From 31c1c7846f6b6e5848bc39a28a65118bfa98e35d Mon Sep 17 00:00:00 2001
From: Andrew Tridgell <tridge@samba.org>
Date: Wed, 1 Sep 2004 04:39:06 +0000
Subject: [PATCH] r2159: converted samba4 over to UTF-16.
I had previously thought this was unnecessary, as windows doesn't use
standards compliant UTF-16, and for filesystem operations treats bytes
as UCS-2, but Bjoern Jacke has pointed out to me that this means we
don't correctly store extended UTF-16 characters as UTF-8 on
disk. This can be seen with (for example) the gothic characters with
codepoints above 64k.
This commit also adds a LOCAL-ICONV torture test that tests the first
1 million codepoints against the system iconv library, and tests 5
million random UTF-16LE buffers for identical error handling to the
system iconv library.
the lib/iconv.c changes need backporting to samba3
(This used to be commit 756f28ac95feaa84b42402723d5f7286865c78db)
---
source4/include/charset.h | 2 +-
source4/lib/charcnv.c | 34 ++--
source4/lib/iconv.c | 264 +++++++++++++++++++++-------
source4/lib/util_unistr.c | 4 +-
source4/libcli/raw/rawrequest.c | 4 +-
source4/librpc/ndr/ndr_basic.c | 8 +-
source4/smb_server/request.c | 2 +-
source4/torture/basic/charset.c | 2 +-
source4/torture/basic/utable.c | 4 +-
source4/torture/config.m4 | 2 +
source4/torture/config.mk | 11 ++
source4/torture/local/iconv.c | 298 ++++++++++++++++++++++++++++++++
source4/torture/torture.c | 5 +-
13 files changed, 542 insertions(+), 98 deletions(-)
create mode 100644 source4/torture/local/iconv.c
diff --git a/source4/include/charset.h b/source4/include/charset.h
index daf10e23c13..f1482ac08c8 100644
--- a/source4/include/charset.h
+++ b/source4/include/charset.h
@@ -20,7 +20,7 @@
*/
/* this defines the charset types used in samba */
-typedef enum {CH_UCS2=0, CH_UNIX=1, CH_DISPLAY=2, CH_DOS=3, CH_UTF8=4, CH_UCS2BE=5} charset_t;
+typedef enum {CH_UTF16=0, CH_UNIX=1, CH_DISPLAY=2, CH_DOS=3, CH_UTF8=4, CH_UTF16BE=5} charset_t;
#define NUM_CHARSETS 6
diff --git a/source4/lib/charcnv.c b/source4/lib/charcnv.c
index 8204a3bcdd8..2109e957d92 100644
--- a/source4/lib/charcnv.c
+++ b/source4/lib/charcnv.c
@@ -47,12 +47,12 @@ static const char *charset_name(charset_t ch)
{
const char *ret = NULL;
- if (ch == CH_UCS2) ret = "UTF-16LE";
+ if (ch == CH_UTF16) ret = "UTF-16LE";
else if (ch == CH_UNIX) ret = lp_unix_charset();
else if (ch == CH_DOS) ret = lp_dos_charset();
else if (ch == CH_DISPLAY) ret = lp_display_charset();
else if (ch == CH_UTF8) ret = "UTF8";
- else if (ch == CH_UCS2BE) ret = "UCS-2BE";
+ else if (ch == CH_UTF16BE) ret = "UTF-16BE";
if (!ret || !*ret) ret = "ASCII";
return ret;
@@ -81,13 +81,13 @@ void init_iconv(void)
/* so that charset_name() works we need to get the UNIX<->UCS2 going
first */
- if (!conv_handles[CH_UNIX][CH_UCS2])
- conv_handles[CH_UNIX][CH_UCS2] = smb_iconv_open(charset_name(CH_UCS2),
+ if (!conv_handles[CH_UNIX][CH_UTF16])
+ conv_handles[CH_UNIX][CH_UTF16] = smb_iconv_open(charset_name(CH_UTF16),
"ASCII");
- if (!conv_handles[CH_UCS2][CH_UNIX])
- conv_handles[CH_UCS2][CH_UNIX] = smb_iconv_open("ASCII",
- charset_name(CH_UCS2));
+ if (!conv_handles[CH_UTF16][CH_UNIX])
+ conv_handles[CH_UTF16][CH_UNIX] = smb_iconv_open("ASCII",
+ charset_name(CH_UTF16));
for (c1=0;c1<NUM_CHARSETS;c1++) {
for (c2=0;c2<NUM_CHARSETS;c2++) {
@@ -293,7 +293,7 @@ size_t unix_strupper(const char *src, size_t srclen, char *dest, size_t destlen)
size_t size;
smb_ucs2_t *buffer;
- size = convert_string_allocate(CH_UNIX, CH_UCS2, src, srclen,
+ size = convert_string_allocate(CH_UNIX, CH_UTF16, src, srclen,
(void **) &buffer);
if (size == -1) {
smb_panic("failed to create UCS2 buffer");
@@ -303,7 +303,7 @@ size_t unix_strupper(const char *src, size_t srclen, char *dest, size_t destlen)
return srclen;
}
- size = convert_string(CH_UCS2, CH_UNIX, buffer, size, dest, destlen);
+ size = convert_string(CH_UTF16, CH_UNIX, buffer, size, dest, destlen);
free(buffer);
return size;
}
@@ -313,7 +313,7 @@ size_t unix_strlower(const char *src, size_t srclen, char *dest, size_t destlen)
size_t size;
smb_ucs2_t *buffer;
- size = convert_string_allocate(CH_UNIX, CH_UCS2, src, srclen,
+ size = convert_string_allocate(CH_UNIX, CH_UTF16, src, srclen,
(void **) &buffer);
if (size == -1) {
smb_panic("failed to create UCS2 buffer");
@@ -322,7 +322,7 @@ size_t unix_strlower(const char *src, size_t srclen, char *dest, size_t destlen)
free(buffer);
return srclen;
}
- size = convert_string(CH_UCS2, CH_UNIX, buffer, size, dest, destlen);
+ size = convert_string(CH_UTF16, CH_UNIX, buffer, size, dest, destlen);
free(buffer);
return size;
}
@@ -461,7 +461,7 @@ ssize_t push_ucs2(const void *base_ptr, void *dest, const char *src, size_t dest
/* ucs2 is always a multiple of 2 bytes */
dest_len &= ~1;
- len += convert_string(CH_UNIX, CH_UCS2, src, src_len, dest, dest_len);
+ len += convert_string(CH_UNIX, CH_UTF16, src, src_len, dest, dest_len);
return len;
}
@@ -480,7 +480,7 @@ ssize_t push_ucs2_talloc(TALLOC_CTX *ctx, smb_ucs2_t **dest, const char *src)
size_t src_len = strlen(src)+1;
*dest = NULL;
- return convert_string_talloc(ctx, CH_UNIX, CH_UCS2, src, src_len, (const void **)dest);
+ return convert_string_talloc(ctx, CH_UNIX, CH_UTF16, src, src_len, (const void **)dest);
}
@@ -498,7 +498,7 @@ ssize_t push_ucs2_allocate(smb_ucs2_t **dest, const char *src)
size_t src_len = strlen(src)+1;
*dest = NULL;
- return convert_string_allocate(CH_UNIX, CH_UCS2, src, src_len, (void **)dest);
+ return convert_string_allocate(CH_UNIX, CH_UTF16, src, src_len, (void **)dest);
}
/**
@@ -603,7 +603,7 @@ size_t pull_ucs2(const void *base_ptr, char *dest, const void *src, size_t dest_
if (src_len != (size_t)-1)
src_len &= ~1;
- ret = convert_string(CH_UCS2, CH_UNIX, src, src_len, dest, dest_len);
+ ret = convert_string(CH_UTF16, CH_UNIX, src, src_len, dest, dest_len);
if (dest_len)
dest[MIN(ret, dest_len-1)] = 0;
@@ -627,7 +627,7 @@ ssize_t pull_ucs2_talloc(TALLOC_CTX *ctx, char **dest, const smb_ucs2_t *src)
{
size_t src_len = (strlen_w(src)+1) * sizeof(smb_ucs2_t);
*dest = NULL;
- return convert_string_talloc(ctx, CH_UCS2, CH_UNIX, src, src_len, (const void **)dest);
+ return convert_string_talloc(ctx, CH_UTF16, CH_UNIX, src, src_len, (const void **)dest);
}
/**
@@ -642,7 +642,7 @@ ssize_t pull_ucs2_allocate(void **dest, const smb_ucs2_t *src)
{
size_t src_len = (strlen_w(src)+1) * sizeof(smb_ucs2_t);
*dest = NULL;
- return convert_string_allocate(CH_UCS2, CH_UNIX, src, src_len, dest);
+ return convert_string_allocate(CH_UTF16, CH_UNIX, src, src_len, dest);
}
/**
diff --git a/source4/lib/iconv.c b/source4/lib/iconv.c
index c869c253cf0..9f6526faa5e 100644
--- a/source4/lib/iconv.c
+++ b/source4/lib/iconv.c
@@ -39,8 +39,8 @@
*
* Note that the only function provided by iconv is conversion between
* characters. It doesn't directly support operations like
- * uppercasing or comparison. We have to convert to UCS-2 and compare
- * there.
+ * uppercasing or comparison. We have to convert to UTF-16LE and
+ * compare there.
*
* @sa Samba Developers Guide
**/
@@ -55,10 +55,11 @@ static size_t iconv_copy (void *,const char **, size_t *, char **, size_t *);
static size_t iconv_swab (void *,const char **, size_t *, char **, size_t *);
static const struct charset_functions const builtin_functions[] = {
- /* windows is really neither UCS-2 not UTF-16 */
+ /* windows is closest to UTF-16 */
{"UCS-2LE", iconv_copy, iconv_copy},
{"UTF-16LE", iconv_copy, iconv_copy},
{"UCS-2BE", iconv_swab, iconv_swab},
+ {"UTF-16BE", iconv_swab, iconv_swab},
/* we include the UTF-8 alias to cope with differing locale settings */
{"UTF8", utf8_pull, utf8_push},
@@ -217,23 +218,23 @@ smb_iconv_t smb_iconv_open(const char *tocode, const char *fromcode)
#endif
/* check for conversion to/from ucs2 */
- if (strcasecmp(fromcode, "UCS-2LE") == 0 && to) {
+ if (strcasecmp(fromcode, "UTF-16LE") == 0 && to) {
ret->direct = to->push;
return ret;
}
- if (strcasecmp(tocode, "UCS-2LE") == 0 && from) {
+ if (strcasecmp(tocode, "UTF-16LE") == 0 && from) {
ret->direct = from->pull;
return ret;
}
#ifdef HAVE_NATIVE_ICONV
- if (strcasecmp(fromcode, "UCS-2LE") == 0) {
+ if (strcasecmp(fromcode, "UTF-16LE") == 0) {
ret->direct = sys_iconv;
ret->cd_direct = ret->cd_push;
ret->cd_push = NULL;
return ret;
}
- if (strcasecmp(tocode, "UCS-2LE") == 0) {
+ if (strcasecmp(tocode, "UTF-16LE") == 0) {
ret->direct = sys_iconv;
ret->cd_direct = ret->cd_pull;
ret->cd_pull = NULL;
@@ -460,100 +461,231 @@ static size_t iconv_copy(void *cd, const char **inbuf, size_t *inbytesleft,
static size_t utf8_pull(void *cd, const char **inbuf, size_t *inbytesleft,
char **outbuf, size_t *outbytesleft)
{
- while (*inbytesleft >= 1 && *outbytesleft >= 2) {
- const uint8_t *c = (const uint8_t *)*inbuf;
- uint8_t *uc = (uint8_t *)*outbuf;
- int len = 1;
+ size_t in_left=*inbytesleft, out_left=*outbytesleft;
+ const uint8_t *c = (const uint8_t *)*inbuf;
+ uint8_t *uc = (uint8_t *)*outbuf;
+ while (in_left >= 1 && out_left >= 2) {
if ((c[0] & 0x80) == 0) {
uc[0] = c[0];
uc[1] = 0;
- } else if ((c[0] & 0xf0) == 0xe0) {
- if (*inbytesleft < 3) {
- DEBUG(0,("short utf8 char\n"));
- goto badseq;
- }
- uc[1] = ((c[0]&0xF)<<4) | ((c[1]>>2)&0xF);
- uc[0] = (c[1]<<6) | (c[2]&0x3f);
- len = 3;
- } else if ((c[0] & 0xe0) == 0xc0) {
- if (*inbytesleft < 2) {
- DEBUG(0,("short utf8 char\n"));
- goto badseq;
+ c += 1;
+ in_left -= 1;
+ out_left -= 2;
+ uc += 2;
+ continue;
+ }
+
+ if ((c[0] & 0xe0) == 0xc0) {
+ if (in_left < 2 ||
+ (c[1] & 0xc0) != 0x80) {
+ errno = EILSEQ;
+ goto error;
}
uc[1] = (c[0]>>2) & 0x7;
uc[0] = (c[0]<<6) | (c[1]&0x3f);
- len = 2;
+ c += 2;
+ in_left -= 2;
+ out_left -= 2;
+ uc += 2;
+ continue;
}
- (*inbuf) += len;
- (*inbytesleft) -= len;
- (*outbytesleft) -= 2;
- (*outbuf) += 2;
+ if ((c[0] & 0xf0) == 0xe0) {
+ if (in_left < 3 ||
+ (c[1] & 0xc0) != 0x80 ||
+ (c[2] & 0xc0) != 0x80) {
+ errno = EILSEQ;
+ goto error;
+ }
+ uc[1] = ((c[0]&0xF)<<4) | ((c[1]>>2)&0xF);
+ uc[0] = (c[1]<<6) | (c[2]&0x3f);
+ c += 3;
+ in_left -= 3;
+ out_left -= 2;
+ uc += 2;
+ continue;
+ }
+
+ if ((c[0] & 0xf8) == 0xf0) {
+ unsigned int codepoint;
+ if (in_left < 4 ||
+ (c[1] & 0xc0) != 0x80 ||
+ (c[2] & 0xc0) != 0x80 ||
+ (c[3] & 0xc0) != 0x80) {
+ errno = EILSEQ;
+ goto error;
+ }
+ codepoint =
+ (c[3]&0x3f) |
+ ((c[2]&0x3f)<<6) |
+ ((c[1]&0x3f)<<12) |
+ ((c[0]&0x7)<<18);
+ if (codepoint < 0x10000) {
+ /* accept UTF-8 characters that are not
+ minimally packed, but pack the result */
+ uc[0] = (codepoint & 0xFF);
+ uc[1] = (codepoint >> 8);
+ c += 4;
+ in_left -= 4;
+ out_left -= 2;
+ uc += 2;
+ continue;
+ }
+
+ codepoint -= 0x10000;
+
+ if (out_left < 4) {
+ errno = E2BIG;
+ goto error;
+ }
+
+ uc[0] = (codepoint>>10) & 0xFF;
+ uc[1] = (codepoint>>18) | 0xd8;
+ uc[2] = codepoint & 0xFF;
+ uc[3] = ((codepoint>>8) & 0x3) | 0xdc;
+ c += 4;
+ in_left -= 4;
+ out_left -= 4;
+ uc += 4;
+ continue;
+ }
+
+ /* we don't handle 5 byte sequences */
+ errno = EINVAL;
+ goto error;
}
- if (*inbytesleft > 0) {
+ if (in_left > 0) {
errno = E2BIG;
- return -1;
+ goto error;
}
-
+
+ *inbytesleft = in_left;
+ *outbytesleft = out_left;
+ *inbuf = c;
+ *outbuf = uc;
return 0;
-badseq:
- errno = EINVAL;
+error:
+ *inbytesleft = in_left;
+ *outbytesleft = out_left;
+ *inbuf = c;
+ *outbuf = uc;
return -1;
}
static size_t utf8_push(void *cd, const char **inbuf, size_t *inbytesleft,
- char **outbuf, size_t *outbytesleft)
+ char **outbuf, size_t *outbytesleft)
{
- while (*inbytesleft >= 2 && *outbytesleft >= 1) {
- uint8_t *c = (uint8_t *)*outbuf;
- const uint8_t *uc = (const uint8_t *)*inbuf;
- int len=1;
+ size_t in_left=*inbytesleft, out_left=*outbytesleft;
+ uint8_t *c = (uint8_t *)*outbuf;
+ const uint8_t *uc = (const uint8_t *)*inbuf;
- if (uc[1] & 0xf8) {
- if (*outbytesleft < 3) {
- DEBUG(0,("short utf8 write\n"));
- goto toobig;
- }
- c[0] = 0xe0 | (uc[1]>>4);
- c[1] = 0x80 | ((uc[1]&0xF)<<2) | (uc[0]>>6);
- c[2] = 0x80 | (uc[0]&0x3f);
- len = 3;
- } else if (uc[1] | (uc[0] & 0x80)) {
- if (*outbytesleft < 2) {
- DEBUG(0,("short utf8 write\n"));
- goto toobig;
- }
- c[0] = 0xc0 | (uc[1]<<2) | (uc[0]>>6);
- c[1] = 0x80 | (uc[0]&0x3f);
- len = 2;
- } else {
+ while (in_left >= 2 && out_left >= 1) {
+ unsigned int codepoint;
+
+ if (uc[1] == 0 && !(uc[0] & 0x80)) {
+ /* simplest case */
c[0] = uc[0];
+ in_left -= 2;
+ out_left -= 1;
+ uc += 2;
+ c += 1;
+ continue;
}
+ if ((uc[1]&0xf8) == 0) {
+ /* next simplest case */
+ if (out_left < 2) {
+ errno = E2BIG;
+ goto error;
+ }
+ c[0] = 0xc0 | (uc[0]>>6) | (uc[1]<<2);
+ c[1] = 0x80 | (uc[0] & 0x3f);
+ in_left -= 2;
+ out_left -= 2;
+ uc += 2;
+ c += 2;
+ continue;
+ }
- (*inbytesleft) -= 2;
- (*outbytesleft) -= len;
- (*inbuf) += 2;
- (*outbuf) += len;
+ if ((uc[1] & 0xfc) == 0xdc) {
+ /* its the second part of a 4 byte sequence. Illegal */
+ if (in_left < 4) {
+ errno = EINVAL;
+ } else {
+ errno = EILSEQ;
+ }
+ goto error;
+ }
+
+ if ((uc[1] & 0xfc) != 0xd8) {
+ codepoint = uc[0] | (uc[1]<<8);
+ if (out_left < 3) {
+ errno = E2BIG;
+ goto error;
+ }
+ c[0] = 0xe0 | (codepoint >> 12);
+ c[1] = 0x80 | ((codepoint >> 6) & 0x3f);
+ c[2] = 0x80 | (codepoint & 0x3f);
+
+ in_left -= 2;
+ out_left -= 3;
+ uc += 2;
+ c += 3;
+ continue;
+ }
+
+ /* its the first part of a 4 byte sequence */
+ if (in_left < 4) {
+ errno = EINVAL;
+ goto error;
+ }
+ if ((uc[3] & 0xfc) != 0xdc) {
