On Tue, Sep 15, 2020 at 05:52:17PM +0200, Christoph Hellwig wrote:
> On Tue, Sep 15, 2020 at 05:43:38PM +0200, Heiko Carstens wrote:
> > + if (user_mode(regs)) {
> > + if (__copy_from_user(code + start - 2, (char __user *)addr, 2))
> > + break;
> > + } else {
> > + if (copy_from_kernel_nofault(code + start - 2, (char *)addr, 2))
> > + break;
> > + }
>
> > for (end = 32; end < 64; end += 2) {
> > addr = regs->psw.addr + end - 32;
> > + if (user_mode(regs)) {
> > + if (__copy_from_user(code + end, (char __user *)addr, 2))
> > + break;
> > + } else {
> > + if (copy_from_kernel_nofault(code + end, (char *)addr, 2))
> > + break;
> > + }
>
> Maybe add a little copy_from_regs helper? That would also get rid
> of the awfully long lines here.
Yes, I'll check how the result looks like, and will probably change
that.
> Also any good reason for the __copy_from_user instead of the normal
> copy_from_user here?
I don't see any reason. If I remember correctly the only difference is
zero padding, and that wouldn't hurt at all here.
Will check and change.
Thank you for taking a look!
