[meta-ti] Kernel Hang I2C1??
Murphy, Dan
dmurphy at ti.com
Thu May 1 14:45:05 PDT 2014
That is so awesome.
I love it when a plan comes together.
Dan
-----------------------------------------------------
Dan Murphy
Texas Instruments Inc.
________________________________________
From: Vanderpool Josh [Joshua.Vanderpool at raymondcorp.com]
Sent: Thursday, May 01, 2014 3:00 PM
To: Murphy, Dan; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
It was input voltage droop, which caused about 2ms of droop on the output of the PMIC...
I now have a successful boot. http://pastebin.com/wM3ArEnX
On to the next task...
________________________________________
From: Murphy, Dan [dmurphy at ti.com]
Sent: Thursday, May 01, 2014 9:19 AM
To: Vanderpool Josh; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
Josh
I am just wondering if there is a CPU freq voltage mismatch here.
Probably should verify the frequency and voltage to the CPU.
Dan
-----------------------------------------------------
Dan Murphy
Texas Instruments Inc.
________________________________________
From: Vanderpool Josh [Joshua.Vanderpool at raymondcorp.com]
Sent: Thursday, May 01, 2014 7:56 AM
To: Murphy, Dan; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
The CCCC is due to me having to power cycle to stop the minicom capture as ttyUSB0 goes away when the hang occurs.
I did remove the wake and sleep stuff in the dts for I2C1
Other than following the guide to disable the PM stuff I did not do anything else in the menuconfig.
The memory is the same as on the BBB and the layout is identical, I did scope the clock and LDQS and the signals look great
-the SDK6.0 U-boot had a memtest and that ran for a long time with out issue.
In u-boot i setup voltages as if it were a BBB. my device is a 800MHz device so i do not go turbo it should read the speed from the device.
-pastebin-> http://pastebin.com/dL0vRvrv
>>This seems a little suspicious.
>>[ 1.469500] cpu cpu0: cpu0 regulator not ready, retry
>>[ 1.474991] platform cpufreq-cpu0.0: Driver cpufreq-cpu0 requests probe deferral
this must come from the dts?
cpus {
cpu at 0 {
cpu0-supply = <&dcdc2_reg>;
};
};
josh
________________________________________
From: Murphy, Dan [dmurphy at ti.com]
Sent: Thursday, May 01, 2014 8:26 AM
To: Vanderpool Josh; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
Josh
It appears that the device rebooted.
And it looks like you are now communicating with the regulator now.
Did you make any other changes beyond disabling PM?
Are the memory timings setup correctly?
Are you supplying enough voltage/current to the CPU?
Does your voltage and OPP match?
This seems a little suspicious.
[ 1.469500] cpu cpu0: cpu0 regulator not ready, retry
[ 1.474991] platform cpufreq-cpu0.0: Driver cpufreq-cpu0 requests probe deferral
Maybe you are setting up a frequency that the voltage does not support. For instance setting
the CPU to OPP100 with an OPP50 voltage.
Dan
-----------------------------------------------------
Dan Murphy
Texas Instruments Inc.
________________________________________
From: Vanderpool Josh [Joshua.Vanderpool at raymondcorp.com]
Sent: Thursday, May 01, 2014 6:40 AM
To: Murphy, Dan; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
I attempted disabling the PM but now i just get a hang with no dump?
My new capture->http://pastebin.com/nf5XThRV
Is there something i can do to get more messages from the kernel during boot to indicate the problem?
Thanks for the continued support.
josh
________________________________________
From: Murphy, Dan [dmurphy at ti.com]
Sent: Wednesday, April 30, 2014 6:10 PM
To: Vanderpool Josh; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
Josh
Disregard this statement I see you are doing it that way
"Plus instead of recreating the whole memory map you the idea was to inherit other dt files to build so your board just becomes a child.
You can always override the defaults in your child dt file.
At the very least you could pull in the am33xx.dtsi which contains the memory map"
-----------------------------------------------------
Dan Murphy
Texas Instruments Inc.
________________________________________
From: Murphy, Dan
Sent: Wednesday, April 30, 2014 5:01 PM
To: Vanderpool Josh; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
Josh
For my BBB the PMIC is on I2C0, I have a device sitting on I2C1 that is working with this configuration and I2C2 is open.
I looked at the tool you used which is not a TI tool it just references the TI SDK mux and pinmux files.
Which TI SDK are you using?
Heres what I have for configurations of the I2C pins for the am335x-bone-common.dtsi file.
i2c0_pins: pinmux_i2c0_pins {
pinctrl-single,pins = <
0x188 (PIN_INPUT_PULLUP | MUX_MODE0) /* i2c0_sda.i2c0_sda */
0x18c (PIN_INPUT_PULLUP | MUX_MODE0) /* i2c0_scl.i2c0_scl */
>;
};
i2c1_pins: pinmux_i2c1_pins {
pinctrl-single,pins = <
0x158 (PIN_INPUT_PULLUP | MUX_MODE2) /* spi0_d1.i2c1_sda */
0x15c (PIN_INPUT_PULLUP | MUX_MODE2) /* spi0_cs0.i2c1_scl */
>;
};
i2c2_pins: pinmux_i2c2_pins {
pinctrl-single,pins = <
0x178 (PIN_INPUT_PULLUP | MUX_MODE3) /* uart1_ctsn.i2c2_sda */
0x17c (PIN_INPUT_PULLUP | MUX_MODE3) /* uart1_rtsn.i2c2_scl */
>;
};
Then in the am335x-boneblack.dts I de-reference the I2C1
&i2c1 {
status = "okay";
pinctrl-names = "default";
pinctrl-0 = <&i2c1_pins>;
clock-frequency = <400000>;
<Place your device here>
};
Plus instead of recreating the whole memory map you the idea was to inherit other dt files to build so your board just becomes a child.
You can always override the defaults in your child dt file.
At the very least you could pull in the am33xx.dtsi which contains the memory map
Dan
-----------------------------------------------------
Dan Murphy
Texas Instruments Inc.
________________________________________
From: Vanderpool Josh [Joshua.Vanderpool at raymondcorp.com]
Sent: Wednesday, April 30, 2014 3:24 PM
To: Murphy, Dan; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
<rant/> confusion is added as the pins on the device start from i2c0 and the software starts from i2c1</rant>
So when I stated i2c1 I was speaking from the hardware standpoint.
>From pinmux.h
MUX_VAL(CONTROL_PADCONF_UART1_RXD, (IEN | OFF | MODE3 )) /* I2C1_SDA_mux2 */\
MUX_VAL(CONTROL_PADCONF_UART1_TXD, (IEN | OFF | MODE3 )) /* I2C1_SCL_mux2 */\
I have gotten the output from a nice tool for DT conversion https://github.com/jluebbe/tools
0x180 0x2b /* uart1_rxd.i2c1_sda_mux2, INPUT | MODE3 */
0x184 0x2b /* uart1_txd.i2c1_scl_mux2, INPUT | MODE3 */
-Josh Vanderpool
-----Original Message-----
From: Murphy, Dan [mailto:dmurphy at ti.com]
Sent: Wednesday, April 30, 2014 4:16 PM
To: Vanderpool Josh; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
OK I2C1 pins should be
i2c1_pins: pinmux_i2c1_pins {
pinctrl-single,pins = <
0x158 (PIN_INPUT_PULLUP | MUX_MODE2) /* uart1_ctsn.i2c2_sda*/
0x15c (PIN_INPUT_PULLUP | MUX_MODE2) /* uart1_rtsn.i2c2_scl */
>;
};
Dan
-----------------------------------------------------
Dan Murphy
Texas Instruments Inc.
________________________________________
From: Vanderpool Josh [Joshua.Vanderpool at raymondcorp.com]
Sent: Wednesday, April 30, 2014 3:13 PM
To: Murphy, Dan; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
Due to my custom board constraints I connected the PMIC and RTC etc to I2C1 at pins. I only mention BBB as I used the same PMIC and eMMC...
I also referred to the am335x-sk schematic during development.
i2c1_pins: pinmux_i2c1_pins {
pinctrl-single,pins = <
0x180 (PIN_INPUT_PULLUP | MUX_MODE3) /* uart1_rx.i2c1_sda */
0x184 (PIN_INPUT_PULLUP | MUX_MODE3) /* uart1_tx.i2c1_scl */
>;
};
-Josh Vanderpool
-----Original Message-----
From: Murphy, Dan [mailto:dmurphy at ti.com]
Sent: Wednesday, April 30, 2014 4:08 PM
To: Vanderpool Josh; Maupin, Chase; meta-ti at yoctoproject.org
Subject: RE: Kernel Hang I2C1??
I guess I am not sure how this device is connected
This device that is being connected to I2C1 on the BBB is off the P9 connector correct?
Or is it hard wired?
Dan
-----------------------------------------------------
Dan Murphy
Texas Instruments Inc.
________________________________________
From: Vanderpool Josh [Joshua.Vanderpool at raymondcorp.com]
Sent: Wednesday, April 30, 2014 3:00 PM
To: Maupin, Chase; meta-ti at yoctoproject.org; Murphy, Dan
Subject: RE: Kernel Hang I2C1??
Chase,
I followed the entry in am335x-evm to add i2c1 rather than i2c0 in the beagle bone black dts.
FYI what ever we find i will add to the post at http://e2e.ti.com/support/arm/sitara_arm/f/791/p/335136/1179969.aspx#1179969
Thanks
-josh
________________________________________
From: Maupin, Chase [chase.maupin at ti.com]
Sent: Wednesday, April 30, 2014 3:40 PM
To: Vanderpool Josh; meta-ti at yoctoproject.org; Murphy, Dan
Subject: RE: Kernel Hang I2C1??
Josh,
Do you have the I2C2 entry in your DT? There are examples in the am33xx.dtsi file.
Sincerely,
Chase Maupin
Integration Team Manager
Linux Core Product Development
e-mail: chase.maupin at ti.com
phone: (214) 567-2950
For support:
Forums - http://e2e.ti.com<http://e2e.ti.com/>
Wiki - http://processors.wiki.ti.com/index.php/Main_Page
From: meta-ti-bounces at yoctoproject.org [mailto:meta-ti-bounces at yoctoproject.org] On Behalf Of Vanderpool Josh
Sent: Wednesday, April 30, 2014 9:03 AM
To: meta-ti at yoctoproject.org
Subject: [meta-ti] Kernel Hang I2C1??
I have a custom board and expect to use i2c1 on the board but it appears to not allow this in the kernel... u-boot was able to be modified to support the hardware port i2c1. Note that that software is 1based and hardware is zero-based...
This is the dump:
* [ 1.482628] ------------[ cut here ]------------
* [ 1.487549] WARNING: CPU: 0 PID: 1 at arch/arm/mach-omap2/omap_device.c:888 omap_device_get_by_hwmod_name+0x94/0xb4()
* [ 1.498803] omap_device_get_by_hwmod_name: no omap_device for i2c1
* [ 1.505365] Modules linked in:
* [ 1.508606] CPU: 0 PID: 1 Comm: swapper Not tainted 3.12.10-ti2013.12.01 #2
* [ 1.515990] Backtrace:
* [ 1.518595] [<c0017880>] (dump_backtrace+0x0/0x10c) from [<c0017a1c>] (show_stack+0x18/0x1c)
* [ 1.527542] r6:c002f564 r5:00000009 r4:dd083e28 r3:00000000
* [ 1.533567] [<c0017a04>] (show_stack+0x0/0x1c) from [<c0580ba4>] (dump_stack+0x20/0x28)
* [ 1.542068] [<c0580b84>] (dump_stack+0x0/0x28) from [<c0046d08>] (warn_slowpath_common+0x70/0x90)
* [ 1.551480] [<c0046c98>] (warn_slowpath_common+0x0/0x90) from [<c0046dcc>] (warn_slowpath_fmt+0x38/0x40)
* [ 1.561528] r8:c076b3fc r7:c0768410 r6:c0c8c9ac r5:c084df08 r4:c06b9fa4
* [ 1.568647] [<c0046d94>] (warn_slowpath_fmt+0x0/0x40) from [<c002f564>] (omap_device_get_by_hwmod_name+0x94/0xb4)
* [ 1.579523] r3:c058be8c r2:c06b6584
* [ 1.583341] [<c002f4d0>] (omap_device_get_by_hwmod_name+0x0/0xb4) from [<c0778280>] (am33xx_pm_init+0x204/0x45c)
* [ 1.594120] r4:00000000
* [ 1.596808] [<c077807c>] (am33xx_pm_init+
The full capture is http://pastebin.com/VyY8CQ0L
I found where roughly where there might be an issue...pm33xx.c as line#395 states it can only use bus i2c1(one based(I assume this is i2c0)) Note I do not need smart reflex or sleep mode at all so any patches to work around this are welcome.
Thanks
Josh Vanderpool
Electrical Engineer
Confidentiality Notice: The preceding e-mail message (including any attachments) contains information that may be confidential, protected by applicable legal privileges, or constitute non-public information. It is intended to be conveyed only to the designated recipient(s). If you are not an intended recipient of this message, please notify the sender by replying to this message and then delete it from your system. Use, dissemination, distribution or reproduction of this message by unintended recipients is not authorized and may be unlawful.
Confidentiality Notice:
The preceding e-mail message (including any attachments) contains information that may be confidential, protected by applicable legal privileges, or constitute non-public information. It is intended to be conveyed only to the designated recipient(s). If you are not an intended recipient of this message, please notify the sender by replying to this message and then delete it from your system. Use, dissemination, distribution or reproduction of this message by unintended recipients is not authorized and may be unlawful.
Confidentiality Notice:
The preceding e-mail message (including any attachments) contains information that may be confidential, protected by applicable legal privileges, or constitute non-public information. It is intended to be conveyed only to the designated recipient(s). If you are not an intended recipient of this message, please notify the sender by replying to this message and then delete it from your system. Use, dissemination, distribution or reproduction of this message by unintended recipients is not authorized and may be unlawful.
Confidentiality Notice:
The preceding e-mail message (including any attachments) contains information that may be confidential, protected by applicable legal privileges, or constitute non-public information. It is intended to be conveyed only to the designated recipient(s). If you are not an intended recipient of this message, please notify the sender by replying to this message and then delete it from your system. Use, dissemination, distribution or reproduction of this message by unintended recipients is
not authorized and may be unlawful.
Confidentiality Notice:
The preceding e-mail message (including any attachments) contains information that may be confidential, protected by applicable legal privileges, or constitute non-public information. It is intended to be conveyed only to the designated recipient(s). If you are not an intended recipient of this message, please notify the sender by replying to this message and then delete it from your system. Use, dissemination, distribution or reproduction of this message by unintended recipients is
not authorized and may be unlawful.
Confidentiality Notice:
The preceding e-mail message (including any attachments) contains information that may be confidential, protected by applicable legal privileges, or constitute non-public information. It is intended to be conveyed only to the designated recipient(s). If you are not an intended recipient of this message, please notify the sender by replying to this message and then delete it from your system. Use, dissemination, distribution or reproduction of this message by unintended recipients is
not authorized and may be unlawful.
Confidentiality Notice:
The preceding e-mail message (including any attachments) contains information that may be confidential, protected by applicable legal privileges, or constitute non-public information. It is intended to be conveyed only to the designated recipient(s). If you are not an intended recipient of this message, please notify the sender by replying to this message and then delete it from your system. Use, dissemination, distribution or reproduction of this message by unintended recipients is
not authorized and may be unlawful.
More information about the meta-ti
mailing list