apparmor/parser/parser_main.c

1202 lines
30 KiB
C
Raw Normal View History

/*
2007-04-11 08:12:51 +00:00
* Copyright (c) 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007
* NOVELL (All rights reserved)
*
* Copyright (c) 2010 - 2013
* Canonical Ltd. (All rights reserved)
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of version 2 of the GNU General Public
* License published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, contact Novell, Inc. or Canonical,
* Ltd.
*/
#include <ctype.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <stdarg.h>
#include <stddef.h>
#include <getopt.h>
#include <errno.h>
#include <fcntl.h>
/* enable the following line to get voluminous debug info */
/* #define DEBUG */
#include <unistd.h>
#include <limits.h>
#include <sys/types.h>
#include <sys/stat.h>
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
#include <sys/wait.h>
parser: Add make variable to build against local or system libapparmor [v3] By default, statically link against the in-tree libapparmor. If the in-tree libapparmor is not yet built, print a helpful error message. To build against the system libapparmor, the USE_SYSTEM make variable can be set on the command line like so: $ make USE_SYSTEM=1 This patch also fixes issues around the inclusion of the apparmor.h header. Previously, the in-tree apparmor.h was always being included even if the parser was being linked against the system libapparmor. It modifies the apparmor.h include path based on the previous patch separating them out in the libapparmor source. This was needed because header file name collisions were already occurring. For source files needing to include apparmor.h, the make targets were also updated to depend on the local apparmor.h when building against the in-tree libapparmor. When building against the system libapparmor, the variable used in the dependency list is empty. Likewise, a libapparmor.a dependency is added to the apparmor_parser target when building against the in-tree apparmor. Patch history: v1: from Tyler Hicks <tyhicks@canonical.com> - initial version v2: revert to altering the include search path rather than including the apparmor.h header directly via cpp arguments, alter the include statements to <sys/apparmor.h> which will work against either in-tree or (default) system paths. v3: convert controlling variable to USE_SYSTEM from SYSTEM_LIBAPPARMOR to unify between the parser and the regression tests. Signed-off-by: Tyler Hicks <tyhicks@canonical.com> Signed-off-by: Steve Beattie <steve@nxnw.org>
2014-01-06 14:46:10 -08:00
#include <sys/apparmor.h>
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
#include "lib.h"
#include "features.h"
#include "parser.h"
#include "parser_version.h"
#include "parser_include.h"
#include "common_optarg.h"
#include "policy_cache.h"
#include "libapparmor_re/apparmor_re.h"
#define OLD_MODULE_NAME "subdomain"
#define PROC_MODULES "/proc/modules"
#define MATCH_FILE "/sys/kernel/security/" MODULE_NAME "/matching"
#define MOUNTED_FS "/proc/mounts"
2007-02-27 02:29:16 +00:00
#define AADFA "pattern=aadfa"
#define PRIVILEGED_OPS (kernel_load)
#define UNPRIVILEGED_OPS (!(PRIVILEGED_OPS))
const char *parser_title = "AppArmor parser";
const char *parser_copyright = "Copyright (C) 1999-2008 Novell Inc.\nCopyright 2009-2012 Canonical Ltd.";
int opt_force_complain = 0;
int binary_input = 0;
int dump_vars = 0;
int dump_expanded_vars = 0;
int show_cache = 0;
int skip_cache = 0;
int skip_read_cache = 0;
int write_cache = 0;
int cond_clear_cache = 1; /* only applies if write is set */
int force_clear_cache = 0; /* force clearing regargless of state */
int create_cache_dir = 0; /* DEPRECATED in favor of write_cache */
int preprocess_only = 0;
int skip_mode_force = 0;
int abort_on_error = 0; /* stop processing profiles if error */
int skip_bad_cache_rebuild = 0;
int mru_skip_cache = 1;
int debug_cache = 0;
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
/* for jobs_max and jobs
* LONG_MAX : no limit
* 0 : auto = detect system processing cores
* n : use that number of processes/threads to compile policy
*/
#define JOBS_AUTO 0
long jobs_max = -8; /* 8 * cpus */
long jobs = JOBS_AUTO; /* default: number of processor cores */
long njobs = 0;
long jobs_scale = 0; /* number of chance to resample online
* cpus. This allows jobs spawning to
* scale when scheduling policy is
* taking cpus off line, and brings
* them back with load
*/
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
bool debug_jobs = false;
struct timespec cache_tstamp, mru_policy_tstamp;
static char *apparmorfs = NULL;
static char *cacheloc = NULL;
static aa_features *features = NULL;
/* Make sure to update BOTH the short and long_options */
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
static const char *short_options = "ad::f:h::rRVvI:b:BCD:NSm:M:qQn:XKTWkL:O:po:j:";
struct option long_options[] = {
{"add", 0, 0, 'a'},
{"binary", 0, 0, 'B'},
{"base", 1, 0, 'b'},
{"subdomainfs", 1, 0, 'f'},
{"help", 2, 0, 'h'},
{"replace", 0, 0, 'r'},
{"reload", 0, 0, 'r'}, /* undocumented reload option == replace */
{"version", 0, 0, 'V'},
{"complain", 0, 0, 'C'},
{"Complain", 0, 0, 'C'}, /* Erk, apparently documented as --Complain */
{"Include", 1, 0, 'I'},
{"remove", 0, 0, 'R'},
2008-06-09 22:15:28 +00:00
{"names", 0, 0, 'N'},
{"stdout", 0, 0, 'S'},
{"ofile", 1, 0, 'o'},
{"match-string", 1, 0, 'm'},
{"features-file", 1, 0, 'M'},
{"quiet", 0, 0, 'q'},
{"skip-kernel-load", 0, 0, 'Q'},
{"verbose", 0, 0, 'v'},
{"namespace", 1, 0, 'n'},
{"readimpliesX", 0, 0, 'X'},
{"skip-cache", 0, 0, 'K'},
{"skip-read-cache", 0, 0, 'T'},
{"write-cache", 0, 0, 'W'},
{"show-cache", 0, 0, 'k'},
{"skip-bad-cache", 0, 0, 129}, /* no short option */
{"purge-cache", 0, 0, 130}, /* no short option */
{"create-cache-dir", 0, 0, 131}, /* no short option */
{"cache-loc", 1, 0, 'L'},
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
{"debug", 2, 0, 'd'},
{"dump", 1, 0, 'D'},
{"Dump", 1, 0, 'D'},
{"optimize", 1, 0, 'O'},
{"Optimize", 1, 0, 'O'},
{"preprocess", 0, 0, 'p'},
{"abort-on-error", 0, 0, 132}, /* no short option */
{"skip-bad-cache-rebuild", 0, 0, 133}, /* no short option */
{"warn", 1, 0, 134}, /* no short option */
{"debug-cache", 0, 0, 135}, /* no short option */
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
{"jobs", 1, 0, 'j'},
{"max-jobs", 1, 0, 136}, /* no short option */
{NULL, 0, 0, 0},
};
static int debug = 0;
void display_version(void)
{
printf("%s version " PARSER_VERSION "\n%s\n", parser_title,
parser_copyright);
}
static void display_usage(const char *command)
{
display_version();
printf("\nUsage: %s [options] [profile]\n\n"
"Options:\n"
"--------\n"
"-a, --add Add apparmor definitions [default]\n"
"-r, --replace Replace apparmor definitions\n"
"-R, --remove Remove apparmor definitions\n"
"-C, --Complain Force the profile into complain mode\n"
2008-06-09 22:15:28 +00:00
"-B, --binary Input is precompiled profile\n"
"-N, --names Dump names of profiles in input.\n"
"-S, --stdout Dump compiled profile to stdout\n"
"-o n, --ofile n Write output to file n\n"
"-b n, --base n Set base dir and cwd\n"
2008-06-09 22:15:28 +00:00
"-I n, --Include n Add n to the search path\n"
"-f n, --subdomainfs n Set location of apparmor filesystem\n"
"-m n, --match-string n Use only features n\n"
"-M n, --features-file n Use only features in file n\n"
"-n n, --namespace n Set Namespace for the profile\n"
"-X, --readimpliesX Map profile read permissions to mr\n"
"-k, --show-cache Report cache hit/miss details\n"
"-K, --skip-cache Do not attempt to load or save cached profiles\n"
"-T, --skip-read-cache Do not attempt to load cached profiles\n"
"-W, --write-cache Save cached profile (force with -T)\n"
" --skip-bad-cache Don't clear cache if out of sync\n"
" --purge-cache Clear cache regardless of its state\n"
" --debug-cache Debug cache file checks\n"
"-L, --cache-loc n Set the location of the profile cache\n"
"-q, --quiet Don't emit warnings\n"
"-v, --verbose Show profile names as they load\n"
"-Q, --skip-kernel-load Do everything except loading into kernel\n"
"-V, --version Display version info and exit\n"
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
"-d [n], --debug Debug apparmor definitions OR [n]\n"
"-p, --preprocess Dump preprocessed profile\n"
"-D [n], --dump Dump internal info for debugging\n"
"-O [n], --Optimize Control dfa optimizations\n"
"-h [cmd], --help[=cmd] Display this text or info about cmd\n"
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
