OsmoTRX + LimeSDR + Pine64

[Alex]

Hi,
I made a simple workaround:
I yust commented out the two “assume“ directives in start function of
Thread.cpp

After recompiling everything seems to work properly, so it's OK, but I
don't know exactly what checks I'm currently removing and their
implications.

Anyway, on a pure user experience basis, I can say it works.

Best regards
Alex

On Sun, 13 Jan 2019, 23:38 Alex <allexander.alex at gmail.com wrote:

> Hi Vadim,
> A ram issue seems strange to me... The board has 2 gb of ram and I also
> successfully run the same config on a 1 gb virtual machine.
>
> Maybe a kernel limitation?
>
>
> Thank you and best regards
> Alex
>
> On Sun, 13 Jan 2019, 23:35 Vadim Yanitskiy <axilirator at gmail.com wrote:
>
>> Hello,
>>
>> > osmo-trx-lms: Threads.cpp:133:
>> > void Thread::start(void* (*)(void*), void*): Assertion `!res' failed.
>> > [...]
>> > Can someone provide me a little light on this?
>>
>> As far as I can see, pthread_attr_setstacksize() fails to set
>> required stack size for a thread. Probably, the amount of RAM
>> is not enough. But in general, this part of code looks dirty:
>>
>> > /** A C++ wrapper for pthread threads.  */
>> > class Thread {
>> >   /* ... */
>> >   // FIXME -- Can this be reduced now?
>> >   size_t mStackSize;
>> >
>> >   /* ... */
>> >   /** Create a thread in a non-running state. */
>> >   Thread(size_t wStackSize = (65536*4)):mThread((pthread_t)0) {
>> >     pthread_attr_init(&mAttrib);    // (pat) moved this here.
>> >     mStackSize=wStackSize;
>> >   }
>> >   /* ... */
>>
>> I am now wondering, where does this magic 65536*4 comes from,
>> and how can we estimate and adjust this properly?
>>
>> With best regards,
>> Vadim Yanitskiy.
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osmocom.org/pipermail/openbsc/attachments/20190114/87b13b87/attachment.htm>