RTEMS crash with mem address not aligned failure
Chris Johns
chrisj at rtems.org
Mon Oct 6 22:45:11 UTC 2014
On 7/10/2014 2:19 am, Joel Sherrill wrote:
>
> On 10/6/2014 10:09 AM, Jiri Gaisler wrote:
>>
>> On 10/06/2014 04:45 PM, Joel Sherrill wrote:
>>> On 10/6/2014 2:03 AM, Chris Johns wrote:
>>>> On 6/10/2014 5:08 pm, Sebastian Huber wrote:
>>>>> On 04/10/14 07:07, Chris Johns wrote:
>>>>>>>> Could this be a newlib issue? We are using:
>>>>>>> Is the memory map ok for the BSP ? It pays when moving to 4.11 from 4.10
>>>>>>> to check the memory map is ok.
>>>>>>>
>>>>>> I can confirm this happens on the SIS bsp in simulator using the
>>>>>> fileio.exe
>>>>>> example. Run the sample, hit a key, start the shell and enter ls.
>>>>> I don't see this problem here. I use sparc-rtems4.11-gcc 4.9.2 20140924
>>>>> with the latest Newlib.
>>>>>
>>>> Do you see it with the RSB tools ?
>>> I have the RSB tools and it doesn't crash for me on sis. I also noticed
>>> that in the report, there is a hello command which is not part
>>> of the standard command set. I wonder if that command or
>>> how it was added is triggering something.
>>>
>>> This may not be the latest but this is what I have installed.
>>>
>>> $ sparc-rtems4.11-gcc --version
>>> sparc-rtems4.11-gcc (GCC) 4.8.3 20140522 (RTEMS
>>> 4.11-RSB-257d1e4378c80968de57ba888e235888adb0e992-1,gcc-4.8.3/newlib-19-Aug-2014)
>> It does crash on sis for me, using RSB from some months ago:
>>
>> $ sparc-rtems4.11-gcc --version
>> sparc-rtems4.11-gcc (GCC) 4.8.2 20131016 (RTEMS 4.11-RSB-527515c38007b0614a2cafed863190438e8375bc-1,gcc-4.8.2/newlib-2.1.0)
>> Copyright (C) 2013 Free Software Foundation, Inc.
>> This is free software; see the source for copying conditions. There is NO
>> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
> Hmm... I wonder if something was fixed in newlib.
I have just rebuild everything ...
gcc version 4.8.3 20140522 (RTEMS
4.11-RSB-bc51abf2519aa82bfb1208ad8844cdbfaad45fc9-1,gcc-4.8.3/newlib-19-Aug-2014)
(GCC)
and the latest rtems git head and I do not see the problem so something
has fixed this.
Chris
More information about the users
mailing list