Fedora and SUSE both pushed Chromium to 156.0.8078.47 with 353 CVEs crammed in, making the browser the day's biggest patch. Debian's OpenJDK update carried the heaviest count on its side with 40 flaws, though a regression fix restoring PGP mail is arguably the more useful piece. Fedora, RHEL, Oracle, and the others shipped routine kernel and daemon patches, with one LibreWAN advisory even listing a mismatched CVE that reporters would likely overlook. If you use any of these distros, patch the browsers first, reboot the servers on the kernel rows, and don't skip the Thunderbird fix.
Chromium's 353 fixes anchor a heavy day of Linux distro patches
Fedora and SUSE shipped the same browser patch within minutes of each other, and both dropped 353 CVEs into it. That's the loudest single event of the day, and it probably isn't the one you need most.
Most of today's news is exactly the kind you install without reading. Kernels got stable bumps. Daemons got sandbox fixes. A browser got 353 patches. Still, a handful of entries deserve a second look, and two of them matter for reasons unrelated to the underlying bugs.
The browsers brought the biggest numbers
Chromium moved to 156.0.8078.47 in Fedora's Tuesday batch and again in SUSE's openSUSE-SU-2026:0360-1. The browser has been clearing large batches of fixes for a while, so 353 in one release is just the latest installment rather than something unprecedented. Memory corruption and use-after-free variants show up across V8, ANGLE, and WebGL, which means the right answer really is "just update."
Firefox is on its own ESR cycle at the same time. Oracle Linux and Red Hat both pushed it to 153.5.0, pulling in a pile of CVEs large enough to make anyone reading the changelog stop halfway through.
The detail worth checking
On the Debian side, OpenJDK takes the biggest hit with 40 CVEs. That kind of count usually makes you open the list and read it twice. The fixes cover sandbox bypass, incomplete certificate checks, information disclosure, and DoS. Worth updating regardless.
Thunderbird is the stand-out here for a weird reason. Debian's earlier fix in DSA-6585-1 actually broke PGP-encrypted mail, so DSA-6585-2 is a regression patch that restores it. If you send encrypted email, this one matters more than the underlying vulns. Older boxes shouldn't be ignored either, since xorg-server clears 15 issues on Freexian's extended line, several of which can escalate to root.
AlmaLinux shipped seven advisories on October 27th. The one worth grabbing no matter which release you host is xdg-dbus-proxy, because it shows up in both 9 and 10. Thunderbird bundles 43 shared flaws, and the kernel carries 15 security fixes plus a patch for a panic that appeared after out-of-bounds reads. The single-issue tail covers brotli, urllib3, and an arbitrary-code-execution flaw in Emacs' Flymake mode.
Then there's the Fedora grab bag. OpenSSH took a DoS fix for oversized decompressed packets, util-linux got a descriptor leak in unshare patched, and Ghostscript closed a -dSAFER sandbox bypass alongside two buffer overflows. The entry most worth a second glance is LibreWAN, where the advisory headline lists CVE-2026-9445 but the changelog points at CVE-2026-94453, an over-TCP use-after-free in IKEv2. Reporters will let that slide. You probably won't.
Oracle's kernel rows are doing most of the carrying again, so schedule the reboot first. The notable repeat is brotli's out-of-bounds write (CVE-2026-111497), which Oracle backported into its Node.js 22 and Node.js 24 bundles. Flatpak path-traversal fixes (CVE-2026-97023 and CVE-2026-97024) and the xdg-dbus-proxy sandbox escape round out the rows worth your time.
Red Hat clearly had a lot of OpenShift to push. The same versions, 4.16.74, 4.14.77, 4.18.59, 4.12.102, and 4.21.40, appear both as bug-fix releases and as packages-and-security advisories. Some headers read Critical while the body rates the security piece Important. Don't get spooked. That split is standard Red Hat bookkeeping: the title covers the whole release, the body covers just the vulnerability, so a Critical header doesn't mean the CVE itself is more dangerous. Brotli also shows up four times across RHEL streams, with emacs twice each. A single flaw funneled through multiple channels looks like a spam problem, but it just means the same fix is being pushed to different customer feeds.
Rocky handed out 20 errata, and exactly one is critical: grafana on Rocky Linux 9. The rest top out at Important or Moderate. None of the notices name a CVSS score or a CVE, so take the advisory IDs, cross them against what you run, and open the links yourself.
Slackware's contribution is narrow and specific. proftpd gained Huffman compression over SSH, but only for SSH compression, closing a hole where a chosen-plaintext attack could target channels sharing a single connection. Slackware points at GitHub issue 2366 for the full writeup. Since the feature is opt-in, your existing setups won't shift unless you deliberately turn it on.
SUSE split its attention across a different kind of mundane. SUSE-SU-2026:4918-1 through 4922 are just rebuilds against the current Go security release, spread across SLE 15 SP4 through SP7. The Tumbleweed moderate run patched one package each, with ImageMagick taking the volume prize at 13 fixes and Xen following with five.
Finally, Ubuntu has a pretty ordinary-looking hole in authd, the authentication daemon that shells out to your cloud identity provider. Ali Mashally found it wasn't validating user identities when matching accounts arriving from providers, which in practice lets a local attacker climb onto someone else's account and read their files. The fix is a single package bump to 0.7.1 followed by a daemon restart, logged under CVE-2026-102631.
