* [PATCH] tree-optimization/99912 - delete trivially dead stmts during DSE
@ 2021-04-29 6:31 Richard Biener
0 siblings, 0 replies; 4+ messages in thread
From: Richard Biener @ 2021-04-29 6:31 UTC (permalink / raw)
To: gcc-patches
DSE performs a backwards walk over stmts removing stores but it
leaves removing resulting dead SSA defs to later passes. This
eats into its own alias walking budget if the removed stores kept
loads live. The following patch adds removal of trivially dead
SSA defs which helps in this situation and reduces the amount of
garbage followup passes need to deal with.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
2021-04-28 Richard Biener <rguenther@suse.de>
PR tree-optimization/99912
* tree-ssa-dse.c (dse_dom_walker::m_need_cfg_cleanup): New.
(dse_dom_walker::todo): Likewise.
(dse_dom_walker::dse_optimize_stmt): Move VDEF check to the
caller.
(dse_dom_walker::before_dom_children): Remove trivially
dead SSA defs and schedule CFG cleanup if we removed all
PHIs in a block.
(pass_dse::execute): Get TODO as computed by the DOM walker
and return it. Wipe dominator info earlier.
* gcc.dg/pr95580.c: Disable DSE.
* gcc.dg/Wrestrict-8.c: Place a use after each memcpy.
* c-c++-common/ubsan/overflow-negate-3.c: Make asms volatile
to prevent them from being removed.
* c-c++-common/ubsan/overflow-sub-4.c: Likewise.
---
.../c-c++-common/ubsan/overflow-negate-3.c | 6 +-
.../c-c++-common/ubsan/overflow-sub-4.c | 6 +-
gcc/testsuite/gcc.dg/Wrestrict-8.c | 4 +-
gcc/testsuite/gcc.dg/pr95580.c | 2 +-
gcc/tree-ssa-dse.c | 71 +++++++++++++++----
5 files changed, 69 insertions(+), 20 deletions(-)
diff --git a/gcc/testsuite/c-c++-common/ubsan/overflow-negate-3.c b/gcc/testsuite/c-c++-common/ubsan/overflow-negate-3.c
index 85acce8e729..6b612fa3fb7 100644
--- a/gcc/testsuite/c-c++-common/ubsan/overflow-negate-3.c
+++ b/gcc/testsuite/c-c++-common/ubsan/overflow-negate-3.c
@@ -8,11 +8,11 @@ main ()
{
int x = INT_MIN;
int y;
- asm ("" : "+g" (x));
+ __asm__ volatile ("" : "+g" (x));
y = -(-x);
- asm ("" : "+g" (y));
+ __asm__ volatile ("" : "+g" (y));
y = -(-INT_MIN);
- asm ("" : "+g" (y));
+ __asm__ volatile ("" : "+g" (y));
}
/* { dg-output "negation of -2147483648 cannot be represented in type 'int'\[^\n\r]*; cast to an unsigned type to negate this value to itself\[^\n\r]*(\n|\r\n|\r)" } */
diff --git a/gcc/testsuite/c-c++-common/ubsan/overflow-sub-4.c b/gcc/testsuite/c-c++-common/ubsan/overflow-sub-4.c
index d3fb9baeec9..14ed5fb93cd 100644
--- a/gcc/testsuite/c-c++-common/ubsan/overflow-sub-4.c
+++ b/gcc/testsuite/c-c++-common/ubsan/overflow-sub-4.c
@@ -9,10 +9,10 @@ main ()
int x = INT_MIN;
int y = 0;
int z;
- asm ("" : "+g" (y));
- asm ("" : "+g" (x));
+ __asm__ volatile ("" : "+g" (y));
+ __asm__ volatile ("" : "+g" (x));
z = y - (-x);
- asm ("" : "+g" (z));
+ __asm__ volatile ("" : "+g" (z));
}
