From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-ed1-f51.google.com (mail-ed1-f51.google.com [209.85.208.51]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id BF9D13E3D8D for ; Thu, 23 Apr 2026 09:17:58 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.208.51 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1776935885; cv=none; b=iIOUxwRUbn8k4EyNZ6+Tc+Gvz9ijatBViDWVsR4eOOgjBbTWQsMneDzTT0h1oljoM8YR1wTQbEizM9RpWnbypi6Bk4kkk2xuhQP6kOpc9lrW6OU/0/EvqKL5S5cDR4OynQ/j9UoifDb+P429sP4Z78k5jBE4VRGSNps+kZqDdJ0= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1776935885; c=relaxed/simple; bh=AZoirHf/5mE5YFev+nodzuKCTioKK86YrjX4J7mCnGY=; h=Date:From:To:Cc:Subject:Message-ID:References:MIME-Version: Content-Type:Content-Disposition:In-Reply-To; b=aoc25O6F1GjmhO23ceKnK/RFDvdfMvkRGh/m+G2cWJ77F5NdmvtN+a2BR/ZVbjsy1jubwES8/FrOFqa2Hw1wwhoI+g1v9KORJAlDLGjjzAIYAdTZlgULJnCodgLV4OnLH693V/lTsnkcrCI2CKmshp2u+6j0ofsQLemEdEcKGww= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=AN52BcQO; arc=none smtp.client-ip=209.85.208.51 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="AN52BcQO" Received: by mail-ed1-f51.google.com with SMTP id 4fb4d7f45d1cf-67148a70f8aso49601a12.1 for ; Thu, 23 Apr 2026 02:17:58 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20251104; t=1776935874; x=1777540674; darn=vger.kernel.org; h=in-reply-to:content-disposition:mime-version:references:message-id :subject:cc:to:from:date:from:to:cc:subject:date:message-id:reply-to; bh=jzw3UnIwHGir8Kj1NlO4m24J9zaCTTh7cxmm2qnVabM=; b=AN52BcQO4Oym4EIS/ybTJU4X0xMmH97zeiNTelL37EawZBYdiu86sX1EIB3vYscAo0 Vcjf64iWxDZFRIgV+BUgPXwjn17ZxWf3x0zvD8PbwUSQdwsXOHXwx6QOZbBNlv+1eTfZ JlZSD8s8hyn9KGC5pumgaqIdPFAm9lkg81pMdNii3VqCoQ6kBqRfR1xVL79QIEpjYl3N KcS7eoYW3iXshaW0YENjeMePfK0vTCHe9LT2oLzygIfBJzJwOQaLeBDNqxkOWERekKAc yRP5mUdzoFDc7O8zjImnYAlKZo+/Gx9l3am2T7A/2JNf7Lh4hZbZweFdIPcyWM0FVNm3 6nWQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1776935874; x=1777540674; h=in-reply-to:content-disposition:mime-version:references:message-id :subject:cc:to:from:date:x-gm-gg:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=jzw3UnIwHGir8Kj1NlO4m24J9zaCTTh7cxmm2qnVabM=; b=en1rwB61cI8v71NjQ9nwXPUN6Ef/yOLNTS9Hs5SyBFed3kHDghqTGvazbvvj+M+ULj /N27fsE+knj1xApVA+KZpRL+w8FIeynZe9s4R/0l78mKUJ/vptZ8DzTclatj/iATFF1r PMi4AjSYAzp6QGFBl9ETTJUXpRiXRRyIRfUfht/6OX8mrizpci4thWT9hWOJxNej9CQ8 gxM5lQ/js1fSrEPIg0fh0yA5gBXQG24++GykJCzQusj0j5jNxWkeMkq99YuXuvD7CNjD 8EBGv0Q59Gz0010L1V51B2osKWj+a3MvyD2QVzHVMHdqGhU2VWbbdUltDSdV0xuXRDuG dI5A== X-Forwarded-Encrypted: i=1; AFNElJ+YMxqJd1eYXP4W0ULrl55ev1olfTIzvuPYSNfmGI4j/peHyHqBznqPIOe36Q2VtJ7eKxMB97MVLBr3rOE=@vger.kernel.org X-Gm-Message-State: AOJu0YzGu86y5sJBeuKaJHDr2G1jxsGjP4hr1OrGuizItWDo380Tmpth jVfKkPWJc/xlBgzqTaCoRlc+jZ3RJGyPeIe/5hFbTpMIyvPxcCkv+fDmbUSZvbEmhg== X-Gm-Gg: AeBDietQLWlgJ7NLUQvxNJIDakfrTQEeLYUVdxLlMpyMwVgqs3iNnL2V6MqAjdRg0W7 CT2V8yMDGaYfBupbv/t75YoC1Ps7cau+S+DhY0tAO91Hge20jBcm2Ih/YdiVHinGTCl/HOpw5dS IXwLXztVRcTl/OJ16AW467J8zRpappQiFBQstHU/enKVEIPgmRNSA3iwmimOc55MfDhAoET0QOF D5AN7dOHLCfThRIUxxk//yyTxUCZW9HnWiuG8C1XurTjR2RS3pxvPuEuJAhF+O7rlxm+k++AsLU 9gQDsaOj9lL0Hp1S7sTvfMkwKZ+2Cdze/idDoIWTwVkl9i6YO2Wj8ZVgahUHlhAP2hbj/PJ76W5 WGkPoFZ5a77G20VMavY/zbF1EuLM777O26/UZBLlk7znRH2WHcom2YZyznv3PZKPxA/Zxj5b7+i cy5m/hcbeMUYTJ/6htJMweskdS3iblEmSgdB0UHwXC+xQkCyb+7RcOwIMwmg/hRVAxbnYzrp2WF