linux-kernel - Re: [PATCH] usb: host: xhci-plat: Enable XHCI_SPURIOUS

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]

Date:	Wed, 26 Jun 2013 15:25:55 +0300
From:	Felipe Balbi <balbi@...com>
To:	George Cherian <george.cherian@...com>
CC:	<balbi@...com>, <linux-usb@...r.kernel.org>,
	<linux-kernel@...r.kernel.org>, <gregkh@...uxfoundation.org>,
	<sarah.a.sharp@...ux.intel.com>
Subject: Re: [PATCH] usb: host: xhci-plat: Enable XHCI_SPURIOUS_SUCCESS quirk
 for xhci-plat

Hi,

On Wed, Jun 26, 2013 at 05:46:29PM +0530, George Cherian wrote:
> >>@@ -25,6 +25,16 @@ static void xhci_plat_quirks(struct device *dev, struct xhci_hcd *xhci)
> >>  	 * dev struct in order to setup MSI
> >>  	 */
> >>  	xhci->quirks |= XHCI_BROKEN_MSI;
> >>+
> >>+	/*
> >>+	 * In some xhci controllers which follows xhci 1.0 spec gives a spurious
> >>+	 * success event after a short transfer. This quirk will ignore such
> >>+	 * spurious event. Hit this issue in synopsis xhci controllers with
> >>+	 * hci_version > 0.96
> >>+	 */
> >>+
> >>+	if (xhci->hci_version > 0x96)
> >>+		xhci->quirks |= XHCI_SPURIOUS_SUCCESS;
> >>  }
> >doesn't look like the correct way to do this. What if enabling that
> >quirk on hosts which don't have the quirk cause problems ?
> For a controller which does not have this issue will never get a
> spurious success for short packet ( and that too for only ISOCH).
> 
> per this commit ad808333d Intel xhci: Ignore spurious successful event.
> 
> This spurious successful event behavior isn't technically disallowed by
> the xHCI specification, so make the xHCI driver just ignore the spurious
> completion event.

still we don't want to enable quirks for devices which aren't quirky :-)

Now how Sarah correctly enables the quirk flag only for the known "bad"
Panther Point device.

> >I would suggest adding a platform_data which (in our case) dwc3 will
> >pass to xhci-plat. Then you can do proper revision detection of the
> >synopsys controller and set the quirk only on the failing hosts.
> >
> >BTW, do you have the STARS number for this errata ?
> No STARS number yet.

once we have it, let's make sure to follow what we do on the dwc3 and
list it in a comment. You already that you found the bug with a Synopsys
controller anyway, might as well point to the fact that there is a
ticket with synopsys to track this issue.

cheers

-- 
balbi

Download attachment "signature.asc" of type "application/pgp-signature" (837 bytes)