* [BFD PATCH] Fix QNX Neutrino core file reading
@ 2003-05-15 19:30 Kris Warkentin
2003-05-15 19:59 ` Daniel Jacobowitz
0 siblings, 1 reply; 2+ messages in thread
From: Kris Warkentin @ 2003-05-15 19:30 UTC (permalink / raw)
To: Gdb-Patches@Sources.Redhat.Com
I finally figured out why our core files weren't showing which thread was
active when it was generated. It seems we weren't setting lwpid correctly
and were not making the .reg section at the correct time. Okay to apply?
cheers,
Kris
Changelog:
* elf.c (elfcore_grok_nto_status): Only set lwpid for the active or
signalled thread.
(elfcore_grok_nto_gregs): Only make .reg section for the active thread.
Index: elf.c
===================================================================
RCS file: /cvs/src/src/bfd/elf.c,v
retrieving revision 1.184
diff -u -r1.184 elf.c
--- elf.c 9 May 2003 10:58:12 -0000 1.184
+++ elf.c 15 May 2003 19:27:31 -0000
@@ -7001,18 +7001,30 @@
char buf[100];
char *name;
asection *sect;
+ short sig;
+ unsigned flags;
/* nto_procfs_status 'pid' field is at offset 0. */
elf_tdata (abfd)->core_pid = bfd_get_32 (abfd, (bfd_byte *) ddata);
- /* nto_procfs_status 'tid' field is at offset 4. */
- elf_tdata (abfd)->core_lwpid = bfd_get_32 (abfd, (bfd_byte *) ddata + 4);
+ /* nto_procfs_status 'tid' field is at offset 4. Pass it back. */
+ *tid = bfd_get_32 (abfd, (bfd_byte *) ddata + 4);
- /* nto_procfs_status 'what' field is at offset 14. */
- elf_tdata (abfd)->core_signal = bfd_get_16 (abfd, (bfd_byte *) ddata +
14);
+ /* nto_procfs_status 'flags' field is at offset 8. */
+ flags = bfd_get_32 (abfd, (bfd_byte *) ddata + 8);
- /* Pass tid back. */
- *tid = elf_tdata (abfd)->core_lwpid;
+ /* nto_procfs_status 'what' field is at offset 14. */
+ if( (sig = bfd_get_16 (abfd, (bfd_byte *) ddata + 14)) > 0 ){
+ elf_tdata (abfd)->core_signal = sig;
+ elf_tdata (abfd)->core_lwpid = *tid;
+ }
+
+ /* _DEBUG_FLAG_CURTID (current thread) is 0x80. Some cores
+ do not come from signals so we make sure we set the current
+ thread just in case. */
+ if (flags & 0x00000080){
+ elf_tdata (abfd)->core_lwpid = *tid;
+ }
/* Make a ".qnx_core_status/%d" section. */
sprintf (buf, ".qnx_core_status/%d", *tid);
@@ -7022,8 +7034,7 @@
return FALSE;
strcpy (name, buf);
- sect = bfd_make_section (abfd, name);
- if (sect == NULL)
+ if ( (sect = bfd_make_section (abfd, name)) == NULL )
return FALSE;
sect->_raw_size = note->descsz;
@@ -7061,7 +7072,12 @@
sect->flags = SEC_HAS_CONTENTS;
sect->alignment_power = 2;
- return elfcore_maybe_make_sect (abfd, ".reg", sect);
+ /* This is the current thread. */
+ if (elf_tdata (abfd)->core_lwpid == tid){
+ return elfcore_maybe_make_sect (abfd, ".reg", sect);
+ }
+
+ return TRUE;
}
#define BFD_QNT_CORE_INFO 7
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [BFD PATCH] Fix QNX Neutrino core file reading
2003-05-15 19:30 [BFD PATCH] Fix QNX Neutrino core file reading Kris Warkentin
@ 2003-05-15 19:59 ` Daniel Jacobowitz
0 siblings, 0 replies; 2+ messages in thread
From: Daniel Jacobowitz @ 2003-05-15 19:59 UTC (permalink / raw)
To: Kris Warkentin; +Cc: Gdb-Patches@Sources.Redhat.Com
On Thu, May 15, 2003 at 03:30:51PM -0400, Kris Warkentin wrote:
> I finally figured out why our core files weren't showing which thread was
> active when it was generated. It seems we weren't setting lwpid correctly
> and were not making the .reg section at the correct time. Okay to apply?
Wrong list - ask on binutils@. Looks reasonable though.
>
> cheers,
>
> Kris
>
> Changelog:
>
> * elf.c (elfcore_grok_nto_status): Only set lwpid for the active or
> signalled thread.
> (elfcore_grok_nto_gregs): Only make .reg section for the active thread.
>
> Index: elf.c
> ===================================================================
> RCS file: /cvs/src/src/bfd/elf.c,v
> retrieving revision 1.184
> diff -u -r1.184 elf.c
> --- elf.c 9 May 2003 10:58:12 -0000 1.184
> +++ elf.c 15 May 2003 19:27:31 -0000
> @@ -7001,18 +7001,30 @@
> char buf[100];
> char *name;
> asection *sect;
> + short sig;
> + unsigned flags;
>
> /* nto_procfs_status 'pid' field is at offset 0. */
> elf_tdata (abfd)->core_pid = bfd_get_32 (abfd, (bfd_byte *) ddata);
>
> - /* nto_procfs_status 'tid' field is at offset 4. */
> - elf_tdata (abfd)->core_lwpid = bfd_get_32 (abfd, (bfd_byte *) ddata + 4);
> + /* nto_procfs_status 'tid' field is at offset 4. Pass it back. */
> + *tid = bfd_get_32 (abfd, (bfd_byte *) ddata + 4);
>
> - /* nto_procfs_status 'what' field is at offset 14. */
> - elf_tdata (abfd)->core_signal = bfd_get_16 (abfd, (bfd_byte *) ddata +
> 14);
> + /* nto_procfs_status 'flags' field is at offset 8. */
> + flags = bfd_get_32 (abfd, (bfd_byte *) ddata + 8);
>
> - /* Pass tid back. */
> - *tid = elf_tdata (abfd)->core_lwpid;
> + /* nto_procfs_status 'what' field is at offset 14. */
> + if( (sig = bfd_get_16 (abfd, (bfd_byte *) ddata + 14)) > 0 ){
> + elf_tdata (abfd)->core_signal = sig;
> + elf_tdata (abfd)->core_lwpid = *tid;
> + }
> +
> + /* _DEBUG_FLAG_CURTID (current thread) is 0x80. Some cores
> + do not come from signals so we make sure we set the current
> + thread just in case. */
> + if (flags & 0x00000080){
> + elf_tdata (abfd)->core_lwpid = *tid;
> + }
>
> /* Make a ".qnx_core_status/%d" section. */
> sprintf (buf, ".qnx_core_status/%d", *tid);
> @@ -7022,8 +7034,7 @@
> return FALSE;
> strcpy (name, buf);
>
> - sect = bfd_make_section (abfd, name);
> - if (sect == NULL)
> + if ( (sect = bfd_make_section (abfd, name)) == NULL )
> return FALSE;
>
> sect->_raw_size = note->descsz;
> @@ -7061,7 +7072,12 @@
> sect->flags = SEC_HAS_CONTENTS;
> sect->alignment_power = 2;
>
> - return elfcore_maybe_make_sect (abfd, ".reg", sect);
> + /* This is the current thread. */
> + if (elf_tdata (abfd)->core_lwpid == tid){
> + return elfcore_maybe_make_sect (abfd, ".reg", sect);
> + }
> +
> + return TRUE;
> }
>
> #define BFD_QNT_CORE_INFO 7
>
>
--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2003-05-15 19:59 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-05-15 19:30 [BFD PATCH] Fix QNX Neutrino core file reading Kris Warkentin
2003-05-15 19:59 ` Daniel Jacobowitz
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox