netdev - Re: [PATCH] usbnet: removed USB

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]

Date:	Tue, 20 Oct 2015 12:24:30 +0200
From:	Oliver Neukum <oneukum@...e.com>
To:	Bjørn Mork <bjorn@...k.no>
Cc:	davem@...emloft.net, netdev@...r.kernel.org
Subject: Re: [PATCH] usbnet: removed USB_USBNET as a config option

On Tue, 2015-10-20 at 10:43 +0200, Bjørn Mork wrote:
> Oliver Neukum <oneukum@...e.com> writes:
> > On Mon, 2015-10-19 at 21:16 +0200, Bjørn Mork wrote:
> >
> >> This looks incomplete.
> >
> > Yes, it is against the earlier patch set which introduces the CDC
> > parser in cdc-acm and cdc-wdm. In hindsight I should have reversed
> > the order of patches.
> 
> Ah, right, I see.
> 
> But that patch has it's own set of problems, as pointed out by the test
> robot:
> 
> All the USBNET based minidrivers (as well as the USB_USBNET
> infrastructure itself) are protected by
> 
> menuconfig USB_NET_DRIVERS
>         tristate "USB Network Adapters"
>         default USB if USB
>         depends on USB && NET
> 
> if USB_NET_DRIVERS
> ..
> 
> 
> but USB_ACM and USB_WDM don't have similar protection.  So you'll have
> to make them select NET and USB_NET_DRIVERS too.

I really would like to avoid that.

> Selecting NET is going to be messy.  It already selects 3 other symbols,
> which you'll have to add as well.  And I have a feeling that the list of
> NET dependencies might grow, which probably explains why there currently
> are no "select NET"'s in the kernel. So "select NET" is most likely out
> of the question.  And I don't think we want to make USB_ACM or USB_WDM
> depend on NET either.  It will not make sense to anyone attempting to
> select these drivers, resulting in infinite confusion.

Well, ACM is a tty driver which in theory makes sense without net,
so I think it would be wrong to demand networking support.

> So we might need to move cdc_parse_cdc_header() out of usbnet.c after
> all. Keeping it there creates an unnecessary "select" mess.  But I don't
> know where to put it...  Either a common CDC helper module or the USB
> core I guess.

A common module means that the overhead of a module would be incurred.
Do you have any other candidates that could go into a common helper
module?

> Sorry for not bringing this up earlier. I know you asked specifically
> about this. I obviously didn't think it thoroughly through before.

Neither did I.

	Regards
		Oliver


--
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