[PATCH] Adeos nanokernel for 2.5.38 2/2: i386 code

Karim Yaghmour (karim@opersys.com)
Sun, 22 Sep 2002 22:59:08 -0400


This is the i386 specific portion of the nanokernel.

Here are the files added:
arch/i386/kernel/adeos.c | 330 ++++++++++++++++++
arch/i386/kernel/ipipe.c | 686 ++++++++++++++++++++++++++++++++++++++
include/asm-i386/adeos.h | 242 +++++++++++++

Here are the files modified:
arch/i386/kernel/Makefile | 3
arch/i386/kernel/apic.c | 9
arch/i386/kernel/bluesmoke.c | 3
arch/i386/kernel/entry.S | 127 ++++++-
arch/i386/kernel/i386_ksyms.c | 21 +
arch/i386/kernel/i8259.c | 36 +
arch/i386/kernel/irq.c | 3
arch/i386/kernel/process.c | 3
arch/i386/kernel/smp.c | 25 +
arch/i386/kernel/time.c | 11
arch/i386/kernel/traps.c | 68 +++
arch/i386/mach-generic/do_timer.h | 9
arch/i386/mach-visws/do_timer.h | 9
arch/i386/mm/fault.c | 6
include/asm-i386/system.h | 49 ++
18 files changed, 1636 insertions, 4 deletions

Again, these modifications ensure that Linux plays nice with the other
domains in the ipipe.

diff -urpN linux-2.5.38/arch/i386/kernel/Makefile linux-2.5.38-adeos/arch/i386/kernel/Makefile
--- linux-2.5.38/arch/i386/kernel/Makefile Sun Sep 22 00:25:02 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/Makefile Sun Sep 22 21:57:18 2002
@@ -4,7 +4,7 @@

EXTRA_TARGETS := head.o init_task.o

-export-objs := mca.o i386_ksyms.o time.o
+export-objs := mca.o i386_ksyms.o time.o ipipe.o adeos.o

obj-y := process.o semaphore.o signal.o entry.o traps.o irq.o vm86.o \
ptrace.o i8259.o ioport.o ldt.o setup.o time.o sys_i386.o \
@@ -26,6 +26,7 @@ obj-$(CONFIG_X86_LOCAL_APIC) += apic.o n
obj-$(CONFIG_X86_IO_APIC) += io_apic.o
obj-$(CONFIG_SOFTWARE_SUSPEND) += suspend.o
obj-$(CONFIG_X86_NUMAQ) += numaq.o
+obj-$(CONFIG_ADEOS) += ipipe.o adeos.o

EXTRA_AFLAGS := -traditional

diff -urpN linux-2.5.38/arch/i386/kernel/adeos.c linux-2.5.38-adeos/arch/i386/kernel/adeos.c
--- linux-2.5.38/arch/i386/kernel/adeos.c Wed Dec 31 19:00:00 1969
+++ linux-2.5.38-adeos/arch/i386/kernel/adeos.c Sun Sep 22 21:57:18 2002
@@ -0,0 +1,330 @@
+/*
+ * linux/arch/i386/kernel/adeos.c
+ *
+ * Copyright (C) 2002 Philippe Gerum.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; either version 2 of the License, or (at your option) any later
+ * version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * This code implements the architecture-dependent ADEOS support.
+ */
+
+#include <linux/config.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/irq.h>
+#include <asm/system.h>
+#include <asm/io.h>
+#include <asm/errno.h>
+
+#ifdef CONFIG_SMP
+
+static volatile unsigned long adeos_cpu_sync_map;
+
+static volatile unsigned long adeos_cpu_lock_map;
+
+static spinlock_t adeos_cpu_barrier = SPIN_LOCK_UNLOCKED;
+
+static void (*adeos_cpu_sync)(void);
+
+void __adeos_do_critical_sync (unsigned irq)
+
+{
+ ipipe_declare_cpuid;
+
+ set_bit(cpuid,&adeos_cpu_sync_map);
+
+ /* Now we are in sync with the lock requestor running on another
+ CPU. Enter a spinning wait until he releases the global
+ lock. */
+ ipipe_hw_spin_lock_simple(&adeos_cpu_barrier);
+
+ /* Got it. Now get out. */
+
+ if (adeos_cpu_sync)
+ /* Call the sync routine if any. */
+ adeos_cpu_sync();
+
+ ipipe_hw_spin_unlock_simple(&adeos_cpu_barrier);
+
+ clear_bit(cpuid,&adeos_cpu_sync_map);
+}
+
+#endif /* CONFIG_SMP */
+
+void __adeos_bootstrap_domain (adomain_t *adp, adattr_t *attr)
+
+{
+ int estacksz = attr->estacksz > 0 ? attr->estacksz : 8192;
+ int **psp = &adp->esp;
+
+ adp->estackbase = (int *)kmalloc(estacksz,GFP_KERNEL);
+
+ if (adp->estackbase == NULL)
+ panic("no memory for domain entry stack");
+
+ adp->esp = adp->estackbase;
+ **psp = 0;
+ *psp = (int *)(((unsigned long)*psp + estacksz - 0x10) & ~0xf);
+ *--(*psp) = 0;
+ *--(*psp) = (int)attr->entry;
+}
+
+void __adeos_cleanup_domain (adomain_t *adp) {
+
+ kfree(adp->estackbase);
+}
+
+/* __adeos_switch_domain() -- Switch domain contexts. The current
+ domain is switched out while the domain pointed by "adp" is
+ switched in. */
+
+#ifdef CONFIG_SMP
+
+adomain_t **__adeos_get_smp_domain_slot (void) {
+ ipipe_declare_cpuid;
+ return adp_cpu_current + cpuid;
+}
+
+void __adeos_switch_domain (adomain_t *adp)
+
+{
+ __asm__ __volatile__( \
+ "pushl %%edx\n\t" \
+ "pushl %%ecx\n\t" \
+ "pushl %%ebx\n\t" \
+ "pushl %%ebp\n\t" \
+ "pushl %%edi\n\t" \
+ "pushl %%esi\n\t" \
+ /* Prepare to switch: build a resume frame for the suspended
+ domain, save it's stack pointer, load the incoming
+ domain's stack pointer, update the global domain
+ descriptor pointer, then finally activate the resume
+ frame of the incoming domain. */
+ "pushl $1f\n\t" \
+ "pushl %%eax\n\t" \
+ "call __adeos_get_smp_domain_slot\n\t" \
+ "popl %%edx\n\t" \
+ "xchg (%%eax), %%edx\n\t" \
+ "movl %%esp, (%%edx)\n\t" \
+ "movl (%%eax), %%eax\n\t" \
+ "movl (%%eax), %%esp\n\t" \
+ "ret\n\t" \
+ /* Call domain switch hook (if any) */
+"1: call __adeos_get_smp_domain_slot\n\t" \
+ "movl (%%eax), %%eax\n\t" \
+ "movl 4(%%eax),%%eax\n\t" \
+ "testl %%eax,%%eax\n\t" \
+ "je 2f\n\t" \
+ "call *%%eax\n\t" \
+ /* Domain resume point */
+"2: popl %%esi\n\t" \
+ "popl %%edi\n\t" \
+ "popl %%ebp\n\t" \
+ "popl %%ebx\n\t" \
+ "popl %%ecx\n\t" \
+ "popl %%edx\n\t" \
+ : /* no output */ \
+ : "a" (adp));
+}
+
+#else /* !CONFIG_SMP */
+
+void __adeos_switch_domain (adomain_t *adp)
+
+{
+ __asm__ __volatile__( \
+ "pushl %%edx\n\t" \
+ "pushl %%ecx\n\t" \
+ "pushl %%ebx\n\t" \
+ "pushl %%ebp\n\t" \
+ "pushl %%edi\n\t" \
+ "pushl %%esi\n\t" \
+ /* Prepare to switch: build a resume frame for the suspended
+ domain, save it's stack pointer, load the incoming
+ domain's stack pointer, update the global domain
+ descriptor pointer, then finally activate the resume
+ frame of the incoming domain. */
+ "movl adp_cpu_current, %%edx\n\t" \
+ "pushl %%edx\n\t" \
+ "pushl $1f\n\t" \
+ "movl %%esp, (%%edx)\n\t" \
+ "movl (%%eax), %%esp\n\t" \
+ "movl %%eax, adp_cpu_current\n\t" \
+ "ret\n\t" \
+ /* Call domain switch hook (if any) */
+"1: popl %%eax\n\t" \
+ "movl 4(%%eax),%%eax\n\t" \
+ "testl %%eax,%%eax\n\t" \
+ "je 2f\n\t" \
+ "call *%%eax\n\t" \
+ /* Domain resume point */
+"2: popl %%esi\n\t" \
+ "popl %%edi\n\t" \
+ "popl %%ebp\n\t" \
+ "popl %%ebx\n\t" \
+ "popl %%ecx\n\t" \
+ "popl %%edx\n\t" \
+ : /* no output */ \
+ : "a" (adp));
+}
+
+#endif /* CONFIG_SMP */
+
+
+/* __adeos_critical_enter() -- Grab the superlock excluding all CPUs
+ but the current one from a critical section. This lock is used when
+ we must enforce a global critical section for a single CPU in a
+ possibly SMP system whichever context the CPUs are running
+ (i.e. interrupt handler or regular thread). */
+
+unsigned long __adeos_critical_enter (void (*syncfn)(void))
+
+{
+ unsigned long flags;
+
+ ipipe_hw_save_flags_and_cli(flags);
+
+#ifdef CONFIG_SMP
+ if (__adeos_online_cpus > 1) /* We might be running a SMP-kernel on a UP box... */
+ {
+ ipipe_declare_cpuid;
+
+ set_bit(cpuid,&adeos_cpu_lock_map);
+
+ while (test_and_set_bit(31,&adeos_cpu_lock_map))
+ {
+ /* Refer to the explanations found in
+ linux/arch/asm-i386/irq.c about
+ SUSPECTED_CPU_OR_CHIPSET_BUG_WORKAROUND for more about
+ this strange loop. */
+ int n = 0;
+ do { cpu_relax(); } while (++n < cpuid);
+ }
+
+ ipipe_hw_spin_lock_simple(&adeos_cpu_barrier);
+
+ adeos_cpu_sync = syncfn;
+
+ /* Send the sync IPI to all processors but the current one. */
+ __adeos_send_IPI_allbutself(ADEOS_CRITICAL_VECTOR);
+
+ while (adeos_cpu_sync_map != (cpu_online_map & ~adeos_cpu_lock_map))
+ cpu_relax();
+ }
+#endif /* CONFIG_SMP */
+
+ return flags;
+}
+
+/* __adeos_critical_exit() -- Release the superlock. */
+
+void __adeos_critical_exit (unsigned long flags)
+
+{
+#ifdef CONFIG_SMP
+ if (__adeos_online_cpus > 1) /* We might be running a SMP-kernel on a UP box... */
+ {
+ ipipe_declare_cpuid;
+
+ ipipe_hw_spin_unlock_simple(&adeos_cpu_barrier);
+
+ while (adeos_cpu_sync_map != 0)
+ cpu_relax();
+
+ if (test_and_clear_bit(cpuid,&adeos_cpu_lock_map))
+ clear_bit(31,&adeos_cpu_lock_map);
+ }
+#endif /* CONFIG_SMP */
+
+ ipipe_hw_restore_flags(flags);
+}
+
+int __adeos_get_sysinfo (adsysinfo_t *info)
+
+{
+#ifdef CONFIG_X86_IO_APIC
+ if (!skip_ioapic_setup && nr_ioapics > 0)
+ info->timer.irq = IO_APIC_VECTOR(0) - FIRST_EXTERNAL_VECTOR;
+ else
+ info->timer.irq = 0;
+#else /* !CONFIG_X86_IO_APIC */
+ info->timer.irq = 0;
+#endif /* CONFIG_X86_IO_APIC */
+
+#ifdef CONFIG_SMP
+ info->ncpus = __adeos_online_cpus;
+#else /* !CONFIG_SMP */
+ info->ncpus = 1;
+#endif /* CONFIG_SMP */
+
+ if (cpu_has_tsc)
+ info->cpufreq = 1000 * cpu_khz;
+ else
+ info->cpufreq = CLOCK_TICK_RATE;
+
+ return 0;
+}
+
+int __adeos_tune_timer (unsigned long ns, int flags)
+
+{
+ unsigned ghz, latch;
+ unsigned long x;
+
+ if (flags & ADEOS_RESET_TIMER)
+ latch = LATCH;
+ else
+ {
+ if (ns < 122071 || ns > (1000 / HZ) * 1000000) /* HZ max, 8khz min */
+ return -EINVAL;
+
+ ghz = 1000000000 / ns;
+ latch = (CLOCK_TICK_RATE + ghz/2) / ghz;
+ }
+
+ x = __adeos_critical_enter(NULL); /* Sync with all CPUs */
+
+ /* Shamelessly lifted from init_IRQ() in i8259.c */
+ outb_p(0x34,0x43); /* binary, mode 2, LSB/MSB, ch 0 */
+ outb_p(latch & 0xff,0x40); /* LSB */
+ outb(latch >> 8,0x40); /* MSB */
+
+ __adeos_critical_exit(x);
+
+ return 0;
+}
+
+void __adeos_set_root_ptd (int key, void *value) {
+
+ current->ptd[key] = value;
+}
+
+void *__adeos_get_root_ptd (int key) {
+
+ return current->ptd[key];
+}
+
+asmlinkage void __adeos_enter_syscall (struct pt_regs *regs) {
+
+ if (__adeos_event_monitors[ADEOS_SYSCALL_PROLOGUE] > 0)
+ ipipe_handle_event(ADEOS_SYSCALL_PROLOGUE,regs);
+}
+
+asmlinkage void __adeos_exit_syscall (void) {
+
+ if (__adeos_event_monitors[ADEOS_SYSCALL_EPILOGUE] > 0)
+ ipipe_handle_event(ADEOS_SYSCALL_EPILOGUE,NULL);
+}
diff -urpN linux-2.5.38/arch/i386/kernel/apic.c linux-2.5.38-adeos/arch/i386/kernel/apic.c
--- linux-2.5.38/arch/i386/kernel/apic.c Sun Sep 22 00:25:34 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/apic.c Sun Sep 22 21:57:18 2002
@@ -1095,7 +1095,10 @@ void smp_apic_timer_interrupt(struct pt_
* NOTE! We'd better ACK the irq immediately,
* because timer handling can be slow.
*/
+#ifndef CONFIG_ADEOS
+ /* IRQ acknowledging is in the hands of the IPIPE exclusively. */
ack_APIC_irq();
+#endif
/*
* update_process_times() expects us to have done irq_enter().
* Besides, if we don't timer interrupts ignore the global
@@ -1120,8 +1123,11 @@ asmlinkage void smp_spurious_interrupt(v
* Spurious interrupts should not be ACKed.
*/
v = apic_read(APIC_ISR + ((SPURIOUS_APIC_VECTOR & ~0x1f) >> 1));
+#ifndef CONFIG_ADEOS
+ /* IRQ acknowledging is in the hands of the IPIPE exclusively. */
if (v & (1 << (SPURIOUS_APIC_VECTOR & 0x1f)))
ack_APIC_irq();
+#endif

