[Mono-dev] [PATCH] Mono port to the MIPS 64-bit

[Ian Dichkovsky]

Hello Mark,

It's good to see you again!

Actually, we was searching earlier for the good example of splitting
MIPS-related sources, but hadn't found it. So, we decided not to split
sources. And now you're giving one more reason against splitting.
Also, these patches could be used for backporting from n64 to o32. It was
easy to implement 64-bit related math operations for n64, but there will be
more effort needed for o32.

And we have prepared ChangeLogs and added them to
https://bugzilla.novell.com/show_bug.cgi?id=497320
ChangeLog.diff
ChangeLog_libgc.diff
ChangeLog_mono_arch.diff
ChangeLog_mono_io-layer.diff
ChangeLog_mono_mini.diff

We will be waiting for your results on bcm1480.

Regards,
Ian


Mark Mason-2 wrote:
> 
> Hello Miguel, Ian,
> 
> I'm working on getting my bcm1480 set back up after my office move (it  
> supports both o32 and n32), after which I can Ian's patches against  
> the o32 port.
> 
> As for splitting MIPS/MIPS64 - I think keeping them together may be  
> more appropriate. The MIPS64 code generator is less of a divergence  
> from MIPS32 than PPC64 is from PPC --- there's a lot of potential code  
> reuse to be had by keeping them together. That's why I'd started the  
> n32 port in the way that I did (and that Ian & company continued with  
> for extending it).
> 
> /Mark
> 
> On Apr 22, 2009, at 10:49 AM, Miguel de Icaza wrote:
> 
>> Hello Ian,
>>
>>    Thanks for doing the work in porting Mono to the MIPS64 platform  
>> and
>> for sending the patches our way.   Thanks to your developers and to
>> SiCortex for funding the port.
>>
>>     Some comments:
>>
>>> We have put 10 patches to the bugzilla.
>>> Bug 497320 - [PATCH] Set of Mono MIPS64-bit port patches
>>> https://bugzilla.novell.com/show_bug.cgi?id=497320
>>>
>>> Our primary goal was porting to the 64-bit platform.  
>>> Unfortunatelly, we
>>> haven't had a chance to test our changes on the box with o32 ABI.  
>>> There is a
>>> probability that we broke something. At least as far as we know,  
>>> the 64-bit
>>> port is the most usable among MIPS ports to this day. We this port  
>>> we have
>>> got quite good results, since regression, extensive and major part  
>>> of Class
>>> Library tests (99,60%) are passed successfully.
>>
>> While reviewing the patches, I noticed that this extends the MIPS port
>> to support MIPS64.   And I am wondering if it would make sense to  
>> follow
>> instead a similar setup to what we have done with PowerPC where we
>> actually have split things in a few places to differentiate PPC from
>> PPC64.
>>
>> For example we use a cpu-ppc.md and a cpu-ppc64.md files
>>
>> Also, we would be happy to give you SVN access so you can merge your
>> patches directly as they get approved, please follow the instructions
>> here:
>>
>> 	http://www.mono-project.com/SVN
>>
>> As I was reviewing the changes to provide some feedback as to what can
>> go in, I found that none of the changes come with ChangeLog entries,  
>> so
>> the rationale for some of the changes is not easy to understand.
>>
>> Would you mind providing ChangeLogs that describe what the changes are
>> to the various files, in particular the ones that affect the  
>> existing 32
>> bit work from Mark?
>>
>> Miguel.
>>
>> _______________________________________________
>> Mono-devel-list mailing list
>> Mono-devel-list at lists.ximian.com
>> http://lists.ximian.com/mailman/listinfo/mono-devel-list
> 
> _______________________________________________
> Mono-devel-list mailing list
> Mono-devel-list at lists.ximian.com
> http://lists.ximian.com/mailman/listinfo/mono-devel-list
> 
> 

-- 
View this message in context: http://www.nabble.com/-PATCH--Mono-port-to-the-MIPS-64-bit-tp23175539p23214505.html
Sent from the Mono - Dev mailing list archive at Nabble.com.