sanchayanmaity/minix
1
0
Fork 0
Code Issues Pull requests Packages Projects Releases Wiki Activity
minix/kernel/arch/i386/mpx386.S

765 lines
19 KiB
ArmAsm
Raw Normal View History

Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
/*
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
* This file, mpx386.s, is included by mpx.s when Minix is compiled for
* 32-bit Intel CPUs. The alternative mpx88.s is compiled for 16-bit CPUs.
*
* This file is part of the lowest layer of the MINIX kernel. (The other part
* is "proc.c".) The lowest layer does process switching and message handling.
* Furthermore it contains the assembler startup code for Minix and the 32-bit
* interrupt handlers. It cooperates with the code in "start.c" to set up a
* good environment for main().
*
* Every transition to the kernel goes through this file. Transitions to the
* kernel may be nested. The initial entry may be with a system call (i.e.,
* send or receive a message), an exception or a hardware interrupt; kernel
* reentries may only be made by hardware interrupts. The count of reentries
* is kept in "k_reenter". It is important for deciding whether to switch to
* the kernel stack and for protecting the message passing code in "proc.c".
*
* For the message passing trap, most of the machine state is saved in the
* proc table. (Some of the registers need not be saved.) Then the stack is
* switched to "k_stack", and interrupts are reenabled. Finally, the system
* call handler (in C) is called. When it returns, interrupts are disabled
* again and the code falls into the restart routine, to finish off held-up
* interrupts and run the process or task whose pointer is in "proc_ptr".
*
* Hardware interrupt handlers do the same, except (1) The entire state must
* be saved. (2) There are too many handlers to do this inline, so the save
* routine is called. A few cycles are saved by pushing the address of the
* appropiate restart routine for a return later. (3) A stack switch is
* avoided when the stack is already switched. (4) The (master) 8259 interrupt
* controller is reenabled centrally in save(). (5) Each interrupt handler
* masks its interrupt line using the 8259 before enabling (other unmasked)
* interrupts, and unmasks it after servicing the interrupt. This limits the
* nest level to the number of lines and protects the handler from itself.
*
* For communication with the boot monitor at startup time some constant
* data are compiled into the beginning of the text segment. This facilitates
* reading the data at the start of the boot process, since only the first
* sector of the file needs to be read.
*
* Some data storage is also allocated at the end of this file. This data
* will be at the start of the data segment of the kernel and will be read
* and modified by the boot monitor before the kernel starts.
*/
NMI watchdog is an awesome feature for debugging locked up kernels. There is not that much use for it on a single CPU, however, deadlock between kernel and system task can be delected. Or a runaway loop. If a kernel gets locked up the timer interrupts don't occure (as all interrupts are disabled in kernel mode). The only chance is to interrupt the kernel by a non-maskable interrupt. This patch generates NMIs using performance counters. It uses the most widely available performace counters. As the performance counters are highly model-specific this patch is not guaranteed to work on every machine. Unfortunately this is also true for KVM :-/ On the other hand adding this feature for other models is not extremely difficult and the framework makes it hopefully easy enough. Depending on the frequency of the CPU an NMI is generated at most about every 0.5s If the cpu's speed is less then 2Ghz it is generated at most every 1s. In general an NMI is generated much less often as the performance counter counts down only if the cpu is not idle. Therefore the overhead of this feature is fairly minimal even if the load is high. Uppon detecting that the kernel is locked up the kernel dumps the state of the kernel registers and panics. Local APIC must be enabled for the watchdog to work. The code is _always_ compiled in, however, it is only enabled if watchdog=<non-zero> is set in the boot monitor. One corner case is serial console debugging. As dumping a lot of stuff to the serial link may take a lot of time, the watchdog does not detect lockups during this time!!! as it would result in too many false positives. 10 nmi have to be handled before the lockup is detected. This means something between ~5s to 10s. Another corner case is that the watchdog is enabled only after the paging is enabled as it would be pure madness to try to get it right.
2010-01-16 21:53:55 +01:00
#include "../../kernel.h" /* configures the kernel */
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
/* sections */
#include <sys/vm_i386.h>
#ifdef __ACK__
.text
begtext:
#ifdef __ACK__
.rom
#else
.data
#endif
begrom:
.data
begdata:
.bss
begbss:
#endif
#include <minix/config.h>
#include <minix/const.h>
#include <minix/com.h>
#include <ibm/interrupt.h>
#include <archconst.h>
#include "../../const.h"
FPU context switching support by Evgeniy Ivanov.
2009-12-02 14:01:48 +01:00
#include "../../proc.h"
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
#include "sconst.h"
/* Selected 386 tss offsets. */
#define TSS3_S_SP0 4
/*
* Exported functions
* Note: in assembly language the .define statement applied to a function name
* is loosely equivalent to a prototype in C code -- it makes it possible to
* link to an entity declared in the assembly code but does not create
* the entity.
*/
.globl restart
.globl reload_cr3
.globl divide_error
.globl single_step_exception
.globl nmi
.globl breakpoint_exception
.globl overflow
.globl bounds_check
.globl inval_opcode
.globl copr_not_available
.globl double_fault
.globl copr_seg_overrun
.globl inval_tss
.globl segment_not_present
.globl stack_exception
.globl general_protection
.globl page_fault
.globl copr_error
FPU context switching support by Evgeniy Ivanov.
2009-12-02 14:01:48 +01:00
.globl alignment_check
.globl machine_check
.globl simd_exception
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
.globl params_size
.globl params_offset
.globl mon_ds
.globl schedcheck
.globl dirtypde
FPU context switching support by Evgeniy Ivanov.
2009-12-02 14:01:48 +01:00
.globl lazy_fpu
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
.globl hwint00 /* handlers for hardware interrupts */
.globl hwint01
.globl hwint02
.globl hwint03
.globl hwint04
.globl hwint05
.globl hwint06
.globl hwint07
.globl hwint08
.globl hwint09
.globl hwint10
.globl hwint11
.globl hwint12
.globl hwint13
.globl hwint14
.globl hwint15
/* Exported variables. */
.globl begbss
.globl begdata
.text
/*===========================================================================*/
/* MINIX */
/*===========================================================================*/
.global MINIX
MINIX:
/* this is the entry point for the MINIX kernel */
jmp over_flags /* skip over the next few bytes */
.short CLICK_SHIFT /* for the monitor: memory granularity */
flags:
/* boot monitor flags:
* call in 386 mode, make bss, make stack,
* load high, don't patch, will return,
* uses generic INT, memory vector,
* new boot code return
*/
.short 0x01FD
nop /* extra byte to sync up disassembler */
over_flags:
/* Set up a C stack frame on the monitor stack. (The monitor sets cs and ds */
/* right. The ss descriptor still references the monitor data segment.) */
movzwl %sp, %esp /* monitor stack is a 16 bit stack */
push %ebp
mov %esp, %ebp
push %esi
push %edi
cmp $0, 4(%ebp) /* monitor return vector is */
je noret /* nonzero if return possible */
incl mon_return
noret:
movl %esp, mon_sp /* save stack pointer for later return */
/* Copy the monitor global descriptor table to the address space of kernel and */
/* switch over to it. Prot_init() can then update it with immediate effect. */
sgdt gdt+GDT_SELECTOR /* get the monitor gdtr */
movl gdt+GDT_SELECTOR+2, %esi /* absolute address of GDT */
mov $gdt, %ebx /* address of kernel GDT */
mov $8*8, %ecx /* copying eight descriptors */
copygdt:
movb %es:(%esi), %al
movb %al, (%ebx)
inc %esi
inc %ebx
loop copygdt
movl gdt+DS_SELECTOR+2, %eax /* base of kernel data */
and $0x00FFFFFF, %eax /* only 24 bits */
add $gdt, %eax /* eax = vir2phys(gdt) */
movl %eax, gdt+GDT_SELECTOR+2 /* set base of GDT */
lgdt gdt+GDT_SELECTOR /* switch over to kernel GDT */
/* Locate boot parameters, set up kernel segment registers and stack. */
mov 8(%ebp), %ebx /* boot parameters offset */
mov 12(%ebp), %edx /* boot parameters length */
mov 16(%ebp), %eax /* address of a.out headers */
movl %eax, aout
mov %ds, %ax /* kernel data */
mov %ax, %es
mov %ax, %fs
mov %ax, %gs
mov %ax, %ss
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
mov $k_boot_stktop, %esp /* set sp to point to the top of kernel stack */
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
/* Save boot parameters into these global variables for i386 code */
movl %edx, params_size
movl %ebx, params_offset
movl $SS_SELECTOR, mon_ds
/* Call C startup code to set up a proper environment to run main(). */
push %edx
push %ebx
push $SS_SELECTOR
push $DS_SELECTOR
push $CS_SELECTOR
call cstart /* cstart(cs, ds, mds, parmoff, parmlen) */
add $5*4, %esp
/* Reload gdtr, idtr and the segment registers to global descriptor table set */
/* up by prot_init(). */
lgdt gdt+GDT_SELECTOR
lidt gdt+IDT_SELECTOR
ljmp $CS_SELECTOR, $csinit
csinit:
movw $DS_SELECTOR, %ax
mov %ax, %ds
mov %ax, %es
mov %ax, %fs
mov %ax, %gs
mov %ax, %ss
movw $TSS_SELECTOR, %ax /* no other TSS is used */
ltr %ax
push $0 /* set flags to known good state */
popf /* esp, clear nested task and int enable */
jmp main /* main() */
/*===========================================================================*/
/* interrupt handlers */
/* interrupt handlers for 386 32-bit protected mode */
/*===========================================================================*/
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
#define PIC_IRQ_HANDLER(irq) \
push $irq ;\
call irq_handle /* intr_handle(irq_handlers[irq]) */ ;\
add $4, %esp ;
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
/*===========================================================================*/
/* hwint00 - 07 */
/*===========================================================================*/
/* Note this is a macro, it just looks like a subroutine. */
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
#define hwint_master(irq) \
TEST_INT_IN_KERNEL(4, 0f) ;\
\
SAVE_PROCESS_CTX(0) ;\
Time accounting based on TSC - as thre are still KERNEL and IDLE entries, time accounting for kernel and idle time works the same as for any other process - everytime we stop accounting for the currently running process, kernel or idle, we read the TSC counter and increment the p_cycles entry. - the process cycles inherently include some of the kernel cycles as we can stop accounting for the process only after we save its context and we start accounting just before we restore its context - this assumes that the system does not scale the CPU frequency which will be true for ... long time ;-)
