Messages in this thread | | | Date | Wed, 2 Jun 2010 09:47:46 -0400 (EDT) | From | Alan Stern <> | Subject | Re: [PATCH 4/6] USB-BKL: Remove BKL use in uhci-debug |
| |
On Tue, 1 Jun 2010, Arnd Bergmann wrote:
> From: Andi Kleen <ak@linux.intel.com> > > BKL was not really needed, just came from earlier push downs.
Yes.
> The only part that's a bit dodgy is the lseek function. Would > need another lock or atomic access to fpos on 32bit? > Better to have a libfs lseek
It doesn't matter. Anyone who tries to do lseeks on this file from two different threads, simultaneously, deserves what they get.
> @@ -539,11 +534,11 @@ static loff_t uhci_debug_lseek(struct file *file, loff_t off, int whence) > new = file->f_pos + off; > break; > } > + > + /* XXX: Can size shrink? */ > if (new < 0 || new > up->size) { > - unlock_kernel(); > return -EINVAL; > } > - unlock_kernel(); > return (file->f_pos = new); > }
This comment isn't needed; the size cannot change after the file has been opened.
Alan Stern
| |