linux/arch/x86/kernel/ptrace.c

734 lines
19 KiB
C
Raw Normal View History

/* By Ross Biro 1/23/92 */
/*
* Pentium III FXSR, SSE support
* Gareth Hughes <gareth@valinux.com>, May 2000
*/
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/mm.h>
#include <linux/smp.h>
#include <linux/errno.h>
#include <linux/ptrace.h>
#include <linux/user.h>
#include <linux/security.h>
#include <linux/audit.h>
#include <linux/seccomp.h>
#include <linux/signal.h>
#include <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/system.h>
#include <asm/processor.h>
#include <asm/i387.h>
#include <asm/debugreg.h>
#include <asm/ldt.h>
#include <asm/desc.h>
#include <asm/prctl.h>
#include <asm/proto.h>
/*
* does not yet catch signals sent when the child dies.
* in exit.c or in signal.c.
*/
/*
* Determines which flags the user has access to [1 = access, 0 = no access].
*/
#define FLAG_MASK_32 ((unsigned long) \
(X86_EFLAGS_CF | X86_EFLAGS_PF | \
X86_EFLAGS_AF | X86_EFLAGS_ZF | \
X86_EFLAGS_SF | X86_EFLAGS_TF | \
X86_EFLAGS_DF | X86_EFLAGS_OF | \
X86_EFLAGS_RF | X86_EFLAGS_AC))
/*
* Determines whether a value may be installed in a segment register.
*/
static inline bool invalid_selector(u16 value)
{
return unlikely(value != 0 && (value & SEGMENT_RPL_MASK) != USER_RPL);
}
#ifdef CONFIG_X86_32
#define FLAG_MASK FLAG_MASK_32
static long *pt_regs_access(struct pt_regs *regs, unsigned long regno)
{
BUILD_BUG_ON(offsetof(struct pt_regs, bx) != 0);
regno >>= 2;
if (regno > FS)
--regno;
return &regs->bx + regno;
}
static u16 get_segment_reg(struct task_struct *task, unsigned long offset)
{
/*
* Returning the value truncates it to 16 bits.
*/
unsigned int retval;
if (offset != offsetof(struct user_regs_struct, gs))
retval = *pt_regs_access(task_pt_regs(task), offset);
else {
retval = task->thread.gs;
if (task == current)
savesegment(gs, retval);
}
return retval;
}
static int set_segment_reg(struct task_struct *task,
unsigned long offset, u16 value)
{
/*
* The value argument was already truncated to 16 bits.
*/
if (invalid_selector(value))
return -EIO;
if (offset != offsetof(struct user_regs_struct, gs))
*pt_regs_access(task_pt_regs(task), offset) = value;
else {
task->thread.gs = value;
if (task == current)
/*
* The user-mode %gs is not affected by
* kernel entry, so we must update the CPU.
*/
loadsegment(gs, value);
}
return 0;
}
static unsigned long debugreg_addr_limit(struct task_struct *task)
{
return TASK_SIZE - 3;
}
#else /* CONFIG_X86_64 */
#define FLAG_MASK (FLAG_MASK_32 | X86_EFLAGS_NT)
static unsigned long *pt_regs_access(struct pt_regs *regs, unsigned long offset)
{
BUILD_BUG_ON(offsetof(struct pt_regs, r15) != 0);
return &regs->r15 + (offset / sizeof(regs->r15));
}
static u16 get_segment_reg(struct task_struct *task, unsigned long offset)
{
/*
* Returning the value truncates it to 16 bits.
*/
unsigned int seg;
switch (offset) {
case offsetof(struct user_regs_struct, fs):
if (task == current) {
/* Older gas can't assemble movq %?s,%r?? */
asm("movl %%fs,%0" : "=r" (seg));
return seg;
}
return task->thread.fsindex;
case offsetof(struct user_regs_struct, gs):
if (task == current) {
asm("movl %%gs,%0" : "=r" (seg));
return seg;
}
return task->thread.gsindex;
case offsetof(struct user_regs_struct, ds):
if (task == current) {
asm("movl %%ds,%0" : "=r" (seg));
return seg;
}
return task->thread.ds;
case offsetof(struct user_regs_struct, es):
if (task == current) {
asm("movl %%es,%0" : "=r" (seg));
return seg;
}
return task->thread.es;
case offsetof(struct user_regs_struct, cs):
case offsetof(struct user_regs_struct, ss):
break;
}
return *pt_regs_access(task_pt_regs(task), offset);
}
static int set_segment_reg(struct task_struct *task,
unsigned long offset, u16 value)
{
/*
* The value argument was already truncated to 16 bits.
*/
if (invalid_selector(value))
return -EIO;
switch (offset) {
case offsetof(struct user_regs_struct,fs):
/*
* If this is setting fs as for normal 64-bit use but
* setting fs_base has implicitly changed it, leave it.
