[PATCH net-next] xfrm: Make xfrm_state timer monotonic

[PATCH net-next] xfrm: Make xfrm_state timer monotonic

[Fan Du]

xfrm_state timer should be independent of system clock change,
so switch to monotonic clock base.

Signed-off-by: Fan Du <fan.du@windriver.com>
---
 net/xfrm/xfrm_state.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index 78f66fa..e57ab07 100644
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -499,7 +499,7 @@ struct xfrm_state *xfrm_state_alloc(struct net *net)
 		INIT_HLIST_NODE(&x->bydst);
 		INIT_HLIST_NODE(&x->bysrc);
 		INIT_HLIST_NODE(&x->byspi);
-		tasklet_hrtimer_init(&x->mtimer, xfrm_timer_handler, CLOCK_REALTIME, HRTIMER_MODE_ABS);
+		tasklet_hrtimer_init(&x->mtimer, xfrm_timer_handler, CLOCK_MONOTONIC, HRTIMER_MODE_ABS);
 		setup_timer(&x->rtimer, xfrm_replay_timer_handler,
 				(unsigned long)x);
 		x->curlft.add_time = get_seconds();
-- 
1.7.9.5

Re: [PATCH net-next] xfrm: Make xfrm_state timer monotonic

[Steffen Klassert]

On Tue, Aug 06, 2013 at 02:57:05PM +0800, Fan Du wrote:
> xfrm_state timer should be independent of system clock change,
> so switch to monotonic clock base.
> 

I think a monotonic timer will reintroduce a bug on suspend/resume
that was fixed by commit 9e0d57fd6
(xfrm: SAD entries do not expire correctly after suspend-resume)

Please make sure that this does not happen.

[PATCHv2 net-next] xfrm: Make xfrm_state timer monotonic

[Fan Du]

On 2013年08月06日 17:22, Steffen Klassert wrote:
> On Tue, Aug 06, 2013 at 02:57:05PM +0800, Fan Du wrote:
>> xfrm_state timer should be independent of system clock change,
>> so switch to monotonic clock base.
>>
>
> I think a monotonic timer will reintroduce a bug on suspend/resume
> that was fixed by commit 9e0d57fd6
> (xfrm: SAD entries do not expire correctly after suspend-resume)
>
> Please make sure that this does not happen.

What about using CLOCK_BOOTTIME? it's monotonic, but includes suspend time as well.


 From 1dae5a22c203224ee46e2a2f95b569a04b74117b Mon Sep 17 00:00:00 2001
From: Fan Du <fan.du@windriver.com>
Date: Mon, 12 Aug 2013 13:33:45 +0800
Subject: [PATCHv2 net-next] xfrm: Make xfrm_state timer monotonic

xfrm_state timer should be independent of system clock change,
so switch to CLOCK_BOOTTIME base which is not only monotonic but
also counting suspend time.

Signed-off-by: Fan Du <fan.du@windriver.com>
---
  net/xfrm/xfrm_state.c |    2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index 78f66fa..1df4436 100644
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -499,7 +499,7 @@ struct xfrm_state *xfrm_state_alloc(struct net *net)
  		INIT_HLIST_NODE(&x->bydst);
  		INIT_HLIST_NODE(&x->bysrc);
  		INIT_HLIST_NODE(&x->byspi);
-		tasklet_hrtimer_init(&x->mtimer, xfrm_timer_handler, CLOCK_REALTIME, HRTIMER_MODE_ABS);
+		tasklet_hrtimer_init(&x->mtimer, xfrm_timer_handler, CLOCK_BOOTTIME, HRTIMER_MODE_ABS);
  		setup_timer(&x->rtimer, xfrm_replay_timer_handler,
  				(unsigned long)x);
  		x->curlft.add_time = get_seconds();
-- 
1.7.9.5




> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

-- 
浮沉随浪只记今朝笑

--fan

Re: [PATCHv2 net-next] xfrm: Make xfrm_state timer monotonic

[Steffen Klassert]

On Mon, Aug 12, 2013 at 01:40:07PM +0800, Fan Du wrote:
> 
> 
> On 2013年08月06日 17:22, Steffen Klassert wrote:
> >On Tue, Aug 06, 2013 at 02:57:05PM +0800, Fan Du wrote:
> >>xfrm_state timer should be independent of system clock change,
> >>so switch to monotonic clock base.
> >>
> >
> >I think a monotonic timer will reintroduce a bug on suspend/resume
> >that was fixed by commit 9e0d57fd6
> >(xfrm: SAD entries do not expire correctly after suspend-resume)
> >
> >Please make sure that this does not happen.
> 
> What about using CLOCK_BOOTTIME? it's monotonic, but includes suspend time as well.

