PDS Kernel Configuration

Merged upstream.

PKGBUILD

@@ -31,7 +31,6 @@ _distro="Arch"
 declare -p -x > current_env
 
 source "$_where"/customization.cfg # load default configuration from file
-source "$_where"/linux-tkg-config/prepare
 
 if [ -e "$_EXT_CONFIG_PATH" ]; then
   msg2 "External configuration file $_EXT_CONFIG_PATH will be used and will override customization.cfg values."
@@ -40,6 +39,8 @@ fi
 
 source current_env
 
+source "$_where"/linux-tkg-config/prepare
+
 # Make sure we're in a clean state
 if [ ! -e "$_where"/BIG_UGLY_FROGMINER ]; then
   _tkg_initscript

customization.cfg

@@ -3,7 +3,7 @@
 # Linux distribution you are using, options are "Arch", "Void", "Ubuntu", "Debian", "Fedora", "Suse", "Gentoo", "Generic".
 # It is automatically set to "Arch" when using PKGBUILD.
 # If left empty, the script will prompt
-_distro=""
+_distro="Arch"
 
 # Kernel Version - Options are "5.4", and from "5.7" to "5.19"
 # you can also set a specific kernel version, e.g. "6.0-rc4" or "5.10.51",
@@ -32,7 +32,7 @@ CUSTOM_GCC_PATH=""
 CUSTOM_LLVM_PATH=""
 
 # Set to true to bypass makepkg.conf and use all available threads for compilation. False will respect your makepkg.conf options.
-_force_all_threads="true"
+_force_all_threads="false"
 
 # Set to true to prevent ccache from being used and set CONFIG_GCC_PLUGINS=y (which needs to be disabled for ccache to work properly)
 _noccache="false"
@@ -46,10 +46,10 @@ _modprobeddb="false"
 _modprobeddb_db_path=~/.config/modprobed.db
 
 # Set to "1" to call make menuconfig, "2" to call make nconfig, "3" to call make xconfig, before building the kernel. Set to false to disable and skip the prompt.
-_menunconfig=""
+_menunconfig="false"
 
 # Set to true to generate a kernel config fragment from your changes in menuconfig/nconfig. Set to false to disable and skip the prompt.
-_diffconfig=""
+_diffconfig="false"
 
 # Set to the file name where the generated config fragment should be written to. Only used if _diffconfig is active.
 _diffconfig_name=""
@@ -90,11 +90,11 @@ _STRIP="true"
 # LEAVE AN EMPTY VALUE TO BE PROMPTED ABOUT FOLLOWING OPTIONS AT BUILD TIME
 
 # CPU scheduler - Options are "upds" (TkG's Undead PDS), "pds", "bmq", "muqss", "cacule" or "cfs" (kernel's default)
-_cpusched=""
+_cpusched="pds"
 
 # Compiler to use - Options are "gcc" or "llvm".
 # For advanced users.
-_compiler=""
+_compiler="gcc"
 
 # Force the use of the LLVM Integrated Assembler whether using LLVM, LTO or not.
 # Set to "1" to enable.
@@ -124,7 +124,7 @@ _preempt_rt_force=""
 # For BMQ:           0: No yield.
 #                    1: Deboost and requeue task. (Default)
 #                    2: Set rq skip task.
-_sched_yield_type=""
+_sched_yield_type="0"
 
 # Round Robin interval is the longest duration two tasks with the same nice level will be delayed for. When CPU time is requested by a task, it receives a time slice equal
 # to the rr_interval in addition to a virtual deadline. When using yield_type 2, a low value can help offset the disadvantages of rescheduling a process that has yielded.
@@ -132,7 +132,7 @@ _sched_yield_type=""
 # PDS default: 4ms"
 # BMQ default: 2ms"
 # Set to "1" for 2ms, "2" for 4ms, "3" for 6ms, "4" for 8ms, or "default" to keep the chosen scheduler defaults.
-_rr_interval=""
+_rr_interval="2"
 
 # Set to "true" to disable FUNCTION_TRACER/GRAPH_TRACER, lowering overhead but limiting debugging and analyzing of kernel functions - Kernel default is "false"
 _ftracedisable="false"
@@ -147,10 +147,10 @@ _misc_adds="true"
 # Full tickless can give higher performances in case you use isolation of CPUs for tasks
 # and it works only when using the nohz_full kernel parameter, otherwise behaves like idle.
 # Just tickless idle perform better for most platforms.
-_tickless=""
+_tickless="2"
 
 # Set to "true" to use ACS override patch - https://wiki.archlinux.org/index.php/PCI_passthrough_via_OVMF#Bypassing_the_IOMMU_groups_.28ACS_override_patch.29 - Kernel default is "false"
-_acs_override=""
+_acs_override="false"
 
 # Set to "true" to add Bcache filesystem support. You'll have to install bcachefs-tools-git from AUR for utilities - https://bcachefs.org/ - If in doubt, set to "false"
 # This can be buggy and isn't recommended on a production machine, also enabling this option will not allow you to enable MGLRU.
@@ -179,13 +179,13 @@ _futex_waitv="false"
 _winesync="false"
 
 # Set to "true" to enable Binder and Ashmem, the kernel modules required to use the android emulator Anbox. ! This doesn't apply to 5.4.y !
-_anbox=""
+_anbox="false"
 
 # A selection of patches from Zen/Liquorix kernel and additional tweaks for a better gaming experience (ZENIFY) - Default is "true"
 _zenify="true"
 
 # compiler optimization level - 1. Optimize for performance (-O2); 2. Optimize harder (-O3); 3. Optimize for size (-Os) - Kernel default is "1"
-_compileroptlevel="1"
+_compileroptlevel="2"
 
 # CPU compiler optimizations - Defaults to prompt at kernel config if left empty
 # AMD CPUs : "k8" "k8sse3" "k10" "barcelona" "bobcat" "jaguar" "bulldozer" "piledriver" "steamroller" "excavator" "zen" "zen2" "zen3" (zen3 opt support depends on GCC11)
@@ -199,7 +199,7 @@ _compileroptlevel="1"
 # - "generic_v2" (depends on GCC11 - to share the package between machines with different CPU µarch supporting at least x86-64-v2
 # - "generic_v3" (depends on GCC11 - to share the package between machines with different CPU µarch supporting at least x86-64-v3
 # - "generic_v4" (depends on GCC11 - to share the package between machines with different CPU µarch supporting at least x86-64-v4
-_processor_opt=""
+_processor_opt="skylake"
 
 # MuQSS only - Make IRQ threading compulsory (FORCE_IRQ_THREADING) - Default is "false"
 _irq_threading="false"
@@ -215,17 +215,17 @@ _cacule_rdb_interval="19"
 _tt_high_hz="false"
 
 # MuQSS and PDS only - SMT (Hyperthreading) aware nice priority and policy support (SMT_NICE) - Kernel default is "true" - You can disable this on non-SMT/HT CPUs for lower overhead
-_smt_nice=""
+_smt_nice="true"
 
 # Trust the CPU manufacturer to initialize Linux's CRNG (RANDOM_TRUST_CPU) - Kernel default is "false"
-_random_trust_cpu="false"
+_random_trust_cpu="true"
 
 # MuQSS only - CPU scheduler runqueue sharing - No sharing (RQ_NONE), SMT (hyperthread) siblings (RQ_SMT), Multicore siblings (RQ_MC), Symmetric Multi-Processing (RQ_SMP), NUMA (RQ_ALL)
 # Valid values are "none", "smt", "mc", "mc-llc"(for zen), "smp", "all" - Kernel default is "smt"
 _runqueue_sharing=""
 
 # Timer frequency - "100" "250" "300" "500" "750" "1000" ("2000" is available for cacule cpusched only) - More options available in kernel config prompt when left empty depending on selected cpusched with the default option pointed with a ">" (2000 for cacule, 100 for muqss and 1000 for other cpu schedulers)
-_timer_freq=""
+_timer_freq="500"
 
 # Default CPU governor - "performance", "ondemand", "schedutil" or leave empty for default (schedutil)
 _default_cpu_gov="ondemand"

install.sh

@@ -44,6 +44,12 @@ plain() {
 
 declare -p -x > current_env
 source customization.cfg
+
+if [ -e "$_EXT_CONFIG_PATH" ]; then
+  msg2 "External configuration file $_EXT_CONFIG_PATH will be used and will override customization.cfg values."
+  source "$_EXT_CONFIG_PATH"
+fi
+
 . current_env
 
 source linux-tkg-config/prepare

linux-tkg-patches/6.0/0009-prjc_v6.0-r0.patch

@@ -1,6 +1,6 @@
-From b53bf730e6bba71ebc0ec8452cc2ca399137090e Mon Sep 17 00:00:00 2001
+From 711a56e8f6314d77141b0f661e6c13c8a2c4dddf Mon Sep 17 00:00:00 2001
 From: Tor Vic <torvic9@mailbox.org>
-Date: Mon, 3 Oct 2022 11:19:50 +0200
+Date: Wed, 16 Nov 2022 11:29:00 +0100
 Subject: [PATCH] Project-C 6.0-rc0-vd
 
