public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r13-5162] hash table: enforce testing is_empty before is_deleted
@ 2023-01-14  0:16 Alexandre Oliva
  0 siblings, 0 replies; only message in thread
From: Alexandre Oliva @ 2023-01-14  0:16 UTC (permalink / raw)
  To: gcc-cvs

https://gcc.gnu.org/g:88679960c2665d87c8354ce35a48aaadbe3f0793

commit r13-5162-g88679960c2665d87c8354ce35a48aaadbe3f0793
Author: Alexandre Oliva <oliva@adacore.com>
Date:   Fri Jan 13 21:15:44 2023 -0300

    hash table: enforce testing is_empty before is_deleted
    
    Existing hash_table traits that use the same representation for empty
    and deleted slots reject marking slots as deleted, and to not pass
    is_deleted for slots that pass is_empty.
    
    Nevertheless, nearly everywhere, we only test for is_deleted after
    checking that !is_empty first.  The one exception was the copy
    constructor, that would fail if traits recognized is_empty slots as
    is_deleted, but then refused to mark_deleted.
    
    This asymmetry is neither necessary nor desirable, and there is a
    theoretical risk that traits might not only fail to refuse to
    mark_deleted, but also return is_deleted for is_empty slots.
    
    This patch introduces checks that detect these potentially problematic
    situations, and reorders the tests in the copy constructor so as to
    use the conventional testing order and thus avoid them.
    
    
    for  gcc/ChangeLog
    
            * hash-table.h (is_deleted): Precheck !is_empty.
            (mark_deleted): Postcheck !is_empty.
            (copy constructor): Test is_empty before is_deleted.

Diff:
---
 gcc/hash-table.h | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

diff --git a/gcc/hash-table.h b/gcc/hash-table.h
index 1d3166504c3..e37625dc315 100644
--- a/gcc/hash-table.h
+++ b/gcc/hash-table.h
@@ -534,6 +534,11 @@ private:
   void expand ();
   static bool is_deleted (value_type &v)
   {
+    /* Traits are supposed to avoid recognizing elements as both empty
+       and deleted, but to fail safe in case custom traits fail to do
+       that, make sure we never test for is_deleted without having
+       first ruled out is_empty.  */
+    gcc_checking_assert (!Descriptor::is_empty (v));
     return Descriptor::is_deleted (v);
   }
 
@@ -545,6 +550,11 @@ private:
   static void mark_deleted (value_type &v)
   {
     Descriptor::mark_deleted (v);
+    /* Traits are supposed to refuse to set elements as deleted if
+       those would be indistinguishable from empty, but to fail safe
+       in case custom traits fail to do that, check that the
+       just-deleted element does not look empty.  */
+    gcc_checking_assert (!Descriptor::is_empty (v));
   }
 
   static void mark_empty (value_type &v)
@@ -700,9 +710,11 @@ hash_table<Descriptor, Lazy, Allocator>::hash_table (const hash_table &h,
       for (size_t i = 0; i < size; ++i)
 	{
 	  value_type &entry = h.m_entries[i];
-	  if (is_deleted (entry))
+	  if (is_empty (entry))
+	    continue;
+	  else if (is_deleted (entry))
 	    mark_deleted (nentries[i]);
-	  else if (!is_empty (entry))
+	  else
 	    new ((void*) (nentries + i)) value_type (entry);
 	}
       m_entries = nentries;

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2023-01-14  0:16 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-01-14  0:16 [gcc r13-5162] hash table: enforce testing is_empty before is_deleted Alexandre Oliva

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).