minix/kernel/arch/i386/sconst.h

152 lines
4 KiB
C
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
#ifndef __SCONST_H__
#define __SCONST_H__
2010-04-02 00:22:33 +02:00
#include "kernel/const.h"
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
/* Miscellaneous constants used in assembler code. */
W = _WORD_SIZE /* Machine word size. */
2005-04-21 16:53:53 +02:00
/* Offsets in struct proc. They MUST match proc.h. */
P_STACKBASE = 0
GSREG = P_STACKBASE
FSREG = GSREG+2 /* 386 introduces FS and GS segments*/
ESREG = FSREG+2
DSREG = ESREG+2
DIREG = DSREG+2
SIREG = DIREG+W
BPREG = SIREG+W
STREG = BPREG+W /* hole for another SP*/
BXREG = STREG+W
DXREG = BXREG+W
CXREG = DXREG+W
AXREG = CXREG+W
RETADR = AXREG+W /* return address for save() call*/
PCREG = RETADR+W
CSREG = PCREG+W
PSWREG = CSREG+W
SPREG = PSWREG+W
SSREG = SPREG+W
P_STACKTOP = SSREG+W
FP_SAVE_AREA_P = P_STACKTOP
P_LDT_SEL = FP_SAVE_AREA_P + 532
P_CR3 = P_LDT_SEL+W
P_CR3_V = P_CR3+4
P_LDT = P_CR3_V+W
P_MISC_FLAGS = P_LDT + 50
Msize = 9 /* size of a message in 32-bit words*/
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
/*
* offset to current process pointer right after trap, we assume we always have
* error code on the stack
*/
#define CURR_PROC_PTR 20
/*
* tests whether the interrupt was triggered in kernel. If so, jump to the
* label. Displacement tell the macro ha far is the CS value saved by the trap
* from the current %esp. The kernel code segment selector has the lower 3 bits
* zeroed
*/
#define TEST_INT_IN_KERNEL(displ, label) \
cmpl $CS_SELECTOR, displ(%esp) ;\
je label ;
/*
* saves the basic interrupt context (no error code) to the process structure
*
* displ is the displacement of %esp from the original stack after trap
* pptr is the process structure pointer
* tmp is an available temporary register
*/
#define SAVE_TRAP_CTX(displ, pptr, tmp) \
movl (0 + displ)(%esp), tmp ;\
movl tmp, PCREG(pptr) ;\
movl (4 + displ)(%esp), tmp ;\
movl tmp, CSREG(pptr) ;\
movl (8 + displ)(%esp), tmp ;\
movl tmp, PSWREG(pptr) ;\
movl (12 + displ)(%esp), tmp ;\
movl tmp, SPREG(pptr) ;\
movl tmp, STREG(pptr) ;\
movl (16 + displ)(%esp), tmp ;\
movl tmp, SSREG(pptr) ;
#define SAVE_SEGS(pptr) \
mov %ds, %ss:DSREG(pptr) ;\
mov %es, %ss:ESREG(pptr) ;\
mov %fs, %ss:FSREG(pptr) ;\
mov %gs, %ss:GSREG(pptr) ;
#define RESTORE_SEGS(pptr) \
movw %ss:DSREG(pptr), %ds ;\
movw %ss:ESREG(pptr), %es ;\
movw %ss:FSREG(pptr), %fs ;\
movw %ss:GSREG(pptr), %gs ;
/*
* restore kernel segments, %ss is kernnel data segment, %cs is aready set and
* %fs, %gs are not used
*/
#define RESTORE_KERNEL_SEGS \
mov %ss, %si ;\
mov %si, %ds ;\
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 %si, %es ;\
movw $0, %si ;\
mov %si, %gs ;\
mov %si, %fs ;
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 SAVE_GP_REGS(pptr) \
mov %eax, %ss:AXREG(pptr) ;\
mov %ecx, %ss:CXREG(pptr) ;\
mov %edx, %ss:DXREG(pptr) ;\
mov %ebx, %ss:BXREG(pptr) ;\
mov %esi, %ss:SIREG(pptr) ;\
mov %edi, %ss:DIREG(pptr) ;
#define RESTORE_GP_REGS(pptr) \
movl %ss:AXREG(pptr), %eax ;\
movl %ss:CXREG(pptr), %ecx ;\
movl %ss:DXREG(pptr), %edx ;\
movl %ss:BXREG(pptr), %ebx ;\
movl %ss:SIREG(pptr), %esi ;\
movl %ss:DIREG(pptr), %edi ;
/*
* save the context of the interrupted process to the structure in the process
* table. It pushses the %ebp to stack to get a scratch register. After %esi is
* saved, we can use it to get the saved %ebp from stack and save it to the
* final location
*
* displ is the stack displacement. In case of an exception, there are two extra
* value on the stack - error code and the exception number
*/
#define SAVE_PROCESS_CTX(displ) \
\
cld /* set the direction flag to a known state */ ;\
\
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
push %ebp ;\
;\
movl (CURR_PROC_PTR + 4 + displ)(%esp), %ebp ;\
\
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 segment registers */ \
SAVE_SEGS(%ebp) ;\
\
SAVE_GP_REGS(%ebp) ;\
pop %esi /* get the orig %ebp and save it */ ;\
mov %esi, %ss:BPREG(%ebp) ;\
\
RESTORE_KERNEL_SEGS ;\
SAVE_TRAP_CTX(displ, %ebp, %esi) ;
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
/*
* clear the IF flag in eflags which are stored somewhere in memory, e.g. on
* stack. iret or popf will load the new value later
*/
#define CLEAR_IF(where) \
mov where, %eax ;\
andl $0xfffffdff, %eax ;\
mov %eax, where ;
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
#endif /* __SCONST_H__ */