netdev - Re: nt: usb: USB_RTL8153_ECM should not default to y

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: <20210917114924.2a7bda93@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
Date:   Fri, 17 Sep 2021 11:49:24 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     Maciej Żenczykowski <zenczykowski@...il.com>
Cc:     Geert Uytterhoeven <geert+renesas@...der.be>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Linux NetDev <netdev@...r.kernel.org>,
        Hayes Wang <hayeswang@...ltek.com>,
        Marek Szyprowski <m.szyprowski@...sung.com>,
        Maciej Zenczykowski <maze@...gle.com>
Subject: Re: nt: usb: USB_RTL8153_ECM should not default to y

On Fri, 17 Sep 2021 19:59:15 +0200 Maciej Żenczykowski wrote:
> I've been browsing some usb ethernet dongle related stuff in the
> kernel (trying to figure out which options to enable in Android 13
> 5.~15 kernels), and I've come across the following patch (see topic,
> full patch quoted below).
> 
> Doesn't it entirely defeat the purpose of the patch it claims to fix
> (and the patch that fixed)?
> Certainly the reasoning provided (in general device drivers should not
> be enabled by default) doesn't jive with me.
> The device driver is CDC_ETHER and AFAICT this is just a compatibility
> option for it.
> 
> Shouldn't it be reverted (ie. the 'default y' line be re-added) ?
> 
> AFAICT the logic should be:
>   if we have CDC ETHER (aka. ECM), but we don't have R8152 then we
> need to have R8153_ECM.
> 
> Alternatively, maybe there shouldn't be a config option for this at all?
> 
> Instead r8153_ecm should simply be part of cdc_ether.ko iff r8152=n
> 
> I'm not knowledgeable enough about Kconfig syntax to know how to
> phrase the logic...
> Maybe there shouldn't be a Kconfig option at all, and just some Makefile if'ery.
> 
> Something like:
> 
> obj-$(CONFIG_USB_RTL8152) += r8152.o
> obj-$(CONFIG_USB_NET_CDCETHER) += cdc_ether.o obj-
> ifndef CONFIG_USB_RTL8152
> obj-$(CONFIG_USB_NET_CDCETHER) += r8153_ecm.o
> endif
> 
> Though it certainly would be nice to use 8153 devices with the
> CDCETHER driver even with the r8152 driver enabled...

Yeah.. more context here:

https://lore.kernel.org/all/7fd014f2-c9a5-e7ec-f1c6-b3e4bb0f6eb6@samsung.com/

default !USB_RTL8152 would be my favorite but that probably doesn't
compute in kconfig land. Or perhaps bring back the 'y' but more clearly
mark it as a sub-option of CDCETHER? It's hard to blame people for
expecting drivers to default to n, we should make it clearer that this
is more of a "make driver X support variation Y", 'cause now it sounds
like a completely standalone driver from the Kconfig wording. At least
to a lay person like myself.