Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
133,267 Commits 606 Branches 0 Tags 3.1 GiB
93293b9ec7
Commit Graph

250 Commits

Author SHA1 Message Date
drochner
a35023d9c5 add a comment telling that i386bsd_sigcontext_addr() returns 
the address of a ucontext_t on NetBSD
(this is a hack of course)
 2005-01-13 12:51:04 +00:00
drochner
c578ca7bb1 wire in the right offsets of eip/esp within a ucontext_t 
(which we use now instead of "struct sigcontext")
 2005-01-13 12:31:09 +00:00
drochner
da4342c647 Signal handlers don't get passed a "struct sigcontext" 
anymore. Look in the ucontext_t for the next frame
instead. Backtrace through a signal handler invocation
works now.
(Since this change is NetBSD specific, this code shouldn't
be in a file which is shared among BSDs.)
 2005-01-13 12:29:41 +00:00
christos
31154055b9 PR/28781: OBATA Akio: gdb -write would core-dump if the executable did not 
have a string table. Check if the string table has been passed to us before
we attempt writing it.
 2004-12-26 16:55:00 +00:00
cube
b5856520ff Back out previous commit. PIDGET() must be used with a ptid_t, and the 
functions that were modified by the previous commit were already passed
the result of PIDGET.  Whatever was intended, it must be re-done
correctly.
 2004-10-16 21:17:51 +00:00
mycroft
2c0f3d7d9b Two changes for better handling of blocked threads with LWPs: 
1) When looking up the active thread, do not convert it to a thread ID; use the
   LWP ID instead.  This causes us to go through the ptrace(2) interface to
   modify registers -- using the kernel state -- rather than through the
   thread library.
2) Do not add blocked threads to the thread list.  These are represented by
   LWPs and will be seen that way instead.
 2004-10-12 22:16:34 +00:00
reinoud
3f13770b97 Finally x86-64 
WARNING: i can't be 100% sure that this patch will work due to the x86-64
code being different from all other archs... maybe it ought to be
streamlined to the other archs?
 2004-10-12 02:38:56 +00:00
reinoud
d1b2149022 Forgot NetBSD/alpha in my ptrace() fix. 2004-10-12 02:31:35 +00:00
reinoud
c90013b67b Pass the LWP id to the ptrace() calls. Passing a zero as it did meant that 
gdb could end up with the wrong registers... all other NetBSD archs
allready did this....
 2004-10-12 01:38:47 +00:00
mycroft
03c7b67710 If we can't read the switchframe from the core dump, fake it, rather than 
just failing.
 2004-10-08 17:17:45 +00:00
ian
ca1f4aea74 Bring in current sim/ppc sources which fix configury to work with 
NetBSD-current, in which statfs() exists but struct statfs does not.
 2004-09-29 06:31:36 +00:00
rearnsha
8cd9c1ad76 Add Thumb breakpoints 2004-08-23 13:54:29 +00:00
nathanw
fd8dc4c93f Remove some unused variables. 2004-08-09 22:39:32 +00:00
nathanw
0ef0b5bfc2 nbsd_thread_unsuspend() doesn't need to get or look at the thread 
info.
 2004-08-09 20:14:22 +00:00
nathanw
82c8abbad2 Fix up a few error messages to correctly report what function they're 
in.
 2004-08-09 20:13:53 +00:00
nathanw
417df974b7 Add rwlock support to "thread sync" command, and clean up some 
whitespace printing.
 2004-07-29 17:57:40 +00:00
nathanw
bb1c6b93e5 Enable scheduler_lock mode. 2004-07-20 21:39:04 +00:00
nathanw
066d05318f Jumbo update to address several issues. 
Go back to activating thread debugging in new_objfile(), as well as
several other places, so that thread operations such as wait() can be
handled inside of the child_ops.create_inferior(). This addresses
spurious SIGTRAPs at the start of debugging threaded programs.

Make use of the libpthread_dbg suspend/resume interface to implement
single-thread stepping.

Be a little more sensible about the way IS_LWP and IS_THREAD work.

Don't try to look up mutex data for a spinlock sync object (ugh,
unions).
 2004-07-20 21:33:44 +00:00
uwe
ae04b1322a Fix the calculation of next_pc for braf/bsrf instructions. 2004-05-28 22:04:36 +00:00
nathanw
411bc3c814 Failure to fetch FP registers from the thread debugging library isn't 
necessairly a problem; the thread may in fact have no FP state.
 2004-05-27 19:15:10 +00:00
matt
3df46d6c37 Fix stack backtracing so not refer to the initial entry frame. 
Default to getting the LR out of the stack if no other way works.

This makes backtraces work on stripped programs.
 2004-05-07 00:07:10 +00:00
nathanw
e30c25106f Remove an extraneous use of a 'retval'. 2004-04-15 19:26:58 +00:00
mrg
d7919c9c21 +TM_FILE= tm-x86-64nbsd.h 
this lets MKCROSSGDB work for ./build.sh -m amd64.
 2004-04-12 07:19:16 +00:00
matt
e5795ee03a Add libraries needed for configure 2004-04-04 20:17:27 +00:00
mrg
df1f7747e1 #endif /* MUST NOT HAVE TOKENS AFTER IT */ 2004-04-02 13:36:24 +00:00
mrg
dd567aa2f5 - (required) const poison 
- fix some function argument type issues (struct _bfd vs bfd)
 2004-04-02 13:34:50 +00:00
mrg
80a13c7a72 labels must have statements:; 2004-04-02 13:33:04 +00:00
matt
5132bfefdc Add support for traversing through signal handler frames. 2004-03-10 01:22:29 +00:00
nathanw
240c6ac9cf Re-implement fetch_kcore_registers() for alpha. 2004-01-02 22:41:17 +00:00
nathanw
243cf32568 Alpha uses software single-step. Lost in the 5.0->5.3 conversion. 
Addresses PR toolchain/23672.
 2003-12-11 21:06:27 +00:00
nathanw
de13a4e603 Catch up alpha_software_single_step() prototype with reality. 2003-12-11 21:02:24 +00:00
matt
4e90dd053f Add kcore and netbsd thread support. 2003-11-14 19:10:00 +00:00
matt
8056b67d49 Add kcore & netbsd thread support 2003-11-14 19:08:49 +00:00
briggs
5e0fd052da Fix from Ian Lance Taylor to allow building a cross-gdb with arm--netbsdelf 
target.  Patch as seen on tech-toolchain@.
 2003-11-14 15:06:15 +00:00
martin
4b0b317837 Remove unused variables. 2003-11-11 16:28:59 +00:00
martin
6354cf63d2 Fix register offset calculations. From Christian Limpach. 2003-11-11 16:24:23 +00:00
mrg
0bcb648614 set the arm breakpoint depending on the byte order. now armeb gdb works. 2003-11-09 02:52:07 +00:00
uwe
af9cced758 The "pref @rN" instruction was introduced in sh3, not sh4, so correct 
its sh_opcode_info entry accordingly.
 2003-11-07 14:43:18 +00:00
hannken
83bd54e877 Add "target kcore" to gdb53/sparc64. 
Taken verbatim from the old gdb in src/gnu/dist/toolchain/gdb.

Approved by: Christos Zoulas <christos@netbsd.org>
 2003-11-04 19:33:39 +00:00
skrll
c1946cebd4 Index fpregsetp->fpr correctly. 2003-11-03 15:59:11 +00:00
hannken
3f72313b8e Add "target kcore" to gdb53/sparc. 
Taken verbatim from the old gdb in src/gnu/dist/toolchain/gdb.

Approved by: Christos Zoulas <zoulas@netbsd.org>
 2003-11-02 20:59:28 +00:00
uwe
1d729d7bc4 First cut at implementing software single stepping for sh3. We need 
it because sh3 ports don't provide PT_STEP.  It is finally possible to
run programs under gdb now, though there are still issues to be
resolved.  It was deemed useful enough to commit early, while the work
is still in progress.  Tested on hpcsh.

Blessed by christos@.
 2003-10-31 00:11:27 +00:00
hannken
0165bf4148 Make it compile again. Blind copy-and-paste from sparc64nbsd-nat.c. 
Approved by: Christos Zoulas <christos@netbsd.org>
 2003-10-27 14:10:54 +00:00
cl
44f9690b65 g/c unused variables 2003-10-27 00:41:09 +00:00
christos
34e2db4a3c Make this link and work again (from Christian Limpach) 2003-10-26 17:45:13 +00:00
simonb
969d07f64f * mips-tdep.c (mips_addr_bits_remove): Don't shift by 32 
From Ian Taylor.
 2003-10-26 00:08:23 +00:00
matt
91f24c5034 Try dealing with sigtramps better. 2003-10-25 19:01:45 +00:00
matt
7819efde3c GDB 5.3 support for VAX 2003-10-25 18:26:07 +00:00
matt
17aa9b0c9b Use gregset_t/fpregset_t since that's what is really wanted. 2003-10-25 18:25:42 +00:00
cl
54641460eb make m68k/gdb53 compile as a crossgdb 
- catch up tm-nbsd.h with gnu/usr.bin/gdb53/arch/m68k/tm.h
- only use nbsd-thread.o in m68k nbsd-host case
 2003-10-25 16:20:58 +00:00