From: Johannes Weiner <hannes@cmpxchg.org>
To: Yosry Ahmed <yosryahmed@google.com>
Cc: Tejun Heo <tj@kernel.org>, Josef Bacik <josef@toxicpanda.com>,
Jens Axboe <axboe@kernel.dk>, Zefan Li <lizefan.x@bytedance.com>,
Michal Hocko <mhocko@kernel.org>,
Roman Gushchin <roman.gushchin@linux.dev>,
Shakeel Butt <shakeelb@google.com>,
Muchun Song <muchun.song@linux.dev>,
Andrew Morton <akpm@linux-foundation.org>,
Vasily Averin <vasily.averin@linux.dev>,
cgroups@vger.kernel.org, linux-block@vger.kernel.org,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
bpf@vger.kernel.org
Subject: Re: [RFC PATCH 3/7] cgroup: rstat: remove cgroup_rstat_flush_irqsafe()
Date: Thu, 23 Mar 2023 11:43:04 -0400 [thread overview]
Message-ID: <20230323154304.GA739026@cmpxchg.org> (raw)
In-Reply-To: <20230323040037.2389095-4-yosryahmed@google.com>
On Thu, Mar 23, 2023 at 04:00:33AM +0000, Yosry Ahmed wrote:
> The naming of cgroup_rstat_flush_irqsafe() can be confusing.
> It can read like "irqsave", which means that it disables
> interrupts throughout, but it actually doesn't. It is just "safe" to
> call from contexts with interrupts disabled.
>
> Furthermore, this is only used today by mem_cgroup_flush_stats(), which
> will be changed by a later patch to optionally allow sleeping. Simplify
> the code and make it easier to reason about by instead passing in an
> explicit @may_sleep argument to cgroup_rstat_flush(), which gets passed
> directly to cgroup_rstat_flush_locked().
>
> Signed-off-by: Yosry Ahmed <yosryahmed@google.com>
> ---
> block/blk-cgroup.c | 2 +-
> include/linux/cgroup.h | 3 +--
> kernel/cgroup/cgroup.c | 4 ++--
> kernel/cgroup/rstat.c | 24 +++++-------------------
> mm/memcontrol.c | 6 +++---
> 5 files changed, 12 insertions(+), 27 deletions(-)
>
> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index bd50b55bdb61..3fe313ce5e6b 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -1043,7 +1043,7 @@ static int blkcg_print_stat(struct seq_file *sf, void *v)
> if (!seq_css(sf)->parent)
> blkcg_fill_root_iostats();
> else
> - cgroup_rstat_flush(blkcg->css.cgroup);
> + cgroup_rstat_flush(blkcg->css.cgroup, true);
>
> rcu_read_lock();
> hlist_for_each_entry_rcu(blkg, &blkcg->blkg_list, blkcg_node) {
> diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
> index 3410aecffdb4..743c345b6a3f 100644
> --- a/include/linux/cgroup.h
> +++ b/include/linux/cgroup.h
> @@ -691,8 +691,7 @@ static inline void cgroup_path_from_kernfs_id(u64 id, char *buf, size_t buflen)
> * cgroup scalable recursive statistics.
> */
> void cgroup_rstat_updated(struct cgroup *cgrp, int cpu);
> -void cgroup_rstat_flush(struct cgroup *cgrp);
> -void cgroup_rstat_flush_irqsafe(struct cgroup *cgrp);
> +void cgroup_rstat_flush(struct cgroup *cgrp, bool may_sleep);
> void cgroup_rstat_flush_hold(struct cgroup *cgrp);
> void cgroup_rstat_flush_release(void);
>
> diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
> index 935e8121b21e..58df0fc379f6 100644
> --- a/kernel/cgroup/cgroup.c
> +++ b/kernel/cgroup/cgroup.c
> @@ -5393,7 +5393,7 @@ static void css_release_work_fn(struct work_struct *work)
> if (ss) {
> /* css release path */
> if (!list_empty(&css->rstat_css_node)) {
> - cgroup_rstat_flush(cgrp);
> + cgroup_rstat_flush(cgrp, true);
> list_del_rcu(&css->rstat_css_node);
> }
>
> @@ -5406,7 +5406,7 @@ static void css_release_work_fn(struct work_struct *work)
> /* cgroup release path */
> TRACE_CGROUP_PATH(release, cgrp);
>
> - cgroup_rstat_flush(cgrp);
> + cgroup_rstat_flush(cgrp, true);
This is an anti-pattern, please don't do this. Naked bool arguments
are a pain to comprehend at the callsite and an easy vector for bugs.
cgroup_rstat_flush_atomic() would be a better name for, well, atomic
callsites.
> diff --git a/kernel/cgroup/rstat.c b/kernel/cgroup/rstat.c
> index af11e28bd055..fe8690bb1e1c 100644
> --- a/kernel/cgroup/rstat.c
> +++ b/kernel/cgroup/rstat.c
> @@ -243,30 +243,16 @@ static bool should_skip_flush(void)
> * This function is safe to call from contexts that disable interrupts, but
> * @may_sleep must be set to false, otherwise the function may block.
> */
> -__bpf_kfunc void cgroup_rstat_flush(struct cgroup *cgrp)
> +__bpf_kfunc void cgroup_rstat_flush(struct cgroup *cgrp, bool may_sleep)
> {
> if (should_skip_flush())
> return;
>
> - might_sleep();
> - spin_lock(&cgroup_rstat_lock);
> - cgroup_rstat_flush_locked(cgrp, true);
> - spin_unlock(&cgroup_rstat_lock);
> -}
> -
> -/**
> - * cgroup_rstat_flush_irqsafe - irqsafe version of cgroup_rstat_flush()
> - * @cgrp: target cgroup
> - *
> - * This function can be called from any context.
> - */
> -void cgroup_rstat_flush_irqsafe(struct cgroup *cgrp)
> -{
> - if (should_skip_flush())
> - return;
> + if (may_sleep)
> + might_sleep();
might_sleep_if()
next prev parent reply other threads:[~2023-03-23 15:43 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-23 4:00 [RFC PATCH 0/7] Make rstat flushing IRQ and sleep friendly Yosry Ahmed
2023-03-23 4:00 ` [RFC PATCH 1/7] cgroup: rstat: only disable interrupts for the percpu lock Yosry Ahmed
2023-03-23 4:29 ` Shakeel Butt
2023-03-23 5:15 ` Yosry Ahmed
2023-03-23 6:33 ` Shakeel Butt
2023-03-23 13:35 ` Yosry Ahmed
2023-03-23 15:40 ` Shakeel Butt
2023-03-23 15:42 ` Yosry Ahmed
2023-03-23 15:46 ` Shakeel Butt
2023-03-23 16:09 ` Shakeel Butt
2023-03-23 16:17 ` Yosry Ahmed
2023-03-23 16:29 ` Shakeel Butt
2023-03-23 16:36 ` Yosry Ahmed
2023-03-23 16:45 ` Shakeel Butt
2023-03-23 16:51 ` Yosry Ahmed
2023-03-23 19:09 ` Shakeel Butt
2023-03-23 17:33 ` Johannes Weiner
2023-03-23 18:09 ` Yosry Ahmed
2023-03-23 18:19 ` Johannes Weiner
2023-03-24 1:39 ` Tejun Heo
2023-03-24 7:22 ` Yosry Ahmed
2023-03-24 14:12 ` Waiman Long
2023-03-24 22:50 ` Yosry Ahmed
2023-03-25 1:54 ` Tejun Heo
2023-03-25 2:17 ` Yosry Ahmed
2023-03-25 4:30 ` Shakeel Butt
2023-03-25 4:37 ` Yosry Ahmed
2023-03-25 4:46 ` Shakeel Butt
2023-03-27 23:23 ` Yosry Ahmed
2023-03-29 18:53 ` Tejun Heo
2023-03-29 19:22 ` Hugh Dickins
2023-03-29 20:00 ` Tejun Heo
2023-03-29 20:38 ` Hugh Dickins
2023-03-30 4:26 ` Yosry Ahmed
2023-03-31 1:51 ` Tejun Heo
2023-03-23 4:00 ` [RFC PATCH 2/7] memcg: do not disable interrupts when holding stats_flush_lock Yosry Ahmed
2023-03-23 4:32 ` Shakeel Butt
2023-03-23 5:16 ` Yosry Ahmed
2023-03-23 4:00 ` [RFC PATCH 3/7] cgroup: rstat: remove cgroup_rstat_flush_irqsafe() Yosry Ahmed
2023-03-23 15:43 ` Johannes Weiner [this message]
2023-03-23 15:45 ` Yosry Ahmed
2023-03-23 4:00 ` [RFC PATCH 4/7] memcg: sleep during flushing stats in safe contexts Yosry Ahmed
2023-03-23 15:56 ` Johannes Weiner
2023-03-23 16:01 ` Yosry Ahmed
2023-03-23 17:27 ` Johannes Weiner
2023-03-23 18:07 ` Yosry Ahmed
2023-03-23 19:35 ` Shakeel Butt
2023-03-23 4:00 ` [RFC PATCH 5/7] vmscan: memcg: sleep when flushing stats during reclaim Yosry Ahmed
2023-03-23 4:00 ` [RFC PATCH 6/7] workingset: memcg: sleep when flushing stats in workingset_refault() Yosry Ahmed
2023-03-23 15:50 ` Johannes Weiner
2023-03-23 16:02 ` Yosry Ahmed
2023-03-23 16:00 ` Johannes Weiner
2023-03-23 16:02 ` Yosry Ahmed
2023-03-23 4:00 ` [RFC PATCH 7/7] memcg: do not modify rstat tree for zero updates Yosry Ahmed
2023-03-23 4:10 ` [RFC PATCH 0/7] Make rstat flushing IRQ and sleep friendly Shakeel Butt
2023-03-23 5:07 ` Yosry Ahmed
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=20230323154304.GA739026@cmpxchg.org \
--to=hannes@cmpxchg.org \
--cc=akpm@linux-foundation.org \
--cc=axboe@kernel.dk \
--cc=bpf@vger.kernel.org \
--cc=cgroups@vger.kernel.org \
--cc=josef@toxicpanda.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=lizefan.x@bytedance.com \
--cc=mhocko@kernel.org \
--cc=muchun.song@linux.dev \
--cc=roman.gushchin@linux.dev \
--cc=shakeelb@google.com \
--cc=tj@kernel.org \
--cc=vasily.averin@linux.dev \
--cc=yosryahmed@google.com \
/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).