vserver 1.9.5.x5
[linux-2.6.git] / fs / proc / base.c
index 434ecb4..28e49e9 100644 (file)
@@ -33,6 +33,7 @@
 #include <linux/security.h>
 #include <linux/ptrace.h>
 #include <linux/vs_network.h>
+#include "internal.h"
 
 /*
  * For hysterical raisins we keep the same inumbers as in the old procfs.
@@ -73,7 +74,12 @@ enum pid_directory_inos {
 #endif
        PROC_TGID_VX_INFO,
        PROC_TGID_IP_INFO,
+#ifdef CONFIG_AUDITSYSCALL
+       PROC_TGID_LOGINUID,
+#endif
        PROC_TGID_FD_DIR,
+       PROC_TGID_OOM_SCORE,
+       PROC_TGID_OOM_ADJUST,
        PROC_TID_INO,
        PROC_TID_STATUS,
        PROC_TID_MEM,
@@ -101,7 +107,12 @@ enum pid_directory_inos {
 #endif
        PROC_TID_VX_INFO,
        PROC_TID_IP_INFO,
+#ifdef CONFIG_AUDITSYSCALL
+       PROC_TID_LOGINUID,
+#endif
        PROC_TID_FD_DIR = 0x8000,       /* 0x8000-0xffff */
+       PROC_TID_OOM_SCORE,
+       PROC_TID_OOM_ADJUST,
 };
 
 struct pid_entry {
@@ -139,6 +150,11 @@ static struct pid_entry tgid_base_stuff[] = {
 #endif
        E(PROC_TGID_VX_INFO,   "vinfo",   S_IFREG|S_IRUGO),
        E(PROC_TGID_IP_INFO,   "ninfo",   S_IFREG|S_IRUGO),
+       E(PROC_TGID_OOM_SCORE, "oom_score",S_IFREG|S_IRUGO),
+       E(PROC_TGID_OOM_ADJUST,"oom_adj", S_IFREG|S_IRUGO|S_IWUSR),
+#ifdef CONFIG_AUDITSYSCALL
+       E(PROC_TGID_LOGINUID, "loginuid", S_IFREG|S_IWUSR|S_IRUGO),
+#endif
        {0,0,NULL,0}
 };
 static struct pid_entry tid_base_stuff[] = {
@@ -166,6 +182,11 @@ static struct pid_entry tid_base_stuff[] = {
 #endif
        E(PROC_TID_VX_INFO,    "vinfo",   S_IFREG|S_IRUGO),
        E(PROC_TID_IP_INFO,    "ninfo",   S_IFREG|S_IRUGO),
+       E(PROC_TID_OOM_SCORE,  "oom_score",S_IFREG|S_IRUGO),
+       E(PROC_TID_OOM_ADJUST, "oom_adj", S_IFREG|S_IRUGO|S_IWUSR),
+#ifdef CONFIG_AUDITSYSCALL
+       E(PROC_TID_LOGINUID, "loginuid", S_IFREG|S_IWUSR|S_IRUGO),
+#endif
        {0,0,NULL,0}
 };
 
@@ -188,20 +209,6 @@ static struct pid_entry tid_attr_stuff[] = {
 
 #undef E
 
-static inline struct task_struct *proc_task(struct inode *inode)
-{
-       return PROC_I(inode)->task;
-}
-
-static inline int proc_type(struct inode *inode)
-{
-       return PROC_I(inode)->type;
-}
-
-int proc_pid_stat(struct task_struct*,char*);
-int proc_pid_status(struct task_struct*,char*);
-int proc_pid_statm(struct task_struct*,char*);
-
 static int proc_fd_link(struct inode *inode, struct dentry **dentry, struct vfsmount **mnt)
 {
        struct task_struct *task = proc_task(inode);
@@ -226,33 +233,6 @@ static int proc_fd_link(struct inode *inode, struct dentry **dentry, struct vfsm
        return -ENOENT;
 }
 
-static int proc_exe_link(struct inode *inode, struct dentry **dentry, struct vfsmount **mnt)
-{
-       struct vm_area_struct * vma;
-       int result = -ENOENT;
-       struct task_struct *task = proc_task(inode);
-       struct mm_struct * mm = get_task_mm(task);
-
-       if (!mm)
-               goto out;
-       down_read(&mm->mmap_sem);
-       vma = mm->mmap;
-       while (vma) {
-               if ((vma->vm_flags & VM_EXECUTABLE) && 
-                   vma->vm_file) {
-                       *mnt = mntget(vma->vm_file->f_vfsmnt);
-                       *dentry = dget(vma->vm_file->f_dentry);
-                       result = 0;
-                       break;
-               }
-               vma = vma->vm_next;
-       }
-       up_read(&mm->mmap_sem);
-       mmput(mm);
-out:
-       return result;
-}
-
 static int proc_cwd_link(struct inode *inode, struct dentry **dentry, struct vfsmount **mnt)
 {
        struct fs_struct *fs;
@@ -351,7 +331,7 @@ static int proc_pid_cmdline(struct task_struct *task, char * buffer)
        if (!mm)
                goto out;
        if (!mm->arg_end)
-               goto out;       /* Shh! No looking before we're done */
+               goto out_mm;    /* Shh! No looking before we're done */
 
        len = mm->arg_end - mm->arg_start;
  
@@ -374,8 +354,8 @@ static int proc_pid_cmdline(struct task_struct *task, char * buffer)
                        res = strnlen(buffer, res);
                }
        }
+out_mm:
        mmput(mm);
-
 out:
        return res;
 }
@@ -409,7 +389,7 @@ static int proc_pid_wchan(struct task_struct *task, char *buffer)
        char *modname;
        const char *sym_name;
        unsigned long wchan, size, offset;
-       char namebuf[128];
+       char namebuf[KSYM_NAME_LEN+1];
 
        wchan = get_wchan(task);
 
@@ -433,6 +413,18 @@ static int proc_pid_schedstat(struct task_struct *task, char *buffer)
 }
 #endif
 
+/* The badness from the OOM killer */
+unsigned long badness(struct task_struct *p, unsigned long uptime);
+static int proc_oom_score(struct task_struct *task, char *buffer)
+{
+       unsigned long points;
+       struct timespec uptime;
+
+       do_posix_clock_monotonic_gettime(&uptime);
+       points = badness(task, uptime.tv_sec);
+       return sprintf(buffer, "%lu\n", points);
+}
+
 /************************************************************************/
 /*                       Here the fs part begins                        */
 /************************************************************************/
@@ -481,7 +473,7 @@ out:
 
 static int proc_permission(struct inode *inode, int mask, struct nameidata *nd)
 {
-       if (vfs_permission(inode, mask) != 0)
+       if (generic_permission(inode, mask, NULL) != 0)
                return -EACCES;
        return proc_check_root(inode);
 }
@@ -706,10 +698,125 @@ static struct file_operations proc_mem_operations = {
        .open           = mem_open,
 };
 
+static ssize_t oom_adjust_read(struct file *file, char *buf,
+                               size_t count, loff_t *ppos)
+{
+       struct task_struct *task = proc_task(file->f_dentry->d_inode);
+       char buffer[8];
+       size_t len;
+       int oom_adjust = task->oomkilladj;
+       loff_t __ppos = *ppos;
+
+       len = sprintf(buffer, "%i\n", oom_adjust);
+       if (__ppos >= len)
+               return 0;
+       if (count > len-__ppos)
+               count = len-__ppos;
+       if (copy_to_user(buf, buffer + __ppos, count))
+               return -EFAULT;
+       *ppos = __ppos + count;
+       return count;
+}
+
+static ssize_t oom_adjust_write(struct file *file, const char *buf,
+                               size_t count, loff_t *ppos)
+{
+       struct task_struct *task = proc_task(file->f_dentry->d_inode);
+       char buffer[8], *end;
+       int oom_adjust;
+
+       if (!capable(CAP_SYS_RESOURCE))
+               return -EPERM;
+       memset(buffer, 0, 8);
+       if (count > 6)
+               count = 6;
+       if (copy_from_user(buffer, buf, count))
+               return -EFAULT;
+       oom_adjust = simple_strtol(buffer, &end, 0);
+       if (oom_adjust < -16 || oom_adjust > 15)
+               return -EINVAL;
+       if (*end == '\n')
+               end++;
+       task->oomkilladj = oom_adjust;
+       if (end - buffer == 0)
+               return -EIO;
+       return end - buffer;
+}
+
+static struct file_operations proc_oom_adjust_operations = {
+       read:           oom_adjust_read,
+       write:          oom_adjust_write,
+};
+
 static struct inode_operations proc_mem_inode_operations = {
        .permission     = proc_permission,
 };
 
