[copy to olsr-dev list]
can we please switch off compiler-optimizations for
all ARM and MIPS-platforms, till the compiler-bug
is filed [lars?] and fixed? Simply omit "-Os" does the trick.
https://lists.olsr.org/pipermail/olsr-dev/2012-August/006046.html
bye, bastian
On 2012-08-29 2:12 PM, Bastian Bittorf wrote:
> [copy to olsr-dev list]
>
> can we please switch off compiler-optimizations for
> all ARM and MIPS-platforms, till the compiler-bug
> is filed [lars?] and fixed? Simply omit "-Os" does the trick.
>
> https://lists.olsr.org/pipermail/olsr-dev/2012-Au
> Disabling all optimization sounds like a horrible idea to me. How
> about
> identifying which specific optimization is causing this issue? If
if you follow the thread, you can see it:
https://lists.olsr.org/pipermail/olsr-dev/2012-August/006056.html
problem1:
for filing a bug-report we need to
On 2012-08-29 5:37 PM, Bastian Bittorf wrote:
>> Disabling all optimization sounds like a horrible idea to me. How
>> about
>> identifying which specific optimization is causing this issue? If
>
> if you follow the thread, you can see it:
> https://lists.olsr.org/pipermail/olsr-dev/2012-August/006
2012.08.29. 17:37 keltezéssel, Bastian Bittorf írta:
>> Disabling all optimization sounds like a horrible idea to me. How
>> about
>> identifying which specific optimization is causing this issue? If
>
> if you follow the thread, you can see it:
> https://lists.olsr.org/pipermail/olsr-dev/2012-Aug
> Only the Truemobile 2300 and the the Asus WL-500G was mentioned in
> the thread
> and both of them is based on BCM4702/4710. An old post [1] on the
> linux-mips
> list indicates that those SoCs may have problems with the
> 'branch-likely'
> instructions, so this might not be a compiler bug at all
2012.08.31. 21:28 keltezéssel, Bastian Bittorf írta:
>> Only the Truemobile 2300 and the the Asus WL-500G was mentioned in
>> the thread
>> and both of them is based on BCM4702/4710. An old post [1] on the
>> linux-mips
>> list indicates that those SoCs may have problems with the
>> 'branch-likely'
Hello Bastian,
On Saturday 01 September 2012 09:25:06 Gabor Juhos wrote:
> 2012.08.31. 21:28 keltezéssel, Bastian Bittorf írta:
> >> Only the Truemobile 2300 and the the Asus WL-500G was mentioned in
> >> the thread
> >> and both of them is based on BCM4702/4710. An old post [1] on the
> >> linux-
2012.09.07. 12:32 keltezéssel, Florian Fainelli írta:
> Hello Bastian,
>
> On Saturday 01 September 2012 09:25:06 Gabor Juhos wrote:
>> 2012.08.31. 21:28 keltezéssel, Bastian Bittorf írta:
Only the Truemobile 2300 and the the Asus WL-500G was mentioned in
the thread
and both of them
sorry for late reply and breaking the ml-thread:
> > You can add '-mno-branch-likely' option to CFLAGS. The option ensures that
> > such
> > instructions will not be emitted by the toolchain.
> Can you test again after r33328? This will need a complete toolchain rebuild.
i tested a fresh build
On 2012-09-21 3:19 PM, Bastian Bittorf wrote:
> sorry for late reply and breaking the ml-thread:
>
>> > You can add '-mno-branch-likely' option to CFLAGS. The option ensures that
>> > such
>> > instructions will not be emitted by the toolchain.
>
>> Can you test again after r33328? This will nee
Have anyone tried other MIPS specific options, like -mfix-r4000?
Felix Fietkau wrote:
On 2012-09-21 3:19 PM, Bastian Bittorf wrote:
> sorry for late reply and breaking the ml-thread:
>
>> > You can add '-mno-branch-likely' option to CFLAGS. The option ensures that
>> > such
>> > instructions wi
> > i tested a fresh build of openwrt r33502.
> > this changed nothing 8-( i will now build an
> > image without "-Os" ...
> If you can get me a very simple test case that I can use to
> reproduce
> this issue, I will look into it.
sorry, no simple testcase. It works without any problems
when i s
On 2012-09-23 9:23 AM, Bastian Bittorf wrote:
>> > i tested a fresh build of openwrt r33502.
>> > this changed nothing 8-( i will now build an
>> > image without "-Os" ...
>
>> If you can get me a very simple test case that I can use to
>> reproduce
>> this issue, I will look into it.
>
> sorry,
> First of all: what's the fastest possible way to see if 'it works'
1)
standard build of package olsr must run on a affected router
with BCM4702/4710 CPU (e.g. Dell Truemobile 2300 or
Asus WL-500G).
2)
Add a 2nd router (arch does not matter) via Ethernet
(we need a second, the problem is that th
15 matches
Mail list logo