On 2011-07-06 at 05:33:58 +0200, Pete Zaitcev <zaitcev@xxxxxxxxxx> wrote:
> On Tue, 5 Jul 2011 13:37:50 +0200
> Tobias Klauser <tklauser@xxxxxxxxxx> wrote:
>
> > Do not bail out in mon_init() if debugfs is not available, instead just
> > print the warning but go ahead loading the binary API.
>
> > --- a/drivers/usb/mon/mon_main.c
> >
> > - if ((rc = mon_text_init()) != 0)
> > - goto err_text;
> > + if (mon_text_init() != 0)
> > + printk(KERN_WARNING TAG ": debugfs is not available\n");
> > if ((rc = mon_bin_init()) != 0)
> > @@ -739,14 +742,9 @@ int __init mon_text_init(void)
> >
> > mondir = debugfs_create_dir("usbmon", usb_debug_root);
> > - if (IS_ERR(mondir)) {
> > - printk(KERN_NOTICE TAG ": debugfs is not available\n");
> > - return -ENODEV;
> > - }
> > - if (mondir == NULL) {
> > - printk(KERN_NOTICE TAG ": unable to create usbmon directory\n");
> > + if (IS_ERR_OR_NULL(mondir))
> > return -ENODEV;
> > - }
> > +
> > mon_dir = mondir;
>
> This looks like going in the right direction, but do we even want the
> warning anymore? Perhaps we ought to retain the split: if debugfs_create_dir
> exited with -ENODEV, skip mon_text_init quietly, otherwise print a
> specific message (my original code is guilty in dumping all error
> returns in one pile).
True. In my naivity it merged together the check for the return value of
debugfs_create_dir, but kerneldoc to debugfs_create_dir says
differently:
* If an error occurs, %NULL will be returned.
*
* If debugfs is not enabled in the kernel, the value -%ENODEV will be
* returned.
I'll implement it as you suggest and send an updated patch.
Thanks
Tobias
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
