On 6/12/07, Stelian Pop <[EMAIL PROTECTED]> wrote:
Le lundi 11 juin 2007 à 16:26 -0500, Nelson Castillo a écrit :
> Hi.
>
> I just added support for user space buffers in kfifo. I found useful
> __kfifo_get_user to copy data to a user buffer in a read call. I didn't
> like the idea of having an extra buffer.
>
> * Is it ok to add this support?
I suppose it is, however:
> + spin_lock_irqsave(fifo->lock, flags);
> +
> + ret = __kfifo_put_user(fifo, buffer, len);
> +
> + spin_unlock_irqrestore(fifo->lock, flags);
[...]
> +int __kfifo_put_user(struct kfifo *fifo, const unsigned char __user *buffer,
> + unsigned int len)
[...]
> + if(copy_from_user(fifo->buffer + (fifo->in & (fifo->size - 1)),
> + buffer, l))
> + return -EFAULT;
[...]
accessing userspace memory with a spinlock taken (moreover an irqsave()
one) is bad bad bad.
Yes.
Perhaps then only the functions __kfifo_put_user and __kfifo_get_user should
be allowed. Those are the ones I'm using now , anyway.
I added the others for completeness but adding them is probably a mistake.
Regards,
Nelson.-
--
http://arhuaco.org
http://emQbit.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
