git/pager.c

256 lines
5.7 KiB
C
Raw Normal View History

#include "cache.h"
#include "run-command.h"
#include "sigchain.h"
#ifndef DEFAULT_PAGER
#define DEFAULT_PAGER "less"
#endif
static struct child_process pager_process = CHILD_PROCESS_INIT;
static const char *pager_program;
pager: don't use unsafe functions in signal handlers Since the commit a3da8821208d (pager: do wait_for_pager on signal death), we call wait_for_pager() in the pager's signal handler. The recent bug report revealed that this causes a deadlock in glibc at aborting "git log" [*1*]. When this happens, git process is left unterminated, and it can't be killed by SIGTERM but only by SIGKILL. The problem is that wait_for_pager() function does more than waiting for pager process's termination, but it does cleanups and printing errors. Unfortunately, the functions that may be used in a signal handler are very limited [*2*]. Particularly, malloc(), free() and the variants can't be used in a signal handler because they take a mutex internally in glibc. This was the cause of the deadlock above. Other than the direct calls of malloc/free, many functions calling malloc/free can't be used. strerror() is such one, either. Also the usage of fflush() and printf() in a signal handler is bad, although it seems working so far. In a safer side, we should avoid them, too. This patch tries to reduce the calls of such functions in signal handlers. wait_for_signal() takes a flag and avoids the unsafe calls. Also, finish_command_in_signal() is introduced for the same reason. There the free() calls are removed, and only waits for the children without whining at errors. [*1*] https://bugzilla.opensuse.org/show_bug.cgi?id=942297 [*2*] http://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html#tag_15_04_03 Signed-off-by: Takashi Iwai <tiwai@suse.de> Reviewed-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-04 17:35:57 +08:00
static void wait_for_pager(int in_signal)
{
pager: don't use unsafe functions in signal handlers Since the commit a3da8821208d (pager: do wait_for_pager on signal death), we call wait_for_pager() in the pager's signal handler. The recent bug report revealed that this causes a deadlock in glibc at aborting "git log" [*1*]. When this happens, git process is left unterminated, and it can't be killed by SIGTERM but only by SIGKILL. The problem is that wait_for_pager() function does more than waiting for pager process's termination, but it does cleanups and printing errors. Unfortunately, the functions that may be used in a signal handler are very limited [*2*]. Particularly, malloc(), free() and the variants can't be used in a signal handler because they take a mutex internally in glibc. This was the cause of the deadlock above. Other than the direct calls of malloc/free, many functions calling malloc/free can't be used. strerror() is such one, either. Also the usage of fflush() and printf() in a signal handler is bad, although it seems working so far. In a safer side, we should avoid them, too. This patch tries to reduce the calls of such functions in signal handlers. wait_for_signal() takes a flag and avoids the unsafe calls. Also, finish_command_in_signal() is introduced for the same reason. There the free() calls are removed, and only waits for the children without whining at errors. [*1*] https://bugzilla.opensuse.org/show_bug.cgi?id=942297 [*2*] http://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html#tag_15_04_03 Signed-off-by: Takashi Iwai <tiwai@suse.de> Reviewed-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-04 17:35:57 +08:00
if (!in_signal) {
fflush(stdout);
fflush(stderr);
}
/* signal EOF to pager */
close(1);
close(2);
pager: don't use unsafe functions in signal handlers Since the commit a3da8821208d (pager: do wait_for_pager on signal death), we call wait_for_pager() in the pager's signal handler. The recent bug report revealed that this causes a deadlock in glibc at aborting "git log" [*1*]. When this happens, git process is left unterminated, and it can't be killed by SIGTERM but only by SIGKILL. The problem is that wait_for_pager() function does more than waiting for pager process's termination, but it does cleanups and printing errors. Unfortunately, the functions that may be used in a signal handler are very limited [*2*]. Particularly, malloc(), free() and the variants can't be used in a signal handler because they take a mutex internally in glibc. This was the cause of the deadlock above. Other than the direct calls of malloc/free, many functions calling malloc/free can't be used. strerror() is such one, either. Also the usage of fflush() and printf() in a signal handler is bad, although it seems working so far. In a safer side, we should avoid them, too. This patch tries to reduce the calls of such functions in signal handlers. wait_for_signal() takes a flag and avoids the unsafe calls. Also, finish_command_in_signal() is introduced for the same reason. There the free() calls are removed, and only waits for the children without whining at errors. [*1*] https://bugzilla.opensuse.org/show_bug.cgi?id=942297 [*2*] http://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html#tag_15_04_03 Signed-off-by: Takashi Iwai <tiwai@suse.de> Reviewed-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-04 17:35:57 +08:00
