kernel.git
11 years agofix-one-mmc-race.patch
Andy Green [Wed, 19 Nov 2008 17:10:52 +0000 (17:10 +0000)]
fix-one-mmc-race.patch

Some boots from Qi trigger a symptom from this interesting race -->

[    2.730000] Unable to handle kernel NULL pointer dereference at virtual address 00000248
[    2.730000] pgd = c0004000
[    2.735000] [00000248] *pgd=00000000
[    2.735000] Internal error: Oops: 5 [#1] PREEMPT
[    2.735000] Modules linked in:
[    2.735000] CPU: 0    Not tainted  (2.6.24-stable10_0c1587137aaf0ee3-mokodev #1071)
[    2.735000] PC is at pcf50633_voltage_set+0x1c/0xfc
[    2.735000] LR is at gta02_glamo_mmc_set_power+0xdc/0x128
[    2.735000] pc : [<c01df570>]    lr : [<c0034324>]    psr: 60000013
[    2.735000] sp : c7c57eb0  ip : c7c57ec8  fp : c7c57ec4
[    2.735000] r10: c7cfca28  r9 : 00000000  r8 : c7c57f68
[    2.735000] r7 : c7cfca68  r6 : c7cfcae0  r5 : 00000c80  r4 : 00000000
[    2.735000] r3 : 00000000  r2 : 00000c80  r1 : 0000000a  r0 : 00000c80
[    2.735000] Flags: nZCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment kernel
[    2.735000] Control: c000717f  Table: 30004000  DAC: 00000017
[    2.735000] Process kmmcd (pid: 102, stack limit = 0xc7c56268)
[    2.735000] Stack: (0xc7c57eb0 to 0xc7c58000)
[    2.735000] 7ea0:                                     c0608c58 00000c80 c7c57edc c7c57ec8
[    2.735000] 7ec0: c0034324 c01df564 c7cfca28 c7cfc800 c7c57f1c c7c57ee0 c0194de0 c0034258
[    2.735000] 7ee0: c7c57f34 c7c57ef0 c01e6230 c005de5c 60000013 c7cfca28 c7cfc800 60000013
[    2.735000] 7f00: c7cfca68 c7c57f68 00000000 c01e6778 c7c57f34 c7c57f20 c01e5d68 c0194da8
[    2.735000] 7f20: c7cfc800 c7cfca08 c7c57f5c c7c57f38 c01e6810 c01e5cbc c0059278 c7c57f48
[    2.735000] 7f40: c02d2ba0 00000002 c7c44420 c7c56000 c7c57f9c c7c57f60 c00592e0 c01e6788
[    2.735000] 7f60: 00000002 c0059278 c0608d74 c04321cc c036e16c 00000000 c7c57fb0 c7c44420
[    2.735000] 7f80: c7c56000 00000000 00000000 00000000 c7c57fd4 c7c57fa0 c005a068 c00591ec
[    2.735000] 7fa0: c02d0624 00000000 c7c4c0e0 c005dc2c c7c57fb0 c7c57fb0 00000000 c7c56000
[    2.735000] 7fc0: c7c44420 c0059f84 c7c57ff4 c7c57fd8 c005db28 c0059f94 00000000 00000000
[    2.735000] 7fe0: 00000000 00000000 00000000 c7c57ff8 c004b170 c005dad8 ffffffff ffffffff
[    2.735000] Backtrace:
[    2.735000] [<c01df554>] (pcf50633_voltage_set+0x0/0xfc) from [<c0034324>] (gta02_glamo_mmc_set_power+0xdc/0x128)
[    2.735000]  r5:00000c80 r4:c0608c58
[    2.735000] [<c0034248>] (gta02_glamo_mmc_set_power+0x0/0x128) from [<c0194de0>] (glamo_mci_set_ios+0x48/0x254)
[    2.735000]  r5:c7cfc800 r4:c7cfca28
[    2.735000] [<c0194d98>] (glamo_mci_set_ios+0x0/0x254) from [<c01e5d68>] (mmc_power_up+0xbc/0x100)
[    2.735000] [<c01e5cac>] (mmc_power_up+0x0/0x100) from [<c01e6810>] (mmc_rescan+0x98/0x1a8)
[    2.735000]  r5:c7cfca08 r4:c7cfc800
[    2.735000] [<c01e6778>] (mmc_rescan+0x0/0x1a8) from [<c00592e0>] (run_workqueue+0x104/0x208)
[    2.735000]  r6:c7c56000 r5:c7c44420 r4:00000002
[    2.735000] [<c00591dc>] (run_workqueue+0x0/0x208) from [<c005a068>] (worker_thread+0xe4/0xf8)
[    2.735000] [<c0059f84>] (worker_thread+0x0/0xf8) from [<c005db28>] (kthread+0x60/0x94)
[    2.735000]  r6:c0059f84 r5:c7c44420 r4:c7c56000
[    2.735000] [<c005dac8>] (kthread+0x0/0x94) from [<c004b170>] (do_exit+0x0/0x6f4)
[    2.735000]  r6:00000000 r5:00000000 r4:00000000
[    2.735000] Code: e351000a e1a04000 e1a00002 8a000032 (e5943248)
[    2.745000] ---[ end trace 123ec1d286354824 ]---

This problem was caused by insufficient timeout waiting for pcf50633 to resume
and broken code to detect timeout exhaustion.

Although I'd like to think it has something to do with mmc resume woes it should make a panic
and subsequent emergency spew on UART2 if that had been the case.

Took the opportunity to move the stuff to show completion of probe to later in the
pcf50633 probe and tighten readiness test.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agogta02-accel-isr-fix-more.patch
Andy Green [Wed, 19 Nov 2008 17:10:52 +0000 (17:10 +0000)]
gta02-accel-isr-fix-more.patch

Fix all of the mismatches Andrzej put us on to.

