Or, better to say, why am I getting -EWOULDBLOCK when trying to bind to existing heap object created in kernel thread wih H_SHARED flag?
On 3/28/08, Gilles Chanteperdrix <[EMAIL PROTECTED]> wrote: > On Fri, Mar 28, 2008 at 4:28 PM, Bosko Radivojevic > <[EMAIL PROTECTED]> wrote: > > Ok, I'll try that. But, is there something wrong with my code/approach > > or is it some kind of arm-specific bug? > > Yes, as I said in my first answer, you are missing the O_CREAT flag in > the call to shm_open. > > The arm-specific bug will happen later (once the shared memory is > mapped). Because the kernel-space and user-space will get different > caches for the same memory area. The solution to this issue is to map > the heap without cache, but to get this, you need modifications of > Xenomai that only exist on my computer for now. So, if you are > interested, I will send a patch. > > -- > Gilles > -- Sent from Gmail for mobile | mobile.google.com _______________________________________________ Xenomai-help mailing list [email protected] https://mail.gna.org/listinfo/xenomai-help
