From: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Date: Wed, 5 Dec 2018 21:45:02 +0000 (+0100)
Subject: journald: set a limit on the number of fields (1k)
X-Git-Tag: v241-rc1~98^2~5
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=052c57f132f04a3cf4148f87561618da1a6908b4;p=thirdparty%2Fsystemd.git

journald: set a limit on the number of fields (1k)

We allocate a iovec entry for each field, so with many short entries,
our memory usage and processing time can be large, even with a relatively
small message size. Let's refuse overly long entries.

CVE-2018-16865
https://bugzilla.redhat.com/show_bug.cgi?id=1653861

What from I can see, the problem is not from an alloca, despite what the CVE
description says, but from the attack multiplication that comes from creating
many very small iovecs: (void* + size_t) for each three bytes of input message.
---

diff --git a/src/journal/journald-native.c b/src/journal/journald-native.c
index e86178ed775..d0fee2a797c 100644
--- a/src/journal/journald-native.c
+++ b/src/journal/journald-native.c
@@ -141,6 +141,11 @@ static int server_process_entry(
                 }
 
                 /* A property follows */
+                if (n > ENTRY_FIELD_COUNT_MAX) {
+                        log_debug("Received an entry that has more than " STRINGIFY(ENTRY_FIELD_COUNT_MAX) " fields, ignoring entry.");
+                        r = 1;
+                        goto finish;
+                }
 
                 /* n existing properties, 1 new, +1 for _TRANSPORT */
                 if (!GREEDY_REALLOC(iovec, m,
diff --git a/src/shared/journal-importer.h b/src/shared/journal-importer.h
index 53354b7c786..7914c0cf5fb 100644
--- a/src/shared/journal-importer.h
+++ b/src/shared/journal-importer.h
@@ -21,6 +21,9 @@
 #endif
 #define LINE_CHUNK 8*1024u
 
+/* The maximum number of fields in an entry */
+#define ENTRY_FIELD_COUNT_MAX 1024
+
 struct iovec_wrapper {
         struct iovec *iovec;
         size_t size_bytes;