diff options
author | Tim Deegan <tim@xen.org> | 2013-03-14 12:39:39 +0000 |
---|---|---|
committer | Ian Jackson <Ian.Jackson@eu.citrix.com> | 2013-03-28 15:51:39 +0000 |
commit | 650354dbc2626b643c12873275ca67782f1382c8 (patch) | |
tree | e0bef56927c688c23470f7a1639e95cfb359f928 /xen/arch/x86/hvm/rtc.c | |
parent | a5c800142cfc82159fcb85b47116cf296caebcc5 (diff) | |
download | xen-650354dbc2626b643c12873275ca67782f1382c8.tar.gz xen-650354dbc2626b643c12873275ca67782f1382c8.tar.bz2 xen-650354dbc2626b643c12873275ca67782f1382c8.zip |
libxl: run libxl__arch_domain_create() much earlier.
Among other things, arch_domain_create() sets the shadow(/hap/p2m)
memory allocation, which must happen after vcpus are assigned (or the
shadow op will fail) but before memory is allocated (or we might run
out of p2m memory).
libxl__build_pre(), which already sets similar things like maxmem,
semes like a reasonable spot for it. That needed a bit of plumbing to
get the right datastructure from the caller.
As a side-effect, the return code from libxl__arch_domain_create() is
no longer ignored.
This bug was analysed in:
From: "Jan Beulich" <JBeulich@xxxxxxxx>
"Re: [Xen-devel] [xen-unstable test] 16788: regressions - FAIL"
Date: Mon, 04 Mar 2013 16:34:53 +0000
http://lists.xen.org/archives/html/xen-devel/2013-03/msg00191.html
Reported-by: Jan Beulich <JBeulich@suse.com>
Signed-off-by: Tim Deegan <tim@xen.org>
Cc: Ian Jackson <ian.jackson@eu.citrix.com>
Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Cc: Ian Campbell <ian.campbell@citrix.com>
Diffstat (limited to 'xen/arch/x86/hvm/rtc.c')
0 files changed, 0 insertions, 0 deletions