*/
if ((value == FS_TLS_SEL && task->thread.fsindex == 0 &&
task->thread.fs != 0) ||
(value == 0 && task->thread.fsindex == FS_TLS_SEL &&
task->thread.fs == 0))
break;
task->thread.fsindex = value;
if (task == current)
loadsegment(fs, task->thread.fsindex);
break;
case offsetof(struct user_regs_struct,gs):
/*
* If this is setting gs as for normal 64-bit use but
* setting gs_base has implicitly changed it, leave it.
*/
if ((value == GS_TLS_SEL && task->thread.gsindex == 0 &&
task->thread.gs != 0) ||
(value == 0 && task->thread.gsindex == GS_TLS_SEL &&
task->thread.gs == 0))
break;
task->thread.gsindex = value;
if (task == current)
load_gs_index(task->thread.gsindex);
break;
case offsetof(struct user_regs_struct,ds):
task->thread.ds = value;
if (task == current)
loadsegment(ds, task->thread.ds);
break;
case offsetof(struct user_regs_struct,es):
task->thread.es = value;
if (task == current)
loadsegment(es, task->thread.es);
break;
/*
* Can't actually change these in 64-bit mode.
*/
case offsetof(struct user_regs_struct,cs):
#ifdef CONFIG_IA32_EMULATION
if (test_tsk_thread_flag(task, TIF_IA32))
task_pt_regs(task)->cs = value;
break;
#endif
case offsetof(struct user_regs_struct,ss):
#ifdef CONFIG_IA32_EMULATION
if (test_tsk_thread_flag(task, TIF_IA32))
task_pt_regs(task)->ss = value;
break;
#endif
}
return 0;
}
static unsigned long debugreg_addr_limit(struct task_struct *task)
{
#ifdef CONFIG_IA32_EMULATION
if (test_tsk_thread_flag(task, TIF_IA32))
return IA32_PAGE_OFFSET - 3;
#endif
return TASK_SIZE64 - 7;
}
#endif /* CONFIG_X86_32 */
static unsigned long get_flags(struct task_struct *task)
{
unsigned long retval = task_pt_regs(task)->flags;
/*
* If the debugger set TF, hide it from the readout.
*/
if (test_tsk_thread_flag(task, TIF_FORCED_TF))
retval &= ~X86_EFLAGS_TF;
return retval;
}
static int set_flags(struct task_struct *task, unsigned long value)
{
struct pt_regs *regs = task_pt_regs(task);
/*
* If the user value contains TF, mark that
* it was not "us" (the debugger) that set it.
* If not, make sure it stays set if we had.
*/
if (value & X86_EFLAGS_TF)
clear_tsk_thread_flag(task, TIF_FORCED_TF);
else if (test_tsk_thread_flag(task, TIF_FORCED_TF))
value |= X86_EFLAGS_TF;
regs->flags = (regs->flags & ~FLAG_MASK) | (value & FLAG_MASK);
return 0;
}
static int putreg(struct task_struct *child,
unsigned long offset, unsigned long value)
{
switch (offset) {
case offsetof(struct user_regs_struct, cs):
case offsetof(struct user_regs_struct, ds):
case offsetof(struct user_regs_struct, es):
case offsetof(struct user_regs_struct, fs):
case offsetof(struct user_regs_struct, gs):
case offsetof(struct user_regs_struct, ss):
return set_segment_reg(child, offset, value);
case offsetof(struct user_regs_struct, flags):
return set_flags(child, value);
#ifdef CONFIG_X86_64
case offsetof(struct user_regs_struct,fs_base):
if (value >= TASK_SIZE_OF(child))
return -EIO;
/*
* When changing the segment base, use do_arch_prctl
* to set either thread.fs or thread.fsindex and the
* corresponding GDT slot.
*/
if (child->thread.fs != value)
return do_arch_prctl(child, ARCH_SET_FS, value);
return 0;
case offsetof(struct user_regs_struct,gs_base):
/*
* Exactly the same here as the %fs handling above.
*/
if (value >= TASK_SIZE_OF(child))
return -EIO;
if (child->thread.gs != value)
return do_arch_prctl(child, ARCH_SET_GS, value);
return 0;
#endif
}
*pt_regs_access(task_pt_regs(child), offset) = value;
return 0;
}
static unsigned long getreg(struct task_struct *task, unsigned long offset)
{
switch (offset) {
case offsetof(struct user_regs_struct, cs):
case offsetof(struct user_regs_struct, ds):
case offsetof(struct user_regs_struct, es):
case offsetof(struct user_regs_struct, fs):
case offsetof(struct user_regs_struct, gs):
case offsetof(struct user_regs_struct, ss):
return get_segment_reg(task, offset);
case offsetof(struct user_regs_struct, flags):
return get_flags(task);
#ifdef CONFIG_X86_64
case offsetof(struct user_regs_struct, fs_base): {
/*
* do_arch_prctl may have used a GDT slot instead of
* the MSR. To userland, it appears the same either
* way, except the %fs segment selector might not be 0.