As I said, I'm open to everything that fixes your problem and does not
introduce a regression. I'll consider applying after some testing
if noone else has objections.

Thanks!

Re: [PATCHv2 net-next] xfrm: Make xfrm_state timer monotonic

[Fan Du]

On 2013年08月13日 15:55, Steffen Klassert wrote:
> On Mon, Aug 12, 2013 at 01:40:07PM +0800, Fan Du wrote:
>>
>>
>> On 2013年08月06日 17:22, Steffen Klassert wrote:
>>> On Tue, Aug 06, 2013 at 02:57:05PM +0800, Fan Du wrote:
>>>> xfrm_state timer should be independent of system clock change,
>>>> so switch to monotonic clock base.
>>>>
>>>
>>> I think a monotonic timer will reintroduce a bug on suspend/resume
>>> that was fixed by commit 9e0d57fd6
>>> (xfrm: SAD entries do not expire correctly after suspend-resume)
>>>
>>> Please make sure that this does not happen.
>>
>> What about using CLOCK_BOOTTIME? it's monotonic, but includes suspend time as well.
>
> As I said, I'm open to everything that fixes your problem and does not
> introduce a regression. I'll consider applying after some testing
> if noone else has objections.

Hi, Steffen

Thanks for your understanding! :)

I happened to bump into CLOCK_BOOTTIME several days ago, so apologize for
eating my words earlier. Changing xfrm_state timer to monotonic does not
solve the problem I've described earlier in:

http://www.spinics.net/lists/netdev/msg245019.html (*1*)

So is there any light of hope for the proposal in (*1*) by using CLOCK_BOOTTIME
instead?


> Thanks!
>
>

-- 
浮沉随浪只记今朝笑

--fan

Re: [PATCHv2 net-next] xfrm: Make xfrm_state timer monotonic

[Steffen Klassert]

On Tue, Aug 13, 2013 at 04:37:04PM +0800, Fan Du wrote:
> 
> 
> On 2013年08月13日 15:55, Steffen Klassert wrote:
> >On Mon, Aug 12, 2013 at 01:40:07PM +0800, Fan Du wrote:
> >>
> >>
> >>On 2013年08月06日 17:22, Steffen Klassert wrote:
> >>>On Tue, Aug 06, 2013 at 02:57:05PM +0800, Fan Du wrote:
> >>>>xfrm_state timer should be independent of system clock change,
> >>>>so switch to monotonic clock base.
> >>>>
> >>>
> >>>I think a monotonic timer will reintroduce a bug on suspend/resume
> >>>that was fixed by commit 9e0d57fd6
> >>>(xfrm: SAD entries do not expire correctly after suspend-resume)
> >>>
> >>>Please make sure that this does not happen.
> >>
> >>What about using CLOCK_BOOTTIME? it's monotonic, but includes suspend time as well.
> >
> >As I said, I'm open to everything that fixes your problem and does not
> >introduce a regression. I'll consider applying after some testing
> >if noone else has objections.
> 
> Hi, Steffen
> 
> Thanks for your understanding! :)
> 
> I happened to bump into CLOCK_BOOTTIME several days ago, so apologize for
> eating my words earlier. Changing xfrm_state timer to monotonic does not
> solve the problem I've described earlier in:
> 
> http://www.spinics.net/lists/netdev/msg245019.html (*1*)
> 
> So is there any light of hope for the proposal in (*1*) by using CLOCK_BOOTTIME
> instead?
> 

Well, you have to convince David in the first place, as he is sitting
upstream from me ;)

Also, I think he is right with his complaint. Why not following the
clock_was_set approach you started with?

Re: [PATCHv2 net-next] xfrm: Make xfrm_state timer monotonic

[Steffen Klassert]

On Mon, Aug 12, 2013 at 01:40:07PM +0800, Fan Du wrote:
> 
> 
> On 2013年08月06日 17:22, Steffen Klassert wrote:
> >On Tue, Aug 06, 2013 at 02:57:05PM +0800, Fan Du wrote:
> >>xfrm_state timer should be independent of system clock change,
> >>so switch to monotonic clock base.
> >>
> >
> >I think a monotonic timer will reintroduce a bug on suspend/resume
> >that was fixed by commit 9e0d57fd6
> >(xfrm: SAD entries do not expire correctly after suspend-resume)
> >
> >Please make sure that this does not happen.
> 
> What about using CLOCK_BOOTTIME? it's monotonic, but includes suspend time as well.
> 
> 
> From 1dae5a22c203224ee46e2a2f95b569a04b74117b Mon Sep 17 00:00:00 2001
> From: Fan Du <fan.du@windriver.com>
> Date: Mon, 12 Aug 2013 13:33:45 +0800
> Subject: [PATCHv2 net-next] xfrm: Make xfrm_state timer monotonic
> 
> xfrm_state timer should be independent of system clock change,
> so switch to CLOCK_BOOTTIME base which is not only monotonic but
> also counting suspend time.
> 
> Signed-off-by: Fan Du <fan.du@windriver.com>
> ---
>  net/xfrm/xfrm_state.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
> index 78f66fa..1df4436 100644
> --- a/net/xfrm/xfrm_state.c
> +++ b/net/xfrm/xfrm_state.c
> @@ -499,7 +499,7 @@ struct xfrm_state *xfrm_state_alloc(struct net *net)
>  		INIT_HLIST_NODE(&x->bydst);
>  		INIT_HLIST_NODE(&x->bysrc);
>  		INIT_HLIST_NODE(&x->byspi);
> -		tasklet_hrtimer_init(&x->mtimer, xfrm_timer_handler, CLOCK_REALTIME, HRTIMER_MODE_ABS);
> +		tasklet_hrtimer_init(&x->mtimer, xfrm_timer_handler, CLOCK_BOOTTIME, HRTIMER_MODE_ABS);
>  		setup_timer(&x->rtimer, xfrm_replay_timer_handler,
>  				(unsigned long)x);
>  		x->curlft.add_time = get_seconds();

This does not apply to my tree, looks like your patch is whitespace damaged.

[PATCHv2 net-next] xfrm: Make xfrm_state timer monotonic

[Fan Du]

xfrm_state timer should be independent of system clock change,
so switch to CLOCK_BOOTTIME base which is not only monotonic but
also counting suspend time.

Thus issue reported in commit: 9e0d57fd6dad37d72a3ca6db00ca8c76f2215454
("xfrm: SAD entries do not expire correctly after suspend-resume")
could ALSO be avoided.

Signed-off-by: Fan Du <fan.du@windriver.com>

v2: Use CLOCK_BOOTTIME to count suspend time, but still monotonic.

---
 net/xfrm/xfrm_state.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index 79bf9a0..0824679 100644
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -486,7 +486,8 @@ struct xfrm_state *xfrm_state_alloc(struct net *net)
 		INIT_HLIST_NODE(&x->bydst);
 		INIT_HLIST_NODE(&x->bysrc);
 		INIT_HLIST_NODE(&x->byspi);
-		tasklet_hrtimer_init(&x->mtimer, xfrm_timer_handler, CLOCK_REALTIME, HRTIMER_MODE_ABS);
+		tasklet_hrtimer_init(&x->mtimer, xfrm_timer_handler,
+					CLOCK_BOOTTIME, HRTIMER_MODE_ABS);
 		setup_timer(&x->rtimer, xfrm_replay_timer_handler,
 				(unsigned long)x);
 		x->curlft.add_time = get_seconds();
-- 
1.7.9.5

Re: [PATCHv2 net-next] xfrm: Make xfrm_state timer monotonic

[David Miller]

From: Fan Du <fan.du@windriver.com>
Date: Thu, 15 Aug 2013 15:49:06 +0800

> xfrm_state timer should be independent of system clock change,
> so switch to CLOCK_BOOTTIME base which is not only monotonic but
> also counting suspend time.
> 
> Thus issue reported in commit: 9e0d57fd6dad37d72a3ca6db00ca8c76f2215454
> ("xfrm: SAD entries do not expire correctly after suspend-resume")
> could ALSO be avoided.
> 
> Signed-off-by: Fan Du <fan.du@windriver.com>
> 
> v2: Use CLOCK_BOOTTIME to count suspend time, but still monotonic.

This seems like a good fix:

Acked-by: David S. Miller <davem@davemloft.net>

Re: [PATCHv2 net-next] xfrm: Make xfrm_state timer monotonic

[Steffen Klassert]

On Thu, Aug 15, 2013 at 02:23:34PM -0700, David Miller wrote:
> From: Fan Du <fan.du@windriver.com>
> Date: Thu, 15 Aug 2013 15:49:06 +0800
> 
> > xfrm_state timer should be independent of system clock change,
> > so switch to CLOCK_BOOTTIME base which is not only monotonic but
> > also counting suspend time.
> > 
> > Thus issue reported in commit: 9e0d57fd6dad37d72a3ca6db00ca8c76f2215454
> > ("xfrm: SAD entries do not expire correctly after suspend-resume")
> > could ALSO be avoided.
> > 
> > Signed-off-by: Fan Du <fan.du@windriver.com>
> > 
> > v2: Use CLOCK_BOOTTIME to count suspend time, but still monotonic.
> 
> This seems like a good fix:
> 
> Acked-by: David S. Miller <davem@davemloft.net>

Applied to ipsec-next, thanks!