Patchwork [doc] compile: missing bits [Re: [PATCH v2 7/9] compile: New 'compile print']

login
register
mail settings
Submitter Jan Kratochvil
Date April 10, 2015, 8:09 a.m.
Message ID <20150410080933.GA8497@host1.jankratochvil.net>
Download mbox | patch
Permalink /patch/6124/
State New
Headers show

Comments

Jan Kratochvil - April 10, 2015, 8:09 a.m.
On Thu, 09 Apr 2015 20:12:21 +0200, Eli Zaretskii wrote:
> > +@value{NGCC} (since version 4.7) stores the options used for compilation
> > +into @code{DW_AT_producer} part of DWARF debugging information according
> > +to the @value{NGCC} option @code{-grecord-gcc-switches}.  One has to
> > +explicitly specify @code{-g} during inferior compilation otherwise
> > +@value{NGCC} produces no DWARF.
> 
> Hmm...  Are you sure -g always produces DWARF?  Maybe we should say
> that for targets where DWARF is not the default, -gdwarf-N should be
> used, and also that -gstabs etc. should _not_ be used?

While you are technically right I do not think it is appropriate to mention it
that way.

When some platform is still using -gstabs by default (is there any?) there are
probably some serious reasons for it - such as some old debugging / crash
analysing tools incompatible with DWARF.  Using -gdwarf-X may do more harm
than good there IMO; unless one really knows what s/he is doing and then one
does not need to read this doc.

I have just added:
	This feature is only relevant for 
	platforms where @code{-g} produces DWARF by default, otherwise one may
	try to enforce DWARF by using @code{-gdwarf-4}.


> > +Specifically @code{PATH} is searched for binaries matching regular expression
> > +@code{@var{ARCH}(-[^-]*)?-@var{OS}-gcc} according to the inferior target being
> > +debugged.  @code{@var{ARCH}} is processor name - multiarch is supported,
[...]
> > +for pattern @code{s390x?}.  @code{OS} is currently supported only for
> >                              ^^^^^^^^^
> 
> @var{os}.

Not sure if @code{@var{os}} is the same as @var{os}, used only @var{os}.


Thanks,
Jan
gdb/doc/
2015-04-10  Jan Kratochvil  <jan.kratochvil@redhat.com>
	    Eli Zaretskii  <eliz@gnu.org>

	* gdb.texinfo (Compiling and Injecting Code): Describe set debug 
	compile, show debug compile.  New subsection Compilation options for 
	the compile command.  New subsection Compiler search for the compile
	command.
Eli Zaretskii - April 10, 2015, 8:31 a.m.
> Date: Fri, 10 Apr 2015 10:09:33 +0200
> From: Jan Kratochvil <jan.kratochvil@redhat.com>
> Cc: Paul_Koning@dell.com, gdb-patches@sourceware.org, pmuldoon@redhat.com
> 
> I have just added:
> 	This feature is only relevant for 
> 	platforms where @code{-g} produces DWARF by default, otherwise one may
> 	try to enforce DWARF by using @code{-gdwarf-4}.

That's good enough, thanks.

> > > +Specifically @code{PATH} is searched for binaries matching regular expression
> > > +@code{@var{ARCH}(-[^-]*)?-@var{OS}-gcc} according to the inferior target being
> > > +debugged.  @code{@var{ARCH}} is processor name - multiarch is supported,
> [...]
> > > +for pattern @code{s390x?}.  @code{OS} is currently supported only for
> > >                              ^^^^^^^^^
> > 
> > @var{os}.
> 
> Not sure if @code{@var{os}} is the same as @var{os}, used only @var{os}.

@var{os} without @code was what I meant, sorry for being unclear.

Thanks, this version is fine with me.
Jan Kratochvil - April 10, 2015, 8:35 a.m.
On Fri, 10 Apr 2015 10:31:02 +0200, Eli Zaretskii wrote:
> Thanks, this version is fine with me.

