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  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]
Date:	Fri, 1 Oct 2010 09:14:22 +1000
From:	Julian Calaby <julian.calaby@...il.com>
To:	Johannes Berg <johannes@...solutions.net>
Cc:	netdev@...r.kernel.org, linux-wireless@...r.kernel.org,
	tgraf@...g.ch
Subject: Re: [RFC 1/2] genetlink: introduce pre_doit/post_doit hooks

On Fri, Oct 1, 2010 at 08:55, Johannes Berg <johannes@...solutions.net> wrote:
> On Fri, 2010-10-01 at 08:51 +1000, Julian Calaby wrote:
>
>> > Come to think of it -- I could get away with a single pointer, since, if
>> > both are assigned,
>> >
>> > user_ptr[0] == wiphy_to_rdev(((netdev *)user_ptr[1])->ieee80211_ptr->wiphy)
>> >
>> > but that's a lot of pointy things, and some functions only have the rdev
>> > so it gets more complex. I think allowing two private pointers is a
>> > decent compromise.
>>
>> Come to think of it -- if someone wanted to have a massive structure
>> with 10 pointers and a set of random data structures, then they could
>> easily create their priv struct and assign it to user_ptr[0], hence
>> rendering my argument null and void.
>
> Oh, well, I thought your argument was that it was arbitrary and not
> really necessary :-)

My argument was more that someone's likely to come up with a scheme
that requires more than 2 pointers, so why not accommodate them from
start with an element for a priv struct - but that requires a new
struct, allocating and freeing it, as well as heap space for it, and a
pointer on the stack to it. (though I'm sure there'll be some sensible
way to make them persistent, but that's another issue)

> Also, this rather cheap, it just needs a bit more stack space in a place
> that isn't typically deeply nested. So if some protocol came around and
> needed three pointers, I'd probably advocate just bumping it to three.
> At some point I might draw a line (10 is probably too much).

I was considering pointing out that a compromise might need to be
made, but I figured you'd already thought of that =)

> But you're right, of course, they can just use the first one and put
> something dynamically allocated into that, if really needed.

Exactly.

Thanks,

-- 

Julian Calaby

Email: julian.calaby@...il.com
Profile: http://www.google.com/profiles/julian.calaby/
.Plan: http://sites.google.com/site/juliancalaby/
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Powered by blists - more mailing lists