The takeaway is simple. Patch the browsers if you use them, reboot the servers on the kernel rows, and give the Thunderbird regression patch special attention if you send encrypted mail.
A Detailed Overview of the Updates
AlmaLinux
AlmaLinux shipped seven security advisories on October 27th, and most are the kind you install without much fuss. Three target the older 8 stream, while two each hit versions 9 and 10. The xdg-dbus-proxy patch is the one that shows up twice, covering both versions 9 and 10, so it's the one worth grabbing no matter which release you run. The rest just depends on which versions you actually host.
The Thunderbird update is the one that looks the most alarming, bundling 43 separate flaws shared with Firefox. Most of them belong to the same familiar family: use-after-free bugs, sandbox escapes, and privilege escalations. The ones that actually bite are the privilege escalations and the full sandbox escapes, where a loaded page can break out entirely. If you open attachments from people you don't completely trust, patch that first. The long list reads worse than it is, though, since it mostly falls into the same three buckets, so the update is thorough without being mysterious.
The kernel advisory is the other heavyweight, carrying 15 security fixes plus a patch for a panic that surfaced after a run of out-of-bounds reads. Those use-after-free and integer-overflow fixes are exactly the sort you want quietly handled, no ceremony.
The remaining four are single-issue patches: a brotli decoder out-of-bounds write, a urllib3 denial-of-service from unbounded memory allocation in its chunk parser, the xdg-dbus-proxy sandbox escape (CVE-2026-94422, split across the two version advisories), and an arbitrary-code-execution flaw in Emacs' Flymake mode. The Emacs case is a gentle reminder that "just editing a file" rarely stays innocuous.
| Advisory ID | Component | AlmaLinux version | What's fixed |
|---|---|---|---|
| ALSA-2026:90365 | kernel | 9 | 15 security fixes (UDP rehash race, FUSE re-lock, net/handshake file-reference leaks, IPVS header reload, SELinux SCTP deref, KVM s390 PCI forwarding, USB use-after-free, SMB setuid clearing, VF MSI clearing, RAID5-PPL use-after-free, KEYS integer overflow, netfilter flowtable ref) plus a fix for a __crypto_xor kernel panic (JIRA:AlmaLinux-257205) |
| ALSA-2026:90449 | brotli | 10 | Out-of-bounds write in the Google Brotli decoder (CVE-2026-111497) |
| ALSA-2026:90759 | python-urllib3 | 8 | Denial-of-service from unbounded memory allocation in the chunk parser (CVE-2026-97689) |
| ALSA-2026:90522 | xdg-dbus-proxy | 9 | Message filtering bypass via reply serial, allowing sandbox escape (CVE-2026-94422) |
| ALSA-2026:90523 | xdg-dbus-proxy | 10 | Same sandbox-escape flaw (CVE-2026-94422) |
| ALSA-2026:90913 | Thunderbird (Firefox) | 8 | 43 shared flaws across use-after-frees, sandbox escapes, privilege escalations, JIT miscompilations, and info disclosure |
| ALSA-2026:90800 | Emacs | 8 | Arbitrary code execution in Flymake mode (CVE-2026-96442) |
Debian GNU/Linux
Debian rolled out another batch of security updates, covering both its regular stable tracks and the extended LTS line that Freexian runs for older releases. Most of the day's fixes chase denial-of-service and memory-safety bugs, but a few cover more interesting ground: cache poisoning, sandbox escapes, and a whole pile of X server flaws.
OpenJDK takes the biggest hit with 40 CVEs. That's the kind of count that makes you open the CVE list and read it twice. The fix addresses sandbox bypass, incomplete certificate checks, information disclosure, and DoS. Worth updating, and the sheer number is a double-take on its own.
The DNS side shows up twice: bind9 (Debian LTS, for bookworm) and pdns (stable/trixie). Both get DoS fixes, and bind9's version also handles cache poisoning, which is the worse outcome when your name server gets tricked into serving bad records.
Thunderbird is the stand-out. The earlier fix in DSA-6585-1 broke PGP-encrypted mail, so DSA-6585-2 is a regression patch that restores it. If you send encrypted email, this one matters more than any of the underlying security fixes.
The two Freexian ELA items keep older releases breathing. clamav (buster and bullseye) dodges a stack of file-processing CVEs, and xorg-server (stretch and buster) clears out the most alarming set of the day. Xorg's list runs to 15 issues, several of which can escalate to root if the X server runs privileged. Cache poisoning and root escalation are exactly the flavors you don't want to leave sitting in an LTS box.
| Package | Advisory | Version | Release | CVE(s) | Impact |
|---|---|---|---|---|---|
| bind9 | DLA-4865-1 | 1:9.18.49-1~deb12u3 | Debian 12 (bookworm) LTS | 10 | Denial of service, cache poisoning |
| php-mongodb | DSA-6589-1 | 2.0.0-1+deb13u2 | Stable (trixie) | 3 | Denial of service |
| openjdk-21 | DSA-6588-1 | 21.0.13+9-1~deb13u2 | Stable (trixie) | 40 | DoS, sandbox bypass, incomplete cert checks, info disclosure |
| pdns | DSA-6587-1 | 4.9.18-0+deb13u1 | Stable (trixie) | 1 | Denial of service |
| thunderbird | DSA-6585-2 | 1:153.5.1esr-1~deb13u2 | Stable (trixie) | Regression (Debian Bug 1152139) | Restores PGP-encrypted email support |
| clamav | ELA-1867-1 | 1.4.6+dfsg-0+deb10u1 (buster), 1.4.6+dfsg-0+deb11u1 (bullseye) | Extended LTS | 13 | DoS (and unspecified impact) on crafted files |
| xorg-server | ELA-1866-1 | 2:1.19.2-1+deb9u24 (stretch), 2:1.20.4-1+deb10u19 (buster) | Extended LTS | 15 | Privilege escalation, DoS, information disclosure |
Fedora Linux
Fedora spent its Tuesday morning handing out security fixes and did not hold back. Both Fedora 43 and Fedora 44 got hit in the same window, though they dropped a couple of minutes apart. Fedora 43 picked up a batch of thirteen packages, and Fedora 44 came away with fifteen.
The one that actually makes a noise is Chromium. The browser moved to 156.0.8078.47 with a mountain of 353 fixes packed into it. That is a number that makes you want to click "just one more update" instead of reading what half of them are. Memory corruption and use-after-free variants show up across pretty much every component, from V8 to WebGL. If you have still not updated, this is the moment.
The kernel bumps on both releases are the usual stable release, this time with a bit of Fedora-specific wrangling to revert the af_alg_restrict sysctl. Nothing dramatic, but the tree got a scattering of fixes and you get them whether you run F43 or F44.
Then comes the grab bag of things you probably did not realize had vulnerabilities. OpenSSH took a denial-of-service fix for oversized decompressed packets, alongside an sftp recursive-copy file-write bug. util-linux got its libmount lockdown tightened and a descriptor leak in unshare patched. Ghostscript closed a -dSAFER sandbox bypass plus two buffer overflows, which is about as thrilling as a document format gets.
The LibreWAN entry is worth a second glance, because the advisory headline lists a CVE that does not match the changelog. The summary text reads CVE-2026-9445, but the change log and references point at CVE-2026-94453, an over-TCP use-after-free in IKEv2. Reporters will let that slide; you probably will not.
Several packages were rebuilt more than they were upgraded. rust-onefetch shows up in both releases, rebuilt against faster-hex 0.10.1 for the RUSTSEC-2026-0306 issue, as do libinstpatch and stargz-snapshotter. The mingw GStreamer variants all shifted to 1.28.8 together, carrying a pile of memory-safety fixes aimed at the Windows port. And Python, for what it is worth, got quarantined from a couple of compromised intermediate certificates plus a tar symlink trick that let writes escape its own directory.
| Package | Fedora | Version | Key fixes |
|---|---|---|---|
| kernel | 43 | 7.2.13-100.fc43 | Stable bump; af_alg_restrict sysctl revert |
| kernel | 44 | 7.2.13-200.fc44 | Stable bump; af_alg_restrict sysctl revert |
| chromium | 43 | 156.0.8078.47-1.fc43 | 353 fixes (CVE-2026-112992 through -113344) |
| flatpak | 43 | 1.16.6-2.fc43 | CVE-2026-96807, CVE-2026-96808 (symlink manipulation / priv-esc) |
| webkitgtk | 43 | 2.54.2-1.fc43 | Perf, process leak, crashes, /dev/dxg sandbox bind |
| eog | 43 | 49.3-2.fc43 | CVE-2026-107651 (PNG heap overflow, arbitrary code exec) |
| libreswan | 43 | 5.4.1-1.fc43 | CVE-2026-77205, -77206, -77207, -94453 |
| openssh | 43 | 10.0p1-14.fc43 | CVE-2026-106585, CVE-2026-106552 |
| util-linux | 43 | 2.41.6-1.fc43 | CVE-2026-76642, -78410, -78408 |
| perl-DBI | 43 | 1.656-1.fc43 | CVE-2026-108518, -111879 |
| rust-onefetch | 43 | 2.28.1-5.fc43 | faster-hex 0.10.1 for RUSTSEC-2026-0306 |
| stargz-snapshotter | 43 | 0.19.0-1.fc43 | CVE-2026-71482, -77395 |
| libinstpatch | 43 | 1.1.8-1.fc43 | CVE-2026-111136 + memory leak |
| ghostscript | 43 | 10.05.1-7.fc43 | CVE-2026-39919, -101258, -103226 |
| python3.12 | 44 | 3.12.16-1.fc44 | CVE-2026-114583, -115557 |
| perl-Alien-Brotli | 44 | 0.2.2-14.fc44 | CVE-2026-111497 |
| brotli | 44 | 1.2.1-1.fc44 | CVE-2026-111497 |
| clevis | 44 | 24-1.fc44 | CVE-2026-109968 through -109971 and -109976 |
| util-linux | 44 | 2.41.6-1.fc44 | CVE-2026-76642, -78410, -78408 |
| mingw-freetype | 44 | 2.14.3-3.fc44 | CVE-2026-95512 |
| perl-DBI | 44 | 1.656-1.fc44 | CVE-2026-108518, -111879 |
| mingw-gstreamer1-plugins-bad-free | 44 | 1.28.8-1.fc44 | CVE-2026-19387 + others (see note) |
| mingw-gstreamer1 | 44 | 1.28.8-1.fc44 | Same GStreamer batch |
| mingw-gstreamer1-plugins-base | 44 | 1.28.8-1.fc44 | Same GStreamer batch |
| mingw-gstreamer1-plugins-good | 44 | 1.28.8-1.fc44 | Same GStreamer batch |
| rust-onefetch | 44 | 2.28.1-5.fc44 | faster-hex 0.10.1 for RUSTSEC-2026-0306 |
| stargz-snapshotter | 44 | 0.19.0-1.fc44 | CVE-2026-71482, -77395 |
| libinstpatch | 44 | 1.1.8-1.fc44 | CVE-2026-111136 + memory leak |
Oracle Linux
Oracle Linux rolled out another batch of errata, and as usual the kernel is doing most of the carrying. Between two kernel updates for Oracle Linux 8 and one for Oracle Linux 9, you are looking at a stack of network, filesystem, and memory-safety fixes. The kernel rows carry the longest CVE lists here, so if you happen to run these machines you may want to schedule the reboot before anything else.
