9075 Improve ZFS pool import/load process and corrupted pool recovery

[unleashed.git] / usr / src / common / smbsrv / smb_string.c

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */
/*
 * Copyright 2010 Sun Microsystems, Inc.  All rights reserved.
 * Use is subject to license terms.
 *
 * Copyright 2014 Nexenta Systems, Inc.  All rights reserved.
 * Copyright (c) 2017 by Delphix. All rights reserved.
 */

#if defined(_KERNEL) || defined(_FAKE_KERNEL)
#include <sys/types.h>
#include <sys/sunddi.h>
#else
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <strings.h>
#endif
#include <sys/u8_textprep.h>
#include <smbsrv/alloc.h>
#include <sys/errno.h>
#include <smbsrv/string.h>
#include <smbsrv/cp_usascii.h>
#include <smbsrv/cp_unicode.h>

#define UNICODE_N_ENTRIES       (sizeof (a_unicode) / sizeof (a_unicode[0]))

/*
 * Global pointer to the current codepage: defaults to ASCII,
 * and a flag indicating whether the codepage is Unicode or ASCII.
 */
static const smb_codepage_t *current_codepage = usascii_codepage;
static boolean_t is_unicode = B_FALSE;

static smb_codepage_t *unicode_codepage = NULL;

static smb_codepage_t *smb_unicode_init(void);

/*
 * strsubst
 *
 * Scan a string replacing all occurrences of orgchar with newchar.
 * Returns a pointer to s, or null of s is null.
 */
char *
strsubst(char *s, char orgchar, char newchar)
{
        char *p = s;

        if (p == 0)
                return (0);

        while (*p) {
                if (*p == orgchar)
                        *p = newchar;
                ++p;
        }

        return (s);
}

/*
 * strcanon
 *
 * Normalize a string by reducing all the repeated characters in
 * buf as defined by class. For example;
 *
 *              char *buf = strdup("/d1//d2//d3\\\\d4\\\\f1.txt");
 *              strcanon(buf, "/\\");
 *
 * Would result in buf containing the following string:
 *
 *              /d1/d2/d3\d4\f1.txt
 *
 * This function modifies the contents of buf in place and returns
 * a pointer to buf.
 */
char *
strcanon(char *buf, const char *class)
{
        char *p = buf;
        char *q = buf;
        char *r;

        while (*p) {
                *q++ = *p;

                if ((r = strchr(class, *p)) != 0) {
                        while (*p == *r)
                                ++p;
                } else
                        ++p;
        }

        *q = '\0';
        return (buf);
}

void
smb_codepage_init(void)
{
        smb_codepage_t *cp;

        if (is_unicode)
                return;

        if ((cp = smb_unicode_init()) != NULL) {
                current_codepage = cp;
                unicode_codepage = cp;
                is_unicode = B_TRUE;
        } else {
                current_codepage = usascii_codepage;
                is_unicode = B_FALSE;
        }
}

void
smb_codepage_fini(void)
{
        if (unicode_codepage != NULL) {
                MEM_FREE("unicode", unicode_codepage);
                unicode_codepage = NULL;
                current_codepage = NULL;
        }
}

/*
 * Determine whether or not a character is an uppercase character.
 * This function operates on the current codepage table. Returns
 * non-zero if the character is uppercase. Otherwise returns zero.
 */
int
smb_isupper(int c)
{
        uint16_t mask = is_unicode ? 0xffff : 0xff;

        return (current_codepage[c & mask].ctype & CODEPAGE_ISUPPER);
}

/*
 * Determine whether or not a character is an lowercase character.
 * This function operates on the current codepage table. Returns
 * non-zero if the character is lowercase. Otherwise returns zero.
 */
int
smb_islower(int c)
{
        uint16_t mask = is_unicode ? 0xffff : 0xff;

        return (current_codepage[c & mask].ctype & CODEPAGE_ISLOWER);
}

/*
 * Convert individual characters to their uppercase equivalent value.
 * If the specified character is lowercase, the uppercase value will
 * be returned. Otherwise the original value will be returned.
 */
int
smb_toupper(int c)
{
        uint16_t mask = is_unicode ? 0xffff : 0xff;

        return (current_codepage[c & mask].upper);
}

/*
 * Convert individual characters to their lowercase equivalent value.
 * If the specified character is uppercase, the lowercase value will
 * be returned. Otherwise the original value will be returned.
 */
int
smb_tolower(int c)
{
        uint16_t mask = is_unicode ? 0xffff : 0xff;

        return (current_codepage[c & mask].lower);
}

/*
 * Convert a string to uppercase using the appropriate codepage. The
 * string is converted in place. A pointer to the string is returned.
 * There is an assumption here that uppercase and lowercase values
 * always result encode to the same length.
 */
char *
smb_strupr(char *s)
{
        smb_wchar_t c;
        char *p = s;

        while (*p) {
                if (smb_isascii(*p)) {
                        *p = smb_toupper(*p);
                        p++;
                } else {
                        if (smb_mbtowc(&c, p, MTS_MB_CHAR_MAX) < 0)
                                return (0);

                        if (c == 0)
                                break;

                        c = smb_toupper(c);
                        p += smb_wctomb(p, c);
                }
        }

        return (s);
}

/*
 * Convert a string to lowercase using the appropriate codepage. The
 * string is converted in place. A pointer to the string is returned.
 * There is an assumption here that uppercase and lowercase values
 * always result encode to the same length.
 */
char *
smb_strlwr(char *s)
{
        smb_wchar_t c;
        char *p = s;

        while (*p) {
                if (smb_isascii(*p)) {
                        *p = smb_tolower(*p);
                        p++;
                } else {
                        if (smb_mbtowc(&c, p, MTS_MB_CHAR_MAX) < 0)
                                return (0);

                        if (c == 0)
                                break;

                        c = smb_tolower(c);
                        p += smb_wctomb(p, c);
                }
        }

        return (s);
}

/*
 * Returns 1 if string contains NO uppercase chars 0 otherwise. However,
 * -1 is returned if "s" is not a valid multi-byte string.
 */
int
smb_isstrlwr(const char *s)
{
        smb_wchar_t c;
        int n;
        const char *p = s;

        while (*p) {
                if (smb_isascii(*p) && smb_isupper(*p))
                        return (0);
                else {
                        if ((n = smb_mbtowc(&c, p, MTS_MB_CHAR_MAX)) < 0)
                                return (-1);

                        if (c == 0)
                                break;

                        if (smb_isupper(c))
                                return (0);

                        p += n;
                }
        }

        return (1);
}

/*
 * Returns 1 if string contains NO lowercase chars 0 otherwise. However,
 * -1 is returned if "s" is not a valid multi-byte string.
 */
int
smb_isstrupr(const char *s)
{
        smb_wchar_t c;
        int n;
        const char *p = s;

        while (*p) {
                if (smb_isascii(*p) && smb_islower(*p))
                        return (0);
                else {
                        if ((n = smb_mbtowc(&c, p, MTS_MB_CHAR_MAX)) < 0)
                                return (-1);

                        if (c == 0)
                                break;

                        if (smb_islower(c))
                                return (0);

                        p += n;
                }
        }

        return (1);
}

/*
 * Compare the null-terminated strings s1 and s2 and return an integer
 * greater than, equal to or less than 0 dependent on whether s1 is
 * lexicographically greater than, equal to or less than s2 after
 * translation of each character to lowercase.  The original strings
 * are not modified.
 *
 * If n is non-zero, at most n bytes are compared.  Otherwise, the strings
 * are compared until a null terminator is encountered.
 *
 * Out:    0 if strings are equal
 *       < 0 if first string < second string
 *       > 0 if first string > second string
 */
int
smb_strcasecmp(const char *s1, const char *s2, size_t n)
{
        int     err = 0;
        int     rc;

        rc = u8_strcmp(s1, s2, n, U8_STRCMP_CI_LOWER, U8_UNICODE_LATEST, &err);
        if (err != 0)
                return (-1);
        return (rc);
}

/*
 * First build a codepage based on cp_unicode.h.  Then build the unicode
 * codepage from this interim codepage by copying the entries over while
 * fixing them and filling in the gaps.
 */
static smb_codepage_t *
smb_unicode_init(void)
{
        smb_codepage_t  *unicode;
        uint32_t        a = 0;
        uint32_t        b = 0;

        unicode = MEM_ZALLOC("unicode", sizeof (smb_codepage_t) << 16);
        if (unicode == NULL)
                return (NULL);

        while (b != 0xffff) {
                /*
                 * If there is a gap in the standard,
                 * fill in the gap with no-case entries.
                 */
                if (UNICODE_N_ENTRIES <= a || a_unicode[a].val > b) {
                        unicode[b].ctype = CODEPAGE_ISNONE;
                        unicode[b].upper = (smb_wchar_t)b;
                        unicode[b].lower = (smb_wchar_t)b;
                        b++;
                        continue;
                }

                /*
                 * Copy the entry and fixup as required.
                 */
                switch (a_unicode[a].ctype) {
                case CODEPAGE_ISNONE:
                        /*
                         * Replace 0xffff in upper/lower fields with its val.
                         */
                        unicode[b].ctype = CODEPAGE_ISNONE;
                        unicode[b].upper = (smb_wchar_t)b;
                        unicode[b].lower = (smb_wchar_t)b;
                        break;
                case CODEPAGE_ISUPPER:
                        /*
                         * Some characters may have case yet not have
                         * case conversion.  Treat them as no-case.
                         */
                        if (a_unicode[a].lower == 0xffff) {
                                unicode[b].ctype = CODEPAGE_ISNONE;
                                unicode[b].upper = (smb_wchar_t)b;
                                unicode[b].lower = (smb_wchar_t)b;
                        } else {
                                unicode[b].ctype = CODEPAGE_ISUPPER;
                                unicode[b].upper = (smb_wchar_t)b;
                                unicode[b].lower = a_unicode[a].lower;
                        }
                        break;
                case CODEPAGE_ISLOWER:
                        /*
                         * Some characters may have case yet not have
                         * case conversion.  Treat them as no-case.
                         */
                        if (a_unicode[a].upper == 0xffff) {
                                unicode[b].ctype = CODEPAGE_ISNONE;
                                unicode[b].upper = (smb_wchar_t)b;
                                unicode[b].lower = (smb_wchar_t)b;
                        } else {
                                unicode[b].ctype = CODEPAGE_ISLOWER;
                                unicode[b].upper = a_unicode[a].upper;
                                unicode[b].lower = (smb_wchar_t)b;
                        }
                        break;
                default:
                        MEM_FREE("unicode", unicode);
                        return (NULL);
                }

                a++;
                b++;
        };

        return (unicode);
}

/*
 * Parse a UNC path (\\server\share\path) into its components.
 * Although a standard UNC path starts with two '\', in DFS
 * all UNC paths start with one '\'. So, this function only
 * checks for one.
 *
 * A valid UNC must at least contain two components i.e. server
 * and share. The path is parsed to:
 *
 * unc_server   server or domain name with no leading/trailing '\'
 * unc_share    share name with no leading/trailing '\'
 * unc_path     relative path to the share with no leading/trailing '\'
 *              it is valid for unc_path to be NULL.
 *
 * Upon successful return of this function, smb_unc_free()
 * MUST be called when returned 'unc' is no longer needed.
 *
 * Returns 0 on success, otherwise returns an errno code.
 */
int
smb_unc_init(const char *path, smb_unc_t *unc)
{
        char *p;

        if (path == NULL || unc == NULL || (*path != '\\' && *path != '/'))
                return (EINVAL);

        bzero(unc, sizeof (smb_unc_t));

#if defined(_KERNEL) || defined(_FAKE_KERNEL)
        unc->unc_buf = smb_mem_strdup(path);
#else
        if ((unc->unc_buf = strdup(path)) == NULL)
                return (ENOMEM);
#endif

        (void) strsubst(unc->unc_buf, '\\', '/');
        (void) strcanon(unc->unc_buf, "/");

        unc->unc_server = unc->unc_buf + 1;
        if (*unc->unc_server == '\0') {
                smb_unc_free(unc);
                return (EINVAL);
        }

        if ((p = strchr(unc->unc_server, '/')) == NULL) {
                smb_unc_free(unc);
                return (EINVAL);
        }

        *p++ = '\0';
        unc->unc_share = p;

        if (*unc->unc_share == '\0') {
                smb_unc_free(unc);
                return (EINVAL);
        }

        unc->unc_path = strchr(unc->unc_share, '/');
        if ((p = unc->unc_path) == NULL)
                return (0);

        unc->unc_path++;
        *p = '\0';

        /* remove the last '/' if any */
        if ((p = strchr(unc->unc_path, '\0')) != NULL) {
                if (*(--p) == '/')
                        *p = '\0';
        }

        return (0);
}

void
smb_unc_free(smb_unc_t *unc)
{
        if (unc == NULL)
                return;

#if defined(_KERNEL) || defined(_FAKE_KERNEL)
        smb_mem_free(unc->unc_buf);
#else
        free(unc->unc_buf);
#endif
        unc->unc_buf = NULL;
}