 ---
@@ -22,7 +22,7 @@ Subject: [PATCH] Project-C 6.0-rc0-vd
  kernel/exit.c                                 |    4 +-
  kernel/locking/rtmutex.c                      |   16 +-
  kernel/sched/Makefile                         |    5 +
- kernel/sched/alt_core.c                       | 7937 +++++++++++++++++
+ kernel/sched/alt_core.c                       | 7959 +++++++++++++++++
  kernel/sched/alt_debug.c                      |   31 +
  kernel/sched/alt_sched.h                      |  645 ++
  kernel/sched/bmq.h                            |  110 +
@@ -43,7 +43,7 @@ Subject: [PATCH] Project-C 6.0-rc0-vd
  kernel/time/hrtimer.c                         |    2 +
  kernel/time/posix-cpu-timers.c                |   10 +-
  kernel/trace/trace_selftest.c                 |    5 +
- 39 files changed, 9245 insertions(+), 23 deletions(-)
+ 39 files changed, 9267 insertions(+), 23 deletions(-)
  create mode 100644 Documentation/scheduler/sched-BMQ.txt
  create mode 100644 kernel/sched/alt_core.c
  create mode 100644 kernel/sched/alt_debug.c
@@ -685,10 +685,10 @@ index 976092b7bd45..31d587c16ec1 100644
  obj-y += build_utility.o
 diff --git a/kernel/sched/alt_core.c b/kernel/sched/alt_core.c
 new file mode 100644
-index 000000000000..f3bac14124c3
+index 000000000000..18dfee354f9b
 --- /dev/null
 +++ b/kernel/sched/alt_core.c
-@@ -0,0 +1,7937 @@
+@@ -0,0 +1,7959 @@
 +/*
 + *  kernel/sched/alt_core.c
 + *
@@ -3111,13 +3111,6 @@ index 000000000000..f3bac14124c3
 +	if (!llist)
 +		return;
 +
-+	/*
-+	 * rq::ttwu_pending racy indication of out-standing wakeups.
-+	 * Races such that false-negatives are possible, since they
-+	 * are shorter lived that false-positives would be.
-+	 */
-+	WRITE_ONCE(rq->ttwu_pending, 0);
-+
 +	rq_lock_irqsave(rq, &rf);
 +	update_rq_clock(rq);
 +
@@ -3131,6 +3124,17 @@ index 000000000000..f3bac14124c3
 +		ttwu_do_activate(rq, p, p->sched_remote_wakeup ? WF_MIGRATED : 0);
 +	}
 +
++	/*
++	 * Must be after enqueueing at least once task such that
++	 * idle_cpu() does not observe a false-negative -- if it does,
++	 * it is possible for select_idle_siblings() to stack a number
++	 * of tasks on this CPU during that window.
++	 *
++	 * It is ok to clear ttwu_pending when another task pending.
++	 * We will receive IPI after local irq enabled and then enqueue it.
++	 * Since now nr_running > 0, idle_cpu() will always get correct result.
++	 */
++	WRITE_ONCE(rq->ttwu_pending, 0);
 +	rq_unlock_irqrestore(rq, &rf);
 +}
 +
@@ -3601,6 +3605,40 @@ index 000000000000..f3bac14124c3
 +	return success;
 +}
 +
++static bool __task_needs_rq_lock(struct task_struct *p)
++{
++	unsigned int state = READ_ONCE(p->__state);
++
++	/*
++	 * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
++	 * the task is blocked. Make sure to check @state since ttwu() can drop
++	 * locks at the end, see ttwu_queue_wakelist().
++	 */
++	if (state == TASK_RUNNING || state == TASK_WAKING)
++		return true;
++
++	/*
++	 * Ensure we load p->on_rq after p->__state, otherwise it would be
++	 * possible to, falsely, observe p->on_rq == 0.
++	 *
++	 * See try_to_wake_up() for a longer comment.
++	 */
++	smp_rmb();
++	if (p->on_rq)
++		return true;
++
++#ifdef CONFIG_SMP
++	/*
++	 * Ensure the task has finished __schedule() and will not be referenced
++	 * anymore. Again, see try_to_wake_up() for a longer comment.
++	 */
++	smp_rmb();
++	smp_cond_load_acquire(&p->on_cpu, !VAL);
++#endif
++
++	return false;
++}
++
 +/**
 + * task_call_func - Invoke a function on task in fixed state
 + * @p: Process for which the function is to be invoked, can be @current.
@@ -3618,28 +3656,12 @@ index 000000000000..f3bac14124c3
 +int task_call_func(struct task_struct *p, task_call_f func, void *arg)
 +{
 +	struct rq *rq = NULL;
-+	unsigned int state;
 +	struct rq_flags rf;
 +	int ret;
 +
 +	raw_spin_lock_irqsave(&p->pi_lock, rf.flags);
 +
-+	state = READ_ONCE(p->__state);
++	if (__task_needs_rq_lock(p))
-+
-+	/*
-+	 * Ensure we load p->on_rq after p->__state, otherwise it would be
-+	 * possible to, falsely, observe p->on_rq == 0.
-+	 *
-+	 * See try_to_wake_up() for a longer comment.
-+	 */
-+	smp_rmb();
-+
-+	/*
-+	 * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
-+	 * the task is blocked. Make sure to check @state since ttwu() can drop
-+	 * locks at the end, see ttwu_queue_wakelist().
-+	 */
-+	if (state == TASK_RUNNING || state == TASK_WAKING || p->on_rq)
 +		rq = __task_rq_lock(p, &rf);
 +
 +	/*
@@ -10130,5 +10152,55 @@ index a2d301f58ced..2ccdede8585c 100644
  	struct wakeup_test_data *x = data;
 
 --
-2.37.3
+2.38.1
 
+diff --git a/kernel/sched/alt_core.c b/kernel/sched/alt_core.c
+index f3bac14124c3..27eafbccf23d 100644
+--- a/kernel/sched/alt_core.c
++++ b/kernel/sched/alt_core.c
+@@ -1448,11 +1448,13 @@ void set_task_cpu(struct task_struct *p, unsigned int new_cpu)
+ 
+ 	WARN_ON_ONCE(is_migration_disabled(p));
+ #endif
+-	if (task_cpu(p) == new_cpu)
+-		return;
++
+ 	trace_sched_migrate_task(p, new_cpu);
+-	rseq_migrate(p);
+-	perf_event_task_migrate(p);
++
++	if (task_cpu(p) != new_cpu) {
++		rseq_migrate(p);
++		perf_event_task_migrate(p);
++	}
+ 
+ 	__set_task_cpu(p, new_cpu);
+ }
+diff --git a/kernel/sched/alt_core.c b/kernel/sched/alt_core.c
+index f3bac14124c3..5678c247c0ab 100644
+--- a/kernel/sched/alt_core.c
++++ b/kernel/sched/alt_core.c
+@@ -810,8 +810,8 @@ unsigned long get_wchan(struct task_struct *p)
+  * Context: rq->lock
+  */
+ #define __SCHED_DEQUEUE_TASK(p, rq, flags)					\
+-	psi_dequeue(p, flags & DEQUEUE_SLEEP);					\
+ 	sched_info_dequeue(rq, p);						\
++	psi_dequeue(p, flags & DEQUEUE_SLEEP);					\
+ 										\
+ 	list_del(&p->sq_node);							\
+ 	if (list_empty(&rq->queue.heads[p->sq_idx])) 				\
+diff --git a/kernel/sched/alt_core.c b/kernel/sched/alt_core.c
+index f3bac14124c3..349a2c92d534 100644
+--- a/kernel/sched/alt_core.c
++++ b/kernel/sched/alt_core.c
+@@ -4404,8 +4404,8 @@ static inline void schedule_debug(struct task_struct *prev, bool preempt)
+ 
+ /*
+  * Compile time debug macro
+- * #define ALT_SCHED_DEBUG
+  */
++#define ALT_SCHED_DEBUG
+ 
+ #ifdef ALT_SCHED_DEBUG
+ void alt_sched_debug(void)

linux-tkg-patches/6.0/0012-misc-additions.patch

@@ -64,140 +64,6 @@ index 2c7171e0b0010..85de313ddec29 100644
  	select CPU_FREQ_GOV_PERFORMANCE
  	help
 
-From 2535fbde890f14c78b750139fcf87d1143850626 Mon Sep 17 00:00:00 2001
-From: Johannes Weiner <hannes@cmpxchg.org>
-Date: Tue, 2 Aug 2022 12:28:11 -0400
-Subject: [PATCH] mm: vmscan: fix extreme overreclaim and swap floods
-
-During proactive reclaim, we sometimes observe severe overreclaim, with
-several thousand times more pages reclaimed than requested.
-
-This trace was obtained from shrink_lruvec() during such an instance:
-
-    prio:0 anon_cost:1141521 file_cost:7767
-    nr_reclaimed:4387406 nr_to_reclaim:1047 (or_factor:4190)
-    nr=[7161123 345 578 1111]
-
-While he reclaimer requested 4M, vmscan reclaimed close to 16G, most of it
-by swapping.  These requests take over a minute, during which the write()
-to memory.reclaim is unkillably stuck inside the kernel.
-
-Digging into the source, this is caused by the proportional reclaim
-bailout logic.  This code tries to resolve a fundamental conflict: to
-reclaim roughly what was requested, while also aging all LRUs fairly and
-in accordance to their size, swappiness, refault rates etc.  The way it
-attempts fairness is that once the reclaim goal has been reached, it stops
-scanning the LRUs with the smaller remaining scan targets, and adjusts the
-remainder of the bigger LRUs according to how much of the smaller LRUs was
-scanned.  It then finishes scanning that remainder regardless of the
-reclaim goal.
-
-This works fine if priority levels are low and the LRU lists are
-comparable in size.  However, in this instance, the cgroup that is
-targeted by proactive reclaim has almost no files left - they've already
-been squeezed out by proactive reclaim earlier - and the remaining anon
-pages are hot.  Anon rotations cause the priority level to drop to 0,
-which results in reclaim targeting all of anon (a lot) and all of file
-(almost nothing).  By the time reclaim decides to bail, it has scanned
-most or all of the file target, and therefor must also scan most or all of
-the enormous anon target.  This target is thousands of times larger than
-the reclaim goal, thus causing the overreclaim.
-
-The bailout code hasn't changed in years, why is this failing now?  The
-most likely explanations are two other recent changes in anon reclaim:
-
-1. Before the series starting with commit 5df741963d52 ("mm: fix LRU
-   balancing effect of new transparent huge pages"), the VM was
-   overall relatively reluctant to swap at all, even if swap was
-   configured. This means the LRU balancing code didn't come into play
-   as often as it does now, and mostly in high pressure situations
-   where pronounced swap activity wouldn't be as surprising.
-
-2. For historic reasons, shrink_lruvec() loops on the scan targets of
-   all LRU lists except the active anon one, meaning it would bail if
-   the only remaining pages to scan were active anon - even if there
-   were a lot of them.
-
-   Before the series starting with commit ccc5dc67340c ("mm/vmscan:
-   make active/inactive ratio as 1:1 for anon lru"), most anon pages
-   would live on the active LRU; the inactive one would contain only a
-   handful of preselected reclaim candidates. After the series, anon
-   gets aged similarly to file, and the inactive list is the default
-   for new anon pages as well, making it often the much bigger list.
-
-   As a result, the VM is now more likely to actually finish large
-   anon targets than before.
-
-Change the code such that only one SWAP_CLUSTER_MAX-sized nudge toward the
-larger LRU lists is made before bailing out on a met reclaim goal.
-
-This fixes the extreme overreclaim problem.
-
-Fairness is more subtle and harder to evaluate.  No obvious misbehavior
-was observed on the test workload, in any case.  Conceptually, fairness
-should primarily be a cumulative effect from regular, lower priority
-scans.  Once the VM is in trouble and needs to escalate scan targets to
-make forward progress, fairness needs to take a backseat.  This is also
-acknowledged by the myriad exceptions in get_scan_count().  This patch
-makes fairness decrease gradually, as it keeps fairness work static over
-increasing priority levels with growing scan targets.  This should make
-more sense - although we may have to re-visit the exact values.
-
-Link: https://lkml.kernel.org/r/20220802162811.39216-1-hannes@cmpxchg.org
-Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
-Reviewed-by: Rik van Riel <riel@surriel.com>
-Acked-by: Mel Gorman <mgorman@techsingularity.net>
-Cc: Hugh Dickins <hughd@google.com>
-Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
-Cc: <stable@vger.kernel.org>
-Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
----
- mm/vmscan.c | 10 ++++------
- 1 file changed, 4 insertions(+), 6 deletions(-)
-
-diff --git a/mm/vmscan.c b/mm/vmscan.c
-index 382dbe97329f33..266eb8cfe93a67 100644
---- a/mm/vmscan.c
-+++ b/mm/vmscan.c
-@@ -2955,8 +2955,8 @@ static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
- 	enum lru_list lru;
- 	unsigned long nr_reclaimed = 0;
- 	unsigned long nr_to_reclaim = sc->nr_to_reclaim;
-+	bool proportional_reclaim;
- 	struct blk_plug plug;
--	bool scan_adjusted;
-
- 	get_scan_count(lruvec, sc, nr);
-
-@@ -2974,8 +2974,8 @@ static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
- 	 * abort proportional reclaim if either the file or anon lru has already
- 	 * dropped to zero at the first pass.
- 	 */
--	scan_adjusted = (!cgroup_reclaim(sc) && !current_is_kswapd() &&
--			 sc->priority == DEF_PRIORITY);
-+	proportional_reclaim = (!cgroup_reclaim(sc) && !current_is_kswapd() &&
-+				sc->priority == DEF_PRIORITY);
-
- 	blk_start_plug(&plug);
- 	while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] ||
-@@ -2995,7 +2995,7 @@ static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
-
- 		cond_resched();
-
--		if (nr_reclaimed < nr_to_reclaim || scan_adjusted)
-+		if (nr_reclaimed < nr_to_reclaim || proportional_reclaim)
- 			continue;
-
- 		/*
-@@ -3046,8 +3046,6 @@ static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
- 		nr_scanned = targets[lru] - nr[lru];
- 		nr[lru] = targets[lru] * (100 - percentage) / 100;
- 		nr[lru] -= min(nr[lru], nr_scanned);
--
--		scan_adjusted = true;
- 	}
- 	blk_finish_plug(&plug);
- 	sc->nr_reclaimed += nr_reclaimed;
 From 430daaab3c78de6bd82f10cfb5a0f016c6e583f6 Mon Sep 17 00:00:00 2001
 From: Desmond Cheong Zhi Xi <desmondcheongzx@gmail.com>
 Date: Mon, 4 Oct 2021 14:07:34 -0400

linux-tkg-patches/6.1/0012-misc-additions.patch

@@ -64,140 +64,6 @@ index 2c7171e0b0010..85de313ddec29 100644
  	select CPU_FREQ_GOV_PERFORMANCE
  	help
 
-From 2535fbde890f14c78b750139fcf87d1143850626 Mon Sep 17 00:00:00 2001
-From: Johannes Weiner <hannes@cmpxchg.org>
-Date: Tue, 2 Aug 2022 12:28:11 -0400
-Subject: [PATCH] mm: vmscan: fix extreme overreclaim and swap floods
-
-During proactive reclaim, we sometimes observe severe overreclaim, with
-several thousand times more pages reclaimed than requested.
-
-This trace was obtained from shrink_lruvec() during such an instance:
-
-    prio:0 anon_cost:1141521 file_cost:7767
-    nr_reclaimed:4387406 nr_to_reclaim:1047 (or_factor:4190)
-    nr=[7161123 345 578 1111]
-
-While he reclaimer requested 4M, vmscan reclaimed close to 16G, most of it
-by swapping.  These requests take over a minute, during which the write()
-to memory.reclaim is unkillably stuck inside the kernel.
-
-Digging into the source, this is caused by the proportional reclaim
-bailout logic.  This code tries to resolve a fundamental conflict: to
-reclaim roughly what was requested, while also aging all LRUs fairly and
-in accordance to their size, swappiness, refault rates etc.  The way it
-attempts fairness is that once the reclaim goal has been reached, it stops
-scanning the LRUs with the smaller remaining scan targets, and adjusts the
-remainder of the bigger LRUs according to how much of the smaller LRUs was
-scanned.  It then finishes scanning that remainder regardless of the
-reclaim goal.
-
-This works fine if priority levels are low and the LRU lists are
-comparable in size.  However, in this instance, the cgroup that is
-targeted by proactive reclaim has almost no files left - they've already
-been squeezed out by proactive reclaim earlier - and the remaining anon
-pages are hot.  Anon rotations cause the priority level to drop to 0,
-which results in reclaim targeting all of anon (a lot) and all of file
-(almost nothing).  By the time reclaim decides to bail, it has scanned
-most or all of the file target, and therefor must also scan most or all of
-the enormous anon target.  This target is thousands of times larger than
-the reclaim goal, thus causing the overreclaim.
-
-The bailout code hasn't changed in years, why is this failing now?  The
-most likely explanations are two other recent changes in anon reclaim:
-
-1. Before the series starting with commit 5df741963d52 ("mm: fix LRU
-   balancing effect of new transparent huge pages"), the VM was
-   overall relatively reluctant to swap at all, even if swap was
-   configured. This means the LRU balancing code didn't come into play
-   as often as it does now, and mostly in high pressure situations
-   where pronounced swap activity wouldn't be as surprising.
-
-2. For historic reasons, shrink_lruvec() loops on the scan targets of
-   all LRU lists except the active anon one, meaning it would bail if
-   the only remaining pages to scan were active anon - even if there
-   were a lot of them.
-
-   Before the series starting with commit ccc5dc67340c ("mm/vmscan:
-   make active/inactive ratio as 1:1 for anon lru"), most anon pages
-   would live on the active LRU; the inactive one would contain only a
-   handful of preselected reclaim candidates. After the series, anon
-   gets aged similarly to file, and the inactive list is the default
-   for new anon pages as well, making it often the much bigger list.
-
-   As a result, the VM is now more likely to actually finish large
-   anon targets than before.
-
-Change the code such that only one SWAP_CLUSTER_MAX-sized nudge toward the
-larger LRU lists is made before bailing out on a met reclaim goal.
-
-This fixes the extreme overreclaim problem.
-
-Fairness is more subtle and harder to evaluate.  No obvious misbehavior
-was observed on the test workload, in any case.  Conceptually, fairness
-should primarily be a cumulative effect from regular, lower priority
-scans.  Once the VM is in trouble and needs to escalate scan targets to
-make forward progress, fairness needs to take a backseat.  This is also
-acknowledged by the myriad exceptions in get_scan_count().  This patch
-makes fairness decrease gradually, as it keeps fairness work static over
-increasing priority levels with growing scan targets.  This should make
-more sense - although we may have to re-visit the exact values.
-
-Link: https://lkml.kernel.org/r/20220802162811.39216-1-hannes@cmpxchg.org
-Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
-Reviewed-by: Rik van Riel <riel@surriel.com>
-Acked-by: Mel Gorman <mgorman@techsingularity.net>
-Cc: Hugh Dickins <hughd@google.com>
-Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
-Cc: <stable@vger.kernel.org>
-Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
----
- mm/vmscan.c | 10 ++++------
- 1 file changed, 4 insertions(+), 6 deletions(-)
-
-diff --git a/mm/vmscan.c b/mm/vmscan.c
-index 382dbe97329f33..266eb8cfe93a67 100644
---- a/mm/vmscan.c
-+++ b/mm/vmscan.c
-@@ -2955,8 +2955,8 @@ static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
- 	enum lru_list lru;
- 	unsigned long nr_reclaimed = 0;
- 	unsigned long nr_to_reclaim = sc->nr_to_reclaim;
-+	bool proportional_reclaim;
- 	struct blk_plug plug;
--	bool scan_adjusted;
-
- 	get_scan_count(lruvec, sc, nr);
-
-@@ -2974,8 +2974,8 @@ static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
- 	 * abort proportional reclaim if either the file or anon lru has already
- 	 * dropped to zero at the first pass.
- 	 */
--	scan_adjusted = (!cgroup_reclaim(sc) && !current_is_kswapd() &&
--			 sc->priority == DEF_PRIORITY);
-+	proportional_reclaim = (!cgroup_reclaim(sc) && !current_is_kswapd() &&
-+				sc->priority == DEF_PRIORITY);
-
- 	blk_start_plug(&plug);
- 	while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] ||
-@@ -2995,7 +2995,7 @@ static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
-
- 		cond_resched();
-
--		if (nr_reclaimed < nr_to_reclaim || scan_adjusted)
-+		if (nr_reclaimed < nr_to_reclaim || proportional_reclaim)
- 			continue;
-
- 		/*
-@@ -3046,8 +3046,6 @@ static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
- 		nr_scanned = targets[lru] - nr[lru];
- 		nr[lru] = targets[lru] * (100 - percentage) / 100;
- 		nr[lru] -= min(nr[lru], nr_scanned);
--
--		scan_adjusted = true;
- 	}
- 	blk_finish_plug(&plug);
- 	sc->nr_reclaimed += nr_reclaimed;
 From 430daaab3c78de6bd82f10cfb5a0f016c6e583f6 Mon Sep 17 00:00:00 2001
 From: Desmond Cheong Zhi Xi <desmondcheongzx@gmail.com>
 Date: Mon, 4 Oct 2021 14:07:34 -0400