Re: [PATCH v4 08/23] multifd: Move iov from pages to params

[Juan Quintela]

"Dr. David Alan Gilbert" <dgilbert@redhat.com> wrote:
> * Juan Quintela (quintela@redhat.com) wrote:
>> This will allow us to reduce the number of system calls on the next patch.
>> 
>> Signed-off-by: Juan Quintela <quintela@redhat.com>
>> ---
>>  migration/multifd.h |  8 ++++++--
>>  migration/multifd.c | 34 ++++++++++++++++++++++++----------
>>  2 files changed, 30 insertions(+), 12 deletions(-)
>> 
>> diff --git a/migration/multifd.h b/migration/multifd.h
>> index e57adc783b..c3f18af364 100644
>> --- a/migration/multifd.h
>> +++ b/migration/multifd.h
>> @@ -62,8 +62,6 @@ typedef struct {
>>      uint64_t packet_num;
>>      /* offset of each page */
>>      ram_addr_t *offset;
>> -    /* pointer to each page */
>> -    struct iovec *iov;
>>      RAMBlock *block;
>>  } MultiFDPages_t;
>>  
>> @@ -110,6 +108,10 @@ typedef struct {
>>      uint64_t num_pages;
>>      /* syncs main thread and channels */
>>      QemuSemaphore sem_sync;
>> +    /* buffers to send */
>> +    struct iovec *iov;
>> +    /* number of iovs used */
>> +    uint32_t iovs_num;
>>      /* used for compression methods */
>>      void *data;
>>  }  MultiFDSendParams;
>> @@ -149,6 +151,8 @@ typedef struct {
>>      uint64_t num_pages;
>>      /* syncs main thread and channels */
>>      QemuSemaphore sem_sync;
>> +    /* buffers to recv */
>> +    struct iovec *iov;
>
> Why is there the asymmetry between send and recv, where the send
> has the iovs_num and the recv doesn't?

When we are sending data, we have the normal page and the iov, so it is
normal_pages + 1.  On reception side, we have to read first the header,
because that is where normal_pages is stored.

I can drop iovs_num on the send side and add a comment, but I think that
the new variable is more descriptive.

Or I can add iovs_num to the recv_side and just do a iovs_num =
normal_pages, but it seems a bit pointless, no?

Later, Juan.