Found-by: Andrzej Zaborowski <balrogg@gmail.com>
Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoFrom: Andrzej Zaborowski <balrogg@gmail.com>
Andy Green [Wed, 19 Nov 2008 17:10:52 +0000 (17:10 +0000)]
From: Andrzej Zaborowski <balrogg@gmail.com>

fix-accel-irq-mismatch.patch

I just found a while to start doing something cool with the
accelerometers but I hit #1613 (both accelerometer nodes can't be read
concurrently for longer than a moment). Turns out to be very silly.
I'll continue the cool stuff another day,
Cheers

11 years agocheckpatch-accel-fixes.patch
Andy Green [Wed, 19 Nov 2008 17:10:51 +0000 (17:10 +0000)]
checkpatch-accel-fixes.patch

style cleanups for accel threshold setting patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agointerface for configuring freefall/wakeup interrupts for the accelerometers
Simon Kagstrom [Wed, 19 Nov 2008 17:10:51 +0000 (17:10 +0000)]
interface for configuring freefall/wakeup interrupts for the accelerometers

Hi!

First: Unfortunately, the freerunner currently wakes up immediately on
suspend when the accelerometer IRQ is selected as a wakeup source. I'm
posting this for comments and if someone else wants to have a look at
this problem.

The patch should be safe to apply even though the sleep portion doesn't
work - as long as it's configured it will not disturb anything.

// Simon
--
lis302dl-configure-wakeup-interrupts.patch

From: simon.kagstrom <simon.kagstrom@gmail.com>

First: Unfortunately, the freerunner currently wakes up immediately on
suspend when the accelerometer IRQ is selected as a wakeup source.

Add configuration of wakeup/freefall interrupts through a sysfs
interface. Configuration is done through echoing a value of the
form

   X Y Z THRESHOLD DURATION SPEC

to freefall_wakeup_1/2. X, Y and Z are threshold values, given as a
value > 0, < 0 or 0 to specify if an interrupt should be generated for
high or low thresholds or neither (off). THRESHOLD specifies the
threshold that must be exceeded. DURATION specifies the time in
milliseconds for which the acceleration should be measured. SPEC is
either '1' or '0' and specifies if the thresholds should be taken all
together or one at a time ('and' or 'or' mode).

Echoing '0' to the file turns off the interrupts.

Example:

  echo "1 1 1 60 60 0" > freefall_wakeup_1   # Turn on x,y,z, 60ms/60 threshold, or-mode
  echo "0" > freefall_wakeup_1 # Turn off interrupt

The hardware supports two simulataneous wakeup sources to be configured,
but the freerunner only connects one of the interrupt outputs. The patch
exports both. Similarly, only the "top" accelerometer can be used as a
wake-up source, and it's not possible to generate DATA_READY interrupts
while the wakeup interrupts are active.

Signed-off-by: Simon Kagstrom <simon.kagstrom@gmail.com>
11 years agoFrom 5ee1ee9e1c8a652b0f9cde72ad5e547db87d4d67 Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:10:51 +0000 (17:10 +0000)]
From 5ee1ee9e1c8a652b0f9cde72ad5e547db87d4d67 Mon Sep 17 00:00:00 2001
Subject: [PATCH] [gta02] Disable hardware ECC unless we get instructed to enable it
     This is restoring the old behavior in regard to ECC. Even if
     hardware ECC was compiled in we didn't use it. Make this a runtime
     option. If the bootloader passes hardware_ecc we will enable the
     hardware ECC for real.

11 years agoFrom ca800f8b234299c56530e56018c031957c5c7ffb Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:10:51 +0000 (17:10 +0000)]
From ca800f8b234299c56530e56018c031957c5c7ffb Mon Sep 17 00:00:00 2001
Subject: [PATCH] [gta02] Disable the NAND debugging as it would print every hardware ECC

11 years agoFrom 98d97ee93af676f7d6d0bf55aaae17e11304598a Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:10:51 +0000 (17:10 +0000)]
From 98d97ee93af676f7d6d0bf55aaae17e11304598a Mon Sep 17 00:00:00 2001
Subject: [PATCH] Revert "s3c2440-nand-disable-hwecc.patch"

This reverts commit 1d89da736ed33d3f7c398fb9f8dfddecb7c7c7a9.

11 years agolis302dl-add-wakeup-defs.patch
Simon Kagstrom [Wed, 19 Nov 2008 17:10:51 +0000 (17:10 +0000)]
lis302dl-add-wakeup-defs.patch

Add definitions for the rest of the wakeup defs and also change FFWUSRC1
to FFWUSRC - there are two of these which are identical.

Signed-off-by: Simon Kagstrom <simon.kagstrom@gmail.com>
11 years agolis302dl-allow-unloading-module.patch
Simon Kagstrom [Wed, 19 Nov 2008 17:10:51 +0000 (17:10 +0000)]
lis302dl-allow-unloading-module.patch

This patch fixes module unloading for the accelerometer (actually module
loading failed before). The two problems were that the interrupt was not
unregistered, and that the device was left in a "strange" state.

Signed-off-by: Simon Kagstrom <simon.kagstrom@gmail.com>
11 years agomach-gta02-spell-fixes.patch
Simon Kagstrom [Wed, 19 Nov 2008 17:10:51 +0000 (17:10 +0000)]
mach-gta02-spell-fixes.patch

Fix spelling error on function name

Signed-off-by: Simon Kagstrom <simon.kagstrom@gmail.com>
11 years agotracking-2.6.27-rc2-fix-fiq.patch
Andy Green [Wed, 19 Nov 2008 17:10:50 +0000 (17:10 +0000)]
tracking-2.6.27-rc2-fix-fiq.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoconfig-2.6.27-rndis.patch
Andy Green [Wed, 19 Nov 2008 17:10:50 +0000 (17:10 +0000)]
config-2.6.27-rndis.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-build-test-for-moredrivers.patch
Andy Green [Wed, 19 Nov 2008 17:10:50 +0000 (17:10 +0000)]
fix-build-test-for-moredrivers.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoadd-s3c-serial-total-meddlings.patch
Andy Green [Wed, 19 Nov 2008 17:10:50 +0000 (17:10 +0000)]
add-s3c-serial-total-meddlings.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agorevert-drivers-serial-s3c2410.c-meddlings.patch
Andy Green [Wed, 19 Nov 2008 17:10:50 +0000 (17:10 +0000)]
revert-drivers-serial-s3c2410.c-meddlings.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc2-include-path-changes.patch
Andy Green [Wed, 19 Nov 2008 17:10:49 +0000 (17:10 +0000)]
tracking-2.6.27-rc2-include-path-changes.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoconfig-s3c2410-serial.patch
Andy Green [Wed, 19 Nov 2008 17:10:48 +0000 (17:10 +0000)]
config-s3c2410-serial.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoconfig-2.6.27-rc2-disable-new-not-needed.patch
Andy Green [Wed, 19 Nov 2008 17:10:48 +0000 (17:10 +0000)]
config-2.6.27-rc2-disable-new-not-needed.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-no-discharging.patch
Andy Green [Wed, 19 Nov 2008 17:10:48 +0000 (17:10 +0000)]
fix-no-discharging.patch

We failed to report status of "discharging", instead reporting
"not charging" even if we knew that the charger was not present.

This patch corrects it and reports "discharging" when charger
is absent.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoadd-includes-from-checkpatch-fixes.patch
Andy Green [Wed, 19 Nov 2008 17:10:48 +0000 (17:10 +0000)]
add-includes-from-checkpatch-fixes.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agocheckpatch-fixes.patch
Andy Green [Wed, 19 Nov 2008 17:10:48 +0000 (17:10 +0000)]
checkpatch-fixes.patch

This cleans out some random externs in C files that checkpatch does not like
and introduces a couple of .h files to contain them.  Plus some other minor
checkpatch style complaints.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agogta01-uart-fifo-trigger-sooner.patch
Mike Westerhof [Wed, 19 Nov 2008 17:10:48 +0000 (17:10 +0000)]
gta01-uart-fifo-trigger-sooner.patch

    Set the UART FIFO to trigger earlier on the GTA01 device to minimize
    UART overruns from the GSM.

Signed-off-by: Mike Westerhof <mwester@dls.net>
11 years agofix-suspend-backlight-timing-gta01.patch
Mike Westerhof [Wed, 19 Nov 2008 17:10:48 +0000 (17:10 +0000)]
fix-suspend-backlight-timing-gta01.patch

    This patch adds the gta01 backlight callback that defers the
    restoring of the backlight until after the jbt driver has
    resumed.  This doesn't eliminate the flashing of the LCD on
    the gta01, but it reduces it considerably.

Signed-off-by: Mike Westerhof <mwester@dls.net>
11 years agogta01-pcf50606-disable-irq-from-suspend-until-resume.patch
Mike Westerhof [Wed, 19 Nov 2008 17:10:47 +0000 (17:10 +0000)]
gta01-pcf50606-disable-irq-from-suspend-until-resume.patch

    This patch is the pcf50606 equivalent of the pcf50633 patch that
    disables interrupts from the chip until after resume is complete.
    In order to ensure no data is lost, the work function is called
    post-resume to process any pending interrupts.

    Most of the code was quite literally re-used from Andy Green's
    original patch.

Signed-off-by: Mike Westerhof <mwester@dls.net>
11 years agosilence-serial-console-gta01.patch
Mike Westerhof [Wed, 19 Nov 2008 17:10:47 +0000 (17:10 +0000)]
silence-serial-console-gta01.patch

    This patch ensures that no console data will go the UART while
    the GSM mux is switched to the GSM.  This is necessary despite
    the code that disables the console due to the fact that the
    console resumes before the neo1973_pm_gsm driver, and consoles
    always resume in the "on" state.

Signed-off-by: Mike Westerhof <mwester@dls.net>
11 years agocommit 830ea3d0c27c0c750b7bf1b56c002ee7943f3edc
Mike Westerhof [Wed, 19 Nov 2008 17:10:47 +0000 (17:10 +0000)]
commit 830ea3d0c27c0c750b7bf1b56c002ee7943f3edc

    gta0x-log-serial-rx-error.patch

    This patch causes a KERN_DEBUG message to be printed each time an
    error status is read from a UART.  This is intended to facilitate
    the reporting of more useful problem and bug reports from users
    in the field.

Signed-off-by: Mike Westerhof <mwester@dls.net>
11 years agogta0x-add-minimal-GSM-flowcontrol.patch
Mike Westerhof [Wed, 19 Nov 2008 17:10:47 +0000 (17:10 +0000)]
gta0x-add-minimal-GSM-flowcontrol.patch

    Add the basic GSM flowcontrol code.

Signed-off-by: Mike Westerhof <mwester@dls.net>
11 years agoconfig-add-udf-support.patch
Andy Green [Wed, 19 Nov 2008 17:10:47 +0000 (17:10 +0000)]
config-add-udf-support.patch

Timo noticed that his compound USB device wasn't getting recognized properly,
it was stuck on the CD-ROM class part of it.  This was accepted by the GTA02
but no scsi device was spawned for it.  He figured out in the end that it is
the lack of a UDF filesystem module that was the problem.

This patch adds UDF and CDROM as a built-in on defconfig-2.6.24, for the
modular configs it just adds it as a module.

Reported-by: Timo Drick <timo@drick.de>
Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoconfig-remove-evbug-module.patch
Andy Green [Wed, 19 Nov 2008 17:10:47 +0000 (17:10 +0000)]
config-remove-evbug-module.patch

Carsten mentioned that evbug module is getting modprobed in
to ASU somehow and making havoc.  This patch stops us building
it because there's no time to figure out why it gets modprobed
in and it doesn't seem to do anything but spew syslog -- and we
never needed it before now.

Reported-by: Carsten Haitzler <raster@openmoko.org>
Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoFrom fa269b44512a03523b164c3cebc20312748c524b Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:10:46 +0000 (17:10 +0000)]
From fa269b44512a03523b164c3cebc20312748c524b Mon Sep 17 00:00:00 2001
Subject: [PATCH] [ar6k] Build the wireless driver without DEBUG
    - Remove DEBUG from the Makefile

    - Do not send events through netlink to userspace. We might
      need to reevaluate this. But we seem to use wireless_send_event
      at the right places.
      (SEND_EVENT_TO_APP)

    - Do not report debug logs to apps
      (REPORT_DEBUG_LOGS_TO_APP)

Signed-Off-By: Holger Freyther <zecke@openmoko.org>
11 years agoadd-build-makerecovery.patch
Andy Green [Wed, 19 Nov 2008 17:10:46 +0000 (17:10 +0000)]
add-build-makerecovery.patch

