make solaris target buildable.
> 1. The ipc check fails since solaris does not define union semun.
> The unix port code actually defines this type itself on solaris. Doing
> the same thing lets it get configured. Then...
> 2. It looks like BSDINSTALL is not defined properly.
BSDINSTALL is defined in make/host.cfg.in as
BSDINSTALL=@INSTALL@
@INSTALL@ is generated by autoconf's standard macro AC_PROG_INSTALL, which
is widely used in almost any autoconf/automake configured package. In case
there is really something wrong with it, then it must be considered a bug
in autoconf.
I can see a doubious fragment in AC_PROG_INSTALL, which is used when no
appropriate bsd-install is found.
Finally Ralf saw a problem with the find on solaris which I also saw and
fixed.
Here is a small patch which allows the m68060 to be used. I have not
tested the FP switching stuff which we know is broken. This is taken
against the libchip snapshot but should merge without problems. If you
have any problems please let me know.
There are other smaller issues such as superscalar enable and cache
control which I have not addressed yet. They are different to all other
m68k processors. These can wait IMO.
Here is a enhanced version of my previous patch. This patch enables
to potentially share the new interrupt management code for all Intel targets
(pc386, go32 and force386) bsp.
Note : this patch is complete only for pc386. It still needs to
be completed for go32 and force386. I carrefully checked
that anything needed is in for force386 (only some function
name changes for IDT manipulation and GDT segment
manipulation). But anyway I will not be able to test any
of theses targets...
returned a buffer which was not zero-filled, the reference count
was not correct. When the application exitted, the "lastClose"
handler was not being called to flush the output. This problem
had manifested itself on a variety of platforms.
The function rtems_termios_dequeue_characters() incorrectly incremented
the buffer pointers when it was invoked and there were no characters
in the ring buffer. This problem had also manifested itself on a
variety of platforms. The symptom was a strange repeating of the
data in the transmitter buffer when the transmitter serial device
was supposed to go idle.
* Added support for bsd "install" ($(BSDINSTALL)) to host.cfg.in, i.e.
the standard "install" program that most packages (including automake)
use. In Makefiles outside of rtems, "install" normally is referenced by
$(INSTALL), but rtems already uses $(INSTALL) for install-if-change,
hence I used $(BSDINSTALL) instead to keep up backward compatibility.
* Removed references to @GREP@ etc. from host.cfg.in, as configure.in
doesn't check for them (Minor cleanup).
* Added installation flags INST*FLAGS to host.cfg.in, which should
replace -m XXXX flags for installation calls.
*Changes to gcc.cfg to enable it to build host programs from multiple
sources files.
Should not disturb existing sources, but neccessary.
* There was a not-so-minor bug in the configuration files: "make
install" and "make debug_install" don't work in all subdirectories!! I
tried to fix this by adding "install" to MTARGETS in main.cfg, which
seems to solve most of the problems. But there still seem to be rare (?)
cases where "make debug_install" still seems to have problems.
* Changes to many host related tool-Makefiles to demonstrate the
abilities of INST*FLAGS, BSDINSTALL and the new rules in gcc.cfg.
..of cause ... but BSDINSTALL is THE standard method to install files
in most program packages besides rtems. This part of the patch fixes
some minor protection setting problems, but doesn't support
TARGET_VARIANTS
NOTE:
I hope you will like the BSDINSTALL, INST*FLAGS stuff. It is a step to
get rid of "install-if-change" and to rely on a more standard
installation procedure. If you don't like BSDINSTALL, removing it from
the patch isn't difficult- just grep for BSDINSTALL and replace
BSDINSTALL with INSTALL or MKDIR.
FINALLY:
I still have another patch pending (well, not a complete patch yet, it's
a partial patch to demonstrate the principle), which adds automatic
rebuilding of files generated by autoconf/configure. At the moment I
don't dare to submit it, because integrating this patch would require to
modify all Makefile.ins because we'd need to add a new "include " line
to each Makefile.in.
* c/src/exec/score/tools/sh - NEW DIRECTORY - contains shgen
Most of it should be self-explanatory. I am a little bit concerned about
host-dependent features (getopt, floating point libraries). This
shouldn't disturb much now, as this tool should be compileable on all
gnu-based hosts and is only applicable for the sh. But in case somebody
complains, we may need to add autoconf checks or even restructurize
parts of rtems (IMO, rtems needs to be restructurized - remember the
"turning rtems upside down" issue).
Haven't had a chance to do an extensive shake-out of 980710, but it
builds just fine on FreeBSD 2.2.5 (after termios is fixed using the
attached patch), and the tests run fine. FYI: FreeBSD doesn't support
System V IPC out of the box, but one only needs to add three options
to the kernel build configuration file, recompile the kernel, and
you're ready.
made no attempt to divide the comments up and place them with just
the appropriate files. Here is an excerpt from Ralf's email:
Changes including comments on changes I made after cycling through
all the targets:
* Added ranlib support. Now all targets use "ranlib" instead of "ar -s"
to build an index for a library. If ranlib isn't detected during
configuration, check if ar -s is working and try "ar -s" instead of
* Removed $(XXX_FOR_TARGET) from make/target.cfg.in, use $(XXX) instead now.
* gcc-target-default.cfg: LINK_XXXX-defines reworked to solve the -l
problem under posix (cf gcc-target-default.cfg)
* rtems-glom replaced by Makefile-rules inside of the wrapup/Makefile.in
that has been using rtems-glom until now.
* Removed CCC and friends in gcc-target-default.cfg, as they have been
breaking CXX support.
* Removed CONFIG.$(TARGET_ARCH).CC lines from several custom/*.cfg
files, because this is now set in custom/default.cfg.
* Added aclocal/ar-s.m4, check whether "ar -s" is working
* Added aclocal/cygwin.m4 and aclocal/exeext.m4.
* Reworked aclocal/canonicalize-tools.m4: Added ar -s check; fixes for
problems when XXX_FOR_TARGET is given via environment variables (didn't
work for gcc until now), adding cygwin check, improved autoconf-cache
handling.
* Removed -l from make rule dependencies. LINK_LIBS is now allowed to
contain -L and -l. LINK_OBJS and LINK_FILES must not contain -L or -l.
gcc28 make-exe rules now link using $(LINK_OBJS) $(LINK_LIBS) => Almost
all custom/*.cfg are modified. This is very likely to break something
because of typos or having missed to edit a file.
Open problems, known bugs, things I didn't do:
* custom/p4000.cfg seems to be out of date and requires to be reviewed.
(JRS NOTE: It is subordinate p4650 and p4600 -- both of which build ok
after minor changes.)
* custom/psim.cfg needs to be reviewed, I added some changes to it, I am
insecure about.
(JRS NOTE: psim had a minor problem endif/endef swapped but runs fine.)
* rtems-glom.in can now be removed.
* gcc*.cfg files "make depend" rules don't honor language specific flags
(e.g CXXFLAGS is ignored for *.cc) - Nothing to worry about now, but may
cause problems for hosts/targets not using gcc or rtems-add-ons that use
external packages.
* AFAIS, the no_bsp BSP can't be build anymore, i.e. configure refused
to configure for it whatever I tried.
* The toplevel and toplevel+1 README files are quite out-dated
* cygwin.m4 isn't of much use for rtems. In most cases (cf.
aclocal/*.m4) it is worked around by directly using $host_os. I think
I'll remove it soon after the next snapshot
* Before release the cygwin patch needs to be tested under cygwin. I may
have broken/missed something (esp. the sed-pattern to convert \\ into /
may be broken).
* You should try to build/run the posix-BSP under solaris - I don't
expect problems, but I am not 100% sure, esp. with regard to ranlib/ar -s.
* You should consider to convert all make/compilers/*.cfg files into
make/compilers/*.cfg.in files and let autoconf generate the *.cfg. This
may help getting rid of some if/then/else statements and help
hard-coding some defines into those files in future and shouldn't
disturb now.
* Not having installed libc.a/libm.a on a host may still break building
rtems, esp. when using -disable-gcc28 as the gcc27-configuration scheme
directly accesses libc.a and libm.a. The problem should not appear when
using gcc28 because it references libc/libm only through -lc and -lm
which may be static or dynamic (I didn't test this).
* shgen is not yet included (I didn't yet have enough time to integrate it).
* I know about a few more configure-probs (esp. cross-checking
--enable-* flags).
+ warn/refuse to configure when --enable-libcdir and
--enable-gcc28 are given.
+ force --enable-libcdir when --disable-gcc28 is given
* Replaced KSHELL with @KSH@ in some shell scripts generated by configure.in.
* Added a dependency to aclocal/*.m4 in the toplevel Makefile => configure
and aclocal.m4 will now be rebuild when any aclocal/*.m4 file is changed
* Some changes to aclocal/gcc-pipe.m4 and aclocal/gcc-specs.m4
* Replaced i[[3456]]86-unknown-freebsd2.[[12]] with i[[3456]]86-*freebsd2.*
in configure.in, as I suppose there might exist a variety of valid vendors
(2nd field of the name-tripple)
* Disabled override MAKEFLAGS in toplevel Makefile.in - Potential
side-effects are not really clear to me.
