The igen/dgen and opc2c tools leak their heap-allocated memory (on
purpose) at program exit, which makes AddressSanitizer fail the tool
execution. This breaks the build, as it makes the tool return a
non-zero exit code.
Fix that by disabling leak detection through the setting of that
environment variable.
I also changed the opc2c rules for m32c to go through a temporary file.
What happened is that the failing opc2c would produce an incomplete file
(probably because ASan exits the process before stdout is flushed).
This meant that further make attempts didn't try to re-create the file,
as it already existed. A "clean" was therefore necessary. This can
also happen in regular builds if the user interrupts the build (^C) in
the middle of the opc2c execution and tries to resume it. Going to a
temporary file avoids this issue.
sim/m32c/ChangeLog:
* Makefile.in: Set ASAN_OPTIONS when running opc2c.
sim/mips/ChangeLog:
* Makefile.in: Set ASAN_OPTIONS when running igen.
sim/mn10300/ChangeLog:
* Makefile.in: Set ASAN_OPTIONS when running igen.
sim/ppc/ChangeLog:
* Makefile.in: Set ASAN_OPTIONS when running igen.
sim/v850/ChangeLog:
* Makefile.in: Set ASAN_OPTIONS when running igen.
Change-Id: I00f21d4dc1aff0ef73471925d41ce7c23e83e082
This simplifies the build a bit (especially for deps in port subdirs),
and avoids recursive make. This in turn speeds up the build, and sets
us up for multi-target.
All the scripts were using this implicitly already, so there's no real
change for them, but we want to call it explicitly as the CPP tool is
used to generate nltvals.def.
This file is quite large and is getting unmanageable. Split it apart
to follow aclocal best practices by putting one-macro-per-file. There
shouldn't be any real functional changes here as can be seen in the
configure script regens.
Rather than hand maintain m4 includes in various autotool files,
use AC_CONFIG_MACRO_DIRS to declare the relevant search paths.
This simplifies the code, makes it more robust, and cleans out
unused logic from configure.
Few arches implement STATE_WATCHPOINTS()->pc while all of them implement
sim_pc_get. Lets switch the sim-watch core for monitoring pc events to
the sim_pc_get API so this module works for all ports, and then we can
delete this old back channel of snooping in the port's cpu state -- the
code needs the pointer to the pc storage so that it can read out bytes
and compare them to the watchrange.
This also fixes the logic on multi-cpu sims by removing the limitation
of only being able to watch CPU0's state.
These settings might have made sense in darker compiler times, but I
think they're largely obsolete now. Looking through the values that
get used in HDEFINES, it's quite limited, and configure itself should
handle them. If we still need something, we can leverage standard
autoconf macros instead, after we get a clear user report.
TDEFINES was never set anywhere and was always empty, so prune that.
This port only had one minor warning left in it, so fix it and then
enable -Werror behavior by deleting the macro call. We'll use the
common default now (which is -Werror).
My recent rewrite of the nltvals generator fixed a bug where SYS_times
was not being exported for v850. But that in turn uncovered this bug
where the SYS_times codepath had a compile error.
Existing ports already have sizeof_pc set to the same size as sim_cia,
so simply make that part of the core code. We already assume this in
places by way of sim_pc_{get,set}, and this is how it's documented in
the sim-base.h API.
There is code to allow sims to pick different register word sizes from
address sizes, but most ports use the defaults for both (32-bits), and
the few that support multiple register sizes never change the address
size (so address defaults to register). I can't think of any machine
where the register hardware size would be larger than the address word
size either. We have ABIs that behave that way (e.g. x32), but the
hardware is still equivalent register sized.
Since we require C11 now, we can assume many headers exist, and
clean up all of the conditional includes. It's not like any of
this code actually accounted for the headers not existing, just
whether we could include them.
The strings.h cleanup is a little nuanced: it isn't in C11, but
every use of it in the codebase will include strings.h only if
string.h doesn't exist. Since we now assume the C11 string.h
exists, we'll never include strings.h, so we can delete it.
We've had this off for a long time because the sim code was way too
full of warnings for it to be feasible. However, I've cleaned things
up significantly from when this was first merged, and we can start to
turn this around.
Change the macro to enable -Werror by default, and allow ports to opt
out. New ports will get it automatically (and we can push back on
them if they try to turn it off).
Also turn it off for the few ports that still hit warnings for me.
All the rest will get the new default, and we'll wait for feedback
if/when new issues come up.
With GDB requiring a C++11 compiler now, this hopefully shouldn't
be a big deal. It's been 10 years since C11 came out, so should
be plenty of time to upgrade.
This will allow us to start cleaning up random header logic and
many of our non-standard custom types.
Rather than stuffing the command line with a bunch of -D flags, start
moving things to config.h which is managed by autoheader. This makes
the makefile a bit simpler and the build output tighter, and it makes
the migration to automake easier as there are fewer vars to juggle.
We'll want to move the other options out too, but it'll take more work.
This was imported from the ppc sim, but that was only used to control
a single file, and that is already governed by the hw models. There's
no need to have a sep configure option here, especially since none of
the other sims are using it. Even when the code is enabled, there's
no runtime overhead.
Currently ports have to call SIM_AC_OPTION_ENVIRONMENT explicitly in
order to make the configure flag available. There's no real reason
to not allow this flag for all ports, so move it to the common sim
macro. This way we get standard behavior across all ports too.
Currently ports have to call SIM_AC_OPTION_ASSERT explicitly in order
to make the configure flag available, which none of them do. There's
no real reason to not allow this flag for all ports, so move it to the
common sim macro. This way we get standard behavior across all ports.
We don't have alternative nltvals.def files, so always symlinking
the targ-vals.def file to it doesn't gain us anything. It does
make the build more complicated though and a pain to convert to
something newer (like automake). Drop the symlinking entirely.
In the future, we'll want to explode this file anyways into the
respective arch dirs so things can be selected dynamically at
runtime, so it's not like we'll be bringing this back.
Currently ports have to call SIM_AC_OPTION_INLINE explicitly in order
to make the configure flag available. There's no real reason to not
allow this flag for all ports, so move it to the common sim macro.
This way we get standard behavior across all ports too.
These options were never exposed for most sims (just the ppc one),
and they are really only useful on 32-bit x86 systems. Considering
modern systems tend to be 64-bit x86_64 and how well modern compilers
are at optimizing code, these have outlived their usefulness.
No other sub directory provides such a configuration option, so
drop it from the sim dir as well. This cleans up a good bit of
code in the process.
If people want to use custom flags for just the sim, they can
still run configure+make by hand in the sim subdir and use the
normal CFLAGS settings.
The common subdir sets up a cconfig.h file to hold checks for the common
code. In practice, most files still end up using config.h instead which
just leads to confusion.
Merge all the configure checks that went into cconfig.h into SIM_AC_COMMON
so we can drop the cconfig.h file altogether. Now there is only a single
config.h file like normal.
The compiler/C library should produce reasonable code for htonl/ntohl,
and at least glibc tries pretty hard to always produce good code for
them. This logic only had support for 32-bit x86 systems anymore, and
it's unlikely people were even opting into this, so drop it all.
Fix a long standing todo where we let getopt write directly to stderr
when an invalid option is passed. Use the sim io funcs instead as they
go through the filtered callbacks that gdb wants.
The --enable-sim-hostendian flag was purely so people had an escape route
for when cross-compiling. This is because historically, AC_C_BIGENDIAN
did not work in those cases. That was fixed a while ago though, so we can
require that macro everywhere now and simplify a good bit of code.
Rather than re-invent endian defines, as well as maintain our own list
of OS & arch-specific includes, punt all that logic in favor of the bfd
ones already set up and maintained elsewhere. We already rely on the
bfd library, so leveraging the endian aspect should be fine.
Pretty much all targets are using this module already, so add it to the
common list of objects. The only oddball out here is cris and that's
because it supports loading via an offset for all the phdrs. We drop
support for that.
No arch is using this anymore, and we want all new ports using the
hardware framework instead. Punt WITH_DEVICES and the two callbacks
device_io_{read,write}_buffer.
We can also punt the tconfig.h file as no port is using it anymore.
This fixes in-tree builds that get confused by picking up the wrong
one (common/ vs <port>/) caused by commit ae7d0cac8c.
Any port that needs to set up a global define can use their own
sim-main.h file that they must provide regardless.
We build & bundle the watchpoint module everywhere, but we don't make
the command line flags available by default. A few targets opted in,
but most did not. Just enable the flag for everyone. Not all targets
will respect the flags (making them nops), but shouldn't be a big deal.
This is how we handle other common modules already.
Fix occurrences of left-shifting negative constants in C code.
sim/arm/ChangeLog:
* thumbemu.c (handle_T2_insn): Fix left shift of negative value.
* armemu.c (handle_v6_insn): Likewise.
sim/avr/ChangeLog:
* interp.c (sign_ext): Fix left shift of negative value.
sim/mips/ChangeLog:
* micromips.igen (process_isa_mode): Fix left shift of negative
value.
sim/msp430/ChangeLog:
* msp430-sim.c (get_op, put_op): Fix left shift of negative value.
sim/v850/ChangeLog:
* simops.c (v850_bins): Fix left shift of negative value.
Having this be a config option doesn't make sense: the code size is
pretty much the same (as all the logic is still active), and if it's
disabled, the sim throws an error if you try to use it. That means
we can't break sims that weren't using it before by enabling it all
the time.
Now that all arches (for the most part) have moved over, move sim-stop.o,
sim-reason.o, and sim-reg.o to the common object list and out of all the
arch ports.
Other than the nice advantage of all sims having to declare one fewer
common function, this also fixes leakage in pretty much every sim.
Many were not freeing any resources, and a few were inconsistent as
to the ones they did. Now we have a single module that takes care of
all the logic for us.
Most of the non-cgen based ones could be deleted outright. The cgen
ones required adding a callback to the arch-specific cleanup func.
The few that still have close callbacks are to manage their internal
state.
We do not convert erc32, m32c, ppc, rl78, or rx as they do not use
the common sim core.
Since every target typedefs this the same way, move it to the common code.
We have to leave Blackfin behind here for now because of inter-dependencies
on types and headers: sim-base.h includes sim-model.h which needs types in
machs.h which needs types in bfim-sim.h which needs SIM_CPU.
Almost every target defines sim_cia the same way -- either using the
address_word type directly, or a type of equivalent size. The only
odd one out is sh64 (who has 32bit address_word and 64bit cia), and
even that case doesn't seem to make sense. We'll put off clean up
though of sh64 and at least set up a sensible default for everyone.
The CIA_{GET,SET} macros serve the same function as CPU_PC_{GET,SET}
except the latter adds a layer of indirection via the sim state. This
lets models set up different functions at runtime and doesn't reach so
directly into the arch-specific cpu state.
It also doesn't make sense to have two sets of macros that do exactly
the same thing, so lets standardize on the one that gets us more.
Now that all the targets are utilizing CPU_PC_{FETCH,STORE}, and the
cpu state is multicore, and the STATE_CPU defines match, we can move
it all to the common code.
