[LITMUS^RT] New release: 2012.3

[Felipe Cerqueira]

Hey, sorry, but I didn't see that cmpxchg64() macro is already implemented
in Litmus 3.0.
So the only code that we need to change at all is include/litmus/litmus.h:

cmpxchg(&tsk_rt(t)->ctrl_page->sched.raw, old.raw, new.raw) == old.raw;

=>

            #ifndef CONFIG_ARM
            return cmpxchg(&tsk_rt(t)->ctrl_page->sched.raw, old.raw,
new.raw) == old.raw;
            #else
                #if __LINUX_ARM_ARCH__ >= 6
                    #ifndef CONFIG_CPU_V6
                        return cmpxchg64(&tsk_rt(t)->ctrl_page->sched.raw,
old.raw, new.raw) == old.raw;
                    #else
                        #error cmpxchg64 not supported in this ARM
architecture
                    #endif
                #else
                    #error cmpxchg64 not supported in this ARM architecture
                #endif
            #endif

It's compiling. I'll send a patch after testing.

2013/1/16 Felipe Cerqueira <felipeqcerqueira at gmail.com>

> No, no, actually the best solution is to put an ifdef for ARM inside
> Litmus code and call cmpxchg64() instead of cmpxchg().
> So we get better compatibility with the newest kernels.
>
> For now we copy cmpxchg64() macro from 3.7.
>
>
> 2013/1/16 Felipe Cerqueira <felipeqcerqueira at gmail.com>
>
>>
>>
>> 2013/1/16 Björn Brandenburg <bbb at mpi-sws.org>
>>
>>>
>>> On Jan 16, 2013, at 5:31 PM, Glenn Elliott <gelliott at cs.unc.edu> wrote:
>>>
>>> >  Also, and probably the biggest issue going forward, testing on ARM is
>>> difficult since ARM kernels are so fragmented.  We have to port Litmus to
>>> each unique kernel devised for each of our ARM platforms.
>>>
>>> One more reason to rebase to a newer Linux version; they consolidated
>>> the ARM support in recent versions.
>>>
>>
>> Unfortunately 64-bit processors will only start to be produced in 2014
>> (Cortex A50 series).
>>
>> Good news though...
>> I found that Linux implements atomic64_cmpxchg() in
>> arch/arm/include/asm/atomic.h, even in the 3.0 kernel. Probably for the
>> older ARMv6k ARM boards.
>> In the first v7 instruction set ARM removed support for LDREXD, and it
>> only got back in v7-AR. The problem is that this is not treated in Linux
>> headers.
>>
>> 3.7 kernel does a little better job by providing access to
>> atomic64_cmpxchg() via another macro:
>>
>> #define cmpxchg64(ptr, o, n)                        \
>>     ((__typeof__(*(ptr)))atomic64_cmpxchg(container_of((ptr),    \
>>                         atomic64_t,        \
>>                         counter),        \
>>                           (unsigned long)(o),    \
>>                           (unsigned long)(n)))
>>
>>
>> /*
>>  * ARMv6 UP and SMP safe atomic ops.  We use load exclusive and
>>  * store exclusive to ensure that these are atomic.  We may loop
>>  * to ensure that the update happens.
>>  */
>> [...]
>>
>> static inline u64 atomic64_cmpxchg(atomic64_t *ptr, u64 old, u64 new)
>> {
>>     u64 oldval;
>>     unsigned long res;
>>
>>     smp_mb();
>>
>>     do {
>>         __asm__ __volatile__("@ atomic64_cmpxchg\n"
>>         "ldrexd        %1, %H1, [%3]\n"
>>         "mov        %0, #0\n"
>>         "teq        %1, %4\n"
>>         "teqeq        %H1, %H4\n"
>>         "strexdeq    %0, %5, %H5, [%3]"
>>         : "=&r" (res), "=&r" (oldval), "+Qo" (ptr->counter)
>>         : "r" (&ptr->counter), "r" (old), "r" (new)
>>         : "cc");
>>     } while (res);
>>
>>     smp_mb();
>>
>>     return oldval;
>> }
>>
>>
>> So, we can simply edit the normal cmpxchg() macro, put a new ifdef
>> CONFIG_ARMV7AR and add support to 64-bit.
>> That way, we won`t need to modify Litmus code.
>>
>> Thanks,
>> Felipe
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.litmus-rt.org/pipermail/litmus-dev/attachments/20130116/1c238ccd/attachment.html>