From c0fa8ea4b9a3fc0043e95622dcc4bf0a96199d57 Mon Sep 17 00:00:00 2001 From: Maks Verver Date: Mon, 9 Jun 2025 20:09:17 +0100 Subject: [PATCH] [CVE-2025-32414] python: Read at most len/4 characters. Fixes #889 by reserving space in the buffer for UTF-8 encoding of text. --- python/libxml.c | 28 ++++++++++++++++++---------- 1 file changed, 18 insertions(+), 10 deletions(-) diff --git a/python/libxml.c b/python/libxml.c index ef630254..f31e080f 100644 --- a/python/libxml.c +++ b/python/libxml.c @@ -287,7 +287,9 @@ xmlPythonFileReadRaw (void * context, char * buffer, int len) { #endif file = (PyObject *) context; if (file == NULL) return(-1); - ret = PyEval_CallMethod(file, (char *) "read", (char *) "(i)", len); + /* When read() returns a string, the length is in characters not bytes, so + request at most len / 4 characters to leave space for UTF-8 encoding. */ + ret = PyEval_CallMethod(file, (char *) "read", (char *) "(i)", len / 4); if (ret == NULL) { printf("xmlPythonFileReadRaw: result is NULL\n"); return(-1); @@ -322,10 +324,12 @@ xmlPythonFileReadRaw (void * context, char * buffer, int len) { Py_DECREF(ret); return(-1); } - if (lenread > len) - memcpy(buffer, data, len); - else - memcpy(buffer, data, lenread); + if (lenread < 0 || lenread > len) { + printf("xmlPythonFileReadRaw: invalid lenread\n"); + Py_DECREF(ret); + return(-1); + } + memcpy(buffer, data, lenread); Py_DECREF(ret); return(lenread); } @@ -352,7 +356,9 @@ xmlPythonFileRead (void * context, char * buffer, int len) { #endif file = (PyObject *) context; if (file == NULL) return(-1); - ret = PyEval_CallMethod(file, (char *) "io_read", (char *) "(i)", len); + /* When read() returns a string, the length is in characters not bytes, so + request at most len / 4 characters to leave space for UTF-8 encoding. */ + ret = PyEval_CallMethod(file, (char *) "io_read", (char *) "(i)", len / 4); if (ret == NULL) { printf("xmlPythonFileRead: result is NULL\n"); return(-1); @@ -387,10 +393,12 @@ xmlPythonFileRead (void * context, char * buffer, int len) { Py_DECREF(ret); return(-1); } - if (lenread > len) - memcpy(buffer, data, len); - else - memcpy(buffer, data, lenread); + if (lenread < 0 || lenread > len) { + printf("xmlPythonFileRead: invalid lenread\n"); + Py_DECREF(ret); + return(-1); + } + memcpy(buffer, data, lenread); Py_DECREF(ret); return(lenread); } -- 2.49.0