/* see sw-dev-man vol 3, chapter 7.4.13.5 */
printk(KERN_INFO "spurious APIC interrupt on CPU#%d, should never happen.\n",
@@ -1142,7 +1148,10 @@ asmlinkage void smp_error_interrupt(void
v = apic_read(APIC_ESR);
apic_write(APIC_ESR, 0);
v1 = apic_read(APIC_ESR);
+#ifndef CONFIG_ADEOS
+ /* IRQ acknowledging is in the hands of the IPIPE exclusively. */
ack_APIC_irq();
+#endif
atomic_inc(&irq_err_count);

/* Here is what the APIC error bits mean:
diff -urpN linux-2.5.38/arch/i386/kernel/bluesmoke.c linux-2.5.38-adeos/arch/i386/kernel/bluesmoke.c
--- linux-2.5.38/arch/i386/kernel/bluesmoke.c Sun Sep 22 00:25:05 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/bluesmoke.c Sun Sep 22 21:57:18 2002
@@ -53,7 +53,10 @@ static void intel_thermal_interrupt(stru
u32 l, h;
unsigned int cpu = smp_processor_id();

+#ifndef CONFIG_ADEOS
+ /* IRQ acknowledging is in the hands of the IPIPE exclusively. */
ack_APIC_irq();
+#endif

rdmsr(MSR_IA32_THERM_STATUS, l, h);
if (l & 1) {
diff -urpN linux-2.5.38/arch/i386/kernel/entry.S linux-2.5.38-adeos/arch/i386/kernel/entry.S
--- linux-2.5.38/arch/i386/kernel/entry.S Sun Sep 22 00:25:04 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/entry.S Sun Sep 22 21:57:18 2002
@@ -71,7 +71,11 @@ NT_MASK = 0x00004000
VM_MASK = 0x00020000

#ifdef CONFIG_PREEMPT
+#ifdef CONFIG_ADEOS
+#define preempt_stop call adeos_stall_ipipe
+#else /* !CONFIG_ADEOS */
#define preempt_stop cli
+#endif /* CONFIG_ADEOS */
#else
#define preempt_stop
#define resume_kernel restore_all
@@ -128,6 +132,12 @@ ENTRY(lcall7)
# gates, which has to be cleaned up later..
pushl %eax
SAVE_ALL
+#ifdef CONFIG_ADEOS
+ movl %esp,%eax # Use a "soon-to-be-clobbered" reg
+ pushl %eax
+ call __adeos_enter_syscall
+ addl $4,%esp
+#endif /* CONFIG_ADEOS */
movl EIP(%esp), %eax # due to call gates, this is eflags, not eip..
movl CS(%esp), %edx # this is eip..
movl EFLAGS(%esp), %ecx # and this is cs..
@@ -142,6 +152,9 @@ ENTRY(lcall7)
pushl $0x7
call *%edx
addl $4, %esp
+#ifdef CONFIG_ADEOS
+ call __adeos_exit_syscall
+#endif /* CONFIG_ADEOS */
popl %eax
jmp resume_userspace

@@ -150,6 +163,12 @@ ENTRY(lcall27)
# gates, which has to be cleaned up later..
pushl %eax
SAVE_ALL
+#ifdef CONFIG_ADEOS
+ movl %esp,%eax # Use a "soon-to-be-clobbered" reg
+ pushl %eax
+ call __adeos_enter_syscall
+ addl $4,%esp
+#endif /* CONFIG_ADEOS */
movl EIP(%esp), %eax # due to call gates, this is eflags, not eip..
movl CS(%esp), %edx # this is eip..
movl EFLAGS(%esp), %ecx # and this is cs..
@@ -164,11 +183,17 @@ ENTRY(lcall27)
pushl $0x27
call *%edx
addl $4, %esp
+#ifdef CONFIG_ADEOS
+ call __adeos_exit_syscall
+#endif /* CONFIG_ADEOS */
popl %eax
jmp resume_userspace


ENTRY(ret_from_fork)
+#ifdef CONFIG_ADEOS
+ sti # The pipeline is still stalled however.
+#endif /* CONFIG_ADEOS */
#if CONFIG_SMP || CONFIG_PREEMPT
# NOTE: this function takes a parameter but it's unused on x86.
call schedule_tail
@@ -187,6 +212,9 @@ ENTRY(ret_from_fork)
ALIGN
ret_from_exception:
preempt_stop
+#ifdef CONFIG_ADEOS
+ /* FIXME: Set marker */
+#endif /* CONFIG_ADEOS */
ret_from_intr:
GET_THREAD_INFO(%ebx)
movl EFLAGS(%esp), %eax # mix EFLAGS and CS
@@ -194,14 +222,22 @@ ret_from_intr:
testl $(VM_MASK | 3), %eax
jz resume_kernel # returning to kernel or vm86-space
ENTRY(resume_userspace)
- cli # make sure we don't miss an interrupt
+#ifdef CONFIG_ADEOS
+ call adeos_stall_ipipe
+#else /* !CONFIG_ADEOS */
+ cli # make sure we don't miss an interrupt
+#endif /* CONFIG_ADEOS */
# setting need_resched or sigpending
# between sampling and the iret
movl TI_FLAGS(%ebx), %ecx
andl $_TIF_WORK_MASK, %ecx # is there any work to be done on
# int/exception return?
jne work_pending
+#ifdef CONFIG_ADEOS
+ jmp unstall_and_restore_all
+#else /* !CONFIG_ADEOS */
jmp restore_all
+#endif /* CONFIG_ADEOS */

#ifdef CONFIG_PREEMPT
ENTRY(resume_kernel)
@@ -211,13 +247,21 @@ need_resched:
movl TI_FLAGS(%ebx), %ecx # need_resched set ?
testb $_TIF_NEED_RESCHED, %cl
jz restore_all
- testl $IF_MASK,EFLAGS(%esp) # interrupts off (execption path) ?
+ testl $IF_MASK,EFLAGS(%esp) # interrupts off (execption path) ? ADEOS FIXME: Test marker
jz restore_all
movl $PREEMPT_ACTIVE,TI_PRE_COUNT(%ebx)
+#ifdef CONFIG_ADEOS
+ call adeos_unstall_ipipe
+#else /* !CONFIG_ADEOS */
sti
+#endif /* CONFIG_ADEOS */
call schedule
movl $0,TI_PRE_COUNT(%ebx)
+#ifdef CONFIG_ADEOS
+ call adeos_stall_ipipe
+#else /* !CONFIG_ADEOS */
cli
+#endif /* CONFIG_ADEOS */
jmp need_resched
#endif

@@ -226,6 +270,14 @@ need_resched:
ENTRY(system_call)
pushl %eax # save orig_eax
SAVE_ALL
+#ifdef CONFIG_ADEOS
+ pushl %eax
+ movl %esp,%ebx # Use a "soon-to-be-clobbered" reg
+ pushl %ebx
+ call __adeos_enter_syscall
+ addl $4,%esp
+ popl %eax
+#endif /* CONFIG_ADEOS */
GET_THREAD_INFO(%ebx)
cmpl $(NR_syscalls), %eax
jae syscall_badsys
@@ -235,13 +287,24 @@ ENTRY(system_call)
syscall_call:
call *sys_call_table(,%eax,4)
movl %eax,EAX(%esp) # store the return value
+#ifdef CONFIG_ADEOS
+ call __adeos_exit_syscall
+#endif /* CONFIG_ADEOS */
syscall_exit:
+#ifdef CONFIG_ADEOS
+ call adeos_stall_ipipe
+#else /* !CONFIG_ADEOS */
cli # make sure we don't miss an interrupt
+#endif /* CONFIG_ADEOS */
# setting need_resched or sigpending
# between sampling and the iret
movl TI_FLAGS(%ebx), %ecx
testw $_TIF_ALLWORK_MASK, %cx # current->work
jne syscall_exit_work
+#ifdef CONFIG_ADEOS
+unstall_and_restore_all:
+ call adeos_unstall_ipipe
+#endif /* CONFIG_ADEOS */
restore_all:
RESTORE_ALL

@@ -252,13 +315,21 @@ work_pending:
jz work_notifysig
work_resched:
call schedule
+#ifdef CONFIG_ADEOS
+ call adeos_stall_ipipe
+#else /* !CONFIG_ADEOS */
cli # make sure we don't miss an interrupt
+#endif /* CONFIG_ADEOS */
# setting need_resched or sigpending
# between sampling and the iret
movl TI_FLAGS(%ebx), %ecx
andl $_TIF_WORK_MASK, %ecx # is there any work to be done other
# than syscall tracing?
+#ifdef CONFIG_ADEOS
+ jz unstall_and_restore_all
+#else /* !CONFIG_ADEOS */
jz restore_all
+#endif /* CONFIG_ADEOS */
testb $_TIF_NEED_RESCHED, %cl
jnz work_resched

@@ -270,7 +341,11 @@ work_notifysig: # deal with pending s
# vm86-space
xorl %edx, %edx
call do_notify_resume
+#ifdef CONFIG_ADEOS
+ jmp unstall_and_restore_all
+#else /* !CONFIG_ADEOS */
jmp restore_all
+#endif /* CONFIG_ADEOS */

ALIGN
work_notifysig_v86:
@@ -280,7 +355,11 @@ work_notifysig_v86:
movl %eax, %esp
xorl %edx, %edx
call do_notify_resume
+#ifdef CONFIG_ADEOS
+ jmp unstall_and_restore_all
+#else /* !CONFIG_ADEOS */
jmp restore_all
+#endif /* CONFIG_ADEOS */

# perform syscall exit tracing
ALIGN
@@ -299,7 +378,11 @@ syscall_trace_entry:
syscall_exit_work:
testb $_TIF_SYSCALL_TRACE, %cl
jz work_pending
+#ifdef CONFIG_ADEOS
+ call adeos_unstall_ipipe
+#else /* !CONFIG_ADEOS */
sti # could let do_syscall_trace() call
+#endif /* CONFIG_ADEOS */
# schedule() instead
movl %esp, %eax
movl $1, %edx
@@ -347,6 +430,46 @@ ENTRY(name) \
/* The include is where all of the SMP etc. interrupts come from */
#include "entry_arch.h"

+#ifdef CONFIG_ADEOS
+
+.data
+ENTRY(ipipe_irq_trampolines)
+.text
+
+vector=0
+ENTRY(ipipe_irq_entries)
+.rept NR_IRQS
+ ALIGN
+1: pushl $vector-256
+ jmp ipipe_irq_common
+.data
+ .long 1b
+.text
+vector=vector+1
+.endr
+
+ ALIGN
+ipipe_irq_common:
+ SAVE_ALL
+ call ipipe_handle_irq
+ testl %eax,%eax
+ jnz 1f
+ popl %ebx
+ popl %ecx
+ popl %edx
+ popl %esi
+ popl %edi
+ popl %ebp
+ popl %eax
+ popl %ds
+ popl %es
+ addl $4, %esp
+ iret
+1: GET_THREAD_INFO(%ebx)
+ jmp ret_from_intr
+
+#endif /* !CONFIG_ADEOS */
+
ENTRY(divide_error)
pushl $0 # no error code
pushl $do_divide_error
diff -urpN linux-2.5.38/arch/i386/kernel/i386_ksyms.c linux-2.5.38-adeos/arch/i386/kernel/i386_ksyms.c
--- linux-2.5.38/arch/i386/kernel/i386_ksyms.c Sun Sep 22 00:25:09 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/i386_ksyms.c Sun Sep 22 21:57:18 2002
@@ -176,3 +176,24 @@ extern int is_sony_vaio_laptop;
EXPORT_SYMBOL(is_sony_vaio_laptop);

EXPORT_SYMBOL(__PAGE_KERNEL);
+
+#ifdef CONFIG_ADEOS
+EXPORT_SYMBOL(ipipe_sync_irqs);
+/* The following are per-platform convenience exports which might be
+ needed by some Adeos domains loaded as kernel modules. */
+extern irq_desc_t irq_desc[];
+EXPORT_SYMBOL(irq_desc);
+extern struct desc_struct idt_table[];
+EXPORT_SYMBOL(idt_table);
+extern struct tty_driver console_driver;
+EXPORT_SYMBOL(console_driver);
+extern void (*kd_mksound)(unsigned int hz, unsigned int ticks);
+EXPORT_SYMBOL(kd_mksound);
+#ifdef CONFIG_SMP
+#ifdef CONFIG_MULTIQUAD
+EXPORT_SYMBOL(logical_apicid_2_cpu);
+#else
+EXPORT_SYMBOL(physical_apicid_2_cpu);
+#endif /* CONFIG_MULTIQUAD */
+#endif /* CONFIG_SMP */
+#endif /* CONFIG_ADEOS */
diff -urpN linux-2.5.38/arch/i386/kernel/i8259.c linux-2.5.38-adeos/arch/i386/kernel/i8259.c
--- linux-2.5.38/arch/i386/kernel/i8259.c Sun Sep 22 00:25:25 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/i8259.c Sun Sep 22 21:57:18 2002
@@ -93,13 +93,21 @@ void disable_8259A_irq(unsigned int irq)
unsigned int mask = 1 << irq;
unsigned long flags;

+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_lock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_lock_irqsave(&i8259A_lock, flags);
+#endif /* CONFIG_ADEOS */
cached_irq_mask |= mask;
if (irq & 8)
outb(cached_A1,0xA1);
else
outb(cached_21,0x21);
+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_unlock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_unlock_irqrestore(&i8259A_lock, flags);
+#endif /* CONFIG_ADEOS */
}

void enable_8259A_irq(unsigned int irq)
@@ -107,13 +115,21 @@ void enable_8259A_irq(unsigned int irq)
unsigned int mask = ~(1 << irq);
unsigned long flags;

+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_lock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_lock_irqsave(&i8259A_lock, flags);
+#endif /* CONFIG_ADEOS */
cached_irq_mask &= mask;
if (irq & 8)
outb(cached_A1,0xA1);
else
outb(cached_21,0x21);
+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_unlock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_unlock_irqrestore(&i8259A_lock, flags);
+#endif /* CONFIG_ADEOS */
}

int i8259A_irq_pending(unsigned int irq)
@@ -122,12 +138,20 @@ int i8259A_irq_pending(unsigned int irq)
unsigned long flags;
int ret;

+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_lock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_lock_irqsave(&i8259A_lock, flags);
+#endif /* CONFIG_ADEOS */
if (irq < 8)
ret = inb(0x20) & mask;
else
ret = inb(0xA0) & (mask >> 8);
+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_unlock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_unlock_irqrestore(&i8259A_lock, flags);
+#endif /* CONFIG_ADEOS */

return ret;
}
@@ -196,12 +220,16 @@ void mask_and_ack_8259A(unsigned int irq

handle_real_irq:
if (irq & 8) {
+#ifndef CONFIG_ADEOS
inb(0xA1); /* DUMMY - (do we need this?) */
+#endif
outb(cached_A1,0xA1);
outb(0x60+(irq&7),0xA0);/* 'Specific EOI' to slave */
outb(0x62,0x20); /* 'Specific EOI' to master-IRQ2 */
} else {
+#ifndef CONFIG_ADEOS
inb(0x21); /* DUMMY - (do we need this?) */
+#endif
outb(cached_21,0x21);
outb(0x60+irq,0x20); /* 'Specific EOI' to master */
}
@@ -267,7 +295,11 @@ void init_8259A(int auto_eoi)
{
unsigned long flags;

+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_lock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_lock_irqsave(&i8259A_lock, flags);
+#endif /* CONFIG_ADEOS */

outb(0xff, 0x21); /* mask all of 8259A-1 */
outb(0xff, 0xA1); /* mask all of 8259A-2 */
@@ -303,7 +335,11 @@ void init_8259A(int auto_eoi)
outb(cached_21, 0x21); /* restore master IRQ mask */
outb(cached_A1, 0xA1); /* restore slave IRQ mask */

+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_unlock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_unlock_irqrestore(&i8259A_lock, flags);
+#endif /* CONFIG_ADEOS */
}

/*
diff -urpN linux-2.5.38/arch/i386/kernel/ipipe.c linux-2.5.38-adeos/arch/i386/kernel/ipipe.c
--- linux-2.5.38/arch/i386/kernel/ipipe.c Wed Dec 31 19:00:00 1969
+++ linux-2.5.38-adeos/arch/i386/kernel/ipipe.c Sun Sep 22 21:57:18 2002
@@ -0,0 +1,686 @@
+/*
+ * linux/arch/i386/kernel/ipipe.c
+ *
+ * Copyright (C) 2002 Philippe Gerum.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; either version 2 of the License, or (at your option) any later
+ * version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * This code implements the ADEOS interrupt pipeline mechanism.
+ */
+
+#include <linux/config.h>
+#include <linux/kernel.h>
+#include <linux/smp.h>
+#include <linux/sched.h>
+#include <linux/irq.h>
+#include <asm/system.h>
+#include <asm/hw_irq.h>
+#include <asm/irq.h>
+#include <asm/desc.h>
+#include <asm/errno.h>
+#include <asm/segment.h>
+#ifdef CONFIG_X86_LOCAL_APIC
+#include <asm/fixmap.h>
+#include <asm/bitops.h>
+#include <asm/mpspec.h>
+#ifdef CONFIG_X86_IO_APIC
+#include <asm/io_apic.h>
+#endif /* CONFIG_X86_IO_APIC */
+#include <asm/apic.h>
+#endif /* CONFIG_X86_LOCAL_APIC */
+
+extern struct desc_struct idt_table[];
+
+extern spinlock_t adeos_pipelock;
+
+extern struct list_head adeos_pipeline;
+
+/* Lifted from traps.c */
+#define ipipe_set_gate(gate_addr,type,dpl,addr) \
+do { \
+ int __d0, __d1; \
+ __asm__ __volatile__ ("movw %%dx,%%ax\n\t" \
+ "movw %4,%%dx\n\t" \
+ "movl %%eax,%0\n\t" \
+ "movl %%edx,%1" \
+ :"=m" (*((long *) (gate_addr))), \
+ "=m" (*(1+(long *) (gate_addr))), "=&a" (__d0), "=&d" (__d1) \
+ :"i" ((short) (0x8000+(dpl<<13)+(type<<8))), \
+ "3" ((char *) (addr)),"2" (__KERNEL_CS << 16)); \
+} while (0)
+
+#define ipipe_set_irq_gate(vector,addr) \
+ipipe_set_gate(idt_table+vector,14,0,addr)
+
+#define ipipe_set_irq_bit(adp,cpuid,irq) \
+do { \
+ set_bit(irq & 0x1f,&(adp)->cpudata[cpuid].irq_pending_lo[irq >> 5]); \
+ set_bit(irq >> 5,&(adp)->cpudata[cpuid].irq_pending_hi); \
+} while(0)
+
+extern void (*ipipe_irq_trampolines[])(void); /* in entry.S now */
+
+#define ipipe_call_asm_irq_handler(adp,irq) \
+ __asm__ __volatile__ ("pushfl\n\t" \
+ "push %%cs\n\t" \
+ "call *%1\n" \
+ : /* no output */ \
+ : "a" (irq), "m" ((adp)->irqs[irq].handler))
+
+#define ipipe_call_c_irq_handler(adp,irq) \
+ __asm__ __volatile__ ("pushl %%ebp\n\t" \
+ "pushl %%edi\n\t" \
+ "pushl %%esi\n\t" \
+ "pushl %%edx\n\t" \
+ "pushl %%ecx\n\t" \
+ "pushl %%ebx\n\t" \
+ "pushl %%eax\n\t" \
+ "call *%1\n\t" \
+ "addl $4,%%esp\n\t" \
+ "popl %%ebx\n\t" \
+ "popl %%ecx\n\t" \
+ "popl %%edx\n\t" \
+ "popl %%esi\n\t" \
+ "popl %%edi\n\t" \
+ "popl %%ebp\n" \
+ : /* no output */ \
+ : "a" (irq), "m" ((adp)->irqs[irq].handler))
+
+static unsigned long ipipe_virtual_irq_map;
+
+static __inline__ unsigned long ffnz (unsigned long word) {
+ /* Derived from bitops.h's ffs() */
+ __asm__("bsfl %1, %0"
+ : "=r" (word)
+ : "r" (word));
+ return word;
+}
+
+/* ipipe_sync_irqs() -- Flush the pending IRQs for the current domain
+ (and processor). This routine flushes the interrupt log (see
+ "Optimistic interrupt protection" from D. Stodolsky et al. for more
+ on the deferred interrupt mecanism). Every interrupt that occurred
+ while the pipeline was stalled gets played. */
+
+void ipipe_sync_irqs (void)
+
+{
+ unsigned long mask, submask;
+ struct adcpudata *cpudata;
+ ipipe_declare_cpuid;
+ int level, rank;
+ adomain_t *adp;
+ unsigned irq;
+
+ adp = adp_cpu_current[cpuid];
+ cpudata = &adp->cpudata[cpuid];
+
+ do
+ {
+ if (test_and_set_bit(IPIPE_SYNC_FLAG,&cpudata->status))
+ return;
+
+ set_bit(IPIPE_STALL_FLAG,&cpudata->status);
+
+ /* In order to update the pending irq mask atomically without
+ fiddling with the processor interrupt state, the irq flag
+ is first cleared, then reasserted if it appears that not
+ all hits have been processed yet. */
+
+ while ((mask = cpudata->irq_pending_hi) != 0)
+ {
+ level = ffnz(mask);
+ clear_bit(level,&cpudata->irq_pending_hi);
+
+ while ((submask = cpudata->irq_pending_lo[level]) != 0)
+ {
+ rank = ffnz(submask);
+ clear_bit(rank,&cpudata->irq_pending_lo[level]);
+ irq = (level << 5) + rank;
+
+ if (!atomic_dec_and_test(&cpudata->irq_hits[irq]))
+ {
+ set_bit(rank,&cpudata->irq_pending_lo[level]);
+ set_bit(level,&cpudata->irq_pending_hi);
+ }
+
+ /* Allow the sync routine to be recursively called on
+ behalf of the IRQ handler or even reentered on
+ behalf of another execution context switched in by
+ the IRQ handler. The latter also means that
+ returning from the switched out context is always
+ safe even if the sync routine has been reentered in
+ the meantime. */
+
+ clear_bit(IPIPE_SYNC_FLAG,&cpudata->status);
+
+ if (test_bit(IPIPE_CALLASM_FLAG,&adp->irqs[irq].control))
+ ipipe_call_asm_irq_handler(adp,irq);
+ else
+ ipipe_call_c_irq_handler(adp,irq);
+
+#ifdef CONFIG_SMP
+ /* If the execution context on entry has migrated to
+ another CPU, exit immediately and let the initial
+ CPU rerun the sync routine in its own context when
+ it sees fit (which it might be currently doing in
+ parallel to this CPU, who knows). */
+
+ if (cpuid != adeos_processor_id())
+ {
+ clear_bit(IPIPE_STALL_FLAG,&cpudata->status);
+ return;
+ }
+#endif /* CONFIG_SMP */
+
+ if (test_and_set_bit(IPIPE_SYNC_FLAG,&cpudata->status))
+ return;
+ }
+ }
+
+ clear_bit(IPIPE_STALL_FLAG,&cpudata->status);
+ clear_bit(IPIPE_SYNC_FLAG,&cpudata->status);
+ }
+ while (cpudata->irq_pending_hi != 0);
+}
+
+int ipipe_ack_common_irq (unsigned irq)
+
+{
+ irq_desc_t *desc = irq_desc + irq;
+ desc->handler->ack(irq);
+ return 1;
+}
+
+int ipipe_ack_apic_irq (unsigned irq) {
+#ifdef CONFIG_X86_LOCAL_APIC
+ ack_APIC_irq();
+#endif
+ return 1;
+}
+
+/* ipipe_handle_irq() -- ADEOS' generic IRQ handler. An optimistic
+ interrupt protection log is maintained here for each domain. */
+
+int ipipe_handle_irq (struct pt_regs regs)
+
+{
+ unsigned irq = regs.orig_eax & 0xff;
+ int acked = regs.orig_eax >= 0;
+ struct list_head *pos;
+ ipipe_declare_cpuid;
+
+ /* First ack, then propagate the interrupt to each domain's log. */
+
+ list_for_each(pos,&adeos_pipeline) {
+
+ adomain_t *_adp = list_entry(pos,adomain_t,link);
+
+ /* For each domain handling the incoming IRQ, mark it as
+ pending in its log. */
+
+ if (test_bit(IPIPE_HANDLE_FLAG,&_adp->irqs[irq].control))
+ {
+ /* Domains that handle this IRQ are polled for
+ acknowledging it by decreasing priority order. */
+
+ if (!acked)
+ acked = _adp->irqs[irq].acknowledge(irq);
+
+ atomic_inc(&_adp->cpudata[cpuid].irq_hits[irq]);
+ ipipe_set_irq_bit(_adp,cpuid,irq);
+ }
+
+ /* If the domain does not want the IRQ to be passed down the
+ interrupt pipe, exit the loop now. */
+
+ if (!test_bit(IPIPE_PASS_FLAG,&_adp->irqs[irq].control))
+ break;
+ }
+
+ /* Now walk the pipeline, yielding control to the highest priority
+ domain that has pending interrupt(s). This search does not go
+ beyond the current domain in the pipeline. To understand this
+ code properly, one must keep in mind that domains having a
+ higher priority than the current one are sleeping on the
+ adeos_suspend_domain() service. In addition, domains having a
+ lower priority have been preempted by an interrupt dispatched
+ to a more prioritary domain. Once the first and most prioritary
+ stage has been selected here, the subsequent stages will be
+ activated in turn when each visited domain calls
+ adeos_suspend_domain() to wake up its neighbour down the
+ pipeline. */
+
+ list_for_each(pos,&adeos_pipeline) {
+
+ adomain_t *_adp = list_entry(pos,adomain_t,link);
+
+ if (test_bit(IPIPE_STALL_FLAG,&_adp->cpudata[cpuid].status))
+ break; /* Stalled stage -- do not go further. */
+
+ if (_adp->cpudata[cpuid].irq_pending_hi != 0)
+ {
+ /* Since the critical IPI might be serviced by the
+ following actions, the current domain might not be
+ linked to the pipeline anymore after its handler
+ returns on SMP boxes, even if the domain remains valid
+ (see adeos_unregister_domain()), so don't make any
+ hazardous assumptions here. */
+
+ if (_adp == adp_cpu_current[cpuid])
+ ipipe_sync_irqs();
+ else
+ {
+ __adeos_switch_domain(_adp);
+ ipipe_reload_cpuid(); /* Processor might have changed. */
+ }
+
+ break;
+ }
+ else if (_adp == adp_cpu_current[cpuid])
+ break;
+ }
+
+ return (adp_cpu_current[cpuid] == adp_root &&
+ !test_bit(IPIPE_STALL_FLAG,&adp_root->cpudata[cpuid].status));
+}
+
+int ipipe_hook_irq (unsigned irq,
+ void (*handler)(unsigned irq),
+ int (*acknowledge)(unsigned irq),
+ unsigned modemask)
+{
+ unsigned long flags;
+
+ ipipe_hw_spin_lock(&adeos_pipelock,flags);
+
+ if (handler != NULL)
+ {
+ if ((modemask & IPIPE_EXCLUSIVE_MASK) != 0)
+ {
+ if (adp_current->irqs[irq].handler != NULL)
+ return -EBUSY;
+ }
+ }
+ else
+ modemask &= ~IPIPE_HANDLE_MASK;
+
+ if (acknowledge == NULL)
+ acknowledge = adp_root->irqs[irq].acknowledge;
+
+ adp_current->irqs[irq].handler = handler;
+ adp_current->irqs[irq].acknowledge = acknowledge;
+ adp_current->irqs[irq].control = modemask;
+
+ if (!ipipe_virtual_irq_p(irq) && handler != NULL)
+ {
+ irq_desc_t *desc = irq_desc + irq;
+
+ ipipe_set_irq_gate(irq + FIRST_EXTERNAL_VECTOR,
+ ipipe_irq_trampolines[irq]);
+
+ if ((modemask & IPIPE_ENABLE_MASK) != 0)
+ desc->handler->enable(irq);
+ }
+
+ ipipe_hw_spin_unlock(&adeos_pipelock,flags);
+
+ return 0;
+}
+
+int ipipe_control_irq (unsigned irq,
+ unsigned clrmask,
+ unsigned setmask)
+{
+ irq_desc_t *desc = irq_desc + irq;
+ unsigned long flags;
+
+ ipipe_hw_spin_lock(&adeos_pipelock,flags);
+
+ if (adp_current->irqs[irq].handler == NULL)
+ setmask &= ~IPIPE_HANDLE_MASK;
+
+ adp_current->irqs[irq].control &= ~clrmask;
+ adp_current->irqs[irq].control |= setmask;
+
+ if ((setmask & IPIPE_ENABLE_MASK) != 0)
+ desc->handler->enable(irq);
+ else if ((clrmask & IPIPE_ENABLE_MASK) != 0)
+ desc->handler->disable(irq);
+
+ ipipe_hw_spin_unlock(&adeos_pipelock,flags);
+
+ return 0;
+}
+
+int ipipe_propagate_irq (unsigned irq)
+
+{
+ struct list_head *ln;
+ unsigned long flags;
+ ipipe_declare_cpuid;
+
+ ipipe_hw_spin_lock(&adeos_pipelock,flags);
+
+ ln = adp_current->link.next;
+
+ while (ln != &adeos_pipeline)
+ {
+ adomain_t *adp = list_entry(ln,adomain_t,link);
+
+ if (test_bit(IPIPE_HANDLE_FLAG,&adp->irqs[irq].control))
+ {
+ atomic_inc(&adp->cpudata[cpuid].irq_hits[irq]);
+ ipipe_set_irq_bit(adp,cpuid,irq);
+ ipipe_hw_spin_unlock(&adeos_pipelock,flags);
+ return 1;
+ }
+
+ ln = adp->link.next;
+ }
+
+ ipipe_hw_spin_unlock(&adeos_pipelock,flags);
+
+ return 0;
+}
+
+unsigned ipipe_alloc_irq (void)
+
+{
+ unsigned long flags, irq = 0;
+ int ipos;
+
+ ipipe_hw_spin_lock(&adeos_pipelock,flags);
+
+ if (ipipe_virtual_irq_map != ~0)
+ {
+ ipos = ffz(ipipe_virtual_irq_map);
+ set_bit(ipos,&ipipe_virtual_irq_map);
+ irq = ipos + IPIPE_VIRQ_BASE;
+ }
+
+ ipipe_hw_spin_unlock(&adeos_pipelock,flags);
+
+ return irq;
+}
+
+void ipipe_free_irq (unsigned irq)
+
+{
+ int ipos = irq - IPIPE_VIRQ_BASE;
+ clear_bit(ipos,&ipipe_virtual_irq_map);
+}
+
+int ipipe_trigger_irq (unsigned irq)
+
+{
+ struct pt_regs regs;
+
+ if (ipipe_virtual_irq_p(irq) &&
+ !test_bit(irq - IPIPE_VIRQ_BASE,&ipipe_virtual_irq_map))
+ return -EINVAL;
+
+ regs.orig_eax = irq; /* Won't be acked */
+ return ipipe_handle_irq(regs);
+}
+
+int ipipe_trigger_ipi (int _cpuid)
+
+{
+#ifdef CONFIG_SMP
+ ipipe_declare_cpuid;
+
+ if (_cpuid == cpuid) /* Self-posting the service interrupt? */
+ ipipe_trigger_irq(ADEOS_SERVICE_IPI);
+ else
+ {
+ unsigned long flags = adeos_test_and_stall_ipipe();
+
+ if (cpuid == ADEOS_OTHER_CPUS)
+ {
+ if (__adeos_online_cpus > 1)
+ /* Send the service IPI to all processors but the current one. */
+ __adeos_send_IPI_allbutself(ADEOS_SERVICE_VECTOR);
+ }
+ else if (_cpuid >= 0 && _cpuid < __adeos_online_cpus)
+ __adeos_send_IPI_other(_cpuid,ADEOS_SERVICE_VECTOR);
+
+ adeos_restore_ipipe(flags);
+ }
+
+ return cpuid != ADEOS_OTHER_CPUS ? 1 : __adeos_online_cpus - 1;
+#else /* !CONFIG_SMP */
+ return 0;
+#endif /* CONFIG_SMP */
+}
+
+void ipipe_init_stage (adomain_t *adp)
+
+{
+ int cpuid, n;
+
+ for (cpuid = 0; cpuid < ADEOS_NR_CPUS; cpuid++)
+ {
+ adp->cpudata[cpuid].irq_pending_hi = 0;
+
+ for (n = 0; n < IPIPE_IRQ_IWORDS; n++)
+ adp->cpudata[cpuid].irq_pending_lo[n] = 0;
+
+ for (n = 0; n < IPIPE_NR_IRQS; n++)
+ atomic_set(&adp->cpudata[cpuid].irq_hits[n],0);
+ }
+
+ for (n = 0; n < IPIPE_NR_IRQS; n++)
+ {
+ adp->irqs[n].acknowledge = NULL;
+ adp->irqs[n].handler = NULL;
+ adp->irqs[n].control = IPIPE_PASS_MASK; /* Pass but don't handle */
+ }
+
+#ifdef CONFIG_SMP
+ ipipe_hook_irq(ADEOS_CRITICAL_IPI,
+ &__adeos_do_critical_sync,
+ &ipipe_ack_apic_irq,
+ IPIPE_HANDLE_MASK); /* Handle but *never* pass */
+#endif /* CONFIG_SMP */
+}
+
+#define BUILD_TRAP_PROTO(trapnr) ipipe_trap_##trapnr(void)
+
+#define BUILD_TRAP_ERRCODE(trapnr) \
+asmlinkage void BUILD_TRAP_PROTO(trapnr); \
+__asm__ ( \
+ "\n" __ALIGN_STR"\n\t" \
+ "ipipe_trap_" #trapnr ":\n\t" \
+ "cld\n\t" \
+ "pushl %es\n\t" \
+ "pushl %ds\n\t" \
+ "pushl %eax\n\t" \
+ "pushl %ebp\n\t" \
+ "pushl %edi\n\t" \
+ "pushl %esi\n\t" \
+ "pushl %edx\n\t" \
+ "pushl %ecx\n\t" \
+ "pushl %ebx\n\t" \
+ "movl $"STR(__KERNEL_DS)",%edx\n\t" \
+ "mov %dx,%ds\n\t" \
+ "mov %dx,%es\n\t" \
+ "movl (__adeos_event_monitors + 4 * " #trapnr "),%eax\n\t" \
+ "testl %eax,%eax\n\t" \
+ "jz 1f\n\t" \
+ "movl %esp,%eax\n\t" \
+ "pushl %eax\n\t" \
+ "pushl $"#trapnr"\n\t" \
+ "call ipipe_handle_event\n\t" \
+ "addl $8,%esp\n\t" \
+ "testl %eax,%eax\n\t" \
+"1: popl %ebx\n\t" \
+ "popl %ecx\n\t" \
+ "popl %edx\n\t" \
+ "popl %esi\n\t" \
+ "popl %edi\n\t" \
+ "popl %ebp\n\t" \
+ "jz 2f\n\t" \
+ "popl %eax\n\t" \
+ "popl %ds\n\t" \
+ "popl %es\n\t" \
+ "addl $4,%esp\n\t" \
+ "iret\n" \
+"2: movl (ipipe_root_traps + 4 * " #trapnr "),%eax\n\t" \
+ "movl 8(%esp),%es\n\t" \
+ "movl %eax,8(%esp)\n\t" \
+ "popl %eax\n\t" \
+ "popl %ds\n\t" \
+ "ret\n");
+
+#define BUILD_TRAP_NOERRCODE(trapnr) \
+asmlinkage void BUILD_TRAP_PROTO(trapnr); \
+__asm__ ( \
+ "\n" __ALIGN_STR"\n\t" \
+ "ipipe_trap_" #trapnr ":\n\t" \
+ "cld\n\t" \
+ "pushl $0\n\t" \
+ "pushl %es\n\t" \
+ "pushl %ds\n\t" \
+ "pushl %eax\n\t" \
+ "pushl %ebp\n\t" \
+ "pushl %edi\n\t" \
+ "pushl %esi\n\t" \
+ "pushl %edx\n\t" \
+ "pushl %ecx\n\t" \
+ "pushl %ebx\n\t" \
+ "movl $"STR(__KERNEL_DS)",%edx\n\t" \
+ "mov %dx,%ds\n\t" \
+ "mov %dx,%es\n\t" \
+ "movl (__adeos_event_monitors + 4 * " #trapnr "),%eax\n\t" \
+ "testl %eax,%eax\n\t" \
+ "jz 1f\n\t" \
+ "movl %esp,%eax\n\t" \
+ "pushl %eax\n\t" \
+ "pushl $"#trapnr"\n\t" \
+ "call ipipe_handle_event\n\t" \
+ "addl $8,%esp\n\t" \
+ "testl %eax,%eax\n\t" \
+"1: popl %ebx\n\t" \
+ "popl %ecx\n\t" \
+ "popl %edx\n\t" \
+ "popl %esi\n\t" \
+ "popl %edi\n\t" \
+ "popl %ebp\n\t" \
+ "jz 2f\n\t" \
+ "popl %eax\n\t" \
+ "popl %ds\n\t" \
+ "popl %es\n\t" \
+ "addl $4,%esp\n\t" \
+ "iret\n" \
+"2: movl (ipipe_root_traps + 4 * " #trapnr "),%eax\n\t" \
+ "movl %eax,12(%esp)\n\t" \
+ "popl %eax\n\t" \
+ "popl %ds\n\t" \
+ "popl %es\n\t" \
+ "ret\n");
+
+void (*ipipe_root_traps[ADEOS_NR_FAULTS])(void);
+
+BUILD_TRAP_NOERRCODE(0x0) BUILD_TRAP_NOERRCODE(0x1) BUILD_TRAP_NOERRCODE(0x2)
+BUILD_TRAP_NOERRCODE(0x3) BUILD_TRAP_NOERRCODE(0x4) BUILD_TRAP_NOERRCODE(0x5)
+BUILD_TRAP_NOERRCODE(0x6) BUILD_TRAP_NOERRCODE(0x7) BUILD_TRAP_ERRCODE(0x8)
+BUILD_TRAP_NOERRCODE(0x9) BUILD_TRAP_ERRCODE(0xa) BUILD_TRAP_ERRCODE(0xb)
+BUILD_TRAP_ERRCODE(0xc) BUILD_TRAP_ERRCODE(0xd) BUILD_TRAP_ERRCODE(0xe)
+BUILD_TRAP_NOERRCODE(0xf) BUILD_TRAP_NOERRCODE(0x10) BUILD_TRAP_ERRCODE(0x11)
+BUILD_TRAP_NOERRCODE(0x12) BUILD_TRAP_NOERRCODE(0x13) BUILD_TRAP_ERRCODE(0x14)
+BUILD_TRAP_ERRCODE(0x15) BUILD_TRAP_ERRCODE(0x16) BUILD_TRAP_ERRCODE(0x17)
+BUILD_TRAP_ERRCODE(0x18) BUILD_TRAP_ERRCODE(0x19) BUILD_TRAP_ERRCODE(0x1a)
+BUILD_TRAP_ERRCODE(0x1b) BUILD_TRAP_ERRCODE(0x1c) BUILD_TRAP_ERRCODE(0x1d)
+BUILD_TRAP_ERRCODE(0x1e) BUILD_TRAP_ERRCODE(0x1f)
+
+#define LIST_TRAP(trapnr) &ipipe_trap_ ## trapnr
+
+void (*ipipe_trap_trampolines[])(void) = {
+ LIST_TRAP(0x0), LIST_TRAP(0x1), LIST_TRAP(0x2), LIST_TRAP(0x3),
+ LIST_TRAP(0x4), LIST_TRAP(0x5), LIST_TRAP(0x6), LIST_TRAP(0x7),
+ LIST_TRAP(0x8), LIST_TRAP(0x9), LIST_TRAP(0xa), LIST_TRAP(0xb),
+ LIST_TRAP(0xc), LIST_TRAP(0xd), LIST_TRAP(0xe), LIST_TRAP(0xf),
+ LIST_TRAP(0x10), LIST_TRAP(0x11), LIST_TRAP(0x12), LIST_TRAP(0x13),
+ LIST_TRAP(0x14), LIST_TRAP(0x15), LIST_TRAP(0x16), LIST_TRAP(0x17),
+ LIST_TRAP(0x18), LIST_TRAP(0x19), LIST_TRAP(0x1a), LIST_TRAP(0x1b),
+ LIST_TRAP(0x1c), LIST_TRAP(0x1d), LIST_TRAP(0x1e), LIST_TRAP(0x1f)
+};
+
+int ipipe_handle_event(unsigned event,
+ struct pt_regs *regs) __attribute__ ((__unused__));
+
+/* ipipe_handle_event() -- Adeos' generic event handler. This routine
+calls the per-domain handlers registered for a given
+exception/event. Each domain before the one which raised the event in
+the pipeline will get a chance to process the event. The latter will
+eventually be allowed to process its own event too if a valid handler
+exists for it. Handler executions are always scheduled by the domain
+which raised the event for the prioritary domains wanting to be
+notified of such event. Note: regs might be NULL. */
+
+int ipipe_handle_event (unsigned event, struct pt_regs *regs)
+
+{
+ struct list_head *pos, *npos;
+ unsigned long flags;
+ ipipe_declare_cpuid;
+ adevinfo_t evinfo;
+ int propagate = 1;
+
+ flags = adeos_test_and_stall_ipipe();
+
+ list_for_each_safe(pos,npos,&adeos_pipeline) {
+
+ adomain_t *adp = list_entry(pos,adomain_t,link);
+
+ if (adp->events[event].handler != NULL)
+ {
+ if (adp == adp_cpu_current[cpuid])
+ {
+ evinfo.domid = adp->domid;
+ evinfo.event = event;
+ evinfo.regs = regs;
+ evinfo.propagate = 0;
+ adp->events[event].handler(&evinfo);
+ propagate = evinfo.propagate;
+ break;
+ }
+
+ adp->cpudata[cpuid].event_data.domid = adp_cpu_current[cpuid]->domid;
+ adp->cpudata[cpuid].event_data.event = event;
+ adp->cpudata[cpuid].event_data.regs = regs;
+ adp->cpudata[cpuid].event_data.propagate = 0;
+ set_bit(IPIPE_XPEND_FLAG,&adp->cpudata[cpuid].status);
+
+ /* Let the prioritary domain process the event. */
+ __adeos_switch_domain(adp);
+
+ if (!adp->cpudata[cpuid].event_data.propagate)
+ {
+ propagate = 0;
+ break;
+ }
+ }
+
+ if (adp == adp_cpu_current[cpuid])
+ break;
+ }
+
+ adeos_restore_ipipe(flags);
+
+ return !propagate;
+}
diff -urpN linux-2.5.38/arch/i386/kernel/irq.c linux-2.5.38-adeos/arch/i386/kernel/irq.c
--- linux-2.5.38/arch/i386/kernel/irq.c Sun Sep 22 00:25:00 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/irq.c Sun Sep 22 21:57:18 2002
@@ -332,7 +332,10 @@ asmlinkage unsigned int do_IRQ(struct pt
irq_enter();
kstat.irqs[cpu][irq]++;
spin_lock(&desc->lock);
+#ifndef CONFIG_ADEOS
+ /* IRQ acknowledging is in the hands of the IPIPE exclusively. */
desc->handler->ack(irq);
+#endif
/*
REPLAY is when Linux resends an IRQ that was dropped earlier
WAITING is used by probe to mark irqs that are being tested
diff -urpN linux-2.5.38/arch/i386/kernel/process.c linux-2.5.38-adeos/arch/i386/kernel/process.c
--- linux-2.5.38/arch/i386/kernel/process.c Sun Sep 22 00:24:57 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/process.c Sun Sep 22 21:57:18 2002
@@ -137,6 +137,9 @@ void cpu_idle (void)
if (!idle)
idle = default_idle;
irq_stat[smp_processor_id()].idle_timestamp = jiffies;
+#ifdef CONFIG_ADEOS
+ adeos_suspend_domain();
+#endif /* CONFIG_ADEOS */
while (!need_resched())
idle();
schedule();
diff -urpN linux-2.5.38/arch/i386/kernel/smp.c linux-2.5.38-adeos/arch/i386/kernel/smp.c
--- linux-2.5.38/arch/i386/kernel/smp.c Sun Sep 22 00:25:00 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/smp.c Sun Sep 22 21:57:18 2002
@@ -159,7 +159,11 @@ static inline void send_IPI_mask_bitmask
unsigned long cfg;
unsigned long flags;

