A longstanding problem with Btrfs subvolumesand duplicate inode numbers was the topic of a late-breaking filesystem sessionat the2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM). The problem hadcropped up in the bcachefs session butJosef Bacik deferred that discussion to this just-created session, which he led. Theproblem is not limited to Btrfs, though, since filesystem snapshots forother filesystems canhave similar kinds of problems.
The5.17.9,5.15.41,5.10.117,5.4.195,4.19.244,4.14.280, and4.9.315stable kernel updates have all been released; each contains another set ofimportant fixes.
Security updates have been issued by Debian (elog, needrestart, openssl, and waitress), Fedora (curl, libxml2, slurm, and vim), Scientific Linux (zlib), SUSE (e2fsprogs, nodejs10, php72, and thunderbird), and Ubuntu (apport, clamav, needrestart, and pcre3).
Bcachefs is a longstanding out-of-treefilesystem that grew out of the bcache cachinglayer that has been in the kernel for nearly ten years. Based on asession led by Kent Overstreet at the2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM), though, it wouldseem that bcachefs is likely to be heading upstream soon. He intends tostart the process toward mainline inclusion over the next six months or so.
The Linux kernel allows processes to share pages in memory, but the pagetables used to control that sharing are not, themselves, shared; as aresult, processes sharing memory maintain duplicate copies of thepage-table data. Normally this duplication imposes little overhead, butthere are situations where it can hurt. At the 2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM), Khaled Aziz(remotely) and Matthew Wilcox led a session to discuss a proposed mechanismto allow those page tables to be shared between cooperating processes.
Version 1.2 of theInkscape drawing tool has been released. New features include multi-pagesupport, editable markers, the ability to flow text around shapes, andmore; see therelease notes for details.
It is perhaps unusual to have a kernel tracing developer leading afilesystem session, Steven Rostedt said, at the beginning of such a session at the2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM). But he was doingso totry to find a good way to dynamically allocate kernel data structuresfor some of the pseudo-filesystems, such as sysfs, debugfs, and tracefs,in the kernel.Avoiding static allocations would save memory, especially on systemsthat are not actually using any of the files in those filesystems.
Software Freedom Conservancy (SFC)has announcedthat it succeeded with its motion in US Federal Court to send the case backto California, where it was originally filed. The suit was filedin October 2021 by SFC, as an owner of Vizio televisions, to getthe company to comply with the GPL on some of the code in the TVs. Back in November, Vizio hadasked to move thecase to Federal Court, because the GPL is only a copyright license(which is a dispute handled at the Federal level) and not a contract (thatcould be adjudicated in state court). Friday's ruling disagreed with that premise:
A new helper library for network filesystems, called netfslib, was the subjectof a filesystem session at the2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM). David Howellsdevelopednetfslib, which was merged for 5.13 a year ago, and led the session.Some filesystems, like AFS and Ceph, are already using some of the servicesthat netfslib provides, while others are starting to look into it.
Security updates have been issued by CentOS (gzip, java-1.8.0-openjdk, java-11-openjdk, and zlib), Debian (adminer, htmldoc, imagemagick, libgoogle-gson-java, lrzip, openjdk-8, openssl, and ruby-nokogiri), Fedora (ecdsautils, et, libxml2, podman, and supertux), Mageia (cairo, clamav, curl, fish, freetype2, golang-github-prometheus-client, python-django-registration, python-nbxmpp, python-waitress, and xmlrpc-c), Red Hat (pcs), SUSE (curl, kernel, pidgin, and webkit2gtk3), and Ubuntu (tiff).
The 5.18-rc7 kernel prepatch has beenreleased for testing. Linus says: "So things continue to be fairlycalm, and as such this is likely the last rc before 5.18 unless somethingbad happens next week".
Memory reclaim in Linux is largely a reactive practice; the kernel tries tofind memory it can repurpose in response to the amount of free memoryfalling too low. Developers have often wondered if a proactive reclaimmechanism might lead to better performance, for some workloads at least,and optimal use of tiered-memory systems will likely require more activereclamation of memory as well. At the 2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM), Davidlohr Buesoled a brief session on the topic.
Compute ExpressLink (CXL) is an upcoming memory technology that is clearly on theminds of Linux memory-management developers; there were five sessionsdedicated to the topic at the 2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM). The first threesessions, on May 3, covered various aspects of memory management in thepresence of CXL. It seems thatCXL may bring some welcome capabilities, especially for cloud-serviceproviders, but that will come at the cost of some headaches on thekernel-development side.
Security updates have been issued by Debian (chromium, postgresql-11, postgresql-13, and waitress), Fedora (curl, java-1.8.0-openjdk-aarch32, keylime, and pcre2), Oracle (gzip and zlib), Red Hat (subversion:1.10), SUSE (clamav, documentation-suse-openstack-cloud, kibana, openstack-keystone, openstack-monasca-notification, e2fsprogs, gzip, and kernel), and Ubuntu (libvorbis and rsyslog).
Many types of kernel changes can be hammered into shape on the mailinglists. There are certain types of patches, however, that have a hard time getting tothe finish line that way; they are sufficiently large and invasive thatthey need an actual gathering of the developers involved. The multi-generational LRU work (MGLRU) falls into thiscategory, which is why it was the subject of a full-hour session at the 2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM). The discussionheld there may well have opened the doors for this code to be merged in thenear future.
Seven new stable kernels were released: 5.17.7, 5.15.39, 5.10.115, 5.4.193, 4.19.242, 4.14.278, and 4.9.313. As usual, they contain importantfixes throughout the tree and users of those series should upgrade.
The kernel's "direct map" makes the entirety of a system's physical memoryavailable in the kernel's virtual address space. Normally, huge pages are used forthis mapping, making it relatively efficient to access. Increasingly,though, there is a need to carve some pages out of the direct map; thissplits up those huge pages and makes the system as a whole less efficient.During a memory-management session at the 2022Linux Storage, Filesystem, Memory-management and BPF Summit (LSFMM),Mike Rapoport led a session on direct-map fragmentation and how it might beavoided.
On May 10, Red Hat announced the release of Red Hat Enterprise Linux 9 (RHEL 9). Not surprisingly, the announcement is rather buzzword-heavy and full of marketing, though there are some technical details scattered in it. The release notes for the RHEL 9 beta are available, which have a lot more information. "The platform will be generally available in the coming weeks."
In a filesystem session at the 2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM), Ted Ts'o broughtup the subject of filesystems that get resized frequently and whether thedefault parameters for filesystem creation should change as a result. Itstems from a conversation that he had with XFS developer DarrickWong, who is experiencing some of the same challenges as ext4 in this area.He outlined the problem and how it comes about, then led the discussion onways to perhaps address it.
Out-of-memory (OOM) situations are dreaded by users, system administrators,and kernel developers alike. Usually, all that is known is that a lot ofmemory is being used somewhere and the system has run out, but the kernel provides little help toanybody trying to figure out where the memory has gone. In a memory-managementsession at the 2022Linux Storage, Filesystem, Memory-management and BPF Summit (LSFMM),Kent Overstreet asked what could be done to improve OOM reports and reducethe pain for all involved.
Over on the Python Software Foundation (PSF) blog, Alex Waygood has a report from this year's Python Language Summit. There are reports from each of the nine sessions, including "Python without the GIL", The 'Faster CPython' project: 3.12 and beyond", "F-Strings in the grammar", lightning talks, and more.
Memory protection keys are a CPU feature that allows additional accessrestrictions to be imposed on regions of memory and changed in a fast andefficient way. Support for protection keys in user space has been in thekernel for some time, but kernel-side protection (often called "protectionkeys supervisor" or PKS) remains unsupported — on x86, at least. At the 2022 LinuxStorage, Filesystem, Memory-management and BPF Summit (LSFMM), IraWeiny provided an update on the state of PKS and led a discussion on whatthe proper in-kernel API for PKS should be.
It would have been surprising indeed if the 2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM) did not include asession working toward solutions to the longstanding problems withget_user_pages(), an internal function that locks user-space pagesin memory for access by the kernel. The issue has, after all, come up numerous timesover the years. This year's event duly contained a session in the jointfilesystem and memory-management track, led by John Hubbard, with a focuson page pinning and how it interacts with filesystems.
In a combined filesystem and memory-management session at the 2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM), Paul McKenneygave an update onthe changes to the read-copy-update (RCU) subsystem that had been made overthe last several years. He started with a quick overview of what RCU isand why it exists at all. He did not go into anyreal depth, though, since many of the topics could take a 90-minute session of theirown, he said, but he did provide some descriptions of the work that has gone intoRCU recently.
The 2022 LinuxStorage, Filesystem, Memory-management and BPF Summit (LSFMM) was thefirst chance for Linux memory-management developers to gather in threeyears. In a session at the end of the first day led by maintainer AndrewMorton, those developers discussed the memory-management developmentprocess. While the overall governance will remain the same, there arenonetheless some significant changes in store for this subsystem.
The Fedora 36release is now available. Improvements include GNOME 42, Waylandsupport by default on systems with NVIDIA graphics, Podman 4.0,Ansible 5, the removal of support for legacy ifcfgconfiguration files, GCC 12, and more; see therelease notes for details.
