[PATCH 1/6] Modify gcc driver for parallel compilation
Richard Biener
richard.guenther@gmail.com
Mon Aug 24 13:17:41 GMT 2020
On Fri, Aug 21, 2020 at 12:00 AM Giuliano Belinassi
<giuliano.belinassi@usp.br> wrote:
>
> Update the driver for parallel compilation. This process work as
> follows:
>
> When calling gcc, the driver will check if the flag
> "-fparallel-jobs" was provided by the user. If yes, then we will
> check what is the desired output, and if it can be parallelized.
> There are the following cases, which is described:
>
> 1. -S or -E was provided: We can't run in parallel, as the output
> can not be easily merged together into one file.
>
> 2. -c was provided: When cc1* forks into multiple processes, it
> must tell the driver where it stored its generated assembler files.
> Therefore we pass a hidden "-fsplit-outputs=filename" to the compiler,
> and we check if "filename" was created by it. If yes, we open it,
> call assembler for each generated asm file
> (this file must not be empty), and link them together with
> partial linking to a single .o file. This process is done for each
> object file in the argument list.
>
> 3. -c was not provided, and the final product will be an binary: Here
> we proceed exactly as 2., but we avoid doing the partial
> linking, feeding the generated object files directly into the final link.
>
> For that to work, we had to heavily modify how the "execute" function
> works, extracting common code which is used multiple times, and
> also detecting when the command is a call to a compiler or an
> assembler, as can be seen in append_split_outputs.
>
> Finally, we added some tests which reflects all cases found when
> bootstrapping the compiler, so development of further features to the
> driver get faster for now on.
Few comments inline, Joseph may want to comment on the overall
structure as driver maintainer (CCed).
I know I asked for the changes on the branch to be squashed but
the diff below is quite unreadable with the ChangeLog not helping
the overall refactoring much. Is it possible to do some of the
factoring/refactoring without any functionality change to make the
actual diff easier to follow?
Thanks,
Richard.
> gcc/ChangeLog
> 2020-08-20 Giuliano Belinassi <giuliano.belinassi@usp.br>
>
> * common.opt (fsplit-outputs): New flag.
> (fparallel-jobs): New flag.
> * gcc.c (extra_arg_storer): New class.
> (have_S): New variable.
> (struct command): Move from execute.
> (is_compiler): New function.
> (is_assembler): New function.
> (get_number_of_args): New function.
> (get_file_by_lines): New function.
> (identify_asm_file): New function.
> (struct infile): New attribute temp_additional_asm.
> (current_infile): New variable.
> (get_path_to_ld): New function.
> (has_hidden_E): New function.
> (sort_asm_files): New function.
> (append_split_outputs): New function.
> (print_command): New function.
> (print_commands): New function.
> (print_argbuf): New function.
> (handle_verbose): Extracted from execute.
> (append_valgrind): Same as above.
> (async_launch_commands): Same as above.
> (await_commands_to_finish): Same as above.
> (split_commands): Same as above.
> (parse_argbuf): Same as above.
> (execute): Refator.
> (fsplit_arg): New function.
> (alloc_infile): Initialize infiles with 0.
> (process_command): Remember when -S was passed.
> (do_spec_on_infiles): Remember current infile being processed.
> (maybe_run_linker): Replace object files when -o is a executable.
> (finalize): Deinitialize temp_object_files.
>
> gcc/testsuite/ChangeLog:
> 20-08-2020 Giuliano Belinassi <giuliano.belinassi@usp.br>
>
> * driver/driver.exp: New test.
> * driver/a.c: New file.
> * driver/b.c: New file.
> * driver/empty.c: New file.
> * driver/foo.c: New file.
> ---
> gcc/common.opt | 4 +
> gcc/gcc.c | 1219 ++++++++++++++++++++++++-------
> gcc/testsuite/driver/a.c | 6 +
> gcc/testsuite/driver/b.c | 6 +
> gcc/testsuite/driver/driver.exp | 80 ++
> gcc/testsuite/driver/empty.c | 0
> gcc/testsuite/driver/foo.c | 7 +
> 7 files changed, 1049 insertions(+), 273 deletions(-)
> create mode 100644 gcc/testsuite/driver/a.c
> create mode 100644 gcc/testsuite/driver/b.c
> create mode 100644 gcc/testsuite/driver/driver.exp
> create mode 100644 gcc/testsuite/driver/empty.c
> create mode 100644 gcc/testsuite/driver/foo.c
>
> diff --git a/gcc/common.opt b/gcc/common.opt
> index 4b08e91859f..4aa3ad8c95b 100644
> --- a/gcc/common.opt
> +++ b/gcc/common.opt
> @@ -3465,4 +3465,8 @@ fipa-ra
> Common Report Var(flag_ipa_ra) Optimization
> Use caller save register across calls if possible.
>
> +fsplit-outputs=
> +Common Joined Var(split_outputs)
> +-fsplit-outputs=<tempfile> Filename in which current Compilation Unit will be split to.
> +
> ; This comment is to ensure we retain the blank line above.
> diff --git a/gcc/gcc.c b/gcc/gcc.c
> index 10bc9881aed..c276a11ca7a 100644
> --- a/gcc/gcc.c
> +++ b/gcc/gcc.c
> @@ -343,6 +343,74 @@ static struct obstack obstack;
>
> static struct obstack collect_obstack;
>
> +/* This is used to store new argv arrays created dinamically to avoid memory
> + leaks. */
> +
> +class extra_arg_storer
> +{
> + public:
> +
> + /* Initialize the vec with a default size. */
> +
> + extra_arg_storer ()
> + {
> + string_vec.create (8);
> + extra_args.create (64);
> + }
> +
> + /* Create new array of strings of size N. */
> + const char **create_new (size_t n)
> + {
> + const char **ret = XNEWVEC (const char *, n);
> + extra_args.safe_push (ret);
> + return ret;
> + }
> +
> + char *create_string (size_t n)
> + {
> + char *ret = XNEWVEC (char, n);
> + string_vec.safe_push (ret);
> + return ret;
> + }
> +
> + void store (char *str)
> + {
> + string_vec.safe_push (str);
> + }
> +
> + ~extra_arg_storer ()
> + {
> + release_extra_args ();
> + release_string_vec ();
> + }
> +
> +
> + private:
> +
> + /* Release all allocated strings. */
> + void release_extra_args ()
> + {
> + size_t i;
> +
> + for (i = 0; i < extra_args.length (); i++)
> + free (extra_args[i]);
> + extra_args.release ();
> + }
> +
> + void release_string_vec ()
> + {
> + size_t i;
> +
> + for (i = 0; i < string_vec.length (); i++)
> + free (string_vec[i]);
> + string_vec.release ();
> + }
> +
> + /* Data structure to hold all arrays. */
> + vec<const char **> extra_args;
> + vec<char *> string_vec;
> +};
> +
> /* Forward declaration for prototypes. */
> struct path_prefix;
> struct prefix_list;
> @@ -1993,6 +2061,9 @@ static int have_o = 0;
> /* Was the option -E passed. */
> static int have_E = 0;
>
> +/* Was the option -S passed. */
> +static int have_S = 0;
> +
> /* Pointer to output file name passed in with -o. */
> static const char *output_file = 0;
>
> @@ -3056,158 +3127,522 @@ add_sysrooted_hdrs_prefix (struct path_prefix *pprefix, const char *prefix,
> require_machine_suffix, os_multilib);
> }
>
> -
> -/* Execute the command specified by the arguments on the current line of spec.
