The rationale for this patch comes from the ROCm port [1], the goal
being to reduce the number of back and forths between GDB and the
target when doing successive operations. I'll start with explaining
the rationale and then go over the implementation. In the ROCm / GPU
world, the term "wave" is somewhat equivalent to a "thread" in GDB.
So if you read if from a GPU stand point, just s/thread/wave/.
ROCdbgapi, the library used by GDB [2] to communicate with the GPU
target, gives the illusion that it's possible for the debugger to
control (start and stop) individual threads. But in reality, this is
not how it works. Under the hood, all threads of a queue are
controlled as a group. To stop one thread in a group of running ones,
the state of all threads is retrieved from the GPU, all threads are
destroyed, and all threads but the one we want to stop are re-created
from the saved state. The net result, from the point of view of GDB,
is that the library stopped one thread. The same thing goes if we
want to resume one thread while others are running: the state of all
running threads is retrieved from the GPU, they are all destroyed, and
they are all re-created, including the thread we want to resume.
This leads to some inefficiencies when combined with how GDB works,
here are two examples:
- Stopping all threads: because the target operates in non-stop mode,
when the user interface mode is all-stop, GDB must stop all threads
individually when presenting a stop. Let's suppose we have 1000
threads and the user does ^C. GDB asks the target to stop one
thread. Behind the scenes, the library retrieves 1000 thread
states and restores the 999 others still running ones. GDB asks
the target to stop another one. The target retrieves 999 thread
states and restores the 998 remaining ones. That means that to
stop 1000 threads, we did 1000 back and forths with the GPU. It
would have been much better to just retrieve the states once and
stop there.
- Resuming with pending events: suppose the 1000 threads hit a
breakpoint at the same time. The breakpoint is conditional and
evaluates to true for the first thread, to false for all others.
GDB pulls one event (for the first thread) from the target, decides
that it should present a stop, so stops all threads using
stop_all_threads. All these other threads have a breakpoint event
to report, which is saved in `thread_info::suspend::waitstatus` for
later. When the user does "continue", GDB resumes that one thread
that did hit the breakpoint. It then processes the pending events
one by one as if they just arrived. It picks one, evaluates the
condition to false, and resumes the thread. It picks another one,
evaluates the condition to false, and resumes the thread. And so
on. In between each resumption, there is a full state retrieval
and re-creation. It would be much nicer if we could wait a little
bit before sending those threads on the GPU, until it processed all
those pending events.
To address this kind of performance issue, ROCdbgapi has a concept
called "forward progress required", which is a boolean state that
allows its user (i.e. GDB) to say "I'm doing a bunch of operations,
you can hold off putting the threads on the GPU until I'm done" (the
"forward progress not required" state). Turning forward progress back
on indicates to the library that all threads that are supposed to be
running should now be really running on the GPU.
It turns out that GDB has a similar concept, though not as general,
commit_resume. One difference is that commit_resume is not stateful:
the target can't look up "does the core need me to schedule resumed
threads for execution right now". It is also specifically linked to
the resume method, it is not used in other contexts. The target
accumulates resumption requests through target_ops::resume calls, and
then commits those resumptions when target_ops::commit_resume is
called. The target has no way to check if it's ok to leave resumed
threads stopped in other target methods.
To bridge the gap, this patch generalizes the commit_resume concept in
GDB to match the forward progress concept of ROCdbgapi. The current
name (commit_resume) can be interpreted as "commit the previous resume
calls". I renamed the concept to "commit_resumed", as in "commit the
threads that are resumed".
In the new version, we have two things:
- the commit_resumed_state field in process_stratum_target: indicates
whether GDB requires target stacks using this target to have
resumed threads committed to the execution target/device. If
false, an execution target is allowed to leave resumed threads
un-committed at the end of whatever method it is executing.
- the commit_resumed target method: called when commit_resumed_state
transitions from false to true. While commit_resumed_state was
false, the target may have left some resumed threads un-committed.
This method being called tells it that it should commit them back
to the execution device.
Let's take the "Stopping all threads" scenario from above and see how
it would work with the ROCm target with this change. Before stopping
all threads, GDB would set the target's commit_resumed_state field to
false. It would then ask the target to stop the first thread. The
target would retrieve all threads' state from the GPU and mark that
one as stopped. Since commit_resumed_state is false, it leaves all
the other threads (still resumed) stopped. GDB would then proceed to
call target_stop for all the other threads. Since resumed threads are
not committed, this doesn't do any back and forth with the GPU.