if (in_signal)
finish_command_in_signal(&pager_process);
else
finish_command(&pager_process);
}
static void wait_for_pager_atexit(void)
{
wait_for_pager(0);
}
static void wait_for_pager_signal(int signo)
{
pager: don't use unsafe functions in signal handlers Since the commit a3da8821208d (pager: do wait_for_pager on signal death), we call wait_for_pager() in the pager's signal handler. The recent bug report revealed that this causes a deadlock in glibc at aborting "git log" [*1*]. When this happens, git process is left unterminated, and it can't be killed by SIGTERM but only by SIGKILL. The problem is that wait_for_pager() function does more than waiting for pager process's termination, but it does cleanups and printing errors. Unfortunately, the functions that may be used in a signal handler are very limited [*2*]. Particularly, malloc(), free() and the variants can't be used in a signal handler because they take a mutex internally in glibc. This was the cause of the deadlock above. Other than the direct calls of malloc/free, many functions calling malloc/free can't be used. strerror() is such one, either. Also the usage of fflush() and printf() in a signal handler is bad, although it seems working so far. In a safer side, we should avoid them, too. This patch tries to reduce the calls of such functions in signal handlers. wait_for_signal() takes a flag and avoids the unsafe calls. Also, finish_command_in_signal() is introduced for the same reason. There the free() calls are removed, and only waits for the children without whining at errors. [*1*] https://bugzilla.opensuse.org/show_bug.cgi?id=942297 [*2*] http://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html#tag_15_04_03 Signed-off-by: Takashi Iwai <tiwai@suse.de> Reviewed-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-04 17:35:57 +08:00
wait_for_pager(1);
sigchain_pop(signo);
raise(signo);
}
static int core_pager_config(const char *var, const char *value, void *data)
{
if (!strcmp(var, "core.pager"))
return git_config_string(&pager_program, var, value);
return 0;
}
static void read_early_config(config_fn_t cb, void *data)
{
git_config_with_options(cb, data, NULL, 1);
/*
* Note that this is a really dirty hack that does the wrong thing in
* many cases. The crux of the problem is that we cannot run
* setup_git_directory() early on in git's setup, so we have no idea if
* we are in a repository or not, and therefore are not sure whether
* and how to read repository-local config.
*
* So if we _aren't_ in a repository (or we are but we would reject its
* core.repositoryformatversion), we'll read whatever is in .git/config
* blindly. Similarly, if we _are_ in a repository, but not at the
* root, we'll fail to find .git/config (because it's really
* ../.git/config, etc). See t7006 for a complete set of failures.
*
* However, we have historically provided this hack because it does
* work some of the time (namely when you are at the top-level of a
* valid repository), and would rarely make things worse (i.e., you do
* not generally have a .git/config file sitting around).
*/
if (!startup_info->have_repository) {
struct git_config_source repo_config;
memset(&repo_config, 0, sizeof(repo_config));
repo_config.file = ".git/config";
git_config_with_options(cb, data, &repo_config, 1);
}
}
const char *git_pager(int stdout_is_tty)
{
const char *pager;
if (!stdout_is_tty)
return NULL;
pager = getenv("GIT_PAGER");
if (!pager) {
if (!pager_program)
read_early_config(core_pager_config, NULL);
pager = pager_program;
}
if (!pager)
pager = getenv("PAGER");
if (!pager)
pager = DEFAULT_PAGER;
if (!*pager || !strcmp(pager, "cat"))
pager = NULL;
return pager;
}
static void setup_pager_env(struct argv_array *env)
{
const char **argv;
int i;
char *pager_env = xstrdup(PAGER_ENV);
int n = split_cmdline(pager_env, &argv);
if (n < 0)
die("malformed build-time PAGER_ENV: %s",
split_cmdline_strerror(n));
for (i = 0; i < n; i++) {
char *cp = strchr(argv[i], '=');
if (!cp)
die("malformed build-time PAGER_ENV");
*cp = '\0';
if (!getenv(argv[i])) {
*cp = '=';
argv_array_push(env, argv[i]);
}
}
free(pager_env);
free(argv);
}
void prepare_pager_args(struct child_process *pager_process, const char *pager)
{
argv_array_push(&pager_process->args, pager);
pager_process->use_shell = 1;
setup_pager_env(&pager_process->env_array);
}
void setup_pager(void)
{
const char *pager = git_pager(isatty(1));
if (!pager)
return;
/*
* force computing the width of the terminal before we redirect
* the standard output to the pager.
*/
(void) term_columns();
setenv("GIT_PAGER_IN_USE", "true", 1);
/* spawn the pager */
prepare_pager_args(&pager_process, pager);
pager_process.in = -1;
argv_array_push(&pager_process.env_array, "GIT_PAGER_IN_USE");
if (start_command(&pager_process))
return;
/* original process continues, but writes to the pipe */
dup2(pager_process.in, 1);
if (isatty(2))
dup2(pager_process.in, 2);
close(pager_process.in);
/* this makes sure that the parent terminates after the pager */
sigchain_push_common(wait_for_pager_signal);
pager: don't use unsafe functions in signal handlers Since the commit a3da8821208d (pager: do wait_for_pager on signal death), we call wait_for_pager() in the pager's signal handler. The recent bug report revealed that this causes a deadlock in glibc at aborting "git log" [*1*]. When this happens, git process is left unterminated, and it can't be killed by SIGTERM but only by SIGKILL. The problem is that wait_for_pager() function does more than waiting for pager process's termination, but it does cleanups and printing errors. Unfortunately, the functions that may be used in a signal handler are very limited [*2*]. Particularly, malloc(), free() and the variants can't be used in a signal handler because they take a mutex internally in glibc. This was the cause of the deadlock above. Other than the direct calls of malloc/free, many functions calling malloc/free can't be used. strerror() is such one, either. Also the usage of fflush() and printf() in a signal handler is bad, although it seems working so far. In a safer side, we should avoid them, too. This patch tries to reduce the calls of such functions in signal handlers. wait_for_signal() takes a flag and avoids the unsafe calls. Also, finish_command_in_signal() is introduced for the same reason. There the free() calls are removed, and only waits for the children without whining at errors. [*1*] https://bugzilla.opensuse.org/show_bug.cgi?id=942297 [*2*] http://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html#tag_15_04_03 Signed-off-by: Takashi Iwai <tiwai@suse.de> Reviewed-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-04 17:35:57 +08:00
atexit(wait_for_pager_atexit);
}
int pager_in_use(void)
{
const char *env;
env = getenv("GIT_PAGER_IN_USE");
return env ? git_config_bool("GIT_PAGER_IN_USE", env) : 0;
}
/*
* Return cached value (if set) or $COLUMNS environment variable (if
* set and positive) or ioctl(1, TIOCGWINSZ).ws_col (if positive),
* and default to 80 if all else fails.
*/
int term_columns(void)
{
static int term_columns_at_startup;
char *col_string;
int n_cols;
if (term_columns_at_startup)
return term_columns_at_startup;
term_columns_at_startup = 80;
col_string = getenv("COLUMNS");
if (col_string && (n_cols = atoi(col_string)) > 0)
term_columns_at_startup = n_cols;
#ifdef TIOCGWINSZ
else {
struct winsize ws;
if (!ioctl(1, TIOCGWINSZ, &ws) && ws.ws_col)
term_columns_at_startup = ws.ws_col;
}
#endif
return term_columns_at_startup;
}
/*
* How many columns do we need to show this number in decimal?
