summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2010-10-26ENGR00132894 usb: clk_get should not be called at IRQ contextPeter Chen
The old implementation will clk_get/clk_put for each clk on/off operation. As some clk on/off operation would be IRQ context. So, it will ocurr the error like "schedule at atomic context". The new implementation will only call clk_get at usb initialization, and clk_put at usb de-initialization. The driver's clock on/off will only call clk_enable/clk_disable. Signed-off-by: Peter Chen <peter.chen@freescale.com>
2010-10-22ENGR00132877 add fixed regulator dcdc_3v15Alan Tull
The i.Mx50EVK (rdp) board has a fixed dc-dc converter. Show it as a fixed regulator so the audio driver will be able to get its voltage. Signed-off-by: Alan Tull <alan.tull@freescale.com>
2010-10-22ENGR00132875-2: MX50EVK:Add platform specific Suspend/Resume supportRanjani Vaidyanathan
EIM IOMUX settings that improve the power numbers break keypad. Move the EIM IOMUX settings to platform specific functions called during suspend/resume. Signed-off-by: Ranjani Vaidyanathan <ra5478@freescale.com>
2010-10-22ENGR00132875-1: Fix EIM IOMUX settings that break keypadRanjani Vaidyanathan
Create a platform specific power management data structure that will facilitate platform specific functions to be called during suspend/resume. Signed-off-by: Ranjani Vaidyanathan <ra5478@freescale.com>
2010-10-22ENGR00132849-2 esdhc: do not use spin_lock when set power on/offAisheng.Dong
Do not use spin_lock when sets power on/off since pmic driver may sleep on this operation. Signed-off-by: Aisheng.Dong <b29396@freescale.com>
2010-10-22ENGR00132849-1 mx50 rdp: set power_mmc to VSDAisheng.Dong
Since VSD voltage is not always on, we need to pass the correct regulator to esdhc driver to dynamically enable it. Signed-off-by: Aisheng.Dong <b29396@freescale.com>
2010-10-22ENGR00132880 Add REGULATOR_CHANGE_STATUS ops mask for mc13892 regulatorsZhou Jingyu
Add REGULATOR_CHANGE_STATUS ops mask for mc13892 regulators Signed-off-by: Zhou Jingyu <Jingyu.Zhou@freescale.com>
2010-10-22ENGR00132885-2 Reverted the commit d6c3005 (ENGR00132615-2)Robby Cai
Previous patch is not stable enough, reverted. Signed-off-by: Aisheng.Dong <b29396@freescale.com>
2010-10-22ENGR00132885-1 Reverted the commit 8281bd8 (ENGR00132615-1)Robby Cai
Previous patch is not stable enough, reverted. Signed-off-by: Aisheng.Dong <b29396@freescale.com>
2010-10-22ENGR00132878 IMX USB:autosuspending caused device can not disconnectHu hui
if enable the auto suspend for the usb device attached to the usb otg host port, after the usb device and it's roothub,hcd auto suspend, disconnecting the usb device can not be realized by the roothub, because the roothub is in the suspend state. Signed-off-by: Hu Hui <b29976@freescale.com>
2010-10-21ENGR00132874 Mx50: Donot change Vcc votlage in LPAPM modeRanjani Vaidyanathan
Vcc voltage needs to be increased before increasing the frequency when exiting from LPAPM mode. Some drivers are initiating the frequency/voltage change request (via clk_enable/clk_disable) in an ISR or tasklet context. Since the regulator API code uses SPI that can sleep during its transactions, the system will crash when sleep is called in an ISR/tasklet context. Hence the current solution is to remove the voltage change code. Signed-off-by: Ranjani Vaidyanathan <ra5478@freescale.com>
2010-10-21ENGR00132832-2 SSI: fix ssi errors caused by dual fifo sdma scriptZeng Zhaoming
MX5 dual fifo breaks mx50 ssi, because new sdma script only support mx51 and mx53. Fix it setting ssi fifo according to cpu model. Signed-off-by: Zeng Zhaoming <b32542@freescale.com>
2010-10-21ENGR00132832-1 MX50 SSI: fix ssi errors caused by mx5 dual fifo enableZeng Zhaoming
Upgrade SDMA script to support SSIs dual fifo breaks mx50 ssi, because mx50 new script not available yet. Fix it by according to cpu model to set correct sdma script address. Signed-off-by: Zeng Zhaoming <b32542@freescale.com>
2010-10-21ENGR00132844 MX5x: Fix bug of UART DMA enable macro usageXinyu Chen
In the MX5x clock code, it's incorrect to use the existence of the macro to determin the uart is DMA enabled or not. Fix this bug by checking the macro value. Signed-off-by: Xinyu Chen <xinyu.chen@freescale.com>
2010-10-21ENGR00132026 MTD: Fix MTD_MODE_RAW write failFrank Li
Kobs-ng use MTD_MODE_RAW write boot stream to NAND flash. mtdchar have not initilized all field of struct mtd_oob_ops in mtd_write when mode is MTD_MODE_RAW. function nand_do_write_ops if (ops->ooboffs && (ops->ooboffs + ops->ooblen > oobmaxlen)) return -EINVAL; ooboffs is not initilized at mtd_write. So random data cause nand_do_write_ops return fail. Signed-off-by: Frank Li <Frank.Li@freescale.com>
2010-10-20ENGR00132805 MX5 MFG: enable MAX173135 and LTC3589Lily Zhang
Enable MAX173135 and LTC3589 drivers in MX5 MFG defconfig. Signed-off-by: Lily Zhang <r58066@freescale.com>
2010-10-20ENGR00132493 RAMDISK: doing write barrier will return -EIOHu hui
when doing write barrier, the data of the bio is NULL, it's OK for ramdisk doing nothing, So the result of the bio must be success. Signed-off-by: Hu Hui <b29976@freescale.com>
2010-10-19ENGR00132153 sgtl5000: do not enforce symmetry if rate is unknownAlan Tull
If a playback and a record stream are started together, the alsa userspace may not not do the necessary ioctl calls for the audio driver stack to know the sample rate and sample size of the first stream before starting the second stream, i.e.: arecord -D plughw:0 -d 10 -f S16_LE -r 44100 -c 2 -traw | aplay -D plughw:0 -f S16_LE -r 44100 -c 2 -traw Normally we want to the driver to make sure that the second stream does not request a different playback rate/bit size from the first stream. But at that point, the driver thinks that the sample rate and sample size are 0. So in that case do not enforce constraints. This still allows for usecases where play is started first and then record is started after a delay while playback continues. Signed-off-by: Alan Tull <alan.tull@freescale.com>
2010-10-19ENGR00132649 SAHARA test causes ext3 file system crashAnish Trivedi
During kernel upgrade, replaced deprecated functions: dmac_clean_range() with clean_dcache_area() and, dmac_inv_range() with invalidate_kernel_vmap_range(), but that results in some test case failures when file system is on the SD card instead of NFS. Quite possibly the cache coherency problems were not showing up during NFS testing. Corrected this by using the proper replacements: dmac_clean_range() replaced by dmac_map_area() and, dmac_inv_range() replaced by dmac_unmap_area(). This fixes the cache coherency issues and all tests pass with file system on SD card as well mounted over NFS. Signed-off-by: Anish Trivedi <anish@freescale.com>
2010-10-19ENGR00132761 MX51: correct the debug early uart's clockXinyu Chen
As uboot changes the PLL1, this causes the UART1 clock PRE/PODF divider settings are not correct. Fix this by set uart's parent to PLL3 as uboot did. Signed-off-by: Xinyu Chen <xinyu.chen@freescale.com>
2010-10-18ENGR00131932-2 ssi, sound: Upgrade SDMA scripts to v1.1.0 versionZeng Zhaoming
New sdma script support double ssi fifo. This feature is not enabled by default in mx51 and mx53 configs. Enable double ssi fifo for audio, if underrun/overrun happens, It is possible to cuase stereo channel swapped and can't be recovered. Signed-off-by: Zeng Zhaoming <b32542@freescale.com>
2010-10-18ENGR00131932-1 MX51, MX53, sdma: Upgrade SDMA scripts to v1.1.0 versionZeng Zhaoming
Upgrade sdma mx51 script to v1.1 http://compass.freescale.net/go/220269616 Upgrade sdma mx53 script to v1.1 http://compass.freescale.net/go/211002731 These new scripts support ssi dual fifo. Pan Qihong-B30266 provide the usage: ----------- mcu_2_ssiapp and ssiapp_2_mcu are supporting SSI tx/rx residing on peripheral DMA region. mcu_2_ssish and ssish_2_mcu are supporting SSI tx/rx residing on shared peripheral DMA region. With all these 4 scripts, will support all the 3 SSIs. ---------- Signed-off-by: Zeng Zhaoming <b32542@freescale.com>
2010-10-18ENGR00132536 usb-host: Clock should be opened before calling ehci_shutdownPeter Chen
When trigger "reboot" command, the driver will call xxx_shutdown. For usb-host, it will write registers. So, it needs to open clock before ehci_shutdown if the usb is in low power mode. Signed-off-by: Peter Chen <peter.chen@freescale.com>
2010-10-18ENGR00132532 mxs-mmc: change serial clock polarityAisheng.Dong
Some mmc cards may not work properly with data change after rising edge of SCK. Changing to use falling edge of SCK to get better compatibility. Signed-off-by: Aisheng.Dong <b29396@freescale.com>
2010-10-18ENGR00132608 I.MX USB:crashed with cable unpluged during data transferHu hui
During data transfer between pc and imx board, if unplug the cable, the udc will enter low power mode and the gadget class driver is disconnected from the udc driver with the ep->desc cleared, this is done in udc irq handler which may interrupte the current udc queue operation, so when queue operation can continue it's work, the ep->desc may be NULL which will cause the NULL pointer exception in kernel. Signed-off-by: Hu Hui <b29976@freescale.com>
2010-10-18ENGR00132563 mx50: enable battery charger on mx50 rdpZhou Jingyu
enable battery charger on mx50 rdp Signed-off-by: Zhou Jingyu <Jingyu.Zhou@freescale.com>
2010-10-16ENGR00132607 MX50: Fix ePxP crash issueRobby Cai
It can be reproduced by loop-running epdc rotation test. The crash reason is the timing to disable ePxP clocks is critical, which was not met before. This issue is fixed by disabling clocks only when ePxP is inactive and no more task is on the wait queue. Signed-off-by: Robby Cai <R63905@freescale.com>
2010-10-15ENGR00124788-2 MX53 ARD: Add Linear PMIC supportNancy Chen
MX53 ARD: Add Linear PMIC support. Signed-off-by: Nancy Chen <Nancy.Chen@freescale.com>
2010-10-15ENGR00124788-1 Add Linear PMIC supportNancy Chen
Add Linear PMIC support. Signed-off-by: Nancy Chen <Nancy.Chen@freescale.com>
2010-10-15ENGR00132674: MX50: Update DDR frequency change code to use new scripts.Ranjani Vaidyanathan
Update the DDR frequency change code to latest the DDR settings for both 24Mhz and 266Mhz. 266MHz latest script version 04 from compass site: http://compass.freescale.net/doc/220496654/Codex_LPDDR2_266MHz.inc 24MHz latest script version 03 from compass site: http://compass.freescale.net/doc/219884330/Codex_LPDDR2_24MHz.inc Signed-off-by: Ranjani Vaidyanathan <ra5478@freescale.com>
2010-10-15ENGR00126326-2 - MAX17135: Add support for temperature sensor monitoringDanny Nold
- MAX17135 re-architected: - Core functionality, including I2C access to the chip, moved into drivers/mfd/max17135-core.c - MAX17135 regulators registered via init() callback in plat-specific code. - All access to registers funnelled through max17135_reg_read/write() apis. - max17135.h moved from regulator/ folder into mfd/ folder. - Support for reading internal and external temperature via sysfs interface, consistent with lm-sensors hwmon interface. Signed-off-by: Danny Nold <dannynold@freescale.com>
2010-10-15ENGR00126326-1 - MSL: Add support for MAX17135 temperature sensor driverDanny Nold
- Support added for MAX17135 temperature sensor driver for both MX50 ARM2 and RDP platforms. - MAX17135 driver re-architected with init() callback function in MSL layer which registers regulators. Signed-off-by: Danny Nold <dannynold@freescale.com>
2010-10-15ENGR00132525-3 sgtl5000: audio clock gatingAlan Tull
Turn off audio clock when possible. Empirical data says that we need to leave the clocks on for 300 mSec after all codec writes are done so schedule work to do that. Signed-off-by: Alan Tull <alan.tull@freescale.com> audio clock gating debug
2010-10-15ENGR00132525-2 sgtl5000: audio clock gatingAlan Tull
Turn off audio clock when possible. Empirical data says that we need to leave the clocks on for 300 mSec after all codec writes are done so schedule work to do that. Signed-off-by: Alan Tull <alan.tull@freescale.com>
2010-10-15ENGR00132525-1 sgtl5000: audio clock gatingAlan Tull
Turn off audio clock when possible. Empirical data says that we need to leave the clocks on for 300 mSec after all codec writes are done so schedule work to do that. This feature supported for i.Mx51 Babbage and i.Mx50 EVK (rdp) boards. Signed-off-by: Alan Tull <alan.tull@freescale.com>
2010-10-15ENGR00132634: MX50: Reduce STOP mode power consumptionRanjani Vaidyanathan
Improve STOP mode power consumption in MX50 by doing the following: 1. Disable regulators that are not being used. 2. Set as many regulators as possible into STANDBY mode. 3. Set EIM, FEC IOMUX settings for minimum power leakage. Signed-off-by: Ranjani Vaidyanathan <ra5478@freescale.com>
2010-10-15ENGR00132656 MX53 ARD: add MFG supportLily Zhang
Add MFG tool support for MX53 ARD. Remove MLB driver in updater defconfig because it is not needed by MFG tool Signed-off-by: Lily Zhang <r58066@freescale.com>
2010-10-15ENGR00132639-2 Power: mxc_pwerkey break mx25 kernel buildHu hui
Driver Part define a new platform data struct, and Put the pmic callback register into platform data structure, then we can cover all pmic Signed-off-by: Hu Hui <b29976@freescale.com>
2010-10-15ENGR00132639-1 Power: mxc_pwerkey break mx25 kernel buildHu hui
MSL Part define a new platform data struct, and Put the pmic callback register into platform data structure, then we can cover all pmic Signed-off-by: Hu Hui <b29976@freescale.com>
2010-10-15ENGR00131580-2 PTP: Support IEEE 1588 interface functionality for MX5Xie Xiaobo
IEEE1588 implementation is a Hardware-assisted method in i.MX5, This supply the 1588 support for MX53 FEC. Signed-off-by: Xie Xiaobo <X.Xie@freescale.com>
2010-10-15ENGR00131580-1 MX53: Add IEEE1588 hardware-assisted device definitionXie Xiaobo
Add IEEE1588 device's resource definition, and set default clock input for ptp RTC. Signed-off-by: Xie Xiaobo <X.Xie@freescale.com>
2010-10-15ENGR00132653 MLB: add regulator checkLily Zhang
Add regulator check in MLB driver. Signed-off-by: Lily Zhang <r58066@freescale.com>
2010-10-15ENGR00132169 MX28 ALSA: arecord from LINE IN has no soundLionel Xu
To fix the problem of no sound when recording from line in. Signed-off-by: Lionel Xu <r63889@freescale.com>
2010-10-15ENGR00132615-2 mx50: delete AHB_HIGH_SET_POINT flag of esdhc clockAisheng.Dong
This is another work-around for esdhc crc error issue when AHB is down to 24Mhz in low power mode. But an expensive one since AHB will go back to 133Mhz and it will switch the entire LP domain frequencies whenever the SDHC is accessed With the workaround of switching esdhc clock to 20Mhz when AHB is 24Mhz, we do not need to have this flag set anymore. Signed-off-by: Aisheng.Dong <b29396@freescale.com>
2010-10-15ENGR00132615-1 esdhc: auto switch clock to 20Mhz when AHB is 24MhzAisheng.Dong
The issue we found in 35 kernel is that esdhc will easily get failed with crc error when ahb bus clock is down to 24Mhz in low power mode. There's still no root cause found yet. This is a workaround that automatically switching the esdhc clock to 20Mhz when AHB is 24 Mhz(LP mode) to prevent this issue to happen and switching back to the max_clk of esdhc when exiting LP mode. Signed-off-by: Aisheng.Dong <b29396@freescale.com>
2010-10-15ENGR00132582 MX50: Add keypad for RD boardFrank Li
Need E-Book ADD On Board. Boot Config (SW5) uses EIM_DA0~7, which are multiplexed with Keypad module. Set SW5 to 'on' to avoid the interference (it appears that some keys are always pressed) after system boot. Signed-off-by: Frank Li <Frank.Li@freescale.com>
2010-10-14ENGR00132334 - EPDC fb: Report PSEUDOCOLOR when framebuffer set to grayscaleDanny Nold
EPDC FB driver currently always reports FB_VISUAL_TRUECOLOR as the visual element of fb_fix_screeninfo. For grayscale formats, it should report FB_VISUAL_PSEUDOCOLOR. Signed-off-by: Danny Nold <dannynold@freescale.com>
2010-10-14ENGR00126942 - EPDC fb: Allow flexible specification of X acceleration memoryDanny Nold
Amount of memory allocated by EPDC FB driver (including that used for X Windows acceleration) changed from a hardcoded 13MB to an amount configurable via kernel parameter. Signed-off-by: Danny Nold <dannynold@freescale.com>
2010-10-14ENGR00132580 allow prediv by 1 for ext1 clkAlan Tull
Not allowed in the spec, but the hardware guy said it was ok. Needed because clock is shared between audio and camera. Signed-off-by: Alan Tull <alan.tull@freescale.com>
2010-10-14ENGR00132594 I.MX USB: the udc isr can't process all normal eventHu hui
the udc isr can not process all normal event, it is caused by the unconditional "goto irq_end" sentence. Signed-off-by: Hu Hui <b29976@freescale.com>
generated by cgit v1.2.3 (git 2.0.4) at 2024-09-16 16:34:18 +0000