"-j n, --jobs n Set the number of compile threads\n"
"--max-jobs n Hard cap on --jobs. Default 8*cpus\n"
"--abort-on-error Abort processing of profiles on first error\n"
"--skip-bad-cache-rebuild Do not try rebuilding the cache if it is rejected by the kernel\n"
"--warn n Enable warnings (see --help=warn)\n"
,command);
}
optflag_table_t warnflag_table[] = {
{ 0, "rule-not-enforced", "warn if a rule is not enforced", WARN_RULE_NOT_ENFORCED },
{ 0, "rule-downgraded", "warn if a rule is downgraded to a lesser but still enforcing rule", WARN_RULE_DOWNGRADED },
{ 0, NULL, NULL, 0 },
};
void display_warn(const char *command)
{
display_version();
printf("\n%s: --warn [Option]\n\n"
"Options:\n"
"--------\n"
,command);
print_flag_table(warnflag_table);
}
/* Treat conf file like options passed on command line
*/
static int getopt_long_file(FILE *f, const struct option *longopts,
char **optarg, int *longindex)
{
static char line[256];
char *pos, *opt, *save;
int i;
for (;;) {
if (!fgets(line, 256, f))
return -1;
pos = line;
while (isblank(*pos))
pos++;
if (*pos == '#')
continue;
opt = strtok_r(pos, " \t\r\n=", &save);
/* blank line */
if (!opt)
continue;
for (i = 0; longopts[i].name &&
strcmp(longopts[i].name, opt) != 0; i++) ;
if (!longopts[i].name) {
PERROR("%s: unknown option (%s) in config file.\n",
progname, opt);
/* skip it */
continue;
}
break;
}
if (longindex)
*longindex = i;
if (*save) {
int len;
while(isblank(*save))
save++;
len = strlen(save) - 1;
if (save[len] == '\n')
save[len] = 0;
}
switch (longopts[i].has_arg) {
case 0:
*optarg = NULL;
break;
case 1:
if (!strlen(save)) {
*optarg = NULL;
return '?';
}
*optarg = save;
break;
case 2:
*optarg = save;
break;
default:
PERROR("%s: internal error bad longopt value\n", progname);
exit(1);
}
if (longopts[i].flag == NULL)
return longopts[i].val;
else
*longopts[i].flag = longopts[i].val;
return 0;
}
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
static long process_jobs_arg(const char *arg, const char *val) {
char *end;
long n;
if (!val || strcmp(val, "auto") == 0)
n = JOBS_AUTO;
else if (strcmp(val, "max") == 0)
n = LONG_MAX;
else {
bool multiple = false;
if (*val == 'x') {
multiple = true;
val++;
}
n = strtol(val, &end, 0);
if (!(*val && val != end && *end == '\0')) {
PERROR("%s: Invalid option %s=%s%s\n", progname, arg, multiple ? "x" : "", val);
exit(1);
}
if (multiple)
n = -n;
}
return n;
}
/* process a single argment from getopt_long
* Returns: 1 if an action arg, else 0
*/
static int process_arg(int c, char *optarg)
{
int count = 0;
switch (c) {
case 0:
PERROR("Assert, in getopt_long handling\n");
display_usage(progname);
exit(0);
break;
case 'a':
count++;
option = OPTION_ADD;
break;
case 'd':
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
if (!optarg) {
debug++;
skip_read_cache = 1;
} else if (strcmp(optarg, "jobs") == 0 ||
strcmp(optarg, "j") == 0) {
debug_jobs = true;
} else {
PERROR("%s: Invalid --debug option '%s'\n",
progname, optarg);
exit(1);
}
break;
case 'h':
if (!optarg) {
display_usage(progname);
} else if (strcmp(optarg, "Dump") == 0 ||
strcmp(optarg, "dump") == 0 ||
strcmp(optarg, "D") == 0) {
display_dump(progname);
} else if (strcmp(optarg, "Optimize") == 0 ||
strcmp(optarg, "optimize") == 0 ||
strcmp(optarg, "O") == 0) {
display_optimize(progname);
} else if (strcmp(optarg, "warn") == 0) {
display_warn(progname);
} else {
PERROR("%s: Invalid --help option %s\n",
progname, optarg);
exit(1);
}
exit(0);
break;
case 'r':
count++;
option = OPTION_REPLACE;
break;
case 'R':
count++;
option = OPTION_REMOVE;
skip_cache = 1;
break;
case 'V':
display_version();
exit(0);
break;
case 'I':
add_search_dir(optarg);
break;
case 'b':
set_base_dir(optarg);
break;
case 'B':
binary_input = 1;
skip_cache = 1;
break;
case 'C':
opt_force_complain = 1;
skip_cache = 1;
break;
case 'N':
count++;
names_only = 1;
skip_cache = 1;
kernel_load = 0;
break;
case 'S':
count++;
option = OPTION_STDOUT;
skip_read_cache = 1;
kernel_load = 0;
break;
case 'o':
count++;
option = OPTION_OFILE;
skip_read_cache = 1;
kernel_load = 0;
ofile = fopen(optarg, "w");
if (!ofile) {
PERROR("%s: Could not open file %s\n",
progname, optarg);
exit(1);
}
break;
case 'f':
apparmorfs = strndup(optarg, PATH_MAX);
break;
case 'D':
skip_read_cache = 1;
if (!optarg) {
