From e63d54db952634c2a3a6a24b6c2a4c6ff66f22bd Mon Sep 17 00:00:00 2001 From: Laszlo Ersek Date: Thu, 27 Feb 2020 22:39:03 +0100 Subject: [PATCH] ArmPlatformPkg: convert LFs to CRLF, expand hard TABs MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit We're going to switch the internal line terminators globally to LF at some point, but until then, let's use CRLF consistently. Convert source files with LFs in them to CRLF, using "unix2dos". "git show -b" prints no code changes for this patch. (I collected all the file name suffixes in this package, with: $ git ls-files -- $PACKAGE | rev | cut -f 1 -d . | sort -u | rev I eliminated those suffixes that didn't stand for text files, then blanket-converted the rest with unix2dos. Finally, picked up the actual changes with git-add.) At the same time, the following file had to undergo TAB expansion: ArmPlatformPkg/Scripts/Ds5/profile.py I used "expand -t 4", conforming to the Indentation section of PEP-8 . Both the CRLF conversion and the TAB expansion are motivated by "PatchCheck.py". "PatchCheck.py" is also the reason why CRLF conversion and TAB expansion have to happen in the same patch. Cc: Ard Biesheuvel Cc: Leif Lindholm Cc: Philippe Mathieu-Daudé Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=1659 Signed-off-by: Laszlo Ersek Message-Id: <20200227213903.13884-4-lersek@redhat.com> Acked-by: Ard Biesheuvel Reviewed-by: Philippe Mathieu-Daude --- ArmPlatformPkg/Scripts/Ds5/profile.py | 656 +++++++++++++------------- 1 file changed, 328 insertions(+), 328 deletions(-) diff --git a/ArmPlatformPkg/Scripts/Ds5/profile.py b/ArmPlatformPkg/Scripts/Ds5/profile.py index f87dee2469..979c6ea2bd 100644 --- a/ArmPlatformPkg/Scripts/Ds5/profile.py +++ b/ArmPlatformPkg/Scripts/Ds5/profile.py @@ -1,328 +1,328 @@ -#!/usr/bin/python - -# -# Copyright (c) 2014, ARM Limited. All rights reserved. -# -# SPDX-License-Identifier: BSD-2-Clause-Patent -# - -import getopt -import operator -import os -import pickle -import sys -from sys import argv -from cStringIO import StringIO - -modules = {} -functions = {} -functions_addr = {} - -def usage(): - print "-t,--trace: Location of the Trace file" - print "-s,--symbols: Location of the symbols and modules" - -def get_address_from_string(address): - return int(address.strip("S:").strip("N:").strip("EL2:").strip("EL1:"), 16) - -def get_module_from_addr(modules, addr): - for key,value in modules.items(): - if (value['start'] <= addr) and (addr <= value['end']): - return key - return None - -def add_cycles_to_function(functions, func_name, addr, cycles): - if func_name != "": - # Check if we are still in the previous function - if add_cycles_to_function.prev_func_name == func_name: - add_cycles_to_function.prev_entry['cycles'] += cycles - return (add_cycles_to_function.prev_func_name, add_cycles_to_function.prev_module_name) - - if func_name in functions.keys(): - for module_name, module_value in functions[func_name].iteritems(): - if (module_value['start'] <= addr) and (addr < module_value['end']): - module_value['cycles'] += cycles - - add_cycles_to_function.prev_func_name = func_name - add_cycles_to_function.prev_module_name = module_name - add_cycles_to_function.prev_entry = module_value - return (func_name, module_name) - elif (module_value['end'] == 0): - module_value['cycles'] += cycles - - add_cycles_to_function.prev_func_name = func_name - add_cycles_to_function.prev_module_name = module_name - add_cycles_to_function.prev_entry = module_value - return (func_name, module_name) - - # Workaround to fix the 'info func' limitation that does not expose the 'static' function - module_name = get_module_from_addr(modules, addr) - functions[func_name] = {} - functions[func_name][module_name] = {} - functions[func_name][module_name]['start'] = 0 - functions[func_name][module_name]['end'] = 0 - functions[func_name][module_name]['cycles'] = cycles - functions[func_name][module_name]['count'] = 0 - - add_cycles_to_function.prev_func_name = func_name - add_cycles_to_function.prev_module_name = module_name - add_cycles_to_function.prev_entry = functions[func_name][module_name] - return (func_name, module_name) - else: - # Check if we are still in the previous function - if (add_cycles_to_function.prev_entry is not None) and (add_cycles_to_function.prev_entry['start'] <= addr) and (addr < add_cycles_to_function.prev_entry['end']): - add_cycles_to_function.prev_entry['cycles'] += cycles - return (add_cycles_to_function.prev_func_name, add_cycles_to_function.prev_module_name) - - # Generate the key for the given address - key = addr & ~0x0FFF - - if key not in functions_addr.keys(): - if 'Unknown' not in functions.keys(): - functions['Unknown'] = {} - if 'Unknown' not in functions['Unknown'].keys(): - functions['Unknown']['Unknown'] = {} - functions['Unknown']['Unknown']['cycles'] = 0 - functions['Unknown']['Unknown']['count'] = 0 - functions['Unknown']['Unknown']['cycles'] += cycles - - add_cycles_to_function.prev_func_name = None - return None - - for func_key, module in functions_addr[key].iteritems(): - for module_key, module_value in module.iteritems(): - if (module_value['start'] <= addr) and (addr < module_value['end']): - module_value['cycles'] += cycles - - # In case o we prefer to fallback on the direct search - add_cycles_to_function.prev_func_name = func_key - add_cycles_to_function.prev_module_name = module_key - add_cycles_to_function.prev_entry = module_value - return (func_key, module_key) - - print "Warning: Function %s @ 0x%x not found" % (func_name, addr) - - add_cycles_to_function.prev_func_name = None - return None - -# Static variables for the previous function -add_cycles_to_function.prev_func_name = None -add_cycles_to_function.prev_entry = None - -def trace_read(): - global trace_process - line = trace.readline() - trace_process += len(line) - return line - -# -# Parse arguments -# -trace_name = None -symbols_file = None - -opts,args = getopt.getopt(sys.argv[1:], "ht:vs:v", ["help","trace=","symbols="]) -if (opts is None) or (not opts): - usage() - sys.exit() - -for o,a in opts: - if o in ("-h","--help"): - usage() - sys.exit() - elif o in ("-t","--trace"): - trace_name = a - elif o in ("-s","--symbols"): - symbols_file = a - else: - assert False, "Unhandled option (%s)" % o - -# -# We try first to see if we run the script from DS-5 -# -try: - from arm_ds.debugger_v1 import Debugger - from arm_ds.debugger_v1 import DebugException - - # Debugger object for accessing the debugger - debugger = Debugger() - - # Initialisation commands - ec = debugger.getExecutionContext(0) - ec.getExecutionService().stop() - ec.getExecutionService().waitForStop() - # in case the execution context reference is out of date - ec = debugger.getExecutionContext(0) - - # - # Get the module name and their memory range - # - info_file = ec.executeDSCommand("info file") - info_file_str = StringIO(info_file) - - line = info_file_str.readline().strip('\n') - while line != '': - if ("Symbols from" in line): - # Get the module name from the line 'Symbols from "/home/...."' - module_name = line.split("\"")[1].split("/")[-1] - modules[module_name] = {} - - # Look for the text section - line = info_file_str.readline().strip('\n') - while (line != '') and ("Symbols from" not in line): - if ("ER_RO" in line): - modules[module_name]['start'] = get_address_from_string(line.split()[0]) - modules[module_name]['end'] = get_address_from_string(line.split()[2]) - line = info_file_str.readline().strip('\n') - break; - if (".text" in line): - modules[module_name]['start'] = get_address_from_string(line.split()[0]) - modules[module_name]['end'] = get_address_from_string(line.split()[2]) - line = info_file_str.readline().strip('\n') - break; - line = info_file_str.readline().strip('\n') - line = info_file_str.readline().strip('\n') - - # - # Get the function name and their memory range - # - info_func = ec.executeDSCommand("info func") - info_func_str = StringIO(info_func) - - # Skip the first line 'Low-level symbols ...' - line = info_func_str.readline().strip('\n') - func_prev = None - while line != '': - # We ignore all the functions after 'Functions in' - if ("Functions in " in line): - line = info_func_str.readline().strip('\n') - while line != '': - line = info_func_str.readline().strip('\n') - line = info_func_str.readline().strip('\n') - continue - - if ("Low-level symbols" in line): - # We need to fixup the last function of the module - if func_prev is not None: - func_prev['end'] = modules[module_name]['end'] - func_prev = None - - line = info_func_str.readline().strip('\n') - continue - - func_name = line.split()[1] - func_start = get_address_from_string(line.split()[0]) - module_name = get_module_from_addr(modules, func_start) - - if func_name not in functions.keys(): - functions[func_name] = {} - functions[func_name][module_name] = {} - functions[func_name][module_name]['start'] = func_start - functions[func_name][module_name]['cycles'] = 0 - functions[func_name][module_name]['count'] = 0 - - # Set the end address of the previous function - if func_prev is not None: - func_prev['end'] = func_start - func_prev = functions[func_name][module_name] - - line = info_func_str.readline().strip('\n') - - # Fixup the last function - func_prev['end'] = modules[module_name]['end'] - - if symbols_file is not None: - pickle.dump((modules, functions), open(symbols_file, "w")) -except: - if symbols_file is None: - print "Error: Symbols file is required when run out of ARM DS-5" - sys.exit() - - (modules, functions) = pickle.load(open(symbols_file, "r")) - -# -# Build optimized table for the functions -# -functions_addr = {} -for func_key, module in functions.iteritems(): - for module_key, module_value in module.iteritems(): - key = module_value['start'] & ~0x0FFF - if key not in functions_addr.keys(): - functions_addr[key] = {} - if func_key not in functions_addr[key].keys(): - functions_addr[key][func_key] = {} - functions_addr[key][func_key][module_key] = module_value - -# -# Process the trace file -# -if trace_name is None: - sys.exit() - -trace = open(trace_name, "r") -trace_size = os.path.getsize(trace_name) -trace_process = 0 - -# Get the column names from the first line -columns = trace_read().split() -column_addr = columns.index('Address') -column_cycles = columns.index('Cycles') -column_function = columns.index('Function') - -line = trace_read() -i = 0 -prev_callee = None -while line: - try: - func_name = line.split('\t')[column_function].strip() - address = get_address_from_string(line.split('\t')[column_addr]) - cycles = int(line.split('\t')[column_cycles]) - callee = add_cycles_to_function(functions, func_name, address, cycles) - if (prev_callee != None) and (prev_callee != callee): - functions[prev_callee[0]][prev_callee[1]]['count'] += 1 - prev_callee = callee - except ValueError: - pass - line = trace_read() - if ((i % 1000000) == 0) and (i != 0): - percent = (trace_process * 100.00) / trace_size - print "Processing file ... (%.2f %%)" % (percent) - i = i + 1 - -# Fixup the last callee -functions[prev_callee[0]][prev_callee[1]]['count'] += 1 - -# -# Process results -# -functions_cycles = {} -all_functions_cycles = {} -total_cycles = 0 - -for func_key, module in functions.iteritems(): - for module_key, module_value in module.iteritems(): - key = "%s/%s" % (module_key, func_key) - functions_cycles[key] = (module_value['cycles'], module_value['count']) - total_cycles += module_value['cycles'] - - if func_key not in