Add unicode support to msdosfs and smbfs; original pathes from imura,

bug fixes by Kuan-Chung Chiu <buganini at gmail dot com>.

Tested by me in production for several days at work.

3 files changed, 109 insertions, 75 deletions

diff --git a/sys/fs/msdosfs/msdosfs_conv.c b/sys/fs/msdosfs/msdosfs_conv.c
index 8dc1d07..fde6512 100644
--- a/sys/fs/msdosfs/msdosfs_conv.c
+++ b/sys/fs/msdosfs/msdosfs_conv.c
@@ -61,9 +61,9 @@
 extern struct iconv_functions *msdosfs_iconv;
 
 static int mbsadjpos(const char **, size_t, size_t, int, int, void *handle);
-static u_int16_t dos2unixchr(const u_char **, size_t *, int, struct msdosfsmount *);
+static u_char * dos2unixchr(const u_char **, size_t *, int, struct msdosfsmount *);
 static u_int16_t unix2doschr(const u_char **, size_t *, struct msdosfsmount *);
-static u_int16_t win2unixchr(u_int16_t, struct msdosfsmount *);
+static u_char * win2unixchr(u_int16_t, struct msdosfsmount *);
 static u_int16_t unix2winchr(const u_char **, size_t *, int, struct msdosfsmount *);
 
 /*
@@ -242,7 +242,7 @@ dos2unixfn(dn, un, lower, pmp)
 {
 	size_t i;
 	int thislong = 0;
-	u_int16_t c;
+	u_char *c;
 
 	/*
 	 * If first char of the filename is SLOT_E5 (0x05), then the real
@@ -259,12 +259,10 @@ dos2unixfn(dn, un, lower, pmp)
 	for (i = 8; i > 0 && *dn != ' ';) {
 		c = dos2unixchr((const u_char **)&dn, &i, lower & LCASE_BASE,
 		    pmp);
-		if (c & 0xff00) {
-			*un++ = c >> 8;
+		while (*c != '\0') {
+			*un++ = *c++;
 			thislong++;
 		}
-		*un++ = c;
-		thislong++;
 	}
 	dn += i;
 
@@ -278,12 +276,10 @@ dos2unixfn(dn, un, lower, pmp)
 		for (i = 3; i > 0 && *dn != ' ';) {
 			c = dos2unixchr((const u_char **)&dn, &i,
 			    lower & LCASE_EXT, pmp);
-			if (c & 0xff00) {
-				*un++ = c >> 8;
+			while (*c != '\0') {
+				*un++ = *c++;
 				thislong++;
 			}
-			*un++ = c;
-			thislong++;
 		}
 	}
 	*un++ = 0;
@@ -652,8 +648,9 @@ win2unixfn(nbp, wep, chksum, pmp)
 	int chksum;
 	struct msdosfsmount *pmp;
 {
+	u_char *c;
 	u_int8_t *cp;
-	u_int8_t *np, name[WIN_CHARS * 2 + 1];
+	u_int8_t *np, name[WIN_CHARS * 3 + 1];
 	u_int16_t code;
 	int i;
 
@@ -686,10 +683,9 @@ win2unixfn(nbp, wep, chksum, pmp)
 			*np = '\0';
 			return -1;
 		default:
-			code = win2unixchr(code, pmp);
-			if (code & 0xff00)
-				*np++ = code >> 8;
-			*np++ = code;
+			c = win2unixchr(code, pmp);
+			while (*c != '\0')
+				*np++ = *c++;
 			break;
 		}
 		cp += 2;
@@ -705,10 +701,9 @@ win2unixfn(nbp, wep, chksum, pmp)
 			*np = '\0';
 			return -1;
 		default:
-			code = win2unixchr(code, pmp);
-			if (code & 0xff00)
-				*np++ = code >> 8;
-			*np++ = code;
+			c = win2unixchr(code, pmp);
+			while (*c != '\0')
+				*np++ = *c++;
 			break;
 		}
 		cp += 2;
@@ -724,10 +719,9 @@ win2unixfn(nbp, wep, chksum, pmp)
 			*np = '\0';
 			return -1;
 		default:
-			code = win2unixchr(code, pmp);
-			if (code & 0xff00)
-				*np++ = code >> 8;
-			*np++ = code;
+			c = win2unixchr(code, pmp);
+			while (*c != '\0')
+				*np++ = *c++;
 			break;
 		}
 		cp += 2;
@@ -817,24 +811,22 @@ mbsadjpos(const char **instr, size_t inlen, size_t outlen, int weight, int flag,
 /*
  * Convert DOS char to Local char
  */
-static u_int16_t
+static u_char *
 dos2unixchr(const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp)
 {
-	u_char c;
-	char *outp, outbuf[3];
-	u_int16_t wc;
+	u_char c, *outp, outbuf[5];
 	size_t len, olen;
 
+	outp = outbuf;
 	if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
-		olen = len = 2;
-		outp = outbuf;
+		olen = len = 4;
 
 		if (lower & (LCASE_BASE | LCASE_EXT))
 			msdosfs_iconv->convchr_case(pmp->pm_d2u, (const char **)instr,
-						  ilen, &outp, &olen, KICONV_LOWER);
+						  ilen, (char **)&outp, &olen, KICONV_LOWER);
 		else
 			msdosfs_iconv->convchr(pmp->pm_d2u, (const char **)instr,
-					     ilen, &outp, &olen);
+					     ilen, (char **)&outp, &olen);
 		len -= olen;
 
 		/*
@@ -843,21 +835,21 @@ dos2unixchr(const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *
 		if (len == 0) {
 			(*ilen)--;
 			(*instr)++;
-			return ('?');
+			*outp++ = '?';
 		}
-
-		wc = 0;
-		while(len--)
-			wc |= (*(outp - len - 1) & 0xff) << (len << 3);
-		return (wc);
+	} else {
+		(*ilen)--;
+		c = *(*instr)++;
+		c = dos2unix[c];
+		if (lower & (LCASE_BASE | LCASE_EXT))
+			c = u2l[c];
+		*outp++ = c;
+		outbuf[1] = '\0';
 	}
 
-	(*ilen)--;
-	c = *(*instr)++;
-	c = dos2unix[c];
-	if (lower & (LCASE_BASE | LCASE_EXT))
-		c = u2l[c];
-	return ((u_int16_t)c);
+	*outp = '\0';
+	outp = outbuf;
+	return (outp);
 }
 
 /*
@@ -940,23 +932,21 @@ unix2doschr(const u_char **instr, size_t *ilen, struct msdosfsmount *pmp)
 /*
  * Convert Windows char to Local char
  */
-static u_int16_t
+static u_char *
 win2unixchr(u_int16_t wc, struct msdosfsmount *pmp)
 {
-	u_char *inp, *outp, inbuf[3], outbuf[3];
+	u_char *inp, *outp, inbuf[3], outbuf[5];
 	size_t ilen, olen, len;
 
-	if (wc == 0)
-		return (0);
-
+	outp = outbuf;
 	if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
 		inbuf[0] = (u_char)(wc>>8);
 		inbuf[1] = (u_char)wc;
 		inbuf[2] = '\0';
 
-		ilen = olen = len = 2;
+		ilen = 2;
+		olen = len = 4;
 		inp = inbuf;
-		outp = outbuf;
 		msdosfs_iconv->convchr(pmp->pm_w2u, (const char **)&inp, &ilen,
 				     (char **)&outp, &olen);
 		len -= olen;
@@ -964,21 +954,15 @@ win2unixchr(u_int16_t wc, struct msdosfsmount *pmp)
 		/*
 		 * return '?' if failed to convert
 		 */
-		if (len == 0) {
-			wc = '?';
-			return (wc);
-		}
-
-		wc = 0;
-		while(len--)
-			wc |= (*(outp - len - 1) & 0xff) << (len << 3);
-		return (wc);
+		if (len == 0)
+			*outp++ = '?';
+	} else {
+		*outp++ = (wc & 0xff00) ? '?' : (u_char)(wc & 0xff);
 	}
 
-	if (wc & 0xff00)
-		wc = '?';
-
-	return (wc);
+	*outp = '\0';
+	outp = outbuf;
+	return (outp);
 }
 
 /*
diff --git a/sys/fs/smbfs/smbfs_smb.c b/sys/fs/smbfs/smbfs_smb.c
index bebc8ee..05162d5 100644
--- a/sys/fs/smbfs/smbfs_smb.c
+++ b/sys/fs/smbfs/smbfs_smb.c
@@ -34,6 +34,7 @@
 #include <sys/vnode.h>
 #include <sys/mbuf.h>
 #include <sys/mount.h>
+#include <sys/endian.h>
 
