vserver 1.9.3
[linux-2.6.git] / include / linux / vserver / sched.h
index fd6bc2a..f5982bb 100644 (file)
@@ -8,24 +8,35 @@
 #include <asm/atomic.h>
 #include <asm/param.h>
 
+struct _vx_ticks {
+       uint64_t user_ticks;            /* token tick events */
+       uint64_t sys_ticks;             /* token tick events */
+       uint64_t hold_ticks;            /* token ticks paused */
+       uint64_t unused[5];             /* cacheline ? */
+};
+
 /* context sub struct */
 
 struct _vx_sched {
-       spinlock_t tokens_lock; /* lock for this structure */
+       atomic_t tokens;                /* number of CPU tokens */
+       spinlock_t tokens_lock;         /* lock for token bucket */
+
+       int fill_rate;                  /* Fill rate: add X tokens... */
+       int interval;                   /* Divisor:   per Y jiffies   */
+       int tokens_min;                 /* Limit:     minimum for unhold */
+       int tokens_max;                 /* Limit:     no more than N tokens */
+       uint32_t jiffies;               /* last time accounted */
 
-       int fill_rate;          /* Fill rate: add X tokens... */
-       int interval;           /* Divisor:   per Y jiffies   */
-       atomic_t tokens;        /* number of CPU tokens in this context */
-       int tokens_min;         /* Limit:     minimum for unhold */
-       int tokens_max;         /* Limit:     no more than N tokens */
-       uint32_t jiffies;       /* add an integral multiple of Y to this */
+       int priority_bias;              /* bias offset for priority */
+       cpumask_t cpus_allowed;         /* cpu mask for context */
 
-       uint64_t ticks;         /* token tick events */
-       cpumask_t cpus_allowed; /* cpu mask for context */
+       struct _vx_ticks cpu[NR_CPUS];
 };
 
 static inline void vx_info_init_sched(struct _vx_sched *sched)
 {
+       int i;
+
        /* scheduling; hard code starting values as constants */
        sched->fill_rate        = 1;
        sched->interval         = 4;
@@ -36,6 +47,13 @@ static inline void vx_info_init_sched(struct _vx_sched *sched)
 
        atomic_set(&sched->tokens, HZ >> 2);
        sched->cpus_allowed     = CPU_MASK_ALL;
+       sched->priority_bias    = 0;
+
+       for_each_cpu(i) {
+               sched->cpu[i].user_ticks        = 0;
+               sched->cpu[i].sys_ticks         = 0;
+               sched->cpu[i].hold_ticks        = 0;
+       }
 }
 
 static inline void vx_info_exit_sched(struct _vx_sched *sched)
@@ -45,20 +63,35 @@ static inline void vx_info_exit_sched(struct _vx_sched *sched)
 
 static inline int vx_info_proc_sched(struct _vx_sched *sched, char *buffer)
 {
-       return sprintf(buffer,
-               "Ticks:\t%16lld\n"
+       int length = 0;
+       int i;
+
+       length += sprintf(buffer,
                "Token:\t\t%8d\n"
                "FillRate:\t%8d\n"
-               "Interval:\t%8d\n"              
+               "Interval:\t%8d\n"
                "TokensMin:\t%8d\n"
                "TokensMax:\t%8d\n"
-               ,(unsigned long long)sched->ticks
+               "PrioBias:\t%8d\n"
                ,atomic_read(&sched->tokens)
                ,sched->fill_rate
                ,sched->interval
                ,sched->tokens_min
                ,sched->tokens_max
+               ,sched->priority_bias
                );
+
+       for_each_online_cpu(i) {
+               length += sprintf(buffer + length,
+                       "cpu %d: %lld %lld %lld\n"
+                       ,i
+                       ,(long long)sched->cpu[i].user_ticks
+                       ,(long long)sched->cpu[i].sys_ticks
+                       ,(long long)sched->cpu[i].hold_ticks
+                       );
+       }
+
+       return length;
 }
 
