linux-kernel - v3.9-rc1 instability on Chromebook Pixel with gmbus irq

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 for Android: free password hash cracker in your pocket

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

Date:	Tue, 5 Mar 2013 18:35:07 -0800
From:	Benson Leung <bleung@...omium.org>
To:	daniel.vetter@...ll.ch
Cc:	Daniel Kurtz <djkurtz@...omium.org>,
	Yufeng Shen <miletus@...omium.org>,
	Olof Johansson <olofj@...omium.org>,
	dri-devel@...ts.freedesktop.org,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: v3.9-rc1 instability on Chromebook Pixel with gmbus irq

Hi Daniel Vetter,

I'm working on touch devices Chromium OS, and I've noticed a
regression between 3.8 and 3.9-rc1, which was posted yesterday.

The hardware in question is a Chromebook Pixel. For this device, we
have i2c input devices: atmel mxt224s touchpad and atmel mxt1664s
touchscreen. The touchpad is on bus 1, "i915 gmbus vga" at 1-004b. The
touchscreen is on bus 2, "i915 gmbus panel" at 2-004a.

I was testing v3.9-rc1 on the Pixel and the touchscreen driver is
being returned -110 (-ETIMEDOUT) on an i2c_transfer after several
seconds of both touch devices working correctly. At the time of the
failure, there are no error messages from GMBUS that I can see, but
the bus never recovers. I can keep interacting with the touchscreen or
touchpad, and the interrupts trigger reads, but all subsequent reads
return -110.

I noticed that between 3.8 and 3.9-rc1, your patch series to add gmbus
irq support was merged. After bisecting, I found that this commit
seems to be causing the timeout problem. Reverting it makes the
problem go away, and the bus is stable.

commit 2c438c0273b76d6cb158f8bdd0aa3ebf66e48a28
Author: Daniel Vetter <daniel.vetter@...ll.ch>
Date:   Sat Dec 1 13:53:46 2012 +0100

    drm/i915: use gmbus irq to wait for gmbus idle

    GMBUS_ACTIVE has inverted sense and so doesn't fit into the
    wait_hw_status helper, hence create a new gmbus_wait_idle functions.
    Also, we only care about the idle irq event and nothing else, which
    allows us to use the wait_event_timeout helper directly without
    jumping through hoops to catch NAKs.

    Since gen2/3 don't have gmbus interrupts, handle them separately with
    the old wait_for macro.

    This shaves another few ms off reading EDID from a hdmi screen on my
    testbox here. EDID reading with interrupt driven gmbus is now as fast
    as with busy-looping gmbus at 28 ms here (with negligible cpu
    overhead).

    Reviewed-by: Imre Deak <imre.deak@...el.com>
    Signed-off-by: Daniel Vetter <daniel.vetter@...ll.ch>

Is there anything I can do to help debug this some more?

--
Benson Leung
Software Engineer, Chrom* OS
bleung@...omium.org
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/