[PATCH v5 1/3] cpu: export lockdep_assert_cpus_held()

[PATCH v5 1/3] cpu: export lockdep_assert_cpus_held()

[Hamza Mahfooz]

If CONFIG_HYPERV=m, lockdep_assert_cpus_held() is undefined for HyperV.
So, export the function so that GPL drivers can use it more broadly.

Cc: Michael Kelley <mhklinux@outlook.com>
Signed-off-by: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com>
---
v5: new to the series
---
 kernel/cpu.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/kernel/cpu.c b/kernel/cpu.c
index b605334f8ee6..d3c848d66908 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -526,6 +526,7 @@ void lockdep_assert_cpus_held(void)
 
 	percpu_rwsem_assert_held(&cpu_hotplug_lock);
 }
+EXPORT_SYMBOL_GPL(lockdep_assert_cpus_held);
 
 #ifdef CONFIG_LOCKDEP
 int lockdep_is_cpus_held(void)
-- 
2.47.1

[PATCH v5 2/3] drivers/hv: introduce vmbus_channel_set_cpu()

[Hamza Mahfooz]

The core functionality in target_cpu_store() is also needed in a
subsequent patch for automatically changing the CPU when taking
a CPU offline. As such, factor out the body of target_cpu_store()
into new function vmbus_channel_set_cpu() that can also be used
elsewhere.

No functional change is intended.

Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Michael Kelley <mhklinux@outlook.com>
Cc: Wei Liu <wei.liu@kernel.org>
Signed-off-by: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com>
---
v2: separate vmbus_channel_set_cpu() changes from
    cpu offlining changes.

v3: address comments from Michael.

v5: return count if ret is 0.
---
 drivers/hv/vmbus_drv.c | 52 +++++++++++++++++++++++++-----------------
 include/linux/hyperv.h |  1 +
 2 files changed, 32 insertions(+), 21 deletions(-)

diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c
index 2892b8da20a5..036be0f6d62d 100644
--- a/drivers/hv/vmbus_drv.c
+++ b/drivers/hv/vmbus_drv.c
@@ -1611,16 +1611,16 @@ static ssize_t target_cpu_show(struct vmbus_channel *channel, char *buf)
 {
 	return sprintf(buf, "%u\n", channel->target_cpu);
 }
-static ssize_t target_cpu_store(struct vmbus_channel *channel,
-				const char *buf, size_t count)
+
+int vmbus_channel_set_cpu(struct vmbus_channel *channel, u32 target_cpu)
 {
-	u32 target_cpu, origin_cpu;
-	ssize_t ret = count;
+	u32 origin_cpu;
+	int ret = 0;
 
-	if (vmbus_proto_version < VERSION_WIN10_V4_1)
-		return -EIO;
+	lockdep_assert_cpus_held();
+	lockdep_assert_held(&vmbus_connection.channel_mutex);
 
-	if (sscanf(buf, "%uu", &target_cpu) != 1)
+	if (vmbus_proto_version < VERSION_WIN10_V4_1)
 		return -EIO;
 
 	/* Validate target_cpu for the cpumask_test_cpu() operation below. */
@@ -1630,22 +1630,17 @@ static ssize_t target_cpu_store(struct vmbus_channel *channel,
 	if (!cpumask_test_cpu(target_cpu, housekeeping_cpumask(HK_TYPE_MANAGED_IRQ)))
 		return -EINVAL;
 
-	/* No CPUs should come up or down during this. */
-	cpus_read_lock();
-
-	if (!cpu_online(target_cpu)) {
-		cpus_read_unlock();
+	if (!cpu_online(target_cpu))
 		return -EINVAL;
-	}
 
 	/*
-	 * Synchronizes target_cpu_store() and channel closure:
+	 * Synchronizes vmbus_channel_set_cpu() and channel closure:
 	 *
 	 * { Initially: state = CHANNEL_OPENED }
 	 *
 	 * CPU1				CPU2
 	 *
-	 * [target_cpu_store()]		[vmbus_disconnect_ring()]
+	 * [vmbus_channel_set_cpu()]	[vmbus_disconnect_ring()]
 	 *
 	 * LOCK channel_mutex		LOCK channel_mutex
 	 * LOAD r1 = state		LOAD r2 = state
@@ -1660,7 +1655,6 @@ static ssize_t target_cpu_store(struct vmbus_channel *channel,
 	 * Note.  The host processes the channel messages "sequentially", in
 	 * the order in which they are received on a per-partition basis.
 	 */
-	mutex_lock(&vmbus_connection.channel_mutex);
 
 	/*
 	 * Hyper-V will ignore MODIFYCHANNEL messages for "non-open" channels;
@@ -1668,17 +1662,17 @@ static ssize_t target_cpu_store(struct vmbus_channel *channel,
 	 */
 	if (channel->state != CHANNEL_OPENED_STATE) {
 		ret = -EIO;
-		goto cpu_store_unlock;
+		goto end;
 	}
 
 	origin_cpu = channel->target_cpu;
 	if (target_cpu == origin_cpu)
-		goto cpu_store_unlock;
+		goto end;
 
 	if (vmbus_send_modifychannel(channel,
 				     hv_cpu_number_to_vp_number(target_cpu))) {
 		ret = -EIO;
-		goto cpu_store_unlock;
+		goto end;
 	}
 
 	/*
@@ -1708,10 +1702,26 @@ static ssize_t target_cpu_store(struct vmbus_channel *channel,
 				origin_cpu, target_cpu);
 	}
 
-cpu_store_unlock:
+end:
+	return ret;
+}
+
+static ssize_t target_cpu_store(struct vmbus_channel *channel,
+				const char *buf, size_t count)
+{
+	u32 target_cpu;
+	ssize_t ret;
+
+	if (sscanf(buf, "%uu", &target_cpu) != 1)
+		return -EIO;
+
+	cpus_read_lock();
+	mutex_lock(&vmbus_connection.channel_mutex);
+	ret = vmbus_channel_set_cpu(channel, target_cpu);
 	mutex_unlock(&vmbus_connection.channel_mutex);
 	cpus_read_unlock();
-	return ret;
+
+	return ret ?: count;
 }
 static VMBUS_CHAN_ATTR(cpu, 0644, target_cpu_show, target_cpu_store);
 
diff --git a/include/linux/hyperv.h b/include/linux/hyperv.h
index 02a226bcf0ed..25e9e982f1b0 100644
--- a/include/linux/hyperv.h
+++ b/include/linux/hyperv.h
@@ -1670,6 +1670,7 @@ int vmbus_send_tl_connect_request(const guid_t *shv_guest_servie_id,
 				  const guid_t *shv_host_servie_id);
 int vmbus_send_modifychannel(struct vmbus_channel *channel, u32 target_vp);
 void vmbus_set_event(struct vmbus_channel *channel);
+int vmbus_channel_set_cpu(struct vmbus_channel *channel, u32 target_cpu);
 
 /* Get the start of the ring buffer. */
 static inline void *
-- 
2.47.1

[PATCH v5 3/3] drivers/hv: add CPU offlining support

[Hamza Mahfooz]

Currently, it is tedious to offline CPUs in a Hyper-V VM since CPUs may
have VMBus channels attached to them that a user would have to manually
rebind elsewhere. So, as made mention of in
commit d570aec0f2154 ("Drivers: hv: vmbus: Synchronize init_vp_index()
vs. CPU hotplug"), rebind channels associated with CPUs that a user is
trying to offline to a new "randomly" selected CPU.

Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Michael Kelley <mhklinux@outlook.com>
Cc: Wei Liu <wei.liu@kernel.org>
Signed-off-by: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com>
---
v2: remove cpus_read_{un,}lock() from hv_pick_new_cpu() and add
    lockdep_assert_cpus_held().

v3: use for_each_cpu_wrap() in hv_pick_new_cpu().

v4: store get_random_u32_below() in start.

v5: style fixes, use target_cpu and set ret to -EBUSY by default.
---
 drivers/hv/hv.c | 72 ++++++++++++++++++++++++++++++++++---------------
 1 file changed, 51 insertions(+), 21 deletions(-)

diff --git a/drivers/hv/hv.c b/drivers/hv/hv.c
index 36d9ba097ff5..fab0690b5c41 100644
--- a/drivers/hv/hv.c
+++ b/drivers/hv/hv.c
@@ -433,13 +433,47 @@ static bool hv_synic_event_pending(void)
 	return pending;
 }
 
+static int hv_pick_new_cpu(struct vmbus_channel *channel)
+{
+	int ret = -EBUSY;
+	int start;
+	int cpu;
+
+	lockdep_assert_cpus_held();
+	lockdep_assert_held(&vmbus_connection.channel_mutex);
+
+	/*
+	 * We can't assume that the relevant interrupts will be sent before
+	 * the cpu is offlined on older versions of hyperv.
+	 */
+	if (vmbus_proto_version < VERSION_WIN10_V5_3)
+		return -EBUSY;
+
+	start = get_random_u32_below(nr_cpu_ids);
+
+	for_each_cpu_wrap(cpu, cpu_online_mask, start) {
+		if (channel->target_cpu == cpu ||
+		    channel->target_cpu == VMBUS_CONNECT_CPU)
+			continue;
+
+		ret = vmbus_channel_set_cpu(channel, cpu);
+		if (!ret)
+			break;
+	}
+
+	if (ret)
+		ret = vmbus_channel_set_cpu(channel, VMBUS_CONNECT_CPU);
+
+	return ret;
+}
+
 /*
  * hv_synic_cleanup - Cleanup routine for hv_synic_init().
  */
 int hv_synic_cleanup(unsigned int cpu)
 {
 	struct vmbus_channel *channel, *sc;
-	bool channel_found = false;
+	int ret = 0;
 
 	if (vmbus_connection.conn_state != CONNECTED)
 		goto always_cleanup;
@@ -456,38 +490,34 @@ int hv_synic_cleanup(unsigned int cpu)
 
 	/*
 	 * Search for channels which are bound to the CPU we're about to
-	 * cleanup.  In case we find one and vmbus is still connected, we
-	 * fail; this will effectively prevent CPU offlining.
-	 *
-	 * TODO: Re-bind the channels to different CPUs.
+	 * cleanup.
 	 */
 	mutex_lock(&vmbus_connection.channel_mutex);
 	list_for_each_entry(channel, &vmbus_connection.chn_list, listentry) {
 		if (channel->target_cpu == cpu) {
-			channel_found = true;
-			break;
+			ret = hv_pick_new_cpu(channel);
+			if (ret) {
+				mutex_unlock(&vmbus_connection.channel_mutex);
+				return ret;
+			}
 		}
 		list_for_each_entry(sc, &channel->sc_list, sc_list) {
 			if (sc->target_cpu == cpu) {
-				channel_found = true;
-				break;
+				ret = hv_pick_new_cpu(sc);
+				if (ret) {
+					mutex_unlock(&vmbus_connection.channel_mutex);
+					return ret;
+				}
 			}
 		}
-		if (channel_found)
-			break;
 	}
 	mutex_unlock(&vmbus_connection.channel_mutex);
 
-	if (channel_found)
-		return -EBUSY;
-
 	/*
-	 * channel_found == false means that any channels that were previously
-	 * assigned to the CPU have been reassigned elsewhere with a call of
-	 * vmbus_send_modifychannel().  Scan the event flags page looking for
-	 * bits that are set and waiting with a timeout for vmbus_chan_sched()
-	 * to process such bits.  If bits are still set after this operation
-	 * and VMBus is connected, fail the CPU offlining operation.
+	 * Scan the event flags page looking for bits that are set and waiting
+	 * with a timeout for vmbus_chan_sched() to process such bits. If bits
+	 * are still set after this operation and VMBus is connected, fail the
+	 * CPU offlining operation.
 	 */
 	if (vmbus_proto_version >= VERSION_WIN10_V4_1 && hv_synic_event_pending())
 		return -EBUSY;
@@ -497,5 +527,5 @@ int hv_synic_cleanup(unsigned int cpu)
 
 	hv_synic_disable_regs(cpu);
 
-	return 0;
+	return ret;
 }
-- 
2.47.1

RE: [PATCH v5 1/3] cpu: export lockdep_assert_cpus_held()

[Michael Kelley]

From: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com> Sent: Friday, January 17, 2025 12:33 PM
> 
> If CONFIG_HYPERV=m, lockdep_assert_cpus_held() is undefined for HyperV.
> So, export the function so that GPL drivers can use it more broadly.
> 
> Cc: Michael Kelley <mhklinux@outlook.com>
> Signed-off-by: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com>
> ---
> v5: new to the series
> ---
>  kernel/cpu.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/kernel/cpu.c b/kernel/cpu.c
> index b605334f8ee6..d3c848d66908 100644
> --- a/kernel/cpu.c
> +++ b/kernel/cpu.c
> @@ -526,6 +526,7 @@ void lockdep_assert_cpus_held(void)
> 
>  	percpu_rwsem_assert_held(&cpu_hotplug_lock);
>  }
> +EXPORT_SYMBOL_GPL(lockdep_assert_cpus_held);
> 
>  #ifdef CONFIG_LOCKDEP
>  int lockdep_is_cpus_held(void)
> --
> 2.47.1

This series should have been posted as "v6" since "v5" was posted on
January 16th.  That notwithstanding,

Reviewed-by: Michael Kelley <mhklinux@outlook.com>
Tested-by: Michael Kelley <mhklinux@outlook.com>

RE: [PATCH v5 2/3] drivers/hv: introduce vmbus_channel_set_cpu()

[Michael Kelley]