2010-02-10 16:36:54 +01:00
push %ebp ;\
call cycles_accounting_stop ;\
add $4, %esp ;\
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
movl $0, %ebp /* for stack trace */ ;\
PIC_IRQ_HANDLER(irq) ;\
movb $END_OF_INT, %al ;\
outb $INT_CTL /* reenable interrupts in master pic */ ;\
jmp restart ;\
\
0: \
pusha ;\
Time accounting based on TSC - as thre are still KERNEL and IDLE entries, time accounting for kernel and idle time works the same as for any other process - everytime we stop accounting for the currently running process, kernel or idle, we read the TSC counter and increment the p_cycles entry. - the process cycles inherently include some of the kernel cycles as we can stop accounting for the process only after we save its context and we start accounting just before we restore its context - this assumes that the system does not scale the CPU frequency which will be true for ... long time ;-)
2010-02-10 16:36:54 +01:00
call cycles_accounting_stop_idle ;\
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
PIC_IRQ_HANDLER(irq) ;\
movb $END_OF_INT, %al ;\
outb $INT_CTL /* reenable interrupts in master pic */ ;\
popa ;\
iret ;
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
/* Each of these entry points is an expansion of the hwint_master macro */
.balign 16
hwint00:
/* Interrupt routine for irq 0 (the clock). */
hwint_master(0)
.balign 16
hwint01:
/* Interrupt routine for irq 1 (keyboard) */
hwint_master(1)
.balign 16
hwint02:
/* Interrupt routine for irq 2 (cascade!) */
hwint_master(2)
.balign 16
hwint03:
/* Interrupt routine for irq 3 (second serial) */
hwint_master(3)
.balign 16
hwint04:
/* Interrupt routine for irq 4 (first serial) */
hwint_master(4)
.balign 16
hwint05:
/* Interrupt routine for irq 5 (XT winchester) */
hwint_master(5)
.balign 16
hwint06:
/* Interrupt routine for irq 6 (floppy) */
hwint_master(6)
.balign 16
hwint07:
/* Interrupt routine for irq 7 (printer) */
hwint_master(7)
/*===========================================================================*/
/* hwint08 - 15 */
/*===========================================================================*/
/* Note this is a macro, it just looks like a subroutine. */
#define hwint_slave(irq) \
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
TEST_INT_IN_KERNEL(4, 0f) ;\
\
SAVE_PROCESS_CTX(0) ;\
Time accounting based on TSC - as thre are still KERNEL and IDLE entries, time accounting for kernel and idle time works the same as for any other process - everytime we stop accounting for the currently running process, kernel or idle, we read the TSC counter and increment the p_cycles entry. - the process cycles inherently include some of the kernel cycles as we can stop accounting for the process only after we save its context and we start accounting just before we restore its context - this assumes that the system does not scale the CPU frequency which will be true for ... long time ;-)
2010-02-10 16:36:54 +01:00
push %ebp ;\
call cycles_accounting_stop ;\
add $4, %esp ;\
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
movl $0, %ebp /* for stack trace */ ;\
PIC_IRQ_HANDLER(irq) ;\
movb $END_OF_INT, %al ;\
outb $INT_CTL /* reenable interrupts in master pic */ ;\
outb $INT2_CTL /* reenable slave 8259 */ ;\
jmp restart ;\
\
0: \
pusha ;\
Time accounting based on TSC - as thre are still KERNEL and IDLE entries, time accounting for kernel and idle time works the same as for any other process - everytime we stop accounting for the currently running process, kernel or idle, we read the TSC counter and increment the p_cycles entry. - the process cycles inherently include some of the kernel cycles as we can stop accounting for the process only after we save its context and we start accounting just before we restore its context - this assumes that the system does not scale the CPU frequency which will be true for ... long time ;-)
2010-02-10 16:36:54 +01:00
call cycles_accounting_stop_idle ;\
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
PIC_IRQ_HANDLER(irq) ;\
movb $END_OF_INT, %al ;\
outb $INT_CTL /* reenable interrupts in master pic */ ;\
outb $INT2_CTL /* reenable slave 8259 */ ;\
popa ;\
iret ;
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
/* Each of these entry points is an expansion of the hwint_slave macro */
.balign 16
hwint08:
/* Interrupt routine for irq 8 (realtime clock) */
hwint_slave(8)
.balign 16
hwint09:
/* Interrupt routine for irq 9 (irq 2 redirected) */
hwint_slave(9)
.balign 16
hwint10:
/* Interrupt routine for irq 10 */
hwint_slave(10)
.balign 16
hwint11:
/* Interrupt routine for irq 11 */
hwint_slave(11)
.balign 16
hwint12:
/* Interrupt routine for irq 12 */
hwint_slave(12)
.balign 16
hwint13:
/* Interrupt routine for irq 13 (FPU exception) */
hwint_slave(13)
.balign 16
hwint14:
/* Interrupt routine for irq 14 (AT winchester) */
hwint_slave(14)
.balign 16
hwint15:
/* Interrupt routine for irq 15 */
hwint_slave(15)
/*
Removal of the system task * Userspace change to use the new kernel calls - _taskcall(SYSTASK...) changed to _kernel_call(...) - int 32 reused for the kernel calls - _do_kernel_call() to make the trap to kernel - kernel_call() to make the actuall kernel call from C using _do_kernel_call() - unlike ipc call the kernel call always succeeds as kernel is always available, however, kernel may return an error * Kernel side implementation of kernel calls - the SYSTEm task does not run, only the proc table entry is preserved - every data_copy(SYSTEM is no data_copy(KERNEL - "locking" is an empty operation now as everything runs in kernel - sys_task() is replaced by kernel_call() which copies the message into kernel, dispatches the call to its handler and finishes by either copying the results back to userspace (if need be) or by suspending the process because of VM - suspended processes are later made runnable once the memory issue is resolved, picked up by the scheduler and only at this time the call is resumed (in fact restarted) which does not need to copy the message from userspace as the message is already saved in the process structure. - no ned for the vmrestart queue, the scheduler will restart the system calls - no special case in do_vmctl(), all requests remove the RTS_VMREQUEST flag
2010-02-09 16:20:09 +01:00
* IPC is only from a process to kernel
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
*/
.align replaced by .balign in mpx386.S
2009-11-13 10:30:45 +01:00
.balign 16
Sys calls are called ipc calls now - the syscalls are pretty much just ipc calls, however, sendrec() is used to implement system task (sys) calls - sendrec() won't be used anymore for this, therefore ipc calls will become pure ipc calls
2010-02-09 16:13:07 +01:00
.globl ipc_entry
ipc_entry:
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
SAVE_PROCESS_CTX(0)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
/* save the pointer to the current process */
push %ebp
/*
* pass the syscall arguments from userspace to the handler.
* SAVE_PROCESS_CTX() does not clobber these registers, they are still
* set as the userspace have set them
*/
push %edx
push %ebx
push %eax
push %ecx
Time accounting based on TSC - as thre are still KERNEL and IDLE entries, time accounting for kernel and idle time works the same as for any other process - everytime we stop accounting for the currently running process, kernel or idle, we read the TSC counter and increment the p_cycles entry. - the process cycles inherently include some of the kernel cycles as we can stop accounting for the process only after we save its context and we start accounting just before we restore its context - this assumes that the system does not scale the CPU frequency which will be true for ... long time ;-)
2010-02-10 16:36:54 +01:00
/* stop user process cycles */
push %ebp
call cycles_accounting_stop
add $4, %esp
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
/* for stack trace */
movl $0, %ebp
Sys calls are called ipc calls now - the syscalls are pretty much just ipc calls, however, sendrec() is used to implement system task (sys) calls - sendrec() won't be used anymore for this, therefore ipc calls will become pure ipc calls
2010-02-09 16:13:07 +01:00
call do_ipc
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
/* restore the current process pointer and save the return value */
add $4 * 4, %esp
pop %esi
mov %eax, AXREG(%esi)
jmp restart
Removal of the system task * Userspace change to use the new kernel calls - _taskcall(SYSTASK...) changed to _kernel_call(...) - int 32 reused for the kernel calls - _do_kernel_call() to make the trap to kernel - kernel_call() to make the actuall kernel call from C using _do_kernel_call() - unlike ipc call the kernel call always succeeds as kernel is always available, however, kernel may return an error * Kernel side implementation of kernel calls - the SYSTEm task does not run, only the proc table entry is preserved - every data_copy(SYSTEM is no data_copy(KERNEL - "locking" is an empty operation now as everything runs in kernel - sys_task() is replaced by kernel_call() which copies the message into kernel, dispatches the call to its handler and finishes by either copying the results back to userspace (if need be) or by suspending the process because of VM - suspended processes are later made runnable once the memory issue is resolved, picked up by the scheduler and only at this time the call is resumed (in fact restarted) which does not need to copy the message from userspace as the message is already saved in the process structure. - no ned for the vmrestart queue, the scheduler will restart the system calls - no special case in do_vmctl(), all requests remove the RTS_VMREQUEST flag
2010-02-09 16:20:09 +01:00
/*
* kernel call is only from a process to kernel
*/
.balign 16
.globl kernel_call_entry
kernel_call_entry:
SAVE_PROCESS_CTX(0)
/* save the pointer to the current process */
push %ebp
/*
* pass the syscall arguments from userspace to the handler.
* SAVE_PROCESS_CTX() does not clobber these registers, they are still
* set as the userspace have set them
*/
push %eax
Time accounting based on TSC - as thre are still KERNEL and IDLE entries, time accounting for kernel and idle time works the same as for any other process - everytime we stop accounting for the currently running process, kernel or idle, we read the TSC counter and increment the p_cycles entry. - the process cycles inherently include some of the kernel cycles as we can stop accounting for the process only after we save its context and we start accounting just before we restore its context - this assumes that the system does not scale the CPU frequency which will be true for ... long time ;-)
2010-02-10 16:36:54 +01:00
/* stop user process cycles */
push %ebp
call cycles_accounting_stop
add $4, %esp
/* for stack trace */
movl $0, %ebp
Removal of the system task * Userspace change to use the new kernel calls - _taskcall(SYSTASK...) changed to _kernel_call(...) - int 32 reused for the kernel calls - _do_kernel_call() to make the trap to kernel - kernel_call() to make the actuall kernel call from C using _do_kernel_call() - unlike ipc call the kernel call always succeeds as kernel is always available, however, kernel may return an error * Kernel side implementation of kernel calls - the SYSTEm task does not run, only the proc table entry is preserved - every data_copy(SYSTEM is no data_copy(KERNEL - "locking" is an empty operation now as everything runs in kernel - sys_task() is replaced by kernel_call() which copies the message into kernel, dispatches the call to its handler and finishes by either copying the results back to userspace (if need be) or by suspending the process because of VM - suspended processes are later made runnable once the memory issue is resolved, picked up by the scheduler and only at this time the call is resumed (in fact restarted) which does not need to copy the message from userspace as the message is already saved in the process structure. - no ned for the vmrestart queue, the scheduler will restart the system calls - no special case in do_vmctl(), all requests remove the RTS_VMREQUEST flag
2010-02-09 16:20:09 +01:00
call kernel_call
/* restore the current process pointer and save the return value */
add $8, %esp
jmp restart
.align replaced by .balign in mpx386.S
2009-11-13 10:30:45 +01:00
.balign 16
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
/*
* called by the exception interrupt vectors. If the exception does not push
* errorcode, we assume that the vector handler pushed 0 instead. Next pushed
* thing is the vector number. From this point on we can continue as if every
* exception pushes an error code
*/
exception_entry:
/*
* check if it is a nested trap by comparing the saved code segment
* descriptor with the kernel CS first
*/
TEST_INT_IN_KERNEL(12, exception_entry_nested)
exception_entry_from_user:
cld
SAVE_PROCESS_CTX(8)
Time accounting based on TSC - as thre are still KERNEL and IDLE entries, time accounting for kernel and idle time works the same as for any other process - everytime we stop accounting for the currently running process, kernel or idle, we read the TSC counter and increment the p_cycles entry. - the process cycles inherently include some of the kernel cycles as we can stop accounting for the process only after we save its context and we start accounting just before we restore its context - this assumes that the system does not scale the CPU frequency which will be true for ... long time ;-)
2010-02-10 16:36:54 +01:00
/* stop user process cycles */
push %ebp
call cycles_accounting_stop
add $4, %esp
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
/* for stack trace clear %ebp */
movl $0, %ebp
/*
* push a pointer to the interrupt state pushed by the cpu and the
* vector number pushed by the vector handler just before calling
* exception_entry and call the exception handler.
*/
push %esp
push $0 /* it's not a nested exception */
call exception_handler
jmp restart
exception_entry_nested:
pusha
mov %esp, %eax
add $(8 * 4), %eax
push %eax
pushl $1 /* it's a nested exception */
call exception_handler
add $8, %esp
popa
/* clear the error code and the exception number */
add $8, %esp
/* resume execution at the point of exception */
iret
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
/*===========================================================================*/
/* restart */
/*===========================================================================*/
restart:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
call schedcheck
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
/* %eax is set by schedcheck() to the process to run */
NMI watchdog is an awesome feature for debugging locked up kernels. There is not that much use for it on a single CPU, however, deadlock between kernel and system task can be delected. Or a runaway loop. If a kernel gets locked up the timer interrupts don't occure (as all interrupts are disabled in kernel mode). The only chance is to interrupt the kernel by a non-maskable interrupt. This patch generates NMIs using performance counters. It uses the most widely available performace counters. As the performance counters are highly model-specific this patch is not guaranteed to work on every machine. Unfortunately this is also true for KVM :-/ On the other hand adding this feature for other models is not extremely difficult and the framework makes it hopefully easy enough. Depending on the frequency of the CPU an NMI is generated at most about every 0.5s If the cpu's speed is less then 2Ghz it is generated at most every 1s. In general an NMI is generated much less often as the performance counter counts down only if the cpu is not idle. Therefore the overhead of this feature is fairly minimal even if the load is high. Uppon detecting that the kernel is locked up the kernel dumps the state of the kernel registers and panics. Local APIC must be enabled for the watchdog to work. The code is _always_ compiled in, however, it is only enabled if watchdog=<non-zero> is set in the boot monitor. One corner case is serial console debugging. As dumping a lot of stuff to the serial link may take a lot of time, the watchdog does not detect lockups during this time!!! as it would result in too many false positives. 10 nmi have to be handled before the lockup is detected. This means something between ~5s to 10s. Another corner case is that the watchdog is enabled only after the paging is enabled as it would be pure madness to try to get it right.
