/* apps/s_socket.c - socket-related functions used by s_client and s_server */ /* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com) * All rights reserved. * * This package is an SSL implementation written * by Eric Young (eay@cryptsoft.com). * The implementation was written so as to conform with Netscapes SSL. * * This library is free for commercial and non-commercial use as long as * the following conditions are aheared to. The following conditions * apply to all code found in this distribution, be it the RC4, RSA, * lhash, DES, etc., code; not just the SSL code. The SSL documentation * included with this distribution is covered by the same copyright terms * except that the holder is Tim Hudson (tjh@cryptsoft.com). * * Copyright remains Eric Young's, and as such any Copyright notices in * the code are not to be removed. * If this package is used in a product, Eric Young should be given attribution * as the author of the parts of the library used. * This can be in the form of a textual message at program startup or * in documentation (online or textual) provided with the package. * * 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 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. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * "This product includes cryptographic software written by * Eric Young (eay@cryptsoft.com)" * The word 'cryptographic' can be left out if the rouines from the library * being used are not cryptographic related :-). * 4. If you include any Windows specific code (or a derivative thereof) from * the apps directory (application code) you must include an acknowledgement: * "This product includes software written by Tim Hudson (tjh@cryptsoft.com)" * * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``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. * * The licence and distribution terms for any publically available version or * derivative of this code cannot be changed. i.e. this code cannot simply be * copied and put under another distribution licence * [including the GNU Public Licence.] */ #include #include #include #include #include /* With IPv6, it looks like Digital has mixed up the proper order of recursive header file inclusion, resulting in the compiler complaining that u_int isn't defined, but only if _POSIX_C_SOURCE is defined, which is needed to have fileno() declared correctly... So let's define u_int */ #if defined(OPENSSL_SYS_VMS_DECC) && !defined(__U_INT) #define __U_INT typedef unsigned int u_int; #endif #define USE_SOCKETS #define NON_MAIN #include "apps.h" #undef USE_SOCKETS #undef NON_MAIN #include "s_apps.h" #include #ifdef FLAT_INC #include "e_os.h" #else #include "../e_os.h" #endif #ifndef OPENSSL_NO_SOCK #if defined(OPENSSL_SYS_NETWARE) && defined(NETWARE_BSDSOCK) #include "netdb.h" #endif static struct hostent *GetHostByName(char *name); #if defined(OPENSSL_SYS_WINDOWS) || (defined(OPENSSL_SYS_NETWARE) && !defined(NETWARE_BSDSOCK)) static void ssl_sock_cleanup(void); #endif static int ssl_sock_init(void); static int init_client_ip(int *sock,unsigned char ip[4], int port, int type); static int init_server(int *sock, int port, int type); static int init_server_long(int *sock, int port,char *ip, int type); static int do_accept(int acc_sock, int *sock, char **host); static int host_ip(char *str, unsigned char ip[4]); #ifdef OPENSSL_SYS_WIN16 #define SOCKET_PROTOCOL 0 /* more microsoft stupidity */ #else #define SOCKET_PROTOCOL IPPROTO_TCP #endif #if defined(OPENSSL_SYS_NETWARE) && !defined(NETWARE_BSDSOCK) static int wsa_init_done=0; #endif #ifdef OPENSSL_SYS_WINDOWS static struct WSAData wsa_state; static int wsa_init_done=0; #ifdef OPENSSL_SYS_WIN16 static HWND