[PATCH-next v2 0/2] ipvs: Fix incorrect use of HK_TYPE_KTHREAD housekeeping cpumask

[PATCH-next v2 0/2] ipvs: Fix incorrect use of HK_TYPE_KTHREAD housekeeping cpumask

[Waiman Long]

 v2:
  - Rebased on top of linux-next

Since commit 041ee6f3727a ("kthread: Rely on HK_TYPE_DOMAIN for preferred
affinity management"), the HK_TYPE_KTHREAD housekeeping cpumask may no
longer be correct in showing the actual CPU affinity of kthreads that
have no predefined CPU affinity. As the ipvs networking code is still
using HK_TYPE_KTHREAD, we need to make HK_TYPE_KTHREAD reflect the
reality.

This patch series makes HK_TYPE_KTHREAD an alias of HK_TYPE_DOMAIN
and uses RCU to protect access to the HK_TYPE_KTHREAD housekeeping
cpumask.

Waiman Long (2):
  sched/isolation: Make HK_TYPE_KTHREAD an alias of HK_TYPE_DOMAIN
  ipvs: Guard access of HK_TYPE_KTHREAD cpumask with RCU

 include/linux/sched/isolation.h |  6 +++++-
 include/net/ip_vs.h             | 20 ++++++++++++++++----
 net/netfilter/ipvs/ip_vs_ctl.c  | 13 ++++++++-----
 3 files changed, 29 insertions(+), 10 deletions(-)

-- 
2.53.0

[PATCH-next v2 1/2] sched/isolation: Make HK_TYPE_KTHREAD an alias of HK_TYPE_DOMAIN

[Waiman Long]

Since commit 041ee6f3727a ("kthread: Rely on HK_TYPE_DOMAIN for preferred
affinity management"), kthreads default to use the HK_TYPE_DOMAIN
cpumask. IOW, it is no longer affected by the setting of the nohz_full
boot kernel parameter.

That means HK_TYPE_KTHREAD should now be an alias of HK_TYPE_DOMAIN
instead of HK_TYPE_KERNEL_NOISE to correctly reflect the current kthread
behavior. Make the change as HK_TYPE_KTHREAD is still being used in
some networking code.

Fixes: 041ee6f3727a ("kthread: Rely on HK_TYPE_DOMAIN for preferred affinity management")
Signed-off-by: Waiman Long <longman@redhat.com>
---
 include/linux/sched/isolation.h | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/include/linux/sched/isolation.h b/include/linux/sched/isolation.h
index dc3975ff1b2e..cf0fd03dd7a2 100644
--- a/include/linux/sched/isolation.h
+++ b/include/linux/sched/isolation.h
@@ -20,6 +20,11 @@ enum hk_type {
 	HK_TYPE_KERNEL_NOISE,
 	HK_TYPE_MAX,
 
+	/*
+	 * HK_TYPE_KTHREAD is now an alias of HK_TYPE_DOMAIN
+	 */
+	HK_TYPE_KTHREAD = HK_TYPE_DOMAIN,
+
 	/*
 	 * The following housekeeping types are only set by the nohz_full
 	 * boot commandline option. So they can share the same value.
@@ -29,7 +34,6 @@ enum hk_type {
 	HK_TYPE_RCU     = HK_TYPE_KERNEL_NOISE,
 	HK_TYPE_MISC    = HK_TYPE_KERNEL_NOISE,
 	HK_TYPE_WQ      = HK_TYPE_KERNEL_NOISE,
-	HK_TYPE_KTHREAD = HK_TYPE_KERNEL_NOISE
 };
 
 #ifdef CONFIG_CPU_ISOLATION
-- 
2.53.0

[PATCH-next v2 2/2] ipvs: Guard access of HK_TYPE_KTHREAD cpumask with RCU

[Waiman Long]

The ip_vs_ctl.c file and the associated ip_vs.h file are the only places
in the kernel where HK_TYPE_KTHREAD cpumask is being retrieved and used.
Now that HK_TYPE_KTHREAD/HK_TYPE_DOMAIN cpumask can be changed at run
time. We need to use RCU to guard access to this cpumask to avoid a
potential UAF problem as the returned cpumask may be freed before it
is being used.

We can replace HK_TYPE_KTHREAD by HK_TYPE_DOMAIN as they are aliases
of each other, but keeping the HK_TYPE_KTHREAD name can highlight the
fact that it is the kthread initiated by ipvs that is being controlled.

Signed-off-by: Waiman Long <longman@redhat.com>
---
 include/net/ip_vs.h            | 20 ++++++++++++++++----
 net/netfilter/ipvs/ip_vs_ctl.c | 13 ++++++++-----
 2 files changed, 24 insertions(+), 9 deletions(-)

diff --git a/include/net/ip_vs.h b/include/net/ip_vs.h
index 72d325c81313..7bda92fd3fe6 100644
--- a/include/net/ip_vs.h
+++ b/include/net/ip_vs.h
@@ -1411,7 +1411,7 @@ static inline int sysctl_run_estimation(struct netns_ipvs *ipvs)
 	return ipvs->sysctl_run_estimation;
 }
 
-static inline const struct cpumask *sysctl_est_cpulist(struct netns_ipvs *ipvs)
+static inline const struct cpumask *__sysctl_est_cpulist(struct netns_ipvs *ipvs)
 {
 	if (ipvs->est_cpulist_valid)
 		return ipvs->sysctl_est_cpulist;
@@ -1529,7 +1529,7 @@ static inline int sysctl_run_estimation(struct netns_ipvs *ipvs)
 	return 1;
 }
 
-static inline const struct cpumask *sysctl_est_cpulist(struct netns_ipvs *ipvs)
+static inline const struct cpumask *__sysctl_est_cpulist(struct netns_ipvs *ipvs)
 {
 	return housekeeping_cpumask(HK_TYPE_KTHREAD);
 }
@@ -1564,6 +1564,18 @@ static inline int sysctl_svc_lfactor(struct netns_ipvs *ipvs)
 	return READ_ONCE(ipvs->sysctl_svc_lfactor);
 }
 
+static inline bool sysctl_est_cpulist_empty(struct netns_ipvs *ipvs)
+{
+	guard(rcu)();
+	return cpumask_empty(__sysctl_est_cpulist(ipvs));
+}
+
+static inline unsigned int sysctl_est_cpulist_weight(struct netns_ipvs *ipvs)
+{
+	guard(rcu)();
+	return cpumask_weight(__sysctl_est_cpulist(ipvs));
+}
+
 /* IPVS core functions
  * (from ip_vs_core.c)
  */
@@ -1895,7 +1907,7 @@ static inline void ip_vs_est_stopped_recalc(struct netns_ipvs *ipvs)
 	/* Stop tasks while cpulist is empty or if disabled with flag */
 	ipvs->est_stopped = !sysctl_run_estimation(ipvs) ||
 			    (ipvs->est_cpulist_valid &&
-			     cpumask_empty(sysctl_est_cpulist(ipvs)));
+			     sysctl_est_cpulist_empty(ipvs));
 #endif
 }
 
@@ -1911,7 +1923,7 @@ static inline bool ip_vs_est_stopped(struct netns_ipvs *ipvs)
 static inline int ip_vs_est_max_threads(struct netns_ipvs *ipvs)
 {
 	unsigned int limit = IPVS_EST_CPU_KTHREADS *
-			     cpumask_weight(sysctl_est_cpulist(ipvs));
+			     sysctl_est_cpulist_weight(ipvs);
 
 	return max(1U, limit);
 }
diff --git a/net/netfilter/ipvs/ip_vs_ctl.c b/net/netfilter/ipvs/ip_vs_ctl.c
index 032425025d88..e253a1ceef48 100644
--- a/net/netfilter/ipvs/ip_vs_ctl.c
+++ b/net/netfilter/ipvs/ip_vs_ctl.c
@@ -2338,11 +2338,14 @@ static int ipvs_proc_est_cpumask_get(const struct ctl_table *table,
 
 	mutex_lock(&ipvs->est_mutex);
 
-	if (ipvs->est_cpulist_valid)
-		mask = *valp;
-	else
-		mask = (struct cpumask *)housekeeping_cpumask(HK_TYPE_KTHREAD);
-	ret = scnprintf(buffer, size, "%*pbl\n", cpumask_pr_args(mask));
+	/* HK_TYPE_KTHREAD cpumask needs RCU protection */
+	scoped_guard(rcu) {
+		if (ipvs->est_cpulist_valid)
+			mask = *valp;
+		else
+			mask = (struct cpumask *)housekeeping_cpumask(HK_TYPE_KTHREAD);
+		ret = scnprintf(buffer, size, "%*pbl\n", cpumask_pr_args(mask));
+	}
 
 	mutex_unlock(&ipvs->est_mutex);
 
-- 
2.53.0

Re: [PATCH-next v2 1/2] sched/isolation: Make HK_TYPE_KTHREAD an alias of HK_TYPE_DOMAIN

[Frederic Weisbecker]

Le Tue, Mar 31, 2026 at 12:50:14PM -0400, Waiman Long a écrit :
> Since commit 041ee6f3727a ("kthread: Rely on HK_TYPE_DOMAIN for preferred
> affinity management"), kthreads default to use the HK_TYPE_DOMAIN
> cpumask. IOW, it is no longer affected by the setting of the nohz_full
> boot kernel parameter.
> 
> That means HK_TYPE_KTHREAD should now be an alias of HK_TYPE_DOMAIN
> instead of HK_TYPE_KERNEL_NOISE to correctly reflect the current kthread
> behavior. Make the change as HK_TYPE_KTHREAD is still being used in
> some networking code.
> 
> Fixes: 041ee6f3727a ("kthread: Rely on HK_TYPE_DOMAIN for preferred affinity management")
> Signed-off-by: Waiman Long <longman@redhat.com>

This makes ipvs_proc_est_cpumask_get() racy because now without RCU locked the
mask pointer can be released at any point.

Other users:

sysctl_est_cpulist() -> ip_vs_est_stopped_recalc() -> ip_vs_est_reload_start()

Here sysctl_est_cpulist() is only invoked if ->est_cpulist_valid
(->est_mutex makes it stable). So housekeeping_cpumask() should not be called.

But ip_vs_est_max_threads() is more complicated. And it's a sign we should
probably call something like ipvs_proc_est_cpumask_set() when the HK_TYPE_DOMAIN
is modified (and ipvs->est_cpulist_valid is 0) in order to update the ipvs
kthreads accordingly.

Thanks.

-- 
Frederic Weisbecker
SUSE Labs

Re: [PATCH-next v2 2/2] ipvs: Guard access of HK_TYPE_KTHREAD cpumask with RCU

[Frederic Weisbecker]

Le Tue, Mar 31, 2026 at 12:50:15PM -0400, Waiman Long a écrit :
> The ip_vs_ctl.c file and the associated ip_vs.h file are the only places
> in the kernel where HK_TYPE_KTHREAD cpumask is being retrieved and used.
> Now that HK_TYPE_KTHREAD/HK_TYPE_DOMAIN cpumask can be changed at run
> time. We need to use RCU to guard access to this cpumask to avoid a
> potential UAF problem as the returned cpumask may be freed before it
> is being used.
> 
> We can replace HK_TYPE_KTHREAD by HK_TYPE_DOMAIN as they are aliases
> of each other, but keeping the HK_TYPE_KTHREAD name can highlight the
> fact that it is the kthread initiated by ipvs that is being controlled.
> 
> Signed-off-by: Waiman Long <longman@redhat.com>

Oh I see you're handling a few concerns here. But it's too late, the previous
patch broke bisection.

