Package "linux-nvidia-tegra"

  linux-nvidia-tegra (5.15.0-1051.51) jammy; urgency=medium

  * jammy/linux-nvidia-tegra: 5.15.0-1051.51 -proposed tracker (LP: #2138013)

  [ Ubuntu-realtime: 5.15.0-1098.107 ]

  * jammy/linux-realtime: 5.15.0-1098.107 -proposed tracker (LP: #2138017)
  [ Ubuntu: 5.15.0-168.178 ]
  * jammy/linux: 5.15.0-168.178 -proposed tracker (LP: #2138043)
  * CVE-2025-40019
    - crypto: essiv - Check ssize for decryption and in-place encryption
  * Black screen when booting 5.15.0-160 (on AMD Lucienne / Cezanne / Navi /
    Renoir / Rembrandt) (LP: #2128729)
    - SAUCE: drm/amd/display: Fix incorrect code path taken in
      amdgpu_dm_atomic_check()
  * CVE-2025-38561
    - ksmbd: fix Preauh_HashValue race condition

 -- Noah Wager <email address hidden> Tue, 20 Jan 2026 10:42:36 -0800

  linux-nvidia-tegra (5.15.0-1050.50) jammy; urgency=medium

  * jammy/linux-nvidia-tegra: 5.15.0-1050.50 -proposed tracker (LP: #2131399)

  [ Ubuntu-realtime: 5.15.0-1097.106 ]

  * jammy/linux-realtime: 5.15.0-1097.106 -proposed tracker (LP: #2131403)
  [ Ubuntu: 5.15.0-164.174 ]
  * jammy/linux: 5.15.0-164.174 -proposed tracker (LP: #2131429)
  * CVE-2024-53218
    - f2fs: fix race in concurrent f2fs_stop_gc_thread
  * CVE-2024-47691
    - f2fs: fix to avoid use-after-free in f2fs_stop_gc_thread()
  * CVE-2025-39993
    - media: rc: fix races with imon_disconnect()
  * i40e driver is triggering VF resets on every link state change
    (LP: #2130552)
    - i40e: avoid redundant VF link state updates
  * CVE-2025-40018
    - ipvs: Defer ip_vs_ftp unregister during netns cleanup
  * CVE-2025-21855
    - ibmvnic: Don't reference skb after sending to VIOS
  * CVE-2024-50067
    - uprobes: encapsulate preparation of uprobe args buffer
    - uprobe: avoid out-of-bounds memory access of fetching args
  * CVE-2024-53090
    - afs: Fix lock recursion
  * CVE-2025-39964
    - crypto: af_alg - Disallow concurrent writes in af_alg_sendmsg
    - crypto: af_alg - Fix incorrect boolean values in af_alg_ctx
  * CVE-2022-49390
    - macsec: fix UAF bug for real_dev

 -- Noah Wager <email address hidden> Tue, 25 Nov 2025 13:18:37 -0800

  linux-nvidia-tegra (5.15.0-1049.49) jammy; urgency=medium

  * jammy/linux-nvidia-tegra: 5.15.0-1049.49 -proposed tracker (LP: #2127542)

  [ Ubuntu-realtime: 5.15.0-1096.105 ]

  * jammy/linux-realtime: 5.15.0-1096.105 -proposed tracker (LP: #2127546)
  * Jammy real-time patch set update: v5.15.193-rt89 (LP: #2130278)
    - rcu/tree: Protect rcu_rdp_is_offloaded() invocations on RT
    - sched: Introduce migratable()
    - arm64: mm: Make arch_faults_on_old_pte() check for migratability
    - printk: rename printk cpulock API and always disable interrupts
    - console: add write_atomic interface
    - kdb: only use atomic consoles for output mirroring
    - serial: 8250: implement write_atomic
    - printk: relocate printk_delay()
    - printk: call boot_delay_msec() in printk_delay()
    - printk: use seqcount_latch for console_seq
    - printk: introduce kernel sync mode
    - printk: move console printing to kthreads
    - printk: add console handover
    - printk: add pr_flush()
    - printk: Enhance the condition check of msleep in pr_flush()
    - sched: Switch wait_task_inactive to HRTIMER_MODE_REL_HARD
    - kthread: Move prio/affinite change into the newly created thread
    - genirq: Move prio assignment into the newly created thread
    - genirq: Disable irqfixup/poll on PREEMPT_RT.
    - efi: Allow efi=runtime
    - mm: Disable zsmalloc on PREEMPT_RT
    - net/core: disable NET_RX_BUSY_POLL on PREEMPT_RT
    - samples/kfifo: Rename read_lock/write_lock
    - crypto: testmgr - Only disable migration in
      crypto_disable_simd_for_test()
    - mm: Allow only SLUB on PREEMPT_RT
    - mm: page_alloc: Use migrate_disable() in drain_local_pages_wq()
    - mm/scatterlist: Replace the !preemptible warning in sg_miter_stop()
    - mm: Disable NUMA_BALANCING_DEFAULT_ENABLED and TRANSPARENT_HUGEPAGE on
      PREEMPT_RT
    - x86/softirq: Disable softirq stacks on PREEMPT_RT
    - Documentation/kcov: Include types.h in the example.
    - Documentation/kcov: Define `ip' in the example.
    - kcov: Allocate per-CPU memory on the relevant node.
    - kcov: Avoid enable+disable interrupts if !in_task().
    - kcov: Replace local_irq_save() with a local_lock_t.
    - gen_stats: Add instead Set the value in __gnet_stats_copy_basic().
    - gen_stats: Add gnet_stats_add_queue().
    - mq, mqprio: Use gnet_stats_add_queue().
    - gen_stats: Move remaining users to gnet_stats_add_queue().
    - u64_stats: Introduce u64_stats_set()
    - net: sched: Protect Qdisc::bstats with u64_stats
    - net: sched: Use _bstats_update/set() instead of raw writes
    - net: sched: Merge Qdisc::bstats and Qdisc::cpu_bstats data types
    - net: sched: Remove Qdisc::running sequence counter
    - net: sched: Allow statistics reads from softirq.
    - net: sched: fix logic error in qdisc_run_begin()
    - net: sched: remove one pair of atomic operations
    - net: stats: Read the statistics in ___gnet_stats_copy_basic() instead of
      adding.
    - net: sched: gred: dynamically allocate tc_gred_qopt_offload
    - sched/rt: Annotate the RT balancing logic irqwork as IRQ_WORK_HARD_IRQ
    - irq_work: Allow irq_work_sync() to sleep if irq_work() no IRQ support.
    - irq_work: Handle some irq_work in a per-CPU thread on PREEMPT_RT
    - irq_work: Also rcuwait for !IRQ_WORK_HARD_IRQ on PREEMPT_RT
    - irq_poll: Use raise_softirq_irqoff() in cpu_dead notifier
    - smp: Wake ksoftirqd on PREEMPT_RT instead do_softirq().
    - fs/namespace: Boost the mount_lock.lock owner instead of spinning on
      PREEMPT_RT.
    - fscache: Use only one fscache_object_cong_wait.
    - sched: Clean up the might_sleep() underscore zoo
    - sched: Make cond_resched_*lock() variants consistent vs. might_sleep()
    - sched: Remove preempt_offset argument from __might_sleep()
    - sched: Cleanup might_sleep() printks
    - sched: Make might_sleep() output less confusing
    - sched: Make RCU nest depth distinct in __might_resched()
    - sched: Make cond_resched_lock() variants RT aware
    - locking/rt: Take RCU nesting into account for __might_resched()
    - sched: Limit the number of task migrations per batch on RT
    - sched: Disable TTWU_QUEUE on RT
    - sched: Move kprobes cleanup out of finish_task_switch()
    - sched: Delay task stack freeing on RT
    - sched: Move mmdrop to RCU on RT
    - cgroup: use irqsave in cgroup_rstat_flush_locked()
    - mm: workingset: replace IRQ-off check with a lockdep assert.
    - jump-label: disable if stop_machine() is used
    - locking: Remove rt_rwlock_is_contended()
    - lockdep/selftests: Avoid using local_lock_{acquire|release}().
    - sched: Trigger warning if ->migration_disabled counter underflows.
    - rtmutex: Add a special case for ww-mutex handling.
    - rtmutex: Add rt_mutex_lock_nest_lock() and rt_mutex_lock_killable().
    - lockdep: Make it RT aware
    - lockdep/selftests: Add rtmutex to the last column
    - lockdep/selftests: Unbalanced migrate_disable() & rcu_read_lock()
    - lockdep/selftests: Skip the softirq related tests on PREEMPT_RT
    - lockdep/selftests: Adapt ww-tests for PREEMPT_RT
    - locking: Allow to include asm/spinlock_types.h from
      linux/spinlock_types_raw.h
    - sched: Make preempt_enable_no_resched() behave like preempt_enable() on
      PREEMPT_RT
    - kernel/sched: add {put|get}_cpu_light()
    - block/mq: do not invoke preempt_disable()
    - md: raid5: Make raid5_percpu handling RT aware
    - scsi/fcoe: Make RT aware.
    - mm/vmalloc: Another preempt disable region which sucks
    - net: Remove preemption disabling in netif_rx()
    - sunrpc: Make svc_xprt_do_enqueue() use get_cpu_light()
    - softirq: Check preemption after reenabling interrupts
    - mm/memcontrol: Disable on PREEMPT_RT
    - signal: Revert ptrace preempt magic
    - ptrace: fix ptrace vs tasklist_lock race
    - fs/dcache: use swait_queue instead of waitqueue
    - fs/dcache: disable preemption on i_dir_seq's write side
    - rcu: Delay RCU-sel

  linux-nvidia-tegra (5.15.0-1048.48) jammy; urgency=medium

  * jammy/linux-nvidia-tegra: 5.15.0-1048.48 -proposed tracker (LP: #2127359)

