On 03/18/2016 05:31 PM, Eric Blake wrote:
> On 03/02/2016 11:55 AM, John Ferlan wrote:
>> Add the functions to add/remove elements from the hashed secret obj list.
>> These will replace secret_driver functions secretAssignDef and secretObjRemove.
>>
>> The virSecretObjListAddLocked will perform the necessary lookups and
>> decide whether to replace an existing hash entry or create a new one.
>> This includes setting up the configPath and base64Path as well as being
>> able to support the caller's need to restore from a previous definition
>> in case something goes wrong in the caller.
>>
>> Need to also move and make global virSecretUsageIDForDef since it'll
>> be used by both secret_driver.c and secret_conf.c
>>
>> Signed-off-by: John Ferlan <jferlan@xxxxxxxxxx>
>> ---
>> src/conf/secret_conf.c | 173 +++++++++++++++++++++++++++++++++++++++++++++
>> src/conf/secret_conf.h | 15 ++++
>> src/libvirt_private.syms | 1 +
>> src/secret/secret_driver.c | 34 ++-------
>> 4 files changed, 196 insertions(+), 27 deletions(-)
>>
>
>> +virSecretObjPtr
>> +virSecretObjListAddLocked(virSecretObjListPtr secrets,
>> + virSecretDefPtr def,
>> + const char *configDir,
>> + virSecretDefPtr *oldDef)
>
>> +
>> + if (secret->def->private && !def->private) {
>> + virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
>> + _("cannot change private flag on existing secret"));
>> + goto cleanup;
>> + }
>
> This says we can't turn an existing public secret into a private one,
> but nothing is protecting us from the converse of someone attempting to
> redefine a private secret into a public one. I think this should be:
>
> if (secret->def->private != def->private)
>
This is a copy of current check in secretDefineXML...
The secret->def->private is our stored one and def->private is the new
one. As I'm reading it, it looks right; however, going with the
suggestion means one could not change either way (e.g. public->private
or private->public). Then again maybe I'm reading your comment wrong.
FWIW: I've changed the name later to "isprivate" to make it clearer (and
to avoid that potential magic word private).
>
>> +
>> + /* Generate the possible configFile and base64File strings
>> + * using the configDir, uuidstr, and appropriate suffix
>> + */
>> + virUUIDFormat(def->uuid, uuidstr);
>> + if (!(configFile = virFileBuildPath(configDir, uuidstr, ".xml")) ||
>> + !(base64File = virFileBuildPath(configDir, uuidstr, ".base64"))) {
>> + VIR_FREE(configFile);
>
> Redundant; the cleanup label also takes care of freeing configFile.
>
Oh right - relic of intermediate edits where goto label didn't exist
Thanks for sticking with this so far -
John
>> + goto cleanup;
>> + }
>> +
>
> Everything else looked okay.
>
--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list