Add PEM_read_bio_ex

The extended function includes a 'flags' argument to allow callers
to specify different requested behaviors.  In particular, callers can
request that temporary storage buffers are allocated from the secure heap,
which could be relevant when loading private key material.

Refactor PEM_read_bio to use BIO_mems instead of BUFs directly,
use some helper routines to reduce the overall function length, and make
some of the checks more reasonable.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1700)

1 files changed, 270 insertions, 141 deletions

diff --git a/crypto/pem/pem_lib.c b/crypto/pem/pem_lib.c
index 3f53fd892d..24320131a4 100644
--- a/crypto/pem/pem_lib.c
+++ b/crypto/pem/pem_lib.c
@@ -228,6 +228,20 @@ static int check_pem(const char *nm, const char *name)
     return 0;
 }
 
+static void pem_free(void *p, unsigned int flags)
+{
+    if (flags & PEM_FLAG_SECURE)
+        OPENSSL_secure_free(p);
+    else
+        OPENSSL_free(p);
+}
+
+static void *pem_malloc(int num, unsigned int flags)
+{
+    return (flags & PEM_FLAG_SECURE) ? OPENSSL_secure_malloc(num)
+                                     : OPENSSL_malloc(num);
+}
+
 int PEM_bytes_read_bio(unsigned char **pdata, long *plen, char **pnm,
                        const char *name, BIO *bp, pem_password_cb *cb,
                        void *u)
@@ -661,177 +675,292 @@ int PEM_read(FILE *fp, char **name, char **header, unsigned char **data,
 }
 #endif
 
-int PEM_read_bio(BIO *bp, char **name, char **header, unsigned char **data,
-                 long *len)
+/* Some helpers for PEM_read_bio_ex(). */
+
+#define isb64(c) (isalnum(c) || (c) == '+' || (c) == '/' || (c) == '=')
+
+static int sanitize_line(char *linebuf, int len, unsigned int flags)
 {
-    EVP_ENCODE_CTX *ctx = EVP_ENCODE_CTX_new();
-    int end = 0, i, k, bl = 0, hl = 0, nohead = 0;
-    char buf[256];
-    BUF_MEM *nameB;
-    BUF_MEM *headerB;
-    BUF_MEM *dataB, *tmpB;
+    int i;
 
-    if (ctx == NULL) {
-        PEMerr(PEM_F_PEM_READ_BIO, ERR_R_MALLOC_FAILURE);
-        return (0);
+    if (flags & PEM_FLAG_EAY_COMPATIBLE) {
+        /* Strip trailing whitespace */
+        while ((len >= 0) && (linebuf[len] <= ' '))
+            len--;
+        /* Go back to whitespace before applying uniform line ending. */
+        len++;
+    } else if (flags & PEM_FLAG_ONLY_B64) {
+        for (i = 0; i < len; ++i) {
+            if (!isb64(linebuf[i]) || linebuf[i] == '\n' || linebuf[i] == '\r')
+                break;
+        }
+        len = i;
+    } else {
+        /* EVP_DecodeBlock strips leading and trailing whitespace, so just strip
+         * control characters in-place and let everything through. */
+        for (i = 0; i < len; ++i) {
+            if (linebuf[i] == '\n' || linebuf[i] == '\r')
+                break;
+            if (iscntrl(linebuf[i]))
+                linebuf[i] = ' ';
+        }
+        len = i;
     }
+    /* The caller allocated LINESIZE+1, so this is safe. */
+    linebuf[len++] = '\n';
+    linebuf[len] = '\0';
+    return len;
+}
 
-    nameB = BUF_MEM_new();
-    headerB = BUF_MEM_new();
-    dataB = BUF_MEM_new();
-    if ((nameB == NULL) || (headerB == NULL) || (dataB == NULL)) {
-        goto err;
+#define LINESIZE 255
+/* Note trailing spaces for begin and end. */
+static const char beginstr[] = "-----BEGIN ";
+static const char endstr[] = "-----END ";
+static const char tailstr[] = "-----\n";
+#define BEGINLEN (sizeof(beginstr) - 1)
+#define ENDLEN (sizeof(endstr) - 1)
+#define TAILLEN (sizeof(tailstr) - 1)
+static int get_name(BIO *bp, char **name, unsigned int flags)
+{
+    char *linebuf;
+    int ret = 0;
+    size_t len;
+
+    /*
+     * Need to hold trailing NUL (accounted for by BIO_gets() and the newline
+     * that will be added by sanitize_line() (the extra '1').
+     */
+    linebuf = pem_malloc(LINESIZE + 1, flags);
+    if (linebuf == NULL) {
+        PEMerr(PEM_F_GET_NAME, ERR_R_MALLOC_FAILURE);
+        return 0;
     }
 
-    buf[254] = '\0';
-    for (;;) {
-        i = BIO_gets(bp, buf, 254);
+    do {
+        len = BIO_gets(bp, linebuf, LINESIZE);
 
-        if (i <= 0) {
-            PEMerr(PEM_F_PEM_READ_BIO, PEM_R_NO_START_LINE);
+        if (len <= 0) {
+            PEMerr(PEM_F_GET_NAME, PEM_R_NO_START_LINE);
             goto err;
         }
 
-        while ((i >= 0) && (buf[i] <= ' '))
-            i--;
-        buf[++i] = '\n';
-        buf[++i] = '\0';
+        /* Strip trailing garbage and standardize ending. */
+        len = sanitize_line(linebuf, len, flags & ~PEM_FLAG_ONLY_B64);
+
+        /* Allow leading empty or non-matching lines. */
+    } while (strncmp(linebuf, beginstr, BEGINLEN) != 0
+             || len < TAILLEN
+             || strncmp(linebuf + len - TAILLEN, tailstr, TAILLEN) != 0);
+    linebuf[len - TAILLEN] = '\0';
+    len = len - BEGINLEN - TAILLEN + 1;
+    *name = pem_malloc(len, flags);
+    if (*name == NULL) {
+        PEMerr(PEM_F_GET_NAME, ERR_R_MALLOC_FAILURE);
+        goto err;
+    }
+    memcpy(*name, linebuf + BEGINLEN, len);
+    ret = 1;
+
+err:
+    pem_free(linebuf, flags);
+    return ret;
+}
+
+/* Keep track of how much of a header we've seen. */
+enum header_status {
+    MAYBE_HEADER,
+    IN_HEADER,
+    POST_HEADER
+};
+
+/**
+ * Extract the optional PEM header, with details on the type of content and
+ * any encryption used on the contents, and the bulk of the data from the bio.
+ * The end of the header is marked by a blank line; if the end-of-input marker
+ * is reached prior to a blank line, there is no header.
+ *
+ * The header and data arguments are BIO** since we may have to swap them
+ * if there is no header, for efficiency.
+ *
+ * We need the name of the PEM-encoded type to verify the end string.
+ */
+static int get_header_and_data(BIO *bp, BIO **header, BIO **data, char *name,
+                               unsigned int flags)
+{
+    BIO *tmp = *header;
+    char *linebuf, *p;
+    int len, line, ret = 0, end = 0;
+    /* 0 if not seen (yet), 1 if reading header, 2 if finished header */
+    enum header_status got_header = MAYBE_HEADER;
+    unsigned int flags_mask;
+    size_t namelen;
+
+    /* Need to hold trailing NUL (accounted for by BIO_gets() and the newline
+     * that will be added by sanitize_line() (the extra '1'). */
+    linebuf = pem_malloc(LINESIZE + 1, flags);
+    if (linebuf == NULL) {
+        PEMerr(PEM_F_GET_HEADER_AND_DATA, ERR_R_MALLOC_FAILURE);
+        return 0;
+    }
 
-        if (strncmp(buf, "-----BEGIN ", 11) == 0) {
-            i = strlen(&(buf[11]));
+    for (line = 0; ; line++) {
+        flags_mask = ~0u;
+        len = BIO_gets(bp, linebuf, LINESIZE);
+        if (len <= 0) {
+            PEMerr(PEM_F_GET_HEADER_AND_DATA, PEM_R_SHORT_HEADER);
+            goto err;
+        }
 
-            if (strncmp(&(buf[11 + i - 6]), "-----\n", 6) != 0)
-                continue;
-            if (!BUF_MEM_grow(nameB, i + 9)) {
-                PEMerr(PEM_F_PEM_READ_BIO, ERR_R_MALLOC_FAILURE);
+        if (got_header == MAYBE_HEADER) {
+            if (memchr(linebuf, ':', len) != NULL)
+                got_header = IN_HEADER;
+        }
+        if (!strncmp(linebuf, endstr, ENDLEN) || got_header == IN_HEADER)
+            flags_mask &= ~PEM_FLAG_ONLY_B64;
+        len = sanitize_line(linebuf, len, flags & flags_mask);
+
+        /* Check for end of header. */
+        if (linebuf[0] == '\n') {
+            if (got_header == POST_HEADER) {
+                /* Another blank line is an error. */
+                PEMerr(PEM_F_GET_HEADER_AND_DATA, PEM_R_BAD_END_LINE);
                 goto err;
             }
-            memcpy(nameB->data, &(buf[11]), i - 6);
-            nameB->data[i - 6] = '\0';
-            break;
+            got_header = POST_HEADER;
+            tmp = *data;
+            continue;
         }
-    }
-    hl = 0;
-    if (!BUF_MEM_grow(headerB, 256)) {
-        PEMerr(PEM_F_PEM_READ_BIO, ERR_R_MALLOC_FAILURE);
-        goto err;
-    }
-    headerB->data[0] = '\0';
-    for (;;) {
-        i = BIO_gets(bp, buf, 254);
-        if (i <= 0)
-            break;
-
-        while ((i >= 0) && (buf[i] <= ' '))
-            i--;
-        buf[++i] = '\n';
-        buf[++i] = '\0';
 
-        if (buf[0] == '\n')
+        /* Check for end of stream (which means there is no header). */
+        if (strncmp(linebuf, endstr, ENDLEN) == 0) {
+            p = linebuf + ENDLEN;
+            namelen = strlen(name);
+            if (strncmp(p, name, namelen) != 0 ||
+                strncmp(p + namelen, tailstr, TAILLEN) != 0) {
+                PEMerr(PEM_F_GET_HEADER_AND_DATA, PEM_R_BAD_END_LINE);
+                goto err;
+            }
+            if (got_header == MAYBE_HEADER) {
+                *header = *data;
+                *data = tmp;
+            }
             break;
-        if (!BUF_MEM_grow(headerB, hl + i + 9)) {
-            PEMerr(PEM_F_PEM_READ_BIO, ERR_R_MALLOC_FAILURE);
+        } else if (end) {
+            /* Malformed input; short line not at end of data. */
+            PEMerr(PEM_F_GET_HEADER_AND_DATA, PEM_R_BAD_END_LINE);
             goto err;
         }
-        if (strncmp(buf, "-----END ", 9) == 0) {
-            nohead = 1;
-            break;
+        /*
+         * Else, a line of text -- could be header or data; we don't
+         * know yet.  Just pass it through.
+         */
+        BIO_puts(tmp, linebuf);
+        /*
+         * Only encrypted files need the line length check applied.
+         */
+        if (got_header == POST_HEADER) {
+            /* 65 includes the trailing newline */
+            if (len > 65)
+                goto err;
+            if (len < 65)
+                end = 1;
         }
-        memcpy(&(headerB->data[hl]), buf, i);
-        headerB->data[hl + i] = '\0';
-        hl += i;
-    }
-
-    bl = 0;
-    if (!BUF_MEM_grow(dataB, 1024)) {
-        PEMerr(PEM_F_PEM_READ_BIO, ERR_R_MALLOC_FAILURE);
-        goto err;
     }
-    dataB->data[0] = '\0';
-    if (!nohead) {
-        for (;;) {
-            i = BIO_gets(bp, buf, 254);
-            if (i <= 0)
-                break;
 
-            while ((i >= 0) && (buf[i] <= ' '))
-                i--;
-            buf[++i] = '\n';
-            buf[++i] = '\0';
+    ret = 1;
+err:
+    pem_free(linebuf, flags);
+    return ret;
+}
 
-            if (i != 65)
-                end = 1;
-            if (strncmp(buf, "-----END ", 9) == 0)
-                break;
-            if (i > 65)
-                break;
-            if (!BUF_MEM_grow_clean(dataB, i + bl + 9)) {
-                PEMerr(PEM_F_PEM_READ_BIO, ERR_R_MALLOC_FAILURE);
-                goto err;
-            }
-            memcpy(&(dataB->data[bl]), buf, i);
-            dataB->data[bl + i] = '\0';
-            bl += i;
-            if (end) {
-                buf[0] = '\0';
-                i = BIO_gets(bp, buf, 254);
-                if (i <= 0)
-                    break;
-
-                while ((i >= 0) && (buf[i] <= ' '))
-                    i--;
-                buf[++i] = '\n';
-                buf[++i] = '\0';
+/**
+ * Read in PEM-formatted data from the given BIO.
+ *
+ * By nature of the PEM format, all content must be printable ASCII (except
+ * for line endings).  Other characters, or lines that are longer than 80
+ * characters, are malformed input and will be rejected.
+ */
+int PEM_read_bio_ex(BIO *bp, char **name_out, char **header,
+                    unsigned char **data, long *len_out, unsigned int flags)
+{
+    EVP_ENCODE_CTX *ctx = EVP_ENCODE_CTX_new();
+    const BIO_METHOD *bmeth;
+    BIO *headerB = NULL, *dataB = NULL;
+    char *name = NULL;
+    int len, taillen, headerlen, ret = 0;
+    BUF_MEM * buf_mem;
 
