On Wed, 6 May 2015, yoma sophian wrote:
> hi alan
>
> 2015-05-06 2:32 GMT+08:00 Alan Stern <st...@rowland.harvard.edu>:
> > On Wed, 6 May 2015, yoma sophian wrote:
> >
> >> hi all:
> >> below are call back functions of usb_storage_driver:
> >> static struct usb_driver usb_storage_driver = {
> >> .name = "usb-storage",
> >> .probe = storage_probe,
> >> .disconnect = usb_stor_disconnect,
> >> .suspend = usb_stor_suspend,
> >> .resume = usb_stor_resume,
> >> .reset_resume = usb_stor_reset_resume,
> >> .pre_reset = usb_stor_pre_reset,
> >> .post_reset = usb_stor_post_reset,
> >> .id_table = usb_storage_usb_ids,
> >> .supports_autosuspend = 1,
> >> .soft_unbind = 1,
> >> };
> >>
> >> my questions are:
> >> 1. where we register the us->suspend_resume_hook in usb_stor_suspend?
> >> I grep the src, but cannot find where the callback is registered.
> >
> > The callback can be registered between the calls to usb_stor_probe1()
> > and usb_stor_probe2().
> I neither find suspend_resume_hook registered in probe1 or probe2.
> in probe1, we do following initialization:
> mutex_init(&(us->dev_mutex));
> us_set_lock_class(&us->dev_mutex, intf);
> init_completion(&us->cmnd_ready);
> init_completion(&(us->notify));
> init_waitqueue_head(&us->delay_wait);
> get_transport(us);
> get_protocol(us);
>
> in probe2, we acquire us_data related resources and register host to
> scsi layer.
> result = usb_stor_acquire_resources(us);
> result = scsi_add_host(us_to_host(us), dev);
>
> it seems we didn't register suspend_resume_hook.
Like I said above, you register the hook _between_ usb_stor_probe1()
and usb_stor_probe2(). Not _inside_ either one.
For example:
result = usb_stor_probe1(&us, intf, id, devlist);
if (result)
return result;
us->suspend_resume_hook = hook_fn;
result = urb_stor_probe2(us);
> >
> >> 2. suppose we are doing the scsi read/write command and system
> >> suddenly store to RAM/DISK, will that scsi command do?
> >> Drop it or keep the unfinished part ?
> >
> > It will wait until the SCSI command is finished and then invoke your
> > callback routine.
> so you mean if we are doing CBW, in/out date and CSW, once hibernate
> happen during the end of CBW, kernel will wait CSW finished then go to
> hibernate?
Yes.
> in the resume case, we will port reset device and re-enumerate storage device.
> if so, shouldn't we free previous resource allocated before hibernate?
What resource are you talking about?
Alan Stern
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
