Re: [PATCH] Btrfs: add a extent ref verify tool V2

[Josef Bacik <jbacik@fb.com>]

On 05/08/2014 07:34 PM, Zach Brown wrote:
>> +#ifdef CONFIG_BTRFS_FS_REF_VERIFY
>> +int btrfs_build_ref_tree(struct btrfs_fs_info *fs_info);
>> +void btrfs_free_ref_cache(struct btrfs_fs_info *fs_info);
>> +int btrfs_ref_tree_mod(struct btrfs_root *root, u64 bytenr, u64 num_bytes,
>> +		       u64 parent, u64 ref_root, u64 owner, u64 offset,
>> +		       int action);
>> +void btrfs_free_ref_tree_range(struct btrfs_fs_info *fs_info, u64 start,
>> +			       u64 len);
>> +#else
>> +
>> +#define btrfs_free_ref_cache(fs_info) do { } while (0)
>> +#define btrfs_ref_tree_mod(root, bytenr, num_bytes, parent, ref_root,	\
>> +			   owner, offset, action) do { } while (0)
>> +#define btrfs_free_ref_tree_range(fs_info, start, len) do { } while (0)
>> +
>> +#endif /* CONFIG_BTRFS_FS_REF_VERIFY */
>> +#endif /* _REF_VERIFY__ */
>
> Don't just omit the arguments when the config isn't enabled.  That can
> let these calls bit rot over time as everyone build tests their changes
> with the config disabled, as they will.
>
> 	int derp;
>
> 	...
>
> 	btrfs_free_ref_cache(derp);
>
> 	...
>
> -	derp++;
>
>
> Will have gcc warn that derp is unused.  They'll delete it and send out
> the patch.  You'll later turn on the config and get undefined derp
> warnings.
>
> inline stubs are the best way out.  They'll still warn if the arg is
> used undefined.  Putting (void)arg; in the macros doesn't warn in that
> case.

Gcc doesn't complain when I build with it off but I can switch to inline 
stubs if that makes you feel better.  Thanks,

Josef