dump_vars = 1;
} else if (strcmp(optarg, "variables") == 0) {
dump_vars = 1;
} else if (strcmp(optarg, "expanded-variables") == 0) {
dump_expanded_vars = 1;
} else if (!handle_flag_table(dumpflag_table, optarg,
&dfaflags)) {
PERROR("%s: Invalid --Dump option %s\n",
progname, optarg);
exit(1);
}
break;
case 'O':
skip_read_cache = 1;
if (!handle_flag_table(optflag_table, optarg,
&dfaflags)) {
PERROR("%s: Invalid --Optimize option %s\n",
progname, optarg);
exit(1);
}
break;
case 'm':
if (aa_features_new_from_string(&features,
optarg, strlen(optarg))) {
fprintf(stderr,
"Failed to parse features string: %m\n");
exit(1);
}
break;
case 'M':
if (aa_features_new(&features, AT_FDCWD, optarg)) {
fprintf(stderr,
"Failed to load features from '%s': %m\n",
optarg);
exit(1);
}
break;
case 'q':
conf_verbose = 0;
conf_quiet = 1;
warnflags = 0;
break;
case 'v':
conf_verbose = 1;
conf_quiet = 0;
break;
case 'n':
profile_ns = strdup(optarg);
break;
case 'X':
read_implies_exec = 1;
break;
case 'K':
skip_cache = 1;
break;
case 'k':
show_cache = 1;
break;
case 'W':
write_cache = 1;
break;
case 'T':
skip_read_cache = 1;
break;
case 129:
cond_clear_cache = 0;
break;
case 130:
force_clear_cache = 1;
break;
case 131:
create_cache_dir = 1;
break;
case 132:
abort_on_error = 1;
break;
case 133:
skip_bad_cache_rebuild = 1;
break;
case 'L':
cacheloc = strdup(optarg);
break;
case 'Q':
kernel_load = 0;
break;
case 'p':
count++;
kernel_load = 0;
skip_cache = 1;
preprocess_only = 1;
skip_mode_force = 1;
break;
case 134:
if (!handle_flag_table(warnflag_table, optarg,
&warnflags)) {
PERROR("%s: Invalid --warn option %s\n",
progname, optarg);
exit(1);
}
break;
case 135:
debug_cache = 1;
break;
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
case 'j':
jobs = process_jobs_arg("-j", optarg);
break;
case 136:
jobs_max = process_jobs_arg("max-jobs", optarg);
break;
default:
display_usage(progname);
exit(1);
break;
}
return count;
}
static int process_args(int argc, char *argv[])
{
int c, o;
int count = 0;
option = OPTION_ADD;
while ((c = getopt_long(argc, argv, short_options, long_options, &o)) != -1)
{
count += process_arg(c, optarg);
}
if (count > 1) {
PERROR("%s: Too many actions given on the command line.\n",
progname);
display_usage(progname);
exit(1);
}
PDEBUG("optind = %d argc = %d\n", optind, argc);
return optind;
}
static int process_config_file(const char *name)
{
char *optarg;
autofclose FILE *f = NULL;
int c, o;
f = fopen(name, "r");
if (!f)
return 0;
while ((c = getopt_long_file(f, long_options, &optarg, &o)) != -1)
process_arg(c, optarg);
return 1;
}
int have_enough_privilege(void)
{
uid_t uid, euid;
uid = getuid();
euid = geteuid();
if (uid != 0 && euid != 0) {
PERROR(_("%s: Sorry. You need root privileges to run this program.\n\n"),
progname);
display_usage(progname);
return EPERM;
}
if (uid != 0 && euid == 0) {
PERROR(_("%s: Warning! You've set this program setuid root.\n"
"Anybody who can run this program can update "
"your AppArmor profiles.\n\n"), progname);
}
return 0;
}
static void set_features_by_match_file(void)
{
autofclose FILE *ms = fopen(MATCH_FILE, "r");
if (ms) {
autofree char *match_string = (char *) malloc(1000);
if (!match_string)
goto no_match;
if (!fgets(match_string, 1000, ms))
goto no_match;
if (strstr(match_string, " perms=c"))
perms_create = 1;
kernel_supports_network = 1;
return;
2007-02-27 02:29:16 +00:00
}
no_match:
perms_create = 1;
}
static void set_supported_features(void)
{
/* has process_args() already assigned a match string? */
if (!features && aa_features_new_from_kernel(&features) == -1) {
set_features_by_match_file();
return;
}
perms_create = 1;
kernel_supports_policydb = aa_features_supports(features, "file");
kernel_supports_network = aa_features_supports(features, "network");
kernel_supports_unix = aa_features_supports(features,
"network/af_unix");
kernel_supports_mount = aa_features_supports(features, "mount");
kernel_supports_dbus = aa_features_supports(features, "dbus");
kernel_supports_signal = aa_features_supports(features, "signal");
kernel_supports_ptrace = aa_features_supports(features, "ptrace");
kernel_supports_setload = aa_features_supports(features,
"policy/set_load");
kernel_supports_diff_encode = aa_features_supports(features,
"policy/diff_encode");
kernel_supports_stacking = aa_features_supports(features,
"domain/stack");
if (aa_features_supports(features, "policy/versions/v7"))
parser: first step implementing fine grained mediation for unix domain sockets This patch implements parsing of fine grained mediation for unix domain sockets, that have abstract and anonymous paths. Sockets with file system paths are handled by regular file access rules. The unix network rules follow the general fine grained network rule pattern of [<qualifiers>] af_name [<access expr>] [<rule conds>] [<local expr>] [<peer expr>] specifically for af_unix this is [<qualifiers>] 'unix' [<access expr>] [<rule conds>] [<local expr>] [<peer expr>] <qualifiers> = [ 'audit' ] [ 'allow' | 'deny' ] <access expr> = ( <access> | <access list> ) <access> = ( 'server' | 'create' | 'bind' | 'listen' | 'accept' | 'connect' | 'shutdown' | 'getattr' | 'setattr' | 'getopt' | 'setopt' | 'send' | 'receive' | 'r' | 'w' | 'rw' ) (some access modes are incompatible with some rules or require additional parameters) <access list> = '(' <access> ( [','] <WS> <access> )* ')' <WS> = white space <rule conds> = ( <type cond> | <protocol cond> )* each cond can appear at most once <type cond> = 'type' '=' ( <AARE> | '(' ( '"' <AARE> '"' | <AARE> )+ ')' ) <protocol cond> = 'protocol' '=' ( <AARE> | '(' ( '"' <AARE> '"' | <AARE> )+ ')' ) <local expr> = ( <path cond> | <attr cond> | <opt cond> )* each cond can appear at most once <peer expr> = 'peer' '=' ( <path cond> | <label cond> )+ each cond can appear at most once <path cond> = 'path' '=' ( <AARE> | '(' '"' <AARE> '"' | <AARE> ')' ) <label cond> = 'label' '=' ( <AARE> | '(' '"' <AARE> '"' | <AARE> ')') <attr cond> = 'attr' '=' ( <AARE> | '(' '"' <AARE> '"' | <AARE> ')' ) <opt cond> = 'opt' '=' ( <AARE> | '(' '"' <AARE> '"' | <AARE> ')' ) <AARE> = ?*[]{}^ ( see man page ) unix domain socket rules are accumulated so that the granted unix socket permissions are the union of all the listed unix rule permissions. unix domain socket rules are broad and general and become more restrictive as further information is specified. Policy may be specified down to the path and label level. The content of the communication is not examined. Some permissions are not compatible with all unix rules. unix socket rule permissions are implied when a rule does not explicitly state an access list. By default if a rule does not have an access list all permissions that are compatible with the specified set of local and peer conditionals are implied. The 'server', 'r', 'w' and 'rw' permissions are aliases for other permissions. server = (create, bind, listen, accept) r = (receive, getattr, getopt) w = (create, connect, send, setattr, setopt) In addition it supports the v7 kernel abi semantics around generic network rules. The v7 abi removes the masking unix and netlink address families from the generic masking and uses fine grained mediation for an address type if supplied. This means that the rules network unix, network netlink, are now enforced instead of ignored. The parser previously could accept these but the kernel would ignore anything written to them. If a network rule is supplied it takes precedence over the finer grained mediation rule. If permission is not granted via a broad network access rule fine grained mediation is applied. Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2014-09-03 13:22:26 -07:00
kernel_abi_version = 7;
else if (aa_features_supports(features, "policy/versions/v6"))
kernel_abi_version = 6;
if (!kernel_supports_diff_encode)
/* clear diff_encode because it is not supported */
dfaflags &= ~DFA_CONTROL_DIFF_ENCODE;
}
int process_binary(int option, aa_kernel_interface *kernel_interface,
const char *profilename)
{
const char *printed_name;
int retval;
printed_name = profilename ? profilename : "stdin";
if (kernel_load) {
if (option == OPTION_ADD) {
retval = profilename ?
aa_kernel_interface_load_policy_from_file(kernel_interface, AT_FDCWD, profilename) :
aa_kernel_interface_load_policy_from_fd(kernel_interface, 0);
if (retval == -1) {
retval = errno;
PERROR(_("Error: Could not load profile %s: %s\n"),
printed_name, strerror(retval));
return retval;
}
} else if (option == OPTION_REPLACE) {
retval = profilename ?
aa_kernel_interface_replace_policy_from_file(kernel_interface, AT_FDCWD, profilename) :
aa_kernel_interface_replace_policy_from_fd(kernel_interface, 0);
if (retval == -1) {
retval = errno;
PERROR(_("Error: Could not replace profile %s: %s\n"),
printed_name, strerror(retval));
return retval;
}
} else {
PERROR(_("Error: Invalid load option specified: %d\n"),
option);
return EINVAL;
}
}
if (conf_verbose) {
switch (option) {
case OPTION_ADD:
printf(_("Cached load succeeded for \"%s\".\n"),
printed_name);
break;
case OPTION_REPLACE:
printf(_("Cached reload succeeded for \"%s\".\n"),
printed_name);
break;
default:
break;
}
}
return 0;
}
void reset_parser(const char *filename)
{
memset(&mru_policy_tstamp, 0, sizeof(mru_policy_tstamp));
memset(&cache_tstamp, 0, sizeof(cache_tstamp));
mru_skip_cache = 1;
free_aliases();
free_symtabs();
free_policies();
reset_include_stack(filename);
}
int test_for_dir_mode(const char *basename, const char *linkdir)
{
int rc = 0;
if (!skip_mode_force) {
autofree char *target = NULL;
if (asprintf(&target, "%s/%s/%s", basedir, linkdir, basename) < 0) {
perror("asprintf");
exit(1);
}
if (access(target, R_OK) == 0)
rc = 1;
}
return rc;
}
int process_profile(int option, aa_kernel_interface *kernel_interface,
const char *profilename, const char *cachedir)
{
int retval = 0;
autofree const char *cachename = NULL;
autofree const char *cachetmpname = NULL;
autoclose int cachetmp = -1;
const char *basename = NULL;
/* per-profile states */
force_complain = opt_force_complain;
if (profilename) {
if ( !(yyin = fopen(profilename, "r")) ) {
PERROR(_("Error: Could not read profile %s: %s.\n"),
profilename, strerror(errno));
return errno;
}
} else {
pwarn("%s: cannot use or update cache, disable, or force-complain via stdin\n", progname);
}
reset_parser(profilename);
if (profilename && option != OPTION_REMOVE) {
/* make decisions about disabled or complain-mode profiles */
basename = strrchr(profilename, '/');
if (basename)
basename++;
else
basename = profilename;
if (test_for_dir_mode(basename, "disable")) {
if (!conf_quiet)
PERROR("Skipping profile in %s/disable: %s\n", basedir, basename);
goto out;
}
if (test_for_dir_mode(basename, "force-complain")) {
PERROR("Warning: found %s in %s/force-complain, forcing complain mode\n", basename, basedir);
force_complain = 1;
}
/* setup cachename and tstamp */
if (!force_complain && !skip_cache) {
cachename = cache_filename(cachedir, basename);
valid_read_cache(cachename);
}
}
if (yyin) {
yyrestart(yyin);
update_mru_tstamp(yyin, profilename ? profilename : "stdin");
}
retval = yyparse();
if (retval != 0)
goto out;
/* Test to see if profile is for another namespace, if so disable
* caching for now
* TODO: Add support for caching profiles in an alternate namespace
* TODO: Add support for embedded namespace defines if they aren't
* removed from the language.
* TODO: test profile->ns NOT profile_ns (must be after parse)
*/
if (profile_ns)
skip_cache = 1;
if (cachename) {
/* Load a binary cache if it exists and is newest */
if (cache_hit(cachename)) {
retval = process_binary(option, kernel_interface,
cachename);
if (!retval || skip_bad_cache_rebuild)
return retval;
}
cachetmp = setup_cache_tmp(&cachetmpname, cachename);
}
if (show_cache)
PERROR("Cache miss: %s\n", profilename ? profilename : "stdin");
if (preprocess_only)
goto out;
if (names_only) {
dump_policy_names();
goto out;
}
if (dump_vars) {
dump_symtab();
goto out;
}
retval = post_process_policy(debug);
if (retval != 0) {
PERROR(_("%s: Errors found in file. Aborting.\n"), progname);
goto out;
}
if (dump_expanded_vars) {
dump_expanded_symtab();
goto out;
}
if (debug > 0) {
printf("----- Debugging built structures -----\n");
dump_policy();
goto out;
}
/* cache file generated by load_policy */
retval = load_policy(option, kernel_interface, cachetmp);
if (retval == 0 && write_cache) {
if (cachetmp == -1) {
unlink(cachetmpname);
PERROR("Warning failed to create cache: %s\n",
basename);
} else {
install_cache(cachetmpname, cachename);
}
}
out:
return retval;
}
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
/* Do not call directly, this is a helper for work_sync, which can handle
* single worker cases and cases were the work queue is optimized away
*
* call only if there are work children to wait on
*/
#define work_sync_one(RESULT) \
do { \
int status; \
wait(&status); \
if (WIFEXITED(status)) \
RESULT(WEXITSTATUS(status)); \
else \
RESULT(ECHILD); \
/* TODO: do we need to handle traced */ \
njobs--; \
if (debug_jobs) \
fprintf(stderr, " JOBS SYNC ONE: result %d, jobs left %ld\n", status, njobs); \
} while (0)
#define work_sync(RESULT) \
do { \
if (debug_jobs) \
fprintf(stderr, "JOBS SYNC: jobs left %ld\n", njobs); \
while (njobs) \
work_sync_one(RESULT); \
} while (0)
#define work_spawn(WORK, RESULT) \
do { \
/* what to do to avoid fork() overhead when single threaded \
if (jobs == 1) { \
// no parallel work so avoid fork() overhead \
RESULT(WORK); \
break; \
}*/ \
if (jobs_scale) { \
long n = sysconf(_SC_NPROCESSORS_ONLN); \
if (n > jobs_max) \
n = jobs_max; \
if (n > jobs) { \
/* reset sample chances - potentially reduce to 0 */ \
jobs_scale = jobs_max - n; \
jobs = n; \
} else \
/* reduce scaling chance by 1 */ \
jobs_scale--; \
} \
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
if (njobs == jobs) { \
/* wait for a child */ \
if (debug_jobs) \
fprintf(stderr, " JOBS SPAWN: waiting (jobs %ld == max %ld) ...\n", njobs, jobs); \
work_sync_one(RESULT); \
} \
\
pid_t child = fork(); \
if (child == 0) { \
/* child - exit work unit with returned value */ \
exit(WORK); \
} else if (child > 0) { \
/* parent */ \
njobs++; \
if (debug_jobs) \
fprintf(stderr, " JOBS SPAWN: created %ld ...\n", njobs); \
} else { \
/* error */ \
if (debug_jobs) \
fprintf(stderr, " JOBS SPAWN: failed error: %d) ...\n", errno); \
RESULT(errno); \
} \
} while (0)
/* sadly C forces us to do this with exit, long_jump or returning error
* from work_spawn and work_sync. We could throw a C++ exception, is it
* worth doing it to avoid the exit here.
