From 4087a6e7fc773ba4eb217dfc196dfe1eee84b25d Mon Sep 17 00:00:00 2001
From: Frans Kaashoek <kaashoek@mit.edu>
Date: Wed, 10 Aug 2022 20:35:42 -0400
Subject: [PATCH 1/4] Read and write p->killed using atomics
---
kernel/console.c | 2 +-
kernel/pipe.c | 4 ++--
kernel/proc.c | 4 ++--
kernel/sysproc.c | 2 +-
kernel/trap.c | 6 +++---
5 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/kernel/console.c b/kernel/console.c
index 23a2d35..b8fa1de 100644
--- a/kernel/console.c
+++ b/kernel/console.c
@@ -89,7 +89,7 @@ consoleread(int user_dst, uint64 dst, int n)
// wait until interrupt handler has put some
// input into cons.buffer.
while(cons.r == cons.w){
- if(myproc()->killed){
+ if(__sync_add_and_fetch(&(myproc()->killed), 0)){
release(&cons.lock);
return -1;
}
diff --git a/kernel/pipe.c b/kernel/pipe.c
index b6eefb9..e438d7e 100644
--- a/kernel/pipe.c
+++ b/kernel/pipe.c
@@ -81,7 +81,7 @@ pipewrite(struct pipe *pi, uint64 addr, int n)
acquire(&pi->lock);
while(i < n){
- if(pi->readopen == 0 || pr->killed){
+ if(pi->readopen == 0 || __sync_add_and_fetch(&pr->killed,0)){
release(&pi->lock);
return -1;
}
@@ -111,7 +111,7 @@ piperead(struct pipe *pi, uint64 addr, int n)
acquire(&pi->lock);
while(pi->nread == pi->nwrite && pi->writeopen){ //DOC: pipe-empty
- if(pr->killed){
+ if(__sync_add_and_fetch(&pr->killed,0)){
release(&pi->lock);
return -1;
}
diff --git a/kernel/proc.c b/kernel/proc.c
index 2d0ffa1..221f0f8 100644
--- a/kernel/proc.c
+++ b/kernel/proc.c
@@ -422,7 +422,7 @@ wait(uint64 addr)
}
// No point waiting if we don't have any children.
- if(!havekids || p->killed){
+ if(!havekids || __sync_add_and_fetch(&p->killed, 0)){
release(&wait_lock);
return -1;
}
@@ -588,7 +588,7 @@ kill(int pid)
for(p = proc; p < &proc[NPROC]; p++){
acquire(&p->lock);
if(p->pid == pid){
- p->killed = 1;
+ __sync_bool_compare_and_swap(&p->killed, 0, 1);
if(p->state == SLEEPING){
// Wake process from sleep().
p->state = RUNNABLE;
diff --git a/kernel/sysproc.c b/kernel/sysproc.c
index e8bcda9..61b715b 100644
--- a/kernel/sysproc.c
+++ b/kernel/sysproc.c
@@ -63,7 +63,7 @@ sys_sleep(void)
acquire(&tickslock);
ticks0 = ticks;
while(ticks - ticks0 < n){
- if(myproc()->killed){
+ if(__sync_add_and_fetch(&(myproc()->killed), 0)){
release(&tickslock);
return -1;
}
diff --git a/kernel/trap.c b/kernel/trap.c
index 75fb3ec..b879f01 100644
--- a/kernel/trap.c
+++ b/kernel/trap.c
@@ -53,7 +53,7 @@ usertrap(void)
if(r_scause() == 8){
// system call
- if(p->killed)
+ if(__sync_add_and_fetch(&p->killed, 0))
exit(-1);
// sepc points to the ecall instruction,
@@ -70,10 +70,10 @@ usertrap(void)
} else {
printf("usertrap(): unexpected scause %p pid=%d\n", r_scause(), p->pid);
printf(" sepc=%p stval=%p\n", r_sepc(), r_stval());
- p->killed = 1;
+ __sync_bool_compare_and_swap(&p->killed, 0, 1);
}
- if(p->killed)
+ if(__sync_add_and_fetch(&p->killed, 0))
exit(-1);
// give up the CPU if this is a timer interrupt.
