Re: PR24511, nm should not mark symbols in .init_array as "t"

[KONRAD Frederic <frederic.konrad@adacore.com>]

Hi Alan,

It seems that this patch has the side effect of marking the symbols in .idata as
"D" instead of "I" previously.  Is that expected?

Thanks,
Fred

Le 5/4/19 à 9:57 AM, Alan Modra a écrit :
> This patch restricts the section names matched in coff_section_type,
> a function that translates section names to symbol type, and arranges
> to translate section flags to symbol type before looking at names.
> The latter change resulted in various test failures due to improper
> section flags being used in tests, and by the plugin support, so fix
> that too.
> 
> The new test fails on many ELF targets that lack .init/fini_array
> in their scripts.  I've just xfailed those.  pru-elf oddly defines
> __init_array_begin rather than __init_array_start.  I've left that
> target as a FAIL, and pj-elf too which reports an error for undefined
> weak symbols.
> 
> bfd/
> 	PR 24511
> 	* syms.c (coff_section_type): Only allow '.', '$' and numeric
> 	following the standard section names.
> 	(bfd_decode_symclass): Prioritize section flag tests in
> 	decode_section_type before name tests in coff_section_type.
> 	* plugin.c (bfd_plugin_canonicalize_symtab): Init fake_section
> 	and fake_common_section using BFD_FAKE_SECTION.  Use "fake" as
> 	their names and choose standard .text section flags for
> 	fake_section.
> ld/
> 	PR 24511
> 	* testsuite/ld-elf/pr14156a.d: Allow for .init/.fini being a
> 	data section on hppa64.
> 	* testsuite/ld-elf/pr14156b.d: Likewise.
> 	* testsuite/ld-scripts/pr18963.t: Map standard sections to set
> 	output section flags.
> 	* testsuite/ld-scripts/sane1.t: Likewise.
> 	* testsuite/ld-elf/init-fini-arrays.s: Reference __init_array_start
> 	and __fini_array_start.  Define __start et al.
> 	* testsuite/ld-elf/pr24511.d: New test.
> 
> diff --git a/bfd/plugin.c b/bfd/plugin.c
> index 8cb44ceb5d..376e92cdb3 100644
> --- a/bfd/plugin.c
> +++ b/bfd/plugin.c
> @@ -530,13 +530,13 @@ bfd_plugin_canonicalize_symtab (bfd *abfd,
>     struct plugin_data_struct *plugin_data = abfd->tdata.plugin_data;
>     long nsyms = plugin_data->nsyms;
>     const struct ld_plugin_symbol *syms = plugin_data->syms;
> -  static asection fake_section;
> -  static asection fake_common_section;
> +  static asection fake_section
> +    = BFD_FAKE_SECTION (fake_section, NULL, "plug", 0,
> +			SEC_ALLOC | SEC_LOAD | SEC_CODE | SEC_HAS_CONTENTS);
> +  static asection fake_common_section
> +    = BFD_FAKE_SECTION (fake_common_section, NULL, "plug", 0, SEC_IS_COMMON);
>     int i;
>   
> -  fake_section.name = ".text";
> -  fake_common_section.flags = SEC_IS_COMMON;
> -
>     for (i = 0; i < nsyms; i++)
>       {
>         asymbol *s = bfd_alloc (abfd, sizeof (asymbol));
> diff --git a/bfd/syms.c b/bfd/syms.c
> index fe7e7dfac8..da1c90d52e 100644
> --- a/bfd/syms.c
> +++ b/bfd/syms.c
> @@ -595,8 +595,9 @@ static const struct section_to_type stt[] =
>   /* Return the single-character symbol type corresponding to
>      section S, or '?' for an unknown COFF section.
>   
> -   Check for any leading string which matches, so .text5 returns
> -   't' as well as .text */
> +   Check for leading strings which match, followed by a number, '.',
> +   or '$' so .text5 matches the .text entry, but .init_array doesn't
> +   match the .init entry.  */
>   
>   static char
>   coff_section_type (const char *s)
> @@ -604,8 +605,12 @@ coff_section_type (const char *s)
>     const struct section_to_type *t;
>   
>     for (t = &stt[0]; t->section; t++)
> -    if (!strncmp (s, t->section, strlen (t->section)))
> -      return t->type;
> +    {
> +      size_t len = strlen (t->section);
> +      if (strncmp (s, t->section, len) == 0
> +	  && memchr (".$0123456789", s[len], 13) != 0)
> +	return t->type;
> +    }
>   
>     return '?';
>   }
