On Tue, Jul 15, 2014 at 10:16:07PM -0400, Paul Moore wrote:
> On Tuesday, July 15, 2014 06:06:11 PM Marcus Meissner wrote:
> > Hi,
> 
> Hello.
> 
> > I currently porting libseccomp to ppc,ppc64,ppc64le,s390,s390x.
> 
> Excellent, thank you.
> 
> > My current work is attached here and should be working already.
> 
> I haven't had a chance to look at your patch yet, but a few administrative 
> notes to consider when submitting patches to this list:
> 
> * Please send patches inline, it makes it easier to review and comment on 
> your 
> work.
> 
> * Please split up large patches into reasonable sized chunks whenever 
> possible; check the git log to get an idea of what is generally acceptable.  
> However, one golden rule when submitting patchsets with multiple patches, 
> each 
> individual patch must not break the build or the functionality of the 
> library.  
> There have been some exceptions in the past, but don't rely on an exception.

Ok, I will try.
 
> * Run "make check" and "make check-syntax", please correct any problems that 
> are reported.

Ok.

> > The testsuite is however giving me trouble.
> > 
> > Due to some endianess problems in the interaction between the testcase
> > generators and the simulator the testsuite will however now show
> > some errors.
> > 
> > The problem is e.g. in 26-sim-arch_all_be_basic, that it emits
> > opcodes in "big endian" as MIPS is the first architecture.
> > 
> > The simulator when testing the big endian platforms does
> > - MIPS succeeds (as it is the "native" arch and no endianess conversion is
> > done) - all other BE architectures get converted endianess from HOST
> > (little) to big, so the opcodes are swapped.
> > 
> > This will result in several new errors currently even on x86_64.
> > Best would probably be to also add -a $ARCH to the testcase bpf stream
> > generator? This would at least make the testsuite pass on x86_64 again.
> 
> In general, for big endian architectures you should be able to duplicate what 
> was done with MIPS.  This will likely involve changes to the regression test 
> driver script as well as some of the programs under tools/.  The git log for 
> those files will likely help you identify what needs to be done, however, if 
> you run into problems please don't hesitate to ask.

> Does that make sense?

Makes sense. I will see that I can adjust the testsuite driver.

Ciao, Marcus

------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
libseccomp-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libseccomp-discuss

Reply via email to