Re: [PATCH v7 43/45] virt: Add SEV-SNP guest driver

From: Brijesh Singh <brijesh.singh@amd.com>
To: Peter Gonda <pgonda@google.com>
Cc: brijesh.singh@amd.com, x86@kernel.org,
	linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
	linux-efi@vger.kernel.org, platform-driver-x86@vger.kernel.org,
	linux-coco@lists.linux.dev, linux-mm@kvack.org,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>, Joerg Roedel <jroedel@suse.de>,
	Tom Lendacky <Thomas.Lendacky@amd.com>,
	"H. Peter Anvin" <hpa@zytor.com>,
	Ard Biesheuvel <ardb@kernel.org>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Sean Christopherson <seanjc@google.com>,
	Vitaly Kuznetsov <vkuznets@redhat.com>,
	Jim Mattson <jmattson@google.com>,
	Andy Lutomirski <luto@kernel.org>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	Sergio Lopez <slp@redhat.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>,
	David Rientjes <rientjes@google.com>,
	Dov Murik <dovmurik@linux.ibm.com>,
	Tobin Feldman-Fitzthum <tobin@ibm.com>,
	Borislav Petkov <bp@alien8.de>,
	Michael Roth <michael.roth@amd.com>,
	Vlastimil Babka <vbabka@suse.cz>,
	"Kirill A . Shutemov" <kirill@shutemov.name>,
	Andi Kleen <ak@linux.intel.com>,
	"Dr . David Alan Gilbert" <dgilbert@redhat.com>,
	tony.luck@intel.com, marcorr@google.com,
	sathyanarayanan.kuppuswamy@linux.intel.com
Subject: Re: [PATCH v7 43/45] virt: Add SEV-SNP guest driver
Date: Fri, 19 Nov 2021 18:28:20 -0600	[thread overview]
Message-ID: <12208d99-ef91-953c-7511-34f7e2cbdceb@amd.com> (raw)
In-Reply-To: <CAMkAt6pAcM-+odnagFTiaY7PPGE1CfAt27x=tG=-4UU9c+dQXA@mail.gmail.com>


On 11/19/21 10:16 AM, Peter Gonda wrote:
> On Thu, Nov 18, 2021 at 10:32 AM Brijesh Singh <brijesh.singh@amd.com> wrote:
>>
>>
>> On 11/17/21 5:34 PM, Peter Gonda wrote:
>>
>>
>>>> +The guest ioctl should be issued on a file descriptor of the /dev/sev-guest device.
>>>> +The ioctl accepts struct snp_user_guest_request. The input and output structure is
>>>> +specified through the req_data and resp_data field respectively. If the ioctl fails
>>>> +to execute due to a firmware error, then fw_err code will be set.
>>> Should way say what it will be set to? Also Sean pointed out on CCP
>>> driver that 0 is strange to set the error to, its a uint so we cannot
>>> do -1 like we did there. What about all FFs?
>>>
>> Sure, all FF's works, I can document and use it.
>>
>>
>>>> +static inline u64 __snp_get_msg_seqno(struct snp_guest_dev *snp_dev)
>>>> +{
>>>> +       u64 count;
>>> I may be overly paranoid here but how about
>>> `lockdep_assert_held(&snp_cmd_mutex);` when writing or reading
>>> directly from this data?
>>>
>> Sure, I can do it.
>>
>> ...
>>
>>>> +
>>>> +       if (rc)
>>>> +               return rc;
>>>> +
>>>> +       rc = verify_and_dec_payload(snp_dev, resp_buf, resp_sz);
>>>> +       if (rc) {
>>>> +               /*
>>>> +                * The verify_and_dec_payload() will fail only if the hypervisor is
>>>> +                * actively modifiying the message header or corrupting the encrypted payload.
>>> modifiying
>>>> +                * This hints that hypervisor is acting in a bad faith. Disable the VMPCK so that
>>>> +                * the key cannot be used for any communication.
>>>> +                */
>>> This looks great, thanks for changes Brijesh. Should we mention in
>>> comment here or at snp_disable_vmpck() the AES-GCM issues with
>>> continuing to use the key? Or will future updaters to this code
>>> understand already?
>>>
>> Sure, I can add comment about the AES-GCM.
>>
>> ...
>>
>>>> +
>>>> +/* See SNP spec SNP_GUEST_REQUEST section for the structure */
>>>> +enum msg_type {
>>>> +       SNP_MSG_TYPE_INVALID = 0,
>>>> +       SNP_MSG_CPUID_REQ,
>>>> +       SNP_MSG_CPUID_RSP,
>>>> +       SNP_MSG_KEY_REQ,
>>>> +       SNP_MSG_KEY_RSP,
>>>> +       SNP_MSG_REPORT_REQ,
>>>> +       SNP_MSG_REPORT_RSP,
>>>> +       SNP_MSG_EXPORT_REQ,
>>>> +       SNP_MSG_EXPORT_RSP,
>>>> +       SNP_MSG_IMPORT_REQ,
>>>> +       SNP_MSG_IMPORT_RSP,
>>>> +       SNP_MSG_ABSORB_REQ,
>>>> +       SNP_MSG_ABSORB_RSP,
>>>> +       SNP_MSG_VMRK_REQ,
>>>> +       SNP_MSG_VMRK_RSP,
>>> Did you want to include MSG_ABSORB_NOMA_REQ and MSG_ABSORB_NOMA_RESP here?
>>>
>> Yes, I can includes those for the completeness.
>>
>> ...
>>
>>>> +struct snp_report_req {
>>>> +       /* message version number (must be non-zero) */
>>>> +       __u8 msg_version;
>>>> +
>>>> +       /* user data that should be included in the report */
>>>> +       __u8 user_data[64];
>>> Are we missing the 'vmpl' field here? Does those default all requests
>>> to be signed with VMPL0? Users might want to change that, they could
>>> be using a paravisor.
>>>
>> Good question, so far I was thinking that guest kernel will provide its
>> vmpl level instead of accepted the vmpl level from the userspace. Do you
>> see a need for a userspace to provide this information ?
> That seems fine. I am just confused because we are just encrypting
> this struct as the payload for the PSP. Doesn't the message require a
> struct that looks like 'snp_report_req_user_data' below?
>
> snp_report_req{
>        /* message version number (must be non-zero) */
>        __u8 msg_version;
>
>       /* user data that should be included in the report */
>        struct snp_report_req_user_data;
> };
>
> struct snp_report_req_user_data {
>   u8 user_data[64];
>   u32 vmpl;
>   u32 reserved;
> };
>
The snp_guest_msg structure is zero'ed before building the hdr and
copying the user provided input, see enc_payload. The patch series was
focused on vmpl-0 only I didn't consider anything other than vmpl-. Let
me work to provide the option for userspace to provide the vmpl as an
input during the request so that we give the flexibility to userspace.


>>
>> thanks