Re-registering a command will delete previous commands of the same name,
running the destroyer for the command object. The Guile destroyer
incorrectly tried to xfree the name and other strings, which is invalid
as they are on the GC heap.
gdb/ChangeLog:
* guile/scm-cmd.c (cmdscm_destroyer): Don't xfree the name and
other strings, as these are on the GC'd heap, and will be
collected along with the smob.
gdb/ChangeLog:
* guile/guile.c (_initialize_guile): Disable automatic
finalization, if Guile offers us that possibility.
* guile/guile.c (call_initialize_gdb_module):
* guile/scm-safe-call.c (gdbscm_with_catch): Arrange to run
finalizers in appropriate places.
* config.in (HAVE_GUILE_MANUAL_FINALIZATION): New definition.
* configure.ac (AC_TRY_LIBGUILE): Add a check for
scm_set_automatic_finalization_enabled.
* configure: Regenerated.
Instead of analyzing the prologue and possibly coming to a wrong
conclusion, this change tries to skip the prologue with the use of
skip_prologue_using_sal. Only if that fails, the prologue analyzer is
invoked as before.
gdb/ChangeLog:
* s390-linux-tdep.c (s390_skip_prologue): Skip the prologue using
SAL, if possible.
For multi-threaded inferiors on S390 GNU/Linux targets, GDB tried to
update the PER info via ptrace() in a newly attached thread before
assuring that the thread is stopped. Depending on the timing, this
could lead to a GDB internal error. The patch defers the PER info
update until just before resuming the thread.
gdb/ChangeLog:
* s390-linux-nat.c (struct arch_lwp_info): New.
(s390_fix_watch_points): Rename to...
(s390_prepare_to_resume): ...this. Skip the PER info update
unless the watch points have changed.
(s390_refresh_per_info, s390_new_thread): New functions.
(s390_insert_watchpoint): Call s390_refresh_per_info instead of
s390_fix_watch_points.
(s390_remove_watchpoint): Likewise.
(_initialize_s390_nat): Reflect renaming of s390_fix_watch_points.
Register s390_prepare_to_resume.
Unfortunately, the Python version of the dg-extract-results.sh script
doesn't produce stable-enough results for GDB. The test messages
appear to end up alpha sorted (losing the original sequence) and also
sorting changes between runs for some reason. That may be tolerable
for GCC, but for GDB, it often renders test results diffing between
different revisions unworkable.
Until that is fixed upstream, delete the script from the GDB tree.
testsuite/ChangeLog:
2015-03-09 Pedro Alves <palves@redhat.com>
* dg-extract-results.py: Delete.
This merges Sergio's fix from GCC:
https://gcc.gnu.org/ml/gcc-patches/2014-12/msg01293.html
gdb/testsuite/ChangeLog:
2015-03-09 Pedro Alves <palves@redhat.com>
Merge dg-extract-results.sh from GCC upstream (r218843).
2014-12-17 Sergio Durigan Junior <sergiodj@redhat.com>
* dg-extract-results.sh: Use --text with grep to avoid issues with
binary files. Fall back to cat -v, if that doesn't work.
Fixes this in C++ mode:
gdb/gdbserver/tracepoint.c: In function ‘void* gdb_agent_helper_thread(void*)’:
gdb/gdbserver/tracepoint.c:7190:47: error: cannot convert ‘sockaddr_un*’ to ‘sockaddr*’ for argument ‘2’ to ‘int accept(int, sockaddr*, socklen_t*)’
fd = accept (listen_fd, &sockaddr, &tmp);
gdb/gdbserver/ChangeLog:
2015-03-09 Pedro Alves <palves@redhat.com>
* tracepoint.c (gdb_agent_helper_thread): Cast '&sockaddr' to
'struct sockaddr' pointer in 'accept' call.
This reverts 366c75fc.
We don't actually need to access the object through
"struct sockaddr *", so we don't need the union:
https://sourceware.org/ml/gdb-patches/2015-03/msg00213.html
gdb/ChangeLog:
2015-03-09 Pedro Alves <palves@redhat.com>
Revert:
2015-03-07 Pedro Alves <palves@redhat.com>
* common/gdb_socket.h: New file.
* ser-tcp.c: Include gdb_socket.h. Don't include netinet/in.h nor
sys/socket.h.
(net_open): Use union gdb_sockaddr_u.
gdb/gdbserver/ChangeLog:
2015-03-09 Pedro Alves <palves@redhat.com>
Revert:
2015-03-07 Pedro Alves <palves@redhat.com>
* gdbreplay.c: No longer include <netinet/in.h>, <sys/socket.h>,
or <winsock2.h> here. Instead include "gdb_socket.h".
(remote_open): Use union gdb_sockaddr_u.
* remote-utils.c: No longer include <netinet/in.h>, <sys/socket.h>
or <winsock2.h> here. Instead include "gdb_socket.h".
(handle_accept_event, remote_prepare): Use union gdb_sockaddr_u.
* tracepoint.c: Include "gdb_socket.h" instead of <sys/socket.h>
or <sys/un.h>.
(init_named_socket, gdb_agent_helper_thread): Use union
gdb_sockaddr_u.
Whoops, these are C specific, but I somehow missed the warnings before:
cc1plus: warning: command line option ‘-Wmissing-prototypes’ is valid for C/ObjC but not for C++ [enabled by default]
cc1plus: warning: command line option ‘-Wdeclaration-after-statement’ is valid for C/ObjC but not for C++ [enabled by default]
cc1plus: warning: command line option ‘-Wmissing-parameter-type’ is valid for C/ObjC but not for C++ [enabled by default]
cc1plus: warning: command line option ‘-Wold-style-declaration’ is valid for C/ObjC but not for C++ [enabled by default]
cc1plus: warning: command line option ‘-Wold-style-definition’ is valid for C/ObjC but not for C++ [enabled by default]
gdb/ChangeLog:
2015-03-07 Pedro Alves <palves@redhat.com>
* configure.ac (build_warnings): Move -Wmissing-prototypes
-Wdeclaration-after-statement -Wmissing-parameter-type
-Wold-style-declaration -Wold-style-definition to the C-specific
set.
