lkml.org 
[lkml]   [2021]   [Nov]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH v7 43/45] virt: Add SEV-SNP guest driver
From
Date


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 ?


thanks

\
 
 \ /
  Last update: 2021-11-18 18:33    [W:0.531 / U:0.900 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site