Cross compiling custom Linux Kenel driver for Rasberry Pi Platform

February 21, 2020February 21, 2020 vivekbhadra1 Comment

Getting Rasberry Pi up and running

I am using a NOOB version for getting up and running with the Rasberry Pi. The version of my Raspberry Pi is 3 Model B (see below):

Format a micro SD card

For formatting the micro SD card I usually use the gparted utility in Ubuntu. Format the micro SD with FAT 32 filesystem format. You need to have a SD card reader for attaching the SD card to the PC. I use the below SD card and the reader for my purposes:

Once you have attached your SD card with the linux PC you are ready to format it. I am using a Ubuntu PC for my purposes. Go to the console and run the below command:

~$ sudo gparted

By default the gparted selects your hard drive (/dev/sda is actually your hard drive, did you know!). Be very cautious, not to play around with your hard drive partitions, you might soon get into a funny situation. So go to the top right corner of the gparted screen and change the drive from /dev/sda to /dev/sdb (if there are no other media attached to the PC). Look at the various field carefully and try to understand what they are. I cannot go into the details of all the fields here. What we are up to is delete everything in the SD card and get it ready for the NOOBS. I can see the below in my SD card (/dev/sdb) at the moment:

You can also see the same partitions with lsblk command on the console, I can see the below:

sda 8:0 0 465.8G 0 disk ├─sda1 8:1 0 512M 0 part /boot/efi └─sda2 8:2 0 465.3G 0 part / sdb 8:16 1 29.7G 0 disk ├─sdb1 8:17 1 2.4G 0 part ├─sdb2 8:18 1 1K 0 part ├─sdb5 8:21 1 32M 0 part ├─sdb6 8:22 1 256M 0 part └─sdb7 8:23 1 27.1G 0 part
As you can see above the SD card has in total 5 partitions /dev/sdb1, /dev/sdb2, /dev/sdb5, /dev/sdb6, /dev/sdb7. I want to get rid of all of these. So what I do in gparted screen is to select/highlight each partition, right click, and then Unmount. Once the selected/highlighted partition has been Unmounted, you can right click and Delete it. Unmount and Delete each of these partitions in the SD card. Now Click on the green Tick at the top of the gparted screen and Apply. At this point you would have unallocated partition like this:

Right click on the Unallocated partition and click on New. Leave everything as default except Filesystems as FAT32, the default is ext4.

After changing the file systems type it should look like the below:

Click Add. Then again click on the green Tick button at the top, then Apply. Once done, you should see something like the below:

Now, close gparted and go back to your favourite file browser and locate the SD card, in my case it shows something like the below, an empty drive:

Get hold of the NOOBS

Go to the below link:

Downloads

Click on the Download ZIP and you will get the latest vesion of the NOOBS. Or if you need a different NOOBS version other than the latest, you can get a copy of that in the below link:

http://downloads.raspberrypi.org/NOOBS/images/

For my purposes I have download the NOOBS_v3_2_1 version (NOOBS_v3_2_1.zip) from the above link. Choose the one you need and move to the next step.

Once you have downloaded the NOOB version you want, extract the NOOBS and you should have a folder called something like NOOBS_*_*. In my case it is NOOBS_v3_2_1. Go inside the folder using your file browser, select and copy everything (Ctrl-A, Ctr-C). Go back to the SD card and do a Ctrl-V. This will copy and paste everything that the NOOBS_v3_2_1 had into the SD card.

Be patient, it takes a while to copy the stuffs across into the SD card. I often make mistake of prematurely eject the drive and nothing works! So be patient! There is a circular progress bar is whirling around at the top on the file browser window, if you have noticed. Once everything copied into the SD card, safely eject the drive and insert it into the SD card slot in Pi.

Now, connect the below to the Pi:

A keyboard.
A monitor/TV
A mouse (very useful)
Connect the Pi using micro USB cable to the PC to power it up.

Once you have powered the Pi, the NOOBS will take care of booting the device and soon you should see something like this on the Monitor:

Select the Raspberry Pi Full option from the menu shown above. You will require at least ~8GB I guess. Other option is to select the LibreELEC_RPi2 which I won’t recommend at this point. After you select one option from the menu, the “install” button will get activated. Click on Install. See the below pictures for reference:

Once you have clicked on the Install the installation starts:

Wait for the installation to finish. After the installation the setup wizard comes up, follow the screens and setup the Country, Language and Timezone etc. Keep the default user account ‘pi’ and password ‘raspberry’. It may ask for restart, do restart.

Find the kernel version of NOOBS installation

Go to Pi and open a terminal. Then type the below:
pi@raspberrypi:~ $ uname -a Linux raspberrypi 4.19.75-v7+ #1270 SMP Tue Sep 24 18:45:11 BST 2019 armv7l GNU/Linux
In the above it shows the kernel version this NOOBS installation has is 4.19.75-v7+. Now, for cross compilation we need the exact version of the Kernel source code, otherwise we will get the below error:
pi@raspberrypi:~ $ sudo insmod GobiNet.ko insmod: ERROR: could not insert module GobiNet.ko: Invalid module formatLinux kernels for Raspberry Pi is located here:

https://github.com/raspberrypi/linux

By default the one coming right now is version 4.19.y. This doesn’t exactly match with the one we need i.e. 4.19.75-v7+. Next you can look into the Branches here:

https://github.com/raspberrypi/linux/branches

However, I cannot see the exact match. So now I look into the Releases tab. This looks like has the one we want:

https://github.com/raspberrypi/linux/releases

In this page I can see the Kernel release which was on the 24th Sep, 2019. This exactly matches with the one in NOOBS:
pi@raspberrypi:~ $ uname -a Linux raspberrypi 4.19.75-v7+ #1270 SMP Tue Sep 24 18:45:11 BST 2019 armv7l GNU/Linux

Download this Kernel source code in your local machine (Ubuntu in my case). The tarball is called linux-raspberrypi-kernel_1.20190925-1.tar.gz. Extract the tarball with the below command:
tar -xvf linux-raspberrypi-kernel_1.20190925-1.tar.gzIn my setup after extraction the linux source is located in the below path:
/home/vbhadra/Downloads/linux-raspberrypi-kernel_1.20190925-1Now, go inside the linux kernel source folder and do th below:
cd /home/vbhadra/Downloads/linux-raspberrypi-kernel_1.20190925-1 make mrproper
This will ensure we have started with a clean kernel source.
Now, come to your Pi again. What we need now is the kernel configuration which was used for the NOOBS kernel compilation. The Kernel configuration now a days kept in the kernel itself as a module. So to grab hold of the kernel .config file do the below in Pi:
pi@raspberrypi:~ $ sudo modprobe configs pi@raspberrypi:~ $ zcat /proc/config.gz > /tmp/.config
Copy this .config file in a sd card and bring it back to the Ubuntu PC where your downloaded Kernel source tree lives. Then copy the .config file into the linux kernel source tree as below:
vbhadra@vbhadra-8200-elit-convertable-microtower:~/Downloads/linux-raspberrypi-kernel_1.20190925-1$ sudo mount /dev/sdb1 ~/usb/ vbhadra@vbhadra-8200-elit-convertable-microtower:~/Downloads/linux-raspberrypi-kernel_1.20190925-1$ ls -la ~/usb/ total 748 -rw-r--r-- 1 vbhadra vbhadra 163003 Sep 26 02:23 .config vbhadra@vbhadra-8200-elit-convertable-microtower:~/Downloads/linux-raspberrypi-kernel_1.20190925-1$ cp ~/usb/.config .

