linux-kernel - Re: [PATCH] KVM: x86: Hoist input checks in kvm_add_msr

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <YJBitcip8/WIsaB9@google.com>
Date:   Mon, 3 May 2021 20:53:09 +0000
From:   Sean Christopherson <seanjc@...gle.com>
To:     Siddharth Chandrasekaran <sidcha@...zon.de>
Cc:     Paolo Bonzini <pbonzini@...hat.com>,
        Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        x86@...nel.org, "H. Peter Anvin" <hpa@...or.com>,
        kvm@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] KVM: x86: Hoist input checks in kvm_add_msr_filter()

On Mon, May 03, 2021, Siddharth Chandrasekaran wrote:
> In ioctl KVM_X86_SET_MSR_FILTER, input from user space is validated
> after a memdup_user(). For invalid inputs we'd memdup and then call
> kfree unnecessarily. Hoist input validation to avoid kfree altogether.
> 
> Signed-off-by: Siddharth Chandrasekaran <sidcha@...zon.de>
> ---
>  arch/x86/kvm/x86.c | 21 ++++++---------------
>  1 file changed, 6 insertions(+), 15 deletions(-)
> 
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index ee0dc58ac3a5..15c20b31cc91 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -5393,11 +5393,16 @@ static int kvm_add_msr_filter(struct kvm_x86_msr_filter *msr_filter,
>  	struct msr_bitmap_range range;
>  	unsigned long *bitmap = NULL;
>  	size_t bitmap_size;
> -	int r;
>  
>  	if (!user_range->nmsrs)
>  		return 0;
>  
> +	if (user_range->flags & ~(KVM_MSR_FILTER_READ | KVM_MSR_FILTER_WRITE))
> +		return -EINVAL;
> +
> +	if (!user_range->flags)
> +		return -EINVAL;
> +
>  	bitmap_size = BITS_TO_LONGS(user_range->nmsrs) * sizeof(long);
>  	if (!bitmap_size || bitmap_size > KVM_MSR_FILTER_MAX_BITMAP_SIZE)
>  		return -EINVAL;
> @@ -5413,24 +5418,10 @@ static int kvm_add_msr_filter(struct kvm_x86_msr_filter *msr_filter,
>  		.bitmap = bitmap,
>  	};
>  
> -	if (range.flags & ~(KVM_MSR_FILTER_READ | KVM_MSR_FILTER_WRITE)) {
> -		r = -EINVAL;
> -		goto err;
> -	}
> -
> -	if (!range.flags) {
> -		r = -EINVAL;
> -		goto err;
> -	}
> -
> -	/* Everything ok, add this range identifier. */
>  	msr_filter->ranges[msr_filter->count] = range;

Might be worth elminating the intermediate "range", too.  Doesn't affect output,
but it would make it a little more obvious that the new range is mostly coming
straight from userspace input.  E.g.

	msr_filter->ranges[msr_filter->count] = (struct msr_bitmap_range) {
		.flags = user_range->flags,
		.base = user_range->base,
		.nmsrs = user_range->nmsrs,
		.bitmap = bitmap,
	};

Either way:

Reviewed-by: Sean Christopherson <seanjc@...gle.com>

>  	msr_filter->count++;
>  
>  	return 0;
> -err:
> -	kfree(bitmap);
> -	return r;
>  }
>  
>  static int kvm_vm_ioctl_set_msr_filter(struct kvm *kvm, void __user *argp)
> --