Re: [PATCH v20 10/15] migration/channel: check that transfer is UNIX socket when "local" set

On 29.07.26 17:09, Peter Xu wrote:
> On Wed, Jul 29, 2026 at 12:12:48PM +0300, Vladimir Sementsov-Ogievskiy wrote:
>> As documented, for "local", the migration channel must be direct
>> UNIX socket connection from source to target. We can't check for
>> it being "direct", but let's at least check that we deal with
>> UNIX socket (fd-passing supported).
>>
>> Signed-off-by: Vladimir Sementsov-Ogievskiy <[email protected]>
>> ---
>>   migration/channel.c | 17 +++++++++++++++++
>>   1 file changed, 17 insertions(+)
>>
>> diff --git a/migration/channel.c b/migration/channel.c
>> index 1e2935f926b..537673c29e9 100644
>> --- a/migration/channel.c
>> +++ b/migration/channel.c
>> @@ -214,6 +214,13 @@ void migration_channel_process_incoming(QIOChannel *ioc)
>>       trace_migration_set_incoming_channel(
>>           ioc, object_get_typename(OBJECT(ioc)));
>>   
>> +    if (migrate_local() && migrate_mode() != MIG_MODE_CPR_EXEC &&
>> +        !qio_channel_has_feature(ioc, QIO_CHANNEL_FEATURE_FD_PASS)) {
>> +        error_setg(&local_err,
>> +                   "local migration requires a UNIX domain socket channel");
>> +        goto out;
>> +    }
>> +
>>       if (migrate_channel_requires_tls_upgrade(ioc)) {
>>           migration_tls_channel_process_incoming(ioc, &local_err);
>>       } else {
>> @@ -241,6 +248,16 @@ void migration_channel_connect_outgoing(MigrationState *s, QIOChannel *ioc)
>>   {
>>       trace_migration_set_outgoing_channel(ioc, object_get_typename(OBJECT(ioc)));
>>   
>> +    if (migrate_local() && migrate_mode() != MIG_MODE_CPR_EXEC &&
>> +        !qio_channel_has_feature(ioc, QIO_CHANNEL_FEATURE_FD_PASS)) {
>> +        Error *local_err = NULL;
>> +
>> +        error_setg(&local_err,
>> +                   "local migration requires a UNIX domain socket channel");
>> +        migration_connect_error_propagate(s, local_err);
>> +        return;
>> +    }
>> +
> 
> We only need to touch migration_capabilities_and_transport_compatible()?
> It should work for both, and fail QMP commands directly might also be
> slightly better.

Hm. Don't remember why I didn't go this way. Will try.

> 
>>       if (migrate_channel_requires_tls_upgrade(ioc)) {
>>           Error *local_err = NULL;
>>   
>> -- 
>> 2.43.0
>>
> 


-- 
Best regards,
Vladimir