*
* atm not all resources maybe cleanedup at exit
*/
int last_error = 0;
void handle_work_result(int retval)
{
if (retval) {
last_error = retval;
if (abort_on_error) {
/* already in abort mode we don't need subsequent
* syncs to do this too
*/
abort_on_error = 0;
work_sync(handle_work_result);
exit(last_error);
}
}
}
static long compute_jobs(long n, long j)
{
if (j == JOBS_AUTO)
j = n;
else if (j < 0)
j = n * j * -1;
return j;
}
static void setup_parallel_compile(void)
{
/* jobs and paralell_max set by default, config or args */
long n = sysconf(_SC_NPROCESSORS_ONLN);
long maxn = sysconf(_SC_NPROCESSORS_CONF);
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
if (n == -1)
/* unable to determine number of processors, default to 1 */
n = 1;
if (maxn == -1)
/* unable to determine number of processors, default to 1 */
maxn = 1;
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
jobs = compute_jobs(n, jobs);
jobs_max = compute_jobs(maxn, jobs_max);
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
if (jobs > jobs_max) {
pwarn("%s: Warning capping number of jobs to %ld * # of cpus == '%ld'",
progname, jobs_max, jobs);
jobs = jobs_max;
} else if (jobs < jobs_max)
/* the bigger the difference the more sample chances given */
jobs_scale = jobs_max + 1 - n;
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
njobs = 0;
if (debug_jobs)
fprintf(stderr, "jobs: %ld\n", jobs);
}
struct dir_cb_data {
aa_kernel_interface *kernel_interface;
const char *dirname; /* name of the parent dir */
const char *cachedir; /* path to the cache sub directory */
};
/* data - pointer to a dir_cb_data */
static int profile_dir_cb(int dirfd unused, const char *name, struct stat *st,
void *data)
{
int rc = 0;
if (!S_ISDIR(st->st_mode) && !is_blacklisted(name, NULL)) {
struct dir_cb_data *cb_data = (struct dir_cb_data *)data;
autofree char *path = NULL;
if (asprintf(&path, "%s/%s", cb_data->dirname, name) < 0)
PERROR(_("Out of memory"));
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
work_spawn(process_profile(option, cb_data->kernel_interface,
path, cb_data->cachedir),
handle_work_result);
}
return rc;
}
/* data - pointer to a dir_cb_data */
static int binary_dir_cb(int dirfd unused, const char *name, struct stat *st,
void *data)
{
int rc = 0;
if (!S_ISDIR(st->st_mode) && !is_blacklisted(name, NULL)) {
struct dir_cb_data *cb_data = (struct dir_cb_data *)data;
autofree char *path = NULL;
if (asprintf(&path, "%s/%s", cb_data->dirname, name) < 0)
PERROR(_("Out of memory"));
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
work_spawn(process_binary(option, cb_data->kernel_interface,
path),
handle_work_result);
}
return rc;
}
static void setup_flags(void)
{
/* Get the match string to determine type of regex support needed */
set_supported_features();
/* Gracefully handle AppArmor kernel without compatibility patch */
if (!features) {
PERROR("Cache read/write disabled: interface file missing. "
"(Kernel needs AppArmor 2.4 compatibility patch.)\n");
write_cache = 0;
skip_read_cache = 1;
return;
}
}
int main(int argc, char *argv[])
{
aa_kernel_interface *kernel_interface = NULL;
aa_policy_cache *policy_cache = NULL;
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
int retval;
int i;
int optind;
/* name of executable, for error reporting and usage display */
progname = argv[0];
init_base_dir();
process_config_file("/etc/apparmor/parser.conf");
optind = process_args(argc, argv);
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
setup_parallel_compile();
setlocale(LC_MESSAGES, "");
bindtextdomain(PACKAGE, LOCALEDIR);
textdomain(PACKAGE);
/* Check to see if we have superuser rights, if we're not
* debugging */
if (!(UNPRIVILEGED_OPS) && ((retval = have_enough_privilege()))) {
return retval;
}
if (!binary_input) parse_default_paths();
setup_flags();
if (!(UNPRIVILEGED_OPS) &&
aa_kernel_interface_new(&kernel_interface, features, apparmorfs) == -1) {
PERROR(_("Warning: unable to find a suitable fs in %s, is it "
"mounted?\nUse --subdomainfs to override.\n"),
MOUNTED_FS);
return 1;
}
if ((!skip_cache && (write_cache || !skip_read_cache)) ||
