On 10/03/20 08:06, bauerchen(陈蒙蒙) wrote: > oh ,yes.Thanks > I want to know if I submit a new fixed patch or just a patch fixed > current problem?? > if a new fixed patch, maybe need a revert ?
Sorry I missed this message. I have already sent a fixed patch, thanks! Paolo > ------------------------------------------------------------------------ > bauerchen(陈蒙蒙) > > > *From:* Igor Mammedov <mailto:imamm...@redhat.com> > *Date:* 2020-03-09 21:19 > *To:* bauerchen(陈蒙蒙) <mailto:bauerc...@tencent.com> > *CC:* borntraeger <mailto:borntrae...@de.ibm.com>; pbonzini > <mailto:pbonz...@redhat.com>; qemu-devel > <mailto:qemu-devel@nongnu.org>; qemu-s390x > <mailto:qemu-s3...@nongnu.org>; mhartmay <mailto:mhart...@linux.ibm.com> > *Subject:* Re: [PATCH] mem-prealloc: initialize cond and > mutex(Internet mail) > On Mon, 9 Mar 2020 11:16:10 +0000 > bauerchen(陈蒙蒙) <bauerc...@tencent.com> wrote: > > > Thanks, in fact,do_touch_pages is called just when vm starts up, > but using init flag and Gonce maybe more elegant ! > > if needed,I can submit a new patch ! > > thanks very much! > > it's called from os_mem_prealloc() -> touch_all_pages() which is called > at least once per an instance of hotsmem backend. So if several backends > are used then it should be called several times. > The same applies when a hostmem backend is added during runtime > (hotplug) > > >
