netdev - Re: [Uclinux-dist-devel] [PATCH v3] add the driver for Analog Devices Blackfin on-chip CAN controllers

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]

Message-ID: <8bd0f97a0912161720o2649dd7eh2725146f4805486d@mail.gmail.com>
Date:	Wed, 16 Dec 2009 20:20:31 -0500
From:	Mike Frysinger <vapier.adi@...il.com>
To:	David Miller <davem@...emloft.net>
Cc:	wg@...ndegger.com, oe@...t.de, netdev@...r.kernel.org,
	socketcan-core@...ts.berlios.de,
	uclinux-dist-devel@...ckfin.uclinux.org
Subject: Re: [Uclinux-dist-devel] [PATCH v3] add the driver for Analog Devices 
	Blackfin on-chip CAN controllers

2009/12/10 David Miller:
> From: Mike Frysinger
>> On Thu, Dec 10, 2009 at 05:48, Wolfgang Grandegger wrote:
>>> Wolfgang Grandegger wrote:
>>>> Mike Frysinger wrote:
>>>>> On Thu, Dec 10, 2009 at 04:11, Wolfgang Grandegger wrote:
>>>>>> Well, I'm still not a friend of the following inline functions,
>>>>>> especially the *one-liners* which are called just *once*. With the usage
>>>>>> of structs they seem even more useless.
>>>>> seems like it would make more sense to not even use the read/write
>>>>> functions either. �,A just declare the regs as volatile and assign/read
>>>>> the struct directly.
>>>>
>>>> Two times no. Don't use volatile and proper accessor functions. See:
>>>>
>>>> http://lxr.linux.no/#linux+v2.6.32/Documentation/volatile-considered-harmful.txt
>>>
>>> I was just wondering if bfin_read/write16 would not be the proper
>>> accessor functions. readw/writew seems to be implemented differently:
>>>
>>> http://lxr.linux.no/#linux+v2.6.32/arch/blackfin/include/asm/io.h#L44
>>>
>>> Puh, they do an cli,nop,nop,sync..sti for the access. This also nicely
>>> shows why accessor functions should be used to access device registers.
>>>
>>> Well, just curious. I don't really know the blackfin arch.
>>
>> the common I/O functions need to account for issues surrounding the
>> bus that has arbitrary devices memory mapped to it.  on-chip devices
>> (like what we're talking about here) do not have these issues and so
>> using the common functions is awful overhead.
>
> Then create special accessors (perhaps with the same names as the
> existing ones, but with "__" prepended) that lack all of the
> interrupt disabling, syncs, etc.
>
> Really it _is_ cleaner and makes your driver look a lot nicer.

it really isnt.  more reasons why io accessors suck:
 - no type checking
   - easy to screw up register sizes (8/16/32)
   - have to add shims for each register that changes sizes between parts
   - no way to mark a register read-only (i.e. const)

unfortunately, we cant push the issue with Blackfin MMRs because of a
minor anomaly that exists on older parts.
-mike
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html