Rust violates GPL by deriving code from libgcc

[xerxes12354]

GPL code means proprietary projects cannot copy it. Rust violates GPL license by reusing GPL code in non-GPL environment.

rust/library/std/src/personality/gcc.rs

Line 80 in a0d0709

// https://github.com/gcc-mirror/gcc/blob/master/libstdc++-v3/libsupc++/eh_personality.cc

This is GPL license violation. Please fix.

[pnkfelix]

cc @Amanieu , as the author of that code.

Based on a cursory review, I think all of this falls under the category of the GCC Runtime Library Exception.

Namely, both of the files linked at those two urls (eh_personality.cc, unwind-c.c) say:

Under Section 7 of GPL version 3, you are granted additional
permissions described in the GCC Runtime Library Exception, version
3.1, as published by the Free Software Foundation.

The GCC Runtime Library Exception is provided here: https://www.gnu.org/licenses/gcc-exception-3.1.en.html

1. Grant of Additional Permission.

You have permission to propagate a work of Target Code formed by combining the Runtime Library with Independent Modules, even if such propagation would otherwise violate the terms of GPLv3, provided that all Target Code was generated by Eligible Compilation Processes. You may then convey such a combination under terms of your choice, consistent with the licensing of the Independent Modules.

A Compilation Process is "Eligible" if it is done using GCC, alone or with other GPL-compatible software, or if it is done without using any work based on GCC. For example, using non-GPL-compatible Software to optimize any GCC intermediate representations would not qualify as an Eligible Compilation Process.

My understanding is that our compilation process here is eligible because it is done without using any work based on GCC itself.

[mqudsi]

My read of the quoted comment is that what the code does is based off what the gcc code does (in lieu of a spec), not "this code is copied from gcc."

[Amanieu]

Our actual implementation is based on LLVM's libcxxabi:

• rust/library/std/src/personality/dwarf/eh.rs

  Line 144 in a0d0709

  unsafe fn read_encoded_pointer(

• https://github.com/llvm/llvm-project/blob/7973346d1672dda58c5ba33918ef38bf707daa42/libcxxabi/src/cxa_personality.cpp#L257

[Noratrieb]

Any updates here? Can it be closed or is there any action that needs to be taken?

[Amanieu]

I'm happy to close this, but I'll leave the final decision to @pnkfelix.

[pnkfelix]

So I tracked down where the comment that raised all this was inserted, in order to understand how we are in a situation where a comment says that some code is based on GCC routines, but nonetheless our current implementation is derived from LLVM according to @Amanieu

The problematic comment was inserted about seven years ago in: 051c2d1

rust/src/libpanic_unwind/gcc.rs

Lines 138 to 140 in 051c2d1

	// Based on GCC's C and C++ personality routines. For reference, see:
	// https://github.com/gcc-mirror/gcc/blob/master/libstdc++-v3/libsupc++/eh_personality.cc
	// https://github.com/gcc-mirror/gcc/blob/trunk/libgcc/unwind-c.c

and that comment was preserved over time as the code was copied from place to place, and has ended up in:

rust/library/std/src/personality/gcc.rs

Lines 80 to 82 in 439292b

	// The following code is based on GCC's C and C++ personality routines. For reference, see:
	// https://github.com/gcc-mirror/gcc/blob/master/libstdc++-v3/libsupc++/eh_personality.cc
	// https://github.com/gcc-mirror/gcc/blob/trunk/libgcc/unwind-c.c

But the thing I don't currently understand is that @Amanieu pointed at a totally different file,

rust/library/std/src/personality/dwarf/eh.rs

Lines 144 to 146 in a0d0709

	unsafe fn read_encoded_pointer(
	reader: &mut DwarfReader,
	context: &EHContext<'_>,

a file which was effectively introduced eight years ago in 6112b22 and is dedicated to parsing DWARF encoded values, but that is ... not related to gcc eh personalities?

• (Update: the two may be somewhat related, since unwinding in some cases may end up making us of parsing DWARF encoded values. But I stand by my claim that these two bits of code should be considered independently from one another, in terms of how they were derived.)

I'm not yet comfortable closing this issue, I don't think I have a clear enough picture of where the gcc.rs personality code came from.

I do see gcc-compatible personality code in the LLVM source tree, or at least evidence of such being added, see e.g. llvm/llvm-project@64142ba, but that post-dates the commits we are talking about...

[pnkfelix]

Actually, looking at 051c2d1 more carefully, I might be convinced that this code was derived from LLVM rather than GCC.

Maybe my mistake was tagging @Amanieu as the author here, when the relevant person to ping is probably @vadimcn as the author of 051c2d1 ?

[pnkfelix]

actually, took yet another look at 051c2d1 and then decided to look at the linked code in gcc.

I'm making the call that it is "obvious" that the Rust implementation is not derived from the GCC one. The two look nothing alike to me. (I say this as someone who used to grade undergraduate homework assignments; I am pretty good at identifying obfuscated similarities between common code bases.)

AFAICT, the reason that the GCC code was referenced there was to explain where this eh personality interface comes from, because we originally just had it as a shallow wrapper around __gcc_personality_sj0, as you can tell from looking at 051c2d1

So, closing as not a copyright violation.

[pnkfelix]

Well, I'm going to reopen because I want the comment in our source code to be revised in some way so that we don't get issues like this filed in the future due to those urls in that comment.