freebsd-dev/sys/kern/subr_trap.c

1179 lines
28 KiB
C
Raw Normal View History

1993-06-12 14:58:17 +00:00
/*-
* Copyright (C) 1994, David Greenman
* Copyright (c) 1990, 1993
* The Regents of the University of California. All rights reserved.
1993-06-12 14:58:17 +00:00
*
* This code is derived from software contributed to Berkeley by
* the University of Utah, and William Jolitz.
*
* 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.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS 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 REGENTS 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.
*
genassym.c: Remove NKMEMCLUSTERS, it is no longer define or used. locores.s: Fix comment on PTDpde and APTDpde to be pde instead of pte Add new equation for calculating location of Sysmap Remove Bill's old #ifdef garbage for counting up memory, that stuff will never be made to work and was just cluttering up the file. Add code that places the PTD, page table pages, and kernel stack below the 640k ISA hole if there is room for it, otherwise put this stuff all at 1MB. This fixes the 28K bogusity in the boot blocks, that can now go away! Fix the caclulation of where first is to be dependent on NKPDE so that we can skip over the above mentioned areas. The 28K thing is now 44K in size due to the increase in kernel virtual memory space, but since we no longer have to worry about that this is no big deal. Use if NNPX > 0 instead of ifdef NPX for floating point code. machdep.c Change the calculation of for the buffer cache to be 20% of all memory above 2MB and add back the upper limit of 2/5's of the VM_KMEM_SIZE so that we do not eat ALL of the kernel memory space on large memory machines, note that this will not even come into effect unless you have more than 32MB. The current buffer cache limit is 6.7MB due to this caclulation. It seems that we where erroniously allocating bufpages pages for buffer_map. buffer_map is UNUSED in this implementation of the buffer cache, but since the map is referenced in several if statements a quick fix was to simply allocate 1 vm page (but no real memory) to it. pmap.h Remove rcsid, don't want them in the kernel files! Removed some cruft inside an #ifdef DEBUGx that caused compiler errors if you where compiling this for debug. Use the #defines for PD_SHIFT and PG_SHIFT in place of constants. trap.c: Remove patch kit header and rcsid, fix $Id$. Now include "npx.h" and use NNPX for controlling the floating point code. Remove a now completly invalid check for a maximum virtual address, the virtual address now ends at 0xFFFFFFFF so there is no more MAX!! (Thanks David, I completly missed that one!) vm_machdep.c Remove patch kit header and rcsid, fix $Id$. Now include "npx.h" and use NNPX for controlling the floating point code. Replace several 0xFE00000 constants with KERNBASE
1993-10-15 10:34:29 +00:00
* from: @(#)trap.c 7.4 (Berkeley) 5/13/91
* $Id: trap.c,v 1.132 1998/12/28 23:02:56 msmith Exp $
1993-06-12 14:58:17 +00:00
*/
/*
1994-06-11 05:13:33 +00:00
* 386 Trap and System call handling
1993-06-12 14:58:17 +00:00
*/
#include "opt_cpu.h"
1996-01-04 21:13:23 +00:00
#include "opt_ddb.h"
#include "opt_ktrace.h"
#include "opt_trap.h"
#include "opt_vm86.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/proc.h>
#include <sys/pioctl.h>
#include <sys/kernel.h>
#include <sys/resourcevar.h>
#include <sys/signalvar.h>
#include <sys/syscall.h>
#include <sys/sysent.h>
#include <sys/uio.h>
#include <sys/vmmeter.h>
1993-06-12 14:58:17 +00:00
#ifdef KTRACE
#include <sys/ktrace.h>
1993-06-12 14:58:17 +00:00
#endif
#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/vm_prot.h>
#include <sys/lock.h>
#include <vm/pmap.h>
#include <vm/vm_kern.h>
#include <vm/vm_map.h>
#include <vm/vm_page.h>
#include <vm/vm_extern.h>
#include <machine/cpu.h>
#include <machine/ipl.h>
#include <machine/md_var.h>
#include <machine/pcb.h>
#ifdef SMP
#include <machine/smp.h>
#endif
#include <machine/tss.h>
1993-06-12 14:58:17 +00:00
#include <i386/isa/intr_machdep.h>
#ifdef POWERFAIL_NMI
1996-09-10 08:32:01 +00:00
#include <sys/syslog.h>
#include <machine/clock.h>
#endif
#ifdef VM86
#include <machine/vm86.h>
#endif
#ifdef DDB
extern int in_Debugger, debugger_on_panic;
#endif
#include "isa.h"
#include "npx.h"
1993-06-12 14:58:17 +00:00
extern struct i386tss common_tss;
1995-12-14 08:21:33 +00:00
int (*pmath_emulate) __P((struct trapframe *));
extern void trap __P((struct trapframe frame));
extern int trapwrite __P((unsigned addr));
extern void syscall __P((struct trapframe frame));
static int trap_pfault __P((struct trapframe *, int, vm_offset_t));
static void trap_fatal __P((struct trapframe *, vm_offset_t));
void dblfault_handler __P((void));
1993-06-12 14:58:17 +00:00
extern inthand_t IDTVEC(syscall);
#define MAX_TRAP_MSG 28
static char *trap_msg[] = {
Fix security holes in sigreturn(), ptrace() and procfs. sigreturn() attempted to check for insecure and fatal eflags and segment selectors, but missed many cases and got the IOPL check back to front. The other syscalls didn't check at all. sys_process.c, machdep.c: Only allow PT_WRITE_U to write to the registers (ordinary and FP). psl.h, locore.s, machdep.c: Eliminate PSL_MBZ, PSL_MBO and PSL_USERCLR. We are not supposed to assume anything about the reserved bits. Use PSL_USERCHANGE and PSL_KERNEL instead. Rename PSL_USERSET to PSL_USER. exception.s: Define a private label for use by doreti when returning to user mode fails. machdep.c: In syscalls, allow changing only the eflags that can be changed on 486's in user mode (no longer attempt to allow benign IOPL changes; allow changing the nasty PSL_NT; don't allow changing the i586 bits). Don't attempt to check all the cases involving invalid selectors and %eip's. Just check for privilege violations and let the invalid things cause a trap. procfs_machdep.c: Call the ptrace register functions to do all the work for reading and writing ordinary registers and for single stepping. trap.c: Ignore traps caused by PSL_NT being set. Previously, users could cause a fatal trap in user mode by setting PSL_NT and executing an iret, and a fatal trap in kernel mode by setting PSL_NT and making a syscall. PSL_NT was cleared too late and not in enough modes to fix the problem. Make all traps in user mode (except T_NMI) nonfatal. Recover from traps caused by attempting to load invalid user registers in doreti by restarting the traps so that they appear to occur in user mode. --- Fix bogons that I noticed while fixing the above: psl.h: Fix some comments. Uniformize idempotency ifdef. exception.s, machdep.c: Remove rsvd[0-14]. rsvd0 hasn't been reserved since the 486 came out. Replace rsvd0 by `align'. rsvd[0-11] used wrong (magic non-unique) trap numbers. Replace rsvd[1-14] by rsvd. locore.s: Enable alignment check flag on 486's and 586's. machdep.c: Use a better type for kstack[]. Use TFREGP() to find the registers. Reformat ptrace functions from SEF to something closer to KNF. procfs_machdep.c: The wrong pointer to the registers got fixed as a side effect. Implement reading and writing of FP registers. /proc/*/*regs now work (only) for processes that are in memory. Clean up comments. trap.c, trap.h: Remove unused trap types.
1995-01-14 13:20:26 +00:00
"", /* 0 unused */
First steps in rewriting locore.s, and making info useful when the machine panics. i386/i386/locore.s: 1) got rid of most .set directives that were being used like #define's, and replaced them with appropriate #define's in the appropriate header files (accessed via genassym). 2) added comments to header inclusions and global definitions, and global variables 3) replaced some hardcoded constants with cpp defines (such as PDESIZE and others) 4) aligned all comments to the same column to make them easier to read 5) moved macro definitions for ENTRY, ALIGN, NOP, etc. to /sys/i386/include/asmacros.h 6) added #ifdef BDE_DEBUGGER around all of Bruce's debugger code 7) added new global '_KERNend' to store last location+1 of kernel 8) cleaned up zeroing of bss so that only bss is zeroed 9) fix zeroing of page tables so that it really does zero them all - not just if they follow the bss. 10) rewrote page table initialization code so that 1) works correctly and 2) write protects the kernel text by default 11) properly initialize the kernel page directory, upages, p0stack PT, and page tables. The previous scheme was more than a bit screwy. 12) change allocation of virtual area of IO hole so that it is fixed at KERNBASE + 0xa0000. The previous scheme put it right after the kernel page tables and then later expected it to be at KERNBASE +0xa0000 13) change multiple bogus settings of user read/write of various areas of kernel VM - including the IO hole; we should never be accessing the IO hole in user mode through the kernel page tables 14) split kernel support routines such as bcopy, bzero, copyin, copyout, etc. into a seperate file 'support.s' 15) split swtch and related routines into a seperate 'swtch.s' 16) split routines related to traps, syscalls, and interrupts into a seperate file 'exception.s' 17) remove some unused global variables from locore that got inserted by Garrett when he pulled them out of some .h files. i386/isa/icu.s: 1) clean up global variable declarations 2) move in declaration of astpending and netisr i386/i386/pmap.c: 1) fix calculation of virtual_avail. It previously was calculated to be right in the middle of the kernel page tables - not a good place to start allocating kernel VM. 2) properly allocate kernel page dir/tables etc out of kernel map - previously only took out 2 pages. i386/i386/machdep.c: 1) modify boot() to print a warning that the system will reboot in PANIC_REBOOT_WAIT_TIME amount of seconds, and let the user abort with a key on the console. The machine will wait for ever if a key is typed before the reboot. The default is 15 seconds, but can be set to 0 to mean don't wait at all, -1 to mean wait forever, or any positive value to wait for that many seconds. 2) print "Rebooting..." just before doing it. kern/subr_prf.c: 1) remove PANICWAIT as it is deprecated by the change to machdep.c i386/i386/trap.c: 1) add table of trap type strings and use it to print a real trap/ panic message rather than just a number. Lot's of work to be done here, but this is the first step. Symbolic traceback is in the TODO. i386/i386/Makefile.i386: 1) add support in to build support.s, exception.s and swtch.s ...and various changes to various header files to make all of the above happen.
1993-11-13 02:25:21 +00:00
"privileged instruction fault", /* 1 T_PRIVINFLT */
Fix security holes in sigreturn(), ptrace() and procfs. sigreturn() attempted to check for insecure and fatal eflags and segment selectors, but missed many cases and got the IOPL check back to front. The other syscalls didn't check at all. sys_process.c, machdep.c: Only allow PT_WRITE_U to write to the registers (ordinary and FP). psl.h, locore.s, machdep.c: Eliminate PSL_MBZ, PSL_MBO and PSL_USERCLR. We are not supposed to assume anything about the reserved bits. Use PSL_USERCHANGE and PSL_KERNEL instead. Rename PSL_USERSET to PSL_USER. exception.s: Define a private label for use by doreti when returning to user mode fails. machdep.c: In syscalls, allow changing only the eflags that can be changed on 486's in user mode (no longer attempt to allow benign IOPL changes; allow changing the nasty PSL_NT; don't allow changing the i586 bits). Don't attempt to check all the cases involving invalid selectors and %eip's. Just check for privilege violations and let the invalid things cause a trap. procfs_machdep.c: Call the ptrace register functions to do all the work for reading and writing ordinary registers and for single stepping. trap.c: Ignore traps caused by PSL_NT being set. Previously, users could cause a fatal trap in user mode by setting PSL_NT and executing an iret, and a fatal trap in kernel mode by setting PSL_NT and making a syscall. PSL_NT was cleared too late and not in enough modes to fix the problem. Make all traps in user mode (except T_NMI) nonfatal. Recover from traps caused by attempting to load invalid user registers in doreti by restarting the traps so that they appear to occur in user mode. --- Fix bogons that I noticed while fixing the above: psl.h: Fix some comments. Uniformize idempotency ifdef. exception.s, machdep.c: Remove rsvd[0-14]. rsvd0 hasn't been reserved since the 486 came out. Replace rsvd0 by `align'. rsvd[0-11] used wrong (magic non-unique) trap numbers. Replace rsvd[1-14] by rsvd. locore.s: Enable alignment check flag on 486's and 586's. machdep.c: Use a better type for kstack[]. Use TFREGP() to find the registers. Reformat ptrace functions from SEF to something closer to KNF. procfs_machdep.c: The wrong pointer to the registers got fixed as a side effect. Implement reading and writing of FP registers. /proc/*/*regs now work (only) for processes that are in memory. Clean up comments. trap.c, trap.h: Remove unused trap types.
1995-01-14 13:20:26 +00:00
"", /* 2 unused */
First steps in rewriting locore.s, and making info useful when the machine panics. i386/i386/locore.s: 1) got rid of most .set directives that were being used like #define's, and replaced them with appropriate #define's in the appropriate header files (accessed via genassym). 2) added comments to header inclusions and global definitions, and global variables 3) replaced some hardcoded constants with cpp defines (such as PDESIZE and others) 4) aligned all comments to the same column to make them easier to read 5) moved macro definitions for ENTRY, ALIGN, NOP, etc. to /sys/i386/include/asmacros.h 6) added #ifdef BDE_DEBUGGER around all of Bruce's debugger code 7) added new global '_KERNend' to store last location+1 of kernel 8) cleaned up zeroing of bss so that only bss is zeroed 9) fix zeroing of page tables so that it really does zero them all - not just if they follow the bss. 10) rewrote page table initialization code so that 1) works correctly and 2) write protects the kernel text by default 11) properly initialize the kernel page directory, upages, p0stack PT, and page tables. The previous scheme was more than a bit screwy. 12) change allocation of virtual area of IO hole so that it is fixed at KERNBASE + 0xa0000. The previous scheme put it right after the kernel page tables and then later expected it to be at KERNBASE +0xa0000 13) change multiple bogus settings of user read/write of various areas of kernel VM - including the IO hole; we should never be accessing the IO hole in user mode through the kernel page tables 14) split kernel support routines such as bcopy, bzero, copyin, copyout, etc. into a seperate file 'support.s' 15) split swtch and related routines into a seperate 'swtch.s' 16) split routines related to traps, syscalls, and interrupts into a seperate file 'exception.s' 17) remove some unused global variables from locore that got inserted by Garrett when he pulled them out of some .h files. i386/isa/icu.s: 1) clean up global variable declarations 2) move in declaration of astpending and netisr i386/i386/pmap.c: 1) fix calculation of virtual_avail. It previously was calculated to be right in the middle of the kernel page tables - not a good place to start allocating kernel VM. 2) properly allocate kernel page dir/tables etc out of kernel map - previously only took out 2 pages. i386/i386/machdep.c: 1) modify boot() to print a warning that the system will reboot in PANIC_REBOOT_WAIT_TIME amount of seconds, and let the user abort with a key on the console. The machine will wait for ever if a key is typed before the reboot. The default is 15 seconds, but can be set to 0 to mean don't wait at all, -1 to mean wait forever, or any positive value to wait for that many seconds. 2) print "Rebooting..." just before doing it. kern/subr_prf.c: 1) remove PANICWAIT as it is deprecated by the change to machdep.c i386/i386/trap.c: 1) add table of trap type strings and use it to print a real trap/ panic message rather than just a number. Lot's of work to be done here, but this is the first step. Symbolic traceback is in the TODO. i386/i386/Makefile.i386: 1) add support in to build support.s, exception.s and swtch.s ...and various changes to various header files to make all of the above happen.
1993-11-13 02:25:21 +00:00
"breakpoint instruction fault", /* 3 T_BPTFLT */
"", /* 4 unused */
Fix security holes in sigreturn(), ptrace() and procfs. sigreturn() attempted to check for insecure and fatal eflags and segment selectors, but missed many cases and got the IOPL check back to front. The other syscalls didn't check at all. sys_process.c, machdep.c: Only allow PT_WRITE_U to write to the registers (ordinary and FP). psl.h, locore.s, machdep.c: Eliminate PSL_MBZ, PSL_MBO and PSL_USERCLR. We are not supposed to assume anything about the reserved bits. Use PSL_USERCHANGE and PSL_KERNEL instead. Rename PSL_USERSET to PSL_USER. exception.s: Define a private label for use by doreti when returning to user mode fails. machdep.c: In syscalls, allow changing only the eflags that can be changed on 486's in user mode (no longer attempt to allow benign IOPL changes; allow changing the nasty PSL_NT; don't allow changing the i586 bits). Don't attempt to check all the cases involving invalid selectors and %eip's. Just check for privilege violations and let the invalid things cause a trap. procfs_machdep.c: Call the ptrace register functions to do all the work for reading and writing ordinary registers and for single stepping. trap.c: Ignore traps caused by PSL_NT being set. Previously, users could cause a fatal trap in user mode by setting PSL_NT and executing an iret, and a fatal trap in kernel mode by setting PSL_NT and making a syscall. PSL_NT was cleared too late and not in enough modes to fix the problem. Make all traps in user mode (except T_NMI) nonfatal. Recover from traps caused by attempting to load invalid user registers in doreti by restarting the traps so that they appear to occur in user mode. --- Fix bogons that I noticed while fixing the above: psl.h: Fix some comments. Uniformize idempotency ifdef. exception.s, machdep.c: Remove rsvd[0-14]. rsvd0 hasn't been reserved since the 486 came out. Replace rsvd0 by `align'. rsvd[0-11] used wrong (magic non-unique) trap numbers. Replace rsvd[1-14] by rsvd. locore.s: Enable alignment check flag on 486's and 586's. machdep.c: Use a better type for kstack[]. Use TFREGP() to find the registers. Reformat ptrace functions from SEF to something closer to KNF. procfs_machdep.c: The wrong pointer to the registers got fixed as a side effect. Implement reading and writing of FP registers. /proc/*/*regs now work (only) for processes that are in memory. Clean up comments. trap.c, trap.h: Remove unused trap types.
1995-01-14 13:20:26 +00:00
"", /* 5 unused */
First steps in rewriting locore.s, and making info useful when the machine panics. i386/i386/locore.s: 1) got rid of most .set directives that were being used like #define's, and replaced them with appropriate #define's in the appropriate header files (accessed via genassym). 2) added comments to header inclusions and global definitions, and global variables 3) replaced some hardcoded constants with cpp defines (such as PDESIZE and others) 4) aligned all comments to the same column to make them easier to read 5) moved macro definitions for ENTRY, ALIGN, NOP, etc. to /sys/i386/include/asmacros.h 6) added #ifdef BDE_DEBUGGER around all of Bruce's debugger code 7) added new global '_KERNend' to store last location+1 of kernel 8) cleaned up zeroing of bss so that only bss is zeroed 9) fix zeroing of page tables so that it really does zero them all - not just if they follow the bss. 10) rewrote page table initialization code so that 1) works correctly and 2) write protects the kernel text by default 11) properly initialize the kernel page directory, upages, p0stack PT, and page tables. The previous scheme was more than a bit screwy. 12) change allocation of virtual area of IO hole so that it is fixed at KERNBASE + 0xa0000. The previous scheme put it right after the kernel page tables and then later expected it to be at KERNBASE +0xa0000 13) change multiple bogus settings of user read/write of various areas of kernel VM - including the IO hole; we should never be accessing the IO hole in user mode through the kernel page tables 14) split kernel support routines such as bcopy, bzero, copyin, copyout, etc. into a seperate file 'support.s' 15) split swtch and related routines into a seperate 'swtch.s' 16) split routines related to traps, syscalls, and interrupts into a seperate file 'exception.s' 17) remove some unused global variables from locore that got inserted by Garrett when he pulled them out of some .h files. i386/isa/icu.s: 1) clean up global variable declarations 2) move in declaration of astpending and netisr i386/i386/pmap.c: 1) fix calculation of virtual_avail. It previously was calculated to be right in the middle of the kernel page tables - not a good place to start allocating kernel VM. 2) properly allocate kernel page dir/tables etc out of kernel map - previously only took out 2 pages. i386/i386/machdep.c: 1) modify boot() to print a warning that the system will reboot in PANIC_REBOOT_WAIT_TIME amount of seconds, and let the user abort with a key on the console. The machine will wait for ever if a key is typed before the reboot. The default is 15 seconds, but can be set to 0 to mean don't wait at all, -1 to mean wait forever, or any positive value to wait for that many seconds. 2) print "Rebooting..." just before doing it. kern/subr_prf.c: 1) remove PANICWAIT as it is deprecated by the change to machdep.c i386/i386/trap.c: 1) add table of trap type strings and use it to print a real trap/ panic message rather than just a number. Lot's of work to be done here, but this is the first step. Symbolic traceback is in the TODO. i386/i386/Makefile.i386: 1) add support in to build support.s, exception.s and swtch.s ...and various changes to various header files to make all of the above happen.
1993-11-13 02:25:21 +00:00
"arithmetic trap", /* 6 T_ARITHTRAP */
"system forced exception", /* 7 T_ASTFLT */
Fix security holes in sigreturn(), ptrace() and procfs. sigreturn() attempted to check for insecure and fatal eflags and segment selectors, but missed many cases and got the IOPL check back to front. The other syscalls didn't check at all. sys_process.c, machdep.c: Only allow PT_WRITE_U to write to the registers (ordinary and FP). psl.h, locore.s, machdep.c: Eliminate PSL_MBZ, PSL_MBO and PSL_USERCLR. We are not supposed to assume anything about the reserved bits. Use PSL_USERCHANGE and PSL_KERNEL instead. Rename PSL_USERSET to PSL_USER. exception.s: Define a private label for use by doreti when returning to user mode fails. machdep.c: In syscalls, allow changing only the eflags that can be changed on 486's in user mode (no longer attempt to allow benign IOPL changes; allow changing the nasty PSL_NT; don't allow changing the i586 bits). Don't attempt to check all the cases involving invalid selectors and %eip's. Just check for privilege violations and let the invalid things cause a trap. procfs_machdep.c: Call the ptrace register functions to do all the work for reading and writing ordinary registers and for single stepping. trap.c: Ignore traps caused by PSL_NT being set. Previously, users could cause a fatal trap in user mode by setting PSL_NT and executing an iret, and a fatal trap in kernel mode by setting PSL_NT and making a syscall. PSL_NT was cleared too late and not in enough modes to fix the problem. Make all traps in user mode (except T_NMI) nonfatal. Recover from traps caused by attempting to load invalid user registers in doreti by restarting the traps so that they appear to occur in user mode. --- Fix bogons that I noticed while fixing the above: psl.h: Fix some comments. Uniformize idempotency ifdef. exception.s, machdep.c: Remove rsvd[0-14]. rsvd0 hasn't been reserved since the 486 came out. Replace rsvd0 by `align'. rsvd[0-11] used wrong (magic non-unique) trap numbers. Replace rsvd[1-14] by rsvd. locore.s: Enable alignment check flag on 486's and 586's. machdep.c: Use a better type for kstack[]. Use TFREGP() to find the registers. Reformat ptrace functions from SEF to something closer to KNF. procfs_machdep.c: The wrong pointer to the registers got fixed as a side effect. Implement reading and writing of FP registers. /proc/*/*regs now work (only) for processes that are in memory. Clean up comments. trap.c, trap.h: Remove unused trap types.
