OpenBSD-4.6/share/man/man1/gcc-local.1

Compare this file to the similar file:
Show the results in this format:

.\" $OpenBSD: gcc-local.1,v 1.19 2008/11/04 23:59:37 martynas Exp $
.\"
.\" Copyright (c) 2002 Marc Espie
.\" Copyright (c) 2003 Anil Madhavapeddy
.\"
.\" All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\"    notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\"    notice, this list of conditions and the following disclaimer in the
.\"    documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE DEVELOPERS ``AS IS'' AND ANY EXPRESS OR
.\" IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
.\" OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
.\" IN NO EVENT SHALL THE DEVELOPERS BE LIABLE FOR ANY DIRECT, INDIRECT,
.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
.\" NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
.\" DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
.\" THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
.\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
.\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
.\"
.Dd $Mdocdate: November 4 2008 $
.Dt GCC-LOCAL 1
.Os
.Sh NAME
.Nm gcc-local
.Nd local modifications to gcc
.Sh DESCRIPTION
Some
.Ox
platforms use a derivative of
.Nm gcc 2.95.3 ,
others use a derivative of
.Nm gcc 3.3.5 .
In both cases,
the
.Nm gcc
software comes with specific modifications for
.Ox .
.Bl -dash
.It
.Nm gcc
does not search under
.Pa /usr/local
for include files nor for libraries:
as a system compiler, it only searches the system paths by default.
.It
On
.Ox ,
the
.Fl pthread
option should be used to link threaded code, isolating the program from
operating system details.
.It
On most architectures,
trampoline code marks the smallest possible area around the trampoline stub
executable using
.Xr mprotect 2 ,
since the stack area is by default non-executable.
.It
On
.Ox ,
by default, trampoline code generation is disabled.
Code requiring trampolines will not compile without
.Fl ftrampolines .
The warning flag
.Fl Wtrampolines
can be used to locate trampoline instances if trampoline generation
is re-enabled.
.It
The
.Fl O2
option does not include
.Fl fstrict-aliasing ,
as this option causes issues on some legacy code.
.Fl fstrict-aliasing
is very unsafe with code that plays tricks with casts, bypassing the
already weak type system of C.
.It
The option
.Fl fno-builtin-<function>
was backported from
.Nm gcc 3.3.5 ,
to
.Nm gcc 2.95.3 ,
and can be used without having to differentiate between
both compilers.
.It
.Nm gcc
recognizes the extra format attribute syslog, to better match
the definition of
.Xr syslog 3 ,
and silence erroneous warnings when used with
.Fl pedantic .
.It
Even in 2.95.3,
.Nm gcc
recognizes the attribute
.Dv __nonnull__ ,
which can be used to mark arguments that can't be
.Dv NULL .
The printf format attribute does not imply
.Dv __nonnull__
for the format.
This allows for correct format checking on the
.Xr err 3
function family.
.It
.Nm gcc
recognizes the extra attribute
.Dv __sentinel__ ,
which can be used to mark varargs function that need a
.Dv NULL
pointer to mark argument termination, like
.Xr execl 3 .
This exposes latent bugs for 64-bit architectures,
where a terminating 0 will expand to a 32-bit int, and not a full-fledged
64-bits pointer.
.It
On some
.Ox
platforms,
.Nm gcc
still uses
.Xr setjmp 3 /
.Xr longjmp 3 -
style exceptions, and so needs extra fixes beyond the pure 2.95.3 release.
.It
On a few
platforms (mostly a.out),
.Nm gcc
uses a linker wrapper to write stubs that call global constructors and
destructors.
Those platforms use
.Nm gcc 2.95.3 ,
and those calls can be traced using
.Fl Wl,-trace-ctors-dtors ,
using
.Xr syslog_r 3 .
.It
On alpha,
.Fl mieee
is enabled by default to enable full compliance with
the IEEE floating point standard,
although the
.Dq inexact
flag is not maintained.
.It
On
.Ox ,
.Nm gcc
comes with the
.Dq ProPolice
stack protection extension, which is enabled by default.
This extension reorders local variable declarations and adds stack consistency
checks at run time, in order to detect stack overflows, and will attempt to
report the problem in the system logs by calling
.Xr syslog 3
with a
.Dv LOG_CRIT
priority message:
.Dq stack overflow in function XXX ,
and abort the faulting process.
It can be turned off using the
.Fl fno-stack-protector
commandline option.
Note that the stack protector relies on some support code in libc.
Stand-alone programs not linked against libc must either provide their own
support bits, or use the
.Fl fno-stack-protector
option.
There is also a
.Fl fstack-protector-all
option, that turns stack protection code on for all functions,
and disables any heuristic that flags some functions as safe.
