Received-SPF: pass (google.com: domain of linux-kernel+bounces-63939-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Date: Tue, 13 Feb 2024 16:51:12 +0000
From: Oliver Upton <oliver.upton@linux.dev>
To: Sebastian Ene <sebastianene@google.com>
Cc: catalin.marinas@arm.com, gshan@redhat.com, james.morse@arm.com,
	mark.rutland@arm.com, maz@kernel.org, rananta@google.com,
	ricarkol@google.com, ryan.roberts@arm.com, shahuang@redhat.com,
	suzuki.poulose@arm.com, will@kernel.org, yuzenghui@huawei.com,
	kvmarm@lists.linux.dev, linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, kernel-team@android.com,
	vdonnefort@google.com
Subject: Re: [PATCH v5 3/4] KVM: arm64: Register ptdump with debugfs on guest
 creation
Message-ID: <ZcueAGbQISkxaTNR@linux.dev>
References: <20240207144832.1017815-2-sebastianene@google.com>
 <20240207144832.1017815-5-sebastianene@google.com>
 <Zcq-NGa-Gd9rVVgF@linux.dev>
 <Zcub87FwaVPkCXQE@google.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Zcub87FwaVPkCXQE@google.com>

On Tue, Feb 13, 2024 at 04:42:27PM +0000, Sebastian Ene wrote:
> On Tue, Feb 13, 2024 at 12:56:20AM +0000, Oliver Upton wrote:
> > On Wed, Feb 07, 2024 at 02:48:32PM +0000, Sebastian Ene wrote:

[...]

> > > +static int kvm_ptdump_guest_open(struct inode *inode, struct file *file)
> > > +{
> > > +	return single_open(file, kvm_ptdump_guest_show, inode->i_private);
> > > +}
> > > +
> > 
> > Shouldn't we take a reference on the KVM struct at open to avoid UAF?
> > 
> > 	struct kvm *kvm = inode->i_private;
> > 
> > 	if (!kvm_get_kvm_safe(kvm))
> > 		return -ENOENT;
> > 
> > Then you can do a put on it at close().
> > 
> 
> Thanks, I though that the kvm_destroy_vm_debugfs will keep spinning if
> there are opened paths to the debugfs entry, but I guess nothing prevents
> that from happening and the kvm struct can be removed behind our back.

kvm_destroy_vm() will get called after the last put() on a kvm struct,
so all debugfs files should be closed too if we're consistent about
this.

-- 
Thanks,
Oliver