Index: stable/10/lib/libfetch/common.c =================================================================== --- stable/10/lib/libfetch/common.c (revision 294193) +++ stable/10/lib/libfetch/common.c (revision 294194) @@ -1,1387 +1,1386 @@ /*- * Copyright (c) 1998-2014 Dag-Erling Smørgrav * Copyright (c) 2013 Michael Gmelin * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer * in this position and unchanged. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. The name of the author may not be used to endorse or promote products * derived from this software without specific prior written permission * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef WITH_SSL #include #endif #include "fetch.h" #include "common.h" /*** Local data **************************************************************/ /* * Error messages for resolver errors */ static struct fetcherr netdb_errlist[] = { #ifdef EAI_NODATA { EAI_NODATA, FETCH_RESOLV, "Host not found" }, #endif { EAI_AGAIN, FETCH_TEMP, "Transient resolver failure" }, { EAI_FAIL, FETCH_RESOLV, "Non-recoverable resolver failure" }, { EAI_NONAME, FETCH_RESOLV, "No address record" }, { -1, FETCH_UNKNOWN, "Unknown resolver error" } }; /* End-of-Line */ static const char ENDL[2] = "\r\n"; /*** Error-reporting functions ***********************************************/ /* * Map error code to string */ static struct fetcherr * fetch_finderr(struct fetcherr *p, int e) { while (p->num != -1 && p->num != e) p++; return (p); } /* * Set error code */ void fetch_seterr(struct fetcherr *p, int e) { p = fetch_finderr(p, e); fetchLastErrCode = p->cat; snprintf(fetchLastErrString, MAXERRSTRING, "%s", p->string); } /* * Set error code according to errno */ void fetch_syserr(void) { switch (errno) { case 0: fetchLastErrCode = FETCH_OK; break; case EPERM: case EACCES: case EROFS: case EAUTH: case ENEEDAUTH: fetchLastErrCode = FETCH_AUTH; break; case ENOENT: case EISDIR: /* XXX */ fetchLastErrCode = FETCH_UNAVAIL; break; case ENOMEM: fetchLastErrCode = FETCH_MEMORY; break; case EBUSY: case EAGAIN: fetchLastErrCode = FETCH_TEMP; break; case EEXIST: fetchLastErrCode = FETCH_EXISTS; break; case ENOSPC: fetchLastErrCode = FETCH_FULL; break; case EADDRINUSE: case EADDRNOTAVAIL: case ENETDOWN: case ENETUNREACH: case ENETRESET: case EHOSTUNREACH: fetchLastErrCode = FETCH_NETWORK; break; case ECONNABORTED: case ECONNRESET: fetchLastErrCode = FETCH_ABORT; break; case ETIMEDOUT: fetchLastErrCode = FETCH_TIMEOUT; break; case ECONNREFUSED: case EHOSTDOWN: fetchLastErrCode = FETCH_DOWN; break; default: fetchLastErrCode = FETCH_UNKNOWN; } snprintf(fetchLastErrString, MAXERRSTRING, "%s", strerror(errno)); } /* * Emit status message */ void fetch_info(const char *fmt, ...) { va_list ap; va_start(ap, fmt); vfprintf(stderr, fmt, ap); va_end(ap); fputc('\n', stderr); } /*** Network-related utility functions ***************************************/ /* * Return the default port for a scheme */ int fetch_default_port(const char *scheme) { struct servent *se; if ((se = getservbyname(scheme, "tcp")) != NULL) return (ntohs(se->s_port)); if (strcasecmp(scheme, SCHEME_FTP) == 0) return (FTP_DEFAULT_PORT); if (strcasecmp(scheme, SCHEME_HTTP) == 0) return (HTTP_DEFAULT_PORT); return (0); } /* * Return the default proxy port for a scheme */ int fetch_default_proxy_port(const char *scheme) { if (strcasecmp(scheme, SCHEME_FTP) == 0) return (FTP_DEFAULT_PROXY_PORT); if (strcasecmp(scheme, SCHEME_HTTP) == 0) return (HTTP_DEFAULT_PROXY_PORT); return (0); } /* * Create a connection for an existing descriptor. */ conn_t * fetch_reopen(int sd) { conn_t *conn; int opt = 1; /* allocate and fill connection structure */ if ((conn = calloc(1, sizeof(*conn))) == NULL) return (NULL); fcntl(sd, F_SETFD, FD_CLOEXEC); setsockopt(sd, SOL_SOCKET, SO_NOSIGPIPE, &opt, sizeof opt); conn->sd = sd; ++conn->ref; return (conn); } /* * Bump a connection's reference count. */ conn_t * fetch_ref(conn_t *conn) { ++conn->ref; return (conn); } /* * Bind a socket to a specific local address */ int fetch_bind(int sd, int af, const char *addr) { struct addrinfo hints, *res, *res0; int err; memset(&hints, 0, sizeof(hints)); hints.ai_family = af; hints.ai_socktype = SOCK_STREAM; hints.ai_protocol = 0; if ((err = getaddrinfo(addr, NULL, &hints, &res0)) != 0) return (-1); for (res = res0; res; res = res->ai_next) if (bind(sd, res->ai_addr, res->ai_addrlen) == 0) return (0); return (-1); } /* * Establish a TCP connection to the specified port on the specified host. */ conn_t * fetch_connect(const char *host, int port, int af, int verbose) { conn_t *conn; char pbuf[10]; const char *bindaddr; struct addrinfo hints, *res, *res0; int sd, err; DEBUG(fprintf(stderr, "---> %s:%d\n", host, port)); if (verbose) fetch_info("looking up %s", host); /* look up host name and set up socket address structure */ snprintf(pbuf, sizeof(pbuf), "%d", port); memset(&hints, 0, sizeof(hints)); hints.ai_family = af; hints.ai_socktype = SOCK_STREAM; hints.ai_protocol = 0; if ((err = getaddrinfo(host, pbuf, &hints, &res0)) != 0) { netdb_seterr(err); return (NULL); } bindaddr = getenv("FETCH_BIND_ADDRESS"); if (verbose) fetch_info("connecting to %s:%d", host, port); /* try to connect */ for (sd = -1, res = res0; res; sd = -1, res = res->ai_next) { if ((sd = socket(res->ai_family, res->ai_socktype, res->ai_protocol)) == -1) continue; if (bindaddr != NULL && *bindaddr != '\0' && fetch_bind(sd, res->ai_family, bindaddr) != 0) { fetch_info("failed to bind to '%s'", bindaddr); close(sd); continue; } if (connect(sd, res->ai_addr, res->ai_addrlen) == 0 && fcntl(sd, F_SETFL, O_NONBLOCK) == 0) break; close(sd); } freeaddrinfo(res0); if (sd == -1) { fetch_syserr(); return (NULL); } if ((conn = fetch_reopen(sd)) == NULL) { fetch_syserr(); close(sd); } return (conn); } #ifdef WITH_SSL /* * Convert characters A-Z to lowercase (intentionally avoid any locale * specific conversions). */ static char fetch_ssl_tolower(char in) { if (in >= 'A' && in <= 'Z') return (in + 32); else return (in); } /* * isalpha implementation that intentionally avoids any locale specific * conversions. */ static int fetch_ssl_isalpha(char in) { return ((in >= 'A' && in <= 'Z') || (in >= 'a' && in <= 'z')); } /* * Check if passed hostnames a and b are equal. */ static int fetch_ssl_hname_equal(const char *a, size_t alen, const char *b, size_t blen) { size_t i; if (alen != blen) return (0); for (i = 0; i < alen; ++i) { if (fetch_ssl_tolower(a[i]) != fetch_ssl_tolower(b[i])) return (0); } return (1); } /* * Check if domain label is traditional, meaning that only A-Z, a-z, 0-9 * and '-' (hyphen) are allowed. Hyphens have to be surrounded by alpha- * numeric characters. Double hyphens (like they're found in IDN a-labels * 'xn--') are not allowed. Empty labels are invalid. */ static int fetch_ssl_is_trad_domain_label(const char *l, size_t len, int wcok) { size_t i; if (!len || l[0] == '-' || l[len-1] == '-') return (0); for (i = 0; i < len; ++i) { if (!isdigit(l[i]) && !fetch_ssl_isalpha(l[i]) && !(l[i] == '*' && wcok) && !(l[i] == '-' && l[i - 1] != '-')) return (0); } return (1); } /* * Check if host name consists only of numbers. This might indicate an IP * address, which is not a good idea for CN wildcard comparison. */ static int fetch_ssl_hname_is_only_numbers(const char *hostname, size_t len) { size_t i; for (i = 0; i < len; ++i) { if (!((hostname[i] >= '0' && hostname[i] <= '9') || hostname[i] == '.')) return (0); } return (1); } /* * Check if the host name h passed matches the pattern passed in m which * is usually part of subjectAltName or CN of a certificate presented to * the client. This includes wildcard matching. The algorithm is based on * RFC6125, sections 6.4.3 and 7.2, which clarifies RFC2818 and RFC3280. */ static int fetch_ssl_hname_match(const char *h, size_t hlen, const char *m, size_t mlen) { int delta, hdotidx, mdot1idx, wcidx; const char *hdot, *mdot1, *mdot2; const char *wc; /* wildcard */ if (!(h && *h && m && *m)) return (0); if ((wc = strnstr(m, "*", mlen)) == NULL) return (fetch_ssl_hname_equal(h, hlen, m, mlen)); wcidx = wc - m; /* hostname should not be just dots and numbers */ if (fetch_ssl_hname_is_only_numbers(h, hlen)) return (0); /* only one wildcard allowed in pattern */ if (strnstr(wc + 1, "*", mlen - wcidx - 1) != NULL) return (0); /* * there must be at least two more domain labels and * wildcard has to be in the leftmost label (RFC6125) */ mdot1 = strnstr(m, ".", mlen); if (mdot1 == NULL || mdot1 < wc || (mlen - (mdot1 - m)) < 4) return (0); mdot1idx = mdot1 - m; mdot2 = strnstr(mdot1 + 1, ".", mlen - mdot1idx - 1); if (mdot2 == NULL || (mlen - (mdot2 - m)) < 2) return (0); /* hostname must contain a dot and not be the 1st char */ hdot = strnstr(h, ".", hlen); if (hdot == NULL || hdot == h) return (0); hdotidx = hdot - h; /* * host part of hostname must be at least as long as * pattern it's supposed to match */ if (hdotidx < mdot1idx) return (0); /* * don't allow wildcards in non-traditional domain names * (IDN, A-label, U-label...) */ if (!fetch_ssl_is_trad_domain_label(h, hdotidx, 0) || !fetch_ssl_is_trad_domain_label(m, mdot1idx, 1)) return (0); /* match domain part (part after first dot) */ if (!fetch_ssl_hname_equal(hdot, hlen - hdotidx, mdot1, mlen - mdot1idx)) return (0); /* match part left of wildcard */ if (!fetch_ssl_hname_equal(h, wcidx, m, wcidx)) return (0); /* match part right of wildcard */ delta = mdot1idx - wcidx - 1; if (!fetch_ssl_hname_equal(hdot - delta, delta, mdot1 - delta, delta)) return (0); /* all tests succeded, it's a match */ return (1); } /* * Get numeric host address info - returns NULL if host was not an IP * address. The caller is responsible for deallocation using * freeaddrinfo(3). */ static struct addrinfo * fetch_ssl_get_numeric_addrinfo(const char *hostname, size_t len) { struct addrinfo hints, *res; char *host; host = (char *)malloc(len + 1); memcpy(host, hostname, len); host[len] = '\0'; memset(&hints, 0, sizeof(hints)); hints.ai_family = PF_UNSPEC; hints.ai_socktype = SOCK_STREAM; hints.ai_protocol = 0; hints.ai_flags = AI_NUMERICHOST; /* port is not relevant for this purpose */ - getaddrinfo(host, "443", &hints, &res); + if (getaddrinfo(host, "443", &hints, &res) != 0) + res = NULL; free(host); return res; } /* * Compare ip address in addrinfo with address passes. */ static int fetch_ssl_ipaddr_match_bin(const struct addrinfo *lhost, const char *rhost, size_t rhostlen) { const void *left; if (lhost->ai_family == AF_INET && rhostlen == 4) { left = (void *)&((struct sockaddr_in*)(void *) lhost->ai_addr)->sin_addr.s_addr; #ifdef INET6 } else if (lhost->ai_family == AF_INET6 && rhostlen == 16) { left = (void *)&((struct sockaddr_in6 *)(void *) lhost->ai_addr)->sin6_addr; #endif } else return (0); return (!memcmp(left, (const void *)rhost, rhostlen) ? 1 : 0); } /* * Compare ip address in addrinfo with host passed. If host is not an IP * address, comparison will fail. */ static int fetch_ssl_ipaddr_match(const struct addrinfo *laddr, const char *r, size_t rlen) { struct addrinfo *raddr; int ret; char *rip; ret = 0; if ((raddr = fetch_ssl_get_numeric_addrinfo(r, rlen)) == NULL) return 0; /* not a numeric host */ if (laddr->ai_family == raddr->ai_family) { if (laddr->ai_family == AF_INET) { rip = (char *)&((struct sockaddr_in *)(void *) raddr->ai_addr)->sin_addr.s_addr; ret = fetch_ssl_ipaddr_match_bin(laddr, rip, 4); #ifdef INET6 } else if (laddr->ai_family == AF_INET6) { rip = (char *)&((struct sockaddr_in6 *)(void *) raddr->ai_addr)->sin6_addr; ret = fetch_ssl_ipaddr_match_bin(laddr, rip, 16); #endif } } freeaddrinfo(raddr); return (ret); } /* * Verify server certificate by subjectAltName. */ static int fetch_ssl_verify_altname(STACK_OF(GENERAL_NAME) *altnames, const char *host, struct addrinfo *ip) { const GENERAL_NAME *name; size_t nslen; int i; const char *ns; for (i = 0; i < sk_GENERAL_NAME_num(altnames); ++i) { #if OPENSSL_VERSION_NUMBER < 0x10000000L /* * This is a workaround, since the following line causes * alignment issues in clang: * name = sk_GENERAL_NAME_value(altnames, i); * OpenSSL explicitly warns not to use those macros * directly, but there isn't much choice (and there * shouldn't be any ill side effects) */ name = (GENERAL_NAME *)SKM_sk_value(void, altnames, i); #else name = sk_GENERAL_NAME_value(altnames, i); #endif ns = (const char *)ASN1_STRING_data(name->d.ia5); nslen = (size_t)ASN1_STRING_length(name->d.ia5); if (name->type == GEN_DNS && ip == NULL && fetch_ssl_hname_match(host, strlen(host), ns, nslen)) return (1); else if (name->type == GEN_IPADD && ip != NULL && fetch_ssl_ipaddr_match_bin(ip, ns, nslen)) return (1); } return (0); } /* * Verify server certificate by CN. */ static int fetch_ssl_verify_cn(X509_NAME *subject, const char *host, struct addrinfo *ip) { ASN1_STRING *namedata; X509_NAME_ENTRY *nameentry; int cnlen, lastpos, loc, ret; unsigned char *cn; ret = 0; lastpos = -1; loc = -1; cn = NULL; /* get most specific CN (last entry in list) and compare */ while ((lastpos = X509_NAME_get_index_by_NID(subject, NID_commonName, lastpos)) != -1) loc = lastpos; if (loc > -1) { nameentry = X509_NAME_get_entry(subject, loc); namedata = X509_NAME_ENTRY_get_data(nameentry); cnlen = ASN1_STRING_to_UTF8(&cn, namedata); if (ip == NULL && fetch_ssl_hname_match(host, strlen(host), cn, cnlen)) ret = 1; else if (ip != NULL && fetch_ssl_ipaddr_match(ip, cn, cnlen)) ret = 1; OPENSSL_free(cn); } return (ret); } /* * Verify that server certificate subjectAltName/CN matches * hostname. First check, if there are alternative subject names. If yes, * those have to match. Only if those don't exist it falls back to * checking the subject's CN. */ static int fetch_ssl_verify_hname(X509 *cert, const char *host) { struct addrinfo *ip; STACK_OF(GENERAL_NAME) *altnames; X509_NAME *subject; int ret; ret = 0; ip = fetch_ssl_get_numeric_addrinfo(host, strlen(host)); altnames = X509_get_ext_d2i(cert, NID_subject_alt_name, NULL, NULL); if (altnames != NULL) { ret = fetch_ssl_verify_altname(altnames, host, ip); } else { subject = X509_get_subject_name(cert); if (subject != NULL) ret = fetch_ssl_verify_cn(subject, host, ip); } if (ip != NULL) freeaddrinfo(ip); if (altnames != NULL) GENERAL_NAMES_free(altnames); return (ret); } /* * Configure transport security layer based on environment. */ static void fetch_ssl_setup_transport_layer(SSL_CTX *ctx, int verbose) { long ssl_ctx_options; - ssl_ctx_options = SSL_OP_ALL | SSL_OP_NO_TICKET; - if (getenv("SSL_ALLOW_SSL2") == NULL) - ssl_ctx_options |= SSL_OP_NO_SSLv2; + ssl_ctx_options = SSL_OP_ALL | SSL_OP_NO_SSLv2 | SSL_OP_NO_TICKET; if (getenv("SSL_ALLOW_SSL3") == NULL) ssl_ctx_options |= SSL_OP_NO_SSLv3; if (getenv("SSL_NO_TLS1") != NULL) ssl_ctx_options |= SSL_OP_NO_TLSv1; if (getenv("SSL_NO_TLS1_1") != NULL) ssl_ctx_options |= SSL_OP_NO_TLSv1_1; if (getenv("SSL_NO_TLS1_2") != NULL) ssl_ctx_options |= SSL_OP_NO_TLSv1_2; if (verbose) fetch_info("SSL options: %lx", ssl_ctx_options); SSL_CTX_set_options(ctx, ssl_ctx_options); } /* * Configure peer verification based on environment. */ #define LOCAL_CERT_FILE "/usr/local/etc/ssl/cert.pem" #define BASE_CERT_FILE "/etc/ssl/cert.pem" static int fetch_ssl_setup_peer_verification(SSL_CTX *ctx, int verbose) { X509_LOOKUP *crl_lookup; X509_STORE *crl_store; const char *ca_cert_file, *ca_cert_path, *crl_file; if (getenv("SSL_NO_VERIFY_PEER") == NULL) { ca_cert_file = getenv("SSL_CA_CERT_FILE"); if (ca_cert_file == NULL && access(LOCAL_CERT_FILE, R_OK) == 0) ca_cert_file = LOCAL_CERT_FILE; if (ca_cert_file == NULL) ca_cert_file = BASE_CERT_FILE; ca_cert_path = getenv("SSL_CA_CERT_PATH"); if (verbose) { fetch_info("Peer verification enabled"); if (ca_cert_file != NULL) fetch_info("Using CA cert file: %s", ca_cert_file); if (ca_cert_path != NULL) fetch_info("Using CA cert path: %s", ca_cert_path); } SSL_CTX_set_verify(ctx, SSL_VERIFY_PEER, fetch_ssl_cb_verify_crt); SSL_CTX_load_verify_locations(ctx, ca_cert_file, ca_cert_path); if ((crl_file = getenv("SSL_CRL_FILE")) != NULL) { if (verbose) fetch_info("Using CRL file: %s", crl_file); crl_store = SSL_CTX_get_cert_store(ctx); crl_lookup = X509_STORE_add_lookup(crl_store, X509_LOOKUP_file()); if (crl_lookup == NULL || !X509_load_crl_file(crl_lookup, crl_file, X509_FILETYPE_PEM)) { fprintf(stderr, "Could not load CRL file %s\n", crl_file); return (0); } X509_STORE_set_flags(crl_store, X509_V_FLAG_CRL_CHECK | X509_V_FLAG_CRL_CHECK_ALL); } } return (1); } /* * Configure client certificate based on environment. */ static int fetch_ssl_setup_client_certificate(SSL_CTX *ctx, int verbose) { const char *client_cert_file, *client_key_file; if ((client_cert_file = getenv("SSL_CLIENT_CERT_FILE")) != NULL) { client_key_file = getenv("SSL_CLIENT_KEY_FILE") != NULL ? getenv("SSL_CLIENT_KEY_FILE") : client_cert_file; if (verbose) { fetch_info("Using client cert file: %s", client_cert_file); fetch_info("Using client key file: %s", client_key_file); } if (SSL_CTX_use_certificate_chain_file(ctx, client_cert_file) != 1) { fprintf(stderr, "Could not load client certificate %s\n", client_cert_file); return (0); } if (SSL_CTX_use_PrivateKey_file(ctx, client_key_file, SSL_FILETYPE_PEM) != 1) { fprintf(stderr, "Could not load client key %s\n", client_key_file); return (0); } } return (1); } /* * Callback for SSL certificate verification, this is called on server * cert verification. It takes no decision, but informs the user in case * verification failed. */ int fetch_ssl_cb_verify_crt(int verified, X509_STORE_CTX *ctx) { X509 *crt; X509_NAME *name; char *str; str = NULL; if (!verified) { if ((crt = X509_STORE_CTX_get_current_cert(ctx)) != NULL && (name = X509_get_subject_name(crt)) != NULL) str = X509_NAME_oneline(name, 0, 0); fprintf(stderr, "Certificate verification failed for %s\n", str != NULL ? str : "no relevant certificate"); OPENSSL_free(str); } return (verified); } #endif /* * Enable SSL on a connection. */ int fetch_ssl(conn_t *conn, const struct url *URL, int verbose) { #ifdef WITH_SSL int ret, ssl_err; X509_NAME *name; char *str; /* Init the SSL library and context */ if (!SSL_library_init()){ fprintf(stderr, "SSL library init failed\n"); return (-1); } SSL_load_error_strings(); conn->ssl_meth = SSLv23_client_method(); conn->ssl_ctx = SSL_CTX_new(conn->ssl_meth); SSL_CTX_set_mode(conn->ssl_ctx, SSL_MODE_AUTO_RETRY); fetch_ssl_setup_transport_layer(conn->ssl_ctx, verbose); if (!fetch_ssl_setup_peer_verification(conn->ssl_ctx, verbose)) return (-1); if (!fetch_ssl_setup_client_certificate(conn->ssl_ctx, verbose)) return (-1); conn->ssl = SSL_new(conn->ssl_ctx); if (conn->ssl == NULL) { fprintf(stderr, "SSL context creation failed\n"); return (-1); } SSL_set_fd(conn->ssl, conn->sd); #if OPENSSL_VERSION_NUMBER >= 0x0090806fL && !defined(OPENSSL_NO_TLSEXT) if (!SSL_set_tlsext_host_name(conn->ssl, __DECONST(struct url *, URL)->host)) { fprintf(stderr, "TLS server name indication extension failed for host %s\n", URL->host); return (-1); } #endif while ((ret = SSL_connect(conn->ssl)) == -1) { ssl_err = SSL_get_error(conn->ssl, ret); if (ssl_err != SSL_ERROR_WANT_READ && ssl_err != SSL_ERROR_WANT_WRITE) { ERR_print_errors_fp(stderr); return (-1); } } conn->ssl_cert = SSL_get_peer_certificate(conn->ssl); if (conn->ssl_cert == NULL) { fprintf(stderr, "No server SSL certificate\n"); return (-1); } if (getenv("SSL_NO_VERIFY_HOSTNAME") == NULL) { if (verbose) fetch_info("Verify hostname"); if (!fetch_ssl_verify_hname(conn->ssl_cert, URL->host)) { fprintf(stderr, "SSL certificate subject doesn't match host %s\n", URL->host); return (-1); } } if (verbose) { fetch_info("%s connection established using %s", SSL_get_version(conn->ssl), SSL_get_cipher(conn->ssl)); name = X509_get_subject_name(conn->ssl_cert); str = X509_NAME_oneline(name, 0, 0); fetch_info("Certificate subject: %s", str); OPENSSL_free(str); name = X509_get_issuer_name(conn->ssl_cert); str = X509_NAME_oneline(name, 0, 0); fetch_info("Certificate issuer: %s", str); OPENSSL_free(str); } return (0); #else (void)conn; (void)verbose; fprintf(stderr, "SSL support disabled\n"); return (-1); #endif } #define FETCH_READ_WAIT -2 #define FETCH_READ_ERROR -1 #define FETCH_READ_DONE 0 #ifdef WITH_SSL static ssize_t fetch_ssl_read(SSL *ssl, char *buf, size_t len) { ssize_t rlen; int ssl_err; rlen = SSL_read(ssl, buf, len); if (rlen < 0) { ssl_err = SSL_get_error(ssl, rlen); if (ssl_err == SSL_ERROR_WANT_READ || ssl_err == SSL_ERROR_WANT_WRITE) { return (FETCH_READ_WAIT); } else { ERR_print_errors_fp(stderr); return (FETCH_READ_ERROR); } } return (rlen); } #endif static ssize_t fetch_socket_read(int sd, char *buf, size_t len) { ssize_t rlen; rlen = read(sd, buf, len); if (rlen < 0) { if (errno == EAGAIN || (errno == EINTR && fetchRestartCalls)) return (FETCH_READ_WAIT); else return (FETCH_READ_ERROR); } return (rlen); } /* * Read a character from a connection w/ timeout */ ssize_t fetch_read(conn_t *conn, char *buf, size_t len) { struct timeval now, timeout, delta; struct pollfd pfd; ssize_t rlen; int deltams; if (fetchTimeout > 0) { gettimeofday(&timeout, NULL); timeout.tv_sec += fetchTimeout; } deltams = INFTIM; memset(&pfd, 0, sizeof pfd); pfd.fd = conn->sd; pfd.events = POLLIN | POLLERR; for (;;) { /* * The socket is non-blocking. Instead of the canonical * poll() -> read(), we do the following: * * 1) call read() or SSL_read(). * 2) if we received some data, return it. * 3) if an error occurred, return -1. * 4) if read() or SSL_read() signaled EOF, return. * 5) if we did not receive any data but we're not at EOF, * call poll(). * * In the SSL case, this is necessary because if we * receive a close notification, we have to call * SSL_read() one additional time after we've read * everything we received. * * In the non-SSL case, it may improve performance (very * slightly) when reading small amounts of data. */ #ifdef WITH_SSL if (conn->ssl != NULL) rlen = fetch_ssl_read(conn->ssl, buf, len); else #endif rlen = fetch_socket_read(conn->sd, buf, len); if (rlen >= 0) { break; } else if (rlen == FETCH_READ_ERROR) { fetch_syserr(); return (-1); } // assert(rlen == FETCH_READ_WAIT); if (fetchTimeout > 0) { gettimeofday(&now, NULL); if (!timercmp(&timeout, &now, >)) { errno = ETIMEDOUT; fetch_syserr(); return (-1); } timersub(&timeout, &now, &delta); deltams = delta.tv_sec * 1000 + delta.tv_usec / 1000;; } errno = 0; pfd.revents = 0; if (poll(&pfd, 1, deltams) < 0) { if (errno == EINTR && fetchRestartCalls) continue; fetch_syserr(); return (-1); } } return (rlen); } /* * Read a line of text from a connection w/ timeout */ #define MIN_BUF_SIZE 1024 int fetch_getln(conn_t *conn) { char *tmp; size_t tmpsize; ssize_t len; char c; if (conn->buf == NULL) { if ((conn->buf = malloc(MIN_BUF_SIZE)) == NULL) { errno = ENOMEM; return (-1); } conn->bufsize = MIN_BUF_SIZE; } conn->buf[0] = '\0'; conn->buflen = 0; do { len = fetch_read(conn, &c, 1); if (len == -1) return (-1); if (len == 0) break; conn->buf[conn->buflen++] = c; if (conn->buflen == conn->bufsize) { tmp = conn->buf; tmpsize = conn->bufsize * 2 + 1; if ((tmp = realloc(tmp, tmpsize)) == NULL) { errno = ENOMEM; return (-1); } conn->buf = tmp; conn->bufsize = tmpsize; } } while (c != '\n'); conn->buf[conn->buflen] = '\0'; DEBUG(fprintf(stderr, "<<< %s", conn->buf)); return (0); } /* * Write to a connection w/ timeout */ ssize_t fetch_write(conn_t *conn, const char *buf, size_t len) { struct iovec iov; iov.iov_base = __DECONST(char *, buf); iov.iov_len = len; return fetch_writev(conn, &iov, 1); } /* * Write a vector to a connection w/ timeout * Note: can modify the iovec. */ ssize_t fetch_writev(conn_t *conn, struct iovec *iov, int iovcnt) { struct timeval now, timeout, delta; struct pollfd pfd; ssize_t wlen, total; int deltams; memset(&pfd, 0, sizeof pfd); if (fetchTimeout) { pfd.fd = conn->sd; pfd.events = POLLOUT | POLLERR; gettimeofday(&timeout, NULL); timeout.tv_sec += fetchTimeout; } total = 0; while (iovcnt > 0) { while (fetchTimeout && pfd.revents == 0) { gettimeofday(&now, NULL); if (!timercmp(&timeout, &now, >)) { errno = ETIMEDOUT; fetch_syserr(); return (-1); } timersub(&timeout, &now, &delta); deltams = delta.tv_sec * 1000 + delta.tv_usec / 1000; errno = 0; pfd.revents = 0; if (poll(&pfd, 1, deltams) < 0) { /* POSIX compliance */ if (errno == EAGAIN) continue; if (errno == EINTR && fetchRestartCalls) continue; return (-1); } } errno = 0; #ifdef WITH_SSL if (conn->ssl != NULL) wlen = SSL_write(conn->ssl, iov->iov_base, iov->iov_len); else #endif wlen = writev(conn->sd, iov, iovcnt); if (wlen == 0) { /* we consider a short write a failure */ /* XXX perhaps we shouldn't in the SSL case */ errno = EPIPE; fetch_syserr(); return (-1); } if (wlen < 0) { if (errno == EINTR && fetchRestartCalls) continue; return (-1); } total += wlen; while (iovcnt > 0 && wlen >= (ssize_t)iov->iov_len) { wlen -= iov->iov_len; iov++; iovcnt--; } if (iovcnt > 0) { iov->iov_len -= wlen; iov->iov_base = __DECONST(char *, iov->iov_base) + wlen; } } return (total); } /* * Write a line of text to a connection w/ timeout */ int fetch_putln(conn_t *conn, const char *str, size_t len) { struct iovec iov[2]; int ret; DEBUG(fprintf(stderr, ">>> %s\n", str)); iov[0].iov_base = __DECONST(char *, str); iov[0].iov_len = len; iov[1].iov_base = __DECONST(char *, ENDL); iov[1].iov_len = sizeof(ENDL); if (len == 0) ret = fetch_writev(conn, &iov[1], 1); else ret = fetch_writev(conn, iov, 2); if (ret == -1) return (-1); return (0); } /* * Close connection */ int fetch_close(conn_t *conn) { int ret; if (--conn->ref > 0) return (0); #ifdef WITH_SSL if (conn->ssl) { SSL_shutdown(conn->ssl); SSL_set_connect_state(conn->ssl); SSL_free(conn->ssl); conn->ssl = NULL; } if (conn->ssl_ctx) { SSL_CTX_free(conn->ssl_ctx); conn->ssl_ctx = NULL; } if (conn->ssl_cert) { X509_free(conn->ssl_cert); conn->ssl_cert = NULL; } #endif ret = close(conn->sd); free(conn->buf); free(conn); return (ret); } /*** Directory-related utility functions *************************************/ int fetch_add_entry(struct url_ent **p, int *size, int *len, const char *name, struct url_stat *us) { struct url_ent *tmp; if (*p == NULL) { *size = 0; *len = 0; } if (*len >= *size - 1) { tmp = realloc(*p, (*size * 2 + 1) * sizeof(**p)); if (tmp == NULL) { errno = ENOMEM; fetch_syserr(); return (-1); } *size = (*size * 2 + 1); *p = tmp; } tmp = *p + *len; snprintf(tmp->name, PATH_MAX, "%s", name); memcpy(&tmp->stat, us, sizeof(*us)); (*len)++; (++tmp)->name[0] = 0; return (0); } /*** Authentication-related utility functions ********************************/ static const char * fetch_read_word(FILE *f) { static char word[1024]; if (fscanf(f, " %1023s ", word) != 1) return (NULL); return (word); } /* * Get authentication data for a URL from .netrc */ int fetch_netrc_auth(struct url *url) { char fn[PATH_MAX]; const char *word; char *p; FILE *f; if ((p = getenv("NETRC")) != NULL) { if (snprintf(fn, sizeof(fn), "%s", p) >= (int)sizeof(fn)) { fetch_info("$NETRC specifies a file name " "longer than PATH_MAX"); return (-1); } } else { if ((p = getenv("HOME")) != NULL) { struct passwd *pwd; if ((pwd = getpwuid(getuid())) == NULL || (p = pwd->pw_dir) == NULL) return (-1); } if (snprintf(fn, sizeof(fn), "%s/.netrc", p) >= (int)sizeof(fn)) return (-1); } if ((f = fopen(fn, "r")) == NULL) return (-1); while ((word = fetch_read_word(f)) != NULL) { if (strcmp(word, "default") == 0) { DEBUG(fetch_info("Using default .netrc settings")); break; } if (strcmp(word, "machine") == 0 && (word = fetch_read_word(f)) != NULL && strcasecmp(word, url->host) == 0) { DEBUG(fetch_info("Using .netrc settings for %s", word)); break; } } if (word == NULL) goto ferr; while ((word = fetch_read_word(f)) != NULL) { if (strcmp(word, "login") == 0) { if ((word = fetch_read_word(f)) == NULL) goto ferr; if (snprintf(url->user, sizeof(url->user), "%s", word) > (int)sizeof(url->user)) { fetch_info("login name in .netrc is too long"); url->user[0] = '\0'; } } else if (strcmp(word, "password") == 0) { if ((word = fetch_read_word(f)) == NULL) goto ferr; if (snprintf(url->pwd, sizeof(url->pwd), "%s", word) > (int)sizeof(url->pwd)) { fetch_info("password in .netrc is too long"); url->pwd[0] = '\0'; } } else if (strcmp(word, "account") == 0) { if ((word = fetch_read_word(f)) == NULL) goto ferr; /* XXX not supported! */ } else { break; } } fclose(f); return (0); ferr: fclose(f); return (-1); } /* * The no_proxy environment variable specifies a set of domains for * which the proxy should not be consulted; the contents is a comma-, * or space-separated list of domain names. A single asterisk will * override all proxy variables and no transactions will be proxied * (for compatability with lynx and curl, see the discussion at * ). */ int fetch_no_proxy_match(const char *host) { const char *no_proxy, *p, *q; size_t h_len, d_len; if ((no_proxy = getenv("NO_PROXY")) == NULL && (no_proxy = getenv("no_proxy")) == NULL) return (0); /* asterisk matches any hostname */ if (strcmp(no_proxy, "*") == 0) return (1); h_len = strlen(host); p = no_proxy; do { /* position p at the beginning of a domain suffix */ while (*p == ',' || isspace((unsigned char)*p)) p++; /* position q at the first separator character */ for (q = p; *q; ++q) if (*q == ',' || isspace((unsigned char)*q)) break; d_len = q - p; if (d_len > 0 && h_len >= d_len && strncasecmp(host + h_len - d_len, p, d_len) == 0) { /* domain name matches */ return (1); } p = q + 1; } while (*q); return (0); } Index: stable/10/lib/libfetch/fetch.3 =================================================================== --- stable/10/lib/libfetch/fetch.3 (revision 294193) +++ stable/10/lib/libfetch/fetch.3 (revision 294194) @@ -1,841 +1,837 @@ .\"- .\" Copyright (c) 1998-2013 Dag-Erling Smørgrav .\" Copyright (c) 2013 Michael Gmelin .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in the .\" documentation and/or other materials provided with the distribution. .\" .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" .\" $FreeBSD$ .\" -.Dd October 15, 2014 +.Dd November 29, 2015 .Dt FETCH 3 .Os .Sh NAME .Nm fetchMakeURL , .Nm fetchParseURL , .Nm fetchFreeURL , .Nm fetchXGetURL , .Nm fetchGetURL , .Nm fetchPutURL , .Nm fetchStatURL , .Nm fetchListURL , .Nm fetchXGet , .Nm fetchGet , .Nm fetchPut , .Nm fetchStat , .Nm fetchList , .Nm fetchXGetFile , .Nm fetchGetFile , .Nm fetchPutFile , .Nm fetchStatFile , .Nm fetchListFile , .Nm fetchXGetHTTP , .Nm fetchGetHTTP , .Nm fetchPutHTTP , .Nm fetchStatHTTP , .Nm fetchListHTTP , .Nm fetchXGetFTP , .Nm fetchGetFTP , .Nm fetchPutFTP , .Nm fetchStatFTP , .Nm fetchListFTP .Nd file transfer functions .Sh LIBRARY .Lb libfetch .Sh SYNOPSIS .In sys/param.h .In stdio.h .In fetch.h .Ft struct url * .Fn fetchMakeURL "const char *scheme" "const char *host" "int port" "const char *doc" "const char *user" "const char *pwd" .Ft struct url * .Fn fetchParseURL "const char *URL" .Ft void .Fn fetchFreeURL "struct url *u" .Ft FILE * .Fn fetchXGetURL "const char *URL" "struct url_stat *us" "const char *flags" .Ft FILE * .Fn fetchGetURL "const char *URL" "const char *flags" .Ft FILE * .Fn fetchPutURL "const char *URL" "const char *flags" .Ft int .Fn fetchStatURL "const char *URL" "struct url_stat *us" "const char *flags" .Ft struct url_ent * .Fn fetchListURL "const char *URL" "const char *flags" .Ft FILE * .Fn fetchXGet "struct url *u" "struct url_stat *us" "const char *flags" .Ft FILE * .Fn fetchGet "struct url *u" "const char *flags" .Ft FILE * .Fn fetchPut "struct url *u" "const char *flags" .Ft int .Fn fetchStat "struct url *u" "struct url_stat *us" "const char *flags" .Ft struct url_ent * .Fn fetchList "struct url *u" "const char *flags" .Ft FILE * .Fn fetchXGetFile "struct url *u" "struct url_stat *us" "const char *flags" .Ft FILE * .Fn fetchGetFile "struct url *u" "const char *flags" .Ft FILE * .Fn fetchPutFile "struct url *u" "const char *flags" .Ft int .Fn fetchStatFile "struct url *u" "struct url_stat *us" "const char *flags" .Ft struct url_ent * .Fn fetchListFile "struct url *u" "const char *flags" .Ft FILE * .Fn fetchXGetHTTP "struct url *u" "struct url_stat *us" "const char *flags" .Ft FILE * .Fn fetchGetHTTP "struct url *u" "const char *flags" .Ft FILE * .Fn fetchPutHTTP "struct url *u" "const char *flags" .Ft int .Fn fetchStatHTTP "struct url *u" "struct url_stat *us" "const char *flags" .Ft struct url_ent * .Fn fetchListHTTP "struct url *u" "const char *flags" .Ft FILE * .Fn fetchXGetFTP "struct url *u" "struct url_stat *us" "const char *flags" .Ft FILE * .Fn fetchGetFTP "struct url *u" "const char *flags" .Ft FILE * .Fn fetchPutFTP "struct url *u" "const char *flags" .Ft int .Fn fetchStatFTP "struct url *u" "struct url_stat *us" "const char *flags" .Ft struct url_ent * .Fn fetchListFTP "struct url *u" "const char *flags" .Sh DESCRIPTION These functions implement a high-level library for retrieving and uploading files using Uniform Resource Locators (URLs). .Pp .Fn fetchParseURL takes a URL in the form of a null-terminated string and splits it into its components function according to the Common Internet Scheme Syntax detailed in RFC1738. A regular expression which produces this syntax is: .Bd -literal :(//((:)?@)?(:)?)?/()? .Ed .Pp If the URL does not seem to begin with a scheme name, the following syntax is assumed: .Bd -literal (((:)?@)?(:)?)?/()? .Ed .Pp Note that some components of the URL are not necessarily relevant to all URL schemes. For instance, the file scheme only needs the and components. .Pp .Fn fetchMakeURL and .Fn fetchParseURL return a pointer to a .Vt url structure, which is defined as follows in .In fetch.h : .Bd -literal #define URL_SCHEMELEN 16 #define URL_USERLEN 256 #define URL_PWDLEN 256 struct url { char scheme[URL_SCHEMELEN+1]; char user[URL_USERLEN+1]; char pwd[URL_PWDLEN+1]; char host[MAXHOSTNAMELEN+1]; int port; char *doc; off_t offset; size_t length; time_t ims_time; }; .Ed .Pp The .Va ims_time field stores the time value for .Li If-Modified-Since HTTP requests. .Pp The pointer returned by .Fn fetchMakeURL or .Fn fetchParseURL should be freed using .Fn fetchFreeURL . .Pp .Fn fetchXGetURL , .Fn fetchGetURL , and .Fn fetchPutURL constitute the recommended interface to the .Nm fetch library. They examine the URL passed to them to determine the transfer method, and call the appropriate lower-level functions to perform the actual transfer. .Fn fetchXGetURL also returns the remote document's metadata in the .Vt url_stat structure pointed to by the .Fa us argument. .Pp The .Fa flags argument is a string of characters which specify transfer options. The meaning of the individual flags is scheme-dependent, and is detailed in the appropriate section below. .Pp .Fn fetchStatURL attempts to obtain the requested document's metadata and fill in the structure pointed to by its second argument. The .Vt url_stat structure is defined as follows in .In fetch.h : .Bd -literal struct url_stat { off_t size; time_t atime; time_t mtime; }; .Ed .Pp If the size could not be obtained from the server, the .Fa size field is set to -1. If the modification time could not be obtained from the server, the .Fa mtime field is set to the epoch. If the access time could not be obtained from the server, the .Fa atime field is set to the modification time. .Pp .Fn fetchListURL attempts to list the contents of the directory pointed to by the URL provided. If successful, it returns a malloced array of .Vt url_ent structures. The .Vt url_ent structure is defined as follows in .In fetch.h : .Bd -literal struct url_ent { char name[PATH_MAX]; struct url_stat stat; }; .Ed .Pp The list is terminated by an entry with an empty name. .Pp The pointer returned by .Fn fetchListURL should be freed using .Fn free . .Pp .Fn fetchXGet , .Fn fetchGet , .Fn fetchPut and .Fn fetchStat are similar to .Fn fetchXGetURL , .Fn fetchGetURL , .Fn fetchPutURL and .Fn fetchStatURL , except that they expect a pre-parsed URL in the form of a pointer to a .Vt struct url rather than a string. .Pp All of the .Fn fetchXGetXXX , .Fn fetchGetXXX and .Fn fetchPutXXX functions return a pointer to a stream which can be used to read or write data from or to the requested document, respectively. Note that although the implementation details of the individual access methods vary, it can generally be assumed that a stream returned by one of the .Fn fetchXGetXXX or .Fn fetchGetXXX functions is read-only, and that a stream returned by one of the .Fn fetchPutXXX functions is write-only. .Sh FILE SCHEME .Fn fetchXGetFile , .Fn fetchGetFile and .Fn fetchPutFile provide access to documents which are files in a locally mounted file system. Only the component of the URL is used. .Pp .Fn fetchXGetFile and .Fn fetchGetFile do not accept any flags. .Pp .Fn fetchPutFile accepts the .Ql a (append to file) flag. If that flag is specified, the data written to the stream returned by .Fn fetchPutFile will be appended to the previous contents of the file, instead of replacing them. .Sh FTP SCHEME .Fn fetchXGetFTP , .Fn fetchGetFTP and .Fn fetchPutFTP implement the FTP protocol as described in RFC959. .Pp If the .Ql P (not passive) flag is specified, an active (rather than passive) connection will be attempted. .Pp The .Ql p flag is supported for compatibility with earlier versions where active connections were the default. It has precedence over the .Ql P flag, so if both are specified, .Nm will use a passive connection. .Pp If the .Ql l (low) flag is specified, data sockets will be allocated in the low (or default) port range instead of the high port range (see .Xr ip 4 ) . .Pp If the .Ql d (direct) flag is specified, .Fn fetchXGetFTP , .Fn fetchGetFTP and .Fn fetchPutFTP will use a direct connection even if a proxy server is defined. .Pp If no user name or password is given, the .Nm fetch library will attempt an anonymous login, with user name "anonymous" and password "anonymous@". .Sh HTTP SCHEME The .Fn fetchXGetHTTP , .Fn fetchGetHTTP and .Fn fetchPutHTTP functions implement the HTTP/1.1 protocol. With a little luck, there is even a chance that they comply with RFC2616 and RFC2617. .Pp If the .Ql d (direct) flag is specified, .Fn fetchXGetHTTP , .Fn fetchGetHTTP and .Fn fetchPutHTTP will use a direct connection even if a proxy server is defined. .Pp If the .Ql i (if-modified-since) flag is specified, and the .Va ims_time field is set in .Vt "struct url" , then .Fn fetchXGetHTTP and .Fn fetchGetHTTP will send a conditional .Li If-Modified-Since HTTP header to only fetch the content if it is newer than .Va ims_time . .Pp Since there seems to be no good way of implementing the HTTP PUT method in a manner consistent with the rest of the .Nm fetch library, .Fn fetchPutHTTP is currently unimplemented. .Sh HTTPS SCHEME Based on HTTP SCHEME. By default the peer is verified using the CA bundle located in .Pa /etc/ssl/cert.pem . The file may contain multiple CA certificates. A common source of a current CA bundle is .Pa \%security/ca_root_nss . .Pp The CA bundle used for peer verification can be changed by setting the environment variables .Ev SSL_CA_CERT_FILE to point to a concatenated bundle of trusted certificates and .Ev SSL_CA_CERT_PATH to point to a directory containing hashes of trusted CAs (see .Xr verify 1 ) . .Pp A certificate revocation list (CRL) can be used by setting the environment variable .Ev SSL_CRL_FILE (see .Xr crl 1 ) . .Pp Peer verification can be disabled by setting the environment variable .Ev SSL_NO_VERIFY_PEER . Note that this also disables CRL checking. .Pp By default the service identity is verified according to the rules detailed in RFC6125 (also known as hostname verification). This feature can be disabled by setting the environment variable .Ev SSL_NO_VERIFY_HOSTNAME . .Pp Client certificate based authentication is supported. The environment variable .Ev SSL_CLIENT_CERT_FILE should be set to point to a file containing key and client certificate to be used in PEM format. In case the key is stored in a separate file, the environment variable .Ev SSL_CLIENT_KEY_FILE can be set to point to the key in PEM format. In case the key uses a password, the user will be prompted on standard input (see .Xr PEM 3 ) . .Pp By default .Nm libfetch allows TLSv1 and newer when negotiating the connecting with the remote peer. You can change this behavior by setting the -.Ev SSL_ALLOW_SSL2 -and .Ev SSL_ALLOW_SSL3 -environment variables to allow SSLv2 and SSLv3, respectively, and +environment variable to allow SSLv3 and .Ev SSL_NO_TLS1 , .Ev SSL_NO_TLS1_1 and .Ev SSL_NO_TLS1_2 to disable TLS 1.0, 1.1 and 1.2 respectively. .Sh AUTHENTICATION Apart from setting the appropriate environment variables and specifying the user name and password in the URL or the .Vt struct url , the calling program has the option of defining an authentication function with the following prototype: .Pp .Ft int .Fn myAuthMethod "struct url *u" .Pp The callback function should fill in the .Fa user and .Fa pwd fields in the provided .Vt struct url and return 0 on success, or any other value to indicate failure. .Pp To register the authentication callback, simply set .Va fetchAuthMethod to point at it. The callback will be used whenever a site requires authentication and the appropriate environment variables are not set. .Pp This interface is experimental and may be subject to change. .Sh RETURN VALUES .Fn fetchParseURL returns a pointer to a .Vt struct url containing the individual components of the URL. If it is unable to allocate memory, or the URL is syntactically incorrect, .Fn fetchParseURL returns a NULL pointer. .Pp The .Fn fetchStat functions return 0 on success and -1 on failure. .Pp All other functions return a stream pointer which may be used to access the requested document, or NULL if an error occurred. .Pp The following error codes are defined in .In fetch.h : .Bl -tag -width 18n .It Bq Er FETCH_ABORT Operation aborted .It Bq Er FETCH_AUTH Authentication failed .It Bq Er FETCH_DOWN Service unavailable .It Bq Er FETCH_EXISTS File exists .It Bq Er FETCH_FULL File system full .It Bq Er FETCH_INFO Informational response .It Bq Er FETCH_MEMORY Insufficient memory .It Bq Er FETCH_MOVED File has moved .It Bq Er FETCH_NETWORK Network error .It Bq Er FETCH_OK No error .It Bq Er FETCH_PROTO Protocol error .It Bq Er FETCH_RESOLV Resolver error .It Bq Er FETCH_SERVER Server error .It Bq Er FETCH_TEMP Temporary error .It Bq Er FETCH_TIMEOUT Operation timed out .It Bq Er FETCH_UNAVAIL File is not available .It Bq Er FETCH_UNKNOWN Unknown error .It Bq Er FETCH_URL Invalid URL .El .Pp The accompanying error message includes a protocol-specific error code and message, e.g.\& "File is not available (404 Not Found)" .Sh ENVIRONMENT .Bl -tag -width ".Ev FETCH_BIND_ADDRESS" .It Ev FETCH_BIND_ADDRESS Specifies a hostname or IP address to which sockets used for outgoing connections will be bound. .It Ev FTP_LOGIN Default FTP login if none was provided in the URL. .It Ev FTP_PASSIVE_MODE If set to .Ql no , forces the FTP code to use active mode. If set to any other value, forces passive mode even if the application requested active mode. .It Ev FTP_PASSWORD Default FTP password if the remote server requests one and none was provided in the URL. .It Ev FTP_PROXY URL of the proxy to use for FTP requests. The document part is ignored. FTP and HTTP proxies are supported; if no scheme is specified, FTP is assumed. If the proxy is an FTP proxy, .Nm libfetch will send .Ql user@host as user name to the proxy, where .Ql user is the real user name, and .Ql host is the name of the FTP server. .Pp If this variable is set to an empty string, no proxy will be used for FTP requests, even if the .Ev HTTP_PROXY variable is set. .It Ev ftp_proxy Same as .Ev FTP_PROXY , for compatibility. .It Ev HTTP_ACCEPT Specifies the value of the .Va Accept header for HTTP requests. If empty, no .Va Accept header is sent. The default is .Dq */* . .It Ev HTTP_AUTH Specifies HTTP authorization parameters as a colon-separated list of items. The first and second item are the authorization scheme and realm respectively; further items are scheme-dependent. Currently, the .Dq basic and .Dq digest authorization methods are supported. .Pp Both methods require two parameters: the user name and password, in that order. .Pp This variable is only used if the server requires authorization and no user name or password was specified in the URL. .It Ev HTTP_PROXY URL of the proxy to use for HTTP requests. The document part is ignored. Only HTTP proxies are supported for HTTP requests. If no port number is specified, the default is 3128. .Pp Note that this proxy will also be used for FTP documents, unless the .Ev FTP_PROXY variable is set. .It Ev http_proxy Same as .Ev HTTP_PROXY , for compatibility. .It Ev HTTP_PROXY_AUTH Specifies authorization parameters for the HTTP proxy in the same format as the .Ev HTTP_AUTH variable. .Pp This variable is used if and only if connected to an HTTP proxy, and is ignored if a user and/or a password were specified in the proxy URL. .It Ev HTTP_REFERER Specifies the referrer URL to use for HTTP requests. If set to .Dq auto , the document URL will be used as referrer URL. .It Ev HTTP_USER_AGENT Specifies the User-Agent string to use for HTTP requests. This can be useful when working with HTTP origin or proxy servers that differentiate between user agents. If defined but empty, no User-Agent header is sent. .It Ev NETRC Specifies a file to use instead of .Pa ~/.netrc -to look up login names and passwords for FTP sites. +to look up login names and passwords for FTP and HTTP sites as well as +HTTP proxies. See .Xr ftp 1 for a description of the file format. -This feature is experimental. .It Ev NO_PROXY Either a single asterisk, which disables the use of proxies altogether, or a comma- or whitespace-separated list of hosts for which proxies should not be used. .It Ev no_proxy Same as .Ev NO_PROXY , for compatibility. -.It Ev SSL_ALLOW_SSL2 -Allow SSL version 2 when negotiating the connection (not recommended). .It Ev SSL_ALLOW_SSL3 Allow SSL version 3 when negotiating the connection (not recommended). .It Ev SSL_CA_CERT_FILE CA certificate bundle containing trusted CA certificates. Default value: .Pa /etc/ssl/cert.pem . .It Ev SSL_CA_CERT_PATH Path containing trusted CA hashes. .It Ev SSL_CLIENT_CERT_FILE PEM encoded client certificate/key which will be used in client certificate authentication. .It Ev SSL_CLIENT_KEY_FILE PEM encoded client key in case key and client certificate are stored separately. .It Ev SSL_CRL_FILE File containing certificate revocation list. .It Ev SSL_NO_TLS1 Do not allow TLS version 1.0 when negotiating the connection. .It Ev SSL_NO_TLS1_1 Do not allow TLS version 1.1 when negotiating the connection. .It Ev SSL_NO_TLS1_2 Do not allow TLS version 1.2 when negotiating the connection. .It Ev SSL_NO_VERIFY_HOSTNAME If set, do not verify that the hostname matches the subject of the certificate presented by the server. .It Ev SSL_NO_VERIFY_PEER If set, do not verify the peer certificate against trusted CAs. .El .Sh EXAMPLES To access a proxy server on .Pa proxy.example.com port 8080, set the .Ev HTTP_PROXY environment variable in a manner similar to this: .Pp .Dl HTTP_PROXY=http://proxy.example.com:8080 .Pp If the proxy server requires authentication, there are two options available for passing the authentication data. The first method is by using the proxy URL: .Pp .Dl HTTP_PROXY=http://:@proxy.example.com:8080 .Pp The second method is by using the .Ev HTTP_PROXY_AUTH environment variable: .Bd -literal -offset indent HTTP_PROXY=http://proxy.example.com:8080 HTTP_PROXY_AUTH=basic:*:: .Ed .Pp To disable the use of a proxy for an HTTP server running on the local host, define .Ev NO_PROXY as follows: .Bd -literal -offset indent NO_PROXY=localhost,127.0.0.1 .Ed .Pp Access HTTPS website without any certificate verification whatsoever: .Bd -literal -offset indent SSL_NO_VERIFY_PEER=1 SSL_NO_VERIFY_HOSTNAME=1 .Ed .Pp Access HTTPS website using client certificate based authentication and a private CA: .Bd -literal -offset indent SSL_CLIENT_CERT_FILE=/path/to/client.pem SSL_CA_CERT_FILE=/path/to/myca.pem .Ed .Sh SEE ALSO .Xr fetch 1 , .Xr ftpio 3 , .Xr ip 4 .Rs .%A J. Postel .%A J. K. Reynolds .%D October 1985 .%B File Transfer Protocol .%O RFC959 .Re .Rs .%A P. Deutsch .%A A. Emtage .%A A. Marine. .%D May 1994 .%T How to Use Anonymous FTP .%O RFC1635 .Re .Rs .%A T. Berners-Lee .%A L. Masinter .%A M. McCahill .%D December 1994 .%T Uniform Resource Locators (URL) .%O RFC1738 .Re .Rs .%A R. Fielding .%A J. Gettys .%A J. Mogul .%A H. Frystyk .%A L. Masinter .%A P. Leach .%A T. Berners-Lee .%D January 1999 .%B Hypertext Transfer Protocol -- HTTP/1.1 .%O RFC2616 .Re .Rs .%A J. Franks .%A P. Hallam-Baker .%A J. Hostetler .%A S. Lawrence .%A P. Leach .%A A. Luotonen .%A L. Stewart .%D June 1999 .%B HTTP Authentication: Basic and Digest Access Authentication .%O RFC2617 .Re .Sh HISTORY The .Nm fetch library first appeared in .Fx 3.0 . .Sh AUTHORS .An -nosplit The .Nm fetch library was mostly written by .An Dag-Erling Sm\(/orgrav Aq des@FreeBSD.org with numerous suggestions and contributions from .An Jordan K. Hubbard Aq jkh@FreeBSD.org , .An Eugene Skepner Aq eu@qub.com , .An Hajimu Umemoto Aq ume@FreeBSD.org , .An Henry Whincup Aq henry@techiebod.com , .An Jukka A. Ukkonen Aq jau@iki.fi , .An Jean-Fran\(,cois Dockes Aq jf@dockes.org , .An Michael Gmelin Aq freebsd@grem.de and others. It replaces the older .Nm ftpio library written by .An Poul-Henning Kamp Aq phk@FreeBSD.org and .An Jordan K. Hubbard Aq jkh@FreeBSD.org . .Pp This manual page was written by .An Dag-Erling Sm\(/orgrav Aq des@FreeBSD.org and .An Michael Gmelin Aq freebsd@grem.de . .Sh BUGS Some parts of the library are not yet implemented. The most notable examples of this are .Fn fetchPutHTTP , .Fn fetchListHTTP , .Fn fetchListFTP and FTP proxy support. .Pp There is no way to select a proxy at run-time other than setting the .Ev HTTP_PROXY or .Ev FTP_PROXY environment variables as appropriate. .Pp .Nm libfetch does not understand or obey 305 (Use Proxy) replies. .Pp Error numbers are unique only within a certain context; the error codes used for FTP and HTTP overlap, as do those used for resolver and system errors. For instance, error code 202 means "Command not implemented, superfluous at this site" in an FTP context and "Accepted" in an HTTP context. .Pp .Fn fetchStatFTP does not check that the result of an MDTM command is a valid date. .Pp In case password protected keys are used for client certificate based authentication the user is prompted for the password on each and every fetch operation. .Pp The man page is incomplete, poorly written and produces badly formatted text. .Pp The error reporting mechanism is unsatisfactory. .Pp Some parts of the code are not fully reentrant. Index: stable/10/lib/libfetch/file.c =================================================================== --- stable/10/lib/libfetch/file.c (revision 294193) +++ stable/10/lib/libfetch/file.c (revision 294194) @@ -1,155 +1,153 @@ /*- * Copyright (c) 1998-2011 Dag-Erling Smørgrav * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer * in this position and unchanged. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. The name of the author may not be used to endorse or promote products * derived from this software without specific prior written permission * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include "fetch.h" #include "common.h" FILE * fetchXGetFile(struct url *u, struct url_stat *us, const char *flags) { FILE *f; if (us && fetchStatFile(u, us, flags) == -1) return (NULL); - f = fopen(u->doc, "r"); + f = fopen(u->doc, "re"); if (f == NULL) { fetch_syserr(); return (NULL); } if (u->offset && fseeko(f, u->offset, SEEK_SET) == -1) { fclose(f); fetch_syserr(); return (NULL); } - fcntl(fileno(f), F_SETFD, FD_CLOEXEC); return (f); } FILE * fetchGetFile(struct url *u, const char *flags) { return (fetchXGetFile(u, NULL, flags)); } FILE * fetchPutFile(struct url *u, const char *flags) { FILE *f; if (CHECK_FLAG('a')) - f = fopen(u->doc, "a"); + f = fopen(u->doc, "ae"); else - f = fopen(u->doc, "w+"); + f = fopen(u->doc, "w+e"); if (f == NULL) { fetch_syserr(); return (NULL); } if (u->offset && fseeko(f, u->offset, SEEK_SET) == -1) { fclose(f); fetch_syserr(); return (NULL); } - fcntl(fileno(f), F_SETFD, FD_CLOEXEC); return (f); } static int fetch_stat_file(const char *fn, struct url_stat *us) { struct stat sb; us->size = -1; us->atime = us->mtime = 0; if (stat(fn, &sb) == -1) { fetch_syserr(); return (-1); } us->size = sb.st_size; us->atime = sb.st_atime; us->mtime = sb.st_mtime; return (0); } int fetchStatFile(struct url *u, struct url_stat *us, const char *flags __unused) { return (fetch_stat_file(u->doc, us)); } struct url_ent * fetchListFile(struct url *u, const char *flags __unused) { struct dirent *de; struct url_stat us; struct url_ent *ue; int size, len; char fn[PATH_MAX], *p; DIR *dir; int l; if ((dir = opendir(u->doc)) == NULL) { fetch_syserr(); return (NULL); } ue = NULL; strncpy(fn, u->doc, sizeof(fn) - 2); fn[sizeof(fn) - 2] = 0; strcat(fn, "/"); p = strchr(fn, 0); l = sizeof(fn) - strlen(fn) - 1; while ((de = readdir(dir)) != NULL) { strncpy(p, de->d_name, l - 1); p[l - 1] = 0; if (fetch_stat_file(fn, &us) == -1) /* should I return a partial result, or abort? */ break; fetch_add_entry(&ue, &size, &len, de->d_name, &us); } return (ue); } Index: stable/10/lib/libfetch/http.c =================================================================== --- stable/10/lib/libfetch/http.c (revision 294193) +++ stable/10/lib/libfetch/http.c (revision 294194) @@ -1,2073 +1,2106 @@ /*- * Copyright (c) 2000-2014 Dag-Erling Smørgrav * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer * in this position and unchanged. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. The name of the author may not be used to endorse or promote products * derived from this software without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include __FBSDID("$FreeBSD$"); /* * The following copyright applies to the base64 code: * *- * Copyright 1997 Massachusetts Institute of Technology * * Permission to use, copy, modify, and distribute this software and * its documentation for any purpose and without fee is hereby * granted, provided that both the above copyright notice and this * permission notice appear in all copies, that both the above * copyright notice and this permission notice appear in all * supporting documentation, and that the name of M.I.T. not be used * in advertising or publicity pertaining to distribution of the * software without specific, written prior permission. M.I.T. makes * no representations about the suitability of this software for any * purpose. It is provided "as is" without express or implied * warranty. * * THIS SOFTWARE IS PROVIDED BY M.I.T. ``AS IS''. M.I.T. DISCLAIMS * ALL EXPRESS OR IMPLIED WARRANTIES WITH REGARD TO THIS SOFTWARE, * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT * SHALL M.I.T. BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef WITH_SSL #include #define MD5Init(c) MD5_Init(c) #define MD5Update(c, data, len) MD5_Update(c, data, len) #define MD5Final(md, c) MD5_Final(md, c) #else #include #endif #include #include #include "fetch.h" #include "common.h" #include "httperr.h" /* Maximum number of redirects to follow */ #define MAX_REDIRECT 20 /* Symbolic names for reply codes we care about */ #define HTTP_OK 200 #define HTTP_PARTIAL 206 #define HTTP_MOVED_PERM 301 #define HTTP_MOVED_TEMP 302 #define HTTP_SEE_OTHER 303 #define HTTP_NOT_MODIFIED 304 #define HTTP_USE_PROXY 305 #define HTTP_TEMP_REDIRECT 307 #define HTTP_PERM_REDIRECT 308 #define HTTP_NEED_AUTH 401 #define HTTP_NEED_PROXY_AUTH 407 #define HTTP_BAD_RANGE 416 #define HTTP_PROTOCOL_ERROR 999 #define HTTP_REDIRECT(xyz) ((xyz) == HTTP_MOVED_PERM \ || (xyz) == HTTP_MOVED_TEMP \ || (xyz) == HTTP_TEMP_REDIRECT \ || (xyz) == HTTP_USE_PROXY \ || (xyz) == HTTP_SEE_OTHER) #define HTTP_ERROR(xyz) ((xyz) > 400 && (xyz) < 599) /***************************************************************************** * I/O functions for decoding chunked streams */ struct httpio { conn_t *conn; /* connection */ int chunked; /* chunked mode */ char *buf; /* chunk buffer */ size_t bufsize; /* size of chunk buffer */ - ssize_t buflen; /* amount of data currently in buffer */ - int bufpos; /* current read offset in buffer */ + size_t buflen; /* amount of data currently in buffer */ + size_t bufpos; /* current read offset in buffer */ int eof; /* end-of-file flag */ int error; /* error flag */ size_t chunksize; /* remaining size of current chunk */ #ifndef NDEBUG size_t total; #endif }; /* * Get next chunk header */ static int http_new_chunk(struct httpio *io) { char *p; if (fetch_getln(io->conn) == -1) return (-1); if (io->conn->buflen < 2 || !isxdigit((unsigned char)*io->conn->buf)) return (-1); for (p = io->conn->buf; *p && !isspace((unsigned char)*p); ++p) { if (*p == ';') break; if (!isxdigit((unsigned char)*p)) return (-1); if (isdigit((unsigned char)*p)) { io->chunksize = io->chunksize * 16 + *p - '0'; } else { io->chunksize = io->chunksize * 16 + 10 + tolower((unsigned char)*p) - 'a'; } } #ifndef NDEBUG if (fetchDebug) { io->total += io->chunksize; if (io->chunksize == 0) fprintf(stderr, "%s(): end of last chunk\n", __func__); else fprintf(stderr, "%s(): new chunk: %lu (%lu)\n", __func__, (unsigned long)io->chunksize, (unsigned long)io->total); } #endif return (io->chunksize); } /* * Grow the input buffer to at least len bytes */ static inline int http_growbuf(struct httpio *io, size_t len) { char *tmp; if (io->bufsize >= len) return (0); if ((tmp = realloc(io->buf, len)) == NULL) return (-1); io->buf = tmp; io->bufsize = len; return (0); } /* * Fill the input buffer, do chunk decoding on the fly */ static ssize_t http_fillbuf(struct httpio *io, size_t len) { ssize_t nbytes; char ch; if (io->error) return (-1); if (io->eof) return (0); + /* not chunked: just fetch the requested amount */ if (io->chunked == 0) { if (http_growbuf(io, len) == -1) return (-1); if ((nbytes = fetch_read(io->conn, io->buf, len)) == -1) { io->error = errno; return (-1); } io->buflen = nbytes; io->bufpos = 0; return (io->buflen); } + /* chunked, but we ran out: get the next chunk header */ if (io->chunksize == 0) { switch (http_new_chunk(io)) { case -1: io->error = EPROTO; return (-1); case 0: io->eof = 1; return (0); } } + /* fetch the requested amount, but no more than the current chunk */ if (len > io->chunksize) len = io->chunksize; if (http_growbuf(io, len) == -1) return (-1); if ((nbytes = fetch_read(io->conn, io->buf, len)) == -1) { io->error = errno; return (-1); } + io->bufpos = 0; io->buflen = nbytes; - io->chunksize -= io->buflen; + io->chunksize -= nbytes; if (io->chunksize == 0) { if (fetch_read(io->conn, &ch, 1) != 1 || ch != '\r' || fetch_read(io->conn, &ch, 1) != 1 || ch != '\n') return (-1); } - io->bufpos = 0; - return (io->buflen); } /* * Read function */ static int http_readfn(void *v, char *buf, int len) { struct httpio *io = (struct httpio *)v; int rlen; if (io->error) return (-1); if (io->eof) return (0); /* empty buffer */ if (!io->buf || io->bufpos == io->buflen) { if ((rlen = http_fillbuf(io, len)) < 0) { if ((errno = io->error) == EINTR) io->error = 0; return (-1); } else if (rlen == 0) { return (0); } } rlen = io->buflen - io->bufpos; if (len < rlen) rlen = len; memcpy(buf, io->buf + io->bufpos, rlen); io->bufpos += rlen; return (rlen); } /* * Write function */ static int http_writefn(void *v, const char *buf, int len) { struct httpio *io = (struct httpio *)v; return (fetch_write(io->conn, buf, len)); } /* * Close function */ static int http_closefn(void *v) { struct httpio *io = (struct httpio *)v; int r; r = fetch_close(io->conn); if (io->buf) free(io->buf); free(io); return (r); } /* * Wrap a file descriptor up */ static FILE * http_funopen(conn_t *conn, int chunked) { struct httpio *io; FILE *f; if ((io = calloc(1, sizeof(*io))) == NULL) { fetch_syserr(); return (NULL); } io->conn = conn; io->chunked = chunked; f = funopen(io, http_readfn, http_writefn, NULL, http_closefn); if (f == NULL) { fetch_syserr(); free(io); return (NULL); } return (f); } /***************************************************************************** * Helper functions for talking to the server and parsing its replies */ /* Header types */ typedef enum { hdr_syserror = -2, hdr_error = -1, hdr_end = 0, hdr_unknown = 1, hdr_content_length, hdr_content_range, hdr_last_modified, hdr_location, hdr_transfer_encoding, hdr_www_authenticate, hdr_proxy_authenticate, } hdr_t; /* Names of interesting headers */ static struct { hdr_t num; const char *name; } hdr_names[] = { { hdr_content_length, "Content-Length" }, { hdr_content_range, "Content-Range" }, { hdr_last_modified, "Last-Modified" }, { hdr_location, "Location" }, { hdr_transfer_encoding, "Transfer-Encoding" }, { hdr_www_authenticate, "WWW-Authenticate" }, { hdr_proxy_authenticate, "Proxy-Authenticate" }, { hdr_unknown, NULL }, }; /* * Send a formatted line; optionally echo to terminal */ static int http_cmd(conn_t *conn, const char *fmt, ...) { va_list ap; size_t len; char *msg; int r; va_start(ap, fmt); len = vasprintf(&msg, fmt, ap); va_end(ap); if (msg == NULL) { errno = ENOMEM; fetch_syserr(); return (-1); } r = fetch_putln(conn, msg, len); free(msg); if (r == -1) { fetch_syserr(); return (-1); } return (0); } /* * Get and parse status line */ static int http_get_reply(conn_t *conn) { char *p; if (fetch_getln(conn) == -1) return (-1); /* * A valid status line looks like "HTTP/m.n xyz reason" where m * and n are the major and minor protocol version numbers and xyz * is the reply code. * Unfortunately, there are servers out there (NCSA 1.5.1, to name * just one) that do not send a version number, so we can't rely * on finding one, but if we do, insist on it being 1.0 or 1.1. * We don't care about the reason phrase. */ if (strncmp(conn->buf, "HTTP", 4) != 0) return (HTTP_PROTOCOL_ERROR); p = conn->buf + 4; if (*p == '/') { if (p[1] != '1' || p[2] != '.' || (p[3] != '0' && p[3] != '1')) return (HTTP_PROTOCOL_ERROR); p += 4; } if (*p != ' ' || !isdigit((unsigned char)p[1]) || !isdigit((unsigned char)p[2]) || !isdigit((unsigned char)p[3])) return (HTTP_PROTOCOL_ERROR); conn->err = (p[1] - '0') * 100 + (p[2] - '0') * 10 + (p[3] - '0'); return (conn->err); } /* * Check a header; if the type matches the given string, return a pointer * to the beginning of the value. */ static const char * http_match(const char *str, const char *hdr) { while (*str && *hdr && tolower((unsigned char)*str++) == tolower((unsigned char)*hdr++)) /* nothing */; if (*str || *hdr != ':') return (NULL); while (*hdr && isspace((unsigned char)*++hdr)) /* nothing */; return (hdr); } /* * Get the next header and return the appropriate symbolic code. We * need to read one line ahead for checking for a continuation line * belonging to the current header (continuation lines start with * white space). * * We get called with a fresh line already in the conn buffer, either * from the previous http_next_header() invocation, or, the first * time, from a fetch_getln() performed by our caller. * * This stops when we encounter an empty line (we dont read beyond the header * area). * * Note that the "headerbuf" is just a place to return the result. Its * contents are not used for the next call. This means that no cleanup * is needed when ie doing another connection, just call the cleanup when * fully done to deallocate memory. */ /* Limit the max number of continuation lines to some reasonable value */ #define HTTP_MAX_CONT_LINES 10 /* Place into which to build a header from one or several lines */ typedef struct { char *buf; /* buffer */ size_t bufsize; /* buffer size */ size_t buflen; /* length of buffer contents */ } http_headerbuf_t; static void init_http_headerbuf(http_headerbuf_t *buf) { buf->buf = NULL; buf->bufsize = 0; buf->buflen = 0; } static void clean_http_headerbuf(http_headerbuf_t *buf) { if (buf->buf) free(buf->buf); init_http_headerbuf(buf); } /* Remove whitespace at the end of the buffer */ static void http_conn_trimright(conn_t *conn) { while (conn->buflen && isspace((unsigned char)conn->buf[conn->buflen - 1])) conn->buflen--; conn->buf[conn->buflen] = '\0'; } static hdr_t http_next_header(conn_t *conn, http_headerbuf_t *hbuf, const char **p) { unsigned int i, len; /* * Have to do the stripping here because of the first line. So * it's done twice for the subsequent lines. No big deal */ http_conn_trimright(conn); if (conn->buflen == 0) return (hdr_end); /* Copy the line to the headerbuf */ if (hbuf->bufsize < conn->buflen + 1) { if ((hbuf->buf = realloc(hbuf->buf, conn->buflen + 1)) == NULL) return (hdr_syserror); hbuf->bufsize = conn->buflen + 1; } strcpy(hbuf->buf, conn->buf); hbuf->buflen = conn->buflen; /* * Fetch possible continuation lines. Stop at 1st non-continuation * and leave it in the conn