greg/xv6
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
xv6/spinlock.c

66 lines
1.2 KiB
C
Raw Normal View History

checkpoint. booting second processor. stack is messed up, but thanks to cliff and plan 9 code, at least boots and gets into C code.
2006-06-22 01:28:57 +00:00
#include "types.h"
#include "defs.h"
#include "x86.h"
disable all interrupts when acquiring lock user program that makes a blocking system call
2006-07-06 21:47:22 +00:00
#include "mmu.h"
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
#include "param.h"
#include "proc.h"
#include "spinlock.h"
checkpoint. booting second processor. stack is messed up, but thanks to cliff and plan 9 code, at least boots and gets into C code.
2006-06-22 01:28:57 +00:00
Changes to allow use of native x86 ELF compilers, which on my Linux 2.4 box using gcc 3.4.6 don't seem to follow the same conventions as the i386-jos-elf-gcc compilers. Can run make 'TOOLPREFIX=' or edit the Makefile. curproc[cpu()] can now be NULL, indicating that no proc is running. This seemed safer to me than having curproc[0] and curproc[1] both pointing at proc[0] potentially. The old implementation of swtch depended on the stack frame layout used inside swtch being okay to return from on the other stack (exactly the V6 you are not expected to understand this). It also could be called in two contexts: at boot time, to schedule the very first process, and later, on behalf of a process, to sleep or schedule some other process. I split this into two functions: scheduler and swtch. The scheduler is now a separate never-returning function, invoked by each cpu once set up. The scheduler looks like: scheduler() { setjmp(cpu.context); pick proc to schedule blah blah blah longjmp(proc.context) } The new swtch is intended to be called only when curproc[cpu()] is not NULL, that is, only on behalf of a user proc. It does: swtch() { if(setjmp(proc.context) == 0) longjmp(cpu.context) } to save the current proc context and then jump over to the scheduler, running on the cpu stack. Similarly the system call stubs are now in assembly in usys.S to avoid needing to know the details of stack frame layout used by the compiler. Also various changes in the debugging prints.
2006-07-11 01:07:40 +00:00
#define DEBUG 0
checkpoint. booting second processor. stack is messed up, but thanks to cliff and plan 9 code, at least boots and gets into C code.
2006-06-22 01:28:57 +00:00
i think my cmpxchg use was wrong in acquire nesting cli/sti: release shouldn't always enable interrupts separate setup of lapic from starting of other cpus, so cpu() works earlier flag to disable locking in console output make locks work even when curproc==0 (still crashes in clock interrupt)
2006-07-12 11:15:38 +00:00
extern use_printf_lock;
Changes to allow use of native x86 ELF compilers, which on my Linux 2.4 box using gcc 3.4.6 don't seem to follow the same conventions as the i386-jos-elf-gcc compilers. Can run make 'TOOLPREFIX=' or edit the Makefile. curproc[cpu()] can now be NULL, indicating that no proc is running. This seemed safer to me than having curproc[0] and curproc[1] both pointing at proc[0] potentially. The old implementation of swtch depended on the stack frame layout used inside swtch being okay to return from on the other stack (exactly the V6 you are not expected to understand this). It also could be called in two contexts: at boot time, to schedule the very first process, and later, on behalf of a process, to sleep or schedule some other process. I split this into two functions: scheduler and swtch. The scheduler is now a separate never-returning function, invoked by each cpu once set up. The scheduler looks like: scheduler() { setjmp(cpu.context); pick proc to schedule blah blah blah longjmp(proc.context) } The new swtch is intended to be called only when curproc[cpu()] is not NULL, that is, only on behalf of a user proc. It does: swtch() { if(setjmp(proc.context) == 0) longjmp(cpu.context) } to save the current proc context and then jump over to the scheduler, running on the cpu stack. Similarly the system call stubs are now in assembly in usys.S to avoid needing to know the details of stack frame layout used by the compiler. Also various changes in the debugging prints.
2006-07-11 01:07:40 +00:00
int getcallerpc(void *v) {
return ((int*)v)[-1];
}
void
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
acquire(struct spinlock * lock)
checkpoint. booting second processor. stack is messed up, but thanks to cliff and plan 9 code, at least boots and gets into C code.
2006-06-22 01:28:57 +00:00
{
i think my cmpxchg use was wrong in acquire nesting cli/sti: release shouldn't always enable interrupts separate setup of lapic from starting of other cpus, so cpu() works earlier flag to disable locking in console output make locks work even when curproc==0 (still crashes in clock interrupt)
2006-07-12 11:15:38 +00:00
unsigned who;
if(curproc[cpu()])
who = (unsigned) curproc[cpu()];
else
who = cpu() + 1;
checkpoint. booting second processor. stack is messed up, but thanks to cliff and plan 9 code, at least boots and gets into C code.
