From: Nick Alcock <nick.alcock@oracle.com>
Date: Mon, 15 Jul 2024 21:08:10 +0000 (+0100)
Subject: libctf: do not deduplicate strings in the header
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=ba66e0cc32d749119485025e7f050021094e1b15;p=thirdparty%2Fbinutils-gdb.git

libctf: do not deduplicate strings in the header

It is unreasonable to expect users to ctf_import the parent before being
able to understand the header -- doubly so because the only string in the
header which is likely to be deduplicable is the parent name, which is the
same in every child, yet without the parent name being *available* in the
child's strtab you cannot call ctf_parent_name to figure out which parent
to import!

libctf/
	* ctf-serialize.c (ctf_preserialize): Prevent deduplication of header string
        fields.
	* ctf-open.c (ctf_set_base): Note this.
	* ctf-string.c (ctf_str_free_atom): Likewise.
---

diff --git a/libctf/ctf-open.c b/libctf/ctf-open.c
index 3ba78c89426..56b1084a6d6 100644
--- a/libctf/ctf-open.c
+++ b/libctf/ctf-open.c
@@ -364,7 +364,12 @@ ctf_set_base (ctf_dict_t *fp, const ctf_header_t *hp, unsigned char *base)
 
   /* Note: before conversion, these will be set to values that will be
      immediately invalidated by the conversion process, but the conversion
-     process will call ctf_set_base() again to fix things up.  */
+     process will call ctf_set_base() again to fix things up.
+
+     These labels are explicitly constrained from being deduplicated (even though
+     .ctf is usually a duplicated name), because they are the key to identifying
+     the parent dict (and determining that this dict is a child) in the first
+     place.  */
 
   if (hp->cth_parlabel != 0)
     fp->ctf_parlabel = ctf_strptr (fp, hp->cth_parlabel);
diff --git a/libctf/ctf-serialize.c b/libctf/ctf-serialize.c
index 19ab8abf63e..52ae41d33ef 100644
--- a/libctf/ctf-serialize.c
+++ b/libctf/ctf-serialize.c
@@ -1048,9 +1048,9 @@ ctf_preserialize (ctf_dict_t *fp)
 
   hdrp = (ctf_header_t *) buf;
   if ((fp->ctf_flags & LCTF_CHILD) && (fp->ctf_parname != NULL))
-    ctf_str_add_ref (fp, fp->ctf_parname, &hdrp->cth_parname);
+    ctf_str_add_no_dedup_ref (fp, fp->ctf_parname, &hdrp->cth_parname);
   if (fp->ctf_cuname != NULL)
-    ctf_str_add_ref (fp, fp->ctf_cuname, &hdrp->cth_cuname);
+    ctf_str_add_no_dedup_ref (fp, fp->ctf_cuname, &hdrp->cth_cuname);
 
   if (ctf_emit_symtypetab_sects (fp, &symstate, &t, objt_size, func_size,
 				 objtidx_size, funcidx_size) < 0)
diff --git a/libctf/ctf-string.c b/libctf/ctf-string.c
index dc72555c38f..f171f12b494 100644
--- a/libctf/ctf-string.c
+++ b/libctf/ctf-string.c
@@ -202,7 +202,9 @@ ctf_str_free_atom (void *a)
    pulled in.)
 
    Note that the *final strtab* may be entirely empty, if all its strings are
-   shared with the parent: the atoms table is a superset.  */
+   shared with the parent: the atoms table is a superset.  (But this will never
+   happen in practice, because some header fields are explicitly never
+   deduplicated.)  */
 int
 ctf_str_create_atoms (ctf_dict_t *fp)
 {