/* { dg-output "negation of -2147483648 cannot be represented in type 'int'\[^\n\r]*; cast to an unsigned type to negate this value to itself\[^\n\r]*(\n|\r\n|\r)" } */
diff --git a/gcc/testsuite/gcc.dg/Wrestrict-8.c b/gcc/testsuite/gcc.dg/Wrestrict-8.c
index 24946b08c3b..62e8bbc3027 100644
--- a/gcc/testsuite/gcc.dg/Wrestrict-8.c
+++ b/gcc/testsuite/gcc.dg/Wrestrict-8.c
@@ -7,7 +7,9 @@ typedef __SIZE_TYPE__ size_t;
extern void* memcpy (void* restrict, const void* restrict, size_t);
-#define T(d, s, n) memcpy (d, s, n)
+void foo (void *);
+
+#define T(d, s, n) do { memcpy (d, s, n); foo (d); } while (0)
struct S1 { char c; } a8_1[8];
diff --git a/gcc/testsuite/gcc.dg/pr95580.c b/gcc/testsuite/gcc.dg/pr95580.c
index 330a3137f1e..77d8150baa8 100644
--- a/gcc/testsuite/gcc.dg/pr95580.c
+++ b/gcc/testsuite/gcc.dg/pr95580.c
@@ -1,6 +1,6 @@
/* PR c/95580 */
/* { dg-do compile } */
-/* { dg-options "-O1 -W -fno-tree-dce" } */
+/* { dg-options "-O1 -W -fno-tree-dce -fno-tree-dse" } */
void bar (void);
diff --git a/gcc/tree-ssa-dse.c b/gcc/tree-ssa-dse.c
index 4967a5a9927..aecf6ab8c46 100644
--- a/gcc/tree-ssa-dse.c
+++ b/gcc/tree-ssa-dse.c
@@ -963,16 +963,25 @@ public:
dse_dom_walker (cdi_direction direction)
: dom_walker (direction),
m_live_bytes (param_dse_max_object_size),
- m_byte_tracking_enabled (false) {}
+ m_byte_tracking_enabled (false),
+ m_need_cfg_cleanup (false) {}
virtual edge before_dom_children (basic_block);
+ unsigned todo () const;
private:
auto_sbitmap m_live_bytes;
bool m_byte_tracking_enabled;
+ bool m_need_cfg_cleanup;
void dse_optimize_stmt (gimple_stmt_iterator *);
};
+unsigned
+dse_dom_walker::todo () const
+{
+ return m_need_cfg_cleanup ? TODO_cleanup_cfg : 0;
+}
+
/* Delete a dead call at GSI, which is mem* call of some kind. */
static void
delete_dead_or_redundant_call (gimple_stmt_iterator *gsi, const char *type)
@@ -1049,11 +1058,6 @@ dse_dom_walker::dse_optimize_stmt (gimple_stmt_iterator *gsi)
{
gimple *stmt = gsi_stmt (*gsi);
- /* If this statement has no virtual defs, then there is nothing
- to do. */
- if (!gimple_vdef (stmt))
- return;
-
/* Don't return early on *this_2(D) ={v} {CLOBBER}. */
if (gimple_has_volatile_ops (stmt)
&& (!gimple_clobber_p (stmt)
@@ -1180,12 +1184,53 @@ dse_dom_walker::before_dom_children (basic_block bb)
for (gsi = gsi_last_bb (bb); !gsi_end_p (gsi);)
{
- dse_optimize_stmt (&gsi);
+ gimple *stmt = gsi_stmt (gsi);
+
+ if (gimple_vdef (stmt))
+ dse_optimize_stmt (&gsi);
+ else if (def_operand_p def_p = single_ssa_def_operand (stmt, SSA_OP_DEF))
+ {
+ /* When we remove dead stores make sure to also delete trivially
+ dead SSA defs. */
+ if (has_zero_uses (DEF_FROM_PTR (def_p))
+ && !gimple_has_side_effects (stmt))
+ {
+ if (dump_file && (dump_flags & TDF_DETAILS))
+ {
+ fprintf (dump_file, " Deleted trivially dead stmt: ");
+ print_gimple_stmt (dump_file, stmt, 0, dump_flags);
+ fprintf (dump_file, "\n");
+ }
+ if (gsi_remove (&gsi, true) && need_eh_cleanup)
+ bitmap_set_bit (need_eh_cleanup, bb->index);
+ release_defs (stmt);
+ }
+ }
if (gsi_end_p (gsi))
gsi = gsi_last_bb (bb);
else
gsi_prev (&gsi);
}
+ bool removed_phi = false;
+ for (gphi_iterator si = gsi_start_phis (bb); !gsi_end_p (si);)
+ {
+ gphi *phi = si.phi ();
+ if (has_zero_uses (gimple_phi_result (phi)))
+ {
+ if (dump_file && (dump_flags & TDF_DETAILS))
+ {
+ fprintf (dump_file, " Deleted trivially dead PHI: ");
+ print_gimple_stmt (dump_file, phi, 0, dump_flags);
+ fprintf (dump_file, "\n");
+ }
+ remove_phi_node (&si, true);
+ removed_phi = true;
+ }
+ else
+ gsi_next (&si);
+ }
+ if (removed_phi && gimple_seq_empty_p (phi_nodes (bb)))
+ m_need_cfg_cleanup = true;
return NULL;
}
@@ -1234,21 +1279,23 @@ pass_dse::execute (function *fun)
/* Dead store elimination is fundamentally a walk of the post-dominator
tree and a backwards walk of statements within each block. */
- dse_dom_walker (CDI_POST_DOMINATORS).walk (fun->cfg->x_exit_block_ptr);
+ dse_dom_walker walker (CDI_POST_DOMINATORS);
+ walker.walk (fun->cfg->x_exit_block_ptr);
+ free_dominance_info (CDI_POST_DOMINATORS);
+
+ unsigned todo = walker.todo ();
/* Removal of stores may make some EH edges dead. Purge such edges from
the CFG as needed. */
if (!bitmap_empty_p (need_eh_cleanup))
{
gimple_purge_all_dead_eh_edges (need_eh_cleanup);
- cleanup_tree_cfg ();
+ todo |= TODO_cleanup_cfg;
}
BITMAP_FREE (need_eh_cleanup);
- /* For now, just wipe the post-dominator information. */
- free_dominance_info (CDI_POST_DOMINATORS);
- return 0;
+ return todo;
}
} // anon namespace
--
2.26.2
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] tree-optimization/99912 - delete trivially dead stmts during DSE
2021-04-27 16:06 ` Prathamesh Kulkarni
@ 2021-04-28 5:55 ` Richard Biener
0 siblings, 0 replies; 4+ messages in thread
From: Richard Biener @ 2021-04-28 5:55 UTC (permalink / raw)
To: Prathamesh Kulkarni; +Cc: gcc Patches
On Tue, 27 Apr 2021, Prathamesh Kulkarni wrote:
> On Tue, 27 Apr 2021 at 19:19, Richard Biener <rguenther@suse.de> wrote:
> >
> > DSE performs a backwards walk over stmts removing stores but it
> > leaves removing resulting dead SSA defs to later passes. This
> > eats into its own alias walking budget if the removed stores kept
> > loads live. The following patch adds removal of trivially dead
> > SSA defs which helps in this situation and reduces the amount of
> > garbage followup passes need to deal with.
> >
> > Bootstrapped on x86_64-unknown-linux-gnu, testing in progress.
> >
> > 2021-04-27 Richard Biener <rguenther@suse.de>
> >
> > PR tree-optimization/99912
> > * tree-ssa-dse.c (dse_dom_walker::m_need_cfg_cleanup): New.
> > (dse_dom_walker::todo): Likewise.
> > (dse_dom_walker::dse_optimize_stmt): Move VDEF check to the
> > caller.
> > (dse_dom_walker::before_dom_children): Remove trivially
> > dead SSA defs and schedule CFG cleanup if we removed all
> > PHIs in a block.
> > (pass_dse::execute): Get TODO as computed by the DOM walker
> > and return it. Wipe dominator info earlier.
> > ---
> > gcc/tree-ssa-dse.c | 65 +++++++++++++++++++++++++++++++++++++---------
> > 1 file changed, 53 insertions(+), 12 deletions(-)
> >
> > diff --git a/gcc/tree-ssa-dse.c b/gcc/tree-ssa-dse.c
> > index 4967a5a9927..f5f39cbe903 100644
> > --- a/gcc/tree-ssa-dse.c
> > +++ b/gcc/tree-ssa-dse.c
> > @@ -963,16 +963,25 @@ public:
> > dse_dom_walker (cdi_direction direction)
> > : dom_walker (direction),
> > m_live_bytes (param_dse_max_object_size),
> > - m_byte_tracking_enabled (false) {}
> > + m_byte_tracking_enabled (false),
> > + m_need_cfg_cleanup (false) {}
> >
> > virtual edge before_dom_children (basic_block);
> > + unsigned todo () const;
> >
> > private:
> > auto_sbitmap m_live_bytes;
> > bool m_byte_tracking_enabled;
> > + bool m_need_cfg_cleanup;
> > void dse_optimize_stmt (gimple_stmt_iterator *);
> > };
> >
> > +unsigned
> > +dse_dom_walker::todo () const
> > +{
> > + return m_need_cfg_cleanup ? TODO_cleanup_cfg : 0;
> > +}
> > +
> > /* Delete a dead call at GSI, which is mem* call of some kind. */
> > static void
> > delete_dead_or_redundant_call (gimple_stmt_iterator *gsi, const char *type)
> > @@ -1049,11 +1058,6 @@ dse_dom_walker::dse_optimize_stmt (gimple_stmt_iterator *gsi)
> > {
> > gimple *stmt = gsi_stmt (*gsi);
> >
> > - /* If this statement has no virtual defs, then there is nothing
> > - to do. */
> > - if (!gimple_vdef (stmt))
> > - return;
> > -
> > /* Don't return early on *this_2(D) ={v} {CLOBBER}. */
> > if (gimple_has_volatile_ops (stmt)
> > && (!gimple_clobber_p (stmt)
> > @@ -1180,12 +1184,47 @@ dse_dom_walker::before_dom_children (basic_block bb)
> >
> > for (gsi = gsi_last_bb (bb); !gsi_end_p (gsi);)
> > {
> > - dse_optimize_stmt (&gsi);
> > + gimple *stmt = gsi_stmt (gsi);
> > +
> > + if (gimple_vdef (stmt))
> > + dse_optimize_stmt (&gsi);
> > + else if (def_operand_p def_p = single_ssa_def_operand (stmt, SSA_OP_DEF))
> > + {
> > + /* When we remove dead stores make sure to also delete trivially
> > + dead SSA defs. */
> > + if (has_zero_uses (DEF_FROM_PTR (def_p))
> > + && !gimple_has_side_effects (stmt))
> > + {
> > + if (dump_file && (dump_flags & TDF_DETAILS))
> > + {
> > + fprintf (dump_file, " Deleted trivially dead stmt: ");
> > + print_gimple_stmt (dump_file, stmt, 0, dump_flags);
> > + fprintf (dump_file, "\n");
> > + }
> > + if (gsi_remove (&gsi, true) && need_eh_cleanup)
> > + bitmap_set_bit (need_eh_cleanup, bb->index);
> > + release_defs (stmt);
> > + }
> > + }
> > if (gsi_end_p (gsi))
> > gsi = gsi_last_bb (bb);
> > else
> > gsi_prev (&gsi);
> > }
> > + bool removed_phi = false;
> > + for (gphi_iterator si = gsi_start_phis (bb); !gsi_end_p (si);)
> > + {
> > + gphi *phi = si.phi ();
> > + if (has_zero_uses (gimple_phi_result (phi)))
> > + {
> > + remove_phi_node (&si, true);
> > + removed_phi = true;
> > + }
> Hi Richard,
> Just curious if this is missing dumping info about removed phi node to
> dump_file ?
Yes, I'll add this. I also have to investigate some testsuite fallout,
so there'll be v2.
Richard.
> Thanks,
> Prathamesh
> > + else
> > + gsi_next (&si);
> > + }
> > + if (removed_phi && gimple_seq_empty_p (phi_nodes (bb)))
> > + m_need_cfg_cleanup = true;
> > return NULL;
> > }
> >
> > @@ -1234,21 +1273,23 @@ pass_dse::execute (function *fun)
> >
> > /* Dead store elimination is fundamentally a walk of the post-dominator
> > tree and a backwards walk of statements within each block. */
> > - dse_dom_walker (CDI_POST_DOMINATORS).walk (fun->cfg->x_exit_block_ptr);
> > + dse_dom_walker walker (CDI_POST_DOMINATORS);
> > + walker.walk (fun->cfg->x_exit_block_ptr);
> > + free_dominance_info (CDI_POST_DOMINATORS);
> > +
> > + unsigned todo = walker.todo ();
> >
> > /* Removal of stores may make some EH edges dead. Purge such edges from
> > the CFG as needed. */
> > if (!bitmap_empty_p (need_eh_cleanup))
> > {
> > gimple_purge_all_dead_eh_edges (need_eh_cleanup);
> > - cleanup_tree_cfg ();
> > + todo |= TODO_cleanup_cfg;
> > }
> >
> > BITMAP_FREE (need_eh_cleanup);
> >
> > - /* For now, just wipe the post-dominator information. */
> > - free_dominance_info (CDI_POST_DOMINATORS);
> > - return 0;
> > + return todo;
> > }
> >
> > } // anon namespace
> > --
> > 2.26.2
>
--
Richard Biener <rguenther@suse.de>
SUSE Software Solutions Germany GmbH, Maxfeldstrasse 5, 90409 Nuernberg,
Germany; GF: Felix Imendörffer; HRB 36809 (AG Nuernberg)
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] tree-optimization/99912 - delete trivially dead stmts during DSE
2021-04-27 13:48 Richard Biener
@ 2021-04-27 16:06 ` Prathamesh Kulkarni
2021-04-28 5:55 ` Richard Biener
0 siblings, 1 reply; 4+ messages in thread
From: Prathamesh Kulkarni @ 2021-04-27 16:06 UTC (permalink / raw)
To: Richard Biener; +Cc: gcc Patches
On Tue, 27 Apr 2021 at 19:19, Richard Biener <rguenther@suse.de> wrote:
>
> DSE performs a backwards walk over stmts removing stores but it
> leaves removing resulting dead SSA defs to later passes. This
> eats into its own alias walking budget if the removed stores kept
> loads live. The following patch adds removal of trivially dead
> SSA defs which helps in this situation and reduces the amount of
> garbage followup passes need to deal with.
>
> Bootstrapped on x86_64-unknown-linux-gnu, testing in progress.
>
> 2021-04-27 Richard Biener <rguenther@suse.de>
>
> PR tree-optimization/99912
> * tree-ssa-dse.c (dse_dom_walker::m_need_cfg_cleanup): New.
> (dse_dom_walker::todo): Likewise.
> (dse_dom_walker::dse_optimize_stmt): Move VDEF check to the
> caller.
> (dse_dom_walker::before_dom_children): Remove trivially
> dead SSA defs and schedule CFG cleanup if we removed all
> PHIs in a block.
> (pass_dse::execute): Get TODO as computed by the DOM walker
> and return it. Wipe dominator info earlier.
> ---
> gcc/tree-ssa-dse.c | 65 +++++++++++++++++++++++++++++++++++++---------
> 1 file changed, 53 insertions(+), 12 deletions(-)
>
> diff --git a/gcc/tree-ssa-dse.c b/gcc/tree-ssa-dse.c
> index 4967a5a9927..f5f39cbe903 100644
> --- a/gcc/tree-ssa-dse.c
> +++ b/gcc/tree-ssa-dse.c
> @@ -963,16 +963,25 @@ public:
> dse_dom_walker (cdi_direction direction)
> : dom_walker (direction),
> m_live_bytes (param_dse_max_object_size),
> - m_byte_tracking_enabled (false) {}
> + m_byte_tracking_enabled (false),
> + m_need_cfg_cleanup (false) {}
>
> virtual edge before_dom_children (basic_block);
> + unsigned todo () const;
>
> private:
> auto_sbitmap m_live_bytes;
> bool m_byte_tracking_enabled;
> + bool m_need_cfg_cleanup;
> void dse_optimize_stmt (gimple_stmt_iterator *);
> };
>
> +unsigned
> +dse_dom_walker::todo () const
> +{
> + return m_need_cfg_cleanup ? TODO_cleanup_cfg : 0;
> +}
> +
> /* Delete a dead call at GSI, which is mem* call of some kind. */
> static void
> delete_dead_or_redundant_call (gimple_stmt_iterator *gsi, const char *type)
> @@ -1049,11 +1058,6 @@ dse_dom_walker::dse_optimize_stmt (gimple_stmt_iterator *gsi)
> {
> gimple *stmt = gsi_stmt (*gsi);
>
> - /* If this statement has no virtual defs, then there is nothing
> - to do. */
> - if (!gimple_vdef (stmt))
> - return;
> -
> /* Don't return early on *this_2(D) ={v} {CLOBBER}. */
> if (gimple_has_volatile_ops (stmt)
> && (!gimple_clobber_p (stmt)
> @@ -1180,12 +1184,47 @@ dse_dom_walker::before_dom_children (basic_block bb)
>
> for (gsi = gsi_last_bb (bb); !gsi_end_p (gsi);)
> {
> - dse_optimize_stmt (&gsi);
> + gimple *stmt = gsi_stmt (gsi);
> +
> + if (gimple_vdef (stmt))
> + dse_optimize_stmt (&gsi);
> + else if (def_operand_p def_p = single_ssa_def_operand (stmt, SSA_OP_DEF))
> + {
> + /* When we remove dead stores make sure to also delete trivially
> + dead SSA defs. */
> + if (has_zero_uses (DEF_FROM_PTR (def_p))
> + && !gimple_has_side_effects (stmt))
> + {
> + if (dump_file && (dump_flags & TDF_DETAILS))
> + {
> + fprintf (dump_file, " Deleted trivially dead stmt: ");
> + print_gimple_stmt (dump_file, stmt, 0, dump_flags);
> + fprintf (dump_file, "\n");
> + }
> + if (gsi_remove (&gsi, true) && need_eh_cleanup)
> + bitmap_set_bit (need_eh_cleanup, bb->index);
> + release_defs (stmt);
> + }
> + }
> if (gsi_end_p (gsi))
> gsi = gsi_last_bb (bb);
> else
> gsi_prev (&gsi);
> }
> + bool removed_phi = false;
> + for (gphi_iterator si = gsi_start_phis (bb); !gsi_end_p (si);)
> + {
> + gphi *phi = si.phi ();
> + if (has_zero_uses (gimple_phi_result (phi)))
> + {
> + remove_phi_node (&si, true);
> + removed_phi = true;
> + }
Hi Richard,
Just curious if this is missing dumping info about removed phi node to
dump_file ?
Thanks,
Prathamesh
> + else
> + gsi_next (&si);
> + }
> + if (removed_phi && gimple_seq_empty_p (phi_nodes (bb)))
> + m_need_cfg_cleanup = true;
> return NULL;
> }
>
> @@ -1234,21 +1273,23 @@ pass_dse::execute (function *fun)
>
> /* Dead store elimination is fundamentally a walk of the post-dominator
> tree and a backwards walk of statements within each block. */
> - dse_dom_walker (CDI_POST_DOMINATORS).walk (fun->cfg->x_exit_block_ptr);
> + dse_dom_walker walker (CDI_POST_DOMINATORS);
> + walker.walk (fun->cfg->x_exit_block_ptr);
> + free_dominance_info (CDI_POST_DOMINATORS);
> +
> + unsigned todo = walker.todo ();
>
> /* Removal of stores may make some EH edges dead. Purge such edges from
> the CFG as needed. */
> if (!bitmap_empty_p (need_eh_cleanup))
> {
> gimple_purge_all_dead_eh_edges (need_eh_cleanup);
> - cleanup_tree_cfg ();
> + todo |= TODO_cleanup_cfg;
> }
>
> BITMAP_FREE (need_eh_cleanup);
>
> - /* For now, just wipe the post-dominator information. */
> - free_dominance_info (CDI_POST_DOMINATORS);
> - return 0;
> + return todo;
> }
>
> } // anon namespace
> --
> 2.26.2
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH] tree-optimization/99912 - delete trivially dead stmts during DSE
@ 2021-04-27 13:48 Richard Biener
2021-04-27 16:06 ` Prathamesh Kulkarni
0 siblings, 1 reply; 4+ messages in thread
From: Richard Biener @ 2021-04-27 13:48 UTC (permalink / raw)
To: gcc-patches
DSE performs a backwards walk over stmts removing stores but it
leaves removing resulting dead SSA defs to later passes. This
eats into its own alias walking budget if the removed stores kept
loads live. The following patch adds removal of trivially dead
SSA defs which helps in this situation and reduces the amount of
garbage followup passes need to deal with.
Bootstrapped on x86_64-unknown-linux-gnu, testing in progress.
2021-04-27 Richard Biener <rguenther@suse.de>
PR tree-optimization/99912
* tree-ssa-dse.c (dse_dom_walker::m_need_cfg_cleanup): New.
(dse_dom_walker::todo): Likewise.
(dse_dom_walker::dse_optimize_stmt): Move VDEF check to the
caller.
(dse_dom_walker::before_dom_children): Remove trivially
dead SSA defs and schedule CFG cleanup if we removed all
PHIs in a block.
(pass_dse::execute): Get TODO as computed by the DOM walker
and return it. Wipe dominator info earlier.
---
gcc/tree-ssa-dse.c | 65 +++++++++++++++++++++++++++++++++++++---------
1 file changed, 53 insertions(+), 12 deletions(-)
diff --git a/gcc/tree-ssa-dse.c b/gcc/tree-ssa-dse.c
index 4967a5a9927..f5f39cbe903 100644
--- a/gcc/tree-ssa-dse.c
+++ b/gcc/tree-ssa-dse.c
@@ -963,16 +963,25 @@ public:
dse_dom_walker (cdi_direction direction)
: dom_walker (direction),
m_live_bytes (param_dse_max_object_size),
- m_byte_tracking_enabled (false) {}
+ m_byte_tracking_enabled (false),
+ m_need_cfg_cleanup (false) {}
virtual edge before_dom_children (basic_block);
+ unsigned todo () const;
private:
auto_sbitmap m_live_bytes;
bool m_byte_tracking_enabled;
+ bool m_need_cfg_cleanup;
void dse_optimize_stmt (gimple_stmt_iterator *);
};
+unsigned
+dse_dom_walker::todo () const
+{
+ return m_need_cfg_cleanup ? TODO_cleanup_cfg : 0;
