Skip to content

R
riscv-gcc-1
Commit 3446405d by Jan Hubicka Committed by Jan Hubicka
Options

final.c: Undo my previous accidental checkin. 

	* final.c: Undo my previous accidental checkin.
	* output.h: Likewise.
	* tm.texi: Likewise.

From-SVN: r44960
parent 823e2b91
Showing with 64 additions and 133 deletions
gcc/ChangeLog
Fri Aug 17 15:41:35 CEST 2001 Jan Hubicka <jh@suse.cz>
* final.c: Undo my previous accidental checkin.
* output.h: Likewise.
* tm.texi: Likewise.
2001-08-16 Richard Henderson <rth@redhat.com>
* varasm.c (named_section_flags): Remove align parameter.
......
gcc/doc/tm.texi
......@@ -7137,20 +7137,6 @@ the target supports DWARF 2 frame unwind information.
This describes commands for alignment.
@table @code
@findex JUMP_ALIGN_AFTER_BARRIER
@item JUMP_ALIGN (@var{label})
The alignment (log base 2) to put in front of @var{label}, which is
a common destination of jumps and has no fallthru incomming edge.
This macro need not be defined if you don't want any special alignment
to be done at such a time. Most machine descriptions do not currently
define the macro.
Unless it's necessary to inspect the @var{label} parameter, it is better
to set the variable @var{align_jumps} in the target's
@code{OVERRIDE_OPTIONS}. Otherwise, you should try to honour the user's
selection in @var{align_jumps} in a @code{JUMP_ALIGN} implementation.
@findex LABEL_ALIGN_AFTER_BARRIER
@item LABEL_ALIGN_AFTER_BARRIER (@var{label})
The alignment (log base 2) to put in front of @var{label}, which follows
......@@ -7160,6 +7146,12 @@ This macro need not be defined if you don't want any special alignment
to be done at such a time. Most machine descriptions do not currently
define the macro.
Unless it's necessary to inspect the @var{label} parameter, it is better
to set the variable @var{align_jumps} in the target's
@code{OVERRIDE_OPTIONS}. Otherwise, you should try to honour the user's
selection in @var{align_jumps} in a @code{LABEL_ALIGN_AFTER_BARRIER}
implementation.
@findex LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP
@item LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP
The maximum number of bytes to skip when applying
......
gcc/final.c
......@@ -632,6 +632,11 @@ static struct label_alignment *label_align;
void
init_insn_lengths ()
{
if (label_align)
{
free (label_align);
label_align = 0;
}
if (uid_shuid)
{
free (uid_shuid);
......@@ -775,19 +780,11 @@ get_attr_length (insn)
#endif
#ifndef LABEL_ALIGN_AFTER_BARRIER
#define LABEL_ALIGN_AFTER_BARRIER(LABEL) 1
#define LABEL_ALIGN_AFTER_BARRIER(LABEL) align_jumps_log
#endif
#ifndef LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP
#define LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP 0
#endif
#ifndef JUMP_ALIGN
#define JUMP_ALIGN(LABEL) align_jumps_log
#endif
#ifndef JUMP_ALIGN_MAX_SKIP
#define JUMP_ALIGN_MAX_SKIP (align_jumps-1)
#define LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP (align_jumps-1)
#endif
#ifndef ADDR_VEC_ALIGN
......@@ -938,85 +935,6 @@ insn_current_reference_address (branch)
}
#endif /* HAVE_ATTR_length */
void
compute_alignments ()
{
int i;
int log, max_skip, max_log;
if (label_align)
{
free (label_align);
label_align = 0;
}
max_labelno = max_label_num ();
min_labelno = get_first_label_num ();
label_align = (struct label_alignment *) xcalloc ((max_labelno - min_labelno + 1),
sizeof (struct label_alignment));
/* If not optimizing or optimizing for size, don't assign any alignments. */
for (i = 0; i < n_basic_blocks; i++)
{
basic_block bb = BASIC_BLOCK (i);
rtx label = bb->head;
int fallthru_frequency = 0, branch_frequency = 0, has_fallthru = 0;
edge e;
if (GET_CODE (label) != CODE_LABEL)
continue;
max_log = LABEL_ALIGN (label);
max_skip = LABEL_ALIGN_MAX_SKIP;
for (e = bb->pred; e; e = e->pred_next)
{
if (e->flags & EDGE_FALLTHRU)
has_fallthru = 1, fallthru_frequency += EDGE_FREQUENCY (e);
else
branch_frequency += EDGE_FREQUENCY (e);
}
/* There are two purposes to align block with no fallthru incomming edge:
1) to avoid fetch stalls when branch destination is near cache boundary
2) to improve cache effciency in case the previous block is not executed
(so it does not need to be in the cache).
We to catch first case, we align frequently executed blocks.
To catch the second, we align blocks that are executed more frequently
than the predecesor and the predecesor is likely to not be executed
when function is called. */
if (!has_fallthru
&& (branch_frequency > BB_FREQ_MAX / 10
|| (bb->frequency > BASIC_BLOCK (i - 1)->frequency * 10
&& (BASIC_BLOCK (i - 1)->frequency
<= ENTRY_BLOCK_PTR->frequency / 2))))
{
log = JUMP_ALIGN (label);
if (max_log < log)
{
max_log = log;
max_skip = JUMP_ALIGN_MAX_SKIP;
}
}
/* In case block is frequent and reached mostly by non-fallthru edge,
align it. It is most likely an first block of loop. */
if (has_fallthru
&& branch_frequency + fallthru_frequency > BB_FREQ_MAX / 10
&& branch_frequency > fallthru_frequency * 5)
{
log = LOOP_ALIGN (label);
if (max_log < log)
{
max_log = log;
max_skip = LOOP_ALIGN_MAX_SKIP;
}
}
LABEL_TO_ALIGNMENT (label) = max_log;
LABEL_TO_MAX_SKIP (label) = max_skip;
}
}
/* Make a pass over all insns and compute their actual lengths by shortening
any branches of variable length if possible. */
......@@ -1054,29 +972,20 @@ shorten_branches (first)
#endif
/* Compute maximum UID and allocate label_align / uid_shuid. */
max_uid = get_max_uid ();
/* We must do some computations even when not actually shortening, in
order to get the alignment information for the labels. */
uid_shuid = (int *) xmalloc (max_uid * sizeof *uid_shuid);
init_insn_lengths ();
if (max_labelno != max_label_num ())
{
int old = max_labelno;
int n_labels;
int n_old_labels;
max_labelno = max_label_num ();
n_labels = max_labelno - min_labelno + 1;
n_old_labels = old - min_labelno + 1;
/* Compute maximum UID and allocate label_align / uid_shuid. */
max_uid = get_max_uid ();
label_align = (struct label_alignment *) xrealloc
(label_align, n_labels * sizeof (struct label_alignment));
max_labelno = max_label_num ();
min_labelno = get_first_label_num ();
label_align = (struct label_alignment *)
xcalloc ((max_labelno - min_labelno + 1), sizeof (struct label_alignment));
if (n_old_labels < n_labels)
memset (label_align + n_old_labels, 0,
(n_labels - n_old_labels) * sizeof (struct label_alignment));
}
uid_shuid = (int *) xmalloc (max_uid * sizeof *uid_shuid);
/* Initialize label_align and set up uid_shuid to be strictly
monotonically rising with insn order. */
......@@ -1103,14 +1012,6 @@ shorten_branches (first)
else if (GET_CODE (insn) == CODE_LABEL)
{
rtx next;
/* Merge in alignments computed by compute_alignments. */
log = LABEL_TO_ALIGNMENT (insn);
if (max_log < log)
{
max_log = log;
max_skip = LABEL_TO_MAX_SKIP (insn);
}
log = LABEL_ALIGN (insn);
if (max_log < log)
......@@ -1162,6 +1063,41 @@ shorten_branches (first)
break;
}
}
/* Again, we allow NOTE_INSN_LOOP_BEG - INSN - CODE_LABEL
sequences in order to handle reorg output efficiently. */
else if (GET_CODE (insn) == NOTE
&& NOTE_LINE_NUMBER (insn) == NOTE_INSN_LOOP_BEG)
{
rtx label;
int nest = 0;
/* Search for the label that starts the loop.
Don't skip past the end of the loop, since that could
lead to putting an alignment where it does not belong.
However, a label after a nested (non-)loop would be OK. */
for (label = insn; label; label = NEXT_INSN (label))
{
if (GET_CODE (label) == NOTE
&& NOTE_LINE_NUMBER (label) == NOTE_INSN_LOOP_BEG)
nest++;
else if (GET_CODE (label) == NOTE
&& NOTE_LINE_NUMBER (label) == NOTE_INSN_LOOP_END
&& --nest == 0)
break;
else if (GET_CODE (label) == CODE_LABEL)
{
log = LOOP_ALIGN (label);
if (max_log < log)
{
max_log = log;
max_skip = LOOP_ALIGN_MAX_SKIP;
}
break;
}
}
}
else
continue;
}
#ifdef HAVE_ATTR_length
......
gcc/output.h
......@@ -20,9 +20,6 @@ along with GNU CC; see the file COPYING. If not, write to
the Free Software Foundation, 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA. */
/* Compute branch alignments based on frequency information in the CFG. */
extern void compute_alignments PARAMS ((void));
/* Initialize data in final at the beginning of a compilation. */
extern void init_final PARAMS ((const char *));
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment