From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <478CBD27.4020501@domain.hid>
Date: Tue, 15 Jan 2008 15:03:19 +0100
From: "Bernhard Michael" <michael.bernhard@domain.hid>
MIME-Version: 1.0
References: <47891672.4050409@domain.hid>
	<2ff1a98a0801140525y2bc4a198o97967585f6de0f7c@domain.hid>
	<478B8249.90303@domain.hid>
	<2ff1a98a0801140752y5e6e47e2tb9980bfb344522a4@domain.hid>
	<478CAB5F.70208@domain.hid>
	<2ff1a98a0801150509w2dae4826r4696cd95c21aecb3@domain.hid>
In-Reply-To: <2ff1a98a0801150509w2dae4826r4696cd95c21aecb3@domain.hid>
Content-Type: text/plain;
	charset="ISO-8859-1"
Content-Transfer-Encoding: 7bit
Subject: Re: [Xenomai-help] Xenomai problems on pxa
List-Id: Help regarding installation and common use of Xenomai
	<xenomai.xenomai.org>
List-Unsubscribe: <https://mail.gna.org/listinfo/xenomai-help>,
	<mailto:xenomai-help-request@domain.hid>
List-Archive: </public/xenomai-help>
List-Post: <mailto:xenomai@xenomai.org>
List-Help: <mailto:xenomai-help-request@domain.hid>
List-Subscribe: <https://mail.gna.org/listinfo/xenomai-help>,
	<mailto:xenomai-help-request@domain.hid>
To: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
Cc: xenomai@xenomai.org

Gilles Chanteperdrix wrote:
> On Jan 15, 2008 1:47 PM, Bernhard Michael <michael.bernhard@domain.hid> wrote:
>> Gilles Chanteperdrix wrote:
>>> It looks like we have two issues:
>>> - one issue is that compiling with the EABI toolchain generates a lot
>>> of unaligned access, which causes a crash, when the fault handler gets
>>> invoked at the wrong time. To solve this issue, I will check if we can
>>> allow the unaligned access to be handled in primary mode under some
>>> condition (it will need at least, a valid linux stack)
>>> - another issue is that there is something wrong with Linux timer
>>> interrupt handling, which causes the soft lockups and weird timer
>>> behaviour.
>>>
>>> Could you try the following:
>>> in the version compiled without EABI, in Linux file
>>> arch/arm/mach-pxa/time.c, function pxa_timer_interrupt, try removing
>>> the do { } while loop in :
>>>
>>>         do {
>>>                 timer_tick();
>>> #ifndef CONFIG_IPIPE
>>>                 OSSR = OSSR_M0;  /* Clear match on timer 0 */
>>> #else /* CONFIG_IPIPE */
>>>                 last_jiffy_time += LATCH;
>>>                 if (__ipipe_mach_timerstolen)
>>>                         next_match = last_jiffy_time + LATCH;
>>>                 else
>>> #endif /* CONFIG_IPIPE */
>>>                 next_match = (OSMR0 += LATCH);
>>>         } while( (signed long)(next_match - OSCR) <= 8 );
>>>
>> Gilles, I removed 'do' and 'while' as you suggested but the kernel does
>> not boot. The last boot messages before the kernel stalls are the following:
>>
>> ...
>> PID hash table entries: 256 (order: 8, 1024 bytes)
>> I-pipe 1.8-03: pipeline enabled.
>> start_kernel(): bug: interrupts were enabled early
>> Console: colour dummy device 80x30
>> Dentry cache hash table entries: 8192 (order: 3, 32768 bytes)
>> Inode-cache hash table entries: 4096 (order: 2, 16384 bytes)
>> Memory: 64MB = 64MB total
>> Memory: 61568KB available (2928K code, 279K data, 104K init)
> 
> Ok, it means that vanilla linux needs the do { } while loop. Let us
> try something else: put back the do and while. And modify the function
> pxa_timer_init. Replace:
> 
> OSMR0 = OSCR + LATCH;
> 
> with:
> 
> last_jiffy_time = OSCR;
> OSMR0 = last_jiffy_time + LATCH;
> 
Ok, this solved the soft lock-up. This change is obvious (once discovered :-) ) 
because 'last_jiffy_time' is initialised nowhere.

Now, I can load the xeno-* modules without problem. It is also working with
the EABI toolchain. Thanks Gilles.

However, the other problems remain :-(

--
Michael