From: Konrad Rzeszutek Wilk <[email protected]>
xen_init_spinlocks() currently calls static_key_slow_inc() before
jump_label_init() is invoked. When CONFIG_JUMP_LABEL is set (which usually is
the case) the effect of this static_key_slow_inc() is deferred until after
jump_label_init(). This is different from when CONFIG_JUMP_LABEL is not set, in
which case the key is set immediately. Thus, depending on the value of config
option, we may observe different behavior.
In addition, when we come to __jump_label_transform() from jump_label_init(),
the key (paravirt_ticketlocks_enabled) is already enabled. On processors where
ideal_nop is not the same as default_nop this will cause a BUG() since it is
expected that before a key is enabled the latter is replaced by the former
during initialization.
To address this problem we need to move
static_key_slow_inc(¶virt_ticketlocks_enabled) so that it is called
after jump_label_init(). We also need to make sure that this is done before
other cpus start to boot. early_initcall appears to be a good place to do so.
(Note that we cannot move whole xen_init_spinlocks() there since pv_lock_ops
need to be set before alternative_instructions() runs.)
Signed-off-by: Konrad Rzeszutek Wilk <[email protected]>
Signed-off-by: Boris Ostrovsky <[email protected]>
---
arch/x86/xen/spinlock.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/arch/x86/xen/spinlock.c b/arch/x86/xen/spinlock.c
index 253f63f..d90628d 100644
--- a/arch/x86/xen/spinlock.c
+++ b/arch/x86/xen/spinlock.c
@@ -267,11 +267,18 @@ void __init xen_init_spinlocks(void)
return;
}
- static_key_slow_inc(¶virt_ticketlocks_enabled);
-
pv_lock_ops.lock_spinning = PV_CALLEE_SAVE(xen_lock_spinning);
pv_lock_ops.unlock_kick = xen_unlock_kick;
}
+static __init int xen_init_spinlocks_jump(void)
+{
+ if (!xen_pvspin)
+ return 0;
+
+ static_key_slow_inc(¶virt_ticketlocks_enabled);
+ return 0;
+}
+early_initcall(xen_init_spinlocks_jump);
static __init int xen_parse_nopvspin(char *arg)
{
--
1.8.1.4
On Thu, 12 Sep 2013 22:29:44 -0400
Boris Ostrovsky <[email protected]> wrote:
> From: Konrad Rzeszutek Wilk <[email protected]>
>
> xen_init_spinlocks() currently calls static_key_slow_inc() before
> jump_label_init() is invoked. When CONFIG_JUMP_LABEL is set (which usually is
> the case) the effect of this static_key_slow_inc() is deferred until after
> jump_label_init(). This is different from when CONFIG_JUMP_LABEL is not set, in
> which case the key is set immediately. Thus, depending on the value of config
> option, we may observe different behavior.
>
> In addition, when we come to __jump_label_transform() from jump_label_init(),
> the key (paravirt_ticketlocks_enabled) is already enabled. On processors where
> ideal_nop is not the same as default_nop this will cause a BUG() since it is
> expected that before a key is enabled the latter is replaced by the former
> during initialization.
>
> To address this problem we need to move
> static_key_slow_inc(¶virt_ticketlocks_enabled) so that it is called
> after jump_label_init(). We also need to make sure that this is done before
> other cpus start to boot. early_initcall appears to be a good place to do so.
> (Note that we cannot move whole xen_init_spinlocks() there since pv_lock_ops
> need to be set before alternative_instructions() runs.)
>
Reviewed-by: Steven Rostedt <[email protected]>
Peter,
This fixes a regression in 3.12 against xen. Please pull and push to
Linus sometime soon.
Thanks,
-- Steve
> Signed-off-by: Konrad Rzeszutek Wilk <[email protected]>
> Signed-off-by: Boris Ostrovsky <[email protected]>
> ---
> arch/x86/xen/spinlock.c | 11 +++++++++--
> 1 file changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/xen/spinlock.c b/arch/x86/xen/spinlock.c
> index 253f63f..d90628d 100644
> --- a/arch/x86/xen/spinlock.c
> +++ b/arch/x86/xen/spinlock.c
> @@ -267,11 +267,18 @@ void __init xen_init_spinlocks(void)
> return;
> }
>
> - static_key_slow_inc(¶virt_ticketlocks_enabled);
> -
> pv_lock_ops.lock_spinning = PV_CALLEE_SAVE(xen_lock_spinning);
> pv_lock_ops.unlock_kick = xen_unlock_kick;
> }
> +static __init int xen_init_spinlocks_jump(void)
> +{
> + if (!xen_pvspin)
> + return 0;
> +
> + static_key_slow_inc(¶virt_ticketlocks_enabled);
> + return 0;
> +}
> +early_initcall(xen_init_spinlocks_jump);
>
> static __init int xen_parse_nopvspin(char *arg)
> {
On 09/12/2013 10:47 PM, Steven Rostedt wrote:
> On Thu, 12 Sep 2013 22:29:44 -0400
> Boris Ostrovsky <[email protected]> wrote:
>
>> From: Konrad Rzeszutek Wilk <[email protected]>
>>
>> xen_init_spinlocks() currently calls static_key_slow_inc() before
>> jump_label_init() is invoked. When CONFIG_JUMP_LABEL is set (which usually is
>> the case) the effect of this static_key_slow_inc() is deferred until after
>> jump_label_init(). This is different from when CONFIG_JUMP_LABEL is not set, in
>> which case the key is set immediately. Thus, depending on the value of config
>> option, we may observe different behavior.