To simplify the implementation of targets, this patch makes it so that
when calling certain target methods, the contract between the core and
the targets guarantees that commit_resumed_state is false. This way,
the target doesn't need two paths, one for commit_resumed_state ==
true and one for commit_resumed_state == false. It can just assert
that commit_resumed_state is false and work with that assumption.
This also helps catch places where we forgot to disable
commit_resumed_state before calling the method, which represents a
probable optimization opportunity. The commit adds assertions in the
target method wrappers (target_resume and friends) to have some
confidence that this contract between the core and the targets is
respected.
The scoped_disable_commit_resumed type is used to disable the commit
resumed state of all process targets on construction, and selectively
re-enable it on destruction (see below for criteria). Note that it
only sets the process_stratum_target::commit_resumed_state flag. A
subsequent call to maybe_call_commit_resumed_all_targets is necessary
to call the commit_resumed method on all target stacks with process
targets that got their commit_resumed_state flag turned back on. This
separation is because we don't want to call the commit_resumed methods
in scoped_disable_commit_resumed's destructor, as they may throw.
On destruction, commit-resumed is not re-enabled for a given target
if:
1. this target has no threads resumed, or
2. this target has at least one resumed thread with a pending status
known to the core (saved in thread_info::suspend::waitstatus).
The first point is not technically necessary, because a proper
commit_resumed implementation would be a no-op if the target has no
resumed threads. But since we have a flag do to a quick check, it
shouldn't hurt.
The second point is more important: together with the
scoped_disable_commit_resumed instance added in fetch_inferior_event,
it makes it so the "Resuming with pending events" described above is
handled efficiently. Here's what happens in that case:
1. The user types "continue".
2. Upon destruction, the scoped_disable_commit_resumed in the
`proceed` function does not enable commit-resumed, as it sees some
threads have pending statuses.
3. fetch_inferior_event is called to handle another event, the
breakpoint hit evaluates to false, and that thread is resumed.
Because there are still more threads with pending statuses, the
destructor of scoped_disable_commit_resumed in
fetch_inferior_event still doesn't enable commit-resumed.
4. Rinse and repeat step 3, until the last pending status is handled
by fetch_inferior_event. In that case,
scoped_disable_commit_resumed's destructor sees there are no more
threads with pending statues, so it asks the target to commit
resumed threads.
This allows us to avoid all unnecessary back and forths, there is a
single commit_resumed call once all pending statuses are processed.
This change required remote_target::remote_stop_ns to learn how to
handle stopping threads that were resumed but pending vCont. The
simplest example where that happens is when using the remote target in
all-stop, but with "maint set target-non-stop on", to force it to
operate in non-stop mode under the hood. If two threads hit a
breakpoint at the same time, GDB will receive two stop replies. It
will present the stop for one thread and save the other one in
thread_info::suspend::waitstatus.
Before this patch, when doing "continue", GDB first resumes the thread
without a pending status:
Sending packet: $vCont;c:p172651.172676#f3
It then consumes the pending status in the next fetch_inferior_event
call:
[infrun] do_target_wait_1: Using pending wait status status->kind = stopped, signal = GDB_SIGNAL_TRAP for Thread 1517137.1517137.
[infrun] target_wait (-1.0.0, status) =
[infrun] 1517137.1517137.0 [Thread 1517137.1517137],
[infrun] status->kind = stopped, signal = GDB_SIGNAL_TRAP
It then realizes it needs to stop all threads to present the stop, so
stops the thread it just resumed:
[infrun] stop_all_threads: Thread 1517137.1517137 not executing
[infrun] stop_all_threads: Thread 1517137.1517174 executing, need stop
remote_stop called
Sending packet: $vCont;t:p172651.172676#04
This is an unnecessary resume/stop. With this patch, we don't commit
resumed threads after proceeding, because of the pending status:
[infrun] maybe_commit_resumed_all_process_targets: not requesting commit-resumed for target extended-remote, a thread has a pending waitstatus
When GDB handles the pending status and stop_all_threads runs, we stop a
resumed but pending vCont thread:
remote_stop_ns: Enqueueing phony stop reply for thread pending vCont-resume (1520940, 1520976, 0)
That thread was never actually resumed on the remote stub / gdbserver,
so we shouldn't send a packet to the remote side asking to stop the
thread.
Note that there are paths that resume the target and then do a
synchronous blocking wait, in sort of nested event loop, via
wait_sync_command_done. For example, inferior function calls, or any
run control command issued from a breakpoint command list. We handle
that making wait_sync_command_one a "sync" point -- force forward
progress, or IOW, force-enable commit-resumed state.
gdb/ChangeLog:
yyyy-mm-dd Simon Marchi <simon.marchi@efficios.com>
Pedro Alves <pedro@palves.net>
* infcmd.c (run_command_1, attach_command, detach_command)
(interrupt_target_1): Use scoped_disable_commit_resumed.