* configure: Regenerate.
gdb/gdbserver/ChangeLog:
2015-03-07 Pedro Alves <palves@redhat.com>
* configure.ac (build_warnings): Move
-Wdeclaration-after-statement to the C-specific set.
* configure: Regenerate.
Building gdbserver in C++ mode shows:
gdb/gdbserver/tracepoint.c: In function ‘void* gdb_agent_helper_thread(void*)’:
gdb/gdbserver/tracepoint.c:7190:47: error: cannot convert ‘sockaddr_un*’ to ‘sockaddr*’ for argument ‘2’ to ‘int accept(int, sockaddr*, socklen_t*)’
fd = accept (listen_fd, &sockaddr, &tmp);
A few places in the tree already have an explicit cast to struct
sockaddr *, but that's a strict aliasing violation. Instead of
propagating invalid code, fix this by using a union instead.
gdb/ChangeLog:
2015-03-07 Pedro Alves <palves@redhat.com>
* common/gdb_socket.h: New file.
* ser-tcp.c: Include gdb_socket.h. Don't include netinet/in.h nor
sys/socket.h.
(net_open): Use union gdb_sockaddr_u.
gdb/gdbserver/ChangeLog:
2015-03-07 Pedro Alves <palves@redhat.com>
* gdbreplay.c: No longer include <netinet/in.h>, <sys/socket.h>,
or <winsock2.h> here. Instead include "gdb_socket.h".
(remote_open): Use union gdb_sockaddr_u.
* remote-utils.c: No longer include <netinet/in.h>, <sys/socket.h>
or <winsock2.h> here. Instead include "gdb_socket.h".
(handle_accept_event, remote_prepare): Use union gdb_sockaddr_u.
* tracepoint.c: Include "gdb_socket.h" instead of <sys/socket.h>
or <sys/un.h>.
(init_named_socket, gdb_agent_helper_thread): Use union
gdb_sockaddr_u.
Although the current TRY/CATCH implementation works in C++ mode too,
it relies on setjmp/longjmp, and longjmp bypasses calling the
destructors of objects on the stack, which is obviously bad for C++.
This patch fixes this by makes TRY/CATCH use real try/catch in C++
mode behind the scenes. The way this is done allows RAII and cleanups
to coexist while we phase out cleanups, instead of requiring a flag
day.
This patch is not strictly necessary until we require a C++ compiler
and start actually using RAII, though I'm all for baby steps, and it
shows my proposed way forward. Putting it in now, allows for easier
experimentation and exposure of potential problems with real C++
exceptions.
gdb/ChangeLog:
2015-03-07 Pedro Alves <palves@redhat.com>
* common/common-exceptions.c [!__cplusplus] (enum catcher_state)
(exceptions_state_mc_action_iter)
(exceptions_state_mc_action_iter_1, exceptions_state_mc_catch):
Don't define.
[__cplusplus] (try_scope_depth): New global.
[__cplusplus] (exception_try_scope_entry)
(exception_try_scope_exit, gdb_exception_sliced_copy)
(exception_rethrow): New functions.
(throw_exception): In C++ mode, throw
gdb_exception_RETURN_MASK_QUIT for RETURN_QUIT and
gdb_exception_RETURN_MASK_ERROR for RETURN_ERROR.
(throw_it): In C++ mode, use try_scope_depth.
* common/common-exceptions.h [!__cplusplus]
(exceptions_state_mc_action_iter)
(exceptions_state_mc_action_iter_1, exceptions_state_mc_catch):
Don't declare.
[__cplusplus] (exception_try_scope_entry)
(exception_try_scope_exit, exception_rethrow): Declare.
[__cplusplus] (struct exception_try_scope): New struct.
[__cplusplus] (TRY, CATCH, END_CATCH): Reimplement on top of real
C++ exceptions.
(struct gdb_exception_RETURN_MASK_ALL)
(struct gdb_exception_RETURN_MASK_ERROR)
(struct gdb_exception_RETURN_MASK_QUIT): New types.
After the previous patch, this is the last remaining use of a volatile
struct gdb_exception. Kill it, as it's troublesome for C++: we can't
assign volatile <-> non-volatile without copy constructors /
assignment operators that do that, which I'd rather avoid.
gdb/ChangeLog:
2015-03-07 Pedro Alves <palves@redhat.com>
* main.c (handle_command_errors): Remove volatile qualifier from
parameter.
All these were caught by actually making TRY/CATCH use try/catch
behind the scenes, which then resulted in the build failing (on x86_64
Fedora 20) because there was code between the try and catch blocks.
gdb/ChangeLog:
2015-03-07 Pedro Alves <palves@redhat.com>
* breakpoint.c (save_breakpoints): Adjust to avoid code between
TRY and CATCH.
* gdbtypes.c (safe_parse_type): Remove empty line.
(types_deeply_equal):
* guile/scm-frame.c (gdbscm_frame_name):
* linux-thread-db.c (find_new_threads_once):
* python/py-breakpoint.c (bppy_get_commands):
* record-btrace.c (record_btrace_insert_breakpoint)
(record_btrace_remove_breakpoint, record_btrace_start_replaying)
(record_btrace_start_replaying): Adjust to avoid code between TRY
and CATCH.
This patch splits the TRY_CATCH macro into three, so that we go from
this:
~~~
volatile gdb_exception ex;
TRY_CATCH (ex, RETURN_MASK_ERROR)
{
}
if (ex.reason < 0)
{
}
~~~
to this:
~~~
TRY
{
}
CATCH (ex, RETURN_MASK_ERROR)
{
}
END_CATCH
~~~
Thus, we'll be getting rid of the local volatile exception object, and
declaring the caught exception in the catch block.
This allows reimplementing TRY/CATCH in terms of C++ exceptions when
building in C++ mode, while still allowing to build GDB in C mode
(using setjmp/longjmp), as a transition step.
TBC, after this patch, is it _not_ valid to have code between the TRY
and the CATCH blocks, like:
TRY
{
}
// some code here.
CATCH (ex, RETURN_MASK_ERROR)
{
}
END_CATCH
Just like it isn't valid to do that with C++'s native try/catch.
By switching to creating the exception object inside the CATCH block
scope, we can get rid of all the explicitly allocated volatile
exception objects all over the tree, and map the CATCH block more
directly to C++'s catch blocks.
The majority of the TRY_CATCH -> TRY+CATCH+END_CATCH conversion was
done with a script, rerun from scratch at every rebase, no manual
editing involved. After the mechanical conversion, a few places
needed manual intervention, to fix preexisting cases where we were
using the exception object outside of the TRY_CATCH block, and cases
where we were using "else" after a 'if (ex.reason) < 0)' [a CATCH
after this patch]. The result was folded into this patch so that GDB
still builds at each incremental step.
END_CATCH is necessary for two reasons:
First, because we name the exception object in the CATCH block, which
requires creating a scope, which in turn must be closed somewhere.
Declaring the exception variable in the initializer field of a for
block, like:
#define CATCH(EXCEPTION, mask) \
for (struct gdb_exception EXCEPTION; \
exceptions_state_mc_catch (&EXCEPTION, MASK); \
EXCEPTION = exception_none)
would avoid needing END_CATCH, but alas, in C mode, we build with C90,
which doesn't allow mixed declarations and code.
Second, because when TRY/CATCH are wired to real C++ try/catch, as
long as we need to handle cleanup chains, even if there's no CATCH
block that wants to catch the exception, we need for stop at every
frame in the unwind chain and run cleanups, then rethrow. That will
be done in END_CATCH.
After we require C++, we'll still need TRY/CATCH/END_CATCH until
cleanups are completely phased out -- TRY/CATCH in C++ mode will
save/restore the current cleanup chain, like in C mode, and END_CATCH
catches otherwise uncaugh exceptions, runs cleanups and rethrows, so
that C++ cleanups and exceptions can coexist.
IMO, this still makes the TRY/CATCH code look a bit more like a
newcomer would expect, so IMO worth it even if we weren't considering
C++.
gdb/ChangeLog.
2015-03-07 Pedro Alves <palves@redhat.com>
* common/common-exceptions.c (struct catcher) <exception>: No
longer a pointer to volatile exception. Now an exception value.
<mask>: Delete field.
(exceptions_state_mc_init): Remove all parameters. Adjust.
(exceptions_state_mc): No longer pop the catcher here.
(exceptions_state_mc_catch): New function.
(throw_exception): Adjust.
* common/common-exceptions.h (exceptions_state_mc_init): Remove
all parameters.
(exceptions_state_mc_catch): Declare.
(TRY_CATCH): Rename to ...
(TRY): ... this. Remove EXCEPTION and MASK parameters.
(CATCH, END_CATCH): New.
All callers adjusted.
gdb/gdbserver/ChangeLog:
2015-03-07 Pedro Alves <palves@redhat.com>
Adjust all callers of TRY_CATCH to use TRY/CATCH/END_CATCH
instead.
More preparation for running the TRY_CATCH->TRY/CATCH conversion
script.
gdb/ChangeLog:
2015-03-07 Tom Tromey <tromey@redhat.com>
* top.c (quit_force): Inline and delete DO_TRY, DO_PRINT_EX.
This normalizes some exception catch blocks that check for ex.reason
to look like this:
~~~
volatile gdb_exception ex;
TRY_CATCH (ex, RETURN_MASK_ALL)
{
...
}
if (ex.reason < 0)
{
...
}
~~~
This is a preparation step for running a script that converts all
TRY_CATCH uses to look like this instead:
~~~
TRY
{
...
}
CATCH (ex, RETURN_MASK_ALL)
{
...
}
END_CATCH
~~~
The motivation for that change is being able to reimplent TRY/CATCH in
terms of C++ try/catch.
This commit makes it so that:
- no condition other than ex.reason < 0 is checked in the if
predicate
- there's no "else" block to check whether no exception was caught
- there's no code between the TRY_CATCH (TRY) block and the
'if (ex.reason < 0)' block (CATCH).
- the exception object is no longer referred to outside the if/catch
block. Note the local volatile exception objects that are
currently defined inside functions that use TRY_CATCH will
disappear. In cases it's more convenient to still refer to the
exception outside the catch block, a new non-volatile local is
added and copy to that object is made within the catch block.
The following patches should make this all clearer.
gdb/ChangeLog:
2015-03-07 Pedro Alves <palves@redhat.com>
* amd64-tdep.c (amd64_frame_cache, amd64_sigtramp_frame_cache)
(amd64_epilogue_frame_cache): Normal exception handling code.
* break-catch-throw.c (check_status_exception_catchpoint)
(re_set_exception_catchpoint): Ditto.
* cli/cli-interp.c (safe_execute_command):
* cli/cli-script.c (script_from_file): Ditto.
* compile/compile-c-symbols.c (generate_c_for_for_one_variable):
Ditto.
* compile/compile-object-run.c (compile_object_run): Ditto.
* cp-abi.c (baseclass_offset): Ditto.
* cp-valprint.c (cp_print_value): Ditto.
* exceptions.c (catch_exceptions_with_msg):
* frame-unwind.c (frame_unwind_try_unwinder): Ditto.
* frame.c (get_frame_address_in_block_if_available): Ditto.
* i386-tdep.c (i386_frame_cache, i386_epilogue_frame_cache)
(i386_sigtramp_frame_cache): Ditto.
* infcmd.c (post_create_inferior): Ditto.
* linespec.c (parse_linespec, find_linespec_symbols):
* p-valprint.c (pascal_object_print_value): Ditto.
* parse.c (parse_expression_for_completion): Ditto.
* python/py-finishbreakpoint.c (bpfinishpy_init): Ditto.
* remote.c (remote_get_noisy_reply): Ditto.
* s390-linux-tdep.c (s390_frame_unwind_cache): Ditto.
* solib-svr4.c (solib_svr4_r_map): Ditto.
I happen to see that show_debug_regs is used as an arithmetic type,
but it should be a boolean,
if (show_debug_regs > 1)
On the other hand, GDB RSP only allows setting it to either 0 or 1,
so it makes no sense to check whether it is greater than 1. This
patch fixes it.
gdb/gdbserver:
2015-03-06 Yao Qi <yao.qi@linaro.org>
* linux-aarch64-low.c (aarch64_insert_point): Use
show_debug_regs as a boolean.
(aarch64_remove_point): Likewise.
This commit introduces a new inline common function "startswith"
which takes two string arguments and returns nonzero if the first
string starts with the second. It also updates the 295 places
where this logic was written out longhand to use the new function.
gdb/ChangeLog:
* common/common-utils.h (startswith): New inline function.
All places where this logic was used updated to use the above.
Trying to fix a permanent breakpoints bug, I broke "next" over a
regular breakpoint. "next" would immediately hit the breakpoint the
program was already stopped at. But, the "next over setup" test
failed to notice this and still issued a pass. That's because the
regex matches "testsuite" in:
Breakpoint 2 at 0x400687: file src/gdb/testsuite/gdb.base/bp-permanent.c, line 46.
gdb/testsuite/ChangeLog:
2015-03-05 Pedro Alves <palves@redhat.com>
* gdb.base/bp-permanent.exp: Tighten "next over setup" regex.
When we find out that a breakpoint is set on top of a program
breakpoint, we mark it as "permanent". E.g.,:
...
if (bp_loc_is_permanent (loc))
{
loc->inserted = 1;
loc->permanent = 1;
}
...
Note we didn't fill in the breakpoint's shadow (shadow_len remains 0).
In case the target claims support for evaluating breakpoint
conditions, GDB sometimes reinserts breakpoints that are already
inserted (to update the conditions on the target side). Since GDB
doesn't know whether the target supports evaluating conditions _of_
software breakpoints (vs hardware breakpoints, etc.) until it actually
tries it, if the target doesn't actually support z0 breakpoints, GDB
ends up reinserting a GDB-managed software/memory breakpoint
(mem-break.c).
And that is the case that is buggy: breakpoints that are marked
inserted contribute their shadows (if any) to the memory returned by
target_read_memory, to mask out breakpoints. Permanent breakpoints
are always marked as inserted. So if the permanent breakpoint doesn't
have a shadow yet in its shadow buffer, but we set shadow_len before
calling target_read_memory, then the still clear shadow_contents
buffer will be used by the breakpoint masking code... And then from
there on, the permanent breakpoint has a broken shadow buffer, and
thus any memory read out of that address will read bogus code, and
many random bad things fall out from that.
The fix is just to set shadow_len at the same time shadow_contents is
set, not one before and another after...
Fixes all gdb.base/bp-permanent.exp FAILs on PPC64 GNU/Linux gdbserver
and probably any other gdbserver port that doesn't do z0 breakpoints.
gdb/ChangeLog:
2015-03-05 Pedro Alves <palves@redhat.com>
PR gdb/18002
* mem-break.c (default_memory_insert_breakpoint): Set shadow_len
after reading the breakpoint's shadow memory.
I forgot to update these target_ops instances when I added these new
hooks.
I confirmed mingw32-w64 builds again at least.
gdb/gdbserver/ChangeLog:
2015-03-05 Pedro Alves <palves@redhat.com>
* lynx-low.c (lynx_target_ops): Install NULL hooks for
stopped_by_sw_breakpoint, supports_stopped_by_sw_breakpoint,
stopped_by_hw_breakpoint, supports_stopped_by_hw_breakpoint.
* nto-low.c (nto_target_ops): Likewise.
* spu-low.c (spu_target_ops): Likewise.
* win32-low.c (win32_target_ops): Likewise.
When interrupting a thread in non-stop vs all-stop, the signal given in
the MI *stopped event is not the same. Currently, mi_expect_interrupt only
accepts the case for non-stop, so this adds the alternative for all-stop.
gdb/testsuite/ChangeLog:
* lib/mi-support.exp (mi_expect_interrupt): Accept
alternative event for when in all-stop mode.
record-btrace was the only target making use of this, and it no longer
uses it.
gdb/ChangeLog:
2015-03-04 Pedro Alves <palves@redhat.com>
* target.h (struct target_ops) <to_decr_pc_after_break>: Delete.
(target_decr_pc_after_break): Delete declaration.
* target.c (default_target_decr_pc_after_break)
(target_decr_pc_after_break): Delete.
* linux-nat.c (check_stopped_by_breakpoint, linux_nat_wait_1): Use
gdbarch_decr_pc_after_break instead of target_decr_pc_after_break.
* linux-thread-db.c (check_event): Likewise.
* infrun.c (adjust_pc_after_break): Likewise.
* darwin-nat.c (cancel_breakpoint): Likewise.
* aix-thread.c (aix_thread_wait): Likewise.
* target-delegates.c: Regenerate.
This patch adjusts gdbserver's Linux backend to tell gdbserver core
(and ultimately GDB) whether a trap was caused by a breakpoint.
It teaches the backend to get that information out of the si_code of
the SIGTRAP siginfo.
gdb/gdbserver/ChangeLog:
2015-03-04 Pedro Alves <palves@redhat.com>
* linux-low.c (check_stopped_by_breakpoint) [USE_SIGTRAP_SIGINFO]:
Decide whether a breakpoint triggered based on the SIGTRAP's
siginfo.si_code.
(thread_still_has_status_pending_p) [USE_SIGTRAP_SIGINFO]: Don't check whether a
breakpoint is inserted if relying on SIGTRAP's siginfo.si_code.
(linux_low_filter_event): Check for breakpoints before checking
watchpoints.
(linux_wait_1): Don't re-increment the PC if relying on SIGTRAP's
siginfo.si_code.
(linux_stopped_by_sw_breakpoint)
(linux_supports_stopped_by_sw_breakpoint)
(linux_stopped_by_hw_breakpoint)
(linux_supports_stopped_by_hw_breakpoint): New functions.
(linux_target_ops): Install new target methods.
This patch teaches the core of gdbserver about the new "swbreak" and
"hwbreak" stop reasons, and adds the necessary hooks a backend needs
to implement to support the feature.
gdb/gdbserver/ChangeLog:
2015-03-04 Pedro Alves <palves@redhat.com>
* remote-utils.c (prepare_resume_reply): Report swbreak/hbreak.
* server.c (swbreak_feature, hwbreak_feature): New globals.
(handle_query) <qSupported>: Handle "swbreak+" and "hwbreak+".
(captured_main): Clear swbreak_feature and hwbreak_feature.
* server.h (swbreak_feature, hwbreak_feature): Declare.
* target.h (struct target_ops) <stopped_by_sw_breakpoint,
supports_stopped_by_sw_breakpoint, stopped_by_hw_breakpoint,
supports_stopped_by_hw_breakpoint>: New fields.
(target_supports_stopped_by_sw_breakpoint)
(target_stopped_by_sw_breakpoint)
(target_supports_stopped_by_hw_breakpoint)
(target_stopped_by_hw_breakpoint): Declare.
This patch adjusts the native Linux target backend to tell the core
whether a trap was caused by a breakpoint.
It teaches the target to get that information out of the si_code of
the SIGTRAP siginfo.
Tested on x86-64 Fedora 20, s390 RHEL 7, and PPC64 Fedora 18. An
earlier version was tested on ARM Fedora 21.
gdb/ChangeLog:
2015-03-04 Pedro Alves <palves@redhat.com>
* linux-nat.c (save_sigtrap): Check for breakpoints before
checking watchpoints.
(status_callback) [USE_SIGTRAP_SIGINFO]: Don't check whether a
breakpoint is inserted if relying on SIGTRAP's siginfo.si_code.
(check_stopped_by_breakpoint) [USE_SIGTRAP_SIGINFO]: Decide whether
a breakpoint triggered based on the SIGTRAP's siginfo.si_code.
(linux_nat_stopped_by_sw_breakpoint)
(linux_nat_supports_stopped_by_sw_breakpoint)
(linux_nat_stopped_by_hw_breakpoint)
(linux_nat_supports_stopped_by_hw_breakpoint): New functions.
(linux_nat_wait_1): Don't re-increment the PC if relying on
SIGTRAP's siginfo->si_code.
(linux_nat_add_target): Install new target methods.
* linux-thread-db.c (check_event): Don't account for breakpoint PC
offset if the target already adjusted the PC.
* nat/linux-ptrace.h (USE_SIGTRAP_SIGINFO): New.
(GDB_ARCH_TRAP_BRKPT): New.
(TRAP_HWBKPT): Define if not already defined.
This adjusts target remote to tell the core whether a trap was caused
by a breakpoint.
To that end, the patch teaches GDB about new RSP stop reasons "T05
swbreak" and "T05 hwbreak", that remote targets report back to GDB,
similarly to how "T05 watch" indicates a stop caused by a watchpoint.
Because targets that can report these events are expected to
themselves adjust the PC after a software breakpoint, these new stop
reasons must only be reported if the stub is talking to a GDB that
understands them. Because of that, the use of the new stop reasons
needs to be handshaked on initial connection, using the qSupported
mechanism. GDB simply sends "swbreak+" in its qSupports query, and
the stub reports back "swbreak+" too.
Because these new stop reasons are required to fix a fundamental
non-stop mode problem, this commit extends the remote non-stop intro
section in the manual, documenting the events as required.
To be clear, GDB will still cope with remote targets that don't
support these new stop reasons; it will behave just like today.
Tested on x86-64 Fedora 20, native and gdbserver.
gdb/ChangeLog:
2015-03-04 Pedro Alves <palves@redhat.com>
* NEWS: Mention the new "swbreak" and "hwbreak" stop reasons.
* remote.c (struct remote_state) <remote_stopped_by_watchpoint_p>:
Delete field.
<stop_reason>: New field.
(PACKET_swbreak_feature, PACKET_hwbreak_feature): New enum values.
(packet_set_cmd_state): New function.
(remote_protocol_features): Register the "swbreak" and "hwbreak"
features.
(remote_query_supported): If not disabled with the corresponding
"set remote foo-packet" command, report support for the swbreak
and hwbreak features.
(struct stop_reply) <remote_stopped_by_watchpoint_p>: Delete
field.
<stop_reason>: New field.
(remote_parse_stop_reply): Handle "swbreak" and "hwbreak".
(remote_wait_as): Adjust.
(remote_stopped_by_sw_breakpoint)
(remote_supports_stopped_by_sw_breakpoint)
(remote_stopped_by_hw_breakpoint)
(remote_supports_stopped_by_hw_breakpoint): New functions.
(remote_stopped_by_watchpoint): New function.
(init_remote_ops): Install them.
