sysfs: fix some bin_vm_ops errors

Commit 86c9508eb1c0ce5aa07b5cf1d36b60c54efc3d7a
"sysfs: don't block indefinitely for unmapped files" in linux-next
crashes the PowerMac G5 when X starts up.  It's caught out by the way
powerpc's pci_mmap of legacy_mem uses shmem_zero_setup(), substituting
a new vma->vm_file whose private_data no longer points to the bin_buffer
(substitution done because some versions of X crash if that mmap fails).

The fix to this is straightforward: the original vm_file is fput() in
that case, so this mmap won't block sysfs at all, so just don't switch
over to bin_vm_ops if vm_file has changed.

But more fixes made before realizing that was the problem:-

It should not be an error if bin_page_mkwrite() finds no underlying
page_mkwrite().

Check that a file already mmap'ed has the same underlying vm_ops
_before_ pointing vma->vm_ops at bin_vm_ops.

If the file being mmap'ed is a shmem/tmpfs file, don't fail the mmap
on CONFIG_NUMA=y, just because that has a set_policy and get_policy:
provide bin_set_policy, bin_get_policy and bin_migrate.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Acked-by: Eric Biederman <ebiederm@aristanetworks.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

diff --git a/fs/sysfs/bin.c b/fs/sysfs/bin.c
index 96cc2bf..07703d3 100644 (file)
--- a/fs/sysfs/bin.c
+++ b/fs/sysfs/bin.c
@@ -241,9 +241,12 @@ static int bin_page_mkwrite(struct vm_area_struct *vma, struct page *page)
        struct sysfs_dirent *attr_sd = file->f_path.dentry->d_fsdata;
        int ret;
 
-       if (!bb->vm_ops || !bb->vm_ops->page_mkwrite)
+       if (!bb->vm_ops)
                return -EINVAL;
 
+       if (!bb->vm_ops->page_mkwrite)
+               return 0;
+
        if (!sysfs_get_active_two(attr_sd))
                return -EINVAL;
 
@@ -273,12 +276,78 @@ static int bin_access(struct vm_area_struct *vma, unsigned long addr,
        return ret;
 }
 
+#ifdef CONFIG_NUMA
+static int bin_set_policy(struct vm_area_struct *vma, struct mempolicy *new)
+{
+       struct file *file = vma->vm_file;
+       struct bin_buffer *bb = file->private_data;
+       struct sysfs_dirent *attr_sd = file->f_path.dentry->d_fsdata;
+       int ret;
+
+       if (!bb->vm_ops || !bb->vm_ops->set_policy)
+               return 0;
+
+       if (!sysfs_get_active_two(attr_sd))
+               return -EINVAL;
+
+       ret = bb->vm_ops->set_policy(vma, new);
+
+       sysfs_put_active_two(attr_sd);
+       return ret;
+}
+
+static struct mempolicy *bin_get_policy(struct vm_area_struct *vma,
+                                       unsigned long addr)
+{
+       struct file *file = vma->vm_file;
+       struct bin_buffer *bb = file->private_data;
+       struct sysfs_dirent *attr_sd = file->f_path.dentry->d_fsdata;
+       struct mempolicy *pol;
+
+       if (!bb->vm_ops || !bb->vm_ops->get_policy)
+               return vma->vm_policy;
+
+       if (!sysfs_get_active_two(attr_sd))
+               return vma->vm_policy;
+
+       pol = bb->vm_ops->get_policy(vma, addr);
+
+       sysfs_put_active_two(attr_sd);
+       return pol;
+}
+
+static int bin_migrate(struct vm_area_struct *vma, const nodemask_t *from,
+                       const nodemask_t *to, unsigned long flags)
+{
+       struct file *file = vma->vm_file;
+       struct bin_buffer *bb = file->private_data;
+       struct sysfs_dirent *attr_sd = file->f_path.dentry->d_fsdata;
+       int ret;
+
+       if (!bb->vm_ops || !bb->vm_ops->migrate)
+               return 0;
+
+       if (!sysfs_get_active_two(attr_sd))
+               return 0;
+
+       ret = bb->vm_ops->migrate(vma, from, to, flags);
+
+       sysfs_put_active_two(attr_sd);
+       return ret;
+}
+#endif
+
 static struct vm_operations_struct bin_vm_ops = {
        .open           = bin_vma_open,
        .close          = bin_vma_close,
        .fault          = bin_fault,
        .page_mkwrite   = bin_page_mkwrite,
        .access         = bin_access,
+#ifdef CONFIG_NUMA
+       .set_policy     = bin_set_policy,
+       .get_policy     = bin_get_policy,
+       .migrate        = bin_migrate,
+#endif
 };
 
 static int mmap(struct file *file, struct vm_area_struct *vma)
@@ -287,7 +356,6 @@ static int mmap(struct file *file, struct vm_area_struct *vma)
        struct sysfs_dirent *attr_sd = file->f_path.dentry->d_fsdata;
        struct bin_attribute *attr = attr_sd->s_bin_attr.bin_attr;
        struct kobject *kobj = attr_sd->s_parent->s_dir.kobj;
-       struct vm_operations_struct *vm_ops;
        int rc;
 
        mutex_lock(&bb->mutex);
@@ -302,24 +370,25 @@ static int mmap(struct file *file, struct vm_area_struct *vma)
                goto out_put;
 
        rc = attr->mmap(kobj, attr, vma);
-       vm_ops = vma->vm_ops;
-       vma->vm_ops = &bin_vm_ops;
        if (rc)
                goto out_put;
 
-       rc = -EINVAL;
-       if (bb->mmapped && bb->vm_ops != vma->vm_ops)
+       /*
+        * PowerPC's pci_mmap of legacy_mem uses shmem_zero_setup()
+        * to satisfy versions of X which crash if the mmap fails: that
+        * substitutes a new vm_file, and we don't then want bin_vm_ops.
+        */
+       if (vma->vm_file != file)
                goto out_put;
 
-#ifdef CONFIG_NUMA
        rc = -EINVAL;
-       if (vm_ops && ((vm_ops->set_policy || vm_ops->get_policy || vm_ops->migrate)))
+       if (bb->mmapped && bb->vm_ops != vma->vm_ops)
                goto out_put;
-#endif
 
        rc = 0;
        bb->mmapped = 1;
-       bb->vm_ops = vm_ops;
+       bb->vm_ops = vma->vm_ops;
+       vma->vm_ops = &bin_vm_ops;
 out_put:
        sysfs_put_active_two(attr_sd);
 out_unlock: