kernel-rt/parallel_xz.sh

27 lines
733 B
Bash
Raw Normal View History

kernel-rt-5.12.0-0.rc1.163.rt9.1 * Fri Mar 05 2021 Juri Lelli <juri.lelli@redhat.com> [5.12.0-0.rc1.163.rt9.1] - Add localversion for -RT release (Thomas Gleixner) - sysfs: Add /sys/kernel/realtime entry (Clark Williams) - genirq: Disable irqpoll on -rt (Ingo Molnar) - signals: Allow RT tasks to cache one sigqueue struct (Thomas Gleixner) - tpm_tis: fix stall after iowrite*()s (Haris Okanovic) - drivers/block/zram: Replace bit spinlocks with rtmutex for -rt (Mike Galbraith) - POWERPC: Allow to enable RT (Sebastian Andrzej Siewior) - powerpc: Avoid recursive header includes (Sebastian Andrzej Siewior) - powerpc/stackprotector: work around stack-guard init from atomic (Sebastian Andrzej Siewior) - powerpc/kvm: Disable in-kernel MPIC emulation for PREEMPT_RT (Bogdan Purcareata) - powerpc/pseries/iommu: Use a locallock instead local_irq_save() (Sebastian Andrzej Siewior) - powerpc: traps: Use PREEMPT_RT (Sebastian Andrzej Siewior) - ARM64: Allow to enable RT (Sebastian Andrzej Siewior) - ARM: Allow to enable RT (Sebastian Andrzej Siewior) - x86: Enable RT also on 32bit (Sebastian Andrzej Siewior) - arm64: fpsimd: Delay freeing memory in fpsimd_flush_thread() (Sebastian Andrzej Siewior) - KVM: arm/arm64: downgrade preempt_disable()d region to migrate_disable() (Josh Cartwright) - genirq: update irq_set_irqchip_state documentation (Josh Cartwright) - ARM: enable irq in translation/section permission fault handlers (Yadi.hu) - tty/serial/pl011: Make the locking work on RT (Thomas Gleixner) - tty/serial/omap: Make the locking RT aware (Thomas Gleixner) - leds: trigger: disable CPU trigger on -RT (Sebastian Andrzej Siewior) - jump-label: disable if stop_machine() is used (Thomas Gleixner) - arch/arm64: Add lazy preempt support (Anders Roxell) - arm: Add support for lazy preemption (Thomas Gleixner) - x86: Support for lazy preemption (Thomas Gleixner) - x86/entry: Use should_resched() in idtentry_exit_cond_resched() (Sebastian Andrzej Siewior) - sched: Add support for lazy preemption (Thomas Gleixner) - mm/scatterlist: Do not disable irqs on RT (Thomas Gleixner) - x86: Allow to enable RT (Sebastian Andrzej Siewior) - cpuset: Convert callback_lock to raw_spinlock_t (Mike Galbraith) - drm/i915/gt: Only disable interrupts for the timeline lock on !force-threaded (Sebastian Andrzej Siewior) - drm/i915: skip DRM_I915_LOW_LEVEL_TRACEPOINTS with NOTRACE (Sebastian Andrzej Siewior) - drm/i915: disable tracing on -RT (Sebastian Andrzej Siewior) - drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates (Mike Galbraith) - drm,radeon,i915: Use preempt_disable/enable_rt() where recommended (Mike Galbraith) - lockdep: disable self-test (Sebastian Andrzej Siewior) - lockdep: selftest: fix warnings due to missing PREEMPT_RT conditionals (Josh Cartwright) - lockdep: selftest: Only do hardirq context test for raw spinlock (Yong Zhang) - lockdep: Make it RT aware (Thomas Gleixner) - net: Remove preemption disabling in netif_rx() (Priyanka Jain) - random: Make it work on rt (Thomas Gleixner) - x86: stackprotector: Avoid random pool on rt (Thomas Gleixner) - panic: skip get_random_bytes for RT_FULL in init_oops_id (Thomas Gleixner) - crypto: cryptd - add a lock instead preempt_disable/local_bh_disable (Sebastian Andrzej Siewior) - crypto: limit more FPU-enabled sections (Sebastian Andrzej Siewior) - irqwork: push most work into softirq context (Sebastian Andrzej Siewior) - net: dev: always take qdisc's busylock in __dev_xmit_skb() (Sebastian Andrzej Siewior) - net: Dequeue in dev_cpu_dead() without the lock (Sebastian Andrzej Siewior) - net: Use skbufhead with raw lock (Thomas Gleixner) - debugobjects: Make RT aware (Thomas Gleixner) - fs: namespace: Use cpu_chill() in trylock loops (Thomas Gleixner) - rt: Introduce cpu_chill() (Thomas Gleixner) - sunrpc: Make svc_xprt_do_enqueue() use get_cpu_light() (Mike Galbraith) - scsi/fcoe: Make RT aware. (Thomas Gleixner) - md: raid5: Make raid5_percpu handling RT aware (Thomas Gleixner) - block/mq: do not invoke preempt_disable() (Sebastian Andrzej Siewior) - mm/vmalloc: Another preempt disable region which sucks (Thomas Gleixner) - rcutorture: Avoid problematic critical section nesting on RT (Scott Wood) - locking: Make spinlock_t and rwlock_t a RCU section on RT (Sebastian Andrzej Siewior) - ptrace: fix ptrace_unfreeze_traced() race with rt-lock (Oleg Nesterov) - ptrace: fix ptrace vs tasklist_lock race (Sebastian Andrzej Siewior) - pid.h: include atomic.h (Grygorii Strashko) - net/core: use local_bh_disable() in netif_rx_ni() (Sebastian Andrzej Siewior) - softirq: Disable softirq stacks for RT (Thomas Gleixner) - softirq: Check preemption after reenabling interrupts (Thomas Gleixner) - sched: Disable TTWU_QUEUE on RT (Thomas Gleixner) - sched: Do not account rcu_preempt_depth on RT in might_sleep() (Thomas Gleixner) - kernel/sched: move stack + kprobe clean up to __put_task_struct() (Sebastian Andrzej Siewior) - sched: Move mmdrop to RCU on RT (Thomas Gleixner) - sched: Limit the number of task migrations per batch (Thomas Gleixner) - wait.h: include atomic.h (Sebastian Andrzej Siewior) - x86: kvm Require const tsc for RT (Thomas Gleixner) - mm/zsmalloc: copy with get_cpu_var() and locking (Mike Galbraith) - mm/memcontrol: Replace local_irq_disable with local locks (Sebastian Andrzej Siewior) - mm/memcontrol: Don't call schedule_work_on in preemption disabled context (Yang Shi) - mm: memcontrol: Provide a local_lock for per-CPU memcg_stock (Sebastian Andrzej Siewior) - mm: slub: Don't enable partial CPU caches on PREEMPT_RT by default (Sebastian Andrzej Siewior) - mm: page_alloc: Use a local_lock instead of explicit local_irq_save(). (Ingo Molnar) - mm: page_alloc: Use migrate_disable() in drain_local_pages_wq() (Sebastian Andrzej Siewior) - mm: slub: Don't resize the location tracking cache on PREEMPT_RT (Sebastian Andrzej Siewior) - mm: slub: Move flush_cpu_slab() invocations __free_slab() invocations out of IRQ context (Sebastian Andrzej Siewior) - mm: slub: Move discard_slab() invocations out of IRQ-off sections (Sebastian Andrzej Siewior) - mm: slub: Enable irqs for __GFP_WAIT (Thomas Gleixner) - mm: slub: Make object_map_lock a raw_spinlock_t (Sebastian Andrzej Siewior) - mm: sl[au]b: Change list_lock to raw_spinlock_t (Thomas Gleixner) - locking: don't check for __LINUX_SPINLOCK_TYPES_H on -RT archs (Sebastian Andrzej Siewior) - trace: Add migrate-disabled counter to tracing output (Thomas Gleixner) - kernel/sched: add {put|get}_cpu_light() (Sebastian Andrzej Siewior) - signal/x86: Delay calling signals in atomic (Oleg Nesterov) - rt: Add local irq locks (Thomas Gleixner) - efi: Allow efi=runtime (Sebastian Andrzej Siewior) - efi: Disable runtime services on RT (Sebastian Andrzej Siewior) - net/core: disable NET_RX_BUSY_POLL on RT (Sebastian Andrzej Siewior) - sched: Disable CONFIG_RT_GROUP_SCHED on RT (Thomas Gleixner) - mm: Allow only SLUB on RT (Ingo Molnar) - kconfig: Disable config options which are not RT compatible (Thomas Gleixner) - net: Properly annotate the try-lock for the seqlock (Sebastian Andrzej