> - When using pipes, this includes several piped-together commands
> - with `|' between them.
> +struct command
> +{
> + const char *prog; /* program name. */
> + const char **argv; /* vector of args. */
> +};
>
> - Return 0 if successful, -1 if failed. */
> +#define EMPTY_CMD(x) (!((x).prog)) /* Is the provided CMD empty? */
> +
> +/* Check if arg is a call to a compiler. Return false if not, true if yes. */
> +
> +static bool
> +is_compiler (const char *arg)
This and is_assembler should somehow magically fall out of
specs processing, specifically
> +{
> + static const char *const compilers[] = {"cc1", "cc1plus", "f771"};
^^ this is incomplete. Of course I don't know how to auto-infer
these but I think it must be possible from somewhere up the
call-chain?
> + const char* ptr = arg;
> +
> + size_t i;
> +
> + /* Jump to last '/' of string. */
> + while (*arg)
> + if (*arg++ == '/')
> + ptr = arg;
> +
> + /* Look if current character seems valid. */
> + gcc_assert (!(*ptr == '\0' || *ptr == '/'));
> +
> + for (i = 0; i < ARRAY_SIZE (compilers); i++)
> + {
> + if (!strcmp (ptr, compilers[i]))
> + return true;
> + }
> +
> + return false;
> +}
> +
> +/* Check if arg is a call to as. Return false if not, true if yes. */
> +
> +static bool
> +is_assembler (const char *arg)
> +{
> + static const char *const assemblers[] = {"as", "gas"};
> + const char* ptr = arg;
> +
> + size_t i;
> +
> + /* Jump to last '/' of string. */
> + while (*arg)
> + if (*arg++ == '/')
> + ptr = arg;
> +
> + /* Look if current character seems valid. */
> + gcc_assert (!(*ptr == '\0' || *ptr == '/'));
> +
> + for (i = 0; i < ARRAY_SIZE (assemblers); i++)
> + {
> + if (!strcmp (ptr, assemblers[i]))
> + return true;
> + }
> +
> + return false;
> +}
> +
> +/* Get argv[] array length. */
>
> static int
> -execute (void)
> +get_number_of_args (const char *argv[])
> +{
> + int argc;
> +
> + for (argc = 0; argv[argc] != NULL; argc++)
> + ;
> +
> + return argc;
> +}
> +
> +static const char *fsplit_arg (extra_arg_storer *);
> +
> +/* Accumulate each line in lines vec. Return true if file exists, false if
> + not. */
> +
> +static bool
> +get_file_by_lines (extra_arg_storer *storer, vec<char *> *lines, const char *name)
> +{
> + int buf_size = 64, len = 0;
> + char *buf = XNEWVEC (char, buf_size);
> +
> +
> + FILE *file = fopen (name, "r");
> +
> + if (!file)
> + return false;
> +
> + while (1)
> + {
> + if (!fgets (buf + len, buf_size, file))
> + {
> + free (buf); /* Release buffer we created unecessarily. */
> + break;
> + }
> +
> + len = strlen (buf);
> + if (buf[len - 1] == '\n') /* Check if we indeed read the entire line. */
> + {
> + buf[len - 1] = '\0';
> + /* Yes. Insert into the lines vector. */
> + lines->safe_push (buf);
> + len = 0;
> +
> + /* Store the created string for future release. */
> + storer->store (buf);
> + buf = XNEWVEC (char, buf_size);
> + }
> + else
> + {
> + /* No. Increase the buffer size and read again. */
> + buf = XRESIZEVEC (char, buf, buf_size * 2);
> + }
> + }
> +
> + if (lines->length () == 0)
> + internal_error ("Empty file: %s", name);
> +
> + fclose (file);
> + return true;
> +}
> +
> +static void
> +identify_asm_file (int argc, const char *argv[],
> + int *infile_pos, int *outfile_pos)
> {
> int i;
> - int n_commands; /* # of command. */
> - char *string;
> - struct pex_obj *pex;
> - struct command
> - {
> - const char *prog; /* program name. */
> - const char **argv; /* vector of args. */
> - };
> - const char *arg;
>
> - struct command *commands; /* each command buffer with above info. */
> + static const char *asm_extension[] = {"s", "S"};
>
> - gcc_assert (!processing_spec_function);
> + bool infile_found = false;
> + bool outfile_found = false;
>
> - if (wrapper_string)
> + for (i = 0; i < argc; i++)
> {
> - string = find_a_file (&exec_prefixes,
> - argbuf[0], X_OK, false);
> - if (string)
> - argbuf[0] = string;
> - insert_wrapper (wrapper_string);
> + const char *arg = argv[i];
> + const char *ext = argv[i];
> + unsigned j;
> +
> + /* Jump to last '.' of string. */
> + while (*arg)
> + if (*arg++ == '.')
> + ext = arg;
> +
> + if (!infile_found)
> + for (j = 0; j < ARRAY_SIZE (asm_extension); ++j)
> + if (!strcmp (ext, asm_extension[j]))
> + {
> + infile_found = true;
> + *infile_pos = i;
> + break;
> + }
> +
> + if (!outfile_found)
> + if (!strcmp (ext, "-o"))
> + {
> + outfile_found = true;
> + *outfile_pos = i+1;
> + }
> +
> + if (infile_found && outfile_found)
> + return;
> }
>
> - /* Count # of piped commands. */
> - for (n_commands = 1, i = 0; argbuf.iterate (i, &arg); i++)
> - if (strcmp (arg, "|") == 0)
> - n_commands++;
> + gcc_assert (infile_found && outfile_found);
>
> - /* Get storage for each command. */
> - commands = (struct command *) alloca (n_commands * sizeof (struct command));
> +}
>
> - /* Split argbuf into its separate piped processes,
> - and record info about each one.
> - Also search for the programs that are to be run. */
> +/* Language is one of three things:
>
> - argbuf.safe_push (0);
> + 1) The name of a real programming language.
> + 2) NULL, indicating that no one has figured out
> + what it is yet.
> + 3) '*', indicating that the file should be passed
> + to the linker. */
> +struct infile
> +{
> + const char *name;
> + const char *language;
> + const char *temp_additional_asm;
> + struct compiler *incompiler;
> + bool compiled;
> + bool preprocessed;
> +};
>
> - commands[0].prog = argbuf[0]; /* first command. */
> - commands[0].argv = argbuf.address ();
> +/* Also a vector of input files specified. */
>
> - if (!wrapper_string)
> +static struct infile *infiles;
> +static struct infile *current_infile = NULL;
> +
> +int n_infiles;
> +
> +static int n_infiles_alloc;
> +
> +static vec<const char *> temp_object_files;
> +
> +/* Get path to the configured ld. */
> +
> +static const char *
> +get_path_to_ld (void)
> +{
> + const char *ret = find_a_file (&exec_prefixes, LINKER_NAME, X_OK, false);
> + if (!ret)
> + ret = "ld";
> +
> + return ret;
> +}
> +
> +/* Check if a hidden -E was passed as argument to something. */
> +
> +static bool
> +has_hidden_E (int argc, const char *argv[])
> +{
> + int i;
> + for (i = 0; i < argc; ++i)
> + if (!strcmp (argv[i], "-E"))
> + return true;
> +
> + return false;
> +}
> +
> +/* Assembler in the container file are inserted as soon as they are ready.
> + Sort them so that builds are reproducible. */
In principle the list of outputs is pre-determined by the
scheduler compiling the partitions - is there any reason
to write the file with the output names only incrementally
rather than in one (sorted) go?
> +static void
> +sort_asm_files (vec <char *> *_lines)
> +{
> + vec <char *> &lines = *_lines;
> + int i, n = lines.length ();
> + char **temp_buf = XALLOCAVEC (char *, n);
> +
> + for (i = 0; i < n; i++)
> + temp_buf[i] = lines[i];
> +
> + for (i = 0; i < n; i++)
> {
> - string = find_a_file (&exec_prefixes, commands[0].prog, X_OK, false);
> - if (string)
> - commands[0].argv[0] = string;
> + char *no_str = strtok (temp_buf[i], " ");
> + char *name = strtok (NULL, "");
> +
> + int pos = atoi (no_str);
> + lines[pos] = name;
> }
> +}
>
> - for (n_commands = 1, i = 0; argbuf.iterate (i, &arg); i++)
> - if (arg && strcmp (arg, "|") == 0)
> - { /* each command. */
> -#if defined (__MSDOS__) || defined (OS2) || defined (VMS)
> - fatal_error (input_location, "%<-pipe%> not supported");
> -#endif
> - argbuf[i] = 0; /* Termination of command args. */
> - commands[n_commands].prog = argbuf[i + 1];
> - commands[n_commands].argv
> - = &(argbuf.address ())[i + 1];
> - string = find_a_file (&exec_prefixes, commands[n_commands].prog,
> - X_OK, false);
> - if (string)
> - commands[n_commands].argv[0] = string;
> - n_commands++;
> - }
> +/* Append -fsplit-output=<tempfile> to all calls to compilers. Return true
> + if a additional call to LD is required to merge the resulting files. */
>
> - /* If -v, print what we are about to do, and maybe query. */
> +static void
> +append_split_outputs (extra_arg_storer *storer,
> + struct command *additional_ld,
> + struct command **_commands,
> + int *_n_commands)
> +{
> + int i;
>
> - if (verbose_flag)
> + struct command *commands = *_commands;
> + int n_commands = *_n_commands;
> +
> + const char **argv;
> + int argc;
> +
> + if (is_compiler (commands[0].prog))
> + {
> + argc = get_number_of_args (commands[0].argv);
> + argv = storer->create_new (argc + 4);
> +
> + memcpy (argv, commands[0].argv, argc * sizeof (const char *));
> +
> + if (!has_hidden_E (argc, commands[0].argv))
> + {
> + const char *extra_argument = fsplit_arg (storer);
> + argv[argc++] = extra_argument;
> + }
> +
> + if (have_c)
> + {
> + argv[argc++] = "-fPIE";
> + argv[argc++] = "-fPIC";
Uh, I think this has to go away - this must be from some early
problems and no longer necessary?
> + }
> +
> + argv[argc] = NULL;
> +
> + commands[0].argv = argv;
> + }
> +
> + else if (is_assembler (commands[0].prog))
> {
> - /* For help listings, put a blank line between sub-processes. */
> - if (print_help_list)
> - fputc ('\n', stderr);
> + vec<char *> additional_asm_files;
> +
> + struct command orig;
> + const char **orig_argv;
> + int orig_argc;
> + const char *orig_obj_file;
> +
> + int infile_pos = -1;
> + int outfile_pos = -1;
> +
> + static const char *path_to_ld = NULL;
> +
> + if (!current_infile->temp_additional_asm)
> + {
> + /* Return because we did not create a additional-asm file for this
> + input. */
> +
> + return;
> + }
> +
> + additional_asm_files.create (2);
> +
> + if (!get_file_by_lines (storer, &additional_asm_files,
> + current_infile->temp_additional_asm))
> + {
> + additional_asm_files.release ();
> + return; /* File not found. This means that cc1* decided not to
> + parallelize. */
> + }
> +
> + sort_asm_files (&additional_asm_files);
> +
> + if (n_commands != 1)
> + fatal_error (input_location,
> + "Auto parallelism is unsupported when piping commands");
> +
> + if (!path_to_ld)
> + path_to_ld = get_path_to_ld ();
> +
> + /* Get original command. */
> + orig = commands[0];
> + orig_argv = commands[0].argv;
> + orig_argc = get_number_of_args (orig.argv);
> +
> +
> + /* Update commands array to include the extra `as' calls. */
> + *_n_commands = additional_asm_files.length ();
> + n_commands = *_n_commands;
> +
> + gcc_assert (n_commands > 0);
> +
> + identify_asm_file (orig_argc, orig_argv, &infile_pos, &outfile_pos);
> +
> + *_commands = XRESIZEVEC (struct command, *_commands, n_commands);
> + commands = *_commands;
>
> - /* Print each piped command as a separate line. */
> for (i = 0; i < n_commands; i++)
> {
> - const char *const *j;
> + const char **argv = storer->create_new (orig_argc + 1);
> + const char *temp_obj = make_temp_file ("additional-obj.o");
> + record_temp_file (temp_obj, true, true);
> + record_temp_file (additional_asm_files[i], true, true);
> +
> + memcpy (argv, orig_argv, (orig_argc + 1) * sizeof (const char *));
> +
> + orig_obj_file = argv[outfile_pos];
> +
> + argv[infile_pos] = additional_asm_files[i];
> + argv[outfile_pos] = temp_obj;
> +
> + commands[i].prog = orig.prog;
> + commands[i].argv = argv;
> +
> + temp_object_files.safe_push (temp_obj);
> + }
> +
> + if (have_c)
> + {
> + unsigned int num_temp_objs = temp_object_files.length ();
> + const char **argv = storer->create_new (num_temp_objs + 5);
> + unsigned int j;
> +
> + argv[0] = path_to_ld;
> + argv[1] = "-o";
> + argv[2] = orig_obj_file;
> + argv[3] = "-r";
> +
> + for (j = 0; j < num_temp_objs; j++)
> + argv[j + 4] = temp_object_files[j];
> + argv[j + 4] = NULL;
> +
> + additional_ld->prog = path_to_ld;
> + additional_ld->argv = argv;
> +
> + if (!have_o)
> + temp_object_files.truncate (0);
> + }
> +
> + additional_asm_files.release ();
> + }
> +}
> +
> +DEBUG_FUNCTION void
> +print_command (struct command *command)
> +{
> + const char **argv;
> +
> + for (argv = command->argv; *argv != NULL; argv++)
> + fprintf (stdout, " %s", *argv);
> + fputc ('\n', stdout);
> +}
> +
> +DEBUG_FUNCTION void
> +print_commands (int n, struct command *commands)
> +{
> + int i;
> +
> + for (i = 0; i < n; i++)
> + print_command (&commands[i]);
> +}
> +
> +DEBUG_FUNCTION void
> +print_argbuf ()
> +{
> + int i;
> + const char *arg;
> +
> + for (i = 0; argbuf.iterate (i, &arg); i++)
> + fprintf (stdout, "%s ", arg);
> + fputc ('\n', stdout);
> +}
> +
> +
> +/* Print what commands will run. Return 0 if success, anything else on
> + error. */
>
> - if (verbose_only_flag)
> +static int
> +handle_verbose (int n_commands, struct command commands[])
> +{
> + int i;
> +
> + /* For help listings, put a blank line between sub-processes. */
> + if (print_help_list)
> + fputc ('\n', stderr);
> +
> + /* Print each piped command as a separate line. */
> + for (i = 0; i < n_commands; i++)
> + {
> + const char *const *j;
> +
> + if (verbose_only_flag)
> + {
> + for (j = commands[i].argv; *j; j++)
> {
> - for (j = commands[i].argv; *j; j++)
> + const char *p;
> + for (p = *j; *p; ++p)
> + if (!ISALNUM ((unsigned char) *p)
> + && *p != '_' && *p != '/' && *p != '-' && *p != '.')
> + break;
> + if (*p || !*j)
> {
> - const char *p;
> + fprintf (stderr, " \"");
> for (p = *j; *p; ++p)
> - if (!ISALNUM ((unsigned char) *p)
> - && *p != '_' && *p != '/' && *p != '-' && *p != '.')
> - break;
> - if (*p || !*j)
> {
> - fprintf (stderr, " \"");
> - for (p = *j; *p; ++p)
> - {
> - if (*p == '"' || *p == '\\' || *p == '$')
> - fputc ('\\', stderr);
> - fputc (*p, stderr);
> - }
> - fputc ('"', stderr);
> + if (*p == '"' || *p == '\\' || *p == '$')
> + fputc ('\\', stderr);
> + fputc (*p, stderr);
> }
> - /* If it's empty, print "". */
> - else if (!**j)
> - fprintf (stderr, " \"\"");
> - else
> - fprintf (stderr, " %s", *j);
> - }
> - }
> - else
> - for (j = commands[i].argv; *j; j++)
> + fputc ('"', stderr);
> + }
> /* If it's empty, print "". */
> - if (!**j)
> + else if (!**j)
> fprintf (stderr, " \"\"");
> else
> fprintf (stderr, " %s", *j);
> -
> - /* Print a pipe symbol after all but the last command. */
> - if (i + 1 != n_commands)
> - fprintf (stderr, " |");
> - fprintf (stderr, "\n");
> + }
> }
> - fflush (stderr);
> - if (verbose_only_flag != 0)
> - {
> - /* verbose_only_flag should act as if the spec was
> - executed, so increment execution_count before
> - returning. This prevents spurious warnings about
> - unused linker input files, etc. */
> - execution_count++;
> - return 0;
> - }
> + else
> + for (j = commands[i].argv; *j; j++)
> + /* If it's empty, print "". */
> + if (!**j)
> + fprintf (stderr, " \"\"");
> + else
> + fprintf (stderr, " %s", *j);
> +
> + /* Print a pipe symbol after all but the last command. */
> + if (i + 1 != n_commands)
> + fprintf (stderr, " |");
> + fprintf (stderr, "\n");
> + }
> + fflush (stderr);
> + if (verbose_only_flag != 0)
> + {
> + /* verbose_only_flag should act as if the spec was
> + executed, so increment execution_count before
> + returning. This prevents spurious warnings about
> + unused linker input files, etc. */
> + execution_count++;
> + return 1;
> + }
> #ifdef DEBUG
> - fnotice (stderr, "\nGo ahead? (y or n) ");
> - fflush (stderr);
> - i = getchar ();
> - if (i != '\n')
> - while (getchar () != '\n')
> - ;
> -
> - if (i != 'y' && i != 'Y')
> - return 0;
> + fnotice (stderr, "\nGo ahead? (y or n) ");
> + fflush (stderr);
> + i = getchar ();
> + if (i != '\n')
> + while (getchar () != '\n')
> + ;
> +
> + if (i != 'y' && i != 'Y')
> + return 1;
> #endif /* DEBUG */
> - }
> +
> + return 0;
> +}
>
> #ifdef ENABLE_VALGRIND_CHECKING
> +
> +/* Append valgrind to each program. */
> +
> +static void
> +append_valgrind (struct obstack *to_be_released,
> + int n_commands, struct command commands[])
> +{
> + int i;
> +
> /* Run the each command through valgrind. To simplify prepending the
> path to valgrind and the option "-q" (for quiet operation unless
> something triggers), we allocate a separate argv array. */
> @@ -3221,7 +3656,7 @@ execute (void)
> for (argc = 0; commands[i].argv[argc] != NULL; argc++)
> ;
>
> - argv = XALLOCAVEC (const char *, argc + 3);
> + argv = obstack_alloc (to_be_released, (argc + 3) * sizeof (const char *));
>
> argv[0] = VALGRIND_PATH;
> argv[1] = "-q";
> @@ -3232,15 +3667,16 @@ execute (void)
> commands[i].argv = argv;
> commands[i].prog = argv[0];
> }
> +}
> #endif
>
> - /* Run each piped subprocess. */
> +/* Launch a list of commands asynchronously. */
>
> - pex = pex_init (PEX_USE_PIPES | ((report_times || report_times_to_file)
> - ? PEX_RECORD_TIMES : 0),
> - progname, temp_filename);
> - if (pex == NULL)
> - fatal_error (input_location, "%<pex_init%> failed: %m");
> +static void
> +async_launch_commands (struct pex_obj *pex,
> + int n_commands, struct command commands[])
> +{
> + int i;
>
> for (i = 0; i < n_commands; i++)
> {
> @@ -3267,151 +3703,341 @@ execute (void)
> }
>
> execution_count++;
> +}
>
> - /* Wait for all the subprocesses to finish. */
>
> - {
> - int *statuses;
> - struct pex_time *times = NULL;
> - int ret_code = 0;
> +/* Wait for all the subprocesses to finish. Return 0 on success, -1 on
> + failure. */
>
> - statuses = (int *) alloca (n_commands * sizeof (int));
> - if (!pex_get_status (pex, n_commands, statuses))
> - fatal_error (input_location, "failed to get exit status: %m");
> +static int
> +await_commands_to_finish (struct pex_obj *pex,
> + int n_commands, struct command commands[])
> +{
>
> - if (report_times || report_times_to_file)
> - {
> - times = (struct pex_time *) alloca (n_commands * sizeof (struct pex_time));
> - if (!pex_get_times (pex, n_commands, times))
> - fatal_error (input_location, "failed to get process times: %m");
> - }
> + int *statuses;
> + struct pex_time *times = NULL;
> + int ret_code = 0, i;
>
> - pex_free (pex);
> + statuses = (int *) alloca (n_commands * sizeof (int));
> + if (!pex_get_status (pex, n_commands, statuses))
> + fatal_error (input_location, "failed to get exit status: %m");
>
> - for (i = 0; i < n_commands; ++i)
> - {
> - int status = statuses[i];
> + if (report_times || report_times_to_file)
> + {
> + times = (struct pex_time *) alloca (n_commands * sizeof (*times));
> + if (!pex_get_times (pex, n_commands, times))
> + fatal_error (input_location, "failed to get process times: %m");
> + }
>
> - if (WIFSIGNALED (status))
> - switch (WTERMSIG (status))
> - {
> - case SIGINT:
> - case SIGTERM:
> - /* SIGQUIT and SIGKILL are not available on MinGW. */
> + for (i = 0; i < n_commands; ++i)
> + {
> + int status = statuses[i];
> +
> + if (WIFSIGNALED (status))
> + switch (WTERMSIG (status))
> + {
> + case SIGINT:
> + case SIGTERM:
> + /* SIGQUIT and SIGKILL are not available on MinGW. */
> #ifdef SIGQUIT
> - case SIGQUIT:
> + case SIGQUIT:
> #endif
> #ifdef SIGKILL
> - case SIGKILL:
> + case SIGKILL:
> #endif
> - /* The user (or environment) did something to the
> - inferior. Making this an ICE confuses the user into
> - thinking there's a compiler bug. Much more likely is
> - the user or OOM killer nuked it. */
> - fatal_error (input_location,
> - "%s signal terminated program %s",
> - strsignal (WTERMSIG (status)),
> - commands[i].prog);
> - break;
> + /* The user (or environment) did something to the
> + inferior. Making this an ICE confuses the user into
> + thinking there's a compiler bug. Much more likely is
> + the user or OOM killer nuked it. */
> + fatal_error (input_location,
> + "%s signal terminated program %s",
> + strsignal (WTERMSIG (status)),
> + commands[i].prog);
> + break;
>
> #ifdef SIGPIPE
> - case SIGPIPE:
> - /* SIGPIPE is a special case. It happens in -pipe mode
> - when the compiler dies before the preprocessor is
> - done, or the assembler dies before the compiler is
> - done. There's generally been an error already, and
> - this is just fallout. So don't generate another
> - error unless we would otherwise have succeeded. */
> - if (signal_count || greatest_status >= MIN_FATAL_STATUS)
> - {
> - signal_count++;
> - ret_code = -1;
> - break;
> - }
> + case SIGPIPE:
> + /* SIGPIPE is a special case. It happens in -pipe mode
> + when the compiler dies before the preprocessor is
> + done, or the assembler dies before the compiler is
> + done. There's generally been an error already, and
> + this is just fallout. So don't generate another
> + error unless we would otherwise have succeeded. */
> + if (signal_count || greatest_status >= MIN_FATAL_STATUS)
> + {
> + signal_count++;
> + ret_code = -1;
> + break;
> + }
> #endif
> - /* FALLTHROUGH */
> + /* FALLTHROUGH. */
>
> - default:
> - /* The inferior failed to catch the signal. */
> - internal_error_no_backtrace ("%s signal terminated program %s",
> - strsignal (WTERMSIG (status)),
> - commands[i].prog);
> - }
> - else if (WIFEXITED (status)
> - && WEXITSTATUS (status) >= MIN_FATAL_STATUS)
> - {
> - /* For ICEs in cc1, cc1obj, cc1plus see if it is
> - reproducible or not. */
> - const char *p;
> - if (flag_report_bug
> - && WEXITSTATUS (status) == ICE_EXIT_CODE