(_initialize_remote): Register new "set/show remote
swbreak-feature-packet" and "set/show remote
swbreak-feature-packet" commands.
gdb/doc/ChangeLog:
2015-03-04 Pedro Alves <palves@redhat.com>
* gdb.texinfo (Remote Configuration): Document the "set/show
remote swbreak-feature-packet" and "set/show remote
hwbreak-feature-packet" commands.
(Packets) <Z0>: Add cross link to the "swbreak" stop reason's
decription.
(Stop Reply Packets): Document the swbreak and hwbreak stop
reasons.
(General Query Packets): Document the swbreak and hwbreak
qSupported features.
(Remote Non-Stop): Explain that swbreak and hwbreak are required.
This adjusts the record targets to tell the core whether a trap was
caused by a breakpoint. Targets that can do this should report
breakpoint traps with the PC already adjusted, so this removes the
re-incrementing record-full was doing.
These targets need to be adjusted before process_stratum targets
beneath are, otherwise target_supports_stopped_by_sw_breakpoint,
etc. would fall through to the target beneath while
recording/replaying, and the core would get confused.
Tested on x86-64 Fedora 20, native and gdbserver.
gdb/ChangeLog:
2015-03-04 Pedro Alves <palves@redhat.com>
* btrace.h: Include target/waitstatus.h.
(struct btrace_thread_info) <stop_reason>: New field.
* record-btrace.c (record_btrace_step_thread): Use
record_check_stopped_by_breakpoint instead of breakpoint_here_p.
(record_btrace_decr_pc_after_break): Delete.
(record_btrace_stopped_by_sw_breakpoint)
(record_btrace_supports_stopped_by_sw_breakpoint)
(record_btrace_stopped_by_hw_breakpoint)
(record_btrace_supports_stopped_by_hw_breakpoint): New functions.
(init_record_btrace_ops): Install them.
* record-full.c (record_full_hw_watchpoint): Delete and replace
with ...
(record_full_stop_reason): ... this throughout.
(record_full_exec_insn): Adjust.
(record_full_wait_1): Adjust. No longer re-increment the PC.
(record_full_wait_1): Adjust. Use
record_check_stopped_by_breakpoint instead of breakpoint_here_p.
(record_full_stopped_by_watchpoint): Adjust.
(record_full_stopped_by_sw_breakpoint)
(record_full_supports_stopped_by_sw_breakpoint)
(record_full_supports_stopped_by_sw_breakpoint)
(record_full_stopped_by_hw_breakpoint)
(record_full_supports_stopped_by_hw_breakpoint): New functions.
(init_record_full_ops, init_record_full_core_ops): Install them.
* record.c (record_check_stopped_by_breakpoint): New function.
* record.h: Include target/waitstatus.h.
(record_check_stopped_by_breakpoint): New declaration.
The moribund locations heuristics are problematic. This patch teaches
GDB about targets that can reliably tell whether a trap was caused by
a software or hardware breakpoint, and thus don't need moribund
locations, thus bypassing all the problems that mechanism has.
The non-stop-fair-events.exp test is frequently failing currently.
E.g., see https://sourceware.org/ml/gdb-testers/2015-q1/msg03148.html.
The root cause is a fundamental problem with moribund locations. For
example, the stepped_breakpoint logic added by af48d08f breaks in this
case (which is what happens with that test):
- Step thread A, no breakpoint is set at PC.
- The kernel doesn't schedule thread A yet.
- Insert breakpoint at A's PC, for some reason (e.g., a step-resume
breakpoint for thread B).
- Kernel finally schedules thread A.
- thread A's stepped_breakpoint flag is not set, even though it now
stepped a breakpoint instruction.
- adjust_pc_after_break gets the PC wrong, because PC == PREV_PC, but
stepped_breakpoint is not set.
We needed the stepped_breakpoint logic to workaround moribund
locations, because otherwise adjust_pc_after_break could apply an
adjustment when it shouldn't just because there _used_ to be a
breakpoint at PC (a moribund breakpoint location). For example, on
x86, that's wrong if the thread really hasn't executed an int3, but
instead executed some other 1-byte long instruction. Getting the PC
adjustment wrong of course leads to the inferior executing the wrong
instruction.
Other problems with moribund locations are:
- if a true SIGTRAP happens to be raised when the program is
executing the PC that used to have a breakpoint, GDB will assume
that is a trap for a breakpoint that has recently been removed, and
thus we miss reporting the random signal to the user.
- to minimize that, we get rid of moribund location after a while.
That while is defined as just a certain number of events being
processed. That number of events sometimes passes by before a
delayed breakpoint is processed, and GDB confuses the trap for a
random signal, thus reporting the random trap. Once the user
resumes the thread, the program crashes because the PC was not
adjusted...
The fix for all this is to bite the bullet and get rid of heuristics
and instead rely on the target knowing accurately what caused the
SIGTRAP. The target/kernel/stub is in the best position to know what
that, because it can e.g. consult priviledged CPU flags GDB has no
access to, or by knowing which exception vector entry was called when
the instruction trapped, etc. Most debug APIs I've seen to date
report breakpoint hits as a distinct event in some fashion. For
example, on the Linux kernel, whether a breakpoint was executed is
exposed to userspace in the si_code field of the SIGTRAP's siginfo.
On Windows, the debug API reports a EXCEPTION_BREAKPOINT exception
code.
We needed to keep around deleted breakpoints in an on-the-side list
(the moribund locations) for two main reasons:
- Know that a SIGTRAP actually is a delayed event for a hit of a
breakpoint that was removed before the event was processed, and
thus should not be reported as a random signal.
- So we still do the decr_pc_after_break adjustment in that case, so
that the thread is resumed at the correct address.
In the new model, if GDB processes an event the target tells is a
breakpoint trap, and GDB doesn't find the corresponding breakpoint in
its breakpoint tables, it means that event is a delayed event for a
breakpoint that has since been removed, and thus the event should be
ignored.
For the decr_pc_after_after issue, it ends up being much simpler that
on targets that can reliably tell whether a breakpoint trapped, for
the breakpoint trap to present the PC already adjusted. Proper
multi-threading support already implies that targets needs to be doing
decr_pc_after_break adjustment themselves, otherwise for example, in
all-stop if two threads hit a breakpoint simultaneously, and the user
does "info threads", he'll see the non-event thread that hit the
breakpoint stopped at the wrong PC.
This way (target adjusts) also ends up eliminating the need for some
awkward re-incrementing of the PC in the record-full and Linux targets
that we do today, and the need for the target_decr_pc_after_break
hook.
If the target always adjusts, then there's a case where GDB needs to
re-increment the PC. Say, on x86, an "int3" instruction that was
explicitly written in the program traps. In this case, GDB should
report a random SIGTRAP signal to the user, with the PC pointing at
the instruction past the int3, just like if GDB was not debugging the
program. The user may well decide to pass the SIGTRAP to the program
because the program being debugged has a SIGTRAP handler that handles
its own breakpoints, and expects the PC to be unadjusted.
Tested on x86-64 Fedora 20.
gdb/ChangeLog:
2015-03-04 Pedro Alves <palves@redhat.com>
* breakpoint.c (need_moribund_for_location_type): New function.
(bpstat_stop_status): Don't skipping checking moribund locations
of breakpoint types which the target tell caused a stop.
(program_breakpoint_here_p): New function, factored out from ...
(bp_loc_is_permanent): ... this.
(update_global_location_list): Don't create a moribund location if
the target supports reporting stops of the type of the removed
breakpoint.
* breakpoint.h (program_breakpoint_here_p): New declaration.
* infrun.c (adjust_pc_after_break): Return early if the target has
already adjusted the PC. Add comments.
(handle_signal_stop): If nothing explains a signal, and the target
tells us the stop was caused by a software breakpoint, check if
there's a breakpoint instruction in the memory. If so, adjust the
PC before presenting the stop to the user. Otherwise, ignore the
trap. If nothing explains a signal, and the target tells us the
stop was caused by a hardware breakpoint, ignore the trap.
* target.h (struct target_ops) <to_stopped_by_sw_breakpoint,
to_supports_stopped_by_sw_breakpoint, to_stopped_by_hw_breakpoint,
to_supports_stopped_by_hw_breakpoint>: New fields.
(target_stopped_by_sw_breakpoint)
(target_supports_stopped_by_sw_breakpoint)
(target_stopped_by_hw_breakpoint)
(target_supports_stopped_by_hw_breakpoint): Define.
* target-delegates.c: Regenerate.
Gary stumbled on this:
(gdb) PASS: gdb.threads/thread-specific-bp.exp: all-stop: continue to end
info threads
Id Target Id Frame
* 1 Thread 0x7ffff7fdb700 (LWP 13717) "thread-specific" end () at /home/gary/work/archer/startswith/src/gdb/testsuite/gdb.threads/thread-specific-bp.c:29
(gdb) FAIL: gdb.threads/thread-specific-bp.exp: all-stop: thread start is gone
info breakpoint
The problem is that "...archer/startswith/src..." has a "start" in it,
which matches the too-lax regex in the test.
Rather than tweaking the regex, we can just remove the whole "info
threads", like we removed similar ones in other files -- GDB nowadays
does this implicitly already, so things should work without it. Thus
removing this even improves testing here a bit.
gdb/testsuite/ChangeLog:
2015-03-04 Pedro Alves <palves@redhat.com>
* gdb.threads/thread-specific-bp.exp: Delete "info threads" test.
This Linuxism has made its way into infrun.c, in the follow-fork code:
inferior_ptid = ptid_build (child_pid, child_pid, 0);
The OS-specific code should fill in the LWPID, TID parts with the
appropriate values, if any, and the core code should not be peeking at
the components of the ptids.
gdb/
2015-03-04 Pedro Alves <palves@redhat.com>
* infrun.c (follow_fork_inferior): Use the whole of the
inferior_ptid and pending_follow.related_pid ptids instead of
building ptids from the process components. Adjust verbose output
to use target_pid_to_str.
* linux-nat.c (linux_child_follow_fork): Use the whole of the
inferior_ptid and pending_follow.related_pid ptids instead of
building ptids from the process components.
These systems (OpenBSD and HP-UX 10.x) already support follow-fork
including the events needed to for "catch fork". This just makes
the upper layers realize this.
gdb/
2015-03-04 Mark Kettenis <kettenis@gnu.org>
* inf-ptrace.c [PT_GET_PROCESS_STATE]
(inf_ptrace_insert_fork_catchpoint): New function.
(inf_ptrace_remove_fork_catchpoint): New function.
(inf_ptrace_target) [PT_GET_PROCESS_STATE]: Install them.
When adding vector register support to GDB, s390_register_name() was
added to suppress the right halves of the first 16 vector registers.
However, that function returned NULL instead of an empty string in such
a case. This leads to an incomplete list of registers returned by
"complete info registers ", because completion stops at the first NULL
return value from user_reg_map_regnum_to_name().
gdb/ChangeLog:
* s390-linux-tdep.c (s390_register_name): Return empty string
instead of NULL for registers that shouldn't be visible.
On some targets each of the assignments "i = 0" in the C source for
"breakpoint-in-ro-region.exp" are compiled to a single instruction.
Then each "si" stops at the beginning of the next source line. But on
some other targets (like s390) such an assignment compiles to multiple
instructions. Then "si" may stop in mid-line, and GDB displays the PC
address in addition to the source line number. This was not considered
by the regexp for this case.
gdb/testsuite/ChangeLog:
* gdb.base/breakpoint-in-ro-region.exp (test_single_step): In the
regexps for GDB's current line display, accept a hex address
preceding the line number.
