libctf: move string deduplication into ctf-archive

This means that any archive containing dicts can get its strings dedupped
together, rather than only those that are ctf_linked.

(For now, we are still constrained to ctf_linked archives, since fixing that
requires further changes to ctf_dedup_strings: but this gives us the first
half of what is necessary.)

libctf/
	* ctf-link.c (ctf_link_write): Move string dedup into...
	* ctf-archive.c (ctf_arc_preserialize): ... this new function.
	(ctf_arc_write_fd): Call it.

libctf/ctf-api.c

@@ -152,7 +152,7 @@ static ctf_list_t open_errors;
    open errors list if NULL): if ERR is nonzero it is the errno to report to the
    debug stream instead of that recorded on fp.  */
 _libctf_printflike_ (4, 5)
-extern void
+void
 ctf_err_warn (ctf_dict_t *fp, int is_warning, int err,
 	      const char *format, ...)
 {
@@ -203,6 +203,18 @@ ctf_err_warn_to_open (ctf_dict_t *fp)
   ctf_list_splice (&open_errors, &fp->ctf_errs_warnings);
 }
 
+/* Copy all the errors/warnings from one fp to another one, and the error code
+   as well.  */
+void
+ctf_err_copy (ctf_dict_t *dest, ctf_dict_t *src)
+{
+  ctf_err_warning_t *cew;
+  for (cew = ctf_list_next (&src->ctf_errs_warnings); cew != NULL;
+       cew = ctf_list_next (cew))
+    ctf_err_warn (dest, cew->cew_is_warning, 0, cew->cew_text);
+  ctf_set_errno (dest, ctf_errno (src));
+}
+
 /* Error-warning reporting: an 'iterator' that returns errors and warnings from
    the error/warning list, in order of emission.  Errors and warnings are popped
    after return: the caller must free the returned error-text pointer.

libctf/ctf-archive.c

@@ -51,6 +51,64 @@ static int ctf_arc_import_parent (const ctf_archive_t *arc, ctf_dict_t *fp,
    and ctfi_symnamedicts.  Never initialized.  */
 static ctf_dict_t enosym;
 
+/* Prepare to serialize everything.  Members of archives have dependencies on
+   each other, because the strtabs and type IDs of children depend on the
+   parent: so we have to work over the archive as a whole to prepare for final
+   serialization.
+
+   Returns zero on success, or an errno, or an ECTF_* value.
+
+   Updates the first dict in the archive with the errno value.  */
+
+static int
+ctf_arc_preserialize (ctf_dict_t **ctf_dicts, ssize_t ctf_dict_cnt)
+{
+  uint64_t old_parent_strlen, all_strlens = 0;
+  ssize_t i;
+  int err;
+
+  ctf_dprintf ("Preserializing dicts.\n");
+
+  /* Preserialize everything, doing everything but strtab generation and things
+     that depend on that.  */
+  for (i = 0; i < ctf_dict_cnt; i++)
+    if (ctf_preserialize (ctf_dicts[i]) < 0)
+      goto err;
+
+  ctf_dprintf ("Deduplicating strings.\n");
+
+  for (i = 0; i < ctf_dict_cnt; i++)
+    all_strlens += ctf_dicts[i]->ctf_str[0].cts_len
+      + ctf_dicts[i]->ctf_str_prov_len;
+
+  /* If linking, deduplicate strings against the children in every dict that has
+     any.  (String deduplication is not yet implemented for non-linked dicts.)  */
+  for (i = 0; i < ctf_dict_cnt; i++)
+    if (ctf_dicts[i]->ctf_flags & LCTF_LINKING && ctf_dicts[i]->ctf_link_outputs)
+      {
+	old_parent_strlen = ctf_dicts[i]->ctf_str[0].cts_len
+	  + ctf_dicts[i]->ctf_str_prov_len;
+
+	if (ctf_dedup_strings (ctf_dicts[i]) < 0)
+	  goto err;
+
+	ctf_dprintf ("Deduplicated strings in archive member %zi: "
+		     "original parent strlen: %zu; original lengths: %zu; "
+		     "final length: %zu.\n", i, (size_t) old_parent_strlen,
+		     (size_t) all_strlens,
+		     (size_t) ctf_dicts[i]->ctf_str_prov_len);
+      }
+
+  return 0;
+
+ err:
+  err = ctf_errno (ctf_dicts[i]);
+  ctf_err_copy (ctf_dicts[0], ctf_dicts[i]);
+  for (i--; i >= 0; i--)
+    ctf_depreserialize (ctf_dicts[i]);
+  return err;
+}
+
 /* Write out a CTF archive to the start of the file referenced by the passed-in
    fd.  The entries in CTF_DICTS are referenced by name: the names are passed in
    the names array, which must have CTF_DICTS entries.
@@ -70,8 +128,15 @@ ctf_arc_write_fd (int fd, ctf_dict_t **ctf_dicts, size_t ctf_dict_cnt,
   char *nametbl = NULL;		/* The name table.  */
   char *np;
   off_t nameoffs;
+  int err;
   struct ctf_archive_modent *modent;
 
