migration_in_progress flag without holding mutex. - Introduce nr_open_wins_progress counter in VAS capabilities struct - This counter tracks the number of open windows are still in progress - The allocate setup window thread closes windows if the migration is set and decrements nr_open_window_progress counter - The migration handler waits for no in-progress open windows. The code flow with the fix is as follows: t1: Allocate and open VAS t2: Migration event window lock vas_pseries_mutex If migration_in_progress set unlock vas_pseries_mutex return open window HCALL nr_open_wins_progress++ // Window opened, but not // added to the list yet unlock vas_pseries_mutex Modify window HCALL migration_in_progress=true setup window lock vas_pseries_mutex Closes all windows from the list While nr_open_wins_progress { unlock vas_pseries_mutex lock vas_pseries_mutex sleep if nr_closed_windows == 0 // Wait if any open window in or migration is not started // progress. The open window // No DLPAR CPU or migration // thread closes the window without add window to the list // adding to the list and return if nr_open_wins_progress-- // the migration is in progress. unlock vas_pseries_mutex return Close VAS window nr_open_wins_progress-- unlock vas_pseries_mutex return -EBUSY lock vas_pseries_mutex } unlock vas_pseries_mutex return Fixes: 37e6764895ef ("powerpc/pseries/vas: Add VAS migration handler") Signed-off-by: Haren Myneni <haren@linux.ibm.com> --- v1 -> v2: - Do not define the migration_in_progress flag as atomic as suggested by Nathan v2 -> v3: - Use wait_event() instead of wait_event_interruptible() so that returns after all windows are closed as suggested by Nathan v3 -> v4: - remove atomic for nr_open_wins_progress counter as suggested by Nathan and Michael Ellerman - Use sleep instead of wait_event_interruptible() to check nr_open_wins_progress counter under mutex. v4 -> v5: - Update the commit message with comments in the code flow and added the second code flow with the fix as suggested by Michael Ellerman --- arch/powerpc/platforms/pseries/vas.c | 51 ++++++++++++++++++++++++---- arch/powerpc/platforms/pseries/vas.h | 2 ++ 2 files changed, 46 insertions(+), 7 deletions(-)