KernelNewbies:

Linux 3.14 has been released on Sun, 30 Mar 2014.

Summary: This release includes the deadline task scheduling policy for real-time tasks, a memory compression mechanism is now considered stable, a port of the locking validator to userspace, ability to store properties such as compression for each inode in Btrfs, trigger support for tracing events, improvements to userspace probing, kernel address space randomization, TCP automatic coalescing of certain kinds of connections, a new network packet scheduler to fight bufferbloat, new drivers and many other small improvements.

1. Prominent features

1.1. Deadline scheduling class for better real-time scheduling

Operating systems traditionally provide scheduling priorities for processes: The higher priority a process has, the more scheduling time that process it can get with respect other processes with lower priorities. In Linux, users usually set scheduling priorities from a value of -20 to 19 using the nice(2) tool (in addition, Linux supports the notion scheduling classes: each class provides different scheduling policies; for example, there is a SCHED_FIFO class with a "first in, first out" policy, and a SCHED_RR with a round-robin policy).

The approach of process priorities is, however, not well suited for real-time tasks. Evidence Srl and the ReTiS Lab have created an alternative designed around real time concepts: deadline scheduling, implemented as a new scheduling policy, SCHED_DEADLINE.

Deadline scheduling gets away with the notion of process priorities. Instead, processes provide three parameters: runtime, period, and deadline. A SCHED_DEADLINE task is guaranteed to receive "runtime" microseconds of execution time every "period" microseconds, and these "runtime" microseconds are available within "deadline" microseconds from the beginning of the period. The task scheduler uses that information to run the process with the earliest deadline, a behavior closer to the requirements needed by real-time systems. For more details about the scheduling algorithms, read the documentation

Recommended LWN article: Deadline scheduling: coming soon?

Recommended page on Wikipedia: SCHED_DEADLINE

Documentation: Documentation/scheduler/sched-deadline.txt

Code: commit 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13

1.2. zram: Memory compression mechanism considered stable

zram provides RAM block devices. Everything written to these block devices gets compressed. If zram block devices are used as swap, when the system tries to move parts of memory to swap it will be effectively moving memory from one part of the RAM to another, except that the data will be compressed before being copied to the destination. This effectively works as a cheap memory compression mechanism to improve responsiveness in systems with limited amounts of memory. Zram is being used by TV companies, Android 4.4, Cyanogenmod, Chrome OS, Lubuntu...

Zram has been in staging since Linux 2.6.33. In this release, zram has been moved out of staging to drivers/block/zram.

Code: commit, commit, commit

1.3. Btrfs: inode properties

This release adds infrastructure in Btrfsto attach name/value pairs to inodes as xattrs. The purpose of these pairs is to store properties for inodes, such as compression. These properties can be inherited, this means when a directory inode has inheritable properties set, these are added to new inodes created under that directory. Subvolumes can also have properties associated with them, and they can be inherited from their parent subvolume. This release adds one specific property implementation, named "compression", whose values can be "lzo" or "zlib" and it's an inheritable property.

Code: commit

1.4. Trigger support for tracing events

The tracing infastructure in the Linux kernel allows to easily register probe functions as events (for more details, see Documentation/trace/events.txt. This release allows these events to conditionally trigger "commands". These commands can take various forms, examples would be enabling or disabling other trace events or invoking a stack trace whenever the trace event is hit. Any given trigger can additionally have an event filter, the command will only be invoked if the event being invoked passes the associated filter.

For example, the following trigger dumps a stacktrace the first 5 times a kmalloc request happens with a size >= 64K: {{{# echo 'stacktrace:5 if bytes_req >= 65536' > \

For more details, see Section 6 in Documentation/trace/events.txt

Recommended LWN article: Triggers for tracing

Code: commit 1, 2, 3, 4,5, 6, 7

1.5. Userspace probes access to all arguments

Userspace probes are a Linux 3.5 feature that allows to set tracing probes in userspace programs at runtime. This release enables to fetch other types of argument for the uprobes: memory, stack, deference, bitfield, retval and file offset. For more details see here.

Code: commit 1, 2, 3

1.6. Userspace locking validator

The Linux kernel has (since 2.6.18) a lock validator that can find locking issues at runtime. This release makes possible to run the Linux locking validator in userspace, making possible to debug locking issues in userspace programs. For more details, see the recommended LWN link.

Recommended LWN article: User-space lockdep

Code: commit 1, 2, 3, 4, 5, 6, 7

1.7. Kernel address space randomization

This release allows to randomize the physical and virtual address at which the kernel image is decompressed, as a security feature that deters exploit attempts relying on knowledge of the location of kernel internals.

Recommended LWN article: Kernel address space layout randomization

Code: 1, 2, 3, 4, 5, 6, 7, 8, 9

1.8. TCP automatic corking

When applications do consecutive small write()/sendmsg() system calls, the Linux kernel will try to coalesce these small writes as much as possible, to lower total amount of sent packets - this feature is called "automatic corking". Automatic corking is done if at least one prior packet for the flow is waiting in Qdisc queues or device transmit queue. Applications can still use TCP_CORK for optimal behavior when they know how/when to uncork their sockets. A new sysctl (/proc/sys/net/ipv4/tcp_autocorking) has been added to control this feature, which defaults to enabled. For benchmarks and more details see the commit link.

Code: commit

1.9. Antibufferbloat: "Proportional Integral controller Enhanced" packet scheduler

Bufferbloat is a phenomenon where excess buffers in the network cause high latency and jitter. As more and more interactive applications (e.g. voice over IP, real-time video streaming and financial transactions) run in the Internet, high latency and jitter degrade application performance. There has been a number of features and improvements in the Linux kernel network stack that try to address this problem.

This release adds a new network packet scheduler: PIE(Proportional Integral controller Enhanced) that can effectively control the average queueing latency to a target value. Simulation results, theoretical analysis and Linux testbed results have shown that PIE can ensure low latency and achieve high link utilization under various congestion situations. The design incurs very small overhead. For more information, please see technical paper about PIE in the IEEE Conference on High Performance Switching and Routing 2013. Also you can refer to the IETF draft submission. All relevant code, documents and test scripts and results can be found at ftp://ftpeng.cisco.com/pie/.

Code: commit

2. Drivers and architectures

All the driver and architecture-specific changes can be found in the Linux_3.14-DriversArch page

3. Core

4. Memory management

5. Block layer

6. File systems

7. Networking

commit

8. Virtualization

9. Security

10. Crypto

11. Tracing/perf

12. Other news sites that track the changes of this release

KernelNewbies: Linux_3.14 (last edited 2017-12-30 01:30:14 by localhost)