*/
unsigned int seg = task->thread.fsindex;
if (task->thread.fs != 0)
return task->thread.fs;
if (task == current)
asm("movl %%fs,%0" : "=r" (seg));
if (seg != FS_TLS_SEL)
return 0;
return get_desc_base(&task->thread.tls_array[FS_TLS]);
}
case offsetof(struct user_regs_struct, gs_base): {
/*
* Exactly the same here as the %fs handling above.
*/
unsigned int seg = task->thread.gsindex;
if (task->thread.gs != 0)
return task->thread.gs;
if (task == current)
asm("movl %%gs,%0" : "=r" (seg));
if (seg != GS_TLS_SEL)
return 0;
return get_desc_base(&task->thread.tls_array[GS_TLS]);
}
#endif
}
return *pt_regs_access(task_pt_regs(task), offset);
}
/*
* This function is trivial and will be inlined by the compiler.
* Having it separates the implementation details of debug
* registers from the interface details of ptrace.
*/
static unsigned long ptrace_get_debugreg(struct task_struct *child, int n)
{
switch (n) {
case 0: return child->thread.debugreg0;
case 1: return child->thread.debugreg1;
case 2: return child->thread.debugreg2;
case 3: return child->thread.debugreg3;
case 6: return child->thread.debugreg6;
case 7: return child->thread.debugreg7;
}
return 0;
}
static int ptrace_set_debugreg(struct task_struct *child,
int n, unsigned long data)
{
int i;
if (unlikely(n == 4 || n == 5))
return -EIO;
if (n < 4 && unlikely(data >= debugreg_addr_limit(child)))
return -EIO;
switch (n) {
case 0: child->thread.debugreg0 = data; break;
case 1: child->thread.debugreg1 = data; break;
case 2: child->thread.debugreg2 = data; break;
case 3: child->thread.debugreg3 = data; break;
case 6:
if ((data & ~0xffffffffUL) != 0)
return -EIO;
child->thread.debugreg6 = data;
break;
case 7:
/*
* Sanity-check data. Take one half-byte at once with
* check = (val >> (16 + 4*i)) & 0xf. It contains the
* R/Wi and LENi bits; bits 0 and 1 are R/Wi, and bits
* 2 and 3 are LENi. Given a list of invalid values,
* we do mask |= 1 << invalid_value, so that
* (mask >> check) & 1 is a correct test for invalid
* values.
*
* R/Wi contains the type of the breakpoint /
* watchpoint, LENi contains the length of the watched
* data in the watchpoint case.
*
* The invalid values are:
* - LENi == 0x10 (undefined), so mask |= 0x0f00. [32-bit]
* - R/Wi == 0x10 (break on I/O reads or writes), so
* mask |= 0x4444.
* - R/Wi == 0x00 && LENi != 0x00, so we have mask |=
* 0x1110.
*
* Finally, mask = 0x0f00 | 0x4444 | 0x1110 == 0x5f54.
*
* See the Intel Manual "System Programming Guide",
* 15.2.4
*
* Note that LENi == 0x10 is defined on x86_64 in long
* mode (i.e. even for 32-bit userspace software, but
* 64-bit kernel), so the x86_64 mask value is 0x5454.
* See the AMD manual no. 24593 (AMD64 System Programming)
*/
#ifdef CONFIG_X86_32
#define DR7_MASK 0x5f54
#else
#define DR7_MASK 0x5554
#endif
data &= ~DR_CONTROL_RESERVED;
for (i = 0; i < 4; i++)
if ((DR7_MASK >> ((data >> (16 + 4*i)) & 0xf)) & 1)
return -EIO;
child->thread.debugreg7 = data;
if (data)
set_tsk_thread_flag(child, TIF_DEBUG);
else
clear_tsk_thread_flag(child, TIF_DEBUG);
break;
}
return 0;
}
/*
* Called by kernel/ptrace.c when detaching..
*
* Make sure the single step bit is not set.