>>
>> In addition, when we come to __jump_label_transform() from jump_label_init(),
>> the key (paravirt_ticketlocks_enabled) is already enabled. On processors where
>> ideal_nop is not the same as default_nop this will cause a BUG() since it is
>> expected that before a key is enabled the latter is replaced by the former
>> during initialization.
>>
>> To address this problem we need to move
>> static_key_slow_inc(¶virt_ticketlocks_enabled) so that it is called
>> after jump_label_init(). We also need to make sure that this is done before
>> other cpus start to boot. early_initcall appears to be a good place to do so.
>> (Note that we cannot move whole xen_init_spinlocks() there since pv_lock_ops
>> need to be set before alternative_instructions() runs.)
>>
> Reviewed-by: Steven Rostedt <[email protected]>
>
> Peter,
>
> This fixes a regression in 3.12 against xen. Please pull and push to
> Linus sometime soon.
Peter,
Are you planning on taking this patch?
Thanks.
-boris
>
> Thanks,
>
> -- Steve
>
>> Signed-off-by: Konrad Rzeszutek Wilk <[email protected]>
>> Signed-off-by: Boris Ostrovsky <[email protected]>
>> ---
>> arch/x86/xen/spinlock.c | 11 +++++++++--
>> 1 file changed, 9 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/x86/xen/spinlock.c b/arch/x86/xen/spinlock.c
>> index 253f63f..d90628d 100644
>> --- a/arch/x86/xen/spinlock.c
>> +++ b/arch/x86/xen/spinlock.c
>> @@ -267,11 +267,18 @@ void __init xen_init_spinlocks(void)
>> return;
>> }
>>
>> - static_key_slow_inc(¶virt_ticketlocks_enabled);
>> -
>> pv_lock_ops.lock_spinning = PV_CALLEE_SAVE(xen_lock_spinning);
>> pv_lock_ops.unlock_kick = xen_unlock_kick;
>> }
>> +static __init int xen_init_spinlocks_jump(void)
>> +{
>> + if (!xen_pvspin)
>> + return 0;
>> +
>> + static_key_slow_inc(¶virt_ticketlocks_enabled);
>> + return 0;
>> +}
>> +early_initcall(xen_init_spinlocks_jump);
>>
>> static __init int xen_parse_nopvspin(char *arg)
>> {
On Tue, Sep 24, 2013 at 12:50:45PM -0700, Boris Ostrovsky wrote:
> On 09/12/2013 10:47 PM, Steven Rostedt wrote:
> >On Thu, 12 Sep 2013 22:29:44 -0400
> >Boris Ostrovsky <[email protected]> wrote:
> >
> >>From: Konrad Rzeszutek Wilk <[email protected]>
> >>
> >>xen_init_spinlocks() currently calls static_key_slow_inc() before
> >>jump_label_init() is invoked. When CONFIG_JUMP_LABEL is set (which usually is
> >>the case) the effect of this static_key_slow_inc() is deferred until after
> >>jump_label_init(). This is different from when CONFIG_JUMP_LABEL is not set, in
> >>which case the key is set immediately. Thus, depending on the value of config
> >>option, we may observe different behavior.
> >>
> >>In addition, when we come to __jump_label_transform() from jump_label_init(),
> >>the key (paravirt_ticketlocks_enabled) is already enabled. On processors where
> >>ideal_nop is not the same as default_nop this will cause a BUG() since it is
> >>expected that before a key is enabled the latter is replaced by the former
> >>during initialization.
> >>
> >>To address this problem we need to move
> >>static_key_slow_inc(¶virt_ticketlocks_enabled) so that it is called
> >>after jump_label_init(). We also need to make sure that this is done before
> >>other cpus start to boot. early_initcall appears to be a good place to do so.
> >>(Note that we cannot move whole xen_init_spinlocks() there since pv_lock_ops
> >>need to be set before alternative_instructions() runs.)
> >>
> >Reviewed-by: Steven Rostedt <[email protected]>
> >
> >Peter,
> >
> >This fixes a regression in 3.12 against xen. Please pull and push to
> >Linus sometime soon.
>
> Peter,
>
> Are you planning on taking this patch?
If he hasn't done yet (and I don't see a tip-bot email so I think the answer
is no) I can do it (as I am back from my mini vacation).
Thanks for the ping!