lisa/android_kernel_xiaomi_sm8350
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

KVM: PPC: Book3S: Fix H_RTAS rets buffer overflow

Browse Source 
commit f62f3c20647ebd5fb6ecb8f0b477b9281c44c10a upstream.

The kvmppc_rtas_hcall() sets the host rtas_args.rets pointer based on
the rtas_args.nargs that was provided by the guest. That guest nargs
value is not range checked, so the guest can cause the host rets pointer
to be pointed outside the args array. The individual rtas function
handlers check the nargs and nrets values to ensure they are correct,
but if they are not, the handlers store a -3 (0xfffffffd) failure
indication in rets[0] which corrupts host memory.

Fix this by testing up front whether the guest supplied nargs and nret
would exceed the array size, and fail the hcall directly without storing
a failure indication to rets[0].

Also expand on a comment about why we kill the guest and try not to
return errors directly if we have a valid rets[0] pointer.

Fixes: 8e591cb720 ("KVM: PPC: Book3S: Add infrastructure to implement kernel-side RTAS calls")
Cc: stable@vger.kernel.org # v3.10+
Reported-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This commit is contained in:
Nicholas Piggin 2021-07-20 20:43:09 +10:00 committed by Greg Kroah-Hartman
parent c968f563cc
commit 2b9ffddd70
1 changed files with 22 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
arch/powerpc/kvm/book3s_rtas.c
View File

@ -240,6 +240,17 @@ int kvmppc_rtas_hcall(struct kvm_vcpu *vcpu)
* value so we can restore it on the way out. * value so we can restore it on the way out.
*/ */
orig_rets = args.rets; orig_rets = args.rets;
if (be32_to_cpu(args.nargs) >= ARRAY_SIZE(args.args)) {
/*
* Don't overflow our args array: ensure there is room for
* at least rets[0] (even if the call specifies 0 nret).
*
* Each handler must then check for the correct nargs and nret
* values, but they may always return failure in rets[0].
*/
rc = -EINVAL;
goto fail;
}
args.rets = &args.args[be32_to_cpu(args.nargs)]; args.rets = &args.args[be32_to_cpu(args.nargs)];
mutex_lock(&vcpu->kvm->arch.rtas_token_lock); mutex_lock(&vcpu->kvm->arch.rtas_token_lock);
@ -267,9 +278,17 @@ int kvmppc_rtas_hcall(struct kvm_vcpu *vcpu)
fail: fail:
/* /*
* We only get here if the guest has called RTAS with a bogus * We only get here if the guest has called RTAS with a bogus
* args pointer. That means we can't get to the args, and so we * args pointer or nargs/nret values that would overflow the
* can't fail the RTAS call. So fail right out to userspace, * array. That means we can't get to the args, and so we can't
* which should kill the guest. * fail the RTAS call. So fail right out to userspace, which
* should kill the guest.
*
* SLOF should actually pass the hcall return value from the
* rtas handler call in r3, so enter_rtas could be modified to
* return a failure indication in r3 and we could return such
* errors to the guest rather than failing to host userspace.
* However old guests that don't test for failure could then
* continue silently after errors, so for now we won't do this.
*/ */
return rc; return rc;
} }