[PATCH][gdb/testsuite] Fix .debug_aranges in watch-loc.c

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

[PATCH][gdb/testsuite] Fix .debug_aranges in watch-loc.c

Tom de Vries
Hi,

When running gdb.dlang/watch-loc.exp with target board cc-with-debug-names, we
run into:
...
FAIL: gdb.dlang/watch-loc.exp: disassemble _Dmain (GDB internal error)
...
in more detail:
...
(gdb) disassemble _Dmain
gdb/dwarf2read.c:5293: internal-error: \
  compunit_symtab* dw2_find_pc_sect_compunit_symtab(objfile*, \
  bound_minimal_symbol, CORE_ADDR, obj_section*, int): \
  Assertion `result != NULL' failed.
...

The problem is that the .debug_aranges section in watch-loc.c contains a
debug_info_offset which is set to 0:
...
asm (
  "      .pushsection    .debug_aranges,\"\",%progbits\n"
  "      .4byte  .Laranges_end - .Laranges_start\n"
  ".Laranges_start:\n"
  "      .2byte  0x2\n"
  "      .4byte  0\n"
...
while the compilation unit at offset 0 in the .debug_section in the executable
is in fact not the compilation unit generated from watch-loc-dw.S.

[ Note: this is a non-trivial test-case.  The file watch-loc-dw.S contains a
.debug_info section, but not an .debug_aranges section or any actual code.
The file watch-loc.c contains code and a .debug_aranges section, but no other
debug section.  So, the intent for the .debug_aranges section in watch-loc.c
is to refer to a compilation unit in the .debug_info section in
watch-loc-dw.S. ]

This happens when linked in object files contain dwarf info and are placed in
the .debug_info section before the compilation units generated from
watch-loc.c and watch-loc-dw.S.

Fix this by defining the debug_info_offset field using a label .Lcu1_begin
that defines the start of an empty .debug_section compilation unit:
...
 asm (
+  "      .pushsection    .debug_info,\"\",%progbits\n"
+  ".Lcu1_begin:"
+  "      .popsection\n"
   "      .pushsection    .debug_aranges,\"\",%progbits\n"
   "      .4byte  .Laranges_end - .Laranges_start \n"
   ".Laranges_start:\n"
   "      .2byte  0x2\n"
-  "      .4byte  0\n"
+  "      .4byte  .Lcu1_begin\n"
...
which during linking merges with the start of the .debug_info section of
watch-loc-dw.S.

Tested on x86_64-linux with native, cc-with-gdb-index and cc-with-debug-names.

OK for trunk?

Thanks,
- Tom

[gdb/testsuite] Fix .debug_aranges in watch-loc.c

gdb/testsuite/ChangeLog:

2019-05-06  Tom de Vries  <[hidden email]>

        PR testsuite/24522
        * gdb.dlang/watch-loc.c: Fix debug_info_offset in .debug_aranges
        section.

---
 gdb/testsuite/gdb.dlang/watch-loc.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/gdb/testsuite/gdb.dlang/watch-loc.c b/gdb/testsuite/gdb.dlang/watch-loc.c
index b0af2fcc67..5dd05b4dfe 100644
--- a/gdb/testsuite/gdb.dlang/watch-loc.c
+++ b/gdb/testsuite/gdb.dlang/watch-loc.c
@@ -38,11 +38,14 @@ main (void)
    generated by GCC.  (.gdb_index includes a gdb-generated map
    instead.)  */
 asm (
+" .pushsection .debug_info,\"\",%progbits \n"
+".Lcu1_begin:"
+" .popsection \n"
 " .pushsection .debug_aranges,\"\",%progbits \n"
 " .4byte .Laranges_end - .Laranges_start \n" // Length of Address Ranges Info
 ".Laranges_start: \n"
 " .2byte 0x2 \n" // DWARF Version
-" .4byte 0 \n" // .Ldebug_info0 - Offset of Compilation Unit Info
+" .4byte .Lcu1_begin\n" // .Ldebug_info0 - Offset of Compilation Unit Info
 " .byte 4 \n" // Size of Address
 " .byte 0 \n" // Size of Segment Descriptor
 " .2byte 0 \n" // Pad to 16 byte boundary
Reply | Threaded
Open this post in threaded view
|

Re: [PATCH][gdb/testsuite] Fix .debug_aranges in watch-loc.c

Jan Kratochvil-2
On Mon, 06 May 2019 18:37:40 +0200, Tom de Vries wrote:
> OK for trunk?

I do not have it reproducible on Fedora 29 x86_64 but I agree it is correct.
I do not have GDB approval permission.


Jan
Reply | Threaded
Open this post in threaded view
|

Re: [PATCH][gdb/testsuite] Fix .debug_aranges in watch-loc.c

Tom de Vries
On 06-05-19 19:21, Jan Kratochvil wrote:
> On Mon, 06 May 2019 18:37:40 +0200, Tom de Vries wrote:
>> OK for trunk?
>
> I do not have it reproducible on Fedora 29 x86_64 but I agree it is correct.
> I do not have GDB approval permission.

Thanks for the review.

With your review done, I consider the patch obvious and I've committed it.

Thanks,
- Tom