Siewior) - net/Qdisc: use a seqlock instead seqcount (Sebastian Andrzej Siewior) - fs/dcache: disable preemption on i_dir_seq's write side (Sebastian Andrzej Siewior) - fs/dcache: use swait_queue instead of waitqueue (Sebastian Andrzej Siewior) - u64_stats: Disable preemption on 32bit-UP/SMP with RT during updates (Sebastian Andrzej Siewior) - xfrm: Use sequence counter with associated spinlock (Ahmed S. Darwish) - mm/memcontrol: Disable preemption in __mod_memcg_lruvec_state() (Sebastian Andrzej Siewior) - mm/vmstat: Protect per cpu variables with preempt disable on RT (Ingo Molnar) - preempt: Provide preempt_*_(no)rt variants (Thomas Gleixner) - signal: Revert ptrace preempt magic (Thomas Gleixner) - locking/rtmutex: Use custom scheduling function for spin-schedule() (Sebastian Andrzej Siewior) - locking/rtmutex: add ww_mutex addon for mutex-rt (Sebastian Andrzej Siewior) - locking/rtmutex: wire up RT's locking (Thomas Gleixner) - locking/rtmutex: add rwlock implementation based on rtmutex (Thomas Gleixner) - locking/rtmutex: add rwsem implementation based on rtmutex (Thomas Gleixner) - locking/rtmutex: add mutex implementation based on rtmutex (Thomas Gleixner) - locking/rtmutex: Allow rt_mutex_trylock() on PREEMPT_RT (Sebastian Andrzej Siewior) - locking/rtmutex: add sleeping lock implementation (Thomas Gleixner) - sched: Add saved_state for tasks blocked on sleeping locks (Thomas Gleixner) - locking/rtmutex: export lockdep-less version of rt_mutex's lock, trylock and unlock (Thomas Gleixner) - locking/rtmutex: Provide rt_mutex_slowlock_locked() (Thomas Gleixner) - locking: split out the rbtree definition (Sebastian Andrzej Siewior) - lockdep: Reduce header files in debug_locks.h (Sebastian Andrzej Siewior) - locking/rtmutex: Avoid include hell (Thomas Gleixner) - locking/spinlock: Split the lock types header (Thomas Gleixner) - locking/rtmutex: Make lock_killable work (Thomas Gleixner) - futex: Fix bug on when a requeued RT task times out (Steven Rostedt) - locking/rtmutex: Handle the various new futex race conditions (Thomas Gleixner) - locking/rtmutex: Remove rt_mutex_timed_lock() (Sebastian Andrzej Siewior) - locking/rtmutex: Move rt_mutex_init() outside of CONFIG_DEBUG_RT_MUTEXES (Sebastian Andrzej Siewior) - locking/rtmutex: Remove output from deadlock detector. (Sebastian Andrzej Siewior) - locking/rtmutex: Remove cruft (Sebastian Andrzej Siewior) - firewire: ohci: Use tasklet_disable_in_atomic() in the two atomic user (Sebastian Andrzej Siewior) - atm: eni: Use tasklet_disable_in_atomic() in the ->send() callback (Sebastian Andrzej Siewior) - PCI: hv: Use tasklet_disable_in_atomic() (Sebastian Andrzej Siewior) - ath9k: Use tasklet_disable_in_atomic(). (Sebastian Andrzej Siewior) - net: sundance: Use tasklet_disable_in_atomic(). (Sebastian Andrzej Siewior) - net: jme: Replace link-change tasklet with a worker (Sebastian Andrzej Siewior) - rcu: Prevent false positive softirq warning on RT (Thomas Gleixner) - tick/sched: Prevent false positive softirq pending warnings on RT (Thomas Gleixner) - softirq: Make softirq control and processing RT aware (Thomas Gleixner) - softirq: Move various protections into inline helpers (Thomas Gleixner) - irqtime: Make accounting correct on RT (Thomas Gleixner) - softirq: Add RT specific softirq accounting (Thomas Gleixner) - tasklets: Prevent tasklet_unlock_spin_wait() deadlock on RT (Thomas Gleixner) - tasklets: Replace spin wait in tasklet_kill() (Peter Zijlstra) - tasklets: Replace spin wait in tasklet_unlock_wait() (Peter Zijlstra) - tasklets: Provide