linux-kernel - Re: ELAN Touchscreen regression in recent 3.12 rc's? (USB)

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: <20131106102838.347b005c@voldemort.scrye.com>
Date:	Wed, 6 Nov 2013 10:28:38 -0700
From:	Kevin Fenzi <kevin@...ye.com>
To:	AceLan Kao <acelan.kao@...onical.com>
Cc:	Jiri Kosina <jkosina@...e.cz>,
	Nikolai Kondrashov <spbnick@...il.com>,
	Josh Boyer <jwboyer@...oraproject.org>,
	"Linux-Kernel@...r. Kernel. Org" <linux-kernel@...r.kernel.org>
Subject: Re: ELAN Touchscreen regression in recent 3.12 rc's? (USB)

On Wed, 6 Nov 2013 09:44:11 +0800
AceLan Kao <acelan.kao@...onical.com> wrote:

> Hi,
> 
> Sorry, no thoughts about this.
> If Kevin would like to try, I can bisect the kernel and build the
> Ubuntu .deb package for him.
> But like I say above, those 4 commits doesn't look like to lead to the
> problem, so I'll try to bisect the whole kernel.
> 
> % git bisect start v3.12-rc7 v3.12-rc6
> Bisecting: 165 revisions left to test after this (roughly 7 steps)
> [edd31476011052d8f6591a3194ba0716b0cea681] bnx2x: Set NETIF_F_HIGHDMA
> unconditionally

Yeah, I can try and do so this weekend. To busy during the week to get
very far on this. 

Also, it seems like on some boots it just never works, but on others It
does finally work: 

[ 3238.722796] usb 2-7: new full-speed USB device number 127 using xhci_hcd
[ 3238.735177] usb 2-7: New USB device found, idVendor=04f3, idProduct=016f
[ 3238.735183] usb 2-7: New USB device strings: Mfr=4, Product=14, SerialNumber=0
[ 3238.735185] usb 2-7: Product: Touchscreen
[ 3238.735187] usb 2-7: Manufacturer: ELAN
[ 3238.735539] usb 2-7: ep 0x2 - rounding interval to 64 microframes, ep desc says 80 microframes
[ 3238.743682] input: ELAN Touchscreen as /devices/pci0000:00/0000:00:14.0/usb2/2-7/2-7:1.0/input/input368
[ 3238.744505] hid-multitouch 0003:04F3:016F.00B4:
input,hiddev0,hidraw1: USB HID v1.10 Device [ELAN Touchscreen] on
usb-0000:00:14.0-7/input0

so, 3238 seconds after boot, with 127 tries it started working on this boot. 

I'll try and isolate it this weekend. Thanks for the attention everyone. 

kevin
--
> 
> Best regards,
> AceLan Kao.
> 
> 2013/11/5 Jiri Kosina <jkosina@...e.cz>:
> > On Wed, 30 Oct 2013, Nikolai Kondrashov wrote:
> >
> >> > It's hard to believe that those quirks will lead to the problem.
> >> > And actually, there are 4 new commits introduced to -rc7, and 3
> >> > of them are quirks.
> >> >
> >> > % git log --pretty=oneline v3.12-rc6..v3.12-rc7 drivers/hid
> >> > 86b84167d4e67372376a57ea9955c5d53dae232f HID: wiimote: add
> >> > LEGO-wiimote VID ad0e669b922c7790182cf19f8015b30e23ad9499 HID:
> >> > Fix unit exponent parsing again
> >> > 684524d35fe8d13be1f2649633e43bd02c96c695 HID: usbhid: quirk for
> >> > SiS Touchscreen
> >> > 8171a67d587a09e14a4949a81e070345fedcf410 HID: usbhid: quirk for
> >> > Synaptics Large Touchccreen
> >> >
> >> > CC'd Nikolai, since his commit changes the protocol.
> >>
> >> My change is very unlikely to produce such problems. It changes
> >> calculation of axes resolution at the time the report descriptor
> >> is processed. The results of the calculation are not used by the
> >> kernel, AFAIK, but only by userspace drivers. The results should
> >> not be used to affect the interactions with the device, but only
> >> the interpretation of the reports (again, in userspace), and even
> >> that is barely done currently.
> >>
> >> Please try reverting that particular commit and see if it affects
> >> the behavior.
> >
> > AceLan,
> >
> > do you have any update, please?
> >
> > --
> > Jiri Kosina
> > SUSE Labs
> 


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