+}
+
/* Delete a dead call at GSI, which is mem* call of some kind. */
static void
delete_dead_or_redundant_call (gimple_stmt_iterator *gsi, const char *type)
@@ -1049,11 +1058,6 @@ dse_dom_walker::dse_optimize_stmt (gimple_stmt_iterator *gsi)
{
gimple *stmt = gsi_stmt (*gsi);
- /* If this statement has no virtual defs, then there is nothing
- to do. */
- if (!gimple_vdef (stmt))
- return;
-
/* Don't return early on *this_2(D) ={v} {CLOBBER}. */
if (gimple_has_volatile_ops (stmt)
&& (!gimple_clobber_p (stmt)
@@ -1180,12 +1184,47 @@ dse_dom_walker::before_dom_children (basic_block bb)
for (gsi = gsi_last_bb (bb); !gsi_end_p (gsi);)
{
- dse_optimize_stmt (&gsi);
+ gimple *stmt = gsi_stmt (gsi);
+
+ if (gimple_vdef (stmt))
+ dse_optimize_stmt (&gsi);
+ else if (def_operand_p def_p = single_ssa_def_operand (stmt, SSA_OP_DEF))
+ {
+ /* When we remove dead stores make sure to also delete trivially
+ dead SSA defs. */
+ if (has_zero_uses (DEF_FROM_PTR (def_p))
+ && !gimple_has_side_effects (stmt))
+ {
+ if (dump_file && (dump_flags & TDF_DETAILS))
+ {
+ fprintf (dump_file, " Deleted trivially dead stmt: ");
+ print_gimple_stmt (dump_file, stmt, 0, dump_flags);
+ fprintf (dump_file, "\n");
+ }
+ if (gsi_remove (&gsi, true) && need_eh_cleanup)
+ bitmap_set_bit (need_eh_cleanup, bb->index);
+ release_defs (stmt);
+ }
+ }
if (gsi_end_p (gsi))
gsi = gsi_last_bb (bb);
else
gsi_prev (&gsi);
}
+ bool removed_phi = false;
+ for (gphi_iterator si = gsi_start_phis (bb); !gsi_end_p (si);)
+ {
+ gphi *phi = si.phi ();
+ if (has_zero_uses (gimple_phi_result (phi)))
+ {
+ remove_phi_node (&si, true);
+ removed_phi = true;
+ }
+ else
+ gsi_next (&si);
+ }
+ if (removed_phi && gimple_seq_empty_p (phi_nodes (bb)))
+ m_need_cfg_cleanup = true;
return NULL;
}
@@ -1234,21 +1273,23 @@ pass_dse::execute (function *fun)
/* Dead store elimination is fundamentally a walk of the post-dominator
tree and a backwards walk of statements within each block. */
- dse_dom_walker (CDI_POST_DOMINATORS).walk (fun->cfg->x_exit_block_ptr);
+ dse_dom_walker walker (CDI_POST_DOMINATORS);
+ walker.walk (fun->cfg->x_exit_block_ptr);
+ free_dominance_info (CDI_POST_DOMINATORS);
+
+ unsigned todo = walker.todo ();
/* Removal of stores may make some EH edges dead. Purge such edges from
the CFG as needed. */
if (!bitmap_empty_p (need_eh_cleanup))
{
gimple_purge_all_dead_eh_edges (need_eh_cleanup);
- cleanup_tree_cfg ();
+ todo |= TODO_cleanup_cfg;
}
BITMAP_FREE (need_eh_cleanup);
- /* For now, just wipe the post-dominator information. */
- free_dominance_info (CDI_POST_DOMINATORS);
- return 0;
+ return todo;
}
} // anon namespace
--
2.26.2
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2021-04-29 6:31 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-04-29 6:31 [PATCH] tree-optimization/99912 - delete trivially dead stmts during DSE Richard Biener
-- strict thread matches above, loose matches on Subject: below --
2021-04-27 13:48 Richard Biener
2021-04-27 16:06 ` Prathamesh Kulkarni
2021-04-28 5:55 ` Richard Biener
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).