From: Kanchan Joshi <joshi.k@samsung.com>
To: Bart Van Assche <bvanassche@acm.org>, Jens Axboe <axboe@kernel.dk>
Cc: linux-block@vger.kernel.org, linux-scsi@vger.kernel.org,
linux-fsdevel@vger.kernel.org,
"Martin K . Petersen" <martin.petersen@oracle.com>,
Christoph Hellwig <hch@lst.de>,
Niklas Cassel <Niklas.Cassel@wdc.com>,
Avri Altman <Avri.Altman@wdc.com>, Bean Huo <huobean@gmail.com>,
Daejun Park <daejun7.park@samsung.com>,
Damien Le Moal <dlemoal@kernel.org>
Subject: Re: [PATCH v2 01/15] block: Make bio_set_ioprio() modify fewer bio->bi_ioprio bits
Date: Thu, 12 Oct 2023 14:19:02 +0530 [thread overview]
Message-ID: <fdf765a0-54a0-a9e9-fffa-3e733c2535b0@samsung.com> (raw)
In-Reply-To: <bfb7e2be-79f8-4f5e-b87e-3045d9c937b4@acm.org>
On 10/11/2023 10:22 PM, Bart Van Assche wrote:
>>> @@ -2926,7 +2926,8 @@ static void bio_set_ioprio(struct bio *bio)
>>> {
>>> /* Nobody set ioprio so far? Initialize it based on task's
>>> nice value */
>>> if (IOPRIO_PRIO_CLASS(bio->bi_ioprio) == IOPRIO_CLASS_NONE)
>>> - bio->bi_ioprio = get_current_ioprio();
>>> + ioprio_set_class_and_level(&bio->bi_ioprio,
>>> + get_current_ioprio());
>>> blkcg_set_ioprio(bio);
>>> }
>>> diff --git a/include/linux/ioprio.h b/include/linux/ioprio.h
>>> index 7578d4f6a969..f2e768ab4b35 100644
>>> --- a/include/linux/ioprio.h
>>> +++ b/include/linux/ioprio.h
>>> @@ -71,4 +71,14 @@ static inline int ioprio_check_cap(int ioprio)
>>> }
>>> #endif /* CONFIG_BLOCK */
>>> +#define IOPRIO_CLASS_LEVEL_MASK ((IOPRIO_CLASS_MASK <<
>>> IOPRIO_CLASS_SHIFT) | \
>>> + (IOPRIO_LEVEL_MASK << 0))
>>> +
>>> +static inline void ioprio_set_class_and_level(u16 *prio, u16
>>> class_level)
>>> +{
>>> + WARN_ON_ONCE(class_level & ~IOPRIO_CLASS_LEVEL_MASK);
>>> + *prio &= ~IOPRIO_CLASS_LEVEL_MASK;
>>> + *prio |= class_level;
>>
>> Return of get_current_ioprio() will touch all 16 bits here. So
>> user-defined value can alter whatever was set in bio by F2FS (patch 4 in
>> this series). Is that not an issue?
>
> The above is incomprehensible to me. Anyway, I will try to answer.
>
> It is not clear to me why it is claimed that "get_current_ioprio() will
> touch all 16 bits here"? The return value of get_current_ioprio() is
> passed to ioprio_set_class_and_level() and that function clears the hint
> bits from the get_current_ioprio() return value.
Function does OR bio->bi_ioprio with whatever is the return of
get_current_ioprio(). So if lifetime bits were set in
get_current_ioprio(), you will end up setting that in bio->bi_ioprio too.
> ioprio_set_class_and_level() preserves the hint bits set by F2FS.
>
>> And what is the user interface you have in mind. Is it ioprio based, or
>> write-hint based or mix of both?
>
> Since the data lifetime is encoded in the hint bits, the hint bits need
> to be set by user space to set a data lifetime.
I asked because more than one way seems to emerge here. Parts of this
series (Patch 4) are taking inode->i_write_hint (and not ioprio value)
and putting that into bio.
I wonder what to expect if application get to send one lifetime with
fcntl (write-hints) and different one with ioprio. Is that not racy?
> In case you would help,
> the blktest test that I wrote to test this functionality is available
> below.
>
> Thanks,
>
> Bart.
>
>
> diff --git a/tests/scsi/097 b/tests/scsi/097
> new file mode 100755
> index 000000000000..01d280021653
> --- /dev/null
> +++ b/tests/scsi/097
> @@ -0,0 +1,62 @@
> +#!/bin/bash
> +# SPDX-License-Identifier: GPL-3.0+
> +# Copyright (C) 2022 Google LLC
> +
> +. tests/zbd/rc
> +. common/null_blk
> +. common/scsi_debug
> +
> +DESCRIPTION="test block data lifetime support"
> +QUICK=1
> +
> +requires() {
> + _have_fio
> + _have_module scsi_debug
> +}
> +
> +test() {
> + echo "Running ${TEST_NAME}"
> +
> + local scsi_debug_params=(
> + delay=0
> + dev_size_mb=1024
> + sector_size=4096
> + )
> + _init_scsi_debug "${scsi_debug_params[@]}" &&
> + local dev="/dev/${SCSI_DEBUG_DEVICES[0]}" fail &&
> + ls -ld "${dev}" >>"${FULL}" &&
> + local i fio_args=(
> + --group_reporting=1
> + --gtod_reduce=1
> + --ioscheduler=none
> + --norandommap
> + --direct=1
> + --rw=randwrite
> + --ioengine=io_uring
> + --time_based
> + ) &&
> + for ((i=1; i<=3; i++)); do
> + fio_args+=(
> + --name=whint"$i"
> + --filename="${dev}"
> + --prio=$((i<<6))
This will not work as prio can only take values between 0-7.
Perhaps you want to use "priohint" to send lifetime.
next prev parent reply other threads:[~2023-10-12 8:49 UTC|newest]
Thread overview: 45+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-05 19:40 [PATCH v2 00/15] Pass data temperature information to UFS devices Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 01/15] block: Make bio_set_ioprio() modify fewer bio->bi_ioprio bits Bart Van Assche
2023-10-06 6:28 ` Kanchan Joshi
2023-10-06 18:20 ` Bart Van Assche
2023-10-10 5:22 ` Kanchan Joshi
2023-10-11 16:52 ` Bart Van Assche
2023-10-12 8:49 ` Kanchan Joshi [this message]
2023-10-12 14:03 ` Niklas Cassel
2023-10-12 17:42 ` Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 02/15] blk-ioprio: Modify " Bart Van Assche
2023-10-06 6:36 ` Kanchan Joshi
2023-10-06 18:25 ` Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 03/15] block: Support data lifetime in the I/O priority bitfield Bart Van Assche
2023-10-06 6:42 ` Kanchan Joshi
2023-10-06 8:19 ` Damien Le Moal
2023-10-06 9:53 ` Niklas Cassel
2023-10-06 18:07 ` Bart Van Assche
2023-10-11 20:51 ` Bart Van Assche
2023-10-12 1:02 ` Damien Le Moal
2023-10-12 18:00 ` Bart Van Assche
2023-10-13 1:08 ` Damien Le Moal
2023-10-13 9:33 ` Niklas Cassel
2023-10-13 21:20 ` Bart Van Assche
2023-10-16 9:20 ` Niklas Cassel
2023-10-16 16:36 ` Bart Van Assche
2023-10-13 20:18 ` Bart Van Assche
2023-10-15 22:22 ` Damien Le Moal
2023-10-16 16:31 ` Bart Van Assche
2023-10-16 6:17 ` Christoph Hellwig
2023-10-16 16:32 ` Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 04/15] fs: Restore write hint support Bart Van Assche
2023-10-10 5:42 ` Kanchan Joshi
2023-10-11 16:56 ` Bart Van Assche
2023-10-16 6:20 ` Christoph Hellwig
2023-10-05 19:40 ` [PATCH v2 05/15] fs/f2fs: Restore the whint_mode mount option Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 06/15] scsi: core: Query the Block Limits Extension VPD page Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 07/15] scsi_proto: Add structures and constants related to I/O groups and streams Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 08/15] sd: Translate data lifetime information Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 09/15] scsi_debug: Reduce code duplication Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 10/15] scsi_debug: Support the block limits extension VPD page Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 11/15] scsi_debug: Rework page code error handling Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 12/15] scsi_debug: Rework subpage " Bart Van Assche
2023-10-05 19:40 ` [PATCH v2 13/15] scsi_debug: Implement the IO Advice Hints Grouping mode page Bart Van Assche
2023-10-05 19:41 ` [PATCH v2 14/15] scsi_debug: Implement GET STREAM STATUS Bart Van Assche
2023-10-05 19:41 ` [PATCH v2 15/15] scsi_debug: Maintain write statistics per group number Bart Van Assche
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=fdf765a0-54a0-a9e9-fffa-3e733c2535b0@samsung.com \
--to=joshi.k@samsung.com \
--cc=Avri.Altman@wdc.com \
--cc=Niklas.Cassel@wdc.com \
--cc=axboe@kernel.dk \
--cc=bvanassche@acm.org \
--cc=daejun7.park@samsung.com \
--cc=dlemoal@kernel.org \
--cc=hch@lst.de \
--cc=huobean@gmail.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-scsi@vger.kernel.org \
--cc=martin.petersen@oracle.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).