7 years agoahci: Get rid of pci_dev argument in ahci_save_initial_config()
Anton Vorontsov [Wed, 3 Mar 2010 17:17:37 +0000 (20:17 +0300)]
ahci: Get rid of pci_dev argument in ahci_save_initial_config()

To make the function generic we have to get rid of "struct pci_dev *",
so let's pass just a "struct devce *".

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agoahci: Factor out PCI specifics from ahci_save_initial_config()
Anton Vorontsov [Wed, 3 Mar 2010 17:17:36 +0000 (20:17 +0300)]
ahci: Factor out PCI specifics from ahci_save_initial_config()

Make ahci_save_initial_config() a bit more generic by introducing
force_port_map and mask_port_map arguments.

Move PCI stuff into ahci_pci_save_initial_config().

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agoahci: Get rid of host->iomap usage
Anton Vorontsov [Wed, 3 Mar 2010 17:17:34 +0000 (20:17 +0300)]
ahci: Get rid of host->iomap usage

Currently the driver uses host->iomap to store all the iomapped BARs
of a PCI device (while AHCI devices actually use just a single memory

We're going to teach AHCI to work with non-PCI buses, so there are two
options to make this work:

1. "fake" host->iomap array for non-PCI devices, and place the needed
   address at iomap[AHCI_PCI_BAR];
2. Get rid of host->iomap usage, instead introduce a private mmio

This patch implements the second option.

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agoahci: pp->active_link is not reliable when FBS is enabled
Shane Huang [Tue, 16 Mar 2010 10:08:55 +0000 (18:08 +0800)]
ahci: pp->active_link is not reliable when FBS is enabled

pp->active_link is not reliable when FBS is enabled.
Both PORT_SCR_ACT and PORT_CMD_ISSUE should be checked
because mixed NCQ and non-NCQ commands may be in flight.

Signed-off-by: Shane Huang <shane.huang@amd.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agolibata: make functions/variables static
Robert Hancock [Wed, 20 Jan 2010 05:03:39 +0000 (23:03 -0600)]
libata: make functions/variables static

Make some variables in ahci and a function in pata_pcmcia static, as found
using sparse.

Signed-off-by: Robert Hancock <hancockrwd@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agoahci: Implement SATA AHCI FIS-based switching support
Shane Huang [Wed, 9 Dec 2009 09:23:04 +0000 (17:23 +0800)]
ahci: Implement SATA AHCI FIS-based switching support

Tested on AMD internal reference board.

Signed-off-by: Shane Huang <shane.huang@amd.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agoahci: disable SNotification capability for ich8
Shaohua Li [Mon, 16 Nov 2009 01:56:05 +0000 (09:56 +0800)]
ahci: disable SNotification capability for ich8

I obseved there is a sata_async_notification() for every ahci
interrupt. But the async notification does nothing (this is hard
disk drive and no pmp). This cause cpu wastes some time on sntf
register access.

It appears ICH AHCI doesn't support SNotification register, but the
controller reports it does. After quirking it, the async notification

PS. it appears all ICH don't support SNotification register from ICH
manual, don't know if we need quirk all ICH. I don't have machines
with all kinds of ICH.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agoahci: Add ifdef wrapper to ahci_gtf_filter_workaround
Markus Trippelsdorf [Fri, 9 Oct 2009 03:41:47 +0000 (05:41 +0200)]
ahci: Add ifdef wrapper to ahci_gtf_filter_workaround

Commit f80ae7e45a0e03da188494c6e947a5c8b0cdfb4a
ahci: filter FPDMA non-zero offset enable for Aspire 3810T
breaks the current git build for configurations that don't define
This adds an ifdef wrapper to ahci_gtf_filter_workaround.

Signed-off-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agoahci: filter FPDMA non-zero offset enable for Aspire 3810T
Tejun Heo [Tue, 15 Sep 2009 19:18:03 +0000 (04:18 +0900)]
ahci: filter FPDMA non-zero offset enable for Aspire 3810T

Curiously, Aspire 3810T issues many SATA feature enable commands via
_GTF, of which one is invalid and another is not supported by the
drive.  In the process, it also enables FPDMA non-zero offset.
However, the feature also needs to be supported and enabled from the
controller and it's wrong to enable it from _GTF unless the controller
can do it by default.

Currently, this ends up enabling FPDMA non-zero offset only on the
drive side leading to NCQ command failures and eventual disabling of
NCQ.  This patch makes libata filter out FPDMA non-zero offset enable
for the machine.

This was reported by Marcus Meissner in bnc#522790.


Reported-by: Marcus Meissner <meissner@novell.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agoahci: display all AHCI 1.3 HBA capability flags (v2)
Robert Hancock [Sun, 20 Sep 2009 23:02:31 +0000 (17:02 -0600)]
ahci: display all AHCI 1.3 HBA capability flags (v2)

Update the AHCI driver to display all of the HBA capabilities defined in the
AHCI 1.3 specification. Some of these are in a new CAP2 (HBA Capabilities
Extended) register which is only defined on AHCI 1.2 or later. The spec says
that undefined registers should always return 0 on read, but to be safe we
assume a value of 0 unless the controller reports AHCI version 1.2 or later.
The value can also be retrieved through sysfs as with the existing capability

For example, on an Intel Ibex Peak (PCH) controller:

ahci 0000:00:1f.2: flags: 64bit ncq sntf stag pm led clo pmp pio slum part ems
sxs apst

We don't do anything special with the new flags yet.

Also, change the code that displays the flags to use the same bit enumerations
that are used to control actual operation.

Signed-off-by: Robert Hancock <hancockrwd@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agoahci: kill @force_restart and refine CLO for ahci_kick_engine()
Shane Huang [Fri, 7 Aug 2009 07:05:52 +0000 (15:05 +0800)]
ahci: kill @force_restart and refine CLO for ahci_kick_engine()

This patch refines ahci_kick_engine() after discussion with Tejun about
FBS(FIS-based switching) support preparation:
a. Kill @force_restart and always kick the engine. The only case where
   @force_restart is zero is when it's called from ahci_p5wdh_hardreset()
   Actually at that point, BSY is pretty much guaranteed to be set.
b. If PMP is attached, ignore busy and always do CLO. (AHCI-1.3 9.2)

Signed-off-by: Shane Huang <shane.huang@amd.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agolibata: Export AHCI capabilities
Matthew Garrett [Fri, 17 Jul 2009 18:13:47 +0000 (19:13 +0100)]
libata: Export AHCI capabilities

AHCI exports various capability bits that may be of interest to userspace
such as whether the BIOS claims a port is hotpluggable or eSATA. Providing
these via sysfs along with the version of the AHCI spec implemented by
the host allows userspace to make policy decisions for things like ALPM.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years ago[libata] add DMA setup FIS auto-activate feature
Shaohua Li [Mon, 27 Jul 2009 01:24:35 +0000 (09:24 +0800)]
[libata] add DMA setup FIS auto-activate feature

Hopefully results in fewer on-the-wire FIS's and no breakage.  We'll see!

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
7 years agoENGR00124224 mx51: Update the new mfg default config according to mx5_defconfig
Peter Chen [Tue, 8 Jun 2010 12:07:01 +0000 (20:07 +0800)]
ENGR00124224 mx51: Update the new mfg default config according to mx5_defconfig

The old imx51_updater_defconfig is out-dated, so delete it,
and mfg firmware will use imx5_updater_defconfig in future.

Signed-off-by: Peter Chen <b29397@freescale.com>
7 years agoENGR00120301 MX51: Fix slow TS issue
Nancy Chen [Mon, 7 Jun 2010 15:45:08 +0000 (10:45 -0500)]
ENGR00120301 MX51: Fix slow TS issue

MX51: Fix slow TS issue.

Signed-off-by: Nancy Chen <Nancy.Chen@freescale.com>
7 years agoENGR00124168 imx28: fix bug that USB gadget driver can't work
Peter Chen [Tue, 8 Jun 2010 03:28:56 +0000 (11:28 +0800)]
ENGR00124168 imx28: fix bug that USB gadget driver can't work

1. USB gadget driver can't work if we insmod usb host module first,
At default, we should not define CONFIG_USB_EHCI_ARC_OTG
(Taking the first usb port as host), otherwise the usb gadget driver
will can't work due to its resource have already claimed by usb0 host.
2. Fixes some warnings and error format for Kconfig
and defconfig

Signed-off-by: Peter Chen <b29397@freescale.com>
7 years agoENGR00124023 v4l2 output:modify the operation of ic-bypass mode
Jason Chen [Fri, 4 Jun 2010 03:24:42 +0000 (11:24 +0800)]
ENGR00124023 v4l2 output:modify the operation of ic-bypass mode

/unit_test//unit_tests/mxc_v4l2_output.out -iw 1920 -ih 1080 -ow 1920
-oh 1080 -d 3 -l 600 -fr 80 /dev/shm/1920_1080_E18_c2.yuv

Error information:
It report the following error:
video4linux video16: unable to update buffer 0 address rc=-13

Fix above issue by modify the operation of ic-bypass mode.

Signed-off-by: Jason Chen <b02280@freescale.com>
7 years agoENGR00124167 [mfg-tool]: Enhancement the UTP protocol
Peter Chen [Fri, 4 Jun 2010 13:18:38 +0000 (21:18 +0800)]
ENGR00124167 [mfg-tool]: Enhancement the UTP protocol

Enhancement the UTP protocol and give more comments
for source code

Signed-off-by: Peter Chen <b29397@freescale.com>
7 years agoENGR00124103 TVE fb driver probe for mode incorrectly
Xinyu Chen [Mon, 7 Jun 2010 02:21:38 +0000 (10:21 +0800)]
ENGR00124103 TVE fb driver probe for mode incorrectly

When passing di1_primay to cmdline on BBG, the modes
matching in TVE probe are successful with fb0's modes.
This causes mode match, enable TVE and do set var to fb0.
Then the fbcon var is set incorrect, fb0 var works incorrect
after resume from suspend when doing fbcon switch.
Just simply add TVE's own mode list to resolve.

Signed-off-by: Xinyu Chen <xinyu.chen@freescale.com>
7 years agoENGR00124105 IPUv3 FB:Support 12-bpp pixel format correctly in sys-fs
Liu Ying [Fri, 4 Jun 2010 17:21:41 +0000 (13:21 -0400)]
ENGR00124105 IPUv3 FB:Support 12-bpp pixel format correctly in sys-fs

mxcfb_check_var() function will set framebuffer's bpp value to be 16,
when the framebuffer's true bpp value is 12. This makes the bpp value in
sys-fs wrong.

Signed-off-by: Liu Ying <b17645@freescale.com>
7 years agoENGR00123731 ipuv3: change the submodules enable timming
Jason Chen [Wed, 2 Jun 2010 08:01:13 +0000 (16:01 +0800)]
ENGR00123731 ipuv3: change the submodules enable timming

change the ipu submodules enable timming.

Signed-off-by: Jason Chen <b02280@freescale.com>
7 years agoENGR00123589: Console can wakeup system while it is not wakeup source
Jeremy.Yao [Tue, 1 Jun 2010 08:44:21 +0000 (16:44 +0800)]
ENGR00123589: Console can wakeup system while it is not wakeup source

Add debug-uart pm function

Signed-off-by: Jeremy Yao <r65161@freescale.com>
7 years agoENGR00122599 MX53: DVFS-core support
Shen Yong [Fri, 16 Apr 2010 03:46:20 +0000 (11:46 +0800)]
ENGR00122599 MX53: DVFS-core support

add dvfs-core platform device and device data

Signed-off-by: Shen Yong <b00984@freescale.com>
7 years agoENGR00123982: Fix GPU clocking scheme on MX51 & MX53.
Ranjani Vaidyanathan-RA5478 [Fri, 28 May 2010 17:57:26 +0000 (12:57 -0500)]
ENGR00123982: Fix GPU clocking scheme on MX51 & MX53.

Add appropriate secondary clocks to GPU2d and GPU3d clocks.
Also fix the naming scheme for the clocks needed by GPU so
that the same clocks can be used on both MX51 and MX53.
Fixed the round_rate function for many clocks so that the rate
returned is not greater than the maximum rate supported for that clock.

Signed-off-by: Ranjani Vaidyanathan-RA5478 <Ranjani.Vaidyanathan@freescale.com>
7 years agoENGR00123838 alsa: fall back to external ram if iram fails
Alan Tull [Tue, 25 May 2010 21:44:26 +0000 (16:44 -0500)]
ENGR00123838 alsa: fall back to external ram if iram fails

If iram_alloc fails, should warn and fall back to external ram.

Signed-off-by: Alan Tull <r80115@freescale.com>
7 years agoENGR00123863 MX53 Clock:Set AXI_B_CLK to be 200MHz
Liu Ying [Wed, 26 May 2010 22:00:45 +0000 (18:00 -0400)]
ENGR00123863 MX53 Clock:Set AXI_B_CLK to be 200MHz

Set AXI_B_CLK to be 200MHz in case it is set to other clock rate in bootloader.
This can make sure IPU has sufficient clock rate.

Signed-off-by: Liu Ying <b17645@freescale.com>
7 years agoENGR00123828 Improved timing for the i.MX23/i.MX28 NAND Flash driver
Patrick Turley [Fri, 7 May 2010 22:50:18 +0000 (17:50 -0500)]
ENGR00123828 Improved timing for the i.MX23/i.MX28 NAND Flash driver

Added improved timing calculations to the driver, and various other
support work.

Signed-off-by: Patrick Turley <patrick.turley@freescale.com>
7 years agoENGR00123808 MX28: Add dual FEC 1588 timer support
Xiexiaobo [Wed, 26 May 2010 11:13:29 +0000 (19:13 +0800)]
ENGR00123808 MX28: Add dual FEC 1588 timer support

The ENET 1588 driver only support one FEC port,
this patch add dual FEC 1588 functions support.

Signed-off-by: Xie Xiaobo <X.Xie@freescale.com>
7 years agoENGR00123841 v4l2 output: crop issue for ic bypass mode
Jason Chen [Wed, 26 May 2010 04:40:01 +0000 (12:40 +0800)]
ENGR00123841 v4l2 output: crop issue for ic bypass mode

If set crop for input image, and ic bypass mode enabled (after crop,
input and output video have the same size), video will play uncorrect if
input is non-interleaving format. Need set u/v offset for this case.

Signed-off-by: Jason Chen <b02280@freescale.com>
7 years agoENGR00123807 v4l2 output: fix ic_bypass case for 720p playback
Jason Chen [Tue, 25 May 2010 11:00:48 +0000 (19:00 +0800)]
ENGR00123807 v4l2 output: fix ic_bypass case for 720p playback

If play a video by using ic bypass mode, and the resolution is
bigger than IC limitation(1024x720), will come out DQUEUE timeout issue.

Signed-off-by: Jason Chen <b02280@freescale.com>
7 years agoENGR00123717 V4L2 capture:Support more YUV pixel formats
Liu Ying [Mon, 24 May 2010 18:56:27 +0000 (14:56 -0400)]
ENGR00123717 V4L2 capture:Support more YUV pixel formats

1) Support YUYV for CSI->MEM capture.
2) Support YUYV and UYVY for CSI->PRP_ENC->MEM capture.
3) Support YUYV for still capture.

Signed-off-by: Liu Ying <b17645@freescale.com>
7 years agoENGR00123770 sgtl5000: use rate_mode to support more sample rates
Alan Tull [Mon, 17 May 2010 21:10:39 +0000 (16:10 -0500)]
ENGR00123770 sgtl5000: use rate_mode to support more sample rates

Use the RATE_MODE field to support more sample rates in hardware,
specifically 8000, 11025, 16000, and 22050 Hz.

Signed-off-by: Alan Tull <r80115@freescale.com>
7 years agoENGR00123744-3 iMX23 enable unique id support
Frank Li [Mon, 24 May 2010 09:27:34 +0000 (17:27 +0800)]
ENGR00123744-3 iMX23 enable unique id support

Add config file

Signed-off-by: Frank Li <Frank.Li@freescale.com>
7 years agoENGR00123744-2 iMX23 enable unique id support
Frank Li [Mon, 24 May 2010 09:27:34 +0000 (17:27 +0800)]
ENGR00123744-2 iMX23 enable unique id support

Add opt uuid support

Signed-off-by: Frank Li <Frank.Li@freescale.com>
7 years agoENGR00123744-1 iMX23 add unique ID support
Frank Li [Mon, 24 May 2010 08:54:32 +0000 (16:54 +0800)]
ENGR00123744-1 iMX23 add unique ID support

copy file from stmp plat

Signed-off-by: Frank Li <Frank.Li@freescale.com>
7 years agoENGR00123741 MX28: Add IOCTL commands to L2 switch driver
Niu Xule [Mon, 24 May 2010 07:49:58 +0000 (15:49 +0800)]
ENGR00123741 MX28: Add IOCTL commands to L2 switch driver

L2 switch driver command line controlling support

Signed-off-by: Niu Xule <b23300@freescale.com>
7 years agoENGR00123734 fix PxP hang issue when streamon and streamoff frequently
Robby Cai [Mon, 24 May 2010 02:40:01 +0000 (10:40 +0800)]
ENGR00123734 fix PxP hang issue when streamon and streamoff frequently

There might be some corners case which causes PxP hang up.
This is just a workaround to fix this.

Signed-off-by: Robby Cai <R63905@freescale.com>
7 years agoENGR00123686 mx5: Fix GPU_2D clock select bitmask
Dinh Nguyen [Thu, 20 May 2010 16:27:28 +0000 (11:27 -0500)]
ENGR00123686  mx5: Fix GPU_2D clock select bitmask

Because of a documentation error, the bitmask for gpu_2d clock select
is defined incorrectly.

Signed-off-by: Dinh Nguyen <Dinh.Nguyen@freescale.com>
7 years agoENGR00122307 MX23 ALSA: Speaker is muted after playback
Lionel Xu [Thu, 20 May 2010 08:04:00 +0000 (16:04 +0800)]
ENGR00122307 MX23 ALSA: Speaker is muted after playback

Speak should not be muted after playback

Signed-off-by: Lionel Xu <r63889@freescale.com>
7 years agoENGR00123436 Clear GPMI LOCK_CS to lower suspend current.
Frank Li [Tue, 18 May 2010 12:00:30 +0000 (20:00 +0800)]
ENGR00123436 Clear GPMI LOCK_CS to lower suspend current.

GPMI LOCK_CS set to low can lower suspend current about 5mA.

Signed-off-by: Frank Li <Frank.Li@freescale.com>
Signed-off-by: Turley Patrick <b18601@freescale.com>
7 years agoENGR00123514 MX28: System hangs when set different cpu freq continuously
Nancy Chen [Wed, 19 May 2010 20:04:59 +0000 (15:04 -0500)]
ENGR00123514 MX28: System hangs when set different cpu freq continuously

MX28: System hangs when set different cpu freq continuously.

Signed-off-by: Nancy Chen <Nancy.Chen@freescale.com>
7 years agoENGR00123438: iMX23 Mount as usb device then go to standby, system halt
Jeremy.Yao [Thu, 20 May 2010 05:08:44 +0000 (13:08 +0800)]
ENGR00123438: iMX23 Mount as usb device then go to standby, system halt

Fix system timer suspend/resume operation in mx23 pm module

Signed-off-by: Jeremy Yao <r65161@freescale.com>
7 years agoENGR00123454 [MX23] FIX LCD no display issue after switch back from TVout
Robby Cai [Wed, 19 May 2010 14:12:05 +0000 (22:12 +0800)]
ENGR00123454 [MX23] FIX LCD no display issue after switch back from TVout

Due to the LCDIF clock caculation problem.

Signed-off-by: Robby Cai <R63905@freescale.com>
7 years agoENGR00123554 mx233/mx28: Enlarge the ramdisk size
Peter Chen [Wed, 19 May 2010 07:33:58 +0000 (15:33 +0800)]
ENGR00123554 mx233/mx28: Enlarge the ramdisk size

The WindowsXP can only recognize the ramdisk size above 8MiB.

Signed-off-by: Peter Chen <b29397@freescale.com>
7 years agoENGR00123491 i.mx53 arc usb host increase tx fifo threshold.
Jun Li [Wed, 19 May 2010 05:20:59 +0000 (13:20 +0800)]
ENGR00123491 i.mx53 arc usb host increase tx fifo threshold.

There are bad turnaround erros when copying data from SD to Udisk
via Host port in i.mx53. Tunning TX fifo fill threshold from 2 to 8.

FIFO Burst Threshold:(Read/Write) [Default = 2] This register controls
the number of data bursts that are posted to the TX latency FIFO in host
mode before the packet begins on to the bus. The minimum value is 2 and
this value should be a low as possible to maximize USB performance. A
higher value can be used in systems with unpredictable latency and/or
insufficient bandwidth where the FIFO may underrun because the data
transferred from the latency FIFO to USB occurs before it can be
replenished from system memory. This value is ignored if the Stream
Disable bit in USBMODE register is set.

Signed-off-by: Li Jun <r65092@freescale.com>
7 years agoENGR00123441 imx23, fix ethernet can not suspend
Zhou Jingyu [Wed, 19 May 2010 02:58:13 +0000 (10:58 +0800)]
ENGR00123441 imx23, fix ethernet can not suspend

fix ethernet can not enter suspend state

Signed-off-by: Zhou Jingyu <Jingyu.Zhou@freescale.com>
7 years agoENGR00123412: iMX23 Copy data from SD to NAND Kernel Warning
Jeremy.Yao [Wed, 19 May 2010 02:42:26 +0000 (10:42 +0800)]
ENGR00123412: iMX23 Copy data from SD to NAND Kernel Warning

Remove unnecessary spin lock.

Signed-off-by: Jeremy Yao <r65161@freescale.com>
7 years agoENGR00123446 Can not get ip address after ifconfig eth0 down
Jason Liu [Tue, 18 May 2010 05:05:44 +0000 (13:05 +0800)]
ENGR00123446 Can not get ip address after ifconfig eth0 down

-Fix the fec_enet_mdio_write function with correct FEC_MMFR_OP_WRITE
-Correct comments "start a read op" inside the fec_enet_mdio_write function
-Need do fec_restart in fec open since we do fec_stop in fec close
-Need do phy_stop in fec close to set the correct phy state machine

Signed-off-by:Jason Liu <r64343@freescale.com>

7 years agoENGR00123389: iMX23 system fail to detect sd remove
Jeremy.Yao [Tue, 18 May 2010 07:46:29 +0000 (15:46 +0800)]
ENGR00123389: iMX23 system fail to detect sd remove

Fix the problem the card is still connected while status shows it's removed

Signed-off-by: Jeremy Yao <r65161@freescale.com>
7 years agoENGR00123559 Android lowmemorykiller judges the number of free pages incorrect.
Xinyu Chen [Tue, 18 May 2010 06:51:18 +0000 (14:51 +0800)]
ENGR00123559 Android lowmemorykiller judges the number of free pages incorrect.

low memory killer has problem to judge the current free pages in system.
It gets total reclaimable anon cache and free pages as it's judging data.
But on our platforms, we have DMA zone there, and normal process can not
allocate memory from DMA. So here, just remove the DMA's
free/anon pages from this data. Let low memory killer works well.

Signed-off-by: Xinyu Chen <xinyu.chen@freescale.com>
7 years agoENGR00123558 v4l2 output: add own workqueue for v4l output
Jason Chen [Tue, 18 May 2010 06:58:01 +0000 (14:58 +0800)]
ENGR00123558 v4l2 output: add own workqueue for v4l output

During v4l2 output playback, there is performance issue after below

UBUNTU: SAUCE: (upstream) netdev/fec: fix performance impact from
mdio poll operation

Use common workqueue cause this issue, so create an own workqueue for
v4l2 output.

Signed-off-by: Jason Chen <b02280@freescale.com>
7 years agonetdev/fec: fix ifconfig eth0 down hang issue
Bryan Wu [Tue, 18 May 2010 04:56:55 +0000 (12:56 +0800)]
netdev/fec: fix ifconfig eth0 down hang issue

BugLink: http://bugs.launchpad.net/bugs/559065
In fec open/close function, we need to use phy_connect and phy_disconnect
operation before we start/stop phy. Otherwise it will cause system hang.

Only call fec_enet_mii_probe() in open function, because the first open
action will cause NULL pointer error.

Signed-off-by: Bryan Wu <bryan.wu@canonical.com>
Acked-by: Tim Gardner <tim.gardner@canonical.com>
Acked-by: Colin King <colin.king@canonical.com>
Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
7 years agoENGR00123527 fix iMX23 udhcpc fail to get ip address
Frank Li [Mon, 17 May 2010 10:43:18 +0000 (18:43 +0800)]
ENGR00123527 fix iMX23 udhcpc fail to get ip address

Miss CONFIG_PACKET at default config

Signed-off-by: Frank Li <Frank.Li@freescale.com>
7 years agoENGR00122965 Improve USB100 feature due to bootloader change.
Frank Li [Thu, 13 May 2010 04:17:36 +0000 (12:17 +0800)]
ENGR00122965 Improve USB100 feature due to bootloader change.

Smalle change due to the bootloader change and fix Mx23 fail
caused by clock module change.

Signed-off-by: Frank Li <Frank.Li@freescale.com>
7 years agoENGR00123512 ipu: fix ipu lib compile error
Jason Chen [Mon, 17 May 2010 07:02:06 +0000 (15:02 +0800)]
ENGR00123512 ipu: fix ipu lib compile error

Fix ipu lib compile error after 4-stripes resize patch pushed.

Signed-off-by: Jason Chen <b02280@freescale.com>
7 years agoENGR00123507 IPUv3:Set suitable IDMAC burst size for UYVY and ABGR32
Liu Ying [Mon, 17 May 2010 14:45:45 +0000 (10:45 -0400)]
ENGR00123507 IPUv3:Set suitable IDMAC burst size for UYVY and ABGR32

1)Set the IDMAC channel's burst size to be 32 pixels if the channel is
in UYVY pixel format.
2)Set the IDMAC channel's burst size to be 16 pixels if the channel is
in ABGR32 pixel format.

Signed-off-by: Liu Ying <b17645@freescale.com>
7 years agoENGR00123294 MX53: 4 stripes algorithm for support resizing for big screen
mark gutman [Wed, 5 May 2010 18:16:44 +0000 (21:16 +0300)]
ENGR00123294  MX53: 4 stripes algorithm for support resizing for big screen

The updated algorithm allows output frame size up to 2048x2048

Signed-off-by: Mark Gutman <Mark.Gutman@freescale.com>
7 years agoENGR00123439 DMFC size: IPU DMFC DP high resulution fail to display small frame
Ran Ferderber [Thu, 13 May 2010 07:24:40 +0000 (10:24 +0300)]
ENGR00123439 DMFC size: IPU DMFC DP high resulution fail to display small frame

update the DMFC size to avoid usage of WAIT4EOT when possible

Signed-off-by: Ran Ferderber <r53561@freescale.com>
7 years agophylib: Move workqueue initialization to a proper place
Anton Vorontsov [Mon, 18 Jan 2010 05:37:16 +0000 (05:37 +0000)]
phylib: Move workqueue initialization to a proper place

commit 541cd3ee00a4fe975b22fac6a3bc846bacef37f7 ("phylib: Fix deadlock
on resume") caused TI DaVinci EMAC ethernet driver to oops upon resume:

 PM: resume of devices complete after 237.098 msecs
 Restarting tasks ... done.
 kernel BUG at kernel/workqueue.c:354!
 Unable to handle kernel NULL pointer dereference at virtual address 00000000
 [<c002c598>] (__bug+0x0/0x2c) from [<c0052a54>] (queue_delayed_work_on+0x74/0xf8)
 [<c00529e0>] (queue_delayed_work_on+0x0/0xf8) from [<c0052b30>] (queue_delayed_work+0x2c/0x30)

The oops pops up because TI DaVinci EMAC driver detaches PHY on
suspend and attaches it back on resume. Attaching makes phylib call
phy_start_machine() that initializes a workqueue. On the other hand,
PHY's resume routine will call phy_start_machine() again, and that
will cause the oops since we just destroyed the already scheduled

This patch fixes the issue by moving workqueue initialization to

p.s. We don't see this oops with ucc_geth and gianfar drivers because
they perform a fine-grained suspend, i.e. they just stop the PHYs
without detaching.

Reported-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Tested-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agophylib: Properly reinitialize PHYs after hibernation
Anton Vorontsov [Wed, 30 Dec 2009 08:23:30 +0000 (08:23 +0000)]
phylib: Properly reinitialize PHYs after hibernation

Since hibernation assumes power loss, we should fully reinitialize
PHYs (including platform fixups), as if PHYs were just attached.

This patch factors phy_init_hw() out of phy_attach_direct(), then
converts mdio_bus to dev_pm_ops and adds an appropriate restore()

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agophylib: Fix deadlock on resume
Anton Vorontsov [Wed, 30 Dec 2009 08:23:28 +0000 (08:23 +0000)]
phylib: Fix deadlock on resume

Sometimes kernel hangs on resume with the following trace:

 ucc_geth e0102000.ucc: resume
 INFO: task bash:1764 blocked for more than 120 seconds.
 "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
 bash          D 0fecf43c     0  1764   1763 0x00000000
 Call Trace:
 [cf9a7c10] [c0012868] ret_from_except+0x0/0x14 (unreliable)
 --- Exception: cf9a7ce0 at __switch_to+0x4c/0x6c
     LR = 0xcf9a7cc0
 [cf9a7cd0] [c0008c14] __switch_to+0x4c/0x6c (unreliable)
 [cf9a7ce0] [c028bcfc] schedule+0x158/0x260
 [cf9a7d10] [c028c720] __mutex_lock_slowpath+0x80/0xd8
 [cf9a7d40] [c01cf388] phy_stop+0x20/0x70
 [cf9a7d50] [c01d514c] ugeth_resume+0x6c/0x13c

Here is why.

On suspend:

- PM core starts suspending devices, ucc_geth_suspend gets called;

- ucc_geth calls phy_stop() on suspend. Note that phy_stop() is
  mostly asynchronous so it doesn't block ucc_geth's suspend routine,
  it just sets PHY_HALTED state and disables PHY's interrupts;

- Suddenly the state machine gets scheduled, it grabs the phydev->lock
  mutex and tries to process the PHY_HALTED state, so it calls
  phydev->adjust_link(phydev->attached_dev). In ucc_geth case
  adjust_link() calls msleep(), which reschedules the code flow back to
  PM core, which now finishes suspend and so we end up sleeping with
  phydev->lock mutex held.

On resume:

- PM core starts resuming devices (notice that nobody rescheduled
  the state machine yet, so the mutex is still held), the core calls
  ucc_geth's resume routine;

- ucc_geth_resume restarts the PHY with phy_stop()/phy_start()
  sequence, and the phy_*() calls are trying to grab the phydev->lock
  mutex. Here comes the deadlock.

This patch fixes the issue by stopping the state machine on suspend
and starting it again on resume.

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agoENGR00123489 Support vpu suspend/resume on mx53
Sammy He [Fri, 14 May 2010 17:12:21 +0000 (01:12 +0800)]
ENGR00123489 Support vpu suspend/resume on mx53

Support vpu suspend/resume on mx53. Don't do power gating and
firmware re-install on mx53.

Signed-off-by: Sammy He <r62914@freescale.com>
7 years agoENGR00123437 iMX28: Add nand support for updater defconfig
Peter Chen [Thu, 13 May 2010 06:33:32 +0000 (14:33 +0800)]
ENGR00123437 iMX28: Add nand support for updater defconfig

1. Add nand support for iMX28 updater defconfig, and
fix bug for compiling error for not adding cpu freq config
2. Build-in ext3 and JBD function for support sd ext3 rootfs

Signed-off-by: Peter Chen <b29397@freescale.com>
7 years agoENGR00123280 MX5x SRTC: Add ioctls for time set notification
Anish Trivedi [Fri, 7 May 2010 14:52:39 +0000 (09:52 -0500)]
ENGR00123280 MX5x SRTC: Add ioctls for time set notification

Updates based on review comments.

Signed-off-by: Anish Trivedi <anish@freescale.com>
7 years agoENGR00122216-3 Build in LDB driver for imx5 platforms
Liu Ying [Wed, 12 May 2010 23:05:23 +0000 (19:05 -0400)]
ENGR00122216-3 Build in LDB driver for imx5 platforms

Build in LDB driver for imx5 platforms.

Signed-off-by: Liu Ying <b17645@freescale.com>
7 years agoENGR00122216-2 Support LDB driver
Liu Ying [Wed, 12 May 2010 23:05:04 +0000 (19:05 -0400)]
ENGR00122216-2 Support LDB driver

This patch adds LDB driver support.

Signed-off-by: Liu Ying <b17645@freescale.com>
7 years agoENGR00122216-1 MX53 MSL:Change for LDB support
Liu Ying [Wed, 12 May 2010 23:04:42 +0000 (19:04 -0400)]
ENGR00122216-1 MX53 MSL:Change for LDB support

This patch includes IOMUX, clock change for LDB
and adds LDB platform device.

Signed-off-by: Liu Ying <b17645@freescale.com>
7 years agoENGR00123078-2 MX28: L2 Switch support
Niu Xule [Wed, 12 May 2010 07:47:41 +0000 (15:47 +0800)]
ENGR00123078-2 MX28: L2 Switch support

L2 Switch support

Signed-off-by: Niu Xule <b23300@freescale.com>
7 years agoENGR00123078-1 MX28: Add L2 Switch resources definition
Niu Xule [Wed, 12 May 2010 07:43:24 +0000 (15:43 +0800)]
ENGR00123078-1 MX28: Add L2 Switch resources definition

L2 Switch resources definition for MX28

Signed-off-by: Niu Xule <b23300@freescale.com>
7 years agoENGR00123400 imx53 arm2: set correct IOMUX for dvi
Jason Chen [Wed, 12 May 2010 08:42:39 +0000 (16:42 +0800)]
ENGR00123400 imx53 arm2: set correct IOMUX for dvi

Set the correct IOMUX for ARM2 board.

Signed-off-by: Jason Chen <b02280@freescale.com>
7 years agoENGR00122780 ipufb: rework display device setting method
Jason Chen [Wed, 12 May 2010 07:36:40 +0000 (15:36 +0800)]
ENGR00122780 ipufb: rework display device setting method

Rework the display command line options to not require new option for
every new display. Add mxcdi0fb/mxcdi1fb option <fmt> and to tell system
what kind of display device is going to use.

Use "di1_primary" to make DI1 the primary display (i.e. fb0).

The display selection is done using video= parameter.

        For example on imx51:

                  di             di_fmt      video_mode
        DVI:      0              RGB24      mxcdi0fb:800x600M-16@60 etc
        DVI-HDMI: 0              RGB24      mxcdi0fb:720P60
        LVDS:     0              LVDS666    mxcdi0fb:XGA
        WVGA lcd: 1              RGB565     mxcdi1fb:800x480M-16@55
        TVE:      1              YUV444     mxcdi1fb:TV-NTSC
        For example on imx53:

                  di             di_fmt      video_mode
        DVI:      0              RGB24      mxcdi0fb:800x600M-16@60 etc
        DVI-HDMI: 0              RGB24      mxcdi0fb:720P60
        WVGA lcd: 0              RGB565     mxcdi0fb:800x480M-16@55
        TVE:      1              YUV444     mxcdi1fb:TV-NTSC

If you want to set NTSC tv as primary display, and make second display
DI0 as DVI 1024x768 resolution, just add

"di1_primary video=mxcdi0fb:RGB24,1024x768M-16@60 video=mxcdi1fb:YUV444,TV-NTSC"

to cmdline.

Signed-off-by: Rob Herring <r.herring@freescale.com>
Signed-off-by: Jason Chen <b02280@freescale.com>
7 years agoENGR00123338 MX23: MTD block device support
Aisheng.Dong [Wed, 12 May 2010 06:50:09 +0000 (14:50 +0800)]
ENGR00123338 MX23: MTD block device support

Add MTD_BLOCK imx23 configuration

Signed-off-by: Aisheng.Dong <b29396@freescale.com>
7 years agoENGR00122962 ipu: don't set axi id to 1 on mx53
Rob Herring [Tue, 27 Apr 2010 22:15:07 +0000 (17:15 -0500)]
ENGR00122962 ipu: don't set axi id to 1 on mx53

Setting the AXI ID on to 1 for sync display channels on MX53 causes
display flickering when system is heavily loaded.

Signed-off-by: Rob Herring <r.herring@freescale.com>
(cherry picked from commit 30df5734fcea4146b60f35238aaf5368f3392b0d)

7 years ago ENGR00123325-2 mx5x: Change default DMA zone to 96M
Sammy He [Tue, 11 May 2010 14:11:39 +0000 (22:11 +0800)]
 ENGR00123325-2 mx5x: Change default DMA zone to 96M

Change default DMA zone to 96M on mx5x

Signed-off-by: Sammy He <r62914@freescale.com>
7 years agoENGR00123325-1 Increase DMA zone to 96M on mx5x
Sammy He [Tue, 11 May 2010 14:08:02 +0000 (22:08 +0800)]
ENGR00123325-1 Increase DMA zone to 96M on mx5x

Increase DMA zone to 96M to fix memory not enough issue with 1080p
video playback.

Signed-off-by: Sammy He <r62914@freescale.com>
7 years agoENGR00123296 MX5: Fix FEC compiling error
Xie Xiaobo [Mon, 10 May 2010 03:31:21 +0000 (11:31 +0800)]
ENGR00123296 MX5: Fix FEC compiling error

1. One variable declaration was controlled by ARCH_MXS, but it was
used by other platforms, So remove the ARCH_MXS flags controlling.
2. Added judgement for FEC 1588 interrupt enablement.

Signed-off-by: Xie Xiaobo <X.Xie@freescale.com>
7 years agoENGR00122832-2 MX28: Support for IEEE 1588 interface functionality
Xiexiaobo [Sun, 9 May 2010 08:27:06 +0000 (16:27 +0800)]
ENGR00122832-2 MX28: Support for IEEE 1588 interface functionality

Supply the 1588 support for MX28 FEC.

Signed-off-by: Xie Xiaobo <X.Xie@freescale.com>
7 years agoENGR00122832-1 MX28: Configure clock source for FEC 1588 timer
Xie Xiaobo [Thu, 6 May 2010 11:02:11 +0000 (19:02 +0800)]
ENGR00122832-1 MX28: Configure clock source for FEC 1588 timer

Select clock for Ethernet 1588 timer and set the divider.

Signed-off-by: Xie Xiaobo <X.Xie@freescale.com>
7 years agoENGR00123293 iMX23 fix ssp clock is too slow
Frank Li [Sat, 8 May 2010 03:47:40 +0000 (11:47 +0800)]
ENGR00123293 iMX23 fix ssp clock is too slow

bypass bit set to wrong possition. So ssp alays use xtal.
mmc use ssp module divider instead of ssp clock source

Signed-off-by: Frank Li <Frank.Li@freescale.com>
7 years agoENGR00123283 MX23: clock driver causes ssp_clk change timeout out of suspend
Robert Lee [Fri, 7 May 2010 20:26:59 +0000 (15:26 -0500)]
ENGR00123283 MX23: clock driver causes ssp_clk change timeout out of suspend

If you boot up and then go into suspend mode, the clk_busy_wait()
 function causes a timeout in the ssp_set_rate function because
 the rate is being set while the clock is disable
 resulting in the busy bit staying false.

Signed-off-by: Robert Lee <robert.lee@freescale.com>
7 years agoENGR00123222 MX23/MX28: Add power optimization functionality to cpufreq
Robert Lee [Fri, 7 May 2010 05:02:09 +0000 (00:02 -0500)]
ENGR00123222 MX23/MX28: Add power optimization functionality to cpufreq

MX23/MX28: Add power optimization functionality to cpufreq
Add working HCLK autoslow interface and functionality
 to cpufreq and clock driver.
Add lowering of x_clk for lowest clock speed state.
MX23: Fix 392MHz voltage value from 1475000 to 1450000.
  This lower value is all that is needed when using 130MHz hclk.
MX23: change emiclk from 120000 to 130910 for 360000 cpu entry.
 130910 allows more bandwidth and avoids momentary halting
 of sdram traffic required between emiclk changes.
MX23/28: fiix problem with previous hclk autoslow
 implementation causing a corruption of the CLKCTRL
 _HBUS register.  The regular hclk divider was getting
 set to 21 or 22 or 23 when the LCD was turned off.
  This very low speed starves the system.

Signed-off-by: Robert Lee <robert.lee@freescale.com>
7 years agoENGR00122684: imx23: add nand support for updater defconfig
Peter Chen [Thu, 6 May 2010 01:40:49 +0000 (09:40 +0800)]
ENGR00122684: imx23: add nand support for updater defconfig

imx23: add nand support for and updater defconfig

Signed-off-by: Peter Chen <b29397@freescale.com>
7 years agoENGR00123170 Preserve NAND bad block indication
Jason Liu [Thu, 6 May 2010 08:47:39 +0000 (16:47 +0800)]
ENGR00123170 Preserve NAND bad block indication

Preserve bad block indication of NAND

Signed-off-by:Jason Liu <r64343@freescale.com>

7 years agoENGR00123197 MX23: Fix CPU freq cannot be changed after LCD is off if mDDR used
Nancy Chen [Thu, 6 May 2010 16:54:34 +0000 (11:54 -0500)]
ENGR00123197 MX23: Fix CPU freq cannot be changed after LCD is off if mDDR used

MX23: Fix CPU frequency can not be changed after LCD is off if mDDR is used.

Signed-off-by: Nancy Chen <Nancy.Chen@freescale.com>
7 years agoENGR00122920: Fix failure to boot from MMC slot 0 on MX51.
Ranjani Vaidyanathan-RA5478 [Tue, 27 Apr 2010 16:04:37 +0000 (11:04 -0500)]
ENGR00122920: Fix failure to boot from MMC slot 0 on MX51.

Disabling PLL3 was causing failure to boot from MMC slot0.
Move SDHC clocks to be sourced from PLL2 instead of PLL3.

Signed-off-by: Ranjani Vaidyanathan-RA5478 <Ranjani.Vaidyanathan@freescale.com>
7 years agoENGR00122789: Fixed long-term video playback issue.
Ranjani Vaidyanathan-RA5478 [Thu, 22 Apr 2010 20:06:01 +0000 (15:06 -0500)]
ENGR00122789: Fixed long-term video playback issue.

Fixed long-term video playback issue due to incorrect increment of
clock usecount.
Fixed various bugs associated with system entering LP-APM mode.
Fixed incorrect enabling of PLL3 that was not allowing system to enter
LP-APM mode if TVE was built in.
EMI_GARB clock needs to be enabled when certain clock dividers are changed.

Signed-off-by: Ranjani Vaidyanathan-RA5478 <Ranjani.Vaidyanathan@freescale.com>
7 years ago ENGR00123128 MX25 wifi small data transfer error in DMA mode
Richard Zhu [Thu, 6 May 2010 08:03:36 +0000 (16:03 +0800)]
 ENGR00123128 MX25 wifi small data transfer error in DMA mode

 Use the DMA mode in normal, and revert to PIO mode when
 there are small data transmissions.

Signed-off-by: Richard Zhu <r65037@freescale.com>
7 years agoENGR00123168 : iMX23 SSP/MMC change clock setting
Jeremy.Yao [Thu, 6 May 2010 08:13:33 +0000 (16:13 +0800)]
ENGR00123168 : iMX23 SSP/MMC change clock setting

Change mmc clock setting to fit new clock driver

Signed-off-by: Jeremy Yao <r65161@freescale.com>
7 years agoENGR00123159: MXS AUART Unset Stick Parity Select
Jeremy.Yao [Thu, 6 May 2010 06:43:08 +0000 (14:43 +0800)]
ENGR00123159: MXS AUART Unset Stick Parity Select

Unset SPS bit, allow normal parity check

Signed-off-by: Jeremy Yao <r65161@freescale.com>
7 years agoENGR00123125 mx23: cleanup and fixes of mach-mx23/clock.c
Robert Lee [Wed, 5 May 2010 03:13:00 +0000 (22:13 -0500)]
ENGR00123125 mx23: cleanup and fixes of mach-mx23/clock.c

- add set_rate functionality for x_clk.
 Added the x_set_rate and x_round_rate functions.  A lower
 x_clk rate will be used for the lowest power 24Mhz state.
 This implementation to be added later into cpufreq but
 going ahead and providing the necessary clock driver
 functionality now.

- add enable/disable functionality to ref clocks
 Assigned the enable and disable function pointers
 of each of the *_ref_clk tables to the mx23 enable
 and disable functions.

- change lcd clock (default) parent to ref_pix
 Self explanatory.  Previously this was incorrectly
 assigned directly to the PLL which would cause
 ref_pix clock to never be gated when not being
 used which causes very small but unnecessary
 additional power usage to occur when the LCD is disabled.

- add hbus autoslow function to be used by cpufreq
 Added this function for eventual used by cpufreq driver.

- fix pll_enable() delay
This delay was changed to match delay requirements given
 in the mx23 reference manual from the definition of the
 PLLCTRL0 POWER bitfield.

- revise emi_set_rate

emi_set_rate had several small issues.
The amount of iram it was asking for was set to an
 arbitrary amount instead of being assigned to the
 size of the assembly code.  Instead, assigning it
 to the size of the assembly code.

When no changes in emiclk speed were being made, we didn't
 need to spend the time allocating iram, disabling fiqs/irqs,
 and jumping to the assembly routines  Added checks to
 minimize the time spent in emi_set_rate if then new
 clock divisors were the same as the old.

Added functionality for switching the parent between
 ref_emi_clk and ref_xtal_clk source based on the clock

- added "set_sys_dependent_parent" functionality to allow
 the reduction some of the peripheral clocks (mainly SSP
 and GPMI) during the lowest power 24MHz cpu_clk state
 was needed to allow the PLL to turn off and achieve
 minimum power usage for that state.  The
 "set_sys_dependent_parent" functionality implements this
 functionality by checking the cpu speed and changing the
 parent as needed (but only when the ref count is zero).

- other minor cleanup

In a couple of locations, I saw the following syntax being used:

> -val |= clkctrl_frac;

This was incorrect because the BM_ definition is already
 shifted to the correct bitfield location of the register.
  Shifting it over further by BP_ is a bug.  This bug
 didn't cause problems in these cases because the BP_
 value happended to be zero, but needed cleaning up anyway.

The mx23_raw_enable was being called without any delays
 to ensure the clock was ready for use.  Added a call
 to check the clock busy bit to ensure the clock is ready
 before being used.

-  Added clock_busy_wait function which replaced all the
 individual busy_wait handling and properly accounts for
 clocks with xtal source (which use a separate xtal
 sourced busy bit).

-  Cleaned up xtal and rtc clocks/sources functionality.

-  Added functionality to SSP and GPMI set_rate function
 to allow changing the parent clock to the xtal if the
 rate is a factor of 24MHz.

Signed-off-by: Robert Lee <robert.lee@freescale.com>
7 years agoENGR00123134-3 MX53 NAND defconfig support
Jason Liu [Thu, 6 May 2010 02:47:07 +0000 (10:47 +0800)]
ENGR00123134-3 MX53 NAND defconfig support

MX53 NAND defconfig support

Signed-off-by:Jason Liu <r64343@freescale.com>

7 years agoENGR00123134-2 MX53 NAND driver support
Jason Liu [Thu, 6 May 2010 02:46:14 +0000 (10:46 +0800)]
ENGR00123134-2 MX53 NAND driver support

MX53 NAND driver support

Signed-off-by:Jason Liu <r64343@freescale.com>

7 years agoENGR00123134-1 MX53 NAND mach support
Jason Liu [Thu, 6 May 2010 02:44:45 +0000 (10:44 +0800)]
ENGR00123134-1 MX53 NAND mach support

MX53 NAND mach support

Signed-off-by:Jason Liu <r64343@freescale.com>

7 years agoENGR00123131 NAND:Get resources from platform data
Jason Liu [Thu, 6 May 2010 02:43:35 +0000 (10:43 +0800)]
ENGR00123131 NAND:Get resources from platform data

Get resources from platform data

Signed-off-by:Jason Liu <r64343@freescale.com>

7 years agoENGR00122142-4 MX28:Dual ENET mach support
Jason [Thu, 22 Apr 2010 06:35:02 +0000 (14:35 +0800)]
ENGR00122142-4  MX28:Dual ENET mach support

MX28:Dual ENET mach support

Signed-off-by:Jason Liu <r64343@freescale.com>

7 years agoENGR00122142-3 fec: Add shared mdio bus for dual MAC
Jason [Thu, 22 Apr 2010 06:31:04 +0000 (14:31 +0800)]
ENGR00122142-3 fec: Add shared mdio bus for dual MAC

MX28 has two MAC,but External PHY can only be configed by MAC0
which means PHY1 should be configed by MAC0 according to IC spec.
This patch enable dual ENET support. If want to test rootfs from
eth1, can add ip=:::::eth1:dhcp in the command line

Signed-off-by: Jason Liu <r64343@freescale.com>
Signed-off-by: Rob Herring <r.herring@freescale.com>
7 years agoENGR00122142-2 imx: fix platforms using fec RMII mode
Rob Herring [Wed, 7 Apr 2010 18:26:04 +0000 (13:26 -0500)]
ENGR00122142-2 imx: fix platforms using fec RMII mode

Add FEC phy interface mode platform data for platforms using RMII.

Signed-off-by: Rob Herring <r.herring@freescale.com>
7 years agoENGR00122142-1 fec: fix RMII with phylib
Rob Herring [Wed, 7 Apr 2010 18:25:33 +0000 (13:25 -0500)]
ENGR00122142-1 fec: fix RMII with phylib

This fixes RMII support which was broken by phylib support.

Signed-off-by: Rob Herring <r.herring@freescale.com>
7 years agoUBUNTU: SAUCE: (upstream) netdev/fec: fix performance impact from mdio poll operation
Bryan Wu [Thu, 22 Apr 2010 06:12:02 +0000 (14:12 +0800)]
UBUNTU: SAUCE: (upstream) netdev/fec: fix performance impact from mdio poll operation

BugLink: http://bugs.launchpad.net/bugs/546649
BugLink: http://bugs.launchpad.net/bugs/457878
After introducing phylib supporting, users experienced performace drop. That is
because of the mdio polling operation of phylib. Use msleep to replace the busy
waiting cpu_relax() and remove the warning message.

Signed-off-by: Bryan Wu <bryan.wu@canonical.com>
Acked-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andy Whitcroft <apw@canonical.com>
7 years agoUBUNTU: SAUCE: (upstream) netdev/fec: fix phy_speed caculating
Bryan Wu [Thu, 22 Apr 2010 05:49:48 +0000 (13:49 +0800)]
UBUNTU: SAUCE: (upstream) netdev/fec: fix phy_speed caculating

BugLink: http://bugs.launchpad.net/bugs/546649
BugLink: http://bugs.launchpad.net/bugs/457878
 - remove duplicated phy_speed caculation
 - fix the phy_speed caculation according to the DataSheet

Signed-off-by: Bryan Wu <bryan.wu@canonical.com>
Acked-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andy Whitcroft <apw@canonical.com>