linux-kernel - Re: [PATCH] Firmware: broadcom sprom: clarifiy SSB dependency

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: <4928549.J7iBRWWaDH@wuerfel>
Date:	Mon, 14 Mar 2016 16:34:09 +0100
From:	Arnd Bergmann <arnd@...db.de>
To:	Rafał Miłecki <zajec5@...il.com>
Cc:	Ralf Baechle <ralf@...ux-mips.org>, Paul Walmsley <paul@...an.com>,
	"linux-mips@...ux-mips.org" <linux-mips@...ux-mips.org>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] Firmware: broadcom sprom: clarifiy SSB dependency

On Monday 14 March 2016 15:37:18 Rafał Miłecki wrote:
> On 14 March 2016 at 15:21, Arnd Bergmann <arnd@...db.de> wrote:
> > The broadcom firmware drvier calls into the ssb SPROM code if that
> > is enabled, but it fails if the SSB code is in a loadable module
> > because the bcm47xx firmware is always built-in:
> >
> > drivers/firmware/built-in.o: In function `bcm47xx_sprom_register_fallbacks':
> > bcm47xx_sprom.c:(.text+0x11c4): undefined reference to `ssb_arch_register_fallback_sprom'
> >
> > This adds a Kconfig dependency to ensure that we cannot turn on the
> > generic sprom support if the ssb sprom is in a module.
> 
> Can you attach your config that triggered this build error? I modified
> condition to the:
> #if IS_BUILTIN(CONFIG_SSB) && IS_ENABLED(CONFIG_SSB_SPROM)
> which I believe should be enough.

>From inspection, I think your solution is sufficient to avoid the error.
I found the bug while travelling and I'm only now catching up on the submissions,
so I must have missed the fact that it was fixed. I looked at the code
to see if additional patches had been applied on top, but I did not
realize that you had modified the driver in place.

> I'm afraid your patch won't allow compiling SPROM driver with BCMA=y
> and SSB as a module.

Correct. The downside of your approach is that it silently stops the machine
from accessing the SSB SPROM when the driver is a module, while still
allowing the generic SPROM code to be built, which may be harder to
figure out for a user than a missing driver.

I don't see a good solution that works either way, so the your latest code looks
good enough. Thanks for the quick reply!

	Arnd