Re: synching with minios

[Anil Madhavapeddy <anil@xxxxxxxxxx>]

Absolutely. It just requires soft floating point support instead of hardware 
calls.  This is available in the ARM backend already, but is an unusual 
requirement for x86.

However, it's a requirement before we can merge Gabor Pali's FreeBSD kernel 
backend into the main tree, so it's on the TODO list to sort out. Note that 
most of the libraries we use only make occasional use of floating point (unless 
they're explicit numerical computation libraries), so this won't be a 
performance hit either.

-anil

On 11 Dec 2012, at 10:03, "Robert N. M. Watson" <robert.watson@xxxxxxxxxxxx> 
wrote:

> It's probably worth observing that unconditional dependence on floating point 
> also caused problems for Mirage running within the FreeBSD kernel, and would 
> also be a problem for embedded microcontroller/etc systems, should the 
> smallness and potential verifiability properties of Mirage become appealing 
> there. It would be lovely to see Mirage, and by implication the OCaml 
> runtime, make dependence on floating point optional.
> 
> Robert
> 
> On 11 Dec 2012, at 09:41, Anil Madhavapeddy <anil@xxxxxxxxxx> wrote:
> 
>> While updating the OCaml runtime to 4.00 (which now boots!), I also looked 
>> at recent MiniOS changes, and found this:
>> http://xenbits.xen.org/hg/xen-unstable.hg/rev/c9f621893a05
>> 
>> ...which adds floating point initialisation to MiniOS also.  This was the 
>> root cause of the floating point exceptions that Balraj was seeing when 
>> running Xen 4.2 on i7 CPUs.  Always worth checking upstream first :)
>> 
>> I'm doing a quick sync of the MiniOS tree with ours too, while I'm here.  
>> It's not quite so "mini" upstream anymore, with many many new drivers that 
>> we don't need, but there are some useful cleanups in the memory management 
>> layer as well.
>> 
>> -anil
>