On 01/03/17 11:26, Torsten Duwe wrote:
> On Fri, Feb 17, 2017 at 11:30:29PM -0700, Sandra Loosemore wrote:
>>>
>>> +@item prolog_pad
>>> +@cindex @code{prolog_pad} function attribute
>>
>> I'm only a documentation maintainer so this is out of my area of
>> responsibility, but I really wish we could rename the attribute and
>> command-line option. Per
>>
>> per https://gcc.gnu.org/codingconventions.html#Spelling
>>
>> the correct spelling is "prologue".
>>
>>> +@cindex extra NOP instructions at the function entry point
>>> +In case the target's text segment can be made writable at run time
>>> +by any means, padding the function entry with a number of NOPs can
>>> +be used to provide a universal tool for instrumentation. Usually,
>>> +prolog padding is enabled globally using the @option{-fprolog-pad=N,M}
>>
>> definitely s/prolog/prologue/ in the running text here.
>
> Well, you're definitely right in both cases.
>
> About 400 occurrences of "prolog" in the source without ChangeLogs,
> mainly in gcc/tree-vect-loop-manip.c and in libgcc/config/libbid/bid_conf.h;
> about 3000 lines with "prologue". However there is a "-mprolog-function"
> switch. One might call this a broken window. I don't want to contribute
> to that.
>
> However, writing some more documentation and being asked for clarity,
> I found it more depicting to talk about the function entry point than
> about the prologue. Also, this is about generic instrumentation, and it
> surely involves NOPs.
>
> So, hereby I'd like to start a small poll for a good name for this feature.
> Anyone with a better idea please speak up now. Otherwise I'll just
> s/prolog/prologue/g.
Hmm, I'd prefer the bike shed to be green :-)
How about --fpatchable-function-entry=<size-spec>?
>
>> The amount of space reserved is expressed as the number of NOP instructions
>> to insert. On targets that have multiple instruction sizes, typically the
>> smallest NOP instruction available for the current CPU mode is used to
>> achieve the finest granularity.
>
> I've made another improvement which makes the code even more robust now.
> +DEF_TARGET_INSN (nop, (void))
> In gcc/target-insns.def. This way I can easily check whether there is a
> (define_insn "nop" ...) in the target md. Currently, all CPUs have it, but
> who knows.
The mid-end already has direct calls to gen_nop with no guards on the
pattern existing, So the compiler won't build without a NOP pattern.
>
> This will also be the default instruction used (It can be overridden
> in the terget hook), so that rule has changed.
>
> So, before the next version, any clever name suggestions?
>
> Torsten
>
R.
