Re: introduce size limit parameter for var-tracking

[Richard Guenther <richard dot guenther at gmail dot com>]

On Tue, Jan 19, 2010 at 1:23 PM, Alexandre Oliva <aoliva@redhat.com> wrote:
> Variable tracking sometimes takes far too long to complete. ?This patch
> provides users with a means to tell the compiler to just give up when
> dataflow analysis is taking too long (or rather too much memory, which
> is correlated).
>
> I'm undecided as to whether or not we should report the final sizes upon
> success, when the parameter is given. ?It was useful for finding out a
> suitable default for Red Hat's gcc 4.4 branch, but reporting is not
> enabled there. ?Should I take it out too?
>
> I haven't introduced a default for the trunk. ?I still hope to speed up
> var-tracking some more, so that we don't have to introduce a cut-off for
> normal testcases, but the approach I've been working on for the last
> month or so isn't working so well :-(
>
> Ok to install?

+                   inform (DECL_SOURCE_LOCATION (cfun->decl),
+                           "variable tracking size limit exceeded
with debug insns, retrying without");

watch long lines.  I'd not refer to "debug insns" but instead
to "-fvar-tracking-assignments" as the latter is a documented
thing people can look up.

-  if (MAY_HAVE_DEBUG_INSNS)
+  /* ??? Remove or document.  */
+  if (success && PARAM_SET_P (PARAM_MAX_VARTRACK_SIZE))
+    inform (DECL_SOURCE_LOCATION (cfun->decl),
+           "variable tracking size: %i", htabsz);
+
+  if (success && MAY_HAVE_DEBUG_INSNS)

I'd indeed remove this hunk.

Ok with that changes.

Thanks,
Richard.

>
>
> --
> Alexandre Oliva, freedom fighter ? ?http://FSFLA.org/~lxoliva/
> You must be the change you wish to see in the world. -- Gandhi
> Be Free! -- http://FSFLA.org/ ? FSF Latin America board member
> Free Software Evangelist ? ? ?Red Hat Brazil Compiler Engineer
>
>