[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: GNAT, executable stack and trampolines



On 07/02/2011 03:51 PM, Ludovic Brenta wrote:
> Reto Buerki wrote on Wed, 15 Jun 2011 15:52:52 +0200:
>> On 05/24/2011 05:12 PM, Reto Buerki wrote:
>>> On 05/24/2011 04:15 PM, Ludovic Brenta wrote:
>>>>
>>>> Reto Buerki wrote:
>>>>> I can't find the file gcc/ada/ChangeLog-2008 in neither GNAT GPL 2010
>>>>> sources nor binary distribution tarballs. Am I missing something?
>>>>
>>>> That's unfortunate but not insurmountable.  You'll have to reconstruct the
>>>> files from the GCC 4.3.0 source distribution and the gcc43.dif patch
>>>> provided by AdaCore.  (The exact version number of GCC should be in a
>>>> README somewhere).
>>>
>>> I did it exactly as described in the README.BUILD file from
>>> gnat-gpl-2010-src.tgz using gcc-4_3-branch-rev160200-src.tgz provided by
>>> AdaCore. Besides that, gcc43.dif does not patch nor create any ChangeLog
>>> files ...
>>
>> I used the opportunity of today's GNAT GPL 2011 release (thanks AdaCore)
>> to analyze the latest gcc-45.dif patch in the source tarball. I wrote a
>> small Ada tool to parse the patch and determine the status of each
>> changelog entry. The output can be found here [1].
>>
>> If I'm not completely mistaken, we are interested in all changelogs in
>> status TBS (To_Be_Submitted). Would a verbose list of these changelogs help?
> 
> Sorry for the late reply.  I reviewed the changes in the .dif file and
> it seems very difficult to me to extract just the parts we need from
> this file without risking major breakage in the compiler.  So, I suggest
> we simply wait for AdaCore to submit this patch and, in the mean time,
> accept that GCC uses trampolines more often than GNAT GPL and therefore
> requires the stack to be executable.

I agree. The format of the .dif file is very unfortunate, it's almost
impossible to determine which parts of the patch belong to which
changelog entry.

- reto


Reply to: