linux-kernel - Re: [PATCH] drivers/block/xsysace - replace in(out)_8/in(out)_be16/in(out)

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, 12 Feb 2013 11:11:53 +0100
From:	Michal Simek <monstr@...str.eu>
To:	Arnd Bergmann <arnd@...db.de>
Cc:	Geert Uytterhoeven <geert@...ux-m68k.org>,
	Grant Likely <grant.likely@...retlab.ca>,
	Alexey Brodkin <Alexey.Brodkin@...opsys.com>,
	Benjamin Herrenschmidt <benh@...nel.crashing.org>,
	Vineet Gupta <Vineet.Gupta1@...opsys.com>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Alan Cox <alan@...rguk.ukuu.org.uk>,
	dahinds@...rs.sourceforge.net
Subject: Re: [PATCH] drivers/block/xsysace - replace in(out)_8/in(out)_be16/in(out)_le16
 with generic iowrite(read)8/16(be)

2013/2/11 Arnd Bergmann <arnd@...db.de>:
> On Monday 11 February 2013, Michal Simek wrote:
>> Unfortunately no. Another is spi/i2c (sysace as we discuss in this
>> thread), probably icap
>> network drivers are ok because they are not shared.
>> Timer when it is moved to clocksource(not important right now)
>> xilinx gpio is using __raw IO functions which is incorrect on arm in
>> connection to barriers.
>>
>> But it reminds me that maybe the easiest solution is not to use endian
>> accessors just use two simple macros which should work on all systems.
>>
>> #define <name>_readreg(offset)          ({__raw_readl(offset); rmb(); })
>> #define <name>_writereg(offset, val)    ({wmb(); __raw_writel(val, offset); })
>>
>> which is probably the faster solution which add minimum additional code
>> to driver and can also remove endian detection code.
>
> I tend to disagree. You should never assume that a device is the same
> endianess as the the CPU, and you should try not to use the __raw_*
> accessors in device drivers either.
>
> In particular, ARM can run both big- and little-endian even though
> big-endian is rarely used, so you need to know the endianess for
> the device you are talking to rather than assume that it knows
> what the CPU does at the time.

For high performance IPs using accessors functions is still problematic
because there will be performance regression it means that
from my point of view there still should be any option to "setup"
proper endians for the driver and it can't be setup at run-time.

Sure the question is if drivers like this should be in the mainline.
But if yes, there should be an option to do it in acceptable way.

Thanks,
Michal

-- 
Michal Simek, Ing. (M.Eng)
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel - Microblaze cpu - http://www.monstr.eu/fdt/
Maintainer of Linux kernel - Xilinx Zynq ARM architecture
Microblaze U-BOOT custodian and responsible for u-boot arm zynq platform
--
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/