* In mvme162.cfg, $(LINK_LIBS) is missing in the CC line in gcc28's make-exe
rule (yet another one I missed to edit). Just append $(LINK_LIBS) to
the "CC" line, like I hopefully did to ALL other custom/*.cfg files.
* the problem with mvme162lx.cfg is a follow-up problem of the
mvme162.cfg-bug.
* mvme162/console and idp/console had variables named Buffer which
conflicted with similarly named variables in some tests.
between CPU32 and CPU32+ cores. Commentary follows:
Unfortunately c/src/exec/score/cpu/m68k/m68k.h incorrectly defines
M68K_HAS_MISALIGNED for the plain old CPU32 (it is correct for the CPU32+).
As a consequence, the recently-relocated m68k memcpy() may still attempt
misaligned memory accesses.
I suggest that until such time as egcs/gcc differentiates these cores
that we invent a new preprocessor symbol, RTEMS__mcpu32p__ for this
purpose, on the assumption that egcs may one day grow a -mcpu32+ option
which will define a __mcpu32p__ symbol (whether this option would also
define __mcpu32__ is yet to be resolved).
BSPs that have a CPU32+ (like gen68360) would for the time being define
RTEMS__mcpu32p__ using -D. The symbol is `RTEMS__mcpu32p__' because
symbols of the form __xxx__ should only be defined by the compiler
itself.
Note that the patch tests for RTEMS__mcpu32p__ *before* __mcpu32__, since
__mcpu32__ is still defined for the CPU32+. It does not change the
gen68360 BSP.
An aside:
Note that in egcs-1.0.3a, the option -m68332 is identical to -mcpu32,
except it defines __mc68332__ as well as __mcpu32__. This is only
for the sake of compatibility. The story with -m68302 is similar;
it defines __mc68302__ and __mc68000__. In my opinion these options
are depreciated and ought to be avoided in RTEMS.
task to be deleted was created via the same API (i.e. were of the object
class created by this API). For example, a POSIX thread calling
the rtems_task_delete(SELF) directive would incorrectly update the RTEMS
object local pointer table.
Jennifer discovered this when moving tests implemented in C using the
Classic RTEMS API into a tree of Ada tests. The Ada tests were implicitly
using POSIX services. This lead to some unexpected behavior.
of seconds since 1988 from RTEMS and not adding in the 1970-1988 correction
factor. Plus removed checks for data/time set since POSIX does not permit
this call to fail. GNAT 3.12 depends on this.
important distinctions between CPU models which are not made by gcc.
These distinctions help give us a more optimized memcpy(). This is important
for message queues and KA9Q.
external interrupt priorities were not being honored. Here is some
of his original report:
using rtems/erc32, I have a problem with interrupt priority when
interrupts occure simultaneously. Erc32 has an interrupt force
register where interrupts can be generated. If more than one
interrupt is generated, the interrupt handlers are scheduled in
the wrong order, i.e. with the lowest priority first.
I have attched a program that generates three interrupts, 0x11, 0x12
and 0x13. Interrupt 0x13 should be handled first, but is actually
handled last. Below is the output from sis:
sis> go
resuming at 0x02000000
RAM size: 4096 K, ROM size: 2048 K
Watchdog disabled
Waitstates = RAM read: 0, RAM write: 0, ROM read: 0, ROM write: 0
Power-down mode enabled
infinite UART baudrate
External interrupt received with vector 0x11
External interrupt received with vector 0x12
External interrupt received with vector 0x13
I have verified that sis generates the interrupts in the correct
order, i.e. 0x13 first, then 0x12 and then 0x11. So the problem
seems to be in the rtems interrupt handler. Do you use the PIL field
in the %psr register to mask lower priority interrupts or are all
external interrupts considered to have the same priority ..?
Here is a description of the fix:
it turned out that lower priority interrupts were not at all masked
off during interrupt handling. I made the following fix to cpu_asm.s:
... fix is in the code ...
There might be a simpler way of doing this, but this works...
Eric Norum per request from Geoffroy Montel:
> The rtems_termios_enqueue_raw_characters function type is void.
> The problem is that I can't return an error message if the input
> buffer is full.
> Could we add a return value?
Sure, but what would you do with the overflow indication? POSIX says,
``when the input limit is reached, the saved characters are thrown away
without notice''.
Anyhow, the change is so small I've done it and enclosed the patch.