> @@ -700,9 +705,9 @@ bfd_decode_symclass (asymbol *symbol)
>       c = 'a';
>     else if (symbol->section)
>       {
> -      c = coff_section_type (symbol->section->name);
> +      c = decode_section_type (symbol->section);
>         if (c == '?')
> -	c = decode_section_type (symbol->section);
> +	c = coff_section_type (symbol->section->name);
>       }
>     else
>       return '?';
> diff --git a/ld/testsuite/ld-elf/init-fini-arrays.s b/ld/testsuite/ld-elf/init-fini-arrays.s
> index 6740ed6793..b8adc29a3c 100644
> --- a/ld/testsuite/ld-elf/init-fini-arrays.s
> +++ b/ld/testsuite/ld-elf/init-fini-arrays.s
> @@ -1,7 +1,20 @@
>    .section .init_array.01000,"aw",%init_array
>    .p2align 2
> - .word 0
> + .weak __init_array_start, ___init_array_start
> + .dc.a __init_array_start
> + .dc.a ___init_array_start
>   
>    .section .fini_array.01000,"aw",%fini_array
>    .p2align 2
> - .word 0
> + .weak __fini_array_start, ___fini_array_start
> + .dc.a __fini_array_start
> + .dc.a ___fini_array_start
> +
> + .text
> + .globl main, _main, start, _start, __start
> +main:
> +_main:
> +start:
> +_start:
> +__start:
> + .dc.a 0
> diff --git a/ld/testsuite/ld-elf/pr14156a.d b/ld/testsuite/ld-elf/pr14156a.d
> index cf38ee1016..535ac3e159 100644
> --- a/ld/testsuite/ld-elf/pr14156a.d
> +++ b/ld/testsuite/ld-elf/pr14156a.d
> @@ -7,10 +7,10 @@
>   #nm: -n
>   
>   #...
> -[0-9a-f]+ T foo
> -[0-9a-f]+ t foo1
> +[0-9a-f]+ [TD] foo
> +[0-9a-f]+ [td] foo1
>   #...
> -[0-9a-f]+ t foo2
> -[0-9a-f]+ t foo3
> -[0-9a-f]+ t last
> +[0-9a-f]+ [td] foo2
> +[0-9a-f]+ [td] foo3
> +[0-9a-f]+ [td] last
>   #pass
> diff --git a/ld/testsuite/ld-elf/pr14156b.d b/ld/testsuite/ld-elf/pr14156b.d
> index f965f74e6e..27da0166f4 100644
> --- a/ld/testsuite/ld-elf/pr14156b.d
> +++ b/ld/testsuite/ld-elf/pr14156b.d
> @@ -7,10 +7,10 @@
>   #nm: -n
>   
>   #...
> -[0-9a-f]+ T foo
> -[0-9a-f]+ t foo1
> +[0-9a-f]+ [TD] foo
> +[0-9a-f]+ [td] foo1
>   #...
> -[0-9a-f]+ t foo2
> -[0-9a-f]+ t foo3
> -[0-9a-f]+ t last
> +[0-9a-f]+ [td] foo2
> +[0-9a-f]+ [td] foo3
> +[0-9a-f]+ [td] last
>   #pass
> diff --git a/ld/testsuite/ld-elf/pr24511.d b/ld/testsuite/ld-elf/pr24511.d
> new file mode 100644
> index 0000000000..f77a43b9e2
> --- /dev/null
> +++ b/ld/testsuite/ld-elf/pr24511.d
> @@ -0,0 +1,18 @@
> +#source: init-fini-arrays.s
> +#ld:
> +#nm: -n
> +# Most targets with their own scripts don't support init/fini_array and
> +# thus don't define __init/fini_array_start.
> +#xfail: avr-*-* cr16-*-* crx-*-* d10v-*-* d30v-*-* dlx-*-* ft32-*-* iq2000-*-*
> +#xfail: m68hc1*-*-* mep-*-* microblaze*-*-elf* s12z-*-* v850-*-* visium-*-*
> +#xfail: xgate-*-* xstormy*-*-*
> +# Some targets with their own scripts haven't kept up with elf.sc and
> +# PROVIDE __init_array_start rather than using PROVIDE_HIDDEN.  These
> +# result in D symbols.  rx-elf makes .init/fini_array SHF_EXECINSTR so
> +# gets t symbols.
> +
> +#...
> +[0-9a-f]+ [dDt] _?__init_array_start
> +#...
> +[0-9a-f]+ [dDt] _?__fini_array_start
> +#pass
> diff --git a/ld/testsuite/ld-scripts/pr18963.t b/ld/testsuite/ld-scripts/pr18963.t
> index b0cd7421eb..830ded78dd 100644
> --- a/ld/testsuite/ld-scripts/pr18963.t
> +++ b/ld/testsuite/ld-scripts/pr18963.t
> @@ -5,16 +5,19 @@ SECTIONS
>     .text :
>     {
>       _start = .;
> +    *(.text)
>       . = 0x10000;
>     }
>     B = .;
>     .data :
>     {
> +    *(.data)
>       . = 0x10000;
>     }
>     C = .;
>     .bss :
>     {
> +    *(.bss)
>       . = 0x10000;
>     }
>     D = A - C + B;
> diff --git a/ld/testsuite/ld-scripts/sane1.t b/ld/testsuite/ld-scripts/sane1.t
> index 037a62c856..90ee9b62d0 100644
> --- a/ld/testsuite/ld-scripts/sane1.t
> +++ b/ld/testsuite/ld-scripts/sane1.t
> @@ -20,6 +20,7 @@ SECTIONS
>       s4 = ABSOLUTE (d1) - 2;
>       s5 = ABSOLUTE (d2) % 5;
>       s6 = ABSOLUTE (d2) / 5;
> +    *(.data)
>     }
>     /DISCARD/ : {*(*)}
>   
>