Re: [PATCH 1/1] security: introduce fs caps

["Serge E. Hallyn" <serue@us.ibm.com>]

Quoting Alexey Dobriyan (adobriyan@gmail.com):
> On Mon, Nov 13, 2006 at 10:43:26AM -0600, Serge E. Hallyn wrote:
> > Quoting Alexey Dobriyan (adobriyan@gmail.com):
> > > On Wed, Nov 08, 2006 at 05:52:03PM -0600, Serge E. Hallyn wrote:
> > > > > > +	__u32 version;
> > > > > > +	__u32 effective;
> > > > > > +	__u32 permitted;
> > > > > > +	__u32 inheritable;
> > > > > > +};
> > > > > > +
> > > > > > +static inline void convert_to_le(struct vfs_cap_data_struct *cap)
> > > > > > +{
> > > > > > +	cap->version = le32_to_cpu(cap->version);
> > > > > > +	cap->effective = le32_to_cpu(cap->effective);
> > > > > > +	cap->permitted = le32_to_cpu(cap->permitted);
> > > > > > +	cap->inheritable = le32_to_cpu(cap->inheritable);
> > > > > > +}
> > > > >
> > > > > This pretty much defeats sparse endian checking. You will get warnings
> > > > > regardless of whether u32 or le32 are used.
> > > >
> > > > But I don't get any sparse warnings with make C=1.  Am I doing something
> > > > wrong?
> > >
> > > You need (temporarily) to use CF:
> > >
> > > 	make C=1 CF=-D__CHECK_ENDIAN__ ...
> >
> > The following patch on top of the existing one eliminates the warning.
> > Does it appear consistent with what you were suggesting?
> 
> > If it is ok, I'll resend the full patch.
> 
> Yes, that's it, modulo:

Thanks, comments integrated with the exception of:

> > --- a/include/linux/capability.h
> > +++ b/include/linux/capability.h
> > +#ifdef CONFIG_SECURITY_FS_CAPABILITIES
> 
> This is exportable header, so no CONFIG_*
> 
> > +#define XATTR_CAPS_SUFFIX "capability"
> > +#define XATTR_NAME_CAPS XATTR_SECURITY_PREFIX XATTR_CAPS_SUFFIX
> > +struct vfs_cap_data_disk {
> > +	__le32 version;
> > +	__le32 effective;
> > +	__le32 permitted;
> > +	__le32 inheritable;
> > +};
> > +#endif
> >
> >  #ifdef __KERNEL__
> >
> >  #include <linux/spinlock.h>
> >  #include <asm/current.h>
> >
> > +#ifdef CONFIG_SECURITY_FS_CAPABILITIES
> > +struct vfs_cap_data {
> > +	__u32 version;
> > +	__u32 effective;
> > +	__u32 permitted;
> > +	__u32 inheritable;
> > +};
> 
> Now you're in kernel, so you can happily use u32 without undescores.

The rest of the file already uses __u32, so for consistency I'd
rather stick with __u32, unless there's a reason why it's really
discouraged.

Will send new patch in just a bit.

thanks,
-serge

> > --- a/security/commoncap.c
> > +++ b/security/commoncap.c
> > @@ -155,7 +147,8 @@ static int set_file_caps(struct linux_bi
> >  {
> >  	struct dentry *dentry;
> >  	ssize_t rc;
> 
> > @@ -178,19 +171,19 @@ static int set_file_caps(struct linux_bi
> >  		return rc;
> >  	}
> >
> > -	if (rc != sizeof(cap_struct)) {
> > +	if (rc != sizeof(dcaps)) {
> >  		printk(KERN_NOTICE "%s: got wrong size for getxattr (%d)\n",
> >  					__FUNCTION__, rc);
> 
> rc is ssize_t, so "%zd".