From: "Matias Bjørling" <mb@lightnvm.io>
To: Igor Konopko <igor.j.konopko@intel.com>,
javier@javigon.com, hans.holmberg@cnexlabs.com
Cc: linux-block@vger.kernel.org
Subject: Re: [PATCH 07/18] lightnvm: pblk: wait for inflight IOs in recovery
Date: Sun, 17 Mar 2019 12:33:39 -0700 [thread overview]
Message-ID: <00eb866a-d2aa-437b-e580-3b0649e657ce@lightnvm.io> (raw)
In-Reply-To: <20190314160428.3559-8-igor.j.konopko@intel.com>
On 3/14/19 9:04 AM, Igor Konopko wrote:
> This patch changes the behaviour of recovery padding in order to
> support a case, when some IOs were already submitted to the drive and
> some next one are not submitted due to error returned.
>
> Currently in case of errors we simply exit the pad function without
> waiting for inflight IOs, which leads to panic on inflight IOs
> completion.
>
> After the changes we always wait for all the inflight IOs before
> exiting the function.
>
> Also, since NVMe has an internal timeout per IO, there is no need to
> introduce additonal one here.
>
> Signed-off-by: Igor Konopko <igor.j.konopko@intel.com>
> ---
> drivers/lightnvm/pblk-recovery.c | 32 +++++++++++++-------------------
> 1 file changed, 13 insertions(+), 19 deletions(-)
>
> diff --git a/drivers/lightnvm/pblk-recovery.c b/drivers/lightnvm/pblk-recovery.c
> index ba1691d..73d5ead 100644
> --- a/drivers/lightnvm/pblk-recovery.c
> +++ b/drivers/lightnvm/pblk-recovery.c
> @@ -200,7 +200,7 @@ static int pblk_recov_pad_line(struct pblk *pblk, struct pblk_line *line,
> rq_ppas = pblk_calc_secs(pblk, left_ppas, 0, false);
> if (rq_ppas < pblk->min_write_pgs) {
> pblk_err(pblk, "corrupted pad line %d\n", line->id);
> - goto fail_free_pad;
> + goto fail_complete;
> }
>
> rq_len = rq_ppas * geo->csecs;
> @@ -209,7 +209,7 @@ static int pblk_recov_pad_line(struct pblk *pblk, struct pblk_line *line,
> PBLK_VMALLOC_META, GFP_KERNEL);
> if (IS_ERR(bio)) {
> ret = PTR_ERR(bio);
> - goto fail_free_pad;
> + goto fail_complete;
> }
>
> bio->bi_iter.bi_sector = 0; /* internal bio */
> @@ -218,8 +218,11 @@ static int pblk_recov_pad_line(struct pblk *pblk, struct pblk_line *line,
> rqd = pblk_alloc_rqd(pblk, PBLK_WRITE_INT);
>
> ret = pblk_alloc_rqd_meta(pblk, rqd);
> - if (ret)
> - goto fail_free_rqd;
> + if (ret) {
> + pblk_free_rqd(pblk, rqd, PBLK_WRITE_INT);
> + bio_put(bio);
> + goto fail_complete;
> + }
>
> rqd->bio = bio;
> rqd->opcode = NVM_OP_PWRITE;
> @@ -266,7 +269,10 @@ static int pblk_recov_pad_line(struct pblk *pblk, struct pblk_line *line,
> if (ret) {
> pblk_err(pblk, "I/O submission failed: %d\n", ret);
> pblk_up_chunk(pblk, rqd->ppa_list[0]);
> - goto fail_free_rqd;
> + kref_put(&pad_rq->ref, pblk_recov_complete);
> + pblk_free_rqd(pblk, rqd, PBLK_WRITE_INT);
> + bio_put(bio);
> + goto fail_complete;
> }
>
> left_line_ppas -= rq_ppas;
> @@ -274,13 +280,9 @@ static int pblk_recov_pad_line(struct pblk *pblk, struct pblk_line *line,
> if (left_ppas && left_line_ppas)
> goto next_pad_rq;
>
> +fail_complete:
> kref_put(&pad_rq->ref, pblk_recov_complete);
> -
> - if (!wait_for_completion_io_timeout(&pad_rq->wait,
> - msecs_to_jiffies(PBLK_COMMAND_TIMEOUT_MS))) {
> - pblk_err(pblk, "pad write timed out\n");
> - ret = -ETIME;
> - }
> + wait_for_completion(&pad_rq->wait);
>
> if (!pblk_line_is_full(line))
> pblk_err(pblk, "corrupted padded line: %d\n", line->id);
> @@ -289,14 +291,6 @@ static int pblk_recov_pad_line(struct pblk *pblk, struct pblk_line *line,
> free_rq:
> kfree(pad_rq);
> return ret;
> -
> -fail_free_rqd:
> - pblk_free_rqd(pblk, rqd, PBLK_WRITE_INT);
> - bio_put(bio);
> -fail_free_pad:
> - kfree(pad_rq);
> - vfree(data);
> - return ret;
> }
>
> static int pblk_pad_distance(struct pblk *pblk, struct pblk_line *line)
>
Hi Igor,
Can you split this patch in two. One that removes the
wait_for_completion_io_timeout (and constant), and another that makes
sure it waits until all inflight IOs are completed?