* infrun.c (do_target_resume): Remove
target_commit_resume call.
(commit_resume_all_targets): Remove.
(maybe_set_commit_resumed_all_targets): New.
(maybe_call_commit_resumed_all_targets): New.
(enable_commit_resumed): New.
(scoped_disable_commit_resumed::scoped_disable_commit_resumed)
(scoped_disable_commit_resumed::~scoped_disable_commit_resumed)
(scoped_disable_commit_resumed::reset)
(scoped_disable_commit_resumed::reset_and_commit)
(scoped_enable_commit_resumed::scoped_enable_commit_resumed)
(scoped_enable_commit_resumed::~scoped_enable_commit_resumed):
New.
(proceed): Use scoped_disable_commit_resumed and
maybe_call_commit_resumed_all_targets.
(fetch_inferior_event): Use scoped_disable_commit_resumed.
* infrun.h (struct scoped_disable_commit_resumed): New.
(maybe_call_commit_resumed_all_process_targets): New.
(struct scoped_enable_commit_resumed): New.
* mi/mi-main.c (exec_continue): Use scoped_disable_commit_resumed.
* process-stratum-target.h (class process_stratum_target):
<commit_resumed_state>: New.
* record-full.c (record_full_wait_1): Change commit_resumed_state
around calling commit_resumed.
* remote.c (class remote_target) <commit_resume>: Rename to...
<commit_resumed>: ... this.
(struct stop_reply): Move up.
(remote_target::commit_resume): Rename to...
(remote_target::commit_resumed): ... this. Check if there is any
thread pending vCont resume.
(remote_target::remote_stop_ns): Generate stop replies for resumed
but pending vCont threads.
(remote_target::wait_ns): Add gdb_assert.
* target-delegates.c: Regenerate.
* target.c (target_wait, target_resume): Assert that the current
process_stratum target isn't in commit-resumed state.
(defer_target_commit_resume): Remove.
(target_commit_resume): Remove.
(target_commit_resumed): New.
(make_scoped_defer_target_commit_resume): Remove.
(target_stop): Assert that the current process_stratum target
isn't in commit-resumed state.
* target.h (struct target_ops) <commit_resume>: Rename to ...
<commit_resumed>: ... this.
(target_commit_resume): Remove.
(target_commit_resumed): New.
(make_scoped_defer_target_commit_resume): Remove.
* top.c (wait_sync_command_done): Use
scoped_enable_commit_resumed.
[1] https://github.com/ROCm-Developer-Tools/ROCgdb/
[2] https://github.com/ROCm-Developer-Tools/ROCdbgapi
Change-Id: I836135531a29214b21695736deb0a81acf8cf566
This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
README for GDB release
This is GDB, the GNU source-level debugger.
A summary of new features is in the file `gdb/NEWS'.
Check the GDB home page at http://www.gnu.org/software/gdb/ for up to
date release information, mailing list links and archives, etc.
GDB's bug tracking data base can be found at
http://www.gnu.org/software/gdb/bugs/
Unpacking and Installation -- quick overview
==========================
The release is provided as a gzipped tar file called
'gdb-VERSION.tar.gz', where VERSION is the version of GDB.
The GDB debugger sources, the generic GNU include
files, the BFD ("binary file description") library, the readline
library, and other libraries all have directories of their own
underneath the gdb-VERSION directory. The idea is that a variety of GNU
tools can share a common copy of these things. Be aware of variation
over time--for example don't try to build GDB with a copy of bfd from
a release other than the GDB release (such as a binutils release),
especially if the releases are more than a few weeks apart.
Configuration scripts and makefiles exist to cruise up and down this
directory tree and automatically build all the pieces in the right
order.
When you unpack the gdb-VERSION.tar.gz file, it will create a
source directory called `gdb-VERSION'.
You can build GDB right in the source directory:
cd gdb-VERSION
./configure --prefix=/usr/local (or wherever you want)
make all install
However, we recommend that an empty directory be used instead.
This way you do not clutter your source tree with binary files
and will be able to create different builds with different
configuration options.
You can build GDB in any empty build directory:
mkdir build
cd build
<full path to your sources>/gdb-VERSION/configure [etc...]
make all install
(Building GDB with DJGPP tools for MS-DOS/MS-Windows is slightly
different; see the file gdb-VERSION/gdb/config/djgpp/README for details.)
This will configure and build all the libraries as well as GDB. If
`configure' can't determine your system type, specify one as its
argument, e.g., `./configure sun4' or `./configure decstation'.
Make sure that your 'configure' line ends in 'gdb-VERSION/configure':
/berman/migchain/source/gdb-VERSION/configure # RIGHT
/berman/migchain/source/gdb-VERSION/gdb/configure # WRONG
The GDB package contains several subdirectories, such as 'gdb',
'bfd', and 'readline'. If your 'configure' line ends in
'gdb-VERSION/gdb/configure', then you are configuring only the gdb
subdirectory, not the whole GDB package. This leads to build errors
such as:
make: *** No rule to make target `../bfd/bfd.h', needed by `gdb.o'. Stop.
If you get other compiler errors during this stage, see the `Reporting
Bugs' section below; there are a few known problems.
GDB's `configure' script has many options to enable or disable
different features or dependencies. These options are not generally
known to the top-level `configure', so if you want to see a complete
list of options, invoke the subdirectory `configure', like:
/berman/migchain/source/gdb-VERSION/gdb/configure --help
(Take note of how this differs from the invocation used to actually
configure the build tree.)
GDB requires a C++11 compiler. If you do not have a
C++11 compiler for your system, you may be able to download and install
the GNU CC compiler. It is available via anonymous FTP from the
directory `ftp://ftp.gnu.org/pub/gnu/gcc'. GDB also requires an ISO
C standard library. The GDB remote server, GDBserver, builds with some
non-ISO standard libraries - e.g. for Windows CE.
GDB can optionally be built against various external libraries.
These dependencies are described below in the "`configure options"
section of this README.
GDB can be used as a cross-debugger, running on a machine of one
type while debugging a program running on a machine of another type.
See below.
More Documentation
******************
All the documentation for GDB comes as part of the machine-readable
distribution. The documentation is written in Texinfo format, which
is a documentation system that uses a single source file to produce
both on-line information and a printed manual. You can use one of the
Info formatting commands to create the on-line version of the
documentation and TeX (or `texi2roff') to typeset the printed version.
GDB includes an already formatted copy of the on-line Info version
of this manual in the `gdb/doc' subdirectory. The main Info file is
`gdb-VERSION/gdb/doc/gdb.info', and it refers to subordinate files
matching `gdb.info*' in the same directory. If necessary, you can
print out these files, or read them with any editor; but they are
easier to read using the `info' subsystem in GNU Emacs or the
standalone `info' program, available as part of the GNU Texinfo
distribution.
If you want to format these Info files yourself, you need one of the
Info formatting programs, such as `texinfo-format-buffer' or
`makeinfo'.
If you have `makeinfo' installed, and are in the top level GDB
source directory (`gdb-VERSION'), you can make the Info file by
typing:
cd gdb/doc
make info
If you want to typeset and print copies of this manual, you need
TeX, a program to print its DVI output files, and `texinfo.tex', the
Texinfo definitions file. This file is included in the GDB
distribution, in the directory `gdb-VERSION/texinfo'.
TeX is a typesetting program; it does not print files directly, but
produces output files called DVI files. To print a typeset document,
you need a program to print DVI files. If your system has TeX
installed, chances are it has such a program. The precise command to
use depends on your system; `lpr -d' is common; another (for PostScript
devices) is `dvips'. The DVI print command may require a file name
without any extension or a `.dvi' extension.
TeX also requires a macro definitions file called `texinfo.tex'.
This file tells TeX how to typeset a document written in Texinfo
format. On its own, TeX cannot read, much less typeset a Texinfo file.
`texinfo.tex' is distributed with GDB and is located in the
`gdb-VERSION/texinfo' directory.
If you have TeX and a DVI printer program installed, you can typeset
and print this manual. First switch to the `gdb' subdirectory of
the main source directory (for example, to `gdb-VERSION/gdb') and then type:
make doc/gdb.dvi
If you prefer to have the manual in PDF format, type this from the
`gdb/doc' subdirectory of the main source directory:
make gdb.pdf
For this to work, you will need the PDFTeX package to be installed.
Installing GDB
**************
GDB comes with a `configure' script that automates the process of
preparing GDB for installation; you can then use `make' to build the
`gdb' program.
The GDB distribution includes all the source code you need for GDB in
a single directory. That directory contains:
`gdb-VERSION/{COPYING,COPYING.LIB}'
Standard GNU license files. Please read them.
`gdb-VERSION/bfd'
source for the Binary File Descriptor library
`gdb-VERSION/config*'
script for configuring GDB, along with other support files
`gdb-VERSION/gdb'
the source specific to GDB itself
`gdb-VERSION/include'
GNU include files
`gdb-VERSION/libiberty'
source for the `-liberty' free software library
`gdb-VERSION/opcodes'
source for the library of opcode tables and disassemblers
`gdb-VERSION/readline'
source for the GNU command-line interface
NOTE: The readline library is compiled for use by GDB, but will
not be installed on your system when "make install" is issued.
`gdb-VERSION/sim'
source for some simulators (ARM, D10V, SPARC, M32R, MIPS, PPC, V850, etc)
`gdb-VERSION/texinfo'
The `texinfo.tex' file, which you need in order to make a printed
manual using TeX.
`gdb-VERSION/etc'
Coding standards, useful files for editing GDB, and other
miscellanea.
Note: the following instructions are for building GDB on Unix or
Unix-like systems. Instructions for building with DJGPP for
MS-DOS/MS-Windows are in the file gdb/config/djgpp/README.
The simplest way to configure and build GDB is to run `configure'
from the `gdb-VERSION' directory.
First switch to the `gdb-VERSION' source directory if you are
not already in it; then run `configure'.
For example:
cd gdb-VERSION
./configure
make
Running `configure' followed by `make' builds the `bfd',
`readline', `mmalloc', and `libiberty' libraries, then `gdb' itself.
The configured source files, and the binaries, are left in the
corresponding source directories.
`configure' is a Bourne-shell (`/bin/sh') script; if your system
does not recognize this automatically when you run a different shell,
you may need to run `sh' on it explicitly:
sh configure
If you run `configure' from a directory that contains source
directories for multiple libraries or programs, `configure' creates
configuration files for every directory level underneath (unless
you tell it not to, with the `--norecursion' option).
You can install `gdb' anywhere; it has no hardwired paths. However,
you should make sure that the shell on your path (named by the `SHELL'
environment variable) is publicly readable. Remember that GDB uses the
shell to start your program--some systems refuse to let GDB debug child
processes whose programs are not readable.
Compiling GDB in another directory
==================================
If you want to run GDB versions for several host or target machines,
you need a different `gdb' compiled for each combination of host and
target. `configure' is designed to make this easy by allowing you to
generate each configuration in a separate subdirectory, rather than in
the source directory. If your `make' program handles the `VPATH'
feature correctly (GNU `make' and SunOS 'make' are two that should),
running `make' in each of these directories builds the `gdb' program
specified there.
To build `gdb' in a separate directory, run `configure' with the
`--srcdir' option to specify where to find the source. (You also need
to specify a path to find `configure' itself from your working
directory. If the path to `configure' would be the same as the
argument to `--srcdir', you can leave out the `--srcdir' option; it
will be assumed.)
For example, you can build GDB in a separate
directory for a Sun 4 like this:
cd gdb-VERSION
mkdir ../gdb-sun4
cd ../gdb-sun4
../gdb-VERSION/configure
make
When `configure' builds a configuration using a remote source
directory, it creates a tree for the binaries with the same structure
(and using the same names) as the tree under the source directory. In
the example, you'd find the Sun 4 library `libiberty.a' in the
directory `gdb-sun4/libiberty', and GDB itself in `gdb-sun4/gdb'.
One popular reason to build several GDB configurations in separate
directories is to configure GDB for cross-compiling (where GDB runs on
one machine--the host--while debugging programs that run on another
machine--the target). You specify a cross-debugging target by giving
the `--target=TARGET' option to `configure'.
When you run `make' to build a program or library, you must run it
in a configured directory--whatever directory you were in when you
called `configure' (or one of its subdirectories).
The `Makefile' that `configure' generates in each source directory
also runs recursively. If you type `make' in a source directory such
as `gdb-VERSION' (or in a separate configured directory configured with
`--srcdir=PATH/gdb-VERSION'), you will build all the required libraries,
and then build GDB.
When you have multiple hosts or targets configured in separate
directories, you can run `make' on them in parallel (for example, if
they are NFS-mounted on each of the hosts); they will not interfere
with each other.
Specifying names for hosts and targets
======================================
The specifications used for hosts and targets in the `configure'
script are based on a three-part naming scheme, but some short
predefined aliases are also supported. The full naming scheme encodes
three pieces of information in the following pattern:
ARCHITECTURE-VENDOR-OS
For example, you can use the alias `sun4' as a HOST argument or in a
`--target=TARGET' option. The equivalent full name is
`sparc-sun-sunos4'.
The `configure' script accompanying GDB does not provide any query
facility to list all supported host and target names or aliases.
`configure' calls the Bourne shell script `config.sub' to map
abbreviations to full names; you can read the script, if you wish, or
you can use it to test your guesses on abbreviations--for example:
% sh config.sub sun4
sparc-sun-sunos4.1.1
% sh config.sub sun3
m68k-sun-sunos4.1.1
% sh config.sub decstation
mips-dec-ultrix4.2
% sh config.sub hp300bsd
m68k-hp-bsd
% sh config.sub i386v
i386-pc-sysv
% sh config.sub i786v
Invalid configuration `i786v': machine `i786v' not recognized
`config.sub' is also distributed in the GDB source directory.
`configure' options
===================
Here is a summary of the `configure' options and arguments that are
most often useful for building GDB. `configure' also has several other
options not listed here. There are many options to gdb's `configure'
script, some of which are only useful in special situation.
*note : (autoconf.info)Running configure scripts, for a full
explanation of `configure'.
configure [--help]
[--prefix=DIR]
[--srcdir=PATH]
[--target=TARGET]
[--host=HOST]
[HOST]
You may introduce options with a single `-' rather than `--' if you
prefer; but you may abbreviate option names if you use `--'. Some
more obscure GDB `configure' options are not listed here.
`--help'
Display a quick summary of how to invoke `configure'.
`-prefix=DIR'
Configure the source to install programs and files under directory
`DIR'.
`--srcdir=PATH'
*Warning: using this option requires GNU `make', or another `make'
that compatibly implements the `VPATH' feature.*
Use this option to make configurations in directories separate
from the GDB source directories. Among other things, you can use
this to build (or maintain) several configurations simultaneously,
in separate directories. `configure' writes configuration
specific files in the current directory, but arranges for them to
use the source in the directory PATH. `configure' will create
directories under the working directory in parallel to the source
directories below PATH.
`--host=HOST'
Configure GDB to run on the specified HOST.
There is no convenient way to generate a list of all available
hosts.
`HOST ...'
Same as `--host=HOST'. If you omit this, GDB will guess; it's
quite accurate.
`--target=TARGET'
Configure GDB for cross-debugging programs running on the specified
TARGET. Without this option, GDB is configured to debug programs
that run on the same machine (HOST) as GDB itself.
There is no convenient way to generate a list of all available
targets.
`--enable-targets=TARGET,TARGET,...'
`--enable-targets=all`
Configure GDB for cross-debugging programs running on the
specified list of targets. The special value `all' configures
GDB for debugging programs running on any target it supports.
`--with-gdb-datadir=PATH'
Set the GDB-specific data directory. GDB will look here for
certain supporting files or scripts. This defaults to the `gdb'
subdirectory of `datadir' (which can be set using `--datadir').
`--with-relocated-sources=DIR'
Sets up the default source path substitution rule so that
directory names recorded in debug information will be
automatically adjusted for any directory under DIR. DIR should
be a subdirectory of GDB's configured prefix, the one mentioned
in the `--prefix' or `--exec-prefix' options to configure. This
option is useful if GDB is supposed to be moved to a different
place after it is built.
`--enable-64-bit-bfd'
Enable 64-bit support in BFD on 32-bit hosts.
`--disable-gdbmi'
Build GDB without the GDB/MI machine interface.
`--enable-tui'
Build GDB with the text-mode full-screen user interface (TUI).
Requires a curses library (ncurses and cursesX are also
supported).
`--with-curses'
Use the curses library instead of the termcap library, for
text-mode terminal operations.
`--with-debuginfod'
Build GDB with libdebuginfod, the debuginfod client library. Used
to automatically fetch source files and separate debug files from
debuginfod servers using the associated executable's build ID.
Enabled by default if libdebuginfod is installed and found at
configure time. debuginfod is packaged with elfutils, starting
with version 0.178. You can get the latest version from
'https://sourceware.org/elfutils/'.
`--with-libunwind-ia64'
Use the libunwind library for unwinding function call stack on ia64
target platforms.
See http://www.nongnu.org/libunwind/index.html for details.
`--with-system-readline'
Use the readline library installed on the host, rather than the
library supplied as part of GDB. Readline 7 or newer is required;
this is enforced by the build system.
`--with-system-zlib
Use the zlib library installed on the host, rather than the
library supplied as part of GDB.
`--with-expat'
Build GDB with Expat, a library for XML parsing. (Done by
default if libexpat is installed and found at configure time.)
This library is used to read XML files supplied with GDB. If it
is unavailable, some features, such as remote protocol memory
maps, target descriptions, and shared library lists, that are
based on XML files, will not be available in GDB. If your host
does not have libexpat installed, you can get the latest version
from `http://expat.sourceforge.net'.
`--with-libiconv-prefix[=DIR]'
Build GDB with GNU libiconv, a character set encoding conversion
library. This is not done by default, as on GNU systems the
`iconv' that is built in to the C library is sufficient. If your
host does not have a working `iconv', you can get the latest
version of GNU iconv from `https://www.gnu.org/software/libiconv/'.
GDB's build system also supports building GNU libiconv as part of
the overall build. See the GDB manual instructions on how to do
this.
`--with-lzma'
Build GDB with LZMA, a compression library. (Done by default if
liblzma is installed and found at configure time.) LZMA is used
by GDB's "mini debuginfo" feature, which is only useful on
platforms using the ELF object file format. If your host does
not have liblzma installed, you can get the latest version from
`https://tukaani.org/xz/'.
`--with-libgmp-prefix=DIR'
Build GDB using the GMP library installed at the directory DIR.
If your host does not have GMP installed, you can get the latest
version at `https://gmplib.org/'.
`--with-mpfr'
Build GDB with GNU MPFR, a library for multiple-precision
floating-point computation with correct rounding. (Done by
default if GNU MPFR is installed and found at configure time.)
This library is used to emulate target floating-point arithmetic
during expression evaluation when the target uses different
floating-point formats than the host. If GNU MPFR is not
available, GDB will fall back to using host floating-point
arithmetic. If your host does not have GNU MPFR installed, you
can get the latest version from `https://www.mpfr.org/'.
`--with-python[=PYTHON]'
Build GDB with Python scripting support. (Done by default if
libpython is present and found at configure time.) Python makes
GDB scripting much more powerful than the restricted CLI
scripting language. If your host does not have Python installed,
you can find it on `http://www.python.org/download/'. The oldest
version of Python supported by GDB is 2.6. The optional argument
PYTHON is used to find the Python headers and libraries. It can
be either the name of a Python executable, or the name of the
directory in which Python is installed.
`--with-guile[=GUILE]'
Build GDB with GNU Guile scripting support. (Done by default if
libguile is present and found at configure time.) If your host
does not have Guile installed, you can find it at
`https://www.gnu.org/software/guile/'. The optional argument
GUILE can be a version number, which will cause `configure' to
try to use that version of Guile; or the file name of a
`pkg-config' executable, which will be queried to find the
information needed to compile and link against Guile.
`--enable-source-highlight'
When printing source code, use source highlighting. This requires
libsource-highlight to be installed and is enabled by default
if the library is found.
`--with-xxhash'
Use libxxhash for hashing. This has no user-visible effect but
speeds up various GDB operations such as symbol loading. Enabled
by default if libxxhash is found.
`--without-included-regex'
Don't use the regex library included with GDB (as part of the
libiberty library). This is the default on hosts with version 2
of the GNU C library.
`--with-sysroot=DIR'
Use DIR as the default system root directory for libraries whose
file names begin with `/lib' or `/usr/lib'. (The value of DIR
can be modified at run time by using the "set sysroot" command.)
If DIR is under the GDB configured prefix (set with `--prefix' or
`--exec-prefix' options), the default system root will be
automatically adjusted if and when GDB is moved to a different
location.
`--with-system-gdbinit=FILE'
Configure GDB to automatically load a system-wide init file.
FILE should be an absolute file name. If FILE is in a directory
under the configured prefix, and GDB is moved to another location
after being built, the location of the system-wide init file will
be adjusted accordingly.
`--with-system-gdbinit-dir=DIR'
Configure GDB to automatically load system-wide init files from
a directory. Files with extensions `.gdb', `.py' (if Python
support is enabled) and `.scm' (if Guile support is enabled) are
supported. DIR should be an absolute directory name. If DIR is
in a directory under the configured prefix, and GDB is moved to
another location after being built, the location of the system-
wide init directory will be adjusted accordingly.
`--enable-build-warnings'
When building the GDB sources, ask the compiler to warn about any
code which looks even vaguely suspicious. It passes many
different warning flags, depending on the exact version of the
compiler you are using.
`--enable-werror'
Treat compiler warnings as werrors. It adds the -Werror flag to
the compiler, which will fail the compilation if the compiler
outputs any warning messages.
`--enable-ubsan'
Enable the GCC undefined behavior sanitizer. By default this is
disabled in GDB releases, but enabled when building from git.
The undefined behavior sanitizer checks for C++ undefined
behavior. It has a performance cost, so if you are looking at
GDB's performance, you should disable it.
