1 /* $NetBSD: linux_machdep.c,v 1.77 2002/05/20 06:22:43 jdolecek Exp $ */ 2 3 /*- 4 * Copyright (c) 1995, 2000 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Frank van der Linden. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 3. All advertising materials mentioning features or use of this software 19 * must display the following acknowledgement: 20 * This product includes software developed by the NetBSD 21 * Foundation, Inc. and its contributors. 22 * 4. Neither the name of The NetBSD Foundation nor the names of its 23 * contributors may be used to endorse or promote products derived 24 * from this software without specific prior written permission. 25 * 26 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 27 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 28 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 29 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 30 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 31 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 32 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 33 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 34 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 35 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 36 * POSSIBILITY OF SUCH DAMAGE. 37 */ 38 39 #include <sys/cdefs.h> 40 __KERNEL_RCSID(0, "$NetBSD: linux_machdep.c,v 1.77 2002/05/20 06:22:43 jdolecek Exp $"); 41 42 #if defined(_KERNEL_OPT) 43 #include "opt_vm86.h" 44 #include "opt_user_ldt.h" 45 #endif 46 47 #include <sys/param.h> 48 #include <sys/systm.h> 49 #include <sys/signalvar.h> 50 #include <sys/kernel.h> 51 #include <sys/map.h> 52 #include <sys/proc.h> 53 #include <sys/user.h> 54 #include <sys/buf.h> 55 #include <sys/reboot.h> 56 #include <sys/conf.h> 57 #include <sys/exec.h> 58 #include <sys/file.h> 59 #include <sys/callout.h> 60 #include <sys/malloc.h> 61 #include <sys/mbuf.h> 62 #include <sys/msgbuf.h> 63 #include <sys/mount.h> 64 #include <sys/vnode.h> 65 #include <sys/device.h> 66 #include <sys/syscallargs.h> 67 #include <sys/filedesc.h> 68 #include <sys/exec_elf.h> 69 #include <sys/disklabel.h> 70 #include <sys/ioctl.h> 71 #include <miscfs/specfs/specdev.h> 72 73 #include <compat/linux/common/linux_types.h> 74 #include <compat/linux/common/linux_signal.h> 75 #include <compat/linux/common/linux_util.h> 76 #include <compat/linux/common/linux_ioctl.h> 77 #include <compat/linux/common/linux_hdio.h> 78 #include <compat/linux/common/linux_exec.h> 79 #include <compat/linux/common/linux_machdep.h> 80 81 #include <compat/linux/linux_syscallargs.h> 82 83 #include <machine/cpu.h> 84 #include <machine/cpufunc.h> 85 #include <machine/psl.h> 86 #include <machine/reg.h> 87 #include <machine/segments.h> 88 #include <machine/specialreg.h> 89 #include <machine/sysarch.h> 90 #include <machine/vm86.h> 91 #include <machine/vmparam.h> 92 93 /* 94 * To see whether wscons is configured (for virtual console ioctl calls). 95 */ 96 #if defined(_KERNEL_OPT) 97 #include "wsdisplay.h" 98 #endif 99 #if (NWSDISPLAY > 0) 100 #include <dev/wscons/wsconsio.h> 101 #include <dev/wscons/wsdisplay_usl_io.h> 102 #if defined(_KERNEL_OPT) 103 #include "opt_xserver.h" 104 #endif 105 #endif 106 107 #ifdef USER_LDT 108 #include <machine/cpu.h> 109 int linux_read_ldt __P((struct proc *, struct linux_sys_modify_ldt_args *, 110 register_t *)); 111 int linux_write_ldt __P((struct proc *, struct linux_sys_modify_ldt_args *, 112 register_t *)); 113 #endif 114 115 #ifdef DEBUG_LINUX 116 #define DPRINTF(a) uprintf a 117 #else 118 #define DPRINTF(a) 119 #endif 120 121 static struct biosdisk_info *fd2biosinfo __P((struct proc *, struct file *)); 122 extern struct disklist *i386_alldisks; 123 extern const char *findblkname __P((int)); 124 125 /* 126 * Deal with some i386-specific things in the Linux emulation code. 127 */ 128 129 void 130 linux_setregs(p, epp, stack) 131 struct proc *p; 132 struct exec_package *epp; 133 u_long stack; 134 { 135 struct pcb *pcb = &p->p_addr->u_pcb; 136 struct trapframe *tf; 137 138 #if NNPX > 0 139 /* If we were using the FPU, forget about it. */ 140 if (npxproc == p) 141 npxdrop(); 142 #endif 143 144 #ifdef USER_LDT 145 pmap_ldt_cleanup(p); 146 #endif 147 148 p->p_md.md_flags &= ~MDP_USEDFPU; 149 pcb->pcb_flags = 0; 150 151 if (i386_use_fxsave) { 152 pcb->pcb_savefpu.sv_xmm.sv_env.en_cw = __Linux_NPXCW__; 153 pcb->pcb_savefpu.sv_xmm.sv_env.en_mxcsr = __INITIAL_MXCSR__; 154 } else 155 pcb->pcb_savefpu.sv_87.sv_env.en_cw = __Linux_NPXCW__; 156 157 tf = p->p_md.md_regs; 158 tf->tf_gs = GSEL(GUDATA_SEL, SEL_UPL); 159 tf->tf_fs = GSEL(GUDATA_SEL, SEL_UPL); 160 tf->tf_es = GSEL(GUDATA_SEL, SEL_UPL); 161 tf->tf_ds = GSEL(GUDATA_SEL, SEL_UPL); 162 tf->tf_edi = 0; 163 tf->tf_esi = 0; 164 tf->tf_ebp = 0; 165 tf->tf_ebx = (int)p->p_psstr; 166 tf->tf_edx = 0; 167 tf->tf_ecx = 0; 168 tf->tf_eax = 0; 169 tf->tf_eip = epp->ep_entry; 170 tf->tf_cs = GSEL(GUCODE_SEL, SEL_UPL); 171 tf->tf_eflags = PSL_USERSET; 172 tf->tf_esp = stack; 173 tf->tf_ss = GSEL(GUDATA_SEL, SEL_UPL); 174 } 175 176 /* 177 * Send an interrupt to process. 178 * 179 * Stack is set up to allow sigcode stored 180 * in u. to call routine, followed by kcall 181 * to sigreturn routine below. After sigreturn 182 * resets the signal mask, the stack, and the 183 * frame pointer, it returns to the user 184 * specified pc, psl. 185 */ 186 187 void 188 linux_sendsig(catcher, sig, mask, code) 189 sig_t catcher; 190 int sig; 191 sigset_t *mask; 192 u_long code; 193 { 194 struct proc *p = curproc; 195 struct trapframe *tf; 196 struct linux_sigframe *fp, frame; 197 int onstack; 198 199 tf = p->p_md.md_regs; 200 201 /* Do we need to jump onto the signal stack? */ 202 onstack = 203 (p->p_sigctx.ps_sigstk.ss_flags & (SS_DISABLE | SS_ONSTACK)) == 0 && 204 (SIGACTION(p, sig).sa_flags & SA_ONSTACK) != 0; 205 206 /* Allocate space for the signal handler context. */ 207 if (onstack) 208 fp = (struct linux_sigframe *)((caddr_t)p->p_sigctx.ps_sigstk.ss_sp + 209 p->p_sigctx.ps_sigstk.ss_size); 210 else 211 fp = (struct linux_sigframe *)tf->tf_esp; 212 fp--; 213 214 /* Build stack frame for signal trampoline. */ 215 frame.sf_handler = catcher; 216 frame.sf_sig = native_to_linux_signo[sig]; 217 218 /* Save register context. */ 219 #ifdef VM86 220 if (tf->tf_eflags & PSL_VM) { 221 frame.sf_sc.sc_gs = tf->tf_vm86_gs; 222 frame.sf_sc.sc_fs = tf->tf_vm86_fs; 223 frame.sf_sc.sc_es = tf->tf_vm86_es; 224 frame.sf_sc.sc_ds = tf->tf_vm86_ds; 225 frame.sf_sc.sc_eflags = get_vflags(p); 226 } else 227 #endif 228 { 229 frame.sf_sc.sc_gs = tf->tf_gs; 230 frame.sf_sc.sc_fs = tf->tf_fs; 231 frame.sf_sc.sc_es = tf->tf_es; 232 frame.sf_sc.sc_ds = tf->tf_ds; 233 frame.sf_sc.sc_eflags = tf->tf_eflags; 234 } 235 frame.sf_sc.sc_edi = tf->tf_edi; 236 frame.sf_sc.sc_esi = tf->tf_esi; 237 frame.sf_sc.sc_ebp = tf->tf_ebp; 238 frame.sf_sc.sc_ebx = tf->tf_ebx; 239 frame.sf_sc.sc_edx = tf->tf_edx; 240 frame.sf_sc.sc_ecx = tf->tf_ecx; 241 frame.sf_sc.sc_eax = tf->tf_eax; 242 frame.sf_sc.sc_eip = tf->tf_eip; 243 frame.sf_sc.sc_cs = tf->tf_cs; 244 frame.sf_sc.sc_esp_at_signal = tf->tf_esp; 245 frame.sf_sc.sc_ss = tf->tf_ss; 246 frame.sf_sc.sc_err = tf->tf_err; 247 frame.sf_sc.sc_trapno = tf->tf_trapno; 248 frame.sf_sc.sc_cr2 = p->p_addr->u_pcb.pcb_cr2; 249 250 /* Save signal stack. */ 251 /* Linux doesn't save the onstack flag in sigframe */ 252 253 /* Save signal mask. */ 254 native_to_linux_old_sigset(&frame.sf_sc.sc_mask, mask); 255 256 if (copyout(&frame, fp, sizeof(frame)) != 0) { 257 /* 258 * Process has trashed its stack; give it an illegal 259 * instruction to halt it in its tracks. 