Checked in:
	e7a8570f75694284ea5bc909f095004f4ef938c6


Jan
Paul_Koning@Dell.com - April 10, 2015, 2:08 p.m.
> On Apr 10, 2015, at 4:09 AM, Jan Kratochvil <jan.kratochvil@redhat.com> wrote:

> 

> ...

> When some platform is still using -gstabs by default (is there any?) ...


In gcc, yes.  But the one I know of isn’t supported in gdb.

	paul

Patch

diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
index d794893..419439a 100644
--- a/gdb/doc/gdb.texinfo
+++ b/gdb/doc/gdb.texinfo
@@ -17223,6 +17223,66 @@  compile file /home/user/example.c
 @end smallexample
 @end table
 
+@noindent
+The process of compiling and injecting the code can be inspected using:
+
+@table @code
+@anchor{set debug compile}
+@item set debug compile
+@cindex compile command debugging info
+Turns on or off display of @value{GDBN} process of compiling and
+injecting the code.  The default is off.
+
+@item show debug compile
+Displays the current state of displaying @value{GDBN} process of
+compiling and injecting the code.
+@end table
+
+@subsection Compilation options for the @code{compile} command
+
+@value{GDBN} needs to specify the right compilation options for the code
+to be injected, in part to make its ABI compatible with the inferior
+and in part to make the injected code compatible with @value{GDBN}'s
+injecting process.
+
+@noindent
+The options used, in increasing precedence:
+
+@table @asis
+@item target architecture and OS options (@code{gdbarch})
+These options depend on target processor type and target operating
+system, usually they specify at least 32-bit (@code{-m32}) or 64-bit
+(@code{-m64}) compilation option.
+
+@item compilation options recorded in the target
+@value{NGCC} (since version 4.7) stores the options used for compilation
+into @code{DW_AT_producer} part of DWARF debugging information according
+to the @value{NGCC} option @code{-grecord-gcc-switches}.  One has to
+explicitly specify @code{-g} during inferior compilation otherwise
+@value{NGCC} produces no DWARF.  This feature is only relevant for
+platforms where @code{-g} produces DWARF by default, otherwise one may
+try to enforce DWARF by using @code{-gdwarf-4}.
+
+@item compilation options set by @code{set compile-args}
+@end table
+
+@noindent
+You can override compilation options using the following command:
+
+@table @code
+@item set compile-args
+@cindex compile command options override
+Set compilation options used for compiling and injecting code with the
+@code{compile} commands.  These options override any conflicting ones
+from the target architecture and/or options stored during inferior
+compilation.
+
+@item show compile-args
+Displays the current state of compilation options override.
+This does not show all the options actually used during compilation,
+use @ref{set debug compile} for that.
+@end table
+
 @subsection Caveats when using the @code{compile} command
 
 There are a few caveats to keep in mind when using the @code{compile}
@@ -17380,6 +17440,24 @@  Access to those variables will generate a compiler error which @value{GDBN}
 will print to the console.
 @end table
 
+@subsection Compiler search for the @code{compile} command
+
+@value{GDBN} needs to find @value{NGCC} for the inferior being debugged which
+may not be obvious for remote targets of different architecture than where
+@value{GDBN} is running.  Environment variable @code{PATH} (@code{PATH} from
+shell that executed @value{GDBN}, not the one set by @value{GDBN}
+command @code{set environment}).  @xref{Environment}.  @code{PATH} on
+@value{GDBN} host is searched for @value{NGCC} binary matching the
+target architecture and operating system.
+
+Specifically @code{PATH} is searched for binaries matching regular expression
+@code{@var{arch}(-[^-]*)?-@var{os}-gcc} according to the inferior target being
+debugged.  @var{arch} is processor name --- multiarch is supported, so for
+example both @code{i386} and @code{x86_64} targets look for pattern
+@code{(x86_64|i.86)} and both @code{s390} and @code{s390x} targets look
+for pattern @code{s390x?}.  @var{os} is currently supported only for
+pattern @code{linux(-gnu)?}.
+
 @node GDB Files
 @chapter @value{GDBN} Files