dt_cc.c

来自「Sun Solaris 10 中的 DTrace 组件的源代码。请参看: htt」· C语言 代码 · 共 2,025 行 · 第 1/4 页

/*
 * Copyright 2005 Sun Microsystems, Inc.  All rights reserved.
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License, Version 1.0 only.
 * See the file usr/src/LICENSING.NOTICE in this distribution or
 * http://www.opensolaris.org/license/ for details.
 */

#pragma ident	"@(#)dt_cc.c	1.14	04/12/18 SMI"

/*
 * DTrace D Language Compiler
 *
 * The code in this source file implements the main engine for the D language
 * compiler.  The driver routine for the compiler is dt_compile(), below.  The
 * compiler operates on either stdio FILEs or in-memory strings as its input
 * and can produce either dtrace_prog_t structures from a D program or a single
 * dtrace_difo_t structure from a D expression.  Multiple entry points are
 * provided as wrappers around dt_compile() for the various input/output pairs.
 * The compiler itself is implemented across the following source files:
 *
 * dt_lex.l - lex scanner
 * dt_grammar.y - yacc grammar
 * dt_parser.c - parse tree creation and semantic checking
 * dt_decl.c - declaration stack processing
 * dt_xlator.c - D translator lookup and creation
 * dt_ident.c - identifier and symbol table routines
 * dt_pragma.c - #pragma processing and D pragmas
 * dt_printf.c - D printf() and printa() argument checking and processing
 * dt_cc.c - compiler driver and dtrace_prog_t construction
 * dt_cg.c - DIF code generator
 * dt_as.c - DIF assembler
 * dt_dof.c - dtrace_prog_t -> DOF conversion
 *
 * Several other source files provide collections of utility routines used by
 * these major files.  The compiler itself is implemented in multiple passes:
 *
 * (1) The input program is scanned and parsed by dt_lex.l and dt_grammar.y
 *     and parse tree nodes are constructed using the routines in dt_parser.c.
 *     This node construction pass is described further in dt_parser.c.
 *
 * (2) The parse tree is "cooked" by assigning each clause a context (see the
 *     routine dt_setcontext(), below) based on its probe description and then
 *     recursively descending the tree performing semantic checking.  The cook
 *     routines are also implemented in dt_parser.c and described there.
 *
 * (3) For actions that are DIF expression statements, the DIF code generator
 *     and assembler are invoked to create a finished DIFO for the statement.
 *
 * (4) The dtrace_prog_t data structures for the program clauses and actions
 *     are built, containing pointers to any DIFOs created in step (3).
 *
 * (5) The caller invokes a routine in dt_dof.c to convert the finished program
 *     into DOF format for use in anonymous tracing or enabling in the kernel.
 *
 * In the implementation, steps 2-4 are intertwined in that they are performed
 * in order for each clause as part of a loop that executes over the clauses.
 *
 * The D compiler currently implements nearly no optimization.  The compiler
 * implements integer constant folding as part of pass (1), and a set of very
 * simple peephole optimizations as part of pass (3).  As with any C compiler,
 * a large number of optimizations are possible on both the intermediate data
 * structures and the generated DIF code.  These possibilities should be
 * investigated in the context of whether they will have any substantive effect
 * on the overall DTrace probe effect before they are undertaken.
 */

#include <sys/types.h>
#include <sys/wait.h>

#include <assert.h>
#include <strings.h>
#include <signal.h>
#include <unistd.h>
#include <stdlib.h>
#include <stdio.h>
#include <errno.h>
#include <ucontext.h>
#include <limits.h>
#include <alloca.h>
#include <ctype.h>

#define	_POSIX_PTHREAD_SEMANTICS
#include <dirent.h>
#undef	_POSIX_PTHREAD_SEMANTICS