260 */ 261 sigexit(p, SIGILL); 262 /* NOTREACHED */ 263 } 264 265 /* 266 * Build context to run handler in. 267 */ 268 tf->tf_gs = GSEL(GUDATA_SEL, SEL_UPL); 269 tf->tf_fs = GSEL(GUDATA_SEL, SEL_UPL); 270 tf->tf_es = GSEL(GUDATA_SEL, SEL_UPL); 271 tf->tf_ds = GSEL(GUDATA_SEL, SEL_UPL); 272 tf->tf_eip = (int)p->p_sigctx.ps_sigcode; 273 tf->tf_cs = GSEL(GUCODE_SEL, SEL_UPL); 274 tf->tf_eflags &= ~(PSL_T|PSL_VM|PSL_AC); 275 tf->tf_esp = (int)fp; 276 tf->tf_ss = GSEL(GUDATA_SEL, SEL_UPL); 277 278 /* Remember that we're now on the signal stack. */ 279 if (onstack) 280 p->p_sigctx.ps_sigstk.ss_flags |= SS_ONSTACK; 281 } 282 283 /* 284 * System call to cleanup state after a signal 285 * has been taken. Reset signal mask and 286 * stack state from context left by sendsig (above). 287 * Return to previous pc and psl as specified by 288 * context left by sendsig. Check carefully to 289 * make sure that the user has not modified the 290 * psl to gain improper privileges or to cause 291 * a machine fault. 292 */ 293 int 294 linux_sys_rt_sigreturn(p, v, retval) 295 struct proc *p; 296 void *v; 297 register_t *retval; 298 { 299 /* XXX XAX write me */ 300 return(ENOSYS); 301 } 302 303 int 304 linux_sys_sigreturn(p, v, retval) 305 struct proc *p; 306 void *v; 307 register_t *retval; 308 { 309 struct linux_sys_sigreturn_args /* { 310 syscallarg(struct linux_sigcontext *) scp; 311 } */ *uap = v; 312 struct linux_sigcontext *scp, context; 313 struct trapframe *tf; 314 sigset_t mask; 315 ssize_t ss_gap; 316 317 /* 318 * The trampoline code hands us the context. 319 * It is unsafe to keep track of it ourselves, in the event that a 320 * program jumps out of a signal handler. 321 */ 322 scp = SCARG(uap, scp); 323 if (copyin((caddr_t)scp, &context, sizeof(*scp)) != 0) 324 return (EFAULT); 325 326 /* Restore register context. */ 327 tf = p->p_md.md_regs; 328 #ifdef VM86 329 if (context.sc_eflags & PSL_VM) { 330 tf->tf_vm86_gs = context.sc_gs; 331 tf->tf_vm86_fs = context.sc_fs; 332 tf->tf_vm86_es = context.sc_es; 333 tf->tf_vm86_ds = context.sc_ds; 334 set_vflags(p, context.sc_eflags); 335 } else 336 #endif 337 { 338 /* 339 * Check for security violations. If we're returning to 340 * protected mode, the CPU will validate the segment registers 341 * automatically and generate a trap on violations. We handle 342 * the trap, rather than doing all of the checking here. 343 */ 344 if (((context.sc_eflags ^ tf->tf_eflags) & PSL_USERSTATIC) != 0 || 345 !USERMODE(context.sc_cs, context.sc_eflags)) 346 return (EINVAL); 347 348 tf->tf_gs = context.sc_gs; 349 tf->tf_fs = context.sc_fs; 350 tf->tf_es = context.sc_es; 351 tf->tf_ds = context.sc_ds; 352 tf->tf_eflags = context.sc_eflags; 353 } 354 tf->tf_edi = context.sc_edi; 355 tf->tf_esi = context.sc_esi; 356 tf->tf_ebp = context.sc_ebp; 357 tf->tf_ebx = context.sc_ebx; 358 tf->tf_edx = context.sc_edx; 359 tf->tf_ecx = context.sc_ecx; 360 tf->tf_eax = context.sc_eax; 361 tf->tf_eip = context.sc_eip; 362 tf->tf_cs = context.sc_cs; 363 tf->tf_esp = context.sc_esp_at_signal; 364 tf->tf_ss = context.sc_ss; 365 366 /* Restore signal stack. */ 367 /* 368 * Linux really does it this way; it doesn't have space in sigframe 369 * to save the onstack flag. 370 */ 371 ss_gap = (ssize_t) 372 ((caddr_t) context.sc_esp_at_signal - (caddr_t) p->p_sigctx.ps_sigstk.ss_sp); 373 if (ss_gap >= 0 && ss_gap < p->p_sigctx.ps_sigstk.ss_size) 374 p->p_sigctx.ps_sigstk.ss_flags |= SS_ONSTACK; 375 else 376 p->p_sigctx.ps_sigstk.ss_flags &= ~SS_ONSTACK; 377 378 /* Restore signal mask. */ 379 linux_old_to_native_sigset(&mask, &context.sc_mask); 380 (void) sigprocmask1(p, SIG_SETMASK, &mask, 0); 381 382 return (EJUSTRETURN); 383 } 384 385 #ifdef USER_LDT 386 387 int 388 linux_read_ldt(p, uap, retval) 389 struct proc *p; 390 struct linux_sys_modify_ldt_args /* { 391 syscallarg(int) func; 392 syscallarg(void *) ptr; 393 syscallarg(size_t) bytecount; 394 } */ *uap; 395 register_t *retval; 396 { 397 struct i386_get_ldt_args gl; 398 int error; 399 caddr_t sg; 400 char *parms; 401 402 DPRINTF(("linux_read_ldt!")); 403 sg = stackgap_init(p, 0); 404 405 gl.start = 0; 406 gl.desc = SCARG(uap, ptr); 407 gl.num = SCARG(uap, bytecount) / sizeof(union descriptor); 408 409 parms = stackgap_alloc(p, &sg, sizeof(gl)); 410 411 if ((error = copyout(&gl, parms, sizeof(gl))) != 0) 412 return (error); 413 414 if ((error = i386_get_ldt(p, parms, retval)) != 0) 415 return (error); 416 417 *retval *= sizeof(union descriptor); 418 return (0); 419 } 420 421 struct linux_ldt_info { 422 u_int entry_number; 423 u_long base_addr; 424 u_int limit; 425 u_int seg_32bit:1; 426 u_int contents:2; 427 u_int read_exec_only:1; 428 u_int limit_in_pages:1; 429 u_int seg_not_present:1; 430 u_int useable:1; 431 }; 432 433 int 434 linux_write_ldt(p, uap, retval) 435 struct proc *p; 436 struct linux_sys_modify_ldt_args /* { 437 syscallarg(int) func; 438 syscallarg(void *) ptr; 439 syscallarg(size_t) bytecount; 440 } */ *uap; 441 register_t *retval; 442 { 443 struct linux_ldt_info ldt_info; 444 struct segment_descriptor sd; 445 struct i386_set_ldt_args sl; 446 int error; 447 caddr_t sg; 448 char *parms; 449 int oldmode = (int)retval[0]; 450 451 DPRINTF(("linux_write_ldt %d\n", oldmode)); 452 if (SCARG(uap, bytecount) != sizeof(ldt_info)) 453 return (EINVAL); 454 if ((error = copyin(SCARG(uap, ptr), &ldt_info, sizeof(ldt_info))) != 0) 455 return error; 456 if (ldt_info.entry_number >= 8192) 457 return (EINVAL); 458 if (ldt_info.contents == 3) { 459 if (oldmode) 460 return (EINVAL); 461 if (ldt_info.seg_not_present) 462 return (EINVAL); 463 } 464 465 if (ldt_info.base_addr == 0 && ldt_info.limit == 0 && 466 (oldmode || (ldt_info.contents == 0 && 467 ldt_info.read_exec_only == 1 && ldt_info.seg_32bit == 0 && 468 ldt_info.limit_in_pages == 0 && ldt_info.seg_not_present == 1 && 469 ldt_info.useable == 0))) { 470 /* this means you should zero the ldt */ 471 (void)memset(&sd, 0, sizeof(sd)); 472 } else { 473 sd.sd_lobase = ldt_info.base_addr & 0xffffff; 474 sd.sd_hibase = (ldt_info.base_addr >> 24) & 0xff; 475 sd.sd_lolimit = ldt_info.limit & 0xffff; 476 sd.sd_hilimit = (ldt_info.limit >> 16) & 0xf; 477 sd.sd_type = 16 | (ldt_info.contents << 2) | 478 (!ldt_info.read_exec_only << 1); 479 sd.sd_dpl = SEL_UPL; 480 sd.sd_p = !ldt_info.seg_not_present; 481 sd.sd_def32 = ldt_info.seg_32bit; 482 sd.sd_gran = ldt_info.limit_in_pages; 483 if (!oldmode) 484 sd.sd_xx = ldt_info.useable; 485 else 486 sd.sd_xx = 0; 487 } 488 sg = stackgap_init(p, 0); 489 sl.start = ldt_info.entry_number; 490 sl.desc = stackgap_alloc(p, &sg, sizeof(sd)); 491 sl.num = 1; 492 493 DPRINTF(("linux_write_ldt: idx=%d, base=0x%lx, limit=0x%x\n", 494 ldt_info.entry_number, ldt_info.base_addr, ldt_info.limit)); 495 496 parms = stackgap_alloc(p, &sg, sizeof(sl)); 497 498 if ((error = copyout(&sd, sl.desc, sizeof(sd))) != 0) 499 return (error); 500 if ((error = copyout(&sl, parms, sizeof(sl))) != 0) 501 return (error); 502 503 if ((error = i386_set_ldt(p, parms, retval)) != 0) 504 return (error); 505 506 *retval = 0; 507 return (0); 508 } 509 510 #endif /* USER_LDT */ 511 512 int 513 linux_sys_modify_ldt(p, v, retval) 514 struct proc *p; 515 void *v; 516 register_t *retval; 517 { 518 struct linux_sys_modify_ldt_args /* { 519 syscallarg(int) func; 520 syscallarg(void *) ptr; 521 syscallarg(size_t) bytecount; 522 } */ *uap = v; 523 524 switch (SCARG(uap, func)) { 525 #ifdef USER_LDT 526 case 0: 527 return linux_read_ldt(p, uap, retval); 528 case 1: 529 retval[0] = 1; 530 return linux_write_ldt(p, uap, retval); 531 case 2: 532 #ifdef notyet 533 return (linux_read_default_ldt(p, uap, retval); 534 #else 535 return (ENOSYS); 536 #endif 537 case 0x11: 538 retval[0] = 0; 539 return linux_write_ldt(p, uap, retval); 540 #endif /* USER_LDT */ 541 542 default: 543 return (ENOSYS); 544 } 545 } 546 547 /* 548 * XXX Pathetic hack to make svgalib work. This will fake the major 549 * device number of an opened VT so that svgalib likes it. grmbl. 550 * Should probably do it 'wrong the right way' and use a mapping 551 * array for all major device numbers, and map linux_mknod too. 552 */ 553 dev_t 554 linux_fakedev(dev, raw) 555 dev_t dev; 556 int raw; 557 { 558 if (raw) { 559 #if (NWSDISPLAY > 0) 560 if (major(dev) == NETBSD_WSCONS_MAJOR) 561 return makedev(LINUX_CONS_MAJOR, (minor(dev) + 1)); 562 #endif 563 } 564 565 return dev; 566 } 567 568 #if (NWSDISPLAY > 0) 569 /* 570 * That's not complete, but enough to get an X server running. 571 */ 572 #define NR_KEYS 128 573 static const u_short plain_map[NR_KEYS] = { 574 0x0200, 0x001b, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, 575 0x0037, 0x0038, 0x0039, 0x0030, 0x002d, 0x003d, 0x007f, 0x0009, 576 0x0b71, 0x0b77, 0x0b65, 0x0b72, 0x0b74, 0x0b79, 0x0b75, 0x0b69, 577 0x0b6f, 0x0b70, 0x005b, 0x005d, 0x0201, 0x0702, 0x0b61, 0x0b73, 578 0x0b64, 0x0b66, 0x0b67, 0x0b68, 0x0b6a, 0x0b6b, 0x0b6c, 0x003b, 579 0x0027, 0x0060, 0x0700, 0x005c, 0x0b7a, 0x0b78, 0x0b63, 0x0b76, 580 0x0b62, 0x0b6e, 0x0b6d, 0x002c, 0x002e, 0x002f, 0x0700, 0x030c, 581 0x0703, 0x0020, 0x0207, 0x0100, 0x0101, 0x0102, 0x0103, 0x0104, 582 0x0105, 0x0106, 0x0107, 0x0108, 0x0109, 0x0208, 0x0209, 0x0307, 583 0x0308, 0x0309, 0x030b, 0x0304, 0x0305, 0x0306, 0x030a, 0x0301, 584 0x0302, 0x0303, 0x0300, 0x0310, 0x0206, 0x0200, 0x003c, 0x010a, 585 0x010b, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 586 0x030e, 0x0702, 0x030d, 0x001c, 0x0701, 0x0205, 0x0114, 0x0603, 587 0x0118, 0x0601, 0x0602, 0x0117, 0x0600, 0x0119, 0x0115, 0x0116, 588 0x011a, 0x010c, 0x010d, 0x011b, 0x011c, 0x0110, 0x0311, 0x011d, 589 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 590 }, shift_map[NR_KEYS] = { 591 0x0200, 0x001b, 0x0021, 0x0040, 0x0023, 0x0024, 0x0025, 0x005e, 592 0x0026, 0x002a, 0x0028, 0x0029, 0x005f, 0x002b, 0x007f, 0x0009, 593 0x0b51, 0x0b57, 0x0b45, 0x0b52, 0x0b54, 0x0b59, 0x0b55, 0x0b49, 594 0x0b4f, 0x0b50, 0x007b, 0x007d, 0x0201, 0x0702, 0x0b41, 0x0b53, 595 0x0b44, 0x0b46, 0x0b47, 0x0b48, 0x0b4a, 0x0b4b, 0x0b4c, 0x003a, 596 0x0022, 0x007e, 0x0700, 0x007c, 0x0b5a, 0x0b58, 0x0b43, 0x0b56, 597 0x0b42, 0x0b4e, 0x0b4d, 0x003c, 0x003e, 0x003f, 0x0700, 0x030c, 598 0x0703, 0x0020, 0x0207, 0x010a, 0x010b, 0x010c, 0x010d, 0x010e, 599 0x010f, 0x0110, 0x0111, 0x0112, 0x0113, 0x0213, 0x0203, 0x0307, 600 0x0308, 0x0309, 0x030b, 0x0304, 0x0305, 0x0306, 0x030a, 0x0301, 601 0x0302, 0x0303, 0x0300, 0x0310, 0x0206, 0x0200, 0x003e, 0x010a, 602 0x010b, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 603 0x030e, 0x0702, 0x030d, 0x0200, 0x0701, 0x0205, 0x0114, 0x0603, 604 0x020b, 0x0601, 0x0602, 0x0117, 0x0600, 0x020a, 0x0115, 0x0116, 605 0x011a, 0x010c, 