force_clear_cache) {
uint16_t max_caches = write_cache && cond_clear_cache ? 1 : 0;
if (!cacheloc && asprintf(&cacheloc, "%s/cache", basedir) == -1) {
PERROR(_("Memory allocation error."));
return 1;
}
if (force_clear_cache) {
if (aa_policy_cache_remove(AT_FDCWD, cacheloc)) {
PERROR(_("Failed to clear cache files (%s): %s\n"),
cacheloc, strerror(errno));
return 1;
}
return 0;
}
if (create_cache_dir)
pwarn(_("The --create-cache-dir option is deprecated. Please use --write-cache.\n"));
retval = aa_policy_cache_new(&policy_cache, features,
AT_FDCWD, cacheloc, max_caches);
if (retval) {
if (errno != ENOENT && errno != EEXIST) {
PERROR(_("Failed setting up policy cache (%s): %s\n"),
cacheloc, strerror(errno));
return 1;
}
if (show_cache) {
if (max_caches > 0)
PERROR("Cache write disabled: Cannot create cache '%s': %m\n",
cacheloc);
else
PERROR("Cache read/write disabled: Policy cache is invalid\n");
}
write_cache = 0;
skip_read_cache = 1;
}
}
retval = last_error = 0;
for (i = optind; i <= argc; i++) {
struct stat stat_file;
if (i < argc && !(profilename = strdup(argv[i]))) {
perror("strdup");
last_error = ENOMEM;
if (abort_on_error)
break;
continue;
}
/* skip stdin if we've seen other command line arguments */
if (i == argc && optind != argc)
continue;
if (profilename && stat(profilename, &stat_file) == -1) {
PERROR("File %s not found, skipping...\n", profilename);
continue;
}
if (profilename && S_ISDIR(stat_file.st_mode)) {
int (*cb)(int dirfd, const char *name, struct stat *st,
void *data);
struct dir_cb_data cb_data;
memset(&cb_data, 0, sizeof(struct dir_cb_data));
cb_data.dirname = profilename;
cb_data.cachedir = cacheloc;
cb_data.kernel_interface = kernel_interface;
cb = binary_input ? binary_dir_cb : profile_dir_cb;
if ((retval = dirat_for_each(AT_FDCWD, profilename,
&cb_data, cb))) {
PDEBUG("Failed loading profiles from %s\n",
profilename);
}
} else if (binary_input) {
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
work_spawn(process_binary(option, kernel_interface,
profilename),
handle_work_result);
} else {
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
work_spawn(process_profile(option, kernel_interface,
profilename, cacheloc),
handle_work_result);
}
if (profilename) free(profilename);
profilename = NULL;
}
parser: add basic support for parallel compiles and loads This adds a basic support for parallel compiles. It uses a fork()/wait model due to the parsers current dependence on global variables and structures. It has been setup in a similar manner to how cilk handles multithreading to make it easy to port to a managed thread model once the parser removes the dependence on global compute structures in the backend. This patch adds two new command line flags -j <n> or --jobs <n> which follows the make syntax of specifying parallel jobs currently defaults to -jauto -j8 or --jobs=8 allows for 8 parallel jobs -jauto or --jobs=auto sets the jobs to the # of cpus -jx4 or --jobs=x4 sets the jobs to # of cpus * 4 -jx1 is equivalent to -jauto Note: unlike make -j must be accompanied by an option --max-jobs=<n> allows setting hard cap on the number of jobs that can be specified by --jobs. It defaults to the number of processors in the system * 8. It supports the "auto" and "max" keywords, and using x<n> for a multiple of the available cpus. additionally the -d flag has been modified to take an optional parameter and --debug=jobs will output debug information for the job control logic. In light testing on one machine the job control logic provides a nice performance boost. On an x86 test machine with 60 profiles in the /etc/apparmor.d/ directory, for the command time apparmor_parser -QT /etc/apparmor.d/ old (equiv of -j1): real 0m10.968s user 0m10.888s sys 0m0.088s ubuntu parallel load using xargs: real 0m8.003s user 0m21.680s sys 0m0.216s -j: real 0m6.547s user 0m17.900s sys 0m0.132s Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Seth Arnold <seth.arnold@canonical.com>
2016-01-13 17:10:57 -08:00
work_sync(handle_work_result);
if (ofile)
fclose(ofile);
aa_policy_cache_unref(policy_cache);
return last_error;
}