linux-kernel - Re: [PATCH 4/6] regmap: add MMIO bus support

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:	Wed, 04 Apr 2012 17:15:42 -0600
From:	Stephen Warren <swarren@...dotorg.org>
To:	Mark Brown <broonie@...nsource.wolfsonmicro.com>
CC:	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 4/6] regmap: add MMIO bus support

On 04/04/2012 04:59 PM, Mark Brown wrote:
> On Wed, Apr 04, 2012 at 03:48:31PM -0600, Stephen Warren wrote:
>> From: Stephen Warren <swarren@...dia.com>
>>
>> This is a basic memory-mapped-IO bus for regmap. It has the following
>> features and limitations:
> 
> I applied these up to here but it was painful as patch 2 didn't apply
> cleanly to -rc1 and git am couldn't find the blobs to use for
> resolution.  Please check things worked out OK but I'm pretty sure they
> did.

The result of merging your topic/mmio and for-next branches is the same
as my local copy of patch 4 I sent, modulo that my local copy still has
regmap_open_file(), but that's not related.

So yes, I think the patches applied fine.

>> * Registers themselves may be 8, 16, 32, or 64-bit. 64-bit is only
>>   supported on 64-bit platforms.
>> * Register offsets are limited to precisely 32-bit.
>> * IO is performed using readl/writel, with no provision for using the
>>   __raw_readl or readl_relaxed variants.
> 
> Also limited native endian register I/O.  It would have been much better
> to fix this in the core - please consider producing followup patches to
> push the code there, though it's far from essential.

I assume the solution here is to:

* Always use __raw_readl/__raw_writel in regmap-mmio.c so that there's
never any endianness conversion, and don't use the endianness conversion
macros when reading/writing the work buffer.
* Implement alternative formatters that format in LE instead of BE
* Add a field to regmap_config indicating which endianness the data
should be in, and use this flag to select the LE/BE formatters.

Does that sound about right?

>> +static int regmap_mmio_gather_write(void *context,
>> +				    const void *reg, size_t reg_size,
>> +				    const void *val, size_t val_size)
>> +{
>> +	struct regmap_mmio_context *ctx = context;
>> +	u32 offset;
>> +
>> +	if (reg_size != 4)
>> +		return -EIO;
> 
> Given that you constrain on registration too this should be BUG_ON(),
> we're seriously confused if we're specifying a different register size
> here and -EIO is going to be a bit obscure.

I wondered about that. I would have been quite happy to repost with that
change. Do you want an incremental patch?
--
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/