From: Aldy Hernandez <aldyh@redhat.com>
To: Jonathan Wakely <jwakely@redhat.com>
Cc: Jakub Jelinek <jakub@redhat.com>,
Andrew MacLeod via Gcc-patches <gcc-patches@gcc.gnu.org>
Subject: Re: std:vec for classes with constructor?
Date: Fri, 7 Aug 2020 20:04:57 +0200 [thread overview]
Message-ID: <CAGm3qMU+g3FZRxDkCkDvNdLQpHgr3MEktz3f6omMSkYZHT8mHQ@mail.gmail.com> (raw)
In-Reply-To: <20200807091729.GO3400@redhat.com>
[-- Attachment #1: Type: text/plain, Size: 1412 bytes --]
Is the attached patch what y'all suggest?
The patch reverts the GCC vec<> usage in ipa-fnsummary.c and
implements the above suggestion.
Bootstraps, tests running...
If tests pass, is this OK?
Thanks.
Aldy
On Fri, Aug 7, 2020 at 11:17 AM Jonathan Wakely via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> On 07/08/20 10:55 +0200, Jakub Jelinek wrote:
> >On Fri, Aug 07, 2020 at 09:34:38AM +0100, Jonathan Wakely via Gcc-patches wrote:
> >> > Now that you say it, vec has a T[1] member so depending on T
> >> > there might be a side-effect as invoking its CTOR? Or it might
> >> > even not compile if there is no default CTOR available... Ick :/
> >>
> >> Right.
> >>
> >> Does the GTY stuff add members to the struct, or otherwise alter its
> >> layout?
> >
> >Or perhaps use offsetof on an alternate structure that should have the same
> >layout.
>
> Yes that's what I was going to suggest next.
>
> >So instead of
> > typedef vec<T, A, vl_embed> vec_embedded;
> > return offsetof (vec_embedded, m_vecdata) + alloc * sizeof (T);
> >use
> > struct alignas (T) U { char data[sizeof (T)]; };
> > typedef vec<U, A, vl_embed> vec_embedded;
>
> static_assert(sizeof(vec_embedded) == sizeof(vec), "");
> static_assert(alignof(vec_embedded) == alignof(vec), "");
>
> > return offsetof (vec_embedded, m_vecdata) + alloc * sizeof (T);
> >where vec_embedded should have the same offset of m_vecdata as vec.
>
[-- Attachment #2: curr.patch --]
[-- Type: text/x-patch, Size: 3342 bytes --]
diff --git a/gcc/ipa-fnsummary.c b/gcc/ipa-fnsummary.c
index 49bab04524b..d853ed8f075 100644
--- a/gcc/ipa-fnsummary.c
+++ b/gcc/ipa-fnsummary.c
@@ -82,7 +82,6 @@ along with GCC; see the file COPYING3. If not see
#include "gimplify.h"
#include "stringpool.h"
#include "attribs.h"
-#include <vector>
#include "tree-into-ssa.h"
/* Summaries. */
@@ -331,7 +330,7 @@ static void
evaluate_conditions_for_known_args (struct cgraph_node *node,
bool inline_p,
vec<tree> known_vals,
- const std::vector<value_range> &known_value_ranges,
+ vec<value_range> known_value_ranges,
vec<ipa_agg_value_set> known_aggs,
clause_t *ret_clause,
clause_t *ret_nonspec_clause)
@@ -446,7 +445,7 @@ evaluate_conditions_for_known_args (struct cgraph_node *node,
continue;
}
}
- if (c->operand_num < (int) known_value_ranges.size ()
+ if (c->operand_num < (int) known_value_ranges.length ()
&& !c->agg_contents
&& !known_value_ranges[c->operand_num].undefined_p ()
&& !known_value_ranges[c->operand_num].varying_p ()
@@ -555,7 +554,7 @@ evaluate_properties_for_edge (struct cgraph_edge *e, bool inline_p,
{
struct cgraph_node *callee = e->callee->ultimate_alias_target ();
class ipa_fn_summary *info = ipa_fn_summaries->get (callee);
- std::vector<value_range> known_value_ranges (32);
+ auto_vec<value_range, 32> known_value_ranges;
class ipa_edge_args *args;
if (clause_ptr)
@@ -630,11 +629,11 @@ evaluate_properties_for_edge (struct cgraph_edge *e, bool inline_p,
i));
if (!vr.undefined_p () && !vr.varying_p ())
{
- if (!known_value_ranges.size ())
+ if (!known_value_ranges.length ())
{
- known_value_ranges.resize (count);
+ known_value_ranges.safe_grow (count);
for (int i = 0; i < count; ++i)
- known_value_ranges[i].set_undefined ();
+ new (&known_value_ranges[i]) value_range ();
}
known_value_ranges[i] = vr;
}
@@ -808,7 +807,7 @@ ipa_fn_summary_t::duplicate (cgraph_node *src,
}
evaluate_conditions_for_known_args (dst, false,
known_vals,
- std::vector<value_range> (),
+ vNULL,
vNULL,
&possible_truths,
/* We are going to specialize,
@@ -3692,8 +3691,7 @@ estimate_ipcp_clone_size_and_time (struct cgraph_node *node,
clause_t clause, nonspec_clause;
/* TODO: Also pass known value ranges. */
- evaluate_conditions_for_known_args (node, false, known_vals,
- std::vector<value_range> (),
+ evaluate_conditions_for_known_args (node, false, known_vals, vNULL,
known_aggs, &clause, &nonspec_clause);
ipa_call_context ctx (node, clause, nonspec_clause,
known_vals, known_contexts,
diff --git a/gcc/vec.h b/gcc/vec.h
index 3ad99b83690..01d7736c520 100644
--- a/gcc/vec.h
+++ b/gcc/vec.h
@@ -1281,7 +1281,10 @@ template<typename T, typename A>
inline size_t
vec<T, A, vl_embed>::embedded_size (unsigned alloc)
{
- typedef vec<T, A, vl_embed> vec_embedded;
+ struct alignas (T) U { char data[sizeof (T)]; };
+ typedef vec<U, A, vl_embed> vec_embedded;
+ static_assert(sizeof(vec_embedded) == sizeof(vec), "");
+ static_assert(alignof(vec_embedded) == alignof(vec), "");
return offsetof (vec_embedded, m_vecdata) + alloc * sizeof (T);
}
next prev parent reply other threads:[~2020-08-07 18:05 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-31 21:44 [patch] multi-range implementation for value_range (irange) Aldy Hernandez
2020-08-05 14:27 ` Gerald Pfeifer
2020-08-05 15:45 ` Aldy Hernandez
2020-08-05 22:43 ` Gerald Pfeifer
2020-08-06 4:00 ` Richard Sandiford
2020-08-10 3:42 ` Martin Liška
2020-08-10 7:44 ` Aldy Hernandez
2020-08-10 10:22 ` Gerald Pfeifer
2020-08-10 10:30 ` Martin Liška
2020-08-10 11:03 ` Aldy Hernandez
2020-08-05 15:09 ` std:vec for classes with constructor? (Was: Re: [patch] multi-range implementation for value_range (irange)) Martin Jambor
2020-08-05 15:41 ` Aldy Hernandez
2020-08-05 16:55 ` Richard Biener
2020-08-24 21:53 ` Jeff Law
2020-08-24 22:03 ` Andrew MacLeod
2020-08-05 16:54 ` Richard Biener
2020-08-06 1:07 ` Andrew MacLeod
2020-08-06 5:16 ` std:vec for classes with constructor? Richard Sandiford
2020-08-06 10:19 ` Jonathan Wakely
2020-08-06 10:31 ` Richard Biener
2020-08-06 10:48 ` Jonathan Wakely
2020-08-06 14:17 ` Aldy Hernandez
2020-08-06 14:35 ` Richard Biener
2020-08-06 14:59 ` Aldy Hernandez
2020-08-06 16:30 ` Jonathan Wakely
2020-08-06 17:58 ` Aldy Hernandez
2020-08-06 19:24 ` Jonathan Wakely
2020-08-07 6:48 ` Richard Biener
2020-08-07 7:57 ` Jonathan Wakely
2020-08-07 8:22 ` Richard Biener
2020-08-07 8:34 ` Jonathan Wakely
2020-08-07 8:54 ` Richard Biener
2020-08-07 8:55 ` Jakub Jelinek
2020-08-07 9:17 ` Jonathan Wakely
2020-08-07 18:04 ` Aldy Hernandez [this message]
2020-08-07 18:33 ` Jakub Jelinek
2020-08-10 12:57 ` Aldy Hernandez
2020-08-10 13:05 ` Aldy Hernandez
2020-08-10 18:03 ` Jakub Jelinek
2020-08-10 13:51 ` Jakub Jelinek
2020-08-10 16:58 ` Richard Biener
2020-08-10 17:18 ` Jakub Jelinek
2020-08-13 11:38 ` r11-2663 causes static_assert failure (was: Re: std:vec for classes with constructor?) Tobias Burnus
2020-08-13 11:52 ` Jakub Jelinek
2020-08-13 12:06 ` r11-2663 causes static_assert failure Tobias Burnus
2020-08-13 12:25 ` Jakub Jelinek
2020-08-13 12:40 ` Jakub Jelinek
2020-08-13 12:46 ` Iain Sandoe
2020-08-13 12:55 ` Iain Sandoe
2020-08-13 13:04 ` Jakub Jelinek
2020-08-13 13:00 ` Jakub Jelinek
2020-08-06 6:57 ` std:vec for classes with constructor? (Was: Re: [patch] multi-range implementation for value_range (irange)) Richard Biener
2020-08-06 10:23 ` Jonathan Wakely
2020-08-06 11:03 ` Aldy Hernandez
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=CAGm3qMU+g3FZRxDkCkDvNdLQpHgr3MEktz3f6omMSkYZHT8mHQ@mail.gmail.com \
--to=aldyh@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=jakub@redhat.com \
--cc=jwakely@redhat.com \
/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).