From: Avri Altman <Avri.Altman@wdc.com>
To: Vincent Whitchurch <vincent.whitchurch@axis.com>,
Ulf Hansson <ulf.hansson@linaro.org>
Cc: "kernel@axis.com" <kernel@axis.com>,
"linux-block@vger.kernel.org" <linux-block@vger.kernel.org>,
"tytso@mit.edu" <tytso@mit.edu>,
"linux-mmc@vger.kernel.org" <linux-mmc@vger.kernel.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: RE: [PATCH] mmc: core: support zeroout using TRIM
Date: Fri, 29 Apr 2022 19:34:10 +0000 [thread overview]
Message-ID: <DM6PR04MB657540F6BB9D02A5BA8A2AE6FCFC9@DM6PR04MB6575.namprd04.prod.outlook.com> (raw)
In-Reply-To: <20220429152118.3617303-1-vincent.whitchurch@axis.com>
> If the device supports TRIM and indicates that it erases to zeros, we can use
> it to support hardware offloading of REQ_OP_WRITE_ZEROES.
>
> Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
I agree that zero out is indeed a privet case of trim, as indicated by ERASED_MEM_CONT.
Looks good to me.
Thanks,
Avri
> ---
>
> Notes:
> https://lore.kernel.org/lkml/20160303182146.GG9772@thunk.org/ seems
> to agree
> that BLKZEROOUT can use TRIM on eMMC.
>
> BLKDISCARD uses DISCARD when available so it can't be used to send
> TRIM.
>
> If TRIM should not be used for BLKZEROOUT for some reason I guess the
> only way
> is to use MMC_IOC_MULTI_CMD like in this commit in mmc-utils but
> that's a
> rather low-level interface:
>
> https://git.kernel.org/pub/scm/utils/mmc/mmc-
> utils.git/commit/?id=43282e80e174cc73b09b81a4d17cb3a7b4dc5cfc
>
> drivers/mmc/core/block.c | 26 ++++++++++++++++++++++----
> drivers/mmc/core/queue.c | 2 ++
> 2 files changed, 24 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c index
> 506dc900f5c7..0398b205a285 100644
> --- a/drivers/mmc/core/block.c
> +++ b/drivers/mmc/core/block.c
> @@ -126,6 +126,7 @@ struct mmc_blk_data {
> #define MMC_BLK_DISCARD BIT(2)
> #define MMC_BLK_SECDISCARD BIT(3)
> #define MMC_BLK_CQE_RECOVERY BIT(4)
> +#define MMC_BLK_TRIM BIT(5)
>
> /*
> * Only set in main mmc_blk_data associated @@ -1090,12 +1091,13
> @@ static void mmc_blk_issue_drv_op(struct mmc_queue *mq, struct
> request *req)
> blk_mq_end_request(req, ret ? BLK_STS_IOERR : BLK_STS_OK); }
>
> -static void mmc_blk_issue_discard_rq(struct mmc_queue *mq, struct
> request *req)
> +static void mmc_blk_issue_erase_rq(struct mmc_queue *mq, struct
> request *req,
> + int type, unsigned int erase_arg)
> {
> struct mmc_blk_data *md = mq->blkdata;
> struct mmc_card *card = md->queue.card;
> unsigned int from, nr;
> - int err = 0, type = MMC_BLK_DISCARD;
> + int err = 0;
> blk_status_t status = BLK_STS_OK;
>
> if (!mmc_can_erase(card)) {
> @@ -1111,13 +1113,13 @@ static void mmc_blk_issue_discard_rq(struct
> mmc_queue *mq, struct request *req)
> if (card->quirks & MMC_QUIRK_INAND_CMD38) {
> err = mmc_switch(card, EXT_CSD_CMD_SET_NORMAL,
> INAND_CMD38_ARG_EXT_CSD,
> - card->erase_arg == MMC_TRIM_ARG ?
> + erase_arg == MMC_TRIM_ARG ?
> INAND_CMD38_ARG_TRIM :
> INAND_CMD38_ARG_ERASE,
> card->ext_csd.generic_cmd6_time);
> }
> if (!err)
> - err = mmc_erase(card, from, nr, card->erase_arg);
> + err = mmc_erase(card, from, nr, erase_arg);
> } while (err == -EIO && !mmc_blk_reset(md, card->host, type));
> if (err)
> status = BLK_STS_IOERR;
> @@ -1127,6 +1129,19 @@ static void mmc_blk_issue_discard_rq(struct
> mmc_queue *mq, struct request *req)
> blk_mq_end_request(req, status); }
>
> +static void mmc_blk_issue_trim_rq(struct mmc_queue *mq, struct request
> +*req) {
> + mmc_blk_issue_erase_rq(mq, req, MMC_BLK_TRIM, MMC_TRIM_ARG);
> }
> +
> +static void mmc_blk_issue_discard_rq(struct mmc_queue *mq, struct
> +request *req) {
> + struct mmc_blk_data *md = mq->blkdata;
> + struct mmc_card *card = md->queue.card;
> +
> + mmc_blk_issue_erase_rq(mq, req, MMC_BLK_DISCARD,
> +card->erase_arg); }
> +
> static void mmc_blk_issue_secdiscard_rq(struct mmc_queue *mq,
> struct request *req) { @@ -2327,6 +2342,9 @@ enum
> mmc_issued mmc_blk_mq_issue_rq(struct mmc_queue *mq, struct request
> *req)
> case REQ_OP_SECURE_ERASE:
> mmc_blk_issue_secdiscard_rq(mq, req);
> break;
> + case REQ_OP_WRITE_ZEROES:
> + mmc_blk_issue_trim_rq(mq, req);
> + break;
> case REQ_OP_FLUSH:
> mmc_blk_issue_flush(mq, req);
> break;
> diff --git a/drivers/mmc/core/queue.c b/drivers/mmc/core/queue.c index
> c69b2d9df6f1..bbe2ea829ea7 100644
> --- a/drivers/mmc/core/queue.c
> +++ b/drivers/mmc/core/queue.c
> @@ -191,6 +191,8 @@ static void mmc_queue_setup_discard(struct
> request_queue *q,
> q->limits.discard_granularity = SECTOR_SIZE;
> if (mmc_can_secure_erase_trim(card))
> blk_queue_flag_set(QUEUE_FLAG_SECERASE, q);
> + if (mmc_can_trim(card) && card->erased_byte == 0)
> + blk_queue_max_write_zeroes_sectors(q, max_discard);
> }
>
> static unsigned short mmc_get_max_segments(struct mmc_host *host)
> --
> 2.34.1
next prev parent reply other threads:[~2022-04-29 19:34 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-04-29 15:21 [PATCH] mmc: core: support zeroout using TRIM Vincent Whitchurch
2022-04-29 19:34 ` Avri Altman [this message]
2022-05-10 11:00 ` Ulf Hansson
2022-10-19 14:54 ` Jon Hunter
2022-10-19 15:07 ` Vincent Whitchurch
2022-10-19 17:33 ` Jon Hunter
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=DM6PR04MB657540F6BB9D02A5BA8A2AE6FCFC9@DM6PR04MB6575.namprd04.prod.outlook.com \
--to=avri.altman@wdc.com \
--cc=kernel@axis.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mmc@vger.kernel.org \
--cc=tytso@mit.edu \
--cc=ulf.hansson@linaro.org \
--cc=vincent.whitchurch@axis.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).