> ---
>  include/net/ip_vs.h            | 20 ++++++++++++++++----
>  net/netfilter/ipvs/ip_vs_ctl.c | 13 ++++++++-----
>  2 files changed, 24 insertions(+), 9 deletions(-)
> 
> diff --git a/include/net/ip_vs.h b/include/net/ip_vs.h
> index 72d325c81313..7bda92fd3fe6 100644
> --- a/include/net/ip_vs.h
> +++ b/include/net/ip_vs.h
> @@ -1411,7 +1411,7 @@ static inline int sysctl_run_estimation(struct netns_ipvs *ipvs)
>  	return ipvs->sysctl_run_estimation;
>  }
>  
> -static inline const struct cpumask *sysctl_est_cpulist(struct netns_ipvs *ipvs)
> +static inline const struct cpumask *__sysctl_est_cpulist(struct netns_ipvs *ipvs)
>  {
>  	if (ipvs->est_cpulist_valid)
>  		return ipvs->sysctl_est_cpulist;
> @@ -1529,7 +1529,7 @@ static inline int sysctl_run_estimation(struct netns_ipvs *ipvs)
>  	return 1;
>  }
>  
> -static inline const struct cpumask *sysctl_est_cpulist(struct netns_ipvs *ipvs)
> +static inline const struct cpumask *__sysctl_est_cpulist(struct netns_ipvs *ipvs)
>  {
>  	return housekeeping_cpumask(HK_TYPE_KTHREAD);
>  }
> @@ -1564,6 +1564,18 @@ static inline int sysctl_svc_lfactor(struct netns_ipvs *ipvs)
>  	return READ_ONCE(ipvs->sysctl_svc_lfactor);
>  }
>  
> +static inline bool sysctl_est_cpulist_empty(struct netns_ipvs *ipvs)
> +{
> +	guard(rcu)();
> +	return cpumask_empty(__sysctl_est_cpulist(ipvs));
> +}
> +
> +static inline unsigned int sysctl_est_cpulist_weight(struct netns_ipvs *ipvs)
> +{
> +	guard(rcu)();
> +	return cpumask_weight(__sysctl_est_cpulist(ipvs));
> +}
> +
>  /* IPVS core functions
>   * (from ip_vs_core.c)
>   */
> @@ -1895,7 +1907,7 @@ static inline void ip_vs_est_stopped_recalc(struct netns_ipvs *ipvs)
>  	/* Stop tasks while cpulist is empty or if disabled with flag */
>  	ipvs->est_stopped = !sysctl_run_estimation(ipvs) ||
>  			    (ipvs->est_cpulist_valid &&
> -			     cpumask_empty(sysctl_est_cpulist(ipvs)));
> +			     sysctl_est_cpulist_empty(ipvs));

It's not needed, if ipvs->est_cpulist_valid, sysctl_est_cpulist() doesn't
refer to housekeeping.

>  #endif
>  }
>  
> @@ -1911,7 +1923,7 @@ static inline bool ip_vs_est_stopped(struct netns_ipvs *ipvs)
>  static inline int ip_vs_est_max_threads(struct netns_ipvs *ipvs)
>  {
>  	unsigned int limit = IPVS_EST_CPU_KTHREADS *
> -			     cpumask_weight(sysctl_est_cpulist(ipvs));
> +			     sysctl_est_cpulist_weight(ipvs);

That probably works for callers ip_vs_start_estimator().

But this is not handling the core issue that related kthreads should be updated,
as is done in ipvs_proc_est_cpumask_set(), when HK_TYPE_DOMAIN mask changes.

>  
>  	return max(1U, limit);
>  }
> diff --git a/net/netfilter/ipvs/ip_vs_ctl.c b/net/netfilter/ipvs/ip_vs_ctl.c
> index 032425025d88..e253a1ceef48 100644
> --- a/net/netfilter/ipvs/ip_vs_ctl.c
> +++ b/net/netfilter/ipvs/ip_vs_ctl.c
> @@ -2338,11 +2338,14 @@ static int ipvs_proc_est_cpumask_get(const struct ctl_table *table,
>  
>  	mutex_lock(&ipvs->est_mutex);
>  
> -	if (ipvs->est_cpulist_valid)
> -		mask = *valp;
> -	else
> -		mask = (struct cpumask *)housekeeping_cpumask(HK_TYPE_KTHREAD);
> -	ret = scnprintf(buffer, size, "%*pbl\n", cpumask_pr_args(mask));
> +	/* HK_TYPE_KTHREAD cpumask needs RCU protection */
> +	scoped_guard(rcu) {
> +		if (ipvs->est_cpulist_valid)
> +			mask = *valp;
> +		else
> +			mask = (struct cpumask *)housekeeping_cpumask(HK_TYPE_KTHREAD);
> +		ret = scnprintf(buffer, size, "%*pbl\n", cpumask_pr_args(mask));
> +	}

And that works.

Thanks.

>  
>  	mutex_unlock(&ipvs->est_mutex);
>  
> -- 
> 2.53.0
> 

-- 
Frederic Weisbecker
SUSE Labs

Re: [PATCH-next v2 2/2] ipvs: Guard access of HK_TYPE_KTHREAD cpumask with RCU

[Waiman Long]

On 4/1/26 8:54 AM, Frederic Weisbecker wrote:
> Le Tue, Mar 31, 2026 at 12:50:15PM -0400, Waiman Long a écrit :
>> The ip_vs_ctl.c file and the associated ip_vs.h file are the only places
>> in the kernel where HK_TYPE_KTHREAD cpumask is being retrieved and used.
>> Now that HK_TYPE_KTHREAD/HK_TYPE_DOMAIN cpumask can be changed at run
>> time. We need to use RCU to guard access to this cpumask to avoid a
>> potential UAF problem as the returned cpumask may be freed before it
>> is being used.
>>
>> We can replace HK_TYPE_KTHREAD by HK_TYPE_DOMAIN as they are aliases
>> of each other, but keeping the HK_TYPE_KTHREAD name can highlight the
>> fact that it is the kthread initiated by ipvs that is being controlled.
>>
>> Signed-off-by: Waiman Long <longman@redhat.com>
> Oh I see you're handling a few concerns here. But it's too late, the previous
> patch broke bisection.
Good point. So I have to either reverse the patch order, or just change 
HK_TYPE_KTHREAD to HK_TYPE_DOMAIN & drop the first one.
>
>> ---
>>   include/net/ip_vs.h            | 20 ++++++++++++++++----
>>   net/netfilter/ipvs/ip_vs_ctl.c | 13 ++++++++-----
>>   2 files changed, 24 insertions(+), 9 deletions(-)
>>
>> diff --git a/include/net/ip_vs.h b/include/net/ip_vs.h
>> index 72d325c81313..7bda92fd3fe6 100644
>> --- a/include/net/ip_vs.h
>> +++ b/include/net/ip_vs.h
>> @@ -1411,7 +1411,7 @@ static inline int sysctl_run_estimation(struct netns_ipvs *ipvs)
>>   	return ipvs->sysctl_run_estimation;
>>   }
>>   
>> -static inline const struct cpumask *sysctl_est_cpulist(struct netns_ipvs *ipvs)
>> +static inline const struct cpumask *__sysctl_est_cpulist(struct netns_ipvs *ipvs)
>>   {
>>   	if (ipvs->est_cpulist_valid)
>>   		return ipvs->sysctl_est_cpulist;
>> @@ -1529,7 +1529,7 @@ static inline int sysctl_run_estimation(struct netns_ipvs *ipvs)
>>   	return 1;
>>   }
>>   
>> -static inline const struct cpumask *sysctl_est_cpulist(struct netns_ipvs *ipvs)
>> +static inline const struct cpumask *__sysctl_est_cpulist(struct netns_ipvs *ipvs)
>>   {
>>   	return housekeeping_cpumask(HK_TYPE_KTHREAD);
>>   }
>> @@ -1564,6 +1564,18 @@ static inline int sysctl_svc_lfactor(struct netns_ipvs *ipvs)
>>   	return READ_ONCE(ipvs->sysctl_svc_lfactor);
>>   }
>>   
>> +static inline bool sysctl_est_cpulist_empty(struct netns_ipvs *ipvs)
>> +{
>> +	guard(rcu)();
>> +	return cpumask_empty(__sysctl_est_cpulist(ipvs));
>> +}
>> +
>> +static inline unsigned int sysctl_est_cpulist_weight(struct netns_ipvs *ipvs)
>> +{
>> +	guard(rcu)();
>> +	return cpumask_weight(__sysctl_est_cpulist(ipvs));
>> +}
>> +
>>   /* IPVS core functions
>>    * (from ip_vs_core.c)
>>    */
>> @@ -1895,7 +1907,7 @@ static inline void ip_vs_est_stopped_recalc(struct netns_ipvs *ipvs)
>>   	/* Stop tasks while cpulist is empty or if disabled with flag */
>>   	ipvs->est_stopped = !sysctl_run_estimation(ipvs) ||
>>   			    (ipvs->est_cpulist_valid &&
>> -			     cpumask_empty(sysctl_est_cpulist(ipvs)));
>> +			     sysctl_est_cpulist_empty(ipvs));
> It's not needed, if ipvs->est_cpulist_valid, sysctl_est_cpulist() doesn't
> refer to housekeeping.
Right.
>>   #endif
>>   }
>>   
>> @@ -1911,7 +1923,7 @@ static inline bool ip_vs_est_stopped(struct netns_ipvs *ipvs)
>>   static inline int ip_vs_est_max_threads(struct netns_ipvs *ipvs)
>>   {
>>   	unsigned int limit = IPVS_EST_CPU_KTHREADS *
>> -			     cpumask_weight(sysctl_est_cpulist(ipvs));
>> +			     sysctl_est_cpulist_weight(ipvs);
> That probably works for callers ip_vs_start_estimator().
>
> But this is not handling the core issue that related kthreads should be updated,
> as is done in ipvs_proc_est_cpumask_set(), when HK_TYPE_DOMAIN mask changes.

If ipvs_proc_est_cpumask_set() has been called, the real affinity should 
be the intersection of the given cpumask and HK_TYPE_DOMAIN cpumask. Is 
that what you are referring to?

Cheers,
Longman

>
>>   
>>   	return max(1U, limit);
>>   }
>> diff --git a/net/netfilter/ipvs/ip_vs_ctl.c b/net/netfilter/ipvs/ip_vs_ctl.c
>> index 032425025d88..e253a1ceef48 100644
>> --- a/net/netfilter/ipvs/ip_vs_ctl.c
>> +++ b/net/netfilter/ipvs/ip_vs_ctl.c
>> @@ -2338,11 +2338,14 @@ static int ipvs_proc_est_cpumask_get(const struct ctl_table *table,
>>   
>>   	mutex_lock(&ipvs->est_mutex);
>>   
>> -	if (ipvs->est_cpulist_valid)
>> -		mask = *valp;
>> -	else
>> -		mask = (struct cpumask *)housekeeping_cpumask(HK_TYPE_KTHREAD);
>> -	ret = scnprintf(buffer, size, "%*pbl\n", cpumask_pr_args(mask));
>> +	/* HK_TYPE_KTHREAD cpumask needs RCU protection */
>> +	scoped_guard(rcu) {
>> +		if (ipvs->est_cpulist_valid)
>> +			mask = *valp;
>> +		else
>> +			mask = (struct cpumask *)housekeeping_cpumask(HK_TYPE_KTHREAD);
>> +		ret = scnprintf(buffer, size, "%*pbl\n", cpumask_pr_args(mask));
>> +	}
> And that works.
>
> Thanks.
>
>>   
>>   	mutex_unlock(&ipvs->est_mutex);
>>   
>> -- 
>> 2.53.0
>>

Re: [PATCH-next v2 0/2] ipvs: Fix incorrect use of HK_TYPE_KTHREAD housekeeping cpumask

[Julian Anastasov]

	Hello,

On Tue, 31 Mar 2026, Waiman Long wrote:

>  v2:
>   - Rebased on top of linux-next
> 
> Since commit 041ee6f3727a ("kthread: Rely on HK_TYPE_DOMAIN for preferred
> affinity management"), the HK_TYPE_KTHREAD housekeeping cpumask may no
> longer be correct in showing the actual CPU affinity of kthreads that
> have no predefined CPU affinity. As the ipvs networking code is still
> using HK_TYPE_KTHREAD, we need to make HK_TYPE_KTHREAD reflect the
> reality.
> 
> This patch series makes HK_TYPE_KTHREAD an alias of HK_TYPE_DOMAIN
> and uses RCU to protect access to the HK_TYPE_KTHREAD housekeeping
> cpumask.
> 
> Waiman Long (2):
>   sched/isolation: Make HK_TYPE_KTHREAD an alias of HK_TYPE_DOMAIN
>   ipvs: Guard access of HK_TYPE_KTHREAD cpumask with RCU

	The patchset looks good to me for nf-next, thanks!

Acked-by: Julian Anastasov <ja@ssi.bg>

	Pablo, Florian, as a bugfix this patchset missed
the chance to be applied before the changes that are in
nf-next in ip_vs.h, there is little fuzz there. If there
is no chance to resolve it somehow, we can apply it
on top of nf-next where it now applies successfully.