From: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com> Sent: Friday, January 17, 2025 12:33 PM
> 
> The core functionality in target_cpu_store() is also needed in a
> subsequent patch for automatically changing the CPU when taking
> a CPU offline. As such, factor out the body of target_cpu_store()
> into new function vmbus_channel_set_cpu() that can also be used
> elsewhere.
> 
> No functional change is intended.
> 
> Cc: Boqun Feng <boqun.feng@gmail.com>
> Cc: Michael Kelley <mhklinux@outlook.com>
> Cc: Wei Liu <wei.liu@kernel.org>
> Signed-off-by: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com>
> ---
> v2: separate vmbus_channel_set_cpu() changes from
>     cpu offlining changes.
> 
> v3: address comments from Michael.
> 
> v5: return count if ret is 0.
> ---
>  drivers/hv/vmbus_drv.c | 52 +++++++++++++++++++++++++-----------------
>  include/linux/hyperv.h |  1 +
>  2 files changed, 32 insertions(+), 21 deletions(-)
> 
> diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c
> index 2892b8da20a5..036be0f6d62d 100644
> --- a/drivers/hv/vmbus_drv.c
> +++ b/drivers/hv/vmbus_drv.c
> @@ -1611,16 +1611,16 @@ static ssize_t target_cpu_show(struct vmbus_channel *channel, char *buf)
>  {
>  	return sprintf(buf, "%u\n", channel->target_cpu);
>  }
> -static ssize_t target_cpu_store(struct vmbus_channel *channel,
> -				const char *buf, size_t count)
> +
> +int vmbus_channel_set_cpu(struct vmbus_channel *channel, u32 target_cpu)
>  {
> -	u32 target_cpu, origin_cpu;
> -	ssize_t ret = count;
> +	u32 origin_cpu;
> +	int ret = 0;
> 
> -	if (vmbus_proto_version < VERSION_WIN10_V4_1)
> -		return -EIO;
> +	lockdep_assert_cpus_held();
> +	lockdep_assert_held(&vmbus_connection.channel_mutex);
> 
> -	if (sscanf(buf, "%uu", &target_cpu) != 1)
> +	if (vmbus_proto_version < VERSION_WIN10_V4_1)
>  		return -EIO;
> 
>  	/* Validate target_cpu for the cpumask_test_cpu() operation below. */
> @@ -1630,22 +1630,17 @@ static ssize_t target_cpu_store(struct vmbus_channel *channel,
>  	if (!cpumask_test_cpu(target_cpu, housekeeping_cpumask(HK_TYPE_MANAGED_IRQ)))
>  		return -EINVAL;
> 
> -	/* No CPUs should come up or down during this. */
> -	cpus_read_lock();
> -
> -	if (!cpu_online(target_cpu)) {
> -		cpus_read_unlock();
> +	if (!cpu_online(target_cpu))
>  		return -EINVAL;
> -	}
> 
>  	/*
> -	 * Synchronizes target_cpu_store() and channel closure:
> +	 * Synchronizes vmbus_channel_set_cpu() and channel closure:
>  	 *
>  	 * { Initially: state = CHANNEL_OPENED }
>  	 *
>  	 * CPU1				CPU2
>  	 *
> -	 * [target_cpu_store()]		[vmbus_disconnect_ring()]
> +	 * [vmbus_channel_set_cpu()]	[vmbus_disconnect_ring()]
>  	 *
>  	 * LOCK channel_mutex		LOCK channel_mutex
>  	 * LOAD r1 = state		LOAD r2 = state
> @@ -1660,7 +1655,6 @@ static ssize_t target_cpu_store(struct vmbus_channel *channel,
>  	 * Note.  The host processes the channel messages "sequentially", in
>  	 * the order in which they are received on a per-partition basis.
>  	 */
> -	mutex_lock(&vmbus_connection.channel_mutex);
> 
>  	/*
>  	 * Hyper-V will ignore MODIFYCHANNEL messages for "non-open" channels;
> @@ -1668,17 +1662,17 @@ static ssize_t target_cpu_store(struct vmbus_channel *channel,
>  	 */
>  	if (channel->state != CHANNEL_OPENED_STATE) {
>  		ret = -EIO;
> -		goto cpu_store_unlock;
> +		goto end;
>  	}
> 
>  	origin_cpu = channel->target_cpu;
>  	if (target_cpu == origin_cpu)
> -		goto cpu_store_unlock;
> +		goto end;
> 
>  	if (vmbus_send_modifychannel(channel,
>  				     hv_cpu_number_to_vp_number(target_cpu))) {
>  		ret = -EIO;
> -		goto cpu_store_unlock;
> +		goto end;
>  	}
> 
>  	/*
> @@ -1708,10 +1702,26 @@ static ssize_t target_cpu_store(struct vmbus_channel *channel,
>  				origin_cpu, target_cpu);
>  	}
> 
> -cpu_store_unlock:
> +end:
> +	return ret;
> +}
> +
> +static ssize_t target_cpu_store(struct vmbus_channel *channel,
> +				const char *buf, size_t count)
> +{
> +	u32 target_cpu;
> +	ssize_t ret;
> +
> +	if (sscanf(buf, "%uu", &target_cpu) != 1)
> +		return -EIO;
> +
> +	cpus_read_lock();
> +	mutex_lock(&vmbus_connection.channel_mutex);
> +	ret = vmbus_channel_set_cpu(channel, target_cpu);
>  	mutex_unlock(&vmbus_connection.channel_mutex);
>  	cpus_read_unlock();
> -	return ret;
> +
> +	return ret ?: count;
>  }
>  static VMBUS_CHAN_ATTR(cpu, 0644, target_cpu_show, target_cpu_store);
> 
> diff --git a/include/linux/hyperv.h b/include/linux/hyperv.h
> index 02a226bcf0ed..25e9e982f1b0 100644
> --- a/include/linux/hyperv.h
> +++ b/include/linux/hyperv.h
> @@ -1670,6 +1670,7 @@ int vmbus_send_tl_connect_request(const guid_t
> *shv_guest_servie_id,
>  				  const guid_t *shv_host_servie_id);
>  int vmbus_send_modifychannel(struct vmbus_channel *channel, u32 target_vp);
>  void vmbus_set_event(struct vmbus_channel *channel);
> +int vmbus_channel_set_cpu(struct vmbus_channel *channel, u32 target_cpu);
> 
>  /* Get the start of the ring buffer. */
>  static inline void *
> --
> 2.47.1

Reviewed-by: Michael Kelley <mhklinux@outlook.com>
Tested-by: Michael Kelley <mhklinux@outlook.com>

RE: [PATCH v5 3/3] drivers/hv: add CPU offlining support

[Michael Kelley]

From: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com> Sent: Friday, January 17, 2025 12:33 PM
> 
> Currently, it is tedious to offline CPUs in a Hyper-V VM since CPUs may
> have VMBus channels attached to them that a user would have to manually
> rebind elsewhere. So, as made mention of in
> commit d570aec0f2154 ("Drivers: hv: vmbus: Synchronize init_vp_index()
> vs. CPU hotplug"), rebind channels associated with CPUs that a user is
> trying to offline to a new "randomly" selected CPU.
> 
> Cc: Boqun Feng <boqun.feng@gmail.com>
> Cc: Michael Kelley <mhklinux@outlook.com>
> Cc: Wei Liu <wei.liu@kernel.org>
> Signed-off-by: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com>
> ---
> v2: remove cpus_read_{un,}lock() from hv_pick_new_cpu() and add
>     lockdep_assert_cpus_held().
> 
> v3: use for_each_cpu_wrap() in hv_pick_new_cpu().
> 
> v4: store get_random_u32_below() in start.
> 
> v5: style fixes, use target_cpu and set ret to -EBUSY by default.
> ---
>  drivers/hv/hv.c | 72 ++++++++++++++++++++++++++++++++++---------------
>  1 file changed, 51 insertions(+), 21 deletions(-)
> 
> diff --git a/drivers/hv/hv.c b/drivers/hv/hv.c
> index 36d9ba097ff5..fab0690b5c41 100644
> --- a/drivers/hv/hv.c
> +++ b/drivers/hv/hv.c
> @@ -433,13 +433,47 @@ static bool hv_synic_event_pending(void)
>  	return pending;
>  }
> 
> +static int hv_pick_new_cpu(struct vmbus_channel *channel)
> +{
> +	int ret = -EBUSY;
> +	int start;
> +	int cpu;
> +
> +	lockdep_assert_cpus_held();
> +	lockdep_assert_held(&vmbus_connection.channel_mutex);
> +
> +	/*
> +	 * We can't assume that the relevant interrupts will be sent before
> +	 * the cpu is offlined on older versions of hyperv.
> +	 */
> +	if (vmbus_proto_version < VERSION_WIN10_V5_3)
> +		return -EBUSY;
> +
> +	start = get_random_u32_below(nr_cpu_ids);
> +
> +	for_each_cpu_wrap(cpu, cpu_online_mask, start) {
> +		if (channel->target_cpu == cpu ||
> +		    channel->target_cpu == VMBUS_CONNECT_CPU)
> +			continue;
> +
> +		ret = vmbus_channel_set_cpu(channel, cpu);
> +		if (!ret)
> +			break;
> +	}
> +
> +	if (ret)
> +		ret = vmbus_channel_set_cpu(channel, VMBUS_CONNECT_CPU);
> +
> +	return ret;
> +}
> +
>  /*
>   * hv_synic_cleanup - Cleanup routine for hv_synic_init().
>   */
>  int hv_synic_cleanup(unsigned int cpu)
>  {
>  	struct vmbus_channel *channel, *sc;
> -	bool channel_found = false;
> +	int ret = 0;
> 
>  	if (vmbus_connection.conn_state != CONNECTED)
>  		goto always_cleanup;
> @@ -456,38 +490,34 @@ int hv_synic_cleanup(unsigned int cpu)
> 
>  	/*
>  	 * Search for channels which are bound to the CPU we're about to
> -	 * cleanup.  In case we find one and vmbus is still connected, we
> -	 * fail; this will effectively prevent CPU offlining.
> -	 *
> -	 * TODO: Re-bind the channels to different CPUs.
> +	 * cleanup.
>  	 */
>  	mutex_lock(&vmbus_connection.channel_mutex);
>  	list_for_each_entry(channel, &vmbus_connection.chn_list, listentry) {
>  		if (channel->target_cpu == cpu) {
> -			channel_found = true;
> -			break;
> +			ret = hv_pick_new_cpu(channel);
> +			if (ret) {
> +				mutex_unlock(&vmbus_connection.channel_mutex);
> +				return ret;
> +			}
>  		}
>  		list_for_each_entry(sc, &channel->sc_list, sc_list) {
>  			if (sc->target_cpu == cpu) {
> -				channel_found = true;
> -				break;
> +				ret = hv_pick_new_cpu(sc);
> +				if (ret) {
> +
> 	mutex_unlock(&vmbus_connection.channel_mutex);
> +					return ret;
> +				}
>  			}
>  		}
> -		if (channel_found)
> -			break;
>  	}
>  	mutex_unlock(&vmbus_connection.channel_mutex);
> 
> -	if (channel_found)
> -		return -EBUSY;
> -
>  	/*
> -	 * channel_found == false means that any channels that were previously
> -	 * assigned to the CPU have been reassigned elsewhere with a call of
> -	 * vmbus_send_modifychannel().  Scan the event flags page looking for
> -	 * bits that are set and waiting with a timeout for vmbus_chan_sched()
> -	 * to process such bits.  If bits are still set after this operation
> -	 * and VMBus is connected, fail the CPU offlining operation.
> +	 * Scan the event flags page looking for bits that are set and waiting
> +	 * with a timeout for vmbus_chan_sched() to process such bits. If bits
> +	 * are still set after this operation and VMBus is connected, fail the
> +	 * CPU offlining operation.
>  	 */
>  	if (vmbus_proto_version >= VERSION_WIN10_V4_1 && hv_synic_event_pending())
>  		return -EBUSY;
> @@ -497,5 +527,5 @@ int hv_synic_cleanup(unsigned int cpu)
> 
>  	hv_synic_disable_regs(cpu);
> 
> -	return 0;
> +	return ret;
>  }
> --
> 2.47.1

