[cfe-dev] [LLVMdev] Proposal to add Bitcode version field to bitcode file wrapper

Mon Oct 6 14:24:03 PDT 2014

One advantage to including the minor version number is that it allows the IR to evolve in more flexible ways (semantic difference without a syntactic difference) but auto-upgrade stays feasible. That is, it can help *avoid* breakages.
Another advantage would be for cases like debug-info metadata, where IIRC there's no backward compatibility at all; we just throw away old stuff. If we had a minor version in there, it becomes more reasonable to support one-minor-version backward compatibility.  (This is a use case that we would be interested in.)
--paulr

From: cfe-dev-bounces at cs.uiuc.edu [mailto:cfe-dev-bounces at cs.uiuc.edu] On Behalf Of Sean Silva
Sent: Monday, October 06, 2014 12:27 PM
To: Yung, Douglas
Cc: cfe-dev at cs.uiuc.edu; llvmdev at cs.uiuc.edu
Subject: Re: [cfe-dev] [LLVMdev] Proposal to add Bitcode version field to bitcode file wrapper

On Sun, Oct 5, 2014 at 8:10 PM, Yung, Douglas <douglas_yung at playstation.sony.com<mailto:douglas_yung at playstation.sony.com>> wrote:
Hi –

I realize the thread has drifted a little, but I wanted to get back to my original proposal. I would like to make a change to the bitcode file wrapper to include the version of llvm that produced the bitcode. I would like to write this version into the unused version field that currently exists. Would there be any objections to this change?

If the version field is currently unused, I don't see the harm in filling it in. It probably would be fine to just use it to store the LLVM major version, so that we can detect incompatibilities. I still would be opposed to using a granularity finer than the "intended" breakage cycle (major versions), since that would give the impression that it is "ok" to break compatibility since it can be detected through the version field.

-- Sean Silva

Since the original wrapper is only emitted for Darwin targets, I ran an experiment where I took bitcode files produced by the official Apple tools and then modified the version field to be non-zero. From my simple tests, there seemed to be no problems with existing tools when the version field is non-zero.

Douglas Yung

From: David Blaikie [mailto:dblaikie at gmail.com<mailto:dblaikie at gmail.com>]
Sent: Sunday, September 28, 2014 1:02
To: Alex Rosenberg
Cc: Greg Bedwell; Yung, Douglas; cfe-dev at cs.uiuc.edu<mailto:cfe-dev at cs.uiuc.edu>; llvmdev at cs.uiuc.edu<mailto:llvmdev at cs.uiuc.edu>
Subject: Re: [cfe-dev] [LLVMdev] Proposal to add Bitcode version field to bitcode file wrapper

On Sat, Sep 27, 2014 at 11:35 PM, Alex Rosenberg <alexr at leftfield.org<mailto:alexr at leftfield.org>> wrote:
How is this use case different from the LTO-supported toolchains shipped by other vendors such as Apple? Do they have this theoretical problem too?

If the issue is solely constrained to debug info metadata, then why not use metadata to describe the format/version of the debug info?

FWIW (I haven't followed the rest of this thread) - that's what we/Apple have done. There's a module flag metadata that specifies the debug info metadata schema version, then the verifier (or some other pass, I forget how it's phrased) can check if the version matches the current LLVM's debug info metadata version, and if it doesn't match, it strips out all the debug info metadata.

Alex

On Sep 27, 2014, at 3:19 AM, Greg Bedwell <gregbedwell at gmail.com<mailto:gregbedwell at gmail.com>> wrote:
As I understand it, the bitcode compatibility promise doesn't extend as far as debug info metadata (happy to be wrong here!).  I think we have a usecase where need to guarantee that debug information from any two arbitrary bitcode files going into an LTO link will result in the expected/correct debug information going into the resulting ELF file; unless we can be sure that this will always work between bitcode files generated by different versions we'd need some way of flagging up an incompatibility and providing useful information on the reason to the user.

