pmc warnings on stable/9

[Ryan Stone]

On Tue, Aug 7, 2012 at 7:16 PM, Sean Bruno <seanbru at yahoo-inc.com> wrote:
> I have no idea the significance, or danger.  When compiling on stable/9
> I have always seen the following WARNINGS.  Can we silence/fix these?
> Or is it supposed to be that way?  :-)
>
> WARNING: hwpmc_soft.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: kern_pmc.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: kern_rwlock.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: kern_sx.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: kern_lock.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: hwpmc_intel.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: hwpmc_tsc.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: hwpmc_amd.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: kern_clock.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: kern_mutex.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: trap.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: hwpmc_x86.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: hwpmc_uncore.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: hwpmc_piv.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: hwpmc_core.c: enum pmc_event has too many values: 1531 > 1023
> WARNING: hwpmc_logging.c: enum pmc_event has too many values: 1531 >
> 1023
> WARNING: hwpmc_mod.c: enum pmc_event has too many values: 1531 > 1023
>
>
> _______________________________________________
> freebsd-hackers at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
> To unsubscribe, send any mail to "freebsd-hackers-unsubscribe at freebsd.org"

These warnings(and the previous one from your other thread) both come
from ctfconvert.  I haven't looked into them in detail but I suspect
that the CTF format is not able to represent things like enums with
more than 1024 entries, or structures with more than a certain number
of members.  I've never seen any actual side-effects of the warnings,
although I suspect that if you tried to use DTrace to examine
variables with the affected types you wouldn't be able to inspect
everything.  As far as I know, DTrace is the only consumer of CTF data
in the kernel.

I have dreams of replacing DTrace's usage of CTF with libdwarf, but
that's little more than idle musings on my part at this stage.