Hot answers tagged

9

eudyptula-boot is quite handy for this; its introductory blog post has more details, but basically it allows you to boot a VM using the kernel you wish to test, and your existing filesystems (using overlayfs). That way you can quickly check a kernel without rebooting, and you still have access to all your files. The only requirement on the kernel being ...


3

The strength comes from all the other things you can do besides loading modules. Basically it gives you a userspace and the possibility of doing all the things you can do from that. An example: I use an initrd to have an encrypted root fs, setting that up requires code that there's no point having in kernel. The "Rationale" section of the Wikipedia page on ...


3

LTS does not mean that there will be no upgrades coming. LTS means that the packages will get minor upgrades to fix bugs and security problems when the 'standard' support for the version has stopped. It gives you the option/possibility to use your server in a safe way without you going to the next version that they support. The next version most often ...


3

ss uses the AF_NETLINK socket layer to talk to the kernel. This is a lower level protocol but allows for data to be transferred very quickly and in large chunks. A quick strace on CentOS 7 shows it sets the transfer window to be 1Mb.


2

The reason is, that the kernel is loaded from flash BEFORE the filesystem gets mounted. You need to connect to the serial port (with USB-SERIAL-CABLE-F) reboot the board and watch the serial output. interrupt the bootloader load a different kernel image into ram (via tftp or similar) transfer the kernel image to flash (do not overwrite the old kernel) ...


2

The Unix kernel has traditionally included some assembly language code.  I haven’t looked at its source code recently, but I suspect that that’s still true. See How does a driver actually communicate with a hardware device? for an overview of that topic.  The answers to that question discuss two kinds of computer architecture.  On a system that uses port-...


2

There are many reason to have an initramfs, some are below. When you need have a separate /usr,/var as some distros depends on having these directories in / When you want to encrypt / but you like to have /boot on a usb stick since you can't have an encrypted /boot When you don't want to have stuff in kernel builtin but instead as module, that way you only ...


2

PIDs do wrap around in normal usage. That's not a problem at all; the kernel ensures that new PIDs don't collide with existing PIDs. Nothing says that PIDs have to be monotomically increasing; process 12345 could easily fork() and have a child process of 5001. In this scenario, yes, a user could potentially use up all process slots and prevent further ...


2

I'm sure someone is still doing this, but back in the days before stuff like ILO/DRAC/etc. became cheap and ubiquitios the best way to get "out of band" access to the console in case of emergencies or an oops was over the serial port. You would mount a Terminal Server in the rack, then run cables to the serial port of your servers. Some BIOSs supported ...


1

The Linux kernel strives to have stable interfaces for everything that applications use. This includes not only system calls but also files under /proc and under select parts of /sys (some parts of /sys are officially unstable and do change, refer to the documentation for details). /proc/net/dev is documented so it's a stable interface. You can count on it ...


1

Total_vm was badly calculated by me and the OOM report is correct. app has allocated 59739 pages which is 233MB. So, this is the correct reason of OOM.


1

The kernel doesn't log module operations in a centralised manner, each module can log whatever it wants (using printk()). The USB module logs "Initializing USB Mass Storage Driver..." etc.; but the FCoE drivers only log messages when errors occur — if everything loads correctly they're silent. Many modules work this way; the tendency is to minimize the ...


1

Since you have build the Gentoo modules yourself, you most probably forgot to remove debug info from them. Try strip --strip-unneeded snd-ua101.ko and see if it makes a difference. Next time you rebuild modules for your system, strip all modules using make INSTALL_MOD_STRIP=1 modules_install


1

You only need to have the public key in your keyring: gpg --keyserver subkeys.pgp.net --recv-keys 0x38DBBDC86092693E (use the long identifier!). If it times out, try again — there are multiple servers, and some of them seem to be having issues currently. apt-key etc. aren't involved in this at all. Once you have the key in your keyring, gpg --verify ...


1

Forget about real mode, that's just a detail of the x86 architecture and it's just there for compatibility with 1980's processors. The processor indeed has a flag that indicates the current privilege level. The details of that flag vary between processor types but to keep things simple just think of it as two settings: user and kernel. There are processor ...


1

I suppose it's true that modern hardware wouldn't mind e.g. a 50MB kernel. You could argue that loading everything as separate modules has not been as important for a while now. However the initial ram system allows bootstrapping of any possible configuration, without needing any special handling in the kernel. Writing kernel code is a Big Deal. The ...


1

The iwl3945 driver is for Intel wifi devices. The Netgear WG111v3 doesn't use that driver. The Thinkpad x60s are shipped with Intel Wifi, so maybe it's a boot time hardware conflict, or the drivers need tweaking. The x60s is supposed to have a switch to turn off wifi. Before booting turn off that switch, then try booting. If that fails, (with the ...


1

fs_usage shows information about file system usage (and related things). Given that, I suspect THROTTLED from fs_usage indicates it is disk access being throttled and not CPU (although THROTTLED is not clearly documented in the man page). OSX reduces I/O throughput of idle applications to give better performance to active ones (see this Ars article) so it's ...


1

You can check seven line,they show you No memory reserved for crash kernel.It's means you need edit grub file ,give crash kernel some space,usually 128 or 256.



Only top voted, non community-wiki answers of a minimum length are eligible