yeahuh / rpms / qemu-kvm

Forked from rpms/qemu-kvm 2 years ago
Clone

Blame SOURCES/kvm-json-parser-fix-handling-of-large-whole-number-value.patch

0a122b
From 2263af4c1921520ddc9177c13a176860c0069fab Mon Sep 17 00:00:00 2001
0a122b
Message-Id: <2263af4c1921520ddc9177c13a176860c0069fab.1387369730.git.minovotn@redhat.com>
0a122b
In-Reply-To: <091eecc4fa42754760dfff393dabcc2b444e9693.1387369730.git.minovotn@redhat.com>
0a122b
References: <091eecc4fa42754760dfff393dabcc2b444e9693.1387369730.git.minovotn@redhat.com>
0a122b
From: Markus Armbruster <armbru@redhat.com>
0a122b
Date: Tue, 10 Dec 2013 15:29:04 +0100
0a122b
Subject: [PATCH 04/21] json-parser: fix handling of large whole number values
0a122b
0a122b
RH-Author: Markus Armbruster <armbru@redhat.com>
0a122b
Message-id: <1386689361-30281-2-git-send-email-armbru@redhat.com>
0a122b
Patchwork-id: 56120
0a122b
O-Subject: [PATCH 7.0 qemu-kvm 01/18] json-parser: fix handling of large whole number values
0a122b
Bugzilla: 997915
0a122b
RH-Acked-by: Laszlo Ersek <lersek@redhat.com>
0a122b
RH-Acked-by: Kevin Wolf <kwolf@redhat.com>
0a122b
RH-Acked-by: Luiz Capitulino <lcapitulino@redhat.com>
0a122b
0a122b
From: Michael Roth <mdroth@linux.vnet.ibm.com>
0a122b
0a122b
Currently our JSON parser assumes that numbers lacking a fractional
0a122b
value are integers and attempts to store them as QInt/int64 values. This
0a122b
breaks in the case where the number overflows/underflows int64 values (which
0a122b
is still valid JSON)
0a122b
0a122b
Fix this by detecting such cases and using a QFloat to store the value
0a122b
instead.
0a122b
0a122b
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
0a122b
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
0a122b
Reviewed-by: Amos Kong <akong@redhat.com>
0a122b
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
0a122b
(cherry picked from commit 3d5b3ec6d460a92245215aaf7b349b0b9e5ffa25)
0a122b
Signed-off-by: Markus Armbruster <armbru@redhat.com>
0a122b
---
0a122b
 qobject/json-parser.c | 26 +++++++++++++++++++++++---
0a122b
 1 file changed, 23 insertions(+), 3 deletions(-)
0a122b
0a122b
Signed-off-by: Michal Novotny <minovotn@redhat.com>
0a122b
---
0a122b
 qobject/json-parser.c | 26 +++++++++++++++++++++++---
0a122b
 1 file changed, 23 insertions(+), 3 deletions(-)
0a122b
0a122b
diff --git a/qobject/json-parser.c b/qobject/json-parser.c
0a122b
index 05279c1..e7947b3 100644
0a122b
--- a/qobject/json-parser.c
0a122b
+++ b/qobject/json-parser.c
0a122b
@@ -640,9 +640,29 @@ static QObject *parse_literal(JSONParserContext *ctxt)
0a122b
     case JSON_STRING:
0a122b
         obj = QOBJECT(qstring_from_escaped_str(ctxt, token));
0a122b
         break;
0a122b
-    case JSON_INTEGER:
0a122b
-        obj = QOBJECT(qint_from_int(strtoll(token_get_value(token), NULL, 10)));
0a122b
-        break;
0a122b
+    case JSON_INTEGER: {
0a122b
+        /* A possibility exists that this is a whole-valued float where the
0a122b
+         * fractional part was left out due to being 0 (.0). It's not a big
0a122b
+         * deal to treat these as ints in the parser, so long as users of the
0a122b
+         * resulting QObject know to expect a QInt in place of a QFloat in
0a122b
+         * cases like these.
0a122b
+         *
0a122b
+         * However, in some cases these values will overflow/underflow a
0a122b
+         * QInt/int64 container, thus we should assume these are to be handled
0a122b
+         * as QFloats/doubles rather than silently changing their values.
0a122b
+         *
0a122b
+         * strtoll() indicates these instances by setting errno to ERANGE
0a122b
+         */
0a122b
+        int64_t value;
0a122b
+
0a122b
+        errno = 0; /* strtoll doesn't set errno on success */
0a122b
+        value = strtoll(token_get_value(token), NULL, 10);
0a122b
+        if (errno != ERANGE) {
0a122b
+            obj = QOBJECT(qint_from_int(value));
0a122b
+            break;
0a122b
+        }
0a122b
+        /* fall through to JSON_FLOAT */
0a122b
+    }
0a122b
     case JSON_FLOAT:
0a122b
         /* FIXME dependent on locale */
0a122b
         obj = QOBJECT(qfloat_from_double(strtod(token_get_value(token), NULL)));
0a122b
-- 
0a122b
1.7.11.7
0a122b