linux-kernel - Re: [PATCH 3/3] Add "no-bus" option for regmap API

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] [day] [month] [year] [list]

Date:	Sat, 29 Dec 2012 11:18:11 -0800
From:	Andrey Smirnov <andrew.smirnov@...il.com>
To:	Mark Brown <broonie@...nsource.wolfsonmicro.com>
Cc:	"andrey.smirnov@...vergeddevices.net" 
	<andrey.smirnov@...vergeddevices.net>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 3/3] Add "no-bus" option for regmap API

>> -     int (*reg_read)(struct regmap *map, unsigned int reg, unsigned int *val);
>> -     int (*reg_write)(struct regmap *map, unsigned int reg, unsigned int val);
>> +     int (*reg_read)(void *context, unsigned int reg, unsigned int *val);
>> +     int (*reg_write)(void *context, unsigned int reg, unsigned int val);
>
> I'd be inclined to just do this in the initial refectoring patches
> rather than rerefactoring here.

I agree. I realized that I would have to have those parameters as
"void *" only after I made first two commits and since I knew that the
first version of the patch would have other problems decided to
postpone fixing this till the second version of the patch.

>
>> +     if (!bus || !bus->fast_io) {
>>               mutex_init(&map->mutex);
>>               map->lock = regmap_lock_mutex;
>>               map->unlock = regmap_unlock_mutex;
>> +     } else {
>> +             spin_lock_init(&map->spinlock);
>> +             map->lock = regmap_lock_spinlock;
>> +             map->unlock = regmap_unlock_spinlock;
>
> It's not immediately obvious to me that no-bus should be forced to use
> mutexes - is there any great reason for tying the two together?

No reason. I'll add provisions for configuring.

>
>> +     if (!bus) {
>> +             map->cache_registers = true;
>> +             goto skip_format_initialization;
>> +     } else {
>> +             map->reg_read = _regmap_bus_read;
>> +     }
>
> Not sure I understand cache_registers here.  Why has this flag been
> added?

As of now when the write operation is performed("_regmap_write" is
being called) the caching is handled either in "_regmap_write" or in
"_regmap_raw_write". The decision which one of the two places it will
be done at is made based on the availability of the
"map->format.format_write". With the addition of "no-bus"
configuration "map->format.format_write" is no longer a valid
criterion for that decision and since caching writes makes sense for
"no-bus" configuration too, I added this variable to serve as flag so
that caching would be handled in "_regmap_write" for "no-bus"
configuration and for busses with "map->format.format_write" and in
"_regmap_raw_write" for all the other cases.

>> + * @reg_read: Optional callback that if filled will be used to perform
>> + *            all the reads from the registers.
>> + * @reg_write: Optional callback that if filled will be used to perform
>> + *             all the writes to the registers.
>
> I'd probably add some comment about not using this in conjunction with
> SPI or I2C.

Will do.
--
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/