2006-06-22 01:28:57 +00:00
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
if(DEBUG) cprintf("cpu%d: acquiring at %x\n", cpu(), getcallerpc(&lock));
i think my cmpxchg use was wrong in acquire nesting cli/sti: release shouldn't always enable interrupts separate setup of lapic from starting of other cpus, so cpu() works earlier flag to disable locking in console output make locks work even when curproc==0 (still crashes in clock interrupt)
2006-07-12 11:15:38 +00:00
if (lock->who == who && lock->locked){
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
lock->count += 1;
} else {
cli();
i think my cmpxchg use was wrong in acquire nesting cli/sti: release shouldn't always enable interrupts separate setup of lapic from starting of other cpus, so cpu() works earlier flag to disable locking in console output make locks work even when curproc==0 (still crashes in clock interrupt)
2006-07-12 11:15:38 +00:00
// if we get the lock, eax will be zero
// if we don't get the lock, eax will be one
while ( cmpxchg(0, 1, &lock->locked) == 1 ) { ; }
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
lock->locker_pc = getcallerpc(&lock);
lock->count = 1;
i think my cmpxchg use was wrong in acquire nesting cli/sti: release shouldn't always enable interrupts separate setup of lapic from starting of other cpus, so cpu() works earlier flag to disable locking in console output make locks work even when curproc==0 (still crashes in clock interrupt)
2006-07-12 11:15:38 +00:00
lock->who = who;
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
}
i think my cmpxchg use was wrong in acquire nesting cli/sti: release shouldn't always enable interrupts separate setup of lapic from starting of other cpus, so cpu() works earlier flag to disable locking in console output make locks work even when curproc==0 (still crashes in clock interrupt)
2006-07-12 11:15:38 +00:00
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
if(DEBUG) cprintf("cpu%d: acquired at %x\n", cpu(), getcallerpc(&lock));
checkpoint. booting second processor. stack is messed up, but thanks to cliff and plan 9 code, at least boots and gets into C code.
2006-06-22 01:28:57 +00:00
}
void
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
release(struct spinlock * lock)
checkpoint. booting second processor. stack is messed up, but thanks to cliff and plan 9 code, at least boots and gets into C code.
2006-06-22 01:28:57 +00:00
{
i think my cmpxchg use was wrong in acquire nesting cli/sti: release shouldn't always enable interrupts separate setup of lapic from starting of other cpus, so cpu() works earlier flag to disable locking in console output make locks work even when curproc==0 (still crashes in clock interrupt)
2006-07-12 11:15:38 +00:00
unsigned who;
if(curproc[cpu()])
who = (unsigned) curproc[cpu()];
else
who = cpu() + 1;
checkpoint. booting second processor. stack is messed up, but thanks to cliff and plan 9 code, at least boots and gets into C code.
2006-06-22 01:28:57 +00:00
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
if(DEBUG) cprintf ("cpu%d: releasing at %x\n", cpu(), getcallerpc(&lock));
Changes to allow use of native x86 ELF compilers, which on my Linux 2.4 box using gcc 3.4.6 don't seem to follow the same conventions as the i386-jos-elf-gcc compilers. Can run make 'TOOLPREFIX=' or edit the Makefile. curproc[cpu()] can now be NULL, indicating that no proc is running. This seemed safer to me than having curproc[0] and curproc[1] both pointing at proc[0] potentially. The old implementation of swtch depended on the stack frame layout used inside swtch being okay to return from on the other stack (exactly the V6 you are not expected to understand this). It also could be called in two contexts: at boot time, to schedule the very first process, and later, on behalf of a process, to sleep or schedule some other process. I split this into two functions: scheduler and swtch. The scheduler is now a separate never-returning function, invoked by each cpu once set up. The scheduler looks like: scheduler() { setjmp(cpu.context); pick proc to schedule blah blah blah longjmp(proc.context) } The new swtch is intended to be called only when curproc[cpu()] is not NULL, that is, only on behalf of a user proc. It does: swtch() { if(setjmp(proc.context) == 0) longjmp(cpu.context) } to save the current proc context and then jump over to the scheduler, running on the cpu stack. Similarly the system call stubs are now in assembly in usys.S to avoid needing to know the details of stack frame layout used by the compiler. Also various changes in the debugging prints.
2006-07-11 01:07:40 +00:00
i think my cmpxchg use was wrong in acquire nesting cli/sti: release shouldn't always enable interrupts separate setup of lapic from starting of other cpus, so cpu() works earlier flag to disable locking in console output make locks work even when curproc==0 (still crashes in clock interrupt)
2006-07-12 11:15:38 +00:00
if(lock->who != who || lock->count < 1 || lock->locked != 1)
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
panic("release");
lock->count -= 1;
if(lock->count < 1){
i think my cmpxchg use was wrong in acquire nesting cli/sti: release shouldn't always enable interrupts separate setup of lapic from starting of other cpus, so cpu() works earlier flag to disable locking in console output make locks work even when curproc==0 (still crashes in clock interrupt)
2006-07-12 11:15:38 +00:00
lock->who = 0;
no more big kernel lock succeeds at usertests.c pipe test
2006-07-12 01:48:35 +00:00
cmpxchg(1, 0, &lock->locked);
sti();
}
}
Reference in a new issue Copy permalink