process.c

linux-2.6.15.6

	if (t->task->mm) {
		unsigned long pgd_cache = 0UL;
		if (test_thread_flag(TIF_32BIT)) {
			struct mm_struct *mm = t->task->mm;
			pgd_t *pgd0 = &mm->pgd[0];
			pud_t *pud0 = pud_offset(pgd0, 0);

			if (pud_none(*pud0)) {
				pmd_t *page = pmd_alloc_one(mm, 0);
				pud_set(pud0, page);
			}
			pgd_cache = get_pgd_cache(pgd0);
		}
		__asm__ __volatile__("stxa %0, [%1] %2\n\t"
				     "membar #Sync"
				     : /* no outputs */
				     : "r" (pgd_cache),
				     "r" (TSB_REG),
				     "i" (ASI_DMMU));
	}
	set_thread_wsaved(0);

	/* Turn off performance counters if on. */
	if (test_and_clear_thread_flag(TIF_PERFCTR)) {
		t->user_cntd0 = t->user_cntd1 = NULL;
		t->pcr_reg = 0;
		write_pcr(0);
	}

	/* Clear FPU register state. */
	t->fpsaved[0] = 0;
	
	if (get_thread_current_ds() != ASI_AIUS)
		set_fs(USER_DS);

	/* Init new signal delivery disposition. */
	clear_thread_flag(TIF_NEWSIGNALS);
}

/* It's a bit more tricky when 64-bit tasks are involved... */
static unsigned long clone_stackframe(unsigned long csp, unsigned long psp)
{
	unsigned long fp, distance, rval;

	if (!(test_thread_flag(TIF_32BIT))) {
		csp += STACK_BIAS;
		psp += STACK_BIAS;
		__get_user(fp, &(((struct reg_window __user *)psp)->ins[6]));
		fp += STACK_BIAS;
	} else
		__get_user(fp, &(((struct reg_window32 __user *)psp)->ins[6]));

	/* Now 8-byte align the stack as this is mandatory in the
	 * Sparc ABI due to how register windows work.  This hides
	 * the restriction from thread libraries etc.  -DaveM
	 */
	csp &= ~7UL;

	distance = fp - psp;
	rval = (csp - distance);
	if (copy_in_user((void __user *) rval, (void __user *) psp, distance))
		rval = 0;
	else if (test_thread_flag(TIF_32BIT)) {
		if (put_user(((u32)csp),
			     &(((struct reg_window32 __user *)rval)->ins[6])))
			rval = 0;
	} else {
		if (put_user(((u64)csp - STACK_BIAS),
			     &(((struct reg_window __user *)rval)->ins[6])))
			rval = 0;
		else
			rval = rval - STACK_BIAS;
	}

	return rval;
}

/* Standard stuff. */
static inline void shift_window_buffer(int first_win, int last_win,
				       struct thread_info *t)
{
	int i;

	for (i = first_win; i < last_win; i++) {
		t->rwbuf_stkptrs[i] = t->rwbuf_stkptrs[i+1];
		memcpy(&t->reg_window[i], &t->reg_window[i+1],
		       sizeof(struct reg_window));
	}
}

void synchronize_user_stack(void)
{
	struct thread_info *t = current_thread_info();
	unsigned long window;

	flush_user_windows();
	if ((window = get_thread_wsaved()) != 0) {
		int winsize = sizeof(struct reg_window);
		int bias = 0;

		if (test_thread_flag(TIF_32BIT))
			winsize = sizeof(struct reg_window32);
		else
			bias = STACK_BIAS;

		window -= 1;
		do {
			unsigned long sp = (t->rwbuf_stkptrs[window] + bias);
			struct reg_window *rwin = &t->reg_window[window];

			if (!copy_to_user((char __user *)sp, rwin, winsize)) {
				shift_window_buffer(window, get_thread_wsaved() - 1, t);
				set_thread_wsaved(get_thread_wsaved() - 1);
			}
		} while (window--);
	}
}

void fault_in_user_windows(void)
{
	struct thread_info *t = current_thread_info();
	unsigned long window;
	int winsize = sizeof(struct reg_window);
	int bias = 0;

	if (test_thread_flag(TIF_32BIT))
		winsize = sizeof(struct reg_window32);
	else
		bias = STACK_BIAS;

	flush_user_windows();
	window = get_thread_wsaved();

	if (window != 0) {
		window -= 1;
		do {
			unsigned long sp = (t->rwbuf_stkptrs[window] + bias);
			struct reg_window *rwin = &t->reg_window[window];

			if (copy_to_user((char __user *)sp, rwin, winsize))
				goto barf;
		} while (window--);
	}
	set_thread_wsaved(0);
	return;

barf:
	set_thread_wsaved(window + 1);
	do_exit(SIGILL);
}

asmlinkage long sparc_do_fork(unsigned long clone_flags,
			      unsigned long stack_start,
			      struct pt_regs *regs,
			      unsigned long stack_size)
{
	int __user *parent_tid_ptr, *child_tid_ptr;

#ifdef CONFIG_COMPAT
	if (test_thread_flag(TIF_32BIT)) {
		parent_tid_ptr = compat_ptr(regs->u_regs[UREG_I2]);
		child_tid_ptr = compat_ptr(regs->u_regs[UREG_I4]);
	} else
#endif
	{
		parent_tid_ptr = (int __user *) regs->u_regs[UREG_I2];
		child_tid_ptr = (int __user *) regs->u_regs[UREG_I4];
	}

	return do_fork(clone_flags, stack_start,
		       regs, stack_size,
		       parent_tid_ptr, child_tid_ptr);
}

/* Copy a Sparc thread.  The fork() return value conventions
 * under SunOS are nothing short of bletcherous:
 * Parent -->  %o0 == childs  pid, %o1 == 0
 * Child  -->  %o0 == parents pid, %o1 == 1
 */
int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
		unsigned long unused,
		struct task_struct *p, struct pt_regs *regs)
{
	struct thread_info *t = p->thread_info;
	char *child_trap_frame;

	/* Calculate offset to stack_frame & pt_regs */
	child_trap_frame = ((char *)t) + (THREAD_SIZE - (TRACEREG_SZ+STACKFRAME_SZ));
	memcpy(child_trap_frame, (((struct sparc_stackf *)regs)-1), (TRACEREG_SZ+STACKFRAME_SZ));

	t->flags = (t->flags & ~((0xffUL << TI_FLAG_CWP_SHIFT) | (0xffUL << TI_FLAG_CURRENT_DS_SHIFT))) |
		(((regs->tstate + 1) & TSTATE_CWP) << TI_FLAG_CWP_SHIFT);
	t->new_child = 1;
	t->ksp = ((unsigned long) child_trap_frame) - STACK_BIAS;
	t->kregs = (struct pt_regs *)(child_trap_frame+sizeof(struct sparc_stackf));
	t->fpsaved[0] = 0;

	if (regs->tstate & TSTATE_PRIV) {
		/* Special case, if we are spawning a kernel thread from
		 * a userspace task (via KMOD, NFS, or similar) we must
		 * disable performance counters in the child because the
		 * address space and protection realm are changing.
		 */
		if (t->flags & _TIF_PERFCTR) {
			t->user_cntd0 = t->user_cntd1 = NULL;
			t->pcr_reg = 0;
			t->flags &= ~_TIF_PERFCTR;
		}
		t->kregs->u_regs[UREG_FP] = t->ksp;
		t->flags |= ((long)ASI_P << TI_FLAG_CURRENT_DS_SHIFT);
		flush_register_windows();
		memcpy((void *)(t->ksp + STACK_BIAS),
		       (void *)(regs->u_regs[UREG_FP] + STACK_BIAS),
		       sizeof(struct sparc_stackf));
		t->kregs->u_regs[UREG_G6] = (unsigned long) t;
		t->kregs->u_regs[UREG_G4] = (unsigned long) t->task;
	} else {
		if (t->flags & _TIF_32BIT) {
			sp &= 0x00000000ffffffffUL;
			regs->u_regs[UREG_FP] &= 0x00000000ffffffffUL;
		}
		t->kregs->u_regs[UREG_FP] = sp;
		t->flags |= ((long)ASI_AIUS << TI_FLAG_CURRENT_DS_SHIFT);
		if (sp != regs->u_regs[UREG_FP]) {
			unsigned long csp;

			csp = clone_stackframe(sp, regs->u_regs[UREG_FP]);
			if (!csp)
				return -EFAULT;
			t->kregs->u_regs[UREG_FP] = csp;
		}
		if (t->utraps)
			t->utraps[0]++;
	}

	/* Set the return value for the child. */
	t->kregs->u_regs[UREG_I0] = current->pid;
	t->kregs->u_regs[UREG_I1] = 1;

	/* Set the second return value for the parent. */
	regs->u_regs[UREG_I1] = 0;

	if (clone_flags & CLONE_SETTLS)
		t->kregs->u_regs[UREG_G7] = regs->u_regs[UREG_I3];

	return 0;
}

/*
 * This is the mechanism for creating a new kernel thread.
 *
 * NOTE! Only a kernel-only process(ie the swapper or direct descendants
 * who haven't done an "execve()") should use this: it will work within
 * a system call from a "real" process, but the process memory space will
 * not be free'd until both the parent and the child have exited.
 */
pid_t kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
{
	long retval;

	/* If the parent runs before fn(arg) is called by the child,
	 * the input registers of this function can be clobbered.
	 * So we stash 'fn' and 'arg' into global registers which
	 * will not be modified by the parent.
	 */
	__asm__ __volatile__("mov %4, %%g2\n\t"	   /* Save FN into global */
			     "mov %5, %%g3\n\t"	   /* Save ARG into global */
			     "mov %1, %%g1\n\t"	   /* Clone syscall nr. */
			     "mov %2, %%o0\n\t"	   /* Clone flags. */
			     "mov 0, %%o1\n\t"	   /* usp arg == 0 */
			     "t 0x6d\n\t"	   /* Linux/Sparc clone(). */
			     "brz,a,pn %%o1, 1f\n\t" /* Parent, just return. */
			     " mov %%o0, %0\n\t"
			     "jmpl %%g2, %%o7\n\t"   /* Call the function. */
			     " mov %%g3, %%o0\n\t"   /* Set arg in delay. */
			     "mov %3, %%g1\n\t"
			     "t 0x6d\n\t"	   /* Linux/Sparc exit(). */
			     /* Notreached by child. */
			     "1:" :
			     "=r" (retval) :
			     "i" (__NR_clone), "r" (flags | CLONE_VM | CLONE_UNTRACED),
			     "i" (__NR_exit),  "r" (fn), "r" (arg) :
			     "g1", "g2", "g3", "o0", "o1", "memory", "cc");
	return retval;
}

/*
 * fill in the user structure for a core dump..
 */
void dump_thread(struct pt_regs * regs, struct user * dump)
{
	/* Only should be used for SunOS and ancient a.out
	 * SparcLinux binaries...  Not worth implementing.
	 */
	memset(dump, 0, sizeof(struct user));
}

typedef struct {
	union {
		unsigned int	pr_regs[32];
		unsigned long	pr_dregs[16];
	} pr_fr;
	unsigned int __unused;
	unsigned int	pr_fsr;
	unsigned char	pr_qcnt;
	unsigned char	pr_q_entrysize;
	unsigned char	pr_en;
	unsigned int	pr_q[64];
} elf_fpregset_t32;

/*
 * fill in the fpu structure for a core dump.
 */
int dump_fpu (struct pt_regs * regs, elf_fpregset_t * fpregs)
{
	unsigned long *kfpregs = current_thread_info()->fpregs;
	unsigned long fprs = current_thread_info()->fpsaved[0];

	if (test_thread_flag(TIF_32BIT)) {
		elf_fpregset_t32 *fpregs32 = (elf_fpregset_t32 *)fpregs;

		if (fprs & FPRS_DL)
			memcpy(&fpregs32->pr_fr.pr_regs[0], kfpregs,
			       sizeof(unsigned int) * 32);
		else
			memset(&fpregs32->pr_fr.pr_regs[0], 0,
			       sizeof(unsigned int) * 32);
		fpregs32->pr_qcnt = 0;
		fpregs32->pr_q_entrysize = 8;
		memset(&fpregs32->pr_q[0], 0,
		       (sizeof(unsigned int) * 64));
		if (fprs & FPRS_FEF) {
			fpregs32->pr_fsr = (unsigned int) current_thread_info()->xfsr[0];
			fpregs32->pr_en = 1;
		} else {
			fpregs32->pr_fsr = 0;
			fpregs32->pr_en = 0;
		}
	} else {
		if(fprs & FPRS_DL)
			memcpy(&fpregs->pr_regs[0], kfpregs,
			       sizeof(unsigned int) * 32);
		else
			memset(&fpregs->pr_regs[0], 0,
			       sizeof(unsigned int) * 32);
		if(fprs & FPRS_DU)
			memcpy(&fpregs->pr_regs[16], kfpregs+16,
			       sizeof(unsigned int) * 32);
		else
			memset(&fpregs->pr_regs[16], 0,
			       sizeof(unsigned int) * 32);
		if(fprs & FPRS_FEF) {
			fpregs->pr_fsr = current_thread_info()->xfsr[0];
			fpregs->pr_gsr = current_thread_info()->gsr[0];
		} else {
			fpregs->pr_fsr = fpregs->pr_gsr = 0;
		}
		fpregs->pr_fprs = fprs;
	}
	return 1;
}

/*
 * sparc_execve() executes a new program after the asm stub has set
 * things up for us.  This should basically do what I want it to.
 */
asmlinkage int sparc_execve(struct pt_regs *regs)
{
	int error, base = 0;
	char *filename;

	/* User register window flush is done by entry.S */

	/* Check for indirect call. */
	if (regs->u_regs[UREG_G1] == 0)
		base = 1;

	filename = getname((char __user *)regs->u_regs[base + UREG_I0]);
	error = PTR_ERR(filename);
	if (IS_ERR(filename))
		goto out;
	error = do_execve(filename,
			  (char __user * __user *)
			  regs->u_regs[base + UREG_I1],
			  (char __user * __user *)
			  regs->u_regs[base + UREG_I2], regs);
	putname(filename);
	if (!error) {
		fprs_write(0);
		current_thread_info()->xfsr[0] = 0;
		current_thread_info()->fpsaved[0] = 0;
		regs->tstate &= ~TSTATE_PEF;
		task_lock(current);
		current->ptrace &= ~PT_DTRACE;
		task_unlock(current);
	}
out:
	return error;
}

unsigned long get_wchan(struct task_struct *task)
{
	unsigned long pc, fp, bias = 0;
	unsigned long thread_info_base;
	struct reg_window *rw;
        unsigned long ret = 0;
	int count = 0; 

	if (!task || task == current ||
            task->state == TASK_RUNNING)
		goto out;

	thread_info_base = (unsigned long) task->thread_info;
	bias = STACK_BIAS;
	fp = task->thread_info->ksp + bias;

	do {
		/* Bogus frame pointer? */
		if (fp < (thread_info_base + sizeof(struct thread_info)) ||
		    fp >= (thread_info_base + THREAD_SIZE))
			break;
		rw = (struct reg_window *) fp;
		pc = rw->ins[7];
		if (!in_sched_functions(pc)) {
			ret = pc;
			goto out;
		}
		fp = rw->ins[6] + bias;
	} while (++count < 16);

out:
	return ret;
}