#include <dt_module.h>
#include <dt_provider.h>
#include <dt_printf.h>
#include <dt_pid.h>
#include <dt_grammar.h>
#include <dt_ident.h>
#include <dt_string.h>
#include <dt_impl.h>

static const dtrace_pattr_t dt_def_pattr = {
{ DTRACE_STABILITY_UNSTABLE, DTRACE_STABILITY_UNSTABLE, DTRACE_CLASS_COMMON },
{ DTRACE_STABILITY_UNSTABLE, DTRACE_STABILITY_UNSTABLE, DTRACE_CLASS_COMMON },
{ DTRACE_STABILITY_UNSTABLE, DTRACE_STABILITY_UNSTABLE, DTRACE_CLASS_COMMON },
{ DTRACE_STABILITY_UNSTABLE, DTRACE_STABILITY_UNSTABLE, DTRACE_CLASS_COMMON },
{ DTRACE_STABILITY_UNSTABLE, DTRACE_STABILITY_UNSTABLE, DTRACE_CLASS_COMMON },
};

static const dtrace_diftype_t dt_void_rtype = {
	DIF_TYPE_CTF, CTF_K_INTEGER, 0, 0, 0
};

static const dtrace_diftype_t dt_int_rtype = {
	DIF_TYPE_CTF, CTF_K_INTEGER, 0, 0, sizeof (uint64_t)
};

/*ARGSUSED*/
static void
dt_idreset(dt_idhash_t *dhp, dt_ident_t *idp, void *ignored)
{
	idp->di_flags &= ~(DT_IDFLG_REF | DT_IDFLG_MOD |
	    DT_IDFLG_DIFR | DT_IDFLG_DIFW);
}

static void
dt_idkill(dt_idhash_t *dhp, dt_ident_t *idp, void *gen)
{
	if (idp->di_gen == (ulong_t)gen)
		dt_idhash_delete(dhp, idp);
}

/*ARGSUSED*/
static void
dt_idpragma(dt_idhash_t *dhp, dt_ident_t *idp, void *ignored)
{
	yylineno = idp->di_lineno;
	xyerror(D_PRAGMA_UNUSED, "unused #pragma %s\n", (char *)idp->di_iarg);
}

static dtrace_stmtdesc_t *
dt_stmt_create(dtrace_hdl_t *dtp, dtrace_ecbdesc_t *edp,
    dtrace_attribute_t descattr, dtrace_attribute_t stmtattr)
{
	dtrace_stmtdesc_t *sdp = dtrace_stmt_create(dtp, edp);

	if (sdp == NULL)
		longjmp(yypcb->pcb_jmpbuf, EDT_NOMEM);

	assert(yypcb->pcb_stmt == NULL);
	yypcb->pcb_stmt = sdp;

	sdp->dtsd_descattr = descattr;
	sdp->dtsd_stmtattr = stmtattr;

	return (sdp);
}

static dtrace_actdesc_t *
dt_stmt_action(dtrace_hdl_t *dtp, dtrace_stmtdesc_t *sdp)
{
	dtrace_actdesc_t *new;

	if ((new = dtrace_stmt_action(dtp, sdp)) == NULL)
		longjmp(yypcb->pcb_jmpbuf, EDT_NOMEM);

	return (new);
}

/*
 * Utility function to determine if a given action description is destructive.
 * The dtdo_destructive bit is set for us by the DIF assembler (see dt_as.c).
 */
static int
dt_action_destructive(const dtrace_actdesc_t *ap)
{
	return (DTRACEACT_ISDESTRUCTIVE(ap->dtad_kind) || (ap->dtad_kind ==
	    DTRACEACT_DIFEXPR && ap->dtad_difo->dtdo_destructive));
}

static void
dt_stmt_append(dtrace_stmtdesc_t *sdp, const dt_node_t *dnp)
{
	dtrace_ecbdesc_t *edp = sdp->dtsd_ecbdesc;
	dtrace_actdesc_t *ap, *tap;
	int commit = 0;
	int speculate = 0;
	int datarec = 0;

	/*
	 * Make sure that the new statement jibes with the rest of the ECB.
	 */
	for (ap = edp->dted_action; ap != NULL; ap = ap->dtad_next) {
		if (ap->dtad_kind == DTRACEACT_COMMIT) {
			if (commit) {
				dnerror(dnp, D_COMM_COMM, "commit( ) may "
				    "not follow commit( )\n");
			}

			if (datarec) {
				dnerror(dnp, D_COMM_DREC, "commit( ) may "
				    "not follow data-recording action(s)\n");
			}

			for (tap = ap; tap != NULL; tap = tap->dtad_next) {
				if (!DTRACEACT_ISAGG(tap->dtad_kind))
					continue;

				dnerror(dnp, D_AGG_COMM, "aggregating actions "
				    "may not follow commit( )\n");
			}

			commit = 1;
			continue;
		}

		if (ap->dtad_kind == DTRACEACT_SPECULATE) {
			if (speculate) {
				dnerror(dnp, D_SPEC_SPEC, "speculate( ) may "
				    "not follow speculate( )\n");
			}

			if (commit) {
				dnerror(dnp, D_SPEC_COMM, "speculate( ) may "
				    "not follow commit( )\n");
			}

			if (datarec) {
				dnerror(dnp, D_SPEC_DREC, "speculate( ) may "
				    "not follow data-recording action(s)\n");
			}

			speculate = 1;
			continue;
		}

		if (DTRACEACT_ISAGG(ap->dtad_kind)) {
			if (speculate) {
				dnerror(dnp, D_AGG_SPEC, "aggregating actions "
				    "may not follow speculate( )\n");
			}

			datarec = 1;
			continue;
		}

		if (speculate) {
			if (dt_action_destructive(ap)) {
				dnerror(dnp, D_ACT_SPEC, "destructive actions "
				    "may not follow speculate( )\n");
			}

			if (ap->dtad_kind == DTRACEACT_EXIT) {
				dnerror(dnp, D_EXIT_SPEC, "exit( ) may not "
				    "follow speculate( )\n");
			}
		}

		/*
		 * Exclude all non data-recording actions.
		 */
		if (dt_action_destructive(ap) ||
		    ap->dtad_kind == DTRACEACT_DISCARD)
			continue;

		if (ap->dtad_kind == DTRACEACT_DIFEXPR &&
		    ap->dtad_difo->dtdo_rtype.dtdt_kind == DIF_TYPE_CTF &&
		    ap->dtad_difo->dtdo_rtype.dtdt_size == 0)
			continue;

		if (commit) {
			dnerror(dnp, D_DREC_COMM, "data-recording actions "
			    "may not follow commit( )\n");
		}

		if (!speculate)
			datarec = 1;
	}

	if (dtrace_stmt_add(yypcb->pcb_hdl, yypcb->pcb_prog, sdp) != 0)
		longjmp(yypcb->pcb_jmpbuf, dtrace_errno(yypcb->pcb_hdl));

	if (yypcb->pcb_stmt == sdp)
		yypcb->pcb_stmt = NULL;
}

/*
 * For the first element of an aggregation tuple or for printa(), we create a
 * simple DIF program that simply returns the immediate value that is the ID
 * of the aggregation itself.  This could be optimized in the future by
 * creating a new in-kernel dtad_kind that just returns an integer.
 */
static void
dt_action_difconst(dtrace_actdesc_t *ap, uint_t id, dtrace_actkind_t kind)
{
	dtrace_difo_t *dp = malloc(sizeof (dtrace_difo_t));

	if (dp == NULL)
		longjmp(yypcb->pcb_jmpbuf, EDT_NOMEM);

	bzero(dp, sizeof (dtrace_difo_t));
	dtrace_difo_hold(dp);

	dp->dtdo_buf = malloc(sizeof (dif_instr_t) * 2);
	dp->dtdo_inttab = malloc(sizeof (uint64_t));

	if (dp->dtdo_buf == NULL || dp->dtdo_inttab == NULL) {
		dtrace_difo_release(dp);
		longjmp(yypcb->pcb_jmpbuf, EDT_NOMEM);
	}

	dp->dtdo_buf[0] = DIF_INSTR_SETX(0, 1); /* setx	DIF_INTEGER[0], %r1 */
	dp->dtdo_buf[1] = DIF_INSTR_RET(1);	/* ret	%r1 */
	dp->dtdo_len = 2;
	dp->dtdo_inttab[0] = id;
	dp->dtdo_intlen = 1;
	dp->dtdo_rtype = dt_int_rtype;

	ap->dtad_difo = dp;
	ap->dtad_kind = kind;
}

static void
dt_action_clear(dtrace_hdl_t *dtp, dt_node_t *dnp, dtrace_stmtdesc_t *sdp)
{
	dt_ident_t *aid;
	dtrace_actdesc_t *ap;
	dt_node_t *anp;

	char n[DT_TYPE_NAMELEN];
	int argc = 0;

	for (anp = dnp->dn_args; anp != NULL; anp = anp->dn_list)
		argc++; /* count up arguments for error messages below */

	if (argc != 1) {
		dnerror(dnp, D_CLEAR_PROTO,
		    "%s( ) prototype mismatch: %d args passed, 1 expected\n",
		    dnp->dn_ident->di_name, argc);
	}

	anp = dnp->dn_args;
	assert(anp != NULL);

	if (anp->dn_kind != DT_NODE_AGG) {
		dnerror(dnp, D_CLEAR_AGGARG,
		    "%s( ) argument #1 is incompatible with prototype:\n"
		    "\tprototype: aggregation\n\t argument: %s\n",
		    dnp->dn_ident->di_name,
		    dt_node_type_name(anp, n, sizeof (n)));
	}

	aid = anp->dn_ident;

	if (aid->di_gen == dtp->dt_gen && !(aid->di_flags & DT_IDFLG_MOD)) {
		dnerror(dnp, D_CLEAR_AGGBAD,
		    "undefined aggregation: @%s\n", aid->di_name);
	}

	ap = dt_stmt_action(dtp, sdp);
	dt_action_difconst(ap, anp->dn_ident->di_id, DTRACEACT_LIBACT);
	ap->dtad_arg = DT_ACT_CLEAR;
}

static void
dt_action_normalize(dtrace_hdl_t *dtp, dt_node_t *dnp, dtrace_stmtdesc_t *sdp)
{
	dt_ident_t *aid;
	dtrace_actdesc_t *ap;
	dt_node_t *anp, *normal;
	int denormal = (strcmp(dnp->dn_ident->di_name, "denormalize") == 0);

	char n[DT_TYPE_NAMELEN];
	int argc = 0;

	for (anp = dnp->dn_args; anp != NULL; anp = anp->dn_list)
		argc++; /* count up arguments for error messages below */

	if ((denormal && argc != 1) || (!denormal && argc != 2)) {
		dnerror(dnp, D_NORMALIZE_PROTO,
		    "%s( ) prototype mismatch: %d args passed, %d expected\n",
		    dnp->dn_ident->di_name, argc, denormal ? 1 : 2);
	}

	anp = dnp->dn_args;
	assert(anp != NULL);

	if (anp->dn_kind != DT_NODE_AGG) {
		dnerror(dnp, D_NORMALIZE_AGGARG,
		    "%s( ) argument #1 is incompatible with prototype:\n"
		    "\tprototype: aggregation\n\t argument: %s\n",
		    dnp->dn_ident->di_name,
		    dt_node_type_name(anp, n, sizeof (n)));
	}

	if ((normal = anp->dn_list) != NULL && !dt_node_is_scalar(normal)) {
		dnerror(dnp, D_NORMALIZE_SCALAR,
		    "%s( ) argument #2 must be of scalar type\n",
		    dnp->dn_ident->di_name);
	}

	aid = anp->dn_ident;

	if (aid->di_gen == dtp->dt_gen && !(aid->di_flags & DT_IDFLG_MOD)) {
		dnerror(dnp, D_NORMALIZE_AGGBAD,
		    "undefined aggregation: @%s\n", aid->di_name);
	}

	ap = dt_stmt_action(dtp, sdp);
	dt_action_difconst(ap, anp->dn_ident->di_id, DTRACEACT_LIBACT);

	if (denormal) {
		ap->dtad_arg = DT_ACT_DENORMALIZE;
		return;
	}

	ap->dtad_arg = DT_ACT_NORMALIZE;

	assert(normal != NULL);
	ap = dt_stmt_action(dtp, sdp);
	dt_cg(yypcb, normal);

	ap->dtad_difo = dt_as(yypcb);
	ap->dtad_kind = DTRACEACT_LIBACT;
	ap->dtad_arg = DT_ACT_NORMALIZE;
}

static void
dt_action_trunc(dtrace_hdl_t *dtp, dt_node_t *dnp, dtrace_stmtdesc_t *sdp)
{
	dt_ident_t *aid;
	dtrace_actdesc_t *ap;
	dt_node_t *anp, *trunc;

	char n[DT_TYPE_NAMELEN];
	int argc = 0;

	for (anp = dnp->dn_args; anp != NULL; anp = anp->dn_list)
		argc++; /* count up arguments for error messages below */

	if (argc > 2 || argc < 1) {
		dnerror(dnp, D_TRUNC_PROTO,
		    "%s( ) prototype mismatch: %d args passed, %s expected\n",
		    dnp->dn_ident->di_name, argc,
		    argc < 1 ? "at least 1" : "no more than 2");
	}

	anp = dnp->dn_args;
	assert(anp != NULL);
	trunc = anp->dn_list;

	if (anp->dn_kind != DT_NODE_AGG) {
		dnerror(dnp, D_TRUNC_AGGARG,
		    "%s( ) argument #1 is incompatible with prototype:\n"
		    "\tprototype: aggregation\n\t argument: %s\n",
		    dnp->dn_ident->di_name,
		    dt_node_type_name(anp, n, sizeof (n)));
	}

	if (argc == 2) {
		assert(trunc != NULL);
		if (!dt_node_is_scalar(trunc)) {
			dnerror(dnp, D_TRUNC_SCALAR,
			    "%s( ) argument #2 must be of scalar type\n",
			    dnp->dn_ident->di_name);
		}
	}

	aid = anp->dn_ident;

	if (aid->di_gen == dtp->dt_gen && !(aid->di_flags & DT_IDFLG_MOD)) {
		dnerror(dnp, D_TRUNC_AGGBAD,
		    "undefined aggregation: @%s\n", aid->di_name);
	}

	ap = dt_stmt_action(dtp, sdp);
	dt_action_difconst(ap, anp->dn_ident->di_id, DTRACEACT_LIBACT);
	ap->dtad_arg = DT_ACT_TRUNC;

	ap = dt_stmt_action(dtp, sdp);

	if (argc == 1) {
		dt_action_difconst(ap, 0, DTRACEACT_LIBACT);
	} else {
		assert(trunc != NULL);
		dt_cg(yypcb, trunc);
		ap->dtad_difo = dt_as(yypcb);
		ap->dtad_kind = DTRACEACT_LIBACT;
	}

	ap->dtad_arg = DT_ACT_TRUNC;
}

static void
dt_action_printa(dtrace_hdl_t *dtp, dt_node_t *dnp, dtrace_stmtdesc_t *sdp)
{
	dt_ident_t *aid, *fid;
	dtrace_actdesc_t *ap;
	const char *format;
	dt_node_t *anp;

	char n[DT_TYPE_NAMELEN];
	int argc = 0, argr = 0;

	for (anp = dnp->dn_args; anp != NULL; anp = anp->dn_list)
		argc++; /* count up arguments for error messages below */

	switch (dnp->dn_args->dn_kind) {
	case DT_NODE_STRING:
		format = dnp->dn_args->dn_string;
		anp = dnp->dn_args->dn_list;
		argr = 2;