linux-kernel - Re: [RFCv2 00/10] xhci: re-work command queue management

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: <CAPcyv4gerLT1uuUre9L4minu6u4HYRFSYD2s-bo_7H9UH64QYQ@mail.gmail.com>
Date:	Wed, 5 Feb 2014 08:48:25 -0800
From:	Dan Williams <dan.j.williams@...el.com>
To:	David Laight <David.Laight@...lab.com>
Cc:	Mathias Nyman <mathias.nyman@...ux.intel.com>,
	"linux-usb@...r.kernel.org" <linux-usb@...r.kernel.org>,
	"sarah.a.sharp@...ux.intel.com" <sarah.a.sharp@...ux.intel.com>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [RFCv2 00/10] xhci: re-work command queue management

On Wed, Feb 5, 2014 at 1:22 AM, David Laight <David.Laight@...lab.com> wrote:
> From: Dan Williams
>> > Adding another list that will have its own set of bugs seems retrograde top me.
>>
>> What bugs?  Please be specific.  The problem to be addressed is not
>> the allocation of commands, but that timeouts of one command eat the
>> timeout periods of subsequent commands.  I'm thinking a
>> single-threaded workqueue better models what the hardware is doing.
>> See my comments in patch 1, but that is orthogonal to how the command
>> contexts are allocated.
>
> No software is bug free.
> The best way to get reasonably bug-free software is the KISS principle
> (Keep It Simple Stupid).

I find this condescending.

Perhaps you did not mean for it to come out that way, but this is far
from a specific technical argument.

> You just seem to be adding a lot of additional complexity.
> Maybe it would look better if more of the code was factored out of the
> call sites.
>
> IIRC at the moment every call site has to:
> 1) find the trb address that will be used (massive layer break).
> 2) actually write the trb (through several layers of function call).
> 3) sort out any timeout (I didn't really look into this bit).
> 4) ring the bell.

I think there is room to push these pre-requisites down.

> ISTM that the call site should just be a single function call.
> If you do that the implementation of the timeouts/completions is
> removed from the callers.

Yes, but I think we need to centralize the context under which
commands are submitted.  The complicating factor is the mix of
synchronous command submission and interrupt driven asynchronous
command queuing.  I think we can simplify it by making it all
submitted from a single event queue context.  I'm investigating if
that is a workable solution...
--
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/