Initial set of changes working to make fish robust against running out of file descriptors

This commit is contained in:
ridiculousfish 2013-01-30 02:22:38 -08:00
parent ea8c6bc15e
commit 1879dc4b59
10 changed files with 181 additions and 83 deletions

View File

@ -65,7 +65,7 @@ static int my_env_set(const wchar_t *key, const wcstring_list_t &val, int scope)
if (is_path_variable(key))
{
/* Fix for https://github.com/fish-shell/fish-shell/issues/199 . Return success if any path setting succeeds. */
bool any_success = false, any_error = false;
bool any_success = false;
for (i=0; i< val.size() ; i++)
{
@ -93,7 +93,6 @@ static int my_env_set(const wchar_t *key, const wcstring_list_t &val, int scope)
}
else
{
any_error = true;
const wchar_t *colon;
append_format(stderr_buffer, _(BUILTIN_SET_PATH_ERROR), L"set", dir, key);
colon = wcschr(dir, L':');

View File

@ -80,7 +80,7 @@ static int is_dead()
static int try_get_socket_once(void)
{
int s, len;
int s;
wchar_t *wdir;
wchar_t *wuname;
@ -128,7 +128,6 @@ static int try_get_socket_once(void)
struct sockaddr_un local = {};
local.sun_family = AF_UNIX;
strncpy(local.sun_path, name.c_str(), (sizeof local.sun_path) - 1);
len = sizeof(local);
if (connect(s, (struct sockaddr *)&local, sizeof local) == -1)
{