 #ifdef USE_MD5_HASH
 #include <sys/md5.h>
@@ -393,6 +394,10 @@ smbfs_smb_setpattr(struct smbnode *np, u_int16_t attr, struct timespec *mtime,
 		if (error)
 			break;
 		mb_put_uint8(mbp, SMB_DT_ASCII);
+		if (SMB_UNICODE_STRINGS(SSTOVC(ssp))) {
+			mb_put_padbyte(mbp);
+			mb_put_uint8(mbp, 0);	/* 1st byte of NULL Unicode char */
+		}
 		mb_put_uint8(mbp, 0);
 		smb_rq_bend(rqp);
 		error = smb_rq_simple(rqp);
@@ -909,6 +914,10 @@ smbfs_smb_search(struct smbfs_fctx *ctx)
 		mb_put_uint16le(mbp, 0);	/* context length */
 		ctx->f_flags &= ~SMBFS_RDD_FINDFIRST;
 	} else {
+		if (SMB_UNICODE_STRINGS(vcp)) {
+			mb_put_padbyte(mbp);
+			mb_put_uint8(mbp, 0);
+		}
 		mb_put_uint8(mbp, 0);	/* file name length */
 		mb_put_uint8(mbp, SMB_DT_VARIABLE);
 		mb_put_uint16le(mbp, SMB_SKEYLEN);
@@ -1069,7 +1078,7 @@ smbfs_smb_trans2find2(struct smbfs_fctx *ctx)
 		mb_put_uint32le(mbp, 0);		/* resume key */
 		mb_put_uint16le(mbp, flags);
 		if (ctx->f_rname)
-			mb_put_mem(mbp, ctx->f_rname, strlen(ctx->f_rname) + 1, MB_MSYSTEM);
+			mb_put_mem(mbp, ctx->f_rname, ctx->f_rnamelen + 1, MB_MSYSTEM);
 		else
 			mb_put_uint8(mbp, 0);	/* resume file name */
 #if 0
@@ -1152,7 +1161,10 @@ static int
 smbfs_findopenLM2(struct smbfs_fctx *ctx, struct smbnode *dnp,
 	const char *wildcard, int wclen, int attr, struct smb_cred *scred)
 {
-	ctx->f_name = malloc(SMB_MAXFNAMELEN, M_SMBFSDATA, M_WAITOK);
+	if (SMB_UNICODE_STRINGS(SSTOVC(ctx->f_ssp))) {
+		ctx->f_name = malloc(SMB_MAXFNAMELEN * 2, M_SMBFSDATA, M_WAITOK);
+	} else
+		ctx->f_name = malloc(SMB_MAXFNAMELEN, M_SMBFSDATA, M_WAITOK);
 	if (ctx->f_name == NULL)
 		return ENOMEM;
 	ctx->f_infolevel = SMB_DIALECT(SSTOVC(ctx->f_ssp)) < SMB_DIALECT_NTLM0_12 ?
@@ -1231,7 +1243,10 @@ smbfs_findnextLM2(struct smbfs_fctx *ctx, int limit)
 		SMBERROR("unexpected info level %d\n", ctx->f_infolevel);
 		return EINVAL;
 	}
-	nmlen = min(size, SMB_MAXFNAMELEN);
+	if (SMB_UNICODE_STRINGS(SSTOVC(ctx->f_ssp))) {
+		nmlen = min(size, SMB_MAXFNAMELEN * 2);
+	} else
+		nmlen = min(size, SMB_MAXFNAMELEN);
 	cp = ctx->f_name;
 	error = md_get_mem(mbp, cp, nmlen, MB_MSYSTEM);
 	if (error)
@@ -1245,8 +1260,12 @@ smbfs_findnextLM2(struct smbfs_fctx *ctx, int limit)
 			return EBADRPC;
 		}
 	}
-	if (nmlen && cp[nmlen - 1] == 0)
-		nmlen--;
+	if (SMB_UNICODE_STRINGS(SSTOVC(ctx->f_ssp))) {
+		if (nmlen > 1 && cp[nmlen - 1] == 0 && cp[nmlen - 2] == 0)
+			nmlen -= 2;
+	} else
+		if (nmlen && cp[nmlen - 1] == 0)
+			nmlen--;
 	if (nmlen == 0)
 		return EBADRPC;
 
