On Wed, Mar 29, 2017 at 10:45:55PM +0000, Legacy, Allain wrote:
> > -----Original Message-----
> > From: Yuanhan Liu [mailto:yuanhan....@linux.intel.com]
> > Sent: Wednesday, March 29, 2017 3:02 AM
>
> <...>
> >
> > OTOH, it's still good to have RTE_LOG_ONCE. It explicitly tells you
> > something
> > is wrong, then you could dump those stats for more info.
>
> Has the implementation of something like RTE_LOG_ONCE been discussed before?
No. It's just something comes to my mind after saw your patch.
> Do you have a proposal on what you would find acceptable as an implementation
> that considers multiples threads, multiple devices, multiple logs per file,
> etc.?
No, that would be way to complex. Something I thought of is quite
simple (like what Linux kernel does):
#define RTE_LOG_ONCE(...) do { \
static int warned; \
if (!warned) { \
RTE_LOG(...); \
warned = 1; \
} \
} while (0)
And yes, I know it's far away from being perfect (and accurate, regarding
to multiple devices and so on as you mentioned). Even though, I think
it's much better than showing nothing when something goes wrong.
--yliu
