CentOS 5 :: Compile Realtime Clock Function To Kernel?
Oct 26, 2009
I am trying to install vmware server on CentOS: In vmware installation guide it says before intalling:
Before you begin, read the following notes and make adjustments to your host system: The realtime clock function must be compiled in your Linux kernel. The parallel port PCstyle hardware option (CONFIG_PARPORT_PC) must be built and loaded as a kernel module (that is, it must be set to m when the kernel is compiled).
I am seeing a 0.3% clock slowdown in kernel 2.6.18-238.9.1.It is giving ntpd conniptions.The problem does not occur in the previous 2.6.18-238.5.1 kernel.
# 2.6.18-238.9.1 [root@blue ~]# uname -a Linux blue 2.6.18-238.9.1.el5 #1 SMP Tue Apr 12 18:10:13 EDT 2011 x86_64 x86_64 x86_64 GNU/Linux[code]....
i need to use the realtime kernel because i use jack audio connection kit so i tried to install the realtime kernel from the install/remove software program. I've found 2 kernels : Kernel-trace and kernel-rt i tried both of 'em but i cannot boot up suse. The first one gets to 3/4 of the loading bar , i hear the welcome sound and then it won't do anything else. The second won't do anything, black screen.
Just got Studio 9.10 running and the RT kernel installed. I have a Prescott P4 3.2ghz (hyperthreading) and i encountered what looks like a common freeze problem. I am guessing my cpu will not be supported in the RT kernel since it is pretty old now.I added nosmp acpi=off during boot at the Grub2 menu and it booted fine and I was able to capture some video great.
My questions is, if I remove the smp support by using the nosmp option, do I still benefit from the RT kernel or would I do just as well with the generic kernel I have installed?Eventually I will upgrade do a new system board and processor but for now I just got some life back in this one by switching to Linux
I succeeded with the 2.6.31.12-rt20 on the Slackware 64 13.0.Kernel config seems fine, i use PAM.I can launch jackd as user with the realtime flag, BUT...Using top and htop, i can see that /etc/security/limits.conf or /etc/security/limits.d/limits.conf or wherever i put this file, priorities and nice are not set.Of course i added this to this file :
I had cloned a centos 5.6 installation from virtualbox virtual machine to physical box. Everything work fine. However, the time showing in os using date command differs from bios time by roughly 4 hours. I am running ntp services which sync the time with another centos server on the network. It appears that some services are using virtual clock and some use physical clock. How do I get rid of virtual clock and only use physical clock?
tty-clock won't compile for me. I'm probably missing something simple, I've never really compiled anything from source that I didn't write do I'm not sure how to do this. Here's the error:
I am trying to get the time of my wall clock from my wireless driver in form of "hh:mm:ss", so I used the function get_rtc_time() inside my driver to get that and it was done. My question: Is it possible to increase the accuracy of this function, for example if I want to get the milliseconds and the Microseconds in additional to the first output? If not, Is there any other function can do that?
I have been using parts of the MRG (kernel-rt, qpid[dc]) package for some time; thanks to Johnny Hughes (hughesjr) for providing the initial binaries. Recently, I discovered that CERN builds MRG for their ScientificLinux distribution, which is a derivative of CentOS.
You can view the package list from their repoview or browse all the packages they support here. I have found these packages to be compatible w/ CentOS. Perhaps something can be done to leverage this work to make MRG available as part of CentOS?It's straight forward to setup a .repo file to pull MRG directly via yum, including updates.
I have read a couple of articles on how dynamic linking works (those stuff about got, plt and lazy binding), and I am still not sure why you need to do dynamic linking in such a complicated way.Suppose your program uses a function in a shared library that needs to be linked dynamically at run time (like a printf). Why can't you statically decide the virtual address of the function at compile time? After all, all you need to do is to enter the page table entry corresponding to the address of the function if the library has been already loaded to a physical page frame.
i want to compile the vanilla kernel 2.6.37-rc3, but i want to obtain a .rpm file. I found this guide long time ago (i used it many times) but it use src.rpm package and the contained kernel.spec file have many lines for adding patches. Someone know where can i download a kernel.spec for vanilla kernel or a guide to obtain an rpm file
I am using AT91SAM9260 and running Linux 2.6.27 on it. Once i sync the Kernel clock and cmos clock with the reference clock and leave it for 1 day, i see a drift of nearly 8-10 sec. The kernel clock is running faster. How can i correct this?
I am using FC 2.6.31 kernel and am tying to bring up a SDIO network card. I am not able to change the SDIO clock speed which is by default 50MHz.I have to bring it down to somewhere around 25Mhz to make the card working smooth.How can we set the host clock speed from our driver in this kernel.
I am developing a function say my_abc() which does the same thing as kernel function abc(). Now I want that instead of using predefined kernel function kernel starts using my function. It might be possible that the kernel function has been used so many places.
When using make menuconfig - under Device Drivers --> Character Devices --> there should be an option with the label "Enhanced Real Time Clock Support" (CONFIG_JS_RTC).
The problem is that this option seems to only show up while using the menu method when other options are either enabled or disabled and I've entirely forgotten what should be what. I swear fingered it out once.
This is on an older computer (P4) so HPET is no good.
You would think that disabling the HPET option would enable the RTC option but that does not appear to be the case.
I understand I can just add the option to the .config file and avoid this hassle but I'm very interested to know how to make this work.
To show my appreciation I will do something nice for you such as call you a nice name or tell you that you are pretty (or ugly if that's what you prefer).
I manage to use the function "time" and "localtime" in gcc.But while I used it in a module in kernel space the above 2 functions are giving error.Even in kernel space time.h header file is not identified. tell the similar time function which can be used in kernel space
I wanted to synchronize both clock from windowsxp and centos. whenever i switch to centos i have to change clock before use is there any method so that i can synch both windows clock and centos clock?
Use python as example, I have 64bit python 2.4.3 yet I can't found rpm for 32bit one so I try to build it.Use default make or make --enable-readline is ok, python make cannot use CFLAGS, can use OPT and LDFLAGS to add -m32, yetit seems can only search for 64bit lib, or else crush.[Moderator edit: Moved from CentOS-4 to CentOS-5 forum.]
I am running a LAMP system with CentOS 5.4.The clock just automatically shifted backwards by 19 hours and 30 minutes crippling some of my reports and probably damaging something else along the way.The router supplying IP to this server is a DD-WRT and shows proper time.By the way what does that mean? time drifted by that 33.667 ms?
I try to setup a HPC cluster with CentOS 5.5. But now there is no Internet connection available in the room where the hardware is located so I set up ntpd server using synchronization with local clock (maybe I do something wrong). Here is my ntp.conf file on a master node (the master node has IP 10.0.1.1, a file server has IP 10.0.1.2 and compute nodes are 10.0.1.3..10.0.1.11, comments are omitted):
Trying to compile the 2.6.36.2 kernel. Its not the first kernel I've compiled, but I've run into a problem I've not come accross before.
While compiling I get this error.
[Code]....
Looks like the headers sys/eventfd.h and linux/virtio_rng.h haven't been selected in my .config file. I just copied my old config from the last kernel I built. This was a while ago (2.6.30), but I thought I'd give it a go anyway.
what config option they are under or what I can do to get rid of this error?