Download the Pi toolchain

You will need the toolchain to cross compile the Kernel and your module for Pi environment. Download the toolchain from the git hub with the below command:
git clone https://github.com/raspberrypi/tools ~/tools
Locate where your tools folder is. I usually create a rpi folder in my home folder and do everything there. So in my setup it is located here:
/home/vbhadra/rpi/tools
Now, you need to point where your tools (cross compiler) is for compiling the kernel. To do that do the below:
export CCPREFIX=/home/vbhadra/rpi/tools/arm-bcm2708/arm-bcm2708-linux-gnueabi/bin/arm-bcm2708-linux-gnueabi-
Compile the kernel to create the Module.symvers file. To compile the kernel do the below in your Ubuntu setup:
make ARCH=arm CROSS\_COMPILE=${CCPREFIX} -j 12
Let the above compile the Kernel which will take a while, wait patiently!

Now check if you have got a file created called in your linux folder:
vbhadra@vbhadra-8200-elit-convertable-microtower:~/Downloads/linux-raspberrypi-kernel_1.20190925-1$ ls -la Module.symvers -rw-r--r-- 1 vbhadra vbhadra 841463 Feb 12 13:31 Module.symversCross

Compile your module

I am trying to cross compile GobiNet and GobiSerial drivers fro SieraWireless LTE device EM7565.

You can grab hold of the driver rom SieraWireless website. This is not about any specific driver and their intricacies. We will only focus on the important bits to cross compile any custom driver for Raspberry Pi (hw version 3 B).

So One you have any driver code which has been compiled for any other platform (possibly) and you want to cross compile it, first thing you should do is to look at the driver Makefile. In this case, I have the below driver Makefile for GobiNet driver:

obj-m := GobiNet.o GobiNet-objs := GobiUSBNet.o QMIDevice.o QMI.o KDIR := /lib/modules/$(shell uname -r)/build PWD := $(shell pwd) OUTPUTDIR=/lib/modules/`uname -r`/kernel/drivers/net/usb/ #KBUILD_CFLAGS += -DQOS_SIMULATE #KBUILD_CFLAGS += -DTX_XMIT_SIERRA -DTX_URB_MONITOR ifdef TX_URB_MONITOR ifeq ($(TX_URB_MONITOR), 1) KBUILD_CFLAGS += -DTX_XMIT_SIERRA -DTX_URB_MONITOR else TX_URB_MONITOR:=0 endif else TX_URB_MONITOR:=0 endif ifeq ($(TX_URB_MONITOR), 1) ccflags-y:=-DTX_URB_MONITOR GobiNet-objs += usbnet_3_0_6.o \ usbnet_2_6_35.o usbnet_3_10_21.o usbnet_3_12_xx.o usbnet_4_4_xx.o endif


RAWIP := 0

ifeq ($(RAWIP), 1)

	ccflags-y:=-DDATA_MODE_RP

endif
#PI_KDIR := ~/k/linux-rpi-3.6.y

#PI_KDIR := /home/vbhadra/Downloads/linux-rpi-4.19.y-rt

PI_KDIR := /home/vbhadra/Downloads/linux-raspberrypi-kernel_1.20190925-1

#PI_CCPREFIX=~/toolchain/rpi/tools-master/arm-bcm2708/arm-bcm2708-linux-gnueabi/bin/arm-bcm2708-linux-gnueabi-

PI_CCPREFIX=/home/vbhadra/rpi/tools/arm-bcm2708/arm-bcm2708-linux-gnueabi/bin/arm-bcm2708-linux-gnueabi- 
OW_KDIR := ~/openwrt/trunk/build_dir/target-mips_r2_uClibc-0.9.33.2/linux-ar71xx_generic/linux-3.8.11

OW_CCPREFIX=~/openwrt/trunk/staging_dir/toolchain-mips_r2_gcc-4.6-linaro_uClibc-0.9.33.2/bin/mips-openwrt-linux-
MARVELL_KDIR := ~/toolchain/qmi_mxwell/kernel-2.6.31

MARVELL_CCPREFIX := ~/toolchain/qmi_mxwell/toolchain/bin/arm-none-linux-gnueabi-
M64_KDIR := ~/toolchain/arm64bit/linux/linux-4.4.1-devel-16.04.1

M64_CCPREFIX := ~/toolchain/arm64bit/be/aarch64ebv8-marvell-linux-gnu-5.2.1_x86_64_20151117/bin/aarch64_be-marvell-linux-gnu-
PPC_KDIR := ~/toolchain/ppc/kernel_ppc

PPC_CCPREFIX :=~/toolchain/ppc/fsl/1.1/sysroots/i686-fslsdk-linux/usr/bin/ppc64e5500-fsl-linux/powerpc64-fsl-linux-
HIKEY9_KDIR := /farm/android/Hikey_9.0_Trunk/kernel/hikey-linaro

HIKEY9_CCPREFIX := /farm/android/Hikey_9.0_Trunk/prebuilts/gcc/linux-x86/aarch64/aarch64-linux-android-4.9/bin/aarch64-linux-android-
all:

	$(MAKE) -C $(KDIR) M=$(PWD) modules
m64:

	$(MAKE) ARCH=arm64 CROSS_COMPILE=${M64_CCPREFIX} -C $(M64_KDIR) M=$(PWD) modules
ppc:

	$(MAKE) ARCH=powerpc CROSS_COMPILE=${PPC_CCPREFIX} -C $(PPC_KDIR) M=$(PWD) modules
marvell:

	$(MAKE) ARCH=arm CROSS_COMPILE=${MARVELL_CCPREFIX} -C $(MARVELL_KDIR) M=$(PWD) modules
pi:

	$(MAKE) ARCH=arm CROSS_COMPILE=${PI_CCPREFIX} -C $(PI_KDIR) M=$(PWD) modules
ow:

	$(MAKE) ARCH=mips CROSS_COMPILE=${OW_CCPREFIX} -C $(OW_KDIR) M=$(PWD) modules
hikey9:

	$(MAKE) ARCH=arm64 CROSS_COMPILE=${HIKEY9_CCPREFIX} -C $(HIKEY9_KDIR) M=$(PWD) modules
install: all

	mkdir -p $(OUTPUTDIR)

	cp -f GobiNet.ko $(OUTPUTDIR)

	depmod

clean: #rm -rf *.o *~ core .depend .*.cmd *.ko *.mod.c .tmp_versions Module.* modules.order .cache.mk *.o.ur-safe make -C $(PI_KDIR) M=$(PWD) clean

The only thing we are interested here is the target, in this file there is a target type called “pi”, that’s the one we want.
Look deeper into this target and associated macros:
pi: $(MAKE) ARCH=arm CROSS_COMPILE=${PI_CCPREFIX} -C $(PI_KDIR) M=$(PWD) modules
Notice the CROSS_COMPILE is set to PI_CCPREFIX which is what you need to fix. In this example we have discussed about the tools we will use for asp Pi Kernel compilation. The same tool will come handy here as well. Point the PI_CCPREFIX to your tool/cross compiler path as below:

PI_CCPREFIX=/home/vbhadra/rpi/tools/arm-bcm2708/arm-bcm2708-linux-gnueabi/bin/arm-bcm2708-linux-gnueabi-
Other important macro is this PI_KDIR. This points to the Kernel folder against which we are going to cross compile. Point this to your previosuly downloaded kernel source folder as below:
PI_KDIR := /home/vbhadra/Downloads/linux-raspberrypi-kernel_1.20190925-1
I also modified the clean target as below:
clean: make -C $(PI_KDIR) M=$(PWD) clean
But this is not necessary.
Save these changes and go back to the driver folder and issue the make clean command to start with a clean slate and then issue make pi command as below:
vbhadra@vbhadra-8200-elit-convertable-microtower:~/rpi/GobiNet$ make clean #rm -rf *.o *~ core .depend .*.cmd *.ko *.mod.c .tmp_versions Module.* modules.order .cache.mk *.o.ur-safe make -C /home/vbhadra/Downloads/linux-raspberrypi-kernel_1.20190925-1 M=/home/vbhadra/rpi/GobiNet clean make[1]: Entering directory '/home/vbhadra/Downloads/linux-raspberrypi-kernel_1.20190925-1' make[1]: Leaving directory '/home/vbhadra/Downloads/linux-raspberrypi-kernel_1.20190925-1' vbhadra@vbhadra-8200-elit-convertable-microtower:~/rpi/GobiNet$ make pi make ARCH=arm CROSS_COMPILE=/home/vbhadra/rpi/tools/arm-bcm2708/arm-bcm2708-linux-gnueabi/bin/arm-bcm2708-linux-gnueabi- -C /home/vbhadra/Downloads/linux-raspberrypi-kernel_1.20190925-1 M=/home/vbhadra/rpi/GobiNet modules make[1]: Entering directory '/home/vbhadra/Downloads/linux-raspberrypi-kernel_1.20190925-1' CC [M] /home/vbhadra/rpi/GobiNet/GobiUSBNet.o In file included from ./include/net/sock.h:58:0, from ./include/net/inet_sock.h:26, from ./include/net/ip.h:31, from /home/vbhadra/rpi/GobiNet/GobiUSBNet.c:94: ./include/linux/memcontrol.h: In function 'task_in_memcg_oom': ./include/linux/memcontrol.h:546:2: warning: return makes integer from pointer without a cast [enabled by default] /home/vbhadra/rpi/GobiNet/GobiUSBNet.c: At top level: /home/vbhadra/rpi/GobiNet/GobiUSBNet.c:112:2: warning: #warning "Remove memcontrol.h task_in_memcg_oom warning : replace 'return p->memcg_in_oom;' to 'return p->memcg_in_oom==NULL ? 0 : 1;' in function task_in_memcg_oom" [-Wcpp] /home/vbhadra/rpi/GobiNet/GobiUSBNet.c:113:2: warning: #warning "Commnet '#define MEMCG_NOT_FIX' above after fix applied." [-Wcpp] CC [M] /home/vbhadra/rpi/GobiNet/QMIDevice.o CC [M] /home/vbhadra/rpi/GobiNet/QMI.o LD [M] /home/vbhadra/rpi/GobiNet/GobiNet.o Building modules, stage 2. MODPOST 1 modules CC /home/vbhadra/rpi/GobiNet/GobiNet.mod.o LD [M] /home/vbhadra/rpi/GobiNet/GobiNet.ko make[1]: Leaving directory '/home/vbhadra/Downloads/linux-raspberrypi-kernel_1.20190925-1'
Check if there is a new kernel module been compiled called GobiNet.ko (in your case your define obj-m name). In this case the driver defined module name was GobiNet in the Makefile as below:
obj-m := GobiNet.o
So I will check if there is a GobiNet.ko been generated:
vbhadra@vbhadra-8200-elit-convertable-microtower:~/rpi/GobiNet$ ls -la GobiNet.ko -rw-r--r-- 1 vbhadra vbhadra 217987 Feb 21 11:17 GobiNet.ko
Copy this .ko file in a sdcard and insert into a Pi. Copy the .ko file from the sdcard to home folder in Pi. Then do the below insert it into the Pi Kernel:
pi@raspberrypi:~ $ sudo insmod GobiNet.ko pi@raspberrypi:~ $ dmesg [ 2526.062621] GobiNet: 2019-11-22/SWI_2.60 [ 2526.064306] usbcore: registered new interface driver GobiNet
As you can see now the cross compiled GobiNet driver is up and running in the Pi Kernel.

The other kernel driver for SieraWireless LTE EM7565 device is called GobiSerial. I have compiled this driver code in the same way. But when I try to insert this in the Pi kernel I get the below error:
pi@raspberrypi:~ $ sudo insmod GobiSerial.ko insmod: ERROR: could not insert module GobiSerial.ko: Unknown symbol in module pi@raspberrypi:~ $ dmesg [ 65.759262] GobiSerial: Unknown symbol usb_serial_generic_open (err -2) [ 65.759290] GobiSerial: Unknown symbol usb_serial_suspend (err -2) [ 65.759328] GobiSerial: Unknown symbol usb_serial_generic_resume (err -2) [ 65.759348] GobiSerial: Unknown symbol usb_serial_deregister_drivers (err -2) [ 65.759407] GobiSerial: Unknown symbol usb_serial_generic_write (err -2) [ 65.759427] GobiSerial: Unknown symbol usb_serial_register_drivers (err -2)
I resolved the problem in the following post here.

Unknown symbol usb_serial* in GobiSerial LTE Linux Kernel driver

February 21, 2020February 21, 2020 vivekbhadra1 Comment

I was working with the SieraWireless LTE device drivers recently and I got into a strange issue with one of the drivers called GobiSerial. I have compiled this driver code in the same way as described here. But when I try to insert this in the Pi kernel I get the below error:
pi@raspberrypi:~ $ sudo insmod GobiSerial.ko insmod: ERROR: could not insert module GobiSerial.ko: Unknown symbol in module pi@raspberrypi:~ $ dmesg [ 65.759262] GobiSerial: Unknown symbol usb_serial_generic_open (err -2) [ 65.759290] GobiSerial: Unknown symbol usb_serial_suspend (err -2) [ 65.759328] GobiSerial: Unknown symbol usb_serial_generic_resume (err -2) [ 65.759348] GobiSerial: Unknown symbol usb_serial_deregister_drivers (err -2) [ 65.759407] GobiSerial: Unknown symbol usb_serial_generic_write (err -2) [ 65.759427] GobiSerial: Unknown symbol usb_serial_register_drivers (err -2)

To debug this issue first I would like to know if the symbols like usb_serial_generic_open exists in the Pi kernel or not. Apparently not, but just to verify that I would use the /proc/kallsyms tool provided by Linux Kernel. This utility prints all the symbols (exported) available in the kernel space. So I would try to grep the missing symbols from all the avialable symbols in the kernel like this:
pi@raspberrypi:~ $ cat /proc/kallsyms | grep -i "usb_serial_generic_open" pi@raspberrypi:~ $
In the above command nothing is returned indicating my hunch the symbols are not present (exported/available) in the Pi Kernel.
Now, by the look of it the symbols should be coming from a Kernel module/driver usb_serial, this is my hunch by the look of the symbols which are missing. My next hunch is some Kernel module are not inserted unless there is a reason for it. Like if you haven’t inserted anything on the USB serial the usb_serial driver won’t be inserted into the kernel. Let’s try this. I will attach the SieraWireless LTE device into the USB port of the Pi. Before I do this I usually clean my Kernel messages buffer as below:
sudo dmesg -c
I do this so that when I insert the device I can only focus on the kernel messages coming out for this device. Now insert the device, wait for a second and then try the dmesg again, and you should be able to see something like this on the console:
[ 252.661876] usb 1-1.3: reset low-speed USB device number 4 using dwc_otg [ 253.511884] usb 1-1.5: reset low-speed USB device number 5 using dwc_otg [ 261.551829] usb 1-1.2: new high-speed USB device number 8 using dwc_otg [ 261.682434] usb 1-1.2: config 1 has an invalid interface number: 8 but max is 3 [ 261.682451] usb 1-1.2: config 1 has no interface number 1 [ 261.683111] usb 1-1.2: New USB device found, idVendor=1199, idProduct=9091, bcdDevice= 0.06 [ 261.683123] usb 1-1.2: New USB device strings: Mfr=1, Product=2, SerialNumber=3 [ 261.683133] usb 1-1.2: Product: Sierra Wireless EM7565 Qualcomm® Snapdragon™ X16 LTE-A [ 261.683143] usb 1-1.2: Manufacturer: Sierra Wireless, Incorporated [ 261.683153] usb 1-1.2: SerialNumber: UF92377624041506 [ 261.686019] QMAP Disabled [ 261.688132] GobiNet 1-1.2:1.8 eth1: register 'GobiNet' at usb-3f980000.usb-1.2, GobiNet Ethernet Device, d2:bd:e1:49:f1:08 [ 261.689996] USB Speed : USB 2.0 [ 261.805791] usbcore: registered new interface driver usbserial_generic [ 261.805832] usbserial: USB Serial support registered for generic [ 261.806590] usbcore: registered new interface driver cdc_wdm [ 261.814689] usbcore: registered new interface driver qcserial [ 261.814756] usbserial: USB Serial support registered for Qualcomm USB modem [ 261.815707] qcserial 1-1.2:1.0: Qualcomm USB modem converter detected [ 261.816223] usb 1-1.2: Qualcomm USB modem converter now attached to ttyUSB0 [ 261.819226] qcserial 1-1.2:1.2: Qualcomm USB modem converter detected [ 261.819635] usb 1-1.2: Qualcomm USB modem converter now attached to ttyUSB1 [ 261.819970] qcserial 1-1.2:1.3: Qualcomm USB modem converter detected [ 261.820236] usb 1-1.2: Qualcomm USB modem converter now attached to ttyUSB2 [ 261.820411] usbcore: registered new interface driver qmi_wwan [ 261.920268] IPv6: ADDRCONF(NETDEV_UP): eth1: link is not ready
Looks like the default linux kernel serial drivers been just inserted. Curious huh? Now, check if our missing symbol is back in the Kernel or not as below:
pi@raspberrypi:~ $ cat /proc/kallsyms | grep -i "usb_serial_generic_open" 00000000 r __ksymtab_usb_serial_generic_open [usbserial] 00000000 r __kstrtab_usb_serial_generic_open [usbserial] 00000000 t usb_serial_generic_open [usbserial]
Hurray!! It is back. Now try inserting the GobiSerial driver module again:
pi@raspberrypi:~ $ sudo insmod GobiSerial.ko pi@raspberrypi:~ $
Seems like the previous error didn’t appear. Lets check the dmesg again, and here we go:
[ 326.647915] usbserial: USB Serial support registered for GobiSerial [ 326.647943] GobiSerial: 2019-11-22/SWI_2.39:GobiSerial

Issue resolved.

How to pipe one bash command output to another within a python script

November 4, 2019November 5, 2019 vivekbhadraLeave a comment

I was recently working with something which required me to build a code base and flash it into the board (hw) and then once that is done open a serial console (where the hw is connected through USB) and issue further commands on the hw. Now, that is a regular job and I wrote a python script which does all the manual works of cleaning, building etc. But flashing the binary onto the hardware is where the problem started coming. The first time I flash the board there is no issue. However, once I opened a serial console, next time when I try to flash the binary the script finds the serial console to be already occupied.

nuttx kernel make menuconfig error in Ubuntu

October 25, 2019October 25, 2019 vivekbhadraLeave a comment

If you are working with nuttx kernel you might have issues trying to bring up the kernel menuconfig with the below command:

make menuconfig

I had to do the below to get it working in a newly installed Ubuntu system:

Make sure you have the right compiler tool chain installed, if not install it

 sudo apt-get install gcc-arm-none-eabi

Make sure you have the nuttx frontend tools installed in your Ubuntu, if not install it as below:

git clone https://bitbucket.org/nuttx/tools.git
cd tools/kconfig-frontends
./configure --prefix=/usr --enable-mconf --disable-nconf --disable-gconf
make
sudo make install

Now try the make menuconfig again from the console and you should be able to see the Kernel menuconfig page somewhat like this: Screenshot from 2019-10-25 10-53-26

How to integrate a custom device driver into nuttx tree

October 23, 2019October 23, 2019 vivekbhadraLeave a comment

I had written a blog on how to integrate a custom driver into Linux kernel tree previously here. Integrating a custom driver in nuttx kernel is pretty similar. But writing this down would probably be a good idea for any future reference. So below are the steps to follow to integrate a custom driver in nuttx.

Useful git command for developers

October 7, 2019November 7, 2019 vivekbhadraLeave a comment

Git log

git log --pretty=format:"%h%x09%an%x09%ad%x09%s"

Adding it to the bash script

echo "alias gitl='git log --pretty=format:"%h%x09%an%x09%ad%x09%s"'" >> ~/.bashrc
source ~/.bashrc

From now on you can type:

gitl
3918a7c8        Vivek Bhadra    Wed Oct 23 14:02:16 2019 +0100  Integrating the esp8266 wifi module in SPRESENSE sdk.
cbce5ee6        Vivek Bhadra    Tue Oct 22 13:16:11 2019 +0100  Fixes.
d4fc23c7        Vivek Bhadra    Wed Oct 16 16:01:03 2019 +0100  Enhanced camera App for better user interactivity.
6cc05f93        Vivek Bhadra    Wed Oct 16 13:47:02 2019 +0100  Customized for delayed capture of 1 mins.

and see the formatted git log, change the format as you would like.

Checking your current git branch

git branch -av

Add it to the .bashrc

echo "alias gitb='git branch -av'" >> ~/.bashrc
source ~/.bashrc

Type:

gitb
* customize_for_12_demo             3918a7c8 Integrating the esp8266 wifi module in XYZ sdk.
  master                            ee8f2ddf Merge pull request #41 from XXX/release-v1.4.1
  remotes/origin/HEAD               -> origin/master
  remotes/origin/auto_build_doxygen a5681292 Updated readme and only build doc from master
  remotes/origin/master             ee8f2ddf Merge pull request #41 from XXX/release-v1.4.1
  remotes/origin/release_1.0.1      ec7f3dab Release tag XYZ_1.0.001 YYY.0.001 release.

on the console to check the git branch you are on.

Creating a patch from multiple commits

git format-patch -x --stdout > patch-ddmmyyy.patch
git format-patch --ignore-space-change --ignore-space-at-eol --ignore-all-space -n --stdout > ~/patch4-LH3-18466-19092019.patch

where n is the number of commits you want to get back.
–ignore-space-change
–ignore-space-at-eol
–ignore-all-space
the above options are self-explanatory.

How to remove untracked files in git?

Print out the list of files which will be removed (dry run)

git clean -n

Delete the untracked files from the repository

git clean -f

How to not display the untracked files with git status?

git status --untracked-files=no


git status -uno

Renaming a local branch in command line

switch to the branch you want to rename

git checkout branch-name

Then rename it as below:

git branch -m new-name

Ubuntu developer’s tips and tricks

October 7, 2019October 7, 2019 vivekbhadraLeave a comment

Stop Ubuntu going into suspended mode

Disable suspend

$ sudo systemctl mask sleep.target suspend.target hibernate.target hybrid-sleep.target
Created symlink /etc/systemd/system/sleep.target → /dev/null.
Created symlink /etc/systemd/system/suspend.target → /dev/null.
Created symlink /etc/systemd/system/hibernate.target → /dev/null.
Created symlink /etc/systemd/system/hybrid-sleep.target → /dev/null.

