This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: new static user probe types
Mark Wielaard <mjw@redhat.com> writes:
> [...]
>> The point of this is to skip any argument-packing work generated by the
>> compiler, which would be inside the "if unlikely" block, right?
(Yes.)
> Partly, but it is sadly not guaranteed by GCC.
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40207
Note that this was for an implementation that relies on static inline
functions. Pure macros are OK.
>> I think this is a lot like some things Mathieu already experimented with
>> and measured in the kernel context. I think he pursued a code-patching
>> flavor that patched an immediate operand because that was measured as
>> faster than having the actual extra load of a simple enabled_p variable.
(Yes, and other tracing widgets from other OS's have discovered the
same thing. Of course that costs per-arch complications instead, thus
the simpler proposal as a first step.)
- FChE