All of lore.kernel.org
 help / color / mirror / Atom feed
From: Masahiro Yamada <masahiroy@kernel.org>
To: Nathan Chancellor <nathan@kernel.org>
Cc: Nick Desaulniers <ndesaulniers@google.com>,
	Kees Cook <keescook@chromium.org>, Shuah Khan <shuah@kernel.org>,
	llvm@lists.linux.dev,
	Linux Kbuild mailing list <linux-kbuild@vger.kernel.org>,
	"open list:DOCUMENTATION" <linux-doc@vger.kernel.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	"open list:KERNEL SELFTEST FRAMEWORK"
	<linux-kselftest@vger.kernel.org>,
	Peter Zijlstra <peterz@infradead.org>
Subject: Re: [PATCH v2] kbuild: Make $(LLVM) more flexible
Date: Wed, 9 Mar 2022 18:33:40 +0900	[thread overview]
Message-ID: <CAK7LNAThknb0=-XhfB6zspke-sNHMEmTbGy8WVeg20ntT72wqA@mail.gmail.com> (raw)
In-Reply-To: <Yid6eS7YV4Oxj+hx@dev-arch.thelio-3990X>

On Wed, Mar 9, 2022 at 12:47 AM Nathan Chancellor <nathan@kernel.org> wrote:
>
> On Mon, Mar 07, 2022 at 11:08:29AM -0800, Nick Desaulniers wrote:
> > On Fri, Mar 4, 2022 at 9:14 AM Nathan Chancellor <nathan@kernel.org> wrote:
> > >
> > > diff --git a/Documentation/kbuild/llvm.rst b/Documentation/kbuild/llvm.rst
> > > index d32616891dcf..68b74416ec48 100644
> > > --- a/Documentation/kbuild/llvm.rst
> > > +++ b/Documentation/kbuild/llvm.rst
> > > @@ -49,17 +49,36 @@ example: ::
> > >  LLVM Utilities
> > >  --------------
> > >
> > > -LLVM has substitutes for GNU binutils utilities. Kbuild supports ``LLVM=1``
> > > -to enable them. ::
> > > -
> > > -       make LLVM=1
> > > -
> > > -They can be enabled individually. The full list of the parameters: ::
> > > +LLVM has substitutes for GNU binutils utilities. They can be enabled individually.
> > > +The full list of supported make variables: ::
> > >
> > >         make CC=clang LD=ld.lld AR=llvm-ar NM=llvm-nm STRIP=llvm-strip \
> > >           OBJCOPY=llvm-objcopy OBJDUMP=llvm-objdump READELF=llvm-readelf \
> > >           HOSTCC=clang HOSTCXX=clang++ HOSTAR=llvm-ar HOSTLD=ld.lld
> > >
> > > +To simplify the above command, Kbuild supports the ``LLVM`` variable: ::
> > > +
> > > +       make LLVM=1
> > > +
> > > +If your LLVM tools are not available in your PATH, you can supply their
> > > +location using the LLVM variable with a trailing slash: ::
> > > +
> > > +       make LLVM=/path/to/llvm/
> > > +
> > > +which will use ``/path/to/llvm/clang``, ``/path/to/llvm/ld.lld``, etc.
> >
> > I don't think we should do this; `PATH=/path/to/llvm/ make LLVM=1`
> > works and (my interpretation of what) Masahiro said "if anyone asks
> > for this, here's how we could do that."  I don't think I've seen an
> > explicit ask for that. I'd rather LLVM= have 2 behaviors than 3, but I
> > won't hold this patch up over that.  Either way:
>
> Right, there has not been an explicit ask for the prefix support yet,
> although I know I personally would use it, but I think that it is worth
> doing now instead of later for a few reasons:
>
> 1. It makes path goofs easier to spot. If you do
>
>      $ PATH=/path/to/llvm:$PATH make LLVM=1 ...
>
>    with a path to LLVM that does not exist (maybe you are bisecting an
>    issue and using a temporary build of LLVM and you forgot the path it
>    was in), you fall back to the LLVM tools that are in other places in
>    your PATH, which is not what the developer intended. I know that I
>    have messed up bisects that way. If you did
>
>      $ make LLVM=/path/to/llvm/
>
>    with a path that does not exist, there will be an error much earlier:
>
>      $ make LLVM=/this/path/does/not/exist/ defconfig
>      /bin/sh: line 1: /this/path/does/not/exist/clang: No such file or directory
>
> 2. It does not take that much more code or documentation to support. It
>    is the same amount of code as the suffix and the documentation is
>    roughly the same amount of lines as well.
>
> 3. If we wait to implement the path-based use of $(LLVM), we have three
>    "sequence" points: the initial support of $(LLVM), the suffix
>    support, and the prefix support. As we are constantly working with
>    various trees, it would make it harder to know what to use when. If
>    we just do it in the same patch, we know 5.18+ can use both of these
>    methods.
>
> However, at the end of the day, we are a team and if you feel like we
> should only have suffix support, I am more than happy to push a v3 that
> does just that and we can revist prefix support in the future. Just let
> me know!


I do not have a strong opinion about this.
(I just mentioned the LLVM=/path/to/llvm/ form because I guessed
somebody would request this sooner or later.)


If you want me to pick up this version, I will apply it with fixing up
a nit pointed out by Kees   (": ::" -> "::")

If you want to send v3, that is fine with me as well.

Please let me know your thoughts.








>
> > Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
> >
> > > +
> > > +If your LLVM tools have a version suffix and you want to test with that
> > > +explicit version rather than the unsuffixed executables like ``LLVM=1``, you
> > > +can pass the suffix using the ``LLVM`` variable: ::
> > > +
> > > +       make LLVM=-14
> > > +
> > > +which will use ``clang-14``, ``ld.lld-14``, etc.
> > > +
> > > +``LLVM=0`` is not the same as omitting ``LLVM`` altogether, it will behave like
> > > +``LLVM=1``.
> >
> > Hmm... I can see someone's build wrappers setting LLVM=1, then them
> > being surprised that appending LLVM=0 doesn't disable LLVM=1 as they
> > might expect.  But Masahiro says let's fix this later which is fine.
>
> Sure, I guess that is a reasonable case to support. I'll see if I can
> come up with something that makes sense after this change lands.
>
> Cheers,
> Nathan



-- 
Best Regards
Masahiro Yamada

  reply	other threads:[~2022-03-09  9:34 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-03-04 17:08 [PATCH v2] kbuild: Make $(LLVM) more flexible Nathan Chancellor
2022-03-04 18:09 ` Kees Cook
2022-03-04 18:15   ` Nathan Chancellor
2022-03-04 18:55     ` Masahiro Yamada
2022-03-07 19:08 ` Nick Desaulniers
2022-03-08 15:47   ` Nathan Chancellor
2022-03-09  9:33     ` Masahiro Yamada [this message]
2022-03-10 17:36       ` Nathan Chancellor
2022-03-18  5:16         ` Masahiro Yamada
2022-03-10  0:36     ` Nick Desaulniers
2022-03-08 16:36   ` Matthew Wilcox
2022-03-09  9:29     ` Masahiro Yamada

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAK7LNAThknb0=-XhfB6zspke-sNHMEmTbGy8WVeg20ntT72wqA@mail.gmail.com' \
    --to=masahiroy@kernel.org \
    --cc=keescook@chromium.org \
    --cc=linux-doc@vger.kernel.org \
    --cc=linux-kbuild@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-kselftest@vger.kernel.org \
    --cc=llvm@lists.linux.dev \
    --cc=nathan@kernel.org \
    --cc=ndesaulniers@google.com \
    --cc=peterz@infradead.org \
    --cc=shuah@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.