x86/xen: don't reset vcpu_info on a cancelled suspend
authorOuyang Zhaowei (Charles) <ouyangzhaowei@huawei.com>
Wed, 6 May 2015 01:47:04 +0000 (09:47 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 31 Jan 2016 19:28:53 +0000 (11:28 -0800)
commit 6a1f513776b78c994045287073e55bae44ed9f8c upstream.

On a cancelled suspend the vcpu_info location does not change (it's
still in the per-cpu area registered by xen_vcpu_setup()).  So do not
call xen_hvm_init_shared_info() which would make the kernel think its
back in the shared info.  With the wrong vcpu_info, events cannot be
received and the domain will hang after a cancelled suspend.

Signed-off-by: Charles Ouyang <ouyangzhaowei@huawei.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
arch/x86/xen/suspend.c

index df0c405..7f664c4 100644 (file)
@@ -34,7 +34,8 @@ static void xen_hvm_post_suspend(int suspend_cancelled)
 {
 #ifdef CONFIG_XEN_PVHVM
        int cpu;
-       xen_hvm_init_shared_info();
+       if (!suspend_cancelled)
+           xen_hvm_init_shared_info();
        xen_callback_vector();
        xen_unplug_emulated_devices();
        if (xen_feature(XENFEAT_hvm_safe_pvclock)) {