Re: [Qemu-devel] [PATCH 5/5] migration: refine ram_save_compressed_page

[Juan Quintela <quintela@redhat.com>]

Liang Li <liang.z.li@intel.com> wrote:
> Use qemu_put_compression_data to do the compression directly
> instead of using do_compress_ram_page, avoid some data copy.
> very small improvement, at the same time, add code to check
> if the compression is successful.
>
> Signed-off-by: Liang Li <liang.z.li@intel.com>
> ---
>  migration/ram.c | 28 +++++++++++++---------------
>  1 file changed, 13 insertions(+), 15 deletions(-)
>
> diff --git a/migration/ram.c b/migration/ram.c
> index 7e62d8d..ec2c0bf 100644
> --- a/migration/ram.c
> +++ b/migration/ram.c
> @@ -927,24 +927,20 @@ static int ram_save_compressed_page(QEMUFile *f, PageSearchStatus *pss,
>                                      uint64_t *bytes_transferred)
>  {
>      int pages = -1;
> -    uint64_t bytes_xmit;
> +    uint64_t bytes_xmit = 0;
>      uint8_t *p;
> -    int ret;
> +    int ret, blen;
>      RAMBlock *block = pss->block;
>      ram_addr_t offset = pss->offset;
>  
>      p = block->host + offset;
>  
> -    bytes_xmit = 0;
>      ret = ram_control_save_page(f, block->offset,
>                                  offset, TARGET_PAGE_SIZE, &bytes_xmit);
>      if (bytes_xmit) {
>          *bytes_transferred += bytes_xmit;
>          pages = 1;
>      }
> -    if (block == last_sent_block) {
> -        offset |= RAM_SAVE_FLAG_CONTINUE;
> -    }
>      if (ret != RAM_SAVE_CONTROL_NOT_SUPP) {
>          if (ret != RAM_SAVE_CONTROL_DELAYED) {
>              if (bytes_xmit > 0) {
> @@ -964,17 +960,19 @@ static int ram_save_compressed_page(QEMUFile *f, PageSearchStatus *pss,
>              flush_compressed_data(f);
>              pages = save_zero_page(f, block, offset, p, bytes_transferred);
>              if (pages == -1) {
> -                set_compress_params(&comp_param[0], block, offset);
> -                /* Use the qemu thread to compress the data to make sure the
> -                 * first page is sent out before other pages
> -                 */
> -                bytes_xmit = do_compress_ram_page(&comp_param[0]);
> -                acct_info.norm_pages++;
> -                qemu_put_qemu_file(f, comp_param[0].file);
> -                *bytes_transferred += bytes_xmit;
> -                pages = 1;
> +                /* Make sure the first page is sent out before other pages */
> +                bytes_xmit = save_page_header(f, block, offset |
> +                                              RAM_SAVE_FLAG_COMPRESS_PAGE);
> +                blen = qemu_put_compression_data(f, p, TARGET_PAGE_SIZE,
> +                                                 migrate_compress_level());
> +                if (blen > 0) {
> +                    *bytes_transferred += bytes_xmit + blen;
> +                    acct_info.norm_pages++;
> +                    pages = 1;
> +                }
>              }
>          } else {
> +            offset |= RAM_SAVE_FLAG_CONTINUE;
>              pages = save_zero_page(f, block, offset, p, bytes_transferred);
>              if (pages == -1) {
>                  pages = compress_page_with_multi_thread(f, block, offset,

You need to decide what to do when blen <= 0.  As said in previous
patch, I can only think of aborting migration.  But ignoring the error
is not an option IMHO.

Later, Juan.