-                break;
-            }
-        }
-    } else {
-        tmpB = headerB;
-        headerB = dataB;
-        dataB = tmpB;
-        bl = hl;
-    }
-    i = strlen(nameB->data);
-    if ((strncmp(buf, "-----END ", 9) != 0) ||
-        (strncmp(nameB->data, &(buf[9]), i) != 0) ||
-        (strncmp(&(buf[9 + i]), "-----\n", 6) != 0)) {
-        PEMerr(PEM_F_PEM_READ_BIO, PEM_R_BAD_END_LINE);
-        goto err;
+    if (ctx == NULL) {
+        PEMerr(PEM_F_PEM_READ_BIO_EX, ERR_R_MALLOC_FAILURE);
+        return 0;
     }
 
-    EVP_DecodeInit(ctx);
-    i = EVP_DecodeUpdate(ctx,
-                         (unsigned char *)dataB->data, &bl,
-                         (unsigned char *)dataB->data, bl);
-    if (i < 0) {
-        PEMerr(PEM_F_PEM_READ_BIO, PEM_R_BAD_BASE64_DECODE);
-        goto err;
+    *len_out = 0;
+    *name_out = *header = NULL;
+    *data = NULL;
+    if ((flags & PEM_FLAG_EAY_COMPATIBLE) && (flags & PEM_FLAG_ONLY_B64)) {
+        /* These two are mutually incompatible; bail out. */
+        PEMerr(PEM_F_PEM_READ_BIO_EX, ERR_R_PASSED_INVALID_ARGUMENT);
+        goto end;
     }
-    i = EVP_DecodeFinal(ctx, (unsigned char *)&(dataB->data[bl]), &k);
-    if (i < 0) {
-        PEMerr(PEM_F_PEM_READ_BIO, PEM_R_BAD_BASE64_DECODE);
-        goto err;
+    bmeth = (flags & PEM_FLAG_SECURE) ? BIO_s_secmem() : BIO_s_mem();
+
+    headerB = BIO_new(bmeth);
+    dataB = BIO_new(bmeth);
+    if (headerB == NULL || dataB == NULL) {
+        PEMerr(PEM_F_PEM_READ_BIO_EX, ERR_R_MALLOC_FAILURE);
+        goto end;
     }
-    bl += k;
 
-    if (bl == 0)
-        goto err;
-    *name = nameB->data;
-    *header = headerB->data;
-    *data = (unsigned char *)dataB->data;
-    *len = bl;
-    OPENSSL_free(nameB);
-    OPENSSL_free(headerB);
-    OPENSSL_free(dataB);
-    EVP_ENCODE_CTX_free(ctx);
-    return (1);
- err:
-    BUF_MEM_free(nameB);
-    BUF_MEM_free(headerB);
-    BUF_MEM_free(dataB);
+    if (!get_name(bp, &name, flags))
+        goto end;
+    if (!get_header_and_data(bp, &headerB, &dataB, name, flags))
+        goto end;
+
+    EVP_DecodeInit(ctx);
+    BIO_get_mem_ptr(dataB, &buf_mem);
+    len = buf_mem->length;
+    if (EVP_DecodeUpdate(ctx, (unsigned char*)buf_mem->data, &len,
+                         (unsigned char*)buf_mem->data, len) < 0
+            || EVP_DecodeFinal(ctx, (unsigned char*)&(buf_mem->data[len]),
+                               &taillen) < 0) {
+        PEMerr(PEM_F_PEM_READ_BIO_EX, PEM_R_BAD_BASE64_DECODE);
+        goto end;
+    }
+    len += taillen;
+    buf_mem->length = len;
+
+    /* There was no data in the PEM file; avoid malloc(0). */
+    if (len == 0)
+        goto end;
+    headerlen = BIO_get_mem_data(headerB, NULL);
+    *header = pem_malloc(headerlen + 1, flags);
+    *data = pem_malloc(len, flags);
+    if (*header == NULL || *data == NULL) {
+        pem_free(*header, flags);
+        pem_free(*data, flags);
+        goto end;
+    }
+    BIO_read(headerB, *header, headerlen);
+    (*header)[headerlen] = '\0';
+    BIO_read(dataB, *data, len);
+    *len_out = len;
+    *name_out = name;
+    name = NULL;
+    ret = 1;
+
+end:
     EVP_ENCODE_CTX_free(ctx);
-    return (0);
+    pem_free(name, flags);
+    BIO_free(headerB);
+    BIO_free(dataB);
+    return ret;
+}
+
+int PEM_read_bio(BIO *bp, char **name, char **header, unsigned char **data,
+                 long *len)
+{
+    return PEM_read_bio_ex(bp, name, header, data, len, PEM_FLAG_EAY_COMPATIBLE);
 }
 
 /*