128
exec.cpp
View File

@ -124,13 +124,14 @@ void exec_close(int fd)
int exec_pipe(int fd[2])
{
int res;
ASSERT_IS_MAIN_THREAD();
int res;
while ((res=pipe(fd)))
{
if (errno != EINTR)
{
wperror(L"pipe");
// caller will call wperror
return res;
}
}
@ -148,6 +149,17 @@ int exec_pipe(int fd[2])
return res;
}
void print_open_fds(void)
{
for (size_t i=0; i < open_fds.size(); ++i)
{
if (open_fds.at(i))
{
fprintf(stderr, "fd %lu\n", i);
}
}
}
/**
Check if the specified fd is used as a part of a pipeline in the
specidied set of IO redirections.
@ -526,9 +538,7 @@ static bool can_use_posix_spawn_for_job(const job_t *job, const process_t *proce
void exec(parser_t &parser, job_t *j)
{
process_t *p;
pid_t pid = 0;
int mypipe[2];
sigset_t chldset;
shared_ptr<io_buffer_t> io_buffer;
@ -559,10 +569,10 @@ void exec(parser_t &parser, job_t *j)
j->io.insert(j->io.begin(), parser.block_io.begin(), parser.block_io.end());
}
const io_buffer_t *input_redirect = 0;
const io_buffer_t *input_redirect = NULL;
for (size_t idx = 0; idx < j->io.size(); idx++)
{
shared_ptr<io_data_t> &io = j->io.at(idx);
const shared_ptr<io_data_t> &io = j->io.at(idx);
if ((io->io_mode == IO_BUFFER))
{
@ -612,11 +622,12 @@ void exec(parser_t &parser, job_t *j)
}
// This is a pipe that the "current" process in our loop below reads from
// Only pipe_read->pipe_fd[0] is used
shared_ptr<io_pipe_t> pipe_read(new io_pipe_t(0, true));
pipe_read->pipe_fd[0] = pipe_read->pipe_fd[1] = -1;
// This is the pipe that the "current" process in our loop below writes to
shared_ptr<io_pipe_t> pipe_write(new io_pipe_t(1, false));
pipe_write->pipe_fd[0] = pipe_write->pipe_fd[1] = -1;
j->io.push_back(pipe_write);
@ -634,7 +645,7 @@ void exec(parser_t &parser, job_t *j)
if (job_get_flag(j, JOB_CONTROL))
{
for (p=j->first_process; p; p = p->next)
for (const process_t *p = j->first_process; p; p = p->next)
{
if (p->type != EXTERNAL)
{
@ -677,6 +688,9 @@ void exec(parser_t &parser, job_t *j)
}
}
/* Make a set of file descriptors that we will need to close */
std::set<int> fds_to_close;
/*
This loop loops over every process_t in the job, starting it as
appropriate. This turns out to be rather complex, since a
@ -685,10 +699,10 @@ void exec(parser_t &parser, job_t *j)
The loop also has to handle pipelining between the jobs.
*/
for (p=j->first_process; p; p = p->next)
for (process_t *p=j->first_process; p; p = p->next)
{
const bool p_wants_pipe = (p->next != NULL);
mypipe[1]=-1;
int mypipe[2] = {-1, -1};
pipe_write->fd = p->pipe_write_fd;
pipe_read->fd = p->pipe_read_fd;
@ -714,6 +728,7 @@ void exec(parser_t &parser, job_t *j)
if (p == j->first_process->next)
{
/* We are the first process that could possibly read from a pipe (aka the second process), so add the pipe read redireciton */
j->io.push_back(pipe_read);
}
@ -726,9 +741,14 @@ void exec(parser_t &parser, job_t *j)
debug(1, PIPE_ERROR);
wperror(L"pipe");
exec_error = true;
job_mark_process_as_failed(j, p);
break;
}
fds_to_close.insert(mypipe[0]);
fds_to_close.insert(mypipe[1]);
// This tells the redirection about the fds, but the redirection does not close them
memcpy(pipe_write->pipe_fd, mypipe, sizeof(int)*2);
}
else
@ -746,7 +766,6 @@ void exec(parser_t &parser, job_t *j)
{
case INTERNAL_FUNCTION:
{
wchar_t * def=0;
int shadows;
@ -757,20 +776,15 @@ void exec(parser_t &parser, job_t *j)
*/
signal_unblock();
wcstring orig_def;
function_get_definition(p->argv0(), &orig_def);
// function_get_named_arguments may trigger autoload, which deallocates the orig_def.
// We should make function_get_definition return a wcstring (but how to handle NULL...)
if (! orig_def.empty())
def = wcsdup(orig_def.c_str());
wcstring def;
bool function_exists = function_get_definition(p->argv0(), &def);
wcstring_list_t named_arguments = function_get_named_arguments(p->argv0());
shadows = function_get_shadows(p->argv0());
signal_block();
if (def == NULL)
if (! function_exists)
{
debug(0, _(L"Unknown function '%ls'"), p->argv0());
break;
@ -778,7 +792,6 @@ void exec(parser_t &parser, job_t *j)
function_block_t *newv = new function_block_t(p, p->argv0(), shadows);
parser.push_block(newv);
/*
set_argv might trigger an event
handler, hence we need to unblock
@ -792,15 +805,26 @@ void exec(parser_t &parser, job_t *j)
if (p->next)
{
// Be careful to handle failure, e.g. too many open fds
io_buffer.reset(io_buffer_t::create(0));
if (io_buffer.get() == NULL)
{
exec_error = true;
job_mark_process_as_failed(j, p);
}
else
{
j->io.push_back(io_buffer);
}
}
internal_exec_helper(parser, def, TOP, j->io);
if (! exec_error)
{
internal_exec_helper(parser, def.c_str(), TOP, j->io);
}
parser.allow_function();
parser.pop_block();
free(def);
break;
}
@ -810,10 +834,21 @@ void exec(parser_t &parser, job_t *j)
if (p->next)
{
io_buffer.reset(io_buffer_t::create(0));
if (io_buffer.get() == NULL)
{
exec_error = true;
job_mark_process_as_failed(j, p);
}
else
{
j->io.push_back(io_buffer);
}
}
if (! exec_error)
{
internal_exec_helper(parser, p->argv0(), TOP, j->io);
}
break;
}
@ -821,7 +856,6 @@ void exec(parser_t &parser, job_t *j)
case INTERNAL_BUILTIN:
{
int builtin_stdin=0;
int fg;
int close_stdin=0;
/*
@ -933,7 +967,7 @@ void exec(parser_t &parser, job_t *j)
builtin_out_redirect = has_fd(j->io, 1);
builtin_err_redirect = has_fd(j->io, 2);
fg = job_get_flag(j, JOB_FOREGROUND);
const int fg = job_get_flag(j, JOB_FOREGROUND);
job_set_flag(j, JOB_FOREGROUND, 0);
signal_unblock();
@ -1300,6 +1334,11 @@ void exec(parser_t &parser, job_t *j)
safe_launch_process _never_ returns...
*/
}
else if (pid < 0)
{
job_mark_process_as_failed(j, p);
exec_error = true;
}
}
@ -1325,24 +1364,29 @@ void exec(parser_t &parser, job_t *j)
previous process_t
*/
if (pipe_read->pipe_fd[0] >= 0)
{
exec_close(pipe_read->pipe_fd[0]);
fds_to_close.erase(pipe_read->pipe_fd[0]);
}
/*
Set up the pipe the next process uses to read from the
current process_t
*/
if (p_wants_pipe)
{
/* The next process will read from this pipe */
assert(p->next != NULL);
pipe_read->pipe_fd[0] = mypipe[0];
/*
If there is a next process in the pipeline, close the
output end of the current pipe (the surrent child
output end of the current pipe (the current child
subprocess already has a copy of the pipe - this makes sure
we don't leak file descriptors either in the shell or in
the children).
*/
if (p->next)
{
exec_close(mypipe[1]);
fds_to_close.erase(mypipe[1]);
}
}
@ -1355,6 +1399,11 @@ void exec(parser_t &parser, job_t *j)
kill(keepalive.pid, SIGKILL);
}
for (std::set<int>::iterator foo = fds_to_close.begin(); foo != fds_to_close.end(); ++foo)
{
fprintf(stderr, "-Malingering %d\n", *foo);
}
signal_unblock();
debug(3, L"Job is constructed");
@ -1368,9 +1417,14 @@ void exec(parser_t &parser, job_t *j)
proc_last_bg_pid = j->pgid;
}
if (!exec_error)
if (! exec_error)
{
job_continue(j, 0);
job_continue(j, false);
}
else
{
/* Mark the errored job as not in the foreground. I can't fully justify whether this is the right place, but it prevents sanity_lose from complaining. */
job_set_flag(j, JOB_FOREGROUND, 0);
}
}
@ -1401,10 +1455,17 @@ static int exec_subshell_internal(const wcstring &cmd, wcstring_list_t *lst)
is_subshell=1;
const shared_ptr<io_buffer_t> io_buffer(io_buffer_t::create(0));
prev_status = proc_get_last_status();
const shared_ptr<io_buffer_t> io_buffer(io_buffer_t::create(0));
// IO buffer creation may fail (e.g. if we have too many open files to make a pipe), so this may be null
if (io_buffer.get() == NULL)
{
status = -1;
}
else
{
parser_t &parser = parser_t::principal_parser();
if (parser.eval(cmd, io_chain_t(io_buffer), SUBST))
{
@ -1416,12 +1477,13 @@ static int exec_subshell_internal(const wcstring &cmd, wcstring_list_t *lst)
}
io_buffer->read();
}
proc_set_last_status(prev_status);
is_subshell = prev_subshell;
if (lst != NULL)
if (lst != NULL && io_buffer.get() != NULL)
{
const char *begin = io_buffer->out_buffer_ptr();
const char *end = begin + io_buffer->out_buffer_size();

