public inbox for jit@gcc.gnu.org
 help / color / mirror / Atom feed
From: David Malcolm <dmalcolm@redhat.com>
To: "Marc Nieper-Wißkirchen" <marc@nieper-wisskirchen.de>, jit@gcc.gnu.org
Subject: Re: alloca and labels as values
Date: Mon, 01 Jan 2018 00:00:00 -0000	[thread overview]
Message-ID: <1545141762.4619.244.camel@redhat.com> (raw)
In-Reply-To: <9bf4ad79-88dd-cac9-babf-c83d59acf7c5@nieper-wisskirchen.de>

On Tue, 2018-12-18 at 08:12 +0100, Marc Nieper-Wißkirchen wrote:
> > > Is `alloca' currently  possible with libgccjit? I haven't found
> 
>  > > anything
>  > > about it in the documentation. If it isn't possible at the
> moment,
>  > > I'd
>  > > love to see it added to libgccjit very soon. (One use case would
> be
>  > > a
>  > > compiler like the Chicken Scheme compiler that employs Cheney-
> on-the-
>  > > MTA
>  > > where the youngest heap generation is allocated on the stack.)
>  >
>  > I haven't tested this, but I believe you ought to be able to
> access it
>  > via the "__builtin_alloca" builtin, via something like:
>  >
>  >
>  > gcc_jit_function *fn_alloca
>  >   = gcc_jit_context_get_builtin_function (ctxt,
> "__builtin_alloca");
>  >
>  >
>  > gcc_jit_rvalue *call =
>  >         gcc_jit_context_new_call (ctxt,
>  >                                   loc,
>  >                                   fn_alloc_a,
>  >                                   1, &size_t_arg);
>  >
>  >  gcc_jit_block_add_assignment (block, loc,
>  >                                dst_lvalue,
>  >                                call);
>  >
>  > ...or something like that.
> 
> 
> That was helpful; thank you very much. I overlooked the procedure 
> `gcc_jit_context_get_builtin_function' in the documentation. Is there
> a 
> way to determine which builtins are supported by a specific version
> of 
> the library or would I have to add a few scripts to my project's 
> `configure.ac <http://configure.ac>' myself?

I think you'd need a configure check.

> -- Marc
> 
> P.S.: Another thing I wasn't able to find so far was how to declare 
> variables as thread local (i.e. GCC's __thread or C11's
> thread_local).

There isn't currently a way to do thread local variables.  I suspect
that something like a variant of gcc_jit_context_new_global could do
it, e.g. "gcc_jit_context_new_thread_local" or somesuch.  Or maybe a
way to mark the result of gcc_jit_context_new_global as being thread-
local.

Dave

  reply	other threads:[~2018-12-18 14:02 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-01-01  0:00 Marc Nieper-Wißkirchen
2018-01-01  0:00 ` David Malcolm [this message]
2019-01-01  0:00   ` Marc Nieper-Wißkirchen
2018-01-01  0:00 Marc Nieper-Wißkirchen
2018-01-01  0:00 ` David Malcolm

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=1545141762.4619.244.camel@redhat.com \
    --to=dmalcolm@redhat.com \
    --cc=jit@gcc.gnu.org \
    --cc=marc@nieper-wisskirchen.de \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).