OqWLfMKd7FGcyFqTzw= X-Received: by 2002:a05:6402:3085:b0:671:dad9:8caf with SMTP id 4fb4d7f45d1cf-6744ed5f907mr227247a12.5.1776935873820; Thu, 23 Apr 2026 02:17:53 -0700 (PDT) Received: from google.com (117.15.199.104.bc.googleusercontent.com. [104.199.15.117]) by smtp.gmail.com with ESMTPSA id a640c23a62f3a-ba451210e49sm655436766b.10.2026.04.23.02.17.52 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 23 Apr 2026 02:17:53 -0700 (PDT) Date: Thu, 23 Apr 2026 09:17:49 +0000 From: Sebastian Ene To: Sudeep Holla Cc: Marc Zyngier , oupton@kernel.org, will@kernel.org, ayrton@google.com, catalin.marinas@arm.com, joey.gouly@arm.com, korneld@google.com, kvmarm@lists.linux.dev, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, android-kvm@google.com, mrigendra.chaubey@gmail.com, perlarsen@google.com, suzuki.poulose@arm.com, yuzenghui@huawei.com, stable@vger.kernel.org Subject: Re: [PATCH] KVM: arm64: Validate the FF-A memory access descriptor placement Message-ID: References: <20260422102540.1433704-1-sebastianene@google.com> <86bjfb18v1.wl-maz@kernel.org> <20260422-jolly-curassow-of-amplitude-25fbaf@sudeepholla> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20260422-jolly-curassow-of-amplitude-25fbaf@sudeepholla> On Wed, Apr 22, 2026 at 08:29:06PM +0100, Sudeep Holla wrote: > On Wed, Apr 22, 2026 at 01:35:55PM +0000, Sebastian Ene wrote: > > On Wed, Apr 22, 2026 at 01:24:02PM +0100, Marc Zyngier wrote: > > > On Wed, 22 Apr 2026 11:25:40 +0100, > > > Sebastian Ene wrote: > > > > > > > > Prevent the pKVM hypervisor from making assumptions that the > > > > endpoint memory access descriptor (EMAD) comes right after the > > > > FF-A memory region header and enforce a strict placement for it > > > > when validating an FF-A memory lend/share transaction. > > > > Hello Marc, > > > > > > > > As I read this, you want to remove a bad assumption... > > > > > > > > > > > Prior to FF-A version 1.1 the header of the memory region > > > > didn't contain an offset to the endpoint memory access descriptor. > > > > The layout of a memory transaction looks like this: > > > > > > > > Field name | Offset > > > > -- 0 > > > > [ Header (ffa_mem_region) |__ ep_mem_offset > > > > EMAD 1 (ffa_mem_region_attributes) | > > > > ] > > > > > > > > Reject the host from specifying a memory access descriptor offset > > > > that is different than the size of the memory region header. > > > > > > And yet you decide that you want to enforce this assumption. I don't > > > understand how you arrive to this conclusion. > > > > > > Looking at the spec, it appears that the offset is *designed* to allow > > > a gap between the header and the EMAD. Refusing to handle a it seems to be a > > > violation of the spec. > > > > > > What am I missing? > > > > While the spec allows the gap to be variable (since version 1.1), the > > arm ff-a driver places it at a fixed position in: > > ffa_mem_region_additional_setup() > > https://elixir.bootlin.com/linux/v7.0/source/drivers/firmware/arm_ffa/driver.c#L671 > > > Hello Sudeep, > That's just the current choice in the driver and can be changed in the future. > > > and makes use of the same assumption in: ffa_mem_desc_offset(). > > https://elixir.bootlin.com/linux/v7.0/source/include/linux/arm_ffa.h#L448 > > Again this is just in the transmit path of the message the driver is > constructing and hence it is a simple choice rather than wrong assumption. > > > The later one seems wrong IMO. because we should compute the offset > > based on the value stored in ep_mem_offset and not adding it up with > > sizeof(struct ffa_mem_region). > > > > Sorry what am I missing as the driver is building these descriptors to > send it across to SPMC, we are populating the field and it will be 0 > before it is initialised Right, what I meant is having something like this since this function is not limited to the driver scope and using it from other components would imply relying on the assumption: 'ep_mem_offset == sizeof(struct ffa_mem_region)'. We will also have to validate that the `ep_mem_offset` doesn't point outside of the mailbox designated buffer. --- diff --git a/include/linux/arm_ffa.h b/include/linux/arm_ffa.h index 81e603839c4a..62d67dae8b70 100644 --- a/include/linux/arm_ffa.h +++ b/include/linux/arm_ffa.h @@ -445,7 +445,7 @@ ffa_mem_desc_offset(struct ffa_mem_region *buf, int count, u32 ffa_version) if (!FFA_MEM_REGION_HAS_EP_MEM_OFFSET(ffa_version)) offset += offsetof(struct ffa_mem_region, ep_mem_offset); else - offset += sizeof(struct ffa_mem_region); + offset += buf->ep_mem_offset; return offset; } --- And then move `ffa_mem_region_additional_setup` to be called earlier before `ffa_mem_desc_offset`: (so that it can setup the value for ep_mem_offset) --- diff --git a/drivers/firmware/arm_ffa/driver.c b/drivers/firmware/arm_ffa/driver.c index f2f94d4d533e..66de59c88aff 100644 --- a/drivers/firmware/arm_ffa/driver.c +++ b/drivers/firmware/arm_ffa/driver.c @@ -691,6 +691,8 @@ ffa_setup_and_transmit(u32 func_id, void *buffer, u32 max_fragsize, mem_region->flags = args->flags; mem_region->sender_id = drv_info->vm_id; mem_region->attributes = ffa_memory_attributes_get(func_id); + + ffa_mem_region_additional_setup(drv_info->version, mem_region); composite_offset = ffa_mem_desc_offset(buffer, args->nattrs, drv_info->version); @@ -708,7 +710,6 @@ ffa_setup_and_transmit(u32 func_id, void *buffer, u32 max_fragsize, } mem_region->handle = 0; mem_region->ep_count = args->nattrs; - ffa_mem_region_additional_setup(drv_info->version, mem_region); --- > > > Maybe this should be the fix instead and not the one in pKVM ? What do > > you think ? > > > > Can you share the diff you have in mind to understand your concern better > or are you referring to this patch itself. Sure, please let me know if you think this is wrong. I might have misunderstood it. > > > The current implementation in pKVM makes use of the > > ffa_mem_desc_offset() to validate the first EMAD. If a compromised host > > places an EMAD at a different offset than sizeof(struct ffa_mem_region), > > then pKVM will not validate that EMAD. > > > > Calling the host as compromised if it chooses a different offset seems bit > of extreme here. I am no sure if I am missing to understand something here. > Sorry for not explaining it, in pKVM model we don't trust the host kernel so we can assume that everything that doesn't pass the hypervisor validation(in this case the ff-a memory transaction) can be a potential attack that wants to compromise EL2. > -- > Regards, > Sudeep Thanks, Sebastian