124 reputation
9
bio website
location
age
visits member for 3 years, 4 months
seen Oct 9 at 22:54

Oct
9
comment How do I delete all of a set of files in a random order?
In zsh, you would use the *.bin(o+functionName) notation as in unix.stackexchange.com/a/9831
Oct
7
comment Why do user space apps need kernel headers?
(Also note that ABI =/= API. The kernel's (external) API includes the symbols referenced in the source code, and the kernel developers have committed to maintain compatibility; unless the symbols are very rarely used in real world software, it would be an unreasonable move for them to make incompatible changes, although no third-party standard mandates the whole of the Linux API and therefore nothing technically impedes it.)
Oct
7
comment Why do user space apps need kernel headers?
(For clarity: note that the kernel has an external ABI, which we are discussing about here, and an internal ABI between the kernel modules, which undergoes frequent and possibly incompatible changes, but it is not seen by userspace and is irrelevant when compiling userspace code.)
Oct
7
comment Why do user space apps need kernel headers?
... it also makes it possible to chroot/LXC into a distro that shipped with a different kernel. etc.
Oct
7
comment Why do user space apps need kernel headers?
@suprjami The ABI obeys the standard SysV ABI, which can be reasonably assumed not to change; among other things, SysV ABI is why you can link against the kernel with your compiler (and version) of choice, not necessarily the very same one that emitted the kernel binary.
Oct
7
comment Why do user space apps need kernel headers?
1. Busybox is designed for embedded and other lightweight systems, where you want to load as less libraries in memory as possible. I have not read the source, but probably BB is basically sidestepping the libraries and link directly against the kernel. 2. No. It couples your userspace binary with a version of the kernel API, which is backwards-compatible: a binary compiled today might not work on an older kernel (depending on the symbols included... read the source code) but will work with future kernels.
Sep
25
comment Why does “ls | wc -l” show the correct number of files in current directory?
Do not parse the output of ls before you read this explanation. Be aware of the pitfalls.
Oct
27
comment Not able to install internet explorer!
Consider Microsoft's official virtual machines with Windows + IE
Oct
27
comment Does a high “used” value for a device in “btrfs filesystem show” pose a problem? Should a balance operation be performed?
With btrfs, don't trust the "regular" df. Just don't use it. See FAQ: Why are there so many ways to check the amount of free space?
Oct
24
comment Is CentOS exactly the same as RHEL?
Indeed, CentOS's FAQ states they don't: "Does CentOS change the upstream Source RPMs? No."
Oct
6
comment Resize an extended partition to the whole drive
Note that some filesystems (btrfs, zfs) allow for adding devices to a filesystem: if you can not extend the partition, you may create a new one and add it as a second device.
Oct
6
comment Resize an extended partition to the whole drive
If you follow frostshutz's idea: after having resized the partition, you might want to resize the filesystem (via resize2fs or analogous program) to extend it over the new area of the partition.
Oct
6
comment Resize an extended partition to the whole drive
What is the exact text of the error message about the partition geometry?
Sep
14
comment How can I monitor disk io?
/sys/block/sda/stat is documented at https://www.kernel.org/doc/Documentation/block/stat.txt
Jun
30
comment How to prevent Windows partition to access(read) Linux partition files?
Or: have the partitions permanently deleted, and pass the offset (-o offset) to mount in the initrd.
Jun
24
comment Should a laptop user switch from ext4 to btrfs?
You found corrupted data twice and still believe the SMART response??
Jun
17
comment How can I solve the issue about mounting my Windows 8 NTFS partition on Linux Mint?
When you write to a filesystem, part of the data is kept in RAM, for quick access. It is sent to the disk later. When an OS is hibernated, that data is put in the hibernation data: if you access the filesystem with another OS, you see a state that may not represent what the first OS has virtually done on the disk. It's a bad habit on NTFS and other journaled filesystems, and even more dangerous on FAT and other non-journaled ones. That's why you are not allowed to write - when the first OS resumes, it will expect the original f.s. state and assume the changes buffered in RAM still apply.
Jun
1
comment Is it good to make a separate partition for /boot?
Are you implying that it was it 5 seconds slower in your experience?
May
30
comment Is it good to make a separate partition for /boot?
It's hard for me to believe that journaling and checksumming make boot noticeably slower. Do you have hard numbers?
May
4
comment Rearrange partitions to install a second Linux distribution
I add (manually) a menu entry in GRUB, to boot it. How: grub: boot from ISO