tasklet_disable_in_atomic() (Thomas Gleixner) - tasklets: Use static inlines for stub implementations (Thomas Gleixner) - softirq: Replace barrier() with cpu_relax() in tasklet_unlock_wait() (Thomas Gleixner) - smp: Wake ksoftirqd on PREEMPT_RT instead do_softirq(). (Sebastian Andrzej Siewior) - tcp: Remove superfluous BH-disable around listening_hash (Sebastian Andrzej Siewior) - net: Move lockdep where it belongs (Thomas Gleixner) - shmem: Use raw_spinlock_t for ->stat_lock (Sebastian Andrzej Siewior) - mm: workingset: replace IRQ-off check with a lockdep assert. (Sebastian Andrzej Siewior) - cgroup: use irqsave in cgroup_rstat_flush_locked() (Sebastian Andrzej Siewior) - kcov: Remove kcov include from sched.h and move it to its users. (Sebastian Andrzej Siewior) - printk: add pr_flush() (John Ogness) - printk: add console handover (John Ogness) - printk: remove deferred printing (John Ogness) - printk: move console printing to kthreads (John Ogness) - printk: introduce kernel sync mode (John Ogness) - printk: change @console_seq to atomic64_t (John Ogness) - printk: combine boot_delay_msec() into printk_delay() (John Ogness) - printk: relocate printk_delay() and vprintk_default() (John Ogness) - serial: 8250: implement write_atomic (John Ogness) - console: add write_atomic interface (John Ogness) - printk: convert @syslog_lock to spin_lock (John Ogness) - printk: remove safe buffers (John Ogness) - printk: track/limit recursion (John Ogness) - printk: console: remove unnecessary safe buffer usage (John Ogness) - printk: kmsg_dump: use kmsg_dump_rewind (John Ogness) - printk: kmsg_dump: remove _nolock() variants (John Ogness) - printk: remove logbuf_lock (John Ogness) - um: synchronize kmsg_dumper (John Ogness) - printk: introduce a kmsg_dump iterator (John Ogness) - printk: add syslog_lock (John Ogness) - printk: use atomic64_t for devkmsg_user.seq (John Ogness) - printk: use seqcount_latch for clear_seq (John Ogness) - printk: introduce CONSOLE_LOG_MAX for improved multi-line support (John Ogness) - printk: consolidate kmsg_dump_get_buffer/syslog_print_all code (John Ogness) - printk: refactor kmsg_dump_get_buffer() (John Ogness) - printk: kmsg_dump: remove unused fields (John Ogness) - printk: limit second loop of syslog_print_all (John Ogness) - powerpc/mm: Move the linear_mapping_mutex to the ifdef where it is used (Sebastian Andrzej Siewior) - notifier: Make atomic_notifiers use raw_spinlock (Valentin Schneider) - genirq: Move prio assignment into the newly created thread (Thomas Gleixner) - kthread: Move prio/affinite change into the newly created thread (Sebastian Andrzej Siewior) - timers: Move clearing of base::timer_running under base::lock (Thomas Gleixner) - highmem: Don't disable preemption on RT in kmap_atomic() (Sebastian Andrzej Siewior) Resolves: rhbz#1891873
2021-03-05 16:23:42 +00:00
#!/bin/sh
# Reads filenames on stdin, xz-compresses each in place.
# Not optimal for "compress relatively few, large files" scenario!
# How many xz's to run in parallel:
procgroup=""
while test "$#" != 0; do
# Get it from -jNUM
N="${1#-j}"
if test "$N" = "$1"; then
# Not -j<something> - warn and ignore
echo "parallel_xz: warning: unrecognized argument: '$1'"
else
procgroup="$N"
fi
shift
done
# If told to use only one cpu:
test "$procgroup" || exec xargs -r xz
test "$procgroup" = 1 && exec xargs -r xz
# xz has some startup cost. If files are really small,
# this cost might be significant. To combat this,
# process several files (in sequence) by each xz process via -n 16:
exec xargs -r -n 16 -P "$procgroup" xz