linux-kernel - Re: [PATCH V2 5/6] x86: add CONFIG_ARM

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: <201207031319.40884.arnd@arndb.de>
Date:	Tue, 3 Jul 2012 13:19:40 +0000
From:	Arnd Bergmann <arnd@...db.de>
To:	Alessandro Rubini <rubini@...dd.com>
Cc:	hpa@...or.com, linux-kernel@...r.kernel.org,
	giancarlo.asnaghi@...com, alan@...ux.intel.com,
	linux@....linux.org.uk, x86@...nel.org, gregkh@...uxfoundation.org,
	linux-arm-kernel@...ts.infradead.org, linux-serial@...r.kernel.org,
	linux-arch@...r.kernel.org
Subject: Re: [PATCH V2 5/6] x86: add CONFIG_ARM_AMBA, selected by STA2X11

On Tuesday 03 July 2012, Alessandro Rubini wrote:
> Peter Anvin:
> > There is no problem with adding ARM or !X86 dependencies to drivers
> > now and fixing them later or as required.
> 
> Ok. This is my summary of the compilation errors I have when
> enabling ARM_AMBA undex x86 and enabling everything that appears
> in "make oldconfig":
> 
> 
> This is a sum up of the errors in all driver that are
> enabled by telling "CONFIG_ARM_AMBA=y" in the x86 config.
> 
>    drivers/dma/pl330.c: register names conflict with arch symbols
>      proposed fix: use proper prefix names
> 
>    drivers/dma/amba-pl08x.c: needs <asm/hardware/pl080.h>
>      proposed fix: move pl080.h to include/linux

Note that there is already an include/linux/amba/pl08x.h.
I would just move the few parts of pl080.h that are actually
needed with global visibility there, and move the rest
to drivers/dma/.

>    drivers/gpio/gpio-pl061.c: uses chained_irq_enter/exit
>      proposed fix: depend on CONFIG_ARM (the function only exists in arm)
> 
>    drivers/mmc/host/mmci.c: uses <asm/sizes.h>
>    drivers/mmc/host/mmci.c: uses readsl/writesl
>      proposed fix: use linux/sizes.h and provide readsl/writesl like others do

Ack on the linux/sizes.h, that definitely makes sense.

I'm not sure I want to spread readsl/writesl beyond the architectures
that already have it. Maybe instead change the driver to use ioread32_rep,
which is already available on all architectures and is defined the
same way as readsl on ARM.

>    drivers/watchdog/sp805_wdt.c: uses writel_relaxed
>      proposed fix: depend on CONFIG_ARM (this is a spear-only cell by now)

This one on the other hand makes sense to be defined on all architectures
IMHO. I don't mind restricting the driver to ARM for now, but having
a generic writel_relaxed would be nice.

	Arnd
--
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/