From: Paolo Bonzini <pbonzini@redhat.com>
Date: Wed, 13 Jun 2012 14:56:31 +0000 (+0200)
Subject: [SCSI] virtio-scsi: unlock during kick
X-Git-Url: https://git.stricted.de/?a=commitdiff_plain;h=b5ee8f2802c559fccb177c0a117f5cd56c1049cc;p=GitHub%2FLineageOS%2Fandroid_kernel_motorola_exynos9610.git

[SCSI] virtio-scsi: unlock during kick

Separate virtqueue_kick_prepare from virtqueue_notify, so that the
expensive vmexit is done without holding the lock.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
---

diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
index 1b3843117268..b0ad5aa6b552 100644
--- a/drivers/scsi/virtio_scsi.c
+++ b/drivers/scsi/virtio_scsi.c
@@ -264,9 +264,11 @@ static int virtscsi_kick_cmd(struct virtio_scsi *vscsi, struct virtqueue *vq,
 
 	ret = virtqueue_add_buf(vq, vscsi->sg, out_num, in_num, cmd, gfp);
 	if (ret >= 0)
-		virtqueue_kick(vq);
+		ret = virtqueue_kick_prepare(vq);
 
 	spin_unlock_irqrestore(&vscsi->vq_lock, flags);
+	if (ret > 0)
+		virtqueue_notify(vq);
 	return ret;
 }