From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <4B8D05B3.1020202@domain.hid>
Date: Tue, 02 Mar 2010 13:33:55 +0100
From: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
MIME-Version: 1.0
References: <E1NmR5I-0003k9-N5@domain.hid> <4B8D0359.4070504@domain.hid>
	<4B8D0466.7040403@domain.hid> <4B8D0542.1020904@domain.hid>
In-Reply-To: <4B8D0542.1020904@domain.hid>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Subject: Re: [Xenomai-core] [Xenomai-git] Jan Kiszka : Native: Fix return
 code of in-kernel rt_cond_wait[_until]
List-Id: Xenomai life and development <xenomai.xenomai.org>
List-Unsubscribe: <https://mail.gna.org/listinfo/xenomai-core>,
	<mailto:xenomai-core-request@domain.hid>
List-Archive: </public/xenomai-core>
List-Post: <mailto:xenomai@xenomai.org>
List-Help: <mailto:xenomai-core-request@domain.hid>
List-Subscribe: <https://mail.gna.org/listinfo/xenomai-core>,
	<mailto:xenomai-core-request@domain.hid>
To: Jan Kiszka <jan.kiszka@domain.hid>
Cc: Xenomai core <Xenomai-core@domain.hid>

Gilles Chanteperdrix wrote:
> Jan Kiszka wrote:
>> Gilles Chanteperdrix wrote:
>>> GIT version control wrote:
>>>> Module: xenomai-jki
>>>> Branch: for-upstream
>>>> Commit: f1dfda551b6997f74141986759932e2723a9f024
>>>> URL:    http://git.xenomai.org/?p=xenomai-jki.git;a=commit;h=f1dfda551b6997f74141986759932e2723a9f024
>>>>
>>>> Author: Jan Kiszka <jan.kiszka@domain.hid>
>>>> Date:   Tue Mar  2 13:17:35 2010 +0100
>>>>
>>>> Native: Fix return code of in-kernel rt_cond_wait[_until]
>>>>
>>>> In rt_cond_wait_inner, do not let rt_cond_wait_epilogue overwrite the
>>>> primary error code of rt_cond_wait_prologue. This restores the in-kernel
>>>> semantics of rt_cond_wait[_until] that were valid before 97323b3287.
>>>>
>>>> Signed-off-by: Jan Kiszka <jan.kiszka@domain.hid>
>>>>
>>>> ---
>>>>
>>>>  ksrc/skins/native/cond.c |    8 ++++----
>>>>  1 files changed, 4 insertions(+), 4 deletions(-)
>>>>
>>>> diff --git a/ksrc/skins/native/cond.c b/ksrc/skins/native/cond.c
>>>> index 10727d1..2dc0069 100644
>>>> --- a/ksrc/skins/native/cond.c
>>>> +++ b/ksrc/skins/native/cond.c
>>>> @@ -472,10 +472,10 @@ static int rt_cond_wait_inner(RT_COND *cond, RT_MUTEX *mutex,
>>>>  	err = rt_cond_wait_prologue(cond, mutex, &lockcnt, 
>>>>  				    timeout_mode, timeout);
>>>>  
>>>> -	if(!err || err == -ETIMEDOUT || err == -EINTR)
>>>> -		do {
>>>> -			err = rt_cond_wait_epilogue(mutex, lockcnt);
>>>> -		} while (err == -EINTR);
>>>> +	if (!err || err == -ETIMEDOUT || err == -EINTR) {
>>>> +		while (rt_cond_wait_epilogue(mutex, lockcnt) == -EINTR)
>>>> +			; /* empty */
>>>> +	}
>>> Not ok. If rt_cond_wait_epilogue returns an error other than -EINTR, we
>>> want this error to be returned.
>> As I said: This is what we used to do in 2.4 (and early 2.5-rc), and due
>> to the absence of spec on this topic, I would vote for restoring old
>> behavior.
> 
> if rt_cond_wait_epilogue returns an error other than -EINTR, there is an
> error to be signaled, so, it should be returned to the user, not 0 or
> -ETIMEDOUT which would let the user think that he got the mutex back,
> whereas the mutex is probably not properly locked.

And if prologue returned -EINTR, and epilogue ran successfully, we do
not want to return -EINTR to the user either.

-- 
					    Gilles.