Thread (26 messages) 26 messages, 6 authors, 2025-01-22

Re: [PATCH v5 6/6] vhost_scsi: Add check for inherit_owner status

From: Mike Christie <michael.christie@oracle.com>
Date: 2025-01-22 20:19:06
Also in: lkml, virtualization

On 12/30/24 6:43 AM, Cindy Lu wrote:
quoted hunk ↗ jump to hunk
The vhost_scsi VHOST_NEW_WORKER requires the inherit_owner
setting to be true. So we need to implement a check for this.

Signed-off-by: Cindy Lu <redacted>
---
 drivers/vhost/scsi.c | 8 ++++++++
 1 file changed, 8 insertions(+)
diff --git a/drivers/vhost/scsi.c b/drivers/vhost/scsi.c
index 718fa4e0b31e..0d63b6b5c852 100644
--- a/drivers/vhost/scsi.c
+++ b/drivers/vhost/scsi.c
@@ -2086,6 +2086,14 @@ vhost_scsi_ioctl(struct file *f,
 			return -EFAULT;
 		return vhost_scsi_set_features(vs, features);
 	case VHOST_NEW_WORKER:
+		/*
+		 * vhost_tasks will account for worker threads under the parent's
+		 * NPROC value but kthreads do not. To avoid userspace overflowing
+		 * the system with worker threads inherit_owner must be true.
+		 */
+		if (!vs->dev.inherit_owner)> +			return -EFAULT;
Why did you keep this here? I had mentioned it belonged in the common code:

https://lore.kernel.org/virtualization/3864ae3b-d6cd-4227-b4bb-56e014c71667@oracle.com/T/#u (local)

because the limitation applies to all drivers. I didn't see a reply about it
but saw you did fix up the comment. Not sure if I missed it or there was
a technical issue you hit.

+		fallthrough;
 	case VHOST_FREE_WORKER:
 	case VHOST_ATTACH_VRING_WORKER:
 	case VHOST_GET_VRING_WORKER:
  
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help