From: John Garry <john.g.garry@oracle.com>
To: Hannes Reinecke <hare@suse.de>,
Bart Van Assche <bvanassche@acm.org>,
"Martin K . Petersen" <martin.petersen@oracle.com>
Cc: linux-scsi@vger.kernel.org,
"James E.J. Bottomley" <James.Bottomley@HansenPartnership.com>
Subject: Re: [PATCH 04/24] scsi: core: Implement reserved command handling
Date: Wed, 16 Apr 2025 09:23:00 +0100 [thread overview]
Message-ID: <49a1e0c0-4e9e-40f4-8a74-ebb63ebccfe2@oracle.com> (raw)
In-Reply-To: <7455bff3-97f6-4156-87e0-197f2d3d8350@suse.de>
On 16/04/2025 08:39, Hannes Reinecke wrote:
>> JFYI, I was working on another solution (different to Hannes') which
>> allows reserved requests pass through scsi_queue_rq(), but I stopped
>> that work when I changed employer.
>>
> Oh, really?
> Do you have the patchset still around?
> I'd be very much interested in that...
Yeah, if you remember, I was working on that for libsas/libata, but
started to get bogged down in how to handle reserved commands for
libsas/libata:
https://lore.kernel.org/linux-scsi/1666693096-180008-1-git-send-email-john.garry@huawei.com/T/#m35ad9a27320077f55c2be5549f036628fe79dcf7
It would be a lot simpler to support for HBAs which plug directly into
the SCSI midlayer - like aacraid.
>
>>> Although it is easy
>>> to create an additional request queue for reserved requests, that
>>> request queue shares its tag set with the SCSI host and hence also the
>>> request submission function. From scsi_host.h:
>>>
>>> struct Scsi_Host {
>>> struct blk_mq_tag_set tag_set;
>>> [ ... ]
>>> };
>>>
>>> From blk-mq.h:
>>>
>>> struct blk_mq_tag_set {
>>> const struct blk_mq_ops *ops;
>>> [ ... ]
>>> };
>>>
>>> Unless someone comes up with an elegant proposal, I will keep the
>>> approach where ufshcd_tag_to_cmd() handles reserved tags and regular
>>> tags differently.
>>>
>>> It should be possible to do this without referencing tags->static_rqs[]
>>> directly from the UFS driver.
>>
>> I'm not sure how that will look, but my preference is to fully
>> implement reserved tags support which can be used by all SCSI LLDs.
>>
> Sure, that was my goal, too.
> But then I got stuck as some drivers (most notably aacraid) use internal
> commands to detect the hardware and allocate SCSI devices, so with my
> approach we need a 'fake' SCSI device for the host to handle that.
> If you have other ideas I'd be very interested in thems.
My solution still involved supporting a fake sdev also for times like
you describe, e.g. shost needs to send reserved commands.
Thanks,
John
next prev parent reply other threads:[~2025-04-16 8:23 UTC|newest]
Thread overview: 65+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-04-03 21:17 [PATCH 00/24] Optimize the hot path in the UFS driver Bart Van Assche
2025-04-03 21:17 ` [PATCH 01/24] scsi: core: Make scsi_cmd_to_rq() accept const arguments Bart Van Assche
2025-04-04 10:39 ` John Garry
2025-04-04 16:23 ` Bart Van Assche
2025-08-18 6:03 ` Hannes Reinecke
2025-08-18 17:17 ` Bart Van Assche
2025-04-03 21:17 ` [PATCH 02/24] scsi: core: Make scsi_cmd_priv() " Bart Van Assche
2025-04-03 21:17 ` [PATCH 03/24] scsi: core: Use scsi_cmd_priv() instead of open-coding it Bart Van Assche
2025-04-03 21:17 ` [PATCH 04/24] scsi: core: Implement reserved command handling Bart Van Assche
2025-04-04 10:49 ` John Garry
2025-04-04 16:34 ` Bart Van Assche
2025-04-04 17:34 ` John Garry
2025-04-14 21:59 ` Bart Van Assche
2025-04-15 17:21 ` Bart Van Assche
2025-04-16 7:16 ` John Garry
2025-04-16 7:39 ` Hannes Reinecke
