Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
* [patch] dw2-ranges.S non-g GAS crash  [Re: spu-as internal error]
       [not found] ` <20080407040340.GF1446@bubble.grove.modra.org>
@ 2008-04-07 11:47   ` Jan Kratochvil
  2008-04-08  5:42     ` Ulrich Weigand
  2008-04-14 19:08     ` Daniel Jacobowitz
  0 siblings, 2 replies; 4+ messages in thread
From: Jan Kratochvil @ 2008-04-07 11:47 UTC (permalink / raw)
  To: Alan Modra; +Cc: Ulrich.Weigand, gdb-patches

[-- Attachment #1: Type: text/plain, Size: 853 bytes --]

On Mon, 07 Apr 2008 06:03:40 +0200, Alan Modra wrote:
> Ulrich Weigand wrote:
> > building the file gdb.dwarf2/dw2-ranges.S from the GDB test suite results
> > in an internal error in spu-as (current SDK CVS version):
> > 
> > /home/uweigand/fsf/gdb-head/gdb/testsuite/gdb.dwarf2/dw2-ranges.S:40:
> > Internal error, aborting at ../../binutils-2.18.50/gas/dwarf2dbg.c line
> > 1675 in out_debug_info
> > Please report this bug.
> 
> Has this testcase ever assembled?  On any target?  I think gas would
> always have aborted unless you pass -g to gas or supply .file
> directives.

It was always built just inside `runtest'.

Requesting a commit approval.  Verified unpatched GDB still FAILs on it.


(Still aware / going to post the fix to the STABS regression by this patch:
	http://sourceware.org/ml/gdb-patches/2007-12/msg00248.html
)


Regards,
Jan

[-- Attachment #2: dw2ranges-file.patch --]
[-- Type: text/plain, Size: 645 bytes --]

2008-04-07  Jan Kratochvil  <jan.kratochvil@redhat.com>

	* gdb.dwarf2/dw2-ranges.S: New stub `.file' directive.

--- ./gdb/testsuite/gdb.dwarf2/dw2-ranges.S	1 Jan 2008 22:53:19 -0000	1.2
+++ ./gdb/testsuite/gdb.dwarf2/dw2-ranges.S	7 Apr 2008 06:24:13 -0000
@@ -19,6 +19,9 @@
    DW_AT_ranges as the linker could place both sections at arbitrary locations.
    */
 
+	/* Such directive is required by GAS for builds without `-g'.  */
+	.file	1 "dw2-ranges.S"
+
 	/* Without this directive GAS will not emit DWARF2 unless we provide an
 	   instruction to assemble.  We want to avoid any instructions to
 	   remain architecture independent.  */

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [patch] dw2-ranges.S non-g GAS crash  [Re: spu-as internal error]
  2008-04-07 11:47   ` [patch] dw2-ranges.S non-g GAS crash [Re: spu-as internal error] Jan Kratochvil
@ 2008-04-08  5:42     ` Ulrich Weigand
  2008-04-08  7:56       ` Jan Kratochvil
  2008-04-14 19:08     ` Daniel Jacobowitz
  1 sibling, 1 reply; 4+ messages in thread
From: Ulrich Weigand @ 2008-04-08  5:42 UTC (permalink / raw)
  To: Jan Kratochvil; +Cc: Alan Modra, Ulrich.Weigand, gdb-patches

Jan Kratochvil wrote:

> > Has this testcase ever assembled?  On any target?  I think gas would
> > always have aborted unless you pass -g to gas or supply .file
> > directives.
> 
> It was always built just inside `runtest'.
> 
> Requesting a commit approval.  Verified unpatched GDB still FAILs on it.

As I just noticed, this is really also a spu-gcc problem: gcc *is* always
called with -g, but spu-gcc does not pass this on to the assembler.

Do you still think we should add the .file directive, or does the test
case make more sense without it?

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [patch] dw2-ranges.S non-g GAS crash  [Re: spu-as internal  error]
  2008-04-08  5:42     ` Ulrich Weigand
@ 2008-04-08  7:56       ` Jan Kratochvil
  0 siblings, 0 replies; 4+ messages in thread
From: Jan Kratochvil @ 2008-04-08  7:56 UTC (permalink / raw)
  To: Ulrich Weigand; +Cc: Alan Modra, Ulrich.Weigand, gdb-patches

On Tue, 08 Apr 2008 00:56:10 +0200, Ulrich Weigand wrote:
> Jan Kratochvil wrote:
> 
> > > Has this testcase ever assembled?  On any target?  I think gas would
> > > always have aborted unless you pass -g to gas or supply .file
> > > directives.
> > 
> > It was always built just inside `runtest'.
> > 
> > Requesting a commit approval.  Verified unpatched GDB still FAILs on it.
...
> Do you still think we should add the .file directive, or does the test
> case make more sense without it?

I think it is not nice the testcase fails to build without `-g' due to an
internal error of GAS.  I was not aware of it.

It is goot to know it is no longer required to be committed.

I requested the approval if it cannot conflict in some way with the debug info
being generated with `-g' but I doubt so.



Regards,
Jan


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [patch] dw2-ranges.S non-g GAS crash  [Re: spu-as internal  error]
  2008-04-07 11:47   ` [patch] dw2-ranges.S non-g GAS crash [Re: spu-as internal error] Jan Kratochvil
  2008-04-08  5:42     ` Ulrich Weigand
@ 2008-04-14 19:08     ` Daniel Jacobowitz
  1 sibling, 0 replies; 4+ messages in thread
From: Daniel Jacobowitz @ 2008-04-14 19:08 UTC (permalink / raw)
  To: Jan Kratochvil; +Cc: Alan Modra, Ulrich.Weigand, gdb-patches

On Mon, Apr 07, 2008 at 08:30:27AM +0200, Jan Kratochvil wrote:
> 2008-04-07  Jan Kratochvil  <jan.kratochvil@redhat.com>
> 
> 	* gdb.dwarf2/dw2-ranges.S: New stub `.file' directive.

I think this is OK.  Thanks.

-- 
Daniel Jacobowitz
CodeSourcery


^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2008-04-14 18:59 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <OF789B1703.6C47E7DB-ONCA257424.0010854A-E9257424.00108D3E@au1.ibm.com>
     [not found] ` <20080407040340.GF1446@bubble.grove.modra.org>
2008-04-07 11:47   ` [patch] dw2-ranges.S non-g GAS crash [Re: spu-as internal error] Jan Kratochvil
2008-04-08  5:42     ` Ulrich Weigand
2008-04-08  7:56       ` Jan Kratochvil
2008-04-14 19:08     ` Daniel Jacobowitz

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox