Re: [PATCH v2 1/9] log to stderr until process is daemonized, even if a logfile is set

[Laine Stump <laine@redhat.com>]

On 2/9/23 12:45 PM, Stefano Brivio wrote:
> On Wed,  8 Feb 2023 12:48:30 -0500
> Laine Stump <laine@redhat.com> wrote:
> 
>> Signed-off-by: Laine Stump <laine@redhat.com>
>> ---
>>   log.c   | 14 +++++++++++++-
>>   log.h   |  1 +
>>   passt.c |  6 ++++--
>>   3 files changed, 18 insertions(+), 3 deletions(-)
>>
>> diff --git a/log.c b/log.c
>> index 468c730..0ab0adf 100644
>> --- a/log.c
>> +++ b/log.c
>> @@ -34,6 +34,7 @@ static int	log_sock = -1;		/* Optional socket to system logger */
>>   static char	log_ident[BUFSIZ];	/* Identifier string for openlog() */
>>   static int	log_mask;		/* Current log priority mask */
>>   static int	log_opt;		/* Options for openlog() */
>> +static int	log_daemon_mode = false; /* true once process is daemonized */
>>   
>>   static int	log_file = -1;		/* Optional log file descriptor */
>>   static size_t	log_size;		/* Maximum log file size in bytes */
>> @@ -67,7 +68,8 @@ void name(const char *format, ...) {					\
>>   	}								\
>>   									\
>>   	if ((setlogmask(0) & LOG_MASK(LOG_DEBUG) ||			\
>> -	     setlogmask(0) == LOG_MASK(LOG_EMERG)) && log_file == -1) {	\
>> +	     setlogmask(0) == LOG_MASK(LOG_EMERG))			\
>> +	     && (log_file == -1 || !log_daemon_mode)) {			\
> 
> This is getting a bit complicated.
> 
> At the moment, LOG_EMERG is abused with the meaning "we don't know
> where/what to log yet", because we didn't process logging options yet.
> 
> Would it be an option to extend this abuse a bit further, and use
> LOG_EMERG to indicate that passt didn't daemonise yet, instead? You
> would just need to move the __setlogmask() calls in main(), and change
> the comment about the first one.

Well, the fact that I didn't understand this was what was being done 
with the log mask kind of indicates this abuse will be misunderstood in 
the long run :-P

But the abuse is already there, so it's not like I'm really making it 
any (much?) worse. It's just increasing the window when the logmask is 
set to LOG_EMERG, while modifying the log macro to not care if log_file 
is opened or not. I'll try that and see how it works out.

> 
> I haven't tested this, and we should be a bit careful with this (check
> what happens with and without running passt from an interactive
> terminal, with and without a log file, etc.).
> 
> A possibly cleaner approach could be to decouple this from the log
> mask, and have an enum instead, to represent logging "states" or
> "modes" -- but I'm not sure we're really going to save much complexity
> with it.

If I'm understanding you, that's kind of what I was doing, it's just 
that there are two modes - false == "logging before being daemonized" 
and true == "logging after being daemonized".

> 
> By the way, the man page should be updated as well.
> 

Good point.