webcookies.c

这是一个同样来自贝尔实验室的和UNIX有着渊源的操作系统, 其简洁的设计和实现易于我们学习和理解

/*
 * Cookie file system.  Allows hget and multiple webfs's to collaborate.
 * Conventionally mounted on /mnt/webcookies.
 */

#include <u.h>
#include <libc.h>
#include <bio.h>
#include <ndb.h>
#include <fcall.h>
#include <thread.h>
#include <9p.h>
#include <ctype.h>

int debug = 0;

typedef struct Cookie Cookie;
typedef struct Jar Jar;

struct Cookie
{
	/* external info */
	char*	name;
	char*	value;
	char*	dom;		/* starts with . */
	char*	path;
	char*	version;
	char*	comment;	/* optional, may be nil */

	uint	expire;		/* time of expiration: ~0 means when webcookies dies */
	int	secure;
	int	explicitdom;	/* dom was explicitly set */
	int	explicitpath;	/* path was explicitly set */
	int	netscapestyle;

	/* internal info */
	int	deleted;
	int	mark;
	int	ondisk;
};

struct Jar
{
	Cookie	*c;
	int	nc;
	int	mc;

	Qid	qid;
	int	dirty;
	char	*file;
	char	*lockfile;
};

struct {
	char	*s;
	int	offset;
	int	ishttp;
} stab[] = {
	"domain",		offsetof(Cookie, dom),		1,
	"path",			offsetof(Cookie, path),		1,
	"name",			offsetof(Cookie, name),		0,
	"value",		offsetof(Cookie, value),	0,
	"comment",		offsetof(Cookie, comment),	1,
	"version",		offsetof(Cookie, version),	1,
};

struct {
	char *s;
	int	offset;
} itab[] = {
	"expire",		offsetof(Cookie, expire),
	"secure",		offsetof(Cookie, secure),
	"explicitdomain",	offsetof(Cookie, explicitdom),
	"explicitpath",		offsetof(Cookie, explicitpath),
	"netscapestyle",	offsetof(Cookie, netscapestyle),
};

#pragma varargck type "J"	Jar*
#pragma varargck type "K"	Cookie*

/* HTTP format */
int
jarfmt(Fmt *fmt)
{
	int i;
	Jar *jar;

	jar = va_arg(fmt->args, Jar*);
	if(jar == nil || jar->nc == 0)
		return fmtstrcpy(fmt, "");

	fmtprint(fmt, "Cookie: ");
	if(jar->c[0].version)
		fmtprint(fmt, "$Version=%s; ", jar->c[0].version);
	for(i=0; i<jar->nc; i++)
		fmtprint(fmt, "%s%s=%s", i ? "; ":"", jar->c[i].name, jar->c[i].value);
	fmtprint(fmt, "\r\n");
	return 0;
}

/* individual cookie */
int
cookiefmt(Fmt *fmt)
{
	int j, k, first;
	char *t;
	Cookie *c;

	c = va_arg(fmt->args, Cookie*);

	first = 1;
	for(j=0; j<nelem(stab); j++){
		t = *(char**)((char*)c+stab[j].offset);
		if(t == nil)
			continue;
		if(first)
			first = 0;
		else
			fmtprint(fmt, " ");
		fmtprint(fmt, "%s=%q", stab[j].s, t);
	}
	for(j=0; j<nelem(itab); j++){
		k = *(int*)((char*)c+itab[j].offset);
		if(k == 0)
			continue;
		if(first)
			first = 0;
		else
			fmtprint(fmt, " ");
		fmtprint(fmt, "%s=%ud", itab[j].s, k);
	}
	return 0;
}

