Re: [PATCH] CRISv10 serial driver rewrite

!MAILaRCHIVE_VOTE_RePLACE
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]
To: Jesper Nilsson <jesper.nilsson@...>
Cc: Andrew Morton <akpm@...>, Mikael Starvik <mikael.starvik@...>, <linux-kernel@...>
Date: Friday, November 2, 2007 - 6:39 am

On 11/02/2007 10:34 AM, Jesper Nilsson wrote:
[...]

Is this enough? Don't you need also spin lock (i.e. spin_lock_irqsave())?


[...]


Aiee, this is nonsense, 0 will never be 1, only signal will stop this, use
completion instead.


also here.


You maybe want to create a function for this deinit invoked from more places in
the new code.


...


Side note, this should be setup_timer without accessing .function.


Is the panic needed here? Can't the cris architecture live without the driver?


bitfileds generate ugly code.


What's the volatile for here? atomic_t?


We know, we have it, don't we?



-- 
Jiri Slaby (jirislaby@gmail.com)
Faculty of Informatics, Masaryk University
-
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:
[PATCH] CRISv10 serial driver rewrite, Jesper Nilsson, (Fri Nov 2, 5:34 am)
Re: [PATCH] CRISv10 serial driver rewrite, Jiri Slaby, (Fri Nov 2, 6:39 am)
Re: [PATCH] CRISv10 serial driver rewrite, Jesper Nilsson, (Mon Nov 5, 7:18 am)
Re: [PATCH] CRISv10 serial driver rewrite, Jiri Slaby, (Mon Nov 5, 7:24 am)