> 
>  include/linux/sched/isolation.h |  6 +++++-
>  include/net/ip_vs.h             | 20 ++++++++++++++++----
>  net/netfilter/ipvs/ip_vs_ctl.c  | 13 ++++++++-----
>  3 files changed, 29 insertions(+), 10 deletions(-)

Regards

--
Julian Anastasov <ja@ssi.bg>

Re: [PATCH-next v2 0/2] ipvs: Fix incorrect use of HK_TYPE_KTHREAD housekeeping cpumask

[Pablo Neira Ayuso]

On Fri, Apr 03, 2026 at 05:15:50PM +0300, Julian Anastasov wrote:
> 
> 	Hello,
> 
> On Tue, 31 Mar 2026, Waiman Long wrote:
> 
> >  v2:
> >   - Rebased on top of linux-next
> > 
> > Since commit 041ee6f3727a ("kthread: Rely on HK_TYPE_DOMAIN for preferred
> > affinity management"), the HK_TYPE_KTHREAD housekeeping cpumask may no
> > longer be correct in showing the actual CPU affinity of kthreads that
> > have no predefined CPU affinity. As the ipvs networking code is still
> > using HK_TYPE_KTHREAD, we need to make HK_TYPE_KTHREAD reflect the
> > reality.
> > 
> > This patch series makes HK_TYPE_KTHREAD an alias of HK_TYPE_DOMAIN
> > and uses RCU to protect access to the HK_TYPE_KTHREAD housekeeping
> > cpumask.
> > 
> > Waiman Long (2):
> >   sched/isolation: Make HK_TYPE_KTHREAD an alias of HK_TYPE_DOMAIN
> >   ipvs: Guard access of HK_TYPE_KTHREAD cpumask with RCU
> 
> 	The patchset looks good to me for nf-next, thanks!
> 
> Acked-by: Julian Anastasov <ja@ssi.bg>
> 
> 	Pablo, Florian, as a bugfix this patchset missed
> the chance to be applied before the changes that are in
> nf-next in ip_vs.h, there is little fuzz there. If there
> is no chance to resolve it somehow, we can apply it
> on top of nf-next where it now applies successfully.

One way to handle this is to follow up with nf-next as you suggest,
then send a backport that applies cleanly for -stable once it is
released.

Else, let me know if I am misunderstanding.

Re: [PATCH-next v2 0/2] ipvs: Fix incorrect use of HK_TYPE_KTHREAD housekeeping cpumask

[Julian Anastasov]

	Hello,

On Fri, 3 Apr 2026, Pablo Neira Ayuso wrote:

> On Fri, Apr 03, 2026 at 05:15:50PM +0300, Julian Anastasov wrote:
> > 
> > 	Hello,
> > 
> > On Tue, 31 Mar 2026, Waiman Long wrote:
> > 
> > >  v2:
> > >   - Rebased on top of linux-next
> > > 
> > > Since commit 041ee6f3727a ("kthread: Rely on HK_TYPE_DOMAIN for preferred
> > > affinity management"), the HK_TYPE_KTHREAD housekeeping cpumask may no
> > > longer be correct in showing the actual CPU affinity of kthreads that
> > > have no predefined CPU affinity. As the ipvs networking code is still
> > > using HK_TYPE_KTHREAD, we need to make HK_TYPE_KTHREAD reflect the
> > > reality.
> > > 
> > > This patch series makes HK_TYPE_KTHREAD an alias of HK_TYPE_DOMAIN
> > > and uses RCU to protect access to the HK_TYPE_KTHREAD housekeeping
> > > cpumask.
> > > 
> > > Waiman Long (2):
> > >   sched/isolation: Make HK_TYPE_KTHREAD an alias of HK_TYPE_DOMAIN
> > >   ipvs: Guard access of HK_TYPE_KTHREAD cpumask with RCU
> > 
> > 	The patchset looks good to me for nf-next, thanks!
> > 
> > Acked-by: Julian Anastasov <ja@ssi.bg>
> > 
> > 	Pablo, Florian, as a bugfix this patchset missed
> > the chance to be applied before the changes that are in
> > nf-next in ip_vs.h, there is little fuzz there. If there
> > is no chance to resolve it somehow, we can apply it
> > on top of nf-next where it now applies successfully.
> 
> One way to handle this is to follow up with nf-next as you suggest,
> then send a backport that applies cleanly for -stable once it is
> released.

	Lets do it this way, thanks!

> Else, let me know if I am misunderstanding.

Regards

--
Julian Anastasov <ja@ssi.bg>