+#ifdef CONFIG_ADEOS
+ ipipe_hw_save_flags_and_cli(flags);
+#else
local_irq_save(flags);
+#endif

/*
* Wait for idle.
@@ -182,7 +186,11 @@ static inline void send_IPI_mask_bitmask
*/
apic_write_around(APIC_ICR, cfg);

+#ifdef CONFIG_ADEOS
+ ipipe_hw_restore_flags(flags);
+#else
local_irq_restore(flags);
+#endif
}

static inline void send_IPI_mask_sequence(int mask, int vector)
@@ -625,7 +633,10 @@ void smp_send_stop(void)
*/
asmlinkage void smp_reschedule_interrupt(void)
{
+#ifndef CONFIG_ADEOS
+ /* IRQ acknowledging is in the hands of the IPIPE exclusively. */
ack_APIC_irq();
+#endif
}

asmlinkage void smp_call_function_interrupt(void)
@@ -634,7 +645,10 @@ asmlinkage void smp_call_function_interr
void *info = call_data->info;
int wait = call_data->wait;

+#ifndef CONFIG_ADEOS
+ /* IRQ acknowledging is in the hands of the IPIPE exclusively. */
ack_APIC_irq();
+#endif
/*
* Notify initiating CPU that I've grabbed the data and am
* about to execute the function
@@ -654,3 +668,14 @@ asmlinkage void smp_call_function_interr
}
}

+#ifdef CONFIG_ADEOS
+
+void __adeos_send_IPI_allbutself (int vector) {
+ send_IPI_allbutself(vector);
+}
+
+void __adeos_send_IPI_other (int cpu, int vector) {
+ send_IPI_mask(1 << cpu,vector);
+}
+
+#endif /* CONFIG_ADEOS */
diff -urpN linux-2.5.38/arch/i386/kernel/time.c linux-2.5.38-adeos/arch/i386/kernel/time.c
--- linux-2.5.38/arch/i386/kernel/time.c Sun Sep 22 00:25:10 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/time.c Sun Sep 22 21:57:18 2002
@@ -370,11 +370,20 @@ static inline void do_timer_interrupt(in
* This will also deassert NMI lines for the watchdog if run
* on an 82489DX-based system.
*/
+#ifdef CONFIG_ADEOS
+ unsigned long flags;
+ ipipe_hw_spin_lock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_lock(&i8259A_lock);
+#endif /* CONFIG_ADEOS */
outb(0x0c, 0x20);
/* Ack the IRQ; AEOI will end it automatically. */
inb(0x20);
+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_unlock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_unlock(&i8259A_lock);
+#endif /* CONFIG_ADEOS */
}
#endif

@@ -450,6 +459,7 @@ void timer_interrupt(int irq, void *dev_

rdtscl(last_tsc_low);

+#ifndef CONFIG_ADEOS
spin_lock(&i8253_lock);
outb_p(0x00, 0x43); /* latch the count ASAP */

@@ -459,6 +469,7 @@ void timer_interrupt(int irq, void *dev_

count = ((LATCH-1) - count) * TICK_SIZE;
delay_at_last_interrupt = (count + LATCH/2) / LATCH;
+#endif /* !CONFIG_ADEOS */
}

do_timer_interrupt(irq, NULL, regs);
diff -urpN linux-2.5.38/arch/i386/kernel/traps.c linux-2.5.38-adeos/arch/i386/kernel/traps.c
--- linux-2.5.38/arch/i386/kernel/traps.c Sun Sep 22 00:25:02 2002
+++ linux-2.5.38-adeos/arch/i386/kernel/traps.c Sun Sep 22 21:57:18 2002
@@ -815,6 +815,72 @@ do { \
* Pentium F0 0F bugfix can have resulted in the mapped
* IDT being write-protected.
*/
+#ifdef CONFIG_ADEOS
+
+extern irq_desc_t irq_desc[];
+
+extern int ipipe_ack_common_irq(unsigned irq);
+
+extern int ipipe_ack_apic_irq(unsigned irq);
+
+extern void (*ipipe_root_traps[])(void);
+
+extern void (*ipipe_trap_trampolines[])(void);
+
+void set_intr_gate(unsigned n, void *addr)
+
+{
+ /* Redirect external interrupts only but the 'TLB invalidate'
+ since we don't allow to virtualize it. */
+ if (n < FIRST_EXTERNAL_VECTOR || n == INVALIDATE_TLB_VECTOR)
+ {
+ if (n == 2 || n == 14) /* Intercept NMI and Page Fault */
+ {
+ ipipe_root_traps[n] = (void (*)(void))addr;
+ addr = (void *)ipipe_trap_trampolines[n];
+ }
+
+ _set_gate(idt_table+n,14,0,addr);
+ }
+ else
+ {
+ unsigned irq = n - FIRST_EXTERNAL_VECTOR;
+ int (*ack)(unsigned);
+
+ if (n < FIRST_DEVICE_VECTOR)
+ ack = &ipipe_ack_common_irq; /* Common ISA interrupts */
+ else
+ ack = &ipipe_ack_apic_irq; /* APIC/SMP interrupts */
+
+ adeos_virtualize_irq(irq,
+ (void (*)(unsigned))addr,
+ ack,
+ IPIPE_CALLASM_MASK|IPIPE_HANDLE_MASK|IPIPE_PASS_MASK);
+ }
+}
+
+static void __init set_trap_gate(unsigned int n, void *addr)
+
+{
+ ipipe_root_traps[n] = (void (*)(void))addr;
+ addr = (void *)ipipe_trap_trampolines[n];
+ _set_gate(idt_table+n,15,0,addr);
+}
+
+static void __init set_system_gate(unsigned int n, void *addr)
+
+{
+ if (n < FIRST_EXTERNAL_VECTOR) /* Trap exceptions only */
+ {
+ ipipe_root_traps[n] = (void (*)(void))addr;
+ addr = (void *)ipipe_trap_trampolines[n];
+ }
+
+ _set_gate(idt_table+n,15,3,addr);
+}
+
+#else /* !CONFIG_ADEOS */
+
void set_intr_gate(unsigned int n, void *addr)
{
_set_gate(idt_table+n,14,0,addr);
@@ -829,6 +895,8 @@ static void __init set_system_gate(unsig
{
_set_gate(idt_table+n,15,3,addr);
}
+
+#endif /* CONFIG_ADEOS */

static void __init set_call_gate(void *a, void *addr)
{
diff -urpN linux-2.5.38/arch/i386/mach-generic/do_timer.h linux-2.5.38-adeos/arch/i386/mach-generic/do_timer.h
--- linux-2.5.38/arch/i386/mach-generic/do_timer.h Sun Sep 22 00:25:18 2002
+++ linux-2.5.38-adeos/arch/i386/mach-generic/do_timer.h Sun Sep 22 21:57:18 2002
@@ -46,13 +46,22 @@ static inline int do_timer_overflow(int
{
int i;

+#ifdef CONFIG_ADEOS
+ unsigned long flags;
+ ipipe_hw_spin_lock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_lock(&i8259A_lock);
+#endif /* CONFIG_ADEOS */
/*
* This is tricky when I/O APICs are used;
* see do_timer_interrupt().
*/
i = inb(0x20);
+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_unlock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_unlock(&i8259A_lock);
+#endif /* CONFIG_ADEOS */

/* assumption about timer being IRQ0 */
if (i & 0x01) {
diff -urpN linux-2.5.38/arch/i386/mach-visws/do_timer.h linux-2.5.38-adeos/arch/i386/mach-visws/do_timer.h
--- linux-2.5.38/arch/i386/mach-visws/do_timer.h Sun Sep 22 00:25:05 2002
+++ linux-2.5.38-adeos/arch/i386/mach-visws/do_timer.h Sun Sep 22 21:57:18 2002
@@ -27,13 +27,22 @@ static inline int do_timer_overflow(int
{
int i;

+#ifdef CONFIG_ADEOS
+ unsigned long flags;
+ ipipe_hw_spin_lock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_lock(&i8259A_lock);
+#endif /* CONFIG_ADEOS */
/*
* This is tricky when I/O APICs are used;
* see do_timer_interrupt().
*/
i = inb(0x20);
+#ifdef CONFIG_ADEOS
+ ipipe_hw_spin_unlock(&i8259A_lock, flags);
+#else /* !CONFIG_ADEOS */
spin_unlock(&i8259A_lock);
+#endif /* CONFIG_ADEOS */

/* assumption about timer being IRQ0 */
if (i & 0x01) {
diff -urpN linux-2.5.38/arch/i386/mm/fault.c linux-2.5.38-adeos/arch/i386/mm/fault.c
--- linux-2.5.38/arch/i386/mm/fault.c Sun Sep 22 00:24:57 2002
+++ linux-2.5.38-adeos/arch/i386/mm/fault.c Sun Sep 22 21:57:18 2002
@@ -157,7 +157,11 @@ asmlinkage void do_page_fault(struct pt_

/* It's safe to allow irq's after cr2 has been saved */
if (regs->eflags & X86_EFLAGS_IF)
- local_irq_enable();
+#ifdef CONFIG_ADEOS
+ ipipe_hw_sti();
+#else
+ local_irq_enable();
+#endif

tsk = current;

diff -urpN linux-2.5.38/include/asm-i386/adeos.h linux-2.5.38-adeos/include/asm-i386/adeos.h
--- linux-2.5.38/include/asm-i386/adeos.h Wed Dec 31 19:00:00 1969
+++ linux-2.5.38-adeos/include/asm-i386/adeos.h Sun Sep 22 21:57:18 2002
@@ -0,0 +1,242 @@
+/*
+ * include/asm-i386/adeos.h
+ *
+ * Copyright (C) 2002 Philippe Gerum.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; either version 2 of the License, or (at your option) any later
+ * version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ */
+
+#ifndef __I386_ADEOS_H
+#define __I386_ADEOS_H
+
+#include <asm/atomic.h>
+#include <linux/version.h>
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,37)
+#include <asm/irq_vectors.h>
+#else
+#include <irq_vectors.h>
+#endif
+#include <asm/siginfo.h>
+#include <asm/ptrace.h>
+#include <linux/list.h>
+#include <linux/threads.h>
+
+#ifdef CONFIG_SMP
+
+#include <asm/fixmap.h>
+#include <asm/apicdef.h>
+
+extern unsigned long cpu_online_map;
+#define smp_num_cpus hweight32(cpu_online_map)
+
+#define ADEOS_NR_CPUS NR_CPUS
+#define ADEOS_CRITICAL_VECTOR 0xe1 /* Used by __adeos_critical_enter/exit */
+#define ADEOS_SERVICE_VECTOR 0xe9
+#define ADEOS_CRITICAL_IPI (ADEOS_CRITICAL_VECTOR - FIRST_EXTERNAL_VECTOR)
+#define ADEOS_SERVICE_IPI (ADEOS_SERVICE_VECTOR - FIRST_EXTERNAL_VECTOR)
+
+static __inline int adeos_smp_apic_id(void) {
+ return GET_APIC_ID(*(unsigned long *)(APIC_BASE+APIC_ID));
+}
+
+/* Cannot include asm/smpboot.h safely, so we define here the little
+ we need to know about CPU/apicid mappings. */
+
+#ifdef CONFIG_MULTIQUAD
+extern volatile int logical_apicid_2_cpu[];
+#define adeos_processor_id() (__adeos_online_cpus > 1 ? logical_apicid_2_cpu[adeos_smp_apic_id()] : 0)
+#else /* !CONFIG_MULTIQUAD */ /* use physical IDs to bootstrap */
+extern volatile int physical_apicid_2_cpu[];
+#define adeos_processor_id() (__adeos_online_cpus > 1 ? physical_apicid_2_cpu[adeos_smp_apic_id()] : 0)
+#endif /* CONFIG_MULTIQUAD */
+
+#define ipipe_declare_cpuid int cpuid = adeos_processor_id()
+#define ipipe_reload_cpuid() do { cpuid = adeos_processor_id(); } while(0)
+
+#else /* !CONFIG_SMP */
+
+#define ADEOS_NR_CPUS 1
+#define smp_num_cpus 1
+#define adeos_processor_id() 0
+/* Array references using this index should be optimized out. */
+#define ipipe_declare_cpuid const int cpuid = 0
+#define ipipe_reload_cpuid() /* nop */
+
+#endif /* CONFIG_SMP */
+
+#define ADEOS_NR_FAULTS (32) /* IDT fault vectors */
+#define ADEOS_NR_EVENTS (ADEOS_NR_FAULTS + 2) /* faults + 2 system events */
+
+/* Pseudo-vectors for syscalls entry/exit (std + lcall7 + lcall27) */
+#define ADEOS_SYSCALL_PROLOGUE 32
+#define ADEOS_SYSCALL_EPILOGUE 33
+
+typedef struct adevinfo {
+
+ unsigned domid;
+ unsigned event;
+ struct pt_regs *regs;
+
+ int propagate; /* Private */
+
+} adevinfo_t;
+
+typedef struct adsysinfo {
+
+ struct {
+ unsigned irq; /* Real timer tick IRQ */
+ } timer;
+
+ int ncpus; /* Number of CPUs on board */
+
+ unsigned long cpufreq; /* CPU frequency (in Hz) */
+
+} adsysinfo_t;
+
+/* Number of virtual IRQs */
+#define IPIPE_NR_VIRQS 32
+/* First virtual IRQ # */
+#define IPIPE_VIRQ_BASE NR_IRQS
+/* IRQs + VIRQs */
+#define IPIPE_NR_IRQS (NR_IRQS + IPIPE_NR_VIRQS)
+/* Number of indirect words needed to map the whole IRQ space. */
+#define IPIPE_IRQ_IWORDS (IPIPE_NR_IRQS / 32)
+
+typedef struct adomain {
+
+ /* -- Section: offset-based references are made on these fields
+ from inline assembly code. Please don't move or reorder. */
+ int *esp; /* Domain stack point */
+ void (*dswitch)(void); /* Domain switch hook */
+ /* -- End of section. */
+
+ int *estackbase;
+ unsigned domid;
+ const char *name;
+ int priority;
+
+ struct adcpudata {
+ unsigned long status;
+ unsigned long irq_pending_hi;
+ unsigned long irq_pending_lo[IPIPE_IRQ_IWORDS];
+ atomic_t irq_hits[IPIPE_NR_IRQS];
+ adevinfo_t event_data;
+ } cpudata[ADEOS_NR_CPUS];
+
+ struct {
+ int (*acknowledge)(unsigned irq);
+ void (*handler)(unsigned irq);
+ unsigned long control;
+ } irqs[IPIPE_NR_IRQS];
+
+ struct {
+ void (*handler)(adevinfo_t *evinfo);
+ } events[ADEOS_NR_EVENTS];
+
+ int ptd_keymax;
+ int ptd_keycount;
+ unsigned long ptd_keymap;
+ void (*ptd_setfun)(int, void *);
+ void *(*ptd_getfun)(int);
+
+ struct list_head link;
+
+} adomain_t;
+
+#define ipipe_virtual_irq_p(irq) ((irq) >= IPIPE_VIRQ_BASE && \
+ (irq) < IPIPE_NR_IRQS)
+
+#define ipipe_hw_spin_lock_init(x) spin_lock_init(x)
+#define ipipe_hw_spin_lock(x,flags) do { ipipe_hw_local_irq_save(flags); spin_lock(x); } while (0)
+#define ipipe_hw_spin_unlock(x,flags) do { spin_unlock(x); ipipe_hw_local_irq_restore(flags); preempt_enable(); } while (0)
+#define ipipe_hw_spin_lock_disable(x) do { ipipe_hw_cli(); spin_lock(x); } while (0)
+#define ipipe_hw_spin_unlock_enable(x) do { spin_unlock(x); ipipe_hw_sti(); preempt_enable(); } while (0)
+#define ipipe_hw_spin_lock_simple(x) spin_lock(x)
+#define ipipe_hw_spin_unlock_simple(x) spin_unlock(x)
+
+#ifndef STR
+#define __STR(x) #x
+#define STR(x) __STR(x)
+#endif
+
+/* Private interface -- Internal use only */
+
+struct adattr;
+
+void __adeos_init(void);
+
+#ifdef CONFIG_PROC_FS
+void __adeos_init_proc(void);
+#endif /* CONFIG_PROC_FS */
+
+void __adeos_bootstrap_domain(adomain_t *adp,
+ struct adattr *attr);
+
+void __adeos_cleanup_domain(adomain_t *adp);
+
+void __adeos_switch_domain(adomain_t *adp);
+
+unsigned long __adeos_critical_enter(void (*sync)(void));
+
+void __adeos_critical_exit(unsigned long flags);
+
+int __adeos_get_sysinfo(adsysinfo_t *info);
+
+int __adeos_tune_timer(unsigned long ns,
+ int flags);
+
+void __adeos_set_root_ptd(int key,
+ void *value);
+
+void *__adeos_get_root_ptd(int key);
+
+void __adeos_do_critical_sync(unsigned irq);
+
+void __adeos_send_IPI_allbutself(int vector);
+
+void __adeos_send_IPI_other(int cpu, int vector);
+
+void ipipe_init_stage(adomain_t *adp);
+
+int ipipe_hook_irq(unsigned irq,
+ void (*handler)(unsigned irq),
+ int (*acknowledge)(unsigned irq),
+ unsigned modemask);
+
+int ipipe_control_irq(unsigned irq,
+ unsigned clrmask,
+ unsigned setmask);
+
+unsigned ipipe_alloc_irq(void);
+
+void ipipe_free_irq(unsigned irq);
+
+int FASTCALL(ipipe_trigger_irq(unsigned irq));
+
+int FASTCALL(ipipe_trigger_ipi(int cpuid));
+
+int FASTCALL(ipipe_propagate_irq(unsigned irq));
+
+void ipipe_sync_irqs(void);
+
+int ipipe_handle_event(unsigned event,
+ struct pt_regs *regs);
+
+extern int __adeos_online_cpus;
+
+extern int __adeos_event_monitors[];
+
+#endif /* !__I386_ADEOS_H */
diff -urpN linux-2.5.38/include/asm-i386/system.h linux-2.5.38-adeos/include/asm-i386/system.h
--- linux-2.5.38/include/asm-i386/system.h Sun Sep 22 00:24:58 2002
+++ linux-2.5.38-adeos/include/asm-i386/system.h Sun Sep 22 21:57:18 2002
@@ -306,6 +306,53 @@ static inline unsigned long __cmpxchg(vo
#define set_wmb(var, value) do { var = value; wmb(); } while (0)

/* interrupt control.. */
+#ifdef CONFIG_ADEOS
+
+#include <linux/adeos.h>
+
+#define local_save_flags(x) ((x) = adeos_test_ipipe())
+#define local_irq_save(x) ((x) = adeos_test_and_stall_ipipe())
+#define local_irq_restore(x) adeos_restore_ipipe(x)
+#define local_irq_disable() adeos_stall_ipipe()
+#define local_irq_enable() adeos_unstall_ipipe()
+#define irqs_disabled() adeos_test_ipipe()
+
+#define safe_halt() \
+__asm__ __volatile__("call adeos_unstall_ipipe; hlt": : :"memory")
+
+#define ipipe_hw_save_flags(x) \
+__asm__ __volatile__("pushfl ; popl %0":"=g" (x): /* no input */)
+
+#define ipipe_hw_save_flags_and_cli(x) \
+__asm__ __volatile__("pushfl ; popl %0 ; cli":"=g" (x): /* no input */)
+
+#define ipipe_hw_restore_flags(x) \
+__asm__ __volatile__("pushl %0 ; popfl": /* no output */ :"g" (x):"memory", "cc")
+
+#define ipipe_hw_cli() \
+__asm__ __volatile__("cli": : :"memory")
+
+#define ipipe_hw_sti() \
+__asm__ __volatile__("sti": : :"memory")
+
+#define ipipe_hw_local_irq_save(x) \
+__asm__ __volatile__("pushfl ; popl %0 ; cli":"=g" (x): /* no input */ :"memory")
+
+#define ipipe_hw_local_irq_restore(x) \
+__asm__ __volatile__("pushl %0 ; popfl": /* no output */ :"g" (x):"memory", "cc")
+
+/*
+ * Temporary defines for UP kernels until all code gets fixed.
+ * (Lifted from linux/interrupt.h since some drivers do not include
+ * it).
+ */
+#if !defined(CONFIG_SMP)
+# define cli() local_irq_disable()
+# define sti() local_irq_enable()
+#endif
+
+#else /* !CONFIG_ADEOS */
+
#define local_save_flags(x) __asm__ __volatile__("pushfl ; popl %0":"=g" (x): /* no input */)
#define local_irq_restore(x) __asm__ __volatile__("pushl %0 ; popfl": /* no output */ :"g" (x):"memory", "cc")
#define local_irq_disable() __asm__ __volatile__("cli": : :"memory")
@@ -322,6 +369,8 @@ static inline unsigned long __cmpxchg(vo

/* For spinlocks etc */
#define local_irq_save(x) __asm__ __volatile__("pushfl ; popl %0 ; cli":"=g" (x): /* no input */ :"memory")
+
+#endif /* CONFIG_ADEOS */

/*
* disable hlt during certain critical i/o operations
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/