[llvm] r188188 - Misc enhancements to LTO:

[Nick Lewycky]

On 12 August 2013 13:29, Shuxin Yang <shuxin.llvm at gmail.com> wrote:

>  On 8/12/13 1:03 PM, Nick Lewycky wrote:
>
> On 12 August 2013 12:22, Shuxin Yang <shuxin.llvm at gmail.com> wrote:
>
>> On 8/12/13 12:16 PM, Nick Lewycky wrote:
>>
>>> Shuxin Yang wrote:
>>>
>>>> Author: shuxin_yang
>>>> Date: Mon Aug 12 13:29:43 2013
>>>> New Revision: 188188
>>>>
>>>> URL: http://llvm.org/viewvc/llvm-project?rev=188188&view=rev
>>>> Log:
>>>> Misc enhancements to LTO:
>>>>
>>>>    1. Add some helper classes for partitions. They are designed in a
>>>>       way such that the top-level LTO driver will not see much
>>>> difference
>>>>       with or without partitioning.
>>>>
>>>>    2. Introduce work-dir. Now all intermediate files generated during
>>>>       LTO phases will be saved under work-dir. User can specify the
>>>> workdir
>>>>       via -lto-workdir=/path/to/dir. By default the work-dir will be
>>>>       erased before linker exit. To keep the workdir, do -lto-keep, or
>>>> -lto-keep=1.
>>>>
>>>>      TODO: Erase the workdir, if the linker exit prematurely.
>>>>        We are currently not able to remove directory on signal. The
>>>> support
>>>>        routines simply ignore directory.
>>>>
>>>>    3. Add one new API lto_codegen_get_files_need_remove().
>>>>       Linker and LTO plugin will communicate via this API about which
>>>> files
>>>>      (including directories) need to removed before linker exit.
>>>>
>>>
>>> Please revert. Adding new flags to libLTO is the wrong direction (in
>>> spite of the ones that exist -- consider those grandfathered in).
>>>
>>  It dose not make sense. Without flags, how do you manage to triage the
>> correctness and performance problem?
>>
>
>  Something else has flags,
>
>
> What are "something else"?  As far as I know, there are only two fall into
> this category:
>
>    - Apple linker, and
>    - GNU gold.
>
>   The former communicate with the libLTO directly with these APIs, while
> GNU gold communicate
> with the libLTO via, which I called adapter.
>

You committed the patch, you tell me. What did this patch intend to change
the behaviour of?

   Directly calling these APIs is really bad idea. I manage to convince the
> black-belt guru Nick @ Apple to
> not directly calling these APIs in the new ld design. The linker's should
> be LTO oblivious, the linker
> should expose symbol-related interface instead of LTO-control interfaces.
>

Directly calling which APIs?

What do you mean by "LTO oblivious"? Oblivious towards whether optimization
is being performed?

>which in turn drives libLTO through the API.
>
> Depending on the what kind of info "something" else need to drive the
> libLTO.
> In general it is very bad idea, if "something else" need micro-management.
>

libLTO is part of the linker that uses it. Having a default setting with
the ability to override it is a sensible convenience for users of libLTO.

Take Apple ld as example,  if I want to change LTO in a way such that I
> don't want to load all module,
> I just want to load summary info. Current APIs are not sufficient. I have
> to modify the API, or add new APIs
> to that matter, in the mean time, I need release the new ld to the user in
> order to accomodate the change.
> that is nightmare.
>

The point of libLTO is to provide an ABI-fixed library, isolating the
linker from llvm's internals. That in turn leads to a few design decisions.
The API is designed to refer to high-level concepts instead of the details
of llvm's actual behaviour. Things like module lazy loading or setting the
datalayout are excluded from the API. Flags are even more private, surely
we should be able to change flags in LLVM's libraries without worrying
about breaking linkers.

If the linker needs to do something where it matters how llvm is
implemented -- you mention loading summary info, I'll assume you mean
lazy-loading the module such that function bodies aren't loaded -- then the
linker doesn't use libLTO at all, but uses llvm directly. Conversely,
libLTO knows all about llvm and will lazy-load .bc files without being
asked to.

Sure, "something else" can control the libLTO, if it want. In my case, if
> "something else" want specify
>  a workdir, then go ahead. Otherwise, the libLTO use default one. Is there
> any wrong here?
>

At a high level that sounds fine to me. The wrong part is using flags to do
it.

>  Adding flags to linker instead, I think that is wrong direction. Linker
>> dose not have data structure which libLTO dose.
>
>
>  This is the discussion to have. What things do you need here which you
> don't think should be exposed through the API, and yet you want to be
> exposed for you?
>
> I actually discuss with Nick @ Apple before.  The conclusion is linker
> must be LTO oblivious,
> it should think in symbol-way, and talk in symbol way (as with GNU gold).
> It would otherwise
>  very very troublesome both for linker and libLTO.
>

And now you're discussing it with me. I also agree that the linker should
communicate primarily in symbols and about symbols with libLTO.

On the other hand, we now have two linkers support LTO. There are different
> way to control
> the libLTO (even for simple task, like save intermediate files), how messy?
>
> I'd like to move all these stuff to libLTO to have a unified control.
>

I have no problem with a unified control.

>  libLTO is intended to be used as a library, it may not get a chance to
>> parse flags.
>>  It has to. Prior to my change, linkers (GNU linker and Apple ld) pass
>> arch to linker, via a function
>> confusingly called, something like "add.*debug.*options".
>
>
>  Can't. If we allow this, every flag in every part of LLVM that libLTO
> links against is baked into the C ABI forever.
>
>  Of course addDebugOptions does allow this, but it's named (and I thought
> documented in the comments) such that anybody using it knows they're using
> a non-stable non-production debugging API. Anybody using addDebugOptions
> for something other than debugging libLTO is living outside the ABI
> guarantees.
>
> addDebugOptions is misnomer. It is also passes essential flags like
> -arch=x86.  Without such flags,
> the LTO dose not even compile.
>

That sounds like a nice bug you've got there! Wouldn't want anything to
happen to it. It'd be a shame if breaks before you manage to add a
liblto_set_arch() function for it.* Honestly, I looked and couldn't find a
-arch flag that libLTO would interpret. How sure are you about this?

In case it isn't completely clear, flags are absolutely right out. Either
you will revert this patch, or I will revert it for you. I'm sorry you
decided to land three things together in one patch, please remember not to
do that in the future.

Nick

* http://www.youtube.com/watch?v=cNZKUozrBl4&t=1m33s
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130812/3f09f953/attachment.html>