+ errno = EILSEQ;
+ goto error;
+ }
+ codepoint = 0x10000 + (uc[2] | ((uc[3] & 0x3)<<8) |
+ (uc[0]<<10) | ((uc[1] & 0x3)<<18));
+
+ if (out_left < 4) {
+ errno = E2BIG;
+ goto error;
+ }
+ c[0] = 0xf0 | (codepoint >> 18);
+ c[1] = 0x80 | ((codepoint >> 12) & 0x3f);
+ c[2] = 0x80 | ((codepoint >> 6) & 0x3f);
+ c[3] = 0x80 | (codepoint & 0x3f);
+
+ in_left -= 4;
+ out_left -= 4;
+ uc += 4;
+ c += 4;
}
- if (*inbytesleft == 1) {
+ if (in_left == 1) {
errno = EINVAL;
- return -1;
+ goto error;
}
- if (*inbytesleft > 1) {
+ if (in_left > 1) {
errno = E2BIG;
- return -1;
+ goto error;
}
+
+ *inbytesleft = in_left;
+ *outbytesleft = out_left;
+ *inbuf = uc;
+ *outbuf = c;
return 0;
-toobig:
- errno = E2BIG;
+error:
+ *inbytesleft = in_left;
+ *outbytesleft = out_left;
+ *inbuf = uc;
+ *outbuf = c;
return -1;
}
diff --git a/source4/lib/util_unistr.c b/source4/lib/util_unistr.c
index 2bd990836ee..63d68fa12e3 100644
--- a/source4/lib/util_unistr.c
+++ b/source4/lib/util_unistr.c
@@ -96,9 +96,9 @@ static int check_dos_char(smb_ucs2_t c)
char buf[10];
smb_ucs2_t c2 = 0;
int len1, len2;
- len1 = convert_string(CH_UCS2, CH_DOS, &c, 2, buf, sizeof(buf));
+ len1 = convert_string(CH_UTF16, CH_DOS, &c, 2, buf, sizeof(buf));
if (len1 == 0) return 0;
- len2 = convert_string(CH_DOS, CH_UCS2, buf, len1, &c2, 2);
+ len2 = convert_string(CH_DOS, CH_UTF16, buf, len1, &c2, 2);
if (len2 != 2) return 0;
return (c == c2);
}
diff --git a/source4/libcli/raw/rawrequest.c b/source4/libcli/raw/rawrequest.c
index 87bbe5a31b2..51c0c0b925a 100644
--- a/source4/libcli/raw/rawrequest.c
+++ b/source4/libcli/raw/rawrequest.c
@@ -526,7 +526,7 @@ static size_t smbcli_req_pull_ucs2(struct smbcli_request *req, TALLOC_CTX *mem_c
return 0;
}
- ret = convert_string_talloc(mem_ctx, CH_UCS2, CH_UNIX, src, src_len2, (const void **)dest);
+ ret = convert_string_talloc(mem_ctx, CH_UTF16, CH_UNIX, src, src_len2, (const void **)dest);
if (ret == -1) {
*dest = NULL;
return 0;
@@ -725,7 +725,7 @@ static size_t smbcli_blob_pull_ucs2(TALLOC_CTX* mem_ctx,
src_len2 += 2;
}
- ret = convert_string_talloc(mem_ctx, CH_UCS2, CH_UNIX, src, src_len2, (const void **)dest);
+ ret = convert_string_talloc(mem_ctx, CH_UTF16, CH_UNIX, src, src_len2, (const void **)dest);
if (ret == -1) {
*dest = NULL;
return 0;
diff --git a/source4/librpc/ndr/ndr_basic.c b/source4/librpc/ndr/ndr_basic.c
index 34a1935c3b0..7f36f7e4ba6 100644
--- a/source4/librpc/ndr/ndr_basic.c
+++ b/source4/librpc/ndr/ndr_basic.c
@@ -421,14 +421,14 @@ NTSTATUS ndr_pull_string(struct ndr_pull *ndr, int ndr_flags, const char **s)
uint32_t len1, ofs, len2;
uint16_t len3;
int ret;
- int chset = CH_UCS2;
+ int chset = CH_UTF16;
if (!(ndr_flags & NDR_SCALARS)) {
return NT_STATUS_OK;
}
if (NDR_BE(ndr)) {
- chset = CH_UCS2BE;
+ chset = CH_UTF16BE;
}
switch (ndr->flags & LIBNDR_STRING_FLAGS) {
@@ -600,14 +600,14 @@ NTSTATUS ndr_push_string(struct ndr_push *ndr, int ndr_flags, const char *s)
{
ssize_t s_len, c_len;
int ret;
- int chset = CH_UCS2;
+ int chset = CH_UTF16;
if (!(ndr_flags & NDR_SCALARS)) {
return NT_STATUS_OK;
}
if (NDR_BE(ndr)) {
- chset = CH_UCS2BE;
+ chset = CH_UTF16BE;
}
s_len = s?strlen(s):0;
