Thank you for you review. I will modify patches according to your opinion. Jinhao Gao
-----Original Message----- From: Michael S. Tsirkin [mailto:m...@redhat.com] Sent: 2020年12月27日 21:20 To: gaojinhao <gaojin...@huawei.com> Cc: qemu-devel@nongnu.org; qemu-...@nongnu.org; Marc-André Lureau <marcandre.lur...@redhat.com>; Stefan Berger <stef...@linux.vnet.ibm.com>; Jason Wang <jasow...@redhat.com>; David Gibson <da...@gibson.dropbear.id.au>; Greg Kurz <gr...@kaod.org>; Juan Quintela <quint...@redhat.com>; Dr . David Alan Gilbert <dgilb...@redhat.com>; Wanghaibin (D) <wanghaibin.w...@huawei.com>; zhukeqian <zhukeqi...@huawei.com> Subject: Re: [PATCH 0/8] Fix memory leak of some device state in migration On Sat, Dec 26, 2020 at 06:33:39PM +0800, g00517791 wrote: > From: Jinhao Gao <gaojin...@huawei.com> > > For some device state having some fields of VMS_ALLOC flag, they don't > free memory allocated for the fields in vmstate_save_state and vmstate > _load_state. We add funcs or sentences of free memory before > allocation of memory or after load of memory to avoid memory leak. > Isn't there a way to handle it centrally? IIUC the issue is repeated loads in case a load fails, right? So can't we do something along the lines of: Signed-off-by: Michael S. Tsirkin <m...@redhat.com> diff --git a/migration/vmstate.c b/migration/vmstate.c index e9d2aef66b..873f76739f 100644 --- a/migration/vmstate.c +++ b/migration/vmstate.c @@ -70,6 +70,7 @@ static void vmstate_handle_alloc(void *ptr, const VMStateField *field, gsize size = vmstate_size(opaque, field); size *= vmstate_n_elems(opaque, field); if (size) { + g_free(*(void **)ptr); *(void **)ptr = g_malloc(size); } } -- MST
