On 27/07/16 11:21, Andrea Bolognani wrote:

On Wed, 2016-07-27 at 08:50 +0200, Erik Skultety wrote:

...

Since commit c4bdff19, the path to the configuration file has been constructed in the following manner: - if no config filename was passed to virConfLoadConfigPath, libvirt.conf was used as default - otherwise the filename was concatenated with "<config_dir>/libvirt/libvirt%s%s.conf" which in admin case resulted in "libvirt-libvirt-admin.conf.conf". Obviously, this non-existent config led to ignoring all user settings in libvirt-admin.conf. This patch requires the config filename to be always provided as an argument with the concatenation being simplified.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1357364

Signed-off-by: Erik Skultety <eskultet@redhat.com> --- src/libvirt.c | 2 +- src/util/virconf.c | 12 ++++-------- 2 files changed, 5 insertions(+), 9 deletions(-)

diff --git a/src/libvirt.c b/src/libvirt.c index 68c8317..52462e3 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -969,7 +969,7 @@ virConnectOpenInternal(const char *name, if (ret == NULL) return NULL;

- if (virConfLoadConfig(&conf, NULL) < 0) + if (virConfLoadConfig(&conf, "libvirt.conf") < 0) goto failed;

if (name && name[0] == '\0') diff --git a/src/util/virconf.c b/src/util/virconf.c index ee54072..3e49f41 100644 --- a/src/util/virconf.c +++ b/src/util/virconf.c @@ -1566,20 +1566,16 @@ virConfLoadConfigPath(const char *name) { char *path; if (geteuid() == 0) { - if (virAsprintf(&path, "%s/libvirt/libvirt%s%s.conf", - SYSCONFDIR, - name ? "-" : "", - name ? name : "") < 0) + if (virAsprintf(&path, "%s/libvirt/%s", + SYSCONFDIR, name) < 0) return NULL; } else { char *userdir = virGetUserConfigDirectory(); if (!userdir) return NULL;

- if (virAsprintf(&path, "%s/libvirt%s%s.conf", - userdir, - name ? "-" : "", - name ? name : "") < 0) { + if (virAsprintf(&path, "%s/%s", + userdir, name) < 0) { VIR_FREE(userdir); return NULL; }

You could also have changed src/libvirt-admin.c to call

virConfLoadConfig(&conf, "admin")

instead, but your solution is clearer.

Yeah, I know I could and I also thought about doing that, but I didn't like it in the end, because imho (and you mentioned that too) being explicit about the config file in this case is somewhat clearer. Also, I'm not sure and I might be talking rubbish but I think I tried to use the approach you mentioned above in the earlier versions of the patch that modified this logic and was discouraged to do it that way in reviews (but I would have to check the history :P to really confirm that...). Anyway, thanks for review, I'll push that in a moment. Erik

ACK

-- Andrea Bolognani / Red Hat / Virtualization