- kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

Can small writes generate a lot of faults?
by Daniel Hilst Selli April 23, 2013

April 23, 2013

My case is this, I'm using collectd [1] with rrdtool [2] to monitor some server. a) When I enable rrdtool plugin I can grab collectd process as top page fault process on top command. PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ nFLT COMMAND 20128 root 18 0 4105m 35m 17m S 7.9 0.3 0:09.62 22k /opt/collectd/sbin/collectd b) Disabling rrdtool plugin dramastically decrease the number of page faults of collectd. c) I know that rrdtool plugin is known to generate a lot of small writes, as stated in [3]. Also I know the parameters to improving collectd's cache usage to save I/O, but this is not the question here.. AFAIK, major page faults are generated when data that is not yet present on RAM is loaded from disk, but in this case data is being write do disk, I can't see how writes can generate faults, but still, it seems that is happening, ... !? Is that possible? Cheers, [1] http://www.collectd.org [2] http://oss.oetiker.ch/rrdtool/ [3] https://collectd.org/wiki/index.php/Inside_the_RRDtool_plugin

4 4

Writing device drivers in Linux: A brief tutorial
by Zack April 23, 2013

April 23, 2013

I thought this might be useful: http://www.freesoftwaremagazine.com/articles/drivers_linux -- Zachary

4 3

Invitation to connect on LinkedIn
by Vivek P. April 23, 2013

April 23, 2013

LinkedIn ------------ Vivek P. requested to add you as a connection on LinkedIn: ------------------------------------------ Deepak, I'd like to add you to my professional network on LinkedIn. - Vivek Accept invitation from Vivek P. http://www.linkedin.com/e/8z2206-hfuyp8yy-a/L90NvsYL3GREvJdnM0f-HnfkWNbSD_6… View profile of Vivek P. http://www.linkedin.com/e/8z2206-hfuyp8yy-a/rso/86727318/UjWT/name/49489491… ------------------------------------------ You are receiving Invitation emails. This email was intended for Deepak Shah. Learn why this is included: http://www.linkedin.com/e/8z2206-hfuyp8yy-a/plh/http%3A%2F%2Fhelp%2Elinkedi… (c) 2012, LinkedIn Corporation. 2029 Stierlin Ct, Mountain View, CA 94043, USA.

2 1

what will kernel_execve() does..?
by sandeep kumar April 23, 2013

April 23, 2013

Hi All I m having a doubt about.. what kernel_execve() does...? As per the comment it should make the kernel thread a user space process. I am doing a kernel bringup. while launching the init process, i was struck in this function... kernel_execve() in arch./arm/kernel/sys_arm.c I am feeling i was struck in some kind of infinite loop here..and program control is not reaching till user space init process's main function. *Can someone tell me, is there any condition that i am missing here, failing which the kernel will struck in infinite loop and doesnt switch to user space?* -- With regards, Sandeep Kumar Anantapalli,

1 0

[HELP] Document: /sys/devices/system/cpu/cpu0/cache/index0/coherency_line_size
by Ren Zhen April 23, 2013

April 23, 2013

Hi everybody, For /sys/devices/system/cpu/cpu0/cache/index0/coherency_line_size, I execute the cmd '#cat coherency_line_size',then it return '64'. My OS is Ubuntu 12.04, CPU is intel core i3. Here is my problem, what is coherency_line_size ? what is its usage? Any help will be appreciated! Sincerely, Ren Zhen

2 1

I2C: kernel & userspace drivers : using i2c-stub
by Sagar Padhye April 23, 2013

April 23, 2013

Hi, [USING : Ubuntu 12.04 LTS] I am trying to learn I2C subsystem. I havent found much of well documented things that can help a newbie on google. I just chose to read the linux kernel documentation txts. I read the way kernel can interact with i2c through both SMBus calls & raw file op calls (ioctl, read, write). It looks good till that point. The doc also tells about i2c stub. which can be used as follows: 1. load the i2c module 2. use i2cset (from the i2c-tools project) to pre-load some data 3. load the target chip driver module 4. observe its behavior in the kernel log I was able to achieve just the 1st step here. I loaded the module successfully. $ sudo modprobe i2c-stub chip_addr=0x023 I got following Kern msg $dmesg ... [658.071706] i2c-stub: Virtual chip at 0x23 So I believe that step 1 is clear. in second step I have no idea how to provide needed inputs to command -> Usage : i2cset [-f] [-y] [-m mask] I2CBUS CHIP-ADDRESS DATA-ADDRESS [VALUE] [MODE] I tried to give I2CBUS = 0, Chip addr = 0x23 & data addr=0 but it gives error Error : Could not open file '/dev/i2c-0' or '/dev/i2c/0' : No such file or directory There is no file at all created in /dev corrosponding to this device. Any idea how to use i2c properly? Is this an ubuntu only problem? Can anybody provide a good tutorial like thing to learn about i2c kernel & userspace driver programming? -- Thanks and Regards, Sagar

3 3

Fwd: [RFC]CPU in SYSFS: what information should be written to /sys/devices/system/cpu/probe to add cpu In X86?
by Ren Zhen April 22, 2013

April 22, 2013

Hi, everybody: What information should be written to /sys/devices/system/cpu/probe to add a cpu In X86? And, what information should be written to /sys/devices/system/cpu/release to remove a cpu In X86? I have read the /Documentation/ABI/testing/sysfs-devices-system-cpu. It says that probe:writes to this file will dynamically add a CPU to the system. Information written to the file to add CPU's is architeture specific. release: writes to this file will dynamically remove a CPU to the system. Information written to the file to remove CPU's is architeture specific. I just want to know what information should be written to those file. Any help will be appreciated! -- Sincerely, Ren Zhen

1 0

Alternative for API system() in multi-threaded app
by Praveen kumar April 22, 2013

April 22, 2013

I was trying to replace system() API in my multi-threaded app as it is suggested not to any calls in multi-threaded app which uses fork()..... When I was going through the man pages of fork()..in the Notes section I found * "Since version 2.3.3, rather than invoking the kernel's fork() system call, the glibc fork() wrapper that is provided as part of the NPTL threading implemenâ* * tation invokes clone(2) with flags that provide the same effect as the traditional system call. The glibc wrapper invokes any fork handlers that have been* * established using pthread_atfork(3)."* Is this mean, I still can use fork() or any other API which uses fork() for its implementation and make sure my app is thread safe or free from dead locks???? Thanks, Praveen R

1 0

Something about Redhat
by Robert Clove April 21, 2013

April 21, 2013

Hi, I want to learn so i found somethings to ask here...... I had an red hat machine i forgot the password so i change it to single user and break the password . NOW, I want to know how to 1)Switch it back to multi user mode? 2)how to know which service are running and how to stop services 3)how to get the machine to receive the dynamic ip i mean how to make dhcp enable. 4)how to make multiple users work at the same time. Thanks

2 4

process context value
by Shraddha Kamat April 21, 2013

April 21, 2013

I was reading about PowerPC-64 address translation and stumbled upon a term 'process context value' - googled but could not find much details about this - please help. regards, Shraddha

1 0