next prev parent reply other threads:[~2019-03-17 19:33 UTC|newest]
Thread overview: 69+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-03-14 16:04 [PATCH 00/18] lightnvm: next set of improvements for 5.2 Igor Konopko
2019-03-14 16:04 ` [PATCH 01/18] lightnvm: pblk: fix warning in pblk_l2p_init() Igor Konopko
2019-03-16 22:29 ` Javier González
2019-03-18 16:25 ` Matias Bjørling
2019-03-14 16:04 ` [PATCH 02/18] lightnvm: pblk: warn when there are opened chunks Igor Konopko
2019-03-16 22:36 ` Javier González
2019-03-17 19:39 ` Matias Bjørling
2019-03-14 16:04 ` [PATCH 03/18] lightnvm: pblk: simplify partial read path Igor Konopko
2019-03-14 21:35 ` Heiner Litz
2019-03-15 9:52 ` Igor Konopko
2019-03-16 22:28 ` Javier González
2019-03-18 12:44 ` Igor Konopko
2019-03-14 16:04 ` [PATCH 04/18] lightnvm: pblk: OOB recovery for closed chunks fix Igor Konopko
2019-03-16 22:43 ` Javier González
2019-03-17 19:24 ` Matias Bjørling
2019-03-18 12:50 ` Igor Konopko
2019-03-18 19:25 ` Javier González
2019-03-14 16:04 ` [PATCH 05/18] lightnvm: pblk: propagate errors when reading meta Igor Konopko
2019-03-16 22:48 ` Javier González
2019-03-18 11:54 ` Hans Holmberg
2019-03-14 16:04 ` [PATCH 06/18] lightnvm: pblk: recover only written metadata Igor Konopko
2019-03-16 23:46 ` Javier González
2019-03-18 12:54 ` Igor Konopko
2019-03-18 15:04 ` Igor Konopko
2019-03-14 16:04 ` [PATCH 07/18] lightnvm: pblk: wait for inflight IOs in recovery Igor Konopko
2019-03-17 19:33 ` Matias Bjørling [this message]
2019-03-18 12:58 ` Igor Konopko
2019-03-14 16:04 ` [PATCH 08/18] lightnvm: pblk: fix spin_unlock order Igor Konopko
2019-03-16 23:49 ` Javier González
2019-03-18 11:55 ` Hans Holmberg
2019-03-14 16:04 ` [PATCH 09/18] lightnvm: pblk: kick writer on write recovery path Igor Konopko
2019-03-16 23:54 ` Javier González
2019-03-18 11:58 ` Hans Holmberg
2019-03-14 16:04 ` [PATCH 10/18] lightnvm: pblk: ensure that emeta is written Igor Konopko
2019-03-17 19:44 ` Matias Bjørling
2019-03-18 13:02 ` Igor Konopko
2019-03-18 18:26 ` Javier González
2019-03-21 13:34 ` Igor Konopko
2019-03-18 7:46 ` Javier González
2019-03-14 16:04 ` [PATCH 11/18] lightnvm: pblk: fix update line wp in OOB recovery Igor Konopko
2019-03-18 6:56 ` Javier González
2019-03-18 13:06 ` Igor Konopko
2019-03-14 16:04 ` [PATCH 12/18] lightnvm: pblk: do not read OOB from emeta region Igor Konopko
2019-03-17 19:56 ` Matias Bjørling
2019-03-18 13:05 ` Igor Konopko
2019-03-14 16:04 ` [PATCH 13/18] lightnvm: pblk: store multiple copies of smeta Igor Konopko
2019-03-18 7:33 ` Javier González
2019-03-18 13:12 ` Igor Konopko
2019-03-14 16:04 ` [PATCH 14/18] lightnvm: pblk: GC error handling Igor Konopko
2019-03-18 7:39 ` Javier González
2019-03-18 12:14 ` Hans Holmberg
2019-03-18 13:22 ` Igor Konopko
2019-03-18 14:14 ` Hans Holmberg
2019-03-14 16:04 ` [PATCH 15/18] lightnvm: pblk: fix in case of lack of lines Igor Konopko
2019-03-18 7:42 ` Javier González
2019-03-18 13:28 ` Igor Konopko
2019-03-18 19:21 ` Javier González
2019-03-21 13:21 ` Igor Konopko
2019-03-22 12:17 ` Hans Holmberg
2019-03-14 16:04 ` [PATCH 16/18] lightnvm: pblk: use nvm_rq_to_ppa_list() Igor Konopko
2019-03-18 7:48 ` Javier González
2019-03-14 16:04 ` [PATCH 17/18] lightnvm: allow to use full device path Igor Konopko
2019-03-18 7:49 ` Javier González
2019-03-18 10:28 ` Hans Holmberg
2019-03-18 13:18 ` Igor Konopko
2019-03-18 14:41 ` Hans Holmberg
2019-03-21 13:18 ` Igor Konopko
2019-03-25 11:40 ` Matias Bjørling
2019-03-14 16:04 ` [PATCH 18/18] lightnvm: track inflight target creations Igor Konopko
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=00eb866a-d2aa-437b-e580-3b0649e657ce@lightnvm.io \
--to=mb@lightnvm.io \
--cc=hans.holmberg@cnexlabs.com \
--cc=igor.j.konopko@intel.com \
--cc=javier@javigon.com \
--cc=linux-block@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).