17
io.cpp
View File

@ -156,6 +156,10 @@ io_buffer_t *io_buffer_t::create(bool is_input)
delete buffer_redirect;
buffer_redirect = NULL;
}
else
{
//fprintf(stderr, "Created pipes {%d, %d} for %p\n", buffer_redirect->pipe_fd[0], buffer_redirect->pipe_fd[1], buffer_redirect);
}
return buffer_redirect;
}
@ -163,16 +167,20 @@ io_buffer_t *io_buffer_t::create(bool is_input)
io_buffer_t::~io_buffer_t()
{
//fprintf(stderr, "Deallocating pipes {%d, %d} for %p\n", this->pipe_fd[0], this->pipe_fd[1], this);
/**
If this is an input buffer, then io_read_buffer will not have
been called, and we need to close the output fd as well.
*/
if (is_input)
if (is_input && pipe_fd[1] >= 0)
{
exec_close(pipe_fd[1]);
}
if (pipe_fd[0] >= 0)
{
exec_close(pipe_fd[0]);
}
/*
Dont free fd for writing. This should already be free'd before
@ -193,6 +201,13 @@ void io_chain_t::remove(const shared_ptr<const io_data_t> &element)
}
}
void io_chain_t::push_back(const shared_ptr<io_data_t> &element)
{
// Ensure we never push back NULL
assert(element.get() != NULL);
std::vector<shared_ptr<io_data_t> >:: push_back(element);
}
void io_remove(io_chain_t &list, const shared_ptr<const io_data_t> &element)
{
list.remove(element);

7
io.h
View File

@ -95,22 +95,22 @@ class io_pipe_t : public io_data_t
protected:
io_pipe_t(io_mode_t m, int f, bool i):
io_data_t(m, f),
pipe_fd(),
is_input(i)
{
pipe_fd[0] = pipe_fd[1] = -1;
}
public:
int pipe_fd[2];
bool is_input;
const bool is_input;
virtual void print() const;
io_pipe_t(int f, bool i):
io_data_t(IO_PIPE, f),
pipe_fd(),
is_input(i)
{
pipe_fd[0] = pipe_fd[1] = -1;
}
};
@ -178,6 +178,7 @@ public:
io_chain_t(const shared_ptr<io_data_t> &);
void remove(const shared_ptr<const io_data_t> &element);
void push_back(const shared_ptr<io_data_t> &element);
shared_ptr<const io_data_t> get_io_for_fd(int fd) const;
shared_ptr<io_data_t> get_io_for_fd(int fd);

View File

@ -1202,11 +1202,14 @@ bool parser_t::job_remove(job_t *j)
void parser_t::job_promote(job_t *job)
{
signal_block();
job_list_t::iterator loc = std::find(my_job_list.begin(), my_job_list.end(), job);
assert(loc != my_job_list.end());
/* Move the job to the beginning */
my_job_list.splice(my_job_list.begin(), my_job_list, loc);
signal_unblock();
}
job_t *parser_t::job_get(job_id_t id)
@ -3755,7 +3758,7 @@ event_block_t::event_block_t(const event_t &evt) :
{
}
function_block_t::function_block_t(process_t *p, const wcstring &n, bool shadows) :
function_block_t::function_block_t(const process_t *p, const wcstring &n, bool shadows) :
block_t(shadows ? FUNCTION_CALL : FUNCTION_CALL_NO_SHADOW),
process(p),
name(n)

View File

@ -164,9 +164,9 @@ struct event_block_t : public block_t
struct function_block_t : public block_t
{
process_t *process;
const process_t *process;
wcstring name;
function_block_t(process_t *p, const wcstring &n, bool shadows);
function_block_t(const process_t *p, const wcstring &n, bool shadows);
};
struct source_block_t : public block_t