2010-01-16 21:53:55 +01:00
mov %eax, %ebp /* will assume P_STACKBASE == 0 */
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
NMI watchdog is an awesome feature for debugging locked up kernels. There is not that much use for it on a single CPU, however, deadlock between kernel and system task can be delected. Or a runaway loop. If a kernel gets locked up the timer interrupts don't occure (as all interrupts are disabled in kernel mode). The only chance is to interrupt the kernel by a non-maskable interrupt. This patch generates NMIs using performance counters. It uses the most widely available performace counters. As the performance counters are highly model-specific this patch is not guaranteed to work on every machine. Unfortunately this is also true for KVM :-/ On the other hand adding this feature for other models is not extremely difficult and the framework makes it hopefully easy enough. Depending on the frequency of the CPU an NMI is generated at most about every 0.5s If the cpu's speed is less then 2Ghz it is generated at most every 1s. In general an NMI is generated much less often as the performance counter counts down only if the cpu is not idle. Therefore the overhead of this feature is fairly minimal even if the load is high. Uppon detecting that the kernel is locked up the kernel dumps the state of the kernel registers and panics. Local APIC must be enabled for the watchdog to work. The code is _always_ compiled in, however, it is only enabled if watchdog=<non-zero> is set in the boot monitor. One corner case is serial console debugging. As dumping a lot of stuff to the serial link may take a lot of time, the watchdog does not detect lockups during this time!!! as it would result in too many false positives. 10 nmi have to be handled before the lockup is detected. This means something between ~5s to 10s. Another corner case is that the watchdog is enabled only after the paging is enabled as it would be pure madness to try to get it right.
2010-01-16 21:53:55 +01:00
/* reconstruct the stack for iret */
movl SSREG(%ebp), %eax
push %eax
movl SPREG(%ebp), %eax
push %eax
movl PSWREG(%ebp), %eax
push %eax
movl CSREG(%ebp), %eax
push %eax
movl PCREG(%ebp), %eax
push %eax
RESTORE_GP_REGS(%ebp)
RESTORE_SEGS(%ebp)
movl %ss:BPREG(%ebp), %ebp
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
iret /* continue process */
/*===========================================================================*/
/* exception handlers */
/*===========================================================================*/
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
#define EXCEPTION_ERR_CODE(vector) \
push $vector ;\
jmp exception_entry
#define EXCEPTION_NO_ERR_CODE(vector) \
pushl $0 ;\
EXCEPTION_ERR_CODE(vector)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
divide_error:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_NO_ERR_CODE(DIVIDE_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
single_step_exception:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_NO_ERR_CODE(DEBUG_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
nmi:
NMI watchdog is an awesome feature for debugging locked up kernels. There is not that much use for it on a single CPU, however, deadlock between kernel and system task can be delected. Or a runaway loop. If a kernel gets locked up the timer interrupts don't occure (as all interrupts are disabled in kernel mode). The only chance is to interrupt the kernel by a non-maskable interrupt. This patch generates NMIs using performance counters. It uses the most widely available performace counters. As the performance counters are highly model-specific this patch is not guaranteed to work on every machine. Unfortunately this is also true for KVM :-/ On the other hand adding this feature for other models is not extremely difficult and the framework makes it hopefully easy enough. Depending on the frequency of the CPU an NMI is generated at most about every 0.5s If the cpu's speed is less then 2Ghz it is generated at most every 1s. In general an NMI is generated much less often as the performance counter counts down only if the cpu is not idle. Therefore the overhead of this feature is fairly minimal even if the load is high. Uppon detecting that the kernel is locked up the kernel dumps the state of the kernel registers and panics. Local APIC must be enabled for the watchdog to work. The code is _always_ compiled in, however, it is only enabled if watchdog=<non-zero> is set in the boot monitor. One corner case is serial console debugging. As dumping a lot of stuff to the serial link may take a lot of time, the watchdog does not detect lockups during this time!!! as it would result in too many false positives. 10 nmi have to be handled before the lockup is detected. This means something between ~5s to 10s. Another corner case is that the watchdog is enabled only after the paging is enabled as it would be pure madness to try to get it right.
2010-01-16 21:53:55 +01:00
#ifndef CONFIG_WATCHDOG
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_NO_ERR_CODE(NMI_VECTOR)
NMI watchdog is an awesome feature for debugging locked up kernels. There is not that much use for it on a single CPU, however, deadlock between kernel and system task can be delected. Or a runaway loop. If a kernel gets locked up the timer interrupts don't occure (as all interrupts are disabled in kernel mode). The only chance is to interrupt the kernel by a non-maskable interrupt. This patch generates NMIs using performance counters. It uses the most widely available performace counters. As the performance counters are highly model-specific this patch is not guaranteed to work on every machine. Unfortunately this is also true for KVM :-/ On the other hand adding this feature for other models is not extremely difficult and the framework makes it hopefully easy enough. Depending on the frequency of the CPU an NMI is generated at most about every 0.5s If the cpu's speed is less then 2Ghz it is generated at most every 1s. In general an NMI is generated much less often as the performance counter counts down only if the cpu is not idle. Therefore the overhead of this feature is fairly minimal even if the load is high. Uppon detecting that the kernel is locked up the kernel dumps the state of the kernel registers and panics. Local APIC must be enabled for the watchdog to work. The code is _always_ compiled in, however, it is only enabled if watchdog=<non-zero> is set in the boot monitor. One corner case is serial console debugging. As dumping a lot of stuff to the serial link may take a lot of time, the watchdog does not detect lockups during this time!!! as it would result in too many false positives. 10 nmi have to be handled before the lockup is detected. This means something between ~5s to 10s. Another corner case is that the watchdog is enabled only after the paging is enabled as it would be pure madness to try to get it right.
