llvm dev - Aug 2020 - LLD: Can we make --warn-backrefs the default?

On Fri, Aug 28, 2020 at 11:16 AM David Blaikie <dblaikie at gmail.com>
wrote:
>
> Would you like to conduct the conversation here, or on the review thread?
(I lean towards having them here, but don't mind if folks feel like it keeps
the noise down & want to more post a notice saying "hey, here's
this thing, if you're interested, go discuss it over there" - more an
optional opt-in rather than requiring people to opt-out via muting the thread,
etc)
Yes, we can conduct the "should we enable --warn-backrefs by default"
conversation here. Since the semantics --warn-backrefs of are a bit
complex, we need a documentation. https://reviews.llvm.org/D86762 is
put up to get wording suggestions. Explicitly adding the people to the
CC list...

FWIW for many code bases, --warn-backrefs should produce no warnings
(error if --fatal-warnings). For some code bases, GNU ld may error
"undefined reference".  --warn-backrefs can catch such problems.

> On Thu, Aug 27, 2020 at 10:15 PM Fangrui Song via llvm-dev <llvm-dev at
lists.llvm.org> wrote:
>>
>> Hi all, LLD's --warn-backrefs is a tool to identify potential
>> incompatible archive selection semantics with traditional Unix linkers.
>> I have improved it (via D77522,D77630 and D77512) to a state where a
>> --warn-backrefs diagnostic almost assuredly means that the link will
>> fail with GNU ld, or the symbol will get different resolution in GNU ld
and LLD.
>>
>> My conclusion is that --warn-backrefs is a very useful layering check
tool.
>> I just wrote a documentation about the advantage (of GNU ld's
archive
>> selection semantics..... But we can do better with --warn-backrefs!
>> GNU ld just reports "undefined reference" with no actionable
feedback
>> about the offending archive)
>>
>> https://reviews.llvm.org/D86762
>>
>> I am wondering whether in the next release we can make --warn-backrefs
>> the default.  I have added many known users to the review.
>> (There is no need for --no-warn-backrefs because
--warn-backrefs-exclude='*' does the same job)
>> _______________________________________________
>> LLVM Developers mailing list
>> llvm-dev at lists.llvm.org
>> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev


-- 
宋方睿

On Mon, Aug 31, 2020 at 1:24 PM Fāng-ruì Sòng <maskray at google.com>
wrote:
> On Fri, Aug 28, 2020 at 11:16 AM David Blaikie <dblaikie at
gmail.com> wrote:
> >
> > Would you like to conduct the conversation here, or on the review
> thread? (I lean towards having them here, but don't mind if folks feel
like
> it keeps the noise down & want to more post a notice saying "hey,
here's
> this thing, if you're interested, go discuss it over there" - more
an
> optional opt-in rather than requiring people to opt-out via muting the
> thread, etc)
>
> Yes, we can conduct the "should we enable --warn-backrefs by
default"
> conversation here. Since the semantics --warn-backrefs of are a bit
> complex, we need a documentation. https://reviews.llvm.org/D86762 is
> put up to get wording suggestions. Explicitly adding the people to the
> CC list...
>
> FWIW for many code bases, --warn-backrefs should produce no warnings
> (error if --fatal-warnings). For some code bases, GNU ld may error
> "undefined reference".  --warn-backrefs can catch such problems.
>
One of the questions raised on the thread there was about different linker
semantics. I assume the "--warn-backrefs by default" we're
discussing is
only related to the ld.lld frontend? Not the Windows linker lld behavior
(or ld64 (old or new) lld behavior)?

>
>
> > On Thu, Aug 27, 2020 at 10:15 PM Fangrui Song via llvm-dev <
> llvm-dev at lists.llvm.org> wrote:
> >>
> >> Hi all, LLD's --warn-backrefs is a tool to identify potential
> >> incompatible archive selection semantics with traditional Unix
linkers.
> >> I have improved it (via D77522,D77630 and D77512) to a state where
a
> >> --warn-backrefs diagnostic almost assuredly means that the link
will
> >> fail with GNU ld, or the symbol will get different resolution in
GNU ld
> and LLD.
> >>
> >> My conclusion is that --warn-backrefs is a very useful layering
check
> tool.
> >> I just wrote a documentation about the advantage (of GNU ld's
archive
> >> selection semantics..... But we can do better with
--warn-backrefs!
> >> GNU ld just reports "undefined reference" with no
actionable feedback
> >> about the offending archive)
> >>
> >> https://reviews.llvm.org/D86762
> >>
> >> I am wondering whether in the next release we can make
--warn-backrefs
> >> the default.  I have added many known users to the review.
> >> (There is no need for --no-warn-backrefs because
> --warn-backrefs-exclude='*' does the same job)
> >> _______________________________________________
> >> LLVM Developers mailing list
> >> llvm-dev at lists.llvm.org
> >> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>
>
>
> --
> 宋方睿
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200831/dfd4564a/attachment.html>

On Mon, Aug 31, 2020 at 1:29 PM David Blaikie <dblaikie at gmail.com>
wrote:
>
>
>
> On Mon, Aug 31, 2020 at 1:24 PM Fāng-ruì Sòng <maskray at google.com>
wrote:
>>
>> On Fri, Aug 28, 2020 at 11:16 AM David Blaikie <dblaikie at
gmail.com> wrote:
>> >
>> > Would you like to conduct the conversation here, or on the review
thread? (I lean towards having them here, but don't mind if folks feel like
it keeps the noise down & want to more post a notice saying "hey,
here's this thing, if you're interested, go discuss it over there"
- more an optional opt-in rather than requiring people to opt-out via muting the
thread, etc)
>>
>> Yes, we can conduct the "should we enable --warn-backrefs by
default"
>> conversation here. Since the semantics --warn-backrefs of are a bit
>> complex, we need a documentation. https://reviews.llvm.org/D86762 is
>> put up to get wording suggestions. Explicitly adding the people to the
>> CC list...
>>
>> FWIW for many code bases, --warn-backrefs should produce no warnings
>> (error if --fatal-warnings). For some code bases, GNU ld may error
>> "undefined reference".  --warn-backrefs can catch such
problems.
>
>
> One of the questions raised on the thread there was about different linker
semantics. I assume the "--warn-backrefs by default" we're
discussing is only related to the ld.lld frontend? Not the Windows linker lld
behavior (or ld64 (old or new) lld behavior)?
Yes, the ELF port (lld/ELF). No other port has implemented
--warn-backrefs. The intion of --warn-backrefs is to capture behavior
differences with GNU ld's ELF ports. If traditional ELF linkers don't
behave like that, I can replace "traditional ELF linkers" in
https://reviews.llvm.org/D86762 with "GNU linkers".
>>
>>
>>
>> > On Thu, Aug 27, 2020 at 10:15 PM Fangrui Song via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
>> >>
>> >> Hi all, LLD's --warn-backrefs is a tool to identify
potential
>> >> incompatible archive selection semantics with traditional Unix
linkers.
>> >> I have improved it (via D77522,D77630 and D77512) to a state
where a
>> >> --warn-backrefs diagnostic almost assuredly means that the
link will
>> >> fail with GNU ld, or the symbol will get different resolution
in GNU ld and LLD.
>> >>
>> >> My conclusion is that --warn-backrefs is a very useful
layering check tool.
>> >> I just wrote a documentation about the advantage (of GNU
ld's archive
>> >> selection semantics..... But we can do better with
--warn-backrefs!
>> >> GNU ld just reports "undefined reference" with no
actionable feedback
>> >> about the offending archive)
>> >>
>> >> https://reviews.llvm.org/D86762
>> >>
>> >> I am wondering whether in the next release we can make
--warn-backrefs
>> >> the default.  I have added many known users to the review.
>> >> (There is no need for --no-warn-backrefs because
--warn-backrefs-exclude='*' does the same job)
>> >> _______________________________________________
>> >> LLVM Developers mailing list
>> >> llvm-dev at lists.llvm.org
>> >> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev