diff options
Diffstat (limited to 'tools/power')
56 files changed, 6957 insertions, 1063 deletions
diff --git a/tools/power/acpi/Makefile b/tools/power/acpi/Makefile new file mode 100644 index 00000000000..e5a3c4be2a1 --- /dev/null +++ b/tools/power/acpi/Makefile @@ -0,0 +1,156 @@ +# tools/power/acpi/Makefile - ACPI tool Makefile +# +# Copyright (c) 2013, Intel Corporation +# Author: Lv Zheng <lv.zheng@intel.com> +# +# This program 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; version 2 +# of the License. + +OUTPUT=./ +ifeq ("$(origin O)", "command line") + OUTPUT := $(O)/ +endif + +ifneq ($(OUTPUT),) +# check that the output directory actually exists +OUTDIR := $(shell cd $(OUTPUT) && /bin/pwd) +$(if $(OUTDIR),, $(error output directory "$(OUTPUT)" does not exist)) +endif + +SUBDIRS = tools/ec + +# --- CONFIGURATION BEGIN --- + +# Set the following to `true' to make a unstripped, unoptimized +# binary. Leave this set to `false' for production use. +DEBUG ?= true + +# make the build silent. Set this to something else to make it noisy again. +V ?= false + +# Prefix to the directories we're installing to +DESTDIR ?= + +# --- CONFIGURATION END --- + +# Directory definitions. These are default and most probably +# do not need to be changed. Please note that DESTDIR is +# added in front of any of them + +bindir ?= /usr/bin +sbindir ?= /usr/sbin +mandir ?= /usr/man + +# Toolchain: what tools do we use, and what options do they need: + +INSTALL = /usr/bin/install -c +INSTALL_PROGRAM = ${INSTALL} +INSTALL_DATA = ${INSTALL} -m 644 +INSTALL_SCRIPT = ${INSTALL_PROGRAM} + +# If you are running a cross compiler, you may want to set this +# to something more interesting, like "arm-linux-". If you want +# to compile vs uClibc, that can be done here as well. +CROSS = #/usr/i386-linux-uclibc/usr/bin/i386-uclibc- +CC = $(CROSS)gcc +LD = $(CROSS)gcc +STRIP = $(CROSS)strip +HOSTCC = gcc + +# check if compiler option is supported +cc-supports = ${shell if $(CC) ${1} -S -o /dev/null -x c /dev/null > /dev/null 2>&1; then echo "$(1)"; fi;} + +# use '-Os' optimization if available, else use -O2 +OPTIMIZATION := $(call cc-supports,-Os,-O2) + +WARNINGS := -Wall +WARNINGS += $(call cc-supports,-Wstrict-prototypes) +WARNINGS += $(call cc-supports,-Wdeclaration-after-statement) + +KERNEL_INCLUDE := ../../../include +ACPICA_INCLUDE := ../../../drivers/acpi/acpica +CFLAGS += -D_LINUX -I$(KERNEL_INCLUDE) -I$(ACPICA_INCLUDE) +CFLAGS += $(WARNINGS) + +ifeq ($(strip $(V)),false) + QUIET=@ + ECHO=@echo +else + QUIET= + ECHO=@\# +endif +export QUIET ECHO + +# if DEBUG is enabled, then we do not strip or optimize +ifeq ($(strip $(DEBUG)),true) + CFLAGS += -O1 -g -DDEBUG + STRIPCMD = /bin/true -Since_we_are_debugging +else + CFLAGS += $(OPTIMIZATION) -fomit-frame-pointer + STRIPCMD = $(STRIP) -s --remove-section=.note --remove-section=.comment +endif + +# --- ACPIDUMP BEGIN --- + +vpath %.c \ + ../../../drivers/acpi/acpica\ + tools/acpidump\ + common\ + os_specific/service_layers + +CFLAGS += -DACPI_DUMP_APP -Itools/acpidump + +DUMP_OBJS = \ + apdump.o\ + apfiles.o\ + apmain.o\ + osunixdir.o\ + osunixmap.o\ + tbprint.o\ + tbxfroot.o\ + utbuffer.o\ + utexcep.o\ + utmath.o\ + utstring.o\ + utxferror.o\ + oslinuxtbl.o\ + cmfsize.o\ + getopt.o + +DUMP_OBJS := $(addprefix $(OUTPUT)tools/acpidump/,$(DUMP_OBJS)) + +$(OUTPUT)acpidump: $(DUMP_OBJS) + $(ECHO) " LD " $@ + $(QUIET) $(LD) $(CFLAGS) $(LDFLAGS) $(DUMP_OBJS) -L$(OUTPUT) -o $@ + $(QUIET) $(STRIPCMD) $@ + +$(OUTPUT)tools/acpidump/%.o: %.c + $(ECHO) " CC " $@ + $(QUIET) $(CC) -c $(CFLAGS) -o $@ $< + +# --- ACPIDUMP END --- + +all: $(OUTPUT)acpidump + echo $(OUTPUT) + +clean: + -find $(OUTPUT) \( -not -type d \) -and \( -name '*~' -o -name '*.[oas]' \) -type f -print \ + | xargs rm -f + -rm -f $(OUTPUT)acpidump + +install-tools: + $(INSTALL) -d $(DESTDIR)${sbindir} + $(INSTALL_PROGRAM) $(OUTPUT)acpidump $(DESTDIR)${sbindir} + +install-man: + $(INSTALL_DATA) -D man/acpidump.8 $(DESTDIR)${mandir}/man8/acpidump.8 + +install: all install-tools install-man + +uninstall: + - rm -f $(DESTDIR)${sbindir}/acpidump + - rm -f $(DESTDIR)${mandir}/man8/acpidump.8 + +.PHONY: all utils install-tools install-man install uninstall clean diff --git a/tools/power/acpi/common/cmfsize.c b/tools/power/acpi/common/cmfsize.c new file mode 100644 index 00000000000..5140e5edae1 --- /dev/null +++ b/tools/power/acpi/common/cmfsize.c @@ -0,0 +1,101 @@ +/****************************************************************************** + * + * Module Name: cfsize - Common get file size function + * + *****************************************************************************/ + +/* + * Copyright (C) 2000 - 2014, Intel Corp. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions, and the following disclaimer, + * without modification. + * 2. Redistributions in binary form must reproduce at minimum a disclaimer + * substantially similar to the "NO WARRANTY" disclaimer below + * ("Disclaimer") and any redistribution must be conditioned upon + * including a substantially similar Disclaimer requirement for further + * binary redistribution. + * 3. Neither the names of the above-listed copyright holders nor the names + * of any contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * Alternatively, this software may be distributed under the terms of the + * GNU General Public License ("GPL") version 2 as published by the Free + * Software Foundation. + * + * NO WARRANTY + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGES. + */ + +#include <acpi/acpi.h> +#include "accommon.h" +#include "acapps.h" +#include <stdio.h> + +#define _COMPONENT ACPI_TOOLS +ACPI_MODULE_NAME("cmfsize") + +/******************************************************************************* + * + * FUNCTION: cm_get_file_size + * + * PARAMETERS: file - Open file descriptor + * + * RETURN: File Size. On error, -1 (ACPI_UINT32_MAX) + * + * DESCRIPTION: Get the size of a file. Uses seek-to-EOF. File must be open. + * Does not disturb the current file pointer. Uses perror for + * error messages. + * + ******************************************************************************/ +u32 cm_get_file_size(FILE * file) +{ + long file_size; + long current_offset; + + /* Save the current file pointer, seek to EOF to obtain file size */ + + current_offset = ftell(file); + if (current_offset < 0) { + goto offset_error; + } + + if (fseek(file, 0, SEEK_END)) { + goto seek_error; + } + + file_size = ftell(file); + if (file_size < 0) { + goto offset_error; + } + + /* Restore original file pointer */ + + if (fseek(file, current_offset, SEEK_SET)) { + goto seek_error; + } + + return ((u32)file_size); + +offset_error: + perror("Could not get file offset"); + return (ACPI_UINT32_MAX); + +seek_error: + perror("Could not seek file"); + return (ACPI_UINT32_MAX); +} diff --git a/tools/power/acpi/common/getopt.c b/tools/power/acpi/common/getopt.c new file mode 100644 index 00000000000..a302f52e4fd --- /dev/null +++ b/tools/power/acpi/common/getopt.c @@ -0,0 +1,239 @@ +/****************************************************************************** + * + * Module Name: getopt + * + *****************************************************************************/ + +/* + * Copyright (C) 2000 - 2014, Intel Corp. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions, and the following disclaimer, + * without modification. + * 2. Redistributions in binary form must reproduce at minimum a disclaimer + * substantially similar to the "NO WARRANTY" disclaimer below + * ("Disclaimer") and any redistribution must be conditioned upon + * including a substantially similar Disclaimer requirement for further + * binary redistribution. + * 3. Neither the names of the above-listed copyright holders nor the names + * of any contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * Alternatively, this software may be distributed under the terms of the + * GNU General Public License ("GPL") version 2 as published by the Free + * Software Foundation. + * + * NO WARRANTY + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGES. + */ + +/* + * ACPICA getopt() implementation + * + * Option strings: + * "f" - Option has no arguments + * "f:" - Option requires an argument + * "f^" - Option has optional single-char sub-options + * "f|" - Option has required single-char sub-options + */ + +#include <stdio.h> +#include <string.h> +#include <acpi/acpi.h> +#include "accommon.h" +#include "acapps.h" + +#define ACPI_OPTION_ERROR(msg, badchar) \ + if (acpi_gbl_opterr) {fprintf (stderr, "%s%c\n", msg, badchar);} + +int acpi_gbl_opterr = 1; +int acpi_gbl_optind = 1; +int acpi_gbl_sub_opt_char = 0; +char *acpi_gbl_optarg; + +static int current_char_ptr = 1; + +/******************************************************************************* + * + * FUNCTION: acpi_getopt_argument + * + * PARAMETERS: argc, argv - from main + * + * RETURN: 0 if an argument was found, -1 otherwise. Sets acpi_gbl_Optarg + * to point to the next argument. + * + * DESCRIPTION: Get the next argument. Used to obtain arguments for the + * two-character options after the original call to acpi_getopt. + * Note: Either the argument starts at the next character after + * the option, or it is pointed to by the next argv entry. + * (After call to acpi_getopt, we need to backup to the previous + * argv entry). + * + ******************************************************************************/ + +int acpi_getopt_argument(int argc, char **argv) +{ + acpi_gbl_optind--; + current_char_ptr++; + + if (argv[acpi_gbl_optind][(int)(current_char_ptr + 1)] != '\0') { + acpi_gbl_optarg = + &argv[acpi_gbl_optind++][(int)(current_char_ptr + 1)]; + } else if (++acpi_gbl_optind >= argc) { + ACPI_OPTION_ERROR("Option requires an argument: -", 'v'); + + current_char_ptr = 1; + return (-1); + } else { + acpi_gbl_optarg = argv[acpi_gbl_optind++]; + } + + current_char_ptr = 1; + return (0); +} + +/******************************************************************************* + * + * FUNCTION: acpi_getopt + * + * PARAMETERS: argc, argv - from main + * opts - options info list + * + * RETURN: Option character or EOF + * + * DESCRIPTION: Get the next option + * + ******************************************************************************/ + +int acpi_getopt(int argc, char **argv, char *opts) +{ + int current_char; + char *opts_ptr; + + if (current_char_ptr == 1) { + if (acpi_gbl_optind >= argc || + argv[acpi_gbl_optind][0] != '-' || + argv[acpi_gbl_optind][1] == '\0') { + return (EOF); + } else if (strcmp(argv[acpi_gbl_optind], "--") == 0) { + acpi_gbl_optind++; + return (EOF); + } + } + + /* Get the option */ + + current_char = argv[acpi_gbl_optind][current_char_ptr]; + + /* Make sure that the option is legal */ + + if (current_char == ':' || + (opts_ptr = strchr(opts, current_char)) == NULL) { + ACPI_OPTION_ERROR("Illegal option: -", current_char); + + if (argv[acpi_gbl_optind][++current_char_ptr] == '\0') { + acpi_gbl_optind++; + current_char_ptr = 1; + } + + return ('?'); + } + + /* Option requires an argument? */ + + if (*++opts_ptr == ':') { + if (argv[acpi_gbl_optind][(int)(current_char_ptr + 1)] != '\0') { + acpi_gbl_optarg = + &argv[acpi_gbl_optind++][(int) + (current_char_ptr + 1)]; + } else if (++acpi_gbl_optind >= argc) { + ACPI_OPTION_ERROR("Option requires an argument: -", + current_char); + + current_char_ptr = 1; + return ('?'); + } else { + acpi_gbl_optarg = argv[acpi_gbl_optind++]; + } + + current_char_ptr = 1; + } + + /* Option has an optional argument? */ + + else if (*opts_ptr == '+') { + if (argv[acpi_gbl_optind][(int)(current_char_ptr + 1)] != '\0') { + acpi_gbl_optarg = + &argv[acpi_gbl_optind++][(int) + (current_char_ptr + 1)]; + } else if (++acpi_gbl_optind >= argc) { + acpi_gbl_optarg = NULL; + } else { + acpi_gbl_optarg = argv[acpi_gbl_optind++]; + } + + current_char_ptr = 1; + } + + /* Option has optional single-char arguments? */ + + else if (*opts_ptr == '^') { + if (argv[acpi_gbl_optind][(int)(current_char_ptr + 1)] != '\0') { + acpi_gbl_optarg = + &argv[acpi_gbl_optind][(int)(current_char_ptr + 1)]; + } else { + acpi_gbl_optarg = "^"; + } + + acpi_gbl_sub_opt_char = acpi_gbl_optarg[0]; + acpi_gbl_optind++; + current_char_ptr = 1; + } + + /* Option has a required single-char argument? */ + + else if (*opts_ptr == '|') { + if (argv[acpi_gbl_optind][(int)(current_char_ptr + 1)] != '\0') { + acpi_gbl_optarg = + &argv[acpi_gbl_optind][(int)(current_char_ptr + 1)]; + } else { + ACPI_OPTION_ERROR + ("Option requires a single-character suboption: -", + current_char); + + current_char_ptr = 1; + return ('?'); + } + + acpi_gbl_sub_opt_char = acpi_gbl_optarg[0]; + acpi_gbl_optind++; + current_char_ptr = 1; + } + + /* Option with no arguments */ + + else { + if (argv[acpi_gbl_optind][++current_char_ptr] == '\0') { + current_char_ptr = 1; + acpi_gbl_optind++; + } + + acpi_gbl_optarg = NULL; + } + + return (current_char); +} diff --git a/tools/power/acpi/man/acpidump.8 b/tools/power/acpi/man/acpidump.8 new file mode 100644 index 00000000000..38f095d86b5 --- /dev/null +++ b/tools/power/acpi/man/acpidump.8 @@ -0,0 +1,120 @@ +.TH ACPIDUMP 8 +.SH NAME +acpidump \- dump a system's ACPI tables to an ASCII file + +.SH SYNOPSIS +.B acpidump +.RI [ options ] +.br + +.SH DESCRIPTION +.B acpidump +dumps the systems ACPI tables to an ASCII file appropriate for +attaching to a bug report. + +Subsequently, they can be processed by utilities in the ACPICA package. + +.SH OPTIONS +acpidump options are as follow: +.TP +.B Options +.TP +.B \-b +Dump tables to binary files +.TP +.B \-c +Dump customized tables +.TP +.B \-h \-? +This help message +.TP +.B \-o <File> +Redirect output to file +.TP +.B \-r <Address> +Dump tables from specified RSDP +.TP +.B \-s +Print table summaries only +.TP +.B \-v +Display version information +.TP +.B \-z +Verbose mode +.TP +.B Table Options +.TP +.B \-a <Address> +Get table via a physical address +.TP +.B \-f <BinaryFile> +Get table via a binary file +.TP +.B \-n <Signature> +Get table via a name/signature +.TP +Invocation without parameters dumps all available tables +.TP +Multiple mixed instances of -a, -f, and -n are supported + +.SH EXAMPLES + +.nf +# acpidump > acpidump.out + +$ acpixtract -a acpidump.out + Acpi table [DSDT] - 15974 bytes written to DSDT.dat + Acpi table [FACS] - 64 bytes written to FACS.dat + Acpi table [FACP] - 116 bytes written to FACP.dat + Acpi table [APIC] - 120 bytes written to APIC.dat + Acpi table [MCFG] - 60 bytes written to MCFG.dat + Acpi table [SSDT] - 444 bytes written to SSDT1.dat + Acpi table [SSDT] - 439 bytes written to SSDT2.dat + Acpi table [SSDT] - 439 bytes written to SSDT3.dat + Acpi table [SSDT] - 439 bytes written to SSDT4.dat + Acpi table [SSDT] - 439 bytes written to SSDT5.dat + Acpi table [RSDT] - 76 bytes written to RSDT.dat + Acpi table [RSDP] - 20 bytes written to RSDP.dat + +$ iasl -d *.dat +... +.fi +creates *.dsl, a human readable form which can be edited +and compiled using iasl. + + +.SH NOTES + +.B "acpidump " +must be run as root. + +.SH REFERENCES +ACPICA: https://acpica.org/ + +.SH FILES +.ta +.nf +/dev/mem +/sys/firmware/acpi/tables/* +/sys/firmware/acpi/tables/dynamic/* +/sys/firmware/efi/systab +.fi + +.SH AUTHOR +.TP +Original by: + Len Brown <len.brown@intel.com> +.TP +Written by: + Chao Guan <chao.guan@intel.com> +.TP +Updated by: + Bob Moore <robert.moore@intel.com> + Lv Zheng <lv.zheng@intel.com> + +.SH SEE ALSO +\&\fIacpixtract\fR\|(8), \fIiasl\fR\|(8). + +.SH COPYRIGHT +COPYRIGHT (c) 2013, Intel Corporation. diff --git a/tools/power/acpi/os_specific/service_layers/oslinuxtbl.c b/tools/power/acpi/os_specific/service_layers/oslinuxtbl.c new file mode 100644 index 00000000000..28c52008e85 --- /dev/null +++ b/tools/power/acpi/os_specific/service_layers/oslinuxtbl.c @@ -0,0 +1,1329 @@ +/****************************************************************************** + * + * Module Name: oslinuxtbl - Linux OSL for obtaining ACPI tables + * + *****************************************************************************/ + +/* + * Copyright (C) 2000 - 2014, Intel Corp. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions, and the following disclaimer, + * without modification. + * 2. Redistributions in binary form must reproduce at minimum a disclaimer + * substantially similar to the "NO WARRANTY" disclaimer below + * ("Disclaimer") and any redistribution must be conditioned upon + * including a substantially similar Disclaimer requirement for further + * binary redistribution. + * 3. Neither the names of the above-listed copyright holders nor the names + * of any contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * Alternatively, this software may be distributed under the terms of the + * GNU General Public License ("GPL") version 2 as published by the Free + * Software Foundation. + * + * NO WARRANTY + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGES. + */ + +#include "acpidump.h" + +#define _COMPONENT ACPI_OS_SERVICES +ACPI_MODULE_NAME("oslinuxtbl") + +#ifndef PATH_MAX +#define PATH_MAX 256 +#endif +/* List of information about obtained ACPI tables */ +typedef struct osl_table_info { + struct osl_table_info *next; + u32 instance; + char signature[ACPI_NAME_SIZE]; + +} osl_table_info; + +/* Local prototypes */ + +static acpi_status osl_table_initialize(void); + +static acpi_status +osl_table_name_from_file(char *filename, char *signature, u32 *instance); + +static acpi_status osl_add_table_to_list(char *signature, u32 instance); + +static acpi_status +osl_read_table_from_file(char *filename, + acpi_size file_offset, + char *signature, struct acpi_table_header **table); + +static acpi_status +osl_map_table(acpi_size address, + char *signature, struct acpi_table_header **table); + +static void osl_unmap_table(struct acpi_table_header *table); + +static acpi_physical_address osl_find_rsdp_via_efi(void); + +static acpi_status osl_load_rsdp(void); + +static acpi_status osl_list_customized_tables(char *directory); + +static acpi_status +osl_get_customized_table(char *pathname, + char *signature, + u32 instance, + struct acpi_table_header **table, + acpi_physical_address * address); + +static acpi_status osl_list_bios_tables(void); + +static acpi_status +osl_get_bios_table(char *signature, + u32 instance, + struct acpi_table_header **table, + acpi_physical_address * address); + +static acpi_status osl_get_last_status(acpi_status default_status); + +/* File locations */ + +#define DYNAMIC_TABLE_DIR "/sys/firmware/acpi/tables/dynamic" +#define STATIC_TABLE_DIR "/sys/firmware/acpi/tables" +#define EFI_SYSTAB "/sys/firmware/efi/systab" + +/* Should we get dynamically loaded SSDTs from DYNAMIC_TABLE_DIR? */ + +u8 gbl_dump_dynamic_tables = TRUE; + +/* Initialization flags */ + +u8 gbl_table_list_initialized = FALSE; + +/* Local copies of main ACPI tables */ + +struct acpi_table_rsdp gbl_rsdp; +struct acpi_table_fadt *gbl_fadt = NULL; +struct acpi_table_rsdt *gbl_rsdt = NULL; +struct acpi_table_xsdt *gbl_xsdt = NULL; + +/* Table addresses */ + +acpi_physical_address gbl_fadt_address = 0; +acpi_physical_address gbl_rsdp_address = 0; + +/* Revision of RSD PTR */ + +u8 gbl_revision = 0; + +struct osl_table_info *gbl_table_list_head = NULL; +u32 gbl_table_count = 0; + +/****************************************************************************** + * + * FUNCTION: osl_get_last_status + * + * PARAMETERS: default_status - Default error status to return + * + * RETURN: Status; Converted from errno. + * + * DESCRIPTION: Get last errno and conver it to acpi_status. + * + *****************************************************************************/ + +static acpi_status osl_get_last_status(acpi_status default_status) +{ + + switch (errno) { + case EACCES: + case EPERM: + + return (AE_ACCESS); + + case ENOENT: + + return (AE_NOT_FOUND); + + case ENOMEM: + + return (AE_NO_MEMORY); + + default: + + return (default_status); + } +} + +/****************************************************************************** + * + * FUNCTION: acpi_os_get_table_by_address + * + * PARAMETERS: address - Physical address of the ACPI table + * table - Where a pointer to the table is returned + * + * RETURN: Status; Table buffer is returned if AE_OK. + * AE_NOT_FOUND: A valid table was not found at the address + * + * DESCRIPTION: Get an ACPI table via a physical memory address. + * + *****************************************************************************/ + +acpi_status +acpi_os_get_table_by_address(acpi_physical_address address, + struct acpi_table_header ** table) +{ + u32 table_length; + struct acpi_table_header *mapped_table; + struct acpi_table_header *local_table = NULL; + acpi_status status = AE_OK; + + /* Get main ACPI tables from memory on first invocation of this function */ + + status = osl_table_initialize(); + if (ACPI_FAILURE(status)) { + return (status); + } + + /* Map the table and validate it */ + + status = osl_map_table(address, NULL, &mapped_table); + if (ACPI_FAILURE(status)) { + return (status); + } + + /* Copy table to local buffer and return it */ + + table_length = ap_get_table_length(mapped_table); + if (table_length == 0) { + status = AE_BAD_HEADER; + goto exit; + } + + local_table = calloc(1, table_length); + if (!local_table) { + status = AE_NO_MEMORY; + goto exit; + } + + ACPI_MEMCPY(local_table, mapped_table, table_length); + +exit: + osl_unmap_table(mapped_table); + *table = local_table; + return (status); +} + +/****************************************************************************** + * + * FUNCTION: acpi_os_get_table_by_name + * + * PARAMETERS: signature - ACPI Signature for desired table. Must be + * a null terminated 4-character string. + * instance - Multiple table support for SSDT/UEFI (0...n) + * Must be 0 for other tables. + * table - Where a pointer to the table is returned + * address - Where the table physical address is returned + * + * RETURN: Status; Table buffer and physical address returned if AE_OK. + * AE_LIMIT: Instance is beyond valid limit + * AE_NOT_FOUND: A table with the signature was not found + * + * NOTE: Assumes the input signature is uppercase. + * + *****************************************************************************/ + +acpi_status +acpi_os_get_table_by_name(char *signature, + u32 instance, + struct acpi_table_header ** table, + acpi_physical_address * address) +{ + acpi_status status; + + /* Get main ACPI tables from memory on first invocation of this function */ + + status = osl_table_initialize(); + if (ACPI_FAILURE(status)) { + return (status); + } + + /* Not a main ACPI table, attempt to extract it from the RSDT/XSDT */ + + if (!gbl_dump_customized_tables) { + + /* Attempt to get the table from the memory */ + + status = + osl_get_bios_table(signature, instance, table, address); + } else { + /* Attempt to get the table from the static directory */ + + status = osl_get_customized_table(STATIC_TABLE_DIR, signature, + instance, table, address); + } + + if (ACPI_FAILURE(status) && status == AE_LIMIT) { + if (gbl_dump_dynamic_tables) { + + /* Attempt to get a dynamic table */ + + status = + osl_get_customized_table(DYNAMIC_TABLE_DIR, + signature, instance, table, + address); + } + } + + return (status); +} + +/****************************************************************************** + * + * FUNCTION: osl_add_table_to_list + * + * PARAMETERS: signature - Table signature + * instance - Table instance + * + * RETURN: Status; Successfully added if AE_OK. + * AE_NO_MEMORY: Memory allocation error + * + * DESCRIPTION: Insert a table structure into OSL table list. + * + *****************************************************************************/ + +static acpi_status osl_add_table_to_list(char *signature, u32 instance) +{ + struct osl_table_info *new_info; + struct osl_table_info *next; + u32 next_instance = 0; + u8 found = FALSE; + + new_info = calloc(1, sizeof(struct osl_table_info)); + if (!new_info) { + return (AE_NO_MEMORY); + } + + ACPI_MOVE_NAME(new_info->signature, signature); + + if (!gbl_table_list_head) { + gbl_table_list_head = new_info; + } else { + next = gbl_table_list_head; + while (1) { + if (ACPI_COMPARE_NAME(next->signature, signature)) { + if (next->instance == instance) { + found = TRUE; + } + if (next->instance >= next_instance) { + next_instance = next->instance + 1; + } + } + + if (!next->next) { + break; + } + next = next->next; + } + next->next = new_info; + } + + if (found) { + if (instance) { + fprintf(stderr, + "%4.4s: Warning unmatched table instance %d, expected %d\n", + signature, instance, next_instance); + } + instance = next_instance; + } + + new_info->instance = instance; + gbl_table_count++; + + return (AE_OK); +} + +/****************************************************************************** + * + * FUNCTION: acpi_os_get_table_by_index + * + * PARAMETERS: index - Which table to get + * table - Where a pointer to the table is returned + * instance - Where a pointer to the table instance no. is + * returned + * address - Where the table physical address is returned + * + * RETURN: Status; Table buffer and physical address returned if AE_OK. + * AE_LIMIT: Index is beyond valid limit + * + * DESCRIPTION: Get an ACPI table via an index value (0 through n). Returns + * AE_LIMIT when an invalid index is reached. Index is not + * necessarily an index into the RSDT/XSDT. + * + *****************************************************************************/ + +acpi_status +acpi_os_get_table_by_index(u32 index, + struct acpi_table_header ** table, + u32 *instance, acpi_physical_address * address) +{ + struct osl_table_info *info; + acpi_status status; + u32 i; + + /* Get main ACPI tables from memory on first invocation of this function */ + + status = osl_table_initialize(); + if (ACPI_FAILURE(status)) { + return (status); + } + + /* Validate Index */ + + if (index >= gbl_table_count) { + return (AE_LIMIT); + } + + /* Point to the table list entry specified by the Index argument */ + + info = gbl_table_list_head; + for (i = 0; i < index; i++) { + info = info->next; + } + + /* Now we can just get the table via the signature */ + + status = acpi_os_get_table_by_name(info->signature, info->instance, + table, address); + + if (ACPI_SUCCESS(status)) { + *instance = info->instance; + } + return (status); +} + +/****************************************************************************** + * + * FUNCTION: osl_find_rsdp_via_efi + * + * PARAMETERS: None + * + * RETURN: RSDP address if found + * + * DESCRIPTION: Find RSDP address via EFI. + * + *****************************************************************************/ + +static acpi_physical_address osl_find_rsdp_via_efi(void) +{ + FILE *file; + char buffer[80]; + unsigned long address = 0; + + file = fopen(EFI_SYSTAB, "r"); + if (file) { + while (fgets(buffer, 80, file)) { + if (sscanf(buffer, "ACPI20=0x%lx", &address) == 1) { + break; + } + } + fclose(file); + } + + return ((acpi_physical_address) (address)); +} + +/****************************************************************************** + * + * FUNCTION: osl_load_rsdp + * + * PARAMETERS: None + * + * RETURN: Status + * + * DESCRIPTION: Scan and load RSDP. + * + *****************************************************************************/ + +static acpi_status osl_load_rsdp(void) +{ + struct acpi_table_header *mapped_table; + u8 *rsdp_address; + acpi_physical_address rsdp_base; + acpi_size rsdp_size; + + /* Get RSDP from memory */ + + rsdp_size = sizeof(struct acpi_table_rsdp); + if (gbl_rsdp_base) { + rsdp_base = gbl_rsdp_base; + } else { + rsdp_base = osl_find_rsdp_via_efi(); + } + + if (!rsdp_base) { + rsdp_base = ACPI_HI_RSDP_WINDOW_BASE; + rsdp_size = ACPI_HI_RSDP_WINDOW_SIZE; + } + + rsdp_address = acpi_os_map_memory(rsdp_base, rsdp_size); + if (!rsdp_address) { + return (osl_get_last_status(AE_BAD_ADDRESS)); + } + + /* Search low memory for the RSDP */ + + mapped_table = ACPI_CAST_PTR(struct acpi_table_header, + acpi_tb_scan_memory_for_rsdp(rsdp_address, + rsdp_size)); + if (!mapped_table) { + acpi_os_unmap_memory(rsdp_address, rsdp_size); + return (AE_NOT_FOUND); + } + + gbl_rsdp_address = + rsdp_base + (ACPI_CAST8(mapped_table) - rsdp_address); + + ACPI_MEMCPY(&gbl_rsdp, mapped_table, sizeof(struct acpi_table_rsdp)); + acpi_os_unmap_memory(rsdp_address, rsdp_size); + + return (AE_OK); +} + +/****************************************************************************** + * + * FUNCTION: osl_can_use_xsdt + * + * PARAMETERS: None + * + * RETURN: TRUE if XSDT is allowed to be used. + * + * DESCRIPTION: This function collects logic that can be used to determine if + * XSDT should be used instead of RSDT. + * + *****************************************************************************/ + +static u8 osl_can_use_xsdt(void) +{ + if (gbl_revision && !acpi_gbl_do_not_use_xsdt) { + return (TRUE); + } else { + return (FALSE); + } +} + +/****************************************************************************** + * + * FUNCTION: osl_table_initialize + * + * PARAMETERS: None + * + * RETURN: Status + * + * DESCRIPTION: Initialize ACPI table data. Get and store main ACPI tables to + * local variables. Main ACPI tables include RSDT, FADT, RSDT, + * and/or XSDT. + * + *****************************************************************************/ + +static acpi_status osl_table_initialize(void) +{ + acpi_status status; + acpi_physical_address address; + + if (gbl_table_list_initialized) { + return (AE_OK); + } + + /* Get RSDP from memory */ + + status = osl_load_rsdp(); + if (ACPI_FAILURE(status)) { + return (status); + } + + /* Get XSDT from memory */ + + if (gbl_rsdp.revision && !gbl_do_not_dump_xsdt) { + if (gbl_xsdt) { + free(gbl_xsdt); + gbl_xsdt = NULL; + } + + gbl_revision = 2; + status = osl_get_bios_table(ACPI_SIG_XSDT, 0, + ACPI_CAST_PTR(struct + acpi_table_header *, + &gbl_xsdt), &address); + if (ACPI_FAILURE(status)) { + return (status); + } + } + + /* Get RSDT from memory */ + + if (gbl_rsdp.rsdt_physical_address) { + if (gbl_rsdt) { + free(gbl_rsdt); + gbl_rsdt = NULL; + } + + status = osl_get_bios_table(ACPI_SIG_RSDT, 0, + ACPI_CAST_PTR(struct + acpi_table_header *, + &gbl_rsdt), &address); + if (ACPI_FAILURE(status)) { + return (status); + } + } + + /* Get FADT from memory */ + + if (gbl_fadt) { + free(gbl_fadt); + gbl_fadt = NULL; + } + + status = osl_get_bios_table(ACPI_SIG_FADT, 0, + ACPI_CAST_PTR(struct acpi_table_header *, + &gbl_fadt), + &gbl_fadt_address); + if (ACPI_FAILURE(status)) { + return (status); + } + + if (!gbl_dump_customized_tables) { + + /* Add mandatory tables to global table list first */ + + status = osl_add_table_to_list(ACPI_RSDP_NAME, 0); + if (ACPI_FAILURE(status)) { + return (status); + } + + status = osl_add_table_to_list(ACPI_SIG_RSDT, 0); + if (ACPI_FAILURE(status)) { + return (status); + } + + if (gbl_revision == 2) { + status = osl_add_table_to_list(ACPI_SIG_XSDT, 0); + if (ACPI_FAILURE(status)) { + return (status); + } + } + + status = osl_add_table_to_list(ACPI_SIG_DSDT, 0); + if (ACPI_FAILURE(status)) { + return (status); + } + + status = osl_add_table_to_list(ACPI_SIG_FACS, 0); + if (ACPI_FAILURE(status)) { + return (status); + } + + /* Add all tables found in the memory */ + + status = osl_list_bios_tables(); + if (ACPI_FAILURE(status)) { + return (status); + } + } else { + /* Add all tables found in the static directory */ + + status = osl_list_customized_tables(STATIC_TABLE_DIR); + if (ACPI_FAILURE(status)) { + return (status); + } + } + + if (gbl_dump_dynamic_tables) { + + /* Add all dynamically loaded tables in the dynamic directory */ + + status = osl_list_customized_tables(DYNAMIC_TABLE_DIR); + if (ACPI_FAILURE(status)) { + return (status); + } + } + + gbl_table_list_initialized = TRUE; + return (AE_OK); +} + +/****************************************************************************** + * + * FUNCTION: osl_list_bios_tables + * + * PARAMETERS: None + * + * RETURN: Status; Table list is initialized if AE_OK. + * + * DESCRIPTION: Add ACPI tables to the table list from memory. + * + * NOTE: This works on Linux as table customization does not modify the + * addresses stored in RSDP/RSDT/XSDT/FADT. + * + *****************************************************************************/ + +static acpi_status osl_list_bios_tables(void) +{ + struct acpi_table_header *mapped_table = NULL; + u8 *table_data; + u8 number_of_tables; + u8 item_size; + acpi_physical_address table_address = 0; + acpi_status status = AE_OK; + u32 i; + + if (osl_can_use_xsdt()) { + item_size = sizeof(u64); + table_data = + ACPI_CAST8(gbl_xsdt) + sizeof(struct acpi_table_header); + number_of_tables = + (u8)((gbl_xsdt->header.length - + sizeof(struct acpi_table_header)) + / item_size); + } else { /* Use RSDT if XSDT is not available */ + + item_size = sizeof(u32); + table_data = + ACPI_CAST8(gbl_rsdt) + sizeof(struct acpi_table_header); + number_of_tables = + (u8)((gbl_rsdt->header.length - + sizeof(struct acpi_table_header)) + / item_size); + } + + /* Search RSDT/XSDT for the requested table */ + + for (i = 0; i < number_of_tables; ++i, table_data += item_size) { + if (osl_can_use_xsdt()) { + table_address = + (acpi_physical_address) (*ACPI_CAST64(table_data)); + } else { + table_address = + (acpi_physical_address) (*ACPI_CAST32(table_data)); + } + + /* Skip NULL entries in RSDT/XSDT */ + + if (!table_address) { + continue; + } + + status = osl_map_table(table_address, NULL, &mapped_table); + if (ACPI_FAILURE(status)) { + return (status); + } + + osl_add_table_to_list(mapped_table->signature, 0); + osl_unmap_table(mapped_table); + } + + return (AE_OK); +} + +/****************************************************************************** + * + * FUNCTION: osl_get_bios_table + * + * PARAMETERS: signature - ACPI Signature for common table. Must be + * a null terminated 4-character string. + * instance - Multiple table support for SSDT/UEFI (0...n) + * Must be 0 for other tables. + * table - Where a pointer to the table is returned + * address - Where the table physical address is returned + * + * RETURN: Status; Table buffer and physical address returned if AE_OK. + * AE_LIMIT: Instance is beyond valid limit + * AE_NOT_FOUND: A table with the signature was not found + * + * DESCRIPTION: Get a BIOS provided ACPI table + * + * NOTE: Assumes the input signature is uppercase. + * + *****************************************************************************/ + +static acpi_status +osl_get_bios_table(char *signature, + u32 instance, + struct acpi_table_header **table, + acpi_physical_address * address) +{ + struct acpi_table_header *local_table = NULL; + struct acpi_table_header *mapped_table = NULL; + u8 *table_data; + u8 number_of_tables; + u8 item_size; + u32 current_instance = 0; + acpi_physical_address table_address = 0; + u32 table_length = 0; + acpi_status status = AE_OK; + u32 i; + + /* Handle special tables whose addresses are not in RSDT/XSDT */ + + if (ACPI_COMPARE_NAME(signature, ACPI_RSDP_NAME) || + ACPI_COMPARE_NAME(signature, ACPI_SIG_RSDT) || + ACPI_COMPARE_NAME(signature, ACPI_SIG_XSDT) || + ACPI_COMPARE_NAME(signature, ACPI_SIG_DSDT) || + ACPI_COMPARE_NAME(signature, ACPI_SIG_FACS)) { + if (instance > 0) { + return (AE_LIMIT); + } + + /* + * Get the appropriate address, either 32-bit or 64-bit. Be very + * careful about the FADT length and validate table addresses. + * Note: The 64-bit addresses have priority. + */ + if (ACPI_COMPARE_NAME(signature, ACPI_SIG_DSDT)) { + if ((gbl_fadt->header.length >= MIN_FADT_FOR_XDSDT) && + gbl_fadt->Xdsdt) { + table_address = + (acpi_physical_address) gbl_fadt->Xdsdt; + } else + if ((gbl_fadt->header.length >= MIN_FADT_FOR_DSDT) + && gbl_fadt->dsdt) { + table_address = + (acpi_physical_address) gbl_fadt->dsdt; + } + } else if (ACPI_COMPARE_NAME(signature, ACPI_SIG_FACS)) { + if ((gbl_fadt->header.length >= MIN_FADT_FOR_XFACS) && + gbl_fadt->Xfacs) { + table_address = + (acpi_physical_address) gbl_fadt->Xfacs; + } else + if ((gbl_fadt->header.length >= MIN_FADT_FOR_FACS) + && gbl_fadt->facs) { + table_address = + (acpi_physical_address) gbl_fadt->facs; + } + } else if (ACPI_COMPARE_NAME(signature, ACPI_SIG_XSDT)) { + if (!gbl_revision) { + return (AE_BAD_SIGNATURE); + } + table_address = + (acpi_physical_address) gbl_rsdp. + xsdt_physical_address; + } else if (ACPI_COMPARE_NAME(signature, ACPI_SIG_RSDT)) { + table_address = + (acpi_physical_address) gbl_rsdp. + rsdt_physical_address; + } else { + table_address = + (acpi_physical_address) gbl_rsdp_address; + signature = ACPI_SIG_RSDP; + } + + /* Now we can get the requested special table */ + + status = osl_map_table(table_address, signature, &mapped_table); + if (ACPI_FAILURE(status)) { + return (status); + } + + table_length = ap_get_table_length(mapped_table); + } else { /* Case for a normal ACPI table */ + + if (osl_can_use_xsdt()) { + item_size = sizeof(u64); + table_data = + ACPI_CAST8(gbl_xsdt) + + sizeof(struct acpi_table_header); + number_of_tables = + (u8)((gbl_xsdt->header.length - + sizeof(struct acpi_table_header)) + / item_size); + } else { /* Use RSDT if XSDT is not available */ + + item_size = sizeof(u32); + table_data = + ACPI_CAST8(gbl_rsdt) + + sizeof(struct acpi_table_header); + number_of_tables = + (u8)((gbl_rsdt->header.length - + sizeof(struct acpi_table_header)) + / item_size); + } + + /* Search RSDT/XSDT for the requested table */ + + for (i = 0; i < number_of_tables; ++i, table_data += item_size) { + if (osl_can_use_xsdt()) { + table_address = + (acpi_physical_address) (*ACPI_CAST64 + (table_data)); + } else { + table_address = + (acpi_physical_address) (*ACPI_CAST32 + (table_data)); + } + + /* Skip NULL entries in RSDT/XSDT */ + + if (!table_address) { + continue; + } + + status = + osl_map_table(table_address, NULL, &mapped_table); + if (ACPI_FAILURE(status)) { + return (status); + } + table_length = mapped_table->length; + + /* Does this table match the requested signature? */ + + if (!ACPI_COMPARE_NAME + (mapped_table->signature, signature)) { + osl_unmap_table(mapped_table); + mapped_table = NULL; + continue; + } + + /* Match table instance (for SSDT/UEFI tables) */ + + if (current_instance != instance) { + osl_unmap_table(mapped_table); + mapped_table = NULL; + current_instance++; + continue; + } + + break; + } + } + + if (!mapped_table) { + return (AE_LIMIT); + } + + if (table_length == 0) { + status = AE_BAD_HEADER; + goto exit; + } + + /* Copy table to local buffer and return it */ + + local_table = calloc(1, table_length); + if (!local_table) { + status = AE_NO_MEMORY; + goto exit; + } + + ACPI_MEMCPY(local_table, mapped_table, table_length); + *address = table_address; + *table = local_table; + +exit: + osl_unmap_table(mapped_table); + return (status); +} + +/****************************************************************************** + * + * FUNCTION: osl_list_customized_tables + * + * PARAMETERS: directory - Directory that contains the tables + * + * RETURN: Status; Table list is initialized if AE_OK. + * + * DESCRIPTION: Add ACPI tables to the table list from a directory. + * + *****************************************************************************/ + +static acpi_status osl_list_customized_tables(char *directory) +{ + void *table_dir; + u32 instance; + char temp_name[ACPI_NAME_SIZE]; + char *filename; + acpi_status status = AE_OK; + + /* Open the requested directory */ + + table_dir = acpi_os_open_directory(directory, "*", REQUEST_FILE_ONLY); + if (!table_dir) { + return (osl_get_last_status(AE_NOT_FOUND)); + } + + /* Examine all entries in this directory */ + + while ((filename = acpi_os_get_next_filename(table_dir))) { + + /* Extract table name and instance number */ + + status = + osl_table_name_from_file(filename, temp_name, &instance); + + /* Ignore meaningless files */ + + if (ACPI_FAILURE(status)) { + continue; + } + + /* Add new info node to global table list */ + + status = osl_add_table_to_list(temp_name, instance); + if (ACPI_FAILURE(status)) { + break; + } + } + + acpi_os_close_directory(table_dir); + return (status); +} + +/****************************************************************************** + * + * FUNCTION: osl_map_table + * + * PARAMETERS: address - Address of the table in memory + * signature - Optional ACPI Signature for desired table. + * Null terminated 4-character string. + * table - Where a pointer to the mapped table is + * returned + * + * RETURN: Status; Mapped table is returned if AE_OK. + * AE_NOT_FOUND: A valid table was not found at the address + * + * DESCRIPTION: Map entire ACPI table into caller's address space. + * + *****************************************************************************/ + +static acpi_status +osl_map_table(acpi_size address, + char *signature, struct acpi_table_header **table) +{ + struct acpi_table_header *mapped_table; + u32 length; + + if (!address) { + return (AE_BAD_ADDRESS); + } + + /* + * Map the header so we can get the table length. + * Use sizeof (struct acpi_table_header) as: + * 1. it is bigger than 24 to include RSDP->Length + * 2. it is smaller than sizeof (struct acpi_table_rsdp) + */ + mapped_table = + acpi_os_map_memory(address, sizeof(struct acpi_table_header)); + if (!mapped_table) { + fprintf(stderr, "Could not map table header at 0x%8.8X%8.8X\n", + ACPI_FORMAT_UINT64(address)); + return (osl_get_last_status(AE_BAD_ADDRESS)); + } + + /* If specified, signature must match */ + + if (signature) { + if (ACPI_VALIDATE_RSDP_SIG(signature)) { + if (!ACPI_VALIDATE_RSDP_SIG(mapped_table->signature)) { + acpi_os_unmap_memory(mapped_table, + sizeof(struct + acpi_table_header)); + return (AE_BAD_SIGNATURE); + } + } else + if (!ACPI_COMPARE_NAME(signature, mapped_table->signature)) + { + acpi_os_unmap_memory(mapped_table, + sizeof(struct acpi_table_header)); + return (AE_BAD_SIGNATURE); + } + } + + /* Map the entire table */ + + length = ap_get_table_length(mapped_table); + acpi_os_unmap_memory(mapped_table, sizeof(struct acpi_table_header)); + if (length == 0) { + return (AE_BAD_HEADER); + } + + mapped_table = acpi_os_map_memory(address, length); + if (!mapped_table) { + fprintf(stderr, + "Could not map table at 0x%8.8X%8.8X length %8.8X\n", + ACPI_FORMAT_UINT64(address), length); + return (osl_get_last_status(AE_INVALID_TABLE_LENGTH)); + } + + (void)ap_is_valid_checksum(mapped_table); + + *table = mapped_table; + return (AE_OK); +} + +/****************************************************************************** + * + * FUNCTION: osl_unmap_table + * + * PARAMETERS: table - A pointer to the mapped table + * + * RETURN: None + * + * DESCRIPTION: Unmap entire ACPI table. + * + *****************************************************************************/ + +static void osl_unmap_table(struct acpi_table_header *table) +{ + if (table) { + acpi_os_unmap_memory(table, ap_get_table_length(table)); + } +} + +/****************************************************************************** + * + * FUNCTION: osl_table_name_from_file + * + * PARAMETERS: filename - File that contains the desired table + * signature - Pointer to 4-character buffer to store + * extracted table signature. + * instance - Pointer to integer to store extracted + * table instance number. + * + * RETURN: Status; Table name is extracted if AE_OK. + * + * DESCRIPTION: Extract table signature and instance number from a table file + * name. + * + *****************************************************************************/ + +static acpi_status +osl_table_name_from_file(char *filename, char *signature, u32 *instance) +{ + + /* Ignore meaningless files */ + + if (strlen(filename) < ACPI_NAME_SIZE) { + return (AE_BAD_SIGNATURE); + } + + /* Extract instance number */ + + if (isdigit((int)filename[ACPI_NAME_SIZE])) { + sscanf(&filename[ACPI_NAME_SIZE], "%d", instance); + } else if (strlen(filename) != ACPI_NAME_SIZE) { + return (AE_BAD_SIGNATURE); + } else { + *instance = 0; + } + + /* Extract signature */ + + ACPI_MOVE_NAME(signature, filename); + return (AE_OK); +} + +/****************************************************************************** + * + * FUNCTION: osl_read_table_from_file + * + * PARAMETERS: filename - File that contains the desired table + * file_offset - Offset of the table in file + * signature - Optional ACPI Signature for desired table. + * A null terminated 4-character string. + * table - Where a pointer to the table is returned + * + * RETURN: Status; Table buffer is returned if AE_OK. + * + * DESCRIPTION: Read a ACPI table from a file. + * + *****************************************************************************/ + +static acpi_status +osl_read_table_from_file(char *filename, + acpi_size file_offset, + char *signature, struct acpi_table_header **table) +{ + FILE *table_file; + struct acpi_table_header header; + struct acpi_table_header *local_table = NULL; + u32 table_length; + s32 count; + acpi_status status = AE_OK; + + /* Open the file */ + + table_file = fopen(filename, "rb"); + if (table_file == NULL) { + fprintf(stderr, "Could not open table file: %s\n", filename); + return (osl_get_last_status(AE_NOT_FOUND)); + } + + fseek(table_file, file_offset, SEEK_SET); + + /* Read the Table header to get the table length */ + + count = fread(&header, 1, sizeof(struct acpi_table_header), table_file); + if (count != sizeof(struct acpi_table_header)) { + fprintf(stderr, "Could not read table header: %s\n", filename); + status = AE_BAD_HEADER; + goto exit; + } + + /* If signature is specified, it must match the table */ + + if (signature) { + if (ACPI_VALIDATE_RSDP_SIG(signature)) { + if (!ACPI_VALIDATE_RSDP_SIG(header.signature)) { + fprintf(stderr, + "Incorrect RSDP signature: found %8.8s\n", + header.signature); + status = AE_BAD_SIGNATURE; + goto exit; + } + } else if (!ACPI_COMPARE_NAME(signature, header.signature)) { + fprintf(stderr, + "Incorrect signature: Expecting %4.4s, found %4.4s\n", + signature, header.signature); + status = AE_BAD_SIGNATURE; + goto exit; + } + } + + table_length = ap_get_table_length(&header); + if (table_length == 0) { + status = AE_BAD_HEADER; + goto exit; + } + + /* Read the entire table into a local buffer */ + + local_table = calloc(1, table_length); + if (!local_table) { + fprintf(stderr, + "%4.4s: Could not allocate buffer for table of length %X\n", + header.signature, table_length); + status = AE_NO_MEMORY; + goto exit; + } + + fseek(table_file, file_offset, SEEK_SET); + + count = fread(local_table, 1, table_length, table_file); + if (count != table_length) { + fprintf(stderr, "%4.4s: Could not read table content\n", + header.signature); + status = AE_INVALID_TABLE_LENGTH; + goto exit; + } + + /* Validate checksum */ + + (void)ap_is_valid_checksum(local_table); + +exit: + fclose(table_file); + *table = local_table; + return (status); +} + +/****************************************************************************** + * + * FUNCTION: osl_get_customized_table + * + * PARAMETERS: pathname - Directory to find Linux customized table + * signature - ACPI Signature for desired table. Must be + * a null terminated 4-character string. + * instance - Multiple table support for SSDT/UEFI (0...n) + * Must be 0 for other tables. + * table - Where a pointer to the table is returned + * address - Where the table physical address is returned + * + * RETURN: Status; Table buffer is returned if AE_OK. + * AE_LIMIT: Instance is beyond valid limit + * AE_NOT_FOUND: A table with the signature was not found + * + * DESCRIPTION: Get an OS customized table. + * + *****************************************************************************/ + +static acpi_status +osl_get_customized_table(char *pathname, + char *signature, + u32 instance, + struct acpi_table_header **table, + acpi_physical_address * address) +{ + void *table_dir; + u32 current_instance = 0; + char temp_name[ACPI_NAME_SIZE]; + char table_filename[PATH_MAX]; + char *filename; + acpi_status status; + + /* Open the directory for customized tables */ + + table_dir = acpi_os_open_directory(pathname, "*", REQUEST_FILE_ONLY); + if (!table_dir) { + return (osl_get_last_status(AE_NOT_FOUND)); + } + + /* Attempt to find the table in the directory */ + + while ((filename = acpi_os_get_next_filename(table_dir))) { + + /* Ignore meaningless files */ + + if (!ACPI_COMPARE_NAME(filename, signature)) { + continue; + } + + /* Extract table name and instance number */ + + status = + osl_table_name_from_file(filename, temp_name, + ¤t_instance); + + /* Ignore meaningless files */ + + if (ACPI_FAILURE(status) || current_instance != instance) { + continue; + } + + /* Create the table pathname */ + + if (instance != 0) { + sprintf(table_filename, "%s/%4.4s%d", pathname, + temp_name, instance); + } else { + sprintf(table_filename, "%s/%4.4s", pathname, + temp_name); + } + break; + } + + acpi_os_close_directory(table_dir); + + if (!filename) { + return (AE_LIMIT); + } + + /* There is no physical address saved for customized tables, use zero */ + + *address = 0; + status = osl_read_table_from_file(table_filename, 0, NULL, table); + + return (status); +} diff --git a/tools/power/acpi/os_specific/service_layers/osunixdir.c b/tools/power/acpi/os_specific/service_layers/osunixdir.c new file mode 100644 index 00000000000..733f9e490fc --- /dev/null +++ b/tools/power/acpi/os_specific/service_layers/osunixdir.c @@ -0,0 +1,204 @@ +/****************************************************************************** + * + * Module Name: osunixdir - Unix directory access interfaces + * + *****************************************************************************/ + +/* + * Copyright (C) 2000 - 2014, Intel Corp. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions, and the following disclaimer, + * without modification. + * 2. Redistributions in binary form must reproduce at minimum a disclaimer + * substantially similar to the "NO WARRANTY" disclaimer below + * ("Disclaimer") and any redistribution must be conditioned upon + * including a substantially similar Disclaimer requirement for further + * binary redistribution. + * 3. Neither the names of the above-listed copyright holders nor the names + * of any contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * Alternatively, this software may be distributed under the terms of the + * GNU General Public License ("GPL") version 2 as published by the Free + * Software Foundation. + * + * NO WARRANTY + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGES. + */ + +#include <acpi/acpi.h> + +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <dirent.h> +#include <fnmatch.h> +#include <ctype.h> +#include <sys/stat.h> + +/* + * Allocated structure returned from os_open_directory + */ +typedef struct external_find_info { + char *dir_pathname; + DIR *dir_ptr; + char temp_buffer[256]; + char *wildcard_spec; + char requested_file_type; + +} external_find_info; + +/******************************************************************************* + * + * FUNCTION: acpi_os_open_directory + * + * PARAMETERS: dir_pathname - Full pathname to the directory + * wildcard_spec - string of the form "*.c", etc. + * + * RETURN: A directory "handle" to be used in subsequent search operations. + * NULL returned on failure. + * + * DESCRIPTION: Open a directory in preparation for a wildcard search + * + ******************************************************************************/ + +void *acpi_os_open_directory(char *dir_pathname, + char *wildcard_spec, char requested_file_type) +{ + struct external_find_info *external_info; + DIR *dir; + + /* Allocate the info struct that will be returned to the caller */ + + external_info = calloc(1, sizeof(struct external_find_info)); + if (!external_info) { + return (NULL); + } + + /* Get the directory stream */ + + dir = opendir(dir_pathname); + if (!dir) { + fprintf(stderr, "Cannot open directory - %s\n", dir_pathname); + free(external_info); + return (NULL); + } + + /* Save the info in the return structure */ + + external_info->wildcard_spec = wildcard_spec; + external_info->requested_file_type = requested_file_type; + external_info->dir_pathname = dir_pathname; + external_info->dir_ptr = dir; + return (external_info); +} + +/******************************************************************************* + * + * FUNCTION: acpi_os_get_next_filename + * + * PARAMETERS: dir_handle - Created via acpi_os_open_directory + * + * RETURN: Next filename matched. NULL if no more matches. + * + * DESCRIPTION: Get the next file in the directory that matches the wildcard + * specification. + * + ******************************************************************************/ + +char *acpi_os_get_next_filename(void *dir_handle) +{ + struct external_find_info *external_info = dir_handle; + struct dirent *dir_entry; + char *temp_str; + int str_len; + struct stat temp_stat; + int err; + + while ((dir_entry = readdir(external_info->dir_ptr))) { + if (!fnmatch + (external_info->wildcard_spec, dir_entry->d_name, 0)) { + if (dir_entry->d_name[0] == '.') { + continue; + } + + str_len = strlen(dir_entry->d_name) + + strlen(external_info->dir_pathname) + 2; + + temp_str = calloc(str_len, 1); + if (!temp_str) { + fprintf(stderr, + "Could not allocate buffer for temporary string\n"); + return (NULL); + } + + strcpy(temp_str, external_info->dir_pathname); + strcat(temp_str, "/"); + strcat(temp_str, dir_entry->d_name); + + err = stat(temp_str, &temp_stat); + if (err == -1) { + fprintf(stderr, + "Cannot stat file (should not happen) - %s\n", + temp_str); + free(temp_str); + return (NULL); + } + + free(temp_str); + + if ((S_ISDIR(temp_stat.st_mode) + && (external_info->requested_file_type == + REQUEST_DIR_ONLY)) + || ((!S_ISDIR(temp_stat.st_mode) + && external_info->requested_file_type == + REQUEST_FILE_ONLY))) { + + /* copy to a temp buffer because dir_entry struct is on the stack */ + + strcpy(external_info->temp_buffer, + dir_entry->d_name); + return (external_info->temp_buffer); + } + } + } + + return (NULL); +} + +/******************************************************************************* + * + * FUNCTION: acpi_os_close_directory + * + * PARAMETERS: dir_handle - Created via acpi_os_open_directory + * + * RETURN: None. + * + * DESCRIPTION: Close the open directory and cleanup. + * + ******************************************************************************/ + +void acpi_os_close_directory(void *dir_handle) +{ + struct external_find_info *external_info = dir_handle; + + /* Close the directory and free allocations */ + + closedir(external_info->dir_ptr); + free(dir_handle); +} diff --git a/tools/power/acpi/os_specific/service_layers/osunixmap.c b/tools/power/acpi/os_specific/service_layers/osunixmap.c new file mode 100644 index 00000000000..99b47b6194a --- /dev/null +++ b/tools/power/acpi/os_specific/service_layers/osunixmap.c @@ -0,0 +1,151 @@ +/****************************************************************************** + * + * Module Name: osunixmap - Unix OSL for file mappings + * + *****************************************************************************/ + +/* + * Copyright (C) 2000 - 2014, Intel Corp. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions, and the following disclaimer, + * without modification. + * 2. Redistributions in binary form must reproduce at minimum a disclaimer + * substantially similar to the "NO WARRANTY" disclaimer below + * ("Disclaimer") and any redistribution must be conditioned upon + * including a substantially similar Disclaimer requirement for further + * binary redistribution. + * 3. Neither the names of the above-listed copyright holders nor the names + * of any contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * Alternatively, this software may be distributed under the terms of the + * GNU General Public License ("GPL") version 2 as published by the Free + * Software Foundation. + * + * NO WARRANTY + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGES. + */ + +#include "acpidump.h" +#include <unistd.h> +#include <sys/mman.h> +#ifdef _free_BSD +#include <sys/param.h> +#endif + +#define _COMPONENT ACPI_OS_SERVICES +ACPI_MODULE_NAME("osunixmap") + +#ifndef O_BINARY +#define O_BINARY 0 +#endif +#ifdef _free_BSD +#define MMAP_FLAGS MAP_SHARED +#else +#define MMAP_FLAGS MAP_PRIVATE +#endif +#define SYSTEM_MEMORY "/dev/mem" +/******************************************************************************* + * + * FUNCTION: acpi_os_get_page_size + * + * PARAMETERS: None + * + * RETURN: Page size of the platform. + * + * DESCRIPTION: Obtain page size of the platform. + * + ******************************************************************************/ +static acpi_size acpi_os_get_page_size(void) +{ + +#ifdef PAGE_SIZE + return PAGE_SIZE; +#else + return sysconf(_SC_PAGESIZE); +#endif +} + +/****************************************************************************** + * + * FUNCTION: acpi_os_map_memory + * + * PARAMETERS: where - Physical address of memory to be mapped + * length - How much memory to map + * + * RETURN: Pointer to mapped memory. Null on error. + * + * DESCRIPTION: Map physical memory into local address space. + * + *****************************************************************************/ + +void *acpi_os_map_memory(acpi_physical_address where, acpi_size length) +{ + u8 *mapped_memory; + acpi_physical_address offset; + acpi_size page_size; + int fd; + + fd = open(SYSTEM_MEMORY, O_RDONLY | O_BINARY); + if (fd < 0) { + fprintf(stderr, "Cannot open %s\n", SYSTEM_MEMORY); + return (NULL); + } + + /* Align the offset to use mmap */ + + page_size = acpi_os_get_page_size(); + offset = where % page_size; + + /* Map the table header to get the length of the full table */ + + mapped_memory = mmap(NULL, (length + offset), PROT_READ, MMAP_FLAGS, + fd, (where - offset)); + if (mapped_memory == MAP_FAILED) { + fprintf(stderr, "Cannot map %s\n", SYSTEM_MEMORY); + close(fd); + return (NULL); + } + + close(fd); + return (ACPI_CAST8(mapped_memory + offset)); +} + +/****************************************************************************** + * + * FUNCTION: acpi_os_unmap_memory + * + * PARAMETERS: where - Logical address of memory to be unmapped + * length - How much memory to unmap + * + * RETURN: None. + * + * DESCRIPTION: Delete a previously created mapping. Where and Length must + * correspond to a previous mapping exactly. + * + *****************************************************************************/ + +void acpi_os_unmap_memory(void *where, acpi_size length) +{ + acpi_physical_address offset; + acpi_size page_size; + + page_size = acpi_os_get_page_size(); + offset = (acpi_physical_address) where % page_size; + munmap((u8 *)where - offset, (length + offset)); +} diff --git a/tools/power/acpi/tools/acpidump/acpidump.h b/tools/power/acpi/tools/acpidump/acpidump.h new file mode 100644 index 00000000000..46f519597fe --- /dev/null +++ b/tools/power/acpi/tools/acpidump/acpidump.h @@ -0,0 +1,130 @@ +/****************************************************************************** + * + * Module Name: acpidump.h - Include file for acpi_dump utility + * + *****************************************************************************/ + +/* + * Copyright (C) 2000 - 2014, Intel Corp. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions, and the following disclaimer, + * without modification. + * 2. Redistributions in binary form must reproduce at minimum a disclaimer + * substantially similar to the "NO WARRANTY" disclaimer below + * ("Disclaimer") and any redistribution must be conditioned upon + * including a substantially similar Disclaimer requirement for further + * binary redistribution. + * 3. Neither the names of the above-listed copyright holders nor the names + * of any contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * Alternatively, this software may be distributed under the terms of the + * GNU General Public License ("GPL") version 2 as published by the Free + * Software Foundation. + * + * NO WARRANTY + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGES. + */ + +/* + * Global variables. Defined in main.c only, externed in all other files + */ +#ifdef _DECLARE_GLOBALS +#define EXTERN +#define INIT_GLOBAL(a,b) a=b +#define DEFINE_ACPI_GLOBALS 1 +#else +#define EXTERN extern +#define INIT_GLOBAL(a,b) a +#endif + +#include <acpi/acpi.h> +#include "accommon.h" +#include "actables.h" + +#include <stdio.h> +#include <fcntl.h> +#include <errno.h> +#include <sys/stat.h> + +/* Globals */ + +EXTERN u8 INIT_GLOBAL(gbl_summary_mode, FALSE); +EXTERN u8 INIT_GLOBAL(gbl_verbose_mode, FALSE); +EXTERN u8 INIT_GLOBAL(gbl_binary_mode, FALSE); +EXTERN u8 INIT_GLOBAL(gbl_dump_customized_tables, FALSE); +EXTERN u8 INIT_GLOBAL(gbl_do_not_dump_xsdt, FALSE); +EXTERN FILE INIT_GLOBAL(*gbl_output_file, NULL); +EXTERN char INIT_GLOBAL(*gbl_output_filename, NULL); +EXTERN u64 INIT_GLOBAL(gbl_rsdp_base, 0); + +/* Globals required for use with ACPICA modules */ + +#ifdef _DECLARE_GLOBALS +u8 acpi_gbl_integer_byte_width = 8; +#endif + +/* Action table used to defer requested options */ + +struct ap_dump_action { + char *argument; + u32 to_be_done; +}; + +#define AP_MAX_ACTIONS 32 + +#define AP_DUMP_ALL_TABLES 0 +#define AP_DUMP_TABLE_BY_ADDRESS 1 +#define AP_DUMP_TABLE_BY_NAME 2 +#define AP_DUMP_TABLE_BY_FILE 3 + +#define AP_MAX_ACPI_FILES 256 /* Prevent infinite loops */ + +/* Minimum FADT sizes for various table addresses */ + +#define MIN_FADT_FOR_DSDT (ACPI_FADT_OFFSET (dsdt) + sizeof (u32)) +#define MIN_FADT_FOR_FACS (ACPI_FADT_OFFSET (facs) + sizeof (u32)) +#define MIN_FADT_FOR_XDSDT (ACPI_FADT_OFFSET (Xdsdt) + sizeof (u64)) +#define MIN_FADT_FOR_XFACS (ACPI_FADT_OFFSET (Xfacs) + sizeof (u64)) + +/* + * apdump - Table get/dump routines + */ +int ap_dump_table_from_file(char *pathname); + +int ap_dump_table_by_name(char *signature); + +int ap_dump_table_by_address(char *ascii_address); + +int ap_dump_all_tables(void); + +u8 ap_is_valid_header(struct acpi_table_header *table); + +u8 ap_is_valid_checksum(struct acpi_table_header *table); + +u32 ap_get_table_length(struct acpi_table_header *table); + +/* + * apfiles - File I/O utilities + */ +int ap_open_output_file(char *pathname); + +int ap_write_to_binary_file(struct acpi_table_header *table, u32 instance); + +struct acpi_table_header *ap_get_table_from_file(char *pathname, + u32 *file_size); diff --git a/tools/power/acpi/tools/acpidump/apdump.c b/tools/power/acpi/tools/acpidump/apdump.c new file mode 100644 index 00000000000..3cac1237836 --- /dev/null +++ b/tools/power/acpi/tools/acpidump/apdump.c @@ -0,0 +1,451 @@ +/****************************************************************************** + * + * Module Name: apdump - Dump routines for ACPI tables (acpidump) + * + *****************************************************************************/ + +/* + * Copyright (C) 2000 - 2014, Intel Corp. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions, and the following disclaimer, + * without modification. + * 2. Redistributions in binary form must reproduce at minimum a disclaimer + * substantially similar to the "NO WARRANTY" disclaimer below + * ("Disclaimer") and any redistribution must be conditioned upon + * including a substantially similar Disclaimer requirement for further + * binary redistribution. + * 3. Neither the names of the above-listed copyright holders nor the names + * of any contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * Alternatively, this software may be distributed under the terms of the + * GNU General Public License ("GPL") version 2 as published by the Free + * Software Foundation. + * + * NO WARRANTY + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGES. + */ + +#include "acpidump.h" + +/* Local prototypes */ + +static int +ap_dump_table_buffer(struct acpi_table_header *table, + u32 instance, acpi_physical_address address); + +/****************************************************************************** + * + * FUNCTION: ap_is_valid_header + * + * PARAMETERS: table - Pointer to table to be validated + * + * RETURN: TRUE if the header appears to be valid. FALSE otherwise + * + * DESCRIPTION: Check for a valid ACPI table header + * + ******************************************************************************/ + +u8 ap_is_valid_header(struct acpi_table_header *table) +{ + + if (!ACPI_VALIDATE_RSDP_SIG(table->signature)) { + + /* Make sure signature is all ASCII and a valid ACPI name */ + + if (!acpi_ut_valid_acpi_name(table->signature)) { + fprintf(stderr, + "Table signature (0x%8.8X) is invalid\n", + *(u32 *)table->signature); + return (FALSE); + } + + /* Check for minimum table length */ + + if (table->length < sizeof(struct acpi_table_header)) { + fprintf(stderr, "Table length (0x%8.8X) is invalid\n", + table->length); + return (FALSE); + } + } + + return (TRUE); +} + +/****************************************************************************** + * + * FUNCTION: ap_is_valid_checksum + * + * PARAMETERS: table - Pointer to table to be validated + * + * RETURN: TRUE if the checksum appears to be valid. FALSE otherwise. + * + * DESCRIPTION: Check for a valid ACPI table checksum. + * + ******************************************************************************/ + +u8 ap_is_valid_checksum(struct acpi_table_header *table) +{ + acpi_status status; + struct acpi_table_rsdp *rsdp; + + if (ACPI_VALIDATE_RSDP_SIG(table->signature)) { + /* + * Checksum for RSDP. + * Note: Other checksums are computed during the table dump. + */ + rsdp = ACPI_CAST_PTR(struct acpi_table_rsdp, table); + status = acpi_tb_validate_rsdp(rsdp); + } else { + status = acpi_tb_verify_checksum(table, table->length); + } + + if (ACPI_FAILURE(status)) { + fprintf(stderr, "%4.4s: Warning: wrong checksum in table\n", + table->signature); + } + + return (AE_OK); +} + +/****************************************************************************** + * + * FUNCTION: ap_get_table_length + * + * PARAMETERS: table - Pointer to the table + * + * RETURN: Table length + * + * DESCRIPTION: Obtain table length according to table signature. + * + ******************************************************************************/ + +u32 ap_get_table_length(struct acpi_table_header *table) +{ + struct acpi_table_rsdp *rsdp; + + /* Check if table is valid */ + + if (!ap_is_valid_header(table)) { + return (0); + } + + if (ACPI_VALIDATE_RSDP_SIG(table->signature)) { + rsdp = ACPI_CAST_PTR(struct acpi_table_rsdp, table); + return (rsdp->length); + } + + /* Normal ACPI table */ + + return (table->length); +} + +/****************************************************************************** + * + * FUNCTION: ap_dump_table_buffer + * + * PARAMETERS: table - ACPI table to be dumped + * instance - ACPI table instance no. to be dumped + * address - Physical address of the table + * + * RETURN: None + * + * DESCRIPTION: Dump an ACPI table in standard ASCII hex format, with a + * header that is compatible with the acpi_xtract utility. + * + ******************************************************************************/ + +static int +ap_dump_table_buffer(struct acpi_table_header *table, + u32 instance, acpi_physical_address address) +{ + u32 table_length; + + table_length = ap_get_table_length(table); + + /* Print only the header if requested */ + + if (gbl_summary_mode) { + acpi_tb_print_table_header(address, table); + return (0); + } + + /* Dump to binary file if requested */ + + if (gbl_binary_mode) { + return (ap_write_to_binary_file(table, instance)); + } + + /* + * Dump the table with header for use with acpixtract utility. + * Note: simplest to just always emit a 64-bit address. acpi_xtract + * utility can handle this. + */ + printf("%4.4s @ 0x%8.8X%8.8X\n", table->signature, + ACPI_FORMAT_UINT64(address)); + + acpi_ut_dump_buffer(ACPI_CAST_PTR(u8, table), table_length, + DB_BYTE_DISPLAY, 0); + printf("\n"); + return (0); +} + +/****************************************************************************** + * + * FUNCTION: ap_dump_all_tables + * + * PARAMETERS: None + * + * RETURN: Status + * + * DESCRIPTION: Get all tables from the RSDT/XSDT (or at least all of the + * tables that we can possibly get). + * + ******************************************************************************/ + +int ap_dump_all_tables(void) +{ + struct acpi_table_header *table; + u32 instance = 0; + acpi_physical_address address; + acpi_status status; + int table_status; + u32 i; + + /* Get and dump all available ACPI tables */ + + for (i = 0; i < AP_MAX_ACPI_FILES; i++) { + status = + acpi_os_get_table_by_index(i, &table, &instance, &address); + if (ACPI_FAILURE(status)) { + + /* AE_LIMIT means that no more tables are available */ + + if (status == AE_LIMIT) { + return (0); + } else if (i == 0) { + fprintf(stderr, + "Could not get ACPI tables, %s\n", + acpi_format_exception(status)); + return (-1); + } else { + fprintf(stderr, + "Could not get ACPI table at index %u, %s\n", + i, acpi_format_exception(status)); + continue; + } + } + + table_status = ap_dump_table_buffer(table, instance, address); + free(table); + + if (table_status) { + break; + } + } + + /* Something seriously bad happened if the loop terminates here */ + + return (-1); +} + +/****************************************************************************** + * + * FUNCTION: ap_dump_table_by_address + * + * PARAMETERS: ascii_address - Address for requested ACPI table + * + * RETURN: Status + * + * DESCRIPTION: Get an ACPI table via a physical address and dump it. + * + ******************************************************************************/ + +int ap_dump_table_by_address(char *ascii_address) +{ + acpi_physical_address address; + struct acpi_table_header *table; + acpi_status status; + int table_status; + u64 long_address; + + /* Convert argument to an integer physical address */ + + status = acpi_ut_strtoul64(ascii_address, 0, &long_address); + if (ACPI_FAILURE(status)) { + fprintf(stderr, "%s: Could not convert to a physical address\n", + ascii_address); + return (-1); + } + + address = (acpi_physical_address) long_address; + status = acpi_os_get_table_by_address(address, &table); + if (ACPI_FAILURE(status)) { + fprintf(stderr, "Could not get table at 0x%8.8X%8.8X, %s\n", + ACPI_FORMAT_UINT64(address), + acpi_format_exception(status)); + return (-1); + } + + table_status = ap_dump_table_buffer(table, 0, address); + free(table); + return (table_status); +} + +/****************************************************************************** + * + * FUNCTION: ap_dump_table_by_name + * + * PARAMETERS: signature - Requested ACPI table signature + * + * RETURN: Status + * + * DESCRIPTION: Get an ACPI table via a signature and dump it. Handles + * multiple tables with the same signature (SSDTs). + * + ******************************************************************************/ + +int ap_dump_table_by_name(char *signature) +{ + char local_signature[ACPI_NAME_SIZE + 1]; + u32 instance; + struct acpi_table_header *table; + acpi_physical_address address; + acpi_status status; + int table_status; + + if (strlen(signature) != ACPI_NAME_SIZE) { + fprintf(stderr, + "Invalid table signature [%s]: must be exactly 4 characters\n", + signature); + return (-1); + } + + /* Table signatures are expected to be uppercase */ + + strcpy(local_signature, signature); + acpi_ut_strupr(local_signature); + + /* To be friendly, handle tables whose signatures do not match the name */ + + if (ACPI_COMPARE_NAME(local_signature, "FADT")) { + strcpy(local_signature, ACPI_SIG_FADT); + } else if (ACPI_COMPARE_NAME(local_signature, "MADT")) { + strcpy(local_signature, ACPI_SIG_MADT); + } + + /* Dump all instances of this signature (to handle multiple SSDTs) */ + + for (instance = 0; instance < AP_MAX_ACPI_FILES; instance++) { + status = acpi_os_get_table_by_name(local_signature, instance, + &table, &address); + if (ACPI_FAILURE(status)) { + + /* AE_LIMIT means that no more tables are available */ + + if (status == AE_LIMIT) { + return (0); + } + + fprintf(stderr, + "Could not get ACPI table with signature [%s], %s\n", + local_signature, acpi_format_exception(status)); + return (-1); + } + + table_status = ap_dump_table_buffer(table, instance, address); + free(table); + + if (table_status) { + break; + } + } + + /* Something seriously bad happened if the loop terminates here */ + + return (-1); +} + +/****************************************************************************** + * + * FUNCTION: ap_dump_table_from_file + * + * PARAMETERS: pathname - File containing the binary ACPI table + * + * RETURN: Status + * + * DESCRIPTION: Dump an ACPI table from a binary file + * + ******************************************************************************/ + +int ap_dump_table_from_file(char *pathname) +{ + struct acpi_table_header *table; + u32 file_size = 0; + int table_status = -1; + + /* Get the entire ACPI table from the file */ + + table = ap_get_table_from_file(pathname, &file_size); + if (!table) { + return (-1); + } + + /* File must be at least as long as the table length */ + + if (table->length > file_size) { + fprintf(stderr, + "Table length (0x%X) is too large for input file (0x%X) %s\n", + table->length, file_size, pathname); + goto exit; + } + + if (gbl_verbose_mode) { + fprintf(stderr, + "Input file: %s contains table [%4.4s], 0x%X (%u) bytes\n", + pathname, table->signature, file_size, file_size); + } + + table_status = ap_dump_table_buffer(table, 0, 0); + +exit: + free(table); + return (table_status); +} + +/****************************************************************************** + * + * FUNCTION: acpi_os* print functions + * + * DESCRIPTION: Used for linkage with ACPICA modules + * + ******************************************************************************/ + +void ACPI_INTERNAL_VAR_XFACE acpi_os_printf(const char *fmt, ...) +{ + va_list args; + + va_start(args, fmt); + vfprintf(stdout, fmt, args); + va_end(args); +} + +void acpi_os_vprintf(const char *fmt, va_list args) +{ + vfprintf(stdout, fmt, args); +} diff --git a/tools/power/acpi/tools/acpidump/apfiles.c b/tools/power/acpi/tools/acpidump/apfiles.c new file mode 100644 index 00000000000..4488accc010 --- /dev/null +++ b/tools/power/acpi/tools/acpidump/apfiles.c @@ -0,0 +1,228 @@ +/****************************************************************************** + * + * Module Name: apfiles - File-related functions for acpidump utility + * + *****************************************************************************/ + +/* + * Copyright (C) 2000 - 2014, Intel Corp. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions, and the following disclaimer, + * without modification. + * 2. Redistributions in binary form must reproduce at minimum a disclaimer + * substantially similar to the "NO WARRANTY" disclaimer below + * ("Disclaimer") and any redistribution must be conditioned upon + * including a substantially similar Disclaimer requirement for further + * binary redistribution. + * 3. Neither the names of the above-listed copyright holders nor the names + * of any contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * Alternatively, this software may be distributed under the terms of the + * GNU General Public License ("GPL") version 2 as published by the Free + * Software Foundation. + * + * NO WARRANTY + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGES. + */ + +#include "acpidump.h" +#include "acapps.h" + +/****************************************************************************** + * + * FUNCTION: ap_open_output_file + * + * PARAMETERS: pathname - Output filename + * + * RETURN: Open file handle + * + * DESCRIPTION: Open a text output file for acpidump. Checks if file already + * exists. + * + ******************************************************************************/ + +int ap_open_output_file(char *pathname) +{ + struct stat stat_info; + FILE *file; + + /* If file exists, prompt for overwrite */ + + if (!stat(pathname, &stat_info)) { + fprintf(stderr, + "Target path already exists, overwrite? [y|n] "); + + if (getchar() != 'y') { + return (-1); + } + } + + /* Point stdout to the file */ + + file = freopen(pathname, "w", stdout); + if (!file) { + perror("Could not open output file"); + return (-1); + } + + /* Save the file and path */ + + gbl_output_file = file; + gbl_output_filename = pathname; + return (0); +} + +/****************************************************************************** + * + * FUNCTION: ap_write_to_binary_file + * + * PARAMETERS: table - ACPI table to be written + * instance - ACPI table instance no. to be written + * + * RETURN: Status + * + * DESCRIPTION: Write an ACPI table to a binary file. Builds the output + * filename from the table signature. + * + ******************************************************************************/ + +int ap_write_to_binary_file(struct acpi_table_header *table, u32 instance) +{ + char filename[ACPI_NAME_SIZE + 16]; + char instance_str[16]; + FILE *file; + size_t actual; + u32 table_length; + + /* Obtain table length */ + + table_length = ap_get_table_length(table); + + /* Construct lower-case filename from the table local signature */ + + if (ACPI_VALIDATE_RSDP_SIG(table->signature)) { + ACPI_MOVE_NAME(filename, ACPI_RSDP_NAME); + } else { + ACPI_MOVE_NAME(filename, table->signature); + } + filename[0] = (char)ACPI_TOLOWER(filename[0]); + filename[1] = (char)ACPI_TOLOWER(filename[1]); + filename[2] = (char)ACPI_TOLOWER(filename[2]); + filename[3] = (char)ACPI_TOLOWER(filename[3]); + filename[ACPI_NAME_SIZE] = 0; + + /* Handle multiple SSDts - create different filenames for each */ + + if (instance > 0) { + sprintf(instance_str, "%u", instance); + strcat(filename, instance_str); + } + + strcat(filename, ACPI_TABLE_FILE_SUFFIX); + + if (gbl_verbose_mode) { + fprintf(stderr, + "Writing [%4.4s] to binary file: %s 0x%X (%u) bytes\n", + table->signature, filename, table->length, + table->length); + } + + /* Open the file and dump the entire table in binary mode */ + + file = fopen(filename, "wb"); + if (!file) { + perror("Could not open output file"); + return (-1); + } + + actual = fwrite(table, 1, table_length, file); + if (actual != table_length) { + perror("Error writing binary output file"); + fclose(file); + return (-1); + } + + fclose(file); + return (0); +} + +/****************************************************************************** + * + * FUNCTION: ap_get_table_from_file + * + * PARAMETERS: pathname - File containing the binary ACPI table + * out_file_size - Where the file size is returned + * + * RETURN: Buffer containing the ACPI table. NULL on error. + * + * DESCRIPTION: Open a file and read it entirely into a new buffer + * + ******************************************************************************/ + +struct acpi_table_header *ap_get_table_from_file(char *pathname, + u32 *out_file_size) +{ + struct acpi_table_header *buffer = NULL; + FILE *file; + u32 file_size; + size_t actual; + + /* Must use binary mode */ + + file = fopen(pathname, "rb"); + if (!file) { + perror("Could not open input file"); + return (NULL); + } + + /* Need file size to allocate a buffer */ + + file_size = cm_get_file_size(file); + if (file_size == ACPI_UINT32_MAX) { + fprintf(stderr, + "Could not get input file size: %s\n", pathname); + goto cleanup; + } + + /* Allocate a buffer for the entire file */ + + buffer = calloc(1, file_size); + if (!buffer) { + fprintf(stderr, + "Could not allocate file buffer of size: %u\n", + file_size); + goto cleanup; + } + + /* Read the entire file */ + + actual = fread(buffer, 1, file_size, file); + if (actual != file_size) { + fprintf(stderr, "Could not read input file: %s\n", pathname); + free(buffer); + buffer = NULL; + goto cleanup; + } + + *out_file_size = file_size; + +cleanup: + fclose(file); + return (buffer); +} diff --git a/tools/power/acpi/tools/acpidump/apmain.c b/tools/power/acpi/tools/acpidump/apmain.c new file mode 100644 index 00000000000..51e8d638db1 --- /dev/null +++ b/tools/power/acpi/tools/acpidump/apmain.c @@ -0,0 +1,351 @@ +/****************************************************************************** + * + * Module Name: apmain - Main module for the acpidump utility + * + *****************************************************************************/ + +/* + * Copyright (C) 2000 - 2014, Intel Corp. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions, and the following disclaimer, + * without modification. + * 2. Redistributions in binary form must reproduce at minimum a disclaimer + * substantially similar to the "NO WARRANTY" disclaimer below + * ("Disclaimer") and any redistribution must be conditioned upon + * including a substantially similar Disclaimer requirement for further + * binary redistribution. + * 3. Neither the names of the above-listed copyright holders nor the names + * of any contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * Alternatively, this software may be distributed under the terms of the + * GNU General Public License ("GPL") version 2 as published by the Free + * Software Foundation. + * + * NO WARRANTY + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGES. + */ + +#define _DECLARE_GLOBALS +#include "acpidump.h" +#include "acapps.h" + +/* + * acpidump - A portable utility for obtaining system ACPI tables and dumping + * them in an ASCII hex format suitable for binary extraction via acpixtract. + * + * Obtaining the system ACPI tables is an OS-specific operation. + * + * This utility can be ported to any host operating system by providing a + * module containing system-specific versions of these interfaces: + * + * acpi_os_get_table_by_address + * acpi_os_get_table_by_index + * acpi_os_get_table_by_name + * + * See the ACPICA Reference Guide for the exact definitions of these + * interfaces. Also, see these ACPICA source code modules for example + * implementations: + * + * source/os_specific/service_layers/oswintbl.c + * source/os_specific/service_layers/oslinuxtbl.c + */ + +/* Local prototypes */ + +static void ap_display_usage(void); + +static int ap_do_options(int argc, char **argv); + +static void ap_insert_action(char *argument, u32 to_be_done); + +/* Table for deferred actions from command line options */ + +struct ap_dump_action action_table[AP_MAX_ACTIONS]; +u32 current_action = 0; + +#define AP_UTILITY_NAME "ACPI Binary Table Dump Utility" +#define AP_SUPPORTED_OPTIONS "?a:bcf:hn:o:r:svxz" + +/****************************************************************************** + * + * FUNCTION: ap_display_usage + * + * DESCRIPTION: Usage message for the acpi_dump utility + * + ******************************************************************************/ + +static void ap_display_usage(void) +{ + + ACPI_USAGE_HEADER("acpidump [options]"); + + ACPI_OPTION("-b", "Dump tables to binary files"); + ACPI_OPTION("-c", "Dump customized tables"); + ACPI_OPTION("-h -?", "This help message"); + ACPI_OPTION("-o <File>", "Redirect output to file"); + ACPI_OPTION("-r <Address>", "Dump tables from specified RSDP"); + ACPI_OPTION("-s", "Print table summaries only"); + ACPI_OPTION("-v", "Display version information"); + ACPI_OPTION("-z", "Verbose mode"); + + printf("\nTable Options:\n"); + + ACPI_OPTION("-a <Address>", "Get table via a physical address"); + ACPI_OPTION("-f <BinaryFile>", "Get table via a binary file"); + ACPI_OPTION("-n <Signature>", "Get table via a name/signature"); + ACPI_OPTION("-x", "Do not use but dump XSDT"); + ACPI_OPTION("-x -x", "Do not use or dump XSDT"); + + printf("\n" + "Invocation without parameters dumps all available tables\n" + "Multiple mixed instances of -a, -f, and -n are supported\n\n"); +} + +/****************************************************************************** + * + * FUNCTION: ap_insert_action + * + * PARAMETERS: argument - Pointer to the argument for this action + * to_be_done - What to do to process this action + * + * RETURN: None. Exits program if action table becomes full. + * + * DESCRIPTION: Add an action item to the action table + * + ******************************************************************************/ + +static void ap_insert_action(char *argument, u32 to_be_done) +{ + + /* Insert action and check for table overflow */ + + action_table[current_action].argument = argument; + action_table[current_action].to_be_done = to_be_done; + + current_action++; + if (current_action > AP_MAX_ACTIONS) { + fprintf(stderr, "Too many table options (max %u)\n", + AP_MAX_ACTIONS); + exit(-1); + } +} + +/****************************************************************************** + * + * FUNCTION: ap_do_options + * + * PARAMETERS: argc/argv - Standard argc/argv + * + * RETURN: Status + * + * DESCRIPTION: Command line option processing. The main actions for getting + * and dumping tables are deferred via the action table. + * + *****************************************************************************/ + +static int ap_do_options(int argc, char **argv) +{ + int j; + acpi_status status; + + /* Command line options */ + + while ((j = acpi_getopt(argc, argv, AP_SUPPORTED_OPTIONS)) != EOF) + switch (j) { + /* + * Global options + */ + case 'b': /* Dump all input tables to binary files */ + + gbl_binary_mode = TRUE; + continue; + + case 'c': /* Dump customized tables */ + + gbl_dump_customized_tables = TRUE; + continue; + + case 'h': + case '?': + + ap_display_usage(); + exit(0); + + case 'o': /* Redirect output to a single file */ + + if (ap_open_output_file(acpi_gbl_optarg)) { + exit(-1); + } + continue; + + case 'r': /* Dump tables from specified RSDP */ + + status = + acpi_ut_strtoul64(acpi_gbl_optarg, 0, + &gbl_rsdp_base); + if (ACPI_FAILURE(status)) { + fprintf(stderr, + "%s: Could not convert to a physical address\n", + acpi_gbl_optarg); + exit(-1); + } + continue; + + case 's': /* Print table summaries only */ + + gbl_summary_mode = TRUE; + continue; + + case 'x': /* Do not use XSDT */ + + if (!acpi_gbl_do_not_use_xsdt) { + acpi_gbl_do_not_use_xsdt = TRUE; + } else { + gbl_do_not_dump_xsdt = TRUE; + } + continue; + + case 'v': /* Revision/version */ + + printf(ACPI_COMMON_SIGNON(AP_UTILITY_NAME)); + exit(0); + + case 'z': /* Verbose mode */ + + gbl_verbose_mode = TRUE; + fprintf(stderr, ACPI_COMMON_SIGNON(AP_UTILITY_NAME)); + continue; + + /* + * Table options + */ + case 'a': /* Get table by physical address */ + + ap_insert_action(acpi_gbl_optarg, + AP_DUMP_TABLE_BY_ADDRESS); + break; + + case 'f': /* Get table from a file */ + + ap_insert_action(acpi_gbl_optarg, + AP_DUMP_TABLE_BY_FILE); + break; + + case 'n': /* Get table by input name (signature) */ + + ap_insert_action(acpi_gbl_optarg, + AP_DUMP_TABLE_BY_NAME); + break; + + default: + + ap_display_usage(); + exit(-1); + } + + /* If there are no actions, this means "get/dump all tables" */ + + if (current_action == 0) { + ap_insert_action(NULL, AP_DUMP_ALL_TABLES); + } + + return (0); +} + +/****************************************************************************** + * + * FUNCTION: main + * + * PARAMETERS: argc/argv - Standard argc/argv + * + * RETURN: Status + * + * DESCRIPTION: C main function for acpidump utility + * + ******************************************************************************/ + +int ACPI_SYSTEM_XFACE main(int argc, char *argv[]) +{ + int status = 0; + struct ap_dump_action *action; + u32 file_size; + u32 i; + + ACPI_DEBUG_INITIALIZE(); /* For debug version only */ + + /* Process command line options */ + + if (ap_do_options(argc, argv)) { + return (-1); + } + + /* Get/dump ACPI table(s) as requested */ + + for (i = 0; i < current_action; i++) { + action = &action_table[i]; + switch (action->to_be_done) { + case AP_DUMP_ALL_TABLES: + + status = ap_dump_all_tables(); + break; + + case AP_DUMP_TABLE_BY_ADDRESS: + + status = ap_dump_table_by_address(action->argument); + break; + + case AP_DUMP_TABLE_BY_NAME: + + status = ap_dump_table_by_name(action->argument); + break; + + case AP_DUMP_TABLE_BY_FILE: + + status = ap_dump_table_from_file(action->argument); + break; + + default: + + fprintf(stderr, + "Internal error, invalid action: 0x%X\n", + action->to_be_done); + return (-1); + } + + if (status) { + return (status); + } + } + + if (gbl_output_file) { + if (gbl_verbose_mode) { + + /* Summary for the output file */ + + file_size = cm_get_file_size(gbl_output_file); + fprintf(stderr, + "Output file %s contains 0x%X (%u) bytes\n\n", + gbl_output_filename, file_size, file_size); + } + + fclose(gbl_output_file); + } + + return (status); +} diff --git a/tools/power/acpi/tools/ec/Makefile b/tools/power/acpi/tools/ec/Makefile new file mode 100644 index 00000000000..b7b0b929bd3 --- /dev/null +++ b/tools/power/acpi/tools/ec/Makefile @@ -0,0 +1,22 @@ +ec_access: ec_access.o + $(ECHO) " LD " $@ + $(QUIET) $(LD) $(CFLAGS) $(LDFLAGS) $< -o $@ + $(QUIET) $(STRIPCMD) $@ + +%.o: %.c + $(ECHO) " CC " $@ + $(QUIET) $(CC) -c $(CFLAGS) -o $@ $< + +all: ec_access + +install: + $(INSTALL) -d $(DESTDIR)${sbindir} + $(INSTALL_PROGRAM) ec_access $(DESTDIR)${sbindir} + +uninstall: + - rm -f $(DESTDIR)${sbindir}/ec_access + +clean: + -rm -f $(OUTPUT)ec_access + +.PHONY: all install uninstall diff --git a/tools/power/acpi/tools/ec/ec_access.c b/tools/power/acpi/tools/ec/ec_access.c new file mode 100644 index 00000000000..6b8aaed44f2 --- /dev/null +++ b/tools/power/acpi/tools/ec/ec_access.c @@ -0,0 +1,238 @@ +/* + * ec_access.c + * + * Copyright (C) 2010 SUSE Linux Products GmbH + * Author: + * Thomas Renninger <trenn@suse.de> + * + * This work is licensed under the terms of the GNU GPL, version 2. + */ + +#include <fcntl.h> +#include <err.h> +#include <stdio.h> +#include <stdlib.h> +#include <libgen.h> +#include <unistd.h> +#include <getopt.h> +#include <stdint.h> +#include <sys/types.h> +#include <sys/stat.h> + + +#define EC_SPACE_SIZE 256 +#define SYSFS_PATH "/sys/kernel/debug/ec/ec0/io" + +/* TBD/Enhancements: + - Provide param for accessing different ECs (not supported by kernel yet) +*/ + +static int read_mode = -1; +static int sleep_time; +static int write_byte_offset = -1; +static int read_byte_offset = -1; +static uint8_t write_value = -1; + +void usage(char progname[], int exit_status) +{ + printf("Usage:\n"); + printf("1) %s -r [-s sleep]\n", basename(progname)); + printf("2) %s -b byte_offset\n", basename(progname)); + printf("3) %s -w byte_offset -v value\n\n", basename(progname)); + + puts("\t-r [-s sleep] : Dump EC registers"); + puts("\t If sleep is given, sleep x seconds,"); + puts("\t re-read EC registers and show changes"); + puts("\t-b offset : Read value at byte_offset (in hex)"); + puts("\t-w offset -v value : Write value at byte_offset"); + puts("\t-h : Print this help\n\n"); + puts("Offsets and values are in hexadecimal number sytem."); + puts("The offset and value must be between 0 and 0xff."); + exit(exit_status); +} + +void parse_opts(int argc, char *argv[]) +{ + int c; + + while ((c = getopt(argc, argv, "rs:b:w:v:h")) != -1) { + + switch (c) { + case 'r': + if (read_mode != -1) + usage(argv[0], EXIT_FAILURE); + read_mode = 1; + break; + case 's': + if (read_mode != -1 && read_mode != 1) + usage(argv[0], EXIT_FAILURE); + + sleep_time = atoi(optarg); + if (sleep_time <= 0) { + sleep_time = 0; + usage(argv[0], EXIT_FAILURE); + printf("Bad sleep time: %s\n", optarg); + } + break; + case 'b': + if (read_mode != -1) + usage(argv[0], EXIT_FAILURE); + read_mode = 1; + read_byte_offset = strtoul(optarg, NULL, 16); + break; + case 'w': + if (read_mode != -1) + usage(argv[0], EXIT_FAILURE); + read_mode = 0; + write_byte_offset = strtoul(optarg, NULL, 16); + break; + case 'v': + write_value = strtoul(optarg, NULL, 16); + break; + case 'h': + usage(argv[0], EXIT_SUCCESS); + default: + fprintf(stderr, "Unknown option!\n"); + usage(argv[0], EXIT_FAILURE); + } + } + if (read_mode == 0) { + if (write_byte_offset < 0 || + write_byte_offset >= EC_SPACE_SIZE) { + fprintf(stderr, "Wrong byte offset 0x%.2x, valid: " + "[0-0x%.2x]\n", + write_byte_offset, EC_SPACE_SIZE - 1); + usage(argv[0], EXIT_FAILURE); + } + if (write_value < 0 || + write_value >= 255) { + fprintf(stderr, "Wrong byte offset 0x%.2x, valid:" + "[0-0xff]\n", write_byte_offset); + usage(argv[0], EXIT_FAILURE); + } + } + if (read_mode == 1 && read_byte_offset != -1) { + if (read_byte_offset < -1 || + read_byte_offset >= EC_SPACE_SIZE) { + fprintf(stderr, "Wrong byte offset 0x%.2x, valid: " + "[0-0x%.2x]\n", + read_byte_offset, EC_SPACE_SIZE - 1); + usage(argv[0], EXIT_FAILURE); + } + } + /* Add additional parameter checks here */ +} + +void dump_ec(int fd) +{ + char buf[EC_SPACE_SIZE]; + char buf2[EC_SPACE_SIZE]; + int byte_off, bytes_read; + + bytes_read = read(fd, buf, EC_SPACE_SIZE); + + if (bytes_read == -1) + err(EXIT_FAILURE, "Could not read from %s\n", SYSFS_PATH); + + if (bytes_read != EC_SPACE_SIZE) + fprintf(stderr, "Could only read %d bytes\n", bytes_read); + + printf(" 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F"); + for (byte_off = 0; byte_off < bytes_read; byte_off++) { + if ((byte_off % 16) == 0) + printf("\n%.2X: ", byte_off); + printf(" %.2x ", (uint8_t)buf[byte_off]); + } + printf("\n"); + + if (!sleep_time) + return; + + printf("\n"); + lseek(fd, 0, SEEK_SET); + sleep(sleep_time); + + bytes_read = read(fd, buf2, EC_SPACE_SIZE); + + if (bytes_read == -1) + err(EXIT_FAILURE, "Could not read from %s\n", SYSFS_PATH); + + if (bytes_read != EC_SPACE_SIZE) + fprintf(stderr, "Could only read %d bytes\n", bytes_read); + + printf(" 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F"); + for (byte_off = 0; byte_off < bytes_read; byte_off++) { + if ((byte_off % 16) == 0) + printf("\n%.2X: ", byte_off); + + if (buf[byte_off] == buf2[byte_off]) + printf(" %.2x ", (uint8_t)buf2[byte_off]); + else + printf("*%.2x ", (uint8_t)buf2[byte_off]); + } + printf("\n"); +} + +void read_ec_val(int fd, int byte_offset) +{ + uint8_t buf; + int error; + + error = lseek(fd, byte_offset, SEEK_SET); + if (error != byte_offset) + err(EXIT_FAILURE, "Cannot set offset to 0x%.2x", byte_offset); + + error = read(fd, &buf, 1); + if (error != 1) + err(EXIT_FAILURE, "Could not read byte 0x%.2x from %s\n", + byte_offset, SYSFS_PATH); + printf("0x%.2x\n", buf); + return; +} + +void write_ec_val(int fd, int byte_offset, uint8_t value) +{ + int error; + + error = lseek(fd, byte_offset, SEEK_SET); + if (error != byte_offset) + err(EXIT_FAILURE, "Cannot set offset to 0x%.2x", byte_offset); + + error = write(fd, &value, 1); + if (error != 1) + err(EXIT_FAILURE, "Cannot write value 0x%.2x to offset 0x%.2x", + value, byte_offset); +} + +int main(int argc, char *argv[]) +{ + int file_mode = O_RDONLY; + int fd; + + parse_opts(argc, argv); + + if (read_mode == 0) + file_mode = O_WRONLY; + else if (read_mode == 1) + file_mode = O_RDONLY; + else + usage(argv[0], EXIT_FAILURE); + + fd = open(SYSFS_PATH, file_mode); + if (fd == -1) + err(EXIT_FAILURE, "%s", SYSFS_PATH); + + if (read_mode) + if (read_byte_offset == -1) + dump_ec(fd); + else if (read_byte_offset < 0 || + read_byte_offset >= EC_SPACE_SIZE) + usage(argv[0], EXIT_FAILURE); + else + read_ec_val(fd, read_byte_offset); + else + write_ec_val(fd, write_byte_offset, write_value); + close(fd); + + exit(EXIT_SUCCESS); +} diff --git a/tools/power/cpupower/.gitignore b/tools/power/cpupower/.gitignore index 8a83dd2ffc1..d42073f1260 100644 --- a/tools/power/cpupower/.gitignore +++ b/tools/power/cpupower/.gitignore @@ -20,3 +20,10 @@ utils/cpufreq-set.o utils/cpufreq-aperf.o cpupower bench/cpufreq-bench +debug/kernel/Module.symvers +debug/i386/centrino-decode +debug/i386/dump_psb +debug/i386/intel_gsic +debug/i386/powernow-k8-decode +debug/x86_64/centrino-decode +debug/x86_64/powernow-k8-decode diff --git a/tools/power/cpupower/Makefile b/tools/power/cpupower/Makefile index e8a03aceceb..2e2ba2efa0d 100644 --- a/tools/power/cpupower/Makefile +++ b/tools/power/cpupower/Makefile @@ -19,6 +19,16 @@ # along with this program; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # +OUTPUT=./ +ifeq ("$(origin O)", "command line") + OUTPUT := $(O)/ +endif + +ifneq ($(OUTPUT),) +# check that the output directory actually exists +OUTDIR := $(shell cd $(OUTPUT) && /bin/pwd) +$(if $(OUTDIR),, $(error output directory "$(OUTPUT)" does not exist)) +endif # --- CONFIGURATION BEGIN --- @@ -52,7 +62,7 @@ LIB_MAJ= 0.0.0 LIB_MIN= 0 PACKAGE = cpupower -PACKAGE_BUGREPORT = cpufreq@vger.kernel.org +PACKAGE_BUGREPORT = linux-pm@vger.kernel.org LANGUAGES = de fr it cs pt @@ -87,6 +97,7 @@ AR = $(CROSS)ar STRIP = $(CROSS)strip RANLIB = $(CROSS)ranlib HOSTCC = gcc +MKDIR = mkdir # Now we set up the build system @@ -95,12 +106,12 @@ HOSTCC = gcc # set up PWD so that older versions of make will work with our build. PWD = $(shell pwd) -GMO_FILES = ${shell for HLANG in ${LANGUAGES}; do echo po/$$HLANG.gmo; done;} +GMO_FILES = ${shell for HLANG in ${LANGUAGES}; do echo $(OUTPUT)po/$$HLANG.gmo; done;} export CROSS CC AR STRIP RANLIB CFLAGS LDFLAGS LIB_OBJS # check if compiler option is supported -cc-supports = ${shell if $(CC) ${1} -S -o /dev/null -xc /dev/null > /dev/null 2>&1; then echo "$(1)"; fi;} +cc-supports = ${shell if $(CC) ${1} -S -o /dev/null -x c /dev/null > /dev/null 2>&1; then echo "$(1)"; fi;} # use '-Os' optimization if available, else use -O2 OPTIMIZATION := $(call cc-supports,-Os,-O2) @@ -117,20 +128,25 @@ UTIL_OBJS = utils/helpers/amd.o utils/helpers/topology.o utils/helpers/msr.o \ utils/helpers/sysfs.o utils/helpers/misc.o utils/helpers/cpuid.o \ utils/helpers/pci.o utils/helpers/bitmask.o \ utils/idle_monitor/nhm_idle.o utils/idle_monitor/snb_idle.o \ + utils/idle_monitor/hsw_ext_idle.o \ utils/idle_monitor/amd_fam14h_idle.o utils/idle_monitor/cpuidle_sysfs.o \ utils/idle_monitor/mperf_monitor.o utils/idle_monitor/cpupower-monitor.o \ utils/cpupower.o utils/cpufreq-info.o utils/cpufreq-set.o \ - utils/cpupower-set.o utils/cpupower-info.o utils/cpuidle-info.o + utils/cpupower-set.o utils/cpupower-info.o utils/cpuidle-info.o \ + utils/cpuidle-set.o + +UTIL_SRC := $(UTIL_OBJS:.o=.c) + +UTIL_OBJS := $(addprefix $(OUTPUT),$(UTIL_OBJS)) UTIL_HEADERS = utils/helpers/helpers.h utils/idle_monitor/cpupower-monitor.h \ utils/helpers/bitmask.h \ utils/idle_monitor/idle_monitors.h utils/idle_monitor/idle_monitors.def -UTIL_SRC := $(UTIL_OBJS:.o=.c) - LIB_HEADERS = lib/cpufreq.h lib/sysfs.h LIB_SRC = lib/cpufreq.c lib/sysfs.c LIB_OBJS = lib/cpufreq.o lib/sysfs.o +LIB_OBJS := $(addprefix $(OUTPUT),$(LIB_OBJS)) CFLAGS += -pipe @@ -168,88 +184,98 @@ endif # the actual make rules -all: libcpupower cpupower $(COMPILE_NLS) $(COMPILE_BENCH) +all: libcpupower $(OUTPUT)cpupower $(COMPILE_NLS) $(COMPILE_BENCH) -lib/%.o: $(LIB_SRC) $(LIB_HEADERS) +$(OUTPUT)lib/%.o: $(LIB_SRC) $(LIB_HEADERS) $(ECHO) " CC " $@ $(QUIET) $(CC) $(CFLAGS) -fPIC -o $@ -c lib/$*.c -libcpupower.so.$(LIB_MAJ): $(LIB_OBJS) +$(OUTPUT)libcpupower.so.$(LIB_MAJ): $(LIB_OBJS) $(ECHO) " LD " $@ $(QUIET) $(CC) -shared $(CFLAGS) $(LDFLAGS) -o $@ \ -Wl,-soname,libcpupower.so.$(LIB_MIN) $(LIB_OBJS) - @ln -sf $@ libcpupower.so - @ln -sf $@ libcpupower.so.$(LIB_MIN) + @ln -sf $(@F) $(OUTPUT)libcpupower.so + @ln -sf $(@F) $(OUTPUT)libcpupower.so.$(LIB_MIN) -libcpupower: libcpupower.so.$(LIB_MAJ) +libcpupower: $(OUTPUT)libcpupower.so.$(LIB_MAJ) # Let all .o files depend on its .c file and all headers # Might be worth to put this into utils/Makefile at some point of time $(UTIL_OBJS): $(UTIL_HEADERS) -.c.o: +$(OUTPUT)%.o: %.c $(ECHO) " CC " $@ $(QUIET) $(CC) $(CFLAGS) -I./lib -I ./utils -o $@ -c $*.c -cpupower: $(UTIL_OBJS) libcpupower.so.$(LIB_MAJ) +$(OUTPUT)cpupower: $(UTIL_OBJS) $(OUTPUT)libcpupower.so.$(LIB_MAJ) $(ECHO) " CC " $@ - $(QUIET) $(CC) $(CFLAGS) $(LDFLAGS) -lcpupower -lrt -lpci -L. -o $@ $(UTIL_OBJS) + $(QUIET) $(CC) $(CFLAGS) $(LDFLAGS) $(UTIL_OBJS) -lcpupower -lrt -lpci -L$(OUTPUT) -o $@ $(QUIET) $(STRIPCMD) $@ -po/$(PACKAGE).pot: $(UTIL_SRC) +$(OUTPUT)po/$(PACKAGE).pot: $(UTIL_SRC) $(ECHO) " GETTEXT " $@ $(QUIET) xgettext --default-domain=$(PACKAGE) --add-comments \ - --keyword=_ --keyword=N_ $(UTIL_SRC) && \ - test -f $(PACKAGE).po && \ - mv -f $(PACKAGE).po po/$(PACKAGE).pot + --keyword=_ --keyword=N_ $(UTIL_SRC) -p $(@D) -o $(@F) -po/%.gmo: po/%.po +$(OUTPUT)po/%.gmo: po/%.po $(ECHO) " MSGFMT " $@ $(QUIET) msgfmt -o $@ po/$*.po create-gmo: ${GMO_FILES} -update-po: po/$(PACKAGE).pot +update-po: $(OUTPUT)po/$(PACKAGE).pot $(ECHO) " MSGMRG " $@ $(QUIET) @for HLANG in $(LANGUAGES); do \ echo -n "Updating $$HLANG "; \ - if msgmerge po/$$HLANG.po po/$(PACKAGE).pot -o \ - po/$$HLANG.new.po; then \ - mv -f po/$$HLANG.new.po po/$$HLANG.po; \ + if msgmerge po/$$HLANG.po $< -o \ + $(OUTPUT)po/$$HLANG.new.po; then \ + mv -f $(OUTPUT)po/$$HLANG.new.po $(OUTPUT)po/$$HLANG.po; \ else \ echo "msgmerge for $$HLANG failed!"; \ - rm -f po/$$HLANG.new.po; \ + rm -f $(OUTPUT)po/$$HLANG.new.po; \ fi; \ done; -compile-bench: libcpupower.so.$(LIB_MAJ) - @V=$(V) confdir=$(confdir) $(MAKE) -C bench +compile-bench: $(OUTPUT)libcpupower.so.$(LIB_MAJ) + @V=$(V) confdir=$(confdir) $(MAKE) -C bench O=$(OUTPUT) + +# we compile into subdirectories. if the target directory is not the +# source directory, they might not exists. So we depend the various +# files onto their directories. +DIRECTORY_DEPS = $(LIB_OBJS) $(UTIL_OBJS) $(GMO_FILES) +$(DIRECTORY_DEPS): | $(sort $(dir $(DIRECTORY_DEPS))) + +# In the second step, we make a rule to actually create these directories +$(sort $(dir $(DIRECTORY_DEPS))): + $(ECHO) " MKDIR " $@ + $(QUIET) $(MKDIR) -p $@ 2>/dev/null clean: - -find . \( -not -type d \) -and \( -name '*~' -o -name '*.[oas]' \) -type f -print \ + -find $(OUTPUT) \( -not -type d \) -and \( -name '*~' -o -name '*.[oas]' \) -type f -print \ | xargs rm -f - -rm -f $(UTIL_BINS) - -rm -f $(IDLE_OBJS) - -rm -f cpupower - -rm -f libcpupower.so* - -rm -rf po/*.gmo po/*.pot - $(MAKE) -C bench clean + -rm -f $(OUTPUT)cpupower + -rm -f $(OUTPUT)libcpupower.so* + -rm -rf $(OUTPUT)po/*.gmo + -rm -rf $(OUTPUT)po/*.pot + $(MAKE) -C bench O=$(OUTPUT) clean install-lib: $(INSTALL) -d $(DESTDIR)${libdir} - $(CP) libcpupower.so* $(DESTDIR)${libdir}/ + $(CP) $(OUTPUT)libcpupower.so* $(DESTDIR)${libdir}/ $(INSTALL) -d $(DESTDIR)${includedir} $(INSTALL_DATA) lib/cpufreq.h $(DESTDIR)${includedir}/cpufreq.h install-tools: $(INSTALL) -d $(DESTDIR)${bindir} - $(INSTALL_PROGRAM) cpupower $(DESTDIR)${bindir} + $(INSTALL_PROGRAM) $(OUTPUT)cpupower $(DESTDIR)${bindir} install-man: $(INSTALL_DATA) -D man/cpupower.1 $(DESTDIR)${mandir}/man1/cpupower.1 $(INSTALL_DATA) -D man/cpupower-frequency-set.1 $(DESTDIR)${mandir}/man1/cpupower-frequency-set.1 $(INSTALL_DATA) -D man/cpupower-frequency-info.1 $(DESTDIR)${mandir}/man1/cpupower-frequency-info.1 + $(INSTALL_DATA) -D man/cpupower-idle-set.1 $(DESTDIR)${mandir}/man1/cpupower-idle-set.1 + $(INSTALL_DATA) -D man/cpupower-idle-info.1 $(DESTDIR)${mandir}/man1/cpupower-idle-info.1 $(INSTALL_DATA) -D man/cpupower-set.1 $(DESTDIR)${mandir}/man1/cpupower-set.1 $(INSTALL_DATA) -D man/cpupower-info.1 $(DESTDIR)${mandir}/man1/cpupower-info.1 $(INSTALL_DATA) -D man/cpupower-monitor.1 $(DESTDIR)${mandir}/man1/cpupower-monitor.1 @@ -257,13 +283,13 @@ install-man: install-gmo: $(INSTALL) -d $(DESTDIR)${localedir} for HLANG in $(LANGUAGES); do \ - echo '$(INSTALL_DATA) -D po/$$HLANG.gmo $(DESTDIR)${localedir}/$$HLANG/LC_MESSAGES/cpupower.mo'; \ - $(INSTALL_DATA) -D po/$$HLANG.gmo $(DESTDIR)${localedir}/$$HLANG/LC_MESSAGES/cpupower.mo; \ + echo '$(INSTALL_DATA) -D $(OUTPUT)po/$$HLANG.gmo $(DESTDIR)${localedir}/$$HLANG/LC_MESSAGES/cpupower.mo'; \ + $(INSTALL_DATA) -D $(OUTPUT)po/$$HLANG.gmo $(DESTDIR)${localedir}/$$HLANG/LC_MESSAGES/cpupower.mo; \ done; install-bench: @#DESTDIR must be set from outside to survive - @sbindir=$(sbindir) bindir=$(bindir) docdir=$(docdir) confdir=$(confdir) $(MAKE) -C bench install + @sbindir=$(sbindir) bindir=$(bindir) docdir=$(docdir) confdir=$(confdir) $(MAKE) -C bench O=$(OUTPUT) install install: all install-lib install-tools install-man $(INSTALL_NLS) $(INSTALL_BENCH) @@ -271,8 +297,12 @@ uninstall: - rm -f $(DESTDIR)${libdir}/libcpupower.* - rm -f $(DESTDIR)${includedir}/cpufreq.h - rm -f $(DESTDIR)${bindir}/utils/cpupower - - rm -f $(DESTDIR)${mandir}/man1/cpufreq-set.1 - - rm -f $(DESTDIR)${mandir}/man1/cpufreq-info.1 + - rm -f $(DESTDIR)${mandir}/man1/cpupower.1 + - rm -f $(DESTDIR)${mandir}/man1/cpupower-frequency-set.1 + - rm -f $(DESTDIR)${mandir}/man1/cpupower-frequency-info.1 + - rm -f $(DESTDIR)${mandir}/man1/cpupower-set.1 + - rm -f $(DESTDIR)${mandir}/man1/cpupower-info.1 + - rm -f $(DESTDIR)${mandir}/man1/cpupower-monitor.1 - for HLANG in $(LANGUAGES); do \ rm -f $(DESTDIR)${localedir}/$$HLANG/LC_MESSAGES/cpupower.mo; \ done; diff --git a/tools/power/cpupower/README b/tools/power/cpupower/README index fd9d4c0d668..1c68f47663b 100644 --- a/tools/power/cpupower/README +++ b/tools/power/cpupower/README @@ -1,6 +1,4 @@ -The cpufrequtils package (homepage: -http://www.kernel.org/pub/linux/utils/kernel/cpufreq/cpufrequtils.html ) -consists of the following elements: +The cpupower package consists of the following elements: requirements ------------ @@ -11,10 +9,10 @@ providing cpuid.h is needed. For both it's not explicitly checked for (yet). -libcpufreq +libcpupower ---------- -"libcpufreq" is a library which offers a unified access method for userspace +"libcpupower" is a library which offers a unified access method for userspace tools and programs to the cpufreq core and drivers in the Linux kernel. This allows for code reduction in userspace tools, a clean implementation of the interaction to the cpufreq core, and support for both the sysfs and proc @@ -28,22 +26,22 @@ make su make install -should suffice on most systems. It builds default libcpufreq, -cpufreq-set and cpufreq-info files and installs them in /usr/lib and -/usr/bin, respectively. If you want to set up the paths differently and/or -want to configure the package to your specific needs, you need to open -"Makefile" with an editor of your choice and edit the block marked -CONFIGURATION. +should suffice on most systems. It builds libcpupower to put in +/usr/lib; cpupower, cpufreq-bench_plot.sh to put in /usr/bin; and +cpufreq-bench to put in /usr/sbin. If you want to set up the paths +differently and/or want to configure the package to your specific +needs, you need to open "Makefile" with an editor of your choice and +edit the block marked CONFIGURATION. THANKS ------ Many thanks to Mattia Dongili who wrote the autotoolization and -libtoolization, the manpages and the italian language file for cpufrequtils; +libtoolization, the manpages and the italian language file for cpupower; to Dave Jones for his feedback and his dump_psb tool; to Bruno Ducrot for his powernow-k8-decode and intel_gsic tools as well as the french language file; and to various others commenting on the previous (pre-)releases of -cpufrequtils. +cpupower. Dominik Brodowski diff --git a/tools/power/cpupower/ToDo b/tools/power/cpupower/ToDo index 874b78b586e..6e8b89f282e 100644 --- a/tools/power/cpupower/ToDo +++ b/tools/power/cpupower/ToDo @@ -3,7 +3,6 @@ ToDos sorted by priority: - Use bitmask functions to parse CPU topology more robust (current implementation has issues on AMD) - Try to read out boost states and frequencies on Intel -- Adjust README - Somewhere saw the ability to read power consumption of RAM from HW on Intel SandyBridge -> another monitor? - Add another c1e debug idle monitor diff --git a/tools/power/cpupower/bench/Makefile b/tools/power/cpupower/bench/Makefile index 2b67606fc3e..7ec7021a29c 100644 --- a/tools/power/cpupower/bench/Makefile +++ b/tools/power/cpupower/bench/Makefile @@ -1,29 +1,36 @@ -LIBS = -L../ -lm -lcpupower +OUTPUT := ./ +ifeq ("$(origin O)", "command line") +ifneq ($(O),) + OUTPUT := $(O)/ +endif +endif -OBJS = main.o parse.o system.o benchmark.o +LIBS = -L../ -L$(OUTPUT) -lm -lcpupower + +OBJS = $(OUTPUT)main.o $(OUTPUT)parse.o $(OUTPUT)system.o $(OUTPUT)benchmark.o CFLAGS += -D_GNU_SOURCE -I../lib -DDEFAULT_CONFIG_FILE=\"$(confdir)/cpufreq-bench.conf\" -%.o : %.c +$(OUTPUT)%.o : %.c $(ECHO) " CC " $@ $(QUIET) $(CC) -c $(CFLAGS) $< -o $@ -cpufreq-bench: $(OBJS) +$(OUTPUT)cpufreq-bench: $(OBJS) $(ECHO) " CC " $@ $(QUIET) $(CC) -o $@ $(CFLAGS) $(OBJS) $(LIBS) -all: cpufreq-bench +all: $(OUTPUT)cpufreq-bench install: mkdir -p $(DESTDIR)/$(sbindir) mkdir -p $(DESTDIR)/$(bindir) mkdir -p $(DESTDIR)/$(docdir) mkdir -p $(DESTDIR)/$(confdir) - install -m 755 cpufreq-bench $(DESTDIR)/$(sbindir)/cpufreq-bench + install -m 755 $(OUTPUT)cpufreq-bench $(DESTDIR)/$(sbindir)/cpufreq-bench install -m 755 cpufreq-bench_plot.sh $(DESTDIR)/$(bindir)/cpufreq-bench_plot.sh install -m 644 README-BENCH $(DESTDIR)/$(docdir)/README-BENCH install -m 755 cpufreq-bench_script.sh $(DESTDIR)/$(docdir)/cpufreq-bench_script.sh install -m 644 example.cfg $(DESTDIR)/$(confdir)/cpufreq-bench.conf clean: - rm -f *.o - rm -f cpufreq-bench + rm -f $(OUTPUT)*.o + rm -f $(OUTPUT)cpufreq-bench diff --git a/tools/power/cpupower/debug/i386/Makefile b/tools/power/cpupower/debug/i386/Makefile index d08cc1ead9b..c05cc0ac80c 100644 --- a/tools/power/cpupower/debug/i386/Makefile +++ b/tools/power/cpupower/debug/i386/Makefile @@ -1,20 +1,41 @@ +OUTPUT=./ +ifeq ("$(origin O)", "command line") + OUTPUT := $(O)/ +endif + +DESTDIR = +bindir = /usr/bin + +INSTALL = /usr/bin/install + + default: all -centrino-decode: centrino-decode.c - $(CC) $(CFLAGS) -o centrino-decode centrino-decode.c +$(OUTPUT)centrino-decode: centrino-decode.c + $(CC) $(CFLAGS) -o $@ centrino-decode.c -dump_psb: dump_psb.c - $(CC) $(CFLAGS) -o dump_psb dump_psb.c +$(OUTPUT)dump_psb: dump_psb.c + $(CC) $(CFLAGS) -o $@ dump_psb.c -intel_gsic: intel_gsic.c - $(CC) $(CFLAGS) -o intel_gsic -llrmi intel_gsic.c +$(OUTPUT)intel_gsic: intel_gsic.c + $(CC) $(CFLAGS) -o $@ -llrmi intel_gsic.c -powernow-k8-decode: powernow-k8-decode.c - $(CC) $(CFLAGS) -o powernow-k8-decode powernow-k8-decode.c +$(OUTPUT)powernow-k8-decode: powernow-k8-decode.c + $(CC) $(CFLAGS) -o $@ powernow-k8-decode.c -all: centrino-decode dump_psb intel_gsic powernow-k8-decode +all: $(OUTPUT)centrino-decode $(OUTPUT)dump_psb $(OUTPUT)intel_gsic $(OUTPUT)powernow-k8-decode clean: - rm -rf centrino-decode dump_psb intel_gsic powernow-k8-decode + rm -rf $(OUTPUT)centrino-decode + rm -rf $(OUTPUT)dump_psb + rm -rf $(OUTPUT)intel_gsic + rm -rf $(OUTPUT)powernow-k8-decode + +install: + $(INSTALL) -d $(DESTDIR)${bindir} + $(INSTALL) $(OUTPUT)centrino-decode $(DESTDIR)${bindir} + $(INSTALL) $(OUTPUT)powernow-k8-decode $(DESTDIR)${bindir} + $(INSTALL) $(OUTPUT)dump_psb $(DESTDIR)${bindir} + $(INSTALL) $(OUTPUT)intel_gsic $(DESTDIR)${bindir} -.PHONY: all default clean +.PHONY: all default clean install diff --git a/tools/power/cpupower/debug/i386/intel_gsic.c b/tools/power/cpupower/debug/i386/intel_gsic.c index 53f5293c9c9..d032c826d42 100644 --- a/tools/power/cpupower/debug/i386/intel_gsic.c +++ b/tools/power/cpupower/debug/i386/intel_gsic.c @@ -66,7 +66,7 @@ int main (void) printf("ecx = 0x%.8x\n", r.ecx); printf("edx = 0x%.8x\n", r.edx); printf("Note also that some BIOS do not support the initial " - "GSIC call, but the newer\nspeeedstep-smi driver may " + "GSIC call, but the newer\nspeedstep-smi driver may " "work.\nFor this, you need to pass some arguments to " "the speedstep-smi driver:\n"); printf("\tsmi_cmd=0x?? smi_port=0x?? smi_sig=1\n"); diff --git a/tools/power/cpupower/debug/kernel/cpufreq-test_tsc.c b/tools/power/cpupower/debug/kernel/cpufreq-test_tsc.c index 66cace601e5..5224ee5b392 100644 --- a/tools/power/cpupower/debug/kernel/cpufreq-test_tsc.c +++ b/tools/power/cpupower/debug/kernel/cpufreq-test_tsc.c @@ -18,19 +18,16 @@ * 5.) if the third value, "diff_pmtmr", changes between 2. and 4., the * TSC-based delay routine on the Linux kernel does not correctly * handle the cpufreq transition. Please report this to - * cpufreq@vger.kernel.org + * linux-pm@vger.kernel.org */ #include <linux/kernel.h> #include <linux/module.h> #include <linux/init.h> #include <linux/delay.h> - +#include <linux/acpi.h> #include <asm/io.h> -#include <acpi/acpi_bus.h> -#include <acpi/acpi_drivers.h> - static int pm_tmr_ioport = 0; /*helper function to safely read acpi pm timesource*/ diff --git a/tools/power/cpupower/debug/x86_64/Makefile b/tools/power/cpupower/debug/x86_64/Makefile index 3326217dd31..1c521452671 100644 --- a/tools/power/cpupower/debug/x86_64/Makefile +++ b/tools/power/cpupower/debug/x86_64/Makefile @@ -1,14 +1,30 @@ +OUTPUT=./ +ifeq ("$(origin O)", "command line") + OUTPUT := $(O)/ +endif + +DESTDIR = +bindir = /usr/bin + +INSTALL = /usr/bin/install + + default: all -centrino-decode: ../i386/centrino-decode.c +$(OUTPUT)centrino-decode: ../i386/centrino-decode.c $(CC) $(CFLAGS) -o $@ $< -powernow-k8-decode: ../i386/powernow-k8-decode.c +$(OUTPUT)powernow-k8-decode: ../i386/powernow-k8-decode.c $(CC) $(CFLAGS) -o $@ $< -all: centrino-decode powernow-k8-decode +all: $(OUTPUT)centrino-decode $(OUTPUT)powernow-k8-decode clean: - rm -rf centrino-decode powernow-k8-decode + rm -rf $(OUTPUT)centrino-decode $(OUTPUT)powernow-k8-decode + +install: + $(INSTALL) -d $(DESTDIR)${bindir} + $(INSTALL) $(OUTPUT)centrino-decode $(DESTDIR)${bindir} + $(INSTALL) $(OUTPUT)powernow-k8-decode $(DESTDIR)${bindir} -.PHONY: all default clean +.PHONY: all default clean install diff --git a/tools/power/cpupower/man/cpupower-frequency-info.1 b/tools/power/cpupower/man/cpupower-frequency-info.1 index bb60a8d1e45..9c85a382e35 100644 --- a/tools/power/cpupower/man/cpupower-frequency-info.1 +++ b/tools/power/cpupower/man/cpupower-frequency-info.1 @@ -1,4 +1,4 @@ -.TH "cpupower-frequency-info" "1" "0.1" "Mattia Dongili" "" +.TH "CPUPOWER\-FREQUENCY\-INFO" "1" "0.1" "" "cpupower Manual" .SH "NAME" .LP cpupower frequency\-info \- Utility to retrieve cpufreq kernel information @@ -50,8 +50,9 @@ Prints out information like provided by the /proc/cpufreq interface in 2.4. and \fB\-m\fR \fB\-\-human\fR human\-readable output for the \-f, \-w, \-s and \-y parameters. .TP -\fB\-h\fR \fB\-\-help\fR -Prints out the help screen. +\fB\-n\fR \fB\-\-no-rounding\fR +Output frequencies and latencies without rounding off values. +.TP .SH "REMARKS" .LP By default only values of core zero are displayed. How to display settings of diff --git a/tools/power/cpupower/man/cpupower-frequency-set.1 b/tools/power/cpupower/man/cpupower-frequency-set.1 index 685f469093a..3eacc8d03d1 100644 --- a/tools/power/cpupower/man/cpupower-frequency-set.1 +++ b/tools/power/cpupower/man/cpupower-frequency-set.1 @@ -1,4 +1,4 @@ -.TH "cpupower-freqency-set" "1" "0.1" "Mattia Dongili" "" +.TH "CPUPOWER\-FREQUENCY\-SET" "1" "0.1" "" "cpupower Manual" .SH "NAME" .LP cpupower frequency\-set \- A small tool which allows to modify cpufreq settings. @@ -26,8 +26,6 @@ specific frequency to be set. Requires userspace governor to be available and lo \fB\-r\fR \fB\-\-related\fR modify all hardware-related CPUs at the same time .TP -\fB\-h\fR \fB\-\-help\fR -Prints out the help screen. .SH "REMARKS" .LP By default values are applied on all cores. How to modify single core diff --git a/tools/power/cpupower/man/cpupower-idle-info.1 b/tools/power/cpupower/man/cpupower-idle-info.1 new file mode 100644 index 00000000000..7b3646adb92 --- /dev/null +++ b/tools/power/cpupower/man/cpupower-idle-info.1 @@ -0,0 +1,91 @@ +.TH "CPUPOWER-IDLE-INFO" "1" "0.1" "" "cpupower Manual" +.SH "NAME" +.LP +cpupower idle\-info \- Utility to retrieve cpu idle kernel information +.SH "SYNTAX" +.LP +cpupower [ \-c cpulist ] idle\-info [\fIoptions\fP] +.SH "DESCRIPTION" +.LP +A tool which prints out per cpu idle information helpful to developers and interested users. +.SH "OPTIONS" +.LP +.TP +\fB\-f\fR \fB\-\-silent\fR +Only print a summary of all available C-states in the system. +.TP +\fB\-e\fR \fB\-\-proc\fR +deprecated. +Prints out idle information in old /proc/acpi/processor/*/power format. This +interface has been removed from the kernel for quite some time, do not let +further code depend on this option, best do not use it. + +.SH IDLE\-INFO DESCRIPTIONS +CPU sleep state statistics and descriptions are retrieved from sysfs files, +exported by the cpuidle kernel subsystem. The kernel only updates these +statistics when it enters or leaves an idle state, therefore on a very idle or +a very busy system, these statistics may not be accurate. They still provide a +good overview about the usage and availability of processor sleep states on +the platform. + +Be aware that the sleep states as exported by the hardware or BIOS and used by +the Linux kernel may not exactly reflect the capabilities of the +processor. This often is the case on the X86 architecture when the acpi_idle +driver is used. It is also possible that the hardware overrules the kernel +requests, due to internal activity monitors or other reasons. +On recent X86 platforms it is often possible to read out hardware registers +which monitor the duration of sleep states the processor resided in. The +cpupower monitor tool (cpupower\-monitor(1)) can be used to show real sleep +state residencies. Please refer to the architecture specific description +section below. + +.SH IDLE\-INFO ARCHITECTURE SPECIFIC DESCRIPTIONS +.SS "X86" +POLL idle state + +If cpuidle is active, X86 platforms have one special idle state. +The POLL idle state is not a real idle state, it does not save any +power. Instead, a busy\-loop is executed doing nothing for a short period of +time. This state is used if the kernel knows that work has to be processed +very soon and entering any real hardware idle state may result in a slight +performance penalty. + +There exist two different cpuidle drivers on the X86 architecture platform: + +"acpi_idle" cpuidle driver + +The acpi_idle cpuidle driver retrieves available sleep states (C\-states) from +the ACPI BIOS tables (from the _CST ACPI function on recent platforms or from +the FADT BIOS table on older ones). +The C1 state is not retrieved from ACPI tables. If the C1 state is entered, +the kernel will call the hlt instruction (or mwait on Intel). + +"intel_idle" cpuidle driver + +In kernel 2.6.36 the intel_idle driver was introduced. +It only serves recent Intel CPUs (Nehalem, Westmere, Sandybridge, Atoms or +newer). On older Intel CPUs the acpi_idle driver is still used (if the BIOS +provides C\-state ACPI tables). +The intel_idle driver knows the sleep state capabilities of the processor and +ignores ACPI BIOS exported processor sleep states tables. + +.SH "REMARKS" +.LP +By default only values of core zero are displayed. How to display settings of +other cores is described in the cpupower(1) manpage in the \-\-cpu option +section. +.SH REFERENCES +http://www.acpi.info/spec.htm +.SH "FILES" +.nf +\fI/sys/devices/system/cpu/cpu*/cpuidle/state*\fP +\fI/sys/devices/system/cpu/cpuidle/*\fP +.fi +.SH "AUTHORS" +.nf +Thomas Renninger <trenn@suse.de> +.fi +.SH "SEE ALSO" +.LP +cpupower(1), cpupower\-monitor(1), cpupower\-info(1), cpupower\-set(1), +cpupower\-idle\-set(1) diff --git a/tools/power/cpupower/man/cpupower-idle-set.1 b/tools/power/cpupower/man/cpupower-idle-set.1 new file mode 100644 index 00000000000..3e6799d7a79 --- /dev/null +++ b/tools/power/cpupower/man/cpupower-idle-set.1 @@ -0,0 +1,77 @@ +.TH "CPUPOWER-IDLE-SET" "1" "0.1" "" "cpupower Manual" +.SH "NAME" +.LP +cpupower idle\-set \- Utility to set cpu idle state specific kernel options +.SH "SYNTAX" +.LP +cpupower [ \-c cpulist ] idle\-info [\fIoptions\fP] +.SH "DESCRIPTION" +.LP +The cpupower idle\-set subcommand allows to set cpu idle, also called cpu +sleep state, specific options offered by the kernel. One example is disabling +sleep states. This can be handy for power vs performance tuning. +.SH "OPTIONS" +.LP +.TP +\fB\-d\fR \fB\-\-disable\fR <STATE_NO> +Disable a specific processor sleep state. +.TP +\fB\-e\fR \fB\-\-enable\fR <STATE_NO> +Enable a specific processor sleep state. +.TP +\fB\-D\fR \fB\-\-disable-by-latency\fR <LATENCY> +Disable all idle states with a equal or higher latency than <LATENCY> +.TP +\fB\-E\fR \fB\-\-enable-all\fR +Enable all idle states if not enabled already. + +.SH "REMARKS" +.LP +Cpuidle Governors Policy on Disabling Sleep States + +.RS 4 +Depending on the used cpuidle governor, implementing the kernel policy +how to choose sleep states, subsequent sleep states on this core, might get +disabled as well. + +There are two cpuidle governors ladder and menu. While the ladder +governor is always available, if CONFIG_CPU_IDLE is selected, the +menu governor additionally requires CONFIG_NO_HZ. + +The behavior and the effect of the disable variable depends on the +implementation of a particular governor. In the ladder governor, for +example, it is not coherent, i.e. if one is disabling a light state, +then all deeper states are disabled as well. Likewise, if one enables a +deep state but a lighter state still is disabled, then this has no effect. +.RE +.LP +Disabling the Lightest Sleep State may not have any Affect + +.RS 4 +If criteria are not met to enter deeper sleep states and the lightest sleep +state is chosen when idle, the kernel may still enter this sleep state, +irrespective of whether it is disabled or not. This is also reflected in +the usage count of the disabled sleep state when using the cpupower idle-info +command. +.RE +.LP +Selecting specific CPU Cores + +.RS 4 +By default processor sleep states of all CPU cores are set. Please refer +to the cpupower(1) manpage in the \-\-cpu option section how to disable +C-states of specific cores. +.RE +.SH "FILES" +.nf +\fI/sys/devices/system/cpu/cpu*/cpuidle/state*\fP +\fI/sys/devices/system/cpu/cpuidle/*\fP +.fi +.SH "AUTHORS" +.nf +Thomas Renninger <trenn@suse.de> +.fi +.SH "SEE ALSO" +.LP +cpupower(1), cpupower\-monitor(1), cpupower\-info(1), cpupower\-set(1), +cpupower\-idle\-info(1) diff --git a/tools/power/cpupower/man/cpupower-info.1 b/tools/power/cpupower/man/cpupower-info.1 index 58e21196f17..340bcd0be7d 100644 --- a/tools/power/cpupower/man/cpupower-info.1 +++ b/tools/power/cpupower/man/cpupower-info.1 @@ -3,7 +3,7 @@ cpupower\-info \- Shows processor power related kernel or hardware configurations .SH SYNOPSIS .ft B -.B cpupower info [ \-b ] [ \-s ] [ \-m ] +.B cpupower info [ \-b ] .SH DESCRIPTION \fBcpupower info \fP shows kernel configurations or processor hardware diff --git a/tools/power/cpupower/man/cpupower-monitor.1 b/tools/power/cpupower/man/cpupower-monitor.1 index d5cfa265c3d..914cbb9d9cd 100644 --- a/tools/power/cpupower/man/cpupower-monitor.1 +++ b/tools/power/cpupower/man/cpupower-monitor.1 @@ -7,11 +7,11 @@ cpupower\-monitor \- Report processor frequency and idle statistics .RB "\-l" .B cpupower monitor -.RB [ "\-m <mon1>," [ "<mon2>,..." ] ] +.RB [ -c ] [ "\-m <mon1>," [ "<mon2>,..." ] ] .RB [ "\-i seconds" ] .br .B cpupower monitor -.RB [ "\-m <mon1>," [ "<mon2>,..." ] ] +.RB [ -c ][ "\-m <mon1>," [ "<mon2>,..." ] ] .RB command .br .SH DESCRIPTION @@ -64,6 +64,17 @@ Only display specific monitors. Use the monitor string(s) provided by \-l option Measure intervall. .RE .PP +\-c +.RS 4 +Schedule the process on every core before starting and ending measuring. +This could be needed for the Idle_Stats monitor when no other MSR based +monitor (has to be run on the core that is measured) is run in parallel. +This is to wake up the processors from deeper sleep states and let the +kernel re +-account its cpuidle (C-state) information before reading the +cpuidle timings from sysfs. +.RE +.PP command .RS 4 Measure idle and frequency characteristics of an arbitrary command/workload. @@ -99,15 +110,23 @@ May work poorly on Linux-2.6.20 through 2.6.29, as the \fBacpi-cpufreq \fP kernel frequency driver periodically cleared aperf/mperf registers in those kernels. -.SS "Nehalem" "SandyBridge" +.SS "Nehalem" "SandyBridge" "HaswellExtended" Intel Core and Package sleep state counters. Threads (hyperthreaded cores) may not be able to enter deeper core states if its sibling is utilized. Deepest package sleep states may in reality show up as machine/platform wide sleep states and can only be entered if all cores are idle. Look up Intel manuals (some are provided in the References section) for further details. - -.SS "Ontario" "Liano" +The monitors are named after the CPU family where the sleep state capabilities +got introduced and may not match exactly the CPU name of the platform. +For example an IvyBridge processor has sleep state capabilities which got +introduced in Nehalem and SandyBridge processor families. +Thus on an IvyBridge processor one will get Nehalem and SandyBridge sleep +state monitors. +HaswellExtended extra package sleep state capabilities are available only in a +specific Haswell (family 0x45) and probably also other future processors. + +.SS "Fam_12h" "Fam_14h" AMD laptop and desktop processor (family 12h and 14h) sleep state counters. The registers are accessed via PCI and therefore can still be read out while cores have been offlined. diff --git a/tools/power/cpupower/man/cpupower-set.1 b/tools/power/cpupower/man/cpupower-set.1 index c4954a9fe4e..2bcc696f449 100644 --- a/tools/power/cpupower/man/cpupower-set.1 +++ b/tools/power/cpupower/man/cpupower-set.1 @@ -3,7 +3,7 @@ cpupower\-set \- Set processor power related kernel or hardware configurations .SH SYNOPSIS .ft B -.B cpupower set [ \-b VAL ] [ \-s VAL ] [ \-m VAL ] +.B cpupower set [ \-b VAL ] .SH DESCRIPTION @@ -55,44 +55,6 @@ Use \fBcpupower -c all info -b\fP to verify. This options needs the msr kernel driver (CONFIG_X86_MSR) loaded. .RE -.PP -\-\-sched\-mc, \-m [ VAL ] -.RE -\-\-sched\-smt, \-s [ VAL ] -.RS 4 -\-\-sched\-mc utilizes cores in one processor package/socket first before -processes are scheduled to other processor packages/sockets. - -\-\-sched\-smt utilizes thread siblings of one processor core first before -processes are scheduled to other cores. - -The impact on power consumption and performance (positiv or negativ) heavily -depends on processor support for deep sleep states, frequency scaling and -frequency boost modes and their dependencies between other thread siblings -and processor cores. - -Taken over from kernel documentation: - -Adjust the kernel's multi-core scheduler support. - -Possible values are: -.RS 2 -0 - No power saving load balance (default value) - -1 - Fill one thread/core/package first for long running threads - -2 - Also bias task wakeups to semi-idle cpu package for power -savings -.RE - -sched_mc_power_savings is dependent upon SCHED_MC, which is -itself architecture dependent. - -sched_smt_power_savings is dependent upon SCHED_SMT, which -is itself architecture dependent. - -The two files are independent of each other. It is possible -that one file may be present without the other. .SH "SEE ALSO" cpupower-info(1), cpupower-monitor(1), powertop(1) diff --git a/tools/power/cpupower/utils/builtin.h b/tools/power/cpupower/utils/builtin.h index c10496fbe3c..2284c8ea4e2 100644 --- a/tools/power/cpupower/utils/builtin.h +++ b/tools/power/cpupower/utils/builtin.h @@ -5,6 +5,7 @@ extern int cmd_set(int argc, const char **argv); extern int cmd_info(int argc, const char **argv); extern int cmd_freq_set(int argc, const char **argv); extern int cmd_freq_info(int argc, const char **argv); +extern int cmd_idle_set(int argc, const char **argv); extern int cmd_idle_info(int argc, const char **argv); extern int cmd_monitor(int argc, const char **argv); diff --git a/tools/power/cpupower/utils/cpufreq-info.c b/tools/power/cpupower/utils/cpufreq-info.c index 28953c9a7bd..b4b90a97662 100644 --- a/tools/power/cpupower/utils/cpufreq-info.c +++ b/tools/power/cpupower/utils/cpufreq-info.c @@ -82,29 +82,42 @@ static void proc_cpufreq_output(void) } } +static int no_rounding; static void print_speed(unsigned long speed) { unsigned long tmp; - if (speed > 1000000) { - tmp = speed % 10000; - if (tmp >= 5000) - speed += 10000; - printf("%u.%02u GHz", ((unsigned int) speed/1000000), - ((unsigned int) (speed%1000000)/10000)); - } else if (speed > 100000) { - tmp = speed % 1000; - if (tmp >= 500) - speed += 1000; - printf("%u MHz", ((unsigned int) speed / 1000)); - } else if (speed > 1000) { - tmp = speed % 100; - if (tmp >= 50) - speed += 100; - printf("%u.%01u MHz", ((unsigned int) speed/1000), - ((unsigned int) (speed%1000)/100)); - } else - printf("%lu kHz", speed); + if (no_rounding) { + if (speed > 1000000) + printf("%u.%06u GHz", ((unsigned int) speed/1000000), + ((unsigned int) speed%1000000)); + else if (speed > 100000) + printf("%u MHz", (unsigned int) speed); + else if (speed > 1000) + printf("%u.%03u MHz", ((unsigned int) speed/1000), + (unsigned int) (speed%1000)); + else + printf("%lu kHz", speed); + } else { + if (speed > 1000000) { + tmp = speed%10000; + if (tmp >= 5000) + speed += 10000; + printf("%u.%02u GHz", ((unsigned int) speed/1000000), + ((unsigned int) (speed%1000000)/10000)); + } else if (speed > 100000) { + tmp = speed%1000; + if (tmp >= 500) + speed += 1000; + printf("%u MHz", ((unsigned int) speed/1000)); + } else if (speed > 1000) { + tmp = speed%100; + if (tmp >= 50) + speed += 100; + printf("%u.%01u MHz", ((unsigned int) speed/1000), + ((unsigned int) (speed%1000)/100)); + } + } return; } @@ -113,26 +126,38 @@ static void print_duration(unsigned long duration) { unsigned long tmp; - if (duration > 1000000) { - tmp = duration % 10000; - if (tmp >= 5000) - duration += 10000; - printf("%u.%02u ms", ((unsigned int) duration/1000000), - ((unsigned int) (duration%1000000)/10000)); - } else if (duration > 100000) { - tmp = duration % 1000; - if (tmp >= 500) - duration += 1000; - printf("%u us", ((unsigned int) duration / 1000)); - } else if (duration > 1000) { - tmp = duration % 100; - if (tmp >= 50) - duration += 100; - printf("%u.%01u us", ((unsigned int) duration/1000), - ((unsigned int) (duration%1000)/100)); - } else - printf("%lu ns", duration); - + if (no_rounding) { + if (duration > 1000000) + printf("%u.%06u ms", ((unsigned int) duration/1000000), + ((unsigned int) duration%1000000)); + else if (duration > 100000) + printf("%u us", ((unsigned int) duration/1000)); + else if (duration > 1000) + printf("%u.%03u us", ((unsigned int) duration/1000), + ((unsigned int) duration%1000)); + else + printf("%lu ns", duration); + } else { + if (duration > 1000000) { + tmp = duration%10000; + if (tmp >= 5000) + duration += 10000; + printf("%u.%02u ms", ((unsigned int) duration/1000000), + ((unsigned int) (duration%1000000)/10000)); + } else if (duration > 100000) { + tmp = duration%1000; + if (tmp >= 500) + duration += 1000; + printf("%u us", ((unsigned int) duration / 1000)); + } else if (duration > 1000) { + tmp = duration%100; + if (tmp >= 50) + duration += 100; + printf("%u.%01u us", ((unsigned int) duration/1000), + ((unsigned int) (duration%1000)/100)); + } else + printf("%lu ns", duration); + } return; } @@ -525,6 +550,7 @@ static struct option info_opts[] = { { .name = "latency", .has_arg = no_argument, .flag = NULL, .val = 'y'}, { .name = "proc", .has_arg = no_argument, .flag = NULL, .val = 'o'}, { .name = "human", .has_arg = no_argument, .flag = NULL, .val = 'm'}, + { .name = "no-rounding", .has_arg = no_argument, .flag = NULL, .val = 'n'}, { }, }; @@ -538,7 +564,8 @@ int cmd_freq_info(int argc, char **argv) int output_param = 0; do { - ret = getopt_long(argc, argv, "oefwldpgrasmyb", info_opts, NULL); + ret = getopt_long(argc, argv, "oefwldpgrasmybn", info_opts, + NULL); switch (ret) { case '?': output_param = '?'; @@ -575,6 +602,9 @@ int cmd_freq_info(int argc, char **argv) } human = 1; break; + case 'n': + no_rounding = 1; + break; default: fprintf(stderr, "invalid or unknown argument\n"); return EXIT_FAILURE; diff --git a/tools/power/cpupower/utils/cpufreq-set.c b/tools/power/cpupower/utils/cpufreq-set.c index dd1539eb8c6..a416de80c55 100644 --- a/tools/power/cpupower/utils/cpufreq-set.c +++ b/tools/power/cpupower/utils/cpufreq-set.c @@ -257,7 +257,7 @@ int cmd_freq_set(int argc, char **argv) print_unknown_arg(); return -EINVAL; } - if ((sscanf(optarg, "%s", gov)) != 1) { + if ((sscanf(optarg, "%19s", gov)) != 1) { print_unknown_arg(); return -EINVAL; } diff --git a/tools/power/cpupower/utils/cpuidle-info.c b/tools/power/cpupower/utils/cpuidle-info.c index b028267c137..75e66de7e7a 100644 --- a/tools/power/cpupower/utils/cpuidle-info.c +++ b/tools/power/cpupower/utils/cpuidle-info.c @@ -22,7 +22,7 @@ static void cpuidle_cpu_output(unsigned int cpu, int verbose) { - int idlestates, idlestate; + unsigned int idlestates, idlestate; char *tmp; printf(_ ("Analyzing CPU %d:\n"), cpu); @@ -31,21 +31,11 @@ static void cpuidle_cpu_output(unsigned int cpu, int verbose) if (idlestates == 0) { printf(_("CPU %u: No idle states\n"), cpu); return; - } else if (idlestates <= 0) { - printf(_("CPU %u: Can't read idle state info\n"), cpu); - return; - } - tmp = sysfs_get_idlestate_name(cpu, idlestates - 1); - if (!tmp) { - printf(_("Could not determine max idle state %u\n"), - idlestates - 1); - return; } printf(_("Number of idle states: %d\n"), idlestates); - printf(_("Available idle states:")); - for (idlestate = 1; idlestate < idlestates; idlestate++) { + for (idlestate = 0; idlestate < idlestates; idlestate++) { tmp = sysfs_get_idlestate_name(cpu, idlestate); if (!tmp) continue; @@ -57,11 +47,15 @@ static void cpuidle_cpu_output(unsigned int cpu, int verbose) if (!verbose) return; - for (idlestate = 1; idlestate < idlestates; idlestate++) { + for (idlestate = 0; idlestate < idlestates; idlestate++) { + int disabled = sysfs_is_idlestate_disabled(cpu, idlestate); + /* Disabled interface not supported on older kernels */ + if (disabled < 0) + disabled = 0; tmp = sysfs_get_idlestate_name(cpu, idlestate); if (!tmp) continue; - printf("%s:\n", tmp); + printf("%s%s:\n", tmp, (disabled) ? " (DISABLED) " : ""); free(tmp); tmp = sysfs_get_idlestate_desc(cpu, idlestate); @@ -106,21 +100,13 @@ static void cpuidle_general_output(void) static void proc_cpuidle_cpu_output(unsigned int cpu) { long max_allowed_cstate = 2000000000; - int cstates, cstate; + unsigned int cstate, cstates; cstates = sysfs_get_idlestate_count(cpu); if (cstates == 0) { - /* - * Go on and print same useless info as you'd see with - * cat /proc/acpi/processor/../power - * printf(_("CPU %u: No C-states available\n"), cpu); - * return; - */ - } else if (cstates <= 0) { - printf(_("CPU %u: Can't read C-state info\n"), cpu); + printf(_("CPU %u: No C-states info\n"), cpu); return; } - /* printf("Cstates: %d\n", cstates); */ printf(_("active state: C0\n")); printf(_("max_cstate: C%u\n"), cstates-1); diff --git a/tools/power/cpupower/utils/cpuidle-set.c b/tools/power/cpupower/utils/cpuidle-set.c new file mode 100644 index 00000000000..d45d8d775c0 --- /dev/null +++ b/tools/power/cpupower/utils/cpuidle-set.c @@ -0,0 +1,181 @@ +#include <unistd.h> +#include <stdio.h> +#include <errno.h> +#include <stdlib.h> +#include <limits.h> +#include <string.h> +#include <ctype.h> + +#include <getopt.h> + +#include "cpufreq.h" +#include "helpers/helpers.h" +#include "helpers/sysfs.h" + +static struct option info_opts[] = { + { .name = "disable", + .has_arg = required_argument, .flag = NULL, .val = 'd'}, + { .name = "enable", + .has_arg = required_argument, .flag = NULL, .val = 'e'}, + { .name = "disable-by-latency", + .has_arg = required_argument, .flag = NULL, .val = 'D'}, + { .name = "enable-all", + .has_arg = no_argument, .flag = NULL, .val = 'E'}, + { }, +}; + + +int cmd_idle_set(int argc, char **argv) +{ + extern char *optarg; + extern int optind, opterr, optopt; + int ret = 0, cont = 1, param = 0, disabled; + unsigned long long latency = 0, state_latency; + unsigned int cpu = 0, idlestate = 0, idlestates = 0; + char *endptr; + + do { + ret = getopt_long(argc, argv, "d:e:ED:", info_opts, NULL); + if (ret == -1) + break; + switch (ret) { + case '?': + param = '?'; + cont = 0; + break; + case 'd': + if (param) { + param = -1; + cont = 0; + break; + } + param = ret; + idlestate = atoi(optarg); + break; + case 'e': + if (param) { + param = -1; + cont = 0; + break; + } + param = ret; + idlestate = atoi(optarg); + break; + case 'D': + if (param) { + param = -1; + cont = 0; + break; + } + param = ret; + latency = strtoull(optarg, &endptr, 10); + if (*endptr != '\0') { + printf(_("Bad latency value: %s\n"), optarg); + exit(EXIT_FAILURE); + } + break; + case 'E': + if (param) { + param = -1; + cont = 0; + break; + } + param = ret; + break; + case -1: + cont = 0; + break; + } + } while (cont); + + switch (param) { + case -1: + printf(_("You can't specify more than one " + "output-specific argument\n")); + exit(EXIT_FAILURE); + case '?': + printf(_("invalid or unknown argument\n")); + exit(EXIT_FAILURE); + } + + /* Default is: set all CPUs */ + if (bitmask_isallclear(cpus_chosen)) + bitmask_setall(cpus_chosen); + + for (cpu = bitmask_first(cpus_chosen); + cpu <= bitmask_last(cpus_chosen); cpu++) { + + if (!bitmask_isbitset(cpus_chosen, cpu)) + continue; + + if (sysfs_is_cpu_online(cpu) != 1) + continue; + + idlestates = sysfs_get_idlestate_count(cpu); + if (idlestates <= 0) + continue; + + switch (param) { + case 'd': + ret = sysfs_idlestate_disable(cpu, idlestate, 1); + if (ret == 0) + printf(_("Idlestate %u disabled on CPU %u\n"), idlestate, cpu); + else if (ret == -1) + printf(_("Idlestate %u not available on CPU %u\n"), + idlestate, cpu); + else if (ret == -2) + printf(_("Idlestate disabling not supported by kernel\n")); + else + printf(_("Idlestate %u not disabled on CPU %u\n"), + idlestate, cpu); + break; + case 'e': + ret = sysfs_idlestate_disable(cpu, idlestate, 0); + if (ret == 0) + printf(_("Idlestate %u enabled on CPU %u\n"), idlestate, cpu); + else if (ret == -1) + printf(_("Idlestate %u not available on CPU %u\n"), + idlestate, cpu); + else if (ret == -2) + printf(_("Idlestate enabling not supported by kernel\n")); + else + printf(_("Idlestate %u not enabled on CPU %u\n"), + idlestate, cpu); + break; + case 'D': + for (idlestate = 0; idlestate < idlestates; idlestate++) { + disabled = sysfs_is_idlestate_disabled + (cpu, idlestate); + state_latency = sysfs_get_idlestate_latency + (cpu, idlestate); + printf("CPU: %u - idlestate %u - state_latency: %llu - latency: %llu\n", + cpu, idlestate, state_latency, latency); + if (disabled == 1 || latency > state_latency) + continue; + ret = sysfs_idlestate_disable + (cpu, idlestate, 1); + if (ret == 0) + printf(_("Idlestate %u disabled on CPU %u\n"), idlestate, cpu); + } + break; + case 'E': + for (idlestate = 0; idlestate < idlestates; idlestate++) { + disabled = sysfs_is_idlestate_disabled + (cpu, idlestate); + if (disabled == 1) { + ret = sysfs_idlestate_disable + (cpu, idlestate, 0); + if (ret == 0) + printf(_("Idlestate %u enabled on CPU %u\n"), idlestate, cpu); + } + } + break; + default: + /* Not reachable with proper args checking */ + printf(_("Invalid or unknown argument\n")); + exit(EXIT_FAILURE); + break; + } + } + return EXIT_SUCCESS; +} diff --git a/tools/power/cpupower/utils/cpupower-info.c b/tools/power/cpupower/utils/cpupower-info.c index 3f68632c28c..136d979e958 100644 --- a/tools/power/cpupower/utils/cpupower-info.c +++ b/tools/power/cpupower/utils/cpupower-info.c @@ -18,8 +18,6 @@ static struct option set_opts[] = { { .name = "perf-bias", .has_arg = optional_argument, .flag = NULL, .val = 'b'}, - { .name = "sched-mc", .has_arg = optional_argument, .flag = NULL, .val = 'm'}, - { .name = "sched-smt", .has_arg = optional_argument, .flag = NULL, .val = 's'}, { }, }; @@ -37,8 +35,6 @@ int cmd_info(int argc, char **argv) union { struct { - int sched_mc:1; - int sched_smt:1; int perf_bias:1; }; int params; @@ -49,23 +45,13 @@ int cmd_info(int argc, char **argv) textdomain(PACKAGE); /* parameter parsing */ - while ((ret = getopt_long(argc, argv, "msb", set_opts, NULL)) != -1) { + while ((ret = getopt_long(argc, argv, "b", set_opts, NULL)) != -1) { switch (ret) { case 'b': if (params.perf_bias) print_wrong_arg_exit(); params.perf_bias = 1; break; - case 'm': - if (params.sched_mc) - print_wrong_arg_exit(); - params.sched_mc = 1; - break; - case 's': - if (params.sched_smt) - print_wrong_arg_exit(); - params.sched_smt = 1; - break; default: print_wrong_arg_exit(); } @@ -78,25 +64,6 @@ int cmd_info(int argc, char **argv) if (bitmask_isallclear(cpus_chosen)) bitmask_setbit(cpus_chosen, 0); - if (params.sched_mc) { - ret = sysfs_get_sched("mc"); - printf(_("System's multi core scheduler setting: ")); - if (ret < 0) - /* if sysfs file is missing it's: errno == ENOENT */ - printf(_("not supported\n")); - else - printf("%d\n", ret); - } - if (params.sched_smt) { - ret = sysfs_get_sched("smt"); - printf(_("System's thread sibling scheduler setting: ")); - if (ret < 0) - /* if sysfs file is missing it's: errno == ENOENT */ - printf(_("not supported\n")); - else - printf("%d\n", ret); - } - /* Add more per cpu options here */ if (!params.perf_bias) return ret; @@ -125,11 +92,12 @@ int cmd_info(int argc, char **argv) if (params.perf_bias) { ret = msr_intel_get_perf_bias(cpu); if (ret < 0) { - printf(_("Could not read perf-bias value\n")); - break; + fprintf(stderr, + _("Could not read perf-bias value[%d]\n"), ret); + exit(EXIT_FAILURE); } else printf(_("perf-bias: %d\n"), ret); } } - return ret; + return 0; } diff --git a/tools/power/cpupower/utils/cpupower-set.c b/tools/power/cpupower/utils/cpupower-set.c index dc4de376211..573c75f8e3f 100644 --- a/tools/power/cpupower/utils/cpupower-set.c +++ b/tools/power/cpupower/utils/cpupower-set.c @@ -18,9 +18,7 @@ #include "helpers/bitmask.h" static struct option set_opts[] = { - { .name = "perf-bias", .has_arg = optional_argument, .flag = NULL, .val = 'b'}, - { .name = "sched-mc", .has_arg = optional_argument, .flag = NULL, .val = 'm'}, - { .name = "sched-smt", .has_arg = optional_argument, .flag = NULL, .val = 's'}, + { .name = "perf-bias", .has_arg = required_argument, .flag = NULL, .val = 'b'}, { }, }; @@ -38,13 +36,11 @@ int cmd_set(int argc, char **argv) union { struct { - int sched_mc:1; - int sched_smt:1; int perf_bias:1; }; int params; } params; - int sched_mc = 0, sched_smt = 0, perf_bias = 0; + int perf_bias = 0; int ret = 0; setlocale(LC_ALL, ""); @@ -52,7 +48,7 @@ int cmd_set(int argc, char **argv) params.params = 0; /* parameter parsing */ - while ((ret = getopt_long(argc, argv, "m:s:b:", + while ((ret = getopt_long(argc, argv, "b:", set_opts, NULL)) != -1) { switch (ret) { case 'b': @@ -66,28 +62,6 @@ int cmd_set(int argc, char **argv) } params.perf_bias = 1; break; - case 'm': - if (params.sched_mc) - print_wrong_arg_exit(); - sched_mc = atoi(optarg); - if (sched_mc < 0 || sched_mc > 2) { - printf(_("--sched-mc param out " - "of range [0-%d]\n"), 2); - print_wrong_arg_exit(); - } - params.sched_mc = 1; - break; - case 's': - if (params.sched_smt) - print_wrong_arg_exit(); - sched_smt = atoi(optarg); - if (sched_smt < 0 || sched_smt > 2) { - printf(_("--sched-smt param out " - "of range [0-%d]\n"), 2); - print_wrong_arg_exit(); - } - params.sched_smt = 1; - break; default: print_wrong_arg_exit(); } @@ -96,19 +70,6 @@ int cmd_set(int argc, char **argv) if (!params.params) print_wrong_arg_exit(); - if (params.sched_mc) { - ret = sysfs_set_sched("mc", sched_mc); - if (ret) - fprintf(stderr, _("Error setting sched-mc %s\n"), - (ret == -ENODEV) ? "not supported" : ""); - } - if (params.sched_smt) { - ret = sysfs_set_sched("smt", sched_smt); - if (ret) - fprintf(stderr, _("Error setting sched-smt %s\n"), - (ret == -ENODEV) ? "not supported" : ""); - } - /* Default is: set all CPUs */ if (bitmask_isallclear(cpus_chosen)) bitmask_setall(cpus_chosen); diff --git a/tools/power/cpupower/utils/cpupower.c b/tools/power/cpupower/utils/cpupower.c index 52bee591c1c..7cdcf88659c 100644 --- a/tools/power/cpupower/utils/cpupower.c +++ b/tools/power/cpupower/utils/cpupower.c @@ -12,17 +12,14 @@ #include <string.h> #include <unistd.h> #include <errno.h> +#include <sys/types.h> +#include <sys/stat.h> +#include <sys/utsname.h> #include "builtin.h" #include "helpers/helpers.h" #include "helpers/bitmask.h" -struct cmd_struct { - const char *cmd; - int (*main)(int, const char **); - int needs_root; -}; - #define ARRAY_SIZE(x) (sizeof(x)/sizeof(x[0])) static int cmd_help(int argc, const char **argv); @@ -43,10 +40,17 @@ int be_verbose; static void print_help(void); +struct cmd_struct { + const char *cmd; + int (*main)(int, const char **); + int needs_root; +}; + static struct cmd_struct commands[] = { { "frequency-info", cmd_freq_info, 0 }, { "frequency-set", cmd_freq_set, 1 }, { "idle-info", cmd_idle_info, 0 }, + { "idle-set", cmd_idle_set, 1 }, { "set", cmd_set, 1 }, { "info", cmd_info, 0 }, { "monitor", cmd_monitor, 0 }, @@ -168,6 +172,8 @@ int main(int argc, const char *argv[]) { const char *cmd; unsigned int i, ret; + struct stat statbuf; + struct utsname uts; cpus_chosen = bitmask_alloc(sysconf(_SC_NPROCESSORS_CONF)); @@ -194,6 +200,15 @@ int main(int argc, const char *argv[]) get_cpu_info(0, &cpupower_cpu_info); run_as_root = !getuid(); + if (run_as_root) { + ret = uname(&uts); + if (!ret && !strcmp(uts.machine, "x86_64") && + stat("/dev/cpu/0/msr", &statbuf) != 0) { + if (system("modprobe msr") == -1) + fprintf(stderr, _("MSR access not available.\n")); + } + } + for (i = 0; i < ARRAY_SIZE(commands); i++) { struct cmd_struct *p = commands + i; diff --git a/tools/power/cpupower/utils/helpers/amd.c b/tools/power/cpupower/utils/helpers/amd.c index 87d5605bdda..6437ef39aee 100644 --- a/tools/power/cpupower/utils/helpers/amd.c +++ b/tools/power/cpupower/utils/helpers/amd.c @@ -112,14 +112,12 @@ int decode_pstates(unsigned int cpu, unsigned int cpu_family, int amd_pci_get_num_boost_states(int *active, int *states) { struct pci_access *pci_acc; - int vendor_id = 0x1022; - int boost_dev_ids[4] = {0x1204, 0x1604, 0x1704, 0}; struct pci_dev *device; uint8_t val = 0; *active = *states = 0; - device = pci_acc_init(&pci_acc, vendor_id, boost_dev_ids); + device = pci_slot_func_init(&pci_acc, 0x18, 4); if (device == NULL) return -ENODEV; diff --git a/tools/power/cpupower/utils/helpers/cpuid.c b/tools/power/cpupower/utils/helpers/cpuid.c index 906895d21cc..93b0aa74ca0 100644 --- a/tools/power/cpupower/utils/helpers/cpuid.c +++ b/tools/power/cpupower/utils/helpers/cpuid.c @@ -158,6 +158,8 @@ out: cpu_info->caps |= CPUPOWER_CAP_HAS_TURBO_RATIO; case 0x2A: /* SNB */ case 0x2D: /* SNB Xeon */ + case 0x3A: /* IVB */ + case 0x3E: /* IVB Xeon */ cpu_info->caps |= CPUPOWER_CAP_HAS_TURBO_RATIO; cpu_info->caps |= CPUPOWER_CAP_IS_SNB; break; diff --git a/tools/power/cpupower/utils/helpers/helpers.h b/tools/power/cpupower/utils/helpers/helpers.h index 2747e738efb..aa9e95486a2 100644 --- a/tools/power/cpupower/utils/helpers/helpers.h +++ b/tools/power/cpupower/utils/helpers/helpers.h @@ -66,8 +66,8 @@ enum cpupower_cpu_vendor {X86_VENDOR_UNKNOWN = 0, X86_VENDOR_INTEL, #define CPUPOWER_CAP_AMD_CBP 0x00000004 #define CPUPOWER_CAP_PERF_BIAS 0x00000008 #define CPUPOWER_CAP_HAS_TURBO_RATIO 0x00000010 -#define CPUPOWER_CAP_IS_SNB 0x00000011 -#define CPUPOWER_CAP_INTEL_IDA 0x00000012 +#define CPUPOWER_CAP_IS_SNB 0x00000020 +#define CPUPOWER_CAP_INTEL_IDA 0x00000040 #define MAX_HW_PSTATES 10 @@ -92,6 +92,14 @@ extern int get_cpu_info(unsigned int cpu, struct cpupower_cpu_info *cpu_info); extern struct cpupower_cpu_info cpupower_cpu_info; /* cpuid and cpuinfo helpers **************************/ +struct cpuid_core_info { + int pkg; + int core; + int cpu; + + /* flags */ + unsigned int is_online:1; +}; /* CPU topology/hierarchy parsing ******************/ struct cpupower_topology { @@ -101,18 +109,12 @@ struct cpupower_topology { unsigned int threads; /* per core */ /* Array gets mallocated with cores entries, holding per core info */ - struct { - int pkg; - int core; - int cpu; - - /* flags */ - unsigned int is_online:1; - } *core_info; + struct cpuid_core_info *core_info; }; extern int get_cpu_topology(struct cpupower_topology *cpu_top); extern void cpu_topology_release(struct cpupower_topology cpu_top); + /* CPU topology/hierarchy parsing ******************/ /* X86 ONLY ****************************************/ @@ -132,8 +134,11 @@ extern unsigned long long msr_intel_get_turbo_ratio(unsigned int cpu); /* PCI stuff ****************************/ extern int amd_pci_get_num_boost_states(int *active, int *states); -extern struct pci_dev *pci_acc_init(struct pci_access **pacc, int vendor_id, - int *dev_ids); +extern struct pci_dev *pci_acc_init(struct pci_access **pacc, int domain, + int bus, int slot, int func, int vendor, + int dev); +extern struct pci_dev *pci_slot_func_init(struct pci_access **pacc, + int slot, int func); /* PCI stuff ****************************/ diff --git a/tools/power/cpupower/utils/helpers/pci.c b/tools/power/cpupower/utils/helpers/pci.c index cd2eb6fe41c..9690798e644 100644 --- a/tools/power/cpupower/utils/helpers/pci.c +++ b/tools/power/cpupower/utils/helpers/pci.c @@ -10,19 +10,24 @@ * **pacc : if a valid pci_dev is returned * *pacc must be passed to pci_acc_cleanup to free it * - * vendor_id : the pci vendor id matching the pci device to access - * dev_ids : device ids matching the pci device to access + * domain: domain + * bus: bus + * slot: slot + * func: func + * vendor: vendor + * device: device + * Pass -1 for one of the six above to match any * * Returns : * struct pci_dev which can be used with pci_{read,write}_* functions * to access the PCI config space of matching pci devices */ -struct pci_dev *pci_acc_init(struct pci_access **pacc, int vendor_id, - int *dev_ids) +struct pci_dev *pci_acc_init(struct pci_access **pacc, int domain, int bus, + int slot, int func, int vendor, int dev) { - struct pci_filter filter_nb_link = { -1, -1, -1, -1, vendor_id, 0}; + struct pci_filter filter_nb_link = { domain, bus, slot, func, + vendor, dev }; struct pci_dev *device; - unsigned int i; *pacc = pci_alloc(); if (*pacc == NULL) @@ -31,14 +36,20 @@ struct pci_dev *pci_acc_init(struct pci_access **pacc, int vendor_id, pci_init(*pacc); pci_scan_bus(*pacc); - for (i = 0; dev_ids[i] != 0; i++) { - filter_nb_link.device = dev_ids[i]; - for (device = (*pacc)->devices; device; device = device->next) { - if (pci_filter_match(&filter_nb_link, device)) - return device; - } + for (device = (*pacc)->devices; device; device = device->next) { + if (pci_filter_match(&filter_nb_link, device)) + return device; } pci_cleanup(*pacc); return NULL; } + +/* Typically one wants to get a specific slot(device)/func of the root domain + and bus */ +struct pci_dev *pci_slot_func_init(struct pci_access **pacc, int slot, + int func) +{ + return pci_acc_init(pacc, 0, 0, slot, func, -1, -1); +} + #endif /* defined(__i386__) || defined(__x86_64__) */ diff --git a/tools/power/cpupower/utils/helpers/sysfs.c b/tools/power/cpupower/utils/helpers/sysfs.c index c6343024a61..851c7a16ca4 100644 --- a/tools/power/cpupower/utils/helpers/sysfs.c +++ b/tools/power/cpupower/utils/helpers/sysfs.c @@ -37,25 +37,6 @@ unsigned int sysfs_read_file(const char *path, char *buf, size_t buflen) return (unsigned int) numread; } -static unsigned int sysfs_write_file(const char *path, - const char *value, size_t len) -{ - int fd; - ssize_t numwrite; - - fd = open(path, O_WRONLY); - if (fd == -1) - return 0; - - numwrite = write(fd, value, len); - if (numwrite < 1) { - close(fd); - return 0; - } - close(fd); - return (unsigned int) numwrite; -} - /* * Detect whether a CPU is online * @@ -108,6 +89,33 @@ int sysfs_is_cpu_online(unsigned int cpu) /* CPUidle idlestate specific /sys/devices/system/cpu/cpuX/cpuidle/ access */ + +/* CPUidle idlestate specific /sys/devices/system/cpu/cpuX/cpuidle/ access */ + +/* + * helper function to check whether a file under "../cpuX/cpuidle/stateX/" dir + * exists. + * For example the functionality to disable c-states was introduced in later + * kernel versions, this function can be used to explicitly check for this + * feature. + * + * returns 1 if the file exists, 0 otherwise. + */ +unsigned int sysfs_idlestate_file_exists(unsigned int cpu, + unsigned int idlestate, + const char *fname) +{ + char path[SYSFS_PATH_MAX]; + struct stat statbuf; + + + snprintf(path, sizeof(path), PATH_TO_CPU "cpu%u/cpuidle/state%u/%s", + cpu, idlestate, fname); + if (stat(path, &statbuf) != 0) + return 0; + return 1; +} + /* * helper function to read file from /sys into given buffer * fname is a relative path under "cpuX/cpuidle/stateX/" dir @@ -140,6 +148,40 @@ unsigned int sysfs_idlestate_read_file(unsigned int cpu, unsigned int idlestate, return (unsigned int) numread; } +/* + * helper function to write a new value to a /sys file + * fname is a relative path under "../cpuX/cpuidle/cstateY/" dir + * + * Returns the number of bytes written or 0 on error + */ +static +unsigned int sysfs_idlestate_write_file(unsigned int cpu, + unsigned int idlestate, + const char *fname, + const char *value, size_t len) +{ + char path[SYSFS_PATH_MAX]; + int fd; + ssize_t numwrite; + + snprintf(path, sizeof(path), PATH_TO_CPU "cpu%u/cpuidle/state%u/%s", + cpu, idlestate, fname); + + fd = open(path, O_WRONLY); + if (fd == -1) + return 0; + + numwrite = write(fd, value, len); + if (numwrite < 1) { + close(fd); + return 0; + } + + close(fd); + + return (unsigned int) numwrite; +} + /* read access to files which contain one numeric value */ enum idlestate_value { @@ -147,6 +189,7 @@ enum idlestate_value { IDLESTATE_POWER, IDLESTATE_LATENCY, IDLESTATE_TIME, + IDLESTATE_DISABLE, MAX_IDLESTATE_VALUE_FILES }; @@ -155,6 +198,7 @@ static const char *idlestate_value_files[MAX_IDLESTATE_VALUE_FILES] = { [IDLESTATE_POWER] = "power", [IDLESTATE_LATENCY] = "latency", [IDLESTATE_TIME] = "time", + [IDLESTATE_DISABLE] = "disable", }; static unsigned long long sysfs_idlestate_get_one_value(unsigned int cpu, @@ -224,8 +268,59 @@ static char *sysfs_idlestate_get_one_string(unsigned int cpu, return result; } +/* + * Returns: + * 1 if disabled + * 0 if enabled + * -1 if idlestate is not available + * -2 if disabling is not supported by the kernel + */ +int sysfs_is_idlestate_disabled(unsigned int cpu, + unsigned int idlestate) +{ + if (sysfs_get_idlestate_count(cpu) <= idlestate) + return -1; + + if (!sysfs_idlestate_file_exists(cpu, idlestate, + idlestate_value_files[IDLESTATE_DISABLE])) + return -2; + return sysfs_idlestate_get_one_value(cpu, idlestate, IDLESTATE_DISABLE); +} + +/* + * Pass 1 as last argument to disable or 0 to enable the state + * Returns: + * 0 on success + * negative values on error, for example: + * -1 if idlestate is not available + * -2 if disabling is not supported by the kernel + * -3 No write access to disable/enable C-states + */ +int sysfs_idlestate_disable(unsigned int cpu, + unsigned int idlestate, + unsigned int disable) +{ + char value[SYSFS_PATH_MAX]; + int bytes_written; + + if (sysfs_get_idlestate_count(cpu) <= idlestate) + return -1; + + if (!sysfs_idlestate_file_exists(cpu, idlestate, + idlestate_value_files[IDLESTATE_DISABLE])) + return -2; + + snprintf(value, SYSFS_PATH_MAX, "%u", disable); + + bytes_written = sysfs_idlestate_write_file(cpu, idlestate, "disable", + value, sizeof(disable)); + if (bytes_written) + return 0; + return -3; +} + unsigned long sysfs_get_idlestate_latency(unsigned int cpu, - unsigned int idlestate) + unsigned int idlestate) { return sysfs_idlestate_get_one_value(cpu, idlestate, IDLESTATE_LATENCY); } @@ -257,7 +352,7 @@ char *sysfs_get_idlestate_desc(unsigned int cpu, unsigned int idlestate) * Negativ in error case * Zero if cpuidle does not export any C-states */ -int sysfs_get_idlestate_count(unsigned int cpu) +unsigned int sysfs_get_idlestate_count(unsigned int cpu) { char file[SYSFS_PATH_MAX]; struct stat statbuf; @@ -362,22 +457,7 @@ char *sysfs_get_cpuidle_driver(void) */ int sysfs_get_sched(const char *smt_mc) { - unsigned long value; - char linebuf[MAX_LINE_LEN]; - char *endp; - char path[SYSFS_PATH_MAX]; - - if (strcmp("mc", smt_mc) && strcmp("smt", smt_mc)) - return -EINVAL; - - snprintf(path, sizeof(path), - PATH_TO_CPU "sched_%s_power_savings", smt_mc); - if (sysfs_read_file(path, linebuf, MAX_LINE_LEN) == 0) - return -1; - value = strtoul(linebuf, &endp, 0); - if (endp == linebuf || errno == ERANGE) - return -1; - return value; + return -ENODEV; } /* @@ -388,21 +468,5 @@ int sysfs_get_sched(const char *smt_mc) */ int sysfs_set_sched(const char *smt_mc, int val) { - char linebuf[MAX_LINE_LEN]; - char path[SYSFS_PATH_MAX]; - struct stat statbuf; - - if (strcmp("mc", smt_mc) && strcmp("smt", smt_mc)) - return -EINVAL; - - snprintf(path, sizeof(path), - PATH_TO_CPU "sched_%s_power_savings", smt_mc); - sprintf(linebuf, "%d", val); - - if (stat(path, &statbuf) != 0) - return -ENODEV; - - if (sysfs_write_file(path, linebuf, MAX_LINE_LEN) == 0) - return -1; - return 0; + return -ENODEV; } diff --git a/tools/power/cpupower/utils/helpers/sysfs.h b/tools/power/cpupower/utils/helpers/sysfs.h index 8cb797bbceb..d28f11fedbd 100644 --- a/tools/power/cpupower/utils/helpers/sysfs.h +++ b/tools/power/cpupower/utils/helpers/sysfs.h @@ -7,8 +7,16 @@ extern unsigned int sysfs_read_file(const char *path, char *buf, size_t buflen); +extern unsigned int sysfs_idlestate_file_exists(unsigned int cpu, + unsigned int idlestate, + const char *fname); + extern int sysfs_is_cpu_online(unsigned int cpu); +extern int sysfs_is_idlestate_disabled(unsigned int cpu, + unsigned int idlestate); +extern int sysfs_idlestate_disable(unsigned int cpu, unsigned int idlestate, + unsigned int disable); extern unsigned long sysfs_get_idlestate_latency(unsigned int cpu, unsigned int idlestate); extern unsigned long sysfs_get_idlestate_usage(unsigned int cpu, @@ -19,7 +27,7 @@ extern char *sysfs_get_idlestate_name(unsigned int cpu, unsigned int idlestate); extern char *sysfs_get_idlestate_desc(unsigned int cpu, unsigned int idlestate); -extern int sysfs_get_idlestate_count(unsigned int cpu); +extern unsigned int sysfs_get_idlestate_count(unsigned int cpu); extern char *sysfs_get_cpuidle_governor(void); extern char *sysfs_get_cpuidle_driver(void); diff --git a/tools/power/cpupower/utils/helpers/topology.c b/tools/power/cpupower/utils/helpers/topology.c index 4eae2c47ba4..c13120af519 100644 --- a/tools/power/cpupower/utils/helpers/topology.c +++ b/tools/power/cpupower/utils/helpers/topology.c @@ -20,9 +20,8 @@ #include <helpers/sysfs.h> /* returns -1 on failure, 0 on success */ -int sysfs_topology_read_file(unsigned int cpu, const char *fname) +static int sysfs_topology_read_file(unsigned int cpu, const char *fname, int *result) { - unsigned long value; char linebuf[MAX_LINE_LEN]; char *endp; char path[SYSFS_PATH_MAX]; @@ -31,20 +30,12 @@ int sysfs_topology_read_file(unsigned int cpu, const char *fname) cpu, fname); if (sysfs_read_file(path, linebuf, MAX_LINE_LEN) == 0) return -1; - value = strtoul(linebuf, &endp, 0); + *result = strtol(linebuf, &endp, 0); if (endp == linebuf || errno == ERANGE) return -1; - return value; + return 0; } -struct cpuid_core_info { - unsigned int pkg; - unsigned int thread; - unsigned int cpu; - /* flags */ - unsigned int is_online:1; -}; - static int __compare(const void *t1, const void *t2) { struct cpuid_core_info *top1 = (struct cpuid_core_info *)t1; @@ -53,9 +44,9 @@ static int __compare(const void *t1, const void *t2) return -1; else if (top1->pkg > top2->pkg) return 1; - else if (top1->thread < top2->thread) + else if (top1->core < top2->core) return -1; - else if (top1->thread > top2->thread) + else if (top1->core > top2->core) return 1; else if (top1->cpu < top2->cpu) return -1; @@ -73,28 +64,42 @@ static int __compare(const void *t1, const void *t2) */ int get_cpu_topology(struct cpupower_topology *cpu_top) { - int cpu, cpus = sysconf(_SC_NPROCESSORS_CONF); + int cpu, last_pkg, cpus = sysconf(_SC_NPROCESSORS_CONF); - cpu_top->core_info = malloc(sizeof(struct cpupower_topology) * cpus); + cpu_top->core_info = malloc(sizeof(struct cpuid_core_info) * cpus); if (cpu_top->core_info == NULL) return -ENOMEM; cpu_top->pkgs = cpu_top->cores = 0; for (cpu = 0; cpu < cpus; cpu++) { cpu_top->core_info[cpu].cpu = cpu; cpu_top->core_info[cpu].is_online = sysfs_is_cpu_online(cpu); - cpu_top->core_info[cpu].pkg = - sysfs_topology_read_file(cpu, "physical_package_id"); - if ((int)cpu_top->core_info[cpu].pkg != -1 && - cpu_top->core_info[cpu].pkg > cpu_top->pkgs) - cpu_top->pkgs = cpu_top->core_info[cpu].pkg; - cpu_top->core_info[cpu].core = - sysfs_topology_read_file(cpu, "core_id"); + if(sysfs_topology_read_file( + cpu, + "physical_package_id", + &(cpu_top->core_info[cpu].pkg)) < 0) + return -1; + if(sysfs_topology_read_file( + cpu, + "core_id", + &(cpu_top->core_info[cpu].core)) < 0) + return -1; } - cpu_top->pkgs++; qsort(cpu_top->core_info, cpus, sizeof(struct cpuid_core_info), __compare); + /* Count the number of distinct pkgs values. This works + because the primary sort of the core_info struct was just + done by pkg value. */ + last_pkg = cpu_top->core_info[0].pkg; + for(cpu = 1; cpu < cpus; cpu++) { + if(cpu_top->core_info[cpu].pkg != last_pkg) { + last_pkg = cpu_top->core_info[cpu].pkg; + cpu_top->pkgs++; + } + } + cpu_top->pkgs++; + /* Intel's cores count is not consecutively numbered, there may * be a core_id of 3, but none of 2. Assume there always is 0 * Get amount of cores by counting duplicates in a package diff --git a/tools/power/cpupower/utils/idle_monitor/amd_fam14h_idle.c b/tools/power/cpupower/utils/idle_monitor/amd_fam14h_idle.c index 202e555988b..2116df9ad83 100644 --- a/tools/power/cpupower/utils/idle_monitor/amd_fam14h_idle.c +++ b/tools/power/cpupower/utils/idle_monitor/amd_fam14h_idle.c @@ -20,8 +20,6 @@ #include "idle_monitor/cpupower-monitor.h" #include "helpers/helpers.h" -/******** PCI parts could go into own file and get shared ***************/ - #define PCI_NON_PC0_OFFSET 0xb0 #define PCI_PC1_OFFSET 0xb4 #define PCI_PC6_OFFSET 0xb8 @@ -82,10 +80,7 @@ static cstate_t amd_fam14h_cstates[AMD_FAM14H_STATE_NUM] = { }; static struct pci_access *pci_acc; -static int pci_vendor_id = 0x1022; -static int pci_dev_ids[2] = {0x1716, 0}; static struct pci_dev *amd_fam14h_pci_dev; - static int nbp1_entered; struct timespec start_time; @@ -286,13 +281,13 @@ struct cpuidle_monitor *amd_fam14h_register(void) if (cpupower_cpu_info.vendor != X86_VENDOR_AMD) return NULL; - if (cpupower_cpu_info.family == 0x14) { - if (cpu_count <= 0 || cpu_count > 2) { - fprintf(stderr, "AMD fam14h: Invalid cpu count: %d\n", - cpu_count); - return NULL; - } - } else + if (cpupower_cpu_info.family == 0x14) + strncpy(amd_fam14h_monitor.name, "Fam_14h", + MONITOR_NAME_LEN - 1); + else if (cpupower_cpu_info.family == 0x12) + strncpy(amd_fam14h_monitor.name, "Fam_12h", + MONITOR_NAME_LEN - 1); + else return NULL; /* We do not alloc for nbp1 machine wide counter */ @@ -303,7 +298,9 @@ struct cpuidle_monitor *amd_fam14h_register(void) sizeof(unsigned long long)); } - amd_fam14h_pci_dev = pci_acc_init(&pci_acc, pci_vendor_id, pci_dev_ids); + /* We need PCI device: Slot 18, Func 6, compare with BKDG + for fam 12h/14h */ + amd_fam14h_pci_dev = pci_slot_func_init(&pci_acc, 0x18, 6); if (amd_fam14h_pci_dev == NULL || pci_acc == NULL) return NULL; @@ -325,7 +322,7 @@ static void amd_fam14h_unregister(void) } struct cpuidle_monitor amd_fam14h_monitor = { - .name = "Ontario", + .name = "", .hw_states = amd_fam14h_cstates, .hw_states_num = AMD_FAM14H_STATE_NUM, .start = amd_fam14h_start, diff --git a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c index 0d6571e418d..c4bae9203a6 100644 --- a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c +++ b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c @@ -39,6 +39,7 @@ static int mode; static int interval = 1; static char *show_monitors_param; static struct cpupower_topology cpu_top; +static unsigned int wake_cpus; /* ToDo: Document this in the manpage */ static char range_abbr[RANGE_MAX] = { 'T', 'C', 'P', 'M', }; @@ -84,7 +85,7 @@ int fill_string_with_spaces(char *s, int n) void print_header(int topology_depth) { int unsigned mon; - int state, need_len, pr_mon_len; + int state, need_len; cstate_t s; char buf[128] = ""; int percent_width = 4; @@ -93,7 +94,6 @@ void print_header(int topology_depth) printf("%s|", buf); for (mon = 0; mon < avail_monitors; mon++) { - pr_mon_len = 0; need_len = monitors[mon]->hw_states_num * (percent_width + 3) - 1; if (mon != 0) { @@ -315,16 +315,28 @@ int fork_it(char **argv) int do_interval_measure(int i) { unsigned int num; + int cpu; + + if (wake_cpus) + for (cpu = 0; cpu < cpu_count; cpu++) + bind_cpu(cpu); for (num = 0; num < avail_monitors; num++) { dprint("HW C-state residency monitor: %s - States: %d\n", monitors[num]->name, monitors[num]->hw_states_num); monitors[num]->start(); } + sleep(i); + + if (wake_cpus) + for (cpu = 0; cpu < cpu_count; cpu++) + bind_cpu(cpu); + for (num = 0; num < avail_monitors; num++) monitors[num]->stop(); + return 0; } @@ -333,7 +345,7 @@ static void cmdline(int argc, char *argv[]) int opt; progname = basename(argv[0]); - while ((opt = getopt(argc, argv, "+li:m:")) != -1) { + while ((opt = getopt(argc, argv, "+lci:m:")) != -1) { switch (opt) { case 'l': if (mode) @@ -352,6 +364,9 @@ static void cmdline(int argc, char *argv[]) mode = show; show_monitors_param = optarg; break; + case 'c': + wake_cpus = 1; + break; default: print_wrong_arg_exit(); } diff --git a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.h b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.h index 9312ee1f2db..9e43f3371fb 100644 --- a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.h +++ b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.h @@ -65,4 +65,21 @@ extern long long timespec_diff_us(struct timespec start, struct timespec end); "could be inaccurate\n"), mes, ov); \ } + +/* Taken over from x86info project sources -> return 0 on success */ +#include <sched.h> +#include <sys/types.h> +#include <unistd.h> +static inline int bind_cpu(int cpu) +{ + cpu_set_t set; + + if (sched_getaffinity(getpid(), sizeof(set), &set) == 0) { + CPU_ZERO(&set); + CPU_SET(cpu, &set); + return sched_setaffinity(getpid(), sizeof(set), &set); + } + return 1; +} + #endif /* __CPUIDLE_INFO_HW__ */ diff --git a/tools/power/cpupower/utils/idle_monitor/hsw_ext_idle.c b/tools/power/cpupower/utils/idle_monitor/hsw_ext_idle.c new file mode 100644 index 00000000000..ebeaba6571a --- /dev/null +++ b/tools/power/cpupower/utils/idle_monitor/hsw_ext_idle.c @@ -0,0 +1,196 @@ +/* + * (C) 2010,2011 Thomas Renninger <trenn@suse.de>, Novell Inc. + * + * Licensed under the terms of the GNU GPL License version 2. + * + * Based on SandyBridge monitor. Implements the new package C-states + * (PC8, PC9, PC10) coming with a specific Haswell (family 0x45) CPU. + */ + +#if defined(__i386__) || defined(__x86_64__) + +#include <stdio.h> +#include <stdint.h> +#include <stdlib.h> +#include <string.h> + +#include "helpers/helpers.h" +#include "idle_monitor/cpupower-monitor.h" + +#define MSR_PKG_C8_RESIDENCY 0x00000630 +#define MSR_PKG_C9_RESIDENCY 0x00000631 +#define MSR_PKG_C10_RESIDENCY 0x00000632 + +#define MSR_TSC 0x10 + +enum intel_hsw_ext_id { PC8 = 0, PC9, PC10, HSW_EXT_CSTATE_COUNT, + TSC = 0xFFFF }; + +static int hsw_ext_get_count_percent(unsigned int self_id, double *percent, + unsigned int cpu); + +static cstate_t hsw_ext_cstates[HSW_EXT_CSTATE_COUNT] = { + { + .name = "PC8", + .desc = N_("Processor Package C8"), + .id = PC8, + .range = RANGE_PACKAGE, + .get_count_percent = hsw_ext_get_count_percent, + }, + { + .name = "PC9", + .desc = N_("Processor Package C9"), + .desc = N_("Processor Package C2"), + .id = PC9, + .range = RANGE_PACKAGE, + .get_count_percent = hsw_ext_get_count_percent, + }, + { + .name = "PC10", + .desc = N_("Processor Package C10"), + .id = PC10, + .range = RANGE_PACKAGE, + .get_count_percent = hsw_ext_get_count_percent, + }, +}; + +static unsigned long long tsc_at_measure_start; +static unsigned long long tsc_at_measure_end; +static unsigned long long *previous_count[HSW_EXT_CSTATE_COUNT]; +static unsigned long long *current_count[HSW_EXT_CSTATE_COUNT]; +/* valid flag for all CPUs. If a MSR read failed it will be zero */ +static int *is_valid; + +static int hsw_ext_get_count(enum intel_hsw_ext_id id, unsigned long long *val, + unsigned int cpu) +{ + int msr; + + switch (id) { + case PC8: + msr = MSR_PKG_C8_RESIDENCY; + break; + case PC9: + msr = MSR_PKG_C9_RESIDENCY; + break; + case PC10: + msr = MSR_PKG_C10_RESIDENCY; + break; + case TSC: + msr = MSR_TSC; + break; + default: + return -1; + }; + if (read_msr(cpu, msr, val)) + return -1; + return 0; +} + +static int hsw_ext_get_count_percent(unsigned int id, double *percent, + unsigned int cpu) +{ + *percent = 0.0; + + if (!is_valid[cpu]) + return -1; + + *percent = (100.0 * + (current_count[id][cpu] - previous_count[id][cpu])) / + (tsc_at_measure_end - tsc_at_measure_start); + + dprint("%s: previous: %llu - current: %llu - (%u)\n", + hsw_ext_cstates[id].name, previous_count[id][cpu], + current_count[id][cpu], cpu); + + dprint("%s: tsc_diff: %llu - count_diff: %llu - percent: %2.f (%u)\n", + hsw_ext_cstates[id].name, + (unsigned long long) tsc_at_measure_end - tsc_at_measure_start, + current_count[id][cpu] - previous_count[id][cpu], + *percent, cpu); + + return 0; +} + +static int hsw_ext_start(void) +{ + int num, cpu; + unsigned long long val; + + for (num = 0; num < HSW_EXT_CSTATE_COUNT; num++) { + for (cpu = 0; cpu < cpu_count; cpu++) { + hsw_ext_get_count(num, &val, cpu); + previous_count[num][cpu] = val; + } + } + hsw_ext_get_count(TSC, &tsc_at_measure_start, 0); + return 0; +} + +static int hsw_ext_stop(void) +{ + unsigned long long val; + int num, cpu; + + hsw_ext_get_count(TSC, &tsc_at_measure_end, 0); + + for (num = 0; num < HSW_EXT_CSTATE_COUNT; num++) { + for (cpu = 0; cpu < cpu_count; cpu++) { + is_valid[cpu] = !hsw_ext_get_count(num, &val, cpu); + current_count[num][cpu] = val; + } + } + return 0; +} + +struct cpuidle_monitor intel_hsw_ext_monitor; + +static struct cpuidle_monitor *hsw_ext_register(void) +{ + int num; + + if (cpupower_cpu_info.vendor != X86_VENDOR_INTEL + || cpupower_cpu_info.family != 6) + return NULL; + + switch (cpupower_cpu_info.model) { + case 0x45: /* HSW */ + break; + default: + return NULL; + } + + is_valid = calloc(cpu_count, sizeof(int)); + for (num = 0; num < HSW_EXT_CSTATE_COUNT; num++) { + previous_count[num] = calloc(cpu_count, + sizeof(unsigned long long)); + current_count[num] = calloc(cpu_count, + sizeof(unsigned long long)); + } + intel_hsw_ext_monitor.name_len = strlen(intel_hsw_ext_monitor.name); + return &intel_hsw_ext_monitor; +} + +void hsw_ext_unregister(void) +{ + int num; + free(is_valid); + for (num = 0; num < HSW_EXT_CSTATE_COUNT; num++) { + free(previous_count[num]); + free(current_count[num]); + } +} + +struct cpuidle_monitor intel_hsw_ext_monitor = { + .name = "HaswellExtended", + .hw_states = hsw_ext_cstates, + .hw_states_num = HSW_EXT_CSTATE_COUNT, + .start = hsw_ext_start, + .stop = hsw_ext_stop, + .do_register = hsw_ext_register, + .unregister = hsw_ext_unregister, + .needs_root = 1, + .overflow_s = 922000000 /* 922337203 seconds TSC overflow + at 20GHz */ +}; +#endif /* defined(__i386__) || defined(__x86_64__) */ diff --git a/tools/power/cpupower/utils/idle_monitor/idle_monitors.def b/tools/power/cpupower/utils/idle_monitor/idle_monitors.def index e3f8d9b2b18..0d6ba4dbb9c 100644 --- a/tools/power/cpupower/utils/idle_monitor/idle_monitors.def +++ b/tools/power/cpupower/utils/idle_monitor/idle_monitors.def @@ -2,6 +2,7 @@ DEF(amd_fam14h) DEF(intel_nhm) DEF(intel_snb) +DEF(intel_hsw_ext) DEF(mperf) #endif DEF(cpuidle_sysfs) diff --git a/tools/power/cpupower/utils/idle_monitor/snb_idle.c b/tools/power/cpupower/utils/idle_monitor/snb_idle.c index a1bc07cd53e..efc8a69c9ab 100644 --- a/tools/power/cpupower/utils/idle_monitor/snb_idle.c +++ b/tools/power/cpupower/utils/idle_monitor/snb_idle.c @@ -150,9 +150,19 @@ static struct cpuidle_monitor *snb_register(void) || cpupower_cpu_info.family != 6) return NULL; - if (cpupower_cpu_info.model != 0x2A - && cpupower_cpu_info.model != 0x2D) + switch (cpupower_cpu_info.model) { + case 0x2A: /* SNB */ + case 0x2D: /* SNB Xeon */ + case 0x3A: /* IVB */ + case 0x3E: /* IVB Xeon */ + case 0x3C: /* HSW */ + case 0x3F: /* HSW */ + case 0x45: /* HSW */ + case 0x46: /* HSW */ + break; + default: return NULL; + } is_valid = calloc(cpu_count, sizeof(int)); for (num = 0; num < SNB_CSTATE_COUNT; num++) { diff --git a/tools/power/x86/turbostat/.gitignore b/tools/power/x86/turbostat/.gitignore new file mode 100644 index 00000000000..7521370d356 --- /dev/null +++ b/tools/power/x86/turbostat/.gitignore @@ -0,0 +1 @@ +turbostat diff --git a/tools/power/x86/turbostat/Makefile b/tools/power/x86/turbostat/Makefile index fd8e1f1297a..d1b3a361e52 100644 --- a/tools/power/x86/turbostat/Makefile +++ b/tools/power/x86/turbostat/Makefile @@ -1,8 +1,22 @@ +CC = $(CROSS_COMPILE)gcc +BUILD_OUTPUT := $(PWD) +PREFIX := /usr +DESTDIR := + turbostat : turbostat.c +CFLAGS += -Wall +CFLAGS += -DMSRHEADER='"../../../../arch/x86/include/uapi/asm/msr-index.h"' + +%: %.c + @mkdir -p $(BUILD_OUTPUT) + $(CC) $(CFLAGS) $< -o $(BUILD_OUTPUT)/$@ +.PHONY : clean clean : - rm -f turbostat + @rm -f $(BUILD_OUTPUT)/turbostat -install : - install turbostat /usr/bin/turbostat - install turbostat.8 /usr/share/man/man8 +install : turbostat + install -d $(DESTDIR)$(PREFIX)/bin + install $(BUILD_OUTPUT)/turbostat $(DESTDIR)$(PREFIX)/bin/turbostat + install -d $(DESTDIR)$(PREFIX)/share/man/man8 + install turbostat.8 $(DESTDIR)$(PREFIX)/share/man/man8 diff --git a/tools/power/x86/turbostat/turbostat.8 b/tools/power/x86/turbostat/turbostat.8 index 555c69a5592..56bfb523c5b 100644 --- a/tools/power/x86/turbostat/turbostat.8 +++ b/tools/power/x86/turbostat/turbostat.8 @@ -4,31 +4,40 @@ turbostat \- Report processor frequency and idle statistics .SH SYNOPSIS .ft B .B turbostat -.RB [ "\-v" ] -.RB [ "\-M MSR#" ] +.RB [ Options ] .RB command .br .B turbostat -.RB [ "\-v" ] -.RB [ "\-M MSR#" ] +.RB [ Options ] .RB [ "\-i interval_sec" ] .SH DESCRIPTION -\fBturbostat \fP reports processor topology, frequency -and idle power state statistics on modern X86 processors. +\fBturbostat \fP reports processor topology, frequency, +idle power-state statistics, temperature and power on modern X86 processors. Either \fBcommand\fP is forked and statistics are printed upon its completion, or statistics are printed periodically. \fBturbostat \fP -requires that the processor +must be run on root, and +minimally requires that the processor supports an "invariant" TSC, plus the APERF and MPERF MSRs. -\fBturbostat \fP will report idle cpu power state residency -on processors that additionally support C-state residency counters. +Additional information is reported depending on hardware counter support. .SS Options +The \fB-p\fP option limits output to the 1st thread in 1st core of each package. +.PP +The \fB-P\fP option limits output to the 1st thread in each Package. +.PP +The \fB-S\fP option limits output to a 1-line System Summary for each interval. +.PP The \fB-v\fP option increases verbosity. .PP -The \fB-M MSR#\fP option dumps the specified MSR, -in addition to the usual frequency and idle statistics. +The \fB-c MSR#\fP option includes the delta of the specified 32-bit MSR counter. +.PP +The \fB-C MSR#\fP option includes the delta of the specified 64-bit MSR counter. +.PP +The \fB-m MSR#\fP option includes the the specified 32-bit MSR value. +.PP +The \fB-M MSR#\fP option includes the the specified 64-bit MSR value. .PP The \fB-i interval_sec\fP option prints statistics every \fiinterval_sec\fP seconds. The default is 5 seconds. @@ -38,14 +47,24 @@ displays the statistics gathered since it was forked. .PP .SH FIELD DESCRIPTIONS .nf -\fBpk\fP processor package number. -\fBcr\fP processor core number. +\fBPackage\fP processor package number. +\fBCore\fP processor core number. \fBCPU\fP Linux CPU (logical processor) number. -\fB%c0\fP percent of the interval that the CPU retired instructions. -\fBGHz\fP average clock rate while the CPU was in c0 state. -\fBTSC\fP average GHz that the TSC ran during the entire interval. -\fB%c1, %c3, %c6\fP show the percentage residency in hardware core idle states. -\fB%pc3, %pc6\fP percentage residency in hardware package idle states. +Note that multiple CPUs per core indicate support for Intel(R) Hyper-Threading Technology. +\fBAVG_MHz\fP number of cycles executed divided by time elapsed. +\fB%Buzy\fP percent of the interval that the CPU retired instructions, aka. % of time in "C0" state. +\fBBzy_MHz\fP average clock rate while the CPU was busy (in "c0" state). +\fBTSC_MHz\fP average MHz that the TSC ran during the entire interval. +\fBCPU%c1, CPU%c3, CPU%c6, CPU%c7\fP show the percentage residency in hardware core idle states. +\fBCoreTmp\fP Degrees Celsius reported by the per-core Digital Thermal Sensor. +\fBPkgTtmp\fP Degrees Celsius reported by the per-package Package Thermal Monitor. +\fBPkg%pc2, Pkg%pc3, Pkg%pc6, Pkg%pc7\fP percentage residency in hardware package idle states. +\fBPkgWatt\fP Watts consumed by the whole package. +\fBCorWatt\fP Watts consumed by the core part of the package. +\fBGFXWatt\fP Watts consumed by the Graphics part of the package -- available only on client processors. +\fBRAMWatt\fP Watts consumed by the DRAM DIMMS -- available only on server processors. +\fBPKG_%\fP percent of the interval that RAPL throttling was active on the Package. +\fBRAM_%\fP percent of the interval that RAPL throttling was active on DRAM. .fi .PP .SH EXAMPLE @@ -53,38 +72,62 @@ Without any parameters, turbostat prints out counters ever 5 seconds. (override interval with "-i sec" option, or specify a command for turbostat to fork). -The first row of statistics reflect the average for the entire system. +The first row of statistics is a summary for the entire system. +For residency % columns, the summary is a weighted average. +For Temperature columns, the summary is the column maximum. +For Watts columns, the summary is a system total. Subsequent rows show per-CPU statistics. .nf -[root@x980]# ./turbostat -cr CPU %c0 GHz TSC %c1 %c3 %c6 %pc3 %pc6 - 0.04 1.62 3.38 0.11 0.00 99.85 0.00 95.07 - 0 0 0.04 1.62 3.38 0.06 0.00 99.90 0.00 95.07 - 0 6 0.02 1.62 3.38 0.08 0.00 99.90 0.00 95.07 - 1 2 0.10 1.62 3.38 0.29 0.00 99.61 0.00 95.07 - 1 8 0.11 1.62 3.38 0.28 0.00 99.61 0.00 95.07 - 2 4 0.01 1.62 3.38 0.01 0.00 99.98 0.00 95.07 - 2 10 0.01 1.61 3.38 0.02 0.00 99.98 0.00 95.07 - 8 1 0.07 1.62 3.38 0.15 0.00 99.78 0.00 95.07 - 8 7 0.03 1.62 3.38 0.19 0.00 99.78 0.00 95.07 - 9 3 0.01 1.62 3.38 0.02 0.00 99.98 0.00 95.07 - 9 9 0.01 1.62 3.38 0.02 0.00 99.98 0.00 95.07 - 10 5 0.01 1.62 3.38 0.13 0.00 99.86 0.00 95.07 - 10 11 0.08 1.62 3.38 0.05 0.00 99.86 0.00 95.07 +[root@ivy]# ./turbostat + Core CPU Avg_MHz %Busy Bzy_MHz TSC_MHz SMI CPU%c1 CPU%c3 CPU%c6 CPU%c7 CoreTmp PkgTmp Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 PkgWatt CorWatt GFXWatt + - - 6 0.36 1596 3492 0 0.59 0.01 99.04 0.00 23 24 23.82 0.01 72.47 0.00 6.40 1.01 0.00 + 0 0 9 0.58 1596 3492 0 0.28 0.01 99.13 0.00 23 24 23.82 0.01 72.47 0.00 6.40 1.01 0.00 + 0 4 1 0.07 1596 3492 0 0.79 + 1 1 10 0.65 1596 3492 0 0.59 0.00 98.76 0.00 23 + 1 5 5 0.28 1596 3492 0 0.95 + 2 2 10 0.66 1596 3492 0 0.41 0.01 98.92 0.00 23 + 2 6 2 0.10 1597 3492 0 0.97 + 3 3 3 0.20 1596 3492 0 0.44 0.00 99.37 0.00 23 + 3 7 5 0.31 1596 3492 0 0.33 .fi .SH VERBOSE EXAMPLE The "-v" option adds verbosity to the output: .nf -GenuineIntel 11 CPUID levels; family:model:stepping 0x6:2c:2 (6:44:2) -12 * 133 = 1600 MHz max efficiency -25 * 133 = 3333 MHz TSC frequency -26 * 133 = 3467 MHz max turbo 4 active cores -26 * 133 = 3467 MHz max turbo 3 active cores -27 * 133 = 3600 MHz max turbo 2 active cores -27 * 133 = 3600 MHz max turbo 1 active cores - +[root@ivy]# turbostat -v +turbostat v3.0 November 23, 2012 - Len Brown <lenb@kernel.org> +CPUID(0): GenuineIntel 13 CPUID levels; family:model:stepping 0x6:3a:9 (6:58:9) +CPUID(6): APERF, DTS, PTM, EPB +RAPL: 851 sec. Joule Counter Range +cpu0: MSR_NHM_PLATFORM_INFO: 0x81010f0012300 +16 * 100 = 1600 MHz max efficiency +35 * 100 = 3500 MHz TSC frequency +cpu0: MSR_NHM_SNB_PKG_CST_CFG_CTL: 0x1e008402 (UNdemote-C3, UNdemote-C1, demote-C3, demote-C1, locked: pkg-cstate-limit=2: pc6-noret) +cpu0: MSR_NHM_TURBO_RATIO_LIMIT: 0x25262727 +37 * 100 = 3700 MHz max turbo 4 active cores +38 * 100 = 3800 MHz max turbo 3 active cores +39 * 100 = 3900 MHz max turbo 2 active cores +39 * 100 = 3900 MHz max turbo 1 active cores +cpu0: MSR_IA32_ENERGY_PERF_BIAS: 0x00000006 (balanced) +cpu0: MSR_RAPL_POWER_UNIT: 0x000a1003 (0.125000 Watts, 0.000015 Joules, 0.000977 sec.) +cpu0: MSR_PKG_POWER_INFO: 0x01e00268 (77 W TDP, RAPL 60 - 0 W, 0.000000 sec.) +cpu0: MSR_PKG_POWER_LIMIT: 0x830000148268 (UNlocked) +cpu0: PKG Limit #1: ENabled (77.000000 Watts, 1.000000 sec, clamp DISabled) +cpu0: PKG Limit #2: ENabled (96.000000 Watts, 0.000977* sec, clamp DISabled) +cpu0: MSR_PP0_POLICY: 0 +cpu0: MSR_PP0_POWER_LIMIT: 0x00000000 (UNlocked) +cpu0: Cores Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled) +cpu0: MSR_PP1_POLICY: 0 +cpu0: MSR_PP1_POWER_LIMIT: 0x00000000 (UNlocked) +cpu0: GFX Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled) +cpu0: MSR_IA32_TEMPERATURE_TARGET: 0x00691400 (105 C) +cpu0: MSR_IA32_PACKAGE_THERM_STATUS: 0x884e0000 (27 C) +cpu0: MSR_IA32_THERM_STATUS: 0x88560000 (19 C +/- 1) +cpu1: MSR_IA32_THERM_STATUS: 0x88560000 (19 C +/- 1) +cpu2: MSR_IA32_THERM_STATUS: 0x88540000 (21 C +/- 1) +cpu3: MSR_IA32_THERM_STATUS: 0x884e0000 (27 C +/- 1) + ... .fi The \fBmax efficiency\fP frequency, a.k.a. Low Frequency Mode, is the frequency available at the minimum package voltage. The \fBTSC frequency\fP is the nominal @@ -100,36 +143,35 @@ eg. Here a cycle soaker is run on 1 CPU (see %c0) for a few seconds until ^C while the other CPUs are mostly idle: .nf -[root@x980 lenb]# ./turbostat cat /dev/zero > /dev/null - -^Ccr CPU %c0 GHz TSC %c1 %c3 %c6 %pc3 %pc6 - 8.49 3.63 3.38 16.23 0.66 74.63 0.00 0.00 - 0 0 1.22 3.62 3.38 32.18 0.00 66.60 0.00 0.00 - 0 6 0.40 3.61 3.38 33.00 0.00 66.60 0.00 0.00 - 1 2 0.11 3.14 3.38 0.19 3.95 95.75 0.00 0.00 - 1 8 0.05 2.88 3.38 0.25 3.95 95.75 0.00 0.00 - 2 4 0.00 3.13 3.38 0.02 0.00 99.98 0.00 0.00 - 2 10 0.00 3.09 3.38 0.02 0.00 99.98 0.00 0.00 - 8 1 0.04 3.50 3.38 14.43 0.00 85.54 0.00 0.00 - 8 7 0.03 2.98 3.38 14.43 0.00 85.54 0.00 0.00 - 9 3 0.00 3.16 3.38 100.00 0.00 0.00 0.00 0.00 - 9 9 99.93 3.63 3.38 0.06 0.00 0.00 0.00 0.00 - 10 5 0.01 2.82 3.38 0.08 0.00 99.91 0.00 0.00 - 10 11 0.02 3.36 3.38 0.06 0.00 99.91 0.00 0.00 -6.950866 sec +root@ivy: turbostat cat /dev/zero > /dev/null +^C + Core CPU Avg_MHz %Busy Bzy_MHz TSC_MHz SMI CPU%c1 CPU%c3 CPU%c6 CPU%c7 CoreTmp PkgTmp Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 PkgWatt CorWatt GFXWatt + - - 496 12.75 3886 3492 0 13.16 0.04 74.04 0.00 36 36 0.00 0.00 0.00 0.00 23.15 17.65 0.00 + 0 0 22 0.57 3830 3492 0 0.83 0.02 98.59 0.00 27 36 0.00 0.00 0.00 0.00 23.15 17.65 0.00 + 0 4 9 0.24 3829 3492 0 1.15 + 1 1 4 0.09 3783 3492 0 99.91 0.00 0.00 0.00 36 + 1 5 3880 99.82 3888 3492 0 0.18 + 2 2 17 0.44 3813 3492 0 0.77 0.04 98.75 0.00 28 + 2 6 12 0.32 3823 3492 0 0.89 + 3 3 16 0.43 3844 3492 0 0.63 0.11 98.84 0.00 30 + 3 7 4 0.11 3827 3492 0 0.94 +30.372243 sec .fi -Above the cycle soaker drives cpu9 up 3.6 Ghz turbo limit +Above the cycle soaker drives cpu5 up its 3.8 GHz turbo limit while the other processors are generally in various states of idle. -Note that cpu3 is an HT sibling sharing core9 -with cpu9, and thus it is unable to get to an idle state -deeper than c1 while cpu9 is busy. +Note that cpu1 and cpu5 are HT siblings within core1. +As cpu5 is very busy, it prevents its sibling, cpu1, +from entering a c-state deeper than c1. + +Note that the Avg_MHz column reflects the total number of cycles executed +divided by the measurement interval. If the %Busy column is 100%, +then the processor was running at that speed the entire interval. +The Avg_MHz multiplied by the %Busy results in the Bzy_MHz -- +which is the average frequency while the processor was executing -- +not including any non-busy idle time. -Note that turbostat reports average GHz of 3.61, while -the arithmetic average of the GHz column above is 3.24. -This is a weighted average, where the weight is %c0. ie. it is the total number of -un-halted cycles elapsed per time divided by the number of CPUs. .SH NOTES .B "turbostat " @@ -145,6 +187,13 @@ may work poorly on Linux-2.6.20 through 2.6.29, as \fBacpi-cpufreq \fPperiodically cleared the APERF and MPERF in those kernels. +If the TSC column does not make sense, then +the other numbers will also make no sense. +Turbostat is lightweight, and its data collection is not atomic. +These issues are usually caused by an extremely short measurement +interval (much less than 1 second), or system activity that prevents +turbostat from being able to run on all CPUS to quickly collect data. + The APERF, MPERF MSRs are defined to count non-halted cycles. Although it is not guaranteed by the architecture, turbostat assumes that they count at TSC rate, which is true on all processors tested to date. @@ -167,6 +216,6 @@ http://www.intel.com/products/processor/manuals/ .SH "SEE ALSO" msr(4), vmstat(8) .PP -.SH AUTHORS +.SH AUTHOR .nf Written by Len Brown <len.brown@intel.com> diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index 310d3dd5e54..d0396af99fa 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -2,7 +2,7 @@ * turbostat -- show CPU frequency and C-state residency * on modern Intel turbo-capable processors. * - * Copyright (c) 2010, Intel Corporation. + * Copyright (c) 2013 Intel Corporation. * Len Brown <len.brown@intel.com> * * This program is free software; you can redistribute it and/or modify it @@ -19,7 +19,11 @@ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. */ +#define _GNU_SOURCE +#include MSRHEADER +#include <stdarg.h> #include <stdio.h> +#include <err.h> #include <unistd.h> #include <sys/types.h> #include <sys/wait.h> @@ -32,442 +36,985 @@ #include <dirent.h> #include <string.h> #include <ctype.h> - -#define MSR_TSC 0x10 -#define MSR_NEHALEM_PLATFORM_INFO 0xCE -#define MSR_NEHALEM_TURBO_RATIO_LIMIT 0x1AD -#define MSR_APERF 0xE8 -#define MSR_MPERF 0xE7 -#define MSR_PKG_C2_RESIDENCY 0x60D /* SNB only */ -#define MSR_PKG_C3_RESIDENCY 0x3F8 -#define MSR_PKG_C6_RESIDENCY 0x3F9 -#define MSR_PKG_C7_RESIDENCY 0x3FA /* SNB only */ -#define MSR_CORE_C3_RESIDENCY 0x3FC -#define MSR_CORE_C6_RESIDENCY 0x3FD -#define MSR_CORE_C7_RESIDENCY 0x3FE /* SNB only */ +#include <sched.h> +#include <cpuid.h> char *proc_stat = "/proc/stat"; unsigned int interval_sec = 5; /* set with -i interval_sec */ unsigned int verbose; /* set with -v */ +unsigned int rapl_verbose; /* set with -R */ +unsigned int rapl_joules; /* set with -J */ +unsigned int thermal_verbose; /* set with -T */ +unsigned int summary_only; /* set with -S */ +unsigned int dump_only; /* set with -s */ unsigned int skip_c0; unsigned int skip_c1; unsigned int do_nhm_cstates; unsigned int do_snb_cstates; +unsigned int do_c8_c9_c10; +unsigned int do_slm_cstates; +unsigned int use_c1_residency_msr; unsigned int has_aperf; -unsigned int units = 1000000000; /* Ghz etc */ +unsigned int has_epb; +unsigned int units = 1000000; /* MHz etc */ unsigned int genuine_intel; unsigned int has_invariant_tsc; unsigned int do_nehalem_platform_info; unsigned int do_nehalem_turbo_ratio_limit; -unsigned int extra_msr_offset; +unsigned int do_ivt_turbo_ratio_limit; +unsigned int extra_msr_offset32; +unsigned int extra_msr_offset64; +unsigned int extra_delta_offset32; +unsigned int extra_delta_offset64; +int do_smi; double bclk; unsigned int show_pkg; unsigned int show_core; unsigned int show_cpu; +unsigned int show_pkg_only; +unsigned int show_core_only; +char *output_buffer, *outp; +unsigned int do_rapl; +unsigned int do_dts; +unsigned int do_ptm; +unsigned int tcc_activation_temp; +unsigned int tcc_activation_temp_override; +double rapl_power_units, rapl_energy_units, rapl_time_units; +double rapl_joule_counter_range; + +#define RAPL_PKG (1 << 0) + /* 0x610 MSR_PKG_POWER_LIMIT */ + /* 0x611 MSR_PKG_ENERGY_STATUS */ +#define RAPL_PKG_PERF_STATUS (1 << 1) + /* 0x613 MSR_PKG_PERF_STATUS */ +#define RAPL_PKG_POWER_INFO (1 << 2) + /* 0x614 MSR_PKG_POWER_INFO */ + +#define RAPL_DRAM (1 << 3) + /* 0x618 MSR_DRAM_POWER_LIMIT */ + /* 0x619 MSR_DRAM_ENERGY_STATUS */ + /* 0x61c MSR_DRAM_POWER_INFO */ +#define RAPL_DRAM_PERF_STATUS (1 << 4) + /* 0x61b MSR_DRAM_PERF_STATUS */ + +#define RAPL_CORES (1 << 5) + /* 0x638 MSR_PP0_POWER_LIMIT */ + /* 0x639 MSR_PP0_ENERGY_STATUS */ +#define RAPL_CORE_POLICY (1 << 6) + /* 0x63a MSR_PP0_POLICY */ + + +#define RAPL_GFX (1 << 7) + /* 0x640 MSR_PP1_POWER_LIMIT */ + /* 0x641 MSR_PP1_ENERGY_STATUS */ + /* 0x642 MSR_PP1_POLICY */ +#define TJMAX_DEFAULT 100 + +#define MAX(a, b) ((a) > (b) ? (a) : (b)) int aperf_mperf_unstable; int backwards_count; char *progname; -int need_reinitialize; - -int num_cpus; - -struct counters { - unsigned long long tsc; /* per thread */ - unsigned long long aperf; /* per thread */ - unsigned long long mperf; /* per thread */ - unsigned long long c1; /* per thread (calculated) */ - unsigned long long c3; /* per core */ - unsigned long long c6; /* per core */ - unsigned long long c7; /* per core */ - unsigned long long pc2; /* per package */ - unsigned long long pc3; /* per package */ - unsigned long long pc6; /* per package */ - unsigned long long pc7; /* per package */ - unsigned long long extra_msr; /* per thread */ - int pkg; - int core; - int cpu; - struct counters *next; -}; -struct counters *cnt_even; -struct counters *cnt_odd; -struct counters *cnt_delta; -struct counters *cnt_average; -struct timeval tv_even; -struct timeval tv_odd; -struct timeval tv_delta; +cpu_set_t *cpu_present_set, *cpu_affinity_set; +size_t cpu_present_setsize, cpu_affinity_setsize; + +struct thread_data { + unsigned long long tsc; + unsigned long long aperf; + unsigned long long mperf; + unsigned long long c1; + unsigned long long extra_msr64; + unsigned long long extra_delta64; + unsigned long long extra_msr32; + unsigned long long extra_delta32; + unsigned int smi_count; + unsigned int cpu_id; + unsigned int flags; +#define CPU_IS_FIRST_THREAD_IN_CORE 0x2 +#define CPU_IS_FIRST_CORE_IN_PACKAGE 0x4 +} *thread_even, *thread_odd; + +struct core_data { + unsigned long long c3; + unsigned long long c6; + unsigned long long c7; + unsigned int core_temp_c; + unsigned int core_id; +} *core_even, *core_odd; + +struct pkg_data { + unsigned long long pc2; + unsigned long long pc3; + unsigned long long pc6; + unsigned long long pc7; + unsigned long long pc8; + unsigned long long pc9; + unsigned long long pc10; + unsigned int package_id; + unsigned int energy_pkg; /* MSR_PKG_ENERGY_STATUS */ + unsigned int energy_dram; /* MSR_DRAM_ENERGY_STATUS */ + unsigned int energy_cores; /* MSR_PP0_ENERGY_STATUS */ + unsigned int energy_gfx; /* MSR_PP1_ENERGY_STATUS */ + unsigned int rapl_pkg_perf_status; /* MSR_PKG_PERF_STATUS */ + unsigned int rapl_dram_perf_status; /* MSR_DRAM_PERF_STATUS */ + unsigned int pkg_temp_c; + +} *package_even, *package_odd; + +#define ODD_COUNTERS thread_odd, core_odd, package_odd +#define EVEN_COUNTERS thread_even, core_even, package_even + +#define GET_THREAD(thread_base, thread_no, core_no, pkg_no) \ + (thread_base + (pkg_no) * topo.num_cores_per_pkg * \ + topo.num_threads_per_core + \ + (core_no) * topo.num_threads_per_core + (thread_no)) +#define GET_CORE(core_base, core_no, pkg_no) \ + (core_base + (pkg_no) * topo.num_cores_per_pkg + (core_no)) +#define GET_PKG(pkg_base, pkg_no) (pkg_base + pkg_no) + +struct system_summary { + struct thread_data threads; + struct core_data cores; + struct pkg_data packages; +} sum, average; + + +struct topo_params { + int num_packages; + int num_cpus; + int num_cores; + int max_cpu_num; + int num_cores_per_pkg; + int num_threads_per_core; +} topo; + +struct timeval tv_even, tv_odd, tv_delta; + +void setup_all_buffers(void); + +int cpu_is_not_present(int cpu) +{ + return !CPU_ISSET_S(cpu, cpu_present_setsize, cpu_present_set); +} +/* + * run func(thread, core, package) in topology order + * skip non-present cpus + */ + +int for_all_cpus(int (func)(struct thread_data *, struct core_data *, struct pkg_data *), + struct thread_data *thread_base, struct core_data *core_base, struct pkg_data *pkg_base) +{ + int retval, pkg_no, core_no, thread_no; + + for (pkg_no = 0; pkg_no < topo.num_packages; ++pkg_no) { + for (core_no = 0; core_no < topo.num_cores_per_pkg; ++core_no) { + for (thread_no = 0; thread_no < + topo.num_threads_per_core; ++thread_no) { + struct thread_data *t; + struct core_data *c; + struct pkg_data *p; + + t = GET_THREAD(thread_base, thread_no, core_no, pkg_no); + + if (cpu_is_not_present(t->cpu_id)) + continue; -unsigned long long get_msr(int cpu, off_t offset) + c = GET_CORE(core_base, core_no, pkg_no); + p = GET_PKG(pkg_base, pkg_no); + + retval = func(t, c, p); + if (retval) + return retval; + } + } + } + return 0; +} + +int cpu_migrate(int cpu) +{ + CPU_ZERO_S(cpu_affinity_setsize, cpu_affinity_set); + CPU_SET_S(cpu, cpu_affinity_setsize, cpu_affinity_set); + if (sched_setaffinity(0, cpu_affinity_setsize, cpu_affinity_set) == -1) + return -1; + else + return 0; +} + +int get_msr(int cpu, off_t offset, unsigned long long *msr) { ssize_t retval; - unsigned long long msr; char pathname[32]; int fd; sprintf(pathname, "/dev/cpu/%d/msr", cpu); fd = open(pathname, O_RDONLY); - if (fd < 0) { - perror(pathname); - need_reinitialize = 1; - return 0; - } + if (fd < 0) + return -1; + + retval = pread(fd, msr, sizeof *msr, offset); + close(fd); - retval = pread(fd, &msr, sizeof msr, offset); - if (retval != sizeof msr) { - fprintf(stderr, "cpu%d pread(..., 0x%zx) = %jd\n", - cpu, offset, retval); - exit(-2); + if (retval != sizeof *msr) { + fprintf(stderr, "%s offset 0x%llx read failed\n", pathname, (unsigned long long)offset); + return -1; } - close(fd); - return msr; + return 0; } +/* + * Example Format w/ field column widths: + * + * Package Core CPU Avg_MHz Bzy_MHz TSC_MHz SMI %Busy CPU_%c1 CPU_%c3 CPU_%c6 CPU_%c7 CoreTmp PkgTmp Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 PkgWatt CorWatt GFXWatt + * 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 + */ + void print_header(void) { if (show_pkg) - fprintf(stderr, "pk"); + outp += sprintf(outp, "Package "); if (show_core) - fprintf(stderr, " cr"); + outp += sprintf(outp, " Core "); if (show_cpu) - fprintf(stderr, " CPU"); - if (do_nhm_cstates) - fprintf(stderr, " %%c0 "); + outp += sprintf(outp, " CPU "); if (has_aperf) - fprintf(stderr, " GHz"); - fprintf(stderr, " TSC"); + outp += sprintf(outp, "Avg_MHz "); if (do_nhm_cstates) - fprintf(stderr, " %%c1"); + outp += sprintf(outp, " %%Busy "); + if (has_aperf) + outp += sprintf(outp, "Bzy_MHz "); + outp += sprintf(outp, "TSC_MHz "); + if (do_smi) + outp += sprintf(outp, " SMI "); + if (extra_delta_offset32) + outp += sprintf(outp, " count 0x%03X ", extra_delta_offset32); + if (extra_delta_offset64) + outp += sprintf(outp, " COUNT 0x%03X ", extra_delta_offset64); + if (extra_msr_offset32) + outp += sprintf(outp, " MSR 0x%03X ", extra_msr_offset32); + if (extra_msr_offset64) + outp += sprintf(outp, " MSR 0x%03X ", extra_msr_offset64); if (do_nhm_cstates) - fprintf(stderr, " %%c3"); + outp += sprintf(outp, " CPU%%c1 "); + if (do_nhm_cstates && !do_slm_cstates) + outp += sprintf(outp, " CPU%%c3 "); if (do_nhm_cstates) - fprintf(stderr, " %%c6"); + outp += sprintf(outp, " CPU%%c6 "); if (do_snb_cstates) - fprintf(stderr, " %%c7"); + outp += sprintf(outp, " CPU%%c7 "); + + if (do_dts) + outp += sprintf(outp, "CoreTmp "); + if (do_ptm) + outp += sprintf(outp, " PkgTmp "); + if (do_snb_cstates) - fprintf(stderr, " %%pc2"); - if (do_nhm_cstates) - fprintf(stderr, " %%pc3"); - if (do_nhm_cstates) - fprintf(stderr, " %%pc6"); + outp += sprintf(outp, "Pkg%%pc2 "); + if (do_nhm_cstates && !do_slm_cstates) + outp += sprintf(outp, "Pkg%%pc3 "); + if (do_nhm_cstates && !do_slm_cstates) + outp += sprintf(outp, "Pkg%%pc6 "); if (do_snb_cstates) - fprintf(stderr, " %%pc7"); - if (extra_msr_offset) - fprintf(stderr, " MSR 0x%x ", extra_msr_offset); + outp += sprintf(outp, "Pkg%%pc7 "); + if (do_c8_c9_c10) { + outp += sprintf(outp, "Pkg%%pc8 "); + outp += sprintf(outp, "Pkg%%pc9 "); + outp += sprintf(outp, "Pk%%pc10 "); + } - putc('\n', stderr); -} + if (do_rapl && !rapl_joules) { + if (do_rapl & RAPL_PKG) + outp += sprintf(outp, "PkgWatt "); + if (do_rapl & RAPL_CORES) + outp += sprintf(outp, "CorWatt "); + if (do_rapl & RAPL_GFX) + outp += sprintf(outp, "GFXWatt "); + if (do_rapl & RAPL_DRAM) + outp += sprintf(outp, "RAMWatt "); + if (do_rapl & RAPL_PKG_PERF_STATUS) + outp += sprintf(outp, " PKG_%% "); + if (do_rapl & RAPL_DRAM_PERF_STATUS) + outp += sprintf(outp, " RAM_%% "); + } else { + if (do_rapl & RAPL_PKG) + outp += sprintf(outp, " Pkg_J "); + if (do_rapl & RAPL_CORES) + outp += sprintf(outp, " Cor_J "); + if (do_rapl & RAPL_GFX) + outp += sprintf(outp, " GFX_J "); + if (do_rapl & RAPL_DRAM) + outp += sprintf(outp, " RAM_W "); + if (do_rapl & RAPL_PKG_PERF_STATUS) + outp += sprintf(outp, " PKG_%% "); + if (do_rapl & RAPL_DRAM_PERF_STATUS) + outp += sprintf(outp, " RAM_%% "); + outp += sprintf(outp, " time "); -void dump_cnt(struct counters *cnt) -{ - if (!cnt) - return; - if (cnt->pkg) fprintf(stderr, "package: %d ", cnt->pkg); - if (cnt->core) fprintf(stderr, "core:: %d ", cnt->core); - if (cnt->cpu) fprintf(stderr, "CPU: %d ", cnt->cpu); - if (cnt->tsc) fprintf(stderr, "TSC: %016llX\n", cnt->tsc); - if (cnt->c3) fprintf(stderr, "c3: %016llX\n", cnt->c3); - if (cnt->c6) fprintf(stderr, "c6: %016llX\n", cnt->c6); - if (cnt->c7) fprintf(stderr, "c7: %016llX\n", cnt->c7); - if (cnt->aperf) fprintf(stderr, "aperf: %016llX\n", cnt->aperf); - if (cnt->pc2) fprintf(stderr, "pc2: %016llX\n", cnt->pc2); - if (cnt->pc3) fprintf(stderr, "pc3: %016llX\n", cnt->pc3); - if (cnt->pc6) fprintf(stderr, "pc6: %016llX\n", cnt->pc6); - if (cnt->pc7) fprintf(stderr, "pc7: %016llX\n", cnt->pc7); - if (cnt->extra_msr) fprintf(stderr, "msr0x%x: %016llX\n", extra_msr_offset, cnt->extra_msr); + } + outp += sprintf(outp, "\n"); } -void dump_list(struct counters *cnt) +int dump_counters(struct thread_data *t, struct core_data *c, + struct pkg_data *p) { - printf("dump_list 0x%p\n", cnt); + outp += sprintf(outp, "t %p, c %p, p %p\n", t, c, p); + + if (t) { + outp += sprintf(outp, "CPU: %d flags 0x%x\n", + t->cpu_id, t->flags); + outp += sprintf(outp, "TSC: %016llX\n", t->tsc); + outp += sprintf(outp, "aperf: %016llX\n", t->aperf); + outp += sprintf(outp, "mperf: %016llX\n", t->mperf); + outp += sprintf(outp, "c1: %016llX\n", t->c1); + outp += sprintf(outp, "msr0x%x: %08llX\n", + extra_delta_offset32, t->extra_delta32); + outp += sprintf(outp, "msr0x%x: %016llX\n", + extra_delta_offset64, t->extra_delta64); + outp += sprintf(outp, "msr0x%x: %08llX\n", + extra_msr_offset32, t->extra_msr32); + outp += sprintf(outp, "msr0x%x: %016llX\n", + extra_msr_offset64, t->extra_msr64); + if (do_smi) + outp += sprintf(outp, "SMI: %08X\n", t->smi_count); + } + + if (c) { + outp += sprintf(outp, "core: %d\n", c->core_id); + outp += sprintf(outp, "c3: %016llX\n", c->c3); + outp += sprintf(outp, "c6: %016llX\n", c->c6); + outp += sprintf(outp, "c7: %016llX\n", c->c7); + outp += sprintf(outp, "DTS: %dC\n", c->core_temp_c); + } - for (; cnt; cnt = cnt->next) - dump_cnt(cnt); + if (p) { + outp += sprintf(outp, "package: %d\n", p->package_id); + outp += sprintf(outp, "pc2: %016llX\n", p->pc2); + outp += sprintf(outp, "pc3: %016llX\n", p->pc3); + outp += sprintf(outp, "pc6: %016llX\n", p->pc6); + outp += sprintf(outp, "pc7: %016llX\n", p->pc7); + outp += sprintf(outp, "pc8: %016llX\n", p->pc8); + outp += sprintf(outp, "pc9: %016llX\n", p->pc9); + outp += sprintf(outp, "pc10: %016llX\n", p->pc10); + outp += sprintf(outp, "Joules PKG: %0X\n", p->energy_pkg); + outp += sprintf(outp, "Joules COR: %0X\n", p->energy_cores); + outp += sprintf(outp, "Joules GFX: %0X\n", p->energy_gfx); + outp += sprintf(outp, "Joules RAM: %0X\n", p->energy_dram); + outp += sprintf(outp, "Throttle PKG: %0X\n", + p->rapl_pkg_perf_status); + outp += sprintf(outp, "Throttle RAM: %0X\n", + p->rapl_dram_perf_status); + outp += sprintf(outp, "PTM: %dC\n", p->pkg_temp_c); + } + + outp += sprintf(outp, "\n"); + + return 0; } -void print_cnt(struct counters *p) +/* + * column formatting convention & formats + */ +int format_counters(struct thread_data *t, struct core_data *c, + struct pkg_data *p) { double interval_float; + char *fmt8; + + /* if showing only 1st thread in core and this isn't one, bail out */ + if (show_core_only && !(t->flags & CPU_IS_FIRST_THREAD_IN_CORE)) + return 0; + + /* if showing only 1st thread in pkg and this isn't one, bail out */ + if (show_pkg_only && !(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) + return 0; interval_float = tv_delta.tv_sec + tv_delta.tv_usec/1000000.0; - /* topology columns, print blanks on 1st (average) line */ - if (p == cnt_average) { + /* topo columns, print blanks on 1st (average) line */ + if (t == &average.threads) { if (show_pkg) - fprintf(stderr, " "); + outp += sprintf(outp, " -"); if (show_core) - fprintf(stderr, " "); + outp += sprintf(outp, " -"); if (show_cpu) - fprintf(stderr, " "); + outp += sprintf(outp, " -"); } else { - if (show_pkg) - fprintf(stderr, "%d", p->pkg); - if (show_core) - fprintf(stderr, "%4d", p->core); + if (show_pkg) { + if (p) + outp += sprintf(outp, "%8d", p->package_id); + else + outp += sprintf(outp, " -"); + } + if (show_core) { + if (c) + outp += sprintf(outp, "%8d", c->core_id); + else + outp += sprintf(outp, " -"); + } if (show_cpu) - fprintf(stderr, "%4d", p->cpu); + outp += sprintf(outp, "%8d", t->cpu_id); } + /* AvgMHz */ + if (has_aperf) + outp += sprintf(outp, "%8.0f", + 1.0 / units * t->aperf / interval_float); + /* %c0 */ if (do_nhm_cstates) { if (!skip_c0) - fprintf(stderr, "%7.2f", 100.0 * p->mperf/p->tsc); + outp += sprintf(outp, "%8.2f", 100.0 * t->mperf/t->tsc); else - fprintf(stderr, " ****"); + outp += sprintf(outp, "********"); } - /* GHz */ - if (has_aperf) { - if (!aperf_mperf_unstable) { - fprintf(stderr, "%5.2f", - 1.0 * p->tsc / units * p->aperf / - p->mperf / interval_float); - } else { - if (p->aperf > p->tsc || p->mperf > p->tsc) { - fprintf(stderr, " ****"); - } else { - fprintf(stderr, "%4.1f*", - 1.0 * p->tsc / - units * p->aperf / - p->mperf / interval_float); - } - } - } + /* BzyMHz */ + if (has_aperf) + outp += sprintf(outp, "%8.0f", + 1.0 * t->tsc / units * t->aperf / t->mperf / interval_float); /* TSC */ - fprintf(stderr, "%5.2f", 1.0 * p->tsc/units/interval_float); + outp += sprintf(outp, "%8.0f", 1.0 * t->tsc/units/interval_float); + + /* SMI */ + if (do_smi) + outp += sprintf(outp, "%8d", t->smi_count); + + /* delta */ + if (extra_delta_offset32) + outp += sprintf(outp, " %11llu", t->extra_delta32); + + /* DELTA */ + if (extra_delta_offset64) + outp += sprintf(outp, " %11llu", t->extra_delta64); + /* msr */ + if (extra_msr_offset32) + outp += sprintf(outp, " 0x%08llx", t->extra_msr32); + + /* MSR */ + if (extra_msr_offset64) + outp += sprintf(outp, " 0x%016llx", t->extra_msr64); if (do_nhm_cstates) { if (!skip_c1) - fprintf(stderr, "%7.2f", 100.0 * p->c1/p->tsc); + outp += sprintf(outp, "%8.2f", 100.0 * t->c1/t->tsc); else - fprintf(stderr, " ****"); + outp += sprintf(outp, "********"); } + + /* print per-core data only for 1st thread in core */ + if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE)) + goto done; + + if (do_nhm_cstates && !do_slm_cstates) + outp += sprintf(outp, "%8.2f", 100.0 * c->c3/t->tsc); if (do_nhm_cstates) - fprintf(stderr, " %6.2f", 100.0 * p->c3/p->tsc); - if (do_nhm_cstates) - fprintf(stderr, " %6.2f", 100.0 * p->c6/p->tsc); + outp += sprintf(outp, "%8.2f", 100.0 * c->c6/t->tsc); if (do_snb_cstates) - fprintf(stderr, " %6.2f", 100.0 * p->c7/p->tsc); + outp += sprintf(outp, "%8.2f", 100.0 * c->c7/t->tsc); + + if (do_dts) + outp += sprintf(outp, "%8d", c->core_temp_c); + + /* print per-package data only for 1st core in package */ + if (!(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) + goto done; + + if (do_ptm) + outp += sprintf(outp, "%8d", p->pkg_temp_c); + if (do_snb_cstates) - fprintf(stderr, " %5.2f", 100.0 * p->pc2/p->tsc); - if (do_nhm_cstates) - fprintf(stderr, " %5.2f", 100.0 * p->pc3/p->tsc); - if (do_nhm_cstates) - fprintf(stderr, " %5.2f", 100.0 * p->pc6/p->tsc); + outp += sprintf(outp, "%8.2f", 100.0 * p->pc2/t->tsc); + if (do_nhm_cstates && !do_slm_cstates) + outp += sprintf(outp, "%8.2f", 100.0 * p->pc3/t->tsc); + if (do_nhm_cstates && !do_slm_cstates) + outp += sprintf(outp, "%8.2f", 100.0 * p->pc6/t->tsc); if (do_snb_cstates) - fprintf(stderr, " %5.2f", 100.0 * p->pc7/p->tsc); - if (extra_msr_offset) - fprintf(stderr, " 0x%016llx", p->extra_msr); - putc('\n', stderr); + outp += sprintf(outp, "%8.2f", 100.0 * p->pc7/t->tsc); + if (do_c8_c9_c10) { + outp += sprintf(outp, "%8.2f", 100.0 * p->pc8/t->tsc); + outp += sprintf(outp, "%8.2f", 100.0 * p->pc9/t->tsc); + outp += sprintf(outp, "%8.2f", 100.0 * p->pc10/t->tsc); + } + + /* + * If measurement interval exceeds minimum RAPL Joule Counter range, + * indicate that results are suspect by printing "**" in fraction place. + */ + if (interval_float < rapl_joule_counter_range) + fmt8 = "%8.2f"; + else + fmt8 = " %6.0f**"; + + if (do_rapl && !rapl_joules) { + if (do_rapl & RAPL_PKG) + outp += sprintf(outp, fmt8, p->energy_pkg * rapl_energy_units / interval_float); + if (do_rapl & RAPL_CORES) + outp += sprintf(outp, fmt8, p->energy_cores * rapl_energy_units / interval_float); + if (do_rapl & RAPL_GFX) + outp += sprintf(outp, fmt8, p->energy_gfx * rapl_energy_units / interval_float); + if (do_rapl & RAPL_DRAM) + outp += sprintf(outp, fmt8, p->energy_dram * rapl_energy_units / interval_float); + if (do_rapl & RAPL_PKG_PERF_STATUS) + outp += sprintf(outp, fmt8, 100.0 * p->rapl_pkg_perf_status * rapl_time_units / interval_float); + if (do_rapl & RAPL_DRAM_PERF_STATUS) + outp += sprintf(outp, fmt8, 100.0 * p->rapl_dram_perf_status * rapl_time_units / interval_float); + } else { + if (do_rapl & RAPL_PKG) + outp += sprintf(outp, fmt8, + p->energy_pkg * rapl_energy_units); + if (do_rapl & RAPL_CORES) + outp += sprintf(outp, fmt8, + p->energy_cores * rapl_energy_units); + if (do_rapl & RAPL_GFX) + outp += sprintf(outp, fmt8, + p->energy_gfx * rapl_energy_units); + if (do_rapl & RAPL_DRAM) + outp += sprintf(outp, fmt8, + p->energy_dram * rapl_energy_units); + if (do_rapl & RAPL_PKG_PERF_STATUS) + outp += sprintf(outp, fmt8, 100.0 * p->rapl_pkg_perf_status * rapl_time_units / interval_float); + if (do_rapl & RAPL_DRAM_PERF_STATUS) + outp += sprintf(outp, fmt8, 100.0 * p->rapl_dram_perf_status * rapl_time_units / interval_float); + outp += sprintf(outp, fmt8, interval_float); + + } +done: + outp += sprintf(outp, "\n"); + + return 0; } -void print_counters(struct counters *counters) +void flush_stdout() { - struct counters *cnt; + fputs(output_buffer, stdout); + fflush(stdout); + outp = output_buffer; +} +void flush_stderr() +{ + fputs(output_buffer, stderr); + outp = output_buffer; +} +void format_all_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p) +{ + static int printed; + + if (!printed || !summary_only) + print_header(); - print_header(); + if (topo.num_cpus > 1) + format_counters(&average.threads, &average.cores, + &average.packages); - if (num_cpus > 1) - print_cnt(cnt_average); + printed = 1; - for (cnt = counters; cnt != NULL; cnt = cnt->next) - print_cnt(cnt); + if (summary_only) + return; + for_all_cpus(format_counters, t, c, p); } -#define SUBTRACT_COUNTER(after, before, delta) (delta = (after - before), (before > after)) +#define DELTA_WRAP32(new, old) \ + if (new > old) { \ + old = new - old; \ + } else { \ + old = 0x100000000 + new - old; \ + } -int compute_delta(struct counters *after, - struct counters *before, struct counters *delta) +void +delta_package(struct pkg_data *new, struct pkg_data *old) { - int errors = 0; - int perf_err = 0; + old->pc2 = new->pc2 - old->pc2; + old->pc3 = new->pc3 - old->pc3; + old->pc6 = new->pc6 - old->pc6; + old->pc7 = new->pc7 - old->pc7; + old->pc8 = new->pc8 - old->pc8; + old->pc9 = new->pc9 - old->pc9; + old->pc10 = new->pc10 - old->pc10; + old->pkg_temp_c = new->pkg_temp_c; + + DELTA_WRAP32(new->energy_pkg, old->energy_pkg); + DELTA_WRAP32(new->energy_cores, old->energy_cores); + DELTA_WRAP32(new->energy_gfx, old->energy_gfx); + DELTA_WRAP32(new->energy_dram, old->energy_dram); + DELTA_WRAP32(new->rapl_pkg_perf_status, old->rapl_pkg_perf_status); + DELTA_WRAP32(new->rapl_dram_perf_status, old->rapl_dram_perf_status); +} - skip_c0 = skip_c1 = 0; +void +delta_core(struct core_data *new, struct core_data *old) +{ + old->c3 = new->c3 - old->c3; + old->c6 = new->c6 - old->c6; + old->c7 = new->c7 - old->c7; + old->core_temp_c = new->core_temp_c; +} - for ( ; after && before && delta; - after = after->next, before = before->next, delta = delta->next) { - if (before->cpu != after->cpu) { - printf("cpu configuration changed: %d != %d\n", - before->cpu, after->cpu); - return -1; - } +/* + * old = new - old + */ +void +delta_thread(struct thread_data *new, struct thread_data *old, + struct core_data *core_delta) +{ + old->tsc = new->tsc - old->tsc; - if (SUBTRACT_COUNTER(after->tsc, before->tsc, delta->tsc)) { - fprintf(stderr, "cpu%d TSC went backwards %llX to %llX\n", - before->cpu, before->tsc, after->tsc); - errors++; - } - /* check for TSC < 1 Mcycles over interval */ - if (delta->tsc < (1000 * 1000)) { - fprintf(stderr, "Insanely slow TSC rate," - " TSC stops in idle?\n"); - fprintf(stderr, "You can disable all c-states" - " by booting with \"idle=poll\"\n"); - fprintf(stderr, "or just the deep ones with" - " \"processor.max_cstate=1\"\n"); - exit(-3); - } - if (SUBTRACT_COUNTER(after->c3, before->c3, delta->c3)) { - fprintf(stderr, "cpu%d c3 counter went backwards %llX to %llX\n", - before->cpu, before->c3, after->c3); - errors++; - } - if (SUBTRACT_COUNTER(after->c6, before->c6, delta->c6)) { - fprintf(stderr, "cpu%d c6 counter went backwards %llX to %llX\n", - before->cpu, before->c6, after->c6); - errors++; - } - if (SUBTRACT_COUNTER(after->c7, before->c7, delta->c7)) { - fprintf(stderr, "cpu%d c7 counter went backwards %llX to %llX\n", - before->cpu, before->c7, after->c7); - errors++; - } - if (SUBTRACT_COUNTER(after->pc2, before->pc2, delta->pc2)) { - fprintf(stderr, "cpu%d pc2 counter went backwards %llX to %llX\n", - before->cpu, before->pc2, after->pc2); - errors++; - } - if (SUBTRACT_COUNTER(after->pc3, before->pc3, delta->pc3)) { - fprintf(stderr, "cpu%d pc3 counter went backwards %llX to %llX\n", - before->cpu, before->pc3, after->pc3); - errors++; - } - if (SUBTRACT_COUNTER(after->pc6, before->pc6, delta->pc6)) { - fprintf(stderr, "cpu%d pc6 counter went backwards %llX to %llX\n", - before->cpu, before->pc6, after->pc6); - errors++; - } - if (SUBTRACT_COUNTER(after->pc7, before->pc7, delta->pc7)) { - fprintf(stderr, "cpu%d pc7 counter went backwards %llX to %llX\n", - before->cpu, before->pc7, after->pc7); - errors++; - } + /* check for TSC < 1 Mcycles over interval */ + if (old->tsc < (1000 * 1000)) + errx(-3, "Insanely slow TSC rate, TSC stops in idle?\n" + "You can disable all c-states by booting with \"idle=poll\"\n" + "or just the deep ones with \"processor.max_cstate=1\""); - perf_err = SUBTRACT_COUNTER(after->aperf, before->aperf, delta->aperf); - if (perf_err) { - fprintf(stderr, "cpu%d aperf counter went backwards %llX to %llX\n", - before->cpu, before->aperf, after->aperf); - } - perf_err |= SUBTRACT_COUNTER(after->mperf, before->mperf, delta->mperf); - if (perf_err) { - fprintf(stderr, "cpu%d mperf counter went backwards %llX to %llX\n", - before->cpu, before->mperf, after->mperf); - } - if (perf_err) { - if (!aperf_mperf_unstable) { - fprintf(stderr, "%s: APERF or MPERF went backwards *\n", progname); - fprintf(stderr, "* Frequency results do not cover entire interval *\n"); - fprintf(stderr, "* fix this by running Linux-2.6.30 or later *\n"); + old->c1 = new->c1 - old->c1; - aperf_mperf_unstable = 1; - } - /* - * mperf delta is likely a huge "positive" number - * can not use it for calculating c0 time - */ - skip_c0 = 1; - skip_c1 = 1; - } + if ((new->aperf > old->aperf) && (new->mperf > old->mperf)) { + old->aperf = new->aperf - old->aperf; + old->mperf = new->mperf - old->mperf; + } else { + + if (!aperf_mperf_unstable) { + fprintf(stderr, "%s: APERF or MPERF went backwards *\n", progname); + fprintf(stderr, "* Frequency results do not cover entire interval *\n"); + fprintf(stderr, "* fix this by running Linux-2.6.30 or later *\n"); + aperf_mperf_unstable = 1; + } /* - * As mperf and tsc collection are not atomic, - * it is possible for mperf's non-halted cycles - * to exceed TSC's all cycles: show c1 = 0% in that case. + * mperf delta is likely a huge "positive" number + * can not use it for calculating c0 time */ - if (delta->mperf > delta->tsc) - delta->c1 = 0; - else /* normal case, derive c1 */ - delta->c1 = delta->tsc - delta->mperf - - delta->c3 - delta->c6 - delta->c7; + skip_c0 = 1; + skip_c1 = 1; + } - if (delta->mperf == 0) - delta->mperf = 1; /* divide by 0 protection */ + if (use_c1_residency_msr) { + /* + * Some models have a dedicated C1 residency MSR, + * which should be more accurate than the derivation below. + */ + } else { /* - * for "extra msr", just copy the latest w/o subtracting + * As counter collection is not atomic, + * it is possible for mperf's non-halted cycles + idle states + * to exceed TSC's all cycles: show c1 = 0% in that case. */ - delta->extra_msr = after->extra_msr; - if (errors) { - fprintf(stderr, "ERROR cpu%d before:\n", before->cpu); - dump_cnt(before); - fprintf(stderr, "ERROR cpu%d after:\n", before->cpu); - dump_cnt(after); - errors = 0; + if ((old->mperf + core_delta->c3 + core_delta->c6 + core_delta->c7) > old->tsc) + old->c1 = 0; + else { + /* normal case, derive c1 */ + old->c1 = old->tsc - old->mperf - core_delta->c3 + - core_delta->c6 - core_delta->c7; } } + + if (old->mperf == 0) { + if (verbose > 1) fprintf(stderr, "cpu%d MPERF 0!\n", old->cpu_id); + old->mperf = 1; /* divide by 0 protection */ + } + + old->extra_delta32 = new->extra_delta32 - old->extra_delta32; + old->extra_delta32 &= 0xFFFFFFFF; + + old->extra_delta64 = new->extra_delta64 - old->extra_delta64; + + /* + * Extra MSR is just a snapshot, simply copy latest w/o subtracting + */ + old->extra_msr32 = new->extra_msr32; + old->extra_msr64 = new->extra_msr64; + + if (do_smi) + old->smi_count = new->smi_count - old->smi_count; +} + +int delta_cpu(struct thread_data *t, struct core_data *c, + struct pkg_data *p, struct thread_data *t2, + struct core_data *c2, struct pkg_data *p2) +{ + /* calculate core delta only for 1st thread in core */ + if (t->flags & CPU_IS_FIRST_THREAD_IN_CORE) + delta_core(c, c2); + + /* always calculate thread delta */ + delta_thread(t, t2, c2); /* c2 is core delta */ + + /* calculate package delta only for 1st core in package */ + if (t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE) + delta_package(p, p2); + return 0; } -void compute_average(struct counters *delta, struct counters *avg) -{ - struct counters *sum; - - sum = calloc(1, sizeof(struct counters)); - if (sum == NULL) { - perror("calloc sum"); - exit(1); - } - - for (; delta; delta = delta->next) { - sum->tsc += delta->tsc; - sum->c1 += delta->c1; - sum->c3 += delta->c3; - sum->c6 += delta->c6; - sum->c7 += delta->c7; - sum->aperf += delta->aperf; - sum->mperf += delta->mperf; - sum->pc2 += delta->pc2; - sum->pc3 += delta->pc3; - sum->pc6 += delta->pc6; - sum->pc7 += delta->pc7; - } - avg->tsc = sum->tsc/num_cpus; - avg->c1 = sum->c1/num_cpus; - avg->c3 = sum->c3/num_cpus; - avg->c6 = sum->c6/num_cpus; - avg->c7 = sum->c7/num_cpus; - avg->aperf = sum->aperf/num_cpus; - avg->mperf = sum->mperf/num_cpus; - avg->pc2 = sum->pc2/num_cpus; - avg->pc3 = sum->pc3/num_cpus; - avg->pc6 = sum->pc6/num_cpus; - avg->pc7 = sum->pc7/num_cpus; - - free(sum); -} - -void get_counters(struct counters *cnt) -{ - for ( ; cnt; cnt = cnt->next) { - cnt->tsc = get_msr(cnt->cpu, MSR_TSC); - if (do_nhm_cstates) - cnt->c3 = get_msr(cnt->cpu, MSR_CORE_C3_RESIDENCY); - if (do_nhm_cstates) - cnt->c6 = get_msr(cnt->cpu, MSR_CORE_C6_RESIDENCY); - if (do_snb_cstates) - cnt->c7 = get_msr(cnt->cpu, MSR_CORE_C7_RESIDENCY); - if (has_aperf) - cnt->aperf = get_msr(cnt->cpu, MSR_APERF); - if (has_aperf) - cnt->mperf = get_msr(cnt->cpu, MSR_MPERF); - if (do_snb_cstates) - cnt->pc2 = get_msr(cnt->cpu, MSR_PKG_C2_RESIDENCY); - if (do_nhm_cstates) - cnt->pc3 = get_msr(cnt->cpu, MSR_PKG_C3_RESIDENCY); - if (do_nhm_cstates) - cnt->pc6 = get_msr(cnt->cpu, MSR_PKG_C6_RESIDENCY); - if (do_snb_cstates) - cnt->pc7 = get_msr(cnt->cpu, MSR_PKG_C7_RESIDENCY); - if (extra_msr_offset) - cnt->extra_msr = get_msr(cnt->cpu, extra_msr_offset); - } -} - -void print_nehalem_info(void) +void clear_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p) +{ + t->tsc = 0; + t->aperf = 0; + t->mperf = 0; + t->c1 = 0; + + t->smi_count = 0; + t->extra_delta32 = 0; + t->extra_delta64 = 0; + + /* tells format_counters to dump all fields from this set */ + t->flags = CPU_IS_FIRST_THREAD_IN_CORE | CPU_IS_FIRST_CORE_IN_PACKAGE; + + c->c3 = 0; + c->c6 = 0; + c->c7 = 0; + c->core_temp_c = 0; + + p->pc2 = 0; + p->pc3 = 0; + p->pc6 = 0; + p->pc7 = 0; + p->pc8 = 0; + p->pc9 = 0; + p->pc10 = 0; + + p->energy_pkg = 0; + p->energy_dram = 0; + p->energy_cores = 0; + p->energy_gfx = 0; + p->rapl_pkg_perf_status = 0; + p->rapl_dram_perf_status = 0; + p->pkg_temp_c = 0; +} +int sum_counters(struct thread_data *t, struct core_data *c, + struct pkg_data *p) +{ + average.threads.tsc += t->tsc; + average.threads.aperf += t->aperf; + average.threads.mperf += t->mperf; + average.threads.c1 += t->c1; + + average.threads.extra_delta32 += t->extra_delta32; + average.threads.extra_delta64 += t->extra_delta64; + + /* sum per-core values only for 1st thread in core */ + if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE)) + return 0; + + average.cores.c3 += c->c3; + average.cores.c6 += c->c6; + average.cores.c7 += c->c7; + + average.cores.core_temp_c = MAX(average.cores.core_temp_c, c->core_temp_c); + + /* sum per-pkg values only for 1st core in pkg */ + if (!(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) + return 0; + + average.packages.pc2 += p->pc2; + average.packages.pc3 += p->pc3; + average.packages.pc6 += p->pc6; + average.packages.pc7 += p->pc7; + average.packages.pc8 += p->pc8; + average.packages.pc9 += p->pc9; + average.packages.pc10 += p->pc10; + + average.packages.energy_pkg += p->energy_pkg; + average.packages.energy_dram += p->energy_dram; + average.packages.energy_cores += p->energy_cores; + average.packages.energy_gfx += p->energy_gfx; + + average.packages.pkg_temp_c = MAX(average.packages.pkg_temp_c, p->pkg_temp_c); + + average.packages.rapl_pkg_perf_status += p->rapl_pkg_perf_status; + average.packages.rapl_dram_perf_status += p->rapl_dram_perf_status; + return 0; +} +/* + * sum the counters for all cpus in the system + * compute the weighted average + */ +void compute_average(struct thread_data *t, struct core_data *c, + struct pkg_data *p) +{ + clear_counters(&average.threads, &average.cores, &average.packages); + + for_all_cpus(sum_counters, t, c, p); + + average.threads.tsc /= topo.num_cpus; + average.threads.aperf /= topo.num_cpus; + average.threads.mperf /= topo.num_cpus; + average.threads.c1 /= topo.num_cpus; + + average.threads.extra_delta32 /= topo.num_cpus; + average.threads.extra_delta32 &= 0xFFFFFFFF; + + average.threads.extra_delta64 /= topo.num_cpus; + + average.cores.c3 /= topo.num_cores; + average.cores.c6 /= topo.num_cores; + average.cores.c7 /= topo.num_cores; + + average.packages.pc2 /= topo.num_packages; + average.packages.pc3 /= topo.num_packages; + average.packages.pc6 /= topo.num_packages; + average.packages.pc7 /= topo.num_packages; + + average.packages.pc8 /= topo.num_packages; + average.packages.pc9 /= topo.num_packages; + average.packages.pc10 /= topo.num_packages; +} + +static unsigned long long rdtsc(void) +{ + unsigned int low, high; + + asm volatile("rdtsc" : "=a" (low), "=d" (high)); + + return low | ((unsigned long long)high) << 32; +} + + +/* + * get_counters(...) + * migrate to cpu + * acquire and record local counters for that cpu + */ +int get_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p) +{ + int cpu = t->cpu_id; + unsigned long long msr; + + if (cpu_migrate(cpu)) { + fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + return -1; + } + + t->tsc = rdtsc(); /* we are running on local CPU of interest */ + + if (has_aperf) { + if (get_msr(cpu, MSR_IA32_APERF, &t->aperf)) + return -3; + if (get_msr(cpu, MSR_IA32_MPERF, &t->mperf)) + return -4; + } + + if (do_smi) { + if (get_msr(cpu, MSR_SMI_COUNT, &msr)) + return -5; + t->smi_count = msr & 0xFFFFFFFF; + } + if (extra_delta_offset32) { + if (get_msr(cpu, extra_delta_offset32, &msr)) + return -5; + t->extra_delta32 = msr & 0xFFFFFFFF; + } + + if (extra_delta_offset64) + if (get_msr(cpu, extra_delta_offset64, &t->extra_delta64)) + return -5; + + if (extra_msr_offset32) { + if (get_msr(cpu, extra_msr_offset32, &msr)) + return -5; + t->extra_msr32 = msr & 0xFFFFFFFF; + } + + if (extra_msr_offset64) + if (get_msr(cpu, extra_msr_offset64, &t->extra_msr64)) + return -5; + + if (use_c1_residency_msr) { + if (get_msr(cpu, MSR_CORE_C1_RES, &t->c1)) + return -6; + } + + /* collect core counters only for 1st thread in core */ + if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE)) + return 0; + + if (do_nhm_cstates && !do_slm_cstates) { + if (get_msr(cpu, MSR_CORE_C3_RESIDENCY, &c->c3)) + return -6; + } + + if (do_nhm_cstates) { + if (get_msr(cpu, MSR_CORE_C6_RESIDENCY, &c->c6)) + return -7; + } + + if (do_snb_cstates) + if (get_msr(cpu, MSR_CORE_C7_RESIDENCY, &c->c7)) + return -8; + + if (do_dts) { + if (get_msr(cpu, MSR_IA32_THERM_STATUS, &msr)) + return -9; + c->core_temp_c = tcc_activation_temp - ((msr >> 16) & 0x7F); + } + + + /* collect package counters only for 1st core in package */ + if (!(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) + return 0; + + if (do_nhm_cstates && !do_slm_cstates) { + if (get_msr(cpu, MSR_PKG_C3_RESIDENCY, &p->pc3)) + return -9; + if (get_msr(cpu, MSR_PKG_C6_RESIDENCY, &p->pc6)) + return -10; + } + if (do_snb_cstates) { + if (get_msr(cpu, MSR_PKG_C2_RESIDENCY, &p->pc2)) + return -11; + if (get_msr(cpu, MSR_PKG_C7_RESIDENCY, &p->pc7)) + return -12; + } + if (do_c8_c9_c10) { + if (get_msr(cpu, MSR_PKG_C8_RESIDENCY, &p->pc8)) + return -13; + if (get_msr(cpu, MSR_PKG_C9_RESIDENCY, &p->pc9)) + return -13; + if (get_msr(cpu, MSR_PKG_C10_RESIDENCY, &p->pc10)) + return -13; + } + if (do_rapl & RAPL_PKG) { + if (get_msr(cpu, MSR_PKG_ENERGY_STATUS, &msr)) + return -13; + p->energy_pkg = msr & 0xFFFFFFFF; + } + if (do_rapl & RAPL_CORES) { + if (get_msr(cpu, MSR_PP0_ENERGY_STATUS, &msr)) + return -14; + p->energy_cores = msr & 0xFFFFFFFF; + } + if (do_rapl & RAPL_DRAM) { + if (get_msr(cpu, MSR_DRAM_ENERGY_STATUS, &msr)) + return -15; + p->energy_dram = msr & 0xFFFFFFFF; + } + if (do_rapl & RAPL_GFX) { + if (get_msr(cpu, MSR_PP1_ENERGY_STATUS, &msr)) + return -16; + p->energy_gfx = msr & 0xFFFFFFFF; + } + if (do_rapl & RAPL_PKG_PERF_STATUS) { + if (get_msr(cpu, MSR_PKG_PERF_STATUS, &msr)) + return -16; + p->rapl_pkg_perf_status = msr & 0xFFFFFFFF; + } + if (do_rapl & RAPL_DRAM_PERF_STATUS) { + if (get_msr(cpu, MSR_DRAM_PERF_STATUS, &msr)) + return -16; + p->rapl_dram_perf_status = msr & 0xFFFFFFFF; + } + if (do_ptm) { + if (get_msr(cpu, MSR_IA32_PACKAGE_THERM_STATUS, &msr)) + return -17; + p->pkg_temp_c = tcc_activation_temp - ((msr >> 16) & 0x7F); + } + return 0; +} + +void print_verbose_header(void) { unsigned long long msr; unsigned int ratio; @@ -475,7 +1022,9 @@ void print_nehalem_info(void) if (!do_nehalem_platform_info) return; - msr = get_msr(0, MSR_NEHALEM_PLATFORM_INFO); + get_msr(0, MSR_NHM_PLATFORM_INFO, &msr); + + fprintf(stderr, "cpu0: MSR_NHM_PLATFORM_INFO: 0x%08llx\n", msr); ratio = (msr >> 40) & 0xFF; fprintf(stderr, "%d * %.0f = %.0f MHz max efficiency\n", @@ -485,13 +1034,130 @@ void print_nehalem_info(void) fprintf(stderr, "%d * %.0f = %.0f MHz TSC frequency\n", ratio, bclk, ratio * bclk); - if (verbose > 1) - fprintf(stderr, "MSR_NEHALEM_PLATFORM_INFO: 0x%llx\n", msr); + get_msr(0, MSR_IA32_POWER_CTL, &msr); + fprintf(stderr, "cpu0: MSR_IA32_POWER_CTL: 0x%08llx (C1E auto-promotion: %sabled)\n", + msr, msr & 0x2 ? "EN" : "DIS"); + + if (!do_ivt_turbo_ratio_limit) + goto print_nhm_turbo_ratio_limits; + + get_msr(0, MSR_IVT_TURBO_RATIO_LIMIT, &msr); + + fprintf(stderr, "cpu0: MSR_IVT_TURBO_RATIO_LIMIT: 0x%08llx\n", msr); + + ratio = (msr >> 56) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 16 active cores\n", + ratio, bclk, ratio * bclk); + + ratio = (msr >> 48) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 15 active cores\n", + ratio, bclk, ratio * bclk); + + ratio = (msr >> 40) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 14 active cores\n", + ratio, bclk, ratio * bclk); + + ratio = (msr >> 32) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 13 active cores\n", + ratio, bclk, ratio * bclk); + + ratio = (msr >> 24) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 12 active cores\n", + ratio, bclk, ratio * bclk); + + ratio = (msr >> 16) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 11 active cores\n", + ratio, bclk, ratio * bclk); + + ratio = (msr >> 8) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 10 active cores\n", + ratio, bclk, ratio * bclk); + + ratio = (msr >> 0) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 9 active cores\n", + ratio, bclk, ratio * bclk); + +print_nhm_turbo_ratio_limits: + get_msr(0, MSR_NHM_SNB_PKG_CST_CFG_CTL, &msr); + +#define SNB_C1_AUTO_UNDEMOTE (1UL << 27) +#define SNB_C3_AUTO_UNDEMOTE (1UL << 28) + + fprintf(stderr, "cpu0: MSR_NHM_SNB_PKG_CST_CFG_CTL: 0x%08llx", msr); + + fprintf(stderr, " (%s%s%s%s%slocked: pkg-cstate-limit=%d: ", + (msr & SNB_C3_AUTO_UNDEMOTE) ? "UNdemote-C3, " : "", + (msr & SNB_C1_AUTO_UNDEMOTE) ? "UNdemote-C1, " : "", + (msr & NHM_C3_AUTO_DEMOTE) ? "demote-C3, " : "", + (msr & NHM_C1_AUTO_DEMOTE) ? "demote-C1, " : "", + (msr & (1 << 15)) ? "" : "UN", + (unsigned int)msr & 7); + + + switch(msr & 0x7) { + case 0: + fprintf(stderr, do_slm_cstates ? "no pkg states" : "pc0"); + break; + case 1: + fprintf(stderr, do_slm_cstates ? "no pkg states" : do_snb_cstates ? "pc2" : "pc0"); + break; + case 2: + fprintf(stderr, do_slm_cstates ? "invalid" : do_snb_cstates ? "pc6-noret" : "pc3"); + break; + case 3: + fprintf(stderr, do_slm_cstates ? "invalid" : "pc6"); + break; + case 4: + fprintf(stderr, do_slm_cstates ? "pc4" : "pc7"); + break; + case 5: + fprintf(stderr, do_slm_cstates ? "invalid" : do_snb_cstates ? "pc7s" : "invalid"); + break; + case 6: + fprintf(stderr, do_slm_cstates ? "pc6" : "invalid"); + break; + case 7: + fprintf(stderr, do_slm_cstates ? "pc7" : "unlimited"); + break; + default: + fprintf(stderr, "invalid"); + } + fprintf(stderr, ")\n"); if (!do_nehalem_turbo_ratio_limit) return; - msr = get_msr(0, MSR_NEHALEM_TURBO_RATIO_LIMIT); + get_msr(0, MSR_NHM_TURBO_RATIO_LIMIT, &msr); + + fprintf(stderr, "cpu0: MSR_NHM_TURBO_RATIO_LIMIT: 0x%08llx\n", msr); + + ratio = (msr >> 56) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 8 active cores\n", + ratio, bclk, ratio * bclk); + + ratio = (msr >> 48) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 7 active cores\n", + ratio, bclk, ratio * bclk); + + ratio = (msr >> 40) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 6 active cores\n", + ratio, bclk, ratio * bclk); + + ratio = (msr >> 32) & 0xFF; + if (ratio) + fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 5 active cores\n", + ratio, bclk, ratio * bclk); ratio = (msr >> 24) & 0xFF; if (ratio) @@ -512,267 +1178,278 @@ void print_nehalem_info(void) if (ratio) fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 1 active cores\n", ratio, bclk, ratio * bclk); - } -void free_counter_list(struct counters *list) +void free_all_buffers(void) { - struct counters *p; + CPU_FREE(cpu_present_set); + cpu_present_set = NULL; + cpu_present_set = 0; - for (p = list; p; ) { - struct counters *free_me; + CPU_FREE(cpu_affinity_set); + cpu_affinity_set = NULL; + cpu_affinity_setsize = 0; - free_me = p; - p = p->next; - free(free_me); - } + free(thread_even); + free(core_even); + free(package_even); + + thread_even = NULL; + core_even = NULL; + package_even = NULL; + + free(thread_odd); + free(core_odd); + free(package_odd); + + thread_odd = NULL; + core_odd = NULL; + package_odd = NULL; + + free(output_buffer); + output_buffer = NULL; + outp = NULL; } -void free_all_counters(void) +/* + * Open a file, and exit on failure + */ +FILE *fopen_or_die(const char *path, const char *mode) { - free_counter_list(cnt_even); - cnt_even = NULL; + FILE *filep = fopen(path, "r"); + if (!filep) + err(1, "%s: open failed", path); + return filep; +} - free_counter_list(cnt_odd); - cnt_odd = NULL; +/* + * Parse a file containing a single int. + */ +int parse_int_file(const char *fmt, ...) +{ + va_list args; + char path[PATH_MAX]; + FILE *filep; + int value; + + va_start(args, fmt); + vsnprintf(path, sizeof(path), fmt, args); + va_end(args); + filep = fopen_or_die(path, "r"); + if (fscanf(filep, "%d", &value) != 1) + err(1, "%s: failed to parse number from file", path); + fclose(filep); + return value; +} - free_counter_list(cnt_delta); - cnt_delta = NULL; +/* + * cpu_is_first_sibling_in_core(cpu) + * return 1 if given CPU is 1st HT sibling in the core + */ +int cpu_is_first_sibling_in_core(int cpu) +{ + return cpu == parse_int_file("/sys/devices/system/cpu/cpu%d/topology/thread_siblings_list", cpu); +} - free_counter_list(cnt_average); - cnt_average = NULL; +/* + * cpu_is_first_core_in_package(cpu) + * return 1 if given CPU is 1st core in package + */ +int cpu_is_first_core_in_package(int cpu) +{ + return cpu == parse_int_file("/sys/devices/system/cpu/cpu%d/topology/core_siblings_list", cpu); } -void insert_counters(struct counters **list, - struct counters *new) +int get_physical_package_id(int cpu) { - struct counters *prev; + return parse_int_file("/sys/devices/system/cpu/cpu%d/topology/physical_package_id", cpu); +} - /* - * list was empty - */ - if (*list == NULL) { - new->next = *list; - *list = new; - return; - } +int get_core_id(int cpu) +{ + return parse_int_file("/sys/devices/system/cpu/cpu%d/topology/core_id", cpu); +} - show_cpu = 1; /* there is more than one CPU */ +int get_num_ht_siblings(int cpu) +{ + char path[80]; + FILE *filep; + int sib1, sib2; + int matches; + char character; + sprintf(path, "/sys/devices/system/cpu/cpu%d/topology/thread_siblings_list", cpu); + filep = fopen_or_die(path, "r"); /* - * insert on front of list. - * It is sorted by ascending package#, core#, cpu# + * file format: + * if a pair of number with a character between: 2 siblings (eg. 1-2, or 1,4) + * otherwinse 1 sibling (self). */ - if (((*list)->pkg > new->pkg) || - (((*list)->pkg == new->pkg) && ((*list)->core > new->core)) || - (((*list)->pkg == new->pkg) && ((*list)->core == new->core) && ((*list)->cpu > new->cpu))) { - new->next = *list; - *list = new; - return; - } + matches = fscanf(filep, "%d%c%d\n", &sib1, &character, &sib2); - prev = *list; + fclose(filep); - while (prev->next && (prev->next->pkg < new->pkg)) { - prev = prev->next; - show_pkg = 1; /* there is more than 1 package */ - } + if (matches == 3) + return 2; + else + return 1; +} - while (prev->next && (prev->next->pkg == new->pkg) - && (prev->next->core < new->core)) { - prev = prev->next; - show_core = 1; /* there is more than 1 core */ - } +/* + * run func(thread, core, package) in topology order + * skip non-present cpus + */ - while (prev->next && (prev->next->pkg == new->pkg) - && (prev->next->core == new->core) - && (prev->next->cpu < new->cpu)) { - prev = prev->next; - } +int for_all_cpus_2(int (func)(struct thread_data *, struct core_data *, + struct pkg_data *, struct thread_data *, struct core_data *, + struct pkg_data *), struct thread_data *thread_base, + struct core_data *core_base, struct pkg_data *pkg_base, + struct thread_data *thread_base2, struct core_data *core_base2, + struct pkg_data *pkg_base2) +{ + int retval, pkg_no, core_no, thread_no; - /* - * insert after "prev" - */ - new->next = prev->next; - prev->next = new; -} + for (pkg_no = 0; pkg_no < topo.num_packages; ++pkg_no) { + for (core_no = 0; core_no < topo.num_cores_per_pkg; ++core_no) { + for (thread_no = 0; thread_no < + topo.num_threads_per_core; ++thread_no) { + struct thread_data *t, *t2; + struct core_data *c, *c2; + struct pkg_data *p, *p2; -void alloc_new_counters(int pkg, int core, int cpu) -{ - struct counters *new; + t = GET_THREAD(thread_base, thread_no, core_no, pkg_no); - if (verbose > 1) - printf("pkg%d core%d, cpu%d\n", pkg, core, cpu); - - new = (struct counters *)calloc(1, sizeof(struct counters)); - if (new == NULL) { - perror("calloc"); - exit(1); - } - new->pkg = pkg; - new->core = core; - new->cpu = cpu; - insert_counters(&cnt_odd, new); - - new = (struct counters *)calloc(1, - sizeof(struct counters)); - if (new == NULL) { - perror("calloc"); - exit(1); - } - new->pkg = pkg; - new->core = core; - new->cpu = cpu; - insert_counters(&cnt_even, new); - - new = (struct counters *)calloc(1, sizeof(struct counters)); - if (new == NULL) { - perror("calloc"); - exit(1); - } - new->pkg = pkg; - new->core = core; - new->cpu = cpu; - insert_counters(&cnt_delta, new); - - new = (struct counters *)calloc(1, sizeof(struct counters)); - if (new == NULL) { - perror("calloc"); - exit(1); - } - new->pkg = pkg; - new->core = core; - new->cpu = cpu; - cnt_average = new; -} + if (cpu_is_not_present(t->cpu_id)) + continue; -int get_physical_package_id(int cpu) -{ - char path[64]; - FILE *filep; - int pkg; + t2 = GET_THREAD(thread_base2, thread_no, core_no, pkg_no); - sprintf(path, "/sys/devices/system/cpu/cpu%d/topology/physical_package_id", cpu); - filep = fopen(path, "r"); - if (filep == NULL) { - perror(path); - exit(1); - } - fscanf(filep, "%d", &pkg); - fclose(filep); - return pkg; -} + c = GET_CORE(core_base, core_no, pkg_no); + c2 = GET_CORE(core_base2, core_no, pkg_no); -int get_core_id(int cpu) -{ - char path[64]; - FILE *filep; - int core; + p = GET_PKG(pkg_base, pkg_no); + p2 = GET_PKG(pkg_base2, pkg_no); - sprintf(path, "/sys/devices/system/cpu/cpu%d/topology/core_id", cpu); - filep = fopen(path, "r"); - if (filep == NULL) { - perror(path); - exit(1); + retval = func(t, c, p, t2, c2, p2); + if (retval) + return retval; + } + } } - fscanf(filep, "%d", &core); - fclose(filep); - return core; + return 0; } /* - * run func(index, cpu) on every cpu in /proc/stat + * run func(cpu) on every cpu in /proc/stat + * return max_cpu number */ - -int for_all_cpus(void (func)(int, int, int)) +int for_all_proc_cpus(int (func)(int)) { FILE *fp; - int cpu_count; + int cpu_num; int retval; - fp = fopen(proc_stat, "r"); - if (fp == NULL) { - perror(proc_stat); - exit(1); - } + fp = fopen_or_die(proc_stat, "r"); retval = fscanf(fp, "cpu %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d\n"); - if (retval != 0) { - perror("/proc/stat format"); - exit(1); - } + if (retval != 0) + err(1, "%s: failed to parse format", proc_stat); - for (cpu_count = 0; ; cpu_count++) { - int cpu; - - retval = fscanf(fp, "cpu%u %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d\n", &cpu); + while (1) { + retval = fscanf(fp, "cpu%u %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d\n", &cpu_num); if (retval != 1) break; - func(get_physical_package_id(cpu), get_core_id(cpu), cpu); + retval = func(cpu_num); + if (retval) { + fclose(fp); + return(retval); + } } fclose(fp); - return cpu_count; + return 0; } void re_initialize(void) { - printf("turbostat: topology changed, re-initializing.\n"); - free_all_counters(); - num_cpus = for_all_cpus(alloc_new_counters); - need_reinitialize = 0; - printf("num_cpus is now %d\n", num_cpus); + free_all_buffers(); + setup_all_buffers(); + printf("turbostat: re-initialized with num_cpus %d\n", topo.num_cpus); } -void dummy(int pkg, int core, int cpu) { return; } + /* - * check to see if a cpu came on-line + * count_cpus() + * remember the last one seen, it will be the max */ -void verify_num_cpus(void) +int count_cpus(int cpu) { - int new_num_cpus; + if (topo.max_cpu_num < cpu) + topo.max_cpu_num = cpu; - new_num_cpus = for_all_cpus(dummy); - - if (new_num_cpus != num_cpus) { - if (verbose) - printf("num_cpus was %d, is now %d\n", - num_cpus, new_num_cpus); - need_reinitialize = 1; - } + topo.num_cpus += 1; + return 0; +} +int mark_cpu_present(int cpu) +{ + CPU_SET_S(cpu, cpu_present_setsize, cpu_present_set); + return 0; } void turbostat_loop() { + int retval; + int restarted = 0; + restart: - get_counters(cnt_even); + restarted++; + + retval = for_all_cpus(get_counters, EVEN_COUNTERS); + if (retval < -1) { + exit(retval); + } else if (retval == -1) { + if (restarted > 1) { + exit(retval); + } + re_initialize(); + goto restart; + } + restarted = 0; gettimeofday(&tv_even, (struct timezone *)NULL); while (1) { - verify_num_cpus(); - if (need_reinitialize) { + if (for_all_proc_cpus(cpu_is_not_present)) { re_initialize(); goto restart; } sleep(interval_sec); - get_counters(cnt_odd); + retval = for_all_cpus(get_counters, ODD_COUNTERS); + if (retval < -1) { + exit(retval); + } else if (retval == -1) { + re_initialize(); + goto restart; + } gettimeofday(&tv_odd, (struct timezone *)NULL); - - compute_delta(cnt_odd, cnt_even, cnt_delta); timersub(&tv_odd, &tv_even, &tv_delta); - compute_average(cnt_delta, cnt_average); - print_counters(cnt_delta); - if (need_reinitialize) { + for_all_cpus_2(delta_cpu, ODD_COUNTERS, EVEN_COUNTERS); + compute_average(EVEN_COUNTERS); + format_all_counters(EVEN_COUNTERS); + flush_stdout(); + sleep(interval_sec); + retval = for_all_cpus(get_counters, EVEN_COUNTERS); + if (retval < -1) { + exit(retval); + } else if (retval == -1) { re_initialize(); goto restart; } - sleep(interval_sec); - get_counters(cnt_even); gettimeofday(&tv_even, (struct timezone *)NULL); - compute_delta(cnt_even, cnt_odd, cnt_delta); timersub(&tv_even, &tv_odd, &tv_delta); - compute_average(cnt_delta, cnt_average); - print_counters(cnt_delta); + for_all_cpus_2(delta_cpu, EVEN_COUNTERS, ODD_COUNTERS); + compute_average(ODD_COUNTERS); + format_all_counters(ODD_COUNTERS); + flush_stdout(); } } @@ -780,19 +1457,15 @@ void check_dev_msr() { struct stat sb; - if (stat("/dev/cpu/0/msr", &sb)) { - fprintf(stderr, "no /dev/cpu/0/msr\n"); - fprintf(stderr, "Try \"# modprobe msr\"\n"); - exit(-5); - } + if (stat("/dev/cpu/0/msr", &sb)) + err(-5, "no /dev/cpu/0/msr\n" + "Try \"# modprobe msr\""); } void check_super_user() { - if (getuid() != 0) { - fprintf(stderr, "must be root\n"); - exit(-6); - } + if (getuid() != 0) + errx(-6, "must be root"); } int has_nehalem_turbo_ratio_limit(unsigned int family, unsigned int model) @@ -812,7 +1485,16 @@ int has_nehalem_turbo_ratio_limit(unsigned int family, unsigned int model) case 0x2A: /* SNB */ case 0x2D: /* SNB Xeon */ case 0x3A: /* IVB */ - case 0x3D: /* IVB Xeon */ + case 0x3E: /* IVB Xeon */ + case 0x3C: /* HSW */ + case 0x3F: /* HSX */ + case 0x45: /* HSW */ + case 0x46: /* HSW */ + case 0x37: /* BYT */ + case 0x4D: /* AVN */ + case 0x3D: /* BDW */ + case 0x4F: /* BDX */ + case 0x56: /* BDX-DE */ return 1; case 0x2E: /* Nehalem-EX Xeon - Beckton */ case 0x2F: /* Westmere-EX Xeon - Eagleton */ @@ -820,6 +1502,346 @@ int has_nehalem_turbo_ratio_limit(unsigned int family, unsigned int model) return 0; } } +int has_ivt_turbo_ratio_limit(unsigned int family, unsigned int model) +{ + if (!genuine_intel) + return 0; + + if (family != 6) + return 0; + + switch (model) { + case 0x3E: /* IVB Xeon */ + return 1; + default: + return 0; + } +} + +/* + * print_epb() + * Decode the ENERGY_PERF_BIAS MSR + */ +int print_epb(struct thread_data *t, struct core_data *c, struct pkg_data *p) +{ + unsigned long long msr; + char *epb_string; + int cpu; + + if (!has_epb) + return 0; + + cpu = t->cpu_id; + + /* EPB is per-package */ + if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE) || !(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) + return 0; + + if (cpu_migrate(cpu)) { + fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + return -1; + } + + if (get_msr(cpu, MSR_IA32_ENERGY_PERF_BIAS, &msr)) + return 0; + + switch (msr & 0x7) { + case ENERGY_PERF_BIAS_PERFORMANCE: + epb_string = "performance"; + break; + case ENERGY_PERF_BIAS_NORMAL: + epb_string = "balanced"; + break; + case ENERGY_PERF_BIAS_POWERSAVE: + epb_string = "powersave"; + break; + default: + epb_string = "custom"; + break; + } + fprintf(stderr, "cpu%d: MSR_IA32_ENERGY_PERF_BIAS: 0x%08llx (%s)\n", cpu, msr, epb_string); + + return 0; +} + +#define RAPL_POWER_GRANULARITY 0x7FFF /* 15 bit power granularity */ +#define RAPL_TIME_GRANULARITY 0x3F /* 6 bit time granularity */ + +double get_tdp(model) +{ + unsigned long long msr; + + if (do_rapl & RAPL_PKG_POWER_INFO) + if (!get_msr(0, MSR_PKG_POWER_INFO, &msr)) + return ((msr >> 0) & RAPL_POWER_GRANULARITY) * rapl_power_units; + + switch (model) { + case 0x37: + case 0x4D: + return 30.0; + default: + return 135.0; + } +} + + +/* + * rapl_probe() + * + * sets do_rapl, rapl_power_units, rapl_energy_units, rapl_time_units + */ +void rapl_probe(unsigned int family, unsigned int model) +{ + unsigned long long msr; + unsigned int time_unit; + double tdp; + + if (!genuine_intel) + return; + + if (family != 6) + return; + + switch (model) { + case 0x2A: + case 0x3A: + case 0x3C: /* HSW */ + case 0x45: /* HSW */ + case 0x46: /* HSW */ + case 0x3D: /* BDW */ + do_rapl = RAPL_PKG | RAPL_CORES | RAPL_CORE_POLICY | RAPL_GFX | RAPL_PKG_POWER_INFO; + break; + case 0x3F: /* HSX */ + case 0x4F: /* BDX */ + case 0x56: /* BDX-DE */ + do_rapl = RAPL_PKG | RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_PKG_PERF_STATUS | RAPL_PKG_POWER_INFO; + break; + case 0x2D: + case 0x3E: + do_rapl = RAPL_PKG | RAPL_CORES | RAPL_CORE_POLICY | RAPL_DRAM | RAPL_PKG_PERF_STATUS | RAPL_DRAM_PERF_STATUS | RAPL_PKG_POWER_INFO; + break; + case 0x37: /* BYT */ + case 0x4D: /* AVN */ + do_rapl = RAPL_PKG | RAPL_CORES ; + break; + default: + return; + } + + /* units on package 0, verify later other packages match */ + if (get_msr(0, MSR_RAPL_POWER_UNIT, &msr)) + return; + + rapl_power_units = 1.0 / (1 << (msr & 0xF)); + if (model == 0x37) + rapl_energy_units = 1.0 * (1 << (msr >> 8 & 0x1F)) / 1000000; + else + rapl_energy_units = 1.0 / (1 << (msr >> 8 & 0x1F)); + + time_unit = msr >> 16 & 0xF; + if (time_unit == 0) + time_unit = 0xA; + + rapl_time_units = 1.0 / (1 << (time_unit)); + + tdp = get_tdp(model); + + rapl_joule_counter_range = 0xFFFFFFFF * rapl_energy_units / tdp; + if (verbose) + fprintf(stderr, "RAPL: %.0f sec. Joule Counter Range, at %.0f Watts\n", rapl_joule_counter_range, tdp); + + return; +} + +int print_thermal(struct thread_data *t, struct core_data *c, struct pkg_data *p) +{ + unsigned long long msr; + unsigned int dts; + int cpu; + + if (!(do_dts || do_ptm)) + return 0; + + cpu = t->cpu_id; + + /* DTS is per-core, no need to print for each thread */ + if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE)) + return 0; + + if (cpu_migrate(cpu)) { + fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + return -1; + } + + if (do_ptm && (t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) { + if (get_msr(cpu, MSR_IA32_PACKAGE_THERM_STATUS, &msr)) + return 0; + + dts = (msr >> 16) & 0x7F; + fprintf(stderr, "cpu%d: MSR_IA32_PACKAGE_THERM_STATUS: 0x%08llx (%d C)\n", + cpu, msr, tcc_activation_temp - dts); + +#ifdef THERM_DEBUG + if (get_msr(cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT, &msr)) + return 0; + + dts = (msr >> 16) & 0x7F; + dts2 = (msr >> 8) & 0x7F; + fprintf(stderr, "cpu%d: MSR_IA32_PACKAGE_THERM_INTERRUPT: 0x%08llx (%d C, %d C)\n", + cpu, msr, tcc_activation_temp - dts, tcc_activation_temp - dts2); +#endif + } + + + if (do_dts) { + unsigned int resolution; + + if (get_msr(cpu, MSR_IA32_THERM_STATUS, &msr)) + return 0; + + dts = (msr >> 16) & 0x7F; + resolution = (msr >> 27) & 0xF; + fprintf(stderr, "cpu%d: MSR_IA32_THERM_STATUS: 0x%08llx (%d C +/- %d)\n", + cpu, msr, tcc_activation_temp - dts, resolution); + +#ifdef THERM_DEBUG + if (get_msr(cpu, MSR_IA32_THERM_INTERRUPT, &msr)) + return 0; + + dts = (msr >> 16) & 0x7F; + dts2 = (msr >> 8) & 0x7F; + fprintf(stderr, "cpu%d: MSR_IA32_THERM_INTERRUPT: 0x%08llx (%d C, %d C)\n", + cpu, msr, tcc_activation_temp - dts, tcc_activation_temp - dts2); +#endif + } + + return 0; +} + +void print_power_limit_msr(int cpu, unsigned long long msr, char *label) +{ + fprintf(stderr, "cpu%d: %s: %sabled (%f Watts, %f sec, clamp %sabled)\n", + cpu, label, + ((msr >> 15) & 1) ? "EN" : "DIS", + ((msr >> 0) & 0x7FFF) * rapl_power_units, + (1.0 + (((msr >> 22) & 0x3)/4.0)) * (1 << ((msr >> 17) & 0x1F)) * rapl_time_units, + (((msr >> 16) & 1) ? "EN" : "DIS")); + + return; +} + +int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p) +{ + unsigned long long msr; + int cpu; + + if (!do_rapl) + return 0; + + /* RAPL counters are per package, so print only for 1st thread/package */ + if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE) || !(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) + return 0; + + cpu = t->cpu_id; + if (cpu_migrate(cpu)) { + fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + return -1; + } + + if (get_msr(cpu, MSR_RAPL_POWER_UNIT, &msr)) + return -1; + + if (verbose) { + fprintf(stderr, "cpu%d: MSR_RAPL_POWER_UNIT: 0x%08llx " + "(%f Watts, %f Joules, %f sec.)\n", cpu, msr, + rapl_power_units, rapl_energy_units, rapl_time_units); + } + if (do_rapl & RAPL_PKG_POWER_INFO) { + + if (get_msr(cpu, MSR_PKG_POWER_INFO, &msr)) + return -5; + + + fprintf(stderr, "cpu%d: MSR_PKG_POWER_INFO: 0x%08llx (%.0f W TDP, RAPL %.0f - %.0f W, %f sec.)\n", + cpu, msr, + ((msr >> 0) & RAPL_POWER_GRANULARITY) * rapl_power_units, + ((msr >> 16) & RAPL_POWER_GRANULARITY) * rapl_power_units, + ((msr >> 32) & RAPL_POWER_GRANULARITY) * rapl_power_units, + ((msr >> 48) & RAPL_TIME_GRANULARITY) * rapl_time_units); + + } + if (do_rapl & RAPL_PKG) { + + if (get_msr(cpu, MSR_PKG_POWER_LIMIT, &msr)) + return -9; + + fprintf(stderr, "cpu%d: MSR_PKG_POWER_LIMIT: 0x%08llx (%slocked)\n", + cpu, msr, (msr >> 63) & 1 ? "": "UN"); + + print_power_limit_msr(cpu, msr, "PKG Limit #1"); + fprintf(stderr, "cpu%d: PKG Limit #2: %sabled (%f Watts, %f* sec, clamp %sabled)\n", + cpu, + ((msr >> 47) & 1) ? "EN" : "DIS", + ((msr >> 32) & 0x7FFF) * rapl_power_units, + (1.0 + (((msr >> 54) & 0x3)/4.0)) * (1 << ((msr >> 49) & 0x1F)) * rapl_time_units, + ((msr >> 48) & 1) ? "EN" : "DIS"); + } + + if (do_rapl & RAPL_DRAM) { + if (get_msr(cpu, MSR_DRAM_POWER_INFO, &msr)) + return -6; + + + fprintf(stderr, "cpu%d: MSR_DRAM_POWER_INFO,: 0x%08llx (%.0f W TDP, RAPL %.0f - %.0f W, %f sec.)\n", + cpu, msr, + ((msr >> 0) & RAPL_POWER_GRANULARITY) * rapl_power_units, + ((msr >> 16) & RAPL_POWER_GRANULARITY) * rapl_power_units, + ((msr >> 32) & RAPL_POWER_GRANULARITY) * rapl_power_units, + ((msr >> 48) & RAPL_TIME_GRANULARITY) * rapl_time_units); + + + if (get_msr(cpu, MSR_DRAM_POWER_LIMIT, &msr)) + return -9; + fprintf(stderr, "cpu%d: MSR_DRAM_POWER_LIMIT: 0x%08llx (%slocked)\n", + cpu, msr, (msr >> 31) & 1 ? "": "UN"); + + print_power_limit_msr(cpu, msr, "DRAM Limit"); + } + if (do_rapl & RAPL_CORE_POLICY) { + if (verbose) { + if (get_msr(cpu, MSR_PP0_POLICY, &msr)) + return -7; + + fprintf(stderr, "cpu%d: MSR_PP0_POLICY: %lld\n", cpu, msr & 0xF); + } + } + if (do_rapl & RAPL_CORES) { + if (verbose) { + + if (get_msr(cpu, MSR_PP0_POWER_LIMIT, &msr)) + return -9; + fprintf(stderr, "cpu%d: MSR_PP0_POWER_LIMIT: 0x%08llx (%slocked)\n", + cpu, msr, (msr >> 31) & 1 ? "": "UN"); + print_power_limit_msr(cpu, msr, "Cores Limit"); + } + } + if (do_rapl & RAPL_GFX) { + if (verbose) { + if (get_msr(cpu, MSR_PP1_POLICY, &msr)) + return -8; + + fprintf(stderr, "cpu%d: MSR_PP1_POLICY: %lld\n", cpu, msr & 0xF); + + if (get_msr(cpu, MSR_PP1_POWER_LIMIT, &msr)) + return -9; + fprintf(stderr, "cpu%d: MSR_PP1_POWER_LIMIT: 0x%08llx (%slocked)\n", + cpu, msr, (msr >> 31) & 1 ? "": "UN"); + print_power_limit_msr(cpu, msr, "GFX Limit"); + } + } + return 0; +} + int is_snb(unsigned int family, unsigned int model) { @@ -829,19 +1851,146 @@ int is_snb(unsigned int family, unsigned int model) switch (model) { case 0x2A: case 0x2D: + case 0x3A: /* IVB */ + case 0x3E: /* IVB Xeon */ + case 0x3C: /* HSW */ + case 0x3F: /* HSW */ + case 0x45: /* HSW */ + case 0x46: /* HSW */ + case 0x3D: /* BDW */ + case 0x4F: /* BDX */ + case 0x56: /* BDX-DE */ + return 1; + } + return 0; +} + +int has_c8_c9_c10(unsigned int family, unsigned int model) +{ + if (!genuine_intel) + return 0; + + switch (model) { + case 0x45: /* HSW */ + case 0x3D: /* BDW */ + return 1; + } + return 0; +} + + +int is_slm(unsigned int family, unsigned int model) +{ + if (!genuine_intel) + return 0; + switch (model) { + case 0x37: /* BYT */ + case 0x4D: /* AVN */ return 1; } return 0; } +#define SLM_BCLK_FREQS 5 +double slm_freq_table[SLM_BCLK_FREQS] = { 83.3, 100.0, 133.3, 116.7, 80.0}; + +double slm_bclk(void) +{ + unsigned long long msr = 3; + unsigned int i; + double freq; + + if (get_msr(0, MSR_FSB_FREQ, &msr)) + fprintf(stderr, "SLM BCLK: unknown\n"); + + i = msr & 0xf; + if (i >= SLM_BCLK_FREQS) { + fprintf(stderr, "SLM BCLK[%d] invalid\n", i); + msr = 3; + } + freq = slm_freq_table[i]; + + fprintf(stderr, "SLM BCLK: %.1f Mhz\n", freq); + + return freq; +} + double discover_bclk(unsigned int family, unsigned int model) { if (is_snb(family, model)) return 100.00; + else if (is_slm(family, model)) + return slm_bclk(); else return 133.33; } +/* + * MSR_IA32_TEMPERATURE_TARGET indicates the temperature where + * the Thermal Control Circuit (TCC) activates. + * This is usually equal to tjMax. + * + * Older processors do not have this MSR, so there we guess, + * but also allow cmdline over-ride with -T. + * + * Several MSR temperature values are in units of degrees-C + * below this value, including the Digital Thermal Sensor (DTS), + * Package Thermal Management Sensor (PTM), and thermal event thresholds. + */ +int set_temperature_target(struct thread_data *t, struct core_data *c, struct pkg_data *p) +{ + unsigned long long msr; + unsigned int target_c_local; + int cpu; + + /* tcc_activation_temp is used only for dts or ptm */ + if (!(do_dts || do_ptm)) + return 0; + + /* this is a per-package concept */ + if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE) || !(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) + return 0; + + cpu = t->cpu_id; + if (cpu_migrate(cpu)) { + fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + return -1; + } + + if (tcc_activation_temp_override != 0) { + tcc_activation_temp = tcc_activation_temp_override; + fprintf(stderr, "cpu%d: Using cmdline TCC Target (%d C)\n", + cpu, tcc_activation_temp); + return 0; + } + + /* Temperature Target MSR is Nehalem and newer only */ + if (!do_nehalem_platform_info) + goto guess; + + if (get_msr(0, MSR_IA32_TEMPERATURE_TARGET, &msr)) + goto guess; + + target_c_local = (msr >> 16) & 0xFF; + + if (verbose) + fprintf(stderr, "cpu%d: MSR_IA32_TEMPERATURE_TARGET: 0x%08llx (%d C)\n", + cpu, msr, target_c_local); + + if (!target_c_local) + goto guess; + + tcc_activation_temp = target_c_local; + + return 0; + +guess: + tcc_activation_temp = TJMAX_DEFAULT; + fprintf(stderr, "cpu%d: Guessing tjMax %d C, Please use -T to specify\n", + cpu, tcc_activation_temp); + + return 0; +} void check_cpuid() { unsigned int eax, ebx, ecx, edx, max_level; @@ -849,16 +1998,16 @@ void check_cpuid() eax = ebx = ecx = edx = 0; - asm("cpuid" : "=a" (max_level), "=b" (ebx), "=c" (ecx), "=d" (edx) : "a" (0)); + __get_cpuid(0, &max_level, &ebx, &ecx, &edx); if (ebx == 0x756e6547 && edx == 0x49656e69 && ecx == 0x6c65746e) genuine_intel = 1; if (verbose) - fprintf(stderr, "%.4s%.4s%.4s ", + fprintf(stderr, "CPUID(0): %.4s%.4s%.4s ", (char *)&ebx, (char *)&edx, (char *)&ecx); - asm("cpuid" : "=a" (fms), "=c" (ecx), "=d" (edx) : "a" (1) : "ebx"); + __get_cpuid(1, &fms, &ebx, &ecx, &edx); family = (fms >> 8) & 0xf; model = (fms >> 4) & 0xf; stepping = fms & 0xf; @@ -869,10 +2018,8 @@ void check_cpuid() fprintf(stderr, "%d CPUID levels; family:model:stepping 0x%x:%x:%x (%d:%d:%d)\n", max_level, family, model, stepping, family, model, stepping); - if (!(edx & (1 << 5))) { - fprintf(stderr, "CPUID: no MSR\n"); - exit(1); - } + if (!(edx & (1 << 5))) + errx(1, "CPUID: no MSR"); /* * check max extended function levels of CPUID. @@ -880,51 +2027,62 @@ void check_cpuid() * This check is valid for both Intel and AMD. */ ebx = ecx = edx = 0; - asm("cpuid" : "=a" (max_level), "=b" (ebx), "=c" (ecx), "=d" (edx) : "a" (0x80000000)); + __get_cpuid(0x80000000, &max_level, &ebx, &ecx, &edx); - if (max_level < 0x80000007) { - fprintf(stderr, "CPUID: no invariant TSC (max_level 0x%x)\n", max_level); - exit(1); - } + if (max_level < 0x80000007) + errx(1, "CPUID: no invariant TSC (max_level 0x%x)", max_level); /* * Non-Stop TSC is advertised by CPUID.EAX=0x80000007: EDX.bit8 * this check is valid for both Intel and AMD */ - asm("cpuid" : "=a" (eax), "=b" (ebx), "=c" (ecx), "=d" (edx) : "a" (0x80000007)); + __get_cpuid(0x80000007, &eax, &ebx, &ecx, &edx); has_invariant_tsc = edx & (1 << 8); - if (!has_invariant_tsc) { - fprintf(stderr, "No invariant TSC\n"); - exit(1); - } + if (!has_invariant_tsc) + errx(1, "No invariant TSC"); /* * APERF/MPERF is advertised by CPUID.EAX=0x6: ECX.bit0 * this check is valid for both Intel and AMD */ - asm("cpuid" : "=a" (eax), "=b" (ebx), "=c" (ecx), "=d" (edx) : "a" (0x6)); + __get_cpuid(0x6, &eax, &ebx, &ecx, &edx); has_aperf = ecx & (1 << 0); - if (!has_aperf) { - fprintf(stderr, "No APERF MSR\n"); - exit(1); - } + do_dts = eax & (1 << 0); + do_ptm = eax & (1 << 6); + has_epb = ecx & (1 << 3); + + if (verbose) + fprintf(stderr, "CPUID(6): %s%s%s%s\n", + has_aperf ? "APERF" : "No APERF!", + do_dts ? ", DTS" : "", + do_ptm ? ", PTM": "", + has_epb ? ", EPB": ""); + + if (!has_aperf) + errx(-1, "No APERF"); do_nehalem_platform_info = genuine_intel && has_invariant_tsc; do_nhm_cstates = genuine_intel; /* all Intel w/ non-stop TSC have NHM counters */ + do_smi = do_nhm_cstates; do_snb_cstates = is_snb(family, model); + do_c8_c9_c10 = has_c8_c9_c10(family, model); + do_slm_cstates = is_slm(family, model); bclk = discover_bclk(family, model); do_nehalem_turbo_ratio_limit = has_nehalem_turbo_ratio_limit(family, model); + do_ivt_turbo_ratio_limit = has_ivt_turbo_ratio_limit(family, model); + rapl_probe(family, model); + + return; } void usage() { - fprintf(stderr, "%s: [-v] [-M MSR#] [-i interval_sec | command ...]\n", - progname); - exit(1); + errx(1, "%s: [-v][-R][-T][-p|-P|-S][-c MSR#][-C MSR#][-m MSR#][-M MSR#][-i interval_sec | command ...]\n", + progname); } @@ -945,6 +2103,200 @@ int open_dev_cpu_msr(int dummy1) return 0; } +void topology_probe() +{ + int i; + int max_core_id = 0; + int max_package_id = 0; + int max_siblings = 0; + struct cpu_topology { + int core_id; + int physical_package_id; + } *cpus; + + /* Initialize num_cpus, max_cpu_num */ + topo.num_cpus = 0; + topo.max_cpu_num = 0; + for_all_proc_cpus(count_cpus); + if (!summary_only && topo.num_cpus > 1) + show_cpu = 1; + + if (verbose > 1) + fprintf(stderr, "num_cpus %d max_cpu_num %d\n", topo.num_cpus, topo.max_cpu_num); + + cpus = calloc(1, (topo.max_cpu_num + 1) * sizeof(struct cpu_topology)); + if (cpus == NULL) + err(1, "calloc cpus"); + + /* + * Allocate and initialize cpu_present_set + */ + cpu_present_set = CPU_ALLOC((topo.max_cpu_num + 1)); + if (cpu_present_set == NULL) + err(3, "CPU_ALLOC"); + cpu_present_setsize = CPU_ALLOC_SIZE((topo.max_cpu_num + 1)); + CPU_ZERO_S(cpu_present_setsize, cpu_present_set); + for_all_proc_cpus(mark_cpu_present); + + /* + * Allocate and initialize cpu_affinity_set + */ + cpu_affinity_set = CPU_ALLOC((topo.max_cpu_num + 1)); + if (cpu_affinity_set == NULL) + err(3, "CPU_ALLOC"); + cpu_affinity_setsize = CPU_ALLOC_SIZE((topo.max_cpu_num + 1)); + CPU_ZERO_S(cpu_affinity_setsize, cpu_affinity_set); + + + /* + * For online cpus + * find max_core_id, max_package_id + */ + for (i = 0; i <= topo.max_cpu_num; ++i) { + int siblings; + + if (cpu_is_not_present(i)) { + if (verbose > 1) + fprintf(stderr, "cpu%d NOT PRESENT\n", i); + continue; + } + cpus[i].core_id = get_core_id(i); + if (cpus[i].core_id > max_core_id) + max_core_id = cpus[i].core_id; + + cpus[i].physical_package_id = get_physical_package_id(i); + if (cpus[i].physical_package_id > max_package_id) + max_package_id = cpus[i].physical_package_id; + + siblings = get_num_ht_siblings(i); + if (siblings > max_siblings) + max_siblings = siblings; + if (verbose > 1) + fprintf(stderr, "cpu %d pkg %d core %d\n", + i, cpus[i].physical_package_id, cpus[i].core_id); + } + topo.num_cores_per_pkg = max_core_id + 1; + if (verbose > 1) + fprintf(stderr, "max_core_id %d, sizing for %d cores per package\n", + max_core_id, topo.num_cores_per_pkg); + if (!summary_only && topo.num_cores_per_pkg > 1) + show_core = 1; + + topo.num_packages = max_package_id + 1; + if (verbose > 1) + fprintf(stderr, "max_package_id %d, sizing for %d packages\n", + max_package_id, topo.num_packages); + if (!summary_only && topo.num_packages > 1) + show_pkg = 1; + + topo.num_threads_per_core = max_siblings; + if (verbose > 1) + fprintf(stderr, "max_siblings %d\n", max_siblings); + + free(cpus); +} + +void +allocate_counters(struct thread_data **t, struct core_data **c, struct pkg_data **p) +{ + int i; + + *t = calloc(topo.num_threads_per_core * topo.num_cores_per_pkg * + topo.num_packages, sizeof(struct thread_data)); + if (*t == NULL) + goto error; + + for (i = 0; i < topo.num_threads_per_core * + topo.num_cores_per_pkg * topo.num_packages; i++) + (*t)[i].cpu_id = -1; + + *c = calloc(topo.num_cores_per_pkg * topo.num_packages, + sizeof(struct core_data)); + if (*c == NULL) + goto error; + + for (i = 0; i < topo.num_cores_per_pkg * topo.num_packages; i++) + (*c)[i].core_id = -1; + + *p = calloc(topo.num_packages, sizeof(struct pkg_data)); + if (*p == NULL) + goto error; + + for (i = 0; i < topo.num_packages; i++) + (*p)[i].package_id = i; + + return; +error: + err(1, "calloc counters"); +} +/* + * init_counter() + * + * set cpu_id, core_num, pkg_num + * set FIRST_THREAD_IN_CORE and FIRST_CORE_IN_PACKAGE + * + * increment topo.num_cores when 1st core in pkg seen + */ +void init_counter(struct thread_data *thread_base, struct core_data *core_base, + struct pkg_data *pkg_base, int thread_num, int core_num, + int pkg_num, int cpu_id) +{ + struct thread_data *t; + struct core_data *c; + struct pkg_data *p; + + t = GET_THREAD(thread_base, thread_num, core_num, pkg_num); + c = GET_CORE(core_base, core_num, pkg_num); + p = GET_PKG(pkg_base, pkg_num); + + t->cpu_id = cpu_id; + if (thread_num == 0) { + t->flags |= CPU_IS_FIRST_THREAD_IN_CORE; + if (cpu_is_first_core_in_package(cpu_id)) + t->flags |= CPU_IS_FIRST_CORE_IN_PACKAGE; + } + + c->core_id = core_num; + p->package_id = pkg_num; +} + + +int initialize_counters(int cpu_id) +{ + int my_thread_id, my_core_id, my_package_id; + + my_package_id = get_physical_package_id(cpu_id); + my_core_id = get_core_id(cpu_id); + + if (cpu_is_first_sibling_in_core(cpu_id)) { + my_thread_id = 0; + topo.num_cores++; + } else { + my_thread_id = 1; + } + + init_counter(EVEN_COUNTERS, my_thread_id, my_core_id, my_package_id, cpu_id); + init_counter(ODD_COUNTERS, my_thread_id, my_core_id, my_package_id, cpu_id); + return 0; +} + +void allocate_output_buffer() +{ + output_buffer = calloc(1, (1 + topo.num_cpus) * 1024); + outp = output_buffer; + if (outp == NULL) + err(-1, "calloc output buffer"); +} + +void setup_all_buffers(void) +{ + topology_probe(); + allocate_counters(&thread_even, &core_even, &package_even); + allocate_counters(&thread_odd, &core_odd, &package_odd); + allocate_output_buffer(); + for_all_proc_cpus(initialize_counters); +} + void turbostat_init() { check_cpuid(); @@ -952,17 +2304,33 @@ void turbostat_init() check_dev_msr(); check_super_user(); - num_cpus = for_all_cpus(alloc_new_counters); + setup_all_buffers(); + + if (verbose) + print_verbose_header(); + + if (verbose) + for_all_cpus(print_epb, ODD_COUNTERS); + + if (verbose) + for_all_cpus(print_rapl, ODD_COUNTERS); + + for_all_cpus(set_temperature_target, ODD_COUNTERS); if (verbose) - print_nehalem_info(); + for_all_cpus(print_thermal, ODD_COUNTERS); } int fork_it(char **argv) { - int retval; pid_t child_pid; - get_counters(cnt_even); + int status; + + status = for_all_cpus(get_counters, EVEN_COUNTERS); + if (status) + exit(status); + /* clear affinity side-effect of get_counters() */ + sched_setaffinity(0, cpu_present_setsize, cpu_present_set); gettimeofday(&tv_even, (struct timezone *)NULL); child_pid = fork(); @@ -970,33 +2338,48 @@ int fork_it(char **argv) /* child */ execvp(argv[0], argv); } else { - int status; /* parent */ - if (child_pid == -1) { - perror("fork"); - exit(1); - } + if (child_pid == -1) + err(1, "fork"); signal(SIGINT, SIG_IGN); signal(SIGQUIT, SIG_IGN); - if (waitpid(child_pid, &status, 0) == -1) { - perror("wait"); - exit(1); - } + if (waitpid(child_pid, &status, 0) == -1) + err(status, "waitpid"); } - get_counters(cnt_odd); + /* + * n.b. fork_it() does not check for errors from for_all_cpus() + * because re-starting is problematic when forking + */ + for_all_cpus(get_counters, ODD_COUNTERS); gettimeofday(&tv_odd, (struct timezone *)NULL); - retval = compute_delta(cnt_odd, cnt_even, cnt_delta); - timersub(&tv_odd, &tv_even, &tv_delta); - compute_average(cnt_delta, cnt_average); - if (!retval) - print_counters(cnt_delta); + for_all_cpus_2(delta_cpu, ODD_COUNTERS, EVEN_COUNTERS); + compute_average(EVEN_COUNTERS); + format_all_counters(EVEN_COUNTERS); + flush_stderr(); fprintf(stderr, "%.6f sec\n", tv_delta.tv_sec + tv_delta.tv_usec/1000000.0); - return 0; + return status; +} + +int get_and_dump_counters(void) +{ + int status; + + status = for_all_cpus(get_counters, ODD_COUNTERS); + if (status) + return status; + + status = for_all_cpus(dump_counters, ODD_COUNTERS); + if (status) + return status; + + flush_stdout(); + + return status; } void cmdline(int argc, char **argv) @@ -1005,19 +2388,48 @@ void cmdline(int argc, char **argv) progname = argv[0]; - while ((opt = getopt(argc, argv, "+vi:M:")) != -1) { + while ((opt = getopt(argc, argv, "+pPsSvi:c:C:m:M:RJT:")) != -1) { switch (opt) { + case 'p': + show_core_only++; + break; + case 'P': + show_pkg_only++; + break; + case 's': + dump_only++; + break; + case 'S': + summary_only++; + break; case 'v': verbose++; break; case 'i': interval_sec = atoi(optarg); break; + case 'c': + sscanf(optarg, "%x", &extra_delta_offset32); + break; + case 'C': + sscanf(optarg, "%x", &extra_delta_offset64); + break; + case 'm': + sscanf(optarg, "%x", &extra_msr_offset32); + break; case 'M': - sscanf(optarg, "%x", &extra_msr_offset); - if (verbose > 1) - fprintf(stderr, "MSR 0x%X\n", extra_msr_offset); + sscanf(optarg, "%x", &extra_msr_offset64); + break; + case 'R': + rapl_verbose++; + break; + case 'T': + tcc_activation_temp_override = atoi(optarg); + break; + case 'J': + rapl_joules++; break; + default: usage(); } @@ -1028,14 +2440,16 @@ int main(int argc, char **argv) { cmdline(argc, argv); - if (verbose > 1) - fprintf(stderr, "turbostat Dec 6, 2010" + if (verbose) + fprintf(stderr, "turbostat v3.7 Feb 6, 2014" " - Len Brown <lenb@kernel.org>\n"); - if (verbose > 1) - fprintf(stderr, "http://userweb.kernel.org/~lenb/acpi/utils/pmtools/turbostat/\n"); turbostat_init(); + /* dump counters and exit */ + if (dump_only) + return get_and_dump_counters(); + /* * if any params left, it must be a command to fork */ diff --git a/tools/power/x86/x86_energy_perf_policy/Makefile b/tools/power/x86/x86_energy_perf_policy/Makefile index f458237fdd7..971c9ffdcb5 100644 --- a/tools/power/x86/x86_energy_perf_policy/Makefile +++ b/tools/power/x86/x86_energy_perf_policy/Makefile @@ -1,8 +1,10 @@ +DESTDIR ?= + x86_energy_perf_policy : x86_energy_perf_policy.c clean : rm -f x86_energy_perf_policy install : - install x86_energy_perf_policy /usr/bin/ - install x86_energy_perf_policy.8 /usr/share/man/man8/ + install x86_energy_perf_policy ${DESTDIR}/usr/bin/ + install x86_energy_perf_policy.8 ${DESTDIR}/usr/share/man/man8/ diff --git a/tools/power/x86/x86_energy_perf_policy/x86_energy_perf_policy.c b/tools/power/x86/x86_energy_perf_policy/x86_energy_perf_policy.c index 33c5c7ee148..40b3e5482f8 100644 --- a/tools/power/x86/x86_energy_perf_policy/x86_energy_perf_policy.c +++ b/tools/power/x86/x86_energy_perf_policy/x86_energy_perf_policy.c @@ -289,7 +289,7 @@ void for_every_cpu(void (func)(int)) "cpu%u %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d\n", &cpu); if (retval != 1) - return; + break; func(cpu); } |