Re: [microblaze-uclinux] [patch] errno corruption

[Yasushi SHOJI <yashi@xxxxxxxxxxxxxxxxx>]

Hi John,

At Tue, 16 Mar 2004 09:45:48 +1000,
John Williams wrote:
> 
> > I have no idea about that.  I'm using the latest dist with your
> > patches applied, unless I totally messed up my env. (and it happens a
> > lot ;p)
> 
> I expect it might be optimisation related - although I'm not sure why 
> you and I would have different optimisation settings in the uClibc build...

ugaaaa, that's right. I have been using -Os instead of -O2!!!

-CFLAGS := $(CPPFLAGS) -Wall -Wstrict-prototypes -Wno-trigraphs -O2 \
+CFLAGS := $(CPPFLAGS) -Wall -Wstrict-prototypes -Wno-trigraphs -Os \

sorry about that! I've totally forgot about it!

> > ps. does anyone know what instruction 'la' stands for?
> 
> "Load Address"

aha! thanks ;)

since it usually used for loading a value at a symbol, i should have
used addk in the second patch instead.

> Anyway, the way forward:  I think we need to digest Dave's earlier email 
> and do it properly - remove these assumptions that are causing the 
> problem.  Basically that means rolling the syscall_return macro into the 
> syscallN declarations, with inline assembly done properly.

basically my second patch + macro move, right?

or do you want to implement in asm, so that we can optimise the way we
want?

> I don't have time to look into it right now - but please feel free to 
> have a go and post the results!  Until then, I'd suggest leaving your 
> first patch in CVS, since it seems to fix the problem  - even if for the 
> wrong reason! :)

if you can tell me how you want, i can generate a patch. ;)

regards,
--
        yashi
___________________________
microblaze-uclinux mailing list
microblaze-uclinux@itee.uq.edu.au
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/