Re-enable suspend

sudo systemctl unmask sleep.target suspend.target hibernate.target hybrid-sleep.target

Introduction to Arduino

October 1, 2019November 20, 2019 vivekbhadraLeave a comment

Install the arduino IDE
1. https://www.arduino.cc/en/Main/Software/
2. Select Linux 64 bit
3. Click Just Download

Unzip the downloaded file as below:
vbhadra@jupiter:~/Downloads$ tar -xf arduino-1.8.10-linux64.tar.xz

Run the arduino IDE
vbhadra@jupiter:~/Downloads$ cd arduino-1.8.10/
vbhadra@jupiter:~/Downloads/arduino-1.8.10$ ./arduino

Configure
1. Tools -> Board -> Board -> Arduino/Genuino Uno
Tools -> Port -> /dev/ttyACM0

Installing and uninstalling .zip user libraries in Arduino IDE
Install
Sketch -> Include Library -> Add .zip library
Uninstall
$ find ~/ -type d -iname Arduino
./Arduino
$ ls -la Arduino/
drwx—— 4 vbhadra vbhadra 4096 Oct 1 11:31 libraries
$ ls -la Arduino/libraries/
readme.txt xxxLPWA/ Vivek/

Remove the library folder you want to uninstall
$ sudo rm -rf Arduino/libraries/xxxLPWA/

Then modify your library code, create a .zip file and install it again.

Compile & verify the code
Sketch -> Verify & Compile
You should see “Done compiling”

Upload the code
Sketch -> Upload
You should see “Done Uploading”.

Troubleshooting
1. Issue 1:
An error occurred while uploading the sketchavrdude: ser_open(): can’t open device “/dev/ttyACM0”: Permission denied

-> Check if the port exist on the linux machine:

ls -l /dev/ttyACM*

-> Change the permission of the port as below:
sudo chmod a+rw /dev/ttyACM0

2. Junk characters in serial port

void setup()
{
  Serial.begin(9600);
  Serial.write("Hello World");
}

3. Mounting the SD card into the /mnt

mount -t vfat /dev/mmcsd0 /mnt

How to work in private branch in GIT

September 10, 2019September 10, 2019 vivekbhadraLeave a comment

Time to time we have to work on source code which are downloaded from a git repository, modify the code, implement the functionality, test it, get the code reviewed and then push it up to the repository. Now, how do we achieve this with the following in mind:

1. I do not want to do a git pull on my local source code base frequently simply because it is cumbersome.

2. I however, want to catch up with the mainline code (let’s call it master branch) so that when I am ready to push the code up I do not have enormous amount of merge conflicts, or conflicting other features which somehow difficult to accommodate at the last minute. So to avoid any surprise I would like to catch up with the mainline (be it master branch) time to time (I do it one a week at least).

3. I do not want to use merge command to catch up with the mainline code as it messes up with the git log.

All these points in mind I would recommend the below working principle:

1. git clone <git repository>

2. git checkout <branch name> (or by default you maybe already in the master branch. This would depend from source code to source code or repository to repository, so better check where you want to base your implementation.)

3. Now, after the above step 2, we are sitting at the branch where we want to start our implementation/bug fixing etc. Let’s call this “master” branch.

4. I would now like to create my private branch called “my_private”. To create a private branch do the below:

git checkout -b my_private

5. Now carry on with your implementations.

Multiple commits in private branch

While developing a new feature you might like to do it in a incremental way, i.e. when a little milestone is achieved you may like to commit it before carrying further. To do this do the below:

git add <file names>

git commit -m “commit message”

Now, if you carry on like this you would end up with multiple commits on your private branch. However, for doing review and pushing it up, it would be nice if all your changes/new implementations comes under one single commit. To achieve this you can use git interactive rebase command as below:

git rebase -i HEAD~n

where n stands for the number of commits (from the latest commits) to be included in the single commit.

This will bring up a screen which will look like as below:


pick 794969c79 Revert all commits that comprise 10x feature
pick 808bc82dd Added database version and product code into DiagnosticsPresenter.
pick 0b4d6973c Implementing Recent Tab
pick 0c284d078 Implementing Recent Tab

# Rebase 75ff6ea39..1f778b29d onto 75ff6ea39 (4 commands)
#
# Commands:
# p, pick = use commit
# r, reword = use commit, but edit the commit message
# e, edit = use commit, but stop for amending
# s, squash = use commit, but meld into previous commit
# f, fixup = like "squash", but discard this commit's log message
# x, exec = run command (the rest of the line) using shell
# d, drop = remove commit
#
# These lines can be re-ordered; they are executed from top to bottom.
#
# If you remove a line here THAT COMMIT WILL BE LOST.
#
# However, if you remove everything, the rebase will be aborted.
#
# Note that empty commits are commented out

Leave the first pick but convert the second and all other picks to squash as below:


pick 794969c79 Revert all commits that comprise 10x feature
squash 808bc82dd Added database version and product code into DiagnosticsPresenter.
squash 0b4d6973c Implementing Recent Tab
squash 0c284d078 Implementing Recent Tab

# Rebase 75ff6ea39..1f778b29d onto 75ff6ea39 (4 commands)
#
# Commands:
# p, pick = use commit
# r, reword = use commit, but edit the commit message
# e, edit = use commit, but stop for amending
# s, squash = use commit, but meld into previous commit
# f, fixup = like "squash", but discard this commit's log message
# x, exec = run command (the rest of the line) using shell
# d, drop = remove commit
#
# These lines can be re-ordered; they are executed from top to bottom.
#
# If you remove a line here THAT COMMIT WILL BE LOST.
#
# However, if you remove everything, the rebase will be aborted.
#
# Note that empty commits are commented out

Write and save, and you will be presented with the screen where you can decide what you would like to have as the new single commit message. Save and exit and you would have all your local commits squished into one single commit as below:


107b084c0 vbhadra Mon Sep 9 10:36:26 2019 +0100 Rework Recents presenter to communicate with STEDs facade instead of other presenter.
1aef95ec0 Andrei Tucma Tue Sep 10 08:30:19 2019 +0100 ITC5 calibration: calibrate popover option only available for itc5
8f761ce9c Robert Fry Mon Sep 9 15:57:09 2019 +0100 Possible fix for internal GPS to show signal strength
475af1405 dlangford Mon Sep 9 15:43:23 2019 +0100 STEDS Message popup design change
eb1eaf342 Algirdas Balciunas Mon Sep 9 14:24:20 2019 +0100 GPS source switching fix
460b05fee Algirdas Balciunas Mon Sep 9 14:11:25 2019 +0100 Fixed ship/cargo values in AIS data dialog
d7a47f77b Algirdas Balciunas Mon Sep 9 10:29:52 2019 +0100 Fix for app slowdown on Atom devices

Now, let’s say you want to catch up with the master branch. Go back to your master branch as below:

git checkout master (remember your base code maybe from a branch which not called master. It doesn’t matter, just go to the branch where you started)

git status (this will tell you how many commits your master is behind the latest commit)

git pull (this will bring all the latest changes delivered by others into your local master branch)

Now, go back to your private branch:

git checkout my_private (move to my_private branch)

git rebase master (this will rebase your my_private branch with the master)


107b084c0 vbhadra Mon Sep 9 10:36:26 2019 +0100 Rework Recents presenter to communicate with STEDs facade instead of other presenter.
1aef95ec0 Andrei Tucma Tue Sep 10 08:30:19 2019 +0100 ITC5 calibration: calibrate popover option only available for itc5
8f761ce9c Robert Fry Mon Sep 9 15:57:09 2019 +0100 Possible fix for internal GPS to show signal strength

Time to push up your changes

1. Go back to the master branch

2. Merge your private branch with master

3. git push


<strong>git checkout master</strong>

<strong>git merge my_private</strong>

<strong>git push</strong>

Customizing Qt Creator 5.12.4

June 19, 2019June 19, 2019 vivekbhadraLeave a comment

Install the custom Toolchain

Download and install the toolchain rpm

sudo alien ~/Downloads/<toolchain-name>.rpm

sudo dpkg -i ./<toolchain-name>.deb

Locate the arm compiler

sudo find / -iname arm-*-gnueabihf-g++ 2>/dev/null

/opt/toolchain/arm-cortexa9_neon-linux-gnueabihf-lh3/bin/arm-cortexa9_neon-linux-gnueabihf-g++

Settings

Go to the Projects Tab on the left pane.

BuildSettings

Click on the Manage Kit buttons.

ManageKits

If you want use your custom compiler and debugger etc then you have to create a manual kit.

A kit is a set of tools like compiler, QtVersion (qmake), debugger, device etc.

Before you can create a manual kit, you have to create place holders for your custom compiler, debugger etc.

Kit

Qt Version

Go tot the Qt Version Tab and select manual. Then click on Add button.

Here we are trying to specify the qmake binary we are going to use for the project.

In my case my qmake is located (comes from installation of the toolchain) at the below location:

/opt/toolchain/arm-cortexa9_neon-linux-gnueabihf-lh3/bin/qmake.

So in the add wizard I navigate to the above location and select qmake binary.

Qt automatically detects it as Qt Version 5.11.2. In the Kit section we are going to refer to this in near future.

Compilers

Next we have to choose our own custom C++ compiler.

In my case my C++11 compiler is located (comes from the toolchain installation) at the below location

/opt/toolchain/arm-cortexa9_neon-linux-gnueabihf-lh3/bin/arm-cortexa9_neon-linux-gnueabihf-g++.

Now, click on the Manual → C++ and click the Add button to start adding your own custom g++ compiler to the list.

This will ask you to provide a name to the custom compiler, give a name, e.g. “G++ for ATOM”, any other string you like.

Put the absolute path of the g++ binary which you want to add under the Compiler path:

/opt/toolchain/arm-cortexa9_neon-linux-gnueabihf-lh3/bin/arm-cortexa9_neon-linux-gnueabihf-g++

In the ABI, select arm-linux-generic-elf-32bit, leave the rest as default.

Apply, ok. Now you have your custom compiler in the list.

Debuggers

Go to the debugger tab.

Click on the Manual and Add.

Provide a suitable name, “ATOM GDB for Cortex A9”.

Specify the absolute path in the Path.

/opt/toolchain/arm-cortexa9_neon-linux-gnueabihf-lh3/bin/arm-cortexa9_neon-linux-gnueabihf-gdb

Apply, Ok.

Devices

In the Devices specify the details of the device where you want to run your generated binaries.

Custom Kit

Now go back to the Kits tab.

Select Manual.

Click on Add.

Provide a name for your custom Kit, e.g. ATOM.

In the Device Type, select Generic Linux Device.

In Device, select your own device, e.g. Custom Atom Device.

Compiler: C++: select your own created compiler from the drop down list.

Debugger: select the one you created previously.

Qt Version: select your own Qt Version as created previously.

Additional build arguments

Under Build & Run on the left hand panel, select the Build Option.

This will show you a loads of option on the right hand pane.

For our project we need to set the environment variable PRODUCT=ATOM.

To add this, click on the Details button along side qmake command under Build & Run.

This will look like below:

In the Additional Argument section specify, PRODUCT=ATOM.

Now, save it and try compiling your code. Make sure you clean it before giving it a try.

How to avoid recompilation of pre-built binaries with cmake

March 8, 2019March 8, 2019 vivekbhadraLeave a comment

How to create and run cppcheck executable from source code in linux machine (C++11 Static Analysis)

November 19, 2018November 19, 2018 vivekbhadraLeave a comment

I do not have administrative rights on a linux server (openSuse Linux) I am working on right now. So installing anything is not an option or at least an easy option. But I wanted to run cppcheck on my new C++11 source code. Here is what I found to achieve that:

How to install vim from tarball?

March 13, 2018March 15, 2018 vivekbhadraLeave a comment

Recently my apt-get install didn’t work in Ubuntu 10.04 and I couldn’t spend time to fix the source list when it didn’t work straight forward.

How to install and use a GIT client in Windows

AsideFebruary 15, 2018April 9, 2018 vivekbhadraLeave a comment

My favorite git repository is github. It is free unless you want to create a private repository. For private repositories you have to pay a monthly price but it is not mandatory. So if you want a free repository of your source code or documents, go to https://github.com/ and create an account.

How to take snapshot and restore a snapshot in VirtualBox 5.2.6

February 14, 2018February 14, 2018 vivekbhadraLeave a comment

Sometimes you may want to take a snapshot of the current state of your virtual machine (in my case Ubuntu Linux) so that you can revert back to it if anything goes wrong in the virtual machine. The feature in VirtualBox which helps in this is called a snapshot.

A snapshot is basically a backup of your virtual machine in its current state.

Why do we need pthread_join?

January 31, 2018February 14, 2018 vivekbhadraLeave a comment

In simple words pthread_join makes the calling thread wait till the newly created thread returns. How this helps?

Difference between re-entrancy and thread-safety

January 31, 2018February 14, 2018 vivekbhadraLeave a comment

Thread-safety of a function refers to the fact whether a function can be safely called from multiple threads simultaneously. Safety here means that even if multiple threads are executing the function simultaneously the data integrity is intact and is not intermingled.

What is memory leak in a C program?

January 30, 2018February 14, 2018 vivekbhadra1 Comment

What is memory leak?

Memory leak is a phenomenon where a running C/C++ program or a running process or thread dynamically allocates memory block from the heap but fails to free the memory block when it no more requires the memory. This happens due to programmatic error where the handle to the allocated memory block gets lost.

How to detect memory leak in c program using valgrind?

January 30, 2018February 14, 2018 vivekbhadra1 Comment

What is memory leak?

Please refer to my previous post for a discussion on What is memory in C a program?

What is virtual function in C++?

January 22, 2018February 14, 2018 vivekbhadraLeave a comment

Let’s say we have a class called Base and a derived class called Derived which is derived from Base class as below:

Why do we need virtual functions in C++?

January 22, 2018February 14, 2018 vivekbhadra1 Comment

Let’s look at the class definition below:


class Base {
public:
void welcome_message() {cout << "Welcome to Base class" << endl;}
};

class Derived : public Base {
public:
void welcome_message() {cout << "Welcome to Derived class" << endl;}
};

Git Rebasing

January 22, 2018February 14, 2018 vivekbhadraLeave a comment

Problem Description

Suppose you have a local GIT repository where dev-stable-topic is the development branch and time to time the releases are made to the customer from a remote GIT repository which has a branch called release-stable from where the selective changes are merge in a controlled way. The remote repository is hosted in a separate linux server.

How to resolve merge conflict in GIT

January 22, 2018February 14, 2018 vivekbhadra1 Comment

Merge Conflict

Often you will encounter merge conflicts while doing rebase operation as described above.

Marketing Management Assignment – Feasibility Analysis: Launching a branch of a company in a new market/country/economic zone

January 12, 2018February 14, 2018 vivekbhadraLeave a comment

Task

In your role as a Marketing Consultant, you have been asked by Wiggo*: a fictitious European supermarket (food retailer) chain (similar to Aldi and Lidl) to research and report on the feasibility of them launching in a new international market (this can be a country or economic zone of your choice and/or one you are familiar with).

How to publish a topic with mqtt broker running on iot.eclipse.org server

December 12, 2017February 14, 2018 vivekbhadraLeave a comment

As part of iot application development for IoT I had to play around with the SSL connection establishment with the remote mqtt servers. The below briefly describes the issue I faced and how did I got over the problem while connecting to a remote mqtt server over SSL.

https://iot.eclipse.org/ is a publicly accessible which runs a MQTT server. As per the tutorial the MQTT server is accessible using the host name io.eclipse.org and the port 1883 in unsecured mode. The server can be securely accessed on the encrypted port 8883.

How to access Ubuntu Linux Virtual Machine in NAT mode using FileZilla from Windows Host

December 8, 2017February 14, 2018 vivekbhadra1 Comment

I was facing an issue recently and it is like this. I run a Ubuntu Linux Virtual Machine on a Windows guest. Because of my specific requirements I need to connect to a specific AP (wifi router) from a Windows machine.

Why cannot connect to linux machine using FileZilla (sftp mode)?

November 24, 2017February 14, 2018 vivekbhadraLeave a comment

I was trying to connect to a newly installed Ubuntu Linux version 12.04 using FileZilla but the application kept on timing out. This can happen either in either of the below two protocols:

FTP or SFTP.

How to install ssh service in Ubuntu Linux (12.04)

November 24, 2017February 14, 2018 vivekbhadra1 Comment

Check if sshd service is running in Ubuntu with the below command:


vbhadra@vbhadra-VirtualBox:~$ ps aux | grep ssh
vbhadra 1702 0.0 0.0 4088 204 ? Ss 11:49 0:00 /usr/bin/ssh-agent /usr/bin/dbus-launch --exit-with-session gnome-session --session=ubuntu
vbhadra 4358 0.0 0.0 4400 824 pts/3 S+ 14:10 0:00 grep --color=auto ssh
vbhadra@vbhadra-VirtualBox:~$ sudo netstat -natp | grep sshd
vbhadra@vbhadra-VirtualBox:~$ sudo netstat -natp | grep ssh
vbhadra@vbhadra-VirtualBox:~$

The above command and their output shows there is no sshd service running on my Ubuntu Linux.

Why cannot connect to Ubutnu linux machine using Filezilla (ftp)?

November 24, 2017February 14, 2018 vivekbhadra3 Comments

I was facing the issue while trying to connect to a newly installed Ubuntu 12.04 running on a Virtual box. I use FileZilla a lot to connect to my linux machine (Virtual machine running on a VirtualBox) for file transfers. My filezilla connection keeps on timing out without much of information.

Installing ftp service in Ubuntu 12.04

November 24, 2017February 14, 2018 vivekbhadra1 Comment

This may be very simple but to keep things handy the below will install ftp service in you Ubutu Linux machine:


vbhadra@vbhadra-VirtualBox:~$ sudo apt-get install vsftpd
Reading package lists... Done
Building dependency tree
Reading state information... Done
The following NEW packages will be installed
vsftpd
0 to upgrade, 1 to newly install, 0 to remove and 318 not to upgrade.
Need to get 130 kB of archives.
After this operation, 353 kB of additional disk space will be used.
Get:1 http://gb.archive.ubuntu.com/ubuntu/ precise/main vsftpd i386 2.3.5-1ubuntu2 [130 kB]
Fetched 130 kB in 0s (927 kB/s)
Preconfiguring packages ...
Selecting previously unselected package vsftpd.
(Reading database ... 183924 files and directories currently installed.)
Unpacking vsftpd (from .../vsftpd_2.3.5-1ubuntu2_i386.deb) ...
Processing triggers for man-db ...
Processing triggers for ureadahead ...
Setting up vsftpd (2.3.5-1ubuntu2) ...
vsftpd start/running, process 4107
vbhadra@vbhadra-VirtualBox:~$

This has been tested on a Ubutnu 12.04 running on a virtual box.

Installing GIT on Ubuntu 12.04

November 24, 2017February 14, 2018 vivekbhadraLeave a comment

Using apt-get install

Use the below command on your ubuntu terminal:

$ sudo apt-get install git

Auto, static and global variable in C/C++ programming

November 17, 2017February 14, 2018 vivekbhadraLeave a comment

Auto variable

An auto variable is a variable which is declared within the scope of a function. We call it local variable as well. Auto variables are automatically created on the stack as soon as the program control enters the function and are destroys as soon as the program control returns from the function.

Let me judge!

November 1, 2017 vivekbhadraLeave a comment

Source: Let me judge!

Let me judge!

AsideNovember 1, 2017September 10, 2019 vivekbhadra1 Comment

My take on people’s amazing judgmental power, the funny and the not so funny side.

How to integrate a custom device driver into kernel tree

AsideJuly 19, 2017February 14, 2018 vivekbhadraLeave a comment

Find out where you want to place the device driver source code. In my case I had to integrate a device driver for a device called ring oscillator. The category of the device was not clear to me and I couldn’t find a suitable place under “linux/drivers” folder except “linux/drivers/misc”. So I decided to place […]

How to create a new package under buildroot

AsideJune 16, 2017February 14, 2018 vivekbhadraLeave a comment

In the earlier post I have explained how to add a kernel module as a new buildroot package. You can check that post here: https://vivekbhadra.wordpress.com/2017/06/15/how-to-addcompile-a-kernel-module-as-a-new-buildroot-package/ In this module I will explain the steps to add a program which is not a kernel module but a user space c program for testing the earlier kernel module.

How to add/compile a kernel module as a new buildroot package

ImageJune 15, 2017February 14, 2018 vivekbhadra1 Comment

Lets say I have a kernel device driver xyz which is coming as an external module and needed to be integrated with the existing software tree. This can be achieved in various different ways. Not going into the details of it, we will try focus on how the external kernel module can be integrated as […]

How to modify kernel command line

ImageJune 9, 2017September 29, 2017 vivekbhadraLeave a comment

[/contact-form] You can modify the kernel command line from the below dts file: ~/repos/solaris_src/linux/arch/mips/boot/dts/xyx/xyz.dts chosen { – bootargs = “console=ttyS0,115200n8 loglevel=8 earlycon=uart8250,mmio32,0x1CC40000,115200 rootwait ro”; + bootargs = “mem=32M console=ttyS0,115200n8 loglevel=8 earlycon=uart8250,mmio32,0x1CC40000,115200 rootwait ro”; };

Write a program to convert a decimal number to its binary equivalent using stack.

January 11, 2017January 11, 2017 vivekbhadraLeave a comment

#include <iostream>
#include <sstream>
#include <stdio.h>
#include <string>

using namespace std;

template <typename T>
class Stack {
private:
    int top;
    T datastore[100];
public:
    Stack() : top(-1) {}
    bool is_stack_empty(void) {return (top == -1) ? true : false;}
    bool is_stack_full(void) {return (top >= 100) ? true : false;}
    void push(T elem)
    {
        if (!is_stack_full()) {
            top++;
            datastore[top] = elem;
        }
    }

    T pop()
    {
        if (is_stack_empty()) {
            cout << "Stack is empty" << endl;
        } else {
            int data = datastore[top];
            top--;
            return data;
        }
    }

    T peek(void)
    {
        if (is_stack_empty()) {
            cout << "Stack is empty" << endl;
        } else {
            int data = datastore[top];
            return data;
        }
    }
};

int main()
{
    Stack<int> s;
    int num;

    do {
        cout << "Enter a number ";
        cin >> num;
        cout << "Number entered is " << num << endl;
        while (num)
        {
            int n = num % 2;
            num = num/2;
            s.push(n);
        }

        cout << "The binary equivalent of the number is " << endl;
        while(!s.is_stack_empty()) {
            cout << s.pop();
        }
        cout << endl;
        cout << "Do you want to continue? (y/n) ";
        getchar();
    } while (getchar() == 'y');

    return 0;
}

Write a program to check if a word is a palindrome using stack.

January 11, 2017January 11, 2017 vivekbhadraLeave a comment

#include <iostream>
#include <sstream>
#include <stdio.h>
#include <string>

using namespace std;

template <typename T>
class Stack {
private:
    int top;
    T datastore[100];
public:
    Stack() : top(-1) {}
    bool is_stack_empty(void) {return (top == -1) ? true : false;}
    bool is_stack_full(void) {return (top >= 100) ? true : false;}
    void push(T elem)
    {
        if (!is_stack_full()) {
            top++;
            datastore[top] = elem;
        }
    }

    T pop()
    {
        if (is_stack_empty()) {
            cout << "Stack is empty" << endl;
        } else {
            int data = datastore[top];
            top--;
            return data;
        }
    }

    T peek(void)
    {
        if (is_stack_empty()) {
            cout << "Stack is empty" << endl;
        } else {
            int data = datastore[top];
            return data;
        }
    }
};

int main()
{
    Stack<char> word;
    string str;
    string pali = "";

    cout << "Enter a word ";
    cin >> str;
    char w;
    for (auto i = 0; i < str.size(); ++i) {
        w = str[i];
        word.push(w);
    }
    while(!word.is_stack_empty()) {
        stringstream ss;
        char c = word.pop();
        ss << c;
        string s = ss.str();
        pali += s;
    }

    if (str == pali)
        cout << "The word is a palindrome " << endl;
    else
        cout << "The word is not a palindrome " << endl;


    return 0;
}

How to import a python script within another python script

September 9, 2016February 14, 2018 vivekbhadraLeave a comment

Let’s say I have a python script called my_module.py in the location ~/junk/my_test/script/lib/. The my_module.py has a very trivial function called my_print() which just prints “Hello World”.

Browse Kernel code with Eclipse

August 18, 2016February 21, 2018 vivekbhadraLeave a comment

This blog will try to demonstrate the steps required to use eclipse for browsing Linux kernel level source code, e.g. Linux kernel tree or device driver codes. If you have already worked with Linux kernel level source code you will know that the Linux kernel is written in C and Linux kernel space device drivers are also written in C. For c/c++ source code the eclipse version I use is called Eclipse-CDT (C/C++ Development Tooling). My development environment is Ubuntu Linux. So most of my references will be based on Ubuntu packages but these steps are pretty much similar in other Linux distros.

How to run an external program from a python script

August 16, 2016February 14, 2018 vivekbhadra1 Comment

This blog will try to demonstrate how a python script can be used to automate the running of an external program with data input file located at a particular location.

Lets say we have the below c++ program which reads from a data file called myData.txt located for this example at user home i.e. ~/. The user home is usually /home/<username> for example /home/vbhadra in my case. Use your linux user name instead while trying it yourself. To find out what is your home directory you can use echo $HOME in linux command prompt. Our objective in this blog is to demonstrate the pythoin script which will run any external program binary with data input. So we will have a less focus on the c++ program.

comepletion_done

April 5, 2016February 14, 2018 vivekbhadra1 Comment

Problem Statement

The main mlme processing kernel thread comes out of the wait_for_completion() immediately without waiting for the complete() to be called on the completion variable. As a result the mlme response was erroneous and the command to check the wifi link status “sudo iw wlan0 link” was returning error.

Working with GDB

February 23, 2016February 14, 2018 vivekbhadraLeave a comment

Using gdb to user space program
vbhadra@ubuntu:~$ sudo gdb ./netlinkUser [sudo] password for vbhadra: GNU gdb (Ubuntu 7.9-1ubuntu1) 7.9 Copyright (C) 2015 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>

Git Cheatsheet

June 3, 2015February 20, 2018 vivekbhadraLeave a comment

Setup cscope for your project

June 1, 2015October 16, 2019 vivekbhadra1 Comment

Run the below script from the top level project directory.

Setup Script

#!/bin/bash
find  -type f -iname '*.c' -o -iname '*.cpp' -o -iname '*.h' -o -iname '*.hpp' -o -iname '*.cc' -o -iname '*.hh'  -type f  > ~/cscope.files
cscope -q -R -b -i ~/cscope.files -f ~/cscope.out
export CSCOPE_DB=~/cscope.out
ctags -R .

Is #AAP hurting Modi?

January 12, 2014 vivekbhadraLeave a comment

It is a myth being spread across by media that #AAP is sweeping the support of the people all across the country to garner more support in future in favour of it for some vested interest or over enthusiasm. But step back and think for a moment, Delhi was the epitome of #AAP movement, it gained its acme of campaign with all those rhetorical promises, by all those theatrics of cutting the electrical wires, burning bill in public. Populism at its best. But even after all these melodrama it couldn’t even stand as the the single largest party which it claimed to be before the election. It reminds me somewhat of the “Singur” moment in West Bengal by Mamta Banerjee. But the difference is that Mamta Banerjee (for good or bad) was able to sweep whole of Bengal, there was no sign of others around, people truly gave their mandate. But here is a party which claims to be the party with a difference (the difference is yet to come out in their practice) has not even managed to be the single largest party in a small/half state like Delhi. And I believe #AAP had a better popular plank for campaign than Mamta Banerjee. So the first impact of #AAP is not even the half of Trinomool Congress.
Now just after the election #AAP has compromised with its basic promise of not aligning with any of the “Chor”/”Bhrasth” (as they say) national parties. But they have already shook hands with Congress. Some would argue on a super technical theorem that this is not an alliance it is a support etc. But think about the support of Mulayam to Congress, that’s exactly the same, supporting from outside and not an alliance. But there is always a quid-pro-quo – Mulayam provides support to a minority govt and in turn centre doesn’t let the CBI go after him and his party leaders, centre turns a blind eye to the riots happening in the state so on and so forth. Now here in Delhi Congress’s immediate target is to defend Modi, cut into anti-congress vote and restrict him to come to power. And once that is accomplished congress can collect a bunch of so called secular parties and form the next govt. And then after what would it demand from #AAP in future that future will tell us, I have no idea. Congress support to a minority govt has been usually maximum 2 years so far. As I strongly believe there is no free lunch this has to be a part of a grand deal, time will tell.
Now coming back to #AAP garnering support across the country. So a new party with some holier than thou agenda’s, with some populist schemes in the list couldn’t even manage to totally woo the already distressed people of a very small state. And after forming the govt the way they have tried to keep their promises so far gives us two points, one they are following the standard populist measures to woo the voters(free water and 50% electricity bill reduction etc) and two that these people are modern day leftists when it comes to economy. I see a lot of people who perhaps voted for them are not happy and have already started raising their voice. So here #AAP in fact loses their voters share rather than extending it. One more point, a percentage of people who voted for AAP in Delhi thinks they want Modi as their next PM in the LS election. So #AAP loses this section of voters too when it comes to LS. The third is interesting. I see a lot of #AAP supporters are pathological Modi haters. A lot of new #AAP supporters would be bad mouthing about Modi all the time, which indicates they are the original Congress supporters and now turning to #AAP as Congress is nowhere in the picture. This section of voters are interesting because rather than cutting into Modi’s vote (which is Congress’s original plan for supporting #AAP) it seems that #AAP is now cutting into Congress’s traditional vote bank. That be the case it is going to help Mr. Modi in coming days than hurting him. It would be difficult for #AAP to translate all their dramatic campaigns in the national platform simply because state to state India is different and they would realise it soon. They would find it extremely difficult follow what they preach as lots of those are just not practical. Unlike Modi #AAP doesn’t have any track records to showcase to the people as their model of governance. The little they have shown has party gone against them already, at least to a section of voter. So however much the media may try to make the king of their choice, at last the the real people’s king would perhaps have the last laugh. Jai Hind!!

Vim Shortcuts for Developers

October 5, 2011November 4, 2019 vivekbhadra

Most of us like to work in multiple files simultaneously. For navigating through the different files in a project VIM provide us with multiple features.The split window command is one of them. Lets talk about it.