/*
 * sort cookies:
 *	- alpha by name
 *	- alpha by domain
 *	- longer paths first, then alpha by path (RFC2109 4.3.4)
 */
int
cookiecmp(Cookie *a, Cookie *b)
{
	int i;

	if((i = strcmp(a->name, b->name)) != 0)
		return i;
	if((i = cistrcmp(a->dom, b->dom)) != 0)
		return i;
	if((i = strlen(b->path) - strlen(a->path)) != 0)
		return i;
	if((i = strcmp(a->path, b->path)) != 0)
		return i;
	return 0;
}

int
exactcookiecmp(Cookie *a, Cookie *b)
{
	int i;

	if((i = cookiecmp(a, b)) != 0)
		return i;
	if((i = strcmp(a->value, b->value)) != 0)
		return i;
	if(a->version || b->version){
		if(!a->version)
			return -1;
		if(!b->version)
			return 1;
		if((i = strcmp(a->version, b->version)) != 0)
			return i;
	}
	if(a->comment || b->comment){
		if(!a->comment)
			return -1;
		if(!b->comment)
			return 1;
		if((i = strcmp(a->comment, b->comment)) != 0)
			return i;
	}
	if((i = b->expire - a->expire) != 0)
		return i;
	if((i = b->secure - a->secure) != 0)
		return i;
	if((i = b->explicitdom - a->explicitdom) != 0)
		return i;
	if((i = b->explicitpath - a->explicitpath) != 0)
		return i;
	if((i = b->netscapestyle - a->netscapestyle) != 0)
		return i;

	return 0;
}

void
freecookie(Cookie *c)
{
	int i;

	for(i=0; i<nelem(stab); i++)
		free(*(char**)((char*)c+stab[i].offset));
}

void
copycookie(Cookie *c)
{
	int i;
	char **ps;

	for(i=0; i<nelem(stab); i++){
		ps = (char**)((char*)c+stab[i].offset);
		if(*ps)
			*ps = estrdup9p(*ps);
	}
}

void
delcookie(Jar *j, Cookie *c)
{
	int i;

	j->dirty = 1;
	i = c - j->c;
	if(i < 0 || i >= j->nc)
		abort();
	c->deleted = 1;
}

void
addcookie(Jar *j, Cookie *c)
{
	int i;

	if(!c->name || !c->value || !c->path || !c->dom){
		fprint(2, "not adding incomplete cookie\n");
		return;
	}

	if(debug)
		fprint(2, "add %K\n", c);

	for(i=0; i<j->nc; i++)
		if(cookiecmp(&j->c[i], c) == 0){
			if(debug)
				fprint(2, "cookie %K matches %K\n", &j->c[i], c);
			if(exactcookiecmp(&j->c[i], c) == 0){
				if(debug)
					fprint(2, "\texactly\n");
				j->c[i].mark = 0;
				return;
			}
			delcookie(j, &j->c[i]);
		}

	j->dirty = 1;
	if(j->nc == j->mc){
		j->mc += 16;
		j->c = erealloc9p(j->c, j->mc*sizeof(Cookie));
	}
	j->c[j->nc] = *c;
	copycookie(&j->c[j->nc]);
	j->nc++;
}

void
purgejar(Jar *j)
{
	int i;

	for(i=j->nc-1; i>=0; i--){
		if(!j->c[i].deleted)
			continue;
		freecookie(&j->c[i]);
		--j->nc;
		j->c[i] = j->c[j->nc];
	}
}

void
addtojar(Jar *jar, char *line, int ondisk)
{
	Cookie c;
	int i, j, nf, *pint;
	char *f[20], *attr, *val, **pstr;
	
	memset(&c, 0, sizeof c);
	c.expire = ~0;
	c.ondisk = ondisk;
	nf = tokenize(line, f, nelem(f));
	for(i=0; i<nf; i++){
		attr = f[i];
		if((val = strchr(attr, '=')) != nil)
			*val++ = '\0';
		else
			val = "";
		/* string attributes */
		for(j=0; j<nelem(stab); j++){
			if(strcmp(stab[j].s, attr) == 0){
				pstr = (char**)((char*)&c+stab[j].offset);
				*pstr = val;
			}
		}
		/* integer attributes */
		for(j=0; j<nelem(itab); j++){
			if(strcmp(itab[j].s, attr) == 0){
				pint = (int*)((char*)&c+itab[j].offset);
				if(val[0]=='\0')
					*pint = 1;
				else
					*pint = strtoul(val, 0, 0);
			}
		}
	}
	if(c.name==nil || c.value==nil || c.dom==nil || c.path==nil){
		if(debug)
			fprint(2, "ignoring fractional cookie %K\n", &c);
		return;
	}
	addcookie(jar, &c);
}

Jar*
newjar(void)
{
	Jar *jar;

	jar = emalloc9p(sizeof(Jar));
	return jar;
}

int
expirejar(Jar *jar, int exiting)
{
	int i, n;
	uint now;

	now = time(0);
	n = 0;
	for(i=0; i<jar->nc; i++){
		if(jar->c[i].expire < now || (exiting && jar->c[i].expire==~0)){
			delcookie(jar, &jar->c[i]);
			n++;
		}
	}
	return n;
}

int
syncjar(Jar *jar)
{
	int i, fd;
	char *line;
	Dir *d;
	Biobuf *b;
	Qid q;

	if(jar->file==nil)
		return 0;

	memset(&q, 0, sizeof q);
	if((d = dirstat(jar->file)) != nil){
		q = d->qid;
		if(d->qid.path != jar->qid.path || d->qid.vers != jar->qid.vers)
			jar->dirty = 1;
		free(d);
	}

	if(jar->dirty == 0)
		return 0;

	fd = -1;
	for(i=0; i<50; i++){
		if((fd = create(jar->lockfile, OWRITE, DMEXCL|0666)) < 0){
			sleep(100);
			continue;
		}
		break;
	}
	if(fd < 0){
		if(debug)
			fprint(2, "open %s: %r", jar->lockfile);
		werrstr("cannot acquire jar lock: %r");
		return -1;
	}

	for(i=0; i<jar->nc; i++)	/* mark is cleared by addcookie */
		jar->c[i].mark = jar->c[i].ondisk;

	if((b = Bopen(jar->file, OREAD)) == nil){
		if(debug)
			fprint(2, "Bopen %s: %r", jar->file);
		werrstr("cannot read cookie file %s: %r", jar->file);
		close(fd);
		return -1;
	}
	for(; (line = Brdstr(b, '\n', 1)) != nil; free(line)){
		if(*line == '#')
			continue;
		addtojar(jar, line, 1);
	}
	Bterm(b);

	for(i=0; i<jar->nc; i++)
		if(jar->c[i].mark)
			delcookie(jar, &jar->c[i]);

	purgejar(jar);

	b = Bopen(jar->file, OWRITE);
	if(b == nil){
		if(debug)
			fprint(2, "Bopen write %s: %r", jar->file);
		close(fd);
		return -1;
	}
	Bprint(b, "# webcookies cookie jar\n");
	Bprint(b, "# comments and non-standard fields will be lost\n");
	for(i=0; i<jar->nc; i++){
		if(jar->c[i].expire == ~0)
			continue;
		Bprint(b, "%K\n", &jar->c[i]);
		jar->c[i].ondisk = 1;
	}
	Bterm(b);

	jar->dirty = 0;
	close(fd);
	if((d = dirstat(jar->file)) != nil){
		jar->qid = d->qid;
		free(d);
	}
	return 0;
}

Jar*
readjar(char *file)
{
	char *lock, *p;
	Jar *jar;

	jar = newjar();
	lock = emalloc9p(strlen(file)+10);
	strcpy(lock, file);
	if((p = strrchr(lock, '/')) != nil)
		p++;
	else
		p = lock;
	memmove(p+2, p, strlen(p)+1);
	p[0] = 'L';
	p[1] = '.';
	jar->lockfile = lock;
	jar->file = file;
	jar->dirty = 1;

	if(syncjar(jar) < 0){
		free(jar->file);
		free(jar->lockfile);
		free(jar);
		return nil;
	}
	return jar;
}

void
closejar(Jar *jar)
{
	int i;

	expirejar(jar, 0);
	if(syncjar(jar) < 0)
		fprint(2, "warning: cannot rewrite cookie jar: %r\n");

	for(i=0; i<jar->nc; i++)
		freecookie(&jar->c[i]);

