linux-kernel - Re: [PATCH 6/7] mpt2sas: store scsi io tracker data in the scsi command / request

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] [day] [month] [year] [list]

Message-ID: <55242E7E.1080705@kernel.dk>
Date:	Tue, 07 Apr 2015 13:22:38 -0600
From:	Jens Axboe <axboe@...nel.dk>
To:	Christoph Hellwig <hch@....de>
CC:	Jens Axboe <axboe@...com>, linux-kernel@...r.kernel.org,
	linux-scsi@...r.kernel.org
Subject: Re: [PATCH 6/7] mpt2sas: store scsi io tracker data in the scsi command
 / request

On 04/07/2015 10:18 AM, Christoph Hellwig wrote:
> On Tue, Apr 07, 2015 at 10:13:23AM -0600, Jens Axboe wrote:
>>> The mq case will also work for the !mq case when you call
>>> scsi_host_find_tag and scsi_cmd_priv.   In general all the mq-specific
>>> codepathes you add should become the default and only one, even if this
>>> requires a lit bit of additional core work.
>>
>> For the core code, I definitely agree. But for this case, in scsi-mq mode,
>> we know that tag == smid - 1. That's not the case if we are not using
>> scsi-mq.
>
> It is if you use the old block tagging code with host-wide tags.
> I guess you'll need to tell mpt2 and mpt3 to use that firs, though.

Yup

>> In general, it'd be great if we could "convert" drivers and not have to
>> support both scsi-mq and legacy mode. Then I could just rip the old code.
>
> Well, the whole point is that you should be able to write a driver like
> your mq version and it should just work. Even better would be if we

It'd be nice to have a host template flag that says "I only run 
scsi-mq", so we didn't have to cater to both cases. Or maybe that 
already exists and I just didn't look hard enough.

> could get rid of the old case entirely for scsi, but for that we need
> a I/O scheduler for blk-mq first :)

I hear ya :)

-- 
Jens Axboe

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/