*/
void ptrace_disable(struct task_struct *child)
{
user_disable_single_step(child);
#ifdef TIF_SYSCALL_EMU
clear_tsk_thread_flag(child, TIF_SYSCALL_EMU);
#endif
}
long arch_ptrace(struct task_struct *child, long request, long addr, long data)
{
int i, ret;
unsigned long __user *datap = (unsigned long __user *)data;
switch (request) {
/* when I and D space are separate, these will need to be fixed. */
case PTRACE_PEEKTEXT: /* read word at location addr. */
case PTRACE_PEEKDATA:
ret = generic_ptrace_peekdata(child, addr, data);
break;
/* read the word at location addr in the USER area. */
case PTRACE_PEEKUSR: {
unsigned long tmp;
ret = -EIO;
if ((addr & (sizeof(data) - 1)) || addr < 0 ||
addr >= sizeof(struct user))
break;
tmp = 0; /* Default return condition */
if (addr < sizeof(struct user_regs_struct))
tmp = getreg(child, addr);
else if (addr >= offsetof(struct user, u_debugreg[0]) &&
addr <= offsetof(struct user, u_debugreg[7])) {
addr -= offsetof(struct user, u_debugreg[0]);
tmp = ptrace_get_debugreg(child, addr / sizeof(data));
}
ret = put_user(tmp, datap);
break;
}
/* when I and D space are separate, this will have to be fixed. */
case PTRACE_POKETEXT: /* write the word at location addr. */
case PTRACE_POKEDATA:
ret = generic_ptrace_pokedata(child, addr, data);
break;
case PTRACE_POKEUSR: /* write the word at location addr in the USER area */
ret = -EIO;
if ((addr & (sizeof(data) - 1)) || addr < 0 ||
addr >= sizeof(struct user))
break;
if (addr < sizeof(struct user_regs_struct))
ret = putreg(child, addr, data);
else if (addr >= offsetof(struct user, u_debugreg[0]) &&
addr <= offsetof(struct user, u_debugreg[7])) {
addr -= offsetof(struct user, u_debugreg[0]);
ret = ptrace_set_debugreg(child,
addr / sizeof(data), data);
}
break;
case PTRACE_GETREGS: { /* Get all gp regs from the child. */
if (!access_ok(VERIFY_WRITE, datap, sizeof(struct user_regs_struct))) {
ret = -EIO;
break;
}
for (i = 0; i < sizeof(struct user_regs_struct); i += sizeof(long)) {
__put_user(getreg(child, i), datap);
datap++;
}
ret = 0;
break;
}
case PTRACE_SETREGS: { /* Set all gp regs in the child. */
unsigned long tmp;
if (!access_ok(VERIFY_READ, datap, sizeof(struct user_regs_struct))) {
ret = -EIO;
break;
}
for (i = 0; i < sizeof(struct user_regs_struct); i += sizeof(long)) {
__get_user(tmp, datap);
putreg(child, i, tmp);
datap++;
}
ret = 0;
break;
}
case PTRACE_GETFPREGS: { /* Get the child FPU state. */
if (!access_ok(VERIFY_WRITE, datap,
sizeof(struct user_i387_struct))) {
ret = -EIO;
break;
}
ret = 0;
if (!tsk_used_math(child))
init_fpu(child);
get_fpregs((struct user_i387_struct __user *)data, child);
break;
}
case PTRACE_SETFPREGS: { /* Set the child FPU state. */
if (!access_ok(VERIFY_READ, datap,
sizeof(struct user_i387_struct))) {
ret = -EIO;
break;
}
set_stopped_child_used_math(child);
set_fpregs(child, (struct user_i387_struct __user *)data);
ret = 0;
break;
}
#ifdef CONFIG_X86_32
case PTRACE_GETFPXREGS: { /* Get the child extended FPU state. */
if (!access_ok(VERIFY_WRITE, datap,
sizeof(struct user_fxsr_struct))) {
ret = -EIO;
break;
}
if (!tsk_used_math(child))
init_fpu(child);
ret = get_fpxregs((struct user_fxsr_struct __user *)data, child);
break;
}
case PTRACE_SETFPXREGS: { /* Set the child extended FPU state. */
if (!access_ok(VERIFY_READ, datap,
sizeof(struct user_fxsr_struct))) {
ret = -EIO;
break;
}
set_stopped_child_used_math(child);
ret = set_fpxregs(child, (struct user_fxsr_struct __user *)data);
break;
}
#endif
#if defined CONFIG_X86_32 || defined CONFIG_IA32_EMULATION
case PTRACE_GET_THREAD_AREA:
if (addr < 0)
return -EIO;
ret = do_get_thread_area(child, addr,
(struct user_desc __user *) data);
break;
case PTRACE_SET_THREAD_AREA:
if (addr < 0)
return -EIO;
ret = do_set_thread_area(child, addr,
(struct user_desc __user *) data, 0);
break;
#endif
#ifdef CONFIG_X86_64
/* normal 64bit interface to access TLS data.
Works just like arch_prctl, except that the arguments
are reversed. */
case PTRACE_ARCH_PRCTL:
ret = do_arch_prctl(child, data, addr);
break;
#endif
default:
ret = ptrace_request(child, request, addr, data);
break;
}
return ret;
}
void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs, int error_code)
{
struct siginfo info;
tsk->thread.trap_no = 1;
tsk->thread.error_code = error_code;
memset(&info, 0, sizeof(info));
info.si_signo = SIGTRAP;
info.si_code = TRAP_BRKPT;
/* User-mode ip? */
info.si_addr = user_mode_vm(regs) ? (void __user *) regs->ip : NULL;
/* Send us the fake SIGTRAP */
force_sig_info(SIGTRAP, &info, tsk);
}
/* notification of system call entry/exit
* - triggered by current->work.syscall_trace
*/
__attribute__((regparm(3)))
[PATCH] UML Support - Ptrace: adds the host SYSEMU support, for UML and general usage Jeff Dike <jdike@addtoit.com>, Paolo 'Blaisorblade' Giarrusso <blaisorblade_spam@yahoo.it>, Bodo Stroesser <bstroesser@fujitsu-siemens.com> Adds a new ptrace(2) mode, called PTRACE_SYSEMU, resembling PTRACE_SYSCALL except that the kernel does not execute the requested syscall; this is useful to improve performance for virtual environments, like UML, which want to run the syscall on their own. In fact, using PTRACE_SYSCALL means stopping child execution twice, on entry and on exit, and each time you also have two context switches; with SYSEMU you avoid the 2nd stop and so save two context switches per syscall. Also, some architectures don't have support in the host for changing the syscall number via ptrace(), which is currently needed to skip syscall execution (UML turns any syscall into getpid() to avoid it being executed on the host). Fixing that is hard, while SYSEMU is easier to implement. * This version of the patch includes some suggestions of Jeff Dike to avoid adding any instructions to the syscall fast path, plus some other little changes, by myself, to make it work even when the syscall is executed with SYSENTER (but I'm unsure about them). It has been widely tested for quite a lot of time. * Various fixed were included to handle the various switches between various states, i.e. when for instance a syscall entry is traced with one of PT_SYSCALL / _SYSEMU / _SINGLESTEP and another one is used on exit. Basically, this is done by remembering which one of them was used even after the call to ptrace_notify(). * We're combining TIF_SYSCALL_EMU with TIF_SYSCALL_TRACE or TIF_SINGLESTEP to make do_syscall_trace() notice that the current syscall was started with SYSEMU on entry, so that no notification ought to be done in the exit path; this is a bit of a hack, so this problem is solved in another way in next patches. * Also, the effects of the patch: "Ptrace - i386: fix Syscall Audit interaction with singlestep" are cancelled; they are restored back in the last patch of this series. Detailed descriptions of the patches doing this kind of processing follow (but I've already summed everything up). * Fix behaviour when changing interception kind #1. In do_syscall_trace(), we check the status of the TIF_SYSCALL_EMU flag only after doing the debugger notification; but the debugger might have changed the status of this flag because he continued execution with PTRACE_SYSCALL, so this is wrong. This patch fixes it by saving the flag status before calling ptrace_notify(). * Fix behaviour when changing interception kind #2: avoid intercepting syscall on return when using SYSCALL again. A guest process switching from using PTRACE_SYSEMU to PTRACE_SYSCALL crashes. The problem is in arch/i386/kernel/entry.S. The current SYSEMU patch inhibits the syscall-handler to be called, but does not prevent do_syscall_trace() to be called after this for syscall completion interception. The appended patch fixes this. It reuses the flag TIF_SYSCALL_EMU to remember "we come from PTRACE_SYSEMU and now are in PTRACE_SYSCALL", since the flag is unused in the depicted situation. * Fix behaviour when changing interception kind #3: avoid intercepting syscall on return when using SINGLESTEP. When testing 2.6.9 and the skas3.v6 patch, with my latest patch and had problems with singlestepping on UML in SKAS with SYSEMU. It looped receiving SIGTRAPs without moving forward. EIP of the traced process was the same for all SIGTRAPs. What's missing is to handle switching from PTRACE_SYSCALL_EMU to PTRACE_SINGLESTEP in a way very similar to what is done for the change from PTRACE_SYSCALL_EMU to PTRACE_SYSCALL_TRACE. I.e., after calling ptrace(PTRACE_SYSEMU), on the return path, the debugger is notified and then wake ups the process; the syscall is executed (or skipped, when do_syscall_trace() returns 0, i.e. when using PTRACE_SYSEMU), and do_syscall_trace() is called again. Since we are on the return path of a SYSEMU'd syscall, if the wake up is performed through ptrace(PTRACE_SYSCALL), we must still avoid notifying the parent of the syscall exit. Now, this behaviour is extended even to resuming with PTRACE_SINGLESTEP. Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-03 22:57:18 +00:00
int do_syscall_trace(struct pt_regs *regs, int entryexit)
{
int is_sysemu = test_thread_flag(TIF_SYSCALL_EMU);
/*
* With TIF_SYSCALL_EMU set we want to ignore TIF_SINGLESTEP for syscall
* interception
*/
int is_singlestep = !is_sysemu && test_thread_flag(TIF_SINGLESTEP);
int ret = 0;
/* do the secure computing check first */
if (!entryexit)
secure_computing(regs->orig_ax);
if (unlikely(current->audit_context)) {
if (entryexit)
audit_syscall_exit(AUDITSC_RESULT(regs->ax),
regs->ax);
/* Debug traps, when using PTRACE_SINGLESTEP, must be sent only
* on the syscall exit path. Normally, when TIF_SYSCALL_AUDIT is
* not used, entry.S will call us only on syscall exit, not
* entry; so when TIF_SYSCALL_AUDIT is used we must avoid
* calling send_sigtrap() on syscall entry.
*
* Note that when PTRACE_SYSEMU_SINGLESTEP is used,
* is_singlestep is false, despite his name, so we will still do
* the correct thing.
*/
else if (is_singlestep)
goto out;
}
if (!(current->ptrace & PT_PTRACED))
goto out;
/* If a process stops on the 1st tracepoint with SYSCALL_TRACE
* and then is resumed with SYSEMU_SINGLESTEP, it will come in
* here. We have to check this and return */
if (is_sysemu && entryexit)
return 0;
[PATCH] UML Support - Ptrace: adds the host SYSEMU support, for UML and general usage Jeff Dike <jdike@addtoit.com>, Paolo 'Blaisorblade' Giarrusso <blaisorblade_spam@yahoo.it>, Bodo Stroesser <bstroesser@fujitsu-siemens.com> Adds a new ptrace(2) mode, called PTRACE_SYSEMU, resembling PTRACE_SYSCALL except that the kernel does not execute the requested syscall; this is useful to improve performance for virtual environments, like UML, which want to run the syscall on their own. In fact, using PTRACE_SYSCALL means stopping child execution twice, on entry and on exit, and each time you also have two context switches; with SYSEMU you avoid the 2nd stop and so save two context switches per syscall. Also, some architectures don't have support in the host for changing the syscall number via ptrace(), which is currently needed to skip syscall execution (UML turns any syscall into getpid() to avoid it being executed on the host). Fixing that is hard, while SYSEMU is easier to implement. * This version of the patch includes some suggestions of Jeff Dike to avoid adding any instructions to the syscall fast path, plus some other little changes, by myself, to make it work even when the syscall is executed with SYSENTER (but I'm unsure about them). It has been widely tested for quite a lot of time. * Various fixed were included to handle the various switches between various states, i.e. when for instance a syscall entry is traced with one of PT_SYSCALL / _SYSEMU / _SINGLESTEP and another one is used on exit. Basically, this is done by remembering which one of them was used even after the call to ptrace_notify(). * We're combining TIF_SYSCALL_EMU with TIF_SYSCALL_TRACE or TIF_SINGLESTEP to make do_syscall_trace() notice that the current syscall was started with SYSEMU on entry, so that no notification ought to be done in the exit path; this is a bit of a hack, so this problem is solved in another way in next patches. * Also, the effects of the patch: "Ptrace - i386: fix Syscall Audit interaction with singlestep" are cancelled; they are restored back in the last patch of this series. Detailed descriptions of the patches doing this kind of processing follow (but I've already summed everything up). * Fix behaviour when changing interception kind #1. In do_syscall_trace(), we check the status of the TIF_SYSCALL_EMU flag only after doing the debugger notification; but the debugger might have changed the status of this flag because he continued execution with PTRACE_SYSCALL, so this is wrong. This patch fixes it by saving the flag status before calling ptrace_notify(). * Fix behaviour when changing interception kind #2: avoid intercepting syscall on return when using SYSCALL again. A guest process switching from using PTRACE_SYSEMU to PTRACE_SYSCALL crashes. The problem is in arch/i386/kernel/entry.S. The current SYSEMU patch inhibits the syscall-handler to be called, but does not prevent do_syscall_trace() to be called after this for syscall completion interception. The appended patch fixes this. It reuses the flag TIF_SYSCALL_EMU to remember "we come from PTRACE_SYSEMU and now are in PTRACE_SYSCALL", since the flag is unused in the depicted situation. * Fix behaviour when changing interception kind #3: avoid intercepting syscall on return when using SINGLESTEP. When testing 2.6.9 and the skas3.v6 patch, with my latest patch and had problems with singlestepping on UML in SKAS with SYSEMU. It looped receiving SIGTRAPs without moving forward. EIP of the traced process was the same for all SIGTRAPs. What's missing is to handle switching from PTRACE_SYSCALL_EMU to PTRACE_SINGLESTEP in a way very similar to what is done for the change from PTRACE_SYSCALL_EMU to PTRACE_SYSCALL_TRACE. I.e., after calling ptrace(PTRACE_SYSEMU), on the return path, the debugger is notified and then wake ups the process; the syscall is executed (or skipped, when do_syscall_trace() returns 0, i.e. when using PTRACE_SYSEMU), and do_syscall_trace() is called again. Since we are on the return path of a SYSEMU'd syscall, if the wake up is performed through ptrace(PTRACE_SYSCALL), we must still avoid notifying the parent of the syscall exit. Now, this behaviour is extended even to resuming with PTRACE_SINGLESTEP. Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-03 22:57:18 +00:00
/* Fake a debug trap */
if (is_singlestep)
