OpenVZ Forum


Home » Mailing lists » Devel » [RFC][PATCH] VPIDs: Virtualization of PIDs (OpenVZ approach)
[RFC][PATCH 5/7] VPIDs: vpid/pid conversion in VPID enabled case [message #1158 is a reply to message #1153] Thu, 02 February 2006 16:30 Go to previous messageGo to previous message
Kirill Korotaev is currently offline  Kirill Korotaev
Messages: 137
Registered: January 2006
Senior Member
This is the main patch which contains all vpid-to-pid conversions
and auxilliary stuff. Virtual pids are distinguished from real ones
by the VPID_BIT bit set. Conversion from vpid to pid and vice versa
is performed in two ways: fast way, when vpid and it's according pid
differ only in VPID_BIT bit set ("linear" case), and more complex way,
when pid may correspond to any vpid ("sparse" case) - in this case we
use a hash-table based mapping.

Note that this patch implies that we have a public vps_info_t pointer
type to represent VPS (otherwise it is useless) so that it can be used
in any virtualisation solution. Virtualization solution should have the
following "interface":
1. vps_info_t has member int id;
2. vps_info_t has member struct task_struct *init_task;
3. the following macros/functions are defined:
a. inside_vps() - returns true if current task is now
inside VPS;
b. task_inside_vps(task_t *) - returns true if task
belongs to VPS;
c. current_vps() - returns vps_info_t for current VPS;
d. task_vps(task_t *) - returns vps_info_t that task
belongs to;
e. set_sparce_vpid(vps_info_t) - switches VPS into state
when "sparce" conversion is used;
f. sparse_vpid(vps_info_t) - returns true if vps is in
"sparce" state and false if it is in "linear";
g. get_vps_tasks_num(vps_info_t) - returns the number of
tasks that belong to VPS.

Kirill

--- ./include/linux/pid.h.vpid_virt 2006-02-02 14:32:41.162807472 +0300
+++ ./include/linux/pid.h 2006-02-02 14:33:17.963212960 +0300
@@ -10,11 +10,17 @@ enum pid_type
PIDTYPE_MAX
};

+#define VPID_BIT 10
+#define VPID_DIV (1 << VPID_BIT)
+
struct pid
{
/* Try to keep pid_chain in the same cacheline as nr for find_pid */
int nr;
struct hlist_node pid_chain;
+#ifdef CONFIG_VIRTUAL_PIDS
+ int vnr;
+#endif
/* list of pids with the same nr, only one of them is in the hash */
struct list_head pid_list;
};
@@ -30,6 +36,52 @@ struct pid
#define comb_pid_to_vpid(pid) (pid)
#define alloc_vpid(pid, vpid) (pid)
#define free_vpid(vpid) do { } while (0)
+#else /* CONFIG_VIRTUAL_PIDS */
+#define __is_virtual_pid(pid) ((pid) & VPID_DIV)
+#define is_virtual_pid(pid) (__is_virtual_pid(pid) || \
+ (((pid) == 1) && inside_vps()))
+
+extern int vpid_to_pid(int pid);
+extern int __vpid_to_pid(int pid);
+extern pid_t pid_type_to_vpid(int type, pid_t pid);
+extern pid_t __pid_type_to_vpid(int type, pid_t pid);
+
+static inline int comb_vpid_to_pid(int vpid)
+{
+ int pid = vpid;
+
+ if (vpid > 0) {
+ pid = vpid_to_pid(vpid);
+ if (unlikely(pid < 0))
+ return 0;
+ } else if (vpid < 0) {
+ pid = vpid_to_pid(-vpid);
+ if (unlikely(pid < 0))
+ return 0;
+ pid = -pid;
+ }
+ return pid;
+}
+
+static inline int comb_pid_to_vpid(int pid)
+{
+ int vpid = pid;
+
+ if (pid > 0) {
+ vpid = pid_type_to_vpid(PIDTYPE_PID, pid);
+ if (unlikely(vpid < 0))
+ return 0;
+ } else if (pid < 0) {
+ vpid = pid_type_to_vpid(PIDTYPE_PGID, -pid);
+ if (unlikely(vpid < 0))
+ return 0;
+ vpid = -vpid;
+ }
+ return vpid;
+}
+
+extern int alloc_vpid(int pid, int vpid);
+extern void free_vpid(int vpid);
#endif