> - && i == 0
> - && (p = strrchr (commands[0].argv[0], DIR_SEPARATOR))
> - && ! strncmp (p + 1, "cc1", 3))
> - try_generate_repro (commands[0].argv);
> - if (WEXITSTATUS (status) > greatest_status)
> - greatest_status = WEXITSTATUS (status);
> - ret_code = -1;
> + default:
> + /* The inferior failed to catch the signal. */
> + internal_error_no_backtrace ("%s signal terminated program %s",
> + strsignal (WTERMSIG (status)),
> + commands[i].prog);
> }
> + else if (WIFEXITED (status)
> + && WEXITSTATUS (status) >= MIN_FATAL_STATUS)
> + {
> + /* For ICEs in cc1, cc1obj, cc1plus see if it is
> + reproducible or not. */
> + const char *p;
> + if (flag_report_bug
> + && WEXITSTATUS (status) == ICE_EXIT_CODE
> + && i == 0
> + && (p = strrchr (commands[0].argv[0], DIR_SEPARATOR))
> + && ! strncmp (p + 1, "cc1", 3))
> + try_generate_repro (commands[0].argv);
> + if (WEXITSTATUS (status) > greatest_status)
> + greatest_status = WEXITSTATUS (status);
> + ret_code = -1;
> + }
>
> - if (report_times || report_times_to_file)
> - {
> - struct pex_time *pt = ×[i];
> - double ut, st;
> + if (report_times || report_times_to_file)
> + {
> + struct pex_time *pt = ×[i];
> + double ut, st;
>
> - ut = ((double) pt->user_seconds
> - + (double) pt->user_microseconds / 1.0e6);
> - st = ((double) pt->system_seconds
> - + (double) pt->system_microseconds / 1.0e6);
> + ut = ((double) pt->user_seconds
> + + (double) pt->user_microseconds / 1.0e6);
> + st = ((double) pt->system_seconds
> + + (double) pt->system_microseconds / 1.0e6);
>
> - if (ut + st != 0)
> - {
> - if (report_times)
> - fnotice (stderr, "# %s %.2f %.2f\n",
> - commands[i].prog, ut, st);
> + if (ut + st != 0)
> + {
> + if (report_times)
> + fnotice (stderr, "# %s %.2f %.2f\n",
> + commands[i].prog, ut, st);
>
> - if (report_times_to_file)
> - {
> - int c = 0;
> - const char *const *j;
> + if (report_times_to_file)
> + {
> + int c = 0;
> + const char *const *j;
>
> - fprintf (report_times_to_file, "%g %g", ut, st);
> + fprintf (report_times_to_file, "%g %g", ut, st);
>
> - for (j = &commands[i].prog; *j; j = &commands[i].argv[++c])
> - {
> - const char *p;
> - for (p = *j; *p; ++p)
> - if (*p == '"' || *p == '\\' || *p == '$'
> - || ISSPACE (*p))
> - break;
> + for (j = &commands[i].prog; *j; j = &commands[i].argv[++c])
> + {
> + const char *p;
> + for (p = *j; *p; ++p)
> + if (*p == '"' || *p == '\\' || *p == '$'
> + || ISSPACE (*p))
> + break;
>
> - if (*p)
> - {
> - fprintf (report_times_to_file, " \"");
> - for (p = *j; *p; ++p)
> - {
> - if (*p == '"' || *p == '\\' || *p == '$')
> - fputc ('\\', report_times_to_file);
> - fputc (*p, report_times_to_file);
> - }
> - fputc ('"', report_times_to_file);
> - }
> - else
> - fprintf (report_times_to_file, " %s", *j);
> - }
> + if (*p)
> + {
> + fprintf (report_times_to_file, " \"");
> + for (p = *j; *p; ++p)
> + {
> + if (*p == '"' || *p == '\\' || *p == '$')
> + fputc ('\\', report_times_to_file);
> + fputc (*p, report_times_to_file);
> + }
> + fputc ('"', report_times_to_file);
> + }
> + else
> + fprintf (report_times_to_file, " %s", *j);
> + }
>
> - fputc ('\n', report_times_to_file);
> - }
> - }
> - }
> + fputc ('\n', report_times_to_file);
> + }
> + }
> + }
> + }
> +
> + return ret_code;
> +}
> +
> +/* Split a single command with pipes into several commands. */
> +
> +static void
> +split_commands (vec<const_char_p> *argbuf_p,
> + int n_commands, struct command commands[])
> +{
> + int i;
> + const char *arg;
> + vec<const_char_p> &argbuf = *argbuf_p;
> +
> + for (n_commands = 1, i = 0; argbuf.iterate (i, &arg); i++)
> + if (arg && strcmp (arg, "|") == 0)
> + { /* each command. */
> + const char *string;
> +#if defined (__MSDOS__) || defined (OS2) || defined (VMS)
> + fatal_error (input_location, "%<-pipe%> not supported");
> +#endif
> + argbuf[i] = 0; /* Termination of command args. */
> + commands[n_commands].prog = argbuf[i + 1];
> + commands[n_commands].argv
> + = &(argbuf.address ())[i + 1];
> + string = find_a_file (&exec_prefixes, commands[n_commands].prog,
> + X_OK, false);
> + if (string)
> + commands[n_commands].argv[0] = string;
> + n_commands++;
> }
> +}
> +
> +struct command *
> +parse_argbuf (vec <const_char_p> *argbuf_p, int *n)
> +{
> + int i, n_commands;
> + vec<const_char_p> &argbuf = *argbuf_p;
> + const char *arg;
> + struct command *commands;
>
> - if (commands[0].argv[0] != commands[0].prog)
> - free (CONST_CAST (char *, commands[0].argv[0]));
> + /* Count # of piped commands. */
> + for (n_commands = 1, i = 0; argbuf.iterate (i, &arg); i++)
> + if (strcmp (arg, "|") == 0)
> + n_commands++;
>
> - return ret_code;
> - }
> + /* Get storage for each command. */
> + commands = XNEWVEC (struct command, n_commands);
> +
> + /* Split argbuf into its separate piped processes,
> + and record info about each one.
> + Also search for the programs that are to be run. */
> +
> + argbuf.safe_push (0);
> +
> + commands[0].prog = argbuf[0]; /* first command. */
> + commands[0].argv = argbuf.address ();
> +
> + split_commands (argbuf_p, n_commands, commands);
> +
> + *n = n_commands;
> + return commands;
> +}
> +
> +/* Execute the command specified by the arguments on the current line of spec.
> + When using pipes, this includes several piped-together commands
> + with `|' between them.
> +
> + Return 0 if successful, -1 if failed. */
> +
> +static int
> +execute (void)
> +{
> + struct pex_obj *pex;
> + struct command *commands; /* each command buffer with program to call
> + and arguments. */
> + int n_commands; /* # of command. */
> + int ret = 0;
> +
> + struct command additional_ld = {NULL, NULL};
> + extra_arg_storer storer;
> +
> + struct command *commands_batch;
> + int n;
> +
> + gcc_assert (!processing_spec_function);
> +
> + if (wrapper_string)
> + {
> + char *string = find_a_file (&exec_prefixes, argbuf[0], X_OK, false);
> + if (string)
> + argbuf[0] = string;
> + insert_wrapper (wrapper_string);
> + }
> +
> + /* Parse the argbuf into several commands. */
> + commands = parse_argbuf (&argbuf, &n_commands);
> +
> + if (!have_S && !have_E && flag_parallel_jobs)
> + append_split_outputs (&storer, &additional_ld, &commands, &n_commands);
> +
> + if (!wrapper_string)
> + {
> + char *string = find_a_file (&exec_prefixes, commands[0].prog,
> + X_OK, false);
> + if (string)
> + commands[0].argv[0] = string;
> + }
> +
> + /* If -v, print what we are about to do, and maybe query. */
> +
> + if (verbose_flag)
> + {
> + int ret_verbose = handle_verbose (n_commands, commands);
> + if (ret_verbose > 0)
> + {
> + ret = 0;
> + goto cleanup;
> + }
> + }
> +
> +#ifdef ENABLE_VALGRIND_CHECKING
> + /* Stack of strings to be released on function return. */
> + struct obstack to_be_released;
> + obstack_init (&to_be_released);
> + append_valgrind (&to_be_released, n_commands, commands);
> +#endif
> +
> + /* FIXME: Interact with GNU Jobserver if necessary. */
> +
> + commands_batch = commands;
> + n = flag_parallel_jobs? 1: n_commands;
> +
> + for (int i = 0; i < n_commands; i += n)
> + {
> + /* Run each piped subprocess. */
> +
> + pex = pex_init (PEX_USE_PIPES | ((report_times || report_times_to_file)
> + ? PEX_RECORD_TIMES : 0),
> + progname, temp_filename);
> + if (pex == NULL)
> + fatal_error (input_location, "%<pex_init%> failed: %m");
> +
> + /* Lauch the commands. */
> + async_launch_commands (pex, n, commands_batch);
> +
> + /* Await them to be done. */
> + ret |= await_commands_to_finish (pex, n, commands_batch);
> +
> + commands_batch = commands_batch + n;
> +
> + /* Cleanup. */
> + pex_free (pex);
> + }
> +
> +
> + if (ret != 0)
> + goto cleanup;
> +
> + /* Run extra ld call. */
> + if (!EMPTY_CMD (additional_ld))
> + {
> + /* If we are here, we must be sure that we had at least two object
> + files to link. */
> + //gcc_assert (n_commands != 1);
> +
> + pex = pex_init (PEX_USE_PIPES | ((report_times || report_times_to_file)
> + ? PEX_RECORD_TIMES : 0),
> + progname, temp_filename);
> +
> + if (verbose_flag)
> + print_command (&additional_ld);
> +
> + async_launch_commands (pex, 1, &additional_ld);
> + ret = await_commands_to_finish (pex, 1, &additional_ld);
> + pex_free (pex);
> + }
> +
> +
> +#ifdef ENABLE_VALGRIND_CHECKING
> + obstack_free (&to_be_released, NULL);
> +#endif
> +
> +cleanup:
> + if (commands[0].argv[0] != commands[0].prog)
> + free (CONST_CAST (char *, commands[0].argv[0]));
> +
> + free (commands);
> +
> + return ret;
> }
> +
>
> /* Find all the switches given to us
> and make a vector describing them.
> @@ -3480,29 +4106,33 @@ static int n_switches_alloc_debug_check[2];
>
> static char *debug_check_temp_file[2];
>
> -/* Language is one of three things:
> -
> - 1) The name of a real programming language.
> - 2) NULL, indicating that no one has figured out
> - what it is yet.
> - 3) '*', indicating that the file should be passed
> - to the linker. */
> -struct infile
> +static const char *
> +fsplit_arg (extra_arg_storer *storer)
> {
> - const char *name;
> - const char *language;
> - struct compiler *incompiler;
> - bool compiled;
> - bool preprocessed;
> -};
> + const char *tempname = make_temp_file ("additional-asm");
> + const char arg[] = "-fsplit-outputs=";
> + char *final;
>
> -/* Also a vector of input files specified. */
> + size_t n = ARRAY_SIZE (arg) + strlen (tempname);
>
> -static struct infile *infiles;
> + gcc_assert (current_infile);
>
> -int n_infiles;
> + current_infile->temp_additional_asm = tempname;
> +
> + /* Remove file, once we may not even need it and create it later. */
> + /* FIXME: This is a little hackish. */
> + remove (tempname);
> +
> + final = storer->create_string (n);
> +
> + strcpy (final, arg);
> + strcat (final, tempname);
> +
> + record_temp_file (tempname, true, true);
> +
> + return final;
> +}
>
> -static int n_infiles_alloc;
>
> /* True if undefined environment variables encountered during spec processing
> are ok to ignore, typically when we're running for --help or --version. */
> @@ -3683,6 +4313,8 @@ alloc_infile (void)
> {
> n_infiles_alloc = 16;
> infiles = XNEWVEC (struct infile, n_infiles_alloc);
> + memset (infiles, 0x00, sizeof (*infiles) * n_infiles_alloc);
> +
> }
> else if (n_infiles_alloc == n_infiles)
> {
> @@ -4648,6 +5280,9 @@ process_command (unsigned int decoded_options_count,
> switch (decoded_options[j].opt_index)
> {
> case OPT_S:
> + have_S = 1;
> + have_c = 1;
> + break;
> case OPT_c:
> case OPT_E:
> have_c = 1;
> @@ -6155,11 +6790,14 @@ do_spec_1 (const char *spec, int inswitch, const char *soft_matched_part)
> open_at_file ();
>
> for (i = 0; (int) i < n_infiles; i++)
> - if (compile_input_file_p (&infiles[i]))
> - {
> - store_arg (infiles[i].name, 0, 0);
> - infiles[i].compiled = true;
> - }
> + {
> + current_infile = &infiles[i];
> + if (compile_input_file_p (current_infile))
> + {
> + store_arg (current_infile->name, 0, 0);
> + current_infile->compiled = true;
> + }
> + }
>
> if (at_file_supplied)
> close_at_file ();
> @@ -6515,7 +7153,7 @@ do_spec_1 (const char *spec, int inswitch, const char *soft_matched_part)
> "%{foo=*:bar%*}%{foo=*:one%*two}"
>
> matches -foo=hello then it will produce:
> -
> +
> barhello onehellotwo
> */
> if (*p == 0 || *p == '}')
> @@ -8642,6 +9280,7 @@ driver::do_spec_on_infiles () const
> for (i = 0; (int) i < n_infiles; i++)
> {
> int this_file_error = 0;
> + current_infile = &infiles[i];
>
> /* Tell do_spec what to substitute for %i. */
>
> @@ -8761,12 +9400,15 @@ driver::do_spec_on_infiles () const
> int i;
>
> for (i = 0; i < n_infiles ; i++)
> - if (infiles[i].incompiler
> - || (infiles[i].language && infiles[i].language[0] != '*'))
> - {
> - set_input (infiles[i].name);
> - break;
> - }
> + {
> + current_infile = &infiles[i];
> + if (infiles[i].incompiler
> + || (infiles[i].language && infiles[i].language[0] != '*'))
> + {
> + set_input (infiles[i].name);
> + break;
> + }
> + }
> }
>
> if (!seen_error ())
> @@ -8788,11 +9430,31 @@ driver::maybe_run_linker (const char *argv0) const
> int linker_was_run = 0;
> int num_linker_inputs;
>
> - /* Determine if there are any linker input files. */
> - num_linker_inputs = 0;
> - for (i = 0; (int) i < n_infiles; i++)
> - if (explicit_link_files[i] || outfiles[i] != NULL)
> - num_linker_inputs++;
> + /* Set outfiles to be the temporary object vector. */
> + const char **outfiles_holder = outfiles;
> + int n_infiles_holder = n_infiles;
> + bool outfiles_switched = false;
> + if (temp_object_files.length () > 0)
> + {
> + /* Insert explicit link files into the temp object vector. */
> +
> + for (i = 0; (int) i < n_infiles; i++)
> + if (explicit_link_files[i] && outfiles[i] != NULL)
> + temp_object_files.safe_push (outfiles[i]);
> +
> + num_linker_inputs = n_infiles = temp_object_files.length ();
> + temp_object_files.safe_push (NULL); /* the NULL sentinel. */
> + outfiles = temp_object_files.address ();
> + }
> + else /* Fall back to the old method. */
> + {
> +
> + /* Determine if there are any linker input files. */
> + num_linker_inputs = 0;
> + for (i = 0; (int) i < n_infiles; i++)
> + if (explicit_link_files[i] || outfiles[i] != NULL)
> + num_linker_inputs++;
> + }
>
> /* Arrange for temporary file names created during linking to take
> on names related with the linker output rather than with the
> @@ -8897,14 +9559,24 @@ driver::maybe_run_linker (const char *argv0) const
> }
>
> /* If options said don't run linker,
> - complain about input files to be given to the linker. */
> + complain about input files to be given to the linker.
> + When fsplit-arg is active, the linker will run and this if
> + will not be triggered. */
>
> - if (! linker_was_run && !seen_error ())
> + if (!outfiles_switched && !linker_was_run && !seen_error ()
> + && temp_object_files.length () == 0)
> for (i = 0; (int) i < n_infiles; i++)
> if (explicit_link_files[i]
> && !(infiles[i].language && infiles[i].language[0] == '*'))
> warning (0, "%s: linker input file unused because linking not done",
> outfiles[i]);
> +
> + if (outfiles_switched)
> + {
> + /* Undo our changes. */
> + outfiles = outfiles_holder;
> + n_infiles = n_infiles_holder;
> + }
> }
>
> /* The end of "main". */
> @@ -10808,6 +11480,7 @@ driver::finalize ()
> linker_options.truncate (0);
> assembler_options.truncate (0);
> preprocessor_options.truncate (0);
> + temp_object_files.truncate (0);
>
> path_prefix_reset (&exec_prefixes);
> path_prefix_reset (&startfile_prefixes);
> diff --git a/gcc/testsuite/driver/a.c b/gcc/testsuite/driver/a.c
> new file mode 100644
> index 00000000000..c6b8c2eb61e
> --- /dev/null
> +++ b/gcc/testsuite/driver/a.c
> @@ -0,0 +1,6 @@
> +int puts (const char *);
> +
> +void a_func (void)
> +{
> + puts ("A test");
> +}
> diff --git a/gcc/testsuite/driver/b.c b/gcc/testsuite/driver/b.c
> new file mode 100644
> index 00000000000..76a2cba0bd9
> --- /dev/null
> +++ b/gcc/testsuite/driver/b.c
> @@ -0,0 +1,6 @@
> +int puts (const char *);
> +
> +void a_func (void)
> +{
> + puts ("Another test");
> +}
> diff --git a/gcc/testsuite/driver/driver.exp b/gcc/testsuite/driver/driver.exp
> new file mode 100644
> index 00000000000..2bbaf07778a
> --- /dev/null
> +++ b/gcc/testsuite/driver/driver.exp
> @@ -0,0 +1,80 @@
> +# Copyright (C) 2008-2020 Free Software Foundation, Inc.
> +
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# 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 GCC; see the file COPYING3. If not see
> +# <http://www.gnu.org/licenses/>.
> +
> +# GCC testsuite that uses the `dg.exp' driver.
> +
> +# Load support procs.
> +load_lib gcc-dg.exp
> +
> +proc check-for-errors { test input } {
> + if { [string equal "$input" ""] } then {
> + pass "$test: std out"
> + } else {
> + fail "$test: std out\n$input"
> + }
> +}
> +
> +if ![check_effective_target_pthread] {
> + return
> +}
> +
> +# If a testcase doesn't have special options, use these.
> +global DEFAULT_CFLAGS
> +if ![info exists DEFAULT_CFLAGS] then {
> + set DEFAULT_CFLAGS " -ansi -pedantic-errors"
> +}
> +
> +# Initialize `dg'.
> +dg-init
> +
> +
> +# Test multi-input compilation
> +check-for-errors "Multi-input Compilation" \
> + [gcc_target_compile "$srcdir/$subdir/a.c $srcdir/$subdir/b.c -c" "" none ""]
> +
> +# Compile file and generate an assembler and object file
> +check-for-errors "Object Generation" \
> + [gcc_target_compile "$srcdir/$subdir/a.c -c" "a.o" none ""]
> +check-for-errors "Object Generation" \
> + [gcc_target_compile "$srcdir/$subdir/b.c -c" "a.o" none ""]
> +check-for-errors "Assembler Generation" \
> + [gcc_target_compile "$srcdir/$subdir/a.c -S" "a.S" none ""]
> +check-for-errors "Assembler Generation" \
> + [gcc_target_compile "$srcdir/$subdir/b.c -S" "b.S" none ""]
> +
> +# Empty file is a valid program
> +check-for-errors "Empty Program" \
> + [gcc_target_compile "$srcdir/$subdir/empty.c -c" "empty.o" none ""]
> +
> +# Test object file passthrough
> +check-for-errors "Object file passthrough" \
> + [gcc_target_compile "$srcdir/$subdir/foo.c a.o" "a.exe" none ""]
> +
> +# Test compilation when assembler is provided
> +check-for-errors "Assembler with Macros" \
> + [gcc_target_compile "a.S -c" "a.o" none ""]
> +
> +# Clean temporary generated files.
> +set temp_files {"a.o" "a.S" "b.o" "b.S" "empty.o"}
> +
> +foreach f $temp_files {
> + if { [file exists $f] } {
> + file delete $f
> + }
> +}
> +
> +# All done.
> +dg-finish
> diff --git a/gcc/testsuite/driver/empty.c b/gcc/testsuite/driver/empty.c
> new file mode 100644
> index 00000000000..e69de29bb2d
> diff --git a/gcc/testsuite/driver/foo.c b/gcc/testsuite/driver/foo.c
> new file mode 100644
> index 00000000000..a18fd2a3b14
> --- /dev/null
> +++ b/gcc/testsuite/driver/foo.c
> @@ -0,0 +1,7 @@
> +void a_func (void);
> +
> +int main()
> +{
> + a_func ();
> + return 0;
> +}
> --
> 2.28.0
>
More information about the Gcc-patches
mailing list