Re: [Qemu-devel] [PATCH] RDMA: Reduce restriction on block length match

[Juan Quintela <quintela@redhat.com>]

"Dr. David Alan Gilbert (git)" <dgilbert@redhat.com> wrote:
> From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
>
> My e4d633207 patch has an over zealous sanity check that checked
> the lengths of the RAM Blocks on source/destination were the same.  This
> isn't true because of the 'used_length' trick for RAM blocks like the
> ACPI table that vary in size.
>
> Prior to that patch RDMA would also fail in this case, but it should
> now work with the changes in the set e4d633207 is in.
>
> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
>
> Fixes: e4d633207c129dc5b7d145240ac4a1997ef3902f
> ---
>  migration/rdma.c | 13 +++++++------
>  1 file changed, 7 insertions(+), 6 deletions(-)
>
> diff --git a/migration/rdma.c b/migration/rdma.c
> index f106b2a..1d094b0 100644
> --- a/migration/rdma.c
> +++ b/migration/rdma.c
> @@ -3338,14 +3338,15 @@ static int qemu_rdma_registration_stop(QEMUFile *f, void *opaque,
>          for (i = 0; i < nb_dest_blocks; i++) {
>              network_to_dest_block(&rdma->dest_blocks[i]);
>  
> -            /* We require that the blocks are in the same order */
> +            /* We require that the blocks are in the same order,
> +             * but the used_length trick for acpi blocks means that
> +             * the destination can validly be larger than the source
> +             */
>              if (rdma->dest_blocks[i].length != local->block[i].length) {

Should we change the check to be that destination is bigger or equal
than source?

With your change, we only remove the check?

Thanks, Juan.


> -                ERROR(errp, "Block %s/%d has a different length %" PRIu64
> -                            "vs %" PRIu64, local->block[i].block_name, i,
> -                            local->block[i].length,
> +                fprintf(stderr, "INFO: Block %s/%d has a different length %"
> +                            PRIu64 "vs %" PRIu64, local->block[i].block_name,
> +                            i, local->block[i].length,
>                              rdma->dest_blocks[i].length);
> -                rdma->error_state = -EINVAL;
> -                return -EINVAL;
>              }
>              local->block[i].remote_host_addr =
>                      rdma->dest_blocks[i].remote_host_addr;