Re: [PATCH v2] eal: redefine logtype values

["De Lara Guarch, Pablo" <pablo.de.lara.guarch@intel.com>]

Hi Olivier,

> -----Original Message-----
> From: Olivier MATZ [mailto:olivier.matz@6wind.com]
> Sent: Wednesday, April 12, 2017 8:24 PM
> To: De Lara Guarch, Pablo
> Cc: thomas.monjalon@6wind.com; dev@dpdk.org
> Subject: Re: [PATCH v2] eal: redefine logtype values
> 
> Hi Pablo,
> 
> 
> On Wed, 12 Apr 2017 16:35:32 +0100
> Pablo de Lara <pablo.de.lara.guarch@intel.com> wrote:

...

> 
> 
> Thanks for spotting this issue. I think there is still a problem with
> the deprecated functions for which we want to keep compat:
> 
> /* Set global log type */
> __rte_deprecated void
> rte_set_log_type(uint32_t type, int enable)
> {
> 	if (type < RTE_LOGTYPE_FIRST_EXT_ID) {
> 		if (enable)
> 			rte_logs.type |= type;
> 		else
> 			rte_logs.type &= ~type;
> 	}
> 
> 	if (enable)
> 		rte_log_set_level(type, 0);
> 	else
> 		rte_log_set_level(type, RTE_LOG_DEBUG);
> }
> 
> /* Get global log type */
> __rte_deprecated uint32_t
> rte_get_log_type(void)
> {
> 	return rte_logs.type;
> }
> 
> 
> There is a problem in these functions because they expect bitmasks
> for the first part.
> 
> I does not look so easy to both fix the issue and keep a full compat
> with previous functions.
> 
> The first solution is your patch + add a shift in rte_set_log_type().
> It would work except if an application uses rte_get_log_type() and masks
> the result with a RTE_LOGTYPE_* (which will not be a bitmask). I think
> it's a rare case.
> 
> The second approach would be to define new constants for the new
> functions and keep the old ones for the old functions. This approach
> is probably more confusing. It also requires to check the doc and
> examples again.
> 
> Any opinion?
> I would prefer the first solution. If you want I can send a patch
> updated based on yours.

Oh yes, I missed that too. First option looks ok to me, so send a patch if you like.