From 53ff9c90f6b1b50e2585d27e41d50f4762548dff Mon Sep 17 00:00:00 2001
From: Marc Fiuczynski <mef@cs.princeton.edu>
Date: Sat, 18 Sep 2004 11:48:31 +0000
Subject: [PATCH] CKRM e15 IO controller framework --- cfq-iosched.c remains
 unchanged for now, hence the controller does nothing until we get a new,
 ported version of this file

---
 arch/i386/kernel/entry.S    |  2 ++
 arch/ppc/kernel/misc.S      |  2 ++
 drivers/block/Makefile      |  3 ++-
 drivers/block/elevator.c    | 10 ++++++++-
 drivers/block/ll_rw_blk.c   | 45 ++++++++++++++++++++++++++++---------
 include/asm-i386/unistd.h   |  4 +++-
 include/asm-ppc/unistd.h    |  4 +++-
 include/asm-x86_64/unistd.h |  6 ++++-
 include/linux/elevator.h    |  5 +++++
 include/linux/fs.h          | 12 ++++++++++
 include/linux/init_task.h   |  1 +
 include/linux/sched.h       |  2 ++
 init/Kconfig                | 13 +++++++++++
 kernel/fork.c               |  1 +
 14 files changed, 94 insertions(+), 16 deletions(-)

diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S
index bb91de327..3ac74183c 100644
--- a/arch/i386/kernel/entry.S
+++ b/arch/i386/kernel/entry.S
@@ -1030,5 +1030,7 @@ ENTRY(sys_call_table)
 	.long sys_mq_notify
 	.long sys_mq_getsetattr
 	.long sys_ni_syscall		/* reserved for kexec */
+	.long sys_ioprio_set
+	.long sys_ioprio_get		/* 285 */
 
 syscall_table_size=(.-sys_call_table)
diff --git a/arch/ppc/kernel/misc.S b/arch/ppc/kernel/misc.S
index 017da4476..81a72414a 100644
--- a/arch/ppc/kernel/misc.S
+++ b/arch/ppc/kernel/misc.S
@@ -1450,3 +1450,5 @@ _GLOBAL(sys_call_table)
 	.long sys_mq_notify
 	.long sys_mq_getsetattr
 	.long sys_ni_syscall		/* 268 reserved for sys_kexec_load */
+	.long sys_ioprio_set
+	.long sys_ioprio_get		
diff --git a/drivers/block/Makefile b/drivers/block/Makefile
index 2654b5b76..c66498bad 100644
--- a/drivers/block/Makefile
+++ b/drivers/block/Makefile
@@ -13,12 +13,13 @@
 # kblockd threads
 #
 
-obj-y	:= elevator.o ll_rw_blk.o ioctl.o genhd.o scsi_ioctl.o
+obj-y	:= elevator.o ll_rw_blk.o ioctl.o genhd.o scsi_ioctl.o ckrm-iostub.o
 
 obj-$(CONFIG_IOSCHED_NOOP)	+= noop-iosched.o
 obj-$(CONFIG_IOSCHED_AS)	+= as-iosched.o
 obj-$(CONFIG_IOSCHED_DEADLINE)	+= deadline-iosched.o
 obj-$(CONFIG_IOSCHED_CFQ)	+= cfq-iosched.o
+obj-$(CONFIG_CKRM_RES_BLKIO)	+= ckrm-io.o
 obj-$(CONFIG_MAC_FLOPPY)	+= swim3.o
 obj-$(CONFIG_BLK_DEV_FD)	+= floppy.o
 obj-$(CONFIG_BLK_DEV_FD98)	+= floppy98.o
diff --git a/drivers/block/elevator.c b/drivers/block/elevator.c
index 35c9385ac..950eb9923 100644
--- a/drivers/block/elevator.c
+++ b/drivers/block/elevator.c
@@ -339,6 +339,14 @@ void elv_put_request(request_queue_t *q, struct request *rq)
 		e->elevator_put_req_fn(q, rq);
 }
 
+void elv_set_congested(request_queue_t *q)
+{
+	elevator_t *e = &q->elevator;
+
+	if (e->elevator_set_congested_fn)
+		e->elevator_set_congested_fn(q);
+}
+
 int elv_may_queue(request_queue_t *q, int rw)
 {
 	elevator_t *e = &q->elevator;
@@ -346,7 +354,7 @@ int elv_may_queue(request_queue_t *q, int rw)
 	if (e->elevator_may_queue_fn)
 		return e->elevator_may_queue_fn(q, rw);
 
