[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] [PATCH 1/1] xen/time: do not decrease steal time after live migration on xen



After guest live migration on xen, steal time in /proc/stat
(cpustat[CPUTIME_STEAL]) might decrease because steal returned by
xen_steal_lock() might be less than this_rq()->prev_steal_time which is
derived from previous return value of xen_steal_clock().

For instance, steal time of each vcpu is 335 before live migration.

cpu  198 0 368 200064 1962 0 0 1340 0 0
cpu0 38 0 81 50063 492 0 0 335 0 0
cpu1 65 0 97 49763 634 0 0 335 0 0
cpu2 38 0 81 50098 462 0 0 335 0 0
cpu3 56 0 107 50138 374 0 0 335 0 0

After live migration, steal time is reduced to 312.

cpu  200 0 370 200330 1971 0 0 1248 0 0
cpu0 38 0 82 50123 500 0 0 312 0 0
cpu1 65 0 97 49832 634 0 0 312 0 0
cpu2 39 0 82 50167 462 0 0 312 0 0
cpu3 56 0 107 50207 374 0 0 312 0 0

The code in this patch is borrowed from do_stolen_accounting() which has
already been removed from linux source code since commit ecb23dc6f2ef
("xen: add steal_clock support on x86"). The core idea of both
do_stolen_accounting() and this patch is to avoid accounting new steal
clock if it is smaller than previous old steal clock.

Similar and more severe issue would impact prior linux 4.8-4.10 as
discussed by Michael Las at
https://0xstubs.org/debugging-a-flaky-cpu-steal-time-counter-on-a-paravirtualized-xen-guest,
which would overflow steal time and lead to 100% st usage in top command
for linux 4.8-4.10. A backport of this patch would fix that issue.

References: 
https://0xstubs.org/debugging-a-flaky-cpu-steal-time-counter-on-a-paravirtualized-xen-guest
Signed-off-by: Dongli Zhang <dongli.zhang@xxxxxxxxxx>
---
 drivers/xen/time.c | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

diff --git a/drivers/xen/time.c b/drivers/xen/time.c
index ac5f23f..2b3a996 100644
--- a/drivers/xen/time.c
+++ b/drivers/xen/time.c
@@ -19,6 +19,8 @@
 /* runstate info updated by Xen */
 static DEFINE_PER_CPU(struct vcpu_runstate_info, xen_runstate);
 
+static DEFINE_PER_CPU(u64, xen_old_steal);
+
 /* return an consistent snapshot of 64-bit time/counter value */
 static u64 get64(const u64 *p)
 {
@@ -83,9 +85,20 @@ bool xen_vcpu_stolen(int vcpu)
 u64 xen_steal_clock(int cpu)
 {
        struct vcpu_runstate_info state;
+       u64 xen_new_steal;
+       s64 steal_delta;
 
        xen_get_runstate_snapshot_cpu(&state, cpu);
-       return state.time[RUNSTATE_runnable] + state.time[RUNSTATE_offline];
+       xen_new_steal = state.time[RUNSTATE_runnable]
+                                       + state.time[RUNSTATE_offline];
+       steal_delta = xen_new_steal - per_cpu(xen_old_steal, cpu);
+
+       if (steal_delta < 0)
+               xen_new_steal = per_cpu(xen_old_steal, cpu);
+       else
+               per_cpu(xen_old_steal, cpu) = xen_new_steal;
+
+       return xen_new_steal;
 }
 
 void xen_setup_runstate_info(int cpu)
-- 
2.7.4


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.