@@ -1330,10 +1349,17 @@ smbfs_findnext(struct smbfs_fctx *ctx, int limit, struct smb_cred *scred)
 			error = smbfs_findnextLM2(ctx, limit);
 		if (error)
 			return error;
-		if ((ctx->f_nmlen == 1 && ctx->f_name[0] == '.') ||
-		    (ctx->f_nmlen == 2 && ctx->f_name[0] == '.' &&
-		     ctx->f_name[1] == '.'))
-			continue;
+		if (SMB_UNICODE_STRINGS(SSTOVC(ctx->f_ssp))) {
+			if ((ctx->f_nmlen == 2 &&
+			     *(u_int16_t *)ctx->f_name == htole16(0x002e)) ||
+			    (ctx->f_nmlen == 4 &&
+			     *(u_int32_t *)ctx->f_name == htole32(0x002e002e)))
+				continue;
+		} else
+			if ((ctx->f_nmlen == 1 && ctx->f_name[0] == '.') ||
+			    (ctx->f_nmlen == 2 && ctx->f_name[0] == '.' &&
+			     ctx->f_name[1] == '.'))
+				continue;
 		break;
 	}
 	smbfs_fname_tolocal(SSTOVC(ctx->f_ssp), ctx->f_name, &ctx->f_nmlen,
diff --git a/sys/fs/smbfs/smbfs_subr.c b/sys/fs/smbfs/smbfs_subr.c
index b775dff..51ee4d5 100644
--- a/sys/fs/smbfs/smbfs_subr.c
+++ b/sys/fs/smbfs/smbfs_subr.c
@@ -130,7 +130,10 @@ smb_fphelp(struct mbchain *mbp, struct smb_vc *vcp, struct smbnode *np,
 		return smb_put_dmem(mbp, vcp, "\\", 2, caseopt);*/
 	while (i--) {
 		np = *--npp;
-		error = mb_put_uint8(mbp, '\\');
+		if (SMB_UNICODE_STRINGS(vcp))
+			error = mb_put_uint16le(mbp, '\\');
+		else
+			error = mb_put_uint8(mbp, '\\');
 		if (error)
 			break;
 		error = smb_put_dmem(mbp, vcp, np->n_name, np->n_nmlen, caseopt);
@@ -148,6 +151,11 @@ smbfs_fullpath(struct mbchain *mbp, struct smb_vc *vcp, struct smbnode *dnp,
 	int caseopt = SMB_CS_NONE;
 	int error;
 
+	if (SMB_UNICODE_STRINGS(vcp)) {
+		error = mb_put_padbyte(mbp);
+		if (error)
+			return error;
+	}
 	if (SMB_DIALECT(vcp) < SMB_DIALECT_LANMAN1_0)
 		caseopt |= SMB_CS_UPPER;
 	if (dnp != NULL) {
@@ -156,7 +164,10 @@ smbfs_fullpath(struct mbchain *mbp, struct smb_vc *vcp, struct smbnode *dnp,
 			return error;
 	}
 	if (name) {
-		error = mb_put_uint8(mbp, '\\');
+		if (SMB_UNICODE_STRINGS(vcp))
+			error = mb_put_uint16le(mbp, '\\');
+		else
+			error = mb_put_uint8(mbp, '\\');
 		if (error)
 			return error;
 		error = smb_put_dmem(mbp, vcp, name, nmlen, caseopt);
@@ -164,6 +175,8 @@ smbfs_fullpath(struct mbchain *mbp, struct smb_vc *vcp, struct smbnode *dnp,
 			return error;
 	}
 	error = mb_put_uint8(mbp, 0);
+	if (SMB_UNICODE_STRINGS(vcp) && error == 0)
+		error = mb_put_uint8(mbp, 0);
 	return error;
 }
 
@@ -191,6 +204,17 @@ smbfs_fname_tolocal(struct smb_vc *vcp, char *name, int *nmlen, int caseopt)
 
 		error = iconv_conv_case
 			(vcp->vc_tolocal, (const char **)&ibuf, &ilen, &obuf, &olen, copt);
+		if (error && SMB_UNICODE_STRINGS(vcp)) {
+			/*
+			 * If using unicode, leaving a file name as it was when
+			 * convert fails will cause a problem because the file name
+			 * will contain NULL.
+			 * Here, put '?' and give converted file name.
+			 */
+			*obuf = '?';
+			olen--;
+			error = 0;
+		}
 		if (!error) {
 			*nmlen = sizeof(outbuf) - olen;
 			memcpy(name, outbuf, *nmlen);