lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date:   Tue, 15 Mar 2022 21:14:14 -0700
From:   Daeho Jeong <daeho43@...il.com>
To:     linux-kernel@...r.kernel.org,
        linux-f2fs-devel@...ts.sourceforge.net, kernel-team@...roid.com
Cc:     Daeho Jeong <daehojeong@...gle.com>
Subject: [PATCH] f2fs: introduce gc_urgent_mid mode

From: Daeho Jeong <daehojeong@...gle.com>

We need a mid level of gc urgent mode to do GC forcibly in a period
of given gc_urgent_sleep_time, but not like using greedy GC approach
and switching to SSR mode such as gc urgent high mode. This can be
used for more aggressive periodic storage clean up.

Signed-off-by: Daeho Jeong <daehojeong@...gle.com>
---
 Documentation/ABI/testing/sysfs-fs-f2fs | 17 +++++++++++------
 fs/f2fs/debug.c                         |  4 +++-
 fs/f2fs/f2fs.h                          |  4 ++++
 fs/f2fs/gc.c                            |  3 +++
 fs/f2fs/sysfs.c                         |  7 +++++++
 5 files changed, 28 insertions(+), 7 deletions(-)

diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs
index 5a5f3c5445f6..9b583dd0298b 100644
--- a/Documentation/ABI/testing/sysfs-fs-f2fs
+++ b/Documentation/ABI/testing/sysfs-fs-f2fs
@@ -297,11 +297,16 @@ Description:	Shows current reserved blocks in system, it may be temporarily
 What:		/sys/fs/f2fs/<disk>/gc_urgent
 Date:		August 2017
 Contact:	"Jaegeuk Kim" <jaegeuk@...nel.org>
-Description:	Do background GC aggressively when set. When gc_urgent = 1,
-		background thread starts to do GC by given gc_urgent_sleep_time
-		interval. When gc_urgent = 2, F2FS will lower the bar of
-		checking idle in order to process outstanding discard commands
-		and GC a little bit aggressively. It is set to 0 by default.
+Description:	Do background GC aggressively when set. Set to 0 by default.
+		gc urgent high(1): does GC forcibly in a period of given
+		gc_urgent_sleep_time and ignores I/O idling check. uses greedy
+		GC approach and turns SSR mode on.
+		gc urgent low(2): lowers the bar of checking I/O idling in
+		order to process outstanding discard commands and GC a
+		little bit aggressively. uses cost benefit GC approach.
+		gc urgent mid(3): does GC forcibly in a period of given
+		gc_urgent_sleep_time and executes a mid level of I/O idling check.
+		uses cost benefit GC approach.
 
 What:		/sys/fs/f2fs/<disk>/gc_urgent_sleep_time
 Date:		August 2017
@@ -532,7 +537,7 @@ Date:		July 2021
 Contact:	"Daeho Jeong" <daehojeong@...gle.com>
 Description:	Show how many segments have been reclaimed by GC during a specific
 		GC mode (0: GC normal, 1: GC idle CB, 2: GC idle greedy,
-		3: GC idle AT, 4: GC urgent high, 5: GC urgent low)
+		3: GC idle AT, 4: GC urgent high, 5: GC urgent low 6: GC urgent mid)
 		You can re-initialize this value to "0".
 
 What:		/sys/fs/f2fs/<disk>/gc_segment_mode
diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index 6d26872c7364..fcdf253cd211 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -476,12 +476,14 @@ static int stat_show(struct seq_file *s, void *v)
 				si->node_segs, si->bg_node_segs);
 		seq_printf(s, "  - Reclaimed segs : Normal (%d), Idle CB (%d), "
 				"Idle Greedy (%d), Idle AT (%d), "
-				"Urgent High (%d), Urgent Low (%d)\n",
+				"Urgent High (%d), Urgent Mid (%d), "
+				"Urgent Low (%d)\n",
 				si->sbi->gc_reclaimed_segs[GC_NORMAL],
 				si->sbi->gc_reclaimed_segs[GC_IDLE_CB],
 				si->sbi->gc_reclaimed_segs[GC_IDLE_GREEDY],
 				si->sbi->gc_reclaimed_segs[GC_IDLE_AT],
 				si->sbi->gc_reclaimed_segs[GC_URGENT_HIGH],
+				si->sbi->gc_reclaimed_segs[GC_URGENT_MID],
 				si->sbi->gc_reclaimed_segs[GC_URGENT_LOW]);
 		seq_printf(s, "Try to move %d blocks (BG: %d)\n", si->tot_blks,
 				si->bg_data_blks + si->bg_node_blks);
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index da729f53daa8..3016b6354a68 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -1313,6 +1313,7 @@ enum {
 	GC_IDLE_AT,
 	GC_URGENT_HIGH,
 	GC_URGENT_LOW,
+	GC_URGENT_MID,
 	MAX_GC_MODE,
 };
 
@@ -2784,6 +2785,9 @@ static inline bool is_idle(struct f2fs_sb_info *sbi, int type)
 	if (is_inflight_io(sbi, type))
 		return false;
 
+	if (sbi->gc_mode == GC_URGENT_MID)
+		return true;
+
 	if (sbi->gc_mode == GC_URGENT_LOW &&
 			(type == DISCARD_TIME || type == GC_TIME))
 		return true;
diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
index 2d53ef121e76..ea5b93b689cd 100644
--- a/fs/f2fs/gc.c
+++ b/fs/f2fs/gc.c
@@ -103,7 +103,10 @@ static int gc_thread_func(void *data)
 				sbi->gc_urgent_high_remaining--;
 			}
 			spin_unlock(&sbi->gc_urgent_high_lock);
+		}
 
+		if (sbi->gc_mode == GC_URGENT_HIGH ||
+				sbi->gc_mode == GC_URGENT_MID) {
 			wait_ms = gc_th->urgent_sleep_time;
 			f2fs_down_write(&sbi->gc_lock);
 			goto do_gc;
diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
index 47efcf233afd..fe29bcb70f46 100644
--- a/fs/f2fs/sysfs.c
+++ b/fs/f2fs/sysfs.c
@@ -468,6 +468,13 @@ static ssize_t __sbi_store(struct f2fs_attr *a,
 			}
 		} else if (t == 2) {
 			sbi->gc_mode = GC_URGENT_LOW;
+		} else if (t == 3) {
+			sbi->gc_mode = GC_URGENT_MID;
+			if (sbi->gc_thread) {
+				sbi->gc_thread->gc_wake = 1;
+				wake_up_interruptible_all(
+					&sbi->gc_thread->gc_wait_queue_head);
+			}
 		} else {
 			return -EINVAL;
 		}
-- 
2.35.1.723.g4982287a31-goog

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