| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
This small helper was used in unssa.c but is useful elsewhere too.
Move it as an inline function to linearize.h and rename it to
'nbr_users()' since it is close to the existing 'has_users()'.
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
|
|
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Signed-off-by: Christopher Li <sparse@chrisli.org>
|
|
A OP_PHISOURCE which is the only user of its operand
can be trivially eliminated. For example, in:
add %r6, ...
...
phisrc %rt, %r6
the phisrc can safely be eliminated if no other instruction use %r6.
With this patch it's rewritten as:
add %rt, ...
...
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Signed-off-by: Christopher Li <sparse@chrisli.org>
|
|
OP_PHI's target can interfer with it's own source swhen defined
in the same basic block. Such interference can create problem
like the 'swap problem' (which only exist if the phi-node are
'processed' sequentially if they're processed in parallel such
problems don't exist) when phi-nodes are destructed. To avoid
such problems OP_PHI are rewritten as OP_COPY.
if an OP_PHI and it's OP_PHISOURCE are in different basic blocks
no such interference is possible and the copy is not needed.
This patch detect such situation and eliminate these unneeded copies.
Note: during unSSA we're removing the OP_PHI & OP_PHISOURCE
but we need to use the def-use chains between them. We must
thus not use kill_instruction() in OP_PHI (this would break
def-use chains and leave stray OP_PHISOURCE), it's enough
to set their bb to NULL.
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Signed-off-by: Christopher Li <sparse@chrisli.org>
|
|
Using the fact that each OP_PHISOURCE is used by a single OP_PHI,
it's easier to rewrite OP_PHISOURCE at the same time as their
associated OP_PHI (because we have access to the new pseudo just created).
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Signed-off-by: Christopher Li <sparse@chrisli.org>
|
|
The unSSA step used to try to maintain the liveness info
while creating the copies but this can't be done so simply
(what is updated is only the liveness for the current bb
while it needs to be done for all concerned bbs).
If/when liveness is needed after this step, it need to be
redone by calling track_pseudo_liveness().
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Signed-off-by: Christopher Li <sparse@chrisli.org>
|
|
Currently, it is a completely "isolated island" from backend
point of view as it also lacks pseudo->def information.
Signed-off-by: Jan Pokorný <pokorny_jan@seznam.cz>
Signed-off-by: Christopher Li <sparse@chrisli.org>
|
|
Hello,
attached are patch, testing input for test-unssa and its outputs before patch
and after patch. Thanks in advance for considering the patch!
Kamil
test:
.L0x7f9fb2030010
<entry-point>
phisrc.32 %phi2(ptr) <- %arg1
br .L0x7f9fb2030130
.L0x7f9fb2030130
copy.32 %r1(ptr) <- %r5(ptr)
br %r1(ptr), .L0x7f9fb2030058, .L0x7f9fb20300e8
.L0x7f9fb2030058
load.32 %r3 <- 0[%r1(ptr)]
phisrc.32 %phi3(ptr) <- %r3
br .L0x7f9fb2030130
.L0x7f9fb20300e8
ret
test:
.L0x7f4a7f7f1010
<entry-point>
copy.32 %r5(ptr) <- %arg1
br .L0x7f4a7f7f1130
.L0x7f4a7f7f1130
copy.32 %r1(ptr) <- %r5(ptr)
br %r1(ptr), .L0x7f4a7f7f1058, .L0x7f4a7f7f10e8
.L0x7f4a7f7f1058
load.32 %r3 <- 0[%r1(ptr)]
copy.32 %r5(ptr) <- %r3
br .L0x7f4a7f7f1130
.L0x7f4a7f7f10e8
ret
>From 66a02fa7cec780fc88d6ef4cce7a1e704928808a Mon Sep 17 00:00:00 2001
From: Kamil Dudka <kdudka@redhat.com>
Date: Sun, 9 Aug 2009 10:22:11 +0200
Subject: [PATCH] unssa: track uses when replacing a phi node
The output of test-unssa is inconsistent for a simple test-case without
this patch:
static void test(void **ptr)
{
while (ptr) {
ptr = *ptr;
}
}
Signed-off-by: Kamil Dudka <kdudka@redhat.com>
Signed-off-by: Christopher Li <sparse@chrisli.org>
|
|
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@looxix.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
I'm not 100% sure it is done correctly, but at least it gives
sensible results.
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@looxix.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
This is an almost total rewrite of unssa().
It work now in two passs:
- first all the phi nodes are replaced by a copy of a new temporary to
the same target as the phi node.
- then, all the phisrc are replaced by one or several copies to the
temporaries just created, one for each phi node the phisrc feed.
It now works on my small tests and on the sparse source files and seems
also work well against the kernel sources.
There remains (at least) two problems:
- I need to update the bb's needs and defines.
- In some files, a few phisrc remains but they have an empty phi_users
and shouldn't have any effect. Maybe they come from dead basic
blocks?
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@looxix.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
For now, it use a simple method but which introduces a lot more copies
than necessary. Can be fixed later.
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@looxix.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