View File

@ -117,6 +117,15 @@ job_iterator_t::job_iterator_t() : job_list(&parser_t::principal_parser().job_li
this->reset();
}
void print_jobs(void)
{
job_iterator_t jobs;
job_t *j;
while ((j = jobs.next()))
{
printf("%p -> %ls -> (foreground %d, complete %d, stopped %d, constructed %d)\n", j, j->command_wcstr(), job_get_flag(j, JOB_FOREGROUND), job_is_completed(j), job_is_stopped(j), job_get_flag(j, JOB_CONSTRUCTED));
}
}
int is_interactive_session=0;
int is_subshell=0;
@ -305,17 +314,21 @@ int job_is_completed(const job_t *j)
}
void job_set_flag(job_t *j, int flag, int set)
void job_set_flag(job_t *j, unsigned int flag, int set)
{
if (set)
{
j->flags |= flag;
}
else
j->flags = j->flags & ((unsigned int)(-1) ^ flag);
{
j->flags &= ~flag;
}
}
int job_get_flag(const job_t *j, int flag)
int job_get_flag(const job_t *j, unsigned int flag)
{
return j->flags&flag?1:0;
return !! (j->flags & flag);
}
int job_signal(job_t *j, int signal)
@ -396,7 +409,10 @@ static void mark_process_status(const job_t *j,
void job_mark_process_as_failed(const job_t *job, process_t *p)
{
/* The given process failed to even lift off (e.g. posix_spawn failed) and so doesn't have a valid pid. Mark it as dead. */
p->completed = 1;
for (process_t *cursor = p; p != NULL; p = p->next)
{
cursor->completed = 1;
}
}
/**
@ -630,7 +646,6 @@ int job_reap(bool interactive)
{
job_t *j = jnext;
jnext = jobs.next();
process_t *p;
/*
If we are reaping only jobs who do not need status messages
@ -642,7 +657,7 @@ int job_reap(bool interactive)
continue;
}
for (p=j->first_process; p; p=p->next)
for (process_t *p = j->first_process; p; p=p->next)
{
int s;
if (!p->completed)
@ -902,7 +917,7 @@ static int select_try(job_t *j)
*/
static void read_try(job_t *j)
{
io_buffer_t *buff=NULL;
io_buffer_t *buff = NULL;
/*
Find the last buffer, which is the one we want to read from
@ -1030,7 +1045,7 @@ static int terminal_return_from_job(job_t *j)
return 1;
}
void job_continue(job_t *j, int cont)
void job_continue(job_t *j, bool cont)
{
/*
Put job first in the job list

8
proc.h
View File

@ -141,9 +141,9 @@ private:
public:
process_t();
~process_t();
/**
Type of process. Can be one of \c EXTERNAL, \c
INTERNAL_BUILTIN, \c INTERNAL_FUNCTION, \c INTERNAL_BLOCK,
@ -453,12 +453,12 @@ extern int no_exec;
/**
Add the specified flag to the bitset of flags for the specified job
*/
void job_set_flag(job_t *j, int flag, int set);
void job_set_flag(job_t *j, unsigned int flag, int set);
/**
Returns one if the specified flag is set in the specified job, 0 otherwise.
*/
int job_get_flag(const job_t *j, int flag);
int job_get_flag(const job_t *j, unsigned int flag);
/**
Sets the status of the last process to exit
@ -515,7 +515,7 @@ int job_is_completed(const job_t *j);
\param j The job
\param cont Whether the function should wait for the job to complete before returning
*/
void job_continue(job_t *j, int cont);
void job_continue(job_t *j, bool cont);
/**
Notify the user about stopped or terminated jobs. Delete terminated

View File

@ -1049,6 +1049,13 @@ static void run_pager(const wcstring &prefix, int is_quoted, const std::vector<c
wcstring prefix_esc;
char *foo;
shared_ptr<io_buffer_t> in(io_buffer_t::create(true));
shared_ptr<io_buffer_t> out(io_buffer_t::create(false));
// The above may fail e.g. if we have too many open fds
if (in.get() == NULL || out.get() == NULL)
return;
wchar_t *escaped_separator;
int has_case_sensitive=0;
@ -1066,7 +1073,6 @@ static void run_pager(const wcstring &prefix, int is_quoted, const std::vector<c
is_quoted?L"-q":L"",
prefix_esc.c_str());
shared_ptr<io_buffer_t> in(io_buffer_t::create(true));
in->fd = 3;
escaped_separator = escape(COMPLETE_SEP_STR, 1);
@ -1133,11 +1139,9 @@ static void run_pager(const wcstring &prefix, int is_quoted, const std::vector<c
in->out_buffer_append(foo, strlen(foo));
free(foo);
term_donate();
shared_ptr<io_buffer_t> out(io_buffer_t::create(false));
out->fd = 4;
term_donate();
parser_t &parser = parser_t::principal_parser();
io_chain_t io_chain;
io_chain.push_back(out);