#define pid_task(elem, type) \
--- ./include/linux/sched.h.vpid_virt 2006-02-02 14:32:41.164807168 +0300
+++ ./include/linux/sched.h 2006-02-02 14:58:53.129832160 +0300
@@ -1327,6 +1327,80 @@ static inline pid_t get_task_ppid(struct
return 0;
return (p->pid > 1 ? p->group_leader->real_parent->tgid : 0);
}
+#else
+static inline pid_t virt_pid(struct task_struct *tsk)
+{
+ return tsk->pids[PIDTYPE_PID].vnr;
+}
+
+static inline pid_t virt_tgid(struct task_struct *tsk)
+{
+ return tsk->pids[PIDTYPE_TGID].vnr;
+}
+
+static inline pid_t virt_pgid(struct task_struct *tsk)
+{
+ return tsk->pids[PIDTYPE_PGID].vnr;
+}
+
+static inline pid_t virt_sid(struct task_struct *tsk)
+{
+ return tsk->pids[PIDTYPE_SID].vnr;
+}
+
+static inline pid_t get_task_pid_ve(struct task_struct *tsk,
+ struct task_struct *ve_tsk)
+{
+ return task_inside_vps(ve_tsk) ? virt_pid(tsk) : tsk->pid;
+}
+
+static inline pid_t get_task_pid(struct task_struct *tsk)
+{
+ return inside_vps() ? virt_pid(tsk) : tsk->pid;
+}
+
+static inline pid_t get_task_tgid(struct task_struct *tsk)
+{
+ return inside_vps() ? virt_tgid(tsk) : tsk->pid;
+}
+
+static inline pid_t get_task_pgid(struct task_struct *tsk)
+{
+ return inside_vps() ? virt_pgid(tsk) : tsk->signal->pgrp;
+}
+
+static inline pid_t get_task_sid(struct task_struct *tsk)
+{
+ return inside_vps() ? virt_sid(tsk) : tsk->signal->session;
+}
+
+static inline int set_virt_pid(struct task_struct *tsk, pid_t pid)
+{
+ tsk->pids[PIDTYPE_PID].vnr = pid;
+ return pid;
+}
+
+static inline void set_virt_tgid(struct task_struct *tsk, pid_t pid)
+{
+ tsk->pids[PIDTYPE_TGID].vnr = pid;
+}
+
+static inline void set_virt_pgid(struct task_struct *tsk, pid_t pid)
+{
+ tsk->pids[PIDTYPE_PGID].vnr = pid;
+}
+
+static inline void set_virt_sid(struct task_struct *tsk, pid_t pid)
+{
+ tsk->pids[PIDTYPE_SID].vnr = pid;
+}
+
+static inline pid_t get_task_ppid(struct task_struct *p)
+{
+ if (!pid_alive(p))
+ return 0;
+ return (get_task_pid(p) > 1) ? get_task_pid(p->real_parent) : 0;
+}
#endif

/* set thread flags in other task's structures
--- ./kernel/pid.c.vpid_virt 2006-02-02 14:15:35.165782728 +0300
+++ ./kernel/pid.c 2006-02-02 14:58:34.632644160 +0300
@@ -27,6 +27,14 @@
#include <linux/bootmem.h>
#include <linux/hash.h>

+#ifdef CONFIG_VIRTUAL_PIDS
+static void __free_vpid(int vpid, struct task_struct *ve_tsk);
+#define PIDMAP_NRFREE (BITS_PER_PAGE / 2)
+#else
+#define __free_vpid(vpid, tsk) do { } while (0)
+#define PIDMAP_NRFREE BITS_PER_PAGE
+#endif
+
#define pid_hashfn(nr) hash_long((unsigned long)nr, pidhash_shift)
static struct hlist_head *pid_hash[PIDTYPE_MAX];
static int pidhash_shift;
@@ -58,7 +66,7 @@ typedef struct pidmap {
} pidmap_t;

static pidmap_t pidmap_array[PIDMAP_ENTRIES] =
- { [ 0 ... PIDMAP_ENTRIES-1 ] = { ATOMIC_INIT(BITS_PER_PAGE), NULL } };
+ { [ 0 ... PIDMAP_ENTRIES-1 ] = { ATOMIC_INIT(PIDMAP_NRFREE), NULL } };

static __cacheline_aligned_in_smp DEFINE_SPINLOCK(pidmap_lock);

@@ -67,6 +75,7 @@ fastcall void free_pidmap(int pid)
pidmap_t *map = pidmap_array + pid / BITS_PER_PAGE;
int offset = pid & BITS_PER_PAGE_MASK;

+ BUG_ON(__is_virtual_pid(pid) || pid == 1);
clear_bit(offset, map->page);
atomic_inc(&map->nr_free);
}
@@ -77,6 +86,8 @@ int alloc_pidmap(void)
pidmap_t *map;

pid = last + 1;
+ if (__is_virtual_pid(pid))
+ pid += VPID_DIV;
if (pid >= pid_max)
pid = RESERVED_PIDS;
offset = pid & BITS_PER_PAGE_MASK;
@@ -107,6 +118,8 @@ int alloc_pidmap(void)
}
offset = find_next_offset(map, offset);
pid = mk_pid(map, offset);
+ if (__is_virtual_pid(pid))
+ pid += VPID_DIV;
/*
* find_next_offset() found a bit, the pid from it
* is in-bounds, and if we fell back to the last
@@ -127,6 +140,8 @@ int alloc_pidmap(void)
break;
}
pid = mk_pid(map, offset);
+ if (__is_virtual_pid(pid))
+ pid += VPID_DIV;
}
return -1;
}
@@ -201,6 +216,7 @@ void fastcall detach_pid(task_t *task, e
if (tmp != type && find_pid(tmp, nr))
return;

+ __free_vpid(task->pids[type].vnr, task);
free_pidmap(nr);
}

@@ -234,6 +250,9 @@ void switch_exec_pids(task_t *leader, ta

leader->pid = leader->tgid = thread->pid;
thread->pid = thread->tgid;
+ set_virt_tgid(leader, virt_pid(thread));
+ set_virt_pid(leader, virt_pid(thread));
+ set_virt_pid(thread, virt_tgid(thread));

attach_pid(thread, PIDTYPE_PID, thread->pid);
attach_pid(thread, PIDTYPE_TGID, thread->tgid);
@@ -247,6 +266,344 @@ void switch_exec_pids(task_t *leader, ta
attach_pid(leader, PIDTYPE_SID, leader->signal->session);
}

+#ifdef CONFIG_VIRTUAL_PIDS
+/* Virtual PID bits.
+ *
+ * At the moment all internal structures in kernel store real global pid.
+ * The only place, where virtual PID is used, is at user frontend. We
+ * remap virtual pids obtained from user to global ones (vpid_to_pid) and
+ * map globals to virtuals before showing them to user (virt_pid_type).
+ *
+ * We hold virtual PIDs inside struct pid, so map global -> virtual is easy.
+ */
+
+pid_t __pid_type_to_vpid(int type, pid_t pid)
+{
+ struct pid * p;
+
+ if (unlikely(is_virtual_pid(pid)))
+ return -1;
+
+ read_lock(&tasklist_lock);
+ p = find_pid(type, pid);
+ if (p) {
+ pid = p->vnr;
+ } else {
+ pid = -1;
+ }
+ read_unlock(&tasklist_lock);
+ return pid;
+}
+
+pid_t pid_type_to_vpid(int type, pid_t pid)
+{
+ int vpid;
+
+ if (unlikely(pid <= 0))
+ return pid;
+
+ BUG_ON(is_virtual_pid(pid));
+
+ if (!inside_vps())
+ return pid;
+
+ vpid = __pid_type_to_vpid(type, pid);
+ if (unlikely(vpid == -1)) {
+ /* It is allowed: global pid can be used everywhere.
+ * This can happen, when kernel remembers stray pids:
+ * signal queues, locks etc.
+ */
+ vpid = pid;
+ }
+ return vpid;
+}
+
+/* To map virtual pids to global we maintain special hash table.
+ *
+ * Mapping entries are allocated when a process with non-trivial
+ * mapping is forked, which is possible only after VE migrated.
+ * Mappings are destroyed, when a global pid is removed from global
+ * pidmap, which means we do not need to refcount mappings.
+ */
+
+static struct hlist_head *vpid_hash;
+
+struct vpid_mapping
+{
+ int pid;
+ int vpid;
+ int vpsid;
+ struct hlist_node link;
+};
+
+static kmem_cache_t *vpid_mapping_cachep;
+
+static inline int vpid_hashfn(int vnr, int vpsid)
+{
+ return hash_long((unsigned long)(vnr + (vpsid << 16)), pidhash_shift);
+}
+
+struct vpid_mapping *__lookup_vpid_mapping(int vnr, int vpsid)
+{
+ struct hlist_node *elem;
+ struct vpid_mapping *map;
+
+ hlist_for_each_entry(map, elem,
+ &vpid_hash[vpid_hashfn(vnr, vpsid)], link) {
+ if (map->vpid == vnr && map-> ...

[ Show the rest of the message ]

Report message to a moderator

[Message index]
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: vzmemcheck displays wrong values?
Next Topic: openvz + ipv6
Goto Forum:
  

[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Sat Aug 30 12:16:38 GMT 2025

Total time taken to generate the page: 0.08005 seconds
Powered by FUDforum Powered by Virtuozzo Containers