URL
https://opencores.org/ocsvn/openrisc/openrisc/trunk
Subversion Repositories openrisc
[/] [openrisc/] [trunk/] [gnu-stable/] [gcc-4.5.1/] [gcc/] [dbgcnt.def] - Rev 280
Go to most recent revision | Compare with Previous | Blame | View Log
/* This file contains the list of the debug counter for GCC.
Copyright (C) 2006, 2007, 2008 Free Software Foundation, Inc.
This file is part of GCC.
GCC is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free
Software Foundation; either version 3, or (at your option) any later
version.
GCC is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
for more details.
You should have received a copy of the GNU General Public License
along with GCC; see the file COPYING3. If not see
<http://www.gnu.org/licenses/>. */
/* A debug counter provides you a way to count an event
and return false after the counter has exceeded the threshold
specified by the option.
What is it used for ?
This is primarily used to speed up the search for the bad transformation
an optimization pass does. By doing a binary search on N,
you can quickly narrow down to one transformation
which is bad, or which triggers the bad behavior downstream
(usually in the form of the badly generated code).
How does it work ?
Every time dbg_cnt(named-counter) is called,
the counter is incremented for the named-counter.
And the incremented value is compared against the threshold (limit)
specified by the option.
dbg_cnt () returns true if it is at or below threshold, and false if above.
How to add a new one ?
To add a new counter, simply add an entry below with some descriptive name,
and add call(s) to dbg_cnt(your-counter-name) in appropriate places.
Usually, you want to control at the finest granularity
any particular transformation can happen.
e.g. for each instruction in a dead code elimination,
or for each copy instruction in register coalescing,
or constant-propagation for each insn,
or a block straightening, etc.
See dce.c for an example. With the dbg_cnt () call in dce.c,
now a developer can use -fdbg-cnt=dce:N
to stop doing the dead code elimination after N times.
How to use it ?
By default, all limits are UINT_MAX.
Since debug count is unsigned int, <= UINT_MAX returns true always.
i.e. dbg_cnt() returns true always regardless of the counter value
(although it still counts the event).
Use -fdbg-cnt=counter1:N,counter2:M,...
which sets the limit for counter1 to N, and the limit for counter2 to M, etc.
e.g. setting a limit to zero will make dbg_cnt () return false *always*.
The following shell file can then be used to binary search for
exact transformation that causes the bug. A second shell script
should be written, say "tryTest", which exits with 1 if the
compiled program fails and exits with 0 if the program succeeds.
This shell script should take 1 parameter, the value to be passed
to set the counter of the compilation command in tryTest. Then,
assuming that the following script is called binarySearch,
the command:
binarySearch tryTest
will automatically find the highest value of the counter for which
the program fails. If tryTest never fails, binarySearch will
produce unpredictable results as it will try to find an upper bound
that does not exist.
When dbgcnt does hits the limit, it writes a comment in the current
dump_file of the form:
***dbgcnt: limit reached for %s.***
Assuming that the dump file is logging the analysis/transformations
it is making, this pinpoints the exact position in the log file
where the problem transformation is being logged.
=====================================
#!/bin/bash
while getopts "l:u:i:" opt
do
case $opt in
l) lb="$OPTARG";;
u) ub="$OPTARG";;
i) init="$OPTARG";;
?) usage; exit 3;;
esac
done
shift $(($OPTIND - 1))
echo $@
cmd=${1+"${@}"}
lb=${lb:=0}
init=${init:=100}
$cmd $lb
lb_val=$?
if [ -z "$ub" ]; then
# find the upper bound
ub=$(($init + $lb))
true
while [ $? -eq $lb_val ]; do
ub=$(($ub * 10))
#ub=`expr $ub \* 10`
$cmd $ub
done
fi
echo command: $cmd
true
while [ `expr $ub - $lb` -gt 1 ]; do
try=$(($lb + ( $ub - $lb ) / 2))
$cmd $try
if [ $? -eq $lb_val ]; then
lb=$try
else
ub=$try
fi
done
echo lbound: $lb
echo ubound: $ub
=====================================
*/
/* Debug counter definitions. */
DEBUG_COUNTER (auto_inc_dec)
DEBUG_COUNTER (ccp)
DEBUG_COUNTER (cfg_cleanup)
DEBUG_COUNTER (cse2_move2add)
DEBUG_COUNTER (cprop)
DEBUG_COUNTER (dce)
DEBUG_COUNTER (dce_fast)
DEBUG_COUNTER (dce_ud)
DEBUG_COUNTER (delete_trivial_dead)
DEBUG_COUNTER (df_byte_scan)
DEBUG_COUNTER (dse)
DEBUG_COUNTER (dse1)
DEBUG_COUNTER (dse2)
DEBUG_COUNTER (gcse2_delete)
DEBUG_COUNTER (global_alloc_at_func)
DEBUG_COUNTER (global_alloc_at_reg)
DEBUG_COUNTER (hoist)
DEBUG_COUNTER (ia64_sched2)
DEBUG_COUNTER (if_conversion)
DEBUG_COUNTER (if_after_combine)
DEBUG_COUNTER (if_after_reload)
DEBUG_COUNTER (local_alloc_for_sched)
DEBUG_COUNTER (postreload_cse)
DEBUG_COUNTER (pre)
DEBUG_COUNTER (pre_insn)
DEBUG_COUNTER (treepre_insert)
DEBUG_COUNTER (sched2_func)
DEBUG_COUNTER (sched_block)
DEBUG_COUNTER (sched_func)
DEBUG_COUNTER (sched_insn)
DEBUG_COUNTER (sched_region)
DEBUG_COUNTER (sel_sched_cnt)
DEBUG_COUNTER (sel_sched_region_cnt)
DEBUG_COUNTER (sel_sched_insn_cnt)
DEBUG_COUNTER (sms_sched_loop)
DEBUG_COUNTER (store_motion)
DEBUG_COUNTER (split_for_sched2)
DEBUG_COUNTER (tail_call)
Go to most recent revision | Compare with Previous | Blame | View Log