[llvm-dev] a lld linker script bug

[Davide Italiano via llvm-dev]

On Sat, Aug 22, 2015 at 3:13 AM, zan jyu Wong via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
> Hi,
>
> I've wrote a test, which is modified from
> lld/test/elf/linkerscript/sections-with-wildcards.tes.
>
> Index: tools/lld/test/elf/linkerscript/filename-with-wildcards.test
> <+>UTF-8
> ===================================================================
> --- tools/lld/test/elf/linkerscript/filename-with-wildcards.test
> (revision )
> +++ tools/lld/test/elf/linkerscript/filename-with-wildcards.test
> (revision )
> @@ -0,0 +1,49 @@
> +/*
> +Tests a linker script that uses the SECTIONS command with rules containing
> +wildcards that matching input object files.
> +*/
> +
> +ENTRY(_start)
> +
> +SECTIONS
> +{
> +  . = 0x500000;
> +  .foo : { *p1.o(.text .rodata*) }
> +  .bar : { *(.text .rodata*) }
> +}
> +
> +/*
> +RUN: mkdir -p %T
> +RUN: yaml2obj -format=elf %p/Inputs/prog1.o.yaml -o=%T/p1.o
> +RUN: yaml2obj -format=elf %p/Inputs/prog2.o.yaml -o=%T/p2.o
> +RUN: yaml2obj -format=elf %p/Inputs/prog3.o.yaml -o=%T/p3.o
> +RUN: cd %T
> +
> +RUN: lld -flavor gnu  -target x86_64 -T %s p1.o p2.o p3.o \
> +RUN:   -static -o %t1
> +RUN: llvm-readobj -s %t1 | FileCheck -check-prefix CHECKSECTIONS %s
> +
> +CHECKSECTIONS:       Index: 1
> +CHECKSECTIONS:       Name: .foo
> +CHECKSECTIONS:       Address: 0x500000
> +CHECKSECTIONS:       Size: 33
> +
> +CHECKSECTIONS:       Index: 2
> +CHECKSECTIONS:       Name: .bar
> +CHECKSECTIONS:       Address: 0x500030
> +CHECKSECTIONS:       Size: 52
> +
> +RUN: llvm-readobj -symbols %t1 | FileCheck -check-prefix CHECKSYMS %s
> +
> +CHECKSYMS:      Name: main
> +CHECKSYMS-NEXT: Value: 0x500000
> +
> +CHECKSYMS:      Name: prog2
> +CHECKSYMS-NEXT: Value: 0x500030
> +
> +CHECKSYMS:      Name: write
> +CHECKSYMS-NEXT: Value: 0x500040
> +
> +CHECKSYMS:      Name: _start
> +CHECKSYMS-NEXT: Value: 0x500048
> +*/
>
> Cheers,
> Huang
>
> On Sat, Aug 22, 2015 at 8:17 AM, Rafael Auler <rafaelauler at gmail.com> wrote:
>>
>> Hi,
>>
>> Thanks for your patch, Huang. It looks good to me. Just one comment:
>> can you write a testcase, similar to the others used for linker script
>> testing, with your example? Alternatively, you can modify
>> lld/test/elf/linkerscript/sections-with-wildcards.test to test your
>> case. This will make your patch complete and ready for commit, and
>> will ensure we do not regress on this bug if this code is ever
>> rewritten.
>>
>> Best regards,
>> Rafael Auler
>>
>> On Fri, Aug 21, 2015 at 8:48 PM, Davide Italiano <davide at freebsd.org>
>> wrote:
>> > On Fri, Aug 21, 2015 at 4:41 AM, zan jyu Wong via llvm-dev
>> > <llvm-dev at lists.llvm.org> wrote:
>> >> Hi,
>> >>
>> >> I've found a crash in lld when using linker script with wildcard
>> >> matching.
>> >> An example linker script:
>> >>
>> >> INPUT(os/main.o os/foo.o os/startup.o)
>> >> OUTPUT(k.bin)
>> >>
>> >> SECTIONS
>> >> {
>> >>     . = 0x0
>> >>     .text : { *startup.o (.text) }
>> >>     .text.2 : { *(.tex*) }
>> >> }
>> >>
>> >> I've wrote up a patch to fix this crash.
>> >>
>> >
>> > The patch looks good to me.
>> > CC:ing Rafael and Michael as they wrote large part of the linker script
>> > support.
>> >
>> >
>> >> Index: tools/lld/lib/ReaderWriter/LinkerScript.cpp
>> >> <+>UTF-8
>> >> ===================================================================
>> >> --- tools/lld/lib/ReaderWriter/LinkerScript.cpp    (revision
>> >> 8570c61c3fce7de2f655a20f8b184efa1bd97c00)
>> >> +++ tools/lld/lib/ReaderWriter/LinkerScript.cpp    (revision )
>> >> @@ -2557,7 +2557,7 @@
>> >>      switch (*j) {
>> >>      case '*':
>> >>        while (!wildcardMatch(pattern.drop_front(j - pattern.begin() +
>> >> 1),
>> >> -                            name.drop_front(i - name.begin() + 1))) {
>> >> +                            name.drop_front(i - name.begin()))) {
>> >>          if (i == name.end())
>> >>            return false;
>> >>          ++i;
>> >> @@ -2565,6 +2565,7 @@
>> >>        break;
>> >>      case '?':
>> >>        // Matches any character
>> >> +      ++i;
>> >>        break;
>> >>      case '[': {
>> >>        // Matches a range of characters specified between brackets
>> >> @@ -2577,20 +2578,22 @@
>> >>          return false;
>> >>
>> >>        j = pattern.begin() + end;
>> >> +      ++i;
>> >>        break;
>> >>      }
>> >>      case '\\':
>> >>        ++j;
>> >>        if (*j != *i)
>> >>          return false;
>> >> +      ++i;
>> >>        break;
>> >>      default:
>> >>        // No wildcard character means we must match exactly the same
>> >> char
>> >>        if (*j != *i)
>> >>          return false;
>> >> +      ++i;
>> >>        break;
>> >>      }
>> >> -    ++i;
>> >>    }
>> >>
>> >>    // If our pattern has't consumed the entire string, it is not a
>> >> match
>> >>
>> >> Cheers,
>> >> Huang
>> >>
>> >> _______________________________________________
>> >> LLVM Developers mailing list
>> >> llvm-dev at lists.llvm.org
>> >> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>> >>
>> >
>> >
>> >
>> > --
>> > Davide
>> >
>> > "There are no solved problems; there are only problems that are more
>> > or less solved" -- Henri Poincare
>
>
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>

I committed the patch and the test as r245792.
Thank you for your contribution.

-- 
Davide

"There are no solved problems; there are only problems that are more
or less solved" -- Henri Poincare