Page MenuHomeFreeBSD
Differential D7974 Diff 20704 sys/arm/linux/linux_machdep.c

Changeset View

Standalone View

View Options

sys/arm/linux/linux_machdep.c

  • This file was added.
/*-
* Copyright (c) 2000 Marcel Moolenaar
* 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 without 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.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/capsicum.h>
#include <sys/file.h>
#include <sys/fcntl.h>
#include <sys/imgact.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mman.h>
#include <sys/mutex.h>
#include <sys/sx.h>
#include <sys/priv.h>
#include <sys/proc.h>
#include <sys/queue.h>
#include <sys/resource.h>
#include <sys/resourcevar.h>
#include <sys/signalvar.h>
#include <sys/syscallsubr.h>
#include <sys/sysproto.h>
#include <sys/unistd.h>
#include <sys/wait.h>
#include <sys/sched.h>
#include <sys/clock.h>
#include <machine/frame.h>
#include <machine/psl.h>
#include <machine/sysarch.h>
#include <vm/vm.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <arm/linux/linux.h>
#include <arm/linux/linux_proto.h>
#include <compat/linux/linux_ipc.h>
#include <compat/linux/linux_misc.h>
#include <compat/linux/linux_signal.h>
#include <compat/linux/linux_util.h>
#include <compat/linux/linux_emul.h>
#include <arm/include/pcb.h> /* needed for pcb definition in linux_set_thread_area */
andrewUnsubmitted
Not Done
Inline Actions

This should be <machine/pcb.h>

andrew: This should be `<machine/pcb.h>`
#include "opt_posix.h"
extern struct sysentvec elf32_freebsd_sysvec; /* defined in i386/i386/elf_machdep.c */
struct l_old_select_argv {
l_int nfds;
l_fd_set *readfds;
l_fd_set *writefds;
l_fd_set *exceptfds;
struct l_timeval *timeout;
};
static int linux_mmap_common(struct thread *td, l_uintptr_t addr,
l_size_t len, l_int prot, l_int flags, l_int fd,
l_loff_t pos);
int
linux_execve(struct thread *td, struct linux_execve_args *args)
{
struct image_args eargs;
char *newpath;
int error;
LCONVPATHEXIST(td, args->path, &newpath);
#ifdef DEBUG
if (ldebug(execve))
printf(ARGS(execve, "%s"), newpath);
#endif
error = exec_copyin_args(&eargs, newpath, UIO_SYSSPACE,
args->argp, args->envp);
free(newpath, M_TEMP);
if (error == 0)
error = linux_common_execve(td, &eargs);
return (error);
}
struct l_ipc_kludge {
struct l_msgbuf *msgp;
l_long msgtyp;
};
int
linux_ipc(struct thread *td, struct linux_ipc_args *args)
{
switch (args->what & 0xFFFF) {
case LINUX_SEMOP: {
struct linux_semop_args a;
a.semid = args->arg1;
a.tsops = args->ptr;
a.nsops = args->arg2;
return (linux_semop(td, &a));
}
case LINUX_SEMGET: {
struct linux_semget_args a;
a.key = args->arg1;
a.nsems = args->arg2;
a.semflg = args->arg3;
return (linux_semget(td, &a));
}
case LINUX_SEMCTL: {
struct linux_semctl_args a;
int error;
a.semid = args->arg1;
a.semnum = args->arg2;
a.cmd = args->arg3;
error = copyin(args->ptr, &a.arg, sizeof(a.arg));
if (error)
return (error);
return (linux_semctl(td, &a));
}
case LINUX_MSGSND: {
struct linux_msgsnd_args a;
a.msqid = args->arg1;
a.msgp = args->ptr;
a.msgsz = args->arg2;
a.msgflg = args->arg3;
return (linux_msgsnd(td, &a));
}
case LINUX_MSGRCV: {
struct linux_msgrcv_args a;
a.msqid = args->arg1;
a.msgsz = args->arg2;
a.msgflg = args->arg3;
if ((args->what >> 16) == 0) {
struct l_ipc_kludge tmp;
int error;
if (args->ptr == NULL)
return (EINVAL);
error = copyin(args->ptr, &tmp, sizeof(tmp));
if (error)
return (error);
a.msgp = tmp.msgp;
a.msgtyp = tmp.msgtyp;
} else {
a.msgp = args->ptr;
a.msgtyp = args->arg5;
}
return (linux_msgrcv(td, &a));
}
case LINUX_MSGGET: {
struct linux_msgget_args a;
a.key = args->arg1;
a.msgflg = args->arg2;
return (linux_msgget(td, &a));
}
case LINUX_MSGCTL: {
struct linux_msgctl_args a;
a.msqid = args->arg1;
a.cmd = args->arg2;
a.buf = args->ptr;
return (linux_msgctl(td, &a));
}
case LINUX_SHMAT: {
struct linux_shmat_args a;
a.shmid = args->arg1;
a.shmaddr = args->ptr;
a.shmflg = args->arg2;
return (linux_shmat(td, &a));
}
case LINUX_SHMDT: {
struct linux_shmdt_args a;
a.shmaddr = args->ptr;
return (linux_shmdt(td, &a));
}
case LINUX_SHMGET: {
struct linux_shmget_args a;
a.key = args->arg1;
a.size = args->arg2;
a.shmflg = args->arg3;
return (linux_shmget(td, &a));
}
case LINUX_SHMCTL: {
struct linux_shmctl_args a;
a.shmid = args->arg1;
a.cmd = args->arg2;
a.buf = args->ptr;
return (linux_shmctl(td, &a));
}
default:
break;
}
return (EINVAL);
}
int
linux_old_select(struct thread *td, struct linux_old_select_args *args)
{
struct l_old_select_argv linux_args;
struct linux_select_args newsel;
int error;
#ifdef DEBUG
if (ldebug(old_select))
printf(ARGS(old_select, "%p"), args->ptr);
#endif
error = copyin(args->ptr, &linux_args, sizeof(linux_args));
if (error)
return (error);
newsel.nfds = linux_args.nfds;
newsel.readfds = linux_args.readfds;
newsel.writefds = linux_args.writefds;
newsel.exceptfds = linux_args.exceptfds;
newsel.timeout = linux_args.timeout;
return (linux_select(td, &newsel));
}
struct timeval32 {
uint32_t tv_sec;
uint32_t tv_usec;
};
int
linux_gettimeofday(struct thread *td, struct linux_gettimeofday_args *uap)
{
struct timeval atv;
struct timeval32 atv32;
struct timezone rtz;
int error = 0;
if (uap->tp) {
microtime(&atv);
atv32.tv_sec = atv.tv_sec;
atv32.tv_usec = atv.tv_usec;
error = copyout(&atv32, uap->tp, sizeof(atv32));
}
if (error == 0 && uap->tzp != NULL) {
rtz.tz_minuteswest = tz_minuteswest;
rtz.tz_dsttime = tz_dsttime;
error = copyout(&rtz, uap->tzp, sizeof(rtz));
}
return (error);
}
int
linux_settimeofday(struct thread *td, struct linux_settimeofday_args *uap)
{
struct timeval32 atv32;
struct timeval atv, *tvp;
struct timezone atz, *tzp;
int error;
if (uap->tp) {
error = copyin(uap->tp, &atv32, sizeof(atv32));
if (error)
return (error);
atv.tv_sec = atv32.tv_sec;
atv.tv_usec = atv32.tv_usec;
tvp = &atv;
} else
tvp = NULL;
if (uap->tzp) {
error = copyin(uap->tzp, &atz, sizeof(atz));
if (error)
return (error);
tzp = &atz;
} else
tzp = NULL;
return (kern_settimeofday(td, tvp, tzp));
}
int
linux_set_cloned_tls(struct thread *td, void *desc)
{
struct l_user_desc info;
int idx, error;
error = copyin(desc, &info, sizeof(struct l_user_desc));
if (error) {
printf(LMSG("copyin failed!"));
} else {
idx = info.entry_number;
/*
* looks like we're getting the idx we returned
* in the set_thread_area() syscall
*/
if (idx != 6 && idx != 3) {
printf(LMSG("resetting idx!"));
idx = 3;
}
/* this doesnt happen in practice */
if (idx == 6) {
/* we might copy out the entry_number as 3 */
info.entry_number = 3;
error = copyout(&info, desc, sizeof(struct l_user_desc));
if (error)
printf(LMSG("copyout failed!"));
}
#ifdef DEBUG
if (ldebug(clone))
printf("Segment created in clone\n");
#endif
}
return (error);
}
int
linux_set_upcall_kse(struct thread *td, register_t stack)
{
if (stack)
td->td_frame->tf_r13 = stack;
/*
* The newly created Linux thread returns
* to the user space by the same path that a parent do.
*/
td->td_frame->tf_r0 = 0;
return (0);
}
#define STACK_SIZE (2 * 1024 * 1024)
#define GUARD_SIZE (4 * PAGE_SIZE)
int
linux_mmap2(struct thread *td, struct linux_mmap2_args *args)
{
#ifdef DEBUG
if (ldebug(mmap2))
printf(ARGS(mmap2, "%p, %d, %d, 0x%08x, %d, %d"),
(void *)args->addr, args->len, args->prot,
args->flags, args->fd, args->pgoff);
#endif
return (linux_mmap_common(td, args->addr, args->len, args->prot,
args->flags, args->fd, (uint64_t)(uint32_t)args->pgoff *
PAGE_SIZE));
}
int
linux_mmap(struct thread *td, struct linux_mmap_args *args)
{
int error;
struct l_mmap_argv linux_args;
error = copyin(args->ptr, &linux_args, sizeof(linux_args));
if (error)
return (error);
#ifdef DEBUG
if (ldebug(mmap))
printf(ARGS(mmap, "%p, %d, %d, 0x%08x, %d, %d"),
(void *)linux_args.addr, linux_args.len, linux_args.prot,
linux_args.flags, linux_args.fd, linux_args.pgoff);
#endif
return (linux_mmap_common(td, linux_args.addr, linux_args.len,
linux_args.prot, linux_args.flags, linux_args.fd,
(uint32_t)linux_args.pgoff));
}
static int
linux_mmap_common(struct thread *td, l_uintptr_t addr, l_size_t len, l_int prot,
l_int flags, l_int fd, l_loff_t pos)
{
struct proc *p = td->td_proc;
struct mmap_args /* {
caddr_t addr;
size_t len;
int prot;
int flags;
int fd;
long pad;
off_t pos;
} */ bsd_args;
int error;
struct file *fp;
cap_rights_t rights;
error = 0;
bsd_args.flags = 0;
fp = NULL;
/*
* Linux mmap(2):
* You must specify exactly one of MAP_SHARED and MAP_PRIVATE
*/
if (!((flags & LINUX_MAP_SHARED) ^ (flags & LINUX_MAP_PRIVATE)))
return (EINVAL);
if (flags & LINUX_MAP_SHARED)
bsd_args.flags |= MAP_SHARED;
if (flags & LINUX_MAP_PRIVATE)
bsd_args.flags |= MAP_PRIVATE;
if (flags & LINUX_MAP_FIXED)
bsd_args.flags |= MAP_FIXED;
if (flags & LINUX_MAP_ANON) {
/* Enforce pos to be on page boundary, then ignore. */
if ((pos & PAGE_MASK) != 0)
return (EINVAL);
pos = 0;
bsd_args.flags |= MAP_ANON;
} else
bsd_args.flags |= MAP_NOSYNC;
if (flags & LINUX_MAP_GROWSDOWN)
bsd_args.flags |= MAP_STACK;
/*
* PROT_READ, PROT_WRITE, or PROT_EXEC implies PROT_READ and PROT_EXEC
* on Linux/i386. We do this to ensure maximum compatibility.
* Linux/ia64 does the same in i386 emulation mode.
*/
bsd_args.prot = prot;
if (bsd_args.prot & (PROT_READ | PROT_WRITE | PROT_EXEC))
bsd_args.prot |= PROT_READ | PROT_EXEC;
/* Linux does not check file descriptor when MAP_ANONYMOUS is set. */
bsd_args.fd = (bsd_args.flags & MAP_ANON) ? -1 : fd;
if (bsd_args.fd != -1) {
/*
* Linux follows Solaris mmap(2) description:
* The file descriptor fildes is opened with
* read permission, regardless of the
* protection options specified.
*
* Checking just CAP_MMAP is fine here, since the real work
* is done in the FreeBSD mmap().
*/
error = fget(td, bsd_args.fd,
cap_rights_init(&rights, CAP_MMAP), &fp);
if (error != 0)
return (error);
if (fp->f_type != DTYPE_VNODE) {
fdrop(fp, td);
return (EINVAL);
}
/* Linux mmap() just fails for O_WRONLY files */
if (!(fp->f_flag & FREAD)) {
fdrop(fp, td);
return (EACCES);
}
fdrop(fp, td);
}
if (flags & LINUX_MAP_GROWSDOWN) {
/*
* The Linux MAP_GROWSDOWN option does not limit auto
* growth of the region. Linux mmap with this option
* takes as addr the inital BOS, and as len, the initial
* region size. It can then grow down from addr without
* limit. However, linux threads has an implicit internal
* limit to stack size of STACK_SIZE. Its just not
* enforced explicitly in linux. But, here we impose
* a limit of (STACK_SIZE - GUARD_SIZE) on the stack
* region, since we can do this with our mmap.
*
* Our mmap with MAP_STACK takes addr as the maximum
* downsize limit on BOS, and as len the max size of
* the region. It them maps the top SGROWSIZ bytes,
* and auto grows the region down, up to the limit
* in addr.
*
* If we don't use the MAP_STACK option, the effect
* of this code is to allocate a stack region of a
* fixed size of (STACK_SIZE - GUARD_SIZE).
*/
if ((caddr_t)PTRIN(addr) + len > p->p_vmspace->vm_maxsaddr) {
/*
* Some linux apps will attempt to mmap
* thread stacks near the top of their
* address space. If their TOS is greater
* than vm_maxsaddr, vm_map_growstack()
* will confuse the thread stack with the
* process stack and deliver a SEGV if they
* attempt to grow the thread stack past their
* current stacksize rlimit. To avoid this,
* adjust vm_maxsaddr upwards to reflect
* the current stacksize rlimit rather
* than the maximum possible stacksize.
* It would be better to adjust the
* mmap'ed region, but some apps do not check
* mmap's return value.
*/
PROC_LOCK(p);
p->p_vmspace->vm_maxsaddr = (char *)USRSTACK -
lim_cur_proc(p, RLIMIT_STACK);
PROC_UNLOCK(p);
}
/*
* This gives us our maximum stack size and a new BOS.
* If we're using VM_STACK, then mmap will just map
* the top SGROWSIZ bytes, and let the stack grow down
* to the limit at BOS. If we're not using VM_STACK
* we map the full stack, since we don't have a way
* to autogrow it.
*/
if (len > STACK_SIZE - GUARD_SIZE) {
bsd_args.addr = (caddr_t)PTRIN(addr);
bsd_args.len = len;
} else {
bsd_args.addr = (caddr_t)PTRIN(addr) -
(STACK_SIZE - GUARD_SIZE - len);
bsd_args.len = STACK_SIZE - GUARD_SIZE;
}
} else {
bsd_args.addr = (caddr_t)PTRIN(addr);
bsd_args.len = len;
}
bsd_args.pos = pos;
#ifdef DEBUG
if (ldebug(mmap))
printf("-> %s(%p, %d, %d, 0x%08x, %d, 0x%x)\n",
__func__,
(void *)bsd_args.addr, bsd_args.len, bsd_args.prot,
bsd_args.flags, bsd_args.fd, (int)bsd_args.pos);
#endif
error = sys_mmap(td, &bsd_args);
#ifdef DEBUG
if (ldebug(mmap))
printf("-> %s() return: 0x%x (0x%08x)\n",
__func__, error, (u_int)td->td_retval[0]);
#endif
return (error);
}
int
linux_mprotect(struct thread *td, struct linux_mprotect_args *uap)
{
struct mprotect_args bsd_args;
bsd_args.addr = uap->addr;
bsd_args.len = uap->len;
bsd_args.prot = uap->prot;
if (bsd_args.prot & (PROT_READ | PROT_WRITE | PROT_EXEC))
bsd_args.prot |= PROT_READ | PROT_EXEC;
return (sys_mprotect(td, &bsd_args));
}
int
linux_sigaction(struct thread *td, struct linux_sigaction_args *args)
{
l_osigaction_t osa;
l_sigaction_t act, oact;
int error;
#ifdef DEBUG
if (ldebug(sigaction))
printf(ARGS(sigaction, "%d, %p, %p"),
args->sig, (void *)args->nsa, (void *)args->osa);
#endif
if (args->nsa != NULL) {
error = copyin(args->nsa, &osa, sizeof(l_osigaction_t));
if (error)
return (error);
act.lsa_handler = osa.lsa_handler;
act.lsa_flags = osa.lsa_flags;
act.lsa_restorer = osa.lsa_restorer;
LINUX_SIGEMPTYSET(act.lsa_mask);
act.lsa_mask.__mask = osa.lsa_mask;
}
error = linux_do_sigaction(td, args->sig, args->nsa ? &act : NULL,
args->osa ? &oact : NULL);
if (args->osa != NULL && !error) {
osa.lsa_handler = oact.lsa_handler;
osa.lsa_flags = oact.lsa_flags;
osa.lsa_restorer = oact.lsa_restorer;
osa.lsa_mask = oact.lsa_mask.__mask;
error = copyout(&osa, args->osa, sizeof(l_osigaction_t));
}
return (error);
}
/*
* Linux has two extra args, restart and oldmask. We dont use these,
* but it seems that "restart" is actually a context pointer that
* enables the signal to happen with a different register set.
*/
int
linux_sigsuspend(struct thread *td, struct linux_sigsuspend_args *args)
{
sigset_t sigmask;
l_sigset_t mask;
#ifdef DEBUG
if (ldebug(sigsuspend))
printf(ARGS(sigsuspend, "%08lx"), (unsigned long)args->mask);
#endif
LINUX_SIGEMPTYSET(mask);
mask.__mask = args->mask;
linux_to_bsd_sigset(&mask, &sigmask);
return (kern_sigsuspend(td, sigmask));
}
int
linux_rt_sigsuspend(struct thread *td, struct linux_rt_sigsuspend_args *uap)
{
l_sigset_t lmask;
sigset_t sigmask;
int error;
#ifdef DEBUG
if (ldebug(rt_sigsuspend))
printf(ARGS(rt_sigsuspend, "%p, %d"),
(void *)uap->newset, uap->sigsetsize);
#endif
if (uap->sigsetsize != sizeof(l_sigset_t))
return (EINVAL);
error = copyin(uap->newset, &lmask, sizeof(l_sigset_t));
if (error)
return (error);
linux_to_bsd_sigset(&lmask, &sigmask);
return (kern_sigsuspend(td, sigmask));
}
int
linux_pause(struct thread *td, struct linux_pause_args *args)
{
struct proc *p = td->td_proc;
sigset_t sigmask;
#ifdef DEBUG
if (ldebug(pause))
printf(ARGS(pause, ""));
#endif
PROC_LOCK(p);
sigmask = td->td_sigmask;
PROC_UNLOCK(p);
return (kern_sigsuspend(td, sigmask));
}
int
linux_sigaltstack(struct thread *td, struct linux_sigaltstack_args *uap)
{
stack_t ss, oss;
l_stack_t lss;
int error;
#ifdef DEBUG
if (ldebug(sigaltstack))
printf(ARGS(sigaltstack, "%p, %p"), uap->uss, uap->uoss);
#endif
if (uap->uss != NULL) {
error = copyin(uap->uss, &lss, sizeof(l_stack_t));
if (error)
return (error);
ss.ss_sp = (char*)lss.ss_sp;
ss.ss_size = (__size_t)lss.ss_size;
ss.ss_flags = linux_to_bsd_sigaltstack(lss.ss_flags);
}
error = kern_sigaltstack(td, (uap->uss != NULL) ? &ss : NULL,
(uap->uoss != NULL) ? &oss : NULL);
if (!error && uap->uoss != NULL) {
lss.ss_sp = oss.ss_sp;
lss.ss_size = (l_size_t )oss.ss_size;
lss.ss_flags = bsd_to_linux_sigaltstack(oss.ss_flags);
error = copyout(&lss, uap->uoss, sizeof(l_stack_t));
}
return (error);
}
int
linux_ftruncate64(struct thread *td, struct linux_ftruncate64_args *args)
{
struct ftruncate_args sa;
#ifdef DEBUG
if (ldebug(ftruncate64))
printf(ARGS(ftruncate64, "%u, %jd"), args->fd,
(intmax_t)args->length);
#endif
sa.fd = args->fd;
sa.length = args->length;
return sys_ftruncate(td, &sa);
}
/* XXX: this wont work with module - convert it */
int
linux_mq_open(struct thread *td, struct linux_mq_open_args *args)
{
#ifdef P1003_1B_MQUEUE
return sys_kmq_open(td, (struct kmq_open_args *) args);
#else
return (ENOSYS);
#endif
}
int
linux_mq_unlink(struct thread *td, struct linux_mq_unlink_args *args)
{
#ifdef P1003_1B_MQUEUE
return sys_kmq_unlink(td, (struct kmq_unlink_args *) args);
#else
return (ENOSYS);
#endif
}
int
linux_mq_timedsend(struct thread *td, struct linux_mq_timedsend_args *args)
{
#ifdef P1003_1B_MQUEUE
return sys_kmq_timedsend(td, (struct kmq_timedsend_args *) args);
#else
return (ENOSYS);
#endif
}
int
linux_mq_timedreceive(struct thread *td, struct linux_mq_timedreceive_args *args)
{
#ifdef P1003_1B_MQUEUE
return sys_kmq_timedreceive(td, (struct kmq_timedreceive_args *) args);
#else
return (ENOSYS);
#endif
}
int
linux_mq_notify(struct thread *td, struct linux_mq_notify_args *args)
{
#ifdef P1003_1B_MQUEUE
return sys_kmq_notify(td, (struct kmq_notify_args *) args);
#else
return (ENOSYS);
#endif
}
int
linux_mq_getsetattr(struct thread *td, struct linux_mq_getsetattr_args *args)
{
#ifdef P1003_1B_MQUEUE
return sys_kmq_setattr(td, (struct kmq_setattr_args *) args);
#else
return (ENOSYS);
#endif
}
int
linux_set_tls(struct thread *td, struct linux_set_tls_args *args)
{
#if __ARM_ARCH >= 6
td->td_pcb->pcb_regs.sf_tpidrurw = (register_t)args->tls;
#else
td->td_md.md_tp = (register_t)args->tls;
#endif
// Do it ??
set_tls(args->tls);
return (0);
}
uint32_t
linux_get_tls(struct thread *td)
{
#if __ARM_ARCH >= 6
return (td->td_pcb->pcb_regs.sf_tpidrurw);
#else
return (td->td_md.md_tp);
#endif
}