On 4/26/22 11:03, Paul Moore wrote:
> On Mon, Apr 25, 2022 at 7:31 PM John Johansen
> <john.johansen@xxxxxxxxxxxxx> wrote:
>> On 4/18/22 07:59, Casey Schaufler wrote:
>>> Replace the timestamp and serial number pair used in audit records
>>> with a structure containing the two elements.
>>>
>>> Signed-off-by: Casey Schaufler <casey@xxxxxxxxxxxxxxxx>
>>> Acked-by: Paul Moore <paul@xxxxxxxxxxxxxx>
>>> ---
>>> kernel/audit.c | 17 +++++++++--------
>>> kernel/audit.h | 12 +++++++++---
>>> kernel/auditsc.c | 22 +++++++++-------------
>>> 3 files changed, 27 insertions(+), 24 deletions(-)
>
> ...
>
>>> diff --git a/kernel/audit.h b/kernel/audit.h
>>> index 4af63e7dde17..260dab6e0e15 100644
>>> --- a/kernel/audit.h
>>> +++ b/kernel/audit.h
>>> @@ -108,10 +114,10 @@ struct audit_context {
>>> AUDIT_CTX_URING, /* in use by io_uring */
>>> } context;
>>> enum audit_state state, current_state;
>>> + struct audit_stamp stamp; /* event identifier */
>>> unsigned int serial; /* serial number for record */
>>
>> shouldn't we be dropping serial from the audit_context, since we have
>> moved it into the audit_stamp?
>
> Unless we make some significant changes to audit_log_start() we still
> need to preserve a timestamp in the audit_context so that regularly
> associated audit records can share a common timestamp (which is what
> groups multiple records into a single "event").
>
sure, but the patch changes things to use ctx->stamp.serial instead of
ctx->serial. Eg. in audit_reset_context() we have
- ctx->serial = 0;
+ ctx->stamp.serial = 0;
I don't see a reason why we need both ctx->serial and ctx->stamp.serial
