xmlconvert.in 14.7 KB
Newer Older
1 2 3 4
#!/usr/bin/perl -wT

#
# EMULAB-COPYRIGHT
5
# Copyright (c) 2000-2004 University of Utah and the Flux Group.
6 7 8 9 10
# All rights reserved.
#

use English;
use Getopt::Std;
11
use XML::Parser;
12 13 14 15 16
    
#
# Convert between XML and DB representation of a virtual experiment.
# Very simple, no DTDs, DOMs, XSLs, etc. Just the facts ...
#
17 18 19 20
# XXX We do not regex the data carefully enough before inserting it into
# the DB. We run quotemeta() over it, but we should be more careful about
# per-slot checks.
#
21 22 23 24 25 26
sub usage()
{
    print STDOUT "Usage: xmlconvert [-x <xmlfile> [-n]] [-d] pid eid\n";
 
    exit(-1);
}
27
my $optlist  = "x:nds";
28
my $fromxml  = 0;
29
my $impotent = 0;
30
my $debug    = 0;
31 32 33
# Results of parsing nse specifications. Therefore different treatment.
# In particular, we don't expect updates to the experiments table
my $simparse = 0;
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52

#
# Configure variables
#
my $TB       = "@prefix@";
my $TBOPS    = "@TBOPSEMAIL@";

# Locals
my $xmlfile;
my $pid;
my $eid;

# This goes at the beginning of the output.
my $XMLHEADER = "<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?>";

#
# These are the virtual tables that make up an experiment.
# Each one could have multiple rows, each of which will be a
# hash table. 
53 54 55 56 57 58 59 60
my %virtual_tables = ("experiments"		=> undef,
		      "virt_nodes"		=> undef,
		      "virt_lans"		=> undef,
		      "virt_trafgens"		=> undef,
		      "virt_agents"		=> undef,
		      "virt_node_desires"	=> undef,
		      "virt_routes"		=> undef,
		      "virt_vtypes"		=> undef,
61
		      "virt_programs"		=> undef,
62 63
		      "nseconfigs"		=> undef,
		      "eventlist"		=> undef);
64 65 66 67

# XXX
# The experiment table is special. Only certain fields are allowed to
# be updated. Not sure what the right approach for this is.
68
# Note that I regex the data before inserting it below.
69
#
70 71 72 73 74 75 76
my %experiment_fields = ("multiplex_factor"		=> 1,
			 "forcelinkdelays"		=> 1,
			 "uselinkdelays"		=> 1,
			 "usewatunnels"			=> 1,
			 "uselatestwadata"		=> 1,
			 "wa_delay_solverweight"	=> 1,
			 "wa_bw_solverweight"		=> 1,
77
			 "wa_plr_solverweight"		=> 1,
78 79
			 "cpu_usage"			=> 1,
			 "mem_usage"			=> 1,
80
			 "veth_encapsulate"		=> 1,
81
			 "sync_server"			=> 1);
82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111

#
# Turn off line buffering on output
#
$| = 1;

#
# Untaint the path
# 
$ENV{'PATH'} = "/bin:/usr/bin:/sbin:/usr/sbin";
delete @ENV{'IFS', 'CDPATH', 'ENV', 'BASH_ENV'};

#
# Testbed Support libraries
#
use lib "@prefix@/lib";
use libdb;
use libtestbed;

#
# Parse command arguments. Once we return from getopts, all that should
# left are the required arguments.
#
%options = ();
if (! getopts($optlist, \%options)) {
    usage();
}
if (defined($options{"d"})) {
    $debug = 1;
}
112 113 114
if (defined($options{"s"})) {
    $simparse = 1;
}
115 116 117 118 119 120 121 122 123 124 125
if (defined($options{"x"})) {
    $fromxml = 1;
    $xmlfile = $options{"x"};

    if ($xmlfile =~ /^([-\w\/\.]+)$/) {
	$xmlfile = $1;
    }
    else {
	fatal("Bad data in argument: $xmlfile.");
    }
    if (defined($options{"n"})) {
126
	$impotent = 1;
127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148
    }    
}
if (@ARGV != 2) {
    usage();
}
$pid   = $ARGV[0];
$eid   = $ARGV[1];

# Taint Check.
if ($pid =~ /^([-\w]+)$/) {
    $pid = $1;
}
else {
    fatal("Bad data in argument: $pid.");
}
if ($eid =~ /^([-\w]+)$/) {
    $eid = $1;
}
else {
    fatal("Bad data in argument: $eid.");
}

149 150 151 152 153 154
my %event_objtypes;
my $query_result = DBQueryFatal("select idx,type from event_objecttypes");
while (my ($idx,$type) = $query_result->fetchrow_array()) {
    $event_objtypes{$type} = $idx;
}

155 156 157 158 159 160 161 162 163 164 165 166 167 168
# Do it
if ($fromxml) {
    readXML($pid, $eid, $xmlfile);
}
else {
    writeXML($pid, $eid);
}
exit(0);

#
# Read in XML and convert to DB representation, doing lots of checks!
# This code is silly. Overly stylized (one tag per line!). Should
# use the XML::Parser package instead. But this was easy and fun for a
# first cut. 
169 170 171 172 173 174 175 176
#
# State variables for parsing code below.
my $current_expt;
my $current_table;
my $current_row;
my $current_slot;
my $current_data;

177 178 179 180 181 182 183 184 185 186
sub readXML($$$) {
    my ($pid, $eid, $xmlfile) = @_;
    my %experiment;

    if ($xmlfile ne "-") {
	open(STDIN, "< $xmlfile")
	    or fatal("opening $xmlfile for STDIN: $!");
    }

    #
187
    # Create a parser.
188
    #
189 190 191 192
    my $parser = new XML::Parser(Style => 'Tree');
    $parser->setHandlers('Start'   => \&StartElement,
			 'End'     => \&EndElement,
			 'Char'    => \&ProcessElement);
193

194 195
    fatal($@)
	if (eval { $parser->parse(*STDIN); return 1; } != 1);
196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216

    # If these are the results of parsing the nse specifications,
    # we don't expect updates to the experiments table
    my %experiments_table;
    if ( ! $simparse ) {

	#
	# Verify. 
	#
	# Must be exactly one experiments table row, and we prune out lots
	# of stuff that is not allowed. Note that we never insert a
	# experiment, but only allow updates of certain values. 
	#
	if (scalar(@{$virtual_tables{"experiments"}}) != 1) {
	    fatal("Must be exactly one experiments table row!");
	}
	%experiments_table = %{@{$virtual_tables{"experiments"}}[0]};
	foreach my $key (keys(%experiments_table)) {
	    delete($experiments_table{$key})
		if (!exists($experiment_fields{$key}));
	}
217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233
    }

    #
    # Okay, a hokey DoS check. Do not allow more than 10000 total rows!
    # Why so many? Well, Rob likes to generate lots of events!
    #
    my $count = 0;
    foreach my $table (keys(%virtual_tables)) {
	$count += scalar(@{$virtual_tables{$table}})
	    if (defined($virtual_tables{$table}));
    }
    if ($count > 10000) {
	fatal("Too many rows of data!");
    }

    #
    # Okay, thats all the checking we do! There is not much that can
234 235
    # screw up the DB just by inserting rows into the allowed set of
    # virtual experiment tables, since we ignore the pid/eid in the xml. 
236
    #
237 238
    # First the experiments table, which gets an update statement, if there
    # is anything to update.
239
    #
240
    if ( (! $simparse) && scalar(keys(%experiments_table))) {
241
	my @setlist = ();
242

243 244
	foreach my $key (keys(%experiments_table)) {
	    my $val = $experiments_table{$key};
245

246 247 248 249
	    if ($val eq "NULL") {
		push(@setlist, "$key=NULL");
	    }
	    else {
250 251 252
		# Sanity check the fields.
		if (TBcheck_dbslot($val, "experiments", $key,
				TBDB_CHECKDBSLOT_WARN|TBDB_CHECKDBSLOT_ERROR)) {
253 254 255
		    push(@setlist, "$key='$val'");
		}
		else {
256 257
		    fatal("Illegal characters in table data: ".
			  "experiments:$key - $val");
258
		}
259
	    }	    
260
	}
261 262 263 264 265 266 267 268
	my $query = "update experiments ".
	            "set " . join(",", @setlist) . " " .
		    "where eid='$eid' and pid='$pid'";

	print "$query\n"
	    if ($debug);
	DBQueryFatal($query)
	    if (!$impotent);
269 270 271 272 273 274 275
    }

    #
    # Now all the other tables, which get inserts. Need to delete all the
    # old info too.
    #
    foreach my $table (keys(%virtual_tables)) {
276 277 278 279
	# Don't want to muck with this table! Done above. 
	next
	    if ($table eq "experiments");

280 281 282 283 284 285
	# Delete only during the initial parsing and not
	# during parsing of nse specifications
	if ( ! $simparse ) {
	    DBQueryFatal("delete from $table ".
			 "where eid='$eid' and pid='$pid'")
		if (!$impotent);
286 287 288 289 290 291 292 293 294 295 296 297
	} else {
	    # The nseconfigs table is special. During a
	    # simparse, we need delete all rows for the
	    # experiment except the one with the vname
	    # 'fullsim'. This row is essentially virtual
	    # info and does not change across swapins
	    # where as the other rows depend on the
	    # mapping
	    if ( !$impotent && ($table eq "nseconfigs") ) {
		DBQueryFatal("delete from $table ". 
		             "where eid='$eid' and pid='$pid' and ".
			     "vname!='fullsim'")
298 299 300 301 302 303 304 305 306 307
	    } elsif ( !$impotent && (($table eq "eventlist") ||
				       ($table eq "virt_agents")) ) {
	        # Both eventlist and virt_agents need to be cleared
		# for NSE event objecttype since entries in this
		# table depend on the particular mapping
		my $nse_objtype = $event_objtypes{"NSE"};
		DBQueryFatal("delete from $table ". 
		             "where pid='$pid' and eid='$eid' and ".
			     "objecttype='$nse_objtype'");
	    } 
308
	}
309 310 311 312 313 314 315 316
	next
	    if (!defined($virtual_tables{$table}));

	foreach my $rowref (@{$virtual_tables{$table}}) {
	    my %rowhash = %{ $rowref };
	    my @fields  = ("pid", "eid");
	    my @values  = ("'$pid'", "'$eid'");

317 318 319 320
	    # If no actual rows, then skip. Might happen.
	    last
		if (! scalar(keys(%rowhash)));

321
	    foreach my $key (keys(%rowhash)) {
322
		my $val = $rowhash{$key};
323 324 325 326

		if ($val eq "NULL") {
		    push(@values, "NULL");
		}
327 328 329
		elsif ($val eq "") {
		    push(@values, "''");
		}
330
		else {
331 332 333 334 335 336 337 338 339 340
		    # Sanity check the fields.
		    if (TBcheck_dbslot($val, $table, $key,
				 TBDB_CHECKDBSLOT_WARN|TBDB_CHECKDBSLOT_ERROR)) {
			push(@values, DBQuoteSpecial($val));
		    }
		    else {
			fatal("Illegal characters in table data: ".
			      "$table:$key - $val");
		    }
		}
341 342
		push(@fields, $key);
	    }
343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358
	    # If we are called after an nseparse, we need to
	    # use replace coz some of the tables such as 
	    # virt_agents and eventlist are not truly
	    # virtual tables. That is coz they contain the
	    # vnode field which is the same as the vname
	    # field in the reserved table. For simulated
	    # nodes, the mapping may change across swapins
	    # and the event may have to be delivered to a
	    # different simhost
	    if ( $simparse ) {
		$query = "replace into $table (" . join(",", @fields) . ") ".
		"values (" . join(",", @values) . ") ";
	    } else {
		$query = "insert into $table (" . join(",", @fields) . ") ".
		"values (" . join(",", @values) . ") ";
	    }
359 360 361 362

	    print "$query\n"
		if ($debug);
	    DBQueryFatal($query)
363
		if (!$impotent);
364 365 366 367 368
	}
    }
    return 0;
}

369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501
#
# XML::Parser routines.
#
#
# Start an element.
# 
sub StartElement ($$$)
{
    my ($expat, $element, %attrs) = @_;

    if ($element eq "virtual_experiment") {
	fatal("Out of sync at experiment start: $element")
	    if (defined($current_expt) ||
		defined($current_table) ||
		defined($current_row) ||
		defined($current_slot));
	$current_expt = "$pid/$eid";
	
	#
	# Sanity check pid/eid.
	#
	if ((exists($attrs{'pid'}) && $attrs{'pid'} ne $pid) ||
	    (exists($attrs{'eid'}) && $attrs{'eid'} ne $eid)) {
	    fatal("pid/eid mismatch!");
	}
    }
    elsif (exists($virtual_tables{$element})) {
	#
	# A new table start.
	#
	fatal("Out of sync at element start: $element")
	    if (!defined($current_expt) ||
		defined($current_table) ||
		defined($current_row) ||
		defined($current_slot));
	$current_table = $element;

	if (! defined($virtual_tables{$element})) {
	    $virtual_tables{$element} = [];
	}
	print "Starting new table: $element\n"
	    if ($debug);
    }
    elsif ($element eq "row") {
	fatal("Out of sync at row start: $element")
	    if (!defined($current_expt) ||
		!defined($current_table) ||
		defined($current_row) ||
		defined($current_slot));
	$current_row = {};
    }
    else {
	fatal("Out of sync at slot start: $element")
	    if (!defined($current_expt) ||
		!defined($current_table) ||
		!defined($current_row) ||
		defined($current_slot));
	$current_slot = $element;
	$current_data = "";
    }
}

#
# End an element.
# 
sub EndElement ($$)
{
    my ($expat, $element) = @_;

    if ($element eq "virtual_experiment") {
	fatal("Out of sync at experiment start: $element")
	    if (!defined($current_expt) ||
		defined($current_table) ||
		defined($current_row) ||
		defined($current_slot));
	undef($current_expt);
    }
    elsif (exists($virtual_tables{$element})) {
	#
	# A table termination.
	#
	fatal("Out of sync at element end: $element")
	    if (!defined($current_expt) ||
		!defined($current_table) ||
		defined($current_row) ||
		defined($current_slot));
	undef($current_table);
    }
    elsif ($element eq "row") {
	fatal("Out of sync at row end: $element")
	    if (!defined($current_expt) ||
		!defined($current_table) ||
		!defined($current_row) ||
		defined($current_slot));

	print "Adding new row to table $current_table\n"
	    if ($debug);
	
	push(@{$virtual_tables{$current_table}}, $current_row);
	undef($current_row);
    }
    else {
	fatal("Out of sync at slot end: $element")
	    if (!defined($current_expt) ||
		!defined($current_table) ||
		!defined($current_row) ||
		!defined($current_slot));
    
	#
	# Always ignore pid/eid.
	#
	if ($current_slot ne "pid" && $current_slot ne "eid") {
	    print "    Entering new slot: $current_slot: $current_data\n"
		if ($debug);
	    $current_row->{$current_slot} = $current_data;
	}
	undef($current_slot);
	undef($current_data);
    }
}

#
# Process stuff inside a slot.
# 
sub ProcessElement ($$)
{
    my ($expat, $string) = @_;

    if (defined($current_slot)) {
	$current_data .= xmldecode($string);
    }
}

502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605
#
# Convert a virtual experiment representation into XML and spit it out.
# The DB holds the data of course.
#
sub writeXML($$) {
    my ($pid, $eid) = @_;

    my $query_result =
	DBQueryFatal("select * from experiments ".
		     "where eid='$eid' and pid='$pid'");

    if (! $query_result->numrows) {
	fatal("No such experiment $pid/$eid exists!");
    }

    spitxml_header();
    spitxml_opentag("virtual_experiment pid='$pid' eid='$eid'", 0);

    # Spit out all of the tables.
    foreach my $table (keys(%virtual_tables)) {
	spitxml_dbrows($table, 1, 
		       "select * from $table ".
		       "where eid='$eid' and pid='$pid'");
    }
    spitxml_closetag("virtual_experiment", 0);

    return 0;
}

#
# Utility functions to pretty print XML output, with specified indentation.
#
sub spitxml_opentag($$)
{
    my ($tag, $level) = @_;

    my $spaces = $level * 2;

    printf("%${spaces}s<%s>\n", "", $tag);
}

sub spitxml_closetag($$)
{
    my ($tag, $level) = @_;

    my $spaces = $level * 2;

    printf("%${spaces}s</%s>\n", "", $tag);
}

sub spitxml_header()
{
    print "$XMLHEADER\n";
}

sub spitxml_entity($$$)
{
    my ($tag, $data, $level) = @_;
    my $spaces = $level * 2;

    $data = "NULL"
	if (!defined($data));

    printf("%${spaces}s<%s>%s</%s>\n", "", $tag, xmlencode($data), $tag);
}

sub spitxml_dbrow($%)
{
    my($level, %dbrow) = @_;

    spitxml_opentag("row", $level);
    
    foreach my $tag (keys(%dbrow)) {
	my $data = $dbrow{$tag};
	spitxml_entity($tag, $data, $level + 1);

    }
    spitxml_closetag("row", $level);
}

sub spitxml_dbrows($$$)
{
    my($tag, $level, $query) = @_;
    my $query_result = DBQueryFatal($query);
    
    if ($query_result->numrows) {
	spitxml_opentag($tag, $level);
	while (my %hashrow = $query_result->fetchhash()) {
	    spitxml_dbrow($level + 1, %hashrow);
	}
	spitxml_closetag($tag, $level);
    }
}

#
# Convert from/to XML special chars. Not many of them ...
# 
sub xmlencode($)
{
    my ($string) = @_;

    my %specialchars = ('&' => '&amp;',
			'<' => '&lt;',
			'>' => '&gt;',
606 607 608
			"'" => '&#39;',
			"]" => '&#93;',
			'"' => '&#34;');
609 610 611 612 613 614 615 616 617 618 619 620

    $string =~ s/([&<>"'])/$specialchars{$1}/ge;
    return $string;
}

sub xmldecode($)
{
    my ($string) = @_;

    my %specialchars = ('&amp;'  => '&',
			'&lt;'   => '<',
			'&gt;'   => '>',
621 622 623
			'&#39;'  => "'",
			'&#93;'  => ']',
			'&#34;'  => '"');
624 625 626 627 628 629 630 631 632 633 634 635 636

    $string =~ s/(&\w+;)/$specialchars{$1}/ge;
    return $string;
}

# Die
sub fatal($)
{
    my ($msg) = @_;

    die("*** $0:\n".
	"    $msg\n");
}