0x010d, 0x011b, 0x011c, 0x0110, 0x0311, 0x011d, 606 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 607 }, altgr_map[NR_KEYS] = { 608 0x0200, 0x0200, 0x0200, 0x0040, 0x0200, 0x0024, 0x0200, 0x0200, 609 0x007b, 0x005b, 0x005d, 0x007d, 0x005c, 0x0200, 0x0200, 0x0200, 610 0x0b71, 0x0b77, 0x0918, 0x0b72, 0x0b74, 0x0b79, 0x0b75, 0x0b69, 611 0x0b6f, 0x0b70, 0x0200, 0x007e, 0x0201, 0x0702, 0x0914, 0x0b73, 612 0x0917, 0x0919, 0x0b67, 0x0b68, 0x0b6a, 0x0b6b, 0x0b6c, 0x0200, 613 0x0200, 0x0200, 0x0700, 0x0200, 0x0b7a, 0x0b78, 0x0916, 0x0b76, 614 0x0915, 0x0b6e, 0x0b6d, 0x0200, 0x0200, 0x0200, 0x0700, 0x030c, 615 0x0703, 0x0200, 0x0207, 0x050c, 0x050d, 0x050e, 0x050f, 0x0510, 616 0x0511, 0x0512, 0x0513, 0x0514, 0x0515, 0x0208, 0x0202, 0x0911, 617 0x0912, 0x0913, 0x030b, 0x090e, 0x090f, 0x0910, 0x030a, 0x090b, 618 0x090c, 0x090d, 0x090a, 0x0310, 0x0206, 0x0200, 0x007c, 0x0516, 619 0x0517, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 620 0x030e, 0x0702, 0x030d, 0x0200, 0x0701, 0x0205, 0x0114, 0x0603, 621 0x0118, 0x0601, 0x0602, 0x0117, 0x0600, 0x0119, 0x0115, 0x0116, 622 0x011a, 0x010c, 0x010d, 0x011b, 0x011c, 0x0110, 0x0311, 0x011d, 623 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 624 }, ctrl_map[NR_KEYS] = { 625 0x0200, 0x0200, 0x0200, 0x0000, 0x001b, 0x001c, 0x001d, 0x001e, 626 0x001f, 0x007f, 0x0200, 0x0200, 0x001f, 0x0200, 0x0008, 0x0200, 627 0x0011, 0x0017, 0x0005, 0x0012, 0x0014, 0x0019, 0x0015, 0x0009, 628 0x000f, 0x0010, 0x001b, 0x001d, 0x0201, 0x0702, 0x0001, 0x0013, 629 0x0004, 0x0006, 0x0007, 0x0008, 0x000a, 0x000b, 0x000c, 0x0200, 630 0x0007, 0x0000, 0x0700, 0x001c, 0x001a, 0x0018, 0x0003, 0x0016, 631 0x0002, 0x000e, 0x000d, 0x0200, 0x020e, 0x007f, 0x0700, 0x030c, 632 0x0703, 0x0000, 0x0207, 0x0100, 0x0101, 0x0102, 0x0103, 0x0104, 633 0x0105, 0x0106, 0x0107, 0x0108, 0x0109, 0x0208, 0x0204, 0x0307, 634 0x0308, 0x0309, 0x030b, 0x0304, 0x0305, 0x0306, 0x030a, 0x0301, 635 0x0302, 0x0303, 0x0300, 0x0310, 0x0206, 0x0200, 0x0200, 0x010a, 636 0x010b, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 637 0x030e, 0x0702, 0x030d, 0x001c, 0x0701, 0x0205, 0x0114, 0x0603, 638 0x0118, 0x0601, 0x0602, 0x0117, 0x0600, 0x0119, 0x0115, 0x0116, 639 0x011a, 0x010c, 0x010d, 0x011b, 0x011c, 0x0110, 0x0311, 0x011d, 640 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 0x0200, 641 }; 642 643 const u_short * const linux_keytabs[] = { 644 plain_map, shift_map, altgr_map, altgr_map, ctrl_map 645 }; 646 #endif 647 648 static struct biosdisk_info * 649 fd2biosinfo(p, fp) 650 struct proc *p; 651 struct file *fp; 652 { 653 struct vnode *vp; 654 const char *blkname; 655 char diskname[16]; 656 int i; 657 struct nativedisk_info *nip; 658 struct disklist *dl = i386_alldisks; 659 660 if (fp->f_type != DTYPE_VNODE) 661 return NULL; 662 vp = (struct vnode *)fp->f_data; 663 664 if (vp->v_type != VBLK) 665 return NULL; 666 667 blkname = findblkname(major(vp->v_rdev)); 668 snprintf(diskname, sizeof diskname, "%s%u", blkname, 669 DISKUNIT(vp->v_rdev)); 670 671 for (i = 0; i < dl->dl_nnativedisks; i++) { 672 nip = &dl->dl_nativedisks[i]; 673 if (strcmp(diskname, nip->ni_devname)) 674 continue; 675 if (nip->ni_nmatches != 0) 676 return &dl->dl_biosdisks[nip->ni_biosmatches[0]]; 677 } 678 679 return NULL; 680 } 681 682 683 /* 684 * We come here in a last attempt to satisfy a Linux ioctl() call 685 */ 686 int 687 linux_machdepioctl(p, v, retval) 688 struct proc *p; 689 void *v; 690 register_t *retval; 691 { 692 struct linux_sys_ioctl_args /* { 693 syscallarg(int) fd; 694 syscallarg(u_long) com; 695 syscallarg(caddr_t) data; 696 } */ *uap = v; 697 struct sys_ioctl_args bia; 698 u_long com; 699 int error, error1; 700 #if (NWSDISPLAY > 0) 701 struct vt_mode lvt; 702 caddr_t bvtp, sg; 703 struct kbentry kbe; 704 #endif 705 struct linux_hd_geometry hdg; 706 struct linux_hd_big_geometry hdg_big; 707 struct biosdisk_info *bip; 708 struct filedesc *fdp; 709 struct file *fp; 710 int fd; 711 struct disklabel label, *labp; 712 struct partinfo partp; 713 int (*ioctlf) __P((struct file *, u_long, caddr_t, struct proc *)); 714 u_long start, biostotal, realtotal; 715 u_char heads, sectors; 716 u_int cylinders; 717 struct ioctl_pt pt; 718 719 fd = SCARG(uap, fd); 720 SCARG(&bia, fd) = fd; 721 SCARG(&bia, data) = SCARG(uap, data); 722 com = SCARG(uap, com); 723 724 fdp = p->p_fd; 725 726 if ((fp = fd_getfile(fdp, fd)) == NULL) 727 return (EBADF); 728 729 switch (com) { 730 #if (NWSDISPLAY > 0) 731 case LINUX_KDGKBMODE: 732 com = KDGKBMODE; 733 break; 734 case LINUX_KDSKBMODE: 735 com = KDSKBMODE; 736 if ((unsigned)SCARG(uap, data) == LINUX_K_MEDIUMRAW) 737 SCARG(&bia, data) = (caddr_t)K_RAW; 738 break; 739 case LINUX_KIOCSOUND: 740 SCARG(&bia, data) = 741 (caddr_t)(((unsigned long)SCARG(&bia, data)) & 0xffff); 742 /* fall through */ 743 case LINUX_KDMKTONE: 744 com = KDMKTONE; 745 break; 746 case LINUX_KDSETMODE: 747 com = KDSETMODE; 748 break; 749 case LINUX_KDGETMODE: 750 /* KD_* values are equal to the wscons numbers */ 751 com = WSDISPLAYIO_GMODE; 752 break; 753 case LINUX_KDENABIO: 754 com = KDENABIO; 755 break; 756 case LINUX_KDDISABIO: 757 com = KDDISABIO; 758 break; 759 case LINUX_KDGETLED: 760 com = KDGETLED; 761 break; 762 case LINUX_KDSETLED: 763 com = KDSETLED; 764 break; 765 case LINUX_VT_OPENQRY: 766 com = VT_OPENQRY; 767 break; 768 case LINUX_VT_GETMODE: 769 SCARG(&bia, com) = VT_GETMODE; 770 if ((error = sys_ioctl(p, &bia, retval))) 771 return error; 772 if ((error = copyin(SCARG(uap, data), (caddr_t)&lvt, 773 sizeof (struct vt_mode)))) 774 return error; 775 lvt.relsig = native_to_linux_signo[lvt.relsig]; 776 lvt.acqsig = native_to_linux_signo[lvt.acqsig]; 777 lvt.frsig = native_to_linux_signo[lvt.frsig]; 778 return copyout((caddr_t)&lvt, SCARG(uap, data), 779 sizeof (struct vt_mode)); 780 case LINUX_VT_SETMODE: 781 com = VT_SETMODE; 782 if ((error = copyin(SCARG(uap, data), (caddr_t)&lvt, 783 sizeof (struct vt_mode)))) 784 return error; 785 lvt.relsig = linux_to_native_signo[lvt.relsig]; 786 lvt.acqsig = linux_to_native_signo[lvt.acqsig]; 787 lvt.frsig = linux_to_native_signo[lvt.frsig]; 788 sg = stackgap_init(p, 0); 789 bvtp = stackgap_alloc(p, &sg, sizeof (struct vt_mode)); 790 if ((error = copyout(&lvt, bvtp, sizeof (struct vt_mode)))) 791 return error; 792 SCARG(&bia, data) = bvtp; 793 break; 794 case LINUX_VT_DISALLOCATE: 795 /* XXX should use WSDISPLAYIO_DELSCREEN */ 796 return 0; 797 case LINUX_VT_RELDISP: 798 com = VT_RELDISP; 799 break; 800 case LINUX_VT_ACTIVATE: 801 com = VT_ACTIVATE; 802 break; 803 case LINUX_VT_WAITACTIVE: 804 com = VT_WAITACTIVE; 805 break; 806 case LINUX_VT_GETSTATE: 807 com = VT_GETSTATE; 808 break; 809 case LINUX_KDGKBTYPE: 810 /* This is what Linux does. */ 811 return (subyte(SCARG(uap, data), KB_101)); 812 case LINUX_KDGKBENT: 813 /* 814 * The Linux KDGKBENT ioctl is different from the 815 * SYSV original. So we handle it in machdep code. 816 * XXX We should use keyboard mapping information 817 * from wsdisplay, but this would be expensive. 