From: "Darrick J. Wong" <djwong@kernel.org>
To: djwong@kernel.org
Cc: linux-xfs@vger.kernel.org, david@fromorbit.com, hch@infradead.org
Subject: [PATCH 10/20] xfs: reduce inactivation delay when quota are tight
Date: Thu, 29 Jul 2021 11:44:48 -0700 [thread overview]
Message-ID: <162758428867.332903.757283672300988786.stgit@magnolia> (raw)
In-Reply-To: <162758423315.332903.16799817941903734904.stgit@magnolia>
From: Darrick J. Wong <djwong@kernel.org>
Implement the same scaling down of inodegc delays when we're tight on
quota.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
fs/xfs/xfs_dquot.h | 10 ++++++
fs/xfs/xfs_icache.c | 86 ++++++++++++++++++++++++++++++++++++++++++++++++---
fs/xfs/xfs_trace.h | 34 ++++++++++++++++++++
3 files changed, 125 insertions(+), 5 deletions(-)
diff --git a/fs/xfs/xfs_dquot.h b/fs/xfs/xfs_dquot.h
index f642884a6834..6b5e3cf40c8b 100644
--- a/fs/xfs/xfs_dquot.h
+++ b/fs/xfs/xfs_dquot.h
@@ -54,6 +54,16 @@ struct xfs_dquot_res {
xfs_qwarncnt_t warnings;
};
+static inline bool
+xfs_dquot_res_over_limits(
+ const struct xfs_dquot_res *qres)
+{
+ if ((qres->softlimit && qres->softlimit < qres->reserved) ||
+ (qres->hardlimit && qres->hardlimit < qres->reserved))
+ return true;
+ return false;
+}
+
/*
* The incore dquot structure
*/
diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c
index 6418e50518f8..7ba80d7bff41 100644
--- a/fs/xfs/xfs_icache.c
+++ b/fs/xfs/xfs_icache.c
@@ -212,6 +212,73 @@ xfs_reclaim_work_queue(
rcu_read_unlock();
}
+/*
+ * Scale down the background work delay if we're close to a quota limit.
+ * Similar to the way that we throttle preallocations, we halve the delay time
+ * for every low free space threshold that isn't met, and we zero it if we're
+ * over the hard limit. Return value is in ms.
+ */
+static inline unsigned int
+xfs_gc_delay_dquot(
+ struct xfs_inode *ip,
+ xfs_dqtype_t type,
+ unsigned int tag,
+ unsigned int delay_ms)
+{
+ struct xfs_dquot *dqp;
+ int64_t freesp;
+ unsigned int shift = 0;
+
+ if (!ip)
+ goto out;
+
+ /*
+ * Leave the delay untouched if there are no quota limits to enforce.
+ * These comparisons are done locklessly because at worst we schedule
+ * background work sooner than necessary.
+ */
+ dqp = xfs_inode_dquot(ip, type);
+ if (!dqp || !xfs_dquot_is_enforced(dqp))
+ goto out;
+
+ if (xfs_dquot_res_over_limits(&dqp->q_ino) ||
+ xfs_dquot_res_over_limits(&dqp->q_rtb)) {
+ trace_xfs_gc_delay_dquot(dqp, tag, 32);
+ return 0;
+ }
+
+ /* no hi watermark, no throttle */
+ if (!dqp->q_prealloc_hi_wmark)
+ goto out;
+
+ /* under the lo watermark, no throttle */
+ if (dqp->q_blk.reserved < dqp->q_prealloc_lo_wmark)
+ goto out;
+
+ /* If we're over the hard limit, run immediately. */
+ if (dqp->q_blk.reserved >= dqp->q_prealloc_hi_wmark) {
+ trace_xfs_gc_delay_dquot(dqp, tag, 32);
+ return 0;
+ }
+
+ /* Scale down the delay if we're close to the soft limits. */
+ freesp = dqp->q_prealloc_hi_wmark - dqp->q_blk.reserved;
+ if (freesp < dqp->q_low_space[XFS_QLOWSP_5_PCNT]) {
+ shift = 2;
+ if (freesp < dqp->q_low_space[XFS_QLOWSP_3_PCNT])
+ shift += 2;
+ if (freesp < dqp->q_low_space[XFS_QLOWSP_1_PCNT])
+ shift += 2;
+ }
+
+ if (shift)
+ trace_xfs_gc_delay_dquot(dqp, tag, shift);
+
+ delay_ms >>= shift;
+out:
+ return delay_ms;
+}
+
/*
* Scale down the background work delay if we're low on free space. Similar to
* the way that we throttle preallocations, we halve the delay time for every
@@ -247,14 +314,17 @@ xfs_gc_delay_freesp(
/*
* Compute the lag between scheduling and executing some kind of background
- * garbage collection work. Return value is in ms.