+#ifdef CONFIG_AUDITSYSCALL
+#define TMPBUFLEN 21
+static ssize_t proc_loginuid_read(struct file * file, char __user * buf,
+                                 size_t count, loff_t *ppos)
+{
+       struct inode * inode = file->f_dentry->d_inode;
+       struct task_struct *task = proc_task(inode);
+       ssize_t length;
+       char tmpbuf[TMPBUFLEN];
+
+       length = scnprintf(tmpbuf, TMPBUFLEN, "%u",
+                               audit_get_loginuid(task->audit_context));
+       return simple_read_from_buffer(buf, count, ppos, tmpbuf, length);
+}
+
+static ssize_t proc_loginuid_write(struct file * file, const char __user * buf,
+                                  size_t count, loff_t *ppos)
+{
+       struct inode * inode = file->f_dentry->d_inode;
+       char *page, *tmp;
+       ssize_t length;
+       struct task_struct *task = proc_task(inode);
+       uid_t loginuid;
+
+       if (!capable(CAP_AUDIT_CONTROL))
+               return -EPERM;
+
+       if (current != task)
+               return -EPERM;
+
+       if (count > PAGE_SIZE)
+               count = PAGE_SIZE;
+
+       if (*ppos != 0) {
+               /* No partial writes. */
+               return -EINVAL;
+       }
+       page = (char*)__get_free_page(GFP_USER);
+       if (!page)
+               return -ENOMEM;
+       length = -EFAULT;
+       if (copy_from_user(page, buf, count))
+               goto out_free_page;
+
+       loginuid = simple_strtoul(page, &tmp, 10);
+       if (tmp == page) {
+               length = -EINVAL;
+               goto out_free_page;
+
+       }
+       length = audit_set_loginuid(task->audit_context, loginuid);
+       if (likely(length == 0))
+               length = count;
+
+out_free_page:
+       free_page((unsigned long) page);
+       return length;
+}
+
+static struct file_operations proc_loginuid_operations = {
+       .read           = proc_loginuid_read,
+       .write          = proc_loginuid_write,
+};
+#endif
+
 static int proc_pid_follow_link(struct dentry *dentry, struct nameidata *nd)
 {
        struct inode *inode = dentry->d_inode;
@@ -788,11 +895,6 @@ static struct inode_operations proc_pid_link_inode_operations = {
        .follow_link    = proc_pid_follow_link
 };
 
-static inline int pid_alive(struct task_struct *p)
-{
-       return p->pids[PIDTYPE_PID].nr != 0;
-}
-
 #define NUMBUF 10
 
 static int proc_readfd(struct file * filp, void * dirent, filldir_t filldir)
@@ -867,7 +969,7 @@ static int proc_pident_readdir(struct file *filp,
        struct inode *inode = dentry->d_inode;
        struct pid_entry *p;
        ino_t ino;
-       int ret;
+       int ret, hide;
 
        ret = -ENOENT;
        if (!pid_alive(proc_task(inode)))
@@ -898,11 +1000,20 @@ static int proc_pident_readdir(struct file *filp,
                        goto out;
                }
                p = ents + i;
+               hide = vx_flags(VXF_INFO_HIDE, 0);
                while (p->name) {
+                       if (hide) {
+                               switch (p->type) {
+                               case PROC_TGID_VX_INFO:
+                               case PROC_TGID_IP_INFO:
+                                       goto skip;
+                               }
+                       }
                        if (filldir(dirent, p->name, p->len, filp->f_pos,
                                    fake_ino(pid, p->type), p->mode >> 12) < 0)
                                goto out;
                        filp->f_pos++;
+               skip:
                        p++;
                }
        }
@@ -1251,6 +1362,9 @@ static struct file_operations proc_tgid_attr_operations;
 static struct inode_operations proc_tgid_attr_inode_operations;
 #endif
 
+extern int proc_pid_vx_info(struct task_struct *, char *);
+extern int proc_pid_nx_info(struct task_struct *, char *);
+
 /* SMP-safe */
 static struct dentry *proc_pident_lookup(struct inode *dir, 
                                         struct dentry *dentry,
@@ -1331,9 +1445,12 @@ static struct dentry *proc_pident_lookup(struct inode *dir,
                        ei->op.proc_read = proc_pid_status;
                        break;
                case PROC_TID_STAT:
+                       inode->i_fop = &proc_info_file_operations;
+                       ei->op.proc_read = proc_tid_stat;
+                       break;
                case PROC_TGID_STAT:
                        inode->i_fop = &proc_info_file_operations;
-                       ei->op.proc_read = proc_pid_stat;
+                       ei->op.proc_read = proc_tgid_stat;
                        break;
                case PROC_TID_CMDLINE:
                case PROC_TGID_CMDLINE:
@@ -1393,26 +1510,49 @@ static struct dentry *proc_pident_lookup(struct inode *dir,
                        inode->i_fop = &proc_info_file_operations;
                        ei->op.proc_read = proc_pid_schedstat;
                        break;
+#endif
+               case PROC_TID_OOM_SCORE:
+               case PROC_TGID_OOM_SCORE:
+                       inode->i_fop = &proc_info_file_operations;
+                       ei->op.proc_read = proc_oom_score;
+                       break;
+               case PROC_TID_OOM_ADJUST:
+               case PROC_TGID_OOM_ADJUST:
+                       inode->i_fop = &proc_oom_adjust_operations;
+                       break;
+#ifdef CONFIG_AUDITSYSCALL
+               case PROC_TID_LOGINUID:
+               case PROC_TGID_LOGINUID:
+                       inode->i_fop = &proc_loginuid_operations;
+                       break;
 #endif
                case PROC_TID_VX_INFO:
                case PROC_TGID_VX_INFO:
+                       if (task_vx_flags(task, VXF_INFO_HIDE, 0))
+                               goto out_noent;
                        inode->i_fop = &proc_info_file_operations;
                        ei->op.proc_read = proc_pid_vx_info;
                        break;
                case PROC_TID_IP_INFO:
                case PROC_TGID_IP_INFO:
+                       if (task_vx_flags(task, VXF_INFO_HIDE, 0))
+                               goto out_noent;
                        inode->i_fop = &proc_info_file_operations;
                        ei->op.proc_read = proc_pid_nx_info;
                        break;
                default:
                        printk("procfs: impossible type (%d)",p->type);
-                       iput(inode);
-                       return ERR_PTR(-EINVAL);
+                       error = -EINVAL;
+                       goto out_put;
        }
        dentry->d_op = &pid_dentry_operations;
        d_add(dentry, inode);
        return NULL;
 
+out_noent:
+       error=-ENOENT;
+out_put:
+       iput(inode);
 out:
        return ERR_PTR(error);
 }
@@ -1496,14 +1636,14 @@ static int proc_self_readlink(struct dentry *dentry, char __user *buffer,
                              int buflen)
 {
        char tmp[30];
-       sprintf(tmp, "%d", vx_map_pid(current->tgid));
+       sprintf(tmp, "%d", vx_map_tgid(current->tgid));
        return vfs_readlink(dentry,buffer,buflen,tmp);
 }
 
 static int proc_self_follow_link(struct dentry *dentry, struct nameidata *nd)
 {
        char tmp[30];
-       sprintf(tmp, "%d", vx_map_pid(current->tgid));
+       sprintf(tmp, "%d", vx_map_tgid(current->tgid));
        return vfs_follow_link(nd,tmp);
 }