On Tue, Mar 04, 2014 at 06:44:01AM -0700, Eric Blake wrote:
On 03/03/2014 10:21 AM, Martin Kletzander wrote:
When domain is started with setting that cannot be done, i.e. those that require cgroups, there is no error reported and it succeeds without any message whatsoever.
When setting with API, virsh, an error is reported, but only due to the fact that no cgroups are mounted (priv->cgroup == NULL).
Given the above it seems reasonable to reject such unsupported settings.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com> ---
Notes: Few questions came to my mind while writing the commit message:
1) Would helper function (macro) be preferred so the code looks cleaner?
What macro do you have in mind?
I haven't thought that through, just from top of my head: #define SESSION_UNSUPP(what) if (!cfg->privileged) { \ virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s %s", \ what, _("is not available in sesison mode")); \ goto cleanup; \ } because the code repeats a lot, but it seems like it would not be that readable. Separate function would do as well.
2) Do we want to allow reading of some of these settings through an API? This would however require our cgroup handling to be reworked.
3) Would new error type (for session-unsupported settings) be any good or it doesn't make sense to create one just for these added messages plus few older ones (just guessing the amount)?
Looks like your use of VIR_ERR_CONFIG_UNSUPPORTED was reasonable.
+ if (!cfg->privileged) { + /* If we have no cgroups than we can have no tunings that + * require them */ + + if (def->mem.hard_limit || def->mem.soft_limit || + def->mem.min_guarantee || def->mem.swap_hard_limit) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("Memory tuning is not available in session mode")); + goto error; + } +
Or maybe we should think about some day adding a daemon that accepts RPC commands for manipulating cgroups on behalf of a session client. But that would be a later and bigger patch.
I was just trying to output a better error message and forgot to add the BZ that started it: https://bugzilla.redhat.com/show_bug.cgi?id=1023366
Seems reasonable to me, but it might be nice to also get Dan's opinion on this one.
Anyway, I'll have a second version which allows at least cpu pinning as setaffinity works and is used when no cgroups are available. Martin