Reviewed-by: Michael Kelley <mhklinux@outlook.com>
Tested-by: Michael Kelley <mhklinux@outlook.com>

Re: [PATCH v5 1/3] cpu: export lockdep_assert_cpus_held()

[Hamza Mahfooz]

ping?

On Wed, Jan 22, 2025 at 03:37:44PM +0000, Michael Kelley wrote:
> From: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com> Sent: Friday, January 17, 2025 12:33 PM
> > 
> > If CONFIG_HYPERV=m, lockdep_assert_cpus_held() is undefined for HyperV.
> > So, export the function so that GPL drivers can use it more broadly.
> > 
> > Cc: Michael Kelley <mhklinux@outlook.com>
> > Signed-off-by: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com>
> > ---
> > v5: new to the series
> > ---
> >  kernel/cpu.c | 1 +
> >  1 file changed, 1 insertion(+)
> > 
> > diff --git a/kernel/cpu.c b/kernel/cpu.c
> > index b605334f8ee6..d3c848d66908 100644
> > --- a/kernel/cpu.c
> > +++ b/kernel/cpu.c
> > @@ -526,6 +526,7 @@ void lockdep_assert_cpus_held(void)
> > 
> >  	percpu_rwsem_assert_held(&cpu_hotplug_lock);
> >  }
> > +EXPORT_SYMBOL_GPL(lockdep_assert_cpus_held);
> > 
> >  #ifdef CONFIG_LOCKDEP
> >  int lockdep_is_cpus_held(void)
> > --
> > 2.47.1
> 
> This series should have been posted as "v6" since "v5" was posted on
> January 16th.  That notwithstanding,
> 
> Reviewed-by: Michael Kelley <mhklinux@outlook.com>
> Tested-by: Michael Kelley <mhklinux@outlook.com>
> 

Re: [PATCH v5 1/3] cpu: export lockdep_assert_cpus_held()

[Thomas Gleixner]

On Fri, Jan 17 2025 at 15:33, Hamza Mahfooz wrote:
> If CONFIG_HYPERV=m, lockdep_assert_cpus_held() is undefined for HyperV.
> So, export the function so that GPL drivers can use it more broadly.
>
> Cc: Michael Kelley <mhklinux@outlook.com>
> Signed-off-by: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com>

Acked-by: Thomas Gleixner <tglx@linutronix.de>

Re: [PATCH v5 1/3] cpu: export lockdep_assert_cpus_held()

[Wei Liu]

On Fri, Jan 17, 2025 at 03:33:06PM -0500, Hamza Mahfooz wrote:
> If CONFIG_HYPERV=m, lockdep_assert_cpus_held() is undefined for HyperV.
> So, export the function so that GPL drivers can use it more broadly.
> 
> Cc: Michael Kelley <mhklinux@outlook.com>
> Signed-off-by: Hamza Mahfooz <hamzamahfooz@linux.microsoft.com>

Series applied to hyperv-next. Thanks.

In the future it is better to write a small cover letter for a patch
series.

Thanks,
Wei.