2010-01-16 21:53:55 +01:00
#else
/*
* We have to be very careful as this interrupt can occur anytime. On
* the other hand, if it interrupts a user process, we will resume the
* same process which makes things a little simpler. We know that we are
* already on kernel stack whenever it happened and we can be
* conservative and save everything as we don't need to be extremely
* efficient as the interrupt is infrequent and some overhead is already
* expected.
*/
/*
* save the important registers. We don't save %cs and %ss and they are
* saved and restored by CPU
*/
pushw %ds
pushw %es
pushw %fs
pushw %gs
pusha
/*
* We cannot be sure about the state of the kernel segment register,
* however, we always set %ds and %es to the same as %ss
*/
mov %ss, %si
mov %si, %ds
mov %si, %es
push %esp
call nmi_watchdog_handler
add $4, %esp
/* restore all the important registers as they were before the trap */
popa
popw %gs
popw %fs
popw %es
popw %ds
iret
#endif
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
breakpoint_exception:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_NO_ERR_CODE(BREAKPOINT_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
overflow:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_NO_ERR_CODE(OVERFLOW_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
bounds_check:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_NO_ERR_CODE(BOUNDS_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
inval_opcode:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_NO_ERR_CODE(INVAL_OP_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
copr_not_available:
FPU context switching support by Evgeniy Ivanov.
2009-12-02 14:01:48 +01:00
TEST_INT_IN_KERNEL(4, copr_not_available_in_kernel)
clts
cld /* set direction flag to a known value */
SAVE_PROCESS_CTX_NON_LAZY(0)
lea P_MISC_FLAGS(%ebp), %ebx
movw (%ebx), %cx
and $MF_FPU_INITIALIZED, %cx
jnz 0f /* jump if FPU is already initialized */
orw $MF_FPU_INITIALIZED, (%ebx)
fninit
Time accounting based on TSC - as thre are still KERNEL and IDLE entries, time accounting for kernel and idle time works the same as for any other process - everytime we stop accounting for the currently running process, kernel or idle, we read the TSC counter and increment the p_cycles entry. - the process cycles inherently include some of the kernel cycles as we can stop accounting for the process only after we save its context and we start accounting just before we restore its context - this assumes that the system does not scale the CPU frequency which will be true for ... long time ;-)
2010-02-10 16:36:54 +01:00
/* stop user process cycles */
push %ebp
call cycles_accounting_stop
add $4, %esp
FPU context switching support by Evgeniy Ivanov.
2009-12-02 14:01:48 +01:00
jmp copr_return
0: /* load FPU context for current process */
mov %ss:FP_SAVE_AREA_P(%ebp), %eax
cmp $0, osfxsr_feature
jz fp_l_no_fxsr /* FXSR is not avaible. */
Cooments to warn not to use certains instructions - gas2ack cannot handle all variants of some instructions. Until this issues is addressed, this patch places a big warning where appropriate. This code is not supposed to change frequently.
2009-12-07 13:01:05 +01:00
/* DO NOT CHANGE THE OPERAND!!! gas2ack does not handle it yet */
FPU context switching support by Evgeniy Ivanov.
2009-12-02 14:01:48 +01:00
fxrstor (%eax)
jmp copr_return
fp_l_no_fxsr:
Cooments to warn not to use certains instructions - gas2ack cannot handle all variants of some instructions. Until this issues is addressed, this patch places a big warning where appropriate. This code is not supposed to change frequently.
2009-12-07 13:01:05 +01:00
/* DO NOT CHANGE THE OPERAND!!! gas2ack does not handle it yet */
FPU context switching support by Evgeniy Ivanov.
2009-12-02 14:01:48 +01:00
frstor (%eax)
copr_return:
orw $MF_USED_FPU, (%ebx) /* fpu was used during last execution */
jmp restart
copr_not_available_in_kernel:
movl $0, (%esp)
panic() cleanup. this change - makes panic() variadic, doing full printf() formatting - no more NO_NUM, and no more separate printf() statements needed to print extra info (or something in hex) before panicing - unifies panic() - same panic() name and usage for everyone - vm, kernel and rest have different names/syntax currently in order to implement their own luxuries, but no longer - throws out the 1st argument, to make source less noisy. the panic() in syslib retrieves the server name from the kernel so it should be clear enough who is panicing; e.g. panic("sigaction failed: %d", errno); looks like: at_wini(73130): panic: sigaction failed: 0 syslib:panic.c: stacktrace: 0x74dc 0x2025 0x100a - throws out report() - printf() is more convenient and powerful - harmonizes/fixes the use of panic() - there were a few places that used printf-style formatting (didn't work) and newlines (messes up the formatting) in panic() - throws out a few per-server panic() functions - cleans up a tie-in of tty with panic() merging printf() and panic() statements to be done incrementally.
2010-03-05 16:05:11 +01:00
call panic
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
double_fault:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_ERR_CODE(DOUBLE_FAULT_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
copr_seg_overrun:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_NO_ERR_CODE(COPROC_SEG_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
inval_tss:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_ERR_CODE(INVAL_TSS_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
segment_not_present:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_ERR_CODE(SEG_NOT_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
stack_exception:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_ERR_CODE(STACK_FAULT_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
general_protection:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_ERR_CODE(PROTECTION_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
page_fault:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_ERR_CODE(PAGE_FAULT_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
copr_error:
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
EXCEPTION_NO_ERR_CODE(COPROC_ERR_VECTOR)
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
FPU context switching support by Evgeniy Ivanov.
2009-12-02 14:01:48 +01:00
alignment_check:
EXCEPTION_NO_ERR_CODE(ALIGNMENT_CHECK_VECTOR)
machine_check:
EXCEPTION_NO_ERR_CODE(MACHINE_CHECK_VECTOR)
simd_exception:
EXCEPTION_NO_ERR_CODE(SIMD_EXCEPTION_VECTOR)
/*===========================================================================*/
/* lazy_fpu */
/*===========================================================================*/
/* void lazy_fpu(struct proc *pptr)
* It's called, when we are on kernel stack.
* Actualy lazy code is just few lines, which check MF_USED_FPU,
* another part is save_init_fpu().
*/
lazy_fpu:
push %ebp
mov %esp, %ebp
push %eax
push %ebx
push %ecx
cmp $0, fpu_presence /* Do we have FPU? */
jz no_fpu_available
mov 8(%ebp), %eax /* Get pptr */
lea P_MISC_FLAGS(%eax), %ebx
movw (%ebx), %cx
and $MF_USED_FPU, %cx
jz 0f /* Don't save FPU */
mov %ss:FP_SAVE_AREA_P(%eax), %eax
cmp $0, osfxsr_feature
jz fp_s_no_fxsr /* FXSR is not avaible. */
Cooments to warn not to use certains instructions - gas2ack cannot handle all variants of some instructions. Until this issues is addressed, this patch places a big warning where appropriate. This code is not supposed to change frequently.
2009-12-07 13:01:05 +01:00
/* DO NOT CHANGE THE OPERAND!!! gas2ack does not handle it yet */
FPU context switching support by Evgeniy Ivanov.
2009-12-02 14:01:48 +01:00
fxsave (%eax)
fninit
jmp fp_saved
fp_s_no_fxsr:
Cooments to warn not to use certains instructions - gas2ack cannot handle all variants of some instructions. Until this issues is addressed, this patch places a big warning where appropriate. This code is not supposed to change frequently.
2009-12-07 13:01:05 +01:00
/* DO NOT CHANGE THE OPERAND!!! gas2ack does not handle it yet */
FPU context switching support by Evgeniy Ivanov.
2009-12-02 14:01:48 +01:00
fnsave (%eax)
fwait /* required for compatibility with processors prior pentium */
fp_saved:
andw $~MF_USED_FPU, (%ebx)
0: mov %cr0, %eax
or $0x00000008, %eax /* Set TS flag */
mov %eax, %cr0
no_fpu_available:
pop %ecx
pop %ebx
pop %eax
pop %ebp
ret
Conversion of kernel assembly from ACK to GNU - .s files removed and replaced by .S as the .S is a standard extension for assembly that needs preprocessing
2009-10-30 17:00:44 +01:00
/*===========================================================================*/
/* reload_cr3 */
/*===========================================================================*/
/* PUBLIC void reload_cr3(void); */
reload_cr3:
push %ebp
mov %esp, %ebp
movl $0, dirtypde
mov %cr3, %eax
mov %eax, %cr3
pop %ebp
ret
/*===========================================================================*/
/* data */
/*===========================================================================*/
#ifdef __ACK__
.rom /* Before the string table please */
#else
.data
#endif
.short 0x526F /* this must be the first data entry (magic #) */
.bss
Complete ovehaul of mode switching code - after a trap to kernel, the code automatically switches to kernel stack, in the future local to the CPU - k_reenter variable replaced by a test whether the CS is kernel cs or not. The information is passed further if needed. Removes a global variable which would need to be cpu local - no need for global variables describing the exception or trap context. This information is kept on stack and a pointer to this structure is passed to the C code as a single structure - removed loadedcr3 variable and its use replaced by reading the %cr3 register - no need to redisable interrupts in restart() as they are already disabled. - unified handling of traps that push and don't push errorcode - removed save() function as the process context is not saved directly to process table but saved as required by the trap code. Essentially it means that save() code is inlined everywhere not only in the exception handling routine - returning from syscall is more arch independent - it sets the retger in C - top of the x86 stack contains the current CPU id and pointer to the currently scheduled process (the one right interrupted) so the mode switch code can find where to save the context without need to use proc_ptr which will be cpu local in the future and therefore difficult to access in assembler and expensive to access in general - some more clean up of level0 code. No need to read-back the argument passed in %eax from the proc structure. The mode switch code does not clobber %the general registers and hence we can just call what is in %eax - many assebly macros in sconst.h as they will be reused by the apic assembly
2009-11-06 10:08:26 +01:00
/*
* this stack is used temporarily for booting only. We switch to a proper kernel
* stack after the first trap to kernel
*/
.globl k_boot_stktop
k_boot_stack:
.space 4096 /* kernel stack */ /* FIXME use macro here */
k_boot_stktop: /* top of kernel stack */
Reference in a new issue Copy permalink