sravan/linux-tkg
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
Files
master
linux-tkg/linux-tkg-patches/6.4/0002-clear-patches.patch
Tk-Glitch feb1867486 Add 6.4 rc initial support
2023-05-15 16:38:06 +02:00

621 lines
21 KiB
Diff
Raw Permalink Blame History

From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Arjan van de Ven <arjan@linux.intel.com>
Date: Mon, 14 Mar 2016 11:10:58 -0600
Subject: [PATCH] pci pme wakeups
Reduce wakeups for PME checks, which are a workaround for miswired
boards (sadly, too many of them) in laptops.
---
drivers/pci/pci.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
index c9338f9..6974fbf 100644
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -62,7 +62,7 @@ struct pci_pme_device {
struct pci_dev *dev;
};
-#define PME_TIMEOUT 1000 /* How long between PME checks */
+#define PME_TIMEOUT 4000 /* How long between PME checks */
static void pci_dev_d3_sleep(struct pci_dev *dev)
{
--
https://clearlinux.org
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Arjan van de Ven <arjan@linux.intel.com>
Date: Sat, 19 Mar 2016 21:32:19 -0400
Subject: [PATCH] intel_idle: tweak cpuidle cstates
Increase target_residency in cpuidle cstate
Tune intel_idle to be a bit less agressive;
Clear linux is cleaner in hygiene (wakupes) than the average linux,
so we can afford changing these in a way that increases
performance while keeping power efficiency
---
drivers/idle/intel_idle.c | 44 +++++++++++++++++++--------------------
1 file changed, 22 insertions(+), 22 deletions(-)
diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
index f449584..c994d24 100644
--- a/drivers/idle/intel_idle.c
+++ b/drivers/idle/intel_idle.c
@@ -531,7 +531,7 @@ static struct cpuidle_state hsw_cstates[] __initdata = {
.desc = "MWAIT 0x01",
.flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_ALWAYS_ENABLE,
.exit_latency = 10,
- .target_residency = 20,
+ .target_residency = 120,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -539,7 +539,7 @@ static struct cpuidle_state hsw_cstates[] __initdata = {
.desc = "MWAIT 0x10",
.flags = MWAIT2flg(0x10) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 33,
- .target_residency = 100,
+ .target_residency = 900,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -547,7 +547,7 @@ static struct cpuidle_state hsw_cstates[] __initdata = {
.desc = "MWAIT 0x20",
.flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 133,
- .target_residency = 400,
+ .target_residency = 1000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -555,7 +555,7 @@ static struct cpuidle_state hsw_cstates[] __initdata = {
.desc = "MWAIT 0x32",
.flags = MWAIT2flg(0x32) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 166,
- .target_residency = 500,
+ .target_residency = 1500,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -563,7 +563,7 @@ static struct cpuidle_state hsw_cstates[] __initdata = {
.desc = "MWAIT 0x40",
.flags = MWAIT2flg(0x40) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 300,
- .target_residency = 900,
+ .target_residency = 2000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -571,7 +571,7 @@ static struct cpuidle_state hsw_cstates[] __initdata = {
.desc = "MWAIT 0x50",
.flags = MWAIT2flg(0x50) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 600,
- .target_residency = 1800,
+ .target_residency = 5000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -579,7 +579,7 @@ static struct cpuidle_state hsw_cstates[] __initdata = {
.desc = "MWAIT 0x60",
.flags = MWAIT2flg(0x60) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 2600,
- .target_residency = 7700,
+ .target_residency = 9000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -599,7 +599,7 @@ static struct cpuidle_state bdw_cstates[] __initdata = {
.desc = "MWAIT 0x01",
.flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_ALWAYS_ENABLE,
.exit_latency = 10,
- .target_residency = 20,
+ .target_residency = 120,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -607,7 +607,7 @@ static struct cpuidle_state bdw_cstates[] __initdata = {
.desc = "MWAIT 0x10",
.flags = MWAIT2flg(0x10) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 40,
- .target_residency = 100,
+ .target_residency = 1000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -615,7 +615,7 @@ static struct cpuidle_state bdw_cstates[] __initdata = {
.desc = "MWAIT 0x20",
.flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 133,
- .target_residency = 400,
+ .target_residency = 1000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -623,7 +623,7 @@ static struct cpuidle_state bdw_cstates[] __initdata = {
.desc = "MWAIT 0x32",
.flags = MWAIT2flg(0x32) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 166,
- .target_residency = 500,
+ .target_residency = 2000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -631,7 +631,7 @@ static struct cpuidle_state bdw_cstates[] __initdata = {
.desc = "MWAIT 0x40",
.flags = MWAIT2flg(0x40) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 300,
- .target_residency = 900,
+ .target_residency = 4000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -639,7 +639,7 @@ static struct cpuidle_state bdw_cstates[] __initdata = {
.desc = "MWAIT 0x50",
.flags = MWAIT2flg(0x50) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 600,
- .target_residency = 1800,
+ .target_residency = 7000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -647,7 +647,7 @@ static struct cpuidle_state bdw_cstates[] __initdata = {
.desc = "MWAIT 0x60",
.flags = MWAIT2flg(0x60) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 2600,
- .target_residency = 7700,
+ .target_residency = 9000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -668,7 +668,7 @@ static struct cpuidle_state skl_cstates[] __initdata = {
.desc = "MWAIT 0x01",
.flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_ALWAYS_ENABLE,
.exit_latency = 10,
- .target_residency = 20,
+ .target_residency = 120,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -676,7 +676,7 @@ static struct cpuidle_state skl_cstates[] __initdata = {
.desc = "MWAIT 0x10",
.flags = MWAIT2flg(0x10) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 70,
- .target_residency = 100,
+ .target_residency = 1000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -684,7 +684,7 @@ static struct cpuidle_state skl_cstates[] __initdata = {
.desc = "MWAIT 0x20",
.flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 85,
- .target_residency = 200,
+ .target_residency = 600,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -692,7 +692,7 @@ static struct cpuidle_state skl_cstates[] __initdata = {
.desc = "MWAIT 0x33",
.flags = MWAIT2flg(0x33) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 124,
- .target_residency = 800,
+ .target_residency = 3000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -700,7 +700,7 @@ static struct cpuidle_state skl_cstates[] __initdata = {
.desc = "MWAIT 0x40",
.flags = MWAIT2flg(0x40) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 200,
- .target_residency = 800,
+ .target_residency = 3200,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -708,7 +708,7 @@ static struct cpuidle_state skl_cstates[] __initdata = {
.desc = "MWAIT 0x50",
.flags = MWAIT2flg(0x50) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 480,
- .target_residency = 5000,
+ .target_residency = 9000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -716,7 +716,7 @@ static struct cpuidle_state skl_cstates[] __initdata = {
.desc = "MWAIT 0x60",
.flags = MWAIT2flg(0x60) | CPUIDLE_FLAG_TLB_FLUSHED,
.exit_latency = 890,
- .target_residency = 5000,
+ .target_residency = 9000,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
@@ -737,7 +737,7 @@ static struct cpuidle_state skx_cstates[] __initdata = {
.desc = "MWAIT 0x01",
.flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_ALWAYS_ENABLE,
.exit_latency = 10,
- .target_residency = 20,
+ .target_residency = 300,
.enter = &intel_idle,
.enter_s2idle = intel_idle_s2idle, },
{
--
https://clearlinux.org
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Arjan van de Ven <arjan@linux.intel.com>
Date: Fri, 6 Jan 2017 15:34:09 +0000
Subject: [PATCH] ipv4/tcp: allow the memory tuning for tcp to go a little
bigger than default
---
net/ipv4/tcp.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 30c1142..4345075 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -4201,8 +4201,8 @@ void __init tcp_init(void)
tcp_init_mem();
/* Set per-socket limits to no more than 1/128 the pressure threshold */
limit = nr_free_buffer_pages() << (PAGE_SHIFT - 7);
- max_wshare = min(4UL*1024*1024, limit);
- max_rshare = min(6UL*1024*1024, limit);
+ max_wshare = min(16UL*1024*1024, limit);
+ max_rshare = min(16UL*1024*1024, limit);
init_net.ipv4.sysctl_tcp_wmem[0] = SK_MEM_QUANTUM;
init_net.ipv4.sysctl_tcp_wmem[1] = 16*1024;
--
https://clearlinux.org
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Arjan van de Ven <arjan@linux.intel.com>
Date: Sun, 18 Feb 2018 23:35:41 +0000
Subject: [PATCH] locking: rwsem: spin faster
tweak rwsem owner spinning a bit
---
kernel/locking/rwsem.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/kernel/locking/rwsem.c b/kernel/locking/rwsem.c
index f11b9bd..1bbfcc1 100644
--- a/kernel/locking/rwsem.c
+++ b/kernel/locking/rwsem.c
@@ -717,6 +717,7 @@ rwsem_spin_on_owner(struct rw_semaphore *sem, unsigned long nonspinnable)
struct task_struct *new, *owner;
unsigned long flags, new_flags;
enum owner_state state;
+ int i = 0;
owner = rwsem_owner_flags(sem, &flags);
state = rwsem_owner_state(owner, flags, nonspinnable);
@@ -750,7 +751,8 @@ rwsem_spin_on_owner(struct rw_semaphore *sem, unsigned long nonspinnable)
break;
}
- cpu_relax();
+ if (i++ > 1000)
+ cpu_relax();
}
rcu_read_unlock();
--
https://clearlinux.org
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Arjan van de Ven <arjan@linux.intel.com>
Date: Thu, 2 Jun 2016 23:36:32 -0500
Subject: [PATCH] initialize ata before graphics
ATA init is the long pole in the boot process, and its asynchronous.
move the graphics init after it so that ata and graphics initialize
in parallel
---
drivers/Makefile | 15 ++++++++-------
1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/drivers/Makefile b/drivers/Makefile
index c0cd1b9..af1e2fb 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -59,15 +59,8 @@ obj-y += char/
# iommu/ comes before gpu as gpu are using iommu controllers
obj-y += iommu/
-# gpu/ comes after char for AGP vs DRM startup and after iommu
-obj-y += gpu/
-
obj-$(CONFIG_CONNECTOR) += connector/
-# i810fb and intelfb depend on char/agp/
-obj-$(CONFIG_FB_I810) += video/fbdev/i810/
-obj-$(CONFIG_FB_INTEL) += video/fbdev/intelfb/
-
obj-$(CONFIG_PARPORT) += parport/
obj-$(CONFIG_NVM) += lightnvm/
obj-y += base/ block/ misc/ mfd/ nfc/
@@ -80,6 +73,14 @@ obj-$(CONFIG_IDE) += ide/
obj-y += scsi/
obj-y += nvme/
obj-$(CONFIG_ATA) += ata/
+
+# gpu/ comes after char for AGP vs DRM startup and after iommu
+obj-y += gpu/
+
+# i810fb and intelfb depend on char/agp/
+obj-$(CONFIG_FB_I810) += video/fbdev/i810/
+obj-$(CONFIG_FB_INTEL) += video/fbdev/intelfb/
+
obj-$(CONFIG_TARGET_CORE) += target/
obj-$(CONFIG_MTD) += mtd/
obj-$(CONFIG_SPI) += spi/
--
https://clearlinux.org
From 676c2dc63592f52b716515573a3a825582a371e9 Mon Sep 17 00:00:00 2001
From: Arjan van de Ven <arjan@linux.intel.com>
Date: Sat, 8 Dec 2018 18:21:32 +0000
Subject: [PATCH 1/9] x86/vdso: Use lfence instead of rep and nop
Signed-off-by: Alexandre Frade <kernel@xanmod.org>
---
arch/x86/include/asm/vdso/processor.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/x86/include/asm/vdso/processor.h b/arch/x86/include/asm/vdso/processor.h
index 57b1a7034c64..e2c45674f989 100644
--- a/arch/x86/include/asm/vdso/processor.h
+++ b/arch/x86/include/asm/vdso/processor.h
@@ -10,7 +10,7 @@
/* REP NOP (PAUSE) is a good thing to insert into busy-wait loops. */
static __always_inline void rep_nop(void)
{
- asm volatile("rep; nop" ::: "memory");
+ asm volatile("lfence" ::: "memory");
}
static __always_inline void cpu_relax(void)
--
2.39.1
From 48dc9669f8db68adc480ffc2698ed8204440e45b Mon Sep 17 00:00:00 2001
From: Arjan van de Ven <arjan@linux.intel.com>
Date: Thu, 13 Dec 2018 01:00:49 +0000
Subject: [PATCH 2/9] sched/wait: Do accept() in LIFO order for cache
efficiency
Signed-off-by: Alexandre Frade <kernel@xanmod.org>
---
include/linux/wait.h | 2 ++
kernel/sched/wait.c | 24 ++++++++++++++++++++++++
net/ipv4/inet_connection_sock.c | 2 +-
3 files changed, 27 insertions(+), 1 deletion(-)
diff --git a/include/linux/wait.h b/include/linux/wait.h
index a0307b516b09..edc21128f387 100644
--- a/include/linux/wait.h
+++ b/include/linux/wait.h
@@ -165,6 +165,7 @@ static inline bool wq_has_sleeper(struct wait_queue_head *wq_head)
extern void add_wait_queue(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry);
extern void add_wait_queue_exclusive(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry);
+extern void add_wait_queue_exclusive_lifo(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry);
extern void add_wait_queue_priority(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry);
extern void remove_wait_queue(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry);
@@ -1192,6 +1193,7 @@ do { \
*/
void prepare_to_wait(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry, int state);
bool prepare_to_wait_exclusive(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry, int state);
+void prepare_to_wait_exclusive_lifo(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry, int state);
long prepare_to_wait_event(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry, int state);
void finish_wait(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry);
long wait_woken(struct wait_queue_entry *wq_entry, unsigned mode, long timeout);
diff --git a/kernel/sched/wait.c b/kernel/sched/wait.c
index 133b74730738..1647fb8662eb 100644
--- a/kernel/sched/wait.c
+++ b/kernel/sched/wait.c
@@ -47,6 +47,17 @@ void add_wait_queue_priority(struct wait_queue_head *wq_head, struct wait_queue_
}
EXPORT_SYMBOL_GPL(add_wait_queue_priority);
+void add_wait_queue_exclusive_lifo(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry)
+{
+ unsigned long flags;
+
+ wq_entry->flags |= WQ_FLAG_EXCLUSIVE;
+ spin_lock_irqsave(&wq_head->lock, flags);
+ __add_wait_queue(wq_head, wq_entry);
+ spin_unlock_irqrestore(&wq_head->lock, flags);
+}
+EXPORT_SYMBOL(add_wait_queue_exclusive_lifo);
+
void remove_wait_queue(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry)
{
unsigned long flags;
@@ -293,6 +304,19 @@ prepare_to_wait_exclusive(struct wait_queue_head *wq_head, struct wait_queue_ent
}
EXPORT_SYMBOL(prepare_to_wait_exclusive);
+void prepare_to_wait_exclusive_lifo(struct wait_queue_head *wq_head, struct wait_queue_entry *wq_entry, int state)
+{
+ unsigned long flags;
+
+ wq_entry->flags |= WQ_FLAG_EXCLUSIVE;
+ spin_lock_irqsave(&wq_head->lock, flags);
+ if (list_empty(&wq_entry->entry))
+ __add_wait_queue(wq_head, wq_entry);
+ set_current_state(state);
+ spin_unlock_irqrestore(&wq_head->lock, flags);
+}
+EXPORT_SYMBOL(prepare_to_wait_exclusive_lifo);
+
void init_wait_entry(struct wait_queue_entry *wq_entry, int flags)
{
wq_entry->flags = flags;
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index f2c43f67187d..9885bfb429a2 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -606,7 +606,7 @@ static int inet_csk_wait_for_connect(struct sock *sk, long timeo)
* having to remove and re-insert us on the wait queue.
*/
for (;;) {
- prepare_to_wait_exclusive(sk_sleep(sk), &wait,
+ prepare_to_wait_exclusive_lifo(sk_sleep(sk), &wait,
TASK_INTERRUPTIBLE);
release_sock(sk);
if (reqsk_queue_empty(&icsk->icsk_accept_queue))
--
2.39.1
From afa213811c5490906caf394b20bb4b616fc6f12a Mon Sep 17 00:00:00 2001
From: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Date: Thu, 25 Aug 2022 15:55:26 -0700
Subject: [PATCH 3/9] sched/fair: Simplify asym_packing logic for SMT sched
groups
When the destination CPU is an SMT sibling and idle, it can only help the
busiest group if all of its other SMT siblings are also idle. Otherwise,
there is not increase in throughput.
It does not matter whether the busiest group has SMT siblings. Simply
check if there are any tasks running on the local group before proceeding.