all_functions_cycles.keys(): - all_functions_cycles[func_key] = (module_value['cycles'], module_value['count']) - else: - all_functions_cycles[func_key] = tuple(map(sum, zip(all_functions_cycles[func_key], (module_value['cycles'], module_value['count'])))) - -sorted_functions_cycles = sorted(functions_cycles.iteritems(), key=operator.itemgetter(1), reverse = True) -sorted_all_functions_cycles = sorted(all_functions_cycles.items(), key=operator.itemgetter(1), reverse = True) - -print -print "----" -for (key,value) in sorted_functions_cycles[:20]: - if value[0] != 0: - print "%s (cycles: %d - %d%%, count: %d)" % (key, value[0], (value[0] * 100) / total_cycles, value[1]) - else: - break; -print "----" -for (key,value) in sorted_all_functions_cycles[:20]: - if value[0] != 0: - print "%s (cycles: %d - %d%%, count: %d)" % (key, value[0], (value[0] * 100) / total_cycles, value[1]) - else: - break; +#!/usr/bin/python + +# +# Copyright (c) 2014, ARM Limited. All rights reserved. +# +# SPDX-License-Identifier: BSD-2-Clause-Patent +# + +import getopt +import operator +import os +import pickle +import sys +from sys import argv +from cStringIO import StringIO + +modules = {} +functions = {} +functions_addr = {} + +def usage(): + print "-t,--trace: Location of the Trace file" + print "-s,--symbols: Location of the symbols and modules" + +def get_address_from_string(address): + return int(address.strip("S:").strip("N:").strip("EL2:").strip("EL1:"), 16) + +def get_module_from_addr(modules, addr): + for key,value in modules.items(): + if (value['start'] <= addr) and (addr <= value['end']): + return key + return None + +def add_cycles_to_function(functions, func_name, addr, cycles): + if func_name != "": + # Check if we are still in the previous function + if add_cycles_to_function.prev_func_name == func_name: + add_cycles_to_function.prev_entry['cycles'] += cycles + return (add_cycles_to_function.prev_func_name, add_cycles_to_function.prev_module_name) + + if func_name in functions.keys(): + for module_name, module_value in functions[func_name].iteritems(): + if (module_value['start'] <= addr) and (addr < module_value['end']): + module_value['cycles'] += cycles + + add_cycles_to_function.prev_func_name = func_name + add_cycles_to_function.prev_module_name = module_name + add_cycles_to_function.prev_entry = module_value + return (func_name, module_name) + elif (module_value['end'] == 0): + module_value['cycles'] += cycles + + add_cycles_to_function.prev_func_name = func_name + add_cycles_to_function.prev_module_name = module_name + add_cycles_to_function.prev_entry = module_value + return (func_name, module_name) + + # Workaround to fix the 'info func' limitation that does not expose the 'static' function + module_name = get_module_from_addr(modules, addr) + functions[func_name] = {} + functions[func_name][module_name] = {} + functions[func_name][module_name]['start'] = 0 + functions[func_name][module_name]['end'] = 0 + functions[func_name][module_name]['cycles'] = cycles + functions[func_name][module_name]['count'] = 0 + + add_cycles_to_function.prev_func_name = func_name + add_cycles_to_function.prev_module_name = module_name + add_cycles_to_function.prev_entry = functions[func_name][module_name] + return (func_name, module_name) + else: + # Check if we are still in the previous function + if (add_cycles_to_function.prev_entry is not None) and (add_cycles_to_function.prev_entry['start'] <= addr) and (addr < add_cycles_to_function.prev_entry['end']): + add_cycles_to_function.prev_entry['cycles'] += cycles + return (add_cycles_to_function.prev_func_name, add_cycles_to_function.prev_module_name) + + # Generate the key for the given address + key = addr & ~0x0FFF + + if key not in functions_addr.keys(): + if 'Unknown' not in functions.keys(): + functions['Unknown'] = {} + if 'Unknown' not in