linux-kernel - Re: [PATCH 0/6] hisi_sas: Expose multiple hw queues for v3 hw as experimental

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <dfbee75d-a223-9ff5-db37-16c00a5667bd@huawei.com>
Date:   Fri, 25 Oct 2019 17:04:38 +0100
From:   John Garry <john.garry@...wei.com>
To:     <jejb@...ux.vnet.ibm.com>, <martin.petersen@...cle.com>
CC:     <linux-scsi@...r.kernel.org>, <linuxarm@...wei.com>,
        <linux-kernel@...r.kernel.org>, <hare@...e.com>,
        <ming.lei@...hat.com>
Subject: Re: [PATCH 0/6] hisi_sas: Expose multiple hw queues for v3 hw as
 experimental

On 24/10/2019 15:21, John Garry wrote:
> This series adds support to expose multiple hw queues for SCSI mid-layer
> as an experimental feature.
> 
> For now it is experimental due to known CPU hotplug issue for managed
> interrupts.
> 
> So now we have two module parameters to enable managed interrupts for v3
> hw driver:
> - auto_affine_msi_experimental:
> Use managed interrupts plus and manage reply map internally. Use
> request tag for IPTT (apart from reserved commands).
> - expose_mq_experimental
> Use managed interrupts plus and expose multipe hw queues. Manage IPTT
> internally with sbitmap.
> 
> Paramater auto_affine_msi_experimental shows better performance (than
> expose_mq_experimental), so we need to maintain it for now to stop
> complaints about performance regression (even though enabling this
> parameter is unsafe).
> 
> I want to remove these module parameters ASAP.
> 
> This series also includes a change to convert the driver to use sbitmap
> where possible for managing IPTT.

Hi Martin,

Can we hold off on this series until Ming has had a look?

Thanks,
John

> 
> This series is based on 5.4 + [0], even though being advertised for
> topic-sas-5.4 dev branch. Sorry for send before that is merged, but I just
> wanted to get these posted.
> 
> [0] https://lore.kernel.org/linux-scsi/1571674935-108326-1-git-send-email-john.garry@huawei.com/T/#t
> 
> 
> John Garry (6):
>    scsi: hisi_sas: Use sbitmap for IPTT management
>    scsi: hisi_sas: Pass scsi_cmnd pointer to hisi_sas_hw.slot_index_alloc
>    scsi: hisi_sas: Add bitmaps_alloc_v3_hw()
>    scsi: hisi_sas: Add slot_index_alloc_v3_hw() and
>      slot_index_free_v3_hw()
>    scsi: hisi_sas: Split interrupt_init_v3_hw()
>    scsi: hisi_sas: Expose multiple hw queues for v3 as experimental
> 
>   drivers/scsi/hisi_sas/hisi_sas.h       |  12 ++-
>   drivers/scsi/hisi_sas/hisi_sas_main.c  | 135 ++++++++++++-------------
>   drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 110 +++++++++++++-------
>   drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 100 ++++++++++++++++--
>   4 files changed, 242 insertions(+), 115 deletions(-)
>