This extended checking has a moderate runtime cost, though.
.It
.Nm gcc
recognizes a new flag,
.Fl Wbounded ,
to perform basic checks on functions which accept buffers and sizes.
An extra attribute,
.Dv __bounded__ ,
has been added to mark functions that can be
checked this way.
.It
.Nm gcc
recognizes a new format attribute, kprintf, to deal with the extra format
arguments
.Ql %b ,
.Ql %r ,
and
.Ql %z
used in the
.Ox
kernel.
.It
.Nm gcc
does not store its version string in objects.
This behavior can be restored with
.Fl fident .
.It
On
.Ox ,
.Nm gcc
will not move variables initialized with the value zero
from the data section to the bss section.
The default behaviour of
.Nm gcc 3.3.5
on other systems is to perform this action, which can be restored for
.Ox
with
.Fl fzero-initialized-in-bss .
.It
On
.Ox ,
.Nm gcc
does not warn for cast expressions used as lvalues outside of
.Fl pedantic .
.It
Even in 2.95.3,
.Nm gcc
recognizes the preprocessor flag
.Fl CC
that lets comments in macros pass through to the output (except in
.Fl traditional
mode).
This is used to allow annotations in macros for
.Xr lint 1 .
.It
On
.Ox ,
.Nm gcc
supports two extra warning options:
.Bl -item
.It
.Fl Wstack-larger-than- Ns Va N
will report functions using more than
.Va N
bytes of stack space for their local variables.
Stack space used for other purposes (such as register window saving,
callee-saved registers, or outbound arguments storage)
is not taken into account for this check.
.It
.Fl Wvariable-decl
will report automatic variable declarations whose size cannot be
determined at compile-time.
.El
.El
.Sh ATTRIBUTES
The
.Dv __bounded__
attribute is used to type-check functions whose parameters pass fixed-length
buffers and their sizes.
The syntax for normal buffers is:
.Pp
.Li __attribute__ ((__bounded__ (
.Dv __buffer__ ,
.Va buffer ,
.Va length
.Li )))
.Pp
where
.Fa buffer
contains the parameter number (starting from 1) of the pointer to the buffer,
and
.Fa length
contains the parameter number of the buffer length argument.
.Pp
.Nm gcc
will emit a warning if the length argument is a constant larger than the
actual size of the buffer.
If the buffer is not a statically declared array of fixed length, no warnings
will be generated.
Refer to
.Xr memcpy 3
for an example of a function with this check.
.Pp
For checking strings, just use
.Dv __string__
instead of
.Dv __buffer__ :
.Pp
.Li __attribute__ ((__bounded__ (
.Dv __string__ ,
.Va buffer ,
.Va length
.Li )))
.Pp
In addition to the checks described above, this also tests if the
.Va length
argument was wrongly derived from a
.Fn sizeof "void *"
operation.
.Xr strlcpy 3
is a good example of a string function with this check.
.Pp
Some functions specify the length as two arguments:
the number of elements and the size of each element.
In this case, use the
.Dv __size__
attribute:
.Pp
.Li __attribute__ ((__bounded__ (
.Dv __size__ ,
.Va buffer ,
.Va nmemb ,
.Va size
.Li )))
.Pp
where
.Va buffer
contains the parameter number of the pointer to the buffer,
.Va nmemb
contains the parameter number of the number of members, and
.Va size
has the parameter number of the size of each element.
The type checks performed by
.Dv __size__
are the same as the
.Dv __buffer__
attribute.
See
.Xr fread 3
for an example of this type of function.
.Pp
If a function accepts a buffer parameter and specifies that it has to be of a
minimum length, the __minbytes__ attribute can be used:
.Pp
.Li __attribute__ ((__bounded__ (
.Dv __minbytes__ ,
.Va buffer ,
.Va minsize
.Li )))
.Pp
where
.Va buffer
contains the parameter number of the pointer to the buffer, and
.Va minsize
specifies the minimum number of bytes that the buffer should be.
.Xr ctime_r 3
is an example of this type of function.
.Pp
If
.Fl Wbounded
is specified with
.Fl Wformat ,
additional checks are performed on
.Xr sscanf 3
format strings.
The
.Ql %s
fields are checked for incorrect bound lengths by checking the size of the
buffer associated with the format argument.
.Sh SEE ALSO
.Xr gcc 1
.Pp
.Pa http://www.research.ibm.com/trl/projects/security/ssp/
.Sh CAVEATS
The
.Fl Wbounded
flag only works with statically allocated fixed-size buffers.
Since it is applied at compile-time, dynamically allocated memory buffers
and non-constant arguments are ignored.