2025-04-16 8:23 ` John Garry [this message]
2025-04-17 21:25 ` Bart Van Assche
2025-04-22 6:54 ` Hannes Reinecke
2025-08-05 22:33 ` Bart Van Assche
2025-08-07 15:40 ` Bart Van Assche
2025-04-16 7:33 ` Hannes Reinecke
2025-08-07 16:35 ` Bart Van Assche
2025-04-03 21:17 ` [PATCH 05/24] scsi: core: Introduce scsi_host_update_can_queue() Bart Van Assche
2025-04-03 21:17 ` [PATCH 06/24] scsi: ufs: core: Change the type of one ufshcd_add_cmd_upiu_trace() argument Bart Van Assche
2025-04-09 5:34 ` Avri Altman
2025-04-14 12:47 ` Peter Wang (王信友)
2025-04-03 21:17 ` [PATCH 07/24] scsi: ufs: core: Only call ufshcd_add_command_trace() for SCSI commands Bart Van Assche
2025-04-09 6:24 ` Avri Altman
2025-04-14 12:47 ` Peter Wang (王信友)
2025-04-03 21:17 ` [PATCH 08/24] scsi: ufs: core: Change the type of one ufshcd_add_command_trace() argument Bart Van Assche
2025-04-09 6:32 ` Avri Altman
2025-04-09 16:37 ` Bart Van Assche
2025-04-14 12:47 ` Peter Wang (王信友)
2025-04-03 21:17 ` [PATCH 09/24] scsi: ufs: core: Change the type of one ufshcd_send_command() argument Bart Van Assche
2025-04-09 6:34 ` Avri Altman
2025-04-14 12:48 ` Peter Wang (王信友)
2025-04-03 21:17 ` [PATCH 10/24] scsi: ufs: core: Only call ufshcd_should_inform_monitor() for SCSI commands Bart Van Assche
2025-04-15 7:34 ` Peter Wang (王信友)
2025-04-03 21:17 ` [PATCH 11/24] scsi: ufs: core: Change the monitor function argument types Bart Van Assche
2025-04-15 7:37 ` Peter Wang (王信友)
2025-04-15 20:10 ` Bart Van Assche
2025-04-03 21:17 ` [PATCH 12/24] scsi: ufs: core: Rework ufshcd_mcq_compl_pending_transfer() Bart Van Assche
2025-04-15 8:00 ` Peter Wang (王信友)
2025-04-15 20:22 ` Bart Van Assche
2025-04-17 12:39 ` Peter Wang (王信友)
2025-04-17 21:33 ` Bart Van Assche
2025-04-22 13:07 ` Peter Wang (王信友)
2025-04-03 21:17 ` [PATCH 13/24] scsi: ufs: core: Rework ufshcd_eh_device_reset_handler() Bart Van Assche
2025-04-03 21:17 ` [PATCH 14/24] scsi: ufs: core: Cache the DMA buffer sizes Bart Van Assche
2025-04-03 21:17 ` [PATCH 15/24] scsi: ufs: core: Add an argument to ufshcd_mcq_decide_queue_depth() Bart Van Assche
2025-04-03 21:18 ` [PATCH 16/24] scsi: ufs: core: Add an argument to ufshcd_alloc_mcq() Bart Van Assche
2025-04-03 21:18 ` [PATCH 17/24] scsi: ufs: core: Call ufshcd_mcq_init() once Bart Van Assche
2025-04-09 6:52 ` Avri Altman
2025-04-09 16:42 ` Bart Van Assche
2025-04-03 21:18 ` [PATCH 18/24] scsi: ufs: core: Allocate the SCSI host earlier Bart Van Assche
2025-04-03 21:18 ` [PATCH 19/24] scsi: ufs: core: Call ufshcd_init_lrb() later Bart Van Assche
2025-04-03 21:18 ` [PATCH 20/24] scsi: ufs: core: Use hba->reserved_slot Bart Van Assche
2025-04-03 21:18 ` [PATCH 21/24] scsi: ufs: core: Allocate the reserved slot as a reserved request Bart Van Assche
2025-04-03 21:18 ` [PATCH 22/24] scsi: ufs: core: Do not clear driver-private command data Bart Van Assche
2025-04-03 21:18 ` [PATCH 23/24] scsi: ufs: core: Optimize the hot path Bart Van Assche
2025-04-03 21:18 ` [PATCH 24/24] scsi: ufs: core: Remove the ufshcd_lrb task_tag member Bart Van Assche
2025-04-04 10:15 ` [PATCH 00/24] Optimize the hot path in the UFS driver John Garry
2025-04-04 17:05 ` Bart Van Assche
2025-04-05 12:03 ` Avri Altman
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=49a1e0c0-4e9e-40f4-8a74-ebb63ebccfe2@oracle.com \
--to=john.g.garry@oracle.com \
--cc=James.Bottomley@HansenPartnership.com \
--cc=bvanassche@acm.org \
--cc=hare@suse.de \
--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).