On 02/21/21 21:43, shuo.a.liu@xxxxxxxxx wrote:
From: Shuo Liu <shuo.a.liu@xxxxxxxxx>
Without cpu hotplug support, vCPU cannot be removed from a Service VM.
Don't expose remove_cpu sysfs when CONFIG_HOTPLUG_CPU disabled.
Signed-off-by: Shuo Liu <shuo.a.liu@xxxxxxxxx>
Acked-by: Randy Dunlap <rdunlap@xxxxxxxxxxxxx> # build-tested
Cc: Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
Cc: Qais Yousef <qais.yousef@xxxxxxx>
---
drivers/virt/acrn/hsm.c | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/drivers/virt/acrn/hsm.c b/drivers/virt/acrn/hsm.c
index 1f6b7c54a1a4..6996ea6219e5 100644
--- a/drivers/virt/acrn/hsm.c
+++ b/drivers/virt/acrn/hsm.c
@@ -404,6 +404,14 @@ static ssize_t remove_cpu_store(struct device *dev,
}
static DEVICE_ATTR_WO(remove_cpu);
+static umode_t acrn_attr_visible(struct kobject *kobj, struct attribute *a, int n)
+{
+ if (a == &dev_attr_remove_cpu.attr)
+ return IS_ENABLED(CONFIG_HOTPLUG_CPU) ? a->mode : 0;
+
+ return a->mode;
+}
+
I can't find this code in Linus' master. But this looks fine from my narrow
PoV. Protecting the attribute with ifdef CONFIG_HOTPLUG_CPU is easier to read
for me, but this doesn't mean this approach is not fine.