 
@@ -70,9 +103,10 @@ static inline int vx_info_proc_sched(struct _vx_sched *sched, char *buffer)
 
 /*  sched vserver commands */
 
-#define VCMD_set_sched         VC_CMD(SCHED, 1, 2)
+#define VCMD_set_sched_v2      VC_CMD(SCHED, 1, 2)
+#define VCMD_set_sched         VC_CMD(SCHED, 1, 3)
 
-struct  vcmd_set_sched_v2 {
+struct vcmd_set_sched_v2 {
        int32_t fill_rate;
        int32_t interval;
        int32_t tokens;
@@ -81,16 +115,38 @@ struct  vcmd_set_sched_v2 {
        uint64_t cpu_mask;
 };
 
+struct vcmd_set_sched_v3 {
+       uint32_t set_mask;
+       int32_t fill_rate;
+       int32_t interval;
+       int32_t tokens;
+       int32_t tokens_min;
+       int32_t tokens_max;
+       int32_t priority_bias;
+};
+
+
+#define VXSM_FILL_RATE         0x0001
+#define VXSM_INTERVAL          0x0002
+#define VXSM_TOKENS            0x0010
+#define VXSM_TOKENS_MIN                0x0020
+#define VXSM_TOKENS_MAX                0x0040
+#define VXSM_PRIO_BIAS         0x0100
+
 #define SCHED_KEEP             (-2)
 
 #ifdef __KERNEL__
 
 extern int vc_set_sched_v1(uint32_t, void __user *);
+extern int vc_set_sched_v2(uint32_t, void __user *);
 extern int vc_set_sched(uint32_t, void __user *);
 
 
 #define VAVAVOOM_RATIO         50
 
+#define MAX_PRIO_BIAS          20
+#define MIN_PRIO_BIAS          -20
+
 #include "context.h"
 
 
@@ -116,24 +172,48 @@ static inline int vx_need_resched(struct task_struct *p)
 {
 #ifdef CONFIG_VSERVER_HARDCPU
        struct vx_info *vxi = p->vx_info;
+#endif
+       int slice = --p->time_slice;
 
+#ifdef CONFIG_VSERVER_HARDCPU
        if (vxi) {
                int tokens;
 
-               p->time_slice--;
-               if (atomic_read(&vxi->vx_usecnt) < 1)
-                       printk("need_resched: p=%p, s=%ld, ref=%d, id=%d/%d\n",
-                               p, p->state, atomic_read(&vxi->vx_usecnt),
-                               vxi->vx_id, p->xid);
                if ((tokens = vx_tokens_avail(vxi)) > 0)
                        vx_consume_token(vxi);
-               return ((p->time_slice == 0) || (tokens < 1));
+               /* for tokens > 0, one token was consumed */
+               if (tokens < 2)
+                       return 1;
        }
 #endif
-       p->time_slice--;
-       return (p->time_slice == 0);
+       return (slice == 0);
+}
+
+
+static inline void vx_onhold_inc(struct vx_info *vxi)
+{
+       int onhold = atomic_read(&vxi->cvirt.nr_onhold);
+
+       atomic_inc(&vxi->cvirt.nr_onhold);
+       if (!onhold)
+               vxi->cvirt.onhold_last = jiffies;
+}
+
+static inline void __vx_onhold_update(struct vx_info *vxi)
+{
+       int cpu = smp_processor_id();
+       uint32_t now = jiffies;
+       uint32_t delta = now - vxi->cvirt.onhold_last;
+
+       vxi->cvirt.onhold_last = now;
+       vxi->sched.cpu[cpu].hold_ticks += delta;
 }
 
+static inline void vx_onhold_dec(struct vx_info *vxi)
+{
+       if (atomic_dec_and_test(&vxi->cvirt.nr_onhold))
+               __vx_onhold_update(vxi);
+}
 
 #endif /* __KERNEL__ */