From: Boris Brezillon <boris.brezillon@free-electrons.com>
To: "Rafał Miłecki" <zajec5@gmail.com>
Cc: "Rafał Miłecki" <rafal@milecki.pl>,
"Richard Weinberger" <richard@nod.at>,
linux-mtd@lists.infradead.org,
"Adrian Hunter" <adrian.hunter@intel.com>,
"Artem Bityutskiy" <dedekind1@gmail.com>
Subject: Re: [PATCH 1/2] ubifs: drop softlimit and delta fields from struct ubifs_wbuf
Date: Wed, 14 Sep 2016 17:28:04 +0200 [thread overview]
Message-ID: <20160914172804.00299d52@bbrezillon> (raw)
In-Reply-To: <20160914172428.3893ff2f@bbrezillon>
On Wed, 14 Sep 2016 17:24:28 +0200
Boris Brezillon <boris.brezillon@free-electrons.com> wrote:
> On Wed, 14 Sep 2016 12:24:34 +0200
> Rafał Miłecki <zajec5@gmail.com> wrote:
>
> > From: Rafał Miłecki <rafal@milecki.pl>
> >
> > Values of these fields are set during init and never modified. They are
> > used (read) in a single function only. There isn't really any reason to
> > keep them in a struct. It only makes struct just a bit bigger without
> > any clear gain.
> >
> > Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
> > ---
> > fs/ubifs/io.c | 18 ++++++++++--------
> > fs/ubifs/ubifs.h | 5 -----
> > 2 files changed, 10 insertions(+), 13 deletions(-)
> >
> > diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
> > index 97be412..4d6ce4a 100644
> > --- a/fs/ubifs/io.c
> > +++ b/fs/ubifs/io.c
> > @@ -452,16 +452,22 @@ static enum hrtimer_restart wbuf_timer_callback_nolock(struct hrtimer *timer)
> > */
> > static void new_wbuf_timer_nolock(struct ubifs_wbuf *wbuf)
> > {
> > + ktime_t softlimit = ktime_set(WBUF_TIMEOUT_SOFTLIMIT, 0);
> > + unsigned long long delta;
> > +
> > + delta = WBUF_TIMEOUT_HARDLIMIT - WBUF_TIMEOUT_SOFTLIMIT;
> > + delta *= 1000000000ULL;
> > +
> > ubifs_assert(!hrtimer_active(&wbuf->timer));
> > + ubifs_assert(delta <= ULONG_MAX);
>
> Hm, do we really want to recalculate delta each time we reschedule the
> wbuf timer?
>
> How about making the softlimit and delta variables static and
> initializing them once?
Nevermind. I just read patch 2, and these delta and softlimit values
have to be recalculated to account for dirty_writeback_interval changes.
>
> >
> > if (wbuf->no_timer)
> > return;
> > dbg_io("set timer for jhead %s, %llu-%llu millisecs",
> > dbg_jhead(wbuf->jhead),
> > - div_u64(ktime_to_ns(wbuf->softlimit), USEC_PER_SEC),
> > - div_u64(ktime_to_ns(wbuf->softlimit) + wbuf->delta,
> > - USEC_PER_SEC));
> > - hrtimer_start_range_ns(&wbuf->timer, wbuf->softlimit, wbuf->delta,
> > + div_u64(ktime_to_ns(softlimit), USEC_PER_SEC),
> > + div_u64(ktime_to_ns(softlimit) + delta, USEC_PER_SEC));
> > + hrtimer_start_range_ns(&wbuf->timer, softlimit, delta,
> > HRTIMER_MODE_REL);
> > }
> >
> > @@ -1059,10 +1065,6 @@ int ubifs_wbuf_init(struct ubifs_info *c, struct ubifs_wbuf *wbuf)
> >
> > hrtimer_init(&wbuf->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
> > wbuf->timer.function = wbuf_timer_callback_nolock;
> > - wbuf->softlimit = ktime_set(WBUF_TIMEOUT_SOFTLIMIT, 0);
> > - wbuf->delta = WBUF_TIMEOUT_HARDLIMIT - WBUF_TIMEOUT_SOFTLIMIT;
> > - wbuf->delta *= 1000000000ULL;
> > - ubifs_assert(wbuf->delta <= ULONG_MAX);
> > return 0;
> > }
> >
> > diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
> > index 4617d45..11bc8fa 100644
> > --- a/fs/ubifs/ubifs.h
> > +++ b/fs/ubifs/ubifs.h
> > @@ -644,9 +644,6 @@ typedef int (*ubifs_lpt_scan_callback)(struct ubifs_info *c,
> > * @io_mutex: serializes write-buffer I/O
> > * @lock: serializes @buf, @lnum, @offs, @avail, @used, @next_ino and @inodes
> > * fields
> > - * @softlimit: soft write-buffer timeout interval
> > - * @delta: hard and soft timeouts delta (the timer expire interval is @softlimit
> > - * and @softlimit + @delta)
> > * @timer: write-buffer timer
> > * @no_timer: non-zero if this write-buffer does not have a timer
> > * @need_sync: non-zero if the timer expired and the wbuf needs sync'ing
> > @@ -675,8 +672,6 @@ struct ubifs_wbuf {
> > int (*sync_callback)(struct ubifs_info *c, int lnum, int free, int pad);
> > struct mutex io_mutex;
> > spinlock_t lock;
> > - ktime_t softlimit;
> > - unsigned long long delta;
> > struct hrtimer timer;
> > unsigned int no_timer:1;
> > unsigned int need_sync:1;
>
>
> ______________________________________________________
> Linux MTD discussion mailing list
> http://lists.infradead.org/mailman/listinfo/linux-mtd/
next prev parent reply other threads:[~2016-09-14 15:28 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-09-14 10:24 [PATCH 1/2] ubifs: drop softlimit and delta fields from struct ubifs_wbuf Rafał Miłecki
2016-09-14 10:24 ` [PATCH 2/2] ubifs: use dirty_writeback_interval value for wbuf timer Rafał Miłecki
2016-09-14 15:48 ` Boris Brezillon
2016-09-14 15:24 ` [PATCH 1/2] ubifs: drop softlimit and delta fields from struct ubifs_wbuf Boris Brezillon
2016-09-14 15:28 ` Boris Brezillon [this message]
2016-09-14 15:40 ` Boris Brezillon
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=20160914172804.00299d52@bbrezillon \
--to=boris.brezillon@free-electrons.com \
--cc=adrian.hunter@intel.com \
--cc=dedekind1@gmail.com \
--cc=linux-mtd@lists.infradead.org \
--cc=rafal@milecki.pl \
--cc=richard@nod.at \
--cc=zajec5@gmail.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