  [ Ubuntu-realtime: 5.15.0-1095.104 ]

  * jammy/linux-realtime: 5.15.0-1095.104 -proposed tracker (LP: #2127363)
  [ Ubuntu: 5.15.0-161.171 ]
  * jammy/linux: 5.15.0-161.171 -proposed tracker (LP: #2127389)
  * VMSCAPE CVE-2025-40300 (LP: #2124105) // CVE-2025-40300
    - Documentation/hw-vuln: Add VMSCAPE documentation
    - x86/vmscape: Enumerate VMSCAPE bug
    - x86/vmscape: Add conditional IBPB mitigation
    - x86/vmscape: Enable the mitigation
    - x86/bugs: Move cpu_bugs_smt_update() down
    - x86/vmscape: Warn when STIBP is disabled with SMT
    - x86/vmscape: Add old Intel CPUs to affected list
  * VMSCAPE CVE-2025-40300 (LP: #2124105)
    - [Config] Enable MITIGATION_VMSCAPE config

 -- Noah Wager <email address hidden> Mon, 20 Oct 2025 07:27:28 -0700

  linux-nvidia-tegra (5.15.0-1047.47) jammy; urgency=medium

  * jammy/linux-nvidia-tegra: 5.15.0-1047.47 -proposed tracker (LP: #2126518)

  * sources list generation using dwarfdump takes up to 0.5hr in build process
    (LP: #2104911)
    - [Packaging] Don't generate list of source files

  * tegra-tests devices check fails due to ubuntu/defconfig config differences
    (LP: #2125414)
    - [Config] nvidia-tegra: Enable AQUANTIA_PHY to pass tegra-tests
    - [Config] nvidia-tegra: Enable HTE to pass tegra-tests

  * apply NVIDIA patches August 22, 2025 (LP: #2121252)
    - NVIDIA: SAUCE: phy: tegra: xusb: Fix unbalanced regulator disable in
      UTMI PHY mode
    - NVIDIA: SAUCE: aquantia: unset and restore aneg intr to suspend
    - NVIDIA: SAUCE: bpmp: changes to fix sleep in atomic context.
    - NVIDIA: SAUCE: i2c: tegra: Do not mark ACPI devices and PREEMPT_RT as
      irq safe
    - NVIDIA: SAUCE: arm64: configs: set CONFIG_SQUASHFS_XATTR
    - NVIDIA: SAUCE: mm: Export mm_trace_rss_stat function

  [ Ubuntu-realtime: 5.15.0-1094.103 ]

  * jammy/linux-realtime: 5.15.0-1094.103 -proposed tracker (LP: #2126522)
  * sources list generation using dwarfdump takes up to 0.5hr in build process
    (LP: #2104911)
    - [Packaging] Don't generate list of source files
  [ Ubuntu: 5.15.0-160.170 ]
  * jammy/linux: 5.15.0-160.170 -proposed tracker (LP: #2126548)
  * [Regression Updates] System hangs when loading audit rules
    (5.15.0-156.166) (LP: #2126434)
    - netlink: avoid infinite retry looping in netlink_unicast()
  [ Ubuntu: 5.15.0-158.168 ]
  * jammy/linux: 5.15.0-158.168 -proposed tracker (LP: #2124104)
  * [UBUNTU 22.04] s390/pci: Handle PCI error codes other than 0x3a
    (LP: #2120344)
    - s390/pci: Handle PCI error codes other than 0x3a
  * sources list generation using dwarfdump takes up to 0.5hr in build process
    (LP: #2104911)
    - [Packaging] Don't generate list of source files
  * CVE-2024-26700
    - drm/amd/display: Fix MST Null Ptr for RV
  * CVE-2023-52593
    - wifi: wfx: fix possible NULL pointer dereference in wfx_set_mfp_ap()
    - wifi: wfx: fix memory leak when starting AP
    - wifi: wfx: repair open network AP mode
  * CVE-2025-38477
    - net/sched: sch_qfq: Fix race condition on qfq_aggregate
    - net/sched: sch_qfq: Avoid triggering might_sleep in atomic context in
      qfq_delete_class
  * CVE-2025-38617
    - net/packet: fix a race in packet_set_ring() and packet_notifier()
  * CVE-2025-38618
    - vsock: Do not allow binding to VMADDR_PORT_ANY
  * Packaging resync (LP: #1786013)
    - [Packaging] resync git-ubuntu-log

 -- Jacob Martin <email address hidden> Thu, 02 Oct 2025 17:51:48 -0500