	free(jar->file);
	free(jar);	
}

/*
 * Domain name matching is per RFC2109, section 2:
 *
 * Hosts names can be specified either as an IP address or a FQHN
 * string.  Sometimes we compare one host name with another.  Host A's
 * name domain-matches host B's if
 *
 * * both host names are IP addresses and their host name strings match
 *   exactly; or
 *
 * * both host names are FQDN strings and their host name strings match
 *   exactly; or
 *
 * * A is a FQDN string and has the form NB, where N is a non-empty name
 *   string, B has the form .B', and B' is a FQDN string.  (So, x.y.com
 *   domain-matches .y.com but not y.com.)
 *
 * Note that domain-match is not a commutative operation: a.b.c.com
 * domain-matches .c.com, but not the reverse.
 *
 * (This does not verify that IP addresses and FQDN's are well-formed.)
 */
int
isdomainmatch(char *name, char *pattern)
{
	int lname, lpattern;

	if(cistrcmp(name, pattern)==0)
		return 1;

	if(strcmp(ipattr(name), "dom")==0 && pattern[0]=='.'){
		lname = strlen(name);
		lpattern = strlen(pattern);
		if(lname >= lpattern && cistrcmp(name+lname-lpattern, pattern)==0)
			return 1;
	}

	return 0;
}

/*
 * RFC2109 4.3.4:
 *	- domain must match
 *	- path in cookie must be a prefix of request path
 *	- cookie must not have expired
 */
int
iscookiematch(Cookie *c, char *dom, char *path, uint now)
{
	return isdomainmatch(dom, c->dom)
		&& strncmp(c->path, path, strlen(c->path))==0
		&& c->expire >= now;
}

/* 
 * Produce a subjar of matching cookies.
 * Secure cookies are only included if secure is set.
 */
Jar*
cookiesearch(Jar *jar, char *dom, char *path, int issecure)
{
	int i;
	Jar *j;
	uint now;

	now = time(0);
	j = newjar();
	for(i=0; i<jar->nc; i++)
		if((issecure || !jar->c[i].secure) && iscookiematch(&jar->c[i], dom, path, now))
			addcookie(j, &jar->c[i]);
	if(j->nc == 0){
		closejar(j);
		werrstr("no cookies found");
		return nil;
	}
	qsort(j->c, j->nc, sizeof(j->c[0]), (int(*)(const void*, const void*))cookiecmp);
	return j;
}

/*
 * RFC2109 4.3.2 security checks
 */
char*
isbadcookie(Cookie *c, char *dom, char *path)
{
	if(strncmp(c->path, path, strlen(c->path)) != 0)
		return "cookie path is not a prefix of the request path";

	if(c->explicitdom && c->dom[0] != '.')
		return "cookie domain doesn't start with dot";

	if(memchr(c->dom+1, '.', strlen(c->dom)-1-1) == nil)
		return "cookie domain doesn't have embedded dots";

	if(!isdomainmatch(dom, c->dom))
		return "request host does not match cookie domain";

	if(strcmp(ipattr(dom), "dom")==0
	&& memchr(dom, '.', strlen(dom)-strlen(c->dom)) != nil)
		return "request host contains dots before cookie domain";

	return 0;
}

/*
 * Sunday, 25-Jan-2002 12:24:36 GMT
 * Sunday, 25 Jan 2002 12:24:36 GMT
 * Sun, 25 Jan 02 12:24:36 GMT
 */
int
isleap(int year)
{
	return year%4==0 && (year%100!=0 || year%400==0);
}

uint
strtotime(char *s)
{
	char *os;
	int i;
	Tm tm;

	static int mday[2][12] = {
		31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31,
		31, 29, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31,
	};
	static char *wday[] = {
		"Sunday", "Monday", "Tuesday", "Wednesday",
		"Thursday", "Friday", "Saturday",
	};
	static char *mon[] = {
		"Jan", "Feb", "Mar", "Apr", "May", "Jun",
		"Jul", "Aug", "Sep", "Oct", "Nov", "Dec",
	};

	os = s;
	/* Sunday, */
	for(i=0; i<nelem(wday); i++){
		if(cistrncmp(s, wday[i], strlen(wday[i])) == 0){
			s += strlen(wday[i]);
			break;
		}
		if(cistrncmp(s, wday[i], 3) == 0){
			s += 3;
			break;
		}
	}
	if(i==nelem(wday)){
		if(debug)
			fprint(2, "bad wday (%s)\n", os);
		return -1;
	}
	if(*s++ != ',' || *s++ != ' '){
		if(debug)