Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.32 as permitted sender) client-ip=23.128.96.32;
Date:   Thu,  5 Oct 2023 19:20:07 -0700
In-Reply-To: <20231002133230.195738-1-michael.roth@amd.com>
Mime-Version: 1.0
References: <20231002133230.195738-1-michael.roth@amd.com>
Message-ID: <169655697465.3534024.7746001565902273346.b4-ty@google.com>
Subject: Re: [PATCH gmem FIXUP] KVM: Don't re-use inodes when creating
 guest_memfd files
From:   Sean Christopherson <seanjc@google.com>
To:     Sean Christopherson <seanjc@google.com>, kvm@vger.kernel.org,
        Michael Roth <michael.roth@amd.com>
Cc:     linux-kernel@vger.kernel.org, Paolo Bonzini <pbonzini@redhat.com>
Content-Type: text/plain; charset="utf-8"
Precedence: bulk

On Mon, 02 Oct 2023 08:32:30 -0500, Michael Roth wrote:
> anon_inode_getfile() uses a singleton inode, which results in the inode
> size changing based with each new KVM_CREATE_GUEST_MEMFD call, which
> can later lead to previously-created guest_memfd files failing bounds
> checks that are later performed when memslots are bound to them. More
> generally, the inode may be associated with other state that cannot be
> shared across multiple guest_memfd instances.
> 
> [...]

Applied to kvm-x86 guest_memfd, thanks!  I added a comment to explain the use
of the "secure" API, there's a non-zero chance we'll forget that wrinkle again
in the future.

[1/1] KVM: Don't re-use inodes when creating guest_memfd files
      https://github.com/kvm-x86/linux/commit/b3bf68b66062

--
https://github.com/kvm-x86/linux/tree/next