This patch adds a script that creates a "recovery kernel" image
from a recovery-uImage-moredrivers-... image.  The use of these
recovery images is explained here:

http://wiki.openmoko.org/wiki/Freerunner_backup_kernel

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-build-hash-name-no-colon-separator.patch
Andy Green [Wed, 19 Nov 2008 17:10:46 +0000 (17:10 +0000)]
fix-build-hash-name-no-colon-separator.patch
Suggested-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Colons are not very friendly for tar... scp has trouble too but
you can override that by prepending any path like ./ ... still
better to dilute the colon down to a _ avoiding - because it can
appear in branch names we use.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoadd-buildkeep-hash-named-copy.patch
Andy Green [Wed, 19 Nov 2008 17:10:46 +0000 (17:10 +0000)]
add-buildkeep-hash-named-copy.patch

Keep a copy of last "moredrivers" config kernel built stamped with
branch and git head at build time.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoadd-build-git-head-info.patch
Andy Green [Wed, 19 Nov 2008 17:10:44 +0000 (17:10 +0000)]
add-build-git-head-info.patch

This patch adds the branch and truncated head has to the version of the
kernel

# cat /proc/version
Linux version 2.6.26-andy-2.6.26:2b3bf342baac52b9-mokodev (agreen@pads.home.warmcat.com) (gcc version 4.1.2) #878 PREEMPT Sun Jul 27 14:45:25 BST 2008

Note this versioning is visible down /lib/modules

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-pcf50633-usbrem-charging-led.patch
Andy Green [Wed, 19 Nov 2008 17:10:04 +0000 (17:10 +0000)]
fix-pcf50633-usbrem-charging-led.patch

If the charger was removed, we fell through back to old
hdq-driven code with normally wrong but slightly random
results for charging LED behaviour in that circumstance

This patch makes us use the tracked charger status
callbacks alone if they are defined in the platform
data.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-glamo-crank-memory-to-90MHz.patch
Andy Green [Wed, 19 Nov 2008 17:10:03 +0000 (17:10 +0000)]
fix-glamo-crank-memory-to-90MHz.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-glamo-turbo-host-interface.patch
Andy Green [Wed, 19 Nov 2008 17:10:03 +0000 (17:10 +0000)]
fix-glamo-turbo-host-interface.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agopcf50606: fix gllin compat link
Cesar Eduardo Barros [Wed, 19 Nov 2008 17:10:03 +0000 (17:10 +0000)]
pcf50606: fix gllin compat link

Uses bus_create_device_link to correctly create the gllin compat link.

Signed-off-by: Cesar Eduardo Barros <cesarb@cesarb.net>
11 years agodevice model: Allow the creation of symlinks on /sys/bus/*/devices
Cesar Eduardo Barros [Wed, 19 Nov 2008 17:10:03 +0000 (17:10 +0000)]
device model: Allow the creation of symlinks on /sys/bus/*/devices

Allows the direct creation of symlinks on /sys/bus/*/devices. This is needed
for a compat symlink from gta01-pm-gps.0 to neo1973-pm-gps.0 on the Openmoko
Neo1973 GTA01.

Signed-off-by: Cesar Eduardo Barros <cesarb@cesarb.net>
11 years agoSigned-Off-Number: Holger Freyther <zecke@openmoko.org>
Andy Green [Wed, 19 Nov 2008 17:10:03 +0000 (17:10 +0000)]
Signed-Off-Number: Holger Freyther <zecke@openmoko.org>

11 years agoconfig-defconfig-2.6.27-for-rc1.patch
Andy Green [Wed, 19 Nov 2008 17:10:03 +0000 (17:10 +0000)]
config-defconfig-2.6.27-for-rc1.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoremove-s3c24xx-serial-resume-dep-gsm-pm.patch
Andy Green [Wed, 19 Nov 2008 17:10:03 +0000 (17:10 +0000)]
remove-s3c24xx-serial-resume-dep-gsm-pm.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-snd_soc_dapm_sync_endpoints.patch
Andy Green [Wed, 19 Nov 2008 17:10:02 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-snd_soc_dapm_sync_endpoints.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-snd_soc_dapm_set_endpoint.patch
Andy Green [Wed, 19 Nov 2008 17:10:02 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-snd_soc_dapm_set_endpoint.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-broken-cfg-uninit-nand.patch
Andy Green [Wed, 19 Nov 2008 17:10:02 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-broken-cfg-uninit-nand.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-s3c_lookup_cpu-mismatch.patch
Andy Green [Wed, 19 Nov 2008 17:10:02 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-s3c_lookup_cpu-mismatch.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-gpio-redef.patch
Andy Green [Wed, 19 Nov 2008 17:10:02 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-gpio-redef.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-ar6000.patch
Andy Green [Wed, 19 Nov 2008 17:10:02 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-ar6000.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-last-2400-ordering.patch
Andy Green [Wed, 19 Nov 2008 17:10:01 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-last-2400-ordering.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-borked-eth-gadget.patch
Andy Green [Wed, 19 Nov 2008 17:10:01 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-borked-eth-gadget.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-tty-not-in-uart_port.patch
Andy Green [Wed, 19 Nov 2008 17:10:01 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-tty-not-in-uart_port.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-asm-semaphore-gone.patch
Andy Green [Wed, 19 Nov 2008 17:10:01 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-asm-semaphore-gone.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-mmc-ultiwrite-gone.patch
Andy Green [Wed, 19 Nov 2008 17:10:01 +0000 (17:10 +0000)]
tracking-2.6.27-mmc-ultiwrite-gone.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-irqtype-falling-glamo.patch
Andy Green [Wed, 19 Nov 2008 17:10:01 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-irqtype-falling-glamo.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-backlight-def-pcf50633.patch
Andy Green [Wed, 19 Nov 2008 17:10:00 +0000 (17:10 +0000)]
fix-backlight-def-pcf50633.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-rc1-pcf50306-defs.patch
Andy Green [Wed, 19 Nov 2008 17:10:00 +0000 (17:10 +0000)]
tracking-2.6.27-rc1-pcf50306-defs.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-gpio-redef-clean.patch
Andy Green [Wed, 19 Nov 2008 17:10:00 +0000 (17:10 +0000)]
tracking-2.6.27-gpio-redef-clean.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.27-irqtype-rename.patch
Andy Green [Wed, 19 Nov 2008 17:10:00 +0000 (17:10 +0000)]
tracking-2.6.27-irqtype-rename.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoconfig-add-defconfig-2.6.27
Andy Green [Wed, 19 Nov 2008 17:09:59 +0000 (17:09 +0000)]
config-add-defconfig-2.6.27

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoconfig-usb-over-ethernet-modul
Andy Green [Wed, 19 Nov 2008 17:09:58 +0000 (17:09 +0000)]
config-usb-over-ethernet-modul

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-add-missing-include.patch
Andy Green [Wed, 19 Nov 2008 17:09:58 +0000 (17:09 +0000)]
fix-add-missing-include.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agouse-gta02-glamo-mci-sd-dynamic-clock.patch
Andy Green [Wed, 19 Nov 2008 17:09:58 +0000 (17:09 +0000)]
use-gta02-glamo-mci-sd-dynamic-clock.patch

This patch uses the new glamo-mci slow clock ratio
patch in order to dynamically reduce SD Card clock
rate when the GPS unit is powered on GTA02.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoadd-glamo-mci-slower-clocking-dynamic-switching.patch
Andy Green [Wed, 19 Nov 2008 17:09:58 +0000 (17:09 +0000)]
add-glamo-mci-slower-clocking-dynamic-switching.patch

This patch gives glamo-mci a concept of a platform-defined
dynamic clock slowing callback.  It means that platform code
can associate some completely external state to decide if
we run the SD clock at normal rate or a rate divided by a
module parameter "sd_slow_ratio", which you can set on
kernel commandline like this:

glamo_mci.sd_slow_ratio=8

you can also change it at runtime by

echo 8 > /sys/module/glamo_mci/parameters/sd_slow_ratio

If no platform callback is defined, then no slow mode
is used.  If it is defined, then the default division
action is / 8, eg, 16MHz normal -> 2MHz slow mode.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-glamo-mci-dont-filter-voltage-change.patch
Andy Green [Wed, 19 Nov 2008 17:09:58 +0000 (17:09 +0000)]
fix-glamo-mci-dont-filter-voltage-change.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-allow-full-sd-voltage-range-selection.patch
Andy Green [Wed, 19 Nov 2008 17:09:58 +0000 (17:09 +0000)]
fix-allow-full-sd-voltage-range-selection.patch

Until now we just drove the SD Card at 3.3V all the time.  But in
fact we can do better, and use a voltage negotiated with the
SD Card itself.

With the shipping 512MB Sandisk SD Card, 2.7V is negotiated which
gives 1.7dBm reduction in power on all the SD Card lines and should
further reduce GPS perturbation during SD Card usage.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-2.6.26-rc9-dai-alsa.patch
Andy Green [Wed, 19 Nov 2008 17:09:58 +0000 (17:09 +0000)]
tracking-2.6.26-rc9-dai-alsa.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-glamo-mci-ensure-more-than-74-clocks-after-power.patch
Andy Green [Wed, 19 Nov 2008 17:09:58 +0000 (17:09 +0000)]
fix-glamo-mci-ensure-more-than-74-clocks-after-power.patch

We are meant to run SD_CLK a little while after power-on for the SD
Card, but with the no idle clock changes we didn't take care about it.

This makes us sleep a little bit before disabling clock if we just
powered up the SD Card.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-glamo-mci-possible-timeout-overflow.patch
Andy Green [Wed, 19 Nov 2008 17:09:57 +0000 (17:09 +0000)]
fix-glamo-mci-possible-timeout-overflow.patch

The MMC stack hands us a timeout calibrated in SD_CLK clocks, but the
Glamo can only deal with up to 65520 clocks of timeout.  If the stack
handed us a request bigger than this, it would just wrap and the
timeout we actually used would be way too short.

With this patch if that happens, we use the longest timeout we can,
65520 clocks and give it our best shot.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-pcf50633-only-do-platform-callback-once-per-event.patch
Andy Green [Wed, 19 Nov 2008 17:09:57 +0000 (17:09 +0000)]
fix-pcf50633-only-do-platform-callback-once-per-event.patch
Reported-by: Holger Freyther <zecke@openmoko.org>
We harmlessly repeated PMU platform callbacks about charging state twice.
Clean it up and leave it to pcf50633_charge_enable() to report once.

Also tidies the sequencing so we set current limit before we enable
charger now.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-pcf50633-remove-charger-curlim-and-enable-apis-from-export.patch
Andy Green [Wed, 19 Nov 2008 17:09:57 +0000 (17:09 +0000)]
fix-pcf50633-remove-charger-curlim-and-enable-apis-from-export.patch

Setting the current limit directly and enabling the charger
isn't anyone's business except pcf50633 driver itself, so these
two functions should not be exported and become static.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-glamo-mci-set-default-drive-level-0.patch
Andy Green [Wed, 19 Nov 2008 17:09:57 +0000 (17:09 +0000)]
fix-glamo-mci-set-default-drive-level-0.patch

Tests on access to SD Card with Glamo drive level "0" show
that it reduces SD_CLK energy at 1.5GHz by 24dBm compared to
drive level 3.  This puts it only 6dB above the background
noise floor compared to 30dB and should make a solution for
GPS trouble with SD Card in.

SD card communication seems unaffected so far on the Sandisk
512MB card we ship.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-bq27000-charger-state-tracking.patch
Andy Green [Wed, 19 Nov 2008 17:09:57 +0000 (17:09 +0000)]
fix-bq27000-charger-state-tracking.patch

Charger trigger stuff goes and asks for POWER_SUPPLY_PROP_STATUS
to figure out what the charger state is.  But until now, we only
reported there what we found out from HDQ, and the HDQ registers
are not updated very often in the coulomb counter, it can be 4
or more second lag before it tells us about what it experiences.

When we react to USB insertion and only after 500ms debounce tell
power_supply stuff that something changed, it most times will
see old pre-USB-insertion state from bq27000 over HDQ at that time
and will report it ain't charging, buggering up the LED trigger
tracking.

This patch maintains distance between bq27000 and pcf50633 by
having platform callbacks in bq27000 that it can use to ask about
definitive charger "online" presence and "activity", whether the
charger says it is charging.  If these callbacks are implemented
(and we implement them in this patch up in mach_gta02.c) then
this information is used in preference to what is found from
HDQ.