Like much of the kernel, the memory-management subsystem is under-documented,and much of the documentation that does exist is less than fully current.At the 2022 LinuxStorage, Filesystem, Memory-management and BPF Summit (LSFMM), MikeRapoport ran a session on memory-management documentation and what can bedone to improve it. The result was a reinvigorated interest indocumentation, but only time will tell what actual improvements will comefrom that interest.
Security updates have been issued by Debian (kicad and qemu), Fedora (thunderbird), Oracle (expat), Red Hat (samba), Slackware (kernel), and SUSE (firefox, ldb, and rsyslog).
The problem of negative dentries accumulating in the dentry cache in anunbounded manner, as we looked at back inApril, came up at the2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM).Negative dentries reflect failed file-name lookups, which are then cached,saving an expensive operation if the file name in question is looked upagain. There is no mechanism to proactively prune back those cacheentries, however, so the cache keeps growing until memory pressure finallycauses the system to forcibly evict some of them, which can make the systemunresponsive for a long time or even cause a soft lockup.
One of the memory-management subsystem's most important jobs is reclaimingunused (or little-used) memory so that it can be put to better use. When it comes toone of the core memory-management data structures — page tables — though,this subsystem often falls down on the job. At the 2022 Linux Storage,Filesystem, Memory-management and BPF Summit (LSFMM), David Hildenbrand led asession on the problems posed by the lack of page-table reclaim andexplored options for improving the situation.
Greg Kroah-Hartman has announced the release of the 5.17.6, 5.15.38, 5.10.114, and 5.4.192 stable kernels. As usual, thesecontain important fixes throughout the tree; users of those series should upgrade.
Security updates have been issued by CentOS (firefox and thunderbird), Debian (ecdsautils and libz-mingw-w64), Fedora (cifs-utils, firefox, galera, git, java-1.8.0-openjdk, java-11-openjdk, java-17-openjdk, java-latest-openjdk, mariadb, maven-shared-utils, mingw-freetype, redis, and seamonkey), Mageia (dcraw, firefox, lighttpd, rsyslog, ruby-nokogiri, and thunderbird), Scientific Linux (thunderbird), SUSE (giflib, kernel, and libwmf), and Ubuntu (dbus and rsyslog).
The 5.18-rc6 kernel prepatch is out fortesting. "Please do go test it all out - because things may lookgood now, but continued testing is the only thing that will makesure."
The GCC project has made the first release of the GCC 12 series, GCC 12.1. As the announcement notes, this month is the 35th anniversary of the GCC 1.0 release. There are lots of changes and fixes in this release, including:
There are certain themes that recur regularly at the Linux Storage,Filesystem, Memory-Management, and BPF Summit; among the most reliable isthe scalability problems posed by the mmap_lock (formerlymmap_sem) lock. This topic has come up in (at least)2013,2018 (twice),and 2019. The 2022 event was noexception, with three consecutive sessions led by Liam Howlett, MichelLespinasse, and Suren Baghdasaryandedicated to the topic. There are improvements on the horizon, but the problemis far from solved.
Security updates have been issued by Debian (dpdk, mruby, openjdk-11, and smarty3), Oracle (thunderbird), Red Hat (thunderbird), SUSE (chromium, libvirt, python-Twisted, and tar), and Ubuntu (cron and jbig2dec).
"Hardware poisoning" is a mechanism for detecting and handling memoryerrors in a running system. When a particular range of memory ceases toremember correctly, it is "poisoned" and further accesses to it willgenerate errors. The kernel has had support forhardware poisoning for over a decade, but that doesn't mean it can't beimproved. At the 2022 Linux Storage,Filesystem, Memory-management and BPF Summit, Yang Shi discussed thechallenges of dealing with hardware poisoning when it affects memory usedfor the page cache.
Security updates have been issued by Debian (firefox-esr), Fedora (firefox, java-1.8.0-openjdk, java-11-openjdk, java-17-openjdk, java-latest-openjdk, recutils, suricata, and zchunk), Oracle (firefox and kernel), Red Hat (firefox), Scientific Linux (firefox), Slackware (mozilla, openssl, and seamonkey), SUSE (apache2-mod_auth_mellon, libvirt, and pgadmin4), and Ubuntu (dpdk, mysql-5.7, networkd-dispatcher, openssl, openssl1.0, sqlite3, and twisted).
There is a lot of work going on right now on speeding up Python; KevinModzelewski gave a presentation at PyCon 2022 on some of thatwork. Much of it has implications for Python programmers in terms of howto best take advantage of these optimizations in their code. He gave anoverview of some of the projects, the kinds of optimizations being workedon, and provided some benchmarks to give a general idea of how much fastervarious Python implementations are getting—and which operations are most affected.
LWN.net