--Greg Bedwell
SN Systems - Sony Computer Entertainment Group

On 27 September 2014 02:24, Sean Silva <chisophugis at gmail.com<mailto:chisophugis at gmail.com>> wrote:

On Fri, Sep 26, 2014 at 5:18 PM, Yung, Douglas <douglas_yung at playstation.sony.com<mailto:douglas_yung at playstation.sony.com>> wrote:
Sorry if I was unclear. There are currently no “known incompatibilities” that I am aware of, although I fully admit to not being an expert on the topic. The idea is that we add versioning information to the bitcode so that if an issue were discovered, it could be easily detected and dealt with.

It sounds like time would be better invested in improving the testing of our bitcode compatibility promise.

-- Sean Silva

Douglas Yung

From: Bob Wilson [mailto:bob.wilson at apple.com<mailto:bob.wilson at apple.com>]
Sent: Friday, September 26, 2014 16:39
To: Yung, Douglas
Cc: llvmdev at cs.uiuc.edu<mailto:llvmdev at cs.uiuc.edu>; cfe-dev at cs.uiuc.edu<mailto:cfe-dev at cs.uiuc.edu>
Subject: Re: [LLVMdev] Proposal to add Bitcode version field to bitcode file wrapper

Bitcode backward compatibility, at least for the current major version, is supposed to make this unnecessary. Can you provide more information about what “known incompatibilities” you’re seeing?

On Sep 26, 2014, at 2:40 PM, Yung, Douglas <douglas_yung at playstation.sony.com<mailto:douglas_yung at playstation.sony.com>> wrote:

Hi,

We would like to add a version number to the bitcode wrapper. This feature would allow easier identification of what compiler produced the bitcode so that known incompatibilities with LTO compilation could be detected. Roughly speaking, this version number would consist of the major, minor and optionally the patch version of the compiler used to produce the bitcode. The version information would be encoded in 4 bytes, with the first byte representing the major version number, the second byte the minor version number, and the third and fourth bytes optionally encoding the patch version or other information. As to where to place this information, we are considering two different possibilities for updating the bitcode wrapper specification.

The first is to simply add a single 32bit wide field at the end of the existing bitcode wrapper format field. This would result in the new structure looking like this:

[Magic_{32}, Version_{32}, Offset_{32}, Size_{32}, CPUType_{32}, BitcodeVersion_{32}]

All of the existing fields would keep their current meanings, and the new field BitcodeVersion is simply appended with the format described in the first paragraph.

A second idea was to use the existing Version field in the bitcode wrapper format to store the bitcode version information. According to the documentation (http://llvm.org/docs/BitCodeFormat.html#bitcode-wrapper-format) this field is currently always set to 0. This would allow us to make use of what is (presumably) an unused field.

As this is a feature that we feel would be beneficial to the community, we wanted to get feedback on the design for our upcoming patches. Any thoughts or opinions on this would be greatly appreciated.

Thanks!

Douglas Yung
_______________________________________________
LLVM Developers mailing list
LLVMdev at cs.uiuc.edu<mailto:LLVMdev at cs.uiuc.edu>         http://llvm.cs.uiuc.edu<http://llvm.cs.uiuc.edu/>
http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev

_______________________________________________
LLVM Developers mailing list
LLVMdev at cs.uiuc.edu<mailto:LLVMdev at cs.uiuc.edu>         http://llvm.cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev

_______________________________________________
cfe-dev mailing list
cfe-dev at cs.uiuc.edu<mailto:cfe-dev at cs.uiuc.edu>
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

_______________________________________________
LLVM Developers mailing list
LLVMdev at cs.uiuc.edu<mailto:LLVMdev at cs.uiuc.edu>         http://llvm.cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev

_______________________________________________
cfe-dev mailing list
cfe-dev at cs.uiuc.edu<mailto:cfe-dev at cs.uiuc.edu>
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

_______________________________________________
cfe-dev mailing list
cfe-dev at cs.uiuc.edu<mailto:cfe-dev at cs.uiuc.edu>
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20141006/bb76dd3e/attachment.html>