All of lore.kernel.org
 help / color / mirror / Atom feed
From: Sergei Shtylyov <sshtylyov@ru.mvista.com>
To: Manuel Lauss <mano@roarinelk.homelinux.net>
Cc: Linux-MIPS <linux-mips@linux-mips.org>
Subject: Re: [PATCH 0/3] Alchemy: platform updates
Date: Sun, 29 Mar 2009 17:03:58 +0400	[thread overview]
Message-ID: <49CF71BE.2070109@ru.mvista.com> (raw)
In-Reply-To: <20090329143802.0a09baca@scarran.roarinelk.net>

Hello.

Manuel Lauss wrote:

>>> Patch overview:
>>>
>>> #1: eliminate alchemy/common/platform.c.  Add platform device
>>>     registration to all boards instead.
>>>   
>>>       
>>   I'm strongly voting against this, as it causes totally unneeded code 
>> duplication. Please don't apply.
>>     
>
> I know it seems like a lot of unecessary duplication.  My reasons for
> doing this are
>
> a) I want to get rid of the various cpu-type specific config symbols,
>    (i.e. I want one kernel binary to run on 2 or more Alchemy boards
>    with different cpu models.  There's no technical reason this
>    shouldn't work.  The OHCI block for instance has different addresses
>    on pre-Au1200 silicon; the easiest solution to that (for me) is
>    to simply move device registration to the boards that need/want it.)
>   

   Well, then create several files (like 
arch/mips/alchemy/common/au1200.c) for SoC specific platfrom device sets 
variations but don't move all platform devices into the board files. 
This will be a cleaner solution.

>    (BTW, what do you think about this?  Please tell me now if I ever
>     have a chance to get your approval on this, so I can stop wasting
>     my and everybody else's time as soon as possible).
>   

  Single kernel binary? If it's at all possible, I am all for it.

> b) the way it is now just seems wrong to me, probably comes from working
>    on SH port where each board registers the devices it needs.  I know
>    this is in no way a valid technical justification; but everytime I
>    look at this file something tells me it is just wrong ;)
>   

  Don't listen to the voices in your head. ;-)

>>>     I realize this is a lot of (needless) code duplication at first,
>>>     but it seems a lot cleaner to me if each board registered the
>>>     devices it needs/wants.  
>>>       
>>    No, it's certainly a step backwards. You could make the common code 
>> more flexible by checking what devices are enabled and registering them 
>> selectively
>>     
>
> Please elaborate.  Shall I check for a config symbol or if bootloader
> has enabled the peripherals?  The latter is IMO stupid since the
> bootloader should only load an OS and not preemptively enable every
> device which might some day be used just because the running OS doesn't
> know what it wants...
>   

   It should check what the board code (or bootloader) has enabled. I 
remember the board code writes some config registers...

> 	Manuel Lauss
>   

WBR, Sergei

  reply	other threads:[~2009-03-29 13:04 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-03-29  9:26 [PATCH 0/3] Alchemy: platform updates Manuel Lauss
2009-03-29  9:27 ` [PATCH 1/3] Alchemy: get rid of common/platform.c Manuel Lauss
2009-03-29  9:27   ` [PATCH 2/3] Alchemy: add RTC device to devboards Manuel Lauss
2009-03-29  9:27     ` [PATCH 3/3] Alchemy: convert to physmap flash Manuel Lauss
2009-04-06  7:57       ` Florian Fainelli
2009-03-29 11:36   ` [PATCH 1/3] Alchemy: get rid of common/platform.c Sergei Shtylyov
2009-03-29 11:37   ` Sergei Shtylyov
2009-03-29 11:35 ` [PATCH 0/3] Alchemy: platform updates Sergei Shtylyov
2009-03-29 12:38   ` Manuel Lauss
2009-03-29 13:03     ` Sergei Shtylyov [this message]
2009-03-29 15:27       ` Kevin Hickey
2009-03-29 15:52         ` Manuel Lauss
2009-03-30  8:39           ` Florian Fainelli
2009-03-30 16:23           ` Ralf Baechle
2009-03-29 15:43       ` Manuel Lauss
2009-03-29 20:48       ` Sergei Shtylyov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=49CF71BE.2070109@ru.mvista.com \
    --to=sshtylyov@ru.mvista.com \
    --cc=linux-mips@linux-mips.org \
    --cc=mano@roarinelk.homelinux.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.