The interesting parts are the ones that repeat across packages. The Brotli out-of-bounds write (CVE-2026-111497) gets fixed in brotli itself for OL8 and OL10, then gets backported into the Node.js 22 and Node.js 24 bundles on OL8 and OL9. That is a decoder gap Oracle clearly did not want to leave open for long. The Flatpak path-traversal fixes (CVE-2026-97023 and CVE-2026-97024) show up on both OL8 and OL9, and the xdg-dbus-proxy sandbox-escape hole (CVE-2026-94422) hits OL9 and OL10. Firefox lands its usual ESR bump to 153.5.0 on OL8 and OL10, which brings the expected pile of security CVEs along for the ride.
A couple of rows deserve a second look for different reasons. The gcc-toolset-15-gcc update is marked Moderate and clears a single libstdc++ CVE, so it is a small ticket. The swtpm entry is a plain bug-fix job that swaps a SELinux tunable for a bool, with no CVE chase at all. Meanwhile the OL8 kernel update flagged Moderate (ELSA-2026-89816) carries essentially one real fix, a KVM/s390 patch, which is unusually quiet for a kernel row. Firefox's security count is the kind of number that makes the security team grin and makes anyone reading the changelog tired.
| Advisory | OS | Package | Severity | What changed |
|---|---|---|---|---|
| ELSA-2026-89859 | OL 8 | flatpak | Important | CVE-2026-97023 (path-traversal deletion), CVE-2026-97024 (fd-relative deploy) |
| ELSA-2026-89756 | OL 8 | firefox | Important | CVE-2026-45186 plus CVE-2026-109704–109782 range; bumps to 153.5.0 ESR |
| ELSA-2026-89083 | OL 8 | kernel 4.18.0-553.178.1 | Important | ~13 CVEs (selinux, ipv4, smb, mptcp, fuse, ipvs, usb, writeback, net/packet) |
| ELSA-2026-90449 | OL 10 | brotli | Important | CVE-2026-111497 (out-of-bounds write in decoder) |
| ELSA-2026-89891 | OL 10 | gcc-toolset-15-gcc | Moderate | CVE-2026-102010 (libstdc++ PR 127656) |
| ELSA-2026-90462 | OL 9 | skopeo | Important | CVE-2025-11395, CVE-2026-78667, CVE-2026-111943 |
| ELSA-2026-89816 | OL 8 | kernel 4.18.0-553.179.1 | Moderate | CVE-2026-74515 (KVM/s390 pci); gfs2 cleanup bundled in |
| ELBA-2026-90801 | OL 10 | swtpm | Bug fix | SELinux policy: bool instead of tunable for virt_use_nfs |
| ELSA-2026-87805 | OL 9 | kernel 5.14.0-687.60.1 | Important | CVE-2025-22063, CVE-2025-38147, CVE-2026-64481, CVE-2026-68202, CVE-2026-68299, CVE-2026-68448, CVE-2026-80530, CVE-2026-80536 |
| ELSA-2026-89758 | OL 10 | firefox | Important | CVE-2026-45186, CVE-2026-77214 plus CVE-2026-109704–109740 range; bumps to 153.5.0 ESR |
| ELSA-2026-500519 | OL 8 | nodejs:22 | Important | CVE-2026-111497 backported into bundled Brotli |
| ELSA-2026-500518 | OL 9 | nodejs:24 | Important | CVE-2026-111497 backported into bundled Brotli |
| ELSA-2026-90492 | OL 8 | libreswan | Important | CVE-2026-77206, CVE-2026-77207, CVE-2026-94453 |
| ELSA-2026-500520 | OL 8 | nodejs:24 | Important | CVE-2026-111497 backported into bundled Brotli |
| ELSA-2026-90494 | OL 9 | python-urllib3 | Important | CVE-2026-97687 (HTTPS proxy TLS override), CVE-2026-97689 (DoS in chunk parser) |
| ELSA-2026-90523 | OL 10 | xdg-dbus-proxy | Important | CVE-2026-94422 (reply-serial bypass lets sandbox be escaped) |
| ELSA-2026-89850 | OL 9 | flatpak | Important | CVE-2026-97023, CVE-2026-97024 |
| ELSA-2026-90522 | OL 9 | xdg-dbus-proxy | Important | CVE-2026-94422 |
| ELSA-2026-90501 | OL 8 | brotli | Important | CVE-2026-111497 |
Oracle Linux
Oracle Linux rolled out another batch of errata, and as usual the kernel is doing most of the carrying. Between two kernel updates for Oracle Linux 8 and one for Oracle Linux 9, you are looking at a stack of network, filesystem, and memory-safety fixes. The kernel rows carry the longest CVE lists here, so if you happen to run these machines you may want to schedule the reboot before anything else.
