[fpc-devel] Successful implementation of inline support for pure assembler routines on x86

[J. Gareth Moreton]

 I would have allowed writing to the stack with inline assembly functions
if a way could be found to ensure that the compiled program behaves the
same way whether the function is inlined or called directly (e.g. through a
direct CALL or a function pointer), which would otherwise require shifting
byte offsets depending on if the function return address was pushed or not,
so I decided to just forbid it completely.

 If you want to inline regular functions with assembler blocks, you'll need
the attached bug fixes.  The rest I'll leave up to you.
 What is Florian's and your vision for Free Pascal? You already have a
cross-platform Object Pascal compiler... what's next for it?  Does any of
my proposals even have a place in that vision, because honestly I don't
know if the following is wanted:

 - pure functions
 - XML node dump (compiler debugging feature) - I personally need it for
pure functions and for fixing #32913.
 - faster and smarter optimization on x86-64

 Gareth aka. Kit

 On Tue 19/03/19 08:04 , Jonas Maebe jonas at freepascal.org sent:
 On 2019-03-18 23:41, J. Gareth Moreton wrote: 

 > I guess there's no place for me here. I don't have the facilities or 
 > knowledge to develop beyond i386 and x86_64, and there's too much 
 > conflict. You win. What am I but a rowdy tyke who doesn't listen to 
 > his elders and betters? 

 I don't consider myself to be better than you, and you giving up is not 
 me "winning". You also don't need to develop beyond i386/x86-64, but you 
 do have to take into account that they exist. This means that unless a 
 feature is inherently specific to a particular platform, it has to be 
 generalisable and can be made to function in a predictable way 
 everywhere (within the bounds of defined behaviour; what that is, is 
 defined by the implementer of the functionality, but it should be 
 possible to specify it and preferably with not too many rules, so it can 
 be understood easily by users and other platform maintainers alike). 

 That is not possible for parsing and trying to understand inline 
 assembly using automated means. It is very fragile, which makes 
 specifying the defined behaviour very hard. It needs a ton of 
 exceptions, many of which will be platform specific, which makes its 
 functionality unpredictable. 

 > I guess I don't understand the philosophy of FPC after all, or can even 
 > program well. 

 This is not about your programming skills, those are quite good. It is, 
 however, indeed very much about FPC's philosophy. That said, this 
 particular feature is something that at least most other compilers (even 
 for other languages) don't implement either, afaik. Most of them even 
 don't support defining pure assembler functions inside the host 
 language's source file, and require you to use external assembler files. 
 They do allow you to inline regular functions that contain inline 
 assembly though. 

 Jonas 
 _______________________________________________ 
 fpc-devel maillist - fpc-devel at lists.freepascal.org [1] 
 http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-devel
[2]">http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-devel 

 

Links:
------
[1] mailto:fpc-devel at lists.freepascal.org
[2] http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-devel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freepascal.org/pipermail/fpc-devel/attachments/20190319/09281043/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: x86_asm_node_fixes.patch
Type: application/octet-stream
Size: 1760 bytes
Desc: not available
URL: <http://lists.freepascal.org/pipermail/fpc-devel/attachments/20190319/09281043/attachment.obj>