linux-kernel - RE: [PATCH 4/5 V2] usb: call pm_runtime_put_sync in pm_runtime_get

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <27240C0AC20F114CBF8149A2696CBE4A24257F@SHSMSX101.ccr.corp.intel.com>
Date:	Fri, 1 Mar 2013 00:59:23 +0000
From:	"Liu, Chuansheng" <chuansheng.liu@...el.com>
To:	"Rafael J. Wysocki" <rjw@...k.pl>
CC:	Alan Stern <stern@...land.harvard.edu>,
	"Li, Fei" <fei.li@...el.com>,
	"gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>,
	"Lan, Tianyu" <tianyu.lan@...el.com>,
	"sarah.a.sharp@...ux.intel.com" <sarah.a.sharp@...ux.intel.com>,
	"linux-usb@...r.kernel.org" <linux-usb@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH 4/5 V2] usb: call pm_runtime_put_sync in
 pm_runtime_get_sync failed case



> -----Original Message-----
> From: Rafael J. Wysocki [mailto:rjw@...k.pl]
> Sent: Friday, March 01, 2013 8:51 AM
> To: Liu, Chuansheng
> Cc: Alan Stern; Li, Fei; gregkh@...uxfoundation.org; Lan, Tianyu;
> sarah.a.sharp@...ux.intel.com; linux-usb@...r.kernel.org;
> linux-kernel@...r.kernel.org
> Subject: Re: [PATCH 4/5 V2] usb: call pm_runtime_put_sync in
> pm_runtime_get_sync failed case
> 
> On Friday, March 01, 2013 12:38:07 AM Liu, Chuansheng wrote:
> >
> > > -----Original Message-----
> > > From: Alan Stern [mailto:stern@...land.harvard.edu]
> > > Sent: Thursday, February 28, 2013 11:17 PM
> > > To: Li, Fei
> > > Cc: gregkh@...uxfoundation.org; Lan, Tianyu;
> sarah.a.sharp@...ux.intel.com;
> > > rjw@...k.pl; linux-usb@...r.kernel.org; linux-kernel@...r.kernel.org; Liu,
> > > Chuansheng
> > > Subject: Re: [PATCH 4/5 V2] usb: call pm_runtime_put_sync in
> > > pm_runtime_get_sync failed case
> > >
> > > On Thu, 28 Feb 2013, Li Fei wrote:
> > >
> > > >
> > > > Even in failed case of pm_runtime_get_sync, the usage_count
> > > > is incremented. In order to keep the usage_count with correct
> > > > value and runtime power management to behave correctly, call
> > > > pm_runtime_put(_sync) in such case.
> > > >
> > > > Signed-off-by Liu Chuansheng <chuansheng.liu@...el.com>
> > > > Signed-off-by: Li Fei <fei.li@...el.com>
> > > > ---
> > > >  drivers/usb/core/hub.c |    3 ++-
> > > >  1 files changed, 2 insertions(+), 1 deletions(-)
> > > >
> > > > diff --git a/drivers/usb/core/hub.c b/drivers/usb/core/hub.c
> > > > index 5480352..f72dede 100644
> > > > --- a/drivers/usb/core/hub.c
> > > > +++ b/drivers/usb/core/hub.c
> > > > @@ -3148,12 +3148,13 @@ int usb_port_resume(struct usb_device
> *udev,
> > > pm_message_t msg)
> > > >
> > > >  	if (port_dev->did_runtime_put) {
> > > >  		status = pm_runtime_get_sync(&port_dev->dev);
> > > > -		port_dev->did_runtime_put = false;
> > > >  		if (status < 0) {
> > > >  			dev_dbg(&udev->dev, "can't resume usb port,
> status %d\n",
> > > >  					status);
> > > > +			pm_runtime_put_sync(&port_dev->dev);
> > > >  			return status;
> > > >  		}
> > > > +		port_dev->did_runtime_put = false;
> > > >  	}
> > >
> > > I don't see much point in this.  After a failed resume, the port's
> > > runtime PM status is undefined.  Whether or not you do a
> > > pm_runtime_put_sync won't make any difference.
> > In case of failed resume, calling pm_runtime_put_sync() is just for decrease
> the dev->power.usage_count,
> > because pm_runtime_get_sync() always increase the
> dev->power.usage_count even failed.
> >
> > If not pairing runtime_get/put, after that case, the device can not enter
> runtime suspend any more due to dev->power.usage_count > 0 always.
> > Is it making sense?
> 
> Well, not really.
> 
> Before returning an error code, rpm_callback() assigns that code to
> dev->power.runtime_error and that will effectively disable runtime PM for dev
> going forward anyway.
Thanks your pointing out.
dev->power.runtime_error!=0 will really block the runtime PM resume/suspend to continue.

But in case of rpm_resume return error when dev->power.disable_depth > 0, the dev->power.runtime_error
is not set yet. Is it the case? Thanks your comments again.

And another case is when user called pm_runtime_set_status to clear the runtime_error after dev->power.runtime_error
is set during pm_runtime_get_sync(), the runtime_resume/suspend() can be tried again? But the dev->power.usage_count is still wrong?
Thanks your correction again:)
> 
> Thanks,
> Rafael
> 
> 
> --
> I speak only for myself.
> Rafael J. Wysocki, Intel Open Source Technology Center.