[Vm-dev] Warning, warning.. new danger!

Igor Stasenko siguctua at gmail.com
Fri Jan 17 18:59:24 UTC 2014

On 17 January 2014 19:03, Eliot Miranda <eliot.miranda at gmail.com> wrote:

> Hi Igor,
> On Fri, Jan 17, 2014 at 5:02 AM, Igor Stasenko <siguctua at gmail.com> wrote:
>> From:
>> http://sourceforge.net/p/fbc/bugs/659/
>> "The GCC devs decided to unilaterally change the Linux x86 ABI [1] [2]
>> [3] [4]. I'm not sure when this happened, but I think it first became a
>> common problem with GCC 4.1. Previously the Linux x86 ABI was the "SysV
>> i386 ABI", which stated that the stack is aligned to a 4-byte boundary on
>> function entry. GCC now assumes by default that the stack is aligned to a
>> 16-byte boundary. This is an very controversial issue with the GCC devs
>> saying they have changed the ABI, and many other people considering the
>> SysV ABI to be the real ABI and GCC to be buggy. GCC Bugzilla is full of
>> flamewars. GCC devs have said "GCC chose to change the *unwritten*standard for the ABI in use for IA32 GNU/Linux" and "The ABI is
>> undocumented; that is reality" [3]"
>> basically, it affects everything which dealing with generated code.
>> like JIT/FFI.
>> Now it explains why NB started crashing on never versions of ubuntu,
>> while worked well before.
> But Mac OS X has had 16-byte alignment for the entirety of Mac OS X on
> x86.  And so Cog has used 16-byte alignment on linux also for a long time
> now.  I don't think this is too much of a departure from the  SysV i386
> ABI.  A change in what registers are callee-saved, what register is the
> result register etc would be disastrous.  There are other changes which
> would be welcome too, such as returning float results through SSE registers
> instead of though the, frankly obsolete, x87 stack.  So I don't think this
> is cause to throw up one's hands, Robbie style ;-)
> i am quite neutral about what ABI to use..
i just don't like that there is no some official way to determine what
actual convention
a binary compatible with .. therefore i forced to assume worst and use most
pessimistic ABI model. :(

> --
> best,
> Eliot

Best regards,
Igor Stasenko.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/vm-dev/attachments/20140117/78d6e5c6/attachment-0001.htm

More information about the Vm-dev mailing list