topWnd=0; static FARPROC lpTopWndProc=NULL; static FARPROC lpTopHookProc=NULL; extern HINSTANCE _hInstance; /* nice global CRT provides */ static LONG FAR PASCAL topHookProc(HWND hwnd, UINT message, WPARAM wParam, LPARAM lParam) { if (hwnd == topWnd) { switch(message) { case WM_DESTROY: case WM_CLOSE: SetWindowLong(topWnd,GWL_WNDPROC,(LONG)lpTopWndProc); ssl_sock_cleanup(); break; } } return CallWindowProc(lpTopWndProc,hwnd,message,wParam,lParam); } static BOOL CALLBACK enumproc(HWND hwnd,LPARAM lParam) { topWnd=hwnd; return(FALSE); } #endif /* OPENSSL_SYS_WIN32 */ #endif /* OPENSSL_SYS_WINDOWS */ #ifdef OPENSSL_SYS_WINDOWS static void ssl_sock_cleanup(void) { if (wsa_init_done) { wsa_init_done=0; #ifndef OPENSSL_SYS_WINCE WSACancelBlockingCall(); #endif WSACleanup(); } } #elif defined(OPENSSL_SYS_NETWARE) && !defined(NETWARE_BSDSOCK) static void sock_cleanup(void) { if (wsa_init_done) { wsa_init_done=0; WSACleanup(); } } #endif static int ssl_sock_init(void) { #ifdef WATT32 extern int _watt_do_exit; _watt_do_exit = 0; if (sock_init()) return (0); #elif defined(OPENSSL_SYS_WINDOWS) if (!wsa_init_done) { int err; #ifdef SIGINT signal(SIGINT,(void (*)(int))ssl_sock_cleanup); #endif wsa_init_done=1; memset(&wsa_state,0,sizeof(wsa_state)); if (WSAStartup(0x0101,&wsa_state)!=0) { err=WSAGetLastError(); BIO_printf(bio_err,"unable to start WINSOCK, error code=%d\n",err); return(0); } #ifdef OPENSSL_SYS_WIN16 EnumTaskWindows(GetCurrentTask(),enumproc,0L); lpTopWndProc=(FARPROC)GetWindowLong(topWnd,GWL_WNDPROC); lpTopHookProc=MakeProcInstance((FARPROC)topHookProc,_hInstance); SetWindowLong(topWnd,GWL_WNDPROC,(LONG)lpTopHookProc); #endif /* OPENSSL_SYS_WIN16 */ } #elif defined(OPENSSL_SYS_NETWARE) && !defined(NETWARE_BSDSOCK) WORD wVerReq; WSADATA wsaData; int err; if (!wsa_init_done) { # ifdef SIGINT signal(SIGINT,(void (*)(int))sock_cleanup); # endif wsa_init_done=1; wVerReq = MAKEWORD( 2, 0 ); err = WSAStartup(wVerReq,&wsaData); if (err != 0) { BIO_printf(bio_err,"unable to start WINSOCK2, error code=%d\n",err); return(0); } } #endif /* OPENSSL_SYS_WINDOWS */ return(1); } int init_client(int *sock, char *host, int port, int type) { unsigned char ip[4]; short p=0; if (!host_ip(host,&(ip[0]))) { return(0); } if (p != 0) port=p; return(init_client_ip(sock,ip,port,type)); } static int init_client_ip(int *sock, unsigned char ip[4], int port, int type) { unsigned long addr; struct sockaddr_in them; int s,i; if (!ssl_sock_init()) return(0); memset((char *)&them,0,sizeof(them)); them.sin_family=AF_INET; them.sin_port=htons((unsigned short)port); addr=(unsigned long) ((unsigned long)ip[0]<<24L)| ((unsigned long)ip[1]<<16L)| ((unsigned long)ip[2]<< 8L)| ((unsigned long)ip[3]); them.sin_addr.s_addr=htonl(addr); if (type == SOCK_STREAM) s=socket(AF_INET,SOCK_STREAM,SOCKET_PROTOCOL); else /* ( type == SOCK_DGRAM) */ s=socket(AF_INET,SOCK_DGRAM,IPPROTO_UDP); if (s == INVALID_SOCKET) { perror("socket"); return(0); } #ifndef OPENSSL_SYS_MPE if (type == SOCK_STREAM) { i=0; i=setsockopt(s,SOL_SOCKET,SO_KEEPALIVE,(char *)&i,sizeof(i)); if (i < 0) { perror("keepalive"); return(0); } } #endif if (connect(s,(struct sockaddr *)&them,sizeof(them)) == -1) { close(s); perror("connect"); return(0); } *sock=s; return(1); } int do_server(int port, int type, int *ret, int (*cb)(char *hostname, int s, unsigned char *conte
// ====================================================================
// Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
// project.
//
// Rights for redistribution and usage in source and binary forms are
// granted according to the OpenSSL license. Warranty of any kind is
// disclaimed.
// ====================================================================