The interesting parts are the ones that repeat across packages. The Brotli out-of-bounds write (CVE-2026-111497) gets fixed in brotli itself for OL8 and OL10, then gets backported into the Node.js 22 and Node.js 24 bundles on OL8 and OL9. That is a decoder gap Oracle clearly did not want to leave open for long. The Flatpak path-traversal fixes (CVE-2026-97023 and CVE-2026-97024) show up on both OL8 and OL9, and the xdg-dbus-proxy sandbox-escape hole (CVE-2026-94422) hits OL9 and OL10. Firefox lands its usual ESR bump to 153.5.0 on OL8 and OL10, which brings the expected pile of security CVEs along for the ride.
A couple of rows deserve a second look for different reasons. The gcc-toolset-15-gcc update is marked Moderate and clears a single libstdc++ CVE, so it is a small ticket. The swtpm entry is a plain bug-fix job that swaps a SELinux tunable for a bool, with no CVE chase at all. Meanwhile the OL8 kernel update flagged Moderate (ELSA-2026-89816) carries essentially one real fix, a KVM/s390 patch, which is unusually quiet for a kernel row. Firefox's security count is the kind of number that makes the security team grin and makes anyone reading the changelog tired.
| Advisory | OS | Package | Severity | What changed |
|---|---|---|---|---|
| ELSA-2026-89859 | OL 8 | flatpak | Important | CVE-2026-97023 (path-traversal deletion), CVE-2026-97024 (fd-relative deploy) |
| ELSA-2026-89756 | OL 8 | firefox | Important | CVE-2026-45186 plus CVE-2026-109704–109782 range; bumps to 153.5.0 ESR |
| ELSA-2026-89083 | OL 8 | kernel 4.18.0-553.178.1 | Important | ~13 CVEs (selinux, ipv4, smb, mptcp, fuse, ipvs, usb, writeback, net/packet) |
| ELSA-2026-90449 | OL 10 | brotli | Important | CVE-2026-111497 (out-of-bounds write in decoder) |
| ELSA-2026-89891 | OL 10 | gcc-toolset-15-gcc | Moderate | CVE-2026-102010 (libstdc++ PR 127656) |
| ELSA-2026-90462 | OL 9 | skopeo | Important | CVE-2025-11395, CVE-2026-78667, CVE-2026-111943 |
| ELSA-2026-89816 | OL 8 | kernel 4.18.0-553.179.1 | Moderate | CVE-2026-74515 (KVM/s390 pci); gfs2 cleanup bundled in |
| ELBA-2026-90801 | OL 10 | swtpm | Bug fix | SELinux policy: bool instead of tunable for virt_use_nfs |
| ELSA-2026-87805 | OL 9 | kernel 5.14.0-687.60.1 | Important | CVE-2025-22063, CVE-2025-38147, CVE-2026-64481, CVE-2026-68202, CVE-2026-68299, CVE-2026-68448, CVE-2026-80530, CVE-2026-80536 |
| ELSA-2026-89758 | OL 10 | firefox | Important | CVE-2026-45186, CVE-2026-77214 plus CVE-2026-109704–109740 range; bumps to 153.5.0 ESR |
| ELSA-2026-500519 | OL 8 | nodejs:22 | Important | CVE-2026-111497 backported into bundled Brotli |
| ELSA-2026-500518 | OL 9 | nodejs:24 | Important | CVE-2026-111497 backported into bundled Brotli |
| ELSA-2026-90492 | OL 8 | libreswan | Important | CVE-2026-77206, CVE-2026-77207, CVE-2026-94453 |
| ELSA-2026-500520 | OL 8 | nodejs:24 | Important | CVE-2026-111497 backported into bundled Brotli |
| ELSA-2026-90494 | OL 9 | python-urllib3 | Important | CVE-2026-97687 (HTTPS proxy TLS override), CVE-2026-97689 (DoS in chunk parser) |
| ELSA-2026-90523 | OL 10 | xdg-dbus-proxy | Important | CVE-2026-94422 (reply-serial bypass lets sandbox be escaped) |
| ELSA-2026-89850 | OL 9 | flatpak | Important | CVE-2026-97023, CVE-2026-97024 |
| ELSA-2026-90522 | OL 9 | xdg-dbus-proxy | Important | CVE-2026-94422 |
| ELSA-2026-90501 | OL 8 | brotli | Important | CVE-2026-111497 |
Red Hat Enterprise Linux
A fresh round of security advisories dropped from Red Hat and Oracle, and if you run anything on the list, it's worth patching. Most ratings land at Important, with several marked Critical. The one with the most teeth is Oracle Linux 10's firefox update, which nails down nearly 30 issues and pushes you onto the 153.5.0 ESR line.
Red Hat clearly had a lot of OpenShift to push. You'll see the same versions, 4.16.74, 4.14.77, 4.18.59, 4.12.102, and 4.21.40, slither out both as bug-fix releases and as packages-and-security advisories. A couple of the OpenShift headers say Critical while the body rates the security piece as Important. That split is standard Red Hat bookkeeping: the title reflects the whole release, the body reflects the vulnerability alone. A Critical header here doesn't mean the CVE is more dangerous.
Brotli appears four times across different RHEL streams, with emacs and rhc showing up twice each. Don't read into it, a single flaw getting patched through multiple channels looks like a spam problem, but it just means the same fix is being funneled to different customer feeds. The rest of the usual suspects are in the mix: java-17-openjdk for RHEL 8, thunderbird for RHEL 10.2, python-tornado, xdg-dbus-proxy, Quarkus, and webkit2gtk3, which is one of the heavier Criticals.
| Advisory | Component | Rated severity | Covers |
|---|---|---|---|
| RHSA-2026:90725 | rhc | Important | RHEL 8.6 Advanced Mission Critical + EUS Long-Life |
| ELSA-2026-89758 | firefox | Important | Oracle Linux 10; fixes ~28 CVEs, bumps to 153.5.0 ESR |
| RHSA-2026:89714 | OpenShift 4.16.74 | Critical (Important in body) | Bug fix + security for 4.16 |
| RHSA-2026:90272 | OpenShift 4.12.102 | Critical | Bug fix + security for 4.12 |
| RHSA-2026:90887 | emacs | Important | RHEL 10 |
| RHSA-2026:89713 | OpenShift 4.16.74 packages | Important | Packages + security for 4.16 |
| RHSA-2026:90895 | thunderbird | Important | RHEL 10.2 EUS |
| RHSA-2026:88352 | java-17-openjdk | Important | RHEL 8.6 AMCUIS/Update Services for SAP/Telecommunications, 8.4 variants, 8.8 EUS, RHEL 8 |
| RHSA-2026:89837 | OpenShift 4.14.77 packages | Important | Packages + security for 4.14 |
| RHSA-2026:89838 | OpenShift 4.14.77 | Critical | Bug fix + security for 4.14 |
| RHSA-2026:90914 | brotli | Important | RHEL 9.2 Update Services for SAP |
| RHSA-2026:90892 | brotli | Important | RHEL 8.4 Advanced Mission Critical + EUS Long-Life |
| RHSA-2026:90311 | Quarkus 3.33.4.SP1 | Important | Red Hat build of Quarkus |
| RHSA-2026:89725 | OpenShift 4.18.59 | Critical (Important in body) | Bug fix + security for 4.18 |
| RHSA-2026:90888 | brotli | Important | RHEL 9.4 Update Services for SAP |
| RHSA-2026:90893 | brotli | Important | RHEL 8.6 Advanced Mission Critical + EUS Long-Life |
| RHSA-2026:90800 | emacs | Important | RHEL 8 |
| RHSA-2026:90777 | python-tornado | Important | RHEL 9.4 Update Services for SAP |
| RHSA-2026:89822 | webkit2gtk3 | Critical | RHEL 9.2 Update Services for SAP |
| RHSA-2026:90726 | rhc | Important | RHEL 8.8 Update Services for SAP + Telecommunications |
| RHSA-2026:89229 | OpenShift 4.21.40 packages | Important | Packages + security for 4.21 |
| RHSA-2026:90522 | xdg-dbus-proxy | Important | RHEL 9.8 EUS |
Rocky Linux
Rocky Linux issued 20 security errata, and exactly one carries a critical rating. It's grafana on Rocky Linux 9 (RLSA-2026:89744), the lone exception in a batch that otherwise tops out at "Important." The advisories span releases 8, 9, and 10.
By severity you've got one critical, thirteen Important, and six Moderate. The affected packages cover the usual suspects for any Linux host: httpd and bind for web and DNS, libreswan for VPN, the container-adjacent trio of skopeo, oci-seccomp-bpf-hook, and flatpak, plus grafana for monitoring and xz for compression. kernel-rt pulls in the real-time kernel, which some of you run for latency-sensitive work.
A few names recur. flatpak gets patched on all three releases, which is unsurprising given how broadly it's used. grafana shows up twice, so watch both: the Critical one on RL9 and the Important one on RL10 if you happen to run both. The xz entry is the one most likely to make you squint, given the compressor's messy backdoor history, even though this advisory doesn't spell out what's actually fixed.