1995-01-14 13:20:26 +00:00
"", /* 8 unused */
"general protection fault", /* 9 T_PROTFLT */
First steps in rewriting locore.s, and making info useful when the machine panics. i386/i386/locore.s: 1) got rid of most .set directives that were being used like #define's, and replaced them with appropriate #define's in the appropriate header files (accessed via genassym). 2) added comments to header inclusions and global definitions, and global variables 3) replaced some hardcoded constants with cpp defines (such as PDESIZE and others) 4) aligned all comments to the same column to make them easier to read 5) moved macro definitions for ENTRY, ALIGN, NOP, etc. to /sys/i386/include/asmacros.h 6) added #ifdef BDE_DEBUGGER around all of Bruce's debugger code 7) added new global '_KERNend' to store last location+1 of kernel 8) cleaned up zeroing of bss so that only bss is zeroed 9) fix zeroing of page tables so that it really does zero them all - not just if they follow the bss. 10) rewrote page table initialization code so that 1) works correctly and 2) write protects the kernel text by default 11) properly initialize the kernel page directory, upages, p0stack PT, and page tables. The previous scheme was more than a bit screwy. 12) change allocation of virtual area of IO hole so that it is fixed at KERNBASE + 0xa0000. The previous scheme put it right after the kernel page tables and then later expected it to be at KERNBASE +0xa0000 13) change multiple bogus settings of user read/write of various areas of kernel VM - including the IO hole; we should never be accessing the IO hole in user mode through the kernel page tables 14) split kernel support routines such as bcopy, bzero, copyin, copyout, etc. into a seperate file 'support.s' 15) split swtch and related routines into a seperate 'swtch.s' 16) split routines related to traps, syscalls, and interrupts into a seperate file 'exception.s' 17) remove some unused global variables from locore that got inserted by Garrett when he pulled them out of some .h files. i386/isa/icu.s: 1) clean up global variable declarations 2) move in declaration of astpending and netisr i386/i386/pmap.c: 1) fix calculation of virtual_avail. It previously was calculated to be right in the middle of the kernel page tables - not a good place to start allocating kernel VM. 2) properly allocate kernel page dir/tables etc out of kernel map - previously only took out 2 pages. i386/i386/machdep.c: 1) modify boot() to print a warning that the system will reboot in PANIC_REBOOT_WAIT_TIME amount of seconds, and let the user abort with a key on the console. The machine will wait for ever if a key is typed before the reboot. The default is 15 seconds, but can be set to 0 to mean don't wait at all, -1 to mean wait forever, or any positive value to wait for that many seconds. 2) print "Rebooting..." just before doing it. kern/subr_prf.c: 1) remove PANICWAIT as it is deprecated by the change to machdep.c i386/i386/trap.c: 1) add table of trap type strings and use it to print a real trap/ panic message rather than just a number. Lot's of work to be done here, but this is the first step. Symbolic traceback is in the TODO. i386/i386/Makefile.i386: 1) add support in to build support.s, exception.s and swtch.s ...and various changes to various header files to make all of the above happen.
1993-11-13 02:25:21 +00:00
"trace trap", /* 10 T_TRCTRAP */
"", /* 11 unused */
"page fault", /* 12 T_PAGEFLT */
Fix security holes in sigreturn(), ptrace() and procfs. sigreturn() attempted to check for insecure and fatal eflags and segment selectors, but missed many cases and got the IOPL check back to front. The other syscalls didn't check at all. sys_process.c, machdep.c: Only allow PT_WRITE_U to write to the registers (ordinary and FP). psl.h, locore.s, machdep.c: Eliminate PSL_MBZ, PSL_MBO and PSL_USERCLR. We are not supposed to assume anything about the reserved bits. Use PSL_USERCHANGE and PSL_KERNEL instead. Rename PSL_USERSET to PSL_USER. exception.s: Define a private label for use by doreti when returning to user mode fails. machdep.c: In syscalls, allow changing only the eflags that can be changed on 486's in user mode (no longer attempt to allow benign IOPL changes; allow changing the nasty PSL_NT; don't allow changing the i586 bits). Don't attempt to check all the cases involving invalid selectors and %eip's. Just check for privilege violations and let the invalid things cause a trap. procfs_machdep.c: Call the ptrace register functions to do all the work for reading and writing ordinary registers and for single stepping. trap.c: Ignore traps caused by PSL_NT being set. Previously, users could cause a fatal trap in user mode by setting PSL_NT and executing an iret, and a fatal trap in kernel mode by setting PSL_NT and making a syscall. PSL_NT was cleared too late and not in enough modes to fix the problem. Make all traps in user mode (except T_NMI) nonfatal. Recover from traps caused by attempting to load invalid user registers in doreti by restarting the traps so that they appear to occur in user mode. --- Fix bogons that I noticed while fixing the above: psl.h: Fix some comments. Uniformize idempotency ifdef. exception.s, machdep.c: Remove rsvd[0-14]. rsvd0 hasn't been reserved since the 486 came out. Replace rsvd0 by `align'. rsvd[0-11] used wrong (magic non-unique) trap numbers. Replace rsvd[1-14] by rsvd. locore.s: Enable alignment check flag on 486's and 586's. machdep.c: Use a better type for kstack[]. Use TFREGP() to find the registers. Reformat ptrace functions from SEF to something closer to KNF. procfs_machdep.c: The wrong pointer to the registers got fixed as a side effect. Implement reading and writing of FP registers. /proc/*/*regs now work (only) for processes that are in memory. Clean up comments. trap.c, trap.h: Remove unused trap types.
1995-01-14 13:20:26 +00:00
"", /* 13 unused */
First steps in rewriting locore.s, and making info useful when the machine panics. i386/i386/locore.s: 1) got rid of most .set directives that were being used like #define's, and replaced them with appropriate #define's in the appropriate header files (accessed via genassym). 2) added comments to header inclusions and global definitions, and global variables 3) replaced some hardcoded constants with cpp defines (such as PDESIZE and others) 4) aligned all comments to the same column to make them easier to read 5) moved macro definitions for ENTRY, ALIGN, NOP, etc. to /sys/i386/include/asmacros.h 6) added #ifdef BDE_DEBUGGER around all of Bruce's debugger code 7) added new global '_KERNend' to store last location+1 of kernel 8) cleaned up zeroing of bss so that only bss is zeroed 9) fix zeroing of page tables so that it really does zero them all - not just if they follow the bss. 10) rewrote page table initialization code so that 1) works correctly and 2) write protects the kernel text by default 11) properly initialize the kernel page directory, upages, p0stack PT, and page tables. The previous scheme was more than a bit screwy. 12) change allocation of virtual area of IO hole so that it is fixed at KERNBASE + 0xa0000. The previous scheme put it right after the kernel page tables and then later expected it to be at KERNBASE +0xa0000 13) change multiple bogus settings of user read/write of various areas of kernel VM - including the IO hole; we should never be accessing the IO hole in user mode through the kernel page tables 14) split kernel support routines such as bcopy, bzero, copyin, copyout, etc. into a seperate file 'support.s' 15) split swtch and related routines into a seperate 'swtch.s' 16) split routines related to traps, syscalls, and interrupts into a seperate file 'exception.s' 17) remove some unused global variables from locore that got inserted by Garrett when he pulled them out of some .h files. i386/isa/icu.s: 1) clean up global variable declarations 2) move in declaration of astpending and netisr i386/i386/pmap.c: 1) fix calculation of virtual_avail. It previously was calculated to be right in the middle of the kernel page tables - not a good place to start allocating kernel VM. 2) properly allocate kernel page dir/tables etc out of kernel map - previously only took out 2 pages. i386/i386/machdep.c: 1) modify boot() to print a warning that the system will reboot in PANIC_REBOOT_WAIT_TIME amount of seconds, and let the user abort with a key on the console. The machine will wait for ever if a key is typed before the reboot. The default is 15 seconds, but can be set to 0 to mean don't wait at all, -1 to mean wait forever, or any positive value to wait for that many seconds. 2) print "Rebooting..." just before doing it. kern/subr_prf.c: 1) remove PANICWAIT as it is deprecated by the change to machdep.c i386/i386/trap.c: 1) add table of trap type strings and use it to print a real trap/ panic message rather than just a number. Lot's of work to be done here, but this is the first step. Symbolic traceback is in the TODO. i386/i386/Makefile.i386: 1) add support in to build support.s, exception.s and swtch.s ...and various changes to various header files to make all of the above happen.
1993-11-13 02:25:21 +00:00
"alignment fault", /* 14 T_ALIGNFLT */
Fix security holes in sigreturn(), ptrace() and procfs. sigreturn() attempted to check for insecure and fatal eflags and segment selectors, but missed many cases and got the IOPL check back to front. The other syscalls didn't check at all. sys_process.c, machdep.c: Only allow PT_WRITE_U to write to the registers (ordinary and FP). psl.h, locore.s, machdep.c: Eliminate PSL_MBZ, PSL_MBO and PSL_USERCLR. We are not supposed to assume anything about the reserved bits. Use PSL_USERCHANGE and PSL_KERNEL instead. Rename PSL_USERSET to PSL_USER. exception.s: Define a private label for use by doreti when returning to user mode fails. machdep.c: In syscalls, allow changing only the eflags that can be changed on 486's in user mode (no longer attempt to allow benign IOPL changes; allow changing the nasty PSL_NT; don't allow changing the i586 bits). Don't attempt to check all the cases involving invalid selectors and %eip's. Just check for privilege violations and let the invalid things cause a trap. procfs_machdep.c: Call the ptrace register functions to do all the work for reading and writing ordinary registers and for single stepping. trap.c: Ignore traps caused by PSL_NT being set. Previously, users could cause a fatal trap in user mode by setting PSL_NT and executing an iret, and a fatal trap in kernel mode by setting PSL_NT and making a syscall. PSL_NT was cleared too late and not in enough modes to fix the problem. Make all traps in user mode (except T_NMI) nonfatal. Recover from traps caused by attempting to load invalid user registers in doreti by restarting the traps so that they appear to occur in user mode. --- Fix bogons that I noticed while fixing the above: psl.h: Fix some comments. Uniformize idempotency ifdef. exception.s, machdep.c: Remove rsvd[0-14]. rsvd0 hasn't been reserved since the 486 came out. Replace rsvd0 by `align'. rsvd[0-11] used wrong (magic non-unique) trap numbers. Replace rsvd[1-14] by rsvd. locore.s: Enable alignment check flag on 486's and 586's. machdep.c: Use a better type for kstack[]. Use TFREGP() to find the registers. Reformat ptrace functions from SEF to something closer to KNF. procfs_machdep.c: The wrong pointer to the registers got fixed as a side effect. Implement reading and writing of FP registers. /proc/*/*regs now work (only) for processes that are in memory. Clean up comments. trap.c, trap.h: Remove unused trap types.
1995-01-14 13:20:26 +00:00
"", /* 15 unused */
"", /* 16 unused */
"", /* 17 unused */
First steps in rewriting locore.s, and making info useful when the machine panics. i386/i386/locore.s: 1) got rid of most .set directives that were being used like #define's, and replaced them with appropriate #define's in the appropriate header files (accessed via genassym). 2) added comments to header inclusions and global definitions, and global variables 3) replaced some hardcoded constants with cpp defines (such as PDESIZE and others) 4) aligned all comments to the same column to make them easier to read 5) moved macro definitions for ENTRY, ALIGN, NOP, etc. to /sys/i386/include/asmacros.h 6) added #ifdef BDE_DEBUGGER around all of Bruce's debugger code 7) added new global '_KERNend' to store last location+1 of kernel 8) cleaned up zeroing of bss so that only bss is zeroed 9) fix zeroing of page tables so that it really does zero them all - not just if they follow the bss. 10) rewrote page table initialization code so that 1) works correctly and 2) write protects the kernel text by default 11) properly initialize the kernel page directory, upages, p0stack PT, and page tables. The previous scheme was more than a bit screwy. 12) change allocation of virtual area of IO hole so that it is fixed at KERNBASE + 0xa0000. The previous scheme put it right after the kernel page tables and then later expected it to be at KERNBASE +0xa0000 13) change multiple bogus settings of user read/write of various areas of kernel VM - including the IO hole; we should never be accessing the IO hole in user mode through the kernel page tables 14) split kernel support routines such as bcopy, bzero, copyin, copyout, etc. into a seperate file 'support.s' 15) split swtch and related routines into a seperate 'swtch.s' 16) split routines related to traps, syscalls, and interrupts into a seperate file 'exception.s' 17) remove some unused global variables from locore that got inserted by Garrett when he pulled them out of some .h files. i386/isa/icu.s: 1) clean up global variable declarations 2) move in declaration of astpending and netisr i386/i386/pmap.c: 1) fix calculation of virtual_avail. It previously was calculated to be right in the middle of the kernel page tables - not a good place to start allocating kernel VM. 2) properly allocate kernel page dir/tables etc out of kernel map - previously only took out 2 pages. i386/i386/machdep.c: 1) modify boot() to print a warning that the system will reboot in PANIC_REBOOT_WAIT_TIME amount of seconds, and let the user abort with a key on the console. The machine will wait for ever if a key is typed before the reboot. The default is 15 seconds, but can be set to 0 to mean don't wait at all, -1 to mean wait forever, or any positive value to wait for that many seconds. 2) print "Rebooting..." just before doing it. kern/subr_prf.c: 1) remove PANICWAIT as it is deprecated by the change to machdep.c i386/i386/trap.c: 1) add table of trap type strings and use it to print a real trap/ panic message rather than just a number. Lot's of work to be done here, but this is the first step. Symbolic traceback is in the TODO. i386/i386/Makefile.i386: 1) add support in to build support.s, exception.s and swtch.s ...and various changes to various header files to make all of the above happen.
1993-11-13 02:25:21 +00:00
"integer divide fault", /* 18 T_DIVIDE */
"non-maskable interrupt trap", /* 19 T_NMI */
"overflow trap", /* 20 T_OFLOW */
"FPU bounds check fault", /* 21 T_BOUND */
"FPU device not available", /* 22 T_DNA */
"double fault", /* 23 T_DOUBLEFLT */
"FPU operand fetch fault", /* 24 T_FPOPFLT */
"invalid TSS fault", /* 25 T_TSSFLT */
"segment not present fault", /* 26 T_SEGNPFLT */
"stack fault", /* 27 T_STKFLT */
"machine check trap", /* 28 T_MCHK */
First steps in rewriting locore.s, and making info useful when the machine panics. i386/i386/locore.s: 1) got rid of most .set directives that were being used like #define's, and replaced them with appropriate #define's in the appropriate header files (accessed via genassym). 2) added comments to header inclusions and global definitions, and global variables 3) replaced some hardcoded constants with cpp defines (such as PDESIZE and others) 4) aligned all comments to the same column to make them easier to read 5) moved macro definitions for ENTRY, ALIGN, NOP, etc. to /sys/i386/include/asmacros.h 6) added #ifdef BDE_DEBUGGER around all of Bruce's debugger code 7) added new global '_KERNend' to store last location+1 of kernel 8) cleaned up zeroing of bss so that only bss is zeroed 9) fix zeroing of page tables so that it really does zero them all - not just if they follow the bss. 10) rewrote page table initialization code so that 1) works correctly and 2) write protects the kernel text by default 11) properly initialize the kernel page directory, upages, p0stack PT, and page tables. The previous scheme was more than a bit screwy. 12) change allocation of virtual area of IO hole so that it is fixed at KERNBASE + 0xa0000. The previous scheme put it right after the kernel page tables and then later expected it to be at KERNBASE +0xa0000 13) change multiple bogus settings of user read/write of various areas of kernel VM - including the IO hole; we should never be accessing the IO hole in user mode through the kernel page tables 14) split kernel support routines such as bcopy, bzero, copyin, copyout, etc. into a seperate file 'support.s' 15) split swtch and related routines into a seperate 'swtch.s' 16) split routines related to traps, syscalls, and interrupts into a seperate file 'exception.s' 17) remove some unused global variables from locore that got inserted by Garrett when he pulled them out of some .h files. i386/isa/icu.s: 1) clean up global variable declarations 2) move in declaration of astpending and netisr i386/i386/pmap.c: 1) fix calculation of virtual_avail. It previously was calculated to be right in the middle of the kernel page tables - not a good place to start allocating kernel VM. 2) properly allocate kernel page dir/tables etc out of kernel map - previously only took out 2 pages. i386/i386/machdep.c: 1) modify boot() to print a warning that the system will reboot in PANIC_REBOOT_WAIT_TIME amount of seconds, and let the user abort with a key on the console. The machine will wait for ever if a key is typed before the reboot. The default is 15 seconds, but can be set to 0 to mean don't wait at all, -1 to mean wait forever, or any positive value to wait for that many seconds. 2) print "Rebooting..." just before doing it. kern/subr_prf.c: 1) remove PANICWAIT as it is deprecated by the change to machdep.c i386/i386/trap.c: 1) add table of trap type strings and use it to print a real trap/ panic message rather than just a number. Lot's of work to be done here, but this is the first step. Symbolic traceback is in the TODO. i386/i386/Makefile.i386: 1) add support in to build support.s, exception.s and swtch.s ...and various changes to various header files to make all of the above happen.
1993-11-13 02:25:21 +00:00
};
1998-04-15 17:47:40 +00:00
static __inline void userret __P((struct proc *p, struct trapframe *frame,
u_quad_t oticks));
#if defined(I586_CPU) && !defined(NO_F00F_HACK)
extern struct gate_descriptor *t_idt;
extern int has_f00f_bug;
#endif
1998-04-15 17:47:40 +00:00
static __inline void
userret(p, frame, oticks)
struct proc *p;
struct trapframe *frame;
u_quad_t oticks;
{
int sig, s;
while ((sig = CURSIG(p)) != 0)
postsig(sig);
#if 0
if (!want_resched &&
(p->p_priority <= p->p_usrpri) &&
(p->p_rtprio.type == RTP_PRIO_NORMAL)) {
int newpriority;
p->p_estcpu += 1;
newpriority = PUSER + p->p_estcpu / 4 + 2 * p->p_nice;
newpriority = min(newpriority, MAXPRI);
p->p_usrpri = newpriority;
}
#endif
p->p_priority = p->p_usrpri;
if (want_resched) {
/*
* Since we are curproc, clock will normally just change
* our priority without moving us from one queue to another
* (since the running process is not on a queue.)
* If that happened after we setrunqueue ourselves but before we
* mi_switch()'ed, we might not be on the queue indicated by
* our priority.
*/
s = splhigh();
setrunqueue(p);
p->p_stats->p_ru.ru_nivcsw++;
mi_switch();
splx(s);
while ((sig = CURSIG(p)) != 0)
postsig(sig);
}
/*
* Charge system time if profiling.
*/
if (p->p_flag & P_PROFIL)
addupc_task(p, frame->tf_eip,
(u_int)(p->p_sticks - oticks) * psratio);
curpriority = p->p_priority;
}
1993-06-12 14:58:17 +00:00
/*
* Exception, fault, and trap interface to the FreeBSD kernel.
* This common code is called from assembly language IDT gate entry
1993-06-12 14:58:17 +00:00
* routines that prepare a suitable stack frame, and restore this
* frame after the exception has been processed.
1993-06-12 14:58:17 +00:00
*/
void
1993-06-12 14:58:17 +00:00
trap(frame)
struct trapframe frame;
{
struct proc *p = curproc;
u_quad_t sticks = 0;
int i = 0, ucode = 0, type, code;
vm_offset_t eva;
if (!(frame.tf_eflags & PSL_I)) {
/*
* Buggy application or kernel code has disabled interrupts
* and then trapped. Enabling interrupts now is wrong, but
* it is better than running with interrupts disabled until
* they are accidentally enabled later.
*/
type = frame.tf_trapno;
if (ISPL(frame.tf_cs) == SEL_UPL || (frame.tf_eflags & PSL_VM))
printf(
"pid %ld (%s): trap %d with interrupts disabled\n",
(long)curproc->p_pid, curproc->p_comm, type);
else if (type != T_BPTFLT && type != T_TRCTRAP)
/*
* XXX not quite right, since this may be for a
* multiple fault in user mode.
*/
printf("kernel trap %d with interrupts disabled\n",
type);
enable_intr();
}
eva = 0;
if (frame.tf_trapno == T_PAGEFLT) {
/*
* For some Cyrix CPUs, %cr2 is clobbered by interrupts.
* This problem is worked around by using an interrupt
* gate for the pagefault handler. We are finally ready
* to read %cr2 and then must reenable interrupts.
*
* XXX this should be in the switch statement, but the
* NO_FOOF_HACK and VM86 goto and ifdefs obfuscate the
* flow of control too much for this to be obviously
* correct.
*/
eva = rcr2();
enable_intr();
}
1993-06-12 14:58:17 +00:00
#if defined(I586_CPU) && !defined(NO_F00F_HACK)
restart:
#endif
1993-06-12 14:58:17 +00:00
type = frame.tf_trapno;
code = frame.tf_err;
1995-05-30 08:16:23 +00:00
#ifdef VM86
if (in_vm86call) {
if (frame.tf_eflags & PSL_VM &&
(type == T_PROTFLT || type == T_STKFLT)) {
i = vm86_emulate((struct vm86frame *)&frame);
if (i != 0)
/*
* returns to original process
*/
vm86_trap((struct vm86frame *)&frame);
return;
}
switch (type) {
/*
* these traps want either a process context, or
* assume a normal userspace trap.
*/
case T_PROTFLT:
case T_SEGNPFLT:
trap_fatal(&frame, eva);
return;
case T_TRCTRAP:
type = T_BPTFLT; /* kernel breakpoint */
/* FALL THROUGH */
}
goto kernel_trap; /* normal kernel trap handling */
}
#endif
if ((ISPL(frame.tf_cs) == SEL_UPL) || (frame.tf_eflags & PSL_VM)) {
/* user trap */
sticks = p->p_sticks;
1997-05-07 20:08:53 +00:00
p->p_md.md_regs = &frame;
switch (type) {
case T_PRIVINFLT: /* privileged instruction fault */
ucode = type;
i = SIGILL;
break;
case T_BPTFLT: /* bpt instruction fault */
case T_TRCTRAP: /* trace trap */
frame.tf_eflags &= ~PSL_T;
i = SIGTRAP;
break;
case T_ARITHTRAP: /* arithmetic trap */
ucode = code;
i = SIGFPE;
break;
case T_ASTFLT: /* Allow process switch */
astoff();
cnt.v_soft++;
if (p->p_flag & P_OWEUPC) {
p->p_flag &= ~P_OWEUPC;
addupc_task(p, p->p_stats->p_prof.pr_addr,
p->p_stats->p_prof.pr_ticks);
}
goto out;
1993-06-12 14:58:17 +00:00
/*
* The following two traps can happen in
* vm86 mode, and, if so, we want to handle
* them specially.
*/
case T_PROTFLT: /* general protection fault */
case T_STKFLT: /* stack fault */
#ifdef VM86
if (frame.tf_eflags & PSL_VM) {
i = vm86_emulate((struct vm86frame *)&frame);
if (i == 0)
goto out;
break;
}
#endif /* VM86 */
/* FALL THROUGH */
case T_SEGNPFLT: /* segment not present fault */
Fix security holes in sigreturn(), ptrace() and procfs. sigreturn() attempted to check for insecure and fatal eflags and segment selectors, but missed many cases and got the IOPL check back to front. The other syscalls didn't check at all. sys_process.c, machdep.c: Only allow PT_WRITE_U to write to the registers (ordinary and FP). psl.h, locore.s, machdep.c: Eliminate PSL_MBZ, PSL_MBO and PSL_USERCLR. We are not supposed to assume anything about the reserved bits. Use PSL_USERCHANGE and PSL_KERNEL instead. Rename PSL_USERSET to PSL_USER. exception.s: Define a private label for use by doreti when returning to user mode fails. machdep.c: In syscalls, allow changing only the eflags that can be changed on 486's in user mode (no longer attempt to allow benign IOPL changes; allow changing the nasty PSL_NT; don't allow changing the i586 bits). Don't attempt to check all the cases involving invalid selectors and %eip's. Just check for privilege violations and let the invalid things cause a trap. procfs_machdep.c: Call the ptrace register functions to do all the work for reading and writing ordinary registers and for single stepping. trap.c: Ignore traps caused by PSL_NT being set. Previously, users could cause a fatal trap in user mode by setting PSL_NT and executing an iret, and a fatal trap in kernel mode by setting PSL_NT and making a syscall. PSL_NT was cleared too late and not in enough modes to fix the problem. Make all traps in user mode (except T_NMI) nonfatal. Recover from traps caused by attempting to load invalid user registers in doreti by restarting the traps so that they appear to occur in user mode. --- Fix bogons that I noticed while fixing the above: psl.h: Fix some comments. Uniformize idempotency ifdef. exception.s, machdep.c: Remove rsvd[0-14]. rsvd0 hasn't been reserved since the 486 came out. Replace rsvd0 by `align'. rsvd[0-11] used wrong (magic non-unique) trap numbers. Replace rsvd[1-14] by rsvd. locore.s: Enable alignment check flag on 486's and 586's. machdep.c: Use a better type for kstack[]. Use TFREGP() to find the registers. Reformat ptrace functions from SEF to something closer to KNF. procfs_machdep.c: The wrong pointer to the registers got fixed as a side effect. Implement reading and writing of FP registers. /proc/*/*regs now work (only) for processes that are in memory. Clean up comments. trap.c, trap.h: Remove unused trap types.
