kernel infinite loop on Kerberos CIFS mount when ticket expires

[Russell Coker]

I mount a CIFS filesystem with Kerberos after logging in with an Active Directory domain account (which creates the Kerberos ticket with a 10 hour expiry - the default for Active Directory). When the ticket expires I get lots of messages like the following and a kernel thread for CIFS using 100% of a CPU core. This problem is not resolved on umount or after any reasonable length of time (I've had machines in that state for over a week).

This happens with Ubuntu 22.04 with the latest kernel as well as with 20.04 with the generic kernel, the OEM kernel, and the HWE kernel.

Nov 3 07:48:56 u3d05ng3 kernel: [81347.750474] CIFS: VFS: Verify user has a krb5 ticket and keyutil
s is installed
Nov 3 07:48:56 u3d05ng3 kernel: [81347.750475] CIFS: VFS: \\AD-6-ACT.sm.local Send error in SessSet
up = -126
Nov 3 07:48:56 u3d05ng3 kernel: [81347.750480] CIFS: VFS: Verify user has a krb5 ticket and keyutil
s is installed
Nov 3 07:48:56 u3d05ng3 kernel: [81347.750481] CIFS: VFS: \\AD-6-ACT.sm.local Send error in SessSet
up = -126
Nov 3 07:48:56 u3d05ng3 kernel: [81347.750485] CIFS: VFS: Verify user has a krb5 ticket and keyutil
s is installed
Nov 3 07:48:56 u3d05ng3 kernel: [81347.750487] CIFS: VFS: \\AD-6-ACT.sm.local Send error in SessSet
up = -126
Nov 3 07:49:01 u3d05ng3 kernel: [81352.754469] SMB2_auth_kerberos: 1423149 callbacks suppressed
Nov 3 07:49:01 u3d05ng3 kernel: [81352.754473] CIFS: VFS: Verify user has a krb5 ticket and keyutil
s is installed
Nov 3 07:49:01 u3d05ng3 kernel: [81352.754477] cifs_setup_session: 1423149 callbacks suppressed
Nov 3 07:49:01 u3d05ng3 kernel: [81352.754479] CIFS: VFS: \\AD-6-ACT.sm.local Send error in SessSet
up = -126
Nov 3 07:49:01 u3d05ng3 kernel: [81352.754484] CIFS: VFS: Verify user has a krb5 ticket and keyutil
s is installed

ProblemType: Bug
DistroRelease: Ubuntu 22.04
Package: linux-image-5.15.0-52-generic 5.15.0-52.58
ProcVersionSignature: Ubuntu 5.15.0-52.58-generic 5.15.60
Uname: Linux 5.15.0-52-generic x86_64
ApportVersion: 2.20.11-0ubuntu82.1
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: russell.coker 2859 F.... pulseaudio
CasperMD5CheckResult: unknown
CurrentDesktop: KDE
Date: Thu Nov 3 11:44:17 2022
InstallationDate: Installed on 2022-05-13 (173 days ago)
InstallationMedia: Ubuntu 20.04.2.0 LTS "Focal Fossa" - Release amd64 (20210209.1)
MachineType: Dell Inc. Precision 3560
ProcFB: 0 i915drmfb
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-5.15.0-52-generic root=UUID=8d80bec3-0520-4486-aac5-2b688f71ce12 ro rootflags=subvol=22.04 nosmt
RelatedPackageVersions:
 linux-restricted-modules-5.15.0-52-generic N/A
 linux-backports-modules-5.15.0-52-generic N/A
 linux-firmware 20220329.git681281e4-0ubuntu3.6
SourcePackage: linux
UpgradeStatus: Upgraded to jammy on 2022-07-01 (124 days ago)
WifiSyslog:

dmi.bios.date: 09/07/2022
dmi.bios.release: 1.23
dmi.bios.vendor: Dell Inc.
dmi.bios.version: 1.23.0
dmi.board.name: 04RH8K
dmi.board.vendor: Dell Inc.
dmi.board.version: A00
dmi.chassis.type: 10
dmi.chassis.vendor: Dell Inc.
dmi.modalias: dmi:bvnDellInc.:bvr1.23.0:bd09/07/2022:br1.23:svnDellInc.:pnPrecision3560:pvr:rvnDellInc.:rn04RH8K:rvrA00:cvnDellInc.:ct10:cvr:sku0A22:
dmi.product.family: Precision
dmi.product.name: Precision 3560
dmi.product.sku: 0A22
dmi.sys.vendor: Dell Inc.

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This change was made by a bot.

[Russell Coker (etbe-sm)]

Using kernels 5.15.0-67-generic and 5.19.0-35-generic I can no longer reproduce the problem of a CPU core going to an infinite busy loop. With these kernels the situation when the ticket expires is that 2 messages are logged every 2 seconds (which seems excessive but isn't a DoS) until the filesystem is umounted OR the user generates a new ticket (which includes running "kinit", logging in again for ssh sessions, and unlocking the screen for GNOME and KDE sessions). When the user generates a new ticket it is noticed by the kernel and the files become available again.

I think that messages are logged too frequently, but the situation otherwise works as expected now.

I don't know when this bug was fixed, the kernel package changelogs have lots of entries about pulling from other sources and I didn't try to chase down the other changelogs.

I don't know which other Ubuntu kernels have this bug fixed. All kernels used by systems I run have it fixed so for me it's a solved problem. Other people may have different experiences.