`--enable-unit-tests[=yes|no]'
Enable (i.e., include) support for unit tests when compiling GDB
and GDBServer. Note that if this option is not passed, GDB will
have selftests if it is a development build, and will *not* have
selftests if it is a non-development build.
`configure' accepts other options, for compatibility with configuring
other GNU tools recursively.
Remote debugging
=================
The files m68k-stub.c, i386-stub.c, and sparc-stub.c are examples
of remote stubs to be used with remote.c. They are designed to run
standalone on an m68k, i386, or SPARC cpu and communicate properly
with the remote.c stub over a serial line.
The directory gdbserver/ contains `gdbserver', a program that
allows remote debugging for Unix applications. GDBserver is only
supported for some native configurations.
The file gdbserver/README includes further notes on GDBserver; in
particular, it explains how to build GDBserver for cross-debugging
(where GDBserver runs on the target machine, which is of a different
architecture than the host machine running GDB).
Reporting Bugs in GDB
=====================
There are several ways of reporting bugs in GDB. The prefered
method is to use the World Wide Web:
http://www.gnu.org/software/gdb/bugs/
As an alternative, the bug report can be submitted, via e-mail, to the
address "bug-gdb@gnu.org".
When submitting a bug, please include the GDB version number, and
how you configured it (e.g., "sun4" or "mach386 host,
i586-intel-synopsys target"). Since GDB supports so many
different configurations, it is important that you be precise about
this. The simplest way to do this is to include the output from these
commands:
% gdb --version
% gdb --config
For more information on how/whether to report bugs, see the
Reporting Bugs chapter of the GDB manual (gdb/doc/gdb.texinfo).
Graphical interface to GDB -- X Windows, MS Windows
==========================
Several graphical interfaces to GDB are available. You should
check:
https://sourceware.org/gdb/wiki/GDB%20Front%20Ends
for an up-to-date list.
Emacs users will very likely enjoy the Grand Unified Debugger mode;
try typing `M-x gdb RET'.
Writing Code for GDB
=====================
There is information about writing code for GDB in the file
`CONTRIBUTE' and at the website:
http://www.gnu.org/software/gdb/
in particular in the wiki.
If you are pondering writing anything but a short patch, especially
take note of the information about copyrights and copyright assignment.
It can take quite a while to get all the paperwork done, so
we encourage you to start that process as soon as you decide you are
planning to work on something, or at least well ahead of when you
think you will be ready to submit the patches.
GDB Testsuite
=============
Included with the GDB distribution is a DejaGNU based testsuite
that can either be used to test your newly built GDB, or for
regression testing a GDB with local modifications.
Running the testsuite requires the prior installation of DejaGNU,
which is generally available via ftp. The directory
ftp://sources.redhat.com/pub/dejagnu/ will contain a recent snapshot.
Once DejaGNU is installed, you can run the tests in one of the
following ways:
(1) cd gdb-VERSION
make check-gdb
or
(2) cd gdb-VERSION/gdb
make check
or
(3) cd gdb-VERSION/gdb/testsuite
make site.exp (builds the site specific file)
runtest -tool gdb GDB=../gdb (or GDB=<somepath> as appropriate)
When using a `make'-based method, you can use the Makefile variable
`RUNTESTFLAGS' to pass flags to `runtest', e.g.:
make RUNTESTFLAGS=--directory=gdb.cp check
If you use GNU make, you can use its `-j' option to run the testsuite
in parallel. This can greatly reduce the amount of time it takes for
the testsuite to run. In this case, if you set `RUNTESTFLAGS' then,
by default, the tests will be run serially even under `-j'. You can
override this and force a parallel run by setting the `make' variable
`FORCE_PARALLEL' to any non-empty value. Note that the parallel `make
check' assumes that you want to run the entire testsuite, so it is not
compatible with some dejagnu options, like `--directory'.
The last method gives you slightly more control in case of problems
with building one or more test executables or if you are using the
testsuite `standalone', without it being part of the GDB source tree.
See the DejaGNU documentation for further details.
Copyright and License Notices
=============================
Most files maintained by the GDB Project contain a copyright notice
as well as a license notice, usually at the start of the file.
To reduce the length of copyright notices, consecutive years in the
copyright notice can be combined into a single range. For instance,
the following list of copyright years...
1986, 1988, 1989, 1991-1993, 1999, 2000, 2007, 2008, 2009, 2010, 2011
... is abbreviated into:
1986, 1988-1989, 1991-1993, 1999-2000, 2007-2011
Every year of each range, inclusive, is a copyrightable year that
could be listed individually.
(this is for editing this file with GNU emacs)
Local Variables:
mode: text
End:
1192f124a3