linux-kernel - Re: [PATCH] usb: dwc2: drd: fix clock gating on USB role switch

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] [day] [month] [year] [list]

Message-ID: <20240904053052.GA8@f902c60e4871>
Date: Wed, 4 Sep 2024 07:30:52 +0200
From: Tomas Marek <tomas.marek@...est.cz>
To: Sergey Shtylyov <s.shtylyov@....ru>
Cc: hminas@...opsys.com, gregkh@...uxfoundation.org,
	Arthur.Petrosyan@...opsys.com, linux-usb@...r.kernel.org,
	linux-kernel@...r.kernel.org, oleg.karfich@...o.com
Subject: Re: [PATCH] usb: dwc2: drd: fix clock gating on USB role switch

Hi Sergey,

thank you for your review.

On Tue, Sep 03, 2024 at 04:20:26PM +0300, Sergey Shtylyov wrote:
> On 9/3/24 12:41 PM, Tomas Marek wrote:
> 
> > The dwc2_handle_usb_suspend_intr() function disables gadget clocks in USB
> > peripheral mode when no other power-down mode is available (introduced by
> > commit 0112b7ce68ea ("usb: dwc2: Update dwc2_handle_usb_suspend_intr function.")).
> > However, the dwc2_drd_role_sw_set() USB role update handler attempts to
> > read DWC2 registers if the USB role has changed while the USB is in suspend
> > mode (when the clocks are gated). This causes the system to hang.
> > 
> > Release gadget clocks before handling the USB role update.
> > 
> > Fixes: 0112b7ce68ea ("usb: dwc2: Update dwc2_handle_usb_suspend_intr function.")
> > 
> > Signed-off-by: Tomas Marek <tomas.marek@...est.cz>
> > ---
> >  drivers/usb/dwc2/drd.c | 12 ++++++++++++
> >  1 file changed, 12 insertions(+)
> > 
> > diff --git a/drivers/usb/dwc2/drd.c b/drivers/usb/dwc2/drd.c
> > index a8605b02115b..ccb33cd1f04b 100644
> > --- a/drivers/usb/dwc2/drd.c
> > +++ b/drivers/usb/dwc2/drd.c
> > @@ -127,6 +127,18 @@ static int dwc2_drd_role_sw_set(struct usb_role_switch *sw, enum usb_role role)
> >  			role = USB_ROLE_DEVICE;
> >  	}
> >  
> > +#if IS_ENABLED(CONFIG_USB_DWC2_PERIPHERAL) || \
> > +	IS_ENABLED(CONFIG_USB_DWC2_DUAL_ROLE)
> > +	if (dwc2_is_device_mode(hsotg)) {
> 
>    Why not avoid #ifdef by doing e.g.:
> 
> 	if ((IS_ENABLED(CONFIG_USB_DWC2_PERIPHERAL) || \
> 	     IS_ENABLED(CONFIG_USB_DWC2_DUAL_ROLE) &&
> 	     dwc2_is_device_mode(hsotg)) {
> 
> > +		if (hsotg->lx_state == DWC2_L2) {
> > +			if (hsotg->params.power_down ==
> > +			    DWC2_POWER_DOWN_PARAM_NONE && hsotg->bus_suspended &&
> > +			    !hsotg->params.no_clock_gating)
> 
>    Well, these *if*s are mergeable too... :-)
> 

OK, why not. In the end all conditions could be merged into one if statement like:

	if ((IS_ENABLED(CONFIG_USB_DWC2_PERIPHERAL) ||
	     IS_ENABLED(CONFIG_USB_DWC2_DUAL_ROLE)) &&
	     dwc2_is_device_mode(hsotg) &&
	     hsotg->lx_state == DWC2_L2 &&
	     hsotg->params.power_down == DWC2_POWER_DOWN_PARAM_NONE &&
	     hsotg->bus_suspended &&
	     !hsotg->params.no_clock_gating)
		dwc2_gadget_exit_clock_gating(hsotg, 0);

I’ll submit a new patch version with this modification.

Best regards

Tomas

> > +				dwc2_gadget_exit_clock_gating(hsotg, 0);
> > +		}
> > +	}
> > +#endif
> > +
> [...]
> 
> MBR, Sergey