From 975f3b31d3fac2c271df3107263df6ae454a98be Mon Sep 17 00:00:00 2001
From: Frans Kaashoek <kaashoek@mit.edu>
Date: Thu, 11 Aug 2022 07:23:17 -0400
Subject: [PATCH 2/4] Clean up using killed()
---
kernel/console.c | 2 +-
kernel/defs.h | 1 +
kernel/pipe.c | 4 ++--
kernel/proc.c | 8 +++++++-
kernel/sysproc.c | 2 +-
kernel/trap.c | 4 ++--
6 files changed, 14 insertions(+), 7 deletions(-)
diff --git a/kernel/console.c b/kernel/console.c
index b8fa1de..d6eb209 100644
--- a/kernel/console.c
+++ b/kernel/console.c
@@ -89,7 +89,7 @@ consoleread(int user_dst, uint64 dst, int n)
// wait until interrupt handler has put some
// input into cons.buffer.
while(cons.r == cons.w){
- if(__sync_add_and_fetch(&(myproc()->killed), 0)){
+ if(killed(myproc())){
release(&cons.lock);
return -1;
}
diff --git a/kernel/defs.h b/kernel/defs.h
index 62b9292..7181d4d 100644
--- a/kernel/defs.h
+++ b/kernel/defs.h
@@ -90,6 +90,7 @@ void proc_mapstacks(pagetable_t);
pagetable_t proc_pagetable(struct proc *);
void proc_freepagetable(pagetable_t, uint64);
int kill(int);
+int killed(struct proc*);
struct cpu* mycpu(void);
struct cpu* getmycpu(void);
struct proc* myproc();
diff --git a/kernel/pipe.c b/kernel/pipe.c
index e438d7e..f6b501a 100644
--- a/kernel/pipe.c
+++ b/kernel/pipe.c
@@ -81,7 +81,7 @@ pipewrite(struct pipe *pi, uint64 addr, int n)
acquire(&pi->lock);
while(i < n){
- if(pi->readopen == 0 || __sync_add_and_fetch(&pr->killed,0)){
+ if(pi->readopen == 0 || killed(pr)){
release(&pi->lock);
return -1;
}
@@ -111,7 +111,7 @@ piperead(struct pipe *pi, uint64 addr, int n)
acquire(&pi->lock);
while(pi->nread == pi->nwrite && pi->writeopen){ //DOC: pipe-empty
- if(__sync_add_and_fetch(&pr->killed,0)){
+ if(killed(pr)){
release(&pi->lock);
return -1;
}
diff --git a/kernel/proc.c b/kernel/proc.c
index 221f0f8..24680b6 100644
--- a/kernel/proc.c
+++ b/kernel/proc.c
@@ -422,7 +422,7 @@ wait(uint64 addr)
}
// No point waiting if we don't have any children.
- if(!havekids || __sync_add_and_fetch(&p->killed, 0)){
+ if(!havekids || killed(p)){
release(&wait_lock);
return -1;
}
@@ -601,6 +601,12 @@ kill(int pid)
return -1;
}
+int
+killed(struct proc *p)
+{
+ return __sync_add_and_fetch(&p->killed, 0);
+}
+
// Copy to either a user address, or kernel address,
// depending on usr_dst.
// Returns 0 on success, -1 on error.
diff --git a/kernel/sysproc.c b/kernel/sysproc.c
index 61b715b..99a36a7 100644
--- a/kernel/sysproc.c
+++ b/kernel/sysproc.c
@@ -63,7 +63,7 @@ sys_sleep(void)
acquire(&tickslock);
ticks0 = ticks;
while(ticks - ticks0 < n){
- if(__sync_add_and_fetch(&(myproc()->killed), 0)){
+ if(killed(myproc())){
release(&tickslock);
return -1;
}
diff --git a/kernel/trap.c b/kernel/trap.c
index b879f01..f895aea 100644
--- a/kernel/trap.c
+++ b/kernel/trap.c
@@ -53,7 +53,7 @@ usertrap(void)
if(r_scause() == 8){
// system call
- if(__sync_add_and_fetch(&p->killed, 0))
+ if(killed(p))
exit(-1);
// sepc points to the ecall instruction,
@@ -73,7 +73,7 @@ usertrap(void)
__sync_bool_compare_and_swap(&p->killed, 0, 1);
}
- if(__sync_add_and_fetch(&p->killed, 0))
+ if(killed(p))
exit(-1);
// give up the CPU if this is a timer interrupt.
From 429c7b717edd4c23d3666327986052b9b6eb29eb Mon Sep 17 00:00:00 2001
From: Frans Kaashoek <kaashoek@mit.edu>
Date: Thu, 11 Aug 2022 08:42:52 -0400
Subject: [PATCH 3/4] Use atomic store_n and load_n
---
kernel/proc.c | 4 ++--
kernel/trap.c | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/kernel/proc.c b/kernel/proc.c
index 24680b6..b4f8a80 100644
--- a/kernel/proc.c
+++ b/kernel/proc.c
@@ -588,7 +588,7 @@ kill(int pid)
for(p = proc; p < &proc[NPROC]; p++){
acquire(&p->lock);
if(p->pid == pid){
- __sync_bool_compare_and_swap(&p->killed, 0, 1);
+ __atomic_store_n(&p->killed, 1, __ATOMIC_SEQ_CST);
if(p->state == SLEEPING){
// Wake process from sleep().
p->state = RUNNABLE;
@@ -604,7 +604,7 @@ kill(int pid)
int
killed(struct proc *p)
{
- return __sync_add_and_fetch(&p->killed, 0);
+ return __atomic_load_n(&p->killed, __ATOMIC_SEQ_CST);
}
// Copy to either a user address, or kernel address,
diff --git a/kernel/trap.c b/kernel/trap.c
index f895aea..1039911 100644
--- a/kernel/trap.c
+++ b/kernel/trap.c
@@ -70,7 +70,7 @@ usertrap(void)
} else {
printf("usertrap(): unexpected scause %p pid=%d\n", r_scause(), p->pid);
printf(" sepc=%p stval=%p\n", r_sepc(), r_stval());
- __sync_bool_compare_and_swap(&p->killed, 0, 1);
+ __atomic_store_n(&p->killed, 1, __ATOMIC_SEQ_CST);
}
if(killed(p))
From 4f716c8550b406c3e4b3e0c21b986ef99bc06c40 Mon Sep 17 00:00:00 2001
From: Frans Kaashoek <kaashoek@mit.edu>
Date: Thu, 11 Aug 2022 14:22:00 -0400
Subject: [PATCH 4/4] Use p->lock to read p->killed
---
kernel/defs.h | 1 +
kernel/proc.c | 17 +++++++++++++++--
kernel/trap.c | 2 +-
3 files changed, 17 insertions(+), 3 deletions(-)
diff --git a/kernel/defs.h b/kernel/defs.h
index 7181d4d..e38ec00 100644
--- a/kernel/defs.h
+++ b/kernel/defs.h
@@ -91,6 +91,7 @@ pagetable_t proc_pagetable(struct proc *);
void proc_freepagetable(pagetable_t, uint64);
int kill(int);
int killed(struct proc*);
+void setkilled(struct proc*);
struct cpu* mycpu(void);
struct cpu* getmycpu(void);
struct proc* myproc();
diff --git a/kernel/proc.c b/kernel/proc.c
index b4f8a80..5e37cb7 100644
--- a/kernel/proc.c
+++ b/kernel/proc.c
@@ -588,7 +588,7 @@ kill(int pid)
for(p = proc; p < &proc[NPROC]; p++){
acquire(&p->lock);
if(p->pid == pid){
- __atomic_store_n(&p->killed, 1, __ATOMIC_SEQ_CST);
+ p->killed = 1;
if(p->state == SLEEPING){
// Wake process from sleep().
p->state = RUNNABLE;
@@ -601,10 +601,23 @@ kill(int pid)
return -1;
}
+void
+setkilled(struct proc *p)
+{
+ acquire(&p->lock);
+ p->killed = 1;
+ release(&p->lock);
+}
+
int
killed(struct proc *p)
{
- return __atomic_load_n(&p->killed, __ATOMIC_SEQ_CST);
+ int k;
+
+ acquire(&p->lock);
+ k = p->killed;
+ release(&p->lock);
+ return k;
}
// Copy to either a user address, or kernel address,
diff --git a/kernel/trap.c b/kernel/trap.c
index 1039911..44c9cdc 100644
--- a/kernel/trap.c
+++ b/kernel/trap.c
@@ -70,7 +70,7 @@ usertrap(void)
} else {
printf("usertrap(): unexpected scause %p pid=%d\n", r_scause(), p->pid);
printf(" sepc=%p stval=%p\n", r_sepc(), r_stval());
- __atomic_store_n(&p->killed, 1, __ATOMIC_SEQ_CST);
+ setkilled(p);
}
if(killed(p))