diff --git a/source4/smb_server/request.c b/source4/smb_server/request.c
index e3f54bde654..645ea105a2b 100644
--- a/source4/smb_server/request.c
+++ b/source4/smb_server/request.c
@@ -468,7 +468,7 @@ static size_t req_pull_ucs2(struct smbsrv_request *req, const char **dest, const
src_len2 += 2;
}
- ret = convert_string_talloc(req->mem_ctx, CH_UCS2, CH_UNIX, src, src_len2, (const void **)dest);
+ ret = convert_string_talloc(req->mem_ctx, CH_UTF16, CH_UNIX, src, src_len2, (const void **)dest);
if (ret == -1) {
*dest = NULL;
diff --git a/source4/torture/basic/charset.c b/source4/torture/basic/charset.c
index 330dcf1707b..879f20617fd 100644
--- a/source4/torture/basic/charset.c
+++ b/source4/torture/basic/charset.c
@@ -50,7 +50,7 @@ static NTSTATUS unicode_open(struct smbcli_tree *tree,
}
SSVAL(ucs_name, i*2, 0);
- i = convert_string_allocate(CH_UCS2, CH_UNIX, ucs_name, (1+u_name_len)*2, (void **)&fname);
+ i = convert_string_allocate(CH_UTF16, CH_UNIX, ucs_name, (1+u_name_len)*2, (void **)&fname);
if (i == -1) {
free(ucs_name);
return NT_STATUS_NO_MEMORY;
diff --git a/source4/torture/basic/utable.c b/source4/torture/basic/utable.c
index f98e1b1b4ad..6faf020ef9a 100644
--- a/source4/torture/basic/utable.c
+++ b/source4/torture/basic/utable.c
@@ -50,7 +50,7 @@ BOOL torture_utable(int dummy)
SSVAL(&c2, 0, c);
fstrcpy(fname, "\\utable\\x");
p = fname+strlen(fname);
- len = convert_string(CH_UCS2, CH_UNIX,
+ len = convert_string(CH_UTF16, CH_UNIX,
&c2, 2,
p, sizeof(fname)-strlen(fname));
p[len] = 0;
@@ -107,7 +107,7 @@ static char *form_name(int c)
p = fname+strlen(fname);
SSVAL(&c2, 0, c);
- len = convert_string(CH_UCS2, CH_UNIX,
+ len = convert_string(CH_UTF16, CH_UNIX,
&c2, 2,
p, sizeof(fname)-strlen(fname));
p[len] = 0;
diff --git a/source4/torture/config.m4 b/source4/torture/config.m4
index e5f1d357e64..fdd5be13551 100644
--- a/source4/torture/config.m4
+++ b/source4/torture/config.m4
@@ -10,6 +10,8 @@ SMB_SUBSYSTEM_MK(TORTURE_RAP,torture/config.mk)
SMB_SUBSYSTEM_MK(TORTURE_AUTH,torture/config.mk)
+SMB_SUBSYSTEM_MK(TORTURE_LOCAL,torture/config.mk)
+
SMB_SUBSYSTEM_MK(TORTURE_NBENCH,torture/config.mk)
SMB_SUBSYSTEM_MK(TORTURE_LDAP,torture/config.mk)
diff --git a/source4/torture/config.mk b/source4/torture/config.mk
index 6d1ddd4d5a7..43a42217f59 100644
--- a/source4/torture/config.mk
+++ b/source4/torture/config.mk
@@ -93,6 +93,16 @@ REQUIRED_SUBSYSTEMS = \
# End SUBSYSTEM TORTURE_AUTH
#################################
+#################################
+# Start SUBSYSTEM TORTURE_LOCAL
+[SUBSYSTEM::TORTURE_LOCAL]
+ADD_OBJ_FILES = \
+ torture/local/iconv.o
+REQUIRED_SUBSYSTEMS = \
+ LIBSMB
+# End SUBSYSTEM TORTURE_LOCAL
+#################################
+
#################################
# Start SUBSYSTEM TORTURE_NBENCH
[SUBSYSTEM::TORTURE_NBENCH]
@@ -125,6 +135,7 @@ REQUIRED_SUBSYSTEMS = \
TORTURE_RPC \
TORTURE_RAP \
TORTURE_AUTH \
+ TORTURE_LOCAL \
TORTURE_NBENCH \
TORTURE_LDAP \
CONFIG \
diff --git a/source4/torture/local/iconv.c b/source4/torture/local/iconv.c
new file mode 100644
index 00000000000..0867be0beef
--- /dev/null
+++ b/source4/torture/local/iconv.c
@@ -0,0 +1,298 @@
+/*
+ Unix SMB/CIFS implementation.
+
+ local testing of iconv routines. This tests the system iconv code against
+ the built-in iconv code
+
+ Copyright (C) Andrew Tridgell 2004
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software
+ Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+*/
+
+#include "includes.h"
+
+/*
+ generate a UTF-16LE buffer for a given unicode codepoint
+*/
+static int gen_codepoint(unsigned int codepoint,
+ char *buf, size_t *size)
+{
+ static iconv_t cd;
+ uint8_t in[4];
+ char *ptr_in;
+ size_t size_in, size_out, ret;
+ if (!cd) {
+ cd = iconv_open("UTF-16LE", "UCS-4LE");
+ }
+
+ in[0] = codepoint & 0xFF;
+ in[1] = (codepoint>>8) & 0xFF;
+ in[2] = (codepoint>>16) & 0xFF;
+ in[3] = (codepoint>>24) & 0xFF;
+
+ ptr_in = in;
+ size_in = 4;
+ size_out = 8;
+
+ ret = iconv(cd, &ptr_in, &size_in, &buf, &size_out);
+
+ *size = 8 - size_out;
+
+ return ret;
+}
+
+
+/*
+ work out the unicode codepoint of the first UTF-8 character in the buffer
+*/
+static unsigned int get_codepoint(char *buf, size_t size)
+{
+ static iconv_t cd;
+ uint8_t out[4];
+ char *ptr_out;
+ size_t size_out, size_in, ret;
+ if (!cd) {
+ cd = iconv_open("UCS-4LE", "UTF-8");
+ }
+
+ size_in = size;
+ ptr_out = out;
+ size_out = sizeof(out);
+ memset(out, 0, sizeof(out));
+
+ ret = iconv(cd, &buf, &size_in, &ptr_out, &size_out);
+
+ return out[0] | (out[1]<<8) | (out[2]<<16) | (out[3]<<24);
+}
+
+/*
+ display a buffer with name prefix
+*/
+static void show_buf(const char *name, uint8_t *buf, size_t size)
+{
+ int i;
+ printf("%s ", name);
+ for (i=0;i<size;i++) {
+ printf("%02x ", buf[i]);
+ }
+ printf("\n");
+}
+
+/*
+ given a UTF-16LE buffer, test the system and built-in iconv code
+ to make sure they do exactly the same thing in converting the buffer
+ to UTF-8, then convert it back again and ensure we get the same buffer
+ back
+*/
+static int test_buffer(uint8_t *inbuf, size_t size)
+{
+ uint8_t buf1[1000], buf2[1000], buf3[1000];
+ size_t outsize1, outsize2, outsize3;
+ const char *ptr_in;
+ char *ptr_out;
+ size_t size_in1, size_in2, size_in3;
+ size_t ret1, ret2, ret3, len1, len2;
+ int ok = 1;
+ int errno1, errno2;
+ static iconv_t cd;
+ static smb_iconv_t cd2, cd3;
+
+ if (!cd) {
+ cd = iconv_open("UTF-8", "UTF-16LE");
+ cd2 = smb_iconv_open("UTF-8", "UTF-16LE");
+ cd3 = smb_iconv_open("UTF-16LE", "UTF-8");
+ }
+
+#if 0
+ int i;
+ for (i=0;i<50;i++) {
+ ptr_in = inbuf;
+ ptr_out = buf1;
+ size_in1 = size;
+ outsize1 = sizeof(buf1);
+
+ memset(ptr_out, 0, outsize1);
+ errno = 0;
+ ret1 = smb_iconv(cd2, &ptr_in, &size_in1, &ptr_out, &outsize1);
+ errno1 = errno;
+ }
+#endif
+
+ ptr_in = inbuf;
+ ptr_out = buf1;
+ size_in1 = size;
+ outsize1 = sizeof(buf1);
+
+ memset(ptr_out, 0, outsize1);
+ errno = 0;
+ ret1 = smb_iconv(cd2, &ptr_in, &size_in1, &ptr_out, &outsize1);
+ errno1 = errno;
+
+ ptr_in = inbuf;
+ ptr_out = buf2;
+ size_in2 = size;
+ outsize2 = sizeof(buf2);
+
+ memset(ptr_out, 0, outsize2);
+ errno = 0;
+ ret2 = iconv(cd, &ptr_in, &size_in2, &ptr_out, &outsize2);
+ errno2 = errno;
+
+ len1 = sizeof(buf1) - outsize1;
+ len2 = sizeof(buf2) - outsize2;
+
+ /* codepoints above 1M are not interesting for now */
+ if (len2 > len1 &&
+ memcmp(buf1, buf2, len1) == 0 &&
+ get_codepoint(buf2+len1, len2-len1) >= (1<<20)) {
+ return ok;
+ }
+ if (len1 > len2 &&
+ memcmp(buf1, buf2, len2) == 0 &&
+ get_codepoint(buf1+len2, len1-len2) >= (1<<20)) {
+ return ok;
+ }
+
+ if (ret1 != ret2) {
+ printf("ret1=%d ret2=%d\n", ret1, ret2);
+ ok = 0;
+ }
+
+ if (errno1 != errno2) {
+ printf("e1=%s e2=%s\n", strerror(errno1), strerror(errno2));
+ show_buf(" rem1:", inbuf+(size-size_in1), size_in1);
+ show_buf(" rem2:", inbuf+(size-size_in2), size_in2);
+ ok = 0;
+ }
+
+ if (outsize1 != outsize2) {
+ printf("\noutsize mismatch outsize1=%d outsize2=%d\n",
+ outsize1, outsize2);
+ ok = 0;
+ }
+
+ if (size_in1 != size_in2) {
+ printf("\nsize_in mismatch size_in1=%d size_in2=%d\n",
+ size_in1, size_in2);
+ ok = 0;
+ }
+
+ if (!ok ||
+ len1 != len2 ||
+ memcmp(buf1, buf2, len1) != 0) {
+ printf("\nsize=%d ret1=%d ret2=%d\n", size, ret1, ret2);
+ show_buf(" IN1:", inbuf, size-size_in1);
+ show_buf(" IN2:", inbuf, size-size_in2);
+ show_buf("OUT1:", buf1, len1);
+ show_buf("OUT2:", buf2, len2);
+ if (len2 > len1 && memcmp(buf1, buf2, len1) == 0) {
+ printf("next codepoint is %u\n", get_codepoint(buf2+len1, len2-len1));
+ }
+ if (len1 > len2 && memcmp(buf1, buf2, len2) == 0) {
+ printf("next codepoint is %u\n", get_codepoint(buf1+len2,len1-len2));
+ }
+
+ ok = 0;
+ }
+
+ if (!ok) return ok;
+
+ size = size - size_in1;
+ ptr_in = buf1;
+ ptr_out = buf3;
+ size_in3 = len1;
+ outsize3 = sizeof(buf3);
+
+ memset(ptr_out, 0, outsize3);
+ ret3 = smb_iconv(cd3, &ptr_in, &size_in3, &ptr_out, &outsize3);
+
+ if (ret3 != 0) {
+ printf("pull failed - %s\n", strerror(errno));
+ ok = 0;
+ }
+
+ if (outsize3 != sizeof(buf3) - size) {
+ printf("wrong outsize3 - %d should be %d\n",
+ outsize3, sizeof(buf3) - size);
+ ok = 0;
+ }
+
+ if (memcmp(buf3, inbuf, size) != 0) {
+ int i;
+ printf("pull bytes mismatch:\n");
+ for (i=0;i<size;i++) {
+ printf("%02x ", inbuf[i]);
+ }
+ printf("\n");
+ for (i=0;i<size;i++) {
+ printf("%02x ", buf3[i]);
+ }
+ printf("\n");
+ ok = 0;
+ }
+
+ return ok;
+}
+
+BOOL torture_local_iconv(int dummy)
+{
+ size_t size;
+ unsigned char inbuf[1000];
+ int ok = 1;
+ unsigned int codepoint, i, c;
+
+ srandom(time(NULL));
+
+ printf("Testing first 1M codepoints\n");
+ for (codepoint=0;ok && codepoint<(1<<20);codepoint++) {
+ if (gen_codepoint(codepoint, inbuf, &size) != 0) {
+ continue;
+ }
+
+ if (codepoint % 100 == 0) {
+ printf("codepoint=%u \r", codepoint);
+ }
+
+ ok = test_buffer(inbuf, size);
+ }
+
+
+ printf("Testing 5M random UTF-16LE sequences\n");
+ for (i=0;ok && i<500000;i++) {
+ if (i % 100 == 0) {
+ printf("i=%u \r", i);
+ }
+
+ size = random() % 100;
+ for (c=0;c<size;c++) {
+ if (random() % 100 < 80) {
+ inbuf[c] = random() % 128;
+ } else {
+ inbuf[c] = random();
+ }
+ if (random() % 10 == 0) {
+ inbuf[c] |= 0xd8;
+ }
+ if (random() % 10 == 0) {
+ inbuf[c] |= 0xdc;
+ }
+ }
+ ok = test_buffer(inbuf, size);
+ }
+
+ return ok == 1;
+}
+
+
diff --git a/source4/torture/torture.c b/source4/torture/torture.c
index 9488c987eca..e3a7d8e5d88 100644
--- a/source4/torture/torture.c
+++ b/source4/torture/torture.c
@@ -4212,8 +4212,9 @@ static struct {
{"RPC-MULTIBIND", torture_multi_bind, 0},
{"RPC-DRSUAPI", torture_rpc_drsuapi, 0},
- /* crypto testers */
- {"CRYPT-NTLMSSP", torture_ntlmssp_self_check, 0},
+ /* local (no server) testers */
+ {"LOCAL-NTLMSSP", torture_ntlmssp_self_check, 0},
+ {"LOCAL-ICONV", torture_local_iconv, 0},
/* ldap testers */
{"LDAP-BASIC", torture_ldap_basic, 0},