linux-kernel - RE: [PATCH 1/2] usb: dwc3: Add avoiding vbus glitch happen during xhci reset

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: <AM5PR0402MB286524243DB67DA15AE4EA42F1980@AM5PR0402MB2865.eurprd04.prod.outlook.com>
Date:   Tue, 22 Jan 2019 02:38:31 +0000
From:   Ran Wang <ran.wang_1@....com>
To:     Rob Herring <robh@...nel.org>
CC:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Mark Rutland <mark.rutland@....com>,
        Felipe Balbi <balbi@...nel.org>,
        "linux-usb@...r.kernel.org" <linux-usb@...r.kernel.org>,
        "devicetree@...r.kernel.org" <devicetree@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH 1/2] usb: dwc3: Add avoiding vbus glitch happen during
 xhci reset

Hi Rob,

On January 22, 2019 09:03, Rob Herring wrote:
> 
> On Wed, Jan 16, 2019 at 06:48:06AM +0000, Ran Wang wrote:
> > When DWC3 is set to host mode by programming register DWC3_GCTL,
> VUBS
> 
> s/VUBS/VBUS/

Yes, will fix it in next version.
 
> > (or its control signal) will on immediately on related Root Hub ports.
> 
> /will on/will turn on/

Got it. Thanks.

> > Then the VUBS will be de-asserted for a little while during xhci reset
> > (conducted by xhci driver) for a little while and back to normal.
> >
> > This VBUS glitch might cause some USB devices emuration fail if kernel
> > boot with them connected. One SW workaround which can fix this is to
> > program all PORTSC[PP] to 0 to turn off VBUS immediately after setting
> > host mode in DWC3 driver(per signal measurement result, it will be too
> > late to do it in xhci-plat.c or xhci.c).
> >
> > Signed-off-by: Ran Wang <ran.wang_1@....com>
> > ---
> >  Documentation/devicetree/bindings/usb/dwc3.txt |    3 +++
> >  1 files changed, 3 insertions(+), 0 deletions(-)
> >
> > diff --git a/Documentation/devicetree/bindings/usb/dwc3.txt
> > b/Documentation/devicetree/bindings/usb/dwc3.txt
> > index 8e5265e..dadb530 100644
> > --- a/Documentation/devicetree/bindings/usb/dwc3.txt
> > +++ b/Documentation/devicetree/bindings/usb/dwc3.txt
> > @@ -106,6 +106,9 @@ Optional properties:
> >  			When just one value, which means INCRX burst
> mode enabled. When
> >  			more than one value, which means undefined length
> INCR burst type
> >  			enabled. The values can be 1, 4, 8, 16, 32, 64, 128
> and 256.
> > + - snps,avoid-vbus-glitch-when-set-host: Power off all Root Hub ports
> immediately
> > +			after setting host mode to avoid vbus (negative)
> glitch happen in later
> > +			xhci reset. And the vbus will back to 5V automatically
> when reset done.
> 
> Can't you imply this from the compatible string. You should have an SoC
> specific compatible.

Sorry, not quite get your point here? Actually I have discussed with Soc design guys
and Felipe, it seems to be DWC3 native behavior rather than SoC specific issue.
https://lkml.org/lkml/2018/11/23/387
https://lkml.org/lkml/2018/12/12/140

I think there could be 2 reasons why we got no report for a long time till now:
1. Most USB devices are not so sensitive to this VBUS glitch, they would works fine.
2. A proper VBUS pump circuit design can successfully filter this glitch out. I have
confirmed this with scope, which means some platforms might resolve issue in
this way already (some might not, of cause).

> Does this even need to be conditional? What would be the harm in doing
> this unconditionally for all DWC3 hosts?

>From the logic level, I don't see a harm to DWC3 hosts if unconditionally apply this.
However, since this is not a only solution (can fix it in HW way), I prefer to let it
controlled by property in DTS node for those questionable board already existing
in market. 

Regards,
Ran