From mboxrd@z Thu Jan 1 00:00:00 1970
From: bugzilla-daemon@bugzilla.kernel.org
Subject: [Bug 197861] Shutting down a VM with Kernel 4.14 will sometime hang
and a reboot is the only way to recover.
Date: Tue, 12 Dec 2017 22:13:30 +0000
Message-ID:
References:
Mime-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8BIT
To: kvm@kernel.org
Return-path:
Received: from mail.wl.linuxfoundation.org ([198.145.29.98]:56162 "EHLO
mail.wl.linuxfoundation.org" rhost-flags-OK-OK-OK-OK)
by vger.kernel.org with ESMTP id S1752320AbdLLWNc (ORCPT
); Tue, 12 Dec 2017 17:13:32 -0500
Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1])
by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 0622F29211
for ; Tue, 12 Dec 2017 22:13:32 +0000 (UTC)
In-Reply-To:
Sender: kvm-owner@vger.kernel.org
List-ID:
https://bugzilla.kernel.org/show_bug.cgi?id=197861
Willem de Bruijn (willemb@google.com) changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |willemb@google.com
--- Comment #6 from Willem de Bruijn (willemb@google.com) ---
Does reverting only that patch resolve the issue?
The logic in it is quite complex, but it is only needed for zerocopy with
MSG_ZEROCOPY. And then only in edge cases.
This is likely not used here. The code is blocking on vhost_net use of
zerocopy. Which does not build skbuffs with zerocopy data in compound pages.
The patch adds checks against shared and cloned skbs that were not present
before. It is not safe to modify skb frags[] on on either, but perhaps this
changed return path causes buffers to not be released, causing
vhost_net_ubuf_put_and_wait to wait seemingly indefinitely.
+ if (skb_shared(skb) || skb_unclone(skb, gfp_mask))
+ return -EINVAL;
--
You are receiving this mail because:
You are watching the assignee of the bug.