buffer */ for (i = 0; i < HTTP_MAX_CONT_LINES; i++) { if (fetch_getln(conn) == -1) return (hdr_syserror); /* * Note: we carry on the idea from the previous version * that a pure whitespace line is equivalent to an empty * one (so it's not continuation and will be handled when * we are called next) */ http_conn_trimright(conn); if (conn->buf[0] != ' ' && conn->buf[0] != "\t"[0]) break; /* Got a continuation line. Concatenate to previous */ len = hbuf->buflen + conn->buflen; if (hbuf->bufsize < len + 1) { len *= 2; if ((hbuf->buf = realloc(hbuf->buf, len + 1)) == NULL) return (hdr_syserror); hbuf->bufsize = len + 1; } strcpy(hbuf->buf + hbuf->buflen, conn->buf); hbuf->buflen += conn->buflen; } /* * We could check for malformed headers but we don't really care. * A valid header starts with a token immediately followed by a * colon; a token is any sequence of non-control, non-whitespace * characters except "()<>@,;:\\\"{}". */ for (i = 0; hdr_names[i].num != hdr_unknown; i++) if ((*p = http_match(hdr_names[i].name, hbuf->buf)) != NULL) return (hdr_names[i].num); return (hdr_unknown); } /************************** * [Proxy-]Authenticate header parsing */ /* * Read doublequote-delimited string into output buffer obuf (allocated * by caller, whose responsibility it is to ensure that it's big enough) * cp points to the first char after the initial '"' * Handles \ quoting * Returns pointer to the first char after the terminating double quote, or * NULL for error. */ static const char * http_parse_headerstring(const char *cp, char *obuf) { for (;;) { switch (*cp) { case 0: /* Unterminated string */ *obuf = 0; return (NULL); case '"': /* Ending quote */ *obuf = 0; return (++cp); case '\\': if (*++cp == 0) { *obuf = 0; return (NULL); } /* FALLTHROUGH */ default: *obuf++ = *cp++; } } } /* Http auth challenge schemes */ typedef enum {HTTPAS_UNKNOWN, HTTPAS_BASIC,HTTPAS_DIGEST} http_auth_schemes_t; /* Data holder for a Basic or Digest challenge. */ typedef struct { http_auth_schemes_t scheme; char *realm; char *qop; char *nonce; char *opaque; char *algo; int stale; int nc; /* Nonce count */ } http_auth_challenge_t; static void init_http_auth_challenge(http_auth_challenge_t *b) { b->scheme = HTTPAS_UNKNOWN; b->realm = b->qop = b->nonce = b->opaque = b->algo = NULL; b->stale = b->nc = 0; } static void clean_http_auth_challenge(http_auth_challenge_t *b) { if (b->realm) free(b->realm); if (b->qop) free(b->qop); if (b->nonce) free(b->nonce); if (b->opaque) free(b->opaque); if (b->algo) free(b->algo); init_http_auth_challenge(b); } /* Data holder for an array of challenges offered in an http response. */ #define MAX_CHALLENGES 10 typedef struct { http_auth_challenge_t *challenges[MAX_CHALLENGES]; int count; /* Number of parsed challenges in the array */ int valid; /* We did parse an authenticate header */ } http_auth_challenges_t; static void init_http_auth_challenges(http_auth_challenges_t *cs) { int i; for (i = 0; i < MAX_CHALLENGES; i++) cs->challenges[i] = NULL; cs->count = cs->valid = 0; } static void clean_http_auth_challenges(http_auth_challenges_t *cs) { int i; /* We rely on non-zero pointers being allocated, not on the count */ for (i = 0; i < MAX_CHALLENGES; i++) { if (cs->challenges[i] != NULL) { clean_http_auth_challenge(cs->challenges[i]); free(cs->challenges[i]); } } init_http_auth_challenges(cs); } /* * Enumeration for lexical elements. Separators will be returned as their own * ascii value */ typedef enum {HTTPHL_WORD=256, HTTPHL_STRING=257, HTTPHL_END=258, HTTPHL_ERROR = 259} http_header_lex_t; /* * Determine what kind of token comes next and return possible value * in buf, which is supposed to have been allocated big enough by * caller. Advance input pointer and return element type. */ static int http_header_lex(const char **cpp, char *buf) { size_t l; /* Eat initial whitespace */ *cpp += strspn(*cpp, " \t"); if (**cpp == 0) return (HTTPHL_END); /* Separator ? */ if (**cpp == ',' || **cpp == '=') return (*((*cpp)++)); /* String ? */ if (**cpp == '"') { *cpp = http_parse_headerstring(++*cpp, buf); if (*cpp == NULL) return (HTTPHL_ERROR); return (HTTPHL_STRING); } /* Read other token, until separator or whitespace */ l = strcspn(*cpp, " \t,="); memcpy(buf, *cpp, l); buf[l] = 0; *cpp += l; return (HTTPHL_WORD); } /* * Read challenges from http xxx-authenticate header and accumulate them * in the challenges list structure. * * Headers with multiple challenges are specified by rfc2617, but * servers (ie: squid) often send them in separate headers instead, * which in turn is forbidden by the http spec (multiple headers with * the same name are only allowed for pure comma-separated lists, see * rfc2616 sec 4.2). * * We support both approaches anyway */ static int http_parse_authenticate(const char *cp, http_auth_challenges_t *cs) { int ret = -1; http_header_lex_t lex; char *key = malloc(strlen(cp) + 1); char *value = malloc(strlen(cp) + 1); char *buf = malloc(strlen(cp) + 1); if (key == NULL || value == NULL || buf == NULL) { fetch_syserr(); goto out; } /* In any case we've seen the header and we set the valid bit */ cs->valid = 1; /* Need word first */ lex = http_header_lex(&cp, key); if (lex != HTTPHL_WORD) goto out; /* Loop on challenges */ for (; cs->count < MAX_CHALLENGES; cs->count++) { cs->challenges[cs->count] = malloc(sizeof(http_auth_challenge_t)); if (cs->challenges[cs->count] == NULL) { fetch_syserr(); goto out; } init_http_auth_challenge(cs->challenges[cs->count]); if (!strcasecmp(key, "basic")) { cs->challenges[cs->count]->scheme = HTTPAS_BASIC; } else if (!strcasecmp(key, "digest")) { cs->challenges[cs->count]->scheme = HTTPAS_DIGEST; } else { cs->challenges[cs->count]->scheme = HTTPAS_UNKNOWN; /* * Continue parsing as basic or digest may * follow, and the syntax is the same for * all. We'll just ignore this one when * looking at the list */ } /* Loop on attributes */ for (;;) { /* Key */ lex = http_header_lex(&cp, key); if (lex != HTTPHL_WORD) goto out; /* Equal sign */ lex = http_header_lex(&cp, buf); if (lex != '=') goto out; /* Value */ lex = http_header_lex(&cp, value); if (lex != HTTPHL_WORD && lex != HTTPHL_STRING) goto out; if (!strcasecmp(key, "realm")) cs->challenges[cs->count]->realm = strdup(value); else if (!strcasecmp(key, "qop")) cs->challenges[cs->count]->qop = strdup(value); else if (!strcasecmp(key, "nonce")) cs->challenges[cs->count]->nonce = strdup(value); else if (!strcasecmp(key, "opaque")) cs->challenges[cs->count]->opaque = strdup(value); else if (!strcasecmp(key, "algorithm")) cs->challenges[cs->count]->algo = strdup(value); else if (!strcasecmp(key, "stale")) cs->challenges[cs->count]->stale = strcasecmp(value, "no"); /* Else ignore unknown attributes */ /* Comma or Next challenge or End */ lex = http_header_lex(&cp, key); /* * If we get a word here, this is the beginning of the * next challenge. Break the attributes loop */ if (lex == HTTPHL_WORD) break; if (lex == HTTPHL_END) { /* End while looking for ',' is normal exit */ cs->count++; ret = 0; goto out; } /* Anything else is an error */ if (lex != ',') goto out; } /* End attributes loop */ } /* End challenge loop */ /* * Challenges max count exceeded. This really can't happen * with normal data, something's fishy -> error */ out: if (key) free(key); if (value) free(value); if (buf) free(buf); return (ret); } /* * Parse a last-modified header */ static int http_parse_mtime(const char *p, time_t *mtime) { char locale[64], *r; struct tm tm; strncpy(locale, setlocale(LC_TIME, NULL), sizeof(locale)); setlocale(LC_TIME, "C"); r = strptime(p, "%a, %d %b %Y %H:%M:%S GMT", &tm); /* * Some proxies use UTC in response, but it should still be * parsed. RFC2616 states GMT and UTC are exactly equal for HTTP. */ if (r == NULL) r = strptime(p, "%a, %d %b %Y %H:%M:%S UTC", &tm); /* XXX should add support for date-2 and date-3 */ setlocale(LC_TIME, locale); if (r == NULL) return (-1); DEBUG(fprintf(stderr, "last modified: [%04d-%02d-%02d " "%02d:%02d:%02d]\n", tm.tm_year + 1900, tm.tm_mon + 1, tm.tm_mday, tm.tm_hour, tm.tm_min, tm.tm_sec)); *mtime = timegm(&tm); return (0); } /* * Parse a content-length header */ static int http_parse_length(const char *p, off_t *length) { off_t len; for (len = 0; *p && isdigit((unsigned char)*p); ++p) len = len * 10 + (*p - '0'); if (*p) return (-1); DEBUG(fprintf(stderr, "content length: [%lld]\n", (long long)len)); *length = len; return (0); } /* * Parse a content-range header */ static int http_parse_range(const char *p, off_t *offset, off_t *length, off_t *size) { off_t first, last, len; if (strncasecmp(p, "bytes ", 6) != 0) return (-1); p += 6; if (*p == '*') { first = last = -1; ++p; } else { for (first = 0; *p && isdigit((unsigned char)*p); ++p) first = first * 10 + *p - '0'; if (*p != '-') return (-1); for (last = 0, ++p; *p && isdigit((unsigned char)*p); ++p) last = last * 10 + *p - '0'; } if (first > last || *p != '/') return (-1); for (len = 0, ++p; *p && isdigit((unsigned char)*p); ++p) len = len * 10 + *p - '0'; if (*p || len < last - first + 1) return (-1); if (first == -1) { DEBUG(fprintf(stderr, "content range: [*/%lld]\n", (long long)len)); *length = 0; } else { DEBUG(fprintf(stderr, "content range: [%lld-%lld/%lld]\n", (long long)first, (long long)last, (long long)len)); *length = last - first + 1; } *offset = first; *size = len; return (0); } /***************************************************************************** * Helper functions for authorization */ /* * Base64 encoding */ static char * http_base64(const char *src) { static const char base64[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZ" "abcdefghijklmnopqrstuvwxyz" "0123456789+/"; char *str, *dst; size_t l; int t, r; l = strlen(src); if ((str = malloc(((l + 2) / 3) * 4 + 1)) == NULL) return (NULL); dst = str; r = 0; while (l >= 3) { t = (src[0] << 16) | (src[1] << 8) | src[2]; dst[0] = base64[(t >> 18) & 0x3f]; dst[1] = base64[(t >> 12) & 0x3f]; dst[2] = base64[(t >> 6) & 0x3f]; dst[3] = base64[(t >> 0) & 0x3f]; src += 3; l -= 3; dst += 4; r += 4; } switch (l) { case 2: t = (src[0] << 16) | (src[1] << 8); dst[0] = base64[(t >> 18) & 0x3f]; dst[1] = base64[(t >> 12) & 0x3f]; dst[2] = base64[(t >> 6) & 0x3f]; dst[3] = '='; dst += 4; r += 4; break; case 1: t = src[0] << 16; dst[0] = base64[(t >> 18) & 0x3f]; dst[1] = base64[(t >> 12) & 0x3f]; dst[2] = dst[3] = '='; dst += 4; r += 4; break; case 0: break; } *dst = 0; return (str); } /* * Extract authorization parameters from environment value. * The value is like scheme:realm:user:pass */ typedef struct { char *scheme; char *realm; char *user; char *password; } http_auth_params_t; static void init_http_auth_params(http_auth_params_t *s) { s->scheme = s->realm = s->user = s->password = NULL; } static void clean_http_auth_params(http_auth_params_t *s) { if (s->scheme) free(s->scheme); if (s->realm) free(s->realm); if (s->user) free(s->user); if (s->password) free(s->password); init_http_auth_params(s); } static int http_authfromenv(const char *p, http_auth_params_t *parms) { int ret = -1; char *v, *ve; char *str = strdup(p); if (str == NULL) { fetch_syserr(); return (-1); } v = str; if ((ve = strchr(v, ':')) == NULL) goto out; *ve = 0; if ((parms->scheme = strdup(v)) == NULL) { fetch_syserr(); goto out; } v = ve + 1; if ((ve = strchr(v, ':')) == NULL) goto out; *ve = 0; if ((parms->realm = strdup(v)) == NULL) { fetch_syserr(); goto out; } v = ve + 1; if ((ve = strchr(v, ':')) == NULL) goto out; *ve = 0; if ((parms->user = strdup(v)) == NULL) { fetch_syserr(); goto out; } v = ve + 1; if ((parms->password = strdup(v)) == NULL) { fetch_syserr(); goto out; } ret = 0; out: if (ret == -1) clean_http_auth_params(parms); if (str) free(str); return (ret); } /* * Digest response: the code to compute the digest is taken from the * sample implementation in RFC2616 */ #define IN const #define OUT #define HASHLEN 16 typedef char HASH[HASHLEN]; #define HASHHEXLEN 32 typedef char HASHHEX[HASHHEXLEN+1]; static const char *hexchars = "0123456789abcdef"; static void CvtHex(IN HASH Bin, OUT HASHHEX Hex) { unsigned short i; unsigned char j; for (i = 0; i < HASHLEN; i++) { j = (Bin[i] >> 4) & 0xf; Hex[i*2] = hexchars[j]; j = Bin[i] & 0xf; Hex[i*2+1] = hexchars[j]; } Hex[HASHHEXLEN] = '\0'; }; /* calculate H(A1) as per spec */ static void DigestCalcHA1( IN char * pszAlg, IN char * pszUserName, IN char * pszRealm, IN char * pszPassword, IN char * pszNonce, IN char * pszCNonce, OUT HASHHEX SessionKey ) { MD5_CTX Md5Ctx; HASH HA1; MD5Init(&Md5Ctx); MD5Update(&Md5Ctx, pszUserName, strlen(pszUserName)); MD5Update(&Md5Ctx, ":", 1); MD5Update(&Md5Ctx, pszRealm, strlen(pszRealm)); MD5Update(&Md5Ctx, ":", 1); MD5Update(&Md5Ctx, pszPassword, strlen(pszPassword)); MD5Final(HA1, &Md5Ctx); if (strcasecmp(pszAlg, "md5-sess") == 0) { MD5Init(&Md5Ctx); MD5Update(&Md5Ctx, HA1, HASHLEN); MD5Update(&Md5Ctx, ":", 1); MD5Update(&Md5Ctx, pszNonce, strlen(pszNonce)); MD5Update(&Md5Ctx, ":", 1); MD5Update(&Md5Ctx, pszCNonce, strlen(pszCNonce)); MD5Final(HA1, &Md5Ctx); } CvtHex(HA1, SessionKey); } /* calculate request-digest/response-digest as per HTTP Digest spec */ static void DigestCalcResponse( IN HASHHEX HA1, /* H(A1) */ IN char * pszNonce, /* nonce from server */ IN char * pszNonceCount, /* 8 hex digits */ IN char * pszCNonce, /* client nonce */ IN char * pszQop, /* qop-value: "", "auth", "auth-int" */ IN char * pszMethod, /* method from the request */ IN char * pszDigestUri, /* requested URL */ IN HASHHEX HEntity, /* H(entity body) if qop="auth-int" */ OUT HASHHEX Response /* request-digest or response-digest */ ) { /* DEBUG(fprintf(stderr, "Calc: HA1[%s] Nonce[%s] qop[%s] method[%s] URI[%s]\n", HA1, pszNonce, pszQop, pszMethod, pszDigestUri));*/ MD5_CTX Md5Ctx; HASH HA2; HASH RespHash; HASHHEX HA2Hex; // calculate H(A2) MD5Init(&Md5Ctx); MD5Update(&Md5Ctx, pszMethod, strlen(pszMethod)); MD5Update(&Md5Ctx, ":", 1); MD5Update(&Md5Ctx, pszDigestUri, strlen(pszDigestUri)); if (strcasecmp(pszQop, "auth-int") == 0) { MD5Update(&Md5Ctx, ":", 1); MD5Update(&Md5Ctx, HEntity, HASHHEXLEN); } MD5Final(HA2, &Md5Ctx); CvtHex(HA2, HA2Hex); // calculate response MD5Init(&Md5Ctx); MD5Update(&Md5Ctx, HA1, HASHHEXLEN); MD5Update(&Md5Ctx, ":", 1); MD5Update(&Md5Ctx, pszNonce, strlen(pszNonce)); MD5Update(&Md5Ctx, ":", 1); if (*pszQop) { MD5Update(&Md5Ctx, pszNonceCount, strlen(pszNonceCount)); MD5Update(&Md5Ctx, ":", 1); MD5Update(&Md5Ctx, pszCNonce, strlen(pszCNonce)); MD5Update(&Md5Ctx, ":", 1); MD5Update(&Md5Ctx, pszQop, strlen(pszQop)); MD5Update(&Md5Ctx, ":", 1); } MD5Update(&Md5Ctx, HA2Hex, HASHHEXLEN); MD5Final(RespHash, &Md5Ctx); CvtHex(RespHash, Response); } /* * Generate/Send a Digest authorization header * This looks like: [Proxy-]Authorization: credentials * * credentials = "Digest" digest-response * digest-response = 1#( username | realm | nonce | digest-uri * | response | [ algorithm ] | [cnonce] | * [opaque] | [message-qop] | * [nonce-count] | [auth-param] ) * username = "username" "=" username-value * username-value = quoted-string * digest-uri = "uri" "=" digest-uri-value * digest-uri-value = request-uri ; As specified by HTTP/1.1 * message-qop = "qop" "=" qop-value * cnonce = "cnonce" "=" cnonce-value * cnonce-value = nonce-value * nonce-count = "nc" "=" nc-value * nc-value = 8LHEX * response = "response" "=" request-digest * request-digest = <"> 32LHEX <"> */ static int http_digest_auth(conn_t *conn, const char *hdr, http_auth_challenge_t *c, http_auth_params_t *parms, struct url *url) { int r; char noncecount[10]; char cnonce[40]; char *options = NULL; if (!c->realm || !c->nonce) { DEBUG(fprintf(stderr, "realm/nonce not set in challenge\n")); return(-1); } if (!c->algo) c->algo = strdup(""); if (asprintf(&options, "%s%s%s%s", *c->algo? ",algorithm=" : "", c->algo, c->opaque? ",opaque=" : "", c->opaque?c->opaque:"")== -1) return (-1); if (!c->qop) { c->qop = strdup(""); *noncecount = 0; *cnonce = 0; } else { c->nc++; sprintf(noncecount, "%08x", c->nc); /* We don't try very hard with the cnonce ... */ sprintf(cnonce, "%x%lx", getpid(), (unsigned long)time(0)); } HASHHEX HA1; DigestCalcHA1(c->algo, parms->user, c->realm, parms->password, c->nonce, cnonce, HA1); DEBUG(fprintf(stderr, "HA1: [%s]\n", HA1)); HASHHEX digest; DigestCalcResponse(HA1, c->nonce, noncecount, cnonce, c->qop, "GET", url->doc, "", digest); if (c->qop[0]) { r = http_cmd(conn, "%s: Digest username=\"%s\",realm=\"%s\"," "nonce=\"%s\",uri=\"%s\",response=\"%s\"," "qop=\"auth\", cnonce=\"%s\", nc=%s%s", hdr, parms->user, c->realm, c->nonce, url->doc, digest, cnonce, noncecount, options); } else { r = http_cmd(conn, "%s: Digest username=\"%s\",realm=\"%s\"," "nonce=\"%s\",uri=\"%s\",response=\"%s\"%s", hdr, parms->user, c->realm, c->nonce, url->doc, digest, options); } if (options) free(options); return (r); } /* * Encode username and password */ static int http_basic_auth(conn_t *conn, const char *hdr, const char *usr, const char *pwd) { char *upw, *auth; int r; DEBUG(fprintf(stderr, "basic: usr: [%s]\n", usr)); DEBUG(fprintf(stderr, "basic: pwd: [%s]\n", pwd)); if (asprintf(&upw, "%s:%s", usr, pwd) == -1) return (-1); auth = http_base64(upw); free(upw); if (auth == NULL) return (-1); r = http_cmd(conn, "%s: Basic %s", hdr, auth); free(auth); return (r); } /* * Chose the challenge to answer and call the appropriate routine to * produce the header. */ static int http_authorize(conn_t *conn, const char *hdr, http_auth_challenges_t *cs, http_auth_params_t *parms, struct url *url) { - http_auth_challenge_t *basic = NULL; http_auth_challenge_t *digest = NULL; int i; /* If user or pass are null we're not happy */ if (!parms->user || !parms->password) { DEBUG(fprintf(stderr, "NULL usr or pass\n")); return (-1); } - /* Look for a Digest and a Basic challenge */ + /* Look for a Digest */ for (i = 0; i < cs->count; i++) { - if (cs->challenges[i]->scheme == HTTPAS_BASIC) - basic = cs->challenges[i]; if (cs->challenges[i]->scheme == HTTPAS_DIGEST) digest = cs->challenges[i]; } /* Error if "Digest" was specified and there is no Digest challenge */ if (!digest && (parms->scheme && !strcasecmp(parms->scheme, "digest"))) { DEBUG(fprintf(stderr, "Digest auth in env, not supported by peer\n")); return (-1); } /* * If "basic" was specified in the environment, or there is no Digest * challenge, do the basic thing. Don't need a challenge for this, * so no need to check basic!=NULL */ if (!digest || (parms->scheme && !strcasecmp(parms->scheme,"basic"))) return (http_basic_auth(conn,hdr,parms->user,parms->password)); /* Else, prefer digest. We just checked that it's not NULL */ return (http_digest_auth(conn, hdr, digest, parms, url)); } /***************************************************************************** * Helper functions for connecting to a server or proxy */ /* * Connect to the correct HTTP server or proxy. */ static conn_t * http_connect(struct url *URL, struct url *purl, const char *flags) { struct url *curl; conn_t *conn; + hdr_t h; + http_headerbuf_t headerbuf; + const char *p; int verbose; int af, val; + int serrno; #ifdef INET6 af = AF_UNSPEC; #else af = AF_INET; #endif verbose = CHECK_FLAG('v'); if (CHECK_FLAG('4')) af = AF_INET; #ifdef INET6 else if (CHECK_FLAG('6')) af = AF_INET6; #endif curl = (purl != NULL) ? purl : URL; if ((conn = fetch_connect(curl->host, curl->port, af, verbose)) == NULL) /* fetch_connect() has already set an error code */ return (NULL); + init_http_headerbuf(&headerbuf); if (strcasecmp(URL->scheme, SCHEME_HTTPS) == 0 && purl) { http_cmd(conn, "CONNECT %s:%d HTTP/1.1", URL->host, URL->port); http_cmd(conn, "Host: %s:%d", URL->host, URL->port); http_cmd(conn, ""); if (http_get_reply(conn) != HTTP_OK) { - fetch_close(conn); - return (NULL); + http_seterr(conn->err); + goto ouch; } - http_get_reply(conn); + /* Read and discard the rest of the proxy response */ + if (fetch_getln(conn) < 0) { + fetch_syserr(); + goto ouch; + } + do { + switch ((h = http_next_header(conn, &headerbuf, &p))) { + case hdr_syserror: + fetch_syserr(); + goto ouch; + case hdr_error: + http_seterr(HTTP_PROTOCOL_ERROR); + goto ouch; + default: + /* ignore */ ; + } + } while (h < hdr_end); } if (strcasecmp(URL->scheme, SCHEME_HTTPS) == 0 && fetch_ssl(conn, URL, verbose) == -1) { fetch_close(conn); /* grrr */ errno = EAUTH; fetch_syserr(); - return (NULL); + goto ouch; } val = 1; setsockopt(conn->sd, IPPROTO_TCP, TCP_NOPUSH, &val, sizeof(val)); + clean_http_headerbuf(&headerbuf); return (conn); +ouch: + serrno = errno; + clean_http_headerbuf(&headerbuf); + fetch_close(conn); + errno = serrno; + return (NULL); } static struct url * http_get_proxy(struct url * url, const char *flags) { struct url *purl; char *p; if (flags != NULL && strchr(flags, 'd') != NULL) return (NULL); if (fetch_no_proxy_match(url->host)) return (NULL); if (((p = getenv("HTTP_PROXY")) || (p = getenv("http_proxy"))) && *p && (purl = fetchParseURL(p))) { if (!*purl->scheme) strcpy(purl->scheme, SCHEME_HTTP); if (!purl->port) purl->port = fetch_default_proxy_port(purl->scheme); if (strcasecmp(purl->scheme, SCHEME_HTTP) == 0) return (purl); fetchFreeURL(purl); } return (NULL); } static void http_print_html(FILE *out, FILE *in) { size_t len; char *line, *p, *q; int comment, tag; comment = tag = 0; while ((line = fgetln(in, &len)) != NULL) { while (len && isspace((unsigned char)line[len - 1])) --len; for (p = q = line; q < line + len; ++q) { if (comment && *q == '-') { if (q + 2 < line + len && strcmp(q, "-->") == 0) { tag = comment = 0; q += 2; } } else if (tag && !comment && *q == '>') { p = q + 1; tag = 0; } else if (!tag && *q == '<') { if (q > p) fwrite(p, q - p, 1, out); tag = 1; if (q + 3 < line + len && strcmp(q, "