Result is if you set the LED trigger like this:

echo bat-charging > /sys/devices/platform/gta02-led.0/leds/gta02-aux:red/trigger

then it lights up properly on USB insertion now, goes away on
removal properly, as as far as I saw, when charging stops too.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agodebug-add-glamo-drive-strength-module-param.patch
Andy Green [Wed, 19 Nov 2008 17:09:57 +0000 (17:09 +0000)]
debug-add-glamo-drive-strength-module-param.patch
Suggested-by: Werner Almesberger <werner@openmoko.org>
This patch allows users to control two additional settings
in Glamo MCI driver from kernel commandline or module
parameters.

First is Glamo drive strength on SD IOs including CLK.
This ranges from 0 (weakest) to 3 (strongest).

echo 0 > /sys/module/glamo_mci/parameters/sd_drive

(Changes to this take effect on next SD Card transaction)

or, from kernel commandline

glamo_mci.sd_drive=0

On tests here with 0 strength, communication to SD card
(shipped 512MB Sandisk) seemed fine, and a dd of 10MB
urandom had the same md5 when written to cache as after
a reboot.  I set the default to 2.

Second is whether we allow SD_CLK when the SD interface
is idle.

# stop the clock when we are idle (default)
echo 0 > /sys/module/glamo_mci/parameters/sd_idleclk

# run the SD clock all the time
echo 1 > /sys/module/glamo_mci/parameters/sd_idleclk

(changes take effect on next SD Card transaction)

From kernel commandline, eg:

glamo_mci.sd_idleclk=1

Normally you don't want to run the SD Clock all the time.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agodebug-move-dev-info-to-dbg.patch
Andy Green [Wed, 19 Nov 2008 17:09:57 +0000 (17:09 +0000)]
debug-move-dev-info-to-dbg.patch
Suggested-by: Sean McNeil <sean@mcneil.com>
To see if some subtle race is involved, Sean has tried
removing syslog traffic during resume and found he was
not seeing the resume crash any more.  We're giving it
a try to see if it changes the behaviour for anyone
else.  It would mean we have a pretty fine race in there
somewhere.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-scard-stop-on-resume.patch
Andy Green [Wed, 19 Nov 2008 17:09:57 +0000 (17:09 +0000)]
fix-scard-stop-on-resume.patch
Reported-by: Ville-Pekka Vainio <vpivaini@cs.helsinki.fi>
The reporter noticed SD Card clock is running again after resume.  After
looking at the code I saw I missed two tricks, this will force it off
after resume and will do better generally depending on what the last SD Card
packet was.

Since bulk read packet is normally last action (which set the clock off even
without this) the old patch worked for normal cases.  But after resume, the last
packet on the wire was not a bulk transfer and we didn't take care about the
clock then.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoadd-limit-sdcard-clk-cmdline.patch
Andy Green [Wed, 19 Nov 2008 17:09:56 +0000 (17:09 +0000)]
add-limit-sdcard-clk-cmdline.patch

This patch allows you to control the maximum clock rate that will
be selected for SD Card access, from the kernel commandline using

 glamo_mci.sd_max_clk=10000000

and also from

 echo 10000000 > /sys/module/glamo_mci/parameters/sd_max_clk

although you have to suspend and resume to make the limit operational
on the actual SD_CLK line.

Clocks that are possible are divided down from ~50MHz, so 25000000,
166666661250000010000000, etc.  With Freerunner A5 revision that
has 100R series resistors in SD Card signals, I didn't get reliable
operation above 16MHz.  With A6 revision the series resistors went
down to 75R, maybe it can work at 25MHz.

Reducing the clock rate is something to try if you find that your
SD Card is not communicating properly with the default speed.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-force-sdcard-clk-off-when-idle.patch
Andy Green [Wed, 19 Nov 2008 17:09:56 +0000 (17:09 +0000)]
fix-force-sdcard-clk-off-when-idle.patch

Existing Glamo bit for stopping SD Card Clock when there is no
transfer taking place does not work.  This patch adds stuff around
the transfer code to force the SD clock up when something is going on
and down when it is idle.  This'll save a little power and noise ;-)

I tested it briefly and was able to SD Boot normally on Sandisk 512M.
Wider testing is appreciated.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agosdio: use interruptible sleep in kthread main loops
Jason Uhlenkott [Wed, 19 Nov 2008 17:09:56 +0000 (17:09 +0000)]
sdio: use interruptible sleep in kthread main loops

My FreeRunner's load average was leveling off at 2.00 as a result of
these two kernel threads:

SDIO Helper   D c02b4500     0   255      2
[<c02b4298>] (schedule+0x0/0x2d4) from [<c02b4040>] (__down+0x100/0x158)
[<c02b3f40>] (__down+0x0/0x158) from [<c02b3e2c>] (__down_failed+0xc/0x20)
 r7:00000000 r6:c01fbd64 r5:c7cb7134 r4:c7d58000
[<c01fa6fc>] (CardDetectHelperFunction+0x0/0x1ac) from [<c01fbd80>] (HelperLaunch+0x1c/0x28)
 r5:c7cb7134 r4:c7cb7134
[<c01fbd64>] (HelperLaunch+0x0/0x28) from [<c005bbd0>] (kthread+0x60/0x94)
 r4:c7d58000
[<c005bb70>] (kthread+0x0/0x94) from [<c0048f7c>] (do_exit+0x0/0x744)
 r6:00000000 r5:00000000 r4:00000000

SDIO Helper   D c02b4500     0   259      2
[<c02b4298>] (schedule+0x0/0x2d4) from [<c02b4040>] (__down+0x100/0x158)
[<c02b3f40>] (__down+0x0/0x158) from [<c02b3e2c>] (__down_failed+0xc/0x20)
 r7:c0382bb4 r6:c0382b34 r5:c7d5a000 r4:00000001
[<c01f9104>] (SDIOIrqHelperFunction+0x0/0x29c) from [<c01fbd80>] (HelperLaunch+0x1c/0x28)
 r8:00000000 r7:00000000 r6:c01fbd64 r5:c0382bb4 r4:c0382bb4
[<c01fbd64>] (HelperLaunch+0x0/0x28) from [<c005bbd0>] (kthread+0x60/0x94)
 r4:c7d5a000
[<c005bb70>] (kthread+0x0/0x94) from [<c0048f7c>] (do_exit+0x0/0x744)
 r6:00000000 r5:00000000 r4:00000000

This fixes them to use interruptible sleep primitives while waiting in
their main loops, as is conventional for kernel threads.  They can't
actually be interrupted since kernel threads ignore all signals, but
by sleeping this way they get classified as long term waiters, and
don't get counted as running for purposes of load average calculation.

