Hi,
Request for comment.
Thanks.
Lee
[RFC 1/2] i2c-intel-mid: I2C FIFO buffer size setting
The FIFO buffer size is different with different CPU stepping
It is better to read the size from board setting
Currently, pre-define it as 32-byte; it is safe for all CPU steppi
[RFC 2/2] i2c-intel-mid: Implement software fragment
There is a problem when xfer size > FIFO buffer size
Implement software fragment in host bus driver
So that I2C slave device driver need not to do modify
diff -ruN a/drivers/i2c/busses/i2c-intel-mid.c
b/drivers/i2c/busses/i
[RFC 1/2] i2c-intel-mid: I2C FIFO buffer size setting
The FIFO buffer size is different with different CPU stepping
It is better to read the size from board setting
Currently, pre-define it as 32-byte; it is safe for all CPU stepping
diff -ruN a/drivers/i2c/busses/i2c-intel-mi
Hi,
Request for comment.
Thanks.
Lee
[RFC] langwell_otg: langwell otg set vbus again if ipc command failed
If a device (for example, hub/HID/storage) is connected when booting,
setting vbus power will fail.
This problem is langwell otg driver does not handle the error status
wh
Hi, Hao
#1 It is very strange.
1-1 re-try in langwell_otg_set_vbus() -> fail
1-2 re-try after ehci driver is ready (put in langwell_otg_set_host())
-> ok.
And I tried replacing intel_scu_ipc_simple_command() by
pmic_scu_ipc_iowrite8() to directly control PMIC, but I got the same
situation.
Hi, Hao
#1 there is a race condition between ipc driver and otg driver.
Otg driver need to init after ipc driver for there is a ipc command in otg
driver.
#2 copy that, the path is a_idle -> a_wait_vrise -> a_wait_bcon -> a_vbus_err
-> a_wait_vfall -> a_idle
Thank you so much.
Lee
-Or
The FIFO buffer size is different with different CPU stepping.
Define it as 32-byte; it is safe for all CPU stepping.
There is a problem when xfer size is greater then FIFO buffer size.
Implement software fragmentation in host bus driver.
So that I2C slave device driver need not to do modify.
Sig
There is a warning message as the following.
In __xlate_proc_name(), slash character is used for subdirectory.
Do not use a name which including slash character when calling
request_irq().
Original idea is from i2c-pca-platform.c.
[4.502961] WARNING: at fs/proc/generic.c:323
__xlate_proc_name+
Fix the following section mismatch warnings.
WARNING: drivers/built-in.o(.data+0xfa10): Section mismatch in reference
from the variable intel_mid_dma_pci to the function
.devinit.text:intel_mid_dma_probe()
The variable intel_mid_dma_pci references
the function __devinit intel_mid_dma_probe()
If th
Hello,
I had sent it to LKML and CC to Dan.
https://lkml.org/lkml/2010/11/30/35
Sincerely,
Major Lee
-Original Message-
From: Koul, Vinod [mailto:vinod.k...@intel.com]
Sent: Tuesday, November 30, 2010 3:58 PM
To: Major Lee/WHQ/Wistron; meego-ker...@meego.com
Subject: RE: [PATCH] inte
There is a possible that lost the last word of a transaction if data is
not ready.
Read again in poll_transfer() to solve this issue when poll_mode is
enabled.
Patch is against mainline (2.6.37-rc5).
Verified on SPI touch screen device.
Signed-off-by: Major Lee
---
dw_spi.c |6 ++
1 fil
Sorry, my fault.
Patch is against AC tree (linux-2.6-mid-ref, 2.6.37-rc5).
Correct it.
Thanks.
Lee
-Original Message-
From: meego-kernel-boun...@lists.meego.com
[mailto:meego-kernel-boun...@lists.meego.com] On Behalf Of
major_...@wistron.com
Sent: Friday, December 10, 2010 10:38 AM
To:
> > +static int ak8974_detect(struct i2c_client *client,
> > + struct i2c_board_info *info)
> First, I think you need to define address_list in i2c_driver.
Otherwise, i2c core will not do auto detection if address_list is not
defined.
>
> And we provide i2c_board_info on Medfiel
Fix the following section mismatch warning.
WARNING: drivers/rtc/built-in.o(.data+0xa0): Section mismatch in
reference from the variable vrtc_mrst_platform_driver to the function
.init.text:vrtc_mrst_platform_probe()
The variable vrtc_mrst_platform_driver references
the function __init vrtc_mrst_p
>
> > Fix the following section mismatch warning.
>
> The patch in the mail was mangled but the logic is right. Applied by
> hand
>
> Thanks
> Alan
>
I am sorry for the "bad" setting of mailing system.
I will check and take care about it.
Lee
___
Me
Hello,
I found system will always return fail when calling gpio_to_irq() in
linux-2.6.37 (kernel-adaptation-mrst-2.6.37 /
kernel-adaptation-oaktrail-2.6.37).
Could anybody confirm the following patch which I pulled?
I apply and verify this patch with my MRST board, seems fine.
Thank you.
Sincerel
Hello,
The following patch fix xfer always return 0 in linux-2.6-mid-ref tree.
To avoid warpping by mail system, I also attach it.
---
i2c-intel-mid: fix xfer always return 0
Signed-off-by: Major Lee
---
i2c-intel-mid.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff -uprN a/dr
Hello,
This problem also confuse me long time ago.
I change module_init(serial_m3110_init) to be late_initcall(serial_m3110_init)
to avoid this problem.
May I ask that if the root cause is conflict between the MRST earlyprintk
console and the real spi max3110 console, is there possible also con
follow penwell_otg, add runtime pm support for langwell_otg.
Signed-off-by: Major Lee
---
langwell_otg.c | 173
++---
1 file changed, 152 insertions(+), 21 deletions(-)
--- linux-2.6.37.bak/drivers/usb/otg/langwell_otg.c 2011-03-02
10:09
Hi, Alan
For hardware design of the platform which in my hand, there is no
charging by USB.
Should I empty the code of langwell_otg_chrg_vbus() when there is
charging by AC only?
I will request support of Intel TPE to verify it.
Thank you.
Sincerely,
Major Lee
> -Original Message-
Hi, Hao
Copy that.
The same reason of "[PATCH] i2c/designware: Do not use legacy PCI power
management".
Thank you.
Sincerely,
Major Lee
> -Original Message-
> From: Wu, Hao [mailto:hao...@intel.com]
> Sent: Wednesday, March 16, 2011 11:11 AM
> To: Major Lee/WHQ/Wistron; meego-ker..
Hi, Hao
I removed pmic_battery_set_charger() in intel_mid_battery, the battery
driver, for our design.
SCU firmware will take care about battery charging/discharging.
In langwell_otg, langwell_otg_chrg_vbus() changes the value of OTG
status control register.
I never change the code of langwell_o
> Could you copy the comment here for me? Thanks.
>
Hi, Hao
The patch for i2c-designware is as attachment which I pick up from
kernel-adaptation-oaktrail.
Or, you can get it from the following link
http://meego.gitorious.org/meego-os-base/kernel-source/blobs/devel-oaktr
ail/patches/linux-2.6.3
Hello,
There are two problems relative to kernel-adaptation-mrst-2.6.37-15.1 on
MRST platform.
1. System failed when booting
The full log is as attachment (log_1.txt).
After marking off
arch/x86/platform/mrst/pmu.c
line 665: "pmu_scu_firmware_debug();",
system works.
2. Compile error
sfi-cpufreq: Temporary fix extract_freq
In my test case (MRST platform, LNW Stepping = C1),
IA32_PERF_STATUS & INTEL_MSR_RANGE == perf->states[index].control + 1,
except highest cpu freq.
Signed-off-by: Major Lee
---
--- linux-2.6.37.bak/arch/x86/kernel/cpu/cpufreq/sfi-cpufreq.c
2011-03-17 11:
I am sorry for there is a problem of mail system;
attach it.
Sincerely,
Major Lee
From: meego-kernel-boun...@lists.meego.com
[mailto:meego-kernel-boun...@lists.meego.com] On Behalf Of
major_...@wistron.com
Sent: Wednesday, March 23, 2011 4
There is no filed called "success indicator" of SFI spec.
Remove the declaration in sfi_processor.h and relative printing message
in sfi-cpufreq.c.
Signed-off-by: Major Lee
---
arch/x86/kernel/cpu/cpufreq/sfi-cpufreq.c |5 ++---
include/linux/sfi_processor.h |1 -
2 files cha
Accroding to 15ec44611904be0dcc97b84c29fbf964e5e2b36f,
"mmc: sdhci: 8-bit bus width changes", add 8-bit bus width
to probe_slot.
And hook platform_8bit_width to support 8-bit bus width.
Signed-off-by: Major Lee
---
sdhci-pci.c | 41 +
1 file changed, 41
> > static struct sdhci_ops sdhci_pci_ops = {
> > .enable_dma = sdhci_pci_enable_dma,
> >+.platform_8bit_width= sdhci_pci_8bit_width,
> > };
> >
> To enable 8bit support, I don't think the new callback
> 'platform_8bit_width' is needed.
> This is because the MMC core will enable th
According to the adapter interrupt comes from PMIC GPIO 0, and
SCU firmware will take care battery charge/discharge scenario.
This patch add adapter interrupt handler, and remove charger
relative function.
Signed-off-by: Major Lee
---
intel_mid_battery.c | 147
+++--
For the I2C adapter number to read EDID in MM platform is 0, add a new
DMI match list to solve this issue.
And, there are 3 I2C adapter numbers in Moorestown platform, using
driver_data to denote which I2C adapter number is selected.
Patch against kernel-adaptation-mrst-2.6.37.6-8.1.
Reported-by:
[PATCH v2] psb_intel_lvds: add support for MM platform to fixi2c adapter
number mismatch issue
v2: fix typo and remove unused callback
Patch against kernel-adaptation-mrst-2.6.37.6-8.1.
For the I2C adapter number to read EDID in MM platform is 0, add a new
DMI match list to solve this issue.
And
Patch against kernel-adaptation-mrst-2.6.37.6-9.2.
(Assume
linux-2.6.36-psb_intel_lvds-add-support-for-mm-to-fix-i2c-adap-num.patch
is applied.)
Pickup linux-2.6.37-gma600-dmicheck.patch from
kernel-adaptation-oaktrail.
Merge psb_intel_lvds-add-support-for-mm-to-fix-i2c-adap-num-v2.patch to
fix ty
If over battery charge current is detected, Renesas PMIC will trigger an
interrupt and reduce charger current automatically.
Do nothing when this interrupt is triggered, and log once in
pmic_battery_read_status().
Adapter detection will be done in adapter detection interrupt handler,
and will als
Add debugfs to check coulomb counter value which reads from PMIC by IPC
command.
(Bug 7312 - battery charge status shows abnormally)
Signed-off-by: Major Lee
---
--- linux-2.6.37.bak/drivers/power/intel_mid_battery.c 2011-05-12
11:57:59.840232580 +0800
+++ linux-2.6.37/drivers/power/intel_mid_b
Issue: "POWER_SUPPLY_CHARGE_NOW is greater than
POWER_SUPPLY_CHARGE_FULL" (Bug 7312).
Update coulomb counter value when user space application writes
battery_capacity_percent.
Do not need to keep updating coulomb counter value for SCU will maintain
it.
This patch also fix a memory leak problem.
Si
Patch against kernel-adaptation-mrst-2.6.37.6-70.1
(linux-2.6.37-sfi-p-states-update.patch is appiled).
cpuinfo_cur_freq is used to show current CPU frequency as obtained from
the hardware, but we can not clarify how to get this value in practice.
Remove get function of sfi_cpufreq_driver for cpui
> > --- linux-2.6.37.bak/arch/x86/kernel/cpu/cpufreq/sfi-cpufreq.c
> > 2011-06-07 17:03:02.677370904 +0800
> > +++ linux-2.6.37/arch/x86/kernel/cpu/cpufreq/sfi-cpufreq.c
> > 2011-06-09 13:24:16.238205000 +0800
> > @@ -364,7 +364,6 @@ static int sfi_cpufreq_cpu_init(struct c
> > if (result)
> >
There is a problem when doing "cat
/sys/class/power_supply/pmic-adap/uevent".
This is caused by a typo; fix it.
<0>[ 72.773481] last sysfs file:
/sys/devices/platform/pmic_battery.17/power_supply/pmic-adap/uevent
<4>[ 72.773574] Pid: 893, comm: cat Tainted: GWC 2.6.37.6 #1
Acer MM 10/
(Bug 10162 - contextkit of Battery API not update the key value.)
When DCDC input line over current detecting, PMIC will change charging
current automatically.
Logging event is enough.
Signed-off-by: Major Lee
---
--- linux-2.6.37.bak/drivers/power/intel_pmic_battery.c 2011-07-21
14:13:32.945077
(Bug 6017 - [PnP]Battery Status Update slowly.)
Notify property is changed.
Signed-off-by: Major Lee
---
--- linux-2.6.37.bak/drivers/power/intel_pmic_battery.c 2011-07-28
08:54:17.453125000 +0800
+++ linux-2.6.37/drivers/power/intel_pmic_battery.c 2011-07-28
08:55:01.054497551 +0800
@@ -109
41 matches
Mail list logo
