Re: [PATCH 3.12-rc7] KVM: Fix modprobe failure for kvm_intel/kvm_amd

[Paolo Bonzini <pbonzini@redhat.com>]

Il 30/10/2013 17:23, Greg KH ha scritto:
>> > static inline struct dentry *debugfs_create_dir(const char *name,
>> >                                                 struct dentry *parent)
>> > {
>> >         return ERR_PTR(-ENODEV);
>> > }
>> > 
>> > which would oops a lot of the current callers.
> It will oops?  Really?  Where?  That shouldn't happen at all.

Doh, it obviously won't because the bogus pointer value will never be
dereferenced by the dummy debugfs_create_file.

>> > Very few places use the currently correct idiom
>> > 
>> > 	if (IS_ERR(root) || !root)
>> > 
>> > but it's very ugly...  Perhaps debugfs_create_dir *should* return an
>> > error-valued pointer after all.
> Or just don't care about the return value, and all will work out just
> fine, right?

Debugfs files would then be created in the debugfs root, which is not nice.

        /* If the parent is not specified, we create it in the root.
         * We need the root dentry to do this, which is in the super 
         * block. A pointer to that is in the struct vfsmount that we
         * have around.
         */
        if (!parent)
                parent = debugfs_mount->mnt_root;

This is all documented right:

/**
 * debugfs_create_file - create a file in the debugfs filesystem
 * ...
 * This function will return a pointer to a dentry if it succeeds.  This
 * pointer must be passed to the debugfs_remove() function when the file is
 * to be removed (no automatic cleanup happens if your module is unloaded,
 * you are responsible here.)  If an error occurs, %NULL will be returned.
 *
 * If debugfs is not enabled in the kernel, the value -%ENODEV will be
 * returned.
 */

so I guess it's just part of the API.

Paolo