[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 26/38] migration: Stop migration immediately in RDMA error paths
From: |
Juan Quintela |
Subject: |
[PULL 26/38] migration: Stop migration immediately in RDMA error paths |
Date: |
Tue, 31 Oct 2023 10:01:30 +0100 |
From: Peter Xu <peterx@redhat.com>
In multiple places, RDMA errors are handled in a strange way, where it only
sets qemu_file_set_error() but not stop the migration immediately.
It's not obvious what will happen later if there is already an error. Make
all such failures stop migration immediately.
Cc: Zhijian Li (Fujitsu) <lizhijian@fujitsu.com>
Cc: Markus Armbruster <armbru@redhat.com>
Cc: Juan Quintela <quintela@redhat.com>
Cc: Fabiano Rosas <farosas@suse.de>
Reported-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231024163933.516546-1-peterx@redhat.com>
---
migration/ram.c | 21 ++++++++++-----------
1 file changed, 10 insertions(+), 11 deletions(-)
diff --git a/migration/ram.c b/migration/ram.c
index 024dedb6b1..20e6153114 100644
--- a/migration/ram.c
+++ b/migration/ram.c
@@ -2973,11 +2973,13 @@ static int ram_save_setup(QEMUFile *f, void *opaque)
ret = rdma_registration_start(f, RAM_CONTROL_SETUP);
if (ret < 0) {
qemu_file_set_error(f, ret);
+ return ret;
}
ret = rdma_registration_stop(f, RAM_CONTROL_SETUP);
if (ret < 0) {
qemu_file_set_error(f, ret);
+ return ret;
}
migration_ops = g_malloc0(sizeof(MigrationOps));
@@ -3043,6 +3045,7 @@ static int ram_save_iterate(QEMUFile *f, void *opaque)
ret = rdma_registration_start(f, RAM_CONTROL_ROUND);
if (ret < 0) {
qemu_file_set_error(f, ret);
+ goto out;
}
t0 = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
@@ -3147,8 +3150,6 @@ static int ram_save_complete(QEMUFile *f, void *opaque)
rs->last_stage = !migration_in_colo_state();
WITH_RCU_READ_LOCK_GUARD() {
- int rdma_reg_ret;
-
if (!migration_in_postcopy()) {
migration_bitmap_sync_precopy(rs, true);
}
@@ -3156,6 +3157,7 @@ static int ram_save_complete(QEMUFile *f, void *opaque)
ret = rdma_registration_start(f, RAM_CONTROL_FINISH);
if (ret < 0) {
qemu_file_set_error(f, ret);
+ return ret;
}
/* try transferring iterative blocks of memory */
@@ -3171,24 +3173,21 @@ static int ram_save_complete(QEMUFile *f, void *opaque)
break;
}
if (pages < 0) {
- ret = pages;
- break;
+ qemu_mutex_unlock(&rs->bitmap_mutex);
+ return pages;
}
}
qemu_mutex_unlock(&rs->bitmap_mutex);
compress_flush_data();
- rdma_reg_ret = rdma_registration_stop(f, RAM_CONTROL_FINISH);
- if (rdma_reg_ret < 0) {
- qemu_file_set_error(f, rdma_reg_ret);
+ ret = rdma_registration_stop(f, RAM_CONTROL_FINISH);
+ if (ret < 0) {
+ qemu_file_set_error(f, ret);
+ return ret;
}
}
- if (ret < 0) {
- return ret;
- }
-
ret = multifd_send_sync_main(rs->pss[RAM_CHANNEL_PRECOPY].pss_channel);
if (ret < 0) {
return ret;
--
2.41.0
- [PULL 11/38] migration: Simplify compress_page_with_multithread(), (continued)
- [PULL 11/38] migration: Simplify compress_page_with_multithread(), Juan Quintela, 2023/10/31
- [PULL 18/38] migration/ram: Fix compilation with -Wshadow=local, Juan Quintela, 2023/10/31
- [PULL 13/38] migration: Create compress_update_rates(), Juan Quintela, 2023/10/31
- [PULL 19/38] migration: rename vmstate_save_needed->vmstate_section_needed, Juan Quintela, 2023/10/31
- [PULL 20/38] migration: set file error on subsection loading, Juan Quintela, 2023/10/31
- [PULL 21/38] qemu-iotests: Filter warnings about block migration being deprecated, Juan Quintela, 2023/10/31
- [PULL 16/38] migration: Merge flush_compressed_data() and compress_flush_data(), Juan Quintela, 2023/10/31
- [PULL 17/38] migration: Rename ram_compressed_pages() to compress_ram_pages(), Juan Quintela, 2023/10/31
- [PULL 22/38] migration: migrate 'inc' command option is deprecated., Juan Quintela, 2023/10/31
- [PULL 23/38] migration: migrate 'blk' command option is deprecated., Juan Quintela, 2023/10/31
- [PULL 26/38] migration: Stop migration immediately in RDMA error paths,
Juan Quintela <=
- [PULL 27/38] qemu-file: Don't increment qemu_file_transferred at qemu_file_fill_buffer, Juan Quintela, 2023/10/31
- [PULL 29/38] qemu_file: total_transferred is not used anymore, Juan Quintela, 2023/10/31
- [PULL 25/38] migration: Deprecate old compression method, Juan Quintela, 2023/10/31
- [PULL 24/38] migration: Deprecate block migration, Juan Quintela, 2023/10/31
- [PULL 28/38] qemu_file: Use a stat64 for qemu_file_transferred, Juan Quintela, 2023/10/31
- [PULL 32/38] qemu-file: Remove _noflush from qemu_file_transferred_noflush(), Juan Quintela, 2023/10/31
- [PULL 34/38] migration: migration_rate_limit_reset() don't need the QEMUFile, Juan Quintela, 2023/10/31
- [PULL 30/38] migration: Use the number of transferred bytes directly, Juan Quintela, 2023/10/31
- [PULL 33/38] migration: migration_transferred_bytes() don't need the QEMUFile, Juan Quintela, 2023/10/31
- [PULL 36/38] migration: Use migration_transferred_bytes(), Juan Quintela, 2023/10/31