Index: head/sys/alpha/linux/linux_sysvec.c =================================================================== --- head/sys/alpha/linux/linux_sysvec.c (revision 55140) +++ head/sys/alpha/linux/linux_sysvec.c (revision 55141) @@ -1,511 +1,511 @@ /*- * Copyright (c) 1994-1996 Søren Schmidt * 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 withough 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. * * $FreeBSD$ */ /* XXX we use functions that might not exist. */ #include "opt_compat.h" #ifndef COMPAT_43 #error "Unable to compile Linux-emulator due to missing COMPAT_43 option!" #endif #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include MALLOC_DEFINE(M_LINUX, "linux", "Linux mode structures"); extern char linux_sigcode[]; extern int linux_szsigcode; extern struct sysent linux_sysent[LINUX_SYS_MAXSYSCALL]; extern struct linker_set linux_ioctl_handler_set; -static int linux_fixup __P((long **stack_base, +static int linux_fixup __P((register_t **stack_base, struct image_params *iparams)); -static int elf_linux_fixup __P((long **stack_base, +static int elf_linux_fixup __P((register_t **stack_base, struct image_params *iparams)); static void linux_prepsyscall __P((struct trapframe *tf, int *args, u_int *code, caddr_t *params)); static void linux_sendsig __P((sig_t catcher, int sig, sigset_t *mask, u_long code)); /* * Linux syscalls return negative errno's, we do positive and map them */ static int bsd_to_linux_errno[ELAST + 1] = { -0, -1, -2, -3, -4, -5, -6, -7, -8, -9, -10, -35, -12, -13, -14, -15, -16, -17, -18, -19, -20, -21, -22, -23, -24, -25, -26, -27, -28, -29, -30, -31, -32, -33, -34, -11,-115,-114, -88, -89, -90, -91, -92, -93, -94, -95, -96, -97, -98, -99, -100,-101,-102,-103,-104,-105,-106,-107,-108,-109, -110,-111, -40, -36,-112,-113, -39, -11, -87,-122, -116, -66, -6, -6, -6, -6, -6, -37, -38, -9, -6, -6, -43, -42, -75, -6, -84 }; int bsd_to_linux_signal[LINUX_SIGTBLSZ] = { LINUX_SIGHUP, LINUX_SIGINT, LINUX_SIGQUIT, LINUX_SIGILL, LINUX_SIGTRAP, LINUX_SIGABRT, 0, LINUX_SIGFPE, LINUX_SIGKILL, LINUX_SIGBUS, LINUX_SIGSEGV, 0, LINUX_SIGPIPE, LINUX_SIGALRM, LINUX_SIGTERM, LINUX_SIGURG, LINUX_SIGSTOP, LINUX_SIGTSTP, LINUX_SIGCONT, LINUX_SIGCHLD, LINUX_SIGTTIN, LINUX_SIGTTOU, LINUX_SIGIO, LINUX_SIGXCPU, LINUX_SIGXFSZ, LINUX_SIGVTALRM, LINUX_SIGPROF, LINUX_SIGWINCH, 0, LINUX_SIGUSR1, LINUX_SIGUSR2 }; int linux_to_bsd_signal[LINUX_SIGTBLSZ] = { SIGHUP, SIGINT, SIGQUIT, SIGILL, SIGTRAP, SIGABRT, SIGBUS, SIGFPE, SIGKILL, SIGUSR1, SIGSEGV, SIGUSR2, SIGPIPE, SIGALRM, SIGTERM, SIGBUS, SIGCHLD, SIGCONT, SIGSTOP, SIGTSTP, SIGTTIN, SIGTTOU, SIGURG, SIGXCPU, SIGXFSZ, SIGVTALRM, SIGPROF, SIGWINCH, SIGIO, SIGURG, 0 }; /* * If FreeBSD & Linux have a difference of opinion about what a trap * means, deal with it here. */ static int translate_traps(int signal, int trap_code) { if (signal != SIGBUS) return signal; switch (trap_code) { case T_PROTFLT: case T_TSSFLT: case T_DOUBLEFLT: case T_PAGEFLT: return SIGSEGV; default: return signal; } } static int -linux_fixup(long **stack_base, struct image_params *imgp) +linux_fixup(register_t **stack_base, struct image_params *imgp) { - long *argv, *envp; + register_t *argv, *envp; argv = *stack_base; envp = *stack_base + (imgp->argc + 1); (*stack_base)--; **stack_base = (intptr_t)(void *)envp; (*stack_base)--; **stack_base = (intptr_t)(void *)argv; (*stack_base)--; **stack_base = imgp->argc; return 0; } static int -elf_linux_fixup(long **stack_base, struct image_params *imgp) +elf_linux_fixup(register_t **stack_base, struct image_params *imgp) { Elf32_Auxargs *args = (Elf32_Auxargs *)imgp->auxargs; - long *pos; + register_t *pos; pos = *stack_base + (imgp->argc + imgp->envc + 2); if (args->trace) { AUXARGS_ENTRY(pos, AT_DEBUG, 1); } if (args->execfd != -1) { AUXARGS_ENTRY(pos, AT_EXECFD, args->execfd); } AUXARGS_ENTRY(pos, AT_PHDR, args->phdr); AUXARGS_ENTRY(pos, AT_PHENT, args->phent); AUXARGS_ENTRY(pos, AT_PHNUM, args->phnum); AUXARGS_ENTRY(pos, AT_PAGESZ, args->pagesz); AUXARGS_ENTRY(pos, AT_FLAGS, args->flags); AUXARGS_ENTRY(pos, AT_ENTRY, args->entry); AUXARGS_ENTRY(pos, AT_BASE, args->base); AUXARGS_ENTRY(pos, AT_UID, imgp->proc->p_cred->p_ruid); AUXARGS_ENTRY(pos, AT_EUID, imgp->proc->p_cred->p_svuid); AUXARGS_ENTRY(pos, AT_GID, imgp->proc->p_cred->p_rgid); AUXARGS_ENTRY(pos, AT_EGID, imgp->proc->p_cred->p_svgid); AUXARGS_ENTRY(pos, AT_NULL, 0); free(imgp->auxargs, M_TEMP); imgp->auxargs = NULL; (*stack_base)--; **stack_base = (long)imgp->argc; return 0; } extern int _ucodesel, _udatasel; /* * Send an interrupt to process. * * Stack is set up to allow sigcode stored * in u. to call routine, followed by kcall * to sigreturn routine below. After sigreturn * resets the signal mask, the stack, and the * frame pointer, it returns to the user * specified pc, psl. */ static void linux_sendsig(sig_t catcher, int sig, sigset_t *mask, u_long code) { register struct proc *p = curproc; register struct trapframe *regs; struct linux_sigframe *fp, frame; struct sigacts *psp = p->p_sigacts; int oonstack; regs = p->p_md.md_regs; oonstack = p->p_sigstk.ss_flags & SS_ONSTACK; #ifdef DEBUG printf("Linux-emul(%ld): linux_sendsig(%p, %d, %p, %lu)\n", (long)p->p_pid, catcher, sig, (void*)mask, code); #endif /* * Allocate space for the signal handler context. */ if ((p->p_flag & P_ALTSTACK) && !oonstack && SIGISMEMBER(psp->ps_sigonstack, sig)) { fp = (struct linux_sigframe *)(p->p_sigstk.ss_sp + p->p_sigstk.ss_size - sizeof(struct linux_sigframe)); p->p_sigstk.ss_flags |= SS_ONSTACK; } else { fp = (struct linux_sigframe *)regs->tf_esp - 1; } /* * grow() will return FALSE if the fp will not fit inside the stack * and the stack can not be grown. useracc will return FALSE * if access is denied. */ if ((grow_stack (p, (int)fp) == FALSE) || !useracc((caddr_t)fp, sizeof (struct linux_sigframe), VM_PROT_WRITE)) { /* * Process has trashed its stack; give it an illegal * instruction to halt it in its tracks. */ SIGACTION(p, SIGILL) = SIG_DFL; SIGDELSET(p->p_sigignore, SIGILL); SIGDELSET(p->p_sigcatch, SIGILL); SIGDELSET(p->p_sigmask, SIGILL); psignal(p, SIGILL); return; } /* * Build the argument list for the signal handler. */ if (p->p_sysent->sv_sigtbl) if (sig <= p->p_sysent->sv_sigsize) sig = p->p_sysent->sv_sigtbl[_SIG_IDX(sig)]; frame.sf_handler = catcher; frame.sf_sig = sig; /* * Build the signal context to be used by sigreturn. */ frame.sf_sc.sc_mask = mask->__bits[0]; frame.sf_sc.sc_gs = rgs(); frame.sf_sc.sc_fs = regs->tf_fs; frame.sf_sc.sc_es = regs->tf_es; frame.sf_sc.sc_ds = regs->tf_ds; frame.sf_sc.sc_edi = regs->tf_edi; frame.sf_sc.sc_esi = regs->tf_esi; frame.sf_sc.sc_ebp = regs->tf_ebp; frame.sf_sc.sc_ebx = regs->tf_ebx; frame.sf_sc.sc_edx = regs->tf_edx; frame.sf_sc.sc_ecx = regs->tf_ecx; frame.sf_sc.sc_eax = regs->tf_eax; frame.sf_sc.sc_eip = regs->tf_eip; frame.sf_sc.sc_cs = regs->tf_cs; frame.sf_sc.sc_eflags = regs->tf_eflags; frame.sf_sc.sc_esp_at_signal = regs->tf_esp; frame.sf_sc.sc_ss = regs->tf_ss; frame.sf_sc.sc_err = regs->tf_err; frame.sf_sc.sc_trapno = code; /* XXX ???? */ if (copyout(&frame, fp, sizeof(frame)) != 0) { /* * Process has trashed its stack; give it an illegal * instruction to halt it in its tracks. */ sigexit(p, SIGILL); /* NOTREACHED */ } /* * Build context to run handler in. */ regs->tf_esp = (int)fp; regs->tf_eip = PS_STRINGS - *(p->p_sysent->sv_szsigcode); regs->tf_eflags &= ~PSL_VM; regs->tf_cs = _ucodesel; regs->tf_ds = _udatasel; regs->tf_es = _udatasel; regs->tf_fs = _udatasel; load_gs(_udatasel); regs->tf_ss = _udatasel; } /* * System call to cleanup state after a signal * has been taken. Reset signal mask and * stack state from context left by sendsig (above). * Return to previous pc and psl as specified by * context left by sendsig. Check carefully to * make sure that the user has not modified the * psl to gain improper privileges or to cause * a machine fault. */ int linux_sigreturn(p, args) struct proc *p; struct linux_sigreturn_args *args; { struct linux_sigcontext *scp, context; register struct trapframe *regs; int eflags; regs = p->p_md.md_regs; #ifdef DEBUG printf("Linux-emul(%ld): linux_sigreturn(%p)\n", (long)p->p_pid, (void *)args->scp); #endif /* * The trampoline code hands us the context. * It is unsafe to keep track of it ourselves, in the event that a * program jumps out of a signal handler. */ scp = SCARG(args,scp); if (copyin((caddr_t)scp, &context, sizeof(*scp)) != 0) return (EFAULT); /* * Check for security violations. */ #define EFLAGS_SECURE(ef, oef) ((((ef) ^ (oef)) & ~PSL_USERCHANGE) == 0) eflags = context.sc_eflags; /* * XXX do allow users to change the privileged flag PSL_RF. The * cpu sets PSL_RF in tf_eflags for faults. Debuggers should * sometimes set it there too. tf_eflags is kept in the signal * context during signal handling and there is no other place * to remember it, so the PSL_RF bit may be corrupted by the * signal handler without us knowing. Corruption of the PSL_RF * bit at worst causes one more or one less debugger trap, so * allowing it is fairly harmless. */ if (!EFLAGS_SECURE(eflags & ~PSL_RF, regs->tf_eflags & ~PSL_RF)) { return(EINVAL); } /* * Don't allow users to load a valid privileged %cs. Let the * hardware check for invalid selectors, excess privilege in * other selectors, invalid %eip's and invalid %esp's. */ #define CS_SECURE(cs) (ISPL(cs) == SEL_UPL) if (!CS_SECURE(context.sc_cs)) { trapsignal(p, SIGBUS, T_PROTFLT); return(EINVAL); } p->p_sigstk.ss_flags &= ~SS_ONSTACK; SIGSETOLD(p->p_sigmask, context.sc_mask); SIG_CANTMASK(p->p_sigmask); /* * Restore signal context. */ /* %gs was restored by the trampoline. */ regs->tf_fs = context.sc_fs; regs->tf_es = context.sc_es; regs->tf_ds = context.sc_ds; regs->tf_edi = context.sc_edi; regs->tf_esi = context.sc_esi; regs->tf_ebp = context.sc_ebp; regs->tf_ebx = context.sc_ebx; regs->tf_edx = context.sc_edx; regs->tf_ecx = context.sc_ecx; regs->tf_eax = context.sc_eax; regs->tf_eip = context.sc_eip; regs->tf_cs = context.sc_cs; regs->tf_eflags = eflags; regs->tf_esp = context.sc_esp_at_signal; regs->tf_ss = context.sc_ss; return (EJUSTRETURN); } static void linux_prepsyscall(struct trapframe *tf, int *args, u_int *code, caddr_t *params) { args[0] = tf->tf_ebx; args[1] = tf->tf_ecx; args[2] = tf->tf_edx; args[3] = tf->tf_esi; args[4] = tf->tf_edi; *params = NULL; /* no copyin */ } struct sysentvec linux_sysvec = { LINUX_SYS_MAXSYSCALL, linux_sysent, 0xff, LINUX_SIGTBLSZ, bsd_to_linux_signal, ELAST + 1, bsd_to_linux_errno, translate_traps, linux_fixup, linux_sendsig, linux_sigcode, &linux_szsigcode, linux_prepsyscall, "Linux a.out", aout_coredump }; struct sysentvec elf_linux_sysvec = { LINUX_SYS_MAXSYSCALL, linux_sysent, 0xff, LINUX_SIGTBLSZ, bsd_to_linux_signal, ELAST + 1, bsd_to_linux_errno, translate_traps, elf_linux_fixup, linux_sendsig, linux_sigcode, &linux_szsigcode, linux_prepsyscall, "Linux ELF", elf_coredump }; static Elf32_Brandinfo linux_brand = { "Linux", "/compat/linux", "/lib/ld-linux.so.1", &elf_linux_sysvec }; static Elf32_Brandinfo linux_glibc2brand = { "Linux", "/compat/linux", "/lib/ld-linux.so.2", &elf_linux_sysvec }; Elf32_Brandinfo *linux_brandlist[] = { &linux_brand, &linux_glibc2brand, NULL }; static int linux_elf_modevent(module_t mod, int type, void *data) { Elf32_Brandinfo **brandinfo; int error; error = 0; switch(type) { case MOD_LOAD: for (brandinfo = &linux_brandlist[0]; *brandinfo != NULL; ++brandinfo) if (elf_insert_brand_entry(*brandinfo) < 0) error = EINVAL; if (error) printf("cannot insert Linux elf brand handler\n"); else { linux_ioctl_register_handlers(&linux_ioctl_handler_set); if (bootverbose) printf("Linux-ELF exec handler installed\n"); } break; case MOD_UNLOAD: linux_ioctl_unregister_handlers(&linux_ioctl_handler_set); for (brandinfo = &linux_brandlist[0]; *brandinfo != NULL; ++brandinfo) if (elf_brand_inuse(*brandinfo)) error = EBUSY; if (error == 0) { for (brandinfo = &linux_brandlist[0]; *brandinfo != NULL; ++brandinfo) if (elf_remove_brand_entry(*brandinfo) < 0) error = EINVAL; } if (error) printf("Could not deinstall ELF interpreter entry\n"); else if (bootverbose) printf("Linux-elf exec handler removed\n"); break; default: break; } return error; } static moduledata_t linux_elf_mod = { "linuxelf", linux_elf_modevent, 0 }; DECLARE_MODULE(linuxelf, linux_elf_mod, SI_SUB_EXEC, SI_ORDER_ANY); Index: head/sys/i386/linux/linux_sysvec.c =================================================================== --- head/sys/i386/linux/linux_sysvec.c (revision 55140) +++ head/sys/i386/linux/linux_sysvec.c (revision 55141) @@ -1,511 +1,511 @@ /*- * Copyright (c) 1994-1996 Søren Schmidt * 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 withough 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. * * $FreeBSD$ */ /* XXX we use functions that might not exist. */ #include "opt_compat.h" #ifndef COMPAT_43 #error "Unable to compile Linux-emulator due to missing COMPAT_43 option!" #endif #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include MALLOC_DEFINE(M_LINUX, "linux", "Linux mode structures"); extern char linux_sigcode[]; extern int linux_szsigcode; extern struct sysent linux_sysent[LINUX_SYS_MAXSYSCALL]; extern struct linker_set linux_ioctl_handler_set; -static int linux_fixup __P((long **stack_base, +static int linux_fixup __P((register_t **stack_base, struct image_params *iparams)); -static int elf_linux_fixup __P((long **stack_base, +static int elf_linux_fixup __P((register_t **stack_base, struct image_params *iparams)); static void linux_prepsyscall __P((struct trapframe *tf, int *args, u_int *code, caddr_t *params)); static void linux_sendsig __P((sig_t catcher, int sig, sigset_t *mask, u_long code)); /* * Linux syscalls return negative errno's, we do positive and map them */ static int bsd_to_linux_errno[ELAST + 1] = { -0, -1, -2, -3, -4, -5, -6, -7, -8, -9, -10, -35, -12, -13, -14, -15, -16, -17, -18, -19, -20, -21, -22, -23, -24, -25, -26, -27, -28, -29, -30, -31, -32, -33, -34, -11,-115,-114, -88, -89, -90, -91, -92, -93, -94, -95, -96, -97, -98, -99, -100,-101,-102,-103,-104,-105,-106,-107,-108,-109, -110,-111, -40, -36,-112,-113, -39, -11, -87,-122, -116, -66, -6, -6, -6, -6, -6, -37, -38, -9, -6, -6, -43, -42, -75, -6, -84 }; int bsd_to_linux_signal[LINUX_SIGTBLSZ] = { LINUX_SIGHUP, LINUX_SIGINT, LINUX_SIGQUIT, LINUX_SIGILL, LINUX_SIGTRAP, LINUX_SIGABRT, 0, LINUX_SIGFPE, LINUX_SIGKILL, LINUX_SIGBUS, LINUX_SIGSEGV, 0, LINUX_SIGPIPE, LINUX_SIGALRM, LINUX_SIGTERM, LINUX_SIGURG, LINUX_SIGSTOP, LINUX_SIGTSTP, LINUX_SIGCONT, LINUX_SIGCHLD, LINUX_SIGTTIN, LINUX_SIGTTOU, LINUX_SIGIO, LINUX_SIGXCPU, LINUX_SIGXFSZ, LINUX_SIGVTALRM, LINUX_SIGPROF, LINUX_SIGWINCH, 0, LINUX_SIGUSR1, LINUX_SIGUSR2 }; int linux_to_bsd_signal[LINUX_SIGTBLSZ] = { SIGHUP, SIGINT, SIGQUIT, SIGILL, SIGTRAP, SIGABRT, SIGBUS, SIGFPE, SIGKILL, SIGUSR1, SIGSEGV, SIGUSR2, SIGPIPE, SIGALRM, SIGTERM, SIGBUS, SIGCHLD, SIGCONT, SIGSTOP, SIGTSTP, SIGTTIN, SIGTTOU, SIGURG, SIGXCPU, SIGXFSZ, SIGVTALRM, SIGPROF, SIGWINCH, SIGIO, SIGURG, 0 }; /* * If FreeBSD & Linux have a difference of opinion about what a trap * means, deal with it here. */ static int translate_traps(int signal, int trap_code) { if (signal != SIGBUS) return signal; switch (trap_code) { case T_PROTFLT: case T_TSSFLT: case T_DOUBLEFLT: case T_PAGEFLT: return SIGSEGV; default: return signal; } } static int -linux_fixup(long **stack_base, struct image_params *imgp) +linux_fixup(register_t **stack_base, struct image_params *imgp) { - long *argv, *envp; + register_t *argv, *envp; argv = *stack_base; envp = *stack_base + (imgp->argc + 1); (*stack_base)--; **stack_base = (intptr_t)(void *)envp; (*stack_base)--; **stack_base = (intptr_t)(void *)argv; (*stack_base)--; **stack_base = imgp->argc; return 0; } static int -elf_linux_fixup(long **stack_base, struct image_params *imgp) +elf_linux_fixup(register_t **stack_base, struct image_params *imgp) { Elf32_Auxargs *args = (Elf32_Auxargs *)imgp->auxargs; - long *pos; + register_t *pos; pos = *stack_base + (imgp->argc + imgp->envc + 2); if (args->trace) { AUXARGS_ENTRY(pos, AT_DEBUG, 1); } if (args->execfd != -1) { AUXARGS_ENTRY(pos, AT_EXECFD, args->execfd); } AUXARGS_ENTRY(pos, AT_PHDR, args->phdr); AUXARGS_ENTRY(pos, AT_PHENT, args->phent); AUXARGS_ENTRY(pos, AT_PHNUM, args->phnum); AUXARGS_ENTRY(pos, AT_PAGESZ, args->pagesz); AUXARGS_ENTRY(pos, AT_FLAGS, args->flags); AUXARGS_ENTRY(pos, AT_ENTRY, args->entry); AUXARGS_ENTRY(pos, AT_BASE, args->base); AUXARGS_ENTRY(pos, AT_UID, imgp->proc->p_cred->p_ruid); AUXARGS_ENTRY(pos, AT_EUID, imgp->proc->p_cred->p_svuid); AUXARGS_ENTRY(pos, AT_GID, imgp->proc->p_cred->p_rgid); AUXARGS_ENTRY(pos, AT_EGID, imgp->proc->p_cred->p_svgid); AUXARGS_ENTRY(pos, AT_NULL, 0); free(imgp->auxargs, M_TEMP); imgp->auxargs = NULL; (*stack_base)--; **stack_base = (long)imgp->argc; return 0; } extern int _ucodesel, _udatasel; /* * Send an interrupt to process. * * Stack is set up to allow sigcode stored * in u. to call routine, followed by kcall * to sigreturn routine below. After sigreturn * resets the signal mask, the stack, and the * frame pointer, it returns to the user * specified pc, psl. */ static void linux_sendsig(sig_t catcher, int sig, sigset_t *mask, u_long code) { register struct proc *p = curproc; register struct trapframe *regs; struct linux_sigframe *fp, frame; struct sigacts *psp = p->p_sigacts; int oonstack; regs = p->p_md.md_regs; oonstack = p->p_sigstk.ss_flags & SS_ONSTACK; #ifdef DEBUG printf("Linux-emul(%ld): linux_sendsig(%p, %d, %p, %lu)\n", (long)p->p_pid, catcher, sig, (void*)mask, code); #endif /* * Allocate space for the signal handler context. */ if ((p->p_flag & P_ALTSTACK) && !oonstack && SIGISMEMBER(psp->ps_sigonstack, sig)) { fp = (struct linux_sigframe *)(p->p_sigstk.ss_sp + p->p_sigstk.ss_size - sizeof(struct linux_sigframe)); p->p_sigstk.ss_flags |= SS_ONSTACK; } else { fp = (struct linux_sigframe *)regs->tf_esp - 1; } /* * grow() will return FALSE if the fp will not fit inside the stack * and the stack can not be grown. useracc will return FALSE * if access is denied. */ if ((grow_stack (p, (int)fp) == FALSE) || !useracc((caddr_t)fp, sizeof (struct linux_sigframe), VM_PROT_WRITE)) { /* * Process has trashed its stack; give it an illegal * instruction to halt it in its tracks. */ SIGACTION(p, SIGILL) = SIG_DFL; SIGDELSET(p->p_sigignore, SIGILL); SIGDELSET(p->p_sigcatch, SIGILL); SIGDELSET(p->p_sigmask, SIGILL); psignal(p, SIGILL); return; } /* * Build the argument list for the signal handler. */ if (p->p_sysent->sv_sigtbl) if (sig <= p->p_sysent->sv_sigsize) sig = p->p_sysent->sv_sigtbl[_SIG_IDX(sig)]; frame.sf_handler = catcher; frame.sf_sig = sig; /* * Build the signal context to be used by sigreturn. */ frame.sf_sc.sc_mask = mask->__bits[0]; frame.sf_sc.sc_gs = rgs(); frame.sf_sc.sc_fs = regs->tf_fs; frame.sf_sc.sc_es = regs->tf_es; frame.sf_sc.sc_ds = regs->tf_ds; frame.sf_sc.sc_edi = regs->tf_edi; frame.sf_sc.sc_esi = regs->tf_esi; frame.sf_sc.sc_ebp = regs->tf_ebp; frame.sf_sc.sc_ebx = regs->tf_ebx; frame.sf_sc.sc_edx = regs->tf_edx; frame.sf_sc.sc_ecx = regs->tf_ecx; frame.sf_sc.sc_eax = regs->tf_eax; frame.sf_sc.sc_eip = regs->tf_eip; frame.sf_sc.sc_cs = regs->tf_cs; frame.sf_sc.sc_eflags = regs->tf_eflags; frame.sf_sc.sc_esp_at_signal = regs->tf_esp; frame.sf_sc.sc_ss = regs->tf_ss; frame.sf_sc.sc_err = regs->tf_err; frame.sf_sc.sc_trapno = code; /* XXX ???? */ if (copyout(&frame, fp, sizeof(frame)) != 0) { /* * Process has trashed its stack; give it an illegal * instruction to halt it in its tracks. */ sigexit(p, SIGILL); /* NOTREACHED */ } /* * Build context to run handler in. */ regs->tf_esp = (int)fp; regs->tf_eip = PS_STRINGS - *(p->p_sysent->sv_szsigcode); regs->tf_eflags &= ~PSL_VM; regs->tf_cs = _ucodesel; regs->tf_ds = _udatasel; regs->tf_es = _udatasel; regs->tf_fs = _udatasel; load_gs(_udatasel); regs->tf_ss = _udatasel; } /* * System call to cleanup state after a signal * has been taken. Reset signal mask and * stack state from context left by sendsig (above). * Return to previous pc and psl as specified by * context left by sendsig. Check carefully to * make sure that the user has not modified the * psl to gain improper privileges or to cause * a machine fault. */ int linux_sigreturn(p, args) struct proc *p; struct linux_sigreturn_args *args; { struct linux_sigcontext *scp, context; register struct trapframe *regs; int eflags; regs = p->p_md.md_regs; #ifdef DEBUG printf("Linux-emul(%ld): linux_sigreturn(%p)\n", (long)p->p_pid, (void *)args->scp); #endif /* * The trampoline code hands us the context. * It is unsafe to keep track of it ourselves, in the event that a * program jumps out of a signal handler. */ scp = SCARG(args,scp); if (copyin((caddr_t)scp, &context, sizeof(*scp)) != 0) return (EFAULT); /* * Check for security violations. */ #define EFLAGS_SECURE(ef, oef) ((((ef) ^ (oef)) & ~PSL_USERCHANGE) == 0) eflags = context.sc_eflags; /* * XXX do allow users to change the privileged flag PSL_RF. The * cpu sets PSL_RF in tf_eflags for faults. Debuggers should * sometimes set it there too. tf_eflags is kept in the signal * context during signal handling and there is no other place * to remember it, so the PSL_RF bit may be corrupted by the * signal handler without us knowing. Corruption of the PSL_RF * bit at worst causes one more or one less debugger trap, so * allowing it is fairly harmless. */ if (!EFLAGS_SECURE(eflags & ~PSL_RF, regs->tf_eflags & ~PSL_RF)) { return(EINVAL); } /* * Don't allow users to load a valid privileged %cs. Let the * hardware check for invalid selectors, excess privilege in * other selectors, invalid %eip's and invalid %esp's. */ #define CS_SECURE(cs) (ISPL(cs) == SEL_UPL) if (!CS_SECURE(context.sc_cs)) { trapsignal(p, SIGBUS, T_PROTFLT); return(EINVAL); } p->p_sigstk.ss_flags &= ~SS_ONSTACK; SIGSETOLD(p->p_sigmask, context.sc_mask); SIG_CANTMASK(p->p_sigmask); /* * Restore signal context. */ /* %gs was restored by the trampoline. */ regs->tf_fs = context.sc_fs; regs->tf_es = context.sc_es; regs->tf_ds = context.sc_ds; regs->tf_edi = context.sc_edi; regs->tf_esi = context.sc_esi; regs->tf_ebp = context.sc_ebp; regs->tf_ebx = context.sc_ebx; regs->tf_edx = context.sc_edx; regs->tf_ecx = context.sc_ecx; regs->tf_eax = context.sc_eax; regs->tf_eip = context.sc_eip; regs->tf_cs = context.sc_cs; regs->tf_eflags = eflags; regs->tf_esp = context.sc_esp_at_signal; regs->tf_ss = context.sc_ss; return (EJUSTRETURN); } static void linux_prepsyscall(struct trapframe *tf, int *args, u_int *code, caddr_t *params) { args[0] = tf->tf_ebx; args[1] = tf->tf_ecx; args[2] = tf->tf_edx; args[3] = tf->tf_esi; args[4] = tf->tf_edi; *params = NULL; /* no copyin */ } struct sysentvec linux_sysvec = { LINUX_SYS_MAXSYSCALL, linux_sysent, 0xff, LINUX_SIGTBLSZ, bsd_to_linux_signal, ELAST + 1, bsd_to_linux_errno, translate_traps, linux_fixup, linux_sendsig, linux_sigcode, &linux_szsigcode, linux_prepsyscall, "Linux a.out", aout_coredump }; struct sysentvec elf_linux_sysvec = { LINUX_SYS_MAXSYSCALL, linux_sysent, 0xff, LINUX_SIGTBLSZ, bsd_to_linux_signal, ELAST + 1, bsd_to_linux_errno, translate_traps, elf_linux_fixup, linux_sendsig, linux_sigcode, &linux_szsigcode, linux_prepsyscall, "Linux ELF", elf_coredump }; static Elf32_Brandinfo linux_brand = { "Linux", "/compat/linux", "/lib/ld-linux.so.1", &elf_linux_sysvec }; static Elf32_Brandinfo linux_glibc2brand = { "Linux", "/compat/linux", "/lib/ld-linux.so.2", &elf_linux_sysvec }; Elf32_Brandinfo *linux_brandlist[] = { &linux_brand, &linux_glibc2brand, NULL }; static int linux_elf_modevent(module_t mod, int type, void *data) { Elf32_Brandinfo **brandinfo; int error; error = 0; switch(type) { case MOD_LOAD: for (brandinfo = &linux_brandlist[0]; *brandinfo != NULL; ++brandinfo) if (elf_insert_brand_entry(*brandinfo) < 0) error = EINVAL; if (error) printf("cannot insert Linux elf brand handler\n"); else { linux_ioctl_register_handlers(&linux_ioctl_handler_set); if (bootverbose) printf("Linux-ELF exec handler installed\n"); } break; case MOD_UNLOAD: linux_ioctl_unregister_handlers(&linux_ioctl_handler_set); for (brandinfo = &linux_brandlist[0]; *brandinfo != NULL; ++brandinfo) if (elf_brand_inuse(*brandinfo)) error = EBUSY; if (error == 0) { for (brandinfo = &linux_brandlist[0]; *brandinfo != NULL; ++brandinfo) if (elf_remove_brand_entry(*brandinfo) < 0) error = EINVAL; } if (error) printf("Could not deinstall ELF interpreter entry\n"); else if (bootverbose) printf("Linux-elf exec handler removed\n"); break; default: break; } return error; } static moduledata_t linux_elf_mod = { "linuxelf", linux_elf_modevent, 0 }; DECLARE_MODULE(linuxelf, linux_elf_mod, SI_SUB_EXEC, SI_ORDER_ANY); Index: head/sys/kern/imgact_elf.c =================================================================== --- head/sys/kern/imgact_elf.c (revision 55140) +++ head/sys/kern/imgact_elf.c (revision 55141) @@ -1,990 +1,990 @@ /*- * Copyright (c) 1995-1996 Søren Schmidt * Copyright (c) 1996 Peter Wemm * 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 withough 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. * * $FreeBSD$ */ #include "opt_rlimit.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include __ElfType(Brandinfo); __ElfType(Auxargs); static int elf_check_header __P((const Elf_Ehdr *hdr)); -static int elf_freebsd_fixup __P((long **stack_base, +static int elf_freebsd_fixup __P((register_t **stack_base, struct image_params *imgp)); static int elf_load_file __P((struct proc *p, const char *file, u_long *addr, u_long *entry)); static int elf_load_section __P((struct proc *p, struct vmspace *vmspace, struct vnode *vp, vm_offset_t offset, caddr_t vmaddr, size_t memsz, size_t filsz, vm_prot_t prot)); static int exec_elf_imgact __P((struct image_params *imgp)); static int elf_trace = 0; SYSCTL_INT(_debug, OID_AUTO, elf_trace, CTLFLAG_RW, &elf_trace, 0, ""); /* * XXX Maximum length of an ELF brand (sysctl wants a statically-allocated * buffer). */ #define MAXBRANDLEN 16 static struct sysentvec elf_freebsd_sysvec = { SYS_MAXSYSCALL, sysent, 0, 0, 0, 0, 0, 0, elf_freebsd_fixup, sendsig, sigcode, &szsigcode, 0, "FreeBSD ELF", elf_coredump }; static Elf_Brandinfo freebsd_brand_info = { "FreeBSD", "", "/usr/libexec/ld-elf.so.1", &elf_freebsd_sysvec }; static Elf_Brandinfo *elf_brand_list[MAX_BRANDS] = { &freebsd_brand_info, NULL, NULL, NULL, NULL, NULL, NULL, NULL }; int elf_insert_brand_entry(Elf_Brandinfo *entry) { int i; for (i=1; ip_sysent == entry->sysvec) return TRUE; } return FALSE; } static int elf_check_header(const Elf_Ehdr *hdr) { if (!IS_ELF(*hdr) || hdr->e_ident[EI_CLASS] != ELF_TARG_CLASS || hdr->e_ident[EI_DATA] != ELF_TARG_DATA || hdr->e_ident[EI_VERSION] != EV_CURRENT) return ENOEXEC; if (!ELF_MACHINE_OK(hdr->e_machine)) return ENOEXEC; if (hdr->e_version != ELF_TARG_VER) return ENOEXEC; return 0; } static int elf_load_section(struct proc *p, struct vmspace *vmspace, struct vnode *vp, vm_offset_t offset, caddr_t vmaddr, size_t memsz, size_t filsz, vm_prot_t prot) { size_t map_len; vm_offset_t map_addr; int error, rv; size_t copy_len; vm_object_t object; vm_offset_t file_addr; vm_offset_t data_buf = 0; object = vp->v_object; error = 0; map_addr = trunc_page((vm_offset_t)vmaddr); file_addr = trunc_page(offset); /* * We have two choices. We can either clear the data in the last page * of an oversized mapping, or we can start the anon mapping a page * early and copy the initialized data into that first page. We * choose the second.. */ if (memsz > filsz) map_len = trunc_page(offset+filsz) - file_addr; else map_len = round_page(offset+filsz) - file_addr; if (map_len != 0) { vm_object_reference(object); vm_map_lock(&vmspace->vm_map); rv = vm_map_insert(&vmspace->vm_map, object, file_addr, /* file offset */ map_addr, /* virtual start */ map_addr + map_len,/* virtual end */ prot, VM_PROT_ALL, MAP_COPY_ON_WRITE | MAP_PREFAULT); vm_map_unlock(&vmspace->vm_map); if (rv != KERN_SUCCESS) { vm_object_deallocate(object); return EINVAL; } /* we can stop now if we've covered it all */ if (memsz == filsz) return 0; } /* * We have to get the remaining bit of the file into the first part * of the oversized map segment. This is normally because the .data * segment in the file is extended to provide bss. It's a neat idea * to try and save a page, but it's a pain in the behind to implement. */ copy_len = (offset + filsz) - trunc_page(offset + filsz); map_addr = trunc_page((vm_offset_t)vmaddr + filsz); map_len = round_page((vm_offset_t)vmaddr + memsz) - map_addr; /* This had damn well better be true! */ if (map_len != 0) { vm_map_lock(&vmspace->vm_map); rv = vm_map_insert(&vmspace->vm_map, NULL, 0, map_addr, map_addr + map_len, VM_PROT_ALL, VM_PROT_ALL, 0); vm_map_unlock(&vmspace->vm_map); if (rv != KERN_SUCCESS) return EINVAL; } if (copy_len != 0) { vm_object_reference(object); rv = vm_map_find(exec_map, object, trunc_page(offset + filsz), &data_buf, PAGE_SIZE, TRUE, VM_PROT_READ, VM_PROT_ALL, MAP_COPY_ON_WRITE | MAP_PREFAULT_PARTIAL); if (rv != KERN_SUCCESS) { vm_object_deallocate(object); return EINVAL; } /* send the page fragment to user space */ error = copyout((caddr_t)data_buf, (caddr_t)map_addr, copy_len); vm_map_remove(exec_map, data_buf, data_buf + PAGE_SIZE); if (error) return (error); } /* * set it to the specified protection */ vm_map_protect(&vmspace->vm_map, map_addr, map_addr + map_len, prot, FALSE); return error; } /* * Load the file "file" into memory. It may be either a shared object * or an executable. * * The "addr" reference parameter is in/out. On entry, it specifies * the address where a shared object should be loaded. If the file is * an executable, this value is ignored. On exit, "addr" specifies * where the file was actually loaded. * * The "entry" reference parameter is out only. On exit, it specifies * the entry point for the loaded file. */ static int elf_load_file(struct proc *p, const char *file, u_long *addr, u_long *entry) { const Elf_Ehdr *hdr = NULL; const Elf_Phdr *phdr = NULL; struct nameidata nd; struct vmspace *vmspace = p->p_vmspace; struct vattr attr; struct image_params image_params, *imgp; vm_prot_t prot; u_long rbase; u_long base_addr = 0; int error, i, numsegs; imgp = &image_params; /* * Initialize part of the common data */ imgp->proc = p; imgp->uap = NULL; imgp->attr = &attr; imgp->firstpage = NULL; imgp->image_header = (char *)kmem_alloc_wait(exec_map, PAGE_SIZE); if (imgp->image_header == NULL) { nd.ni_vp = NULL; error = ENOMEM; goto fail; } NDINIT(&nd, LOOKUP, LOCKLEAF|FOLLOW, UIO_SYSSPACE, file, p); if ((error = namei(&nd)) != 0) { nd.ni_vp = NULL; goto fail; } NDFREE(&nd, NDF_ONLY_PNBUF); imgp->vp = nd.ni_vp; /* * Check permissions, modes, uid, etc on the file, and "open" it. */ error = exec_check_permissions(imgp); if (error) { VOP_UNLOCK(nd.ni_vp, 0, p); goto fail; } error = exec_map_first_page(imgp); VOP_UNLOCK(nd.ni_vp, 0, p); if (error) goto fail; hdr = (const Elf_Ehdr *)imgp->image_header; if ((error = elf_check_header(hdr)) != 0) goto fail; if (hdr->e_type == ET_DYN) rbase = *addr; else if (hdr->e_type == ET_EXEC) rbase = 0; else { error = ENOEXEC; goto fail; } /* Only support headers that fit within first page for now */ if ((hdr->e_phoff > PAGE_SIZE) || (hdr->e_phoff + hdr->e_phentsize * hdr->e_phnum) > PAGE_SIZE) { error = ENOEXEC; goto fail; } phdr = (const Elf_Phdr *)(imgp->image_header + hdr->e_phoff); for (i = 0, numsegs = 0; i < hdr->e_phnum; i++) { if (phdr[i].p_type == PT_LOAD) { /* Loadable segment */ prot = 0; if (phdr[i].p_flags & PF_X) prot |= VM_PROT_EXECUTE; if (phdr[i].p_flags & PF_W) prot |= VM_PROT_WRITE; if (phdr[i].p_flags & PF_R) prot |= VM_PROT_READ; if ((error = elf_load_section(p, vmspace, nd.ni_vp, phdr[i].p_offset, (caddr_t)phdr[i].p_vaddr + rbase, phdr[i].p_memsz, phdr[i].p_filesz, prot)) != 0) goto fail; /* * Establish the base address if this is the * first segment. */ if (numsegs == 0) base_addr = trunc_page(phdr[i].p_vaddr + rbase); numsegs++; } } *addr = base_addr; *entry=(unsigned long)hdr->e_entry + rbase; fail: if (imgp->firstpage) exec_unmap_first_page(imgp); if (imgp->image_header) kmem_free_wakeup(exec_map, (vm_offset_t)imgp->image_header, PAGE_SIZE); if (nd.ni_vp) vrele(nd.ni_vp); return error; } static char fallback_elf_brand[MAXBRANDLEN+1] = { "none" }; SYSCTL_STRING(_kern, OID_AUTO, fallback_elf_brand, CTLFLAG_RW, fallback_elf_brand, sizeof(fallback_elf_brand), "ELF brand of last resort"); static int exec_elf_imgact(struct image_params *imgp) { const Elf_Ehdr *hdr = (const Elf_Ehdr *) imgp->image_header; const Elf_Phdr *phdr; Elf_Auxargs *elf_auxargs = NULL; struct vmspace *vmspace; vm_prot_t prot; u_long text_size = 0, data_size = 0; u_long text_addr = 0, data_addr = 0; u_long addr, entry = 0, proghdr = 0; int error, i; const char *interp = NULL; Elf_Brandinfo *brand_info; const char *brand; char path[MAXPATHLEN]; /* * Do we have a valid ELF header ? */ if (elf_check_header(hdr) != 0 || hdr->e_type != ET_EXEC) return -1; /* * From here on down, we return an errno, not -1, as we've * detected an ELF file. */ if ((hdr->e_phoff > PAGE_SIZE) || (hdr->e_phoff + hdr->e_phentsize * hdr->e_phnum) > PAGE_SIZE) { /* Only support headers in first page for now */ return ENOEXEC; } phdr = (const Elf_Phdr*)(imgp->image_header + hdr->e_phoff); /* * From this point on, we may have resources that need to be freed. */ if ((error = exec_extract_strings(imgp)) != 0) goto fail; exec_new_vmspace(imgp); vmspace = imgp->proc->p_vmspace; for (i = 0; i < hdr->e_phnum; i++) { switch(phdr[i].p_type) { case PT_LOAD: /* Loadable segment */ prot = 0; if (phdr[i].p_flags & PF_X) prot |= VM_PROT_EXECUTE; if (phdr[i].p_flags & PF_W) prot |= VM_PROT_WRITE; if (phdr[i].p_flags & PF_R) prot |= VM_PROT_READ; if ((error = elf_load_section(imgp->proc, vmspace, imgp->vp, phdr[i].p_offset, (caddr_t)phdr[i].p_vaddr, phdr[i].p_memsz, phdr[i].p_filesz, prot)) != 0) goto fail; /* * Is this .text or .data ?? * * We only handle one each of those yet XXX */ if (hdr->e_entry >= phdr[i].p_vaddr && hdr->e_entry <(phdr[i].p_vaddr+phdr[i].p_memsz)) { text_addr = trunc_page(phdr[i].p_vaddr); text_size = round_page(phdr[i].p_memsz + phdr[i].p_vaddr - text_addr); entry = (u_long)hdr->e_entry; } else { data_addr = trunc_page(phdr[i].p_vaddr); data_size = round_page(phdr[i].p_memsz + phdr[i].p_vaddr - data_addr); } break; case PT_INTERP: /* Path to interpreter */ if (phdr[i].p_filesz > MAXPATHLEN || phdr[i].p_offset + phdr[i].p_filesz > PAGE_SIZE) { error = ENOEXEC; goto fail; } interp = imgp->image_header + phdr[i].p_offset; break; case PT_PHDR: /* Program header table info */ proghdr = phdr[i].p_vaddr; break; default: break; } } vmspace->vm_tsize = text_size >> PAGE_SHIFT; vmspace->vm_taddr = (caddr_t)(uintptr_t)text_addr; vmspace->vm_dsize = data_size >> PAGE_SHIFT; vmspace->vm_daddr = (caddr_t)(uintptr_t)data_addr; addr = ELF_RTLD_ADDR(vmspace); imgp->entry_addr = entry; /* If the executable has a brand, search for it in the brand list. */ brand_info = NULL; brand = (const char *)&hdr->e_ident[EI_BRAND]; if (brand[0] != '\0') { for (i = 0; i < MAX_BRANDS; i++) { Elf_Brandinfo *bi = elf_brand_list[i]; if (bi != NULL && strcmp(brand, bi->brand) == 0) { brand_info = bi; break; } } } /* Lacking a known brand, search for a recognized interpreter. */ if (brand_info == NULL && interp != NULL) { for (i = 0; i < MAX_BRANDS; i++) { Elf_Brandinfo *bi = elf_brand_list[i]; if (bi != NULL && strcmp(interp, bi->interp_path) == 0) { brand_info = bi; break; } } } /* Lacking a recognized interpreter, try the default brand */ if (brand_info == NULL && fallback_elf_brand[0] != '\0') { for (i = 0; i < MAX_BRANDS; i++) { Elf_Brandinfo *bi = elf_brand_list[i]; if (bi != NULL && strcmp(fallback_elf_brand, bi->brand) == 0) { brand_info = bi; break; } } } #ifdef __alpha__ /* XXX - Assume FreeBSD on the alpha. */ if (brand_info == NULL) brand_info = &freebsd_brand_info; #endif if (brand_info == NULL) { if (brand[0] == 0) uprintf("ELF binary type not known." " Use \"brandelf\" to brand it.\n"); else uprintf("ELF binary type \"%.*s\" not known.\n", EI_NIDENT - EI_BRAND, brand); error = ENOEXEC; goto fail; } imgp->proc->p_sysent = brand_info->sysvec; if (interp != NULL) { snprintf(path, sizeof(path), "%s%s", brand_info->emul_path, interp); if ((error = elf_load_file(imgp->proc, path, &addr, &imgp->entry_addr)) != 0) { if ((error = elf_load_file(imgp->proc, interp, &addr, &imgp->entry_addr)) != 0) { uprintf("ELF interpreter %s not found\n", path); goto fail; } } } /* * Construct auxargs table (used by the fixup routine) */ elf_auxargs = malloc(sizeof(Elf_Auxargs), M_TEMP, M_WAITOK); elf_auxargs->execfd = -1; elf_auxargs->phdr = proghdr; elf_auxargs->phent = hdr->e_phentsize; elf_auxargs->phnum = hdr->e_phnum; elf_auxargs->pagesz = PAGE_SIZE; elf_auxargs->base = addr; elf_auxargs->flags = 0; elf_auxargs->entry = entry; elf_auxargs->trace = elf_trace; imgp->auxargs = elf_auxargs; imgp->interpreted = 0; /* don't allow modifying the file while we run it */ imgp->vp->v_flag |= VTEXT; fail: return error; } static int -elf_freebsd_fixup(long **stack_base, struct image_params *imgp) +elf_freebsd_fixup(register_t **stack_base, struct image_params *imgp) { Elf_Auxargs *args = (Elf_Auxargs *)imgp->auxargs; - long *pos; + register_t *pos; pos = *stack_base + (imgp->argc + imgp->envc + 2); if (args->trace) { AUXARGS_ENTRY(pos, AT_DEBUG, 1); } if (args->execfd != -1) { AUXARGS_ENTRY(pos, AT_EXECFD, args->execfd); } AUXARGS_ENTRY(pos, AT_PHDR, args->phdr); AUXARGS_ENTRY(pos, AT_PHENT, args->phent); AUXARGS_ENTRY(pos, AT_PHNUM, args->phnum); AUXARGS_ENTRY(pos, AT_PAGESZ, args->pagesz); AUXARGS_ENTRY(pos, AT_FLAGS, args->flags); AUXARGS_ENTRY(pos, AT_ENTRY, args->entry); AUXARGS_ENTRY(pos, AT_BASE, args->base); AUXARGS_ENTRY(pos, AT_NULL, 0); free(imgp->auxargs, M_TEMP); imgp->auxargs = NULL; (*stack_base)--; suword(*stack_base, (long) imgp->argc); return 0; } /* * Code for generating ELF core dumps. */ typedef void (*segment_callback) __P((vm_map_entry_t, void *)); /* Closure for cb_put_phdr(). */ struct phdr_closure { Elf_Phdr *phdr; /* Program header to fill in */ Elf_Off offset; /* Offset of segment in core file */ }; /* Closure for cb_size_segment(). */ struct sseg_closure { int count; /* Count of writable segments. */ size_t size; /* Total size of all writable segments. */ }; static void cb_put_phdr __P((vm_map_entry_t, void *)); static void cb_size_segment __P((vm_map_entry_t, void *)); static void each_writable_segment __P((struct proc *, segment_callback, void *)); static int elf_corehdr __P((struct proc *, struct vnode *, struct ucred *, int, void *, size_t)); static void elf_puthdr __P((struct proc *, void *, size_t *, const prstatus_t *, const prfpregset_t *, const prpsinfo_t *, int)); static void elf_putnote __P((void *, size_t *, const char *, int, const void *, size_t)); extern int osreldate; int elf_coredump(p, vp, limit) register struct proc *p; register struct vnode *vp; off_t limit; { register struct ucred *cred = p->p_ucred; int error = 0; struct sseg_closure seginfo; void *hdr; size_t hdrsize; /* Size the program segments. */ seginfo.count = 0; seginfo.size = 0; each_writable_segment(p, cb_size_segment, &seginfo); /* * Calculate the size of the core file header area by making * a dry run of generating it. Nothing is written, but the * size is calculated. */ hdrsize = 0; elf_puthdr((struct proc *)NULL, (void *)NULL, &hdrsize, (const prstatus_t *)NULL, (const prfpregset_t *)NULL, (const prpsinfo_t *)NULL, seginfo.count); if (hdrsize + seginfo.size >= limit) return (EFAULT); /* * Allocate memory for building the header, fill it up, * and write it out. */ hdr = malloc(hdrsize, M_TEMP, M_WAITOK); if (hdr == NULL) { return EINVAL; } error = elf_corehdr(p, vp, cred, seginfo.count, hdr, hdrsize); /* Write the contents of all of the writable segments. */ if (error == 0) { Elf_Phdr *php; off_t offset; int i; php = (Elf_Phdr *)((char *)hdr + sizeof(Elf_Ehdr)) + 1; offset = hdrsize; for (i = 0; i < seginfo.count; i++) { error = vn_rdwr(UIO_WRITE, vp, (caddr_t)php->p_vaddr, php->p_filesz, offset, UIO_USERSPACE, IO_NODELOCKED|IO_UNIT, cred, (int *)NULL, p); if (error != 0) break; offset += php->p_filesz; php++; } } free(hdr, M_TEMP); return error; } /* * A callback for each_writable_segment() to write out the segment's * program header entry. */ static void cb_put_phdr(entry, closure) vm_map_entry_t entry; void *closure; { struct phdr_closure *phc = (struct phdr_closure *)closure; Elf_Phdr *phdr = phc->phdr; phc->offset = round_page(phc->offset); phdr->p_type = PT_LOAD; phdr->p_offset = phc->offset; phdr->p_vaddr = entry->start; phdr->p_paddr = 0; phdr->p_filesz = phdr->p_memsz = entry->end - entry->start; phdr->p_align = PAGE_SIZE; phdr->p_flags = 0; if (entry->protection & VM_PROT_READ) phdr->p_flags |= PF_R; if (entry->protection & VM_PROT_WRITE) phdr->p_flags |= PF_W; if (entry->protection & VM_PROT_EXECUTE) phdr->p_flags |= PF_X; phc->offset += phdr->p_filesz; phc->phdr++; } /* * A callback for each_writable_segment() to gather information about * the number of segments and their total size. */ static void cb_size_segment(entry, closure) vm_map_entry_t entry; void *closure; { struct sseg_closure *ssc = (struct sseg_closure *)closure; ssc->count++; ssc->size += entry->end - entry->start; } /* * For each writable segment in the process's memory map, call the given * function with a pointer to the map entry and some arbitrary * caller-supplied data. */ static void each_writable_segment(p, func, closure) struct proc *p; segment_callback func; void *closure; { vm_map_t map = &p->p_vmspace->vm_map; vm_map_entry_t entry; for (entry = map->header.next; entry != &map->header; entry = entry->next) { vm_object_t obj; if ((entry->eflags & MAP_ENTRY_IS_SUB_MAP) || (entry->protection & (VM_PROT_READ|VM_PROT_WRITE)) != (VM_PROT_READ|VM_PROT_WRITE)) continue; if ((obj = entry->object.vm_object) == NULL) continue; /* Find the deepest backing object. */ while (obj->backing_object != NULL) obj = obj->backing_object; /* Ignore memory-mapped devices and such things. */ if (obj->type != OBJT_DEFAULT && obj->type != OBJT_SWAP && obj->type != OBJT_VNODE) continue; (*func)(entry, closure); } } /* * Write the core file header to the file, including padding up to * the page boundary. */ static int elf_corehdr(p, vp, cred, numsegs, hdr, hdrsize) struct proc *p; struct vnode *vp; struct ucred *cred; int numsegs; size_t hdrsize; void *hdr; { size_t off; prstatus_t status; prfpregset_t fpregset; prpsinfo_t psinfo; /* Gather the information for the header. */ bzero(&status, sizeof status); status.pr_version = PRSTATUS_VERSION; status.pr_statussz = sizeof(prstatus_t); status.pr_gregsetsz = sizeof(gregset_t); status.pr_fpregsetsz = sizeof(fpregset_t); status.pr_osreldate = osreldate; status.pr_cursig = p->p_sig; status.pr_pid = p->p_pid; fill_regs(p, &status.pr_reg); fill_fpregs(p, &fpregset); bzero(&psinfo, sizeof psinfo); psinfo.pr_version = PRPSINFO_VERSION; psinfo.pr_psinfosz = sizeof(prpsinfo_t); strncpy(psinfo.pr_fname, p->p_comm, MAXCOMLEN); /* XXX - We don't fill in the command line arguments properly yet. */ strncpy(psinfo.pr_psargs, p->p_comm, PRARGSZ); /* Fill in the header. */ bzero(hdr, hdrsize); off = 0; elf_puthdr(p, hdr, &off, &status, &fpregset, &psinfo, numsegs); /* Write it to the core file. */ return vn_rdwr(UIO_WRITE, vp, hdr, hdrsize, (off_t)0, UIO_SYSSPACE, IO_NODELOCKED|IO_UNIT, cred, NULL, p); } static void elf_puthdr(struct proc *p, void *dst, size_t *off, const prstatus_t *status, const prfpregset_t *fpregset, const prpsinfo_t *psinfo, int numsegs) { size_t ehoff; size_t phoff; size_t noteoff; size_t notesz; ehoff = *off; *off += sizeof(Elf_Ehdr); phoff = *off; *off += (numsegs + 1) * sizeof(Elf_Phdr); noteoff = *off; elf_putnote(dst, off, "FreeBSD", NT_PRSTATUS, status, sizeof *status); elf_putnote(dst, off, "FreeBSD", NT_FPREGSET, fpregset, sizeof *fpregset); elf_putnote(dst, off, "FreeBSD", NT_PRPSINFO, psinfo, sizeof *psinfo); notesz = *off - noteoff; /* Align up to a page boundary for the program segments. */ *off = round_page(*off); if (dst != NULL) { Elf_Ehdr *ehdr; Elf_Phdr *phdr; struct phdr_closure phc; /* * Fill in the ELF header. */ ehdr = (Elf_Ehdr *)((char *)dst + ehoff); ehdr->e_ident[EI_MAG0] = ELFMAG0; ehdr->e_ident[EI_MAG1] = ELFMAG1; ehdr->e_ident[EI_MAG2] = ELFMAG2; ehdr->e_ident[EI_MAG3] = ELFMAG3; ehdr->e_ident[EI_CLASS] = ELF_CLASS; ehdr->e_ident[EI_DATA] = ELF_DATA; ehdr->e_ident[EI_VERSION] = EV_CURRENT; ehdr->e_ident[EI_PAD] = 0; strncpy(ehdr->e_ident + EI_BRAND, "FreeBSD", EI_NIDENT - EI_BRAND); ehdr->e_type = ET_CORE; ehdr->e_machine = ELF_ARCH; ehdr->e_version = EV_CURRENT; ehdr->e_entry = 0; ehdr->e_phoff = phoff; ehdr->e_flags = 0; ehdr->e_ehsize = sizeof(Elf_Ehdr); ehdr->e_phentsize = sizeof(Elf_Phdr); ehdr->e_phnum = numsegs + 1; ehdr->e_shentsize = sizeof(Elf_Shdr); ehdr->e_shnum = 0; ehdr->e_shstrndx = SHN_UNDEF; /* * Fill in the program header entries. */ phdr = (Elf_Phdr *)((char *)dst + phoff); /* The note segement. */ phdr->p_type = PT_NOTE; phdr->p_offset = noteoff; phdr->p_vaddr = 0; phdr->p_paddr = 0; phdr->p_filesz = notesz; phdr->p_memsz = 0; phdr->p_flags = 0; phdr->p_align = 0; phdr++; /* All the writable segments from the program. */ phc.phdr = phdr; phc.offset = *off; each_writable_segment(p, cb_put_phdr, &phc); } } static void elf_putnote(void *dst, size_t *off, const char *name, int type, const void *desc, size_t descsz) { Elf_Note note; note.n_namesz = strlen(name) + 1; note.n_descsz = descsz; note.n_type = type; if (dst != NULL) bcopy(¬e, (char *)dst + *off, sizeof note); *off += sizeof note; if (dst != NULL) bcopy(name, (char *)dst + *off, note.n_namesz); *off += roundup2(note.n_namesz, sizeof(Elf_Size)); if (dst != NULL) bcopy(desc, (char *)dst + *off, note.n_descsz); *off += roundup2(note.n_descsz, sizeof(Elf_Size)); } /* * Tell kern_execve.c about it, with a little help from the linker. */ static struct execsw elf_execsw = {exec_elf_imgact, "ELF"}; EXEC_SET(elf, elf_execsw); Index: head/sys/kern/kern_exec.c =================================================================== --- head/sys/kern/kern_exec.c (revision 55140) +++ head/sys/kern/kern_exec.c (revision 55141) @@ -1,791 +1,791 @@ /* * Copyright (c) 1993, David Greenman * 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. * * $FreeBSD$ */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include MALLOC_DEFINE(M_PARGS, "proc-args", "Process arguments"); -static long *exec_copyout_strings __P((struct image_params *)); +static register_t *exec_copyout_strings __P((struct image_params *)); static long ps_strings = PS_STRINGS; SYSCTL_LONG(_kern, KERN_PS_STRINGS, ps_strings, CTLFLAG_RD, &ps_strings, ""); static long usrstack = USRSTACK; SYSCTL_LONG(_kern, KERN_USRSTACK, usrstack, CTLFLAG_RD, &usrstack, ""); u_long ps_arg_cache_limit = PAGE_SIZE / 16; SYSCTL_LONG(_kern, OID_AUTO, ps_arg_cache_limit, CTLFLAG_RW, &ps_arg_cache_limit, ""); int ps_argsopen = 1; SYSCTL_INT(_kern, OID_AUTO, ps_argsopen, CTLFLAG_RW, &ps_argsopen, 0, ""); /* * Each of the items is a pointer to a `const struct execsw', hence the * double pointer here. */ static const struct execsw **execsw; #ifndef _SYS_SYSPROTO_H_ struct execve_args { char *fname; char **argv; char **envv; }; #endif /* * execve() system call. */ int execve(p, uap) struct proc *p; register struct execve_args *uap; { struct nameidata nd, *ndp; - long *stack_base; + register_t *stack_base; int error, len, i; struct image_params image_params, *imgp; struct vattr attr; imgp = &image_params; /* * Initialize part of the common data */ imgp->proc = p; imgp->uap = uap; imgp->attr = &attr; imgp->argc = imgp->envc = 0; imgp->argv0 = NULL; imgp->entry_addr = 0; imgp->vmspace_destroyed = 0; imgp->interpreted = 0; imgp->interpreter_name[0] = '\0'; imgp->auxargs = NULL; imgp->vp = NULL; imgp->firstpage = NULL; imgp->ps_strings = 0; /* * Allocate temporary demand zeroed space for argument and * environment strings */ imgp->stringbase = (char *)kmem_alloc_wait(exec_map, ARG_MAX + PAGE_SIZE); if (imgp->stringbase == NULL) { error = ENOMEM; goto exec_fail; } imgp->stringp = imgp->stringbase; imgp->stringspace = ARG_MAX; imgp->image_header = imgp->stringbase + ARG_MAX; /* * Translate the file name. namei() returns a vnode pointer * in ni_vp amoung other things. */ ndp = &nd; NDINIT(ndp, LOOKUP, LOCKLEAF | FOLLOW | SAVENAME, UIO_USERSPACE, uap->fname, p); interpret: error = namei(ndp); if (error) { kmem_free_wakeup(exec_map, (vm_offset_t)imgp->stringbase, ARG_MAX + PAGE_SIZE); goto exec_fail; } imgp->vp = ndp->ni_vp; imgp->fname = uap->fname; /* * Check file permissions (also 'opens' file) */ error = exec_check_permissions(imgp); if (error) { VOP_UNLOCK(imgp->vp, 0, p); goto exec_fail_dealloc; } error = exec_map_first_page(imgp); VOP_UNLOCK(imgp->vp, 0, p); if (error) goto exec_fail_dealloc; /* * Loop through list of image activators, calling each one. * If there is no match, the activator returns -1. If there * is a match, but there was an error during the activation, * the error is returned. Otherwise 0 means success. If the * image is interpreted, loop back up and try activating * the interpreter. */ for (i = 0; execsw[i]; ++i) { if (execsw[i]->ex_imgact) error = (*execsw[i]->ex_imgact)(imgp); else continue; if (error == -1) continue; if (error) goto exec_fail_dealloc; if (imgp->interpreted) { exec_unmap_first_page(imgp); /* free name buffer and old vnode */ NDFREE(ndp, NDF_ONLY_PNBUF); vrele(ndp->ni_vp); /* set new name to that of the interpreter */ NDINIT(ndp, LOOKUP, LOCKLEAF | FOLLOW | SAVENAME, UIO_SYSSPACE, imgp->interpreter_name, p); goto interpret; } break; } /* If we made it through all the activators and none matched, exit. */ if (error == -1) { error = ENOEXEC; goto exec_fail_dealloc; } /* * Copy out strings (args and env) and initialize stack base */ stack_base = exec_copyout_strings(imgp); p->p_vmspace->vm_minsaddr = (char *)stack_base; /* * If custom stack fixup routine present for this process * let it do the stack setup. * Else stuff argument count as first item on stack */ if (p->p_sysent->sv_fixup) (*p->p_sysent->sv_fixup)(&stack_base, imgp); else suword(--stack_base, imgp->argc); /* * For security and other reasons, the file descriptor table cannot * be shared after an exec. */ if (p->p_fd->fd_refcnt > 1) { struct filedesc *tmp; tmp = fdcopy(p); fdfree(p); p->p_fd = tmp; } /* Stop profiling */ stopprofclock(p); /* close files on exec */ fdcloseexec(p); /* reset caught signals */ execsigs(p); /* name this process - nameiexec(p, ndp) */ len = min(ndp->ni_cnd.cn_namelen,MAXCOMLEN); bcopy(ndp->ni_cnd.cn_nameptr, p->p_comm, len); p->p_comm[len] = 0; /* * mark as execed, wakeup the process that vforked (if any) and tell * it that it now has its own resources back */ p->p_flag |= P_EXEC; if (p->p_pptr && (p->p_flag & P_PPWAIT)) { p->p_flag &= ~P_PPWAIT; wakeup((caddr_t)p->p_pptr); } /* * Implement image setuid/setgid. * * Don't honor setuid/setgid if the filesystem prohibits it or if * the process is being traced. */ if ((((attr.va_mode & VSUID) && p->p_ucred->cr_uid != attr.va_uid) || ((attr.va_mode & VSGID) && p->p_ucred->cr_gid != attr.va_gid)) && (imgp->vp->v_mount->mnt_flag & MNT_NOSUID) == 0 && (p->p_flag & P_TRACED) == 0) { /* * Turn off syscall tracing for set-id programs, except for * root. */ if (p->p_tracep && suser(p)) { p->p_traceflag = 0; vrele(p->p_tracep); p->p_tracep = NULL; } /* * Set the new credentials. */ p->p_ucred = crcopy(p->p_ucred); if (attr.va_mode & VSUID) p->p_ucred->cr_uid = attr.va_uid; if (attr.va_mode & VSGID) p->p_ucred->cr_gid = attr.va_gid; setsugid(p); } else { if (p->p_ucred->cr_uid == p->p_cred->p_ruid && p->p_ucred->cr_gid == p->p_cred->p_rgid) p->p_flag &= ~P_SUGID; } /* * Implement correct POSIX saved-id behavior. */ p->p_cred->p_svuid = p->p_ucred->cr_uid; p->p_cred->p_svgid = p->p_ucred->cr_gid; /* * Store the vp for use in procfs */ if (p->p_textvp) /* release old reference */ vrele(p->p_textvp); VREF(ndp->ni_vp); p->p_textvp = ndp->ni_vp; /* * If tracing the process, trap to debugger so breakpoints * can be set before the program executes. */ STOPEVENT(p, S_EXEC, 0); if (p->p_flag & P_TRACED) psignal(p, SIGTRAP); /* clear "fork but no exec" flag, as we _are_ execing */ p->p_acflag &= ~AFORK; /* Set values passed into the program in registers. */ setregs(p, imgp->entry_addr, (u_long)(uintptr_t)stack_base, imgp->ps_strings); /* Free any previous argument cache */ if (p->p_args && --p->p_args->ar_ref == 0) FREE(p->p_args, M_PARGS); p->p_args = NULL; /* Cache arguments if they fit inside our allowance */ i = imgp->endargs - imgp->stringbase; if (ps_arg_cache_limit >= i + sizeof(struct pargs)) { MALLOC(p->p_args, struct pargs *, sizeof(struct pargs) + i, M_PARGS, M_WAITOK); p->p_args->ar_ref = 1; p->p_args->ar_length = i; bcopy(imgp->stringbase, p->p_args->ar_args, i); } exec_fail_dealloc: /* * free various allocated resources */ if (imgp->firstpage) exec_unmap_first_page(imgp); if (imgp->stringbase != NULL) kmem_free_wakeup(exec_map, (vm_offset_t)imgp->stringbase, ARG_MAX + PAGE_SIZE); if (imgp->vp) { NDFREE(ndp, NDF_ONLY_PNBUF); vrele(imgp->vp); } if (error == 0) return (0); exec_fail: if (imgp->vmspace_destroyed) { /* sorry, no more process anymore. exit gracefully */ exit1(p, W_EXITCODE(0, SIGABRT)); /* NOT REACHED */ return(0); } else { return(error); } } int exec_map_first_page(imgp) struct image_params *imgp; { int s, rv, i; int initial_pagein; vm_page_t ma[VM_INITIAL_PAGEIN]; vm_object_t object; if (imgp->firstpage) { exec_unmap_first_page(imgp); } object = imgp->vp->v_object; s = splvm(); ma[0] = vm_page_grab(object, 0, VM_ALLOC_NORMAL | VM_ALLOC_RETRY); if ((ma[0]->valid & VM_PAGE_BITS_ALL) != VM_PAGE_BITS_ALL) { initial_pagein = VM_INITIAL_PAGEIN; if (initial_pagein > object->size) initial_pagein = object->size; for (i = 1; i < initial_pagein; i++) { if ((ma[i] = vm_page_lookup(object, i)) != NULL) { if ((ma[i]->flags & PG_BUSY) || ma[i]->busy) break; if (ma[i]->valid) break; vm_page_busy(ma[i]); } else { ma[i] = vm_page_alloc(object, i, VM_ALLOC_NORMAL); if (ma[i] == NULL) break; } } initial_pagein = i; rv = vm_pager_get_pages(object, ma, initial_pagein, 0); ma[0] = vm_page_lookup(object, 0); if ((rv != VM_PAGER_OK) || (ma[0] == NULL) || (ma[0]->valid == 0)) { if (ma[0]) { vm_page_protect(ma[0], VM_PROT_NONE); vm_page_free(ma[0]); } splx(s); return EIO; } } vm_page_wire(ma[0]); vm_page_wakeup(ma[0]); splx(s); pmap_kenter((vm_offset_t) imgp->image_header, VM_PAGE_TO_PHYS(ma[0])); imgp->firstpage = ma[0]; return 0; } void exec_unmap_first_page(imgp) struct image_params *imgp; { if (imgp->firstpage) { pmap_kremove((vm_offset_t) imgp->image_header); vm_page_unwire(imgp->firstpage, 1); imgp->firstpage = NULL; } } /* * Destroy old address space, and allocate a new stack * The new stack is only SGROWSIZ large because it is grown * automatically in trap.c. */ int exec_new_vmspace(imgp) struct image_params *imgp; { int error; struct vmspace *vmspace = imgp->proc->p_vmspace; caddr_t stack_addr = (caddr_t) (USRSTACK - MAXSSIZ); vm_map_t map = &vmspace->vm_map; imgp->vmspace_destroyed = 1; /* * Blow away entire process VM, if address space not shared, * otherwise, create a new VM space so that other threads are * not disrupted */ if (vmspace->vm_refcnt == 1) { if (vmspace->vm_shm) shmexit(imgp->proc); pmap_remove_pages(vmspace_pmap(vmspace), 0, VM_MAXUSER_ADDRESS); vm_map_remove(map, 0, VM_MAXUSER_ADDRESS); } else { vmspace_exec(imgp->proc); vmspace = imgp->proc->p_vmspace; map = &vmspace->vm_map; } /* Allocate a new stack */ error = vm_map_stack (&vmspace->vm_map, (vm_offset_t)stack_addr, (vm_size_t)MAXSSIZ, VM_PROT_ALL, VM_PROT_ALL, 0); if (error) return (error); /* vm_ssize and vm_maxsaddr are somewhat antiquated concepts in the * VM_STACK case, but they are still used to monitor the size of the * process stack so we can check the stack rlimit. */ vmspace->vm_ssize = SGROWSIZ >> PAGE_SHIFT; vmspace->vm_maxsaddr = (char *)USRSTACK - MAXSSIZ; return(0); } /* * Copy out argument and environment strings from the old process * address space into the temporary string buffer. */ int exec_extract_strings(imgp) struct image_params *imgp; { char **argv, **envv; char *argp, *envp; int error; size_t length; /* * extract arguments first */ argv = imgp->uap->argv; if (argv) { argp = (caddr_t) (intptr_t) fuword(argv); if (argp == (caddr_t) -1) return (EFAULT); if (argp) argv++; if (imgp->argv0) argp = imgp->argv0; if (argp) { do { if (argp == (caddr_t) -1) return (EFAULT); if ((error = copyinstr(argp, imgp->stringp, imgp->stringspace, &length))) { if (error == ENAMETOOLONG) return(E2BIG); return (error); } imgp->stringspace -= length; imgp->stringp += length; imgp->argc++; } while ((argp = (caddr_t) (intptr_t) fuword(argv++))); } } imgp->endargs = imgp->stringp; /* * extract environment strings */ envv = imgp->uap->envv; if (envv) { while ((envp = (caddr_t) (intptr_t) fuword(envv++))) { if (envp == (caddr_t) -1) return (EFAULT); if ((error = copyinstr(envp, imgp->stringp, imgp->stringspace, &length))) { if (error == ENAMETOOLONG) return(E2BIG); return (error); } imgp->stringspace -= length; imgp->stringp += length; imgp->envc++; } } return (0); } /* * Copy strings out to the new process address space, constructing * new arg and env vector tables. Return a pointer to the base * so that it can be used as the initial stack pointer. */ -long * +register_t * exec_copyout_strings(imgp) struct image_params *imgp; { int argc, envc; char **vectp; char *stringp, *destp; - long *stack_base; + register_t *stack_base; struct ps_strings *arginfo; int szsigcode; /* * Calculate string base and vector table pointers. * Also deal with signal trampoline code for this exec type. */ arginfo = (struct ps_strings *)PS_STRINGS; szsigcode = *(imgp->proc->p_sysent->sv_szsigcode); destp = (caddr_t)arginfo - szsigcode - SPARE_USRSPACE - roundup((ARG_MAX - imgp->stringspace), sizeof(char *)); /* * install sigcode */ if (szsigcode) copyout(imgp->proc->p_sysent->sv_sigcode, ((caddr_t)arginfo - szsigcode), szsigcode); /* * If we have a valid auxargs ptr, prepare some room * on the stack. */ if (imgp->auxargs) /* * The '+ 2' is for the null pointers at the end of each of the * arg and env vector sets, and 'AT_COUNT*2' is room for the * ELF Auxargs data. */ vectp = (char **)(destp - (imgp->argc + imgp->envc + 2 + AT_COUNT*2) * sizeof(char*)); else /* * The '+ 2' is for the null pointers at the end of each of the * arg and env vector sets */ vectp = (char **) (destp - (imgp->argc + imgp->envc + 2) * sizeof(char*)); /* * vectp also becomes our initial stack base */ - stack_base = (long *)vectp; + stack_base = (register_t *)vectp; stringp = imgp->stringbase; argc = imgp->argc; envc = imgp->envc; /* * Copy out strings - arguments and environment. */ copyout(stringp, destp, ARG_MAX - imgp->stringspace); /* * Fill in "ps_strings" struct for ps, w, etc. */ suword(&arginfo->ps_argvstr, (long)(intptr_t)vectp); suword(&arginfo->ps_nargvstr, argc); /* * Fill in argument portion of vector table. */ for (; argc > 0; --argc) { suword(vectp++, (long)(intptr_t)destp); while (*stringp++ != 0) destp++; destp++; } /* a null vector table pointer seperates the argp's from the envp's */ suword(vectp++, 0); suword(&arginfo->ps_envstr, (long)(intptr_t)vectp); suword(&arginfo->ps_nenvstr, envc); /* * Fill in environment portion of vector table. */ for (; envc > 0; --envc) { suword(vectp++, (long)(intptr_t)destp); while (*stringp++ != 0) destp++; destp++; } /* end of vector table is a null pointer */ suword(vectp, 0); return (stack_base); } /* * Check permissions of file to execute. * Return 0 for success or error code on failure. */ int exec_check_permissions(imgp) struct image_params *imgp; { struct proc *p = imgp->proc; struct vnode *vp = imgp->vp; struct vattr *attr = imgp->attr; int error; /* Get file attributes */ error = VOP_GETATTR(vp, attr, p->p_ucred, p); if (error) return (error); /* * 1) Check if file execution is disabled for the filesystem that this * file resides on. * 2) Insure that at least one execute bit is on - otherwise root * will always succeed, and we don't want to happen unless the * file really is executable. * 3) Insure that the file is a regular file. */ if ((vp->v_mount->mnt_flag & MNT_NOEXEC) || ((attr->va_mode & 0111) == 0) || (attr->va_type != VREG)) { return (EACCES); } /* * Zero length files can't be exec'd */ if (attr->va_size == 0) return (ENOEXEC); /* * Check for execute permission to file based on current credentials. */ error = VOP_ACCESS(vp, VEXEC, p->p_ucred, p); if (error) return (error); /* * Check number of open-for-writes on the file and deny execution * if there are any. */ if (vp->v_writecount) return (ETXTBSY); /* * Call filesystem specific open routine (which does nothing in the * general case). */ error = VOP_OPEN(vp, FREAD, p->p_ucred, p); if (error) return (error); return (0); } /* * Exec handler registration */ int exec_register(execsw_arg) const struct execsw *execsw_arg; { const struct execsw **es, **xs, **newexecsw; int count = 2; /* New slot and trailing NULL */ if (execsw) for (es = execsw; *es; es++) count++; newexecsw = malloc(count * sizeof(*es), M_TEMP, M_WAITOK); if (newexecsw == NULL) return ENOMEM; xs = newexecsw; if (execsw) for (es = execsw; *es; es++) *xs++ = *es; *xs++ = execsw_arg; *xs = NULL; if (execsw) free(execsw, M_TEMP); execsw = newexecsw; return 0; } int exec_unregister(execsw_arg) const struct execsw *execsw_arg; { const struct execsw **es, **xs, **newexecsw; int count = 1; if (execsw == NULL) panic("unregister with no handlers left?\n"); for (es = execsw; *es; es++) { if (*es == execsw_arg) break; } if (*es == NULL) return ENOENT; for (es = execsw; *es; es++) if (*es != execsw_arg) count++; newexecsw = malloc(count * sizeof(*es), M_TEMP, M_WAITOK); if (newexecsw == NULL) return ENOMEM; xs = newexecsw; for (es = execsw; *es; es++) if (*es != execsw_arg) *xs++ = *es; *xs = NULL; if (execsw) free(execsw, M_TEMP); execsw = newexecsw; return 0; }