*/
decimal_width: avoid integer overflow The decimal_width function originally appeared in blame.c as "lineno_width", and was designed for calculating the print-width of small-ish integer values (line numbers in text files). In ec7ff5b, it was made into a reusable function, and in dc801e7, we started using it to align diffstats. Binary files in a diffstat show byte counts rather than line numbers, meaning they can be quite large (e.g., consider adding or removing a 2GB file). decimal_width is not up to the challenge for two reasons: 1. It takes the value as an "int", whereas large files may easily surpass this. The value may be truncated, in which case we will produce an incorrect value. 2. It counts "up" by repeatedly multiplying another integer by 10 until it surpasses the value. This can cause an infinite loop when the value is close to the largest representable integer. For example, consider using a 32-bit signed integer, and a value of 2,140,000,000 (just shy of 2^31-1). We will count up and eventually see that 1,000,000,000 is smaller than our value. The next step would be to multiply by 10 and see that 10,000,000,000 is too large, ending the loop. But we can't represent that value, and we have signed overflow. This is technically undefined behavior, but a common behavior is to lose the high bits, in which case our iterator will certainly be less than the number. So we'll keep multiplying, overflow again, and so on. This patch changes the argument to a uintmax_t (the same type we use to store the diffstat information for binary filese), and counts "down" by repeatedly dividing our value by 10. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-02-05 16:14:19 +08:00
int decimal_width(uintmax_t number)
{
decimal_width: avoid integer overflow The decimal_width function originally appeared in blame.c as "lineno_width", and was designed for calculating the print-width of small-ish integer values (line numbers in text files). In ec7ff5b, it was made into a reusable function, and in dc801e7, we started using it to align diffstats. Binary files in a diffstat show byte counts rather than line numbers, meaning they can be quite large (e.g., consider adding or removing a 2GB file). decimal_width is not up to the challenge for two reasons: 1. It takes the value as an "int", whereas large files may easily surpass this. The value may be truncated, in which case we will produce an incorrect value. 2. It counts "up" by repeatedly multiplying another integer by 10 until it surpasses the value. This can cause an infinite loop when the value is close to the largest representable integer. For example, consider using a 32-bit signed integer, and a value of 2,140,000,000 (just shy of 2^31-1). We will count up and eventually see that 1,000,000,000 is smaller than our value. The next step would be to multiply by 10 and see that 10,000,000,000 is too large, ending the loop. But we can't represent that value, and we have signed overflow. This is technically undefined behavior, but a common behavior is to lose the high bits, in which case our iterator will certainly be less than the number. So we'll keep multiplying, overflow again, and so on. This patch changes the argument to a uintmax_t (the same type we use to store the diffstat information for binary filese), and counts "down" by repeatedly dividing our value by 10. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-02-05 16:14:19 +08:00
int width;
decimal_width: avoid integer overflow The decimal_width function originally appeared in blame.c as "lineno_width", and was designed for calculating the print-width of small-ish integer values (line numbers in text files). In ec7ff5b, it was made into a reusable function, and in dc801e7, we started using it to align diffstats. Binary files in a diffstat show byte counts rather than line numbers, meaning they can be quite large (e.g., consider adding or removing a 2GB file). decimal_width is not up to the challenge for two reasons: 1. It takes the value as an "int", whereas large files may easily surpass this. The value may be truncated, in which case we will produce an incorrect value. 2. It counts "up" by repeatedly multiplying another integer by 10 until it surpasses the value. This can cause an infinite loop when the value is close to the largest representable integer. For example, consider using a 32-bit signed integer, and a value of 2,140,000,000 (just shy of 2^31-1). We will count up and eventually see that 1,000,000,000 is smaller than our value. The next step would be to multiply by 10 and see that 10,000,000,000 is too large, ending the loop. But we can't represent that value, and we have signed overflow. This is technically undefined behavior, but a common behavior is to lose the high bits, in which case our iterator will certainly be less than the number. So we'll keep multiplying, overflow again, and so on. This patch changes the argument to a uintmax_t (the same type we use to store the diffstat information for binary filese), and counts "down" by repeatedly dividing our value by 10. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-02-05 16:14:19 +08:00
for (width = 1; number >= 10; width++)
number /= 10;
return width;
}
struct pager_command_config_data {
const char *cmd;
int want;
char *value;
};
static int pager_command_config(const char *var, const char *value, void *vdata)
{
struct pager_command_config_data *data = vdata;
const char *cmd;
if (skip_prefix(var, "pager.", &cmd) && !strcmp(cmd, data->cmd)) {
int b = git_config_maybe_bool(var, value);
if (b >= 0)
data->want = b;
else {
data->want = 1;
data->value = xstrdup(value);
}
}
return 0;
}
/* returns 0 for "no pager", 1 for "use pager", and -1 for "not specified" */
int check_pager_config(const char *cmd)
{
struct pager_command_config_data data;
data.cmd = cmd;
data.want = -1;
data.value = NULL;
read_early_config(pager_command_config, &data);
if (data.value)
pager_program = data.value;
return data.want;
}