For the "multiple targets" test in catch-syscall.exp, set the 'arch1'
variable to a valid string.
gdb/testsuite/ChangeLog:
* gdb.base/catch-syscall.exp (test_catch_syscall_multi_arch): Set
the 'arch1' variable for "s390*-linux*" targets.
This patch fixes a typo that caused the wrong syscall XML file to be
used for s390x targets.
gdb/ChangeLog:
* s390-linux-tdep.c (s390_gdbarch_init): Use the correct syscall
XML file for 64-bit targets.
This fixes:
> gdb compile failed, /gdb/testsuite/gdb.threads/clone-thread_db.c: In function 'main':
> /gdb/testsuite/gdb.threads/clone-thread_db.c:67:3: warning: implicit declaration of function 'alarm' [-Wimplicit-function-declaration]
> alarm (300);
> ^
> /gdb/testsuite/gdb.threads/clone-thread_db.c:69:3: warning: implicit declaration of function 'pthread_create' [-Wimplicit-function-declaration]
> pthread_create (&child, NULL, thread_fn, NULL);
> ^
> /gdb/testsuite/gdb.threads/clone-thread_db.c:70:3: warning: implicit declaration of function 'pthread_join' [-Wimplicit-function-declaration]
> pthread_join (child);
> ^
And then adding the missing headers revealed the pthread_join call was
incorrect. This probably fixes the crash we see on ppc64be, e.g., at
https://sourceware.org/ml/gdb-testers/2015-q1/msg04415.html
the logs there show:
...
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x3fffb7ff54a0 (LWP 9275)]
0x00003fffb7f3ce74 in .pthread_join () from /lib64/libpthread.so.0
(gdb) FAIL: gdb.threads/clone-thread_db.exp: continue to end
...
Tested on x86_64 Fedora 20.
gdb/testsuite/
2015-03-04 Pedro Alves <palves@redhat.com>
* gdb.threads/clone-thread_db.c: Include unistd.h and pthread.h.
(main): Pass missing retval argument to pthread_join call.
find_default_create_inferior and find_default_attach were removed in b3ccfe11.
gdb/ChangeLog:
* target.h (find_default_create_inferior): Remove declaration.
(find_default_attach): Likewise.
In this case, we want to resume the entire process and not an
individual thread.
gdb/
2015-03-03 Pedro Alves <palves@redhat.com>
* inf-ptrace.c (inf_ptrace_resume): Remove spurious whitespace.
Use ptid_get_pid to get the overall process id when resuming all
threads.
Ref: https://sourceware.org/ml/gdb-patches/2015-03/msg00060.html
The record-btrace target can hit an assertion here:
Breakpoint 1, record_btrace_fetch_registers (ops=0x974bfc0 <record_btrace_ops>,
regcache=0x9a0a798, regno=8) at gdb/record-btrace.c:1202
1202 gdb_assert (tp != NULL);
(gdb) p regcache->ptid
$3 = {pid = 23856, lwp = 0, tid = 0}
The problem is that the linux-nat layer converts the ptid to a
single-process ptid before passing the request down to the inf-ptrace
layer, which loses information, and then record-btrace can't find the
corresponding thread in GDB's thread list:
(gdb) bt
#0 record_btrace_fetch_registers (ops=0x974bfc0 <record_btrace_ops>, regcache=0x9a0a798, regno=8)
at gdb/record-btrace.c:1202
#1 0x083f4ee2 in delegate_fetch_registers (self=0x974bfc0 <record_btrace_ops>, arg1=0x9a0a798,
arg2=8) at gdb/target-delegates.c:149
#2 0x08406562 in target_fetch_registers (regcache=0x9a0a798, regno=8)
at gdb/target.c:3279
#3 0x08355255 in regcache_raw_read (regcache=0x9a0a798, regnum=8,
buf=0xbfffe6c0 "¨\003\222\tÀ8kIøæÿ¿HO5\b\035]")
at gdb/regcache.c:643
#4 0x083558a7 in regcache_cooked_read (regcache=0x9a0a798, regnum=8,
buf=0xbfffe6c0 "¨\003\222\tÀ8kIøæÿ¿HO5\b\035]")
at gdb/regcache.c:734
#5 0x08355de3 in regcache_cooked_read_unsigned (regcache=0x9a0a798, regnum=8, val=0xbfffe738)
at gdb/regcache.c:838
#6 0x0827a106 in i386_linux_resume (ops=0x9737ca0 <linux_ops_saved>, ptid=..., step=1,
signal=GDB_SIGNAL_0) at gdb/i386-linux-nat.c:670
#7 0x08280c12 in linux_resume_one_lwp (lp=0x9a0a5b8, step=1, signo=GDB_SIGNAL_0)
at gdb/linux-nat.c:1529
#8 0x08281281 in linux_nat_resume (ops=0x98da608, ptid=..., step=1, signo=GDB_SIGNAL_0)
at gdb/linux-nat.c:1708
#9 0x0850738e in record_btrace_resume (ops=0x98da608, ptid=..., step=1, signal=GDB_SIGNAL_0)
at gdb/record-btrace.c:1760
...
The fix is just to not lose information, and let the intact ptid reach
record-btrace.c.
Tested on x86-64 Fedora 20, -m32.
gdb/ChangeLog:
2015-03-03 Pedro Alves <palves@redhat.com>
* i386-linux-nat.c (i386_linux_resume): Get the ptrace PID out of
the lwp field of ptid. Pass the full ptid to get_thread_regcache.
* inf-ptrace.c (get_ptrace_pid): New function.
(inf_ptrace_resume): Use it.
* linux-nat.c (linux_resume_one_lwp): Pass the LWP's ptid ummodified
to the lower layer.
Andy Wingo
Yao Qi
Andreas Arnez
Pedro Alves
Tom Tromey
Gary Benson
Mark Kettenis
Simon Marchi
Philippe Proulx