Cc: Ben Segall <bsegall@google.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Tim C. Chen <tim.c.chen@intel.com>
Cc: Valentin Schneider <vschneid@redhat.com>
Cc: x86@kernel.org
Cc: linux-kernel@vger.kernel.org
Reviewed-by: Len Brown <len.brown@intel.com>
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Alexandre Frade <kernel@xanmod.org>
---
kernel/sched/fair.c | 29 +++++++++--------------------
1 file changed, 9 insertions(+), 20 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 345cc5e9fa6e..60f9690a5626 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -8921,12 +8921,10 @@ static bool asym_smt_can_pull_tasks(int dst_cpu, struct sd_lb_stats *sds,
struct sched_group *sg)
{
#ifdef CONFIG_SCHED_SMT
- bool local_is_smt, sg_is_smt;
+ bool local_is_smt;
int sg_busy_cpus;
local_is_smt = sds->local->flags & SD_SHARE_CPUCAPACITY;
- sg_is_smt = sg->flags & SD_SHARE_CPUCAPACITY;
-
sg_busy_cpus = sgs->group_weight - sgs->idle_cpus;
if (!local_is_smt) {
@@ -8947,25 +8945,16 @@ static bool asym_smt_can_pull_tasks(int dst_cpu, struct sd_lb_stats *sds,
return sched_asym_prefer(dst_cpu, sg->asym_prefer_cpu);
}
- /* @dst_cpu has SMT siblings. */
-
- if (sg_is_smt) {
- int local_busy_cpus = sds->local->group_weight -
- sds->local_stat.idle_cpus;
- int busy_cpus_delta = sg_busy_cpus - local_busy_cpus;
-
- if (busy_cpus_delta == 1)
- return sched_asym_prefer(dst_cpu, sg->asym_prefer_cpu);
-
- return false;
- }
-
/*
- * @sg does not have SMT siblings. Ensure that @sds::local does not end
- * up with more than one busy SMT sibling and only pull tasks if there
- * are not busy CPUs (i.e., no CPU has running tasks).
+ * @dst_cpu has SMT siblings. When both @dst_cpu and the busiest core
+ * have one or more busy siblings, moving tasks between them results
+ * in the same throughput. Only if all the siblings of @dst_cpu are
+ * idle throughput can increase.
+ *
+ * If the difference in the number of busy CPUs is two or more, let
+ * find_busiest_group() take care of it.
*/
- if (!sds->local_stat.sum_nr_running)
+ if (sg_busy_cpus == 1 && !sds->local_stat.sum_nr_running)
return sched_asym_prefer(dst_cpu, sg->asym_prefer_cpu);
return false;
--
2.39.1
From a1f627fd10ced4f5eeae678bc4ba96ea7fa01d7e Mon Sep 17 00:00:00 2001
From: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Date: Thu, 25 Aug 2022 15:55:28 -0700
Subject: [PATCH 4/9] sched/fair: Let lower-priority CPUs do active balancing
When more than one SMT siblings of a physical core are busy, an idle CPU
of lower priority can help.
Indicate that the low priority CPU can do active balancing from the high-
priority CPU only if they belong to separate cores.
Cc: Ben Segall <bsegall@google.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Tim C. Chen <tim.c.chen@intel.com>
Cc: Valentin Schneider <vschneid@redhat.com>
Cc: x86@kernel.org
Cc: linux-kernel@vger.kernel.org
Reviewed-by: Len Brown <len.brown@intel.com>
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Alexandre Frade <kernel@xanmod.org>
---
kernel/sched/fair.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 60f9690a5626..67b0eacad0e9 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -10176,9 +10176,14 @@ asym_active_balance(struct lb_env *env)
* ASYM_PACKING needs to force migrate tasks from busy but
* lower priority CPUs in order to pack all tasks in the
* highest priority CPUs.
+ *
+ * If the busy CPU has higher priority but is an SMT sibling
+ * in which other SMT siblings are also busy, a lower-priority
+ * CPU in a separate core can help.
*/
return env->idle != CPU_NOT_IDLE && (env->sd->flags & SD_ASYM_PACKING) &&
- sched_asym_prefer(env->dst_cpu, env->src_cpu);
+ (sched_asym_prefer(env->dst_cpu, env->src_cpu) ||
+ !(env->sd->flags & SD_SHARE_CPUCAPACITY));
}
static inline bool
--
2.39.1
Reference in New Issue View Git Blame Copy Permalink