1995-01-14 13:20:26 +00:00
case T_TSSFLT: /* invalid TSS fault */
case T_DOUBLEFLT: /* double fault */
default:
ucode = code + BUS_SEGM_FAULT ;
i = SIGBUS;
break;
1993-06-12 14:58:17 +00:00
case T_PAGEFLT: /* page fault */
i = trap_pfault(&frame, TRUE, eva);
if (i == -1)
return;
#if defined(I586_CPU) && !defined(NO_F00F_HACK)
if (i == -2)
goto restart;
#endif
if (i == 0)
goto out;
1993-06-12 14:58:17 +00:00
ucode = T_PAGEFLT;
break;
1993-06-12 14:58:17 +00:00
case T_DIVIDE: /* integer divide fault */
ucode = FPE_INTDIV_TRAP;
i = SIGFPE;
break;
1993-06-12 14:58:17 +00:00
#if NISA > 0
case T_NMI:
#ifdef POWERFAIL_NMI
goto handle_powerfail;
#else /* !POWERFAIL_NMI */
#ifdef DDB
/* NMI can be hooked up to a pushbutton for debugging */
printf ("NMI ... going to debugger\n");
if (kdb_trap (type, 0, &frame))
return;
#endif /* DDB */
/* machine/parity/power fail/"kitchen sink" faults */
if (isa_nmi(code) == 0) return;
panic("NMI indicates hardware failure");
#endif /* POWERFAIL_NMI */
#endif /* NISA > 0 */
1993-06-12 14:58:17 +00:00
case T_OFLOW: /* integer overflow fault */
ucode = FPE_INTOVF_TRAP;
i = SIGFPE;
break;
1993-06-12 14:58:17 +00:00
case T_BOUND: /* bounds check fault */
ucode = FPE_SUBRNG_TRAP;
i = SIGFPE;
break;
1993-06-12 14:58:17 +00:00
case T_DNA:
#if NNPX > 0
/* if a transparent fault (due to context switch "late") */
if (npxdna())
return;
#endif
1995-12-14 08:21:33 +00:00
if (!pmath_emulate) {
i = SIGFPE;
ucode = FPE_FPU_NP_TRAP;
break;
}
i = (*pmath_emulate)(&frame);
if (i == 0) {
if (!(frame.tf_eflags & PSL_T))
return;
frame.tf_eflags &= ~PSL_T;
i = SIGTRAP;
}
/* else ucode = emulator_only_knows() XXX */
break;
case T_FPOPFLT: /* FPU operand fetch fault */
ucode = T_FPOPFLT;
i = SIGILL;
break;
}
} else {
#ifdef VM86
kernel_trap:
#endif
/* kernel trap */
switch (type) {
case T_PAGEFLT: /* page fault */
(void) trap_pfault(&frame, FALSE, eva);
return;
case T_DNA:
#if NNPX > 0
/*
* The kernel is apparently using npx for copying.
* XXX this should be fatal unless the kernel has
* registered such use.
*/
if (npxdna())
return;
#endif
break;
case T_PROTFLT: /* general protection fault */
case T_SEGNPFLT: /* segment not present fault */
Fix security holes in sigreturn(), ptrace() and procfs. sigreturn() attempted to check for insecure and fatal eflags and segment selectors, but missed many cases and got the IOPL check back to front. The other syscalls didn't check at all. sys_process.c, machdep.c: Only allow PT_WRITE_U to write to the registers (ordinary and FP). psl.h, locore.s, machdep.c: Eliminate PSL_MBZ, PSL_MBO and PSL_USERCLR. We are not supposed to assume anything about the reserved bits. Use PSL_USERCHANGE and PSL_KERNEL instead. Rename PSL_USERSET to PSL_USER. exception.s: Define a private label for use by doreti when returning to user mode fails. machdep.c: In syscalls, allow changing only the eflags that can be changed on 486's in user mode (no longer attempt to allow benign IOPL changes; allow changing the nasty PSL_NT; don't allow changing the i586 bits). Don't attempt to check all the cases involving invalid selectors and %eip's. Just check for privilege violations and let the invalid things cause a trap. procfs_machdep.c: Call the ptrace register functions to do all the work for reading and writing ordinary registers and for single stepping. trap.c: Ignore traps caused by PSL_NT being set. Previously, users could cause a fatal trap in user mode by setting PSL_NT and executing an iret, and a fatal trap in kernel mode by setting PSL_NT and making a syscall. PSL_NT was cleared too late and not in enough modes to fix the problem. Make all traps in user mode (except T_NMI) nonfatal. Recover from traps caused by attempting to load invalid user registers in doreti by restarting the traps so that they appear to occur in user mode. --- Fix bogons that I noticed while fixing the above: psl.h: Fix some comments. Uniformize idempotency ifdef. exception.s, machdep.c: Remove rsvd[0-14]. rsvd0 hasn't been reserved since the 486 came out. Replace rsvd0 by `align'. rsvd[0-11] used wrong (magic non-unique) trap numbers. Replace rsvd[1-14] by rsvd. locore.s: Enable alignment check flag on 486's and 586's. machdep.c: Use a better type for kstack[]. Use TFREGP() to find the registers. Reformat ptrace functions from SEF to something closer to KNF. procfs_machdep.c: The wrong pointer to the registers got fixed as a side effect. Implement reading and writing of FP registers. /proc/*/*regs now work (only) for processes that are in memory. Clean up comments. trap.c, trap.h: Remove unused trap types.
1995-01-14 13:20:26 +00:00
/*
* Invalid segment selectors and out of bounds
* %eip's and %esp's can be set up in user mode.
* This causes a fault in kernel mode when the
* kernel tries to return to user mode. We want
* to get this fault so that we can fix the
* problem here and not have to check all the
* selectors and pointers when the user changes
* them.
*/
#define MAYBE_DORETI_FAULT(where, whereto) \
do { \
if (frame.tf_eip == (int)where) { \
frame.tf_eip = (int)whereto; \
return; \
} \
} while (0)
if (intr_nesting_level == 0) {
/*
* Invalid %fs's and %gs's can be created using
* procfs or PT_SETREGS or by invalidating the
* underlying LDT entry. This causes a fault
* in kernel mode when the kernel attempts to
* switch contexts. Lose the bad context
* (XXX) so that we can continue, and generate
* a signal.
*/
if (frame.tf_eip == (int)cpu_switch_load_fs) {
curpcb->pcb_fs = 0;
psignal(p, SIGBUS);
return;
}
if (frame.tf_eip == (int)cpu_switch_load_gs) {
curpcb->pcb_gs = 0;
psignal(p, SIGBUS);
return;
}
Fix security holes in sigreturn(), ptrace() and procfs. sigreturn() attempted to check for insecure and fatal eflags and segment selectors, but missed many cases and got the IOPL check back to front. The other syscalls didn't check at all. sys_process.c, machdep.c: Only allow PT_WRITE_U to write to the registers (ordinary and FP). psl.h, locore.s, machdep.c: Eliminate PSL_MBZ, PSL_MBO and PSL_USERCLR. We are not supposed to assume anything about the reserved bits. Use PSL_USERCHANGE and PSL_KERNEL instead. Rename PSL_USERSET to PSL_USER. exception.s: Define a private label for use by doreti when returning to user mode fails. machdep.c: In syscalls, allow changing only the eflags that can be changed on 486's in user mode (no longer attempt to allow benign IOPL changes; allow changing the nasty PSL_NT; don't allow changing the i586 bits). Don't attempt to check all the cases involving invalid selectors and %eip's. Just check for privilege violations and let the invalid things cause a trap. procfs_machdep.c: Call the ptrace register functions to do all the work for reading and writing ordinary registers and for single stepping. trap.c: Ignore traps caused by PSL_NT being set. Previously, users could cause a fatal trap in user mode by setting PSL_NT and executing an iret, and a fatal trap in kernel mode by setting PSL_NT and making a syscall. PSL_NT was cleared too late and not in enough modes to fix the problem. Make all traps in user mode (except T_NMI) nonfatal. Recover from traps caused by attempting to load invalid user registers in doreti by restarting the traps so that they appear to occur in user mode. --- Fix bogons that I noticed while fixing the above: psl.h: Fix some comments. Uniformize idempotency ifdef. exception.s, machdep.c: Remove rsvd[0-14]. rsvd0 hasn't been reserved since the 486 came out. Replace rsvd0 by `align'. rsvd[0-11] used wrong (magic non-unique) trap numbers. Replace rsvd[1-14] by rsvd. locore.s: Enable alignment check flag on 486's and 586's. machdep.c: Use a better type for kstack[]. Use TFREGP() to find the registers. Reformat ptrace functions from SEF to something closer to KNF. procfs_machdep.c: The wrong pointer to the registers got fixed as a side effect. Implement reading and writing of FP registers. /proc/*/*regs now work (only) for processes that are in memory. Clean up comments. trap.c, trap.h: Remove unused trap types.
1995-01-14 13:20:26 +00:00
MAYBE_DORETI_FAULT(doreti_iret,
doreti_iret_fault);
MAYBE_DORETI_FAULT(doreti_popl_ds,
doreti_popl_ds_fault);
MAYBE_DORETI_FAULT(doreti_popl_es,
doreti_popl_es_fault);
if (curpcb && curpcb->pcb_onfault) {
frame.tf_eip = (int)curpcb->pcb_onfault;
return;
}
}
break;
Fix security holes in sigreturn(), ptrace() and procfs. sigreturn() attempted to check for insecure and fatal eflags and segment selectors, but missed many cases and got the IOPL check back to front. The other syscalls didn't check at all. sys_process.c, machdep.c: Only allow PT_WRITE_U to write to the registers (ordinary and FP). psl.h, locore.s, machdep.c: Eliminate PSL_MBZ, PSL_MBO and PSL_USERCLR. We are not supposed to assume anything about the reserved bits. Use PSL_USERCHANGE and PSL_KERNEL instead. Rename PSL_USERSET to PSL_USER. exception.s: Define a private label for use by doreti when returning to user mode fails. machdep.c: In syscalls, allow changing only the eflags that can be changed on 486's in user mode (no longer attempt to allow benign IOPL changes; allow changing the nasty PSL_NT; don't allow changing the i586 bits). Don't attempt to check all the cases involving invalid selectors and %eip's. Just check for privilege violations and let the invalid things cause a trap. procfs_machdep.c: Call the ptrace register functions to do all the work for reading and writing ordinary registers and for single stepping. trap.c: Ignore traps caused by PSL_NT being set. Previously, users could cause a fatal trap in user mode by setting PSL_NT and executing an iret, and a fatal trap in kernel mode by setting PSL_NT and making a syscall. PSL_NT was cleared too late and not in enough modes to fix the problem. Make all traps in user mode (except T_NMI) nonfatal. Recover from traps caused by attempting to load invalid user registers in doreti by restarting the traps so that they appear to occur in user mode. --- Fix bogons that I noticed while fixing the above: psl.h: Fix some comments. Uniformize idempotency ifdef. exception.s, machdep.c: Remove rsvd[0-14]. rsvd0 hasn't been reserved since the 486 came out. Replace rsvd0 by `align'. rsvd[0-11] used wrong (magic non-unique) trap numbers. Replace rsvd[1-14] by rsvd. locore.s: Enable alignment check flag on 486's and 586's. machdep.c: Use a better type for kstack[]. Use TFREGP() to find the registers. Reformat ptrace functions from SEF to something closer to KNF. procfs_machdep.c: The wrong pointer to the registers got fixed as a side effect. Implement reading and writing of FP registers. /proc/*/*regs now work (only) for processes that are in memory. Clean up comments. trap.c, trap.h: Remove unused trap types.
1995-01-14 13:20:26 +00:00
case T_TSSFLT:
/*
* PSL_NT can be set in user mode and isn't cleared
* automatically when the kernel is entered. This
* causes a TSS fault when the kernel attempts to
* `iret' because the TSS link is uninitialized. We
* want to get this fault so that we can fix the
* problem here and not every time the kernel is
* entered.
*/
if (frame.tf_eflags & PSL_NT) {
frame.tf_eflags &= ~PSL_NT;
return;
}
break;
case T_TRCTRAP: /* trace trap */
if (frame.tf_eip == (int)IDTVEC(syscall)) {
/*
* We've just entered system mode via the
* syscall lcall. Continue single stepping
* silently until the syscall handler has
* saved the flags.