None of these notices hand you a CVSS score or a CVE name. Those sit behind the errata links, so take the table below, cross it against what you actually run, and patch from there.
| Advisory ID | Package | Severity | Release | Scope |
|---|---|---|---|---|
| RLSA-2026:89761 | skopeo | Important | 10 | Security |
| RLSA-2026:89905 | freeipmi | Moderate | 10 | Security + bug fixes/enhancements |
| RLSA-2026:89752 | oci-seccomp-bpf-hook | Important | 10 | Security |
| RLSA-2026:89923 | libreswan | Important | 10 | Security |
| RLSA-2026:89732 | grafana | Important | 10 | Security |
| RLSA-2026:89849 | flatpak | Important | 10 | Security |
| RLSA-2026:89860 | httpd | Important | 10 | Security |
| RLSA-2026:89755 | lrzsz | Important | 9 | Security + bug fixes/enhancements |
| RLSA-2026:89744 | grafana | Critical | 9 | Security |
| RLSA-2026:89742 | python3.9 | Important | 9 | Security |
| RLSA-2026:89749 | xz | Moderate | 9 | Security |
| RLSA-2026:89748 | libreswan | Important | 9 | Security |
| RLSA-2026:89746 | bind9.18 | Moderate | 9 | Security |
| RLSA-2026:89745 | bind | Moderate | 9 | Security |
| RLSA-2026:89850 | flatpak | Important | 9 | Security |
| RLSA-2026:89753 | freeipmi | Moderate | 9 | Security + bug fixes/enhancements |
| RLSA-2026:89814 | kernel-rt | Moderate | 8 | Security |
| RLSA-2026:89859 | flatpak | Important | 8 | Security |
| RLSA-2026:89730 | libX11 | Important | 8 | Security |
| RLSA-2026:89747 | xorg-x11-server | Important | 8 | Security |
Slackware Linux
Slackware has pushed a security fix for proftpd, and if you run it, this is worth a couple of minutes to get onto.
The update lands on both Slackware 15.0 and -current, both sitting at 1.3.9e now. What actually changed? ProFTPD gained the ability to use Huffman compression over SSH, but only for SSH compression, which closes a hole where a chosen-plaintext attack could target concurrent channels sharing the same SSH connection. In other words, if multiple channels were piggybacking on one connection, they were potentially exposed. Slackware points at GitHub issue 2366 for the full detail.
Nothing earth-shattering in the change itself, but the fact that it's opt-in means your existing setups won't shift under you unless you deliberately turn it on.
| Slackware version | Architecture | Package | MD5 |
|---|---|---|---|
| 15.0 | i586 | proftpd-1.3.9e-i586-1_slack15.0.txz | 0520380846c0214b6755778b96dbb7d6 |
| 15.0 | x86_64 | proftpd-1.3.9e-x86_64-1_slack15.0.txz | 3f44012a5e5a4dddd2b9b80dcd590944 |
| -current | i686 | proftpd-1.3.9e-i686-1.txz | 9710872e479d2cfa884ac6e6ebb76bb7 |
| -current | x86_64 | proftpd-1.3.9e-x86_64-1.txz | 1b91a0a1ece3a496b9b88fa013af9091 |
SUSE Linux
openSUSE and SUSE pushed a batch of security fixes this round, and one entry is doing most of the talking. Announcement openSUSE-SU-2026:0360-1 is rated important and pushes Chromium 156 (156.0.8078.47) to stable for openSUSE Backports SLE-15-SP7, plugging 353 separate vulnerabilities. That is a staggering count of CVEs for what is essentially a browser bump. Most are the usual suspects: memory corruption, use-after-free, and out-of-bounds reads across ANGLE, V8, and GPU rendering, plus a dozen other subsystems. A gn build-tool update rides along in the same release, though its changelog reads more like an internal tidy-up than anything you'd actually notice.
The rest of the batch is more mundane, falling into two camps. Four more important announcements, suseconnect-ng (4918) and three podman releases (4920, 4921, 4922), are simply rebuilds against the current Go security release, spread across SLE 15 SP4 through SP7. There is little to celebrate, but they are the boring kind of patch you would rather have installed.
The larger group is a run of moderate Tumbleweed updates for the original release media, each patching one package. ImageMagick takes the volume prize with 13 fixes, Xen follows with five, and libsoup carries two. Twisted, corepack, psqlODBC, azure-storage-azcopy, M2Crypto, and matrix-synapse each fix just one.
| Announcement ID | Package(s) | Rating | Vulnerabilities | Platform |
|---|---|---|---|---|
| openSUSE-SU-2026:0360-1 | chromium, gn | important | 353 CVEs | openSUSE Backports SLE-15-SP7 |
| SUSE-SU-2026:4918-1 | suseconnect-ng | important | Go rebuild, no CVE listed | SLE 15 SP5 (several variants) |
| SUSE-SU-2026:4920-1 | podman | important | Go rebuild, no CVE listed | SLE 15 SP5/SP6/SP7 + others |
| SUSE-SU-2026:4921-1 | podman | important | Go rebuild, no CVE listed | SLE 15 SP4 |
| SUSE-SU-2026:4922-1 | podman | important | Go rebuild, no CVE listed | openSUSE Leap 15.3, SUSE Storage 7.1 |
| openSUSE-SU-2026:12132-1 | libsoup-2_4-1 | moderate | 2 (CVE-2026-103399, CVE-2026-15713) | openSUSE Tumbleweed |
| openSUSE-SU-2026:12133-1 | matrix-synapse | moderate | 1 (CVE-2026-69248) | openSUSE Tumbleweed |
| openSUSE-SU-2026:12134-1 | corepack24 | moderate | 1 (CVE-2026-54272) | openSUSE Tumbleweed |
| openSUSE-SU-2026:12135-1 | psqlODBC | moderate | 1 (CVE-2026-109448) | openSUSE Tumbleweed |
| openSUSE-SU-2026:12136-1 | python-M2Crypto-doc | moderate | 1 (CVE-2026-109087) | openSUSE Tumbleweed |
| openSUSE-SU-2026:12137-1 | python-Twisted-doc | moderate | 1 (CVE-2026-106454) | openSUSE Tumbleweed |
| openSUSE-SU-2026:12138-1 | xen | moderate | 5 (CVE-2026-79607-79611) | openSUSE Tumbleweed |
| openSUSE-SU-2026:12139-1 | ImageMagick | moderate | 13 | openSUSE Tumbleweed |
| openSUSE-SU-2026:12140-1 | azure-storage-azcopy | moderate | 1 (CVE-2026-78669) | openSUSE Tumbleweed |
| openSUSE-SU-2026:12141-1 | grok-build | moderate | 3 (CVE-2026-109349-109351) | openSUSE Tumbleweed |
Ubuntu Linux
Ubuntu just dropped a pretty ordinary-looking hole in authd, the authentication daemon that shells out to your cloud identity provider. If you run Ubuntu 26.10 off that setup, it's worth a glance. Ali Mashally found that authd wasn't properly validating user identities when it matched accounts arriving from identity providers, which in practice means a local attacker could potentially climb onto someone else's account and read their files. Not ideal on a machine you were fairly sure was locked up.
The repair is a single-package affair: bump authd to version 0.7.1, then restart the daemon so the patch actually bites. It's logged under CVE-2026-102631, and the full writeup lives on ubuntu.com if you want to see exactly how the mismatch played out.
| Release | Package | New version | CVE | After the update |
|---|---|---|---|---|
| Ubuntu 26.10 | authd | 0.7.1 | CVE-2026-102631 | Restart authd so the fix takes effect |
How to apply these Linux security updates
Before running any update commands, check which services are currently active on your system. If Nginx or Apache is handling live traffic, schedule a brief maintenance window or use rolling restarts to minimize downtime during the patching process. Desktop users can usually apply these fixes by opening a terminal and running the standard package manager command for their distribution followed by an upgrade flag. A reboot will be necessary if the kernel received updates to ensure the new security modules load correctly.
Power users who rely on command-line tools like jq should verify the patch level after installation. Regression bugs can occasionally break scripts that depend on specific JSON parsing behavior, so a quick test run is worth the few minutes it takes. If you use PackageKit or other GUI package managers and prefer to skip them because they sometimes hang or try to install junk, do not let that stop you from running the command-line equivalent to get these critical patches applied.
Applying these patches requires distribution-specific package management commands. RHEL-based systems typically use dnf update or yum update, while Debian and Ubuntu rely on apt upgrade. SUSE users should run zypper patch to properly address all security advisories, and Slackware administrators can manage updates with upgradepkg or slackpkg. After executing the commands, a reboot is usually necessary for kernel changes to take effect. Finally, review your package manager’s logs to verify that all patches installed successfully and no dependencies were disrupted.
Debian/Ubuntu (apt)
The first thing to do is refresh the local package index; running sudo apt update contacts all configured repositories and pulls in the newest lists of available versions. Skipping this step leaves the system blind to any recent uploads, which explains why “upgrade” sometimes claims there’s nothing to do even after a security advisory has been published. Once the index is current, invoke sudo apt upgrade -y; the -y flag answers every prompt automatically so the process doesn’t pause for user input. This command upgrades all installed packages that have newer versions in the repositories while preserving configuration files.
sudo apt update sudo apt upgrade -y
Fedora/RedHat/Rocky/Alma/Oracle (dnf or yum)
On modern Fedora and recent Red Hat derivatives, dnf is the package manager; older RHEL releases still rely on yum. Begin with a check‑update operation—sudo dnf check-update or sudo yum check-update—to see exactly which packages are awaiting an upgrade. This preview step can be useful for spotting unexpected kernel bumps before they land. To actually apply the updates, run sudo dnf upgrade -y (or sudo yum update if you prefer the older tool). The upgrade command pulls down the new binaries and runs any necessary post‑install scripts, such as rebuilding initramfs when a kernel changes.
sudo dnf check-update sudo dnf upgrade -y
or on older releases
sudo yum check-update sudo yum update
SUSE (zypper)
SUSE’s command line front‑end is called zypper. First execute sudo zypper refresh so that the metadata for all enabled repos gets updated; without this, zypper will happily report “No updates available” even though newer packages sit on the mirror. After a fresh refresh, issue sudo zypper update -y; this upgrades every package to the latest version in the configured repositories and automatically handles service restarts when required.
sudo zypper refresh sudo zypper update -y
Slackware (slackpkg and pkgtool)
Slackware doesn’t have a single unified updater, but the official way to pull updates is through slackpkg. Start with sudo slackpkg update to download the newest package list from the chosen mirror. Then run sudo slackpkg upgrade-all; this command walks through each installed package and replaces it with the most recent build available in the official repository. For users who prefer a more granular approach, specifying a package name after upgrade limits the operation to that single item. When dealing with community‑maintained repositories, pkgtool takes over: a combined sudo pkgtool update && sudo pkgtool upgrade will sync and apply updates from the mirrors listed in /etc/slackpkg/mirrors.
sudo slackpkg update sudo slackpkg upgrade-all
