This is a note to let you know that I've just added the patch titled

    s390/time: fix sched_clock() overflow

to the 3.0-stable tree which can be found at:
    
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
     s390-time-fix-sched_clock-overflow.patch
and it can be found in the queue-3.0 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree,
please let <[email protected]> know about it.


>From ed4f20943cd4c7b55105c04daedf8d63ab6d499c Mon Sep 17 00:00:00 2001
From: Heiko Carstens <[email protected]>
Date: Mon, 14 Jan 2013 16:55:55 +0100
Subject: s390/time: fix sched_clock() overflow

From: Heiko Carstens <[email protected]>

commit ed4f20943cd4c7b55105c04daedf8d63ab6d499c upstream.

Converting a 64 Bit TOD format value to nanoseconds means that the value
must be divided by 4.096. In order to achieve that we multiply with 125
and divide by 512.
When used within sched_clock() this triggers an overflow after appr.
417 days. Resulting in a sched_clock() return value that is much smaller
than previously and therefore may cause all sort of weird things in
subsystems that rely on a monotonic sched_clock() behaviour.

To fix this implement a tod_to_ns() helper function which converts TOD
values without overflow and call this function from both places that
open coded the conversion: sched_clock() and kvm_s390_handle_wait().

Reviewed-by: Martin Schwidefsky <[email protected]>
Signed-off-by: Heiko Carstens <[email protected]>
Signed-off-by: Martin Schwidefsky <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

---
 arch/s390/include/asm/timex.h |   28 ++++++++++++++++++++++++++++
 arch/s390/kernel/time.c       |    2 +-
 arch/s390/kvm/interrupt.c     |    2 +-
 3 files changed, 30 insertions(+), 2 deletions(-)

--- a/arch/s390/include/asm/timex.h
+++ b/arch/s390/include/asm/timex.h
@@ -126,4 +126,32 @@ static inline unsigned long long get_clo
        return get_clock_xt() - sched_clock_base_cc;
 }
 
+/**
+ * tod_to_ns - convert a TOD format value to nanoseconds
+ * @todval: to be converted TOD format value
+ * Returns: number of nanoseconds that correspond to the TOD format value
+ *
+ * Converting a 64 Bit TOD format value to nanoseconds means that the value
+ * must be divided by 4.096. In order to achieve that we multiply with 125
+ * and divide by 512:
+ *
+ *    ns = (todval * 125) >> 9;
+ *
+ * In order to avoid an overflow with the multiplication we can rewrite this.
+ * With a split todval == 2^32 * th + tl (th upper 32 bits, tl lower 32 bits)
+ * we end up with
+ *
+ *    ns = ((2^32 * th + tl) * 125 ) >> 9;
+ * -> ns = (2^23 * th * 125) + ((tl * 125) >> 9);
+ *
+ */
+static inline unsigned long long tod_to_ns(unsigned long long todval)
+{
+       unsigned long long ns;
+
+       ns = ((todval >> 32) << 23) * 125;
+       ns += ((todval & 0xffffffff) * 125) >> 9;
+       return ns;
+}
+
 #endif
--- a/arch/s390/kernel/time.c
+++ b/arch/s390/kernel/time.c
@@ -63,7 +63,7 @@ static DEFINE_PER_CPU(struct clock_event
  */
 unsigned long long notrace __kprobes sched_clock(void)
 {
-       return (get_clock_monotonic() * 125) >> 9;
+       return tod_to_ns(get_clock_monotonic());
 }
 
 /*
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -358,7 +358,7 @@ int kvm_s390_handle_wait(struct kvm_vcpu
                return 0;
        }
 
-       sltime = ((vcpu->arch.sie_block->ckc - now)*125)>>9;
+       sltime = tod_to_ns(vcpu->arch.sie_block->ckc - now);
 
        hrtimer_start(&vcpu->arch.ckc_timer, ktime_set (0, sltime) , 
HRTIMER_MODE_REL);
        VCPU_EVENT(vcpu, 5, "enabled wait via clock comparator: %llx ns", 
sltime);


Patches currently in stable-queue which might be from [email protected] 
are

queue-3.0/s390-time-fix-sched_clock-overflow.patch
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to