*/
return;
}
if (frame.tf_eip == (int)IDTVEC(syscall) + 1) {
/*
* The syscall handler has now saved the
* flags. Stop single stepping it.
*/
frame.tf_eflags &= ~PSL_T;
return;
}
/*
* Fall through.
*/
case T_BPTFLT:
/*
* If DDB is enabled, let it handle the debugger trap.
* Otherwise, debugger traps "can't happen".
*/
#ifdef DDB
if (kdb_trap (type, 0, &frame))
1993-06-12 14:58:17 +00:00
return;
#endif
break;
1995-05-30 08:16:23 +00:00
#if NISA > 0
case T_NMI:
#ifdef POWERFAIL_NMI
#ifndef TIMER_FREQ
# define TIMER_FREQ 1193182
#endif
handle_powerfail:
{
static unsigned lastalert = 0;
if(time_second - lastalert > 10)
{
log(LOG_WARNING, "NMI: power fail\n");
sysbeep(TIMER_FREQ/880, hz);
lastalert = time_second;
}
return;
}
#else /* !POWERFAIL_NMI */
#ifdef DDB
/* NMI can be hooked up to a pushbutton for debugging */
printf ("NMI ... going to debugger\n");
if (kdb_trap (type, 0, &frame))
return;
#endif /* DDB */
/* machine/parity/power fail/"kitchen sink" faults */
if (isa_nmi(code) == 0) return;
/* FALL THROUGH */
#endif /* POWERFAIL_NMI */
#endif /* NISA > 0 */
}
trap_fatal(&frame, eva);
return;
1993-06-12 14:58:17 +00:00
}
/* Translate fault for emulators (e.g. Linux) */
if (*p->p_sysent->sv_transtrap)
i = (*p->p_sysent->sv_transtrap)(i, type);
1993-06-12 14:58:17 +00:00
trapsignal(p, i, ucode);
#ifdef DEBUG
if (type <= MAX_TRAP_MSG) {
1995-05-30 08:16:23 +00:00
uprintf("fatal process exception: %s",
trap_msg[type]);
if ((type == T_PAGEFLT) || (type == T_PROTFLT))
uprintf(", fault VA = 0x%lx", (u_long)eva);
uprintf("\n");
}
#endif
1993-06-12 14:58:17 +00:00
out:
userret(p, &frame, sticks);
}
#ifdef notyet
/*
* This version doesn't allow a page fault to user space while
* in the kernel. The rest of the kernel needs to be made "safe"
* before this can be used. I think the only things remaining
* to be made safe are the iBCS2 code and the process tracing/
* debugging code.
*/
static int
trap_pfault(frame, usermode, eva)
struct trapframe *frame;
int usermode;
vm_offset_t eva;
{
vm_offset_t va;
struct vmspace *vm = NULL;
vm_map_t map = 0;
int rv = 0;
vm_prot_t ftype;
struct proc *p = curproc;
if (frame->tf_err & PGEX_W)
ftype = VM_PROT_READ | VM_PROT_WRITE;
else
ftype = VM_PROT_READ;
va = trunc_page(eva);
if (va < VM_MIN_KERNEL_ADDRESS) {
vm_offset_t v;
vm_page_t mpte;
if (p == NULL ||
(!usermode && va < VM_MAXUSER_ADDRESS &&
(intr_nesting_level != 0 || curpcb == NULL ||
curpcb->pcb_onfault == NULL))) {
trap_fatal(frame, eva);
return (-1);
}
/*
* This is a fault on non-kernel virtual memory.
* vm is initialized above to NULL. If curproc is NULL
* or curproc->p_vmspace is NULL the fault is fatal.
*/
vm = p->p_vmspace;
if (vm == NULL)
goto nogo;
map = &vm->vm_map;
/*
* Keep swapout from messing with us during this
* critical time.
*/
++p->p_lock;
/*
* Grow the stack if necessary
*/
#ifndef VM_STACK
if ((caddr_t)va > vm->vm_maxsaddr && va < USRSTACK) {
if (!grow(p, va)) {
rv = KERN_FAILURE;
--p->p_lock;
goto nogo;
}
}
#else
/* grow_stack returns false only if va falls into
* a growable stack region and the stack growth
* fails. It returns true if va was not within
* a growable stack region, or if the stack
* growth succeeded.
*/
if (!grow_stack (p, va)) {
rv = KERN_FAILURE;
--p->p_lock;
goto nogo;
}
#endif
/* Fault in the user page: */
rv = vm_fault(map, va, ftype,
(ftype & VM_PROT_WRITE) ? VM_FAULT_DIRTY : 0);
--p->p_lock;
} else {
/*
* Don't allow user-mode faults in kernel address space.
*/
if (usermode)
goto nogo;
/*
* Since we know that kernel virtual address addresses
* always have pte pages mapped, we just have to fault
* the page.
*/
rv = vm_fault(kernel_map, va, ftype, FALSE);
}
if (rv == KERN_SUCCESS)
return (0);
nogo:
if (!usermode) {
if (intr_nesting_level == 0 && curpcb && curpcb->pcb_onfault) {
frame->tf_eip = (int)curpcb->pcb_onfault;
return (0);
}
trap_fatal(frame, eva);
return (-1);
}
/* kludge to pass faulting virtual address to sendsig */
frame->tf_err = eva;
return((rv == KERN_PROTECTION_FAILURE) ? SIGBUS : SIGSEGV);
}
#endif
int
trap_pfault(frame, usermode, eva)
struct trapframe *frame;
int usermode;
vm_offset_t eva;
{
vm_offset_t va;
struct vmspace *vm = NULL;
vm_map_t map = 0;
int rv = 0;
vm_prot_t ftype;
struct proc *p = curproc;
va = trunc_page(eva);
if (va >= KERNBASE) {
/*
* Don't allow user-mode faults in kernel address space.
* An exception: if the faulting address is the invalid
* instruction entry in the IDT, then the Intel Pentium
* F00F bug workaround was triggered, and we need to
* treat it is as an illegal instruction, and not a page
* fault.
*/
#if defined(I586_CPU) && !defined(NO_F00F_HACK)
if ((eva == (unsigned int)&t_idt[6]) && has_f00f_bug) {
frame->tf_trapno = T_PRIVINFLT;
return -2;
}
#endif
if (usermode)
goto nogo;
map = kernel_map;
} else {
/*
* This is a fault on non-kernel virtual memory.
* vm is initialized above to NULL. If curproc is NULL
* or curproc->p_vmspace is NULL the fault is fatal.
*/
if (p != NULL)
vm = p->p_vmspace;
if (vm == NULL)
goto nogo;
map = &vm->vm_map;
}
if (frame->tf_err & PGEX_W)
ftype = VM_PROT_READ | VM_PROT_WRITE;
else
ftype = VM_PROT_READ;
if (map != kernel_map) {
1993-06-12 14:58:17 +00:00
/*
* Keep swapout from messing with us during this
* critical time.
1993-06-12 14:58:17 +00:00
*/
++p->p_lock;
/*
* Grow the stack if necessary
*/
#ifndef VM_STACK
if ((caddr_t)va > vm->vm_maxsaddr && va < USRSTACK) {
if (!grow(p, va)) {
rv = KERN_FAILURE;
--p->p_lock;
goto nogo;
}
}
#else
/* grow_stack returns false only if va falls into
* a growable stack region and the stack growth
* fails. It returns true if va was not within
* a growable stack region, or if the stack
* growth succeeded.
*/
if (!grow_stack (p, va)) {
rv = KERN_FAILURE;
--p->p_lock;
goto nogo;
}
#endif
/* Fault in the user page: */
rv = vm_fault(map, va, ftype,
(ftype & VM_PROT_WRITE) ? VM_FAULT_DIRTY : 0);
--p->p_lock;
} else {
/*
* Don't have to worry about process locking or stacks in the kernel.
*/
rv = vm_fault(map, va, ftype, FALSE);
1993-06-12 14:58:17 +00:00
}
if (rv == KERN_SUCCESS)
return (0);
nogo:
if (!usermode) {
if (intr_nesting_level == 0 && curpcb && curpcb->pcb_onfault) {
frame->tf_eip = (int)curpcb->pcb_onfault;
return (0);
1993-06-12 14:58:17 +00:00
}
trap_fatal(frame, eva);
return (-1);
1993-06-12 14:58:17 +00:00
}
/* kludge to pass faulting virtual address to sendsig */
frame->tf_err = eva;
return((rv == KERN_PROTECTION_FAILURE) ? SIGBUS : SIGSEGV);
}
static void
trap_fatal(frame, eva)
struct trapframe *frame;
vm_offset_t eva;
{
int code, type, ss, esp;
struct soft_segment_descriptor softseg;
code = frame->tf_err;
type = frame->tf_trapno;
sdtossd(&gdt[IDXSEL(frame->tf_cs & 0xffff)].sd, &softseg);
if (type <= MAX_TRAP_MSG)
printf("\n\nFatal trap %d: %s while in %s mode\n",
type, trap_msg[type],
frame->tf_eflags & PSL_VM ? "vm86" :
ISPL(frame->tf_cs) == SEL_UPL ? "user" : "kernel");
#ifdef SMP
/* three seperate prints in case of a trap on an unmapped page */
printf("mp_lock = %08x; ", mp_lock);
printf("cpuid = %d; ", cpuid);
printf("lapic.id = %08x\n", lapic.id);
#endif
if (type == T_PAGEFLT) {
printf("fault virtual address = 0x%x\n", eva);
printf("fault code = %s %s, %s\n",
code & PGEX_U ? "user" : "supervisor",
code & PGEX_W ? "write" : "read",
code & PGEX_P ? "protection violation" : "page not present");
}
printf("instruction pointer = 0x%x:0x%x\n",
frame->tf_cs & 0xffff, frame->tf_eip);
if ((ISPL(frame->tf_cs) == SEL_UPL) || (frame->tf_eflags & PSL_VM)) {
ss = frame->tf_ss & 0xffff;
esp = frame->tf_esp;
} else {
ss = GSEL(GDATA_SEL, SEL_KPL);
esp = (int)&frame->tf_esp;
}
printf("stack pointer = 0x%x:0x%x\n", ss, esp);
printf("frame pointer = 0x%x:0x%x\n", ss, frame->tf_ebp);
printf("code segment = base 0x%x, limit 0x%x, type 0x%x\n",
softseg.ssd_base, softseg.ssd_limit, softseg.ssd_type);
printf(" = DPL %d, pres %d, def32 %d, gran %d\n",
softseg.ssd_dpl, softseg.ssd_p, softseg.ssd_def32,
softseg.ssd_gran);
printf("processor eflags = ");
if (frame->tf_eflags & PSL_T)
printf("trace trap, ");
if (frame->tf_eflags & PSL_I)
printf("interrupt enabled, ");
if (frame->tf_eflags & PSL_NT)
printf("nested task, ");
if (frame->tf_eflags & PSL_RF)
printf("resume, ");
if (frame->tf_eflags & PSL_VM)
printf("vm86, ");
printf("IOPL = %d\n", (frame->tf_eflags & PSL_IOPL) >> 12);
printf("current process = ");
if (curproc) {
printf("%lu (%s)\n",
(u_long)curproc->p_pid, curproc->p_comm ?
curproc->p_comm : "");
} else {
printf("Idle\n");
}
printf("interrupt mask = ");
if ((cpl & net_imask) == net_imask)
printf("net ");
if ((cpl & tty_imask) == tty_imask)
printf("tty ");
if ((cpl & bio_imask) == bio_imask)
printf("bio ");
if ((cpl & cam_imask) == cam_imask)
printf("cam ");
if (cpl == 0)
printf("none");
#ifdef SMP
/**
* XXX FIXME:
* we probably SHOULD have stopped the other CPUs before now!
* another CPU COULD have been touching cpl at this moment...
*/
printf(" <- SMP: XXX");
#endif
printf("\n");
#ifdef KDB
if (kdb_trap(&psl))
return;
#endif
#ifdef DDB
if ((debugger_on_panic || in_Debugger) && kdb_trap(type, 0, frame))
return;
#endif
printf("trap number = %d\n", type);
if (type <= MAX_TRAP_MSG)
panic(trap_msg[type]);
else
panic("unknown/reserved trap");
1993-06-12 14:58:17 +00:00
}
/*
* Double fault handler. Called when a fault occurs while writing
* a frame for a trap/exception onto the stack. This usually occurs
* when the stack overflows (such is the case with infinite recursion,
* for example).
*
* XXX Note that the current PTD gets replaced by IdlePTD when the
* task switch occurs. This means that the stack that was active at
* the time of the double fault is not available at <kstack> unless
1995-12-19 14:47:41 +00:00
* the machine was idle when the double fault occurred. The downside
* of this is that "trace <ebp>" in ddb won't work.
*/
void
dblfault_handler()
{
printf("\nFatal double fault:\n");
printf("eip = 0x%x\n", common_tss.tss_eip);
printf("esp = 0x%x\n", common_tss.tss_esp);
printf("ebp = 0x%x\n", common_tss.tss_ebp);
#ifdef SMP
/* three seperate prints in case of a trap on an unmapped page */
printf("mp_lock = %08x; ", mp_lock);
printf("cpuid = %d; ", cpuid);
printf("lapic.id = %08x\n", lapic.id);
#endif
panic("double fault");
}
1993-06-12 14:58:17 +00:00
/*
* Compensate for 386 brain damage (missing URKR).
* This is a little simpler than the pagefault handler in trap() because
* it the page tables have already been faulted in and high addresses
* are thrown out early for other reasons.
1993-06-12 14:58:17 +00:00
*/
int trapwrite(addr)
unsigned addr;
{
struct proc *p;
vm_offset_t va;
struct vmspace *vm;
int rv;
1993-06-12 14:58:17 +00:00
va = trunc_page((vm_offset_t)addr);
/*
* XXX - MAX is END. Changed > to >= for temp. fix.
*/
if (va >= VM_MAXUSER_ADDRESS)
return (1);
p = curproc;
vm = p->p_vmspace;
++p->p_lock;
#ifndef VM_STACK
if ((caddr_t)va >= vm->vm_maxsaddr && va < USRSTACK) {
if (!grow(p, va)) {
--p->p_lock;
return (1);
}
}
#else
if (!grow_stack (p, va)) {
--p->p_lock;
return (1);
}
#endif
/*
* fault the data page
*/
rv = vm_fault(&vm->vm_map, va, VM_PROT_READ|VM_PROT_WRITE, VM_FAULT_DIRTY);
--p->p_lock;
if (rv != KERN_SUCCESS)
return 1;
return (0);
1993-06-12 14:58:17 +00:00
}
/*
* System call request from POSIX system call gate interface to kernel.
1993-06-12 14:58:17 +00:00
* Like trap(), argument is call by reference.
*/
void
1993-06-12 14:58:17 +00:00
syscall(frame)
struct trapframe frame;
1993-06-12 14:58:17 +00:00
{
caddr_t params;
int i;
struct sysent *callp;
struct proc *p = curproc;
u_quad_t sticks;
int error;
int args[8];
u_int code;
1993-06-12 14:58:17 +00:00
#ifdef DIAGNOSTIC
if (ISPL(frame.tf_cs) != SEL_UPL)
1993-06-12 14:58:17 +00:00
panic("syscall");
#endif
sticks = p->p_sticks;
1997-05-07 20:08:53 +00:00
p->p_md.md_regs = &frame;
params = (caddr_t)frame.tf_esp + sizeof(int);
code = frame.tf_eax;
Mega-commit for Linux emulator update.. This has been stress tested under netscape-2.0 for Linux running all the Java stuff. The scrollbars are now working, at least on my machine. (whew! :-) I'm uncomfortable with the size of this commit, but it's too inter-dependant to easily seperate out. The main changes: COMPAT_LINUX is *GONE*. Most of the code has been moved out of the i386 machine dependent section into the linux emulator itself. The int 0x80 syscall code was almost identical to the lcall 7,0 code and a minor tweak allows them to both be used with the same C code. All kernels can now just modload the lkm and it'll DTRT without having to rebuild the kernel first. Like IBCS2, you can statically compile it in with "options LINUX". A pile of new syscalls implemented, including getdents(), llseek(), readv(), writev(), msync(), personality(). The Linux-ELF libraries want to use some of these. linux_select() now obeys Linux semantics, ie: returns the time remaining of the timeout value rather than leaving it the original value. Quite a few bugs removed, including incorrect arguments being used in syscalls.. eg: mixups between passing the sigset as an int, vs passing it as a pointer and doing a copyin(), missing return values, unhandled cases, SIOC* ioctls, etc. The build for the code has changed. i386/conf/files now knows how to build linux_genassym and generate linux_assym.h on the fly. Supporting changes elsewhere in the kernel: The user-mode signal trampoline has moved from the U area to immediately below the top of the stack (below PS_STRINGS). This allows the different binary emulations to have their own signal trampoline code (which gets rid of the hardwired syscall 103 (sigreturn on BSD, syslog on Linux)) and so that the emulator can provide the exact "struct sigcontext *" argument to the program's signal handlers. The sigstack's "ss_flags" now uses SS_DISABLE and SS_ONSTACK flags, which have the same values as the re-used SA_DISABLE and SA_ONSTACK which are intended for sigaction only. This enables the support of a SA_RESETHAND flag to sigaction to implement the gross SYSV and Linux SA_ONESHOT signal semantics where the signal handler is reset when it's triggered. makesyscalls.sh no longer appends the struct sysentvec on the end of the generated init_sysent.c code. It's a lot saner to have it in a seperate file rather than trying to update the structure inside the awk script. :-) At exec time, the dozen bytes or so of signal trampoline code are copied to the top of the user's stack, rather than obtaining the trampoline code the old way by getting a clone of the parent's user area. This allows Linux and native binaries to freely exec each other without getting trampolines mixed up.
1996-03-02 19:38:20 +00:00
if (p->p_sysent->sv_prepsyscall) {
(*p->p_sysent->sv_prepsyscall)(&frame, args, &code, &params);
} else {
/*
Mega-commit for Linux emulator update.. This has been stress tested under netscape-2.0 for Linux running all the Java stuff. The scrollbars are now working, at least on my machine. (whew! :-) I'm uncomfortable with the size of this commit, but it's too inter-dependant to easily seperate out. The main changes: COMPAT_LINUX is *GONE*. Most of the code has been moved out of the i386 machine dependent section into the linux emulator itself. The int 0x80 syscall code was almost identical to the lcall 7,0 code and a minor tweak allows them to both be used with the same C code. All kernels can now just modload the lkm and it'll DTRT without having to rebuild the kernel first. Like IBCS2, you can statically compile it in with "options LINUX". A pile of new syscalls implemented, including getdents(), llseek(), readv(), writev(), msync(), personality(). The Linux-ELF libraries want to use some of these. linux_select() now obeys Linux semantics, ie: returns the time remaining of the timeout value rather than leaving it the original value. Quite a few bugs removed, including incorrect arguments being used in syscalls.. eg: mixups between passing the sigset as an int, vs passing it as a pointer and doing a copyin(), missing return values, unhandled cases, SIOC* ioctls, etc. The build for the code has changed. i386/conf/files now knows how to build linux_genassym and generate linux_assym.h on the fly. Supporting changes elsewhere in the kernel: The user-mode signal trampoline has moved from the U area to immediately below the top of the stack (below PS_STRINGS). This allows the different binary emulations to have their own signal trampoline code (which gets rid of the hardwired syscall 103 (sigreturn on BSD, syslog on Linux)) and so that the emulator can provide the exact "struct sigcontext *" argument to the program's signal handlers. The sigstack's "ss_flags" now uses SS_DISABLE and SS_ONSTACK flags, which have the same values as the re-used SA_DISABLE and SA_ONSTACK which are intended for sigaction only. This enables the support of a SA_RESETHAND flag to sigaction to implement the gross SYSV and Linux SA_ONESHOT signal semantics where the signal handler is reset when it's triggered. makesyscalls.sh no longer appends the struct sysentvec on the end of the generated init_sysent.c code. It's a lot saner to have it in a seperate file rather than trying to update the structure inside the awk script. :-) At exec time, the dozen bytes or so of signal trampoline code are copied to the top of the user's stack, rather than obtaining the trampoline code the old way by getting a clone of the parent's user area. This allows Linux and native binaries to freely exec each other without getting trampolines mixed up.
1996-03-02 19:38:20 +00:00
* Need to check if this is a 32 bit or 64 bit syscall.
*/
Mega-commit for Linux emulator update.. This has been stress tested under netscape-2.0 for Linux running all the Java stuff. The scrollbars are now working, at least on my machine. (whew! :-) I'm uncomfortable with the size of this commit, but it's too inter-dependant to easily seperate out. The main changes: COMPAT_LINUX is *GONE*. Most of the code has been moved out of the i386 machine dependent section into the linux emulator itself. The int 0x80 syscall code was almost identical to the lcall 7,0 code and a minor tweak allows them to both be used with the same C code. All kernels can now just modload the lkm and it'll DTRT without having to rebuild the kernel first. Like IBCS2, you can statically compile it in with "options LINUX". A pile of new syscalls implemented, including getdents(), llseek(), readv(), writev(), msync(), personality(). The Linux-ELF libraries want to use some of these. linux_select() now obeys Linux semantics, ie: returns the time remaining of the timeout value rather than leaving it the original value. Quite a few bugs removed, including incorrect arguments being used in syscalls.. eg: mixups between passing the sigset as an int, vs passing it as a pointer and doing a copyin(), missing return values, unhandled cases, SIOC* ioctls, etc. The build for the code has changed. i386/conf/files now knows how to build linux_genassym and generate linux_assym.h on the fly. Supporting changes elsewhere in the kernel: The user-mode signal trampoline has moved from the U area to immediately below the top of the stack (below PS_STRINGS). This allows the different binary emulations to have their own signal trampoline code (which gets rid of the hardwired syscall 103 (sigreturn on BSD, syslog on Linux)) and so that the emulator can provide the exact "struct sigcontext *" argument to the program's signal handlers. The sigstack's "ss_flags" now uses SS_DISABLE and SS_ONSTACK flags, which have the same values as the re-used SA_DISABLE and SA_ONSTACK which are intended for sigaction only. This enables the support of a SA_RESETHAND flag to sigaction to implement the gross SYSV and Linux SA_ONESHOT signal semantics where the signal handler is reset when it's triggered. makesyscalls.sh no longer appends the struct sysentvec on the end of the generated init_sysent.c code. It's a lot saner to have it in a seperate file rather than trying to update the structure inside the awk script. :-) At exec time, the dozen bytes or so of signal trampoline code are copied to the top of the user's stack, rather than obtaining the trampoline code the old way by getting a clone of the parent's user area. This allows Linux and native binaries to freely exec each other without getting trampolines mixed up.
1996-03-02 19:38:20 +00:00
if (code == SYS_syscall) {
/*
* Code is first argument, followed by actual args.
*/
code = fuword(params);
params += sizeof(int);
} else if (code == SYS___syscall) {
/*
* Like syscall, but code is a quad, so as to maintain
* quad alignment for the rest of the arguments.
*/
code = fuword(params);
params += sizeof(quad_t);
}
1993-06-12 14:58:17 +00:00
}
if (p->p_sysent->sv_mask)
code &= p->p_sysent->sv_mask;
1995-05-30 08:16:23 +00:00
if (code >= p->p_sysent->sv_size)
callp = &p->p_sysent->sv_table[0];
else
callp = &p->p_sysent->sv_table[code];
1993-06-12 14:58:17 +00:00
Mega-commit for Linux emulator update.. This has been stress tested under netscape-2.0 for Linux running all the Java stuff. The scrollbars are now working, at least on my machine. (whew! :-) I'm uncomfortable with the size of this commit, but it's too inter-dependant to easily seperate out. The main changes: COMPAT_LINUX is *GONE*. Most of the code has been moved out of the i386 machine dependent section into the linux emulator itself. The int 0x80 syscall code was almost identical to the lcall 7,0 code and a minor tweak allows them to both be used with the same C code. All kernels can now just modload the lkm and it'll DTRT without having to rebuild the kernel first. Like IBCS2, you can statically compile it in with "options LINUX". A pile of new syscalls implemented, including getdents(), llseek(), readv(), writev(), msync(), personality(). The Linux-ELF libraries want to use some of these. linux_select() now obeys Linux semantics, ie: returns the time remaining of the timeout value rather than leaving it the original value. Quite a few bugs removed, including incorrect arguments being used in syscalls.. eg: mixups between passing the sigset as an int, vs passing it as a pointer and doing a copyin(), missing return values, unhandled cases, SIOC* ioctls, etc. The build for the code has changed. i386/conf/files now knows how to build linux_genassym and generate linux_assym.h on the fly. Supporting changes elsewhere in the kernel: The user-mode signal trampoline has moved from the U area to immediately below the top of the stack (below PS_STRINGS). This allows the different binary emulations to have their own signal trampoline code (which gets rid of the hardwired syscall 103 (sigreturn on BSD, syslog on Linux)) and so that the emulator can provide the exact "struct sigcontext *" argument to the program's signal handlers. The sigstack's "ss_flags" now uses SS_DISABLE and SS_ONSTACK flags, which have the same values as the re-used SA_DISABLE and SA_ONSTACK which are intended for sigaction only. This enables the support of a SA_RESETHAND flag to sigaction to implement the gross SYSV and Linux SA_ONESHOT signal semantics where the signal handler is reset when it's triggered. makesyscalls.sh no longer appends the struct sysentvec on the end of the generated init_sysent.c code. It's a lot saner to have it in a seperate file rather than trying to update the structure inside the awk script. :-) At exec time, the dozen bytes or so of signal trampoline code are copied to the top of the user's stack, rather than obtaining the trampoline code the old way by getting a clone of the parent's user area. This allows Linux and native binaries to freely exec each other without getting trampolines mixed up.
1996-03-02 19:38:20 +00:00
if (params && (i = callp->sy_narg * sizeof(int)) &&
1993-06-12 14:58:17 +00:00
(error = copyin(params, (caddr_t)args, (u_int)i))) {
#ifdef KTRACE
if (KTRPOINT(p, KTR_SYSCALL))
ktrsyscall(p->p_tracep, code, callp->sy_narg, args);
1993-06-12 14:58:17 +00:00
#endif
goto bad;
1993-06-12 14:58:17 +00:00
}
#ifdef KTRACE
if (KTRPOINT(p, KTR_SYSCALL))
ktrsyscall(p->p_tracep, code, callp->sy_narg, args);
1993-06-12 14:58:17 +00:00
#endif
p->p_retval[0] = 0;
p->p_retval[1] = frame.tf_edx;
STOPEVENT(p, S_SCE, callp->sy_narg);
error = (*callp->sy_call)(p, args);
switch (error) {
case 0:
1993-06-12 14:58:17 +00:00
/*
* Reinitialize proc pointer `p' as it may be different
* if this is a child returning from fork syscall.
1993-06-12 14:58:17 +00:00
*/
p = curproc;
frame.tf_eax = p->p_retval[0];
frame.tf_edx = p->p_retval[1];
frame.tf_eflags &= ~PSL_C;
break;
1993-06-12 14:58:17 +00:00
case ERESTART:
/*
Mega-commit for Linux emulator update.. This has been stress tested under netscape-2.0 for Linux running all the Java stuff. The scrollbars are now working, at least on my machine. (whew! :-) I'm uncomfortable with the size of this commit, but it's too inter-dependant to easily seperate out. The main changes: COMPAT_LINUX is *GONE*. Most of the code has been moved out of the i386 machine dependent section into the linux emulator itself. The int 0x80 syscall code was almost identical to the lcall 7,0 code and a minor tweak allows them to both be used with the same C code. All kernels can now just modload the lkm and it'll DTRT without having to rebuild the kernel first. Like IBCS2, you can statically compile it in with "options LINUX". A pile of new syscalls implemented, including getdents(), llseek(), readv(), writev(), msync(), personality(). The Linux-ELF libraries want to use some of these. linux_select() now obeys Linux semantics, ie: returns the time remaining of the timeout value rather than leaving it the original value. Quite a few bugs removed, including incorrect arguments being used in syscalls.. eg: mixups between passing the sigset as an int, vs passing it as a pointer and doing a copyin(), missing return values, unhandled cases, SIOC* ioctls, etc. The build for the code has changed. i386/conf/files now knows how to build linux_genassym and generate linux_assym.h on the fly. Supporting changes elsewhere in the kernel: The user-mode signal trampoline has moved from the U area to immediately below the top of the stack (below PS_STRINGS). This allows the different binary emulations to have their own signal trampoline code (which gets rid of the hardwired syscall 103 (sigreturn on BSD, syslog on Linux)) and so that the emulator can provide the exact "struct sigcontext *" argument to the program's signal handlers. The sigstack's "ss_flags" now uses SS_DISABLE and SS_ONSTACK flags, which have the same values as the re-used SA_DISABLE and SA_ONSTACK which are intended for sigaction only. This enables the support of a SA_RESETHAND flag to sigaction to implement the gross SYSV and Linux SA_ONESHOT signal semantics where the signal handler is reset when it's triggered. makesyscalls.sh no longer appends the struct sysentvec on the end of the generated init_sysent.c code. It's a lot saner to have it in a seperate file rather than trying to update the structure inside the awk script. :-) At exec time, the dozen bytes or so of signal trampoline code are copied to the top of the user's stack, rather than obtaining the trampoline code the old way by getting a clone of the parent's user area. This allows Linux and native binaries to freely exec each other without getting trampolines mixed up.
1996-03-02 19:38:20 +00:00
* Reconstruct pc, assuming lcall $X,y is 7 bytes,
* int 0x80 is 2 bytes. We saved this in tf_err.
*/
Mega-commit for Linux emulator update.. This has been stress tested under netscape-2.0 for Linux running all the Java stuff. The scrollbars are now working, at least on my machine. (whew! :-) I'm uncomfortable with the size of this commit, but it's too inter-dependant to easily seperate out. The main changes: COMPAT_LINUX is *GONE*. Most of the code has been moved out of the i386 machine dependent section into the linux emulator itself. The int 0x80 syscall code was almost identical to the lcall 7,0 code and a minor tweak allows them to both be used with the same C code. All kernels can now just modload the lkm and it'll DTRT without having to rebuild the kernel first. Like IBCS2, you can statically compile it in with "options LINUX". A pile of new syscalls implemented, including getdents(), llseek(), readv(), writev(), msync(), personality(). The Linux-ELF libraries want to use some of these. linux_select() now obeys Linux semantics, ie: returns the time remaining of the timeout value rather than leaving it the original value. Quite a few bugs removed, including incorrect arguments being used in syscalls.. eg: mixups between passing the sigset as an int, vs passing it as a pointer and doing a copyin(), missing return values, unhandled cases, SIOC* ioctls, etc. The build for the code has changed. i386/conf/files now knows how to build linux_genassym and generate linux_assym.h on the fly. Supporting changes elsewhere in the kernel: The user-mode signal trampoline has moved from the U area to immediately below the top of the stack (below PS_STRINGS). This allows the different binary emulations to have their own signal trampoline code (which gets rid of the hardwired syscall 103 (sigreturn on BSD, syslog on Linux)) and so that the emulator can provide the exact "struct sigcontext *" argument to the program's signal handlers. The sigstack's "ss_flags" now uses SS_DISABLE and SS_ONSTACK flags, which have the same values as the re-used SA_DISABLE and SA_ONSTACK which are intended for sigaction only. This enables the support of a SA_RESETHAND flag to sigaction to implement the gross SYSV and Linux SA_ONESHOT signal semantics where the signal handler is reset when it's triggered. makesyscalls.sh no longer appends the struct sysentvec on the end of the generated init_sysent.c code. It's a lot saner to have it in a seperate file rather than trying to update the structure inside the awk script. :-) At exec time, the dozen bytes or so of signal trampoline code are copied to the top of the user's stack, rather than obtaining the trampoline code the old way by getting a clone of the parent's user area. This allows Linux and native binaries to freely exec each other without getting trampolines mixed up.
1996-03-02 19:38:20 +00:00
frame.tf_eip -= frame.tf_err;
break;
case EJUSTRETURN:
break;
default:
bad:
if (p->p_sysent->sv_errsize)
if (error >= p->p_sysent->sv_errsize)
error = -1; /* XXX */
1995-05-30 08:16:23 +00:00
else
error = p->p_sysent->sv_errtbl[error];
frame.tf_eax = error;
frame.tf_eflags |= PSL_C;
break;
1993-06-12 14:58:17 +00:00
}
if ((frame.tf_eflags & PSL_T) && !(frame.tf_eflags & PSL_VM)) {
/* Traced syscall. */
frame.tf_eflags &= ~PSL_T;
trapsignal(p, SIGTRAP, 0);
}
userret(p, &frame, sticks);
1993-06-12 14:58:17 +00:00
#ifdef KTRACE
if (KTRPOINT(p, KTR_SYSRET))
ktrsysret(p->p_tracep, code, error, p->p_retval[0]);
1993-06-12 14:58:17 +00:00
#endif
/*
* This works because errno is findable through the
* register set. If we ever support an emulation where this
* is not the case, this code will need to be revisited.
*/
STOPEVENT(p, S_SCX, code);
1993-06-12 14:58:17 +00:00
}
The biggie: Get rid of the UPAGES from the top of the per-process address space. (!) Have each process use the kernel stack and pcb in the kvm space. Since the stacks are at a different address, we cannot copy the stack at fork() and allow the child to return up through the function call tree to return to user mode - create a new execution context and have the new process begin executing from cpu_switch() and go to user mode directly. In theory this should speed up fork a bit. Context switch the tss_esp0 pointer in the common tss. This is a lot simpler since than swithching the gdt[GPROC0_SEL].sd.sd_base pointer to each process's tss since the esp0 pointer is a 32 bit pointer, and the sd_base setting is split into three different bit sections at non-aligned boundaries and requires a lot of twiddling to reset. The 8K of memory at the top of the process space is now empty, and unmapped (and unmappable, it's higher than VM_MAXUSER_ADDRESS). Simplity the pmap code to manage process contexts, we no longer have to double map the UPAGES, this simplifies and should measuably speed up fork(). The following parts came from John Dyson: Set PG_G on the UPAGES that are now in kernel context, and invalidate them when swapping them out. Move the upages object (upobj) from the vmspace to the proc structure. Now that the UPAGES (pcb and kernel stack) are out of user space, make rfork(..RFMEM..) do what was intended by sharing the vmspace entirely via reference counting rather than simply inheriting the mappings.
1997-04-07 07:16:06 +00:00
/*
* Simplified back end of syscall(), used when returning from fork()
* directly into user mode.
*/
void
fork_return(p, frame)
struct proc *p;
struct trapframe frame;
{
frame.tf_eax = 0; /* Child returns zero */
frame.tf_eflags &= ~PSL_C; /* success */
frame.tf_edx = 1;
The biggie: Get rid of the UPAGES from the top of the per-process address space. (!) Have each process use the kernel stack and pcb in the kvm space. Since the stacks are at a different address, we cannot copy the stack at fork() and allow the child to return up through the function call tree to return to user mode - create a new execution context and have the new process begin executing from cpu_switch() and go to user mode directly. In theory this should speed up fork a bit. Context switch the tss_esp0 pointer in the common tss. This is a lot simpler since than swithching the gdt[GPROC0_SEL].sd.sd_base pointer to each process's tss since the esp0 pointer is a 32 bit pointer, and the sd_base setting is split into three different bit sections at non-aligned boundaries and requires a lot of twiddling to reset. The 8K of memory at the top of the process space is now empty, and unmapped (and unmappable, it's higher than VM_MAXUSER_ADDRESS). Simplity the pmap code to manage process contexts, we no longer have to double map the UPAGES, this simplifies and should measuably speed up fork(). The following parts came from John Dyson: Set PG_G on the UPAGES that are now in kernel context, and invalidate them when swapping them out. Move the upages object (upobj) from the vmspace to the proc structure. Now that the UPAGES (pcb and kernel stack) are out of user space, make rfork(..RFMEM..) do what was intended by sharing the vmspace entirely via reference counting rather than simply inheriting the mappings.
1997-04-07 07:16:06 +00:00
userret(p, &frame, 0);
#ifdef KTRACE
if (KTRPOINT(p, KTR_SYSRET))
ktrsysret(p->p_tracep, SYS_fork, 0, 0);
#endif
}