818 */ 819 if ((error = copyin(SCARG(uap, data), &kbe, 820 sizeof(struct kbentry)))) 821 return (error); 822 if (kbe.kb_table >= sizeof(linux_keytabs) / sizeof(u_short *) 823 || kbe.kb_index >= NR_KEYS) 824 return (EINVAL); 825 kbe.kb_value = linux_keytabs[kbe.kb_table][kbe.kb_index]; 826 return (copyout(&kbe, SCARG(uap, data), 827 sizeof(struct kbentry))); 828 #endif 829 case LINUX_HDIO_GETGEO: 830 case LINUX_HDIO_GETGEO_BIG: 831 /* 832 * Try to mimic Linux behaviour: return the BIOS geometry 833 * if possible (extending its # of cylinders if it's beyond 834 * the 1023 limit), fall back to the MI geometry (i.e. 835 * the real geometry) if not found, by returning an 836 * error. See common/linux_hdio.c 837 */ 838 FILE_USE(fp); 839 bip = fd2biosinfo(p, fp); 840 ioctlf = fp->f_ops->fo_ioctl; 841 error = ioctlf(fp, DIOCGDEFLABEL, (caddr_t)&label, p); 842 error1 = ioctlf(fp, DIOCGPART, (caddr_t)&partp, p); 843 FILE_UNUSE(fp, p); 844 if (error != 0 && error1 != 0) 845 return error1; 846 labp = error != 0 ? &label : partp.disklab; 847 start = error1 != 0 ? partp.part->p_offset : 0; 848 if (bip != NULL && bip->bi_head != 0 && bip->bi_sec != 0 849 && bip->bi_cyl != 0) { 850 heads = bip->bi_head; 851 sectors = bip->bi_sec; 852 cylinders = bip->bi_cyl; 853 biostotal = heads * sectors * cylinders; 854 realtotal = labp->d_ntracks * labp->d_nsectors * 855 labp->d_ncylinders; 856 if (realtotal > biostotal) 857 cylinders = realtotal / (heads * sectors); 858 } else { 859 heads = labp->d_ntracks; 860 cylinders = labp->d_ncylinders; 861 sectors = labp->d_nsectors; 862 } 863 if (com == LINUX_HDIO_GETGEO) { 864 hdg.start = start; 865 hdg.heads = heads; 866 hdg.cylinders = cylinders; 867 hdg.sectors = sectors; 868 return copyout(&hdg, SCARG(uap, data), sizeof hdg); 869 } else { 870 hdg_big.start = start; 871 hdg_big.heads = heads; 872 hdg_big.cylinders = cylinders; 873 hdg_big.sectors = sectors; 874 return copyout(&hdg_big, SCARG(uap, data), 875 sizeof hdg_big); 876 } 877 878 default: 879 /* 880 * Unknown to us. If it's on a device, just pass it through 881 * using PTIOCLINUX, the device itself might be able to 882 * make some sense of it. 883 * XXX hack: if the function returns EJUSTRETURN, 884 * it has stuffed a sysctl return value in pt.data. 885 */ 886 FILE_USE(fp); 887 ioctlf = fp->f_ops->fo_ioctl; 888 pt.com = SCARG(uap, com); 889 pt.data = SCARG(uap, data); 890 error = ioctlf(fp, PTIOCLINUX, (caddr_t)&pt, p); 891 FILE_UNUSE(fp, p); 892 if (error == EJUSTRETURN) { 893 retval[0] = (register_t)pt.data; 894 error = 0; 895 } 896 897 if (error == ENOTTY) 898 DPRINTF(("linux_machdepioctl: invalid ioctl %08lx\n", 899 com)); 900 return error; 901 } 902 SCARG(&bia, com) = com; 903 return sys_ioctl(p, &bia, retval); 904 } 905 906 /* 907 * Set I/O permissions for a process. Just set the maximum level 908 * right away (ignoring the argument), otherwise we would have 909 * to rely on I/O permission maps, which are not implemented. 910 */ 911 int 912 linux_sys_iopl(p, v, retval) 913 struct proc *p; 914 void *v; 915 register_t *retval; 916 { 917 #if 0 918 struct linux_sys_iopl_args /* { 919 syscallarg(int) level; 920 } */ *uap = v; 921 #endif 922 struct trapframe *fp = p->p_md.md_regs; 923 924 if (suser(p->p_ucred, &p->p_acflag) != 0) 925 return EPERM; 926 fp->tf_eflags |= PSL_IOPL; 927 *retval = 0; 928 return 0; 929 } 930 931 /* 932 * See above. If a root process tries to set access to an I/O port, 933 * just let it have the whole range. 934 */ 935 int 936 linux_sys_ioperm(p, v, retval) 937 struct proc *p; 938 void *v; 939 register_t *retval; 940 { 941 struct linux_sys_ioperm_args /* { 942 syscallarg(unsigned int) lo; 943 syscallarg(unsigned int) hi; 944 syscallarg(int) val; 945 } */ *uap = v; 946 struct trapframe *fp = p->p_md.md_regs; 947 948 if (suser(p->p_ucred, &p->p_acflag) != 0) 949 return EPERM; 950 if (SCARG(uap, val)) 951 fp->tf_eflags |= PSL_IOPL; 952 *retval = 0; 953 return 0; 954 } 955