[PATCH v2] Add string cache and use it in cooked index

[PATCH v2] Add string cache and use it in cooked index

[Tom Tromey]

The cooked index needs to allocate names in some cases -- when
canonicalizing or when synthesizing Ada package names.  This process
currently uses a vector of unique_ptrs to manage the memory.

Another series I'm writing adds another spot where this allocation
must be done, and examining the result showed that certain names were
allocated multiple times.

To clean this up, this patch introduces a string cache object and
changes the cooked indexer to use it.  I considered using bcache here,
but bcache doesn't work as nicely with string_view -- because bcache
is fundamentally memory-based, a temporary copy of the contents must
be made to ensure that bcache can see the trailing \0.  Furthermore,
writing a custom class lets us avoid another copy when canonicalizing
C++ names.
---
 gdb/dwarf2/cooked-index.c |  16 ++---
 gdb/dwarf2/cooked-index.h |   3 +-
 gdbsupport/string-set.h   | 129 ++++++++++++++++++++++++++++++++++++++
 3 files changed, 135 insertions(+), 13 deletions(-)
 create mode 100644 gdbsupport/string-set.h

diff --git a/gdb/dwarf2/cooked-index.c b/gdb/dwarf2/cooked-index.c
index 6612585649f..9d3a4b03489 100644
--- a/gdb/dwarf2/cooked-index.c
+++ b/gdb/dwarf2/cooked-index.c
@@ -368,13 +368,11 @@ cooked_index_shard::handle_gnat_encoded_entry
       cooked_index_entry *last = (cooked_index_entry *) *slot;
       if (last == nullptr || last->per_cu != entry->per_cu)
 	{
-	  gdb::unique_xmalloc_ptr<char> new_name
-	    = make_unique_xstrndup (name.data (), name.length ());
+	  const char *new_name = m_names.insert (name);
 	  last = create (entry->die_offset, DW_TAG_module,
-			 IS_SYNTHESIZED, language_ada, new_name.get (), parent,
+			 IS_SYNTHESIZED, language_ada, new_name, parent,
 			 entry->per_cu);
 	  last->canonical = last->name;
-	  m_names.push_back (std::move (new_name));
 	  new_entries.push_back (last);
 	  *slot = last;
 	}
@@ -383,9 +381,7 @@ cooked_index_shard::handle_gnat_encoded_entry
     }
 
   entry->set_parent (parent);
-  auto new_canon = make_unique_xstrndup (tail.data (), tail.length ());
-  entry->canonical = new_canon.get ();
-  m_names.push_back (std::move (new_canon));
+  entry->canonical = m_names.insert (tail);
 }
 
 /* See cooked-index.h.  */
@@ -503,10 +499,7 @@ cooked_index_shard::finalize (const parent_map_map *parent_maps)
 	      if (canon_name == nullptr)
 		entry->canonical = entry->name;
 	      else
-		{
-		  entry->canonical = canon_name.get ();
-		  m_names.push_back (std::move (canon_name));
-		}
+		entry->canonical = m_names.insert (std::move (canon_name));
 	      *slot = entry;
 	    }
 	  else
@@ -526,7 +519,6 @@ cooked_index_shard::finalize (const parent_map_map *parent_maps)
   m_entries.insert (m_entries.end (), new_gnat_entries.begin (),
 		    new_gnat_entries.end ());
 
-  m_names.shrink_to_fit ();
   m_entries.shrink_to_fit ();
   std::sort (m_entries.begin (), m_entries.end (),
 	     [] (const cooked_index_entry *a, const cooked_index_entry *b)
diff --git a/gdb/dwarf2/cooked-index.h b/gdb/dwarf2/cooked-index.h
index f6586359770..c6911c23781 100644
--- a/gdb/dwarf2/cooked-index.h
+++ b/gdb/dwarf2/cooked-index.h
@@ -32,6 +32,7 @@
 #include "dwarf2/read.h"
 #include "dwarf2/parent-map.h"
 #include "gdbsupport/range-chain.h"
+#include "gdbsupport/string-set.h"
 #include "complaints.h"
 
 #if CXX_STD_THREAD
@@ -367,7 +368,7 @@ class cooked_index_shard
   /* The addrmap.  This maps address ranges to dwarf2_per_cu objects.  */
   addrmap_fixed *m_addrmap = nullptr;
   /* Storage for canonical names.  */
-  std::vector<gdb::unique_xmalloc_ptr<char>> m_names;
+  gdb::string_set m_names;
 };
 
 using cooked_index_shard_up = std::unique_ptr<cooked_index_shard>;
diff --git a/gdbsupport/string-set.h b/gdbsupport/string-set.h
new file mode 100644
index 00000000000..c2507454650
--- /dev/null
+++ b/gdbsupport/string-set.h
@@ -0,0 +1,129 @@
+/* String-interning set
+
+   Copyright (C) 2025 Free Software Foundation, Inc.
+
+   This file is part of GDB.
+
+   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 this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+#ifndef GDBSUPPORT_STRING_SET_H
+#define GDBSUPPORT_STRING_SET_H
+
+#include "gdbsupport/unordered_set.h"
+#include <string_view>
+
+namespace gdb
+{
+
+/* This is a string-interning set.  It manages storage for strings,
+   ensuring that just a single copy of a given string is kept.  The
+   underlying C string will remain valid for the lifetime of this
+   object.  */
+
+class string_set
+{
+public:
+
+  /* Insert STR into this set.  Returns a pointer to the interned
+     string.  */
+  const char *insert (const char *str)
+  {
+    /* We need to take the length to hash the string anyway, so it's
+       convenient to just wrap it here.  */
+    return insert (std::string_view (str));
+  }
+
+  /* An overload accepting a string view.  */
+  const char *insert (std::string_view str)
+  {
+    return m_set.insert (str).first->get ();
+  }
+
+  /* An overload that takes ownership of the string.  */
+  const char *insert (gdb::unique_xmalloc_ptr<char> str)
+  {
+    return m_set.insert (local_string (std::move (str))).first->get ();
+  }
+
+private:
+
+  /* The type of string we store.  Note that we do not store
+     std::string here to avoid the small-string optimization
+     invalidating a pointer on rehash.  */
+  struct local_string
+  {
+    explicit local_string (std::string_view str)
+      : contents (xstrndup (str.data (), str.size ())),
+	len (str.size ())
+    { }
+
+    explicit local_string (gdb::unique_xmalloc_ptr<char> str)
+      : contents (std::move (str)),
+	len (strlen (contents.get ()))
+    { }
+
+    const char *get () const
+    { return contents.get (); }
+
+    std::string_view as_view () const
+    { return std::string_view (contents.get (), len); }
+
+    /* \0-terminated string contents.  */
+    gdb::unique_xmalloc_ptr<char> contents;
+    /* Length of the string.  */
+    size_t len;
+  };
+
+  /* Equality object for the set.  */
+  struct str_eq
+  {
+    using is_transparent = void;
+
+    bool operator() (std::string_view lhs, const local_string &rhs)
+      const noexcept
+    {
+      return lhs == rhs.as_view ();
+    }
+
+    bool operator() (const local_string &lhs, const local_string &rhs)
+      const noexcept
+    {
+      return (*this) (lhs.as_view (), rhs);
+    }
+  };
+
+  /* Hash object for the set.  */
+  struct str_hash
+  {
+    using is_transparent = void;
+    using is_avalanching = void;
+
+    bool operator() (const local_string &rhs) const noexcept
+    {
+      return (*this) (rhs.as_view ());
+    }
+
+    bool operator() (std::string_view rhs) const noexcept
+    {
+      return ankerl::unordered_dense::hash<std::string_view> () (rhs);
+    }
+  };
+
+  /* The strings.  */
+  gdb::unordered_set<local_string, str_hash, str_eq> m_set;
+};
+
+}
+
+#endif /* GDBSUPPORT_STRING_SET_H */
-- 
2.46.1

Re: [PATCH v2] Add string cache and use it in cooked index

[Simon Marchi]

On 3/10/25 3:38 PM, Tom Tromey wrote:
> The cooked index needs to allocate names in some cases -- when
> canonicalizing or when synthesizing Ada package names.  This process
> currently uses a vector of unique_ptrs to manage the memory.
> 
> Another series I'm writing adds another spot where this allocation
> must be done, and examining the result showed that certain names were
> allocated multiple times.
> 
> To clean this up, this patch introduces a string cache object and
> changes the cooked indexer to use it.  I considered using bcache here,
> but bcache doesn't work as nicely with string_view -- because bcache
> is fundamentally memory-based, a temporary copy of the contents must
> be made to ensure that bcache can see the trailing \0.  Furthermore,
> writing a custom class lets us avoid another copy when canonicalizing
> C++ names.

Thanks, LGTM.

Approved-By: Simon Marchi <simon.marchi@efficios.com>

I was thinking that maybe this woud actually be a good use case for an
obstack, since free'ing would probably be faster than freeing the
strings individually.  But we have cases where the caller passes a
unique_xmalloc_ptr, in two cases we'd need an extra copy into the
obstack.

Simon

Re: [PATCH v2] Add string cache and use it in cooked index

[Tom Tromey]

>>>>> "Simon" == Simon Marchi <simark@simark.ca> writes:

Simon> I was thinking that maybe this woud actually be a good use case for an
Simon> obstack, since free'ing would probably be faster than freeing the
Simon> strings individually.  But we have cases where the caller passes a
Simon> unique_xmalloc_ptr, in two cases we'd need an extra copy into the
Simon> obstack.

Yeah.  Maybe it would still be better, I didn't try that approach.  I
tend to suspect the speed of scanning is more important than the speed
of destroying the objects but I didn't measure it or anything.  Anyway
it's easily done if someone needs it.

Tom