-	return 0;
+	return 1;
 }
 
 void elv_completed_request(request_queue_t *q, struct request *rq)
diff --git a/drivers/block/ll_rw_blk.c b/drivers/block/ll_rw_blk.c
index 5a570baa6..49ff5e0b7 100644
--- a/drivers/block/ll_rw_blk.c
+++ b/drivers/block/ll_rw_blk.c
@@ -1594,6 +1594,10 @@ static struct request *get_request(request_queue_t *q, int rw, int gfp_mask)
 	struct io_context *ioc = get_io_context(gfp_mask);
 
 	spin_lock_irq(q->queue_lock);
+
+	if (!elv_may_queue(q, rw))
+		goto out_lock;
+
 	if (rl->count[rw]+1 >= q->nr_requests) {
 		/*
 		 * The queue will fill after this allocation, so set it as
@@ -1607,15 +1611,12 @@ static struct request *get_request(request_queue_t *q, int rw, int gfp_mask)
 		}
 	}
 
-	if (blk_queue_full(q, rw)
-			&& !ioc_batching(ioc) && !elv_may_queue(q, rw)) {
-		/*
-		 * The queue is full and the allocating process is not a
-		 * "batcher", and not exempted by the IO scheduler
-		 */
-		spin_unlock_irq(q->queue_lock);
-		goto out;
-	}
+	/*
+	 * The queue is full and the allocating process is not a
+	 * "batcher", and not exempted by the IO scheduler
+	 */
+	if (blk_queue_full(q, rw) && !ioc_batching(ioc))
+		goto out_lock;
 
 	rl->count[rw]++;
 	if (rl->count[rw] >= queue_congestion_on_threshold(q))
@@ -1633,8 +1634,7 @@ static struct request *get_request(request_queue_t *q, int rw, int gfp_mask)
 		 */
 		spin_lock_irq(q->queue_lock);
 		freed_request(q, rw);
-		spin_unlock_irq(q->queue_lock);
-		goto out;
+		goto out_lock;
 	}
 
 	if (ioc_batching(ioc))
@@ -1664,6 +1664,11 @@ static struct request *get_request(request_queue_t *q, int rw, int gfp_mask)
 out:
 	put_io_context(ioc);
 	return rq;
+out_lock:
+	if (!rq)
+		elv_set_congested(q);
+	spin_unlock_irq(q->queue_lock);
+	goto out;
 }
 
 /*
@@ -3168,3 +3173,21 @@ void blk_unregister_queue(struct gendisk *disk)
 		kobject_put(&disk->kobj);
 	}
 }
+
+asmlinkage int sys_ioprio_set(int ioprio)
+{
+	if (ioprio < IOPRIO_IDLE || ioprio > IOPRIO_RT)
+		return -EINVAL;
+	if (ioprio == IOPRIO_RT && !capable(CAP_SYS_ADMIN))
+		return -EACCES;
+
+	printk("%s: set ioprio %d\n", current->comm, ioprio);
+	current->ioprio = ioprio;
+	return 0;
+}
+
+asmlinkage int sys_ioprio_get(void)
+{
+	return current->ioprio;
+}
+
diff --git a/include/asm-i386/unistd.h b/include/asm-i386/unistd.h
index 72b388bf5..d27db1931 100644
--- a/include/asm-i386/unistd.h
+++ b/include/asm-i386/unistd.h
@@ -289,8 +289,10 @@
 #define __NR_mq_notify		(__NR_mq_open+4)
 #define __NR_mq_getsetattr	(__NR_mq_open+5)
 #define __NR_sys_kexec_load	283
+#define __NR_ioprio_set		284
+#define __NR_ioprio_get		285
 
-#define NR_syscalls 284
+#define NR_syscalls 286
 
 #ifndef __KERNEL_SYSCALLS_NO_ERRNO__
 /* user-visible error numbers are in the range -1 - -124: see <asm-i386/errno.h> */
diff --git a/include/asm-ppc/unistd.h b/include/asm-ppc/unistd.h
index 21774ed93..64e443d47 100644
--- a/include/asm-ppc/unistd.h
+++ b/include/asm-ppc/unistd.h
@@ -273,8 +273,10 @@
 #define __NR_mq_notify		266
 #define __NR_mq_getsetattr	267
 #define __NR_kexec_load		268
+#define __NR_ioprio_set		269
+#define __NR_ioprio_get		270
 
-#define __NR_syscalls		269
+#define __NR_syscalls		271
 
 #define __NR(n)	#n
 
diff --git a/include/asm-x86_64/unistd.h b/include/asm-x86_64/unistd.h
index 311e25a4f..81e4e85ba 100644
--- a/include/asm-x86_64/unistd.h
+++ b/include/asm-x86_64/unistd.h
@@ -552,8 +552,12 @@ __SYSCALL(__NR_mq_notify, sys_mq_notify)
 __SYSCALL(__NR_mq_getsetattr, sys_mq_getsetattr)
 #define __NR_kexec_load 	246
 __SYSCALL(__NR_kexec_load, sys_ni_syscall)
+#define __NR_ioprio_set		247
+__SYSCALL(__NR_ioprio_set, sys_ioprio_set);
+#define __NR_ioprio_get		248
+__SYSCALL(__NR_ioprio_get, sys_ioprio_get);
 
-#define __NR_syscall_max __NR_kexec_load
+#define __NR_syscall_max __NR_ioprio_get
 #ifndef __NO_STUBS
 
 /* user-visible error numbers are in the range -1 - -4095 */
diff --git a/include/linux/elevator.h b/include/linux/elevator.h
index 27e8183f4..b42a9c4e2 100644
--- a/include/linux/elevator.h
+++ b/include/linux/elevator.h
@@ -17,6 +17,7 @@ typedef void (elevator_requeue_req_fn) (request_queue_t *, struct request *);
 typedef struct request *(elevator_request_list_fn) (request_queue_t *, struct request *);
 typedef void (elevator_completed_req_fn) (request_queue_t *, struct request *);
 typedef int (elevator_may_queue_fn) (request_queue_t *, int);
+typedef void (elevator_set_congested_fn) (request_queue_t *);
 
 typedef int (elevator_set_req_fn) (request_queue_t *, struct request *, int);
 typedef void (elevator_put_req_fn) (request_queue_t *, struct request *);
@@ -45,6 +46,7 @@ struct elevator_s
 	elevator_put_req_fn *elevator_put_req_fn;
 
 	elevator_may_queue_fn *elevator_may_queue_fn;
+	elevator_set_congested_fn *elevator_set_congested_fn;
 
 	elevator_init_fn *elevator_init_fn;
 	elevator_exit_fn *elevator_exit_fn;
@@ -74,6 +76,7 @@ extern struct request *elv_latter_request(request_queue_t *, struct request *);
 extern int elv_register_queue(request_queue_t *q);
 extern void elv_unregister_queue(request_queue_t *q);
 extern int elv_may_queue(request_queue_t *, int);
+extern void elv_set_congested(request_queue_t *);
 extern void elv_completed_request(request_queue_t *, struct request *);
 extern int elv_set_request(request_queue_t *, struct request *, int);
 extern void elv_put_request(request_queue_t *, struct request *);
@@ -119,4 +122,6 @@ extern int elv_try_last_merge(request_queue_t *, struct bio *);
 #define ELEVATOR_INSERT_BACK	2
 #define ELEVATOR_INSERT_SORT	3
 
+#define RQ_ELV_DATA(rq)		(rq)->elevator_private
+
 #endif
diff --git a/include/linux/fs.h b/include/linux/fs.h
index e83d8e4dd..996b7378e 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -1592,5 +1592,17 @@ static inline void free_secdata(void *secdata)
 { }
 #endif	/* CONFIG_SECURITY */
 
+/* io priorities */
+
+#define IOPRIO_NR      21
+
+#define IOPRIO_IDLE	0
+#define IOPRIO_NORM	10
+#define IOPRIO_RT	20
+
+asmlinkage int sys_ioprio_set(int ioprio);
+asmlinkage int sys_ioprio_get(void);
+
+
 #endif /* __KERNEL__ */
 #endif /* _LINUX_FS_H */
diff --git a/include/linux/init_task.h b/include/linux/init_task.h
index 9bcf2db6d..3a43ff49b 100644
--- a/include/linux/init_task.h
+++ b/include/linux/init_task.h
@@ -116,6 +116,7 @@ extern struct group_info init_groups;
 	.vx_info	= NULL,						\
 	.nid		= 0,						\
 	.nx_info	= NULL,						\
+	.ioprio		= IOPRIO_NORM,					\
 }
 
 
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 93f3c3230..2253b0cc6 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -570,6 +570,8 @@ struct task_struct {
 
 	struct io_context *io_context;
 
+	int ioprio;
+
 	unsigned long ptrace_message;
 	siginfo_t *last_siginfo; /* For ptrace use.  */
 
diff --git a/init/Kconfig b/init/Kconfig
index 89ec58c3f..b73e02043 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -178,6 +178,19 @@ config CKRM_CPU_MONITOR
 	
 	  Say N if unsure, Y to use the feature.
 
+config CKRM_RES_BLKIO
+	tristate " Disk I/O Resource Controller"
+	depends on CKRM_TYPE_TASKCLASS && IOSCHED_CFQ
+	default m
+	help
+	  Provides a resource controller for best-effort block I/O 
+	  bandwidth control. The controller attempts this by proportional 
+	  servicing of requests in the I/O scheduler. However, seek
+	  optimizations and reordering by device drivers/disk controllers may
+	  alter the actual bandwidth delivered to a class.
+	
+	  Say N if unsure, Y to use the feature.
+
 config CKRM_TYPE_SOCKETCLASS
 	bool "Class Manager for socket groups"
 	depends on CKRM
diff --git a/kernel/fork.c b/kernel/fork.c
index df85a9daa..144311e8c 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1128,6 +1128,7 @@ struct task_struct *copy_process(unsigned long clone_flags,
 	} else
 		link_pid(p, p->pids + PIDTYPE_TGID, &p->group_leader->pids[PIDTYPE_TGID].pid);
 
+	p->ioprio = current->ioprio;
 	nr_threads++;
 	/* p is copy of current */
 	vxi = p->vx_info;
-- 
2.47.0