linux-kernel - Re: [PATCH 1/1] UIO: Add a write() function to enable/disable interrupts

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: <7b9198260805241600m7a1c8dcva6cb6b41a4d2bacb@mail.gmail.com>
Date:	Sun, 25 May 2008 00:00:55 +0100
From:	"Tom Spink" <tspink@...il.com>
To:	"Thomas Gleixner" <tglx@...utronix.de>
Cc:	"Greg KH" <gregkh@...e.de>, "Hans J. Koch" <hjk@...utronix.de>,
	"Uwe Kleine-K??nig" <Uwe.Kleine-Koenig@...i.com>,
	linux-kernel@...r.kernel.org,
	"Jan Altenberg" <jan.altenberg@...utronix.de>,
	"Magnus Damm" <magnus.damm@...il.com>
Subject: Re: [PATCH 1/1] UIO: Add a write() function to enable/disable interrupts

2008/5/24 Thomas Gleixner <tglx@...utronix.de>:
> On Sat, 24 May 2008, Tom Spink wrote:
>> 2008/5/24 Thomas Gleixner <tglx@...utronix.de>:
>> > It makes a certain amount of sense to use write. You hold the device
>> > file descriptor anyway for the read (wait for interrupt) operation,
>> > so using the same file descriptor is not a too bad idea:
>>
>> What do you think about my ioctl idea, earlier in the thread?
>
> I think it's a pretty bad idea.

<grin>

>
>> >    while (!stop) {
>> >
>> >        /* wait for interrupt */
>> >        read(fd);
>> >
>> >        do_stuff();
>> >
>> >        /*reenable interrupt */
>> >        write(fd);
>> >    }
>>
>> So, instead of write, you'd use ioctl(fd, ...).
>
> And what's the actual gain ?

Simpler implementation, simpler use and future-proofing (in the sense
that ->write is no longer tied to this operation)

>
>> > I thought about using a sysfs entry for a while, but looking at the
>> > actual use case made the write() solution a more natural choice.
>>
>> I thought ioctl would be more natural, as [en,dis]abling interrupts is
>> a "controlling" operation :-)
>
> Oh no. We are not going to open the bottomless pit of ioctls in
> UIO. Once we have an ioctl channel in place we have the same mess
> which we want to avoid in the first place.
>
> Also when a driver needs more than the obvious interrupt wait /
> control functions (which are pretty symetric btw.) aside of the
> mmapped access to the device then it does not belong into the category
> of an UIO driver.

Fair enough :-) symmetry is good.  This is pretty much the response I
got from Hans.

> Thanks,
>
>        tglx

-- 
Tom Spink
--
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/