[Bugs] [Bug 1377584] memory leak problems are found in daemon:glusterd, server: glusterfsd and client:glusterfs

bugzilla at redhat.com bugzilla at redhat.com
Mon Dec 12 06:17:13 UTC 2016


https://bugzilla.redhat.com/show_bug.cgi?id=1377584

George <george.lian at nokia.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |george.lian at nokia.com



--- Comment #12 from George <george.lian at nokia.com> ---
diff: 
--- inode.c
+++ inode_fix.c
@@ -1582,13 +1582,13 @@
         new->xl = xl;
         new->ctxcount = xl->graph->xl_count + 1;

-        new->lru_limit = lru_limit;
-
         new->hashsize = 14057; /* TODO: Random Number?? */

         /* In case FUSE is initing the inode table. */
         if (lru_limit == 0)
                 lru_limit = DEFAULT_INODE_MEMPOOL_ENTRIES;
+            
+        new->lru_limit = lru_limit;

         new->inode_pool = mem_pool_new (inode_t, lru_limit);

the above change will affect function in inode_table_prune:
                while (table->lru_limit
                       && table->lru_size > (table->lru_limit)) {

                        entry = list_entry (table->lru.next, inode_t, list);

                        table->lru_size--;
                        __inode_retire (entry);

                        ret++;
                }

it will lead to __inode_retire without notify the kernel, and while kerenl do
fuse_forget, it finally will lead to invalid memory access and finally lead to
coredump, isn't it?

-- 
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug https://bugzilla.redhat.com/token.cgi?t=SvmI4XtX3J&a=cc_unsubscribe


More information about the Bugs mailing list