.ident  "rc4-ia64.S, Version 2.0"
.ident  "IA-64 ISA artwork by Andy Polyakov <appro@fy.chalmers.se>"

// What's wrong with compiler generated code? Because of the nature of
// C language, compiler doesn't [dare to] reorder load and stores. But
// being memory-bound, RC4 should benefit from reorder [on in-order-
// execution core such as IA-64]. But what can we reorder? At the very
// least we can safely reorder references to key schedule in respect
// to input and output streams. Secondly, from the first [close] glance
// it appeared that it's possible to pull up some references to
// elements of the key schedule itself. Original rationale ["prior
// loads are not safe only for "degenerated" key schedule, when some
// elements equal to the same value"] was kind of sloppy. I should have
// formulated as it really was: if we assume that pulling up reference
// to key[x+1] is not safe, then it would mean that key schedule would
// "degenerate," which is never the case. The problem is that this
// holds true in respect to references to key[x], but not to key[y].
// Legitimate "collisions" do occur within every 256^2 bytes window.
// Fortunately there're enough free instruction slots to keep prior
// reference to key[x+1], detect "collision" and compensate for it.
// All this without sacrificing a single clock cycle:-) Throughput is
// ~210MBps on 900MHz CPU, which is is >3x faster than gcc generated
// code and +30% - if compared to HP-UX C. Unrolling loop below should
// give >30% on top of that...

.text
.explicit

#if defined(_HPUX_SOURCE) && !defined(_LP64)
# define ADDP	addp4
#else
# define ADDP	add
#endif

#ifndef SZ
#define SZ	4	// this is set to sizeof(RC4_INT)
#endif
// SZ==4 seems to be optimal. At least SZ==8 is not any faster, not for
// assembler implementation, while SZ==1 code is ~30% slower.
#if SZ==1	// RC4_INT is unsigned char
# define	LDKEY	ld1
# define	STKEY	st1
# define	OFF	0
#elif SZ==4	// RC4_INT is unsigned int
# define	LDKEY	ld4
# define	STKEY	st4
# define	OFF	2
#elif SZ==8	// RC4_INT is unsigned long
# define	LDKEY	ld8
# define	STKEY	st8
# define	OFF	3
#endif

out=r8;		// [expanded] output pointer
inp=r9;		// [expanded] output pointer
prsave=r10;
key=r28;	// [expanded] pointer to RC4_KEY
ksch=r29;	// (key->data+255)[&~(sizeof(key->data)-1)]
xx=r30;
yy=r31;

// void RC4(RC4_KEY *key,size_t len,const void *inp,void *out);
.global	RC4#
.proc	RC4#
.align	32
.skip	16
RC4:
	.prologue
	.fframe 0
	.save   ar.pfs,r2
	.save	ar.lc,r3
	.save	pr,prsave
{ .mii;	alloc	r2=ar.pfs,4,12,0,16
	mov	prsave=pr
	ADDP	key=0,in0		};;
{ .mib;	cmp.eq	p6,p0=0,in1			// len==0?
	mov	r3=ar.lc
(p6)	br.ret.spnt.many	b0	};;	// emergency exit

	.body
	.rotr	dat[4],key_x[4],tx[2],rnd[2],key_y[2],ty[1];

{ .mib;	LDKEY	xx=[key],SZ			// load key->x
	add	in1=-1,in1			// adjust len for loop counter
	nop.b	0			}
{ .mib;	ADDP	inp=0,in2
	ADDP	out=0,in3
	brp.loop.imp	.Ltop,.Lexit-16	};;
{ .mmi;	LDKEY	yy=[key]			// load key->y
	add	ksch=SZ,key
	mov	ar.lc=in1		}
{ .mmi;	mov	key_y[1]=r0			// guarantee inequality
						// in first iteration
	add	xx=1,xx
	mov	pr.rot=1<<16		};;
{ .mii;	nop.m	0
	dep	key_x[1]=xx,r0,OFF,8
	mov	ar.ec=3			};;	// note that epilogue counter
						// is off by 1. I compensate
						// for this at exit...
.Ltop:
// The loop is scheduled for 4*(n+2) spin-rate on Itanium 2, which
// theoretically gives asymptotic performance of clock frequency
// divided by 4 bytes per seconds, or 400MBps on 1.6GHz CPU. This is
// for sizeof(RC4_INT)==4. For smaller RC4_INT STKEY inadvertently
// splits the last bundle and you end up with 5*n spin-rate:-(
// Originally the loop was scheduled for 3*n and relied on key
// schedule to be aligned at 256*sizeof(RC4_INT) boundary. But
// *(out++)=dat, which maps to st1, had same effect [inadvertent
// bundle split] and holded the loop back. Rescheduling for 4*n
// made it possible to eliminate dependence on specific alignment
// and allow OpenSSH keep "abusing" our API. Reaching for 3*n would
// require unrolling, sticking to variable shift instruction for
// collecting output [to avoid starvation for integer shifter] and
// copying of key schedule to controlled place in stack [so that
// deposit instruction can serve as substitute for whole
// key->data+((x&255)<<log2(sizeof(key->data[0])))]...
{ .mmi;	(p19)	st1	[out]=dat[3],1			// *(out++)=dat
	(p16)	add	xx=1,xx				// x++
	(p18)	dep	rnd[1]=rnd[1],r0,OFF,8	}	// ((tx+ty)&255)<<OFF
{ .mmi;	(p16)	add	key_x[1]=ksch,key_x[1]		// &key[xx&255]
	(p17)	add	key_y[1]=ksch,key_y[1]	};;	// &key[yy&255]	
{ .mmi;	(p16)	LDKEY	tx[0]=[key_x[1]]		// tx=key[xx]
	(p17)	LDKEY	ty[0]=[key_y[1]]		// ty=key[yy]	
	(p16)	dep	key_x[0]=xx,r0,OFF,8	}	// (xx&255)<<OFF
{ .mmi;	(p18)	add	rnd[1]=ksch,rnd[1]		// &key[(tx+ty)&255]
	(p16)	cmp.ne.unc p20,p21=key_x[1],key_y[1] };;
{ .mmi;	(p18)	LDKEY	rnd[1]=[rnd[1]]			// rnd=key[(tx+ty)&255]
	(p16)	ld1	dat[0]=[inp],1		}	// dat=*(inp++)
.pred.rel	"mutex",p20,p21
{ .mmi;	(p21)	add	yy=yy,tx[1]			// (p16)
	(p20)	add	yy=yy,tx[0]			// (p16) y+=tx
	(p21)	mov	tx[0]=tx[1]		};;	// (p16)
{ .mmi;	(p17)	STKEY	[key_y[1]]=tx[1]		// key[yy]=tx
	(p17)	STKEY	[key_x[2]]=ty[0]		// key[xx]=ty
	(p16)	dep	key_y[0]=yy,r0,OFF,8	}	// &key[yy&255]
{ .mmb;	(p17)	add	rnd[0]=tx[1],ty[0]		// tx+=ty
	(p18)	xor	dat[2]=dat[2],rnd[1]		// dat^=rnd
	br.ctop.sptk	.Ltop			};;
.Lexit:
{ .mib;	STKEY	[key]=yy,-SZ			// save key->y
	mov	pr=prsave,0x1ffff
	nop.b	0			}
{ .mib;	st1	[out]=dat[3],1			// compensate for truncated
						// epilogue counter
	add	xx=-1,xx
	nop.b	0			};;
{ .mib;	STKEY	[key]=xx			// save key->x
	mov	ar.lc=r3
	br.ret.sptk.many	b0	};;
.endp	RC4#
generated by cgit v1.2.3 (git 2.25.1) at 2025-05-19 18:00:46 +0000