From d59c40cc483729f2e67c80e58df769ad19976fe9 Mon Sep 17 00:00:00 2001 From: Leonardo Bras Date: Mon, 11 Jul 2022 18:11:13 -0300 Subject: migration/multifd: Report to user when zerocopy not working MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Some errors, like the lack of Scatter-Gather support by the network interface(NETIF_F_SG) may cause sendmsg(...,MSG_ZEROCOPY) to fail on using zero-copy, which causes it to fall back to the default copying mechanism. After each full dirty-bitmap scan there should be a zero-copy flush happening, which checks for errors each of the previous calls to sendmsg(...,MSG_ZEROCOPY). If all of them failed to use zero-copy, then increment dirty_sync_missed_zero_copy migration stat to let the user know about it. Signed-off-by: Leonardo Bras Reviewed-by: Daniel P. Berrangé Acked-by: Peter Xu Message-Id: <20220711211112.18951-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert --- migration/ram.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'migration/ram.h') diff --git a/migration/ram.h b/migration/ram.h index 5d90945a6e..c7af65ac74 100644 --- a/migration/ram.h +++ b/migration/ram.h @@ -89,4 +89,6 @@ void ram_write_tracking_prepare(void); int ram_write_tracking_start(void); void ram_write_tracking_stop(void); +void dirty_sync_missed_zero_copy(void); + #endif -- cgit v1.2.3