Messages in this thread | | | Date | Tue, 11 Nov 2008 15:25:27 -0700 | From | Jonathan Corbet <> | Subject | Re: [PATCH 3/4] add ksm kernel shared memory driver |
| |
On Wed, 12 Nov 2008 00:17:39 +0200 Izik Eidus <ieidus@redhat.com> wrote:
> >> +static int ksm_dev_open(struct inode *inode, struct file *filp) > >> +{ > >> + try_module_get(THIS_MODULE); > >> + return 0; > >> +} > >> + > >> +static int ksm_dev_release(struct inode *inode, struct file *filp) > >> +{ > >> + module_put(THIS_MODULE); > >> + return 0; > >> +} > >> + > >> +static struct file_operations ksm_chardev_ops = { > >> + .open = ksm_dev_open, > >> + .release = ksm_dev_release, > >> + .unlocked_ioctl = ksm_dev_ioctl, > >> + .compat_ioctl = ksm_dev_ioctl, > >> +}; > >> > > > > Why do you roll your own module reference counting? Is there a > > reason you don't just set .owner and let the VFS handle it? > > > > Yes, I am taking get_task_mm() if the module will be removed before i > free the mms, things will go wrong
But...if you set .owner, the VFS will do the try_module_get() *before* calling into your module (as an added bonus, it will actually check the return value too). All you've succeeded in doing here is adding a microscopic race to the module reference counting; otherwise the end result is the same.
jon
| |