This is intended as a minimal fix.  In the longer term, it'd probably
make sense to replace the semaphores with completions or something,
and to do away with some of these StudlyCapped wrapper functions.

Signed-off-by: Jason Uhlenkott <jasonuhl@jasonuhl.org>
--

This is untested, but what could possibly go wrong?  ;)

11 years agointroduce-BANKCON-meddling-sysfs.patch
Andy Green [Wed, 19 Nov 2008 17:09:56 +0000 (17:09 +0000)]
introduce-BANKCON-meddling-sysfs.patch

A few questions have been flying around about how optimal
our waitstates are for various things including Glamo.

This patch introduces new sysfs nodes

/sys/devices/platform/neo1973-memconfig.0/BANKCON0
...
/sys/devices/platform/neo1973-memconfig.0/BANKCON7

If you cat them you get translated info about bus speed on
that chip select, eg,

# cat /sys/devices/platform/neo1973-memconfig.0/BANKCON1
BANKCON1 = 0x00000A40
 Type = ROM / SRAM
 PMC  = normal (1 data)
 Tacp = 2 clocks
 Tcah = 0 clocks
 Tcoh = 1 clock
 Tacc = 3 clocks
 Tcos = 1 clock
 Tacs = 0 clocks

You can write them in hex too

# echo 0x200 > /sys/devices/platform/neo1973-memconfig.0/BANKCON1

The write format for BANKCON0 - 5 looks like this

 b1..b0   PMC  Page Mode Config
 b3..b2   Tacp Page Mode Access Cycle
 b5..b4   Tcah Address hold after CS deasserted
 b7..b6   Tcoh CS hold after OE deasserted
 b10..b8  Tacc Access Cycle Period
 b12..b11 Tcos CS setup before OE asserted
 b14..b13 Tacs Address setup before CS asserted

BANKCON 6 and 7 have two extra bits

 b16..b15  MT  Memory type (00=ROM/SRAM, 11=DRAM)

If it's ROM/SRAM, the rest of the bits are as described above.
For DRAM

 b1..b0   SCAN Column address number
 b3..b2   RAS to CAS delay

The patch is intended to let people experiement on their own.  But
of course you will crash things for sure if the timing is wrong, and
you can also trash SD Card data if you make Glamo unstable, so remove
it or remount ro first.  Other horrible things are possible, but
because the settings aren't sticky, you should always be able to
recover by either normal reboot usually or at worst NOR boot and then
dfu.  Most likely you will just crash your session and have to reboot
if your settings are bad, but consider yourself warned bad things are
possible. :-)

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agochange-backlight-level-not-forced-up-on-resume.patch
Andy Green [Wed, 19 Nov 2008 17:09:56 +0000 (17:09 +0000)]
change-backlight-level-not-forced-up-on-resume.patch

We at least always enabled backlight on resume, this patch
changes us to set backlight back to last requested backlight
brightness level on resume.  Note it means that you can
resume with screen blanked, but it should come back if that
happened with touchscreen action as usual.

/sys/class/backlight/pcf50633-bl/actual_brightness
and
/sys/class/backlight/pcf50633-bl/brightness

seem to agree after resume when reportedly they didn't before.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-no-uart-leak-when-gps-off.patch
Andy Green [Wed, 19 Nov 2008 17:09:56 +0000 (17:09 +0000)]
fix-no-uart-leak-when-gps-off.patch

During the suspend current reduction campaign on suspend I
forced the GPS UART to be GPIO and to drive 0 into the GPS
unit so we would not burn current there.  On resume it lets
the pins act as UARTs again.  But really, we should do this
all the time that the GPS unit is off, lest we leak it
enough power to hold internal state and make trouble.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoadd-pcf50633-allow-force-charger-type.patch
Andy Green [Wed, 19 Nov 2008 17:09:56 +0000 (17:09 +0000)]
add-pcf50633-allow-force-charger-type.patch

This patch adds a sysfs node:

/sys/class/i2c-adapter/i2c-0/0-0073/force_usb_limit_dangerous

it allows to force the charging limit regardless of the one chosen
by pcf50633 kernel driver.  As such, if you write a charging limit
here that is not suitable for the power source, and the power source
is not current limited on its side, it could draw more current than
your power source can handle, burn down you house, etc.

If you're certain that your power supply can handle it, you can use
this on your own responsibility to make the amount drawn by the
PMU match what you believed your power supply could handle.

Example usage, in case where you have a dumb 500mA USB charger that
does not have the ID resistor:

# cat /sys/class/i2c-adapter/i2c-0/0-0073/charger_type
host/500mA usb mode 100mA   <=== dumb charger does not ennumerate us
# echo 500 > /sys/class/i2c-adapter/i2c-0/0-0073/force_usb_limit_dangerous
# cat /sys/class/i2c-adapter/i2c-0/0-0073/charger_type
host/500mA usb mode 500mA

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-aux-key-level-by-gtaxx.patch
Andy Green [Wed, 19 Nov 2008 17:09:55 +0000 (17:09 +0000)]
fix-aux-key-level-by-gtaxx.patch
Reported-by: Mickey Lauer <mickey@openmoko.org>
AUX level detection is inverted based on GTA01 or 02

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agofix-pcf50633-mask-second-on-resume.patch
Andy Green [Wed, 19 Nov 2008 17:09:55 +0000 (17:09 +0000)]
fix-pcf50633-mask-second-on-resume.patch

We leave SECOND unmasked on resume, it's like the
situation at probe() time, but there it makes us
turn SECOND off after coldplug action.  So we need
to act like after that has happened, not exactly
like what we do at probe / init time.

