memory allocated for guest before QEMU is demonized and then mapped
later in guest's addess space after it is demonized, leads to EPT
violation and QEMU abort.
To avoid this and similar issues switch to deamonized mode early
before processing applying/processing other options.
Signed-off-by: Igor Mammedov <imamm...@redhat.com>
---
I haven't digged in kvm itsefl yet why EPT violation happens,
but fix on QEMU side looks trivial so it won't hurt to use it anyway.
there is simpler reporoducer without migration:
start guest with
'-object memory-ram,id=foo,size=1G'
when guest is up and running hotplug dimm device using 'foo' memdev
as backend.
---
vl.c | 15 ++++++++-------
1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/vl.c b/vl.c
index ca9db1c..19f731d 100644
--- a/vl.c
+++ b/vl.c
@@ -4012,6 +4012,14 @@ int main(int argc, char **argv, char **envp)
}
loc_set_none();
+ os_daemonize();
+
+ if (pid_file && qemu_create_pidfile(pid_file) != 0) {
+ os_pidfile_error();
+ exit(1);
+ }
+
+
if (qemu_init_main_loop()) {
fprintf(stderr, "qemu_init_main_loop failed\n");
exit(1);
@@ -4249,13 +4257,6 @@ int main(int argc, char **argv, char **envp)
}
#endif
- os_daemonize();
-
- if (pid_file && qemu_create_pidfile(pid_file) != 0) {
- os_pidfile_error();
- exit(1);
- }
-
/* store value for the future use */
qemu_opt_set_number(qemu_find_opts_singleton("memory"), "size", ram_size);
--
1.9.0