send_sigtrap(current, regs, 0);
if (!test_thread_flag(TIF_SYSCALL_TRACE) && !is_sysemu)
goto out;
/* the 0x80 provides a way for the tracing parent to distinguish
between a syscall stop and SIGTRAP delivery */
[PATCH] UML Support - Ptrace: adds the host SYSEMU support, for UML and general usage Jeff Dike <jdike@addtoit.com>, Paolo 'Blaisorblade' Giarrusso <blaisorblade_spam@yahoo.it>, Bodo Stroesser <bstroesser@fujitsu-siemens.com> Adds a new ptrace(2) mode, called PTRACE_SYSEMU, resembling PTRACE_SYSCALL except that the kernel does not execute the requested syscall; this is useful to improve performance for virtual environments, like UML, which want to run the syscall on their own. In fact, using PTRACE_SYSCALL means stopping child execution twice, on entry and on exit, and each time you also have two context switches; with SYSEMU you avoid the 2nd stop and so save two context switches per syscall. Also, some architectures don't have support in the host for changing the syscall number via ptrace(), which is currently needed to skip syscall execution (UML turns any syscall into getpid() to avoid it being executed on the host). Fixing that is hard, while SYSEMU is easier to implement. * This version of the patch includes some suggestions of Jeff Dike to avoid adding any instructions to the syscall fast path, plus some other little changes, by myself, to make it work even when the syscall is executed with SYSENTER (but I'm unsure about them). It has been widely tested for quite a lot of time. * Various fixed were included to handle the various switches between various states, i.e. when for instance a syscall entry is traced with one of PT_SYSCALL / _SYSEMU / _SINGLESTEP and another one is used on exit. Basically, this is done by remembering which one of them was used even after the call to ptrace_notify(). * We're combining TIF_SYSCALL_EMU with TIF_SYSCALL_TRACE or TIF_SINGLESTEP to make do_syscall_trace() notice that the current syscall was started with SYSEMU on entry, so that no notification ought to be done in the exit path; this is a bit of a hack, so this problem is solved in another way in next patches. * Also, the effects of the patch: "Ptrace - i386: fix Syscall Audit interaction with singlestep" are cancelled; they are restored back in the last patch of this series. Detailed descriptions of the patches doing this kind of processing follow (but I've already summed everything up). * Fix behaviour when changing interception kind #1. In do_syscall_trace(), we check the status of the TIF_SYSCALL_EMU flag only after doing the debugger notification; but the debugger might have changed the status of this flag because he continued execution with PTRACE_SYSCALL, so this is wrong. This patch fixes it by saving the flag status before calling ptrace_notify(). * Fix behaviour when changing interception kind #2: avoid intercepting syscall on return when using SYSCALL again. A guest process switching from using PTRACE_SYSEMU to PTRACE_SYSCALL crashes. The problem is in arch/i386/kernel/entry.S. The current SYSEMU patch inhibits the syscall-handler to be called, but does not prevent do_syscall_trace() to be called after this for syscall completion interception. The appended patch fixes this. It reuses the flag TIF_SYSCALL_EMU to remember "we come from PTRACE_SYSEMU and now are in PTRACE_SYSCALL", since the flag is unused in the depicted situation. * Fix behaviour when changing interception kind #3: avoid intercepting syscall on return when using SINGLESTEP. When testing 2.6.9 and the skas3.v6 patch, with my latest patch and had problems with singlestepping on UML in SKAS with SYSEMU. It looped receiving SIGTRAPs without moving forward. EIP of the traced process was the same for all SIGTRAPs. What's missing is to handle switching from PTRACE_SYSCALL_EMU to PTRACE_SINGLESTEP in a way very similar to what is done for the change from PTRACE_SYSCALL_EMU to PTRACE_SYSCALL_TRACE. I.e., after calling ptrace(PTRACE_SYSEMU), on the return path, the debugger is notified and then wake ups the process; the syscall is executed (or skipped, when do_syscall_trace() returns 0, i.e. when using PTRACE_SYSEMU), and do_syscall_trace() is called again. Since we are on the return path of a SYSEMU'd syscall, if the wake up is performed through ptrace(PTRACE_SYSCALL), we must still avoid notifying the parent of the syscall exit. Now, this behaviour is extended even to resuming with PTRACE_SINGLESTEP. Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-03 22:57:18 +00:00
/* Note that the debugger could change the result of test_thread_flag!*/
ptrace_notify(SIGTRAP | ((current->ptrace & PT_TRACESYSGOOD) ? 0x80:0));
/*
* this isn't the same as continuing with a signal, but it will do
* for normal use. strace only continues with a signal if the