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoFrom 3a32be40f78404d5f1185f0b3d6b5632381cb33f Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:09:55 +0000 (17:09 +0000)]
From 3a32be40f78404d5f1185f0b3d6b5632381cb33f Mon Sep 17 00:00:00 2001
Subject: [PATCH] [neo1973 leds] Move from mutex to spinlock because we may not use mutexes
    The led triggers may call set_brightness from atomic contexts. As
    mutex_lock calls might_sleep and sleeping is not allowed in atomic contexts
    we have to switch to spinlocks here.

Signed-Off-By: Holger Freyther <zecke@openmoko.org>
11 years agoFrom cede5c6c9b06ecbb0f7f2df7b7070092b87ddaf8 Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:09:55 +0000 (17:09 +0000)]
From cede5c6c9b06ecbb0f7f2df7b7070092b87ddaf8 Mon Sep 17 00:00:00 2001
Subject: [PATCH] [pcf50633] Avoid ooops on start with inserted usb cable
    The pcf50633_global might not be initialized when we get the first
    usb interrupt. We would oops inside the dev_err because we made up
    a struct device.

Signed-Off-By: Holger Freyther <zecke@openmoko.org>
11 years agocommit 5f42e24d361cd83178fe8da9d68efbf41a011483
Mike Westerhof [Wed, 19 Nov 2008 17:09:55 +0000 (17:09 +0000)]
commit 5f42e24d361cd83178fe8da9d68efbf41a011483

    Add missing initialization for the touchscreen driver for the
    gta01 platform.

Signed-off-by: Mike Westerhof <mwester@dls.net>
11 years agoRemove some bits of nspy + GSM flow control patches that leaked into stable
Mike Westerhof [Wed, 19 Nov 2008 17:09:54 +0000 (17:09 +0000)]
Remove some bits of nspy + GSM flow control patches that leaked into stable

Signed-off-by: Mike Westerhof <mwester@dls.net>
11 years agoadd-ar6k-wake-interrupt.patch
Matt [Wed, 19 Nov 2008 17:09:54 +0000 (17:09 +0000)]
add-ar6k-wake-interrupt.patch

Signed-off-by: Matt Hsu <matt_hsu@openmoko.org>
- add an interrupt for ar6k wifi module

11 years agochange-remove-kernel-charging-led-drive.patch
Andy Green [Wed, 19 Nov 2008 17:09:53 +0000 (17:09 +0000)]
change-remove-kernel-charging-led-drive.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agoFrom c221bb27c8e22daa451e26353140777223d397d2 Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:09:53 +0000 (17:09 +0000)]
From c221bb27c8e22daa451e26353140777223d397d2 Mon Sep 17 00:00:00 2001
Subject: [PATCH] [pcf50633] Report more events to userspace using the default callback

Signed-Off-By: Holger Freyther <zecke@openmoko.org>
11 years agoFrom 5718bde77ed1a75e0fd2cdf5e099e66121d10c0a Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:09:53 +0000 (17:09 +0000)]
From 5718bde77ed1a75e0fd2cdf5e099e66121d10c0a Mon Sep 17 00:00:00 2001
Subject: [PATCH] [battery] Make the bq27000 send an uevent when the charging state possible changed
     Remove the todo entries from the pcf50633, make the mach-gta02
     call the bq27000 driver from the pmu callback.

11 years agoFrom 000450f1ad2c713d2345a872fdf44f5dd3702e1b Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:09:53 +0000 (17:09 +0000)]
From 000450f1ad2c713d2345a872fdf44f5dd3702e1b Mon Sep 17 00:00:00 2001
Subject: [PATCH] [janitor] make checkpatch.pl happy

11 years agoFrom 683ef8067815f6ba0ede73fa71973823726213a3 Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:09:53 +0000 (17:09 +0000)]
From 683ef8067815f6ba0ede73fa71973823726213a3 Mon Sep 17 00:00:00 2001
Subject: [PATCH] [janitor] Make checkpatch happy on the header files

11 years agoFrom ae3f72fc608fcd0a98a980a335ac4dc7ad95b221 Mon Sep 17 00:00:00 2001
Holger Freyther [Wed, 19 Nov 2008 17:09:53 +0000 (17:09 +0000)]
From ae3f72fc608fcd0a98a980a335ac4dc7ad95b221 Mon Sep 17 00:00:00 2001
Subject: [PATCH] [bq27000] Make the checkpatch.pl happy

11 years agoFrom 119f4e02ba81cffe4dbc88d8ff667048ad28d925 Mon Sep 17 00:00:00 2001
Andrzej Zaborowski [Wed, 19 Nov 2008 17:09:53 +0000 (17:09 +0000)]
From 119f4e02ba81cffe4dbc88d8ff667048ad28d925 Mon Sep 17 00:00:00 2001
Subject: [PATCH] Hacky CONFIG_NO_IDLE_HZ (dyn-tick) support for S3C24xx.

11 years agotracking-2.6.26-rc7-repeat-cdev-removal-pcf50633.patch
Andy Green [Wed, 19 Nov 2008 17:09:52 +0000 (17:09 +0000)]
tracking-2.6.26-rc7-repeat-cdev-removal-pcf50633.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agotracking-defconfig-2.6.26-rc7.patch
Andy Green [Wed, 19 Nov 2008 17:09:52 +0000 (17:09 +0000)]
tracking-defconfig-2.6.26-rc7.patch

Signed-off-by: Andy Green <andy@openmoko.com>
11 years agointroduce-resume-exception-capture.patch
Andy Green [Wed, 19 Nov 2008 17:09:52 +0000 (17:09 +0000)]
introduce-resume-exception-capture.patch

This patch introduces a new resume debugging concept: if we
get an OOPS inbetween starting suspend and finishing resume, it
uses a new "emergency spew" device similar to BUT NOT REQUIRING
CONFIG_DEBUG_LL to dump the syslog buffer and then the OOPS
on the debug device defined by the existing CONFIG_DEBUG_S3C_UART
index.  But neither CONFIG_DEBUG_LL nor the S3C low level configs
are needed to use this feature.

Another difference between this feature and CONFIG_DEBUG_LL is that
it does not affect resume timing, ordering or UART traffic UNLESS
there is an OOPS during resume.

The patch adds three global exports, one to say if we are inside
suspend / resume, and two callbacks for printk() to use to init
and dump the emergency data.  The callbacks are set in s3c serial
device init, but the whole structure is arch independent.

Signed-off-by: Andy Green <andy@openmoko.com>