Incorrect bitfield aliasing with Tree SSA


Incorrect bitfield aliasing with Tree SSA
\ Adam Nemet (15 Jun 2007)
. \ Daniel Berlin (15 Jun 2007)
. . \ Adam Nemet (15 Jun 2007)
. . . \ Daniel Berlin (15 Jun 2007)
. . . . \ Eric Botcazou (16 Jun 2007)
. . . . . \ Daniel Berlin (16 Jun 2007)
. . . . . . \ Eric Botcazou (16 Jun 2007)
. . . . . . . \ Daniel Berlin (17 Jun 2007)
. . . . . . . . \ Eric Botcazou (17 Jun 2007)
. . . . . . . . . \ Adam Nemet (17 Jun 2007)
. . . . . . . . . . \ Eric Botcazou (18 Jun 2007)
. . . . . . \ (Richard Kenner) (17 Jun 2007)
. . . . . . . \ Daniel Berlin (17 Jun 2007)
. . . . . . . . \ Eric Botcazou (17 Jun 2007)
. . . . . . . . . \ Daniel Berlin (17 Jun 2007)
. . . . . . . . . . \ Laurent GUERBY (17 Jun 2007)
. . . . . . . . . . . \ Daniel Berlin (17 Jun 2007)
. . . . . . . . . . . . \ Laurent GUERBY (17 Jun 2007)
. . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . \ Eric Botcazou (18 Jun 2007)
. . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . \ Eric Botcazou (18 Jun 2007)
. . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . \ Adam Nemet (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . \ Adam Nemet (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . . . \ Adam Nemet (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . . . . . . . . . . \ Daniel Berlin (18 Jun 2007)
. . . . . . . . . . . . \ (Richard Kenner) (18 Jun 2007)
. . . . . . . . \ (Richard Kenner) (17 Jun 2007)

5 msg	Resuming SPEC performance tracking at RedHat
62 msg	Some thoughts about steerring commitee work
7 msg	Libmudflap for sh-elf toolchain cannot access e...
34 msg	RFC: Make dllimport/dllexport imply default vis...
10 msg	Re: [patch,committed] Make Fortran maintainers ...
1 msg	RFA: (dataflow?) testsuite regression since las...
10 msg	Please fork soft-fp from libc
3 msg	How to submit Intel BID library patch?
2 msg	maybe configure.ac problem ?
1 msg	[PATCH] Fix breakage
2 msg	r125698 breaks bootstrap on trunk for i686-pc-l...
2 msg	Diego Novillo appointed middle-end maintainer a...
2 msg	Ian Taylor appoined non-algorithmic GWP
1 msg	Bootstrap failure on trunk
4 msg	Merge PTR_PLUS branch?
1 msg	gcc-4.2-20070613 is now available
2 msg	Backtrace() called inside a signal handler trap...
2 msg	Generating a phi node
2 msg	Blackfin eh broken with dataflow merge

Subject:	Incorrect bitfield aliasing with Tree SSA
Group:	Gcc
From:	Adam Nemet
Date:	15 Jun 2007

get_alias_set and internally record_component_aliases makes
assumptions about the IR that are only valid in RTL. As a consequence
the constant 1 is propagated into the function call in the example
below (I found this issue with 4.1 but it is also present on trunk):

struct s
{
long long a:12;
long long b:12;
long long c:40;
};

struct s s, *p = &s;

f ()
{
p->a = 1;
s.a = 2;
s.b = 3;
use (p->a, s.b);
}

or with VOPS:

# VUSE <p_7(D)>
p.0_1 = p;
# SMT.7_9 = VDEF <SMT.7_8(D)> <----- missing SFT.2 VDEF
p.0_1->a = 1;
# SFT.2_11 = VDEF <SFT.2_10(D)>
s.a = 2;
# SFT.1_13 = VDEF <SFT.1_12(D)>
s.b = 3;
# p_14 = VDEF <p_7(D)>
# SFT.1_15 = VDEF <SFT.1_13>
# SFT.2_16 = VDEF <SFT.2_11>
# SMT.7_17 = VDEF <SMT.7_9>
use (1, 3) [tail call];

In RTL the expmed.c functions extract_bit_field and store_bit_field
change the alias set of a bit-extraction or bit-insertion expression
of a memory operand to 0 because the actual RTL generated will access
adjacent fields. Therefore there is no need to represent alias-subset
relationship between non-addressable fields and their containing
structure and record_component_aliases will skip such fields.

First I was trying to remove this assumption even for RTL but that
increased memory consumption on the cc1 preprocessed files by 1%.

The other idea I had was to make this check dependent on the IR type.
The problem is that the function tree_register_cfg_hooks setting the
IR type is called only after gimplification whereas get_alias_set is
called during. Would it be acceptable to call tree_register_cfg_hooks
earlier? Hopefully, this will be a minimal change and would make it
easy to backport it to 4.1 and 4.2.

Another option would be to have a Tree-SSA-specific version of
get_alias_set and change all the invocations in the tree-* modules.

Do people have other recommendations or a preference?

Adam