+ * garbage collection work. Return value is in ms. If an inode is passed in,
+ * its dquots will be considered in the lag computation.
*/
static inline unsigned int
xfs_gc_delay_ms(
struct xfs_mount *mp,
+ struct xfs_inode *ip,
unsigned int tag)
{
unsigned int default_ms;
+ unsigned int udelay, gdelay, pdelay, fdelay;
switch (tag) {
case XFS_ICI_INODEGC_TAG:
@@ -272,7 +342,12 @@ xfs_gc_delay_ms(
return 0;
}
- return xfs_gc_delay_freesp(mp, tag, default_ms);
+ udelay = xfs_gc_delay_dquot(ip, XFS_DQTYPE_USER, tag, default_ms);
+ gdelay = xfs_gc_delay_dquot(ip, XFS_DQTYPE_GROUP, tag, default_ms);
+ pdelay = xfs_gc_delay_dquot(ip, XFS_DQTYPE_PROJ, tag, default_ms);
+ fdelay = xfs_gc_delay_freesp(mp, tag, default_ms);
+
+ return min(min(udelay, gdelay), min(pdelay, fdelay));
}
/*
@@ -308,7 +383,7 @@ xfs_inodegc_queue(
if (radix_tree_tagged(&mp->m_perag_tree, XFS_ICI_INODEGC_TAG)) {
unsigned int delay;
- delay = xfs_gc_delay_ms(mp, XFS_ICI_INODEGC_TAG);
+ delay = xfs_gc_delay_ms(mp, ip, XFS_ICI_INODEGC_TAG);
trace_xfs_inodegc_queue(mp, delay);
queue_delayed_work(mp->m_gc_workqueue, &mp->m_inodegc_work,
msecs_to_jiffies(delay));
@@ -323,6 +398,7 @@ xfs_inodegc_queue(
static void
xfs_gc_requeue_now(
struct xfs_mount *mp,
+ struct xfs_inode *ip,
unsigned int tag)
{
struct delayed_work *dwork;
@@ -347,7 +423,7 @@ xfs_gc_requeue_now(
if (!radix_tree_tagged(&mp->m_perag_tree, tag))
goto unlock;
- if (xfs_gc_delay_ms(mp, tag) == default_ms)
+ if (xfs_gc_delay_ms(mp, ip, tag) == default_ms)
goto unlock;
trace_xfs_gc_requeue_now(mp, tag);
@@ -378,7 +454,7 @@ xfs_perag_set_inode_tag(
pag->pag_ici_needs_inactive++;
if (was_tagged) {
- xfs_gc_requeue_now(mp, tag);
+ xfs_gc_requeue_now(mp, ip, tag);
return;
}
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index 2092a8542862..001fd202dbfb 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -213,6 +213,40 @@ TRACE_EVENT(xfs_inodegc_requeue_mempressure,
__entry->caller_ip)
);
+TRACE_EVENT(xfs_gc_delay_dquot,
+ TP_PROTO(struct xfs_dquot *dqp, unsigned int tag, unsigned int shift),
+ TP_ARGS(dqp, tag, shift),
+ TP_STRUCT__entry(
+ __field(dev_t, dev)
+ __field(u32, id)
+ __field(xfs_dqtype_t, type)
+ __field(unsigned int, tag)
+ __field(unsigned int, shift)
+ __field(unsigned long long, reserved)
+ __field(unsigned long long, hi_mark)
+ __field(unsigned long long, lo_mark)
+ ),
+ TP_fast_assign(
+ __entry->dev = dqp->q_mount->m_super->s_dev;
+ __entry->id = dqp->q_id;
+ __entry->type = dqp->q_type;
+ __entry->reserved = dqp->q_blk.reserved;
+ __entry->hi_mark = dqp->q_prealloc_hi_wmark;
+ __entry->lo_mark = dqp->q_prealloc_lo_wmark;
+ __entry->tag = tag;
+ __entry->shift = shift;
+ ),
+ TP_printk("dev %d:%d tag %u shift %u dqid 0x%x dqtype %s reserved %llu hi %llu lo %llu",
+ MAJOR(__entry->dev), MINOR(__entry->dev),
+ __entry->tag,
+ __entry->shift,
+ __entry->id,
+ __print_flags(__entry->type, "|", XFS_DQTYPE_STRINGS),
+ __entry->reserved,
+ __entry->hi_mark,
+ __entry->lo_mark)
+);
+
TRACE_EVENT(xfs_gc_delay_fdblocks,
TP_PROTO(struct xfs_mount *mp, unsigned int tag, unsigned int shift),
TP_ARGS(mp, tag, shift),
next prev parent reply other threads:[~2021-07-29 18:44 UTC|newest]
Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-29 18:43 [PATCHSET v8 00/20] xfs: deferred inode inactivation Darrick J. Wong
2021-07-29 18:43 ` [PATCH 01/20] xfs: move xfs_inactive call to xfs_inode_mark_reclaimable Darrick J. Wong
2021-07-29 18:44 ` [PATCH 02/20] xfs: detach dquots from inode if we don't need to inactivate it Darrick J. Wong
2021-07-29 18:44 ` [PATCH 03/20] xfs: defer inode inactivation to a workqueue Darrick J. Wong
2021-07-30 4:24 ` Dave Chinner
2021-07-31 4:21 ` Darrick J. Wong
2021-08-01 21:49 ` Dave Chinner
2021-08-01 23:47 ` Dave Chinner
2021-08-03 8:34 ` [PATCH, alternative] xfs: per-cpu deferred inode inactivation queues Dave Chinner
2021-08-03 20:20 ` Darrick J. Wong
2021-08-04 3:20 ` [PATCH, alternative v2] " Darrick J. Wong
2021-08-04 10:03 ` [PATCH] xfs: inodegc needs to stop before freeze Dave Chinner
2021-08-04 12:37 ` Dave Chinner
2021-08-04 10:46 ` [PATCH] xfs: don't run inodegc flushes when inodegc is not active Dave Chinner
2021-08-04 16:20 ` Darrick J. Wong
2021-08-04 11:09 ` [PATCH, alternative v2] xfs: per-cpu deferred inode inactivation queues Dave Chinner
2021-08-04 15:59 ` Darrick J. Wong
2021-08-04 21:35 ` Dave Chinner
2021-08-04 11:49 ` [PATCH, pre-03/20 #1] xfs: introduce CPU hotplug infrastructure Dave Chinner
2021-08-04 11:50 ` [PATCH, pre-03/20 #2] xfs: introduce all-mounts list for cpu hotplug notifications Dave Chinner
2021-08-04 16:06 ` Darrick J. Wong
2021-08-04 21:17 ` Dave Chinner
2021-08-04 11:52 ` [PATCH, post-03/20 1/1] xfs: hook up inodegc to CPU dead notification Dave Chinner
2021-08-04 16:19 ` Darrick J. Wong
2021-08-04 21:48 ` Dave Chinner
2021-07-29 18:44 ` [PATCH 04/20] xfs: throttle inode inactivation queuing on memory reclaim Darrick J. Wong
2021-07-29 18:44 ` [PATCH 05/20] xfs: don't throttle memory reclaim trying to queue inactive inodes Darrick J. Wong
2021-07-29 18:44 ` [PATCH 06/20] xfs: throttle inodegc queuing on backlog Darrick J. Wong
2021-08-02 0:45 ` Dave Chinner
2021-08-02 1:30 ` Dave Chinner
2021-07-29 18:44 ` [PATCH 07/20] xfs: queue inodegc worker immediately when memory is tight Darrick J. Wong
2021-07-29 18:44 ` [PATCH 08/20] xfs: expose sysfs knob to control inode inactivation delay Darrick J. Wong
2021-07-29 18:44 ` [PATCH 09/20] xfs: reduce inactivation delay when free space is tight Darrick J. Wong
2021-07-29 18:44 ` Darrick J. Wong [this message]
2021-07-29 18:44 ` [PATCH 11/20] xfs: reduce inactivation delay when realtime extents are tight Darrick J. Wong
2021-07-29 18:44 ` [PATCH 12/20] xfs: inactivate inodes any time we try to free speculative preallocations Darrick J. Wong
2021-07-29 18:45 ` [PATCH 13/20] xfs: flush inode inactivation work when compiling usage statistics Darrick J. Wong
2021-07-29 18:45 ` [PATCH 14/20] xfs: parallelize inode inactivation Darrick J. Wong
2021-08-02 0:55 ` Dave Chinner
2021-08-02 21:33 ` Darrick J. Wong
2021-07-29 18:45 ` [PATCH 15/20] xfs: reduce inactivation delay when AG free space are tight Darrick J. Wong
2021-07-29 18:45 ` [PATCH 16/20] xfs: queue inodegc worker immediately on backlog Darrick J. Wong
2021-07-29 18:45 ` [PATCH 17/20] xfs: don't run speculative preallocation gc when fs is frozen Darrick J. Wong
2021-07-29 18:45 ` [PATCH 18/20] xfs: scale speculative preallocation gc delay based on free space Darrick J. Wong
2021-07-29 18:45 ` [PATCH 19/20] xfs: use background worker pool when transactions can't get " Darrick J. Wong
2021-07-29 18:45 ` [PATCH 20/20] xfs: avoid buffer deadlocks when walking fs inodes Darrick J. Wong
2021-08-02 10:35 ` [PATCHSET v8 00/20] xfs: deferred inode inactivation Dave Chinner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=162758428867.332903.757283672300988786.stgit@magnolia \
--to=djwong@kernel.org \
--cc=david@fromorbit.com \
--cc=hch@infradead.org \
--cc=linux-xfs@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).