[libvirt] [PATCH] Properly advertise cpuselection guest capability
There's no sense in advertising cpuselection capability when host CPU is not properly detected and advertised in host capabilities. --- src/qemu/qemu_conf.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 3fb0074..ff5bd43 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -892,7 +892,8 @@ qemudCapsInitGuest(virCapsPtr caps, guest->arch.defaultInfo.emulator_mtime = binary_mtime; - if (qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 + if (caps->host.cpu + && qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 && ncpus > 0 && !virCapabilitiesAddGuestFeature(guest, "cpuselection", 1, 0)) goto error; -- 1.7.0.4
On Tue, Apr 06, 2010 at 05:54:58PM +0200, Jiri Denemark wrote:
There's no sense in advertising cpuselection capability when host CPU is not properly detected and advertised in host capabilities. --- src/qemu/qemu_conf.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 3fb0074..ff5bd43 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -892,7 +892,8 @@ qemudCapsInitGuest(virCapsPtr caps,
guest->arch.defaultInfo.emulator_mtime = binary_mtime;
- if (qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 + if (caps->host.cpu + && qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 && ncpus > 0 && !virCapabilitiesAddGuestFeature(guest, "cpuselection", 1, 0)) goto error;
We usually put && on end of line, and I really prefer fully parenthesized tests expressions, so I would go for a bit of cleanup too. ACK Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel@veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/
guest->arch.defaultInfo.emulator_mtime = binary_mtime;
- if (qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 + if (caps->host.cpu + && qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 && ncpus > 0 && !virCapabilitiesAddGuestFeature(guest, "cpuselection", 1, 0)) goto error;
We usually put && on end of line, Yeah, usually, although not always :-) I prefer it this way as you don't have to look at the end of line to check if that line is part of the condition or not. But I don't really care too much and I can change it.
and I really prefer fully parenthesized tests expressions Hmm, I don't :-) Because you can see the difference if you mistakenly type = instead of == there (well, not in this exact case, but in general):
if (x = 0) vs. if ((x = 0)) In the first case gcc would warn you but in the second one it wouldn't. So I prefer extra parentheses to be put only around assignments not tests to reveal this kind of typos. Jirka
On Wed, Apr 07, 2010 at 11:29:07AM +0200, Jiri Denemark wrote:
guest->arch.defaultInfo.emulator_mtime = binary_mtime;
- if (qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 + if (caps->host.cpu + && qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 && ncpus > 0 && !virCapabilitiesAddGuestFeature(guest, "cpuselection", 1, 0)) goto error;
We usually put && on end of line, Yeah, usually, although not always :-) I prefer it this way as you don't have to look at the end of line to check if that line is part of the condition or not. But I don't really care too much and I can change it.
and I really prefer fully parenthesized tests expressions Hmm, I don't :-) Because you can see the difference if you mistakenly type = instead of == there (well, not in this exact case, but in general):
if (x = 0) vs. if ((x = 0))
In the first case gcc would warn you but in the second one it wouldn't. So I prefer extra parentheses to be put only around assignments not tests to reveal this kind of typos.
If you're afraid of == turning into = then do what peopel suggest which is to swap the arguments (foo() = bar) or (0 = bar) will both be caught and even more drastically by compilers, but the expression will be readable/understandable without trying to remember what is the order of priorities for operators in C ! Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel@veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/
- if (qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 + if (caps->host.cpu + && qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 && ncpus > 0 && !virCapabilitiesAddGuestFeature(guest, "cpuselection", 1, 0)) goto error;
We usually put && on end of line
OK, I fixed this,
and I really prefer fully parenthesized tests expressions
didn't fix this (also because it's actually pretty rare in libvirt code),
ACK
and pushed. Jirka
participants (2)
-
Daniel Veillard -
Jiri Denemark