functions['Unknown'].keys(): + functions['Unknown']['Unknown'] = {} + functions['Unknown']['Unknown']['cycles'] = 0 + functions['Unknown']['Unknown']['count'] = 0 + functions['Unknown']['Unknown']['cycles'] += cycles + + add_cycles_to_function.prev_func_name = None + return None + + for func_key, module in functions_addr[key].iteritems(): + for module_key, module_value in module.iteritems(): + if (module_value['start'] <= addr) and (addr < module_value['end']): + module_value['cycles'] += cycles + + # In case o we prefer to fallback on the direct search + add_cycles_to_function.prev_func_name = func_key + add_cycles_to_function.prev_module_name = module_key + add_cycles_to_function.prev_entry = module_value + return (func_key, module_key) + + print "Warning: Function %s @ 0x%x not found" % (func_name, addr) + + add_cycles_to_function.prev_func_name = None + return None + +# Static variables for the previous function +add_cycles_to_function.prev_func_name = None +add_cycles_to_function.prev_entry = None + +def trace_read(): + global trace_process + line = trace.readline() + trace_process += len(line) + return line + +# +# Parse arguments +# +trace_name = None +symbols_file = None + +opts,args = getopt.getopt(sys.argv[1:], "ht:vs:v", ["help","trace=","symbols="]) +if (opts is None) or (not opts): + usage() + sys.exit() + +for o,a in opts: + if o in ("-h","--help"): + usage() + sys.exit() + elif o in ("-t","--trace"): + trace_name = a + elif o in ("-s","--symbols"): + symbols_file = a + else: + assert False, "Unhandled option (%s)" % o + +# +# We try first to see if we run the script from DS-5 +# +try: + from arm_ds.debugger_v1 import Debugger + from arm_ds.debugger_v1 import DebugException + + # Debugger object for accessing the debugger + debugger = Debugger() + + # Initialisation commands + ec = debugger.getExecutionContext(0) + ec.getExecutionService().stop() + ec.getExecutionService().waitForStop() + # in case the execution context reference is out of date + ec = debugger.getExecutionContext(0) + + # + # Get the module name and their memory range + # + info_file = ec.executeDSCommand("info file") + info_file_str = StringIO(info_file) + + line = info_file_str.readline().strip('\n') + while line != '': + if ("Symbols from" in line): + # Get the module name from the line 'Symbols from "/home/...."' + module_name = line.split("\"")[1].split("/")[-1] + modules[module_name] = {} + + # Look for the text section + line = info_file_str.readline().strip('\n') + while (line != '') and ("Symbols from" not in line): + if ("ER_RO" in line): + modules[module_name]['start'] = get_address_from_string(line.split()[0]) + modules[module_name]['end'] = get_address_from_string(line.split()[2]) + line = info_file_str.readline().strip('\n') + break; + if (".text" in line): + modules[module_name]['start'] = get_address_from_string(line.split()[0]) + modules[module_name]['end'] = get_address_from_string(line.split()[2]) + line = info_file_str.readline().strip('\n') + break; + line = info_file_str.readline().strip('\n') + line = info_file_str.readline().strip('\n') + + # + # Get the function name and their memory range + # + info_func = ec.executeDSCommand("info func") + info_func_str = StringIO(info_func) + + # Skip the first line 'Low-level symbols ...' + line = info_func_str.readline().strip('\n') + func_prev = None + while line != '': + # We ignore all the functions after 'Functions in' + if ("Functions in " in line): + line = info_func_str.readline().strip('\n') + while line != '': + line = info_func_str.readline().strip('\n') + line = info_func_str.readline().strip('\n') + continue + + if ("Low-level symbols" in line): + # We need to fixup the last function of the module + if func_prev is not None: + func_prev['end'] = modules[module_name]['end'] + func_prev = None + + line = info_func_str.readline().strip('\n') + continue + + func_name = line.split()[1] + func_start = get_address_from_string(line.split()[0]) + module_name = get_module_from_addr(modules, func_start) + + if func_name not in functions.keys(): + functions[func_name] = {} + functions[func_name][module_name] = {} + functions[func_name][module_name]['start'] = func_start + functions[func_name][module_name]['cycles'] = 0 + functions[func_name][module_name]['count'] = 0 + + # Set the end address of the previous function + if func_prev is not None: + func_prev['end'] = func_start + func_prev = functions[func_name][module_name] + + line = info_func_str.readline().strip('\n') + + # Fixup the last function + func_prev['end'] = modules[module_name]['end'] + + if symbols_file is not None: + pickle.dump((modules, functions), open(symbols_file, "w")) +except: + if symbols_file is None: + print "Error: Symbols file is required when run out of ARM DS-5" + sys.exit() + + (modules, functions) = pickle.load(open(symbols_file, "r")) + +# +# Build optimized table for the functions +# +functions_addr = {} +for func_key, module in functions.iteritems(): + for module_key, module_value in module.iteritems(): + key = module_value['start'] & ~0x0FFF + if key not in functions_addr.keys(): + functions_addr[key] = {} + if func_key not in functions_addr[key].keys(): + functions_addr[key][func_key] = {} + functions_addr[key][func_key][module_key] = module_value + +# +# Process the trace file +# +if trace_name is None: + sys.exit() + +trace = open(trace_name, "r") +trace_size = os.path.getsize(trace_name) +trace_process = 0 + +# Get the column names from the first line +columns = trace_read().split() +column_addr = columns.index('Address') +column_cycles = columns.index('Cycles') +column_function = columns.index('Function') + +line = trace_read() +i = 0 +prev_callee = None +while line: + try: + func_name = line.split('\t')[column_function].strip() + address = get_address_from_string(line.split('\t')[column_addr]) + cycles = int(line.split('\t')[column_cycles]) + callee = add_cycles_to_function(functions, func_name, address, cycles) + if (prev_callee != None) and (prev_callee != callee): + functions[prev_callee[0]][prev_callee[1]]['count'] += 1 + prev_callee = callee + except ValueError: + pass + line = trace_read() + if ((i % 1000000) == 0) and (i != 0): + percent = (trace_process * 100.00) / trace_size + print "Processing file ... (%.2f %%)" % (percent) + i = i + 1 + +# Fixup the last callee +functions[prev_callee[0]][prev_callee[1]]['count'] += 1 + +# +# Process results +# +functions_cycles = {} +all_functions_cycles = {} +total_cycles = 0 + +for func_key, module in functions.iteritems(): + for module_key, module_value in module.iteritems(): + key = "%s/%s" % (module_key, func_key) + functions_cycles[key] = (module_value['cycles'], module_value['count']) + total_cycles += module_value['cycles'] + + if func_key not in all_functions_cycles.keys(): + all_functions_cycles[func_key] = (module_value['cycles'], module_value['count']) + else: + all_functions_cycles[func_key] = tuple(map(sum, zip(all_functions_cycles[func_key], (module_value['cycles'], module_value['count'])))) + +sorted_functions_cycles = sorted(functions_cycles.iteritems(), key=operator.itemgetter(1), reverse = True) +sorted_all_functions_cycles = sorted(all_functions_cycles.items(), key=operator.itemgetter(1), reverse = True) + +print +print "----" +for (key,value) in sorted_functions_cycles[:20]: + if value[0] != 0: + print "%s (cycles: %d - %d%%, count: %d)" % (key, value[0], (value[0] * 100) / total_cycles, value[1]) + else: + break; +print "----" +for (key,value) in sorted_all_functions_cycles[:20]: + if value[0] != 0: + print "%s (cycles: %d - %d%%, count: %d)" % (key, value[0], (value[0] * 100) / total_cycles, value[1]) + else: + break; -- 2.39.2