* stopping signal is not SIGTRAP. -brl
*/
if (current->exit_code) {
send_sig(current->exit_code, current, 1);
current->exit_code = 0;
}
[PATCH] UML Support - Ptrace: adds the host SYSEMU support, for UML and general usage Jeff Dike <jdike@addtoit.com>, Paolo 'Blaisorblade' Giarrusso <blaisorblade_spam@yahoo.it>, Bodo Stroesser <bstroesser@fujitsu-siemens.com> Adds a new ptrace(2) mode, called PTRACE_SYSEMU, resembling PTRACE_SYSCALL except that the kernel does not execute the requested syscall; this is useful to improve performance for virtual environments, like UML, which want to run the syscall on their own. In fact, using PTRACE_SYSCALL means stopping child execution twice, on entry and on exit, and each time you also have two context switches; with SYSEMU you avoid the 2nd stop and so save two context switches per syscall. Also, some architectures don't have support in the host for changing the syscall number via ptrace(), which is currently needed to skip syscall execution (UML turns any syscall into getpid() to avoid it being executed on the host). Fixing that is hard, while SYSEMU is easier to implement. * This version of the patch includes some suggestions of Jeff Dike to avoid adding any instructions to the syscall fast path, plus some other little changes, by myself, to make it work even when the syscall is executed with SYSENTER (but I'm unsure about them). It has been widely tested for quite a lot of time. * Various fixed were included to handle the various switches between various states, i.e. when for instance a syscall entry is traced with one of PT_SYSCALL / _SYSEMU / _SINGLESTEP and another one is used on exit. Basically, this is done by remembering which one of them was used even after the call to ptrace_notify(). * We're combining TIF_SYSCALL_EMU with TIF_SYSCALL_TRACE or TIF_SINGLESTEP to make do_syscall_trace() notice that the current syscall was started with SYSEMU on entry, so that no notification ought to be done in the exit path; this is a bit of a hack, so this problem is solved in another way in next patches. * Also, the effects of the patch: "Ptrace - i386: fix Syscall Audit interaction with singlestep" are cancelled; they are restored back in the last patch of this series. Detailed descriptions of the patches doing this kind of processing follow (but I've already summed everything up). * Fix behaviour when changing interception kind #1. In do_syscall_trace(), we check the status of the TIF_SYSCALL_EMU flag only after doing the debugger notification; but the debugger might have changed the status of this flag because he continued execution with PTRACE_SYSCALL, so this is wrong. This patch fixes it by saving the flag status before calling ptrace_notify(). * Fix behaviour when changing interception kind #2: avoid intercepting syscall on return when using SYSCALL again. A guest process switching from using PTRACE_SYSEMU to PTRACE_SYSCALL crashes. The problem is in arch/i386/kernel/entry.S. The current SYSEMU patch inhibits the syscall-handler to be called, but does not prevent do_syscall_trace() to be called after this for syscall completion interception. The appended patch fixes this. It reuses the flag TIF_SYSCALL_EMU to remember "we come from PTRACE_SYSEMU and now are in PTRACE_SYSCALL", since the flag is unused in the depicted situation. * Fix behaviour when changing interception kind #3: avoid intercepting syscall on return when using SINGLESTEP. When testing 2.6.9 and the skas3.v6 patch, with my latest patch and had problems with singlestepping on UML in SKAS with SYSEMU. It looped receiving SIGTRAPs without moving forward. EIP of the traced process was the same for all SIGTRAPs. What's missing is to handle switching from PTRACE_SYSCALL_EMU to PTRACE_SINGLESTEP in a way very similar to what is done for the change from PTRACE_SYSCALL_EMU to PTRACE_SYSCALL_TRACE. I.e., after calling ptrace(PTRACE_SYSEMU), on the return path, the debugger is notified and then wake ups the process; the syscall is executed (or skipped, when do_syscall_trace() returns 0, i.e. when using PTRACE_SYSEMU), and do_syscall_trace() is called again. Since we are on the return path of a SYSEMU'd syscall, if the wake up is performed through ptrace(PTRACE_SYSCALL), we must still avoid notifying the parent of the syscall exit. Now, this behaviour is extended even to resuming with PTRACE_SINGLESTEP. Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-03 22:57:18 +00:00
ret = is_sysemu;
out:
if (unlikely(current->audit_context) && !entryexit)
audit_syscall_entry(AUDIT_ARCH_I386, regs->orig_ax,
regs->bx, regs->cx, regs->dx, regs->si);
if (ret == 0)
return 0;
regs->orig_ax = -1; /* force skip of syscall restarting */
if (unlikely(current->audit_context))
audit_syscall_exit(AUDITSC_RESULT(regs->ax), regs->ax);
return 1;
}