+  /* Prepare by serializing everything.  Done first because it allocates a lot
+     of space and thus is more likely to fail.  */
+  if (ctf_dict_cnt > 0 &&
+      (err = ctf_arc_preserialize (ctf_dicts, ctf_dict_cnt)) < 0)
+    return err;
+
   ctf_dprintf ("Writing CTF archive with %lu files\n",
 	       (unsigned long) ctf_dict_cnt);
 

libctf/ctf-impl.h

@@ -792,6 +792,7 @@ _libctf_printflike_ (4, 5)
 extern void ctf_err_warn (ctf_dict_t *, int is_warning, int err,
 			  const char *, ...);
 extern void ctf_err_warn_to_open (ctf_dict_t *);
+extern void ctf_err_copy (ctf_dict_t *dest, ctf_dict_t *src);
 extern void ctf_assert_fail_internal (ctf_dict_t *, const char *,
 				      size_t, const char *);
 extern const char *ctf_link_input_name (ctf_dict_t *);

libctf/ctf-link.c

@@ -2052,7 +2052,6 @@ ctf_link_write (ctf_dict_t *fp, size_t *size, size_t threshold)
   long fsize;
   const char *errloc;
   unsigned char *buf = NULL;
-  uint64_t old_parent_strlen, all_strlens = 0;
 
   memset (&arg, 0, sizeof (ctf_name_list_accum_cb_arg_t));
   arg.fp = fp;
@@ -2123,41 +2122,6 @@ ctf_link_write (ctf_dict_t *fp, size_t *size, size_t threshold)
   memmove (&(arg.files[1]), arg.files, sizeof (ctf_dict_t *) * (arg.i));
   arg.files[0] = fp;
 
-  /* Preserialize everything, doing everything but strtab generation and things that
-     depend on that.  */
-  for (i = 0; i < arg.i + 1; i++)
-    {
-      if (ctf_preserialize (arg.files[i]) < 0)
-	{
-	  errno = ctf_errno (arg.files[i]);
-	  for (i--; i >= 0; i--)
-	    ctf_depreserialize (arg.files[i]);
-	  errloc = "preserialization";
-	  goto err_no;
-	}
-    }
-
-  ctf_dprintf ("Deduplicating strings.\n");
-
-  for (i = 0; i < arg.i; i++)
-    all_strlens += arg.files[i]->ctf_str[0].cts_len
-      + arg.files[i]->ctf_str_prov_len;
-  old_parent_strlen = arg.files[0]->ctf_str[0].cts_len
-    + arg.files[0]->ctf_str_prov_len;
-
-  if (ctf_dedup_strings (fp) < 0)
-    {
-      for (i = 0; i < arg.i + 1; i++)
-	ctf_depreserialize (arg.files[i]);
-      errloc = "string deduplication";
-      goto err_str_dedup;
-    }
-
-  ctf_dprintf ("Deduplicated strings: original parent strlen: %zu; "
-	       "original lengths: %zu; final length: %zu.\n",
-	       (size_t) old_parent_strlen, (size_t) all_strlens,
-	       (size_t) arg.files[0]->ctf_str_prov_len);
-
   if ((f = tmpfile ()) == NULL)
     {
       errloc = "tempfile creation";
@@ -2168,9 +2132,8 @@ ctf_link_write (ctf_dict_t *fp, size_t *size, size_t threshold)
 			       (const char **) arg.names,
 			       threshold)) < 0)
     {
-      errloc = "archive writing";
+      errloc = NULL;				/* errno is set for us.  */
-      errno = err;
+      goto err_set;
-      goto err_no;
     }
 
   if (fseek (f, 0, SEEK_END) < 0)
@@ -2229,7 +2192,7 @@ ctf_link_write (ctf_dict_t *fp, size_t *size, size_t threshold)
 
  err_no:
   ctf_set_errno (fp, errno);
- err_str_dedup:
+ err_set:
   /* Turn off the is-linking flag on all the dicts in this link, as above.  */
   for (i = 0; i < arg.i; i++)
     {
@@ -2251,7 +2214,8 @@ ctf_link_write (ctf_dict_t *fp, size_t *size, size_t threshold)
 	free (arg.dynames[i]);
       free (arg.dynames);
     }
-  ctf_err_warn (fp, 0, 0, _("cannot write archive in link: %s failure"),
+  if (errloc)
-		errloc);
+    ctf_err_warn (fp, 0, 0, _("cannot write archive in link: %s failure"),
+		  errloc);
   return NULL;
 }