Re: [PATCH 1/4] uaccess: add copy_word_from_user

[Steven Rostedt]

On Wed, 25 Feb 2009, Andrew Morton wrote:

> btw,
> 
> On Wed, 25 Feb 2009 15:30:08 -0500
> Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
> 
> > +int copy_word_from_user(void *to, const void __user *from,
> > +			unsigned int copy, unsigned int read,
> > +			unsigned int *copied, int skip)
> 
> You presently have this not-exported-to-modules.  Was that deliberate?

Nope, I didn't even think about it. I usually do the EXPORT_SYMBOL when
I have a need to export to a module. Not saying I left it out on purpose.
I left it out simply because I did not think about using it in a module 
:-/

> 
> There are arguments either way.  Lately we've tended to take the
> position that a whole interface either is or is not wholly exported. 
> The uaccess functions are exported, so this one should be as well.
> 
> We can of course do that later on, when there's a user - I have no
> particular preference personally.
> 
> <looks at probe_kernel_read and probe_kernel_write>
> 
> These are really part of the uaccess interface too.  I don't see a need
> for both lib/uaccess.c and mm/maccess.c?

Ah, I didn't notice that file. I could move this to that file, or move the 
maccess to this file. I just thought lib would be a better location.
What would you suggest?

> 
> probe_kernel_read() and probe_kernel_write() are EXPORT_SYMBOL_GPL,
> whereas the rest of the uaccess interface is EXPORT_SYMBOL.  Ho hum.

Looks like another clean up. I guess probe_kernel is because it is more 
kernel internals and should not be used by binary modules? Although 
nothing prevents a kernel module from using the uaccess code for the 
kernel. Or even make a fault section themselves.

-- Steve

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/