Switch branches/tags
Nothing to show
Commits on Jan 7, 2012
  1. vfs: fix up ENOIOCTLCMD error handling

    torvalds authored and jdkoreclipse committed Jan 5, 2012
    We're doing some odd things there, which already messes up various users
    (see the net/socket.c code that this removes), and it was going to add
    yet more crud to the block layer because of the incorrect error code
    ENOIOCTLCMD is not an error return that should be returned to user mode
    from the "ioctl()" system call, but it should *not* be translated as
    EINVAL ("Invalid argument").  It should be translated as ENOTTY
    ("Inappropriate ioctl for device").
    That EINVAL confusion has apparently so permeated some code that the
    block layer actually checks for it, which is sad.  We continue to do so
    for now, but add a big comment about how wrong that is, and we should
    remove it entirely eventually.  In the meantime, this tries to keep the
    changes localized to just the EINVAL -> ENOTTY fix, and removing code
    that makes it harder to do the right thing.
    Signed-off-by: Linus Torvalds <>
  2. update defconfig

    jdkoreclipse committed Jan 7, 2012
  3. Add jRCU by Joe Korty

    Shankar authored and jdkoreclipse committed Dec 29, 2011
  4. Bug fixes for the user undervolt code. Replaced a lot of ugly code

    imoseyon authored and jdkoreclipse committed Jan 4, 2012
    with more ugly code. HAHA
    There was a rather crazy bug where at boot it could calibrate the
    350 and lower freqs to super high voltages. It seems SmartReflex
    is cable of pushing a lot of juice at any frequency and you don't
    want to force calibration when that happens. Bad JUJU.
    At least that's what I think based on my limited understanding.
  5. User undervolting fix by force calibration of SmartReflex.

    imoseyon authored and jdkoreclipse committed Dec 31, 2011
    There are some ugly hacks in there but the hacks are harmless,
    I think. Since force recalibration only takes effect for the
    frequency that you're currently on, I implemented a full
    cycle through of all valid frequencies, waiting 2 sec in
  6. Added SLQB memory allocator.

    Ezekeel authored and jdkoreclipse committed Dec 17, 2011
  7. make it easy to build

    jdkoreclipse committed Dec 20, 2011
  8. cpufreq updates from 3.2

    jdkoreclipse committed Dec 19, 2011
  9. compiler optimizations

    jdkoreclipse committed Dec 21, 2011
  10. Enable SGX OPP3

    1076 authored and jdkoreclipse committed Dec 7, 2011
  11. Reduce 1.4GHz to 1380 mV

    1076 authored and jdkoreclipse committed Dec 7, 2011
  12. 1.4GHz overclock

    1076 authored and jdkoreclipse committed Dec 7, 2011
  13. Add userspace undervolting

    1076 authored and jdkoreclipse committed Dec 7, 2011
  14. patch to 3.0.16

    jdkoreclipse committed Jan 7, 2012
Commits on Dec 15, 2011
  1. OMAP4: PM: omap4_trigger_ioctrl: use proper mask for proper delay

    Girish S Ghongdemath authored and arve-android committed Dec 4, 2011
    PRM_IO_PMCTRL register has more fields beyond WUCLK_STATUS.
    Shifting the bits does not guarantee that the resultant value
    represents the WUCLK_STATUS (a single bit). Instead, mask the
    same to be sure about picking up just the register field we need.
    Without this, we may not manage to get the test to pass and it
    times out after 1+ms in every attempt. This wrongly forces the
    prcm_interrupt_handler to be delayed by 1ms when IOPAD wakeup
    is triggered. This causes the IO daisy chain blind window to
    be abnormally long, instead of a few us.
    Measurements show that time between rising and falling edge
    of WUCLIN is:
    * without fix: 1600us
    * with fix: 1.2us
    Also the latch is really complete when the WUCLKIN pulse goes
    low, so we need to wait until WUCLK_STATUS goes low.
    Change-Id: Ib0a052968796181b4b252afc13bca8f12cdd940b
    [ parallel fix]
    Signed-off-by: Djamil Elaidi <>
    Signed-off-by: Girish S Ghongdemath <>
  2. ARM: omap4: gpio: Use pad wakeupevent to detect pulses for edge trigg…

    arve-android committed Dec 14, 2011
    …ered gpios
    When the gpio controller is off we use the pad wakeup to wake up.
    The current code only detects edges if the state of the gpio when
    we wake up is different from the state we read before we turned
    the gpio controller off. This will miss short pulses. We now use
    the pad wakeupevent status to try to detect these pulses.
    If the interrupt is configured for dual edge triggering this
    will generate an interrupt for any pulse detected. If the
    interrupt is configured for single edge detect we only
    generate an interrupt if the old state matches the starting
    state for the enabled edge detect. This will still miss a
    pulse when the interrupt is configured for only the second
    edge in that pulse. We may also falsely generate an interrupt
    if an edge of the wrong polarity occured before reading the
    initial state (since the pad wakeup is enabled before then).
    Change-Id: Id9a68cd97b0281a7debb2f37d845979894b63bcc
    Signed-off-by: Arve Hjønnevåg <>
  3. ARM: omap4: gpio: Fix omap2_gpio_set_wakeupenables to have an effect

    arve-android committed Dec 14, 2011
    The wakeupenable bits set by this function do not take effect unless
    omap4_trigger_ioctrl is called. This function cannot be safely called
    after any gpio bank is disabled however, so enable the pad wakeup for
    all the gpio banks, call omap4_trigger_ioctrl, then disable the gpio
    Change-Id: I4dd5aac70ebfebe9187fc4058d6470e03f89035b
    Signed-off-by: Arve Hjønnevåg <>
  4. Revert "Revert "ARM: omap4: gpio: Fix lost gpio edge interrutps in lo…

    arve-android committed Dec 15, 2011
    …w power idle modes.""
    This reverts commit 6e26e77.
Commits on Dec 14, 2011
  1. Revert "HACK: ARM: tuna: disable L3_SLEEP when MHL is active"

    colincross committed Dec 14, 2011
    This reverts commit bdc7da3.
    Change-Id: Ic99fe7def849258cf4d217a5944abc647575ce5b
  2. omap: dispc: force L3_1 CD to NOSLEEP when dispc module is active.

    Avinash.H.M authored and colincross committed Oct 11, 2011
    With the DSS FIFO optimizations, ramdom lockups and reboots are seen. It has
    been identified that L3_1 CD is idling and not responding to the traffic
    initiated by DSS.  The Workaround suggested by Hardware team is to keep the
    L3_1 CD in NO_SLEEP mode, when DSS is active.  Once DSS module is switched to
    idle mode, put L3_1 CD to HW_AUTO.
    Change-Id: I28ae362ad330a79a493927575c9570462c4303a7
    Signed-off-by: Akash Choudhari <>
    Signed-off-by: Avinash.H.M. <>
    Signed-off-by: Sunita Nadampalli <>
    CC: Todd Poynor <>
    CC: Colin Cross <>
    CC: Nishanth Menon <>
    CC: Girish S G <>
    CC: Mahesh <>
    CC: Dandawate Saket <>
    CC: Lajos Molnar <>
  3. HACK: video: ti_hdmi: print warning on invalid AKSV

    Erik Gilling
    Erik Gilling committed Dec 14, 2011
    Change-Id: I6851fec6e53d9aac3a0a58eb9ea4bc75d4783ea3
    Signed-off-by: Erik Gilling <>
  4. Revert "HACK: OMAP4: HDCP: enable HDCP debug messages"

    Erik Gilling
    Erik Gilling committed Dec 14, 2011
    This reverts commit 4a9df85.
  5. modem_if: io: initialize skb->mac_header before